initial commit, 4.5 stable

thirdparty/icu4c/common/unicode/appendable.h

@@ -0,0 +1,239 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+/*
+*******************************************************************************
+*   Copyright (C) 2011-2012, International Business Machines
+*   Corporation and others.  All Rights Reserved.
+*******************************************************************************
+*   file name:  appendable.h
+*   encoding:   UTF-8
+*   tab size:   8 (not used)
+*   indentation:4
+*
+*   created on: 2010dec07
+*   created by: Markus W. Scherer
+*/
+
+#ifndef __APPENDABLE_H__
+#define __APPENDABLE_H__
+
+/**
+ * \file
+ * \brief C++ API: Appendable class: Sink for Unicode code points and 16-bit code units (char16_ts).
+ */
+
+#include "unicode/utypes.h"
+
+#if U_SHOW_CPLUSPLUS_API
+
+#include "unicode/uobject.h"
+
+U_NAMESPACE_BEGIN
+
+class UnicodeString;
+
+/**
+ * Base class for objects to which Unicode characters and strings can be appended.
+ * Combines elements of Java Appendable and ICU4C ByteSink.
+ *
+ * This class can be used in APIs where it does not matter whether the actual destination is
+ * a UnicodeString, a char16_t[] array, a UnicodeSet, or any other object
+ * that receives and processes characters and/or strings.
+ *
+ * Implementation classes must implement at least appendCodeUnit(char16_t).
+ * The base class provides default implementations for the other methods.
+ *
+ * The methods do not take UErrorCode parameters.
+ * If an error occurs (e.g., out-of-memory),
+ * in addition to returning false from failing operations,
+ * the implementation must prevent unexpected behavior (e.g., crashes)
+ * from further calls and should make the error condition available separately
+ * (e.g., store a UErrorCode, make/keep a UnicodeString bogus).
+ * @stable ICU 4.8
+ */
+class U_COMMON_API Appendable : public UObject {
+public:
+    /**
+     * Destructor.
+     * @stable ICU 4.8
+     */
+    ~Appendable();
+
+    /**
+     * Appends a 16-bit code unit.
+     * @param c code unit
+     * @return true if the operation succeeded
+     * @stable ICU 4.8
+     */
+    virtual UBool appendCodeUnit(char16_t c) = 0;
+
+    /**
+     * Appends a code point.
+     * The default implementation calls appendCodeUnit(char16_t) once or twice.
+     * @param c code point 0..0x10ffff
+     * @return true if the operation succeeded
+     * @stable ICU 4.8
+     */
+    virtual UBool appendCodePoint(UChar32 c);
+
+    /**
+     * Appends a string.
+     * The default implementation calls appendCodeUnit(char16_t) for each code unit.
+     * @param s string, must not be nullptr if length!=0
+     * @param length string length, or -1 if NUL-terminated
+     * @return true if the operation succeeded
+     * @stable ICU 4.8
+     */
+    virtual UBool appendString(const char16_t *s, int32_t length);
+
+    /**
+     * Tells the object that the caller is going to append roughly
+     * appendCapacity char16_ts. A subclass might use this to pre-allocate
+     * a larger buffer if necessary.
+     * The default implementation does nothing. (It always returns true.)
+     * @param appendCapacity estimated number of char16_ts that will be appended
+     * @return true if the operation succeeded
+     * @stable ICU 4.8
+     */
+    virtual UBool reserveAppendCapacity(int32_t appendCapacity);
+
+    /**
+     * Returns a writable buffer for appending and writes the buffer's capacity to
+     * *resultCapacity. Guarantees *resultCapacity>=minCapacity.
+     * May return a pointer to the caller-owned scratch buffer which must have
+     * scratchCapacity>=minCapacity.
+     * The returned buffer is only valid until the next operation
+     * on this Appendable.
+     *
+     * After writing at most *resultCapacity char16_ts, call appendString() with the
+     * pointer returned from this function and the number of char16_ts written.
+     * Many appendString() implementations will avoid copying char16_ts if this function
+     * returned an internal buffer.
+     *
+     * Partial usage example:
+     * \code
+     *  int32_t capacity;
+     *  char16_t* buffer = app.getAppendBuffer(..., &capacity);
+     *  ... Write n char16_ts into buffer, with n <= capacity.
+     *  app.appendString(buffer, n);
+     * \endcode
+     * In many implementations, that call to append will avoid copying char16_ts.
+     *
+     * If the Appendable allocates or reallocates an internal buffer, it should use
+     * the desiredCapacityHint if appropriate.
+     * If a caller cannot provide a reasonable guess at the desired capacity,
+     * it should pass desiredCapacityHint=0.
+     *
+     * If a non-scratch buffer is returned, the caller may only pass
+     * a prefix to it to appendString().
+     * That is, it is not correct to pass an interior pointer to appendString().
+     *
+     * The default implementation always returns the scratch buffer.
+     *
+     * @param minCapacity required minimum capacity of the returned buffer;
+     *                    must be non-negative
+     * @param desiredCapacityHint desired capacity of the returned buffer;
+     *                            must be non-negative
+     * @param scratch default caller-owned buffer
+     * @param scratchCapacity capacity of the scratch buffer
+     * @param resultCapacity pointer to an integer which will be set to the
+     *                       capacity of the returned buffer
+     * @return a buffer with *resultCapacity>=minCapacity
+     * @stable ICU 4.8
+     */
+    virtual char16_t *getAppendBuffer(int32_t minCapacity,
+                                   int32_t desiredCapacityHint,
+                                   char16_t *scratch, int32_t scratchCapacity,
+                                   int32_t *resultCapacity);
+};
+
+/**
+ * An Appendable implementation which writes to a UnicodeString.
+ *
+ * This class is not intended for public subclassing.
+ * @stable ICU 4.8
+ */
+class U_COMMON_API UnicodeStringAppendable : public Appendable {
+public:
+    /**
+     * Aliases the UnicodeString (keeps its reference) for writing.
+     * @param s The UnicodeString to which this Appendable will write.
+     * @stable ICU 4.8
+     */
+    explicit UnicodeStringAppendable(UnicodeString &s) : str(s) {}
+
+    /**
+     * Destructor.
+     * @stable ICU 4.8
+     */
+    ~UnicodeStringAppendable();
+
+    /**
+     * Appends a 16-bit code unit to the string.
+     * @param c code unit
+     * @return true if the operation succeeded
+     * @stable ICU 4.8
+     */
+    virtual UBool appendCodeUnit(char16_t c) override;
+
+    /**
+     * Appends a code point to the string.
+     * @param c code point 0..0x10ffff
+     * @return true if the operation succeeded
+     * @stable ICU 4.8
+     */
+    virtual UBool appendCodePoint(UChar32 c) override;
+
+    /**
+     * Appends a string to the UnicodeString.
+     * @param s string, must not be nullptr if length!=0
+     * @param length string length, or -1 if NUL-terminated
+     * @return true if the operation succeeded
+     * @stable ICU 4.8
+     */
+    virtual UBool appendString(const char16_t *s, int32_t length) override;
+
+    /**
+     * Tells the UnicodeString that the caller is going to append roughly
+     * appendCapacity char16_ts.
+     * @param appendCapacity estimated number of char16_ts that will be appended
+     * @return true if the operation succeeded
+     * @stable ICU 4.8
+     */
+    virtual UBool reserveAppendCapacity(int32_t appendCapacity) override;
+
+    /**
+     * Returns a writable buffer for appending and writes the buffer's capacity to
+     * *resultCapacity. Guarantees *resultCapacity>=minCapacity.
+     * May return a pointer to the caller-owned scratch buffer which must have
+     * scratchCapacity>=minCapacity.
+     * The returned buffer is only valid until the next write operation
+     * on the UnicodeString.
+     *
+     * For details see Appendable::getAppendBuffer().
+     *
+     * @param minCapacity required minimum capacity of the returned buffer;
+     *                    must be non-negative
+     * @param desiredCapacityHint desired capacity of the returned buffer;
+     *                            must be non-negative
+     * @param scratch default caller-owned buffer
+     * @param scratchCapacity capacity of the scratch buffer
+     * @param resultCapacity pointer to an integer which will be set to the
+     *                       capacity of the returned buffer
+     * @return a buffer with *resultCapacity>=minCapacity
+     * @stable ICU 4.8
+     */
+    virtual char16_t *getAppendBuffer(int32_t minCapacity,
+                                   int32_t desiredCapacityHint,
+                                   char16_t *scratch, int32_t scratchCapacity,
+                                   int32_t *resultCapacity) override;
+
+private:
+    UnicodeString &str;
+};
+
+U_NAMESPACE_END
+
+#endif /* U_SHOW_CPLUSPLUS_API */
+
+#endif  // __APPENDABLE_H__

thirdparty/icu4c/common/unicode/brkiter.h

@@ -0,0 +1,672 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+/*
+********************************************************************************
+*   Copyright (C) 1997-2016, International Business Machines
+*   Corporation and others.  All Rights Reserved.
+********************************************************************************
+*
+* File brkiter.h
+*
+* Modification History:
+*
+*   Date        Name        Description
+*   02/18/97    aliu        Added typedef for TextCount.  Made DONE const.
+*   05/07/97    aliu        Fixed DLL declaration.
+*   07/09/97    jfitz       Renamed BreakIterator and interface synced with JDK
+*   08/11/98    helena      Sync-up JDK1.2.
+*   01/13/2000  helena      Added UErrorCode parameter to createXXXInstance methods.
+********************************************************************************
+*/
+
+#ifndef BRKITER_H
+#define BRKITER_H
+
+#include "unicode/utypes.h"
+
+/**
+ * \file
+ * \brief C++ API: Break Iterator.
+ */
+
+#include "unicode/utypes.h"
+
+#if U_SHOW_CPLUSPLUS_API
+
+#if UCONFIG_NO_BREAK_ITERATION
+
+U_NAMESPACE_BEGIN
+
+/*
+ * Allow the declaration of APIs with pointers to BreakIterator
+ * even when break iteration is removed from the build.
+ */
+class BreakIterator;
+
+U_NAMESPACE_END
+
+#else
+
+#include "unicode/uobject.h"
+#include "unicode/unistr.h"
+#include "unicode/chariter.h"
+#include "unicode/locid.h"
+#include "unicode/ubrk.h"
+#include "unicode/strenum.h"
+#include "unicode/utext.h"
+#include "unicode/umisc.h"
+
+U_NAMESPACE_BEGIN
+
+class CharString;
+
+/**
+ * The BreakIterator class implements methods for finding the location
+ * of boundaries in text. BreakIterator is an abstract base class.
+ * Instances of BreakIterator maintain a current position and scan over
+ * text returning the index of characters where boundaries occur.
+ * <p>
+ * Line boundary analysis determines where a text string can be broken
+ * when line-wrapping. The mechanism correctly handles punctuation and
+ * hyphenated words.
+ * <p>
+ * Sentence boundary analysis allows selection with correct
+ * interpretation of periods within numbers and abbreviations, and
+ * trailing punctuation marks such as quotation marks and parentheses.
+ * <p>
+ * Word boundary analysis is used by search and replace functions, as
+ * well as within text editing applications that allow the user to
+ * select words with a double click. Word selection provides correct
+ * interpretation of punctuation marks within and following
+ * words. Characters that are not part of a word, such as symbols or
+ * punctuation marks, have word-breaks on both sides.
+ * <p>
+ * Character boundary analysis allows users to interact with
+ * characters as they expect to, for example, when moving the cursor
+ * through a text string. Character boundary analysis provides correct
+ * navigation of through character strings, regardless of how the
+ * character is stored.  For example, an accented character might be
+ * stored as a base character and a diacritical mark. What users
+ * consider to be a character can differ between languages.
+ * <p>
+ * The text boundary positions are found according to the rules
+ * described in Unicode Standard Annex #29, Text Boundaries, and
+ * Unicode Standard Annex #14, Line Breaking Properties.  These
+ * are available at http://www.unicode.org/reports/tr14/ and
+ * http://www.unicode.org/reports/tr29/.
+ * <p>
+ * In addition to the C++ API defined in this header file, a
+ * plain C API with equivalent functionality is defined in the
+ * file ubrk.h
+ * <p>
+ * Code snippets illustrating the use of the Break Iterator APIs
+ * are available in the ICU User Guide,
+ * https://unicode-org.github.io/icu/userguide/boundaryanalysis/
+ * and in the sample program icu/source/samples/break/break.cpp
+ *
+ */
+class U_COMMON_API BreakIterator : public UObject {
+public:
+    /**
+     *  destructor
+     *  @stable ICU 2.0
+     */
+    virtual ~BreakIterator();
+
+    /**
+     * Return true if another object is semantically equal to this
+     * one. The other object should be an instance of the same subclass of
+     * BreakIterator. Objects of different subclasses are considered
+     * unequal.
+     * <P>
+     * Return true if this BreakIterator is at the same position in the
+     * same text, and is the same class and type (word, line, etc.) of
+     * BreakIterator, as the argument.  Text is considered the same if
+     * it contains the same characters, it need not be the same
+     * object, and styles are not considered.
+     * @stable ICU 2.0
+     */
+    virtual bool operator==(const BreakIterator&) const = 0;
+
+    /**
+     * Returns the complement of the result of operator==
+     * @param rhs The BreakIterator to be compared for inequality
+     * @return the complement of the result of operator==
+     * @stable ICU 2.0
+     */
+    bool operator!=(const BreakIterator& rhs) const { return !operator==(rhs); }
+
+    /**
+     * Return a polymorphic copy of this object.  This is an abstract
+     * method which subclasses implement.
+     * @stable ICU 2.0
+     */
+    virtual BreakIterator* clone() const = 0;
+
+    /**
+     * Return a polymorphic class ID for this object. Different subclasses
+     * will return distinct unequal values.
+     * @stable ICU 2.0
+     */
+    virtual UClassID getDynamicClassID() const override = 0;
+
+    /**
+     * Return a CharacterIterator over the text being analyzed.
+     * @stable ICU 2.0
+     */
+    virtual CharacterIterator& getText() const = 0;
+
+    /**
+      *  Get a UText for the text being analyzed.
+      *  The returned UText is a shallow clone of the UText used internally
+      *  by the break iterator implementation.  It can safely be used to
+      *  access the text without impacting any break iterator operations,
+      *  but the underlying text itself must not be altered.
+      *
+      * @param fillIn A UText to be filled in.  If nullptr, a new UText will be
+      *           allocated to hold the result.
+      * @param status receives any error codes.
+      * @return   The current UText for this break iterator.  If an input
+      *           UText was provided, it will always be returned.
+      * @stable ICU 3.4
+      */
+     virtual UText *getUText(UText *fillIn, UErrorCode &status) const = 0;
+
+    /**
+     * Change the text over which this operates. The text boundary is
+     * reset to the start.
+     *
+     * The BreakIterator will retain a reference to the supplied string.
+     * The caller must not modify or delete the text while the BreakIterator
+     * retains the reference.
+     *
+     * @param text The UnicodeString used to change the text.
+     * @stable ICU 2.0
+     */
+    virtual void  setText(const UnicodeString &text) = 0;
+
+    /**
+     * Reset the break iterator to operate over the text represented by
+     * the UText.  The iterator position is reset to the start.
+     *
+     * This function makes a shallow clone of the supplied UText.  This means
+     * that the caller is free to immediately close or otherwise reuse the
+     * Utext that was passed as a parameter, but that the underlying text itself
+     * must not be altered while being referenced by the break iterator.
+     *
+     * All index positions returned by break iterator functions are
+     * native indices from the UText. For example, when breaking UTF-8
+     * encoded text, the break positions returned by next(), previous(), etc.
+     * will be UTF-8 string indices, not UTF-16 positions.
+     *
+     * @param text The UText used to change the text.
+     * @param status receives any error codes.
+     * @stable ICU 3.4
+     */
+    virtual void  setText(UText *text, UErrorCode &status) = 0;
+
+    /**
+     * Change the text over which this operates. The text boundary is
+     * reset to the start.
+     * Note that setText(UText *) provides similar functionality to this function,
+     * and is more efficient.
+     * @param it The CharacterIterator used to change the text.
+     * @stable ICU 2.0
+     */
+    virtual void  adoptText(CharacterIterator* it) = 0;
+
+    enum {
+        /**
+         * DONE is returned by previous() and next() after all valid
+         * boundaries have been returned.
+         * @stable ICU 2.0
+         */
+        DONE = static_cast<int32_t>(-1)
+    };
+
+    /**
+     * Sets the current iteration position to the beginning of the text, position zero.
+     * @return The offset of the beginning of the text, zero.
+     * @stable ICU 2.0
+     */
+    virtual int32_t first() = 0;
+
+    /**
+     * Set the iterator position to the index immediately BEYOND the last character in the text being scanned.
+     * @return The index immediately BEYOND the last character in the text being scanned.
+     * @stable ICU 2.0
+     */
+    virtual int32_t last() = 0;
+
+    /**
+     * Set the iterator position to the boundary preceding the current boundary.
+     * @return The character index of the previous text boundary or DONE if all
+     * boundaries have been returned.
+     * @stable ICU 2.0
+     */
+    virtual int32_t previous() = 0;
+
+    /**
+     * Advance the iterator to the boundary following the current boundary.
+     * @return The character index of the next text boundary or DONE if all
+     * boundaries have been returned.
+     * @stable ICU 2.0
+     */
+    virtual int32_t next() = 0;
+
+    /**
+     * Return character index of the current iterator position within the text.
+     * @return The boundary most recently returned.
+     * @stable ICU 2.0
+     */
+    virtual int32_t current() const = 0;
+
+    /**
+     * Advance the iterator to the first boundary following the specified offset.
+     * The value returned is always greater than the offset or
+     * the value BreakIterator.DONE
+     * @param offset the offset to begin scanning.
+     * @return The first boundary after the specified offset.
+     * @stable ICU 2.0
+     */
+    virtual int32_t following(int32_t offset) = 0;
+
+    /**
+     * Set the iterator position to the first boundary preceding the specified offset.
+     * The value returned is always smaller than the offset or
+     * the value BreakIterator.DONE
+     * @param offset the offset to begin scanning.
+     * @return The first boundary before the specified offset.
+     * @stable ICU 2.0
+     */
+    virtual int32_t preceding(int32_t offset) = 0;
+
+    /**
+     * Return true if the specified position is a boundary position.
+     * As a side effect, the current position of the iterator is set
+     * to the first boundary position at or following the specified offset.
+     * @param offset the offset to check.
+     * @return True if "offset" is a boundary position.
+     * @stable ICU 2.0
+     */
+    virtual UBool isBoundary(int32_t offset) = 0;
+
+    /**
+     * Set the iterator position to the nth boundary from the current boundary
+     * @param n the number of boundaries to move by.  A value of 0
+     * does nothing.  Negative values move to previous boundaries
+     * and positive values move to later boundaries.
+     * @return The new iterator position, or
+     * DONE if there are fewer than |n| boundaries in the specified direction.
+     * @stable ICU 2.0
+     */
+    virtual int32_t next(int32_t n) = 0;
+
+   /**
+     * For RuleBasedBreakIterators, return the status tag from the break rule
+     * that determined the boundary at the current iteration position.
+     * <p>
+     * For break iterator types that do not support a rule status,
+     * a default value of 0 is returned.
+     * <p>
+     * @return the status from the break rule that determined the boundary at
+     *         the current iteration position.
+     * @see RuleBaseBreakIterator::getRuleStatus()
+     * @see UWordBreak
+     * @stable ICU 52
+     */
+    virtual int32_t getRuleStatus() const;
+
+   /**
+    * For RuleBasedBreakIterators, get the status (tag) values from the break rule(s)
+    * that determined the boundary at the current iteration position.
+    * <p>
+    * For break iterator types that do not support rule status,
+    * no values are returned.
+    * <p>
+    * The returned status value(s) are stored into an array provided by the caller.
+    * The values are stored in sorted (ascending) order.
+    * If the capacity of the output array is insufficient to hold the data,
+    *  the output will be truncated to the available length, and a
+    *  U_BUFFER_OVERFLOW_ERROR will be signaled.
+    * <p>
+    * @see RuleBaseBreakIterator::getRuleStatusVec
+    *
+    * @param fillInVec an array to be filled in with the status values.
+    * @param capacity  the length of the supplied vector.  A length of zero causes
+    *                  the function to return the number of status values, in the
+    *                  normal way, without attempting to store any values.
+    * @param status    receives error codes.
+    * @return          The number of rule status values from rules that determined
+    *                  the boundary at the current iteration position.
+    *                  In the event of a U_BUFFER_OVERFLOW_ERROR, the return value
+    *                  is the total number of status values that were available,
+    *                  not the reduced number that were actually returned.
+    * @see getRuleStatus
+    * @stable ICU 52
+    */
+    virtual int32_t getRuleStatusVec(int32_t *fillInVec, int32_t capacity, UErrorCode &status);
+
+    /**
+     * Create BreakIterator for word-breaks using the given locale.
+     * Returns an instance of a BreakIterator implementing word breaks.
+     * WordBreak is useful for word selection (ex. double click)
+     * @param where the locale.
+     * @param status the error code
+     * @return A BreakIterator for word-breaks.  The UErrorCode& status
+     * parameter is used to return status information to the user.
+     * To check whether the construction succeeded or not, you should check
+     * the value of U_SUCCESS(err).  If you wish more detailed information, you
+     * can check for informational error results which still indicate success.
+     * U_USING_FALLBACK_WARNING indicates that a fall back locale was used.  For
+     * example, 'de_CH' was requested, but nothing was found there, so 'de' was
+     * used.  U_USING_DEFAULT_WARNING indicates that the default locale data was
+     * used; neither the requested locale nor any of its fall back locales
+     * could be found.
+     * The caller owns the returned object and is responsible for deleting it.
+     * @stable ICU 2.0
+     */
+    static BreakIterator* U_EXPORT2
+    createWordInstance(const Locale& where, UErrorCode& status);
+
+    /**
+     * Create BreakIterator for line-breaks using specified locale.
+     * Returns an instance of a BreakIterator implementing line breaks. Line
+     * breaks are logically possible line breaks, actual line breaks are
+     * usually determined based on display width.
+     * LineBreak is useful for word wrapping text.
+     * @param where the locale.
+     * @param status The error code.
+     * @return A BreakIterator for line-breaks.  The UErrorCode& status
+     * parameter is used to return status information to the user.
+     * To check whether the construction succeeded or not, you should check
+     * the value of U_SUCCESS(err).  If you wish more detailed information, you
+     * can check for informational error results which still indicate success.
+     * U_USING_FALLBACK_WARNING indicates that a fall back locale was used.  For
+     * example, 'de_CH' was requested, but nothing was found there, so 'de' was
+     * used.  U_USING_DEFAULT_WARNING indicates that the default locale data was
+     * used; neither the requested locale nor any of its fall back locales
+     * could be found.
+     * The caller owns the returned object and is responsible for deleting it.
+     * @stable ICU 2.0
+     */
+    static BreakIterator* U_EXPORT2
+    createLineInstance(const Locale& where, UErrorCode& status);
+
+    /**
+     * Create BreakIterator for character-breaks using specified locale
+     * Returns an instance of a BreakIterator implementing character breaks.
+     * Character breaks are boundaries of combining character sequences.
+     * @param where the locale.
+     * @param status The error code.
+     * @return A BreakIterator for character-breaks.  The UErrorCode& status
+     * parameter is used to return status information to the user.
+     * To check whether the construction succeeded or not, you should check
+     * the value of U_SUCCESS(err).  If you wish more detailed information, you
+     * can check for informational error results which still indicate success.
+     * U_USING_FALLBACK_WARNING indicates that a fall back locale was used.  For
+     * example, 'de_CH' was requested, but nothing was found there, so 'de' was
+     * used.  U_USING_DEFAULT_WARNING indicates that the default locale data was
+     * used; neither the requested locale nor any of its fall back locales
+     * could be found.
+     * The caller owns the returned object and is responsible for deleting it.
+     * @stable ICU 2.0
+     */
+    static BreakIterator* U_EXPORT2
+    createCharacterInstance(const Locale& where, UErrorCode& status);
+
+    /**
+     * Create BreakIterator for sentence-breaks using specified locale
+     * Returns an instance of a BreakIterator implementing sentence breaks.
+     * @param where the locale.
+     * @param status The error code.
+     * @return A BreakIterator for sentence-breaks.  The UErrorCode& status
+     * parameter is used to return status information to the user.
+     * To check whether the construction succeeded or not, you should check
+     * the value of U_SUCCESS(err).  If you wish more detailed information, you
+     * can check for informational error results which still indicate success.
+     * U_USING_FALLBACK_WARNING indicates that a fall back locale was used.  For
+     * example, 'de_CH' was requested, but nothing was found there, so 'de' was
+     * used.  U_USING_DEFAULT_WARNING indicates that the default locale data was
+     * used; neither the requested locale nor any of its fall back locales
+     * could be found.
+     * The caller owns the returned object and is responsible for deleting it.
+     * @stable ICU 2.0
+     */
+    static BreakIterator* U_EXPORT2
+    createSentenceInstance(const Locale& where, UErrorCode& status);
+
+#ifndef U_HIDE_DEPRECATED_API
+    /**
+     * Create BreakIterator for title-casing breaks using the specified locale
+     * Returns an instance of a BreakIterator implementing title breaks.
+     * The iterator returned locates title boundaries as described for
+     * Unicode 3.2 only. For Unicode 4.0 and above title boundary iteration,
+     * please use a word boundary iterator. See {@link #createWordInstance }.
+     *
+     * @param where the locale.
+     * @param status The error code.
+     * @return A BreakIterator for title-breaks.  The UErrorCode& status
+     * parameter is used to return status information to the user.
+     * To check whether the construction succeeded or not, you should check
+     * the value of U_SUCCESS(err).  If you wish more detailed information, you
+     * can check for informational error results which still indicate success.
+     * U_USING_FALLBACK_WARNING indicates that a fall back locale was used.  For
+     * example, 'de_CH' was requested, but nothing was found there, so 'de' was
+     * used.  U_USING_DEFAULT_WARNING indicates that the default locale data was
+     * used; neither the requested locale nor any of its fall back locales
+     * could be found.
+     * The caller owns the returned object and is responsible for deleting it.
+     * @deprecated ICU 64 Use createWordInstance instead.
+     */
+    static BreakIterator* U_EXPORT2
+    createTitleInstance(const Locale& where, UErrorCode& status);
+#endif /* U_HIDE_DEPRECATED_API */
+
+    /**
+     * Get the set of Locales for which TextBoundaries are installed.
+     * <p><b>Note:</b> this will not return locales added through the register
+     * call. To see the registered locales too, use the getAvailableLocales
+     * function that returns a StringEnumeration object </p>
+     * @param count the output parameter of number of elements in the locale list
+     * @return available locales
+     * @stable ICU 2.0
+     */
+    static const Locale* U_EXPORT2 getAvailableLocales(int32_t& count);
+
+    /**
+     * Get name of the object for the desired Locale, in the desired language.
+     * @param objectLocale must be from getAvailableLocales.
+     * @param displayLocale specifies the desired locale for output.
+     * @param name the fill-in parameter of the return value
+     * Uses best match.
+     * @return user-displayable name
+     * @stable ICU 2.0
+     */
+    static UnicodeString& U_EXPORT2 getDisplayName(const Locale& objectLocale,
+                                         const Locale& displayLocale,
+                                         UnicodeString& name);
+
+    /**
+     * Get name of the object for the desired Locale, in the language of the
+     * default locale.
+     * @param objectLocale must be from getMatchingLocales
+     * @param name the fill-in parameter of the return value
+     * @return user-displayable name
+     * @stable ICU 2.0
+     */
+    static UnicodeString& U_EXPORT2 getDisplayName(const Locale& objectLocale,
+                                         UnicodeString& name);
+
+#ifndef U_FORCE_HIDE_DEPRECATED_API
+    /**
+     * Deprecated functionality. Use clone() instead.
+     *
+     * Thread safe client-buffer-based cloning operation
+     *    Do NOT call delete on a safeclone, since 'new' is not used to create it.
+     * @param stackBuffer user allocated space for the new clone. If nullptr new memory will be allocated.
+     * If buffer is not large enough, new memory will be allocated.
+     * @param BufferSize reference to size of allocated space.
+     * If BufferSize == 0, a sufficient size for use in cloning will
+     * be returned ('pre-flighting')
+     * If BufferSize is not enough for a stack-based safe clone,
+     * new memory will be allocated.
+     * @param status to indicate whether the operation went on smoothly or there were errors
+     *  An informational status value, U_SAFECLONE_ALLOCATED_ERROR, is used if any allocations were
+     *  necessary.
+     * @return pointer to the new clone
+     *
+     * @deprecated ICU 52. Use clone() instead.
+     */
+    virtual BreakIterator *  createBufferClone(void *stackBuffer,
+                                               int32_t &BufferSize,
+                                               UErrorCode &status) = 0;
+#endif  // U_FORCE_HIDE_DEPRECATED_API
+
+#ifndef U_HIDE_DEPRECATED_API
+
+    /**
+     *   Determine whether the BreakIterator was created in user memory by
+     *   createBufferClone(), and thus should not be deleted.  Such objects
+     *   must be closed by an explicit call to the destructor (not delete).
+     * @deprecated ICU 52. Always delete the BreakIterator.
+     */
+    inline UBool isBufferClone();
+
+#endif /* U_HIDE_DEPRECATED_API */
+
+#if !UCONFIG_NO_SERVICE
+    /**
+     * Register a new break iterator of the indicated kind, to use in the given locale.
+     * The break iterator will be adopted.  Clones of the iterator will be returned
+     * if a request for a break iterator of the given kind matches or falls back to
+     * this locale.
+     * Because ICU may choose to cache BreakIterators internally, this must
+     * be called at application startup, prior to any calls to
+     * BreakIterator::createXXXInstance to avoid undefined behavior.
+     * @param toAdopt the BreakIterator instance to be adopted
+     * @param locale the Locale for which this instance is to be registered
+     * @param kind the type of iterator for which this instance is to be registered
+     * @param status the in/out status code, no special meanings are assigned
+     * @return a registry key that can be used to unregister this instance
+     * @stable ICU 2.4
+     */
+    static URegistryKey U_EXPORT2 registerInstance(BreakIterator* toAdopt,
+                                        const Locale& locale,
+                                        UBreakIteratorType kind,
+                                        UErrorCode& status);
+
+    /**
+     * Unregister a previously-registered BreakIterator using the key returned from the
+     * register call.  Key becomes invalid after a successful call and should not be used again.
+     * The BreakIterator corresponding to the key will be deleted.
+     * Because ICU may choose to cache BreakIterators internally, this should
+     * be called during application shutdown, after all calls to
+     * BreakIterator::createXXXInstance to avoid undefined behavior.
+     * @param key the registry key returned by a previous call to registerInstance
+     * @param status the in/out status code, no special meanings are assigned
+     * @return true if the iterator for the key was successfully unregistered
+     * @stable ICU 2.4
+     */
+    static UBool U_EXPORT2 unregister(URegistryKey key, UErrorCode& status);
+
+    /**
+     * Return a StringEnumeration over the locales available at the time of the call,
+     * including registered locales.
+     * @return a StringEnumeration over the locales available at the time of the call
+     * @stable ICU 2.4
+     */
+    static StringEnumeration* U_EXPORT2 getAvailableLocales();
+#endif
+
+    /**
+     * Returns the locale for this break iterator. Two flavors are available: valid and
+     * actual locale.
+     * @stable ICU 2.8
+     */
+    Locale getLocale(ULocDataLocaleType type, UErrorCode& status) const;
+
+#ifndef U_HIDE_INTERNAL_API
+    /** Get the locale for this break iterator object. You can choose between valid and actual locale.
+     *  @param type type of the locale we're looking for (valid or actual)
+     *  @param status error code for the operation
+     *  @return the locale
+     *  @internal
+     */
+    const char *getLocaleID(ULocDataLocaleType type, UErrorCode& status) const;
+#endif  /* U_HIDE_INTERNAL_API */
+
+    /**
+     *  Set the subject text string upon which the break iterator is operating
+     *  without changing any other aspect of the matching state.
+     *  The new and previous text strings must have the same content.
+     *
+     *  This function is intended for use in environments where ICU is operating on
+     *  strings that may move around in memory.  It provides a mechanism for notifying
+     *  ICU that the string has been relocated, and providing a new UText to access the
+     *  string in its new position.
+     *
+     *  Note that the break iterator implementation never copies the underlying text
+     *  of a string being processed, but always operates directly on the original text
+     *  provided by the user. Refreshing simply drops the references to the old text
+     *  and replaces them with references to the new.
+     *
+     *  Caution:  this function is normally used only by very specialized,
+     *  system-level code.  One example use case is with garbage collection that moves
+     *  the text in memory.
+     *
+     * @param input      The new (moved) text string.
+     * @param status     Receives errors detected by this function.
+     * @return           *this
+     *
+     * @stable ICU 49
+     */
+    virtual BreakIterator &refreshInputText(UText *input, UErrorCode &status) = 0;
+
+ private:
+    static BreakIterator* buildInstance(const Locale& loc, const char *type, UErrorCode& status);
+    static BreakIterator* createInstance(const Locale& loc, int32_t kind, UErrorCode& status);
+    static BreakIterator* makeInstance(const Locale& loc, int32_t kind, UErrorCode& status);
+
+    friend class ICUBreakIteratorFactory;
+    friend class ICUBreakIteratorService;
+
+protected:
+    // Do not enclose protected default/copy constructors with #ifndef U_HIDE_INTERNAL_API
+    // or else the compiler will create a public ones.
+    /** @internal */
+    BreakIterator();
+    /** @internal */
+    BreakIterator (const BreakIterator &other);
+#ifndef U_HIDE_INTERNAL_API
+    /** @internal */
+    BreakIterator (const Locale& valid, const Locale &actual);
+    /** @internal. Assignment Operator, used by RuleBasedBreakIterator. */
+    BreakIterator &operator = (const BreakIterator &other);
+#endif  /* U_HIDE_INTERNAL_API */
+
+private:
+
+    /** @internal (private) */
+    CharString* actualLocale = nullptr;
+    CharString* validLocale = nullptr;
+    CharString* requestLocale = nullptr;
+};
+
+#ifndef U_HIDE_DEPRECATED_API
+
+inline UBool BreakIterator::isBufferClone()
+{
+    return false;
+}
+
+#endif /* U_HIDE_DEPRECATED_API */
+
+U_NAMESPACE_END
+
+#endif /* #if !UCONFIG_NO_BREAK_ITERATION */
+
+#endif /* U_SHOW_CPLUSPLUS_API */
+
+#endif // BRKITER_H
+//eof

thirdparty/icu4c/common/unicode/bytestream.h

@@ -0,0 +1,307 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+// Copyright (C) 2009-2012, International Business Machines
+// Corporation and others. All Rights Reserved.
+//
+// Copyright 2007 Google Inc. All Rights Reserved.
+// Author: sanjay@google.com (Sanjay Ghemawat)
+//
+// Abstract interface that consumes a sequence of bytes (ByteSink).
+//
+// Used so that we can write a single piece of code that can operate
+// on a variety of output string types.
+//
+// Various implementations of this interface are provided:
+//   ByteSink:
+//      CheckedArrayByteSink    Write to a flat array, with bounds checking
+//      StringByteSink          Write to an STL string
+
+// This code is a contribution of Google code, and the style used here is
+// a compromise between the original Google code and the ICU coding guidelines.
+// For example, data types are ICU-ified (size_t,int->int32_t),
+// and API comments doxygen-ified, but function names and behavior are
+// as in the original, if possible.
+// Assertion-style error handling, not available in ICU, was changed to
+// parameter "pinning" similar to UnicodeString.
+//
+// In addition, this is only a partial port of the original Google code,
+// limited to what was needed so far. The (nearly) complete original code
+// is in the ICU svn repository at icuhtml/trunk/design/strings/contrib
+// (see ICU ticket 6765, r25517).
+
+#ifndef __BYTESTREAM_H__
+#define __BYTESTREAM_H__
+
+/**
+ * \file
+ * \brief C++ API: Interface for writing bytes, and implementation classes.
+ */
+
+#include "unicode/utypes.h"
+
+#if U_SHOW_CPLUSPLUS_API
+
+#include "unicode/uobject.h"
+#include "unicode/std_string.h"
+
+U_NAMESPACE_BEGIN
+
+/**
+ * A ByteSink can be filled with bytes.
+ * @stable ICU 4.2
+ */
+class U_COMMON_API ByteSink : public UMemory {
+public:
+  /**
+   * Default constructor.
+   * @stable ICU 4.2
+   */
+  ByteSink() { }
+  /**
+   * Virtual destructor.
+   * @stable ICU 4.2
+   */
+  virtual ~ByteSink();
+
+  /**
+   * Append "bytes[0,n-1]" to this.
+   * @param bytes the pointer to the bytes
+   * @param n the number of bytes; must be non-negative
+   * @stable ICU 4.2
+   */
+  virtual void Append(const char* bytes, int32_t n) = 0;
+
+  /**
+   * Appends n bytes to this. Same as Append().
+   * Call AppendU8() with u8"string literals" which are const char * in C++11
+   * but const char8_t * in C++20.
+   * If the compiler does support char8_t as a distinct type,
+   * then an AppendU8() overload for that is defined and will be chosen.
+   *
+   * @param bytes the pointer to the bytes
+   * @param n the number of bytes; must be non-negative
+   * @stable ICU 67
+   */
+  inline void AppendU8(const char* bytes, int32_t n) {
+    Append(bytes, n);
+  }
+
+#if defined(__cpp_char8_t) || defined(U_IN_DOXYGEN)
+  /**
+   * Appends n bytes to this. Same as Append() but for a const char8_t * pointer.
+   * Call AppendU8() with u8"string literals" which are const char * in C++11
+   * but const char8_t * in C++20.
+   * If the compiler does support char8_t as a distinct type,
+   * then this AppendU8() overload for that is defined and will be chosen.
+   *
+   * @param bytes the pointer to the bytes
+   * @param n the number of bytes; must be non-negative
+   * @stable ICU 67
+   */
+  inline void AppendU8(const char8_t* bytes, int32_t n) {
+    Append(reinterpret_cast<const char*>(bytes), n);
+  }
+#endif
+
+  /**
+   * Returns a writable buffer for appending and writes the buffer's capacity to
+   * *result_capacity. Guarantees *result_capacity>=min_capacity.
+   * May return a pointer to the caller-owned scratch buffer which must have
+   * scratch_capacity>=min_capacity.
+   * The returned buffer is only valid until the next operation
+   * on this ByteSink.
+   *
+   * After writing at most *result_capacity bytes, call Append() with the
+   * pointer returned from this function and the number of bytes written.
+   * Many Append() implementations will avoid copying bytes if this function
+   * returned an internal buffer.
+   *
+   * Partial usage example:
+   *  int32_t capacity;
+   *  char* buffer = sink->GetAppendBuffer(..., &capacity);
+   *  ... Write n bytes into buffer, with n <= capacity.
+   *  sink->Append(buffer, n);
+   * In many implementations, that call to Append will avoid copying bytes.
+   *
+   * If the ByteSink allocates or reallocates an internal buffer, it should use
+   * the desired_capacity_hint if appropriate.
+   * If a caller cannot provide a reasonable guess at the desired capacity,
+   * it should pass desired_capacity_hint=0.
+   *
+   * If a non-scratch buffer is returned, the caller may only pass
+   * a prefix to it to Append().
+   * That is, it is not correct to pass an interior pointer to Append().
+   *
+   * The default implementation always returns the scratch buffer.
+   *
+   * @param min_capacity required minimum capacity of the returned buffer;
+   *                     must be non-negative
+   * @param desired_capacity_hint desired capacity of the returned buffer;
+   *                              must be non-negative
+   * @param scratch default caller-owned buffer
+   * @param scratch_capacity capacity of the scratch buffer
+   * @param result_capacity pointer to an integer which will be set to the
+   *                        capacity of the returned buffer
+   * @return a buffer with *result_capacity>=min_capacity
+   * @stable ICU 4.2
+   */
+  virtual char* GetAppendBuffer(int32_t min_capacity,
+                                int32_t desired_capacity_hint,
+                                char* scratch, int32_t scratch_capacity,
+                                int32_t* result_capacity);
+
+  /**
+   * Flush internal buffers.
+   * Some byte sinks use internal buffers or provide buffering
+   * and require calling Flush() at the end of the stream.
+   * The ByteSink should be ready for further Append() calls after Flush().
+   * The default implementation of Flush() does nothing.
+   * @stable ICU 4.2
+   */
+  virtual void Flush();
+
+private:
+  ByteSink(const ByteSink &) = delete;
+  ByteSink &operator=(const ByteSink &) = delete;
+};
+
+// -------------------------------------------------------------
+// Some standard implementations
+
+/** 
+ * Implementation of ByteSink that writes to a flat byte array,
+ * with bounds-checking:
+ * This sink will not write more than capacity bytes to outbuf.
+ * If more than capacity bytes are Append()ed, then excess bytes are ignored,
+ * and Overflowed() will return true.
+ * Overflow does not cause a runtime error.
+ * @stable ICU 4.2
+ */
+class U_COMMON_API CheckedArrayByteSink : public ByteSink {
+public:
+  /**
+   * Constructs a ByteSink that will write to outbuf[0..capacity-1].
+   * @param outbuf buffer to write to
+   * @param capacity size of the buffer
+   * @stable ICU 4.2
+   */
+  CheckedArrayByteSink(char* outbuf, int32_t capacity);
+  /**
+   * Destructor.
+   * @stable ICU 4.2
+   */
+  virtual ~CheckedArrayByteSink();
+  /**
+   * Returns the sink to its original state, without modifying the buffer.
+   * Useful for reusing both the buffer and the sink for multiple streams.
+   * Resets the state to NumberOfBytesWritten()=NumberOfBytesAppended()=0
+   * and Overflowed()=false.
+   * @return *this
+   * @stable ICU 4.6
+   */
+  virtual CheckedArrayByteSink& Reset();
+  /**
+   * Append "bytes[0,n-1]" to this.
+   * @param bytes the pointer to the bytes
+   * @param n the number of bytes; must be non-negative
+   * @stable ICU 4.2
+   */
+  virtual void Append(const char* bytes, int32_t n) override;
+  /**
+   * Returns a writable buffer for appending and writes the buffer's capacity to
+   * *result_capacity. For details see the base class documentation.
+   * @param min_capacity required minimum capacity of the returned buffer;
+   *                     must be non-negative
+   * @param desired_capacity_hint desired capacity of the returned buffer;
+   *                              must be non-negative
+   * @param scratch default caller-owned buffer
+   * @param scratch_capacity capacity of the scratch buffer
+   * @param result_capacity pointer to an integer which will be set to the
+   *                        capacity of the returned buffer
+   * @return a buffer with *result_capacity>=min_capacity
+   * @stable ICU 4.2
+   */
+  virtual char* GetAppendBuffer(int32_t min_capacity,
+                                int32_t desired_capacity_hint,
+                                char* scratch, int32_t scratch_capacity,
+                                int32_t* result_capacity) override;
+  /**
+   * Returns the number of bytes actually written to the sink.
+   * @return number of bytes written to the buffer
+   * @stable ICU 4.2
+   */
+  int32_t NumberOfBytesWritten() const { return size_; }
+  /**
+   * Returns true if any bytes were discarded, i.e., if there was an
+   * attempt to write more than 'capacity' bytes.
+   * @return true if more than 'capacity' bytes were Append()ed
+   * @stable ICU 4.2
+   */
+  UBool Overflowed() const { return overflowed_; }
+  /**
+   * Returns the number of bytes appended to the sink.
+   * If Overflowed() then NumberOfBytesAppended()>NumberOfBytesWritten()
+   * else they return the same number.
+   * @return number of bytes written to the buffer
+   * @stable ICU 4.6
+   */
+  int32_t NumberOfBytesAppended() const { return appended_; }
+private:
+  char* outbuf_;
+  const int32_t capacity_;
+  int32_t size_;
+  int32_t appended_;
+  UBool overflowed_;
+
+  CheckedArrayByteSink() = delete;
+  CheckedArrayByteSink(const CheckedArrayByteSink &) = delete;
+  CheckedArrayByteSink &operator=(const CheckedArrayByteSink &) = delete;
+};
+
+/** 
+ * Implementation of ByteSink that writes to a "string".
+ * The StringClass is usually instantiated with a std::string.
+ * @stable ICU 4.2
+ */
+template<typename StringClass>
+class StringByteSink : public ByteSink {
+ public:
+  /**
+   * Constructs a ByteSink that will append bytes to the dest string.
+   * @param dest pointer to string object to append to
+   * @stable ICU 4.2
+   */
+  StringByteSink(StringClass* dest) : dest_(dest) { }
+  /**
+   * Constructs a ByteSink that reserves append capacity and will append bytes to the dest string.
+   * 
+   * @param dest pointer to string object to append to
+   * @param initialAppendCapacity capacity beyond dest->length() to be reserve()d
+   * @stable ICU 60
+   */
+  StringByteSink(StringClass* dest, int32_t initialAppendCapacity) : dest_(dest) {
+    if (initialAppendCapacity > 0 &&
+        static_cast<uint32_t>(initialAppendCapacity) > dest->capacity() - dest->length()) {
+      dest->reserve(dest->length() + initialAppendCapacity);
+    }
+  }
+  /**
+   * Append "bytes[0,n-1]" to this.
+   * @param data the pointer to the bytes
+   * @param n the number of bytes; must be non-negative
+   * @stable ICU 4.2
+   */
+  virtual void Append(const char* data, int32_t n) override { dest_->append(data, n); }
+ private:
+  StringClass* dest_;
+
+  StringByteSink() = delete;
+  StringByteSink(const StringByteSink &) = delete;
+  StringByteSink &operator=(const StringByteSink &) = delete;
+};
+
+U_NAMESPACE_END
+
+#endif /* U_SHOW_CPLUSPLUS_API */
+
+#endif  // __BYTESTREAM_H__

thirdparty/icu4c/common/unicode/bytestrie.h

@@ -0,0 +1,568 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+/*
+*******************************************************************************
+*   Copyright (C) 2010-2012, International Business Machines
+*   Corporation and others.  All Rights Reserved.
+*******************************************************************************
+*   file name:  bytestrie.h
+*   encoding:   UTF-8
+*   tab size:   8 (not used)
+*   indentation:4
+*
+*   created on: 2010sep25
+*   created by: Markus W. Scherer
+*/
+
+#ifndef __BYTESTRIE_H__
+#define __BYTESTRIE_H__
+
+/**
+ * \file
+ * \brief C++ API: Trie for mapping byte sequences to integer values.
+ */
+
+#include "unicode/utypes.h"
+
+#if U_SHOW_CPLUSPLUS_API
+
+#include "unicode/stringpiece.h"
+#include "unicode/uobject.h"
+#include "unicode/ustringtrie.h"
+
+class BytesTrieTest;
+
+U_NAMESPACE_BEGIN
+
+class ByteSink;
+class BytesTrieBuilder;
+class CharString;
+class UVector32;
+
+/**
+ * Light-weight, non-const reader class for a BytesTrie.
+ * Traverses a byte-serialized data structure with minimal state,
+ * for mapping byte sequences to non-negative integer values.
+ *
+ * This class owns the serialized trie data only if it was constructed by
+ * the builder's build() method.
+ * The public constructor and the copy constructor only alias the data (only copy the pointer).
+ * There is no assignment operator.
+ *
+ * This class is not intended for public subclassing.
+ * @stable ICU 4.8
+ */
+class U_COMMON_API BytesTrie : public UMemory {
+public:
+    /**
+     * Constructs a BytesTrie reader instance.
+     *
+     * The trieBytes must contain a copy of a byte sequence from the BytesTrieBuilder,
+     * starting with the first byte of that sequence.
+     * The BytesTrie object will not read more bytes than
+     * the BytesTrieBuilder generated in the corresponding build() call.
+     *
+     * The array is not copied/cloned and must not be modified while
+     * the BytesTrie object is in use.
+     *
+     * @param trieBytes The byte array that contains the serialized trie.
+     * @stable ICU 4.8
+     */
+    BytesTrie(const void *trieBytes)
+            : ownedArray_(nullptr), bytes_(static_cast<const uint8_t *>(trieBytes)),
+              pos_(bytes_), remainingMatchLength_(-1) {}
+
+    /**
+     * Destructor.
+     * @stable ICU 4.8
+     */
+    ~BytesTrie();
+
+    /**
+     * Copy constructor, copies the other trie reader object and its state,
+     * but not the byte array which will be shared. (Shallow copy.)
+     * @param other Another BytesTrie object.
+     * @stable ICU 4.8
+     */
+    BytesTrie(const BytesTrie &other)
+            : ownedArray_(nullptr), bytes_(other.bytes_),
+              pos_(other.pos_), remainingMatchLength_(other.remainingMatchLength_) {}
+
+    /**
+     * Resets this trie to its initial state.
+     * @return *this
+     * @stable ICU 4.8
+     */
+    BytesTrie &reset() {
+        pos_=bytes_;
+        remainingMatchLength_=-1;
+        return *this;
+    }
+
+    /**
+     * Returns the state of this trie as a 64-bit integer.
+     * The state value is never 0.
+     *
+     * @return opaque state value
+     * @see resetToState64
+     * @stable ICU 65
+     */
+    uint64_t getState64() const {
+        return (static_cast<uint64_t>(remainingMatchLength_ + 2) << kState64RemainingShift) |
+            static_cast<uint64_t>(pos_ - bytes_);
+    }
+
+    /**
+     * Resets this trie to the saved state.
+     * Unlike resetToState(State), the 64-bit state value
+     * must be from getState64() from the same trie object or
+     * from one initialized the exact same way.
+     * Because of no validation, this method is faster.
+     *
+     * @param state The opaque trie state value from getState64().
+     * @return *this
+     * @see getState64
+     * @see resetToState
+     * @see reset
+     * @stable ICU 65
+     */
+    BytesTrie &resetToState64(uint64_t state) {
+        remainingMatchLength_ = static_cast<int32_t>(state >> kState64RemainingShift) - 2;
+        pos_ = bytes_ + (state & kState64PosMask);
+        return *this;
+    }
+
+    /**
+     * BytesTrie state object, for saving a trie's current state
+     * and resetting the trie back to this state later.
+     * @stable ICU 4.8
+     */
+    class State : public UMemory {
+    public:
+        /**
+         * Constructs an empty State.
+         * @stable ICU 4.8
+         */
+        State() { bytes=nullptr; }
+    private:
+        friend class BytesTrie;
+
+        const uint8_t *bytes;
+        const uint8_t *pos;
+        int32_t remainingMatchLength;
+    };
+
+    /**
+     * Saves the state of this trie.
+     * @param state The State object to hold the trie's state.
+     * @return *this
+     * @see resetToState
+     * @stable ICU 4.8
+     */
+    const BytesTrie &saveState(State &state) const {
+        state.bytes=bytes_;
+        state.pos=pos_;
+        state.remainingMatchLength=remainingMatchLength_;
+        return *this;
+    }
+
+    /**
+     * Resets this trie to the saved state.
+     * If the state object contains no state, or the state of a different trie,
+     * then this trie remains unchanged.
+     * @param state The State object which holds a saved trie state.
+     * @return *this
+     * @see saveState
+     * @see reset
+     * @stable ICU 4.8
+     */
+    BytesTrie &resetToState(const State &state) {
+        if(bytes_==state.bytes && bytes_!=nullptr) {
+            pos_=state.pos;
+            remainingMatchLength_=state.remainingMatchLength;
+        }
+        return *this;
+    }
+
+    /**
+     * Determines whether the byte sequence so far matches, whether it has a value,
+     * and whether another input byte can continue a matching byte sequence.
+     * @return The match/value Result.
+     * @stable ICU 4.8
+     */
+    UStringTrieResult current() const;
+
+    /**
+     * Traverses the trie from the initial state for this input byte.
+     * Equivalent to reset().next(inByte).
+     * @param inByte Input byte value. Values -0x100..-1 are treated like 0..0xff.
+     *               Values below -0x100 and above 0xff will never match.
+     * @return The match/value Result.
+     * @stable ICU 4.8
+     */
+    inline UStringTrieResult first(int32_t inByte) {
+        remainingMatchLength_=-1;
+        if(inByte<0) {
+            inByte+=0x100;
+        }
+        return nextImpl(bytes_, inByte);
+    }
+
+    /**
+     * Traverses the trie from the current state for this input byte.
+     * @param inByte Input byte value. Values -0x100..-1 are treated like 0..0xff.
+     *               Values below -0x100 and above 0xff will never match.
+     * @return The match/value Result.
+     * @stable ICU 4.8
+     */
+    UStringTrieResult next(int32_t inByte);
+
+    /**
+     * Traverses the trie from the current state for this byte sequence.
+     * Equivalent to
+     * \code
+     * Result result=current();
+     * for(each c in s)
+     *   if(!USTRINGTRIE_HAS_NEXT(result)) return USTRINGTRIE_NO_MATCH;
+     *   result=next(c);
+     * return result;
+     * \endcode
+     * @param s A string or byte sequence. Can be nullptr if length is 0.
+     * @param length The length of the byte sequence. Can be -1 if NUL-terminated.
+     * @return The match/value Result.
+     * @stable ICU 4.8
+     */
+    UStringTrieResult next(const char *s, int32_t length);
+
+    /**
+     * Returns a matching byte sequence's value if called immediately after
+     * current()/first()/next() returned USTRINGTRIE_INTERMEDIATE_VALUE or USTRINGTRIE_FINAL_VALUE.
+     * getValue() can be called multiple times.
+     *
+     * Do not call getValue() after USTRINGTRIE_NO_MATCH or USTRINGTRIE_NO_VALUE!
+     * @return The value for the byte sequence so far.
+     * @stable ICU 4.8
+     */
+    inline int32_t getValue() const {
+        const uint8_t *pos=pos_;
+        int32_t leadByte=*pos++;
+        // U_ASSERT(leadByte>=kMinValueLead);
+        return readValue(pos, leadByte>>1);
+    }
+
+    /**
+     * Determines whether all byte sequences reachable from the current state
+     * map to the same value.
+     * @param uniqueValue Receives the unique value, if this function returns true.
+     *                    (output-only)
+     * @return true if all byte sequences reachable from the current state
+     *         map to the same value.
+     * @stable ICU 4.8
+     */
+    inline UBool hasUniqueValue(int32_t &uniqueValue) const {
+        const uint8_t *pos=pos_;
+        // Skip the rest of a pending linear-match node.
+        return pos!=nullptr && findUniqueValue(pos+remainingMatchLength_+1, false, uniqueValue);
+    }
+
+    /**
+     * Finds each byte which continues the byte sequence from the current state.
+     * That is, each byte b for which it would be next(b)!=USTRINGTRIE_NO_MATCH now.
+     * @param out Each next byte is appended to this object.
+     *            (Only uses the out.Append(s, length) method.)
+     * @return the number of bytes which continue the byte sequence from here
+     * @stable ICU 4.8
+     */
+    int32_t getNextBytes(ByteSink &out) const;
+
+    /**
+     * Iterator for all of the (byte sequence, value) pairs in a BytesTrie.
+     * @stable ICU 4.8
+     */
+    class U_COMMON_API Iterator : public UMemory {
+    public:
+        /**
+         * Iterates from the root of a byte-serialized BytesTrie.
+         * @param trieBytes The trie bytes.
+         * @param maxStringLength If 0, the iterator returns full strings/byte sequences.
+         *                        Otherwise, the iterator returns strings with this maximum length.
+         * @param errorCode Standard ICU error code. Its input value must
+         *                  pass the U_SUCCESS() test, or else the function returns
+         *                  immediately. Check for U_FAILURE() on output or use with
+         *                  function chaining. (See User Guide for details.)
+         * @stable ICU 4.8
+         */
+        Iterator(const void *trieBytes, int32_t maxStringLength, UErrorCode &errorCode);
+
+        /**
+         * Iterates from the current state of the specified BytesTrie.
+         * @param trie The trie whose state will be copied for iteration.
+         * @param maxStringLength If 0, the iterator returns full strings/byte sequences.
+         *                        Otherwise, the iterator returns strings with this maximum length.
+         * @param errorCode Standard ICU error code. Its input value must
+         *                  pass the U_SUCCESS() test, or else the function returns
+         *                  immediately. Check for U_FAILURE() on output or use with
+         *                  function chaining. (See User Guide for details.)
+         * @stable ICU 4.8
+         */
+        Iterator(const BytesTrie &trie, int32_t maxStringLength, UErrorCode &errorCode);
+
+        /**
+         * Destructor.
+         * @stable ICU 4.8
+         */
+        ~Iterator();
+
+        /**
+         * Resets this iterator to its initial state.
+         * @return *this
+         * @stable ICU 4.8
+         */
+        Iterator &reset();
+
+        /**
+         * @return true if there are more elements.
+         * @stable ICU 4.8
+         */
+        UBool hasNext() const;
+
+        /**
+         * Finds the next (byte sequence, value) pair if there is one.
+         *
+         * If the byte sequence is truncated to the maximum length and does not
+         * have a real value, then the value is set to -1.
+         * In this case, this "not a real value" is indistinguishable from
+         * a real value of -1.
+         * @param errorCode Standard ICU error code. Its input value must
+         *                  pass the U_SUCCESS() test, or else the function returns
+         *                  immediately. Check for U_FAILURE() on output or use with
+         *                  function chaining. (See User Guide for details.)
+         * @return true if there is another element.
+         * @stable ICU 4.8
+         */
+        UBool next(UErrorCode &errorCode);
+
+        /**
+         * @return The NUL-terminated byte sequence for the last successful next().
+         * @stable ICU 4.8
+         */
+        StringPiece getString() const;
+        /**
+         * @return The value for the last successful next().
+         * @stable ICU 4.8
+         */
+        int32_t getValue() const { return value_; }
+
+    private:
+        UBool truncateAndStop();
+
+        const uint8_t *branchNext(const uint8_t *pos, int32_t length, UErrorCode &errorCode);
+
+        const uint8_t *bytes_;
+        const uint8_t *pos_;
+        const uint8_t *initialPos_;
+        int32_t remainingMatchLength_;
+        int32_t initialRemainingMatchLength_;
+
+        CharString *str_;
+        int32_t maxLength_;
+        int32_t value_;
+
+        // The stack stores pairs of integers for backtracking to another
+        // outbound edge of a branch node.
+        // The first integer is an offset from bytes_.
+        // The second integer has the str_->length() from before the node in bits 15..0,
+        // and the remaining branch length in bits 24..16. (Bits 31..25 are unused.)
+        // (We could store the remaining branch length minus 1 in bits 23..16 and not use bits 31..24,
+        // but the code looks more confusing that way.)
+        UVector32 *stack_;
+    };
+
+private:
+    friend class BytesTrieBuilder;
+    friend class ::BytesTrieTest;
+
+    /**
+     * Constructs a BytesTrie reader instance.
+     * Unlike the public constructor which just aliases an array,
+     * this constructor adopts the builder's array.
+     * This constructor is only called by the builder.
+     */
+    BytesTrie(void *adoptBytes, const void *trieBytes)
+            : ownedArray_(static_cast<uint8_t *>(adoptBytes)),
+              bytes_(static_cast<const uint8_t *>(trieBytes)),
+              pos_(bytes_), remainingMatchLength_(-1) {}
+
+    // No assignment operator.
+    BytesTrie &operator=(const BytesTrie &other) = delete;
+
+    inline void stop() {
+        pos_=nullptr;
+    }
+
+    // Reads a compact 32-bit integer.
+    // pos is already after the leadByte, and the lead byte is already shifted right by 1.
+    static int32_t readValue(const uint8_t *pos, int32_t leadByte);
+    static inline const uint8_t *skipValue(const uint8_t *pos, int32_t leadByte) {
+        // U_ASSERT(leadByte>=kMinValueLead);
+        if(leadByte>=(kMinTwoByteValueLead<<1)) {
+            if(leadByte<(kMinThreeByteValueLead<<1)) {
+                ++pos;
+            } else if(leadByte<(kFourByteValueLead<<1)) {
+                pos+=2;
+            } else {
+                pos+=3+((leadByte>>1)&1);
+            }
+        }
+        return pos;
+    }
+    static inline const uint8_t *skipValue(const uint8_t *pos) {
+        int32_t leadByte=*pos++;
+        return skipValue(pos, leadByte);
+    }
+
+    // Reads a jump delta and jumps.
+    static const uint8_t *jumpByDelta(const uint8_t *pos);
+
+    static inline const uint8_t *skipDelta(const uint8_t *pos) {
+        int32_t delta=*pos++;
+        if(delta>=kMinTwoByteDeltaLead) {
+            if(delta<kMinThreeByteDeltaLead) {
+                ++pos;
+            } else if(delta<kFourByteDeltaLead) {
+                pos+=2;
+            } else {
+                pos+=3+(delta&1);
+            }
+        }
+        return pos;
+    }
+
+    static inline UStringTrieResult valueResult(int32_t node) {
+        return static_cast<UStringTrieResult>(USTRINGTRIE_INTERMEDIATE_VALUE - (node & kValueIsFinal));
+    }
+
+    // Handles a branch node for both next(byte) and next(string).
+    UStringTrieResult branchNext(const uint8_t *pos, int32_t length, int32_t inByte);
+
+    // Requires remainingLength_<0.
+    UStringTrieResult nextImpl(const uint8_t *pos, int32_t inByte);
+
+    // Helper functions for hasUniqueValue().
+    // Recursively finds a unique value (or whether there is not a unique one)
+    // from a branch.
+    static const uint8_t *findUniqueValueFromBranch(const uint8_t *pos, int32_t length,
+                                                    UBool haveUniqueValue, int32_t &uniqueValue);
+    // Recursively finds a unique value (or whether there is not a unique one)
+    // starting from a position on a node lead byte.
+    static UBool findUniqueValue(const uint8_t *pos, UBool haveUniqueValue, int32_t &uniqueValue);
+
+    // Helper functions for getNextBytes().
+    // getNextBytes() when pos is on a branch node.
+    static void getNextBranchBytes(const uint8_t *pos, int32_t length, ByteSink &out);
+    static void append(ByteSink &out, int c);
+
+    // BytesTrie data structure
+    //
+    // The trie consists of a series of byte-serialized nodes for incremental
+    // string/byte sequence matching. The root node is at the beginning of the trie data.
+    //
+    // Types of nodes are distinguished by their node lead byte ranges.
+    // After each node, except a final-value node, another node follows to
+    // encode match values or continue matching further bytes.
+    //
+    // Node types:
+    //  - Value node: Stores a 32-bit integer in a compact, variable-length format.
+    //    The value is for the string/byte sequence so far.
+    //    One node bit indicates whether the value is final or whether
+    //    matching continues with the next node.
+    //  - Linear-match node: Matches a number of bytes.
+    //  - Branch node: Branches to other nodes according to the current input byte.
+    //    The node byte is the length of the branch (number of bytes to select from)
+    //    minus 1. It is followed by a sub-node:
+    //    - If the length is at most kMaxBranchLinearSubNodeLength, then
+    //      there are length-1 (key, value) pairs and then one more comparison byte.
+    //      If one of the key bytes matches, then the value is either a final value for
+    //      the string/byte sequence so far, or a "jump" delta to the next node.
+    //      If the last byte matches, then matching continues with the next node.
+    //      (Values have the same encoding as value nodes.)
+    //    - If the length is greater than kMaxBranchLinearSubNodeLength, then
+    //      there is one byte and one "jump" delta.
+    //      If the input byte is less than the sub-node byte, then "jump" by delta to
+    //      the next sub-node which will have a length of length/2.
+    //      (The delta has its own compact encoding.)
+    //      Otherwise, skip the "jump" delta to the next sub-node
+    //      which will have a length of length-length/2.
+
+    // Node lead byte values.
+
+    // 00..0f: Branch node. If node!=0 then the length is node+1, otherwise
+    // the length is one more than the next byte.
+
+    // For a branch sub-node with at most this many entries, we drop down
+    // to a linear search.
+    static const int32_t kMaxBranchLinearSubNodeLength=5;
+
+    // 10..1f: Linear-match node, match 1..16 bytes and continue reading the next node.
+    static const int32_t kMinLinearMatch=0x10;
+    static const int32_t kMaxLinearMatchLength=0x10;
+
+    // 20..ff: Variable-length value node.
+    // If odd, the value is final. (Otherwise, intermediate value or jump delta.)
+    // Then shift-right by 1 bit.
+    // The remaining lead byte value indicates the number of following bytes (0..4)
+    // and contains the value's top bits.
+    static const int32_t kMinValueLead=kMinLinearMatch+kMaxLinearMatchLength;  // 0x20
+    // It is a final value if bit 0 is set.
+    static const int32_t kValueIsFinal=1;
+
+    // Compact value: After testing bit 0, shift right by 1 and then use the following thresholds.
+    static const int32_t kMinOneByteValueLead=kMinValueLead/2;  // 0x10
+    static const int32_t kMaxOneByteValue=0x40;  // At least 6 bits in the first byte.
+
+    static const int32_t kMinTwoByteValueLead=kMinOneByteValueLead+kMaxOneByteValue+1;  // 0x51
+    static const int32_t kMaxTwoByteValue=0x1aff;
+
+    static const int32_t kMinThreeByteValueLead=kMinTwoByteValueLead+(kMaxTwoByteValue>>8)+1;  // 0x6c
+    static const int32_t kFourByteValueLead=0x7e;
+
+    // A little more than Unicode code points. (0x11ffff)
+    static const int32_t kMaxThreeByteValue=((kFourByteValueLead-kMinThreeByteValueLead)<<16)-1;
+
+    static const int32_t kFiveByteValueLead=0x7f;
+
+    // Compact delta integers.
+    static const int32_t kMaxOneByteDelta=0xbf;
+    static const int32_t kMinTwoByteDeltaLead=kMaxOneByteDelta+1;  // 0xc0
+    static const int32_t kMinThreeByteDeltaLead=0xf0;
+    static const int32_t kFourByteDeltaLead=0xfe;
+    static const int32_t kFiveByteDeltaLead=0xff;
+
+    static const int32_t kMaxTwoByteDelta=((kMinThreeByteDeltaLead-kMinTwoByteDeltaLead)<<8)-1;  // 0x2fff
+    static const int32_t kMaxThreeByteDelta=((kFourByteDeltaLead-kMinThreeByteDeltaLead)<<16)-1;  // 0xdffff
+
+    // For getState64():
+    // The remainingMatchLength_ is -1..14=(kMaxLinearMatchLength=0x10)-2
+    // so we need at least 5 bits for that.
+    // We add 2 to store it as a positive value 1..16=kMaxLinearMatchLength.
+    static constexpr int32_t kState64RemainingShift = 59;
+    static constexpr uint64_t kState64PosMask = (UINT64_C(1) << kState64RemainingShift) - 1;
+
+    uint8_t *ownedArray_;
+
+    // Fixed value referencing the BytesTrie bytes.
+    const uint8_t *bytes_;
+
+    // Iterator variables.
+
+    // Pointer to next trie byte to read. nullptr if no more matches.
+    const uint8_t *pos_;
+    // Remaining length of a linear-match node, minus 1. Negative if not in such a node.
+    int32_t remainingMatchLength_;
+};
+
+U_NAMESPACE_END
+
+#endif /* U_SHOW_CPLUSPLUS_API */
+
+#endif  // __BYTESTRIE_H__

thirdparty/icu4c/common/unicode/bytestriebuilder.h

@@ -0,0 +1,193 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+/*
+*******************************************************************************
+*   Copyright (C) 2010-2016, International Business Machines
+*   Corporation and others.  All Rights Reserved.
+*******************************************************************************
+*   file name:  bytestriebuilder.h
+*   encoding:   UTF-8
+*   tab size:   8 (not used)
+*   indentation:4
+*
+*   created on: 2010sep25
+*   created by: Markus W. Scherer
+*/
+
+/**
+ * \file
+ * \brief C++ API: Builder for icu::BytesTrie
+ */
+
+#ifndef __BYTESTRIEBUILDER_H__
+#define __BYTESTRIEBUILDER_H__
+
+#include "unicode/utypes.h"
+
+#if U_SHOW_CPLUSPLUS_API
+
+#include "unicode/bytestrie.h"
+#include "unicode/stringpiece.h"
+#include "unicode/stringtriebuilder.h"
+
+class BytesTrieTest;
+
+U_NAMESPACE_BEGIN
+
+class BytesTrieElement;
+class CharString;
+/**
+ * Builder class for BytesTrie.
+ *
+ * This class is not intended for public subclassing.
+ * @stable ICU 4.8
+ */
+class U_COMMON_API BytesTrieBuilder : public StringTrieBuilder {
+public:
+    /**
+     * Constructs an empty builder.
+     * @param errorCode Standard ICU error code.
+     * @stable ICU 4.8
+     */
+    BytesTrieBuilder(UErrorCode &errorCode);
+
+    /**
+     * Destructor.
+     * @stable ICU 4.8
+     */
+    virtual ~BytesTrieBuilder();
+
+    /**
+     * Adds a (byte sequence, value) pair.
+     * The byte sequence must be unique.
+     * The bytes will be copied; the builder does not keep
+     * a reference to the input StringPiece or its data().
+     * @param s The input byte sequence.
+     * @param value The value associated with this byte sequence.
+     * @param errorCode Standard ICU error code. Its input value must
+     *                  pass the U_SUCCESS() test, or else the function returns
+     *                  immediately. Check for U_FAILURE() on output or use with
+     *                  function chaining. (See User Guide for details.)
+     * @return *this
+     * @stable ICU 4.8
+     */
+    BytesTrieBuilder &add(StringPiece s, int32_t value, UErrorCode &errorCode);
+
+    /**
+     * Builds a BytesTrie for the add()ed data.
+     * Once built, no further data can be add()ed until clear() is called.
+     *
+     * A BytesTrie cannot be empty. At least one (byte sequence, value) pair
+     * must have been add()ed.
+     *
+     * This method passes ownership of the builder's internal result array to the new trie object.
+     * Another call to any build() variant will re-serialize the trie.
+     * After clear() has been called, a new array will be used as well.
+     * @param buildOption Build option, see UStringTrieBuildOption.
+     * @param errorCode Standard ICU error code. Its input value must
+     *                  pass the U_SUCCESS() test, or else the function returns
+     *                  immediately. Check for U_FAILURE() on output or use with
+     *                  function chaining. (See User Guide for details.)
+     * @return A new BytesTrie for the add()ed data.
+     * @stable ICU 4.8
+     */
+    BytesTrie *build(UStringTrieBuildOption buildOption, UErrorCode &errorCode);
+
+    /**
+     * Builds a BytesTrie for the add()ed data and byte-serializes it.
+     * Once built, no further data can be add()ed until clear() is called.
+     *
+     * A BytesTrie cannot be empty. At least one (byte sequence, value) pair
+     * must have been add()ed.
+     *
+     * Multiple calls to buildStringPiece() return StringPieces referring to the
+     * builder's same byte array, without rebuilding.
+     * If buildStringPiece() is called after build(), the trie will be
+     * re-serialized into a new array (because build() passes on ownership).
+     * If build() is called after buildStringPiece(), the trie object returned
+     * by build() will become the owner of the underlying string for the
+     * previously returned StringPiece.
+     * After clear() has been called, a new array will be used as well.
+     * @param buildOption Build option, see UStringTrieBuildOption.
+     * @param errorCode Standard ICU error code. Its input value must
+     *                  pass the U_SUCCESS() test, or else the function returns
+     *                  immediately. Check for U_FAILURE() on output or use with
+     *                  function chaining. (See User Guide for details.)
+     * @return A StringPiece which refers to the byte-serialized BytesTrie for the add()ed data.
+     * @stable ICU 4.8
+     */
+    StringPiece buildStringPiece(UStringTrieBuildOption buildOption, UErrorCode &errorCode);
+
+    /**
+     * Removes all (byte sequence, value) pairs.
+     * New data can then be add()ed and a new trie can be built.
+     * @return *this
+     * @stable ICU 4.8
+     */
+    BytesTrieBuilder &clear();
+
+private:
+    friend class ::BytesTrieTest;
+
+    BytesTrieBuilder(const BytesTrieBuilder &other) = delete;  // no copy constructor
+    BytesTrieBuilder &operator=(const BytesTrieBuilder &other) = delete;  // no assignment operator
+
+    void buildBytes(UStringTrieBuildOption buildOption, UErrorCode &errorCode);
+
+    virtual int32_t getElementStringLength(int32_t i) const override;
+    virtual char16_t getElementUnit(int32_t i, int32_t byteIndex) const override;
+    virtual int32_t getElementValue(int32_t i) const override;
+
+    virtual int32_t getLimitOfLinearMatch(int32_t first, int32_t last, int32_t byteIndex) const override;
+
+    virtual int32_t countElementUnits(int32_t start, int32_t limit, int32_t byteIndex) const override;
+    virtual int32_t skipElementsBySomeUnits(int32_t i, int32_t byteIndex, int32_t count) const override;
+    virtual int32_t indexOfElementWithNextUnit(int32_t i, int32_t byteIndex, char16_t byte) const override;
+
+    virtual UBool matchNodesCanHaveValues() const override { return false; }
+
+    virtual int32_t getMaxBranchLinearSubNodeLength() const override { return BytesTrie::kMaxBranchLinearSubNodeLength; }
+    virtual int32_t getMinLinearMatch() const override { return BytesTrie::kMinLinearMatch; }
+    virtual int32_t getMaxLinearMatchLength() const override { return BytesTrie::kMaxLinearMatchLength; }
+
+    /**
+     * @internal (private)
+     */
+    class BTLinearMatchNode : public LinearMatchNode {
+    public:
+        BTLinearMatchNode(const char *units, int32_t len, Node *nextNode);
+        virtual bool operator==(const Node &other) const override;
+        virtual void write(StringTrieBuilder &builder) override;
+    private:
+        const char *s;
+    };
+    
+    virtual Node *createLinearMatchNode(int32_t i, int32_t byteIndex, int32_t length,
+                                        Node *nextNode) const override;
+
+    UBool ensureCapacity(int32_t length);
+    virtual int32_t write(int32_t byte) override;
+    int32_t write(const char *b, int32_t length);
+    virtual int32_t writeElementUnits(int32_t i, int32_t byteIndex, int32_t length) override;
+    virtual int32_t writeValueAndFinal(int32_t i, UBool isFinal) override;
+    virtual int32_t writeValueAndType(UBool hasValue, int32_t value, int32_t node) override;
+    virtual int32_t writeDeltaTo(int32_t jumpTarget) override;
+    static int32_t internalEncodeDelta(int32_t i, char intBytes[]);
+
+    CharString *strings;  // Pointer not object so we need not #include internal charstr.h.
+    BytesTrieElement *elements;
+    int32_t elementsCapacity;
+    int32_t elementsLength;
+
+    // Byte serialization of the trie.
+    // Grows from the back: bytesLength measures from the end of the buffer!
+    char *bytes;
+    int32_t bytesCapacity;
+    int32_t bytesLength;
+};
+
+U_NAMESPACE_END
+
+#endif /* U_SHOW_CPLUSPLUS_API */
+
+#endif  // __BYTESTRIEBUILDER_H__

thirdparty/icu4c/common/unicode/caniter.h

@@ -0,0 +1,215 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+/*
+ *******************************************************************************
+ * Copyright (C) 1996-2014, International Business Machines Corporation and
+ * others. All Rights Reserved.
+ *******************************************************************************
+ */
+
+#ifndef CANITER_H
+#define CANITER_H
+
+#include "unicode/utypes.h"
+
+#if U_SHOW_CPLUSPLUS_API
+
+#if !UCONFIG_NO_NORMALIZATION
+
+#include "unicode/uobject.h"
+#include "unicode/unistr.h"
+
+/**
+ * \file
+ * \brief C++ API: Canonical Iterator
+ */
+ 
+/** Should permutation skip characters with combining class zero
+ *  Should be either true or false. This is a compile time option
+ *  @stable ICU 2.4
+ */
+#ifndef CANITER_SKIP_ZEROES
+#define CANITER_SKIP_ZEROES true
+#endif
+
+U_NAMESPACE_BEGIN
+
+class Hashtable;
+class Normalizer2;
+class Normalizer2Impl;
+
+/**
+ * This class allows one to iterate through all the strings that are canonically equivalent to a given
+ * string. For example, here are some sample results:
+Results for: {LATIN CAPITAL LETTER A WITH RING ABOVE}{LATIN SMALL LETTER D}{COMBINING DOT ABOVE}{COMBINING CEDILLA}
+1: \\u0041\\u030A\\u0064\\u0307\\u0327
+ = {LATIN CAPITAL LETTER A}{COMBINING RING ABOVE}{LATIN SMALL LETTER D}{COMBINING DOT ABOVE}{COMBINING CEDILLA}
+2: \\u0041\\u030A\\u0064\\u0327\\u0307
+ = {LATIN CAPITAL LETTER A}{COMBINING RING ABOVE}{LATIN SMALL LETTER D}{COMBINING CEDILLA}{COMBINING DOT ABOVE}
+3: \\u0041\\u030A\\u1E0B\\u0327
+ = {LATIN CAPITAL LETTER A}{COMBINING RING ABOVE}{LATIN SMALL LETTER D WITH DOT ABOVE}{COMBINING CEDILLA}
+4: \\u0041\\u030A\\u1E11\\u0307
+ = {LATIN CAPITAL LETTER A}{COMBINING RING ABOVE}{LATIN SMALL LETTER D WITH CEDILLA}{COMBINING DOT ABOVE}
+5: \\u00C5\\u0064\\u0307\\u0327
+ = {LATIN CAPITAL LETTER A WITH RING ABOVE}{LATIN SMALL LETTER D}{COMBINING DOT ABOVE}{COMBINING CEDILLA}
+6: \\u00C5\\u0064\\u0327\\u0307
+ = {LATIN CAPITAL LETTER A WITH RING ABOVE}{LATIN SMALL LETTER D}{COMBINING CEDILLA}{COMBINING DOT ABOVE}
+7: \\u00C5\\u1E0B\\u0327
+ = {LATIN CAPITAL LETTER A WITH RING ABOVE}{LATIN SMALL LETTER D WITH DOT ABOVE}{COMBINING CEDILLA}
+8: \\u00C5\\u1E11\\u0307
+ = {LATIN CAPITAL LETTER A WITH RING ABOVE}{LATIN SMALL LETTER D WITH CEDILLA}{COMBINING DOT ABOVE}
+9: \\u212B\\u0064\\u0307\\u0327
+ = {ANGSTROM SIGN}{LATIN SMALL LETTER D}{COMBINING DOT ABOVE}{COMBINING CEDILLA}
+10: \\u212B\\u0064\\u0327\\u0307
+ = {ANGSTROM SIGN}{LATIN SMALL LETTER D}{COMBINING CEDILLA}{COMBINING DOT ABOVE}
+11: \\u212B\\u1E0B\\u0327
+ = {ANGSTROM SIGN}{LATIN SMALL LETTER D WITH DOT ABOVE}{COMBINING CEDILLA}
+12: \\u212B\\u1E11\\u0307
+ = {ANGSTROM SIGN}{LATIN SMALL LETTER D WITH CEDILLA}{COMBINING DOT ABOVE}
+ *<br>Note: the code is intended for use with small strings, and is not suitable for larger ones,
+ * since it has not been optimized for that situation.
+ * Note, CanonicalIterator is not intended to be subclassed.
+ * @author M. Davis
+ * @author C++ port by V. Weinstein
+ * @stable ICU 2.4
+ */
+class U_COMMON_API CanonicalIterator final : public UObject {
+public:
+    /**
+     * Construct a CanonicalIterator object
+     * @param source    string to get results for
+     * @param status    Fill-in parameter which receives the status of this operation.
+     * @stable ICU 2.4
+     */
+    CanonicalIterator(const UnicodeString &source, UErrorCode &status);
+
+    /** Destructor
+     *  Cleans pieces
+     * @stable ICU 2.4
+     */
+    virtual ~CanonicalIterator();
+
+    /**
+     * Gets the NFD form of the current source we are iterating over.
+     * @return gets the source: NOTE: it is the NFD form of source
+     * @stable ICU 2.4
+     */
+    UnicodeString getSource();
+
+    /**
+     * Resets the iterator so that one can start again from the beginning.
+     * @stable ICU 2.4
+     */
+    void reset();
+
+    /**
+     * Get the next canonically equivalent string.
+     * <br><b>Warning: The strings are not guaranteed to be in any particular order.</b>
+     * @return the next string that is canonically equivalent. A bogus string is returned when
+     * the iteration is done.
+     * @stable ICU 2.4
+     */
+    UnicodeString next();
+
+    /**
+     * Set a new source for this iterator. Allows object reuse.
+     * @param newSource     the source string to iterate against. This allows the same iterator to be used
+     *                     while changing the source string, saving object creation.
+     * @param status        Fill-in parameter which receives the status of this operation.
+     * @stable ICU 2.4
+     */
+    void setSource(const UnicodeString &newSource, UErrorCode &status);
+
+#ifndef U_HIDE_INTERNAL_API
+    /**
+     * Dumb recursive implementation of permutation.
+     * TODO: optimize
+     * @param source     the string to find permutations for
+     * @param skipZeros  determine if skip zeros
+     * @param result     the results in a set.
+     * @param status       Fill-in parameter which receives the status of this operation.
+     * @param depth     depth of the call.
+     * @internal
+     */
+    static void U_EXPORT2 permute(UnicodeString &source, UBool skipZeros, Hashtable *result, UErrorCode &status, int32_t depth=0);
+#endif  /* U_HIDE_INTERNAL_API */
+
+    /**
+     * ICU "poor man's RTTI", returns a UClassID for this class.
+     *
+     * @stable ICU 2.2
+     */
+    static UClassID U_EXPORT2 getStaticClassID();
+
+    /**
+     * ICU "poor man's RTTI", returns a UClassID for the actual class.
+     *
+     * @stable ICU 2.2
+     */
+    virtual UClassID getDynamicClassID() const override;
+
+private:
+    // ===================== PRIVATES ==============================
+    // private default constructor
+    CanonicalIterator() = delete;
+
+
+    /**
+     * Copy constructor. Private for now.
+     * @internal (private)
+     */
+    CanonicalIterator(const CanonicalIterator& other) = delete;
+
+    /**
+     * Assignment operator. Private for now.
+     * @internal (private)
+     */
+    CanonicalIterator& operator=(const CanonicalIterator& other) = delete;
+
+    // fields
+    UnicodeString source;
+    UBool done;
+
+    // 2 dimensional array holds the pieces of the string with
+    // their different canonically equivalent representations
+    UnicodeString **pieces;
+    int32_t pieces_length;
+    int32_t *pieces_lengths;
+
+    // current is used in iterating to combine pieces
+    int32_t *current;
+    int32_t current_length;
+
+    // transient fields
+    UnicodeString buffer;
+
+    const Normalizer2 *nfd;
+    const Normalizer2Impl *nfcImpl;
+
+    // we have a segment, in NFD. Find all the strings that are canonically equivalent to it.
+    UnicodeString *getEquivalents(const UnicodeString &segment, int32_t &result_len, UErrorCode &status); //private String[] getEquivalents(String segment)
+
+    //Set getEquivalents2(String segment);
+    Hashtable *getEquivalents2(Hashtable *fillinResult, const char16_t *segment, int32_t segLen, UErrorCode &status);
+    //Hashtable *getEquivalents2(const UnicodeString &segment, int32_t segLen, UErrorCode &status);
+
+    /**
+     * See if the decomposition of cp2 is at segment starting at segmentPos
+     * (with canonical rearrangement!)
+     * If so, take the remainder, and return the equivalents
+     */
+    //Set extract(int comp, String segment, int segmentPos, StringBuffer buffer);
+    Hashtable *extract(Hashtable *fillinResult, UChar32 comp, const char16_t *segment, int32_t segLen, int32_t segmentPos, UErrorCode &status);
+    //Hashtable *extract(UChar32 comp, const UnicodeString &segment, int32_t segLen, int32_t segmentPos, UErrorCode &status);
+
+    void cleanPieces();
+
+};
+
+U_NAMESPACE_END
+
+#endif /* #if !UCONFIG_NO_NORMALIZATION */
+
+#endif /* U_SHOW_CPLUSPLUS_API */
+
+#endif

thirdparty/icu4c/common/unicode/casemap.h

@@ -0,0 +1,497 @@
+// © 2017 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+
+// casemap.h
+// created: 2017jan12 Markus W. Scherer
+
+#ifndef __CASEMAP_H__
+#define __CASEMAP_H__
+
+#include "unicode/utypes.h"
+
+#if U_SHOW_CPLUSPLUS_API
+
+#include "unicode/stringpiece.h"
+#include "unicode/uobject.h"
+
+/**
+ * \file
+ * \brief C++ API: Low-level C++ case mapping functions.
+ */
+
+U_NAMESPACE_BEGIN
+
+class BreakIterator;
+class ByteSink;
+class Edits;
+
+/**
+ * Low-level C++ case mapping functions.
+ *
+ * @stable ICU 59
+ */
+class U_COMMON_API CaseMap final : public UMemory {
+public:
+    /**
+     * Lowercases a UTF-16 string and optionally records edits.
+     * Casing is locale-dependent and context-sensitive.
+     * The result may be longer or shorter than the original.
+     * The source string and the destination buffer must not overlap.
+     *
+     * @param locale    The locale ID. ("" = root locale, nullptr = default locale.)
+     * @param options   Options bit set, usually 0. See U_OMIT_UNCHANGED_TEXT and U_EDITS_NO_RESET.
+     * @param src       The original string.
+     * @param srcLength The length of the original string. If -1, then src must be NUL-terminated.
+     * @param dest      A buffer for the result string. The result will be NUL-terminated if
+     *                  the buffer is large enough.
+     *                  The contents is undefined in case of failure.
+     * @param destCapacity The size of the buffer (number of char16_ts). If it is 0, then
+     *                  dest may be nullptr and the function will only return the length of the result
+     *                  without writing any of the result string.
+     * @param edits     Records edits for index mapping, working with styled text,
+     *                  and getting only changes (if any).
+     *                  The Edits contents is undefined if any error occurs.
+     *                  This function calls edits->reset() first unless
+     *                  options includes U_EDITS_NO_RESET. edits can be nullptr.
+     * @param errorCode Reference to an in/out error code value
+     *                  which must not indicate a failure before the function call.
+     * @return The length of the result string, if successful.
+     *         When the result would be longer than destCapacity,
+     *         the full length is returned and a U_BUFFER_OVERFLOW_ERROR is set.
+     *
+     * @see u_strToLower
+     * @stable ICU 59
+     */
+     static int32_t toLower(
+            const char *locale, uint32_t options,
+            const char16_t *src, int32_t srcLength,
+            char16_t *dest, int32_t destCapacity, Edits *edits,
+            UErrorCode &errorCode);
+
+    /**
+     * Uppercases a UTF-16 string and optionally records edits.
+     * Casing is locale-dependent and context-sensitive.
+     * The result may be longer or shorter than the original.
+     * The source string and the destination buffer must not overlap.
+     *
+     * @param locale    The locale ID. ("" = root locale, nullptr = default locale.)
+     * @param options   Options bit set, usually 0. See U_OMIT_UNCHANGED_TEXT and U_EDITS_NO_RESET.
+     * @param src       The original string.
+     * @param srcLength The length of the original string. If -1, then src must be NUL-terminated.
+     * @param dest      A buffer for the result string. The result will be NUL-terminated if
+     *                  the buffer is large enough.
+     *                  The contents is undefined in case of failure.
+     * @param destCapacity The size of the buffer (number of char16_ts). If it is 0, then
+     *                  dest may be nullptr and the function will only return the length of the result
+     *                  without writing any of the result string.
+     * @param edits     Records edits for index mapping, working with styled text,
+     *                  and getting only changes (if any).
+     *                  The Edits contents is undefined if any error occurs.
+     *                  This function calls edits->reset() first unless
+     *                  options includes U_EDITS_NO_RESET. edits can be nullptr.
+     * @param errorCode Reference to an in/out error code value
+     *                  which must not indicate a failure before the function call.
+     * @return The length of the result string, if successful.
+     *         When the result would be longer than destCapacity,
+     *         the full length is returned and a U_BUFFER_OVERFLOW_ERROR is set.
+     *
+     * @see u_strToUpper
+     * @stable ICU 59
+     */
+    static int32_t toUpper(
+            const char *locale, uint32_t options,
+            const char16_t *src, int32_t srcLength,
+            char16_t *dest, int32_t destCapacity, Edits *edits,
+            UErrorCode &errorCode);
+
+#if !UCONFIG_NO_BREAK_ITERATION
+
+    /**
+     * Titlecases a UTF-16 string and optionally records edits.
+     * Casing is locale-dependent and context-sensitive.
+     * The result may be longer or shorter than the original.
+     * The source string and the destination buffer must not overlap.
+     *
+     * Titlecasing uses a break iterator to find the first characters of words
+     * that are to be titlecased. It titlecases those characters and lowercases
+     * all others. (This can be modified with options bits.)
+     *
+     * @param locale    The locale ID. ("" = root locale, nullptr = default locale.)
+     * @param options   Options bit set, usually 0. See U_OMIT_UNCHANGED_TEXT, U_EDITS_NO_RESET,
+     *                  U_TITLECASE_NO_LOWERCASE,
+     *                  U_TITLECASE_NO_BREAK_ADJUSTMENT, U_TITLECASE_ADJUST_TO_CASED,
+     *                  U_TITLECASE_WHOLE_STRING, U_TITLECASE_SENTENCES.
+     * @param iter      A break iterator to find the first characters of words that are to be titlecased.
+     *                  It is set to the source string (setText())
+     *                  and used one or more times for iteration (first() and next()).
+     *                  If nullptr, then a word break iterator for the locale is used
+     *                  (or something equivalent).
+     * @param src       The original string.
+     * @param srcLength The length of the original string. If -1, then src must be NUL-terminated.
+     * @param dest      A buffer for the result string. The result will be NUL-terminated if
+     *                  the buffer is large enough.
+     *                  The contents is undefined in case of failure.
+     * @param destCapacity The size of the buffer (number of char16_ts). If it is 0, then
+     *                  dest may be nullptr and the function will only return the length of the result
+     *                  without writing any of the result string.
+     * @param edits     Records edits for index mapping, working with styled text,
+     *                  and getting only changes (if any).
+     *                  The Edits contents is undefined if any error occurs.
+     *                  This function calls edits->reset() first unless
+     *                  options includes U_EDITS_NO_RESET. edits can be nullptr.
+     * @param errorCode Reference to an in/out error code value
+     *                  which must not indicate a failure before the function call.
+     * @return The length of the result string, if successful.
+     *         When the result would be longer than destCapacity,
+     *         the full length is returned and a U_BUFFER_OVERFLOW_ERROR is set.
+     *
+     * @see u_strToTitle
+     * @see ucasemap_toTitle
+     * @stable ICU 59
+     */
+    static int32_t toTitle(
+            const char *locale, uint32_t options, BreakIterator *iter,
+            const char16_t *src, int32_t srcLength,
+            char16_t *dest, int32_t destCapacity, Edits *edits,
+            UErrorCode &errorCode);
+
+#endif  // UCONFIG_NO_BREAK_ITERATION
+
+    /**
+     * Case-folds a UTF-16 string and optionally records edits.
+     *
+     * Case folding is locale-independent and not context-sensitive,
+     * but there is an option for whether to include or exclude mappings for dotted I
+     * and dotless i that are marked with 'T' in CaseFolding.txt.
+     *
+     * The result may be longer or shorter than the original.
+     * The source string and the destination buffer must not overlap.
+     *
+     * @param options   Options bit set, usually 0. See U_OMIT_UNCHANGED_TEXT, U_EDITS_NO_RESET,
+     *                  U_FOLD_CASE_DEFAULT, U_FOLD_CASE_EXCLUDE_SPECIAL_I.
+     * @param src       The original string.
+     * @param srcLength The length of the original string. If -1, then src must be NUL-terminated.
+     * @param dest      A buffer for the result string. The result will be NUL-terminated if
+     *                  the buffer is large enough.
+     *                  The contents is undefined in case of failure.
+     * @param destCapacity The size of the buffer (number of char16_ts). If it is 0, then
+     *                  dest may be nullptr and the function will only return the length of the result
+     *                  without writing any of the result string.
+     * @param edits     Records edits for index mapping, working with styled text,
+     *                  and getting only changes (if any).
+     *                  The Edits contents is undefined if any error occurs.
+     *                  This function calls edits->reset() first unless
+     *                  options includes U_EDITS_NO_RESET. edits can be nullptr.
+     * @param errorCode Reference to an in/out error code value
+     *                  which must not indicate a failure before the function call.
+     * @return The length of the result string, if successful.
+     *         When the result would be longer than destCapacity,
+     *         the full length is returned and a U_BUFFER_OVERFLOW_ERROR is set.
+     *
+     * @see u_strFoldCase
+     * @stable ICU 59
+     */
+    static int32_t fold(
+            uint32_t options,
+            const char16_t *src, int32_t srcLength,
+            char16_t *dest, int32_t destCapacity, Edits *edits,
+            UErrorCode &errorCode);
+
+    /**
+     * Lowercases a UTF-8 string and optionally records edits.
+     * Casing is locale-dependent and context-sensitive.
+     * The result may be longer or shorter than the original.
+     *
+     * @param locale    The locale ID. ("" = root locale, nullptr = default locale.)
+     * @param options   Options bit set, usually 0. See U_OMIT_UNCHANGED_TEXT and U_EDITS_NO_RESET.
+     * @param src       The original string.
+     * @param sink      A ByteSink to which the result string is written.
+     *                  sink.Flush() is called at the end.
+     * @param edits     Records edits for index mapping, working with styled text,
+     *                  and getting only changes (if any).
+     *                  The Edits contents is undefined if any error occurs.
+     *                  This function calls edits->reset() first unless
+     *                  options includes U_EDITS_NO_RESET. edits can be nullptr.
+     * @param errorCode Reference to an in/out error code value
+     *                  which must not indicate a failure before the function call.
+     *
+     * @see ucasemap_utf8ToLower
+     * @stable ICU 60
+     */
+    static void utf8ToLower(
+            const char *locale, uint32_t options,
+            StringPiece src, ByteSink &sink, Edits *edits,
+            UErrorCode &errorCode);
+
+    /**
+     * Uppercases a UTF-8 string and optionally records edits.
+     * Casing is locale-dependent and context-sensitive.
+     * The result may be longer or shorter than the original.
+     *
+     * @param locale    The locale ID. ("" = root locale, nullptr = default locale.)
+     * @param options   Options bit set, usually 0. See U_OMIT_UNCHANGED_TEXT and U_EDITS_NO_RESET.
+     * @param src       The original string.
+     * @param sink      A ByteSink to which the result string is written.
+     *                  sink.Flush() is called at the end.
+     * @param edits     Records edits for index mapping, working with styled text,
+     *                  and getting only changes (if any).
+     *                  The Edits contents is undefined if any error occurs.
+     *                  This function calls edits->reset() first unless
+     *                  options includes U_EDITS_NO_RESET. edits can be nullptr.
+     * @param errorCode Reference to an in/out error code value
+     *                  which must not indicate a failure before the function call.
+     *
+     * @see ucasemap_utf8ToUpper
+     * @stable ICU 60
+     */
+    static void utf8ToUpper(
+            const char *locale, uint32_t options,
+            StringPiece src, ByteSink &sink, Edits *edits,
+            UErrorCode &errorCode);
+
+#if !UCONFIG_NO_BREAK_ITERATION
+
+    /**
+     * Titlecases a UTF-8 string and optionally records edits.
+     * Casing is locale-dependent and context-sensitive.
+     * The result may be longer or shorter than the original.
+     *
+     * Titlecasing uses a break iterator to find the first characters of words
+     * that are to be titlecased. It titlecases those characters and lowercases
+     * all others. (This can be modified with options bits.)
+     *
+     * @param locale    The locale ID. ("" = root locale, nullptr = default locale.)
+     * @param options   Options bit set, usually 0. See U_OMIT_UNCHANGED_TEXT, U_EDITS_NO_RESET,
+     *                  U_TITLECASE_NO_LOWERCASE,
+     *                  U_TITLECASE_NO_BREAK_ADJUSTMENT, U_TITLECASE_ADJUST_TO_CASED,
+     *                  U_TITLECASE_WHOLE_STRING, U_TITLECASE_SENTENCES.
+     * @param iter      A break iterator to find the first characters of words that are to be titlecased.
+     *                  It is set to the source string (setUText())
+     *                  and used one or more times for iteration (first() and next()).
+     *                  If nullptr, then a word break iterator for the locale is used
+     *                  (or something equivalent).
+     * @param src       The original string.
+     * @param sink      A ByteSink to which the result string is written.
+     *                  sink.Flush() is called at the end.
+     * @param edits     Records edits for index mapping, working with styled text,
+     *                  and getting only changes (if any).
+     *                  The Edits contents is undefined if any error occurs.
+     *                  This function calls edits->reset() first unless
+     *                  options includes U_EDITS_NO_RESET. edits can be nullptr.
+     * @param errorCode Reference to an in/out error code value
+     *                  which must not indicate a failure before the function call.
+     *
+     * @see ucasemap_utf8ToTitle
+     * @stable ICU 60
+     */
+    static void utf8ToTitle(
+            const char *locale, uint32_t options, BreakIterator *iter,
+            StringPiece src, ByteSink &sink, Edits *edits,
+            UErrorCode &errorCode);
+
+#endif  // UCONFIG_NO_BREAK_ITERATION
+
+    /**
+     * Case-folds a UTF-8 string and optionally records edits.
+     *
+     * Case folding is locale-independent and not context-sensitive,
+     * but there is an option for whether to include or exclude mappings for dotted I
+     * and dotless i that are marked with 'T' in CaseFolding.txt.
+     *
+     * The result may be longer or shorter than the original.
+     *
+     * @param options   Options bit set, usually 0. See U_OMIT_UNCHANGED_TEXT and U_EDITS_NO_RESET.
+     * @param src       The original string.
+     * @param sink      A ByteSink to which the result string is written.
+     *                  sink.Flush() is called at the end.
+     * @param edits     Records edits for index mapping, working with styled text,
+     *                  and getting only changes (if any).
+     *                  The Edits contents is undefined if any error occurs.
+     *                  This function calls edits->reset() first unless
+     *                  options includes U_EDITS_NO_RESET. edits can be nullptr.
+     * @param errorCode Reference to an in/out error code value
+     *                  which must not indicate a failure before the function call.
+     *
+     * @see ucasemap_utf8FoldCase
+     * @stable ICU 60
+     */
+    static void utf8Fold(
+            uint32_t options,
+            StringPiece src, ByteSink &sink, Edits *edits,
+            UErrorCode &errorCode);
+
+    /**
+     * Lowercases a UTF-8 string and optionally records edits.
+     * Casing is locale-dependent and context-sensitive.
+     * The result may be longer or shorter than the original.
+     * The source string and the destination buffer must not overlap.
+     *
+     * @param locale    The locale ID. ("" = root locale, nullptr = default locale.)
+     * @param options   Options bit set, usually 0. See U_OMIT_UNCHANGED_TEXT and U_EDITS_NO_RESET.
+     * @param src       The original string.
+     * @param srcLength The length of the original string. If -1, then src must be NUL-terminated.
+     * @param dest      A buffer for the result string. The result will be NUL-terminated if
+     *                  the buffer is large enough.
+     *                  The contents is undefined in case of failure.
+     * @param destCapacity The size of the buffer (number of bytes). If it is 0, then
+     *                  dest may be nullptr and the function will only return the length of the result
+     *                  without writing any of the result string.
+     * @param edits     Records edits for index mapping, working with styled text,
+     *                  and getting only changes (if any).
+     *                  The Edits contents is undefined if any error occurs.
+     *                  This function calls edits->reset() first unless
+     *                  options includes U_EDITS_NO_RESET. edits can be nullptr.
+     * @param errorCode Reference to an in/out error code value
+     *                  which must not indicate a failure before the function call.
+     * @return The length of the result string, if successful.
+     *         When the result would be longer than destCapacity,
+     *         the full length is returned and a U_BUFFER_OVERFLOW_ERROR is set.
+     *
+     * @see ucasemap_utf8ToLower
+     * @stable ICU 59
+     */
+    static int32_t utf8ToLower(
+            const char *locale, uint32_t options,
+            const char *src, int32_t srcLength,
+            char *dest, int32_t destCapacity, Edits *edits,
+            UErrorCode &errorCode);
+
+    /**
+     * Uppercases a UTF-8 string and optionally records edits.
+     * Casing is locale-dependent and context-sensitive.
+     * The result may be longer or shorter than the original.
+     * The source string and the destination buffer must not overlap.
+     *
+     * @param locale    The locale ID. ("" = root locale, nullptr = default locale.)
+     * @param options   Options bit set, usually 0. See U_OMIT_UNCHANGED_TEXT and U_EDITS_NO_RESET.
+     * @param src       The original string.
+     * @param srcLength The length of the original string. If -1, then src must be NUL-terminated.
+     * @param dest      A buffer for the result string. The result will be NUL-terminated if
+     *                  the buffer is large enough.
+     *                  The contents is undefined in case of failure.
+     * @param destCapacity The size of the buffer (number of bytes). If it is 0, then
+     *                  dest may be nullptr and the function will only return the length of the result
+     *                  without writing any of the result string.
+     * @param edits     Records edits for index mapping, working with styled text,
+     *                  and getting only changes (if any).
+     *                  The Edits contents is undefined if any error occurs.
+     *                  This function calls edits->reset() first unless
+     *                  options includes U_EDITS_NO_RESET. edits can be nullptr.
+     * @param errorCode Reference to an in/out error code value
+     *                  which must not indicate a failure before the function call.
+     * @return The length of the result string, if successful.
+     *         When the result would be longer than destCapacity,
+     *         the full length is returned and a U_BUFFER_OVERFLOW_ERROR is set.
+     *
+     * @see ucasemap_utf8ToUpper
+     * @stable ICU 59
+     */
+    static int32_t utf8ToUpper(
+            const char *locale, uint32_t options,
+            const char *src, int32_t srcLength,
+            char *dest, int32_t destCapacity, Edits *edits,
+            UErrorCode &errorCode);
+
+#if !UCONFIG_NO_BREAK_ITERATION
+
+    /**
+     * Titlecases a UTF-8 string and optionally records edits.
+     * Casing is locale-dependent and context-sensitive.
+     * The result may be longer or shorter than the original.
+     * The source string and the destination buffer must not overlap.
+     *
+     * Titlecasing uses a break iterator to find the first characters of words
+     * that are to be titlecased. It titlecases those characters and lowercases
+     * all others. (This can be modified with options bits.)
+     *
+     * @param locale    The locale ID. ("" = root locale, nullptr = default locale.)
+     * @param options   Options bit set, usually 0. See U_OMIT_UNCHANGED_TEXT, U_EDITS_NO_RESET,
+     *                  U_TITLECASE_NO_LOWERCASE,
+     *                  U_TITLECASE_NO_BREAK_ADJUSTMENT, U_TITLECASE_ADJUST_TO_CASED,
+     *                  U_TITLECASE_WHOLE_STRING, U_TITLECASE_SENTENCES.
+     * @param iter      A break iterator to find the first characters of words that are to be titlecased.
+     *                  It is set to the source string (setUText())
+     *                  and used one or more times for iteration (first() and next()).
+     *                  If nullptr, then a word break iterator for the locale is used
+     *                  (or something equivalent).
+     * @param src       The original string.
+     * @param srcLength The length of the original string. If -1, then src must be NUL-terminated.
+     * @param dest      A buffer for the result string. The result will be NUL-terminated if
+     *                  the buffer is large enough.
+     *                  The contents is undefined in case of failure.
+     * @param destCapacity The size of the buffer (number of bytes). If it is 0, then
+     *                  dest may be nullptr and the function will only return the length of the result
+     *                  without writing any of the result string.
+     * @param edits     Records edits for index mapping, working with styled text,
+     *                  and getting only changes (if any).
+     *                  The Edits contents is undefined if any error occurs.
+     *                  This function calls edits->reset() first unless
+     *                  options includes U_EDITS_NO_RESET. edits can be nullptr.
+     * @param errorCode Reference to an in/out error code value
+     *                  which must not indicate a failure before the function call.
+     * @return The length of the result string, if successful.
+     *         When the result would be longer than destCapacity,
+     *         the full length is returned and a U_BUFFER_OVERFLOW_ERROR is set.
+     *
+     * @see ucasemap_utf8ToTitle
+     * @stable ICU 59
+     */
+    static int32_t utf8ToTitle(
+            const char *locale, uint32_t options, BreakIterator *iter,
+            const char *src, int32_t srcLength,
+            char *dest, int32_t destCapacity, Edits *edits,
+            UErrorCode &errorCode);
+
+#endif  // UCONFIG_NO_BREAK_ITERATION
+
+    /**
+     * Case-folds a UTF-8 string and optionally records edits.
+     *
+     * Case folding is locale-independent and not context-sensitive,
+     * but there is an option for whether to include or exclude mappings for dotted I
+     * and dotless i that are marked with 'T' in CaseFolding.txt.
+     *
+     * The result may be longer or shorter than the original.
+     * The source string and the destination buffer must not overlap.
+     *
+     * @param options   Options bit set, usually 0. See U_OMIT_UNCHANGED_TEXT, U_EDITS_NO_RESET,
+     *                  U_FOLD_CASE_DEFAULT, U_FOLD_CASE_EXCLUDE_SPECIAL_I.
+     * @param src       The original string.
+     * @param srcLength The length of the original string. If -1, then src must be NUL-terminated.
+     * @param dest      A buffer for the result string. The result will be NUL-terminated if
+     *                  the buffer is large enough.
+     *                  The contents is undefined in case of failure.
+     * @param destCapacity The size of the buffer (number of bytes). If it is 0, then
+     *                  dest may be nullptr and the function will only return the length of the result
+     *                  without writing any of the result string.
+     * @param edits     Records edits for index mapping, working with styled text,
+     *                  and getting only changes (if any).
+     *                  The Edits contents is undefined if any error occurs.
+     *                  This function calls edits->reset() first unless
+     *                  options includes U_EDITS_NO_RESET. edits can be nullptr.
+     * @param errorCode Reference to an in/out error code value
+     *                  which must not indicate a failure before the function call.
+     * @return The length of the result string, if successful.
+     *         When the result would be longer than destCapacity,
+     *         the full length is returned and a U_BUFFER_OVERFLOW_ERROR is set.
+     *
+     * @see ucasemap_utf8FoldCase
+     * @stable ICU 59
+     */
+    static int32_t utf8Fold(
+            uint32_t options,
+            const char *src, int32_t srcLength,
+            char *dest, int32_t destCapacity, Edits *edits,
+            UErrorCode &errorCode);
+
+private:
+    CaseMap() = delete;
+    CaseMap(const CaseMap &other) = delete;
+    CaseMap &operator=(const CaseMap &other) = delete;
+};
+
+U_NAMESPACE_END
+
+#endif /* U_SHOW_CPLUSPLUS_API */
+
+#endif  // __CASEMAP_H__

thirdparty/icu4c/common/unicode/char16ptr.h

@@ -0,0 +1,453 @@
+// © 2017 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+
+// char16ptr.h
+// created: 2017feb28 Markus W. Scherer
+
+#ifndef __CHAR16PTR_H__
+#define __CHAR16PTR_H__
+
+#include "unicode/utypes.h"
+
+#if U_SHOW_CPLUSPLUS_API || U_SHOW_CPLUSPLUS_HEADER_API
+
+#include <cstddef>
+#include <string_view>
+#include <type_traits>
+
+#endif
+
+/**
+ * \file
+ * \brief C++ API: char16_t pointer wrappers with
+ *        implicit conversion from bit-compatible raw pointer types.
+ *        Also conversion functions from char16_t * to UChar * and OldUChar *.
+ */
+
+/**
+ * \def U_ALIASING_BARRIER
+ * Barrier for pointer anti-aliasing optimizations even across function boundaries.
+ * @internal
+ */
+#ifdef U_ALIASING_BARRIER
+    // Use the predefined value.
+#elif (defined(__clang__) || defined(__GNUC__)) && U_PLATFORM != U_PF_BROWSER_NATIVE_CLIENT
+#   define U_ALIASING_BARRIER(ptr) asm volatile("" : : "rm"(ptr) : "memory")
+#elif defined(U_IN_DOXYGEN)
+#   define U_ALIASING_BARRIER(ptr)
+#endif
+
+// ICU DLL-exported
+#if U_SHOW_CPLUSPLUS_API
+
+U_NAMESPACE_BEGIN
+
+/**
+ * char16_t * wrapper with implicit conversion from distinct but bit-compatible pointer types.
+ * @stable ICU 59
+ */
+class U_COMMON_API Char16Ptr final {
+public:
+    /**
+     * Copies the pointer.
+     * @param p pointer
+     * @stable ICU 59
+     */
+    inline Char16Ptr(char16_t *p);
+#if !U_CHAR16_IS_TYPEDEF
+    /**
+     * Converts the pointer to char16_t *.
+     * @param p pointer to be converted
+     * @stable ICU 59
+     */
+    inline Char16Ptr(uint16_t *p);
+#endif
+#if U_SIZEOF_WCHAR_T==2 || defined(U_IN_DOXYGEN)
+    /**
+     * Converts the pointer to char16_t *.
+     * (Only defined if U_SIZEOF_WCHAR_T==2.)
+     * @param p pointer to be converted
+     * @stable ICU 59
+     */
+    inline Char16Ptr(wchar_t *p);
+#endif
+    /**
+     * nullptr constructor.
+     * @param p nullptr
+     * @stable ICU 59
+     */
+    inline Char16Ptr(std::nullptr_t p);
+    /**
+     * Destructor.
+     * @stable ICU 59
+     */
+    inline ~Char16Ptr();
+
+    /**
+     * Pointer access.
+     * @return the wrapped pointer
+     * @stable ICU 59
+     */
+    inline char16_t *get() const;
+    /**
+     * char16_t pointer access via type conversion (e.g., static_cast).
+     * @return the wrapped pointer
+     * @stable ICU 59
+     */
+    inline operator char16_t *() const { return get(); }
+
+private:
+    Char16Ptr() = delete;
+
+#ifdef U_ALIASING_BARRIER
+    template<typename T> static char16_t *cast(T *t) {
+        U_ALIASING_BARRIER(t);
+        return reinterpret_cast<char16_t *>(t);
+    }
+
+    char16_t *p_;
+#else
+    union {
+        char16_t *cp;
+        uint16_t *up;
+        wchar_t *wp;
+    } u_;
+#endif
+};
+
+/// \cond
+#ifdef U_ALIASING_BARRIER
+
+Char16Ptr::Char16Ptr(char16_t *p) : p_(p) {}
+#if !U_CHAR16_IS_TYPEDEF
+Char16Ptr::Char16Ptr(uint16_t *p) : p_(cast(p)) {}
+#endif
+#if U_SIZEOF_WCHAR_T==2
+Char16Ptr::Char16Ptr(wchar_t *p) : p_(cast(p)) {}
+#endif
+Char16Ptr::Char16Ptr(std::nullptr_t p) : p_(p) {}
+Char16Ptr::~Char16Ptr() {
+    U_ALIASING_BARRIER(p_);
+}
+
+char16_t *Char16Ptr::get() const { return p_; }
+
+#else
+
+Char16Ptr::Char16Ptr(char16_t *p) { u_.cp = p; }
+#if !U_CHAR16_IS_TYPEDEF
+Char16Ptr::Char16Ptr(uint16_t *p) { u_.up = p; }
+#endif
+#if U_SIZEOF_WCHAR_T==2
+Char16Ptr::Char16Ptr(wchar_t *p) { u_.wp = p; }
+#endif
+Char16Ptr::Char16Ptr(std::nullptr_t p) { u_.cp = p; }
+Char16Ptr::~Char16Ptr() {}
+
+char16_t *Char16Ptr::get() const { return u_.cp; }
+
+#endif
+/// \endcond
+
+/**
+ * const char16_t * wrapper with implicit conversion from distinct but bit-compatible pointer types.
+ * @stable ICU 59
+ */
+class U_COMMON_API ConstChar16Ptr final {
+public:
+    /**
+     * Copies the pointer.
+     * @param p pointer
+     * @stable ICU 59
+     */
+    inline ConstChar16Ptr(const char16_t *p);
+#if !U_CHAR16_IS_TYPEDEF
+    /**
+     * Converts the pointer to char16_t *.
+     * @param p pointer to be converted
+     * @stable ICU 59
+     */
+    inline ConstChar16Ptr(const uint16_t *p);
+#endif
+#if U_SIZEOF_WCHAR_T==2 || defined(U_IN_DOXYGEN)
+    /**
+     * Converts the pointer to char16_t *.
+     * (Only defined if U_SIZEOF_WCHAR_T==2.)
+     * @param p pointer to be converted
+     * @stable ICU 59
+     */
+    inline ConstChar16Ptr(const wchar_t *p);
+#endif
+    /**
+     * nullptr constructor.
+     * @param p nullptr
+     * @stable ICU 59
+     */
+    inline ConstChar16Ptr(const std::nullptr_t p);
+
+    /**
+     * Destructor.
+     * @stable ICU 59
+     */
+    inline ~ConstChar16Ptr();
+
+    /**
+     * Pointer access.
+     * @return the wrapped pointer
+     * @stable ICU 59
+     */
+    inline const char16_t *get() const;
+    /**
+     * char16_t pointer access via type conversion (e.g., static_cast).
+     * @return the wrapped pointer
+     * @stable ICU 59
+     */
+    inline operator const char16_t *() const { return get(); }
+
+private:
+    ConstChar16Ptr() = delete;
+
+#ifdef U_ALIASING_BARRIER
+    template<typename T> static const char16_t *cast(const T *t) {
+        U_ALIASING_BARRIER(t);
+        return reinterpret_cast<const char16_t *>(t);
+    }
+
+    const char16_t *p_;
+#else
+    union {
+        const char16_t *cp;
+        const uint16_t *up;
+        const wchar_t *wp;
+    } u_;
+#endif
+};
+
+/// \cond
+#ifdef U_ALIASING_BARRIER
+
+ConstChar16Ptr::ConstChar16Ptr(const char16_t *p) : p_(p) {}
+#if !U_CHAR16_IS_TYPEDEF
+ConstChar16Ptr::ConstChar16Ptr(const uint16_t *p) : p_(cast(p)) {}
+#endif
+#if U_SIZEOF_WCHAR_T==2
+ConstChar16Ptr::ConstChar16Ptr(const wchar_t *p) : p_(cast(p)) {}
+#endif
+ConstChar16Ptr::ConstChar16Ptr(const std::nullptr_t p) : p_(p) {}
+ConstChar16Ptr::~ConstChar16Ptr() {
+    U_ALIASING_BARRIER(p_);
+}
+
+const char16_t *ConstChar16Ptr::get() const { return p_; }
+
+#else
+
+ConstChar16Ptr::ConstChar16Ptr(const char16_t *p) { u_.cp = p; }
+#if !U_CHAR16_IS_TYPEDEF
+ConstChar16Ptr::ConstChar16Ptr(const uint16_t *p) { u_.up = p; }
+#endif
+#if U_SIZEOF_WCHAR_T==2
+ConstChar16Ptr::ConstChar16Ptr(const wchar_t *p) { u_.wp = p; }
+#endif
+ConstChar16Ptr::ConstChar16Ptr(const std::nullptr_t p) { u_.cp = p; }
+ConstChar16Ptr::~ConstChar16Ptr() {}
+
+const char16_t *ConstChar16Ptr::get() const { return u_.cp; }
+
+#endif
+/// \endcond
+
+U_NAMESPACE_END
+
+#endif  // U_SHOW_CPLUSPLUS_API
+
+// Usable in header-only definitions
+#if U_SHOW_CPLUSPLUS_API || U_SHOW_CPLUSPLUS_HEADER_API
+
+namespace U_ICU_NAMESPACE_OR_INTERNAL {
+
+#ifndef U_FORCE_HIDE_INTERNAL_API
+/** @internal */
+template<typename T, typename = std::enable_if_t<std::is_same_v<T, UChar>>>
+inline const char16_t *uprv_char16PtrFromUChar(const T *p) {
+    if constexpr (std::is_same_v<UChar, char16_t>) {
+        return p;
+    } else {
+#if U_SHOW_CPLUSPLUS_API
+        return ConstChar16Ptr(p).get();
+#else
+#ifdef U_ALIASING_BARRIER
+        U_ALIASING_BARRIER(p);
+#endif
+        return reinterpret_cast<const char16_t *>(p);
+#endif
+    }
+}
+#if !U_CHAR16_IS_TYPEDEF && (!defined(_LIBCPP_VERSION) || _LIBCPP_VERSION < 180000)
+/** @internal */
+inline const char16_t *uprv_char16PtrFromUint16(const uint16_t *p) {
+#if U_SHOW_CPLUSPLUS_API
+    return ConstChar16Ptr(p).get();
+#else
+#ifdef U_ALIASING_BARRIER
+    U_ALIASING_BARRIER(p);
+#endif
+    return reinterpret_cast<const char16_t *>(p);
+#endif
+}
+#endif
+#if U_SIZEOF_WCHAR_T==2
+/** @internal */
+inline const char16_t *uprv_char16PtrFromWchar(const wchar_t *p) {
+#if U_SHOW_CPLUSPLUS_API
+    return ConstChar16Ptr(p).get();
+#else
+#ifdef U_ALIASING_BARRIER
+    U_ALIASING_BARRIER(p);
+#endif
+    return reinterpret_cast<const char16_t *>(p);
+#endif
+}
+#endif
+#endif
+
+/**
+ * Converts from const char16_t * to const UChar *.
+ * Includes an aliasing barrier if available.
+ * @param p pointer
+ * @return p as const UChar *
+ * @stable ICU 59
+ */
+inline const UChar *toUCharPtr(const char16_t *p) {
+#ifdef U_ALIASING_BARRIER
+    U_ALIASING_BARRIER(p);
+#endif
+    return reinterpret_cast<const UChar *>(p);
+}
+
+/**
+ * Converts from char16_t * to UChar *.
+ * Includes an aliasing barrier if available.
+ * @param p pointer
+ * @return p as UChar *
+ * @stable ICU 59
+ */
+inline UChar *toUCharPtr(char16_t *p) {
+#ifdef U_ALIASING_BARRIER
+    U_ALIASING_BARRIER(p);
+#endif
+    return reinterpret_cast<UChar *>(p);
+}
+
+/**
+ * Converts from const char16_t * to const OldUChar *.
+ * Includes an aliasing barrier if available.
+ * @param p pointer
+ * @return p as const OldUChar *
+ * @stable ICU 59
+ */
+inline const OldUChar *toOldUCharPtr(const char16_t *p) {
+#ifdef U_ALIASING_BARRIER
+    U_ALIASING_BARRIER(p);
+#endif
+    return reinterpret_cast<const OldUChar *>(p);
+}
+
+/**
+ * Converts from char16_t * to OldUChar *.
+ * Includes an aliasing barrier if available.
+ * @param p pointer
+ * @return p as OldUChar *
+ * @stable ICU 59
+ */
+inline OldUChar *toOldUCharPtr(char16_t *p) {
+#ifdef U_ALIASING_BARRIER
+    U_ALIASING_BARRIER(p);
+#endif
+    return reinterpret_cast<OldUChar *>(p);
+}
+
+}  // U_ICU_NAMESPACE_OR_INTERNAL
+
+#endif  // U_SHOW_CPLUSPLUS_API || U_SHOW_CPLUSPLUS_HEADER_API
+
+// ICU DLL-exported
+#if U_SHOW_CPLUSPLUS_API
+
+U_NAMESPACE_BEGIN
+
+#ifndef U_FORCE_HIDE_INTERNAL_API
+/**
+ * Is T convertible to a std::u16string_view or some other 16-bit string view?
+ * @internal
+ */
+template<typename T>
+constexpr bool ConvertibleToU16StringView =
+    std::is_convertible_v<T, std::u16string_view>
+#if !U_CHAR16_IS_TYPEDEF && (!defined(_LIBCPP_VERSION) || _LIBCPP_VERSION < 180000)
+    || std::is_convertible_v<T, std::basic_string_view<uint16_t>>
+#endif
+#if U_SIZEOF_WCHAR_T==2
+    || std::is_convertible_v<T, std::wstring_view>
+#endif
+    ;
+
+namespace internal {
+/**
+ * Pass-through overload.
+ * @internal
+ */
+inline std::u16string_view toU16StringView(std::u16string_view sv) { return sv; }
+
+#if !U_CHAR16_IS_TYPEDEF && (!defined(_LIBCPP_VERSION) || _LIBCPP_VERSION < 180000)
+/**
+ * Basically undefined behavior but sometimes necessary conversion
+ * from std::basic_string_view<uint16_t> to std::u16string_view.
+ * @internal
+ */
+inline std::u16string_view toU16StringView(std::basic_string_view<uint16_t> sv) {
+    return { ConstChar16Ptr(sv.data()), sv.length() };
+}
+#endif
+
+#if U_SIZEOF_WCHAR_T==2
+/**
+ * Basically undefined behavior but sometimes necessary conversion
+ * from std::wstring_view to std::u16string_view.
+ * @internal
+ */
+inline std::u16string_view toU16StringView(std::wstring_view sv) {
+    return { ConstChar16Ptr(sv.data()), sv.length() };
+}
+#endif
+
+/**
+ * Pass-through overload.
+ * @internal
+ */
+template <typename T,
+          typename = typename std::enable_if_t<!std::is_pointer_v<std::remove_reference_t<T>>>>
+inline std::u16string_view toU16StringViewNullable(const T& text) {
+    return toU16StringView(text);
+}
+
+/**
+ * In case of nullptr, return an empty view.
+ * @internal
+ */
+template <typename T,
+          typename = typename std::enable_if_t<std::is_pointer_v<std::remove_reference_t<T>>>,
+          typename = void>
+inline std::u16string_view toU16StringViewNullable(const T& text) {
+    if (text == nullptr) return {};  // For backward compatibility.
+    return toU16StringView(text);
+}
+
+}  // internal
+#endif  // U_FORCE_HIDE_INTERNAL_API
+
+U_NAMESPACE_END
+
+#endif  // U_SHOW_CPLUSPLUS_API
+
+#endif  // __CHAR16PTR_H__

thirdparty/icu4c/common/unicode/chariter.h

@@ -0,0 +1,734 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+/*
+********************************************************************
+*
+*   Copyright (C) 1997-2011, International Business Machines
+*   Corporation and others.  All Rights Reserved.
+*
+********************************************************************
+*/
+
+#ifndef CHARITER_H
+#define CHARITER_H
+
+#include "unicode/utypes.h"
+
+#if U_SHOW_CPLUSPLUS_API
+
+#include "unicode/uobject.h"
+#include "unicode/unistr.h"
+/**
+ * \file
+ * \brief C++ API: Character Iterator
+ */
+ 
+U_NAMESPACE_BEGIN
+/**
+ * Abstract class that defines an API for forward-only iteration
+ * on text objects.
+ * This is a minimal interface for iteration without random access
+ * or backwards iteration. It is especially useful for wrapping
+ * streams with converters into an object for collation or
+ * normalization.
+ *
+ * <p>Characters can be accessed in two ways: as code units or as
+ * code points.
+ * Unicode code points are 21-bit integers and are the scalar values
+ * of Unicode characters. ICU uses the type UChar32 for them.
+ * Unicode code units are the storage units of a given
+ * Unicode/UCS Transformation Format (a character encoding scheme).
+ * With UTF-16, all code points can be represented with either one
+ * or two code units ("surrogates").
+ * String storage is typically based on code units, while properties
+ * of characters are typically determined using code point values.
+ * Some processes may be designed to work with sequences of code units,
+ * or it may be known that all characters that are important to an
+ * algorithm can be represented with single code units.
+ * Other processes will need to use the code point access functions.</p>
+ *
+ * <p>ForwardCharacterIterator provides nextPostInc() to access
+ * a code unit and advance an internal position into the text object,
+ * similar to a <code>return text[position++]</code>.<br>
+ * It provides next32PostInc() to access a code point and advance an internal
+ * position.</p>
+ *
+ * <p>next32PostInc() assumes that the current position is that of
+ * the beginning of a code point, i.e., of its first code unit.
+ * After next32PostInc(), this will be true again.
+ * In general, access to code units and code points in the same
+ * iteration loop should not be mixed. In UTF-16, if the current position
+ * is on a second code unit (Low Surrogate), then only that code unit
+ * is returned even by next32PostInc().</p>
+ *
+ * <p>For iteration with either function, there are two ways to
+ * check for the end of the iteration. When there are no more
+ * characters in the text object:
+ * <ul>
+ * <li>The hasNext() function returns false.</li>
+ * <li>nextPostInc() and next32PostInc() return DONE
+ *     when one attempts to read beyond the end of the text object.</li>
+ * </ul>
+ *
+ * Example:
+ * \code 
+ * void function1(ForwardCharacterIterator &it) {
+ *     UChar32 c;
+ *     while(it.hasNext()) {
+ *         c=it.next32PostInc();
+ *         // use c
+ *     }
+ * }
+ *
+ * void function1(ForwardCharacterIterator &it) {
+ *     char16_t c;
+ *     while((c=it.nextPostInc())!=ForwardCharacterIterator::DONE) {
+ *         // use c
+ *      }
+ *  }
+ * \endcode
+ * </p>
+ *
+ * @stable ICU 2.0
+ */
+class U_COMMON_API ForwardCharacterIterator : public UObject {
+public:
+    /**
+     * Value returned by most of ForwardCharacterIterator's functions
+     * when the iterator has reached the limits of its iteration.
+     * @stable ICU 2.0
+     */
+    enum { DONE = 0xffff };
+    
+    /**
+     * Destructor.  
+     * @stable ICU 2.0
+     */
+    virtual ~ForwardCharacterIterator();
+    
+    /**
+     * Returns true when both iterators refer to the same
+     * character in the same character-storage object.  
+     * @param that The ForwardCharacterIterator to be compared for equality
+     * @return true when both iterators refer to the same
+     * character in the same character-storage object
+     * @stable ICU 2.0
+     */
+    virtual bool operator==(const ForwardCharacterIterator& that) const = 0;
+    
+    /**
+     * Returns true when the iterators refer to different
+     * text-storage objects, or to different characters in the
+     * same text-storage object.  
+     * @param that The ForwardCharacterIterator to be compared for inequality
+     * @return true when the iterators refer to different
+     * text-storage objects, or to different characters in the
+     * same text-storage object
+     * @stable ICU 2.0
+     */
+    inline bool operator!=(const ForwardCharacterIterator& that) const;
+    
+    /**
+     * Generates a hash code for this iterator.  
+     * @return the hash code.
+     * @stable ICU 2.0
+     */
+    virtual int32_t hashCode() const = 0;
+
+    /**
+     * Returns a UClassID for this ForwardCharacterIterator ("poor man's
+     * RTTI").<P> Despite the fact that this function is public,
+     * DO NOT CONSIDER IT PART OF CHARACTERITERATOR'S API! 
+     * @return a UClassID for this ForwardCharacterIterator 
+     * @stable ICU 2.0
+     */
+    virtual UClassID getDynamicClassID() const override = 0;
+
+    /**
+     * Gets the current code unit for returning and advances to the next code unit
+     * in the iteration range
+     * (toward endIndex()).  If there are
+     * no more code units to return, returns DONE.
+     * @return the current code unit.
+     * @stable ICU 2.0
+     */
+    virtual char16_t nextPostInc() = 0;
+
+    /**
+     * Gets the current code point for returning and advances to the next code point
+     * in the iteration range
+     * (toward endIndex()).  If there are
+     * no more code points to return, returns DONE.
+     * @return the current code point.
+     * @stable ICU 2.0
+     */
+    virtual UChar32 next32PostInc() = 0;
+
+    /**
+     * Returns false if there are no more code units or code points
+     * at or after the current position in the iteration range.
+     * This is used with nextPostInc() or next32PostInc() in forward
+     * iteration.
+     * @returns false if there are no more code units or code points
+     * at or after the current position in the iteration range.
+     * @stable ICU 2.0
+     */
+    virtual UBool        hasNext() = 0;
+    
+protected:
+    /** Default constructor to be overridden in the implementing class. @stable ICU 2.0*/
+    ForwardCharacterIterator();
+    
+    /** Copy constructor to be overridden in the implementing class. @stable ICU 2.0*/
+    ForwardCharacterIterator(const ForwardCharacterIterator &other);
+    
+    /**
+     * Assignment operator to be overridden in the implementing class.
+     * @stable ICU 2.0
+     */
+    ForwardCharacterIterator &operator=(const ForwardCharacterIterator&) { return *this; }
+};
+
+/**
+ * Abstract class that defines an API for iteration
+ * on text objects.
+ * This is an interface for forward and backward iteration
+ * and random access into a text object.
+ *
+ * <p>The API provides backward compatibility to the Java and older ICU
+ * CharacterIterator classes but extends them significantly:
+ * <ol>
+ * <li>CharacterIterator is now a subclass of ForwardCharacterIterator.</li>
+ * <li>While the old API functions provided forward iteration with
+ *     "pre-increment" semantics, the new one also provides functions
+ *     with "post-increment" semantics. They are more efficient and should
+ *     be the preferred iterator functions for new implementations.
+ *     The backward iteration always had "pre-decrement" semantics, which
+ *     are efficient.</li>
+ * <li>Just like ForwardCharacterIterator, it provides access to
+ *     both code units and code points. Code point access versions are available
+ *     for the old and the new iteration semantics.</li>
+ * <li>There are new functions for setting and moving the current position
+ *     without returning a character, for efficiency.</li>
+ * </ol>
+ *
+ * See ForwardCharacterIterator for examples for using the new forward iteration
+ * functions. For backward iteration, there is also a hasPrevious() function
+ * that can be used analogously to hasNext().
+ * The old functions work as before and are shown below.</p>
+ *
+ * <p>Examples for some of the new functions:</p>
+ *
+ * Forward iteration with hasNext():
+ * \code
+ * void forward1(CharacterIterator &it) {
+ *     UChar32 c;
+ *     for(it.setToStart(); it.hasNext();) {
+ *         c=it.next32PostInc();
+ *         // use c
+ *     }
+ *  }
+ * \endcode
+ * Forward iteration more similar to loops with the old forward iteration,
+ * showing a way to convert simple for() loops:
+ * \code
+ * void forward2(CharacterIterator &it) {
+ *     char16_t c;
+ *     for(c=it.firstPostInc(); c!=CharacterIterator::DONE; c=it.nextPostInc()) {
+ *          // use c
+ *      }
+ * }
+ * \endcode
+ * Backward iteration with setToEnd() and hasPrevious():
+ * \code
+ *  void backward1(CharacterIterator &it) {
+ *      UChar32 c;
+ *      for(it.setToEnd(); it.hasPrevious();) {
+ *         c=it.previous32();
+ *          // use c
+ *      }
+ *  }
+ * \endcode
+ * Backward iteration with a more traditional for() loop:
+ * \code
+ * void backward2(CharacterIterator &it) {
+ *     char16_t c;
+ *     for(c=it.last(); c!=CharacterIterator::DONE; c=it.previous()) {
+ *         // use c
+ *      }
+ *  }
+ * \endcode
+ *
+ * Example for random access:
+ * \code
+ *  void random(CharacterIterator &it) {
+ *      // set to the third code point from the beginning
+ *      it.move32(3, CharacterIterator::kStart);
+ *      // get a code point from here without moving the position
+ *      UChar32 c=it.current32();
+ *      // get the position
+ *      int32_t pos=it.getIndex();
+ *      // get the previous code unit
+ *      char16_t u=it.previous();
+ *      // move back one more code unit
+ *      it.move(-1, CharacterIterator::kCurrent);
+ *      // set the position back to where it was
+ *      // and read the same code point c and move beyond it
+ *      it.setIndex(pos);
+ *      if(c!=it.next32PostInc()) {
+ *          exit(1); // CharacterIterator inconsistent
+ *      }
+ *  }
+ * \endcode
+ *
+ * <p>Examples, especially for the old API:</p>
+ *
+ * Function processing characters, in this example simple output
+ * <pre>
+ * \code
+ *  void processChar( char16_t c )
+ *  {
+ *      cout << " " << c;
+ *  }
+ * \endcode
+ * </pre>
+ * Traverse the text from start to finish
+ * <pre> 
+ * \code
+ *  void traverseForward(CharacterIterator& iter)
+ *  {
+ *      for(char16_t c = iter.first(); c != CharacterIterator::DONE; c = iter.next()) {
+ *          processChar(c);
+ *      }
+ *  }
+ * \endcode
+ * </pre>
+ * Traverse the text backwards, from end to start
+ * <pre>
+ * \code
+ *  void traverseBackward(CharacterIterator& iter)
+ *  {
+ *      for(char16_t c = iter.last(); c != CharacterIterator::DONE; c = iter.previous()) {
+ *          processChar(c);
+ *      }
+ *  }
+ * \endcode
+ * </pre>
+ * Traverse both forward and backward from a given position in the text. 
+ * Calls to notBoundary() in this example represents some additional stopping criteria.
+ * <pre>
+ * \code
+ * void traverseOut(CharacterIterator& iter, int32_t pos)
+ * {
+ *      char16_t c;
+ *      for (c = iter.setIndex(pos);
+ *      c != CharacterIterator::DONE && (Unicode::isLetter(c) || Unicode::isDigit(c));
+ *          c = iter.next()) {}
+ *      int32_t end = iter.getIndex();
+ *      for (c = iter.setIndex(pos);
+ *          c != CharacterIterator::DONE && (Unicode::isLetter(c) || Unicode::isDigit(c));
+ *          c = iter.previous()) {}
+ *      int32_t start = iter.getIndex() + 1;
+ *  
+ *      cout << "start: " << start << " end: " << end << endl;
+ *      for (c = iter.setIndex(start); iter.getIndex() < end; c = iter.next() ) {
+ *          processChar(c);
+ *     }
+ *  }
+ * \endcode
+ * </pre>
+ * Creating a StringCharacterIterator and calling the test functions
+ * <pre>
+ * \code
+ *  void CharacterIterator_Example( void )
+ *   {
+ *       cout << endl << "===== CharacterIterator_Example: =====" << endl;
+ *       UnicodeString text("Ein kleiner Satz.");
+ *       StringCharacterIterator iterator(text);
+ *       cout << "----- traverseForward: -----------" << endl;
+ *       traverseForward( iterator );
+ *       cout << endl << endl << "----- traverseBackward: ----------" << endl;
+ *       traverseBackward( iterator );
+ *       cout << endl << endl << "----- traverseOut: ---------------" << endl;
+ *       traverseOut( iterator, 7 );
+ *       cout << endl << endl << "-----" << endl;
+ *   }
+ * \endcode
+ * </pre>
+ *
+ * @stable ICU 2.0
+ */
+class U_COMMON_API CharacterIterator : public ForwardCharacterIterator {
+public:
+    /**
+     * Origin enumeration for the move() and move32() functions.
+     * @stable ICU 2.0
+     */
+    enum EOrigin { kStart, kCurrent, kEnd };
+
+    /**
+     * Destructor.
+     * @stable ICU 2.0
+     */
+    virtual ~CharacterIterator();
+
+    /**
+     * Returns a pointer to a new CharacterIterator of the same
+     * concrete class as this one, and referring to the same
+     * character in the same text-storage object as this one.  The
+     * caller is responsible for deleting the new clone.  
+     * @return a pointer to a new CharacterIterator
+     * @stable ICU 2.0
+     */
+    virtual CharacterIterator* clone() const = 0;
+
+    /**
+     * Sets the iterator to refer to the first code unit in its
+     * iteration range, and returns that code unit.
+     * This can be used to begin an iteration with next().
+     * @return the first code unit in its iteration range.
+     * @stable ICU 2.0
+     */
+    virtual char16_t first() = 0;
+
+    /**
+     * Sets the iterator to refer to the first code unit in its
+     * iteration range, returns that code unit, and moves the position
+     * to the second code unit. This is an alternative to setToStart()
+     * for forward iteration with nextPostInc().
+     * @return the first code unit in its iteration range.
+     * @stable ICU 2.0
+     */
+    virtual char16_t firstPostInc();
+
+    /**
+     * Sets the iterator to refer to the first code point in its
+     * iteration range, and returns that code unit,
+     * This can be used to begin an iteration with next32().
+     * Note that an iteration with next32PostInc(), beginning with,
+     * e.g., setToStart() or firstPostInc(), is more efficient.
+     * @return the first code point in its iteration range.
+     * @stable ICU 2.0
+     */
+    virtual UChar32 first32() = 0;
+
+    /**
+     * Sets the iterator to refer to the first code point in its
+     * iteration range, returns that code point, and moves the position
+     * to the second code point. This is an alternative to setToStart()
+     * for forward iteration with next32PostInc().
+     * @return the first code point in its iteration range.
+     * @stable ICU 2.0
+     */
+    virtual UChar32 first32PostInc();
+
+    /**
+     * Sets the iterator to refer to the first code unit or code point in its
+     * iteration range. This can be used to begin a forward
+     * iteration with nextPostInc() or next32PostInc().
+     * @return the start position of the iteration range
+     * @stable ICU 2.0
+     */
+    inline int32_t    setToStart();
+
+    /**
+     * Sets the iterator to refer to the last code unit in its
+     * iteration range, and returns that code unit.
+     * This can be used to begin an iteration with previous().
+     * @return the last code unit.
+     * @stable ICU 2.0
+     */
+    virtual char16_t last() = 0;
+
+    /**
+     * Sets the iterator to refer to the last code point in its
+     * iteration range, and returns that code unit.
+     * This can be used to begin an iteration with previous32().
+     * @return the last code point.
+     * @stable ICU 2.0
+     */
+    virtual UChar32 last32() = 0;
+
+    /**
+     * Sets the iterator to the end of its iteration range, just behind
+     * the last code unit or code point. This can be used to begin a backward
+     * iteration with previous() or previous32().
+     * @return the end position of the iteration range
+     * @stable ICU 2.0
+     */
+    inline int32_t    setToEnd();
+
+    /**
+     * Sets the iterator to refer to the "position"-th code unit
+     * in the text-storage object the iterator refers to, and
+     * returns that code unit.  
+     * @param position the "position"-th code unit in the text-storage object
+     * @return the "position"-th code unit.
+     * @stable ICU 2.0
+     */
+    virtual char16_t         setIndex(int32_t position) = 0;
+
+    /**
+     * Sets the iterator to refer to the beginning of the code point
+     * that contains the "position"-th code unit
+     * in the text-storage object the iterator refers to, and
+     * returns that code point.
+     * The current position is adjusted to the beginning of the code point
+     * (its first code unit).
+     * @param position the "position"-th code unit in the text-storage object
+     * @return the "position"-th code point.
+     * @stable ICU 2.0
+     */
+    virtual UChar32       setIndex32(int32_t position) = 0;
+
+    /**
+     * Returns the code unit the iterator currently refers to. 
+     * @return the current code unit. 
+     * @stable ICU 2.0
+     */
+    virtual char16_t current() const = 0;
+
+    /**
+     * Returns the code point the iterator currently refers to.  
+     * @return the current code point.
+     * @stable ICU 2.0
+     */
+    virtual UChar32 current32() const = 0;
+
+    /**
+     * Advances to the next code unit in the iteration range
+     * (toward endIndex()), and returns that code unit.  If there are
+     * no more code units to return, returns DONE.
+     * @return the next code unit.
+     * @stable ICU 2.0
+     */
+    virtual char16_t next() = 0;
+
+    /**
+     * Advances to the next code point in the iteration range
+     * (toward endIndex()), and returns that code point.  If there are
+     * no more code points to return, returns DONE.
+     * Note that iteration with "pre-increment" semantics is less
+     * efficient than iteration with "post-increment" semantics
+     * that is provided by next32PostInc().
+     * @return the next code point.
+     * @stable ICU 2.0
+     */
+    virtual UChar32 next32() = 0;
+
+    /**
+     * Advances to the previous code unit in the iteration range
+     * (toward startIndex()), and returns that code unit.  If there are
+     * no more code units to return, returns DONE.  
+     * @return the previous code unit.
+     * @stable ICU 2.0
+     */
+    virtual char16_t previous() = 0;
+
+    /**
+     * Advances to the previous code point in the iteration range
+     * (toward startIndex()), and returns that code point.  If there are
+     * no more code points to return, returns DONE. 
+     * @return the previous code point. 
+     * @stable ICU 2.0
+     */
+    virtual UChar32 previous32() = 0;
+
+    /**
+     * Returns false if there are no more code units or code points
+     * before the current position in the iteration range.
+     * This is used with previous() or previous32() in backward
+     * iteration.
+     * @return false if there are no more code units or code points
+     * before the current position in the iteration range, return true otherwise.
+     * @stable ICU 2.0
+     */
+    virtual UBool        hasPrevious() = 0;
+
+    /**
+     * Returns the numeric index in the underlying text-storage
+     * object of the character returned by first().  Since it's
+     * possible to create an iterator that iterates across only
+     * part of a text-storage object, this number isn't
+     * necessarily 0.  
+     * @returns the numeric index in the underlying text-storage
+     * object of the character returned by first().
+     * @stable ICU 2.0
+     */
+    inline int32_t startIndex() const;
+
+    /**
+     * Returns the numeric index in the underlying text-storage
+     * object of the position immediately BEYOND the character
+     * returned by last().  
+     * @return the numeric index in the underlying text-storage
+     * object of the position immediately BEYOND the character
+     * returned by last().
+     * @stable ICU 2.0
+     */
+    inline int32_t endIndex() const;
+
+    /**
+     * Returns the numeric index in the underlying text-storage
+     * object of the character the iterator currently refers to
+     * (i.e., the character returned by current()).  
+     * @return the numeric index in the text-storage object of 
+     * the character the iterator currently refers to
+     * @stable ICU 2.0
+     */
+    inline int32_t getIndex() const;
+
+    /**
+     * Returns the length of the entire text in the underlying
+     * text-storage object.
+     * @return the length of the entire text in the text-storage object
+     * @stable ICU 2.0
+     */
+    inline int32_t           getLength() const;
+
+    /**
+     * Moves the current position relative to the start or end of the
+     * iteration range, or relative to the current position itself.
+     * The movement is expressed in numbers of code units forward
+     * or backward by specifying a positive or negative delta.
+     * @param delta the position relative to origin. A positive delta means forward;
+     * a negative delta means backward.
+     * @param origin Origin enumeration {kStart, kCurrent, kEnd}
+     * @return the new position
+     * @stable ICU 2.0
+     */
+    virtual int32_t      move(int32_t delta, EOrigin origin) = 0;
+
+    /**
+     * Moves the current position relative to the start or end of the
+     * iteration range, or relative to the current position itself.
+     * The movement is expressed in numbers of code points forward
+     * or backward by specifying a positive or negative delta.
+     * @param delta the position relative to origin. A positive delta means forward;
+     * a negative delta means backward.
+     * @param origin Origin enumeration {kStart, kCurrent, kEnd}
+     * @return the new position
+     * @stable ICU 2.0
+     */
+#ifdef move32
+     // One of the system headers right now is sometimes defining a conflicting macro we don't use
+#undef move32
+#endif
+    virtual int32_t      move32(int32_t delta, EOrigin origin) = 0;
+
+    /**
+     * Copies the text under iteration into the UnicodeString
+     * referred to by "result".  
+     * @param result Receives a copy of the text under iteration.  
+     * @stable ICU 2.0
+     */
+    virtual void            getText(UnicodeString&  result) = 0;
+
+protected:
+    /**
+     * Empty constructor.
+     * @stable ICU 2.0
+     */
+    CharacterIterator();
+
+    /**
+     * Constructor, just setting the length field in this base class.
+     * @stable ICU 2.0
+     */
+    CharacterIterator(int32_t length);
+
+    /**
+     * Constructor, just setting the length and position fields in this base class.
+     * @stable ICU 2.0
+     */
+    CharacterIterator(int32_t length, int32_t position);
+
+    /**
+     * Constructor, just setting the length, start, end, and position fields in this base class.
+     * @stable ICU 2.0
+     */
+    CharacterIterator(int32_t length, int32_t textBegin, int32_t textEnd, int32_t position);
+  
+    /**
+     * Copy constructor.
+     *
+     * @param that The CharacterIterator to be copied
+     * @stable ICU 2.0
+     */
+    CharacterIterator(const CharacterIterator &that);
+
+    /**
+     * Assignment operator.  Sets this CharacterIterator to have the same behavior,
+     * as the one passed in.
+     * @param that The CharacterIterator passed in.
+     * @return the newly set CharacterIterator.
+     * @stable ICU 2.0
+     */
+    CharacterIterator &operator=(const CharacterIterator &that);
+
+    /**
+     * Base class text length field.
+     * Necessary this for correct getText() and hashCode().
+     * @stable ICU 2.0
+     */
+    int32_t textLength;
+
+    /**
+     * Base class field for the current position.
+     * @stable ICU 2.0
+     */
+    int32_t  pos;
+
+    /**
+     * Base class field for the start of the iteration range.
+     * @stable ICU 2.0
+     */
+    int32_t  begin;
+
+    /**
+     * Base class field for the end of the iteration range.
+     * @stable ICU 2.0
+     */
+    int32_t  end;
+};
+
+inline bool
+ForwardCharacterIterator::operator!=(const ForwardCharacterIterator& that) const {
+    return !operator==(that);
+}
+
+inline int32_t
+CharacterIterator::setToStart() {
+    return move(0, kStart);
+}
+
+inline int32_t
+CharacterIterator::setToEnd() {
+    return move(0, kEnd);
+}
+
+inline int32_t
+CharacterIterator::startIndex() const {
+    return begin;
+}
+
+inline int32_t
+CharacterIterator::endIndex() const {
+    return end;
+}
+
+inline int32_t
+CharacterIterator::getIndex() const {
+    return pos;
+}
+
+inline int32_t
+CharacterIterator::getLength() const {
+    return textLength;
+}
+
+U_NAMESPACE_END
+
+#endif /* U_SHOW_CPLUSPLUS_API */
+
+#endif

thirdparty/icu4c/common/unicode/dbbi.h

@@ -0,0 +1,48 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+/*
+**********************************************************************
+*   Copyright (C) 1999-2006,2013 IBM Corp. All rights reserved.
+**********************************************************************
+*   Date        Name        Description
+*   12/1/99    rgillam     Complete port from Java.
+*   01/13/2000 helena      Added UErrorCode to ctors.
+**********************************************************************
+*/
+
+#ifndef DBBI_H
+#define DBBI_H
+
+#include "unicode/utypes.h"
+
+#if U_SHOW_CPLUSPLUS_API
+
+#include "unicode/rbbi.h"
+
+#if !UCONFIG_NO_BREAK_ITERATION
+
+/**
+ * \file
+ * \brief C++ API: Dictionary Based Break Iterator
+ */
+ 
+U_NAMESPACE_BEGIN
+
+#ifndef U_HIDE_DEPRECATED_API
+/**
+ * An obsolete subclass of RuleBasedBreakIterator. Handling of dictionary-
+ * based break iteration has been folded into the base class. This class
+ * is deprecated as of ICU 3.6.
+ * @deprecated ICU 3.6
+ */
+typedef RuleBasedBreakIterator DictionaryBasedBreakIterator;
+
+#endif  /* U_HIDE_DEPRECATED_API */
+
+U_NAMESPACE_END
+
+#endif /* #if !UCONFIG_NO_BREAK_ITERATION */
+
+#endif /* U_SHOW_CPLUSPLUS_API */
+
+#endif

thirdparty/icu4c/common/unicode/docmain.h

@@ -0,0 +1,247 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+/********************************************************************
+ * COPYRIGHT: 
+ * Copyright (c) 1997-2012, International Business Machines Corporation and
+ * others. All Rights Reserved.
+ *
+ *  FILE NAME: DOCMAIN.h
+ *
+ *   Date          Name        Description
+ *   12/11/2000    Ram        Creation.
+ */
+
+/**
+ * \file
+ * \brief (Non API- contains Doxygen definitions)
+ *
+ * This file contains documentation for Doxygen and does not have
+ * any significance with respect to C or C++ API
+ */
+
+/*! \mainpage
+ *
+ * \section API API Reference Usage
+ * 
+ * <h3>C++ Programmers:</h3>
+ * <p>Use <a href="hierarchy.html">Class Hierarchy</a> or <a href="classes.html"> Alphabetical List </a>
+ * or <a href="annotated.html"> Compound List</a>
+ * to find the class you are interested in. For example, to find BreakIterator,
+ * you can go to the <a href="classes.html"> Alphabetical List</a>, then click on
+ * "BreakIterator". Once you are at the class, you will find an inheritance
+ * chart, a list of the public members, a detailed description of the class,
+ * then detailed member descriptions.</p>
+ * 
+ * <h3>C Programmers:</h3>
+ * <p>Use <a href="#Module">Module List</a> or <a href="globals_u.html">File Members</a>
+ * to find a list of all the functions and constants.
+ * For example, to find BreakIterator functions you would click on
+ * <a href="files.html"> File List</a>,
+ * then find "ubrk.h" and click on it. You will find descriptions of Defines,
+ * Typedefs, Enumerations, and Functions, with detailed descriptions below.
+ * If you want to find a specific function, such as ubrk_next(), then click
+ * first on <a href="globals.html"> File Members</a>, then use your browser
+ * Find dialog to search for "ubrk_next()".</p>
+ *
+ *
+ * <h3>API References for Previous Releases</h3>
+ * <p>The API References for each release of ICU are also available as
+ * a zip file from the ICU 
+ * <a href="https://icu.unicode.org/download">download page</a>.</p>
+ *
+ * <hr>
+ *
+ * <h2>Architecture (User's Guide)</h2>
+ * <ul>
+ *   <li><a href="https://unicode-org.github.io/icu/userguide/">Introduction</a></li>
+ *   <li><a href="https://unicode-org.github.io/icu/userguide/i18n">Internationalization</a></li>
+ *   <li><a href="https://unicode-org.github.io/icu/userguide/design">Locale Model, Multithreading, Error Handling, etc.</a></li>
+ *   <li><a href="https://unicode-org.github.io/icu/userguide/conversion">Conversion</a></li>
+ * </ul>
+ *
+ * <hr>
+ *\htmlonly <h2><a NAME="Module">Module List</a></h2> \endhtmlonly
+ * <table border="1" cols="3" align="center">
+ *   <tr>
+ *     <td><strong>Module Name</strong></td>
+ *     <td><strong>C</strong></td>
+ *     <td><strong>C++</strong></td>
+ *   </tr>
+ *   <tr>
+ *     <td>Basic Types and Constants</td>
+ *     <td>utypes.h</td>
+ *     <td>utypes.h</td>
+ *   </tr>
+ *   <tr>
+ *     <td>Strings and Character Iteration</td>
+ *     <td>ustring.h, utf8.h, utf16.h, icu::StringPiece, UText, UCharIterator, icu::ByteSink</td>
+ *     <td>icu::UnicodeString, icu::CharacterIterator, icu::Appendable, icu::StringPiece,icu::ByteSink</td>
+ *   </tr>
+ *   <tr>
+ *     <td>Unicode Character<br/>Properties and Names</td>
+ *     <td>uchar.h, uscript.h</td>
+ *     <td>C API</td>
+ *   </tr>
+ *   <tr>
+ *     <td>Sets of Unicode Code Points and Strings</td>
+ *     <td>uset.h</td>
+ *     <td>icu::UnicodeSet</td>
+ *   </tr>
+ *   <tr>
+ *     <td>Maps from Unicode Code Points to Integer Values</td>
+ *     <td>ucptrie.h, umutablecptrie.h</td>
+ *     <td>C API</td>
+ *   </tr>
+ *   <tr>
+ *     <td>Maps from Strings to Integer Values</td>
+ *     <td>(no C API)</td>
+ *     <td>icu::BytesTrie, icu::UCharsTrie</td>
+ *   </tr>
+ *   <tr>
+ *     <td>Codepage Conversion</td>
+ *     <td>ucnv.h, ucnvsel.h</td>
+ *     <td>C API</td>
+ *   </tr>
+ *   <tr>
+ *     <td>Codepage Detection</td>
+ *     <td>ucsdet.h</td>
+ *     <td>C API</td>
+ *   </tr>
+ *   <tr>
+ *     <td>Unicode Text Compression</td>
+ *     <td>ucnv.h<br/>(encoding name "SCSU" or "BOCU-1")</td>
+ *     <td>C API</td>
+ *   </tr>
+ *   <tr>
+ *     <td>Locales </td>
+ *     <td>uloc.h, ulocale.h, ulocbuilder.h</a></td>
+ *     <td>icu::Locale, icu::LocaleBuilder, icu::LocaleMatcher</td>
+ *   </tr>
+ *   <tr>
+ *     <td>Resource Bundles</td>
+ *     <td>ures.h</td>
+ *     <td>icu::ResourceBundle</td>
+ *   </tr>
+ *   <tr>
+ *     <td>Normalization</td>
+ *     <td>unorm2.h</td>
+ *     <td>icu::Normalizer2</td>
+ *   </tr>
+ *   <tr>
+ *     <td>Calendars and Time Zones</td>
+ *     <td>ucal.h</td>
+ *     <td>icu::Calendar, icu::TimeZone</td>
+ *   </tr>
+ *   <tr>
+ *     <td>Date and Time Formatting</td>
+ *     <td>udat.h</td>
+ *     <td>icu::DateFormat</td>
+ *   </tr>
+ *   <tr>
+ *     <td>Relative Date and Time Formatting</td>
+ *     <td>ureldatefmt.h</td>
+ *     <td>icu::RelativeDateTimeFormatter</td>
+ *   </tr>
+ *   <tr>
+ *     <td>Message Formatting</td>
+ *     <td>umsg.h</td>
+ *     <td>icu::MessageFormat</td>
+ *   </tr>
+ *   <tr>
+ *     <td>Message Formatting 2<br/>(technology preview)</td>
+ *     <td>(no C API)</td>
+ *     <td>icu::message2::MessageFormatter</td>
+ *   </tr>
+ *   <tr>
+ *     <td>List Formatting</td>
+ *     <td>ulistformatter.h</td>
+ *     <td>icu::ListFormatter</td>
+ *   </tr>
+ *   <tr>
+ *     <td>Number Formatting<br/>(includes currency and unit formatting)</td>
+ *     <td>unumberformatter.h, unum.h, usimplenumberformatter.h</td>
+ *     <td>icu::number::NumberFormatter (ICU 60+) or icu::NumberFormat (older versions)<br>icu::number::SimpleNumberFormatter (ICU 73+)</td>
+ *   </tr>
+ *   <tr>
+ *     <td>Number Range Formatting<br />(includes currency and unit ranges)</td>
+ *     <td>unumberrangeformatter.h</td>
+ *     <td>icu::number::NumberRangeFormatter</td>
+ *   </tr>
+ *   <tr>
+ *     <td>Number Spellout<br/>(Rule Based Number Formatting)</td>
+ *     <td>unum.h<br/>(use UNUM_SPELLOUT)</td>
+ *     <td>icu::RuleBasedNumberFormat</td>
+ *   </tr>
+ *   <tr>
+ *     <td>Text Transformation<br/>(Transliteration)</td>
+ *     <td>utrans.h</td>
+ *     <td>icu::Transliterator</td>
+ *   </tr>
+ *   <tr>
+ *     <td>Bidirectional Algorithm</td>
+ *     <td>ubidi.h, ubiditransform.h</td>
+ *     <td>C API</td>
+ *   </tr>
+ *   <tr>
+ *     <td>Arabic Shaping</td>
+ *     <td>ushape.h</td>
+ *     <td>C API</td>
+ *   </tr>
+ *   <tr>
+ *     <td>Collation</td>
+ *     <td>ucol.h</td>
+ *     <td>icu::Collator</td>
+ *   </tr>
+ *   <tr>
+ *     <td>String Searching</td>
+ *     <td>usearch.h</td>
+ *     <td>icu::StringSearch</td>
+ *   </tr>
+ *   <tr>
+ *     <td>Index Characters/<br/>Bucketing for Sorted Lists</td>
+ *     <td>(no C API)</td>
+ *     <td>icu::AlphabeticIndex</td>
+ *   </tr>
+ *   <tr>
+ *     <td>Text Boundary Analysis<br/>(Break Iteration)</td>
+ *     <td>ubrk.h</td>
+ *     <td>icu::BreakIterator</td>
+ *   </tr>
+ *   <tr>
+ *     <td>Regular Expressions</td>
+ *     <td>uregex.h</td>
+ *     <td>icu::RegexPattern, icu::RegexMatcher</td>
+ *   </tr>
+ *   <tr>
+ *     <td>StringPrep</td>
+ *     <td>usprep.h</td>
+ *     <td>C API</td>
+ *   </tr>
+ *   <tr>
+ *     <td>International Domain Names in Applications:<br/>
+ *         UTS #46 in C/C++, IDNA2003 only via C API</td>
+ *     <td>uidna.h</td>
+ *     <td>idna.h</td>
+ *   </tr>
+ *   <tr>
+ *     <td>Identifier Spoofing & Confusability</td>
+ *     <td>uspoof.h</td>
+ *     <td>C API</td>
+ *   <tr>
+ *     <td>Universal Time Scale</td>
+ *     <td>utmscale.h</td>
+ *     <td>C API</td>
+ *   </tr>
+ *   <tr>
+ *     <td>Paragraph Layout / Complex Text Layout</td>
+ *     <td>playout.h</td>
+ *     <td>icu::ParagraphLayout</td>
+ *   </tr>
+ *   <tr>
+ *     <td>ICU I/O</td>
+ *     <td>ustdio.h</td>
+ *     <td>ustream.h</td>
+ *   </tr>
+ * </table>
+ * <i>This main page is generated from docmain.h</i>
+ */

thirdparty/icu4c/common/unicode/dtintrv.h

@@ -0,0 +1,163 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+/*
+*******************************************************************************
+* Copyright (C) 2008-2009, International Business Machines Corporation and
+* others. All Rights Reserved.
+*******************************************************************************
+*
+* File DTINTRV.H 
+*
+*******************************************************************************
+*/
+
+#ifndef __DTINTRV_H__
+#define __DTINTRV_H__
+
+#include "unicode/utypes.h"
+
+#if U_SHOW_CPLUSPLUS_API
+
+#include "unicode/uobject.h"
+
+/**
+ * \file
+ * \brief C++ API: Date Interval data type
+ */
+
+U_NAMESPACE_BEGIN
+
+
+/**
+ * This class represents a date interval.
+ * It is a pair of UDate representing from UDate 1 to UDate 2.
+ * @stable ICU 4.0
+**/
+class U_COMMON_API DateInterval : public UObject {
+public:
+
+    /** 
+     * Construct a DateInterval given a from date and a to date.
+     * @param fromDate  The from date in date interval.
+     * @param toDate    The to date in date interval.
+     * @stable ICU 4.0
+     */
+    DateInterval(UDate fromDate, UDate toDate);
+
+    /**
+     * destructor
+     * @stable ICU 4.0
+     */
+    virtual ~DateInterval();
+ 
+    /** 
+     * Get the from date.
+     * @return  the from date in dateInterval.
+     * @stable ICU 4.0
+     */
+    inline UDate getFromDate() const;
+
+    /** 
+     * Get the to date.
+     * @return  the to date in dateInterval.
+     * @stable ICU 4.0
+     */
+    inline UDate getToDate() const;
+
+
+    /**
+     * Return the class ID for this class. This is useful only for comparing to
+     * a return value from getDynamicClassID(). For example:
+     * <pre>
+     * .   Base* polymorphic_pointer = createPolymorphicObject();
+     * .   if (polymorphic_pointer->getDynamicClassID() ==
+     * .       derived::getStaticClassID()) ...
+     * </pre>
+     * @return          The class ID for all objects of this class.
+     * @stable ICU 4.0
+     */
+    static UClassID U_EXPORT2 getStaticClassID();
+
+    /**
+     * Returns a unique class ID POLYMORPHICALLY. Pure virtual override. This
+     * method is to implement a simple version of RTTI, since not all C++
+     * compilers support genuine RTTI. Polymorphic operator==() and clone()
+     * methods call this method.
+     *
+     * @return          The class ID for this object. All objects of a
+     *                  given class have the same class ID.  Objects of
+     *                  other classes have different class IDs.
+     * @stable ICU 4.0
+     */
+    virtual UClassID getDynamicClassID() const override;
+
+    /**
+     * Copy constructor.
+     * @stable ICU 4.0
+     */
+    DateInterval(const DateInterval& other);
+
+    /**
+     * Default assignment operator
+     * @stable ICU 4.0
+     */
+    DateInterval& operator=(const DateInterval&);
+
+    /**
+     * Equality operator.
+     * @return true if the two DateIntervals are the same
+     * @stable ICU 4.0
+     */
+    virtual bool operator==(const DateInterval& other) const;
+
+    /**
+     * Non-equality operator
+     * @return true if the two DateIntervals are not the same
+     * @stable ICU 4.0
+     */
+    inline bool operator!=(const DateInterval& other) const;
+
+
+    /**
+     * clone this object. 
+     * The caller owns the result and should delete it when done.
+     * @return a cloned DateInterval
+     * @stable ICU 4.0
+     */
+     virtual DateInterval* clone() const;
+
+private:
+    /** 
+     * Default constructor, not implemented.
+     */
+    DateInterval() = delete;
+
+    UDate fromDate;
+    UDate toDate;
+
+} ;// end class DateInterval
+
+
+inline UDate 
+DateInterval::getFromDate() const { 
+    return fromDate; 
+}
+
+
+inline UDate 
+DateInterval::getToDate() const { 
+    return toDate; 
+}
+
+
+inline bool
+DateInterval::operator!=(const DateInterval& other) const { 
+    return ( !operator==(other) );
+}
+
+
+U_NAMESPACE_END
+
+#endif /* U_SHOW_CPLUSPLUS_API */
+
+#endif

thirdparty/icu4c/common/unicode/edits.h

@@ -0,0 +1,531 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+
+// edits.h
+// created: 2016dec30 Markus W. Scherer
+
+#ifndef __EDITS_H__
+#define __EDITS_H__
+
+#include "unicode/utypes.h"
+
+#if U_SHOW_CPLUSPLUS_API
+
+#include "unicode/uobject.h"
+
+/**
+ * \file
+ * \brief C++ API: C++ class Edits for low-level string transformations on styled text.
+ */
+
+U_NAMESPACE_BEGIN
+
+class UnicodeString;
+
+/**
+ * Records lengths of string edits but not replacement text. Supports replacements, insertions, deletions
+ * in linear progression. Does not support moving/reordering of text.
+ *
+ * There are two types of edits: <em>change edits</em> and <em>no-change edits</em>. Add edits to
+ * instances of this class using {@link #addReplace(int32_t, int32_t)} (for change edits) and
+ * {@link #addUnchanged(int32_t)} (for no-change edits). Change edits are retained with full granularity,
+ * whereas adjacent no-change edits are always merged together. In no-change edits, there is a one-to-one
+ * mapping between code points in the source and destination strings.
+ *
+ * After all edits have been added, instances of this class should be considered immutable, and an
+ * {@link Edits::Iterator} can be used for queries.
+ *
+ * There are four flavors of Edits::Iterator:
+ *
+ * <ul>
+ * <li>{@link #getFineIterator()} retains full granularity of change edits.
+ * <li>{@link #getFineChangesIterator()} retains full granularity of change edits, and when calling
+ * next() on the iterator, skips over no-change edits (unchanged regions).
+ * <li>{@link #getCoarseIterator()} treats adjacent change edits as a single edit. (Adjacent no-change
+ * edits are automatically merged during the construction phase.)
+ * <li>{@link #getCoarseChangesIterator()} treats adjacent change edits as a single edit, and when
+ * calling next() on the iterator, skips over no-change edits (unchanged regions).
+ * </ul>
+ *
+ * For example, consider the string "abcßDeF", which case-folds to "abcssdef". This string has the
+ * following fine edits:
+ * <ul>
+ * <li>abc ⇨ abc (no-change)
+ * <li>ß ⇨ ss (change)
+ * <li>D ⇨ d (change)
+ * <li>e ⇨ e (no-change)
+ * <li>F ⇨ f (change)
+ * </ul>
+ * and the following coarse edits (note how adjacent change edits get merged together):
+ * <ul>
+ * <li>abc ⇨ abc (no-change)
+ * <li>ßD ⇨ ssd (change)
+ * <li>e ⇨ e (no-change)
+ * <li>F ⇨ f (change)
+ * </ul>
+ *
+ * The "fine changes" and "coarse changes" iterators will step through only the change edits when their
+ * `Edits::Iterator::next()` methods are called. They are identical to the non-change iterators when
+ * their `Edits::Iterator::findSourceIndex()` or `Edits::Iterator::findDestinationIndex()`
+ * methods are used to walk through the string.
+ *
+ * For examples of how to use this class, see the test `TestCaseMapEditsIteratorDocs` in
+ * UCharacterCaseTest.java.
+ *
+ * An Edits object tracks a separate UErrorCode, but ICU string transformation functions
+ * (e.g., case mapping functions) merge any such errors into their API's UErrorCode.
+ *
+ * @stable ICU 59
+ */
+class U_COMMON_API Edits final : public UMemory {
+public:
+    /**
+     * Constructs an empty object.
+     * @stable ICU 59
+     */
+    Edits() :
+            array(stackArray), capacity(STACK_CAPACITY), length(0), delta(0), numChanges(0),
+            errorCode_(U_ZERO_ERROR) {}
+    /**
+     * Copy constructor.
+     * @param other source edits
+     * @stable ICU 60
+     */
+    Edits(const Edits &other) :
+            array(stackArray), capacity(STACK_CAPACITY), length(other.length),
+            delta(other.delta), numChanges(other.numChanges),
+            errorCode_(other.errorCode_) {
+        copyArray(other);
+    }
+    /**
+     * Move constructor, might leave src empty.
+     * This object will have the same contents that the source object had.
+     * @param src source edits
+     * @stable ICU 60
+     */
+    Edits(Edits &&src) noexcept :
+            array(stackArray), capacity(STACK_CAPACITY), length(src.length),
+            delta(src.delta), numChanges(src.numChanges),
+            errorCode_(src.errorCode_) {
+        moveArray(src);
+    }
+
+    /**
+     * Destructor.
+     * @stable ICU 59
+     */
+    ~Edits();
+
+    /**
+     * Assignment operator.
+     * @param other source edits
+     * @return *this
+     * @stable ICU 60
+     */
+    Edits &operator=(const Edits &other);
+
+    /**
+     * Move assignment operator, might leave src empty.
+     * This object will have the same contents that the source object had.
+     * The behavior is undefined if *this and src are the same object.
+     * @param src source edits
+     * @return *this
+     * @stable ICU 60
+     */
+    Edits &operator=(Edits &&src) noexcept;
+
+    /**
+     * Resets the data but may not release memory.
+     * @stable ICU 59
+     */
+    void reset() noexcept;
+
+    /**
+     * Adds a no-change edit: a record for an unchanged segment of text.
+     * Normally called from inside ICU string transformation functions, not user code.
+     * @stable ICU 59
+     */
+    void addUnchanged(int32_t unchangedLength);
+    /**
+     * Adds a change edit: a record for a text replacement/insertion/deletion.
+     * Normally called from inside ICU string transformation functions, not user code.
+     * @stable ICU 59
+     */
+    void addReplace(int32_t oldLength, int32_t newLength);
+    /**
+     * Sets the UErrorCode if an error occurred while recording edits.
+     * Preserves older error codes in the outErrorCode.
+     * Normally called from inside ICU string transformation functions, not user code.
+     * @param outErrorCode Set to an error code if it does not contain one already
+     *                  and an error occurred while recording edits.
+     *                  Otherwise unchanged.
+     * @return true if U_FAILURE(outErrorCode)
+     * @stable ICU 59
+     */
+    UBool copyErrorTo(UErrorCode &outErrorCode) const;
+
+    /**
+     * How much longer is the new text compared with the old text?
+     * @return new length minus old length
+     * @stable ICU 59
+     */
+    int32_t lengthDelta() const { return delta; }
+    /**
+     * @return true if there are any change edits
+     * @stable ICU 59
+     */
+    UBool hasChanges() const { return numChanges != 0; }
+
+    /**
+     * @return the number of change edits
+     * @stable ICU 60
+     */
+    int32_t numberOfChanges() const { return numChanges; }
+
+    /**
+     * Access to the list of edits.
+     *
+     * At any moment in time, an instance of this class points to a single edit: a "window" into a span
+     * of the source string and the corresponding span of the destination string. The source string span
+     * starts at {@link #sourceIndex()} and runs for {@link #oldLength()} chars; the destination string
+     * span starts at {@link #destinationIndex()} and runs for {@link #newLength()} chars.
+     *
+     * The iterator can be moved between edits using the `next()`, `findSourceIndex(int32_t, UErrorCode &)`,
+     * and `findDestinationIndex(int32_t, UErrorCode &)` methods.
+     * Calling any of these methods mutates the iterator to make it point to the corresponding edit.
+     *
+     * For more information, see the documentation for {@link Edits}.
+     *
+     * @see getCoarseIterator
+     * @see getFineIterator
+     * @stable ICU 59
+     */
+    struct U_COMMON_API Iterator final : public UMemory {
+        /**
+         * Default constructor, empty iterator.
+         * @stable ICU 60
+         */
+        Iterator() :
+                array(nullptr), index(0), length(0),
+                remaining(0), onlyChanges_(false), coarse(false),
+                dir(0), changed(false), oldLength_(0), newLength_(0),
+                srcIndex(0), replIndex(0), destIndex(0) {}
+        /**
+         * Copy constructor.
+         * @stable ICU 59
+         */
+        Iterator(const Iterator &other) = default;
+        /**
+         * Assignment operator.
+         * @stable ICU 59
+         */
+        Iterator &operator=(const Iterator &other) = default;
+
+        /**
+         * Advances the iterator to the next edit.
+         * @param errorCode ICU error code. Its input value must pass the U_SUCCESS() test,
+         *                  or else the function returns immediately. Check for U_FAILURE()
+         *                  on output or use with function chaining. (See User Guide for details.)
+         * @return true if there is another edit
+         * @stable ICU 59
+         */
+        UBool next(UErrorCode &errorCode) { return next(onlyChanges_, errorCode); }
+
+        /**
+         * Moves the iterator to the edit that contains the source index.
+         * The source index may be found in a no-change edit
+         * even if normal iteration would skip no-change edits.
+         * Normal iteration can continue from a found edit.
+         *
+         * The iterator state before this search logically does not matter.
+         * (It may affect the performance of the search.)
+         *
+         * The iterator state after this search is undefined
+         * if the source index is out of bounds for the source string.
+         *
+         * @param i source index
+         * @param errorCode ICU error code. Its input value must pass the U_SUCCESS() test,
+         *                  or else the function returns immediately. Check for U_FAILURE()
+         *                  on output or use with function chaining. (See User Guide for details.)
+         * @return true if the edit for the source index was found
+         * @stable ICU 59
+         */
+        UBool findSourceIndex(int32_t i, UErrorCode &errorCode) {
+            return findIndex(i, true, errorCode) == 0;
+        }
+
+        /**
+         * Moves the iterator to the edit that contains the destination index.
+         * The destination index may be found in a no-change edit
+         * even if normal iteration would skip no-change edits.
+         * Normal iteration can continue from a found edit.
+         *
+         * The iterator state before this search logically does not matter.
+         * (It may affect the performance of the search.)
+         *
+         * The iterator state after this search is undefined
+         * if the source index is out of bounds for the source string.
+         *
+         * @param i destination index
+         * @param errorCode ICU error code. Its input value must pass the U_SUCCESS() test,
+         *                  or else the function returns immediately. Check for U_FAILURE()
+         *                  on output or use with function chaining. (See User Guide for details.)
+         * @return true if the edit for the destination index was found
+         * @stable ICU 60
+         */
+        UBool findDestinationIndex(int32_t i, UErrorCode &errorCode) {
+            return findIndex(i, false, errorCode) == 0;
+        }
+
+        /**
+         * Computes the destination index corresponding to the given source index.
+         * If the source index is inside a change edit (not at its start),
+         * then the destination index at the end of that edit is returned,
+         * since there is no information about index mapping inside a change edit.
+         *
+         * (This means that indexes to the start and middle of an edit,
+         * for example around a grapheme cluster, are mapped to indexes
+         * encompassing the entire edit.
+         * The alternative, mapping an interior index to the start,
+         * would map such an interval to an empty one.)
+         *
+         * This operation will usually but not always modify this object.
+         * The iterator state after this search is undefined.
+         *
+         * @param i source index
+         * @param errorCode ICU error code. Its input value must pass the U_SUCCESS() test,
+         *                  or else the function returns immediately. Check for U_FAILURE()
+         *                  on output or use with function chaining. (See User Guide for details.)
+         * @return destination index; undefined if i is not 0..string length
+         * @stable ICU 60
+         */
+        int32_t destinationIndexFromSourceIndex(int32_t i, UErrorCode &errorCode);
+
+        /**
+         * Computes the source index corresponding to the given destination index.
+         * If the destination index is inside a change edit (not at its start),
+         * then the source index at the end of that edit is returned,
+         * since there is no information about index mapping inside a change edit.
+         *
+         * (This means that indexes to the start and middle of an edit,
+         * for example around a grapheme cluster, are mapped to indexes
+         * encompassing the entire edit.
+         * The alternative, mapping an interior index to the start,
+         * would map such an interval to an empty one.)
+         *
+         * This operation will usually but not always modify this object.
+         * The iterator state after this search is undefined.
+         *
+         * @param i destination index
+         * @param errorCode ICU error code. Its input value must pass the U_SUCCESS() test,
+         *                  or else the function returns immediately. Check for U_FAILURE()
+         *                  on output or use with function chaining. (See User Guide for details.)
+         * @return source index; undefined if i is not 0..string length
+         * @stable ICU 60
+         */
+        int32_t sourceIndexFromDestinationIndex(int32_t i, UErrorCode &errorCode);
+
+        /**
+         * Returns whether the edit currently represented by the iterator is a change edit.
+         *
+         * @return true if this edit replaces oldLength() units with newLength() different ones.
+         *         false if oldLength units remain unchanged.
+         * @stable ICU 59
+         */
+        UBool hasChange() const { return changed; }
+
+        /**
+         * The length of the current span in the source string, which starts at {@link #sourceIndex}.
+         *
+         * @return the number of units in the original string which are replaced or remain unchanged.
+         * @stable ICU 59
+         */
+        int32_t oldLength() const { return oldLength_; }
+
+        /**
+         * The length of the current span in the destination string, which starts at
+         * {@link #destinationIndex}, or in the replacement string, which starts at
+         * {@link #replacementIndex}.
+         *
+         * @return the number of units in the modified string, if hasChange() is true.
+         *         Same as oldLength if hasChange() is false.
+         * @stable ICU 59
+         */
+        int32_t newLength() const { return newLength_; }
+
+        /**
+         * The start index of the current span in the source string; the span has length
+         * {@link #oldLength}.
+         *
+         * @return the current index into the source string
+         * @stable ICU 59
+         */
+        int32_t sourceIndex() const { return srcIndex; }
+
+        /**
+         * The start index of the current span in the replacement string; the span has length
+         * {@link #newLength}. Well-defined only if the current edit is a change edit.
+         *
+         * The *replacement string* is the concatenation of all substrings of the destination
+         * string corresponding to change edits.
+         *
+         * This method is intended to be used together with operations that write only replacement
+         * characters (e.g. operations specifying the \ref U_OMIT_UNCHANGED_TEXT option).
+         * The source string can then be modified in-place.
+         *
+         * @return the current index into the replacement-characters-only string,
+         *         not counting unchanged spans
+         * @stable ICU 59
+         */
+        int32_t replacementIndex() const {
+            // TODO: Throw an exception if we aren't in a change edit?
+            return replIndex;
+        }
+
+        /**
+         * The start index of the current span in the destination string; the span has length
+         * {@link #newLength}.
+         *
+         * @return the current index into the full destination string
+         * @stable ICU 59
+         */
+        int32_t destinationIndex() const { return destIndex; }
+
+#ifndef U_HIDE_INTERNAL_API
+        /**
+         * A string representation of the current edit represented by the iterator for debugging. You
+         * should not depend on the contents of the return string.
+         * @internal
+         */
+        UnicodeString& toString(UnicodeString& appendTo) const;
+#endif  // U_HIDE_INTERNAL_API
+
+    private:
+        friend class Edits;
+
+        Iterator(const uint16_t *a, int32_t len, UBool oc, UBool crs);
+
+        int32_t readLength(int32_t head);
+        void updateNextIndexes();
+        void updatePreviousIndexes();
+        UBool noNext();
+        UBool next(UBool onlyChanges, UErrorCode &errorCode);
+        UBool previous(UErrorCode &errorCode);
+        /** @return -1: error or i<0; 0: found; 1: i>=string length */
+        int32_t findIndex(int32_t i, UBool findSource, UErrorCode &errorCode);
+
+        const uint16_t *array;
+        int32_t index, length;
+        // 0 if we are not within compressed equal-length changes.
+        // Otherwise the number of remaining changes, including the current one.
+        int32_t remaining;
+        UBool onlyChanges_, coarse;
+
+        int8_t dir;  // iteration direction: back(<0), initial(0), forward(>0)
+        UBool changed;
+        int32_t oldLength_, newLength_;
+        int32_t srcIndex, replIndex, destIndex;
+    };
+
+    /**
+     * Returns an Iterator for coarse-grained change edits
+     * (adjacent change edits are treated as one).
+     * Can be used to perform simple string updates.
+     * Skips no-change edits.
+     * @return an Iterator that merges adjacent changes.
+     * @stable ICU 59
+     */
+    Iterator getCoarseChangesIterator() const {
+        return Iterator(array, length, true, true);
+    }
+
+    /**
+     * Returns an Iterator for coarse-grained change and no-change edits
+     * (adjacent change edits are treated as one).
+     * Can be used to perform simple string updates.
+     * Adjacent change edits are treated as one edit.
+     * @return an Iterator that merges adjacent changes.
+     * @stable ICU 59
+     */
+    Iterator getCoarseIterator() const {
+        return Iterator(array, length, false, true);
+    }
+
+    /**
+     * Returns an Iterator for fine-grained change edits
+     * (full granularity of change edits is retained).
+     * Can be used for modifying styled text.
+     * Skips no-change edits.
+     * @return an Iterator that separates adjacent changes.
+     * @stable ICU 59
+     */
+    Iterator getFineChangesIterator() const {
+        return Iterator(array, length, true, false);
+    }
+
+    /**
+     * Returns an Iterator for fine-grained change and no-change edits
+     * (full granularity of change edits is retained).
+     * Can be used for modifying styled text.
+     * @return an Iterator that separates adjacent changes.
+     * @stable ICU 59
+     */
+    Iterator getFineIterator() const {
+        return Iterator(array, length, false, false);
+    }
+
+    /**
+     * Merges the two input Edits and appends the result to this object.
+     *
+     * Consider two string transformations (for example, normalization and case mapping)
+     * where each records Edits in addition to writing an output string.<br>
+     * Edits ab reflect how substrings of input string a
+     * map to substrings of intermediate string b.<br>
+     * Edits bc reflect how substrings of intermediate string b
+     * map to substrings of output string c.<br>
+     * This function merges ab and bc such that the additional edits
+     * recorded in this object reflect how substrings of input string a
+     * map to substrings of output string c.
+     *
+     * If unrelated Edits are passed in where the output string of the first
+     * has a different length than the input string of the second,
+     * then a U_ILLEGAL_ARGUMENT_ERROR is reported.
+     *
+     * @param ab reflects how substrings of input string a
+     *     map to substrings of intermediate string b.
+     * @param bc reflects how substrings of intermediate string b
+     *     map to substrings of output string c.
+     * @param errorCode ICU error code. Its input value must pass the U_SUCCESS() test,
+     *                  or else the function returns immediately. Check for U_FAILURE()
+     *                  on output or use with function chaining. (See User Guide for details.)
+     * @return *this, with the merged edits appended
+     * @stable ICU 60
+     */
+    Edits &mergeAndAppend(const Edits &ab, const Edits &bc, UErrorCode &errorCode);
+
+private:
+    void releaseArray() noexcept;
+    Edits &copyArray(const Edits &other);
+    Edits &moveArray(Edits &src) noexcept;
+
+    void setLastUnit(int32_t last) { array[length - 1] = static_cast<uint16_t>(last); }
+    int32_t lastUnit() const { return length > 0 ? array[length - 1] : 0xffff; }
+
+    void append(int32_t r);
+    UBool growArray();
+
+    static const int32_t STACK_CAPACITY = 100;
+    uint16_t *array;
+    int32_t capacity;
+    int32_t length;
+    int32_t delta;
+    int32_t numChanges;
+    UErrorCode errorCode_;
+    uint16_t stackArray[STACK_CAPACITY];
+};
+
+U_NAMESPACE_END
+
+#endif /* U_SHOW_CPLUSPLUS_API */
+
+#endif  // __EDITS_H__

thirdparty/icu4c/common/unicode/enumset.h

@@ -0,0 +1,69 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+/*
+******************************************************************************
+*
+*   Copyright (C) 2012,2014 International Business Machines
+*   Corporation and others.  All Rights Reserved.
+*
+******************************************************************************
+*/
+
+/**
+ * \file
+ * \brief C++: internal template EnumSet<>
+ */
+
+#ifndef ENUMSET_H
+#define ENUMSET_H
+
+#include "unicode/utypes.h"
+
+#if U_SHOW_CPLUSPLUS_API
+
+U_NAMESPACE_BEGIN
+
+/* Can't use #ifndef U_HIDE_INTERNAL_API for the entire EnumSet class, needed in .h file declarations */
+/**
+ * enum bitset for boolean fields. Similar to Java EnumSet<>. 
+ * Needs to range check. Used for private instance variables.
+ * @internal
+ * \cond
+ */
+template<typename T, uint32_t minValue, uint32_t limitValue>
+class EnumSet {
+public:
+    inline EnumSet() : fBools(0) {}
+    inline EnumSet(const EnumSet<T,minValue,limitValue>& other) : fBools(other.fBools) {}
+    inline ~EnumSet() {}
+#ifndef U_HIDE_INTERNAL_API
+    inline void clear() { fBools=0; }
+    inline void add(T toAdd) { set(toAdd, 1); }
+    inline void remove(T toRemove) { set(toRemove, 0); }
+    inline int32_t contains(T toCheck) const { return get(toCheck); }
+    inline void set(T toSet, int32_t v) { fBools=(fBools&(~flag(toSet)))|(v?(flag(toSet)):0); }
+    inline int32_t get(T toCheck) const { return (fBools & flag(toCheck))?1:0; }
+    inline UBool isValidEnum(T toCheck) const {  return (toCheck>=minValue&&toCheck<limitValue); }
+    inline UBool isValidValue(int32_t v) const { return (v==0||v==1); }
+    inline const EnumSet<T,minValue,limitValue>& operator=(const EnumSet<T,minValue,limitValue>& other) {
+        fBools = other.fBools;
+        return *this;
+    }
+  
+    inline uint32_t getAll() const {
+        return fBools; 
+    }
+#endif /* U_HIDE_INTERNAL_API */
+
+private:
+    inline uint32_t flag(T toCheck) const { return (1<<(toCheck-minValue)); }
+private:
+    uint32_t fBools;
+};
+
+/** \endcond */
+
+U_NAMESPACE_END
+
+#endif /* U_SHOW_CPLUSPLUS_API */
+#endif /* ENUMSET_H */

thirdparty/icu4c/common/unicode/errorcode.h

@@ -0,0 +1,144 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+/*
+*******************************************************************************
+*
+*   Copyright (C) 2009-2011, International Business Machines
+*   Corporation and others.  All Rights Reserved.
+*
+*******************************************************************************
+*   file name:  errorcode.h
+*   encoding:   UTF-8
+*   tab size:   8 (not used)
+*   indentation:4
+*
+*   created on: 2009mar10
+*   created by: Markus W. Scherer
+*/
+
+#ifndef __ERRORCODE_H__
+#define __ERRORCODE_H__
+
+/**
+ * \file 
+ * \brief C++ API: ErrorCode class intended to make it easier to use
+ *                 ICU C and C++ APIs from C++ user code.
+ */
+
+#include "unicode/utypes.h"
+
+#if U_SHOW_CPLUSPLUS_API
+
+#include "unicode/uobject.h"
+
+U_NAMESPACE_BEGIN
+
+/**
+ * Wrapper class for UErrorCode, with conversion operators for direct use
+ * in ICU C and C++ APIs.
+ * Intended to be used as a base class, where a subclass overrides
+ * the handleFailure() function so that it throws an exception,
+ * does an assert(), logs an error, etc.
+ * This is not an abstract base class. This class can be used and instantiated
+ * by itself, although it will be more useful when subclassed.
+ *
+ * Features:
+ * - The constructor initializes the internal UErrorCode to U_ZERO_ERROR,
+ *   removing one common source of errors.
+ * - Same use in C APIs taking a UErrorCode * (pointer)
+ *   and C++ taking UErrorCode & (reference) via conversion operators.
+ * - Possible automatic checking for success when it goes out of scope.
+ *
+ * Note: For automatic checking for success in the destructor, a subclass
+ * must implement such logic in its own destructor because the base class
+ * destructor cannot call a subclass function (like handleFailure()).
+ * The ErrorCode base class destructor does nothing.
+ *
+ * Note also: While it is possible for a destructor to throw an exception,
+ * it is generally unsafe to do so. This means that in a subclass the destructor
+ * and the handleFailure() function may need to take different actions.
+ *
+ * Sample code:
+ * \code
+ *   class IcuErrorCode: public icu::ErrorCode {
+ *   public:
+ *     virtual ~IcuErrorCode() {  // should be defined in .cpp as "key function"
+ *       // Safe because our handleFailure() does not throw exceptions.
+ *       if(isFailure()) { handleFailure(); }
+ *     }
+ *   protected:
+ *     virtual void handleFailure() const {
+ *       log_failure(u_errorName(errorCode));
+ *       exit(errorCode);
+ *     }
+ *   };
+ *   IcuErrorCode error_code;
+ *   UConverter *cnv = ucnv_open("Shift-JIS", error_code);
+ *   length = ucnv_fromUChars(dest, capacity, src, length, error_code);
+ *   ucnv_close(cnv);
+ *   // IcuErrorCode destructor checks for success.
+ * \endcode
+ *
+ * @stable ICU 4.2
+ */
+class U_COMMON_API ErrorCode: public UMemory {
+public:
+    /**
+     * Default constructor. Initializes its UErrorCode to U_ZERO_ERROR.
+     * @stable ICU 4.2
+     */
+    ErrorCode() : errorCode(U_ZERO_ERROR) {}
+    /** Destructor, does nothing. See class documentation for details. @stable ICU 4.2 */
+    virtual ~ErrorCode();
+    /** Conversion operator, returns a reference. @stable ICU 4.2 */
+    operator UErrorCode & () { return errorCode; }
+    /** Conversion operator, returns a pointer. @stable ICU 4.2 */
+    operator UErrorCode * () { return &errorCode; }
+    /** Tests for U_SUCCESS(). @stable ICU 4.2 */
+    UBool isSuccess() const { return U_SUCCESS(errorCode); }
+    /** Tests for U_FAILURE(). @stable ICU 4.2 */
+    UBool isFailure() const { return U_FAILURE(errorCode); }
+    /** Returns the UErrorCode value. @stable ICU 4.2 */
+    UErrorCode get() const { return errorCode; }
+    /** Sets the UErrorCode value. @stable ICU 4.2 */
+    void set(UErrorCode value) { errorCode=value; }
+    /** Returns the UErrorCode value and resets it to U_ZERO_ERROR. @stable ICU 4.2 */
+    UErrorCode reset();
+    /**
+     * Asserts isSuccess().
+     * In other words, this method checks for a failure code,
+     * and the base class handles it like this:
+     * \code
+     *   if(isFailure()) { handleFailure(); }
+     * \endcode
+     * @stable ICU 4.4
+     */
+    void assertSuccess() const;
+    /**
+     * Return a string for the UErrorCode value.
+     * The string will be the same as the name of the error code constant
+     * in the UErrorCode enum.
+     * @stable ICU 4.4
+     */
+    const char* errorName() const;
+
+protected:
+    /**
+     * Internal UErrorCode, accessible to subclasses.
+     * @stable ICU 4.2
+     */
+    UErrorCode errorCode;
+    /**
+     * Called by assertSuccess() if isFailure() is true.
+     * A subclass should override this function to deal with a failure code:
+     * Throw an exception, log an error, terminate the program, or similar.
+     * @stable ICU 4.2
+     */
+    virtual void handleFailure() const {}
+};
+
+U_NAMESPACE_END
+
+#endif /* U_SHOW_CPLUSPLUS_API */
+
+#endif  // __ERRORCODE_H__

thirdparty/icu4c/common/unicode/filteredbrk.h

@@ -0,0 +1,152 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+/*
+********************************************************************************
+*   Copyright (C) 1997-2015, International Business Machines
+*   Corporation and others.  All Rights Reserved.
+********************************************************************************
+*/
+
+#ifndef FILTEREDBRK_H
+#define FILTEREDBRK_H
+
+#include "unicode/utypes.h"
+
+#if U_SHOW_CPLUSPLUS_API
+
+#include "unicode/brkiter.h"
+
+#if !UCONFIG_NO_BREAK_ITERATION && !UCONFIG_NO_FILTERED_BREAK_ITERATION
+
+U_NAMESPACE_BEGIN
+
+/**
+ * \file
+ * \brief C++ API: FilteredBreakIteratorBuilder
+ */
+
+/**
+ * The BreakIteratorFilter is used to modify the behavior of a BreakIterator
+ *  by constructing a new BreakIterator which suppresses certain segment boundaries.
+ *  See  http://www.unicode.org/reports/tr35/tr35-general.html#Segmentation_Exceptions .
+ *  For example, a typical English Sentence Break Iterator would break on the space
+ *  in the string "Mr. Smith" (resulting in two segments),
+ *  but with "Mr." as an exception, a filtered break iterator
+ *  would consider the string "Mr. Smith" to be a single segment.
+ *
+ * @stable ICU 56
+ */
+class U_COMMON_API FilteredBreakIteratorBuilder : public UObject {
+ public:
+  /**
+   *  destructor.
+   * @stable ICU 56
+   */
+  virtual ~FilteredBreakIteratorBuilder();
+
+  /**
+   * Construct a FilteredBreakIteratorBuilder based on rules in a locale.
+   * The rules are taken from CLDR exception data for the locale,
+   *  see http://www.unicode.org/reports/tr35/tr35-general.html#Segmentation_Exceptions
+   *  This is the equivalent of calling createInstance(UErrorCode&)
+   *    and then repeatedly calling addNoBreakAfter(...) with the contents
+   *    of the CLDR exception data.
+   * @param where the locale.
+   * @param status The error code.
+   * @return the new builder
+   * @stable ICU 56
+   */
+  static FilteredBreakIteratorBuilder *createInstance(const Locale& where, UErrorCode& status);
+
+#ifndef U_HIDE_DEPRECATED_API
+  /**
+   * This function has been deprecated in favor of createEmptyInstance, which has
+   * identical behavior.
+   * @param status The error code.
+   * @return the new builder
+   * @deprecated ICU 60 use createEmptyInstance instead
+   * @see createEmptyInstance()
+   */
+  static FilteredBreakIteratorBuilder *createInstance(UErrorCode &status);
+#endif  /* U_HIDE_DEPRECATED_API */
+
+  /**
+   * Construct an empty FilteredBreakIteratorBuilder.
+   * In this state, it will not suppress any segment boundaries.
+   * @param status The error code.
+   * @return the new builder
+   * @stable ICU 60
+   */
+  static FilteredBreakIteratorBuilder *createEmptyInstance(UErrorCode &status);
+
+  /**
+   * Suppress a certain string from being the end of a segment.
+   * For example, suppressing "Mr.", then segments ending in "Mr." will not be returned
+   * by the iterator.
+   * @param string the string to suppress, such as "Mr."
+   * @param status error code
+   * @return returns true if the string was not present and now added,
+   * false if the call was a no-op because the string was already being suppressed.
+   * @stable ICU 56
+   */
+  virtual UBool suppressBreakAfter(const UnicodeString& string, UErrorCode& status) = 0;
+
+  /**
+   * Stop suppressing a certain string from being the end of the segment.
+   * This function does not create any new segment boundaries, but only serves to un-do
+   * the effect of earlier calls to suppressBreakAfter, or to un-do the effect of
+   * locale data which may be suppressing certain strings.
+   * @param string the exception to remove
+   * @param status error code
+   * @return returns true if the string was present and now removed,
+   * false if the call was a no-op because the string was not being suppressed.
+   * @stable ICU 56
+   */
+  virtual UBool unsuppressBreakAfter(const UnicodeString& string, UErrorCode& status) = 0;
+
+#ifndef U_FORCE_HIDE_DEPRECATED_API
+  /**
+   * This function has been deprecated in favor of wrapIteratorWithFilter()
+   * The behavior is identical.
+   * @param adoptBreakIterator the break iterator to adopt
+   * @param status error code
+   * @return the new BreakIterator, owned by the caller.
+   * @deprecated ICU 60 use wrapIteratorWithFilter() instead
+   * @see wrapBreakIteratorWithFilter()
+   */
+  virtual BreakIterator *build(BreakIterator* adoptBreakIterator, UErrorCode& status) = 0;
+#endif  // U_FORCE_HIDE_DEPRECATED_API
+
+  /**
+   * Wrap (adopt) an existing break iterator in a new filtered instance.
+   * The resulting BreakIterator is owned by the caller.
+   * The BreakIteratorFilter may be destroyed before the BreakIterator is destroyed.
+   * Note that the adoptBreakIterator is adopted by the new BreakIterator
+   * and should no longer be used by the caller.
+   * The FilteredBreakIteratorBuilder may be reused.
+   * This function is an alias for build()
+   * @param adoptBreakIterator the break iterator to adopt
+   * @param status error code
+   * @return the new BreakIterator, owned by the caller.
+   * @stable ICU 60
+   */
+  inline BreakIterator *wrapIteratorWithFilter(BreakIterator* adoptBreakIterator, UErrorCode& status) {
+    return build(adoptBreakIterator, status);
+  }
+
+ protected:
+  /**
+   * For subclass use
+   * @stable ICU 56
+   */
+  FilteredBreakIteratorBuilder();
+};
+
+
+U_NAMESPACE_END
+
+#endif // #if !UCONFIG_NO_BREAK_ITERATION && !UCONFIG_NO_FILTERED_BREAK_ITERATION
+
+#endif /* U_SHOW_CPLUSPLUS_API */
+
+#endif // #ifndef FILTEREDBRK_H

thirdparty/icu4c/common/unicode/icudataver.h

@@ -0,0 +1,43 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+/*
+******************************************************************************
+*
+*   Copyright (C) 2009-2013, International Business Machines
+*   Corporation and others.  All Rights Reserved.
+*
+******************************************************************************
+*/
+
+
+/**
+ * \file
+ * \brief C API: access to ICU Data Version number
+ */
+
+#ifndef __ICU_DATA_VER_H__
+#define __ICU_DATA_VER_H__
+
+#include "unicode/utypes.h"
+
+/**
+ * @stable ICU 49
+ */
+#define U_ICU_VERSION_BUNDLE "icuver"
+
+/**
+ * @stable ICU 49
+ */
+#define U_ICU_DATA_KEY "DataVersion"
+
+/**
+ * Retrieves the data version from icuver and stores it in dataVersionFillin.
+ * 
+ * @param dataVersionFillin icuver data version information to be filled in if not-null
+ * @param status stores the error code from the calls to resource bundle
+ * 
+ * @stable ICU 49
+ */
+U_CAPI void U_EXPORT2 u_getDataVersion(UVersionInfo dataVersionFillin, UErrorCode *status);
+
+#endif

thirdparty/icu4c/common/unicode/icuplug.h

@@ -0,0 +1,391 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+/*
+******************************************************************************
+*
+*   Copyright (C) 2009-2015, International Business Machines
+*   Corporation and others.  All Rights Reserved.
+*
+******************************************************************************
+*
+*  FILE NAME : icuplug.h
+*
+*   Date         Name        Description
+*   10/29/2009   sl          New.
+******************************************************************************
+*/
+
+/**
+ * \file
+ * \brief C API: ICU Plugin API 
+ *
+ * <h2>C API: ICU Plugin API</h2>
+ *
+ * <p>C API allowing run-time loadable modules that extend or modify ICU functionality.</p>
+ *
+ * <h3>Loading and Configuration</h3>
+ *
+ * <p>At ICU startup time, the environment variable "ICU_PLUGINS" will be 
+ * queried for a directory name.  If it is not set, the preprocessor symbol 
+ * "DEFAULT_ICU_PLUGINS" will be checked for a default value.</p>
+ *
+ * <p>Within the above-named directory, the file  "icuplugins##.txt" will be 
+ * opened, if present, where ## is the major+minor number of the currently 
+ * running ICU (such as, 44 for ICU 4.4, thus icuplugins44.txt)</p>
+ *
+ * <p>The configuration file has this format:</p>
+ *
+ * <ul>
+ * <li>Hash (#) begins a comment line</li>
+ * 
+ * <li>Non-comment lines have two or three components:
+ * LIBRARYNAME     ENTRYPOINT     [ CONFIGURATION .. ]</li>
+ *
+ * <li>Tabs or spaces separate the three items.</li>
+ *
+ * <li>LIBRARYNAME is the name of a shared library, either a short name if 
+ * it is on the loader path,  or a full pathname.</li>
+ *
+ * <li>ENTRYPOINT is the short (undecorated) symbol name of the plugin's 
+ * entrypoint, as above.</li>
+ *
+ * <li>CONFIGURATION is the entire rest of the line . It's passed as-is to 
+ * the plugin.</li>
+ * </ul>
+ *
+ * <p>An example configuration file is, in its entirety:</p>
+ *
+ * \code
+ * # this is icuplugins44.txt
+ * testplug.dll    myPlugin        hello=world
+ * \endcode
+ * <p>Plugins are categorized as "high" or "low" level.  Low level are those 
+ * which must be run BEFORE high level plugins, and before any operations 
+ * which cause ICU to be 'initialized'.  If a plugin is low level but 
+ * causes ICU to allocate memory or become initialized, that plugin is said 
+ * to cause a 'level change'. </p>
+ *
+ * <p>At load time, ICU first queries all plugins to determine their level, 
+ * then loads all 'low' plugins first, and then loads all 'high' plugins.  
+ * Plugins are otherwise loaded in the order listed in the configuration file.</p>
+ * 
+ * <h3>Implementing a Plugin</h3>
+ * \code
+ * U_CAPI UPlugTokenReturn U_EXPORT2 
+ * myPlugin (UPlugData *plug, UPlugReason reason, UErrorCode *status) {
+ *   if(reason==UPLUG_REASON_QUERY) {
+ *      uplug_setPlugName(plug, "Simple Plugin");
+ *      uplug_setPlugLevel(plug, UPLUG_LEVEL_HIGH);
+ *    } else if(reason==UPLUG_REASON_LOAD) {
+ *       ... Set up some ICU things here.... 
+ *    } else if(reason==UPLUG_REASON_UNLOAD) {
+ *       ... unload, clean up ...
+ *    }
+ *   return UPLUG_TOKEN;
+ *  }
+ * \endcode
+ *
+ * <p>The UPlugData*  is an opaque pointer to the plugin-specific data, and is 
+ * used in all other API calls.</p>
+ *
+ * <p>The API contract is:</p>
+ * <ol><li>The plugin MUST always return UPLUG_TOKEN as a return value- to 
+ * indicate that it is a valid plugin.</li>
+ *
+ * <li>When the 'reason' parameter is set to UPLUG_REASON_QUERY,  the 
+ * plugin MUST call uplug_setPlugLevel() to indicate whether it is a high 
+ * level or low level plugin.</li>
+ *
+ * <li>When the 'reason' parameter is UPLUG_REASON_QUERY, the plugin 
+ * SHOULD call uplug_setPlugName to indicate a human readable plugin name.</li></ol>
+ * 
+ *
+ * \internal ICU 4.4 Technology Preview
+ */
+
+
+#ifndef ICUPLUG_H
+#define ICUPLUG_H
+
+#include "unicode/utypes.h"
+
+
+#if UCONFIG_ENABLE_PLUGINS || defined(U_IN_DOXYGEN)
+
+
+
+/* === Basic types === */
+
+#ifndef U_HIDE_INTERNAL_API
+struct UPlugData;
+/**
+ * @{
+ * Typedef for opaque structure passed to/from a plugin. 
+ * Use the APIs to access it.
+ * @internal ICU 4.4 Technology Preview
+ */
+typedef struct UPlugData UPlugData;
+
+/** @} */
+
+/**
+ * Random Token to identify a valid ICU plugin. Plugins must return this 
+ * from the entrypoint.
+ * @internal ICU 4.4 Technology Preview
+ */
+#define UPLUG_TOKEN 0x54762486
+
+/**
+ * Max width of names, symbols, and configuration strings
+ * @internal ICU 4.4 Technology Preview
+ */
+#define UPLUG_NAME_MAX              100
+
+
+/**
+ * Return value from a plugin entrypoint. 
+ * Must always be set to UPLUG_TOKEN
+ * @see UPLUG_TOKEN
+ * @internal ICU 4.4 Technology Preview
+ */
+typedef uint32_t UPlugTokenReturn;
+
+/**
+ * Reason code for the entrypoint's call
+ * @internal ICU 4.4 Technology Preview
+ */
+typedef enum {
+    UPLUG_REASON_QUERY = 0,     /**< The plugin is being queried for info. **/
+    UPLUG_REASON_LOAD = 1,     /**< The plugin is being loaded. **/
+    UPLUG_REASON_UNLOAD = 2,   /**< The plugin is being unloaded. **/
+    /**
+     * Number of known reasons.
+     * @internal The numeric value may change over time, see ICU ticket #12420.
+     */
+    UPLUG_REASON_COUNT
+} UPlugReason;
+
+
+/**
+ * Level of plugin loading
+ *     INITIAL:  UNKNOWN
+ *       QUERY:   INVALID ->  { LOW | HIGH }
+ *     ERR -> INVALID
+ * @internal ICU 4.4 Technology Preview
+ */
+typedef enum {
+    UPLUG_LEVEL_INVALID = 0,     /**< The plugin is invalid, hasn't called uplug_setLevel, or can't load. **/
+    UPLUG_LEVEL_UNKNOWN = 1,     /**< The plugin is waiting to be installed. **/
+    UPLUG_LEVEL_LOW     = 2,     /**< The plugin must be called before u_init completes **/
+    UPLUG_LEVEL_HIGH    = 3,     /**< The plugin can run at any time. **/
+    /**
+     * Number of known levels.
+     * @internal The numeric value may change over time, see ICU ticket #12420.
+     */
+    UPLUG_LEVEL_COUNT
+} UPlugLevel;
+
+/**
+ * Entrypoint for an ICU plugin.
+ * @param plug the UPlugData handle.
+ * @param reason the reason code for the entrypoint's call.
+ * @param status Standard ICU error code. Its input value must
+ *               pass the U_SUCCESS() test, or else the function returns
+ *               immediately. Check for U_FAILURE() on output or use with
+ *               function chaining. (See User Guide for details.)
+ * @return A valid plugin must return UPLUG_TOKEN
+ * @internal ICU 4.4 Technology Preview
+ */
+typedef UPlugTokenReturn (U_EXPORT2 UPlugEntrypoint) (
+                  UPlugData *plug,
+                  UPlugReason reason,
+                  UErrorCode *status);
+
+/* === Needed for Implementing === */
+
+/**
+ * Request that this plugin not be unloaded at cleanup time.
+ * This is appropriate for plugins which cannot be cleaned up.
+ * @see u_cleanup()
+ * @param plug plugin
+ * @param dontUnload  set true if this plugin can't be unloaded
+ * @internal ICU 4.4 Technology Preview
+ */
+U_CAPI void U_EXPORT2 
+uplug_setPlugNoUnload(UPlugData *plug, UBool dontUnload);
+
+/**
+ * Set the level of this plugin.
+ * @param plug plugin data handle
+ * @param level the level of this plugin
+ * @internal ICU 4.4 Technology Preview
+ */
+U_CAPI void U_EXPORT2
+uplug_setPlugLevel(UPlugData *plug, UPlugLevel level);
+
+/**
+ * Get the level of this plugin.
+ * @param plug plugin data handle
+ * @return the level of this plugin
+ * @internal ICU 4.4 Technology Preview
+ */
+U_CAPI UPlugLevel U_EXPORT2
+uplug_getPlugLevel(UPlugData *plug);
+
+/**
+ * Get the lowest level of plug which can currently load.
+ * For example, if UPLUG_LEVEL_LOW is returned, then low level plugins may load
+ * if UPLUG_LEVEL_HIGH is returned, then only high level plugins may load.
+ * @return the lowest level of plug which can currently load
+ * @internal ICU 4.4 Technology Preview
+ */
+U_CAPI UPlugLevel U_EXPORT2
+uplug_getCurrentLevel(void);
+
+
+/**
+ * Get plug load status
+ * @return The error code of this plugin's load attempt.
+ * @internal ICU 4.4 Technology Preview
+ */
+U_CAPI UErrorCode U_EXPORT2
+uplug_getPlugLoadStatus(UPlugData *plug); 
+
+/**
+ * Set the human-readable name of this plugin.
+ * @param plug plugin data handle
+ * @param name the name of this plugin. The first UPLUG_NAME_MAX characters willi be copied into a new buffer.
+ * @internal ICU 4.4 Technology Preview
+ */
+U_CAPI void U_EXPORT2
+uplug_setPlugName(UPlugData *plug, const char *name);
+
+/**
+ * Get the human-readable name of this plugin.
+ * @param plug plugin data handle
+ * @return the name of this plugin
+ * @internal ICU 4.4 Technology Preview
+ */
+U_CAPI const char * U_EXPORT2
+uplug_getPlugName(UPlugData *plug);
+
+/**
+ * Return the symbol name for this plugin, if known.
+ * @param plug plugin data handle
+ * @return the symbol name, or NULL
+ * @internal ICU 4.4 Technology Preview
+ */
+U_CAPI const char * U_EXPORT2
+uplug_getSymbolName(UPlugData *plug);
+
+/**
+ * Return the library name for this plugin, if known.
+ * @param plug plugin data handle
+ * @param status error code
+ * @return the library name, or NULL
+ * @internal ICU 4.4 Technology Preview
+ */
+U_CAPI const char * U_EXPORT2
+uplug_getLibraryName(UPlugData *plug, UErrorCode *status);
+
+/**
+ * Return the library used for this plugin, if known.
+ * Plugins could use this to load data out of their 
+ * @param plug plugin data handle
+ * @return the library, or NULL
+ * @internal ICU 4.4 Technology Preview
+ */
+U_CAPI void * U_EXPORT2
+uplug_getLibrary(UPlugData *plug);
+
+/**
+ * Return the plugin-specific context data.
+ * @param plug plugin data handle
+ * @return the context, or NULL if not set
+ * @internal ICU 4.4 Technology Preview
+ */
+U_CAPI void * U_EXPORT2
+uplug_getContext(UPlugData *plug);
+
+/**
+ * Set the plugin-specific context data.
+ * @param plug plugin data handle
+ * @param context new context to set
+ * @internal ICU 4.4 Technology Preview
+ */
+U_CAPI void U_EXPORT2
+uplug_setContext(UPlugData *plug, void *context);
+
+
+/**
+ * Get the configuration string, if available.
+ * The string is in the platform default codepage.
+ * @param plug plugin data handle
+ * @return configuration string, or else null.
+ * @internal ICU 4.4 Technology Preview
+ */
+U_CAPI const char * U_EXPORT2
+uplug_getConfiguration(UPlugData *plug);
+
+/**
+ * Return all currently installed plugins, from newest to oldest
+ * Usage Example:
+ * \code
+ *    UPlugData *plug = NULL;
+ *    while(plug=uplug_nextPlug(plug)) {
+ *        ... do something with 'plug' ...
+ *    }
+ * \endcode
+ * Not thread safe- do not call while plugs are added or removed.
+ * @param prior pass in 'NULL' to get the first (most recent) plug, 
+ *  otherwise pass the value returned on a prior call to uplug_nextPlug
+ * @return the next oldest plugin, or NULL if no more.
+ * @internal ICU 4.4 Technology Preview
+ */
+U_CAPI UPlugData* U_EXPORT2
+uplug_nextPlug(UPlugData *prior);
+
+/**
+ * Inject a plugin as if it were loaded from a library.
+ * This is useful for testing plugins. 
+ * Note that it will have a 'NULL' library pointer associated
+ * with it, and therefore no llibrary will be closed at cleanup time.
+ * Low level plugins may not be able to load, as ordering can't be enforced.
+ * @param entrypoint entrypoint to install
+ * @param config user specified configuration string, if available, or NULL.
+ * @param status error result
+ * @return the new UPlugData associated with this plugin, or NULL if error.
+ * @internal ICU 4.4 Technology Preview
+ */
+U_CAPI UPlugData* U_EXPORT2
+uplug_loadPlugFromEntrypoint(UPlugEntrypoint *entrypoint, const char *config, UErrorCode *status);
+
+
+/**
+ * Inject a plugin from a library, as if the information came from a config file.
+ * Low level plugins may not be able to load, and ordering can't be enforced.
+ * @param libName DLL name to load
+ * @param sym symbol of plugin (UPlugEntrypoint function)
+ * @param config configuration string, or NULL
+ * @param status error result
+ * @return the new UPlugData associated with this plugin, or NULL if error.
+ * @internal ICU 4.4 Technology Preview
+ */
+U_CAPI UPlugData* U_EXPORT2
+uplug_loadPlugFromLibrary(const char *libName, const char *sym, const char *config, UErrorCode *status);
+
+/**
+ * Remove a plugin. 
+ * Will request the plugin to be unloaded, and close the library if needed
+ * @param plug plugin handle to close
+ * @param status error result
+ * @internal ICU 4.4 Technology Preview
+ */
+U_CAPI void U_EXPORT2
+uplug_removePlug(UPlugData *plug, UErrorCode *status);
+#endif  /* U_HIDE_INTERNAL_API */
+
+#endif /* UCONFIG_ENABLE_PLUGINS */
+
+#endif /* _ICUPLUG */
+

thirdparty/icu4c/common/unicode/idna.h

@@ -0,0 +1,333 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+/*
+*******************************************************************************
+*   Copyright (C) 2010-2012, International Business Machines
+*   Corporation and others.  All Rights Reserved.
+*******************************************************************************
+*   file name:  idna.h
+*   encoding:   UTF-8
+*   tab size:   8 (not used)
+*   indentation:4
+*
+*   created on: 2010mar05
+*   created by: Markus W. Scherer
+*/
+
+#ifndef __IDNA_H__
+#define __IDNA_H__
+
+/**
+ * \file
+ * \brief C++ API: Internationalizing Domain Names in Applications (IDNA)
+ */
+
+#include "unicode/utypes.h"
+
+#if U_SHOW_CPLUSPLUS_API
+
+#if !UCONFIG_NO_IDNA
+
+#include "unicode/bytestream.h"
+#include "unicode/stringpiece.h"
+#include "unicode/uidna.h"
+#include "unicode/unistr.h"
+
+U_NAMESPACE_BEGIN
+
+class IDNAInfo;
+
+/**
+ * Abstract base class for IDNA processing.
+ * See http://www.unicode.org/reports/tr46/
+ * and http://www.ietf.org/rfc/rfc3490.txt
+ *
+ * The IDNA class is not intended for public subclassing.
+ *
+ * This C++ API currently only implements UTS #46.
+ * The uidna.h C API implements both UTS #46 (functions using UIDNA service object)
+ * and IDNA2003 (functions that do not use a service object).
+ * @stable ICU 4.6
+ */
+class U_COMMON_API IDNA : public UObject {
+public:
+    /**
+     * Destructor.
+     * @stable ICU 4.6
+     */
+    ~IDNA();
+
+    /**
+     * Returns an IDNA instance which implements UTS #46.
+     * Returns an unmodifiable instance, owned by the caller.
+     * Cache it for multiple operations, and delete it when done.
+     * The instance is thread-safe, that is, it can be used concurrently.
+     *
+     * UTS #46 defines Unicode IDNA Compatibility Processing,
+     * updated to the latest version of Unicode and compatible with both
+     * IDNA2003 and IDNA2008.
+     *
+     * The worker functions use transitional processing, including deviation mappings,
+     * unless UIDNA_NONTRANSITIONAL_TO_ASCII or UIDNA_NONTRANSITIONAL_TO_UNICODE
+     * is used in which case the deviation characters are passed through without change.
+     * <b>Unicode 15.1 UTS #46 deprecated transitional processing.</b>
+     *
+     * Disallowed characters are mapped to U+FFFD.
+     *
+     * For available options see the uidna.h header.
+     * Operations with the UTS #46 instance do not support the
+     * UIDNA_ALLOW_UNASSIGNED option.
+     *
+     * By default, the UTS #46 implementation allows all ASCII characters (as valid or mapped).
+     * When the UIDNA_USE_STD3_RULES option is used, ASCII characters other than
+     * letters, digits, hyphen (LDH) and dot/full stop are disallowed and mapped to U+FFFD.
+     *
+     * @param options Bit set to modify the processing and error checking.
+     *                These should include UIDNA_DEFAULT, or
+     *                UIDNA_NONTRANSITIONAL_TO_ASCII | UIDNA_NONTRANSITIONAL_TO_UNICODE.
+     *                See option bit set values in uidna.h.
+     * @param errorCode Standard ICU error code. Its input value must
+     *                  pass the U_SUCCESS() test, or else the function returns
+     *                  immediately. Check for U_FAILURE() on output or use with
+     *                  function chaining. (See User Guide for details.)
+     * @return the UTS #46 IDNA instance, if successful
+     * @stable ICU 4.6
+     */
+    static IDNA *
+    createUTS46Instance(uint32_t options, UErrorCode &errorCode);
+
+    /**
+     * Converts a single domain name label into its ASCII form for DNS lookup.
+     * If any processing step fails, then info.hasErrors() will be true and
+     * the result might not be an ASCII string.
+     * The label might be modified according to the types of errors.
+     * Labels with severe errors will be left in (or turned into) their Unicode form.
+     *
+     * The UErrorCode indicates an error only in exceptional cases,
+     * such as a U_MEMORY_ALLOCATION_ERROR.
+     *
+     * @param label Input domain name label
+     * @param dest Destination string object
+     * @param info Output container of IDNA processing details.
+     * @param errorCode Standard ICU error code. Its input value must
+     *                  pass the U_SUCCESS() test, or else the function returns
+     *                  immediately. Check for U_FAILURE() on output or use with
+     *                  function chaining. (See User Guide for details.)
+     * @return dest
+     * @stable ICU 4.6
+     */
+    virtual UnicodeString &
+    labelToASCII(const UnicodeString &label, UnicodeString &dest,
+                 IDNAInfo &info, UErrorCode &errorCode) const = 0;
+
+    /**
+     * Converts a single domain name label into its Unicode form for human-readable display.
+     * If any processing step fails, then info.hasErrors() will be true.
+     * The label might be modified according to the types of errors.
+     *
+     * The UErrorCode indicates an error only in exceptional cases,
+     * such as a U_MEMORY_ALLOCATION_ERROR.
+     *
+     * @param label Input domain name label
+     * @param dest Destination string object
+     * @param info Output container of IDNA processing details.
+     * @param errorCode Standard ICU error code. Its input value must
+     *                  pass the U_SUCCESS() test, or else the function returns
+     *                  immediately. Check for U_FAILURE() on output or use with
+     *                  function chaining. (See User Guide for details.)
+     * @return dest
+     * @stable ICU 4.6
+     */
+    virtual UnicodeString &
+    labelToUnicode(const UnicodeString &label, UnicodeString &dest,
+                   IDNAInfo &info, UErrorCode &errorCode) const = 0;
+
+    /**
+     * Converts a whole domain name into its ASCII form for DNS lookup.
+     * If any processing step fails, then info.hasErrors() will be true and
+     * the result might not be an ASCII string.
+     * The domain name might be modified according to the types of errors.
+     * Labels with severe errors will be left in (or turned into) their Unicode form.
+     *
+     * The UErrorCode indicates an error only in exceptional cases,
+     * such as a U_MEMORY_ALLOCATION_ERROR.
+     *
+     * @param name Input domain name
+     * @param dest Destination string object
+     * @param info Output container of IDNA processing details.
+     * @param errorCode Standard ICU error code. Its input value must
+     *                  pass the U_SUCCESS() test, or else the function returns
+     *                  immediately. Check for U_FAILURE() on output or use with
+     *                  function chaining. (See User Guide for details.)
+     * @return dest
+     * @stable ICU 4.6
+     */
+    virtual UnicodeString &
+    nameToASCII(const UnicodeString &name, UnicodeString &dest,
+                IDNAInfo &info, UErrorCode &errorCode) const = 0;
+
+    /**
+     * Converts a whole domain name into its Unicode form for human-readable display.
+     * If any processing step fails, then info.hasErrors() will be true.
+     * The domain name might be modified according to the types of errors.
+     *
+     * The UErrorCode indicates an error only in exceptional cases,
+     * such as a U_MEMORY_ALLOCATION_ERROR.
+     *
+     * @param name Input domain name
+     * @param dest Destination string object
+     * @param info Output container of IDNA processing details.
+     * @param errorCode Standard ICU error code. Its input value must
+     *                  pass the U_SUCCESS() test, or else the function returns
+     *                  immediately. Check for U_FAILURE() on output or use with
+     *                  function chaining. (See User Guide for details.)
+     * @return dest
+     * @stable ICU 4.6
+     */
+    virtual UnicodeString &
+    nameToUnicode(const UnicodeString &name, UnicodeString &dest,
+                  IDNAInfo &info, UErrorCode &errorCode) const = 0;
+
+    // UTF-8 versions of the processing methods ---------------------------- ***
+
+    /**
+     * Converts a single domain name label into its ASCII form for DNS lookup.
+     * UTF-8 version of labelToASCII(), same behavior.
+     *
+     * @param label Input domain name label
+     * @param dest Destination byte sink; Flush()ed if successful
+     * @param info Output container of IDNA processing details.
+     * @param errorCode Standard ICU error code. Its input value must
+     *                  pass the U_SUCCESS() test, or else the function returns
+     *                  immediately. Check for U_FAILURE() on output or use with
+     *                  function chaining. (See User Guide for details.)
+     * @return dest
+     * @stable ICU 4.6
+     */
+    virtual void
+    labelToASCII_UTF8(StringPiece label, ByteSink &dest,
+                      IDNAInfo &info, UErrorCode &errorCode) const;
+
+    /**
+     * Converts a single domain name label into its Unicode form for human-readable display.
+     * UTF-8 version of labelToUnicode(), same behavior.
+     *
+     * @param label Input domain name label
+     * @param dest Destination byte sink; Flush()ed if successful
+     * @param info Output container of IDNA processing details.
+     * @param errorCode Standard ICU error code. Its input value must
+     *                  pass the U_SUCCESS() test, or else the function returns
+     *                  immediately. Check for U_FAILURE() on output or use with
+     *                  function chaining. (See User Guide for details.)
+     * @return dest
+     * @stable ICU 4.6
+     */
+    virtual void
+    labelToUnicodeUTF8(StringPiece label, ByteSink &dest,
+                       IDNAInfo &info, UErrorCode &errorCode) const;
+
+    /**
+     * Converts a whole domain name into its ASCII form for DNS lookup.
+     * UTF-8 version of nameToASCII(), same behavior.
+     *
+     * @param name Input domain name
+     * @param dest Destination byte sink; Flush()ed if successful
+     * @param info Output container of IDNA processing details.
+     * @param errorCode Standard ICU error code. Its input value must
+     *                  pass the U_SUCCESS() test, or else the function returns
+     *                  immediately. Check for U_FAILURE() on output or use with
+     *                  function chaining. (See User Guide for details.)
+     * @return dest
+     * @stable ICU 4.6
+     */
+    virtual void
+    nameToASCII_UTF8(StringPiece name, ByteSink &dest,
+                     IDNAInfo &info, UErrorCode &errorCode) const;
+
+    /**
+     * Converts a whole domain name into its Unicode form for human-readable display.
+     * UTF-8 version of nameToUnicode(), same behavior.
+     *
+     * @param name Input domain name
+     * @param dest Destination byte sink; Flush()ed if successful
+     * @param info Output container of IDNA processing details.
+     * @param errorCode Standard ICU error code. Its input value must
+     *                  pass the U_SUCCESS() test, or else the function returns
+     *                  immediately. Check for U_FAILURE() on output or use with
+     *                  function chaining. (See User Guide for details.)
+     * @return dest
+     * @stable ICU 4.6
+     */
+    virtual void
+    nameToUnicodeUTF8(StringPiece name, ByteSink &dest,
+                      IDNAInfo &info, UErrorCode &errorCode) const;
+};
+
+class UTS46;
+
+/**
+ * Output container for IDNA processing errors.
+ * The IDNAInfo class is not suitable for subclassing.
+ * @stable ICU 4.6
+ */
+class U_COMMON_API IDNAInfo : public UMemory {
+public:
+    /**
+     * Constructor for stack allocation.
+     * @stable ICU 4.6
+     */
+    IDNAInfo() : errors(0), labelErrors(0), isTransDiff(false), isBiDi(false), isOkBiDi(true) {}
+    /**
+     * Were there IDNA processing errors?
+     * @return true if there were processing errors
+     * @stable ICU 4.6
+     */
+    UBool hasErrors() const { return errors!=0; }
+    /**
+     * Returns a bit set indicating IDNA processing errors.
+     * See UIDNA_ERROR_... constants in uidna.h.
+     * @return bit set of processing errors
+     * @stable ICU 4.6
+     */
+    uint32_t getErrors() const { return errors; }
+    /**
+     * Returns true if transitional and nontransitional processing produce different results.
+     * This is the case when the input label or domain name contains
+     * one or more deviation characters outside a Punycode label (see UTS #46).
+     * <ul>
+     * <li>With nontransitional processing, such characters are
+     * copied to the destination string.
+     * <li>With transitional processing, such characters are
+     * mapped (sharp s/sigma) or removed (joiner/nonjoiner).
+     * </ul>
+     * @return true if transitional and nontransitional processing produce different results
+     * @stable ICU 4.6
+     */
+    UBool isTransitionalDifferent() const { return isTransDiff; }
+
+private:
+    friend class UTS46;
+
+    IDNAInfo(const IDNAInfo &other) = delete;  // no copying
+    IDNAInfo &operator=(const IDNAInfo &other) = delete;  // no copying
+
+    void reset() {
+        errors=labelErrors=0;
+        isTransDiff=false;
+        isBiDi=false;
+        isOkBiDi=true;
+    }
+
+    uint32_t errors, labelErrors;
+    UBool isTransDiff;
+    UBool isBiDi;
+    UBool isOkBiDi;
+};
+
+U_NAMESPACE_END
+
+#endif  // UCONFIG_NO_IDNA
+
+#endif /* U_SHOW_CPLUSPLUS_API */
+
+#endif  // __IDNA_H__

thirdparty/icu4c/common/unicode/localebuilder.h

@@ -0,0 +1,309 @@
+// © 2018 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+#ifndef __LOCALEBUILDER_H__
+#define __LOCALEBUILDER_H__
+
+#include "unicode/utypes.h"
+
+#if U_SHOW_CPLUSPLUS_API
+
+#include "unicode/locid.h"
+#include "unicode/localematcher.h"
+#include "unicode/stringpiece.h"
+#include "unicode/uobject.h"
+
+/**
+ * \file
+ * \brief C++ API: Builder API for Locale
+ */
+
+U_NAMESPACE_BEGIN
+class CharString;
+
+/**
+ * <code>LocaleBuilder</code> is used to build instances of <code>Locale</code>
+ * from values configured by the setters.  Unlike the <code>Locale</code>
+ * constructors, the <code>LocaleBuilder</code> checks if a value configured by a
+ * setter satisfies the syntax requirements defined by the <code>Locale</code>
+ * class.  A <code>Locale</code> object created by a <code>LocaleBuilder</code> is
+ * well-formed and can be transformed to a well-formed IETF BCP 47 language tag
+ * without losing information.
+ *
+ * <p>The following example shows how to create a <code>Locale</code> object
+ * with the <code>LocaleBuilder</code>.
+ * <blockquote>
+ * <pre>
+ *     UErrorCode status = U_ZERO_ERROR;
+ *     Locale aLocale = LocaleBuilder()
+ *                          .setLanguage("sr")
+ *                          .setScript("Latn")
+ *                          .setRegion("RS")
+ *                          .build(status);
+ *     if (U_SUCCESS(status)) {
+ *       // ...
+ *     }
+ * </pre>
+ * </blockquote>
+ *
+ * <p>LocaleBuilders can be reused; <code>clear()</code> resets all
+ * fields to their default values.
+ *
+ * <p>LocaleBuilder tracks errors in an internal UErrorCode. For all setters,
+ * except setLanguageTag and setLocale, LocaleBuilder will return immediately
+ * if the internal UErrorCode is in error state.
+ * To reset internal state and error code, call clear method.
+ * The setLanguageTag and setLocale method will first clear the internal
+ * UErrorCode, then track the error of the validation of the input parameter
+ * into the internal UErrorCode.
+ *
+ * @stable ICU 64
+ */
+class U_COMMON_API LocaleBuilder : public UObject {
+public:
+    /**
+     * Constructs an empty LocaleBuilder. The default value of all
+     * fields, extensions, and private use information is the
+     * empty string.
+     *
+     * @stable ICU 64
+     */
+    LocaleBuilder();
+
+    /**
+     * Destructor
+     * @stable ICU 64
+     */
+    virtual ~LocaleBuilder();
+
+    /**
+     * Resets the <code>LocaleBuilder</code> to match the provided
+     * <code>locale</code>.  Existing state is discarded.
+     *
+     * <p>All fields of the locale must be well-formed.
+     * <p>This method clears the internal UErrorCode.
+     *
+     * @param locale the locale
+     * @return This builder.
+     *
+     * @stable ICU 64
+     */
+    LocaleBuilder& setLocale(const Locale& locale);
+
+    /**
+     * Resets the LocaleBuilder to match the provided IETF BCP 47 language tag.
+     * Discards the existing state.
+     * The empty string causes the builder to be reset, like {@link #clear}.
+     * Legacy language tags (marked as “Type: grandfathered” in BCP 47)
+     * are converted to their canonical form before being processed.
+     * Otherwise, the <code>language tag</code> must be well-formed,
+     * or else the build() method will later report an U_ILLEGAL_ARGUMENT_ERROR.
+     *
+     * <p>This method clears the internal UErrorCode.
+     *
+     * @param tag the language tag, defined as IETF BCP 47 language tag.
+     * @return This builder.
+     * @stable ICU 64
+     */
+    LocaleBuilder& setLanguageTag(StringPiece tag);
+
+    /**
+     * Sets the language.  If <code>language</code> is the empty string, the
+     * language in this <code>LocaleBuilder</code> is removed. Otherwise, the
+     * <code>language</code> must be well-formed, or else the build() method will
+     * later report an U_ILLEGAL_ARGUMENT_ERROR.
+     *
+     * <p>The syntax of language value is defined as
+     * [unicode_language_subtag](http://www.unicode.org/reports/tr35/tr35.html#unicode_language_subtag).
+     *
+     * @param language the language
+     * @return This builder.
+     * @stable ICU 64
+     */
+    LocaleBuilder& setLanguage(StringPiece language);
+
+    /**
+     * Sets the script. If <code>script</code> is the empty string, the script in
+     * this <code>LocaleBuilder</code> is removed.
+     * Otherwise, the <code>script</code> must be well-formed, or else the build()
+     * method will later report an U_ILLEGAL_ARGUMENT_ERROR.
+     *
+     * <p>The script value is a four-letter script code as
+     * [unicode_script_subtag](http://www.unicode.org/reports/tr35/tr35.html#unicode_script_subtag)
+     * defined by ISO 15924
+     *
+     * @param script the script
+     * @return This builder.
+     * @stable ICU 64
+     */
+    LocaleBuilder& setScript(StringPiece script);
+
+    /**
+     * Sets the region.  If region is the empty string, the region in this
+     * <code>LocaleBuilder</code> is removed. Otherwise, the <code>region</code>
+     * must be well-formed, or else the build() method will later report an
+     * U_ILLEGAL_ARGUMENT_ERROR.
+     *
+     * <p>The region value is defined by
+     *  [unicode_region_subtag](http://www.unicode.org/reports/tr35/tr35.html#unicode_region_subtag)
+     * as a two-letter ISO 3166 code or a three-digit UN M.49 area code.
+     *
+     * <p>The region value in the <code>Locale</code> created by the
+     * <code>LocaleBuilder</code> is always normalized to upper case.
+     *
+     * @param region the region
+     * @return This builder.
+     * @stable ICU 64
+     */
+    LocaleBuilder& setRegion(StringPiece region);
+
+    /**
+     * Sets the variant.  If variant is the empty string, the variant in this
+     * <code>LocaleBuilder</code> is removed.  Otherwise, the <code>variant</code>
+     * must be well-formed, or else the build() method will later report an
+     * U_ILLEGAL_ARGUMENT_ERROR.
+     *
+     * <p><b>Note:</b> This method checks if <code>variant</code>
+     * satisfies the
+     * [unicode_variant_subtag](http://www.unicode.org/reports/tr35/tr35.html#unicode_variant_subtag)
+     * syntax requirements, and normalizes the value to lowercase letters. However,
+     * the <code>Locale</code> class does not impose any syntactic
+     * restriction on variant. To set an ill-formed variant, use a Locale constructor.
+     * If there are multiple unicode_variant_subtag, the caller must concatenate
+     * them with '-' as separator (ex: "foobar-fibar").
+     *
+     * @param variant the variant
+     * @return This builder.
+     * @stable ICU 64
+     */
+    LocaleBuilder& setVariant(StringPiece variant);
+
+    /**
+     * Sets the extension for the given key. If the value is the empty string,
+     * the extension is removed.  Otherwise, the <code>key</code> and
+     * <code>value</code> must be well-formed, or else the build() method will
+     * later report an U_ILLEGAL_ARGUMENT_ERROR.
+     *
+     * <p><b>Note:</b> The key ('u') is used for the Unicode locale extension.
+     * Setting a value for this key replaces any existing Unicode locale key/type
+     * pairs with those defined in the extension.
+     *
+     * <p><b>Note:</b> The key ('x') is used for the private use code. To be
+     * well-formed, the value for this key needs only to have subtags of one to
+     * eight alphanumeric characters, not two to eight as in the general case.
+     *
+     * @param key the extension key
+     * @param value the extension value
+     * @return This builder.
+     * @stable ICU 64
+     */
+    LocaleBuilder& setExtension(char key, StringPiece value);
+
+    /**
+     * Sets the Unicode locale keyword type for the given key. If the type
+     * StringPiece is constructed with a nullptr, the keyword is removed.
+     * If the type is the empty string, the keyword is set without type subtags.
+     * Otherwise, the key and type must be well-formed, or else the build()
+     * method will later report an U_ILLEGAL_ARGUMENT_ERROR.
+     *
+     * <p>Keys and types are converted to lower case.
+     *
+     * <p><b>Note</b>:Setting the 'u' extension via {@link #setExtension}
+     * replaces all Unicode locale keywords with those defined in the
+     * extension.
+     *
+     * @param key the Unicode locale key
+     * @param type the Unicode locale type
+     * @return This builder.
+     * @stable ICU 64
+     */
+    LocaleBuilder& setUnicodeLocaleKeyword(
+        StringPiece key, StringPiece type);
+
+    /**
+     * Adds a unicode locale attribute, if not already present, otherwise
+     * has no effect.  The attribute must not be empty string and must be
+     * well-formed or U_ILLEGAL_ARGUMENT_ERROR will be set to status
+     * during the build() call.
+     *
+     * @param attribute the attribute
+     * @return This builder.
+     * @stable ICU 64
+     */
+    LocaleBuilder& addUnicodeLocaleAttribute(StringPiece attribute);
+
+    /**
+     * Removes a unicode locale attribute, if present, otherwise has no
+     * effect.  The attribute must not be empty string and must be well-formed
+     * or U_ILLEGAL_ARGUMENT_ERROR will be set to status during the build() call.
+     *
+     * <p>Attribute comparison for removal is case-insensitive.
+     *
+     * @param attribute the attribute
+     * @return This builder.
+     * @stable ICU 64
+     */
+    LocaleBuilder& removeUnicodeLocaleAttribute(StringPiece attribute);
+
+    /**
+     * Resets the builder to its initial, empty state.
+     * <p>This method clears the internal UErrorCode.
+     *
+     * @return this builder
+     * @stable ICU 64
+     */
+    LocaleBuilder& clear();
+
+    /**
+     * Resets the extensions to their initial, empty state.
+     * Language, script, region and variant are unchanged.
+     *
+     * @return this builder
+     * @stable ICU 64
+     */
+    LocaleBuilder& clearExtensions();
+
+    /**
+     * Returns an instance of <code>Locale</code> created from the fields set
+     * on this builder.
+     * If any set methods or during the build() call require memory allocation
+     * but fail U_MEMORY_ALLOCATION_ERROR will be set to status.
+     * If any of the fields set by the setters are not well-formed, the status
+     * will be set to U_ILLEGAL_ARGUMENT_ERROR. The state of the builder will
+     * not change after the build() call and the caller is free to keep using
+     * the same builder to build more locales.
+     *
+     * @return a new Locale
+     * @stable ICU 64
+     */
+    Locale build(UErrorCode& status);
+
+    /**
+     * Sets the UErrorCode if an error occurred while recording sets.
+     * Preserves older error codes in the outErrorCode.
+     * @param outErrorCode Set to an error code that occurred while setting subtags.
+     *                  Unchanged if there is no such error or if outErrorCode
+     *                  already contained an error.
+     * @return true if U_FAILURE(outErrorCode)
+     * @stable ICU 65
+     */
+    UBool copyErrorTo(UErrorCode &outErrorCode) const;
+
+private:
+    friend class LocaleMatcher::Result;
+
+    void copyExtensionsFrom(const Locale& src, UErrorCode& errorCode);
+
+    UErrorCode status_;
+    char language_[9];
+    char script_[5];
+    char region_[4];
+    CharString *variant_;  // Pointer not object so we need not #include internal charstr.h.
+    icu::Locale *extensions_;  // Pointer not object. Storage for all other fields.
+
+};
+
+U_NAMESPACE_END
+
+#endif /* U_SHOW_CPLUSPLUS_API */
+
+#endif  // __LOCALEBUILDER_H__

thirdparty/icu4c/common/unicode/localematcher.h

@@ -0,0 +1,710 @@
+// © 2019 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+
+// localematcher.h
+// created: 2019may08 Markus W. Scherer
+
+#ifndef __LOCALEMATCHER_H__
+#define __LOCALEMATCHER_H__
+
+#include "unicode/utypes.h"
+
+#if U_SHOW_CPLUSPLUS_API
+
+#include <optional>
+
+#include "unicode/locid.h"
+#include "unicode/stringpiece.h"
+#include "unicode/uobject.h"
+
+/**
+ * \file
+ * \brief C++ API: Locale matcher: User's desired locales vs. application's supported locales.
+ */
+
+/**
+ * Builder option for whether the language subtag or the script subtag is most important.
+ *
+ * @see LocaleMatcher::Builder#setFavorSubtag(ULocMatchFavorSubtag)
+ * @stable ICU 65
+ */
+enum ULocMatchFavorSubtag {
+    /**
+     * Language differences are most important, then script differences, then region differences.
+     * (This is the default behavior.)
+     *
+     * @stable ICU 65
+     */
+    ULOCMATCH_FAVOR_LANGUAGE,
+    /**
+     * Makes script differences matter relatively more than language differences.
+     *
+     * @stable ICU 65
+     */
+    ULOCMATCH_FAVOR_SCRIPT
+};
+#ifndef U_IN_DOXYGEN
+typedef enum ULocMatchFavorSubtag ULocMatchFavorSubtag;
+#endif
+
+/**
+ * Builder option for whether all desired locales are treated equally or
+ * earlier ones are preferred.
+ *
+ * @see LocaleMatcher::Builder#setDemotionPerDesiredLocale(ULocMatchDemotion)
+ * @stable ICU 65
+ */
+enum ULocMatchDemotion {
+    /**
+     * All desired locales are treated equally.
+     *
+     * @stable ICU 65
+     */
+    ULOCMATCH_DEMOTION_NONE,
+    /**
+     * Earlier desired locales are preferred.
+     *
+     * <p>From each desired locale to the next,
+     * the distance to any supported locale is increased by an additional amount
+     * which is at least as large as most region mismatches.
+     * A later desired locale has to have a better match with some supported locale
+     * due to more than merely having the same region subtag.
+     *
+     * <p>For example: <code>Supported={en, sv}  desired=[en-GB, sv]</code>
+     * yields <code>Result(en-GB, en)</code> because
+     * with the demotion of sv its perfect match is no better than
+     * the region distance between the earlier desired locale en-GB and en=en-US.
+     *
+     * <p>Notes:
+     * <ul>
+     *   <li>In some cases, language and/or script differences can be as small as
+     *       the typical region difference. (Example: sr-Latn vs. sr-Cyrl)
+     *   <li>It is possible for certain region differences to be larger than usual,
+     *       and larger than the demotion.
+     *       (As of CLDR 35 there is no such case, but
+     *        this is possible in future versions of the data.)
+     * </ul>
+     *
+     * @stable ICU 65
+     */
+    ULOCMATCH_DEMOTION_REGION
+};
+#ifndef U_IN_DOXYGEN
+typedef enum ULocMatchDemotion ULocMatchDemotion;
+#endif
+
+/**
+ * Builder option for whether to include or ignore one-way (fallback) match data.
+ * The LocaleMatcher uses CLDR languageMatch data which includes fallback (oneway=true) entries.
+ * Sometimes it is desirable to ignore those.
+ *
+ * <p>For example, consider a web application with the UI in a given language,
+ * with a link to another, related web app.
+ * The link should include the UI language, and the target server may also use
+ * the client’s Accept-Language header data.
+ * The target server has its own list of supported languages.
+ * One may want to favor UI language consistency, that is,
+ * if there is a decent match for the original UI language, we want to use it,
+ * but not if it is merely a fallback.
+ *
+ * @see LocaleMatcher::Builder#setDirection(ULocMatchDirection)
+ * @stable ICU 67
+ */
+enum ULocMatchDirection {
+    /**
+     * Locale matching includes one-way matches such as Breton→French. (default)
+     *
+     * @stable ICU 67
+     */
+    ULOCMATCH_DIRECTION_WITH_ONE_WAY,
+    /**
+     * Locale matching limited to two-way matches including e.g. Danish↔Norwegian
+     * but ignoring one-way matches.
+     *
+     * @stable ICU 67
+     */
+    ULOCMATCH_DIRECTION_ONLY_TWO_WAY
+};
+#ifndef U_IN_DOXYGEN
+typedef enum ULocMatchDirection ULocMatchDirection;
+#endif
+
+struct UHashtable;
+
+U_NAMESPACE_BEGIN
+
+struct LSR;
+
+class LikelySubtags;
+class LocaleDistance;
+class LocaleLsrIterator;
+class UVector;
+
+/**
+ * Immutable class that picks the best match between a user's desired locales and
+ * an application's supported locales.
+ * Movable but not copyable.
+ *
+ * <p>Example:
+ * <pre>
+ * UErrorCode errorCode = U_ZERO_ERROR;
+ * LocaleMatcher matcher = LocaleMatcher::Builder().setSupportedLocales("fr, en-GB, en").build(errorCode);
+ * Locale *bestSupported = matcher.getBestLocale(Locale.US, errorCode);  // "en"
+ * </pre>
+ *
+ * <p>A matcher takes into account when languages are close to one another,
+ * such as Danish and Norwegian,
+ * and when regional variants are close, like en-GB and en-AU as opposed to en-US.
+ *
+ * <p>If there are multiple supported locales with the same (language, script, region)
+ * likely subtags, then the current implementation returns the first of those locales.
+ * It ignores variant subtags (except for pseudolocale variants) and extensions.
+ * This may change in future versions.
+ *
+ * <p>For example, the current implementation does not distinguish between
+ * de, de-DE, de-Latn, de-1901, de-u-co-phonebk.
+ *
+ * <p>If you prefer one equivalent locale over another, then provide only the preferred one,
+ * or place it earlier in the list of supported locales.
+ *
+ * <p>Otherwise, the order of supported locales may have no effect on the best-match results.
+ * The current implementation compares each desired locale with supported locales
+ * in the following order:
+ * 1. Default locale, if supported;
+ * 2. CLDR "paradigm locales" like en-GB and es-419;
+ * 3. other supported locales.
+ * This may change in future versions.
+ *
+ * <p>Often a product will just need one matcher instance, built with the languages
+ * that it supports. However, it may want multiple instances with different
+ * default languages based on additional information, such as the domain.
+ *
+ * <p>This class is not intended for public subclassing.
+ *
+ * @stable ICU 65
+ */
+class U_COMMON_API LocaleMatcher : public UMemory {
+public:
+    /**
+     * Data for the best-matching pair of a desired and a supported locale.
+     * Movable but not copyable.
+     *
+     * @stable ICU 65
+     */
+    class U_COMMON_API Result : public UMemory {
+    public:
+        /**
+         * Move constructor; might modify the source.
+         * This object will have the same contents that the source object had.
+         *
+         * @param src Result to move contents from.
+         * @stable ICU 65
+         */
+        Result(Result &&src) noexcept;
+
+        /**
+         * Destructor.
+         *
+         * @stable ICU 65
+         */
+        ~Result();
+
+        /**
+         * Move assignment; might modify the source.
+         * This object will have the same contents that the source object had.
+         *
+         * @param src Result to move contents from.
+         * @stable ICU 65
+         */
+        Result &operator=(Result &&src) noexcept;
+
+        /**
+         * Returns the best-matching desired locale.
+         * nullptr if the list of desired locales is empty or if none matched well enough.
+         *
+         * @return the best-matching desired locale, or nullptr.
+         * @stable ICU 65
+         */
+        inline const Locale *getDesiredLocale() const { return desiredLocale; }
+
+        /**
+         * Returns the best-matching supported locale.
+         * If none matched well enough, this is the default locale.
+         * The default locale is nullptr if Builder::setNoDefaultLocale() was called,
+         * or if the list of supported locales is empty and no explicit default locale is set.
+         *
+         * @return the best-matching supported locale, or nullptr.
+         * @stable ICU 65
+         */
+        inline const Locale *getSupportedLocale() const { return supportedLocale; }
+
+        /**
+         * Returns the index of the best-matching desired locale in the input Iterable order.
+         * -1 if the list of desired locales is empty or if none matched well enough.
+         *
+         * @return the index of the best-matching desired locale, or -1.
+         * @stable ICU 65
+         */
+        inline int32_t getDesiredIndex() const { return desiredIndex; }
+
+        /**
+         * Returns the index of the best-matching supported locale in the
+         * constructor’s or builder’s input order (“set” Collection plus “added” locales).
+         * If the matcher was built from a locale list string, then the iteration order is that
+         * of a LocalePriorityList built from the same string.
+         * -1 if the list of supported locales is empty or if none matched well enough.
+         *
+         * @return the index of the best-matching supported locale, or -1.
+         * @stable ICU 65
+         */
+        inline int32_t getSupportedIndex() const { return supportedIndex; }
+
+        /**
+         * Takes the best-matching supported locale and adds relevant fields of the
+         * best-matching desired locale, such as the -t- and -u- extensions.
+         * May replace some fields of the supported locale.
+         * The result is the locale that should be used for date and number formatting, collation, etc.
+         * Returns the root locale if getSupportedLocale() returns nullptr.
+         *
+         * <p>Example: desired=ar-SA-u-nu-latn, supported=ar-EG, resolved locale=ar-SA-u-nu-latn
+         *
+         * @return a locale combining the best-matching desired and supported locales.
+         * @stable ICU 65
+         */
+        Locale makeResolvedLocale(UErrorCode &errorCode) const;
+
+    private:
+        Result(const Locale *desired, const Locale *supported,
+               int32_t desIndex, int32_t suppIndex, UBool owned) :
+                desiredLocale(desired), supportedLocale(supported),
+                desiredIndex(desIndex), supportedIndex(suppIndex),
+                desiredIsOwned(owned) {}
+
+        Result(const Result &other) = delete;
+        Result &operator=(const Result &other) = delete;
+
+        const Locale *desiredLocale;
+        const Locale *supportedLocale;
+        int32_t desiredIndex;
+        int32_t supportedIndex;
+        UBool desiredIsOwned;
+
+        friend class LocaleMatcher;
+    };
+
+    /**
+     * LocaleMatcher builder.
+     * Movable but not copyable.
+     *
+     * @stable ICU 65
+     */
+    class U_COMMON_API Builder : public UMemory {
+    public:
+        /**
+         * Constructs a builder used in chaining parameters for building a LocaleMatcher.
+         *
+         * @return a new Builder object
+         * @stable ICU 65
+         */
+        Builder() {}
+
+        /**
+         * Move constructor; might modify the source.
+         * This builder will have the same contents that the source builder had.
+         *
+         * @param src Builder to move contents from.
+         * @stable ICU 65
+         */
+        Builder(Builder &&src) noexcept;
+
+        /**
+         * Destructor.
+         *
+         * @stable ICU 65
+         */
+        ~Builder();
+
+        /**
+         * Move assignment; might modify the source.
+         * This builder will have the same contents that the source builder had.
+         *
+         * @param src Builder to move contents from.
+         * @stable ICU 65
+         */
+        Builder &operator=(Builder &&src) noexcept;
+
+        /**
+         * Parses an Accept-Language string
+         * (<a href="https://tools.ietf.org/html/rfc2616#section-14.4">RFC 2616 Section 14.4</a>),
+         * such as "af, en, fr;q=0.9", and sets the supported locales accordingly.
+         * Allows whitespace in more places but does not allow "*".
+         * Clears any previously set/added supported locales first.
+         *
+         * @param locales the Accept-Language string of locales to set
+         * @return this Builder object
+         * @stable ICU 65
+         */
+        Builder &setSupportedLocalesFromListString(StringPiece locales);
+
+        /**
+         * Copies the supported locales, preserving iteration order.
+         * Clears any previously set/added supported locales first.
+         * Duplicates are allowed, and are not removed.
+         *
+         * @param locales the list of locale
+         * @return this Builder object
+         * @stable ICU 65
+         */
+        Builder &setSupportedLocales(Locale::Iterator &locales);
+
+        /**
+         * Copies the supported locales from the begin/end range, preserving iteration order.
+         * Clears any previously set/added supported locales first.
+         * Duplicates are allowed, and are not removed.
+         *
+         * Each of the iterator parameter values must be an
+         * input iterator whose value is convertible to const Locale &.
+         *
+         * @param begin Start of range.
+         * @param end Exclusive end of range.
+         * @return this Builder object
+         * @stable ICU 65
+         */
+        template<typename Iter>
+        Builder &setSupportedLocales(Iter begin, Iter end) {
+            if (U_FAILURE(errorCode_)) { return *this; }
+            clearSupportedLocales();
+            while (begin != end) {
+                addSupportedLocale(*begin++);
+            }
+            return *this;
+        }
+
+        /**
+         * Copies the supported locales from the begin/end range, preserving iteration order.
+         * Calls the converter to convert each *begin to a Locale or const Locale &.
+         * Clears any previously set/added supported locales first.
+         * Duplicates are allowed, and are not removed.
+         *
+         * Each of the iterator parameter values must be an
+         * input iterator whose value is convertible to const Locale &.
+         *
+         * @param begin Start of range.
+         * @param end Exclusive end of range.
+         * @param converter Converter from *begin to const Locale & or compatible.
+         * @return this Builder object
+         * @stable ICU 65
+         */
+        template<typename Iter, typename Conv>
+        Builder &setSupportedLocalesViaConverter(Iter begin, Iter end, Conv converter) {
+            if (U_FAILURE(errorCode_)) { return *this; }
+            clearSupportedLocales();
+            while (begin != end) {
+                addSupportedLocale(converter(*begin++));
+            }
+            return *this;
+        }
+
+        /**
+         * Adds another supported locale.
+         * Duplicates are allowed, and are not removed.
+         *
+         * @param locale another locale
+         * @return this Builder object
+         * @stable ICU 65
+         */
+        Builder &addSupportedLocale(const Locale &locale);
+
+        /**
+         * Sets no default locale.
+         * There will be no explicit or implicit default locale.
+         * If there is no good match, then the matcher will return nullptr for the
+         * best supported locale.
+         *
+         * @stable ICU 68
+         */
+        Builder &setNoDefaultLocale();
+
+        /**
+         * Sets the default locale; if nullptr, or if it is not set explicitly,
+         * then the first supported locale is used as the default locale.
+         * There is no default locale at all (nullptr will be returned instead)
+         * if setNoDefaultLocale() is called.
+         *
+         * @param defaultLocale the default locale (will be copied)
+         * @return this Builder object
+         * @stable ICU 65
+         */
+        Builder &setDefaultLocale(const Locale *defaultLocale);
+
+        /**
+         * If ULOCMATCH_FAVOR_SCRIPT, then the language differences are smaller than script
+         * differences.
+         * This is used in situations (such as maps) where
+         * it is better to fall back to the same script than a similar language.
+         *
+         * @param subtag the subtag to favor
+         * @return this Builder object
+         * @stable ICU 65
+         */
+        Builder &setFavorSubtag(ULocMatchFavorSubtag subtag);
+
+        /**
+         * Option for whether all desired locales are treated equally or
+         * earlier ones are preferred (this is the default).
+         *
+         * @param demotion the demotion per desired locale to set.
+         * @return this Builder object
+         * @stable ICU 65
+         */
+        Builder &setDemotionPerDesiredLocale(ULocMatchDemotion demotion);
+
+        /**
+         * Option for whether to include or ignore one-way (fallback) match data.
+         * By default, they are included.
+         *
+         * @param matchDirection the match direction to set.
+         * @return this Builder object
+         * @stable ICU 67
+         */
+        Builder &setDirection(ULocMatchDirection matchDirection) {
+            if (U_SUCCESS(errorCode_)) {
+                direction_ = matchDirection;
+            }
+            return *this;
+        }
+
+        /**
+         * Sets the maximum distance for an acceptable match.
+         * The matcher will return a match for a pair of locales only if
+         * they match at least as well as the pair given here.
+         *
+         * For example, setMaxDistance(en-US, en-GB) limits matches to ones where the
+         * (desired, support) locales have a distance no greater than a region subtag difference.
+         * This is much stricter than the CLDR default.
+         *
+         * The details of locale matching are subject to changes in
+         * CLDR data and in the algorithm.
+         * Specifying a maximum distance in relative terms via a sample pair of locales
+         * insulates from changes that affect all distance metrics similarly,
+         * but some changes will necessarily affect relative distances between
+         * different pairs of locales.
+         *
+         * @param desired the desired locale for distance comparison.
+         * @param supported the supported locale for distance comparison.
+         * @return this Builder object
+         * @stable ICU 68
+         */
+        Builder &setMaxDistance(const Locale &desired, const Locale &supported);
+
+        /**
+         * Sets the UErrorCode if an error occurred while setting parameters.
+         * Preserves older error codes in the outErrorCode.
+         *
+         * @param outErrorCode Set to an error code if it does not contain one already
+         *                  and an error occurred while setting parameters.
+         *                  Otherwise unchanged.
+         * @return true if U_FAILURE(outErrorCode)
+         * @stable ICU 65
+         */
+        UBool copyErrorTo(UErrorCode &outErrorCode) const;
+
+        /**
+         * Builds and returns a new locale matcher.
+         * This builder can continue to be used.
+         *
+         * @param errorCode ICU error code. Its input value must pass the U_SUCCESS() test,
+         *                  or else the function returns immediately. Check for U_FAILURE()
+         *                  on output or use with function chaining. (See User Guide for details.)
+         * @return LocaleMatcher
+         * @stable ICU 65
+         */
+        LocaleMatcher build(UErrorCode &errorCode) const;
+
+    private:
+        friend class LocaleMatcher;
+
+        Builder(const Builder &other) = delete;
+        Builder &operator=(const Builder &other) = delete;
+
+        void clearSupportedLocales();
+        bool ensureSupportedLocaleVector();
+
+        UErrorCode errorCode_ = U_ZERO_ERROR;
+        UVector *supportedLocales_ = nullptr;
+        int32_t thresholdDistance_ = -1;
+        ULocMatchDemotion demotion_ = ULOCMATCH_DEMOTION_REGION;
+        Locale *defaultLocale_ = nullptr;
+        bool withDefault_ = true;
+        ULocMatchFavorSubtag favor_ = ULOCMATCH_FAVOR_LANGUAGE;
+        ULocMatchDirection direction_ = ULOCMATCH_DIRECTION_WITH_ONE_WAY;
+        Locale *maxDistanceDesired_ = nullptr;
+        Locale *maxDistanceSupported_ = nullptr;
+    };
+
+    // FYI No public LocaleMatcher constructors in C++; use the Builder.
+
+    /**
+     * Move copy constructor; might modify the source.
+     * This matcher will have the same settings that the source matcher had.
+     * @param src source matcher
+     * @stable ICU 65
+     */
+    LocaleMatcher(LocaleMatcher &&src) noexcept;
+
+    /**
+     * Destructor.
+     * @stable ICU 65
+     */
+    ~LocaleMatcher();
+
+    /**
+     * Move assignment operator; might modify the source.
+     * This matcher will have the same settings that the source matcher had.
+     * The behavior is undefined if *this and src are the same object.
+     * @param src source matcher
+     * @return *this
+     * @stable ICU 65
+     */
+    LocaleMatcher &operator=(LocaleMatcher &&src) noexcept;
+
+    /**
+     * Returns the supported locale which best matches the desired locale.
+     *
+     * @param desiredLocale Typically a user's language.
+     * @param errorCode ICU error code. Its input value must pass the U_SUCCESS() test,
+     *                  or else the function returns immediately. Check for U_FAILURE()
+     *                  on output or use with function chaining. (See User Guide for details.)
+     * @return the best-matching supported locale.
+     * @stable ICU 65
+     */
+    const Locale *getBestMatch(const Locale &desiredLocale, UErrorCode &errorCode) const;
+
+    /**
+     * Returns the supported locale which best matches one of the desired locales.
+     *
+     * @param desiredLocales Typically a user's languages, in order of preference (descending).
+     * @param errorCode ICU error code. Its input value must pass the U_SUCCESS() test,
+     *                  or else the function returns immediately. Check for U_FAILURE()
+     *                  on output or use with function chaining. (See User Guide for details.)
+     * @return the best-matching supported locale.
+     * @stable ICU 65
+     */
+    const Locale *getBestMatch(Locale::Iterator &desiredLocales, UErrorCode &errorCode) const;
+
+    /**
+     * Parses an Accept-Language string
+     * (<a href="https://tools.ietf.org/html/rfc2616#section-14.4">RFC 2616 Section 14.4</a>),
+     * such as "af, en, fr;q=0.9",
+     * and returns the supported locale which best matches one of the desired locales.
+     * Allows whitespace in more places but does not allow "*".
+     *
+     * @param desiredLocaleList Typically a user's languages, as an Accept-Language string.
+     * @param errorCode ICU error code. Its input value must pass the U_SUCCESS() test,
+     *                  or else the function returns immediately. Check for U_FAILURE()
+     *                  on output or use with function chaining. (See User Guide for details.)
+     * @return the best-matching supported locale.
+     * @stable ICU 65
+     */
+    const Locale *getBestMatchForListString(StringPiece desiredLocaleList, UErrorCode &errorCode) const;
+
+    /**
+     * Returns the best match between the desired locale and the supported locales.
+     * If the result's desired locale is not nullptr, then it is the address of the input locale.
+     * It has not been cloned.
+     *
+     * @param desiredLocale Typically a user's language.
+     * @param errorCode ICU error code. Its input value must pass the U_SUCCESS() test,
+     *                  or else the function returns immediately. Check for U_FAILURE()
+     *                  on output or use with function chaining. (See User Guide for details.)
+     * @return the best-matching pair of the desired and a supported locale.
+     * @stable ICU 65
+     */
+    Result getBestMatchResult(const Locale &desiredLocale, UErrorCode &errorCode) const;
+
+    /**
+     * Returns the best match between the desired and supported locales.
+     * If the result's desired locale is not nullptr, then it is a clone of
+     * the best-matching desired locale. The Result object owns the clone.
+     *
+     * @param desiredLocales Typically a user's languages, in order of preference (descending).
+     * @param errorCode ICU error code. Its input value must pass the U_SUCCESS() test,
+     *                  or else the function returns immediately. Check for U_FAILURE()
+     *                  on output or use with function chaining. (See User Guide for details.)
+     * @return the best-matching pair of a desired and a supported locale.
+     * @stable ICU 65
+     */
+    Result getBestMatchResult(Locale::Iterator &desiredLocales, UErrorCode &errorCode) const;
+
+    /**
+     * Returns true if the pair of locales matches acceptably.
+     * This is influenced by Builder options such as setDirection(), setFavorSubtag(),
+     * and setMaxDistance().
+     *
+     * @param desired The desired locale.
+     * @param supported The supported locale.
+     * @param errorCode ICU error code. Its input value must pass the U_SUCCESS() test,
+     *                  or else the function returns immediately. Check for U_FAILURE()
+     *                  on output or use with function chaining. (See User Guide for details.)
+     * @return true if the pair of locales matches acceptably.
+     * @stable ICU 68
+     */
+    UBool isMatch(const Locale &desired, const Locale &supported, UErrorCode &errorCode) const;
+
+#ifndef U_HIDE_INTERNAL_API
+    /**
+     * Returns a fraction between 0 and 1, where 1 means that the languages are a
+     * perfect match, and 0 means that they are completely different.
+     *
+     * <p>This is mostly an implementation detail, and the precise values may change over time.
+     * The implementation may use either the maximized forms or the others ones, or both.
+     * The implementation may or may not rely on the forms to be consistent with each other.
+     *
+     * <p>Callers should construct and use a matcher rather than match pairs of locales directly.
+     *
+     * @param desired Desired locale.
+     * @param supported Supported locale.
+     * @param errorCode ICU error code. Its input value must pass the U_SUCCESS() test,
+     *                  or else the function returns immediately. Check for U_FAILURE()
+     *                  on output or use with function chaining. (See User Guide for details.)
+     * @return value between 0 and 1, inclusive.
+     * @internal (has a known user)
+     */
+    double internalMatch(const Locale &desired, const Locale &supported, UErrorCode &errorCode) const;
+#endif  // U_HIDE_INTERNAL_API
+
+private:
+    LocaleMatcher(const Builder &builder, UErrorCode &errorCode);
+    LocaleMatcher(const LocaleMatcher &other) = delete;
+    LocaleMatcher &operator=(const LocaleMatcher &other) = delete;
+
+    int32_t putIfAbsent(const LSR &lsr, int32_t i, int32_t suppLength, UErrorCode &errorCode);
+
+    std::optional<int32_t> getBestSuppIndex(LSR desiredLSR, LocaleLsrIterator *remainingIter, UErrorCode &errorCode) const;
+
+    const LikelySubtags &likelySubtags;
+    const LocaleDistance &localeDistance;
+    int32_t thresholdDistance;
+    int32_t demotionPerDesiredLocale;
+    ULocMatchFavorSubtag favorSubtag;
+    ULocMatchDirection direction;
+
+    // These are in input order.
+    const Locale ** supportedLocales;
+    LSR *lsrs;
+    int32_t supportedLocalesLength;
+    // These are in preference order: 1. Default locale 2. paradigm locales 3. others.
+    UHashtable *supportedLsrToIndex;  // Map<LSR, Integer>
+    // Array versions of the supportedLsrToIndex keys and values.
+    // The distance lookup loops over the supportedLSRs and returns the index of the best match.
+    const LSR **supportedLSRs;
+    int32_t *supportedIndexes;
+    int32_t supportedLSRsLength;
+    Locale *ownedDefaultLocale;
+    const Locale *defaultLocale;
+};
+
+U_NAMESPACE_END
+
+#endif  // U_SHOW_CPLUSPLUS_API
+#endif  // __LOCALEMATCHER_H__

thirdparty/icu4c/common/unicode/localpointer.h

@@ -0,0 +1,609 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+/*
+*******************************************************************************
+*
+*   Copyright (C) 2009-2016, International Business Machines
+*   Corporation and others.  All Rights Reserved.
+*
+*******************************************************************************
+*   file name:  localpointer.h
+*   encoding:   UTF-8
+*   tab size:   8 (not used)
+*   indentation:4
+*
+*   created on: 2009nov13
+*   created by: Markus W. Scherer
+*/
+
+#ifndef __LOCALPOINTER_H__
+#define __LOCALPOINTER_H__
+
+/**
+ * \file
+ * \brief C++ API: "Smart pointers" for use with and in ICU4C C++ code.
+ *
+ * These classes are inspired by
+ * - std::auto_ptr
+ * - boost::scoped_ptr & boost::scoped_array
+ * - Taligent Safe Pointers (TOnlyPointerTo)
+ *
+ * but none of those provide for all of the goals for ICU smart pointers:
+ * - Smart pointer owns the object and releases it when it goes out of scope.
+ * - No transfer of ownership via copy/assignment to reduce misuse. Simpler & more robust.
+ * - ICU-compatible: No exceptions.
+ * - Need to be able to orphan/release the pointer and its ownership.
+ * - Need variants for normal C++ object pointers, C++ arrays, and ICU C service objects.
+ *
+ * For details see https://icu.unicode.org/design/cpp/scoped_ptr
+ */
+
+#include "unicode/utypes.h"
+
+#if U_SHOW_CPLUSPLUS_API
+
+#include <memory>
+
+U_NAMESPACE_BEGIN
+
+/**
+ * "Smart pointer" base class; do not use directly: use LocalPointer etc.
+ *
+ * Base class for smart pointer classes that do not throw exceptions.
+ *
+ * Do not use this base class directly, since it does not delete its pointer.
+ * A subclass must implement methods that delete the pointer:
+ * Destructor and adoptInstead().
+ *
+ * There is no operator T *() provided because the programmer must decide
+ * whether to use getAlias() (without transfer of ownership) or orphan()
+ * (with transfer of ownership and NULLing of the pointer).
+ *
+ * @see LocalPointer
+ * @see LocalArray
+ * @see U_DEFINE_LOCAL_OPEN_POINTER
+ * @stable ICU 4.4
+ */
+template<typename T>
+class LocalPointerBase {
+public:
+    // No heap allocation. Use only on the stack.
+    static void* U_EXPORT2 operator new(size_t) = delete;
+    static void* U_EXPORT2 operator new[](size_t) = delete;
+#if U_HAVE_PLACEMENT_NEW
+    static void* U_EXPORT2 operator new(size_t, void*) = delete;
+#endif
+
+    /**
+     * Constructor takes ownership.
+     * @param p simple pointer to an object that is adopted
+     * @stable ICU 4.4
+     */
+    explicit LocalPointerBase(T *p=nullptr) : ptr(p) {}
+    /**
+     * Destructor deletes the object it owns.
+     * Subclass must override: Base class does nothing.
+     * @stable ICU 4.4
+     */
+    ~LocalPointerBase() { /* delete ptr; */ }
+    /**
+     * nullptr check.
+     * @return true if ==nullptr
+     * @stable ICU 4.4
+     */
+    UBool isNull() const { return ptr==nullptr; }
+    /**
+     * nullptr check.
+     * @return true if !=nullptr
+     * @stable ICU 4.4
+     */
+    UBool isValid() const { return ptr!=nullptr; }
+    /**
+     * Comparison with a simple pointer, so that existing code
+     * with ==nullptr need not be changed.
+     * @param other simple pointer for comparison
+     * @return true if this pointer value equals other
+     * @stable ICU 4.4
+     */
+    bool operator==(const T *other) const { return ptr==other; }
+    /**
+     * Comparison with a simple pointer, so that existing code
+     * with !=nullptr need not be changed.
+     * @param other simple pointer for comparison
+     * @return true if this pointer value differs from other
+     * @stable ICU 4.4
+     */
+    bool operator!=(const T *other) const { return ptr!=other; }
+    /**
+     * Access without ownership change.
+     * @return the pointer value
+     * @stable ICU 4.4
+     */
+    T *getAlias() const { return ptr; }
+    /**
+     * Access without ownership change.
+     * @return the pointer value as a reference
+     * @stable ICU 4.4
+     */
+    T &operator*() const { return *ptr; }
+    /**
+     * Access without ownership change.
+     * @return the pointer value
+     * @stable ICU 4.4
+     */
+    T *operator->() const { return ptr; }
+    /**
+     * Gives up ownership; the internal pointer becomes nullptr.
+     * @return the pointer value;
+     *         caller becomes responsible for deleting the object
+     * @stable ICU 4.4
+     */
+    T *orphan() {
+        T *p=ptr;
+        ptr=nullptr;
+        return p;
+    }
+    /**
+     * Deletes the object it owns,
+     * and adopts (takes ownership of) the one passed in.
+     * Subclass must override: Base class does not delete the object.
+     * @param p simple pointer to an object that is adopted
+     * @stable ICU 4.4
+     */
+    void adoptInstead(T *p) {
+        // delete ptr;
+        ptr=p;
+    }
+protected:
+    /**
+     * Actual pointer.
+     * @internal
+     */
+    T *ptr;
+private:
+    // No comparison operators with other LocalPointerBases.
+    bool operator==(const LocalPointerBase<T> &other) = delete;
+    bool operator!=(const LocalPointerBase<T> &other) = delete;
+    // No ownership sharing: No copy constructor, no assignment operator.
+    LocalPointerBase(const LocalPointerBase<T> &other) = delete;
+    void operator=(const LocalPointerBase<T> &other) = delete;
+};
+
+/**
+ * "Smart pointer" class, deletes objects via the standard C++ delete operator.
+ * For most methods see the LocalPointerBase base class.
+ *
+ * Usage example:
+ * \code
+ * LocalPointer<UnicodeString> s(new UnicodeString((UChar32)0x50005));
+ * int32_t length=s->length();  // 2
+ * char16_t lead=s->charAt(0);  // 0xd900
+ * if(some condition) { return; }  // no need to explicitly delete the pointer
+ * s.adoptInstead(new UnicodeString((char16_t)0xfffc));
+ * length=s->length();  // 1
+ * // no need to explicitly delete the pointer
+ * \endcode
+ *
+ * @see LocalPointerBase
+ * @stable ICU 4.4
+ */
+template<typename T>
+class LocalPointer : public LocalPointerBase<T> {
+public:
+    using LocalPointerBase<T>::operator*;
+    using LocalPointerBase<T>::operator->;
+    /**
+     * Constructor takes ownership.
+     * @param p simple pointer to an object that is adopted
+     * @stable ICU 4.4
+     */
+    explicit LocalPointer(T *p=nullptr) : LocalPointerBase<T>(p) {}
+    /**
+     * Constructor takes ownership and reports an error if nullptr.
+     *
+     * This constructor is intended to be used with other-class constructors
+     * that may report a failure UErrorCode,
+     * so that callers need to check only for U_FAILURE(errorCode)
+     * and not also separately for isNull().
+     *
+     * @param p simple pointer to an object that is adopted
+     * @param errorCode in/out UErrorCode, set to U_MEMORY_ALLOCATION_ERROR
+     *     if p==nullptr and no other failure code had been set
+     * @stable ICU 55
+     */
+    LocalPointer(T *p, UErrorCode &errorCode) : LocalPointerBase<T>(p) {
+        if(p==nullptr && U_SUCCESS(errorCode)) {
+            errorCode=U_MEMORY_ALLOCATION_ERROR;
+        }
+    }
+    /**
+     * Move constructor, leaves src with isNull().
+     * @param src source smart pointer
+     * @stable ICU 56
+     */
+    LocalPointer(LocalPointer<T> &&src) noexcept : LocalPointerBase<T>(src.ptr) {
+        src.ptr=nullptr;
+    }
+
+    /**
+     * Constructs a LocalPointer from a C++11 std::unique_ptr.
+     * The LocalPointer steals the object owned by the std::unique_ptr.
+     *
+     * This constructor works via move semantics. If your std::unique_ptr is
+     * in a local variable, you must use std::move.
+     *
+     * @param p The std::unique_ptr from which the pointer will be stolen.
+     * @stable ICU 64
+     */
+    explicit LocalPointer(std::unique_ptr<T> &&p)
+        : LocalPointerBase<T>(p.release()) {}
+
+    /**
+     * Destructor deletes the object it owns.
+     * @stable ICU 4.4
+     */
+    ~LocalPointer() {
+        delete LocalPointerBase<T>::ptr;
+    }
+    /**
+     * Move assignment operator, leaves src with isNull().
+     * The behavior is undefined if *this and src are the same object.
+     * @param src source smart pointer
+     * @return *this
+     * @stable ICU 56
+     */
+    LocalPointer<T> &operator=(LocalPointer<T> &&src) noexcept {
+        delete LocalPointerBase<T>::ptr;
+        LocalPointerBase<T>::ptr=src.ptr;
+        src.ptr=nullptr;
+        return *this;
+    }
+
+    /**
+     * Move-assign from an std::unique_ptr to this LocalPointer.
+     * Steals the pointer from the std::unique_ptr.
+     *
+     * @param p The std::unique_ptr from which the pointer will be stolen.
+     * @return *this
+     * @stable ICU 64
+     */
+    LocalPointer<T> &operator=(std::unique_ptr<T> &&p) noexcept {
+        adoptInstead(p.release());
+        return *this;
+    }
+
+    /**
+     * Swap pointers.
+     * @param other other smart pointer
+     * @stable ICU 56
+     */
+    void swap(LocalPointer<T> &other) noexcept {
+        T *temp=LocalPointerBase<T>::ptr;
+        LocalPointerBase<T>::ptr=other.ptr;
+        other.ptr=temp;
+    }
+    /**
+     * Non-member LocalPointer swap function.
+     * @param p1 will get p2's pointer
+     * @param p2 will get p1's pointer
+     * @stable ICU 56
+     */
+    friend inline void swap(LocalPointer<T> &p1, LocalPointer<T> &p2) noexcept {
+        p1.swap(p2);
+    }
+    /**
+     * Deletes the object it owns,
+     * and adopts (takes ownership of) the one passed in.
+     * @param p simple pointer to an object that is adopted
+     * @stable ICU 4.4
+     */
+    void adoptInstead(T *p) {
+        delete LocalPointerBase<T>::ptr;
+        LocalPointerBase<T>::ptr=p;
+    }
+    /**
+     * Deletes the object it owns,
+     * and adopts (takes ownership of) the one passed in.
+     *
+     * If U_FAILURE(errorCode), then the current object is retained and the new one deleted.
+     *
+     * If U_SUCCESS(errorCode) but the input pointer is nullptr,
+     * then U_MEMORY_ALLOCATION_ERROR is set,
+     * the current object is deleted, and nullptr is set.
+     *
+     * @param p simple pointer to an object that is adopted
+     * @param errorCode in/out UErrorCode, set to U_MEMORY_ALLOCATION_ERROR
+     *     if p==nullptr and no other failure code had been set
+     * @stable ICU 55
+     */
+    void adoptInsteadAndCheckErrorCode(T *p, UErrorCode &errorCode) {
+        if(U_SUCCESS(errorCode)) {
+            delete LocalPointerBase<T>::ptr;
+            LocalPointerBase<T>::ptr=p;
+            if(p==nullptr) {
+                errorCode=U_MEMORY_ALLOCATION_ERROR;
+            }
+        } else {
+            delete p;
+        }
+    }
+
+    /**
+     * Conversion operator to a C++11 std::unique_ptr.
+     * Disowns the object and gives it to the returned std::unique_ptr.
+     *
+     * This operator works via move semantics. If your LocalPointer is
+     * in a local variable, you must use std::move.
+     *
+     * @return An std::unique_ptr owning the pointer previously owned by this
+     *         icu::LocalPointer.
+     * @stable ICU 64
+     */
+    operator std::unique_ptr<T> () && {
+        return std::unique_ptr<T>(LocalPointerBase<T>::orphan());
+    }
+};
+
+/**
+ * "Smart pointer" class, deletes objects via the C++ array delete[] operator.
+ * For most methods see the LocalPointerBase base class.
+ * Adds operator[] for array item access.
+ *
+ * Usage example:
+ * \code
+ * LocalArray<UnicodeString> a(new UnicodeString[2]);
+ * a[0].append((char16_t)0x61);
+ * if(some condition) { return; }  // no need to explicitly delete the array
+ * a.adoptInstead(new UnicodeString[4]);
+ * a[3].append((char16_t)0x62).append((char16_t)0x63).reverse();
+ * // no need to explicitly delete the array
+ * \endcode
+ *
+ * @see LocalPointerBase
+ * @stable ICU 4.4
+ */
+template<typename T>
+class LocalArray : public LocalPointerBase<T> {
+public:
+    using LocalPointerBase<T>::operator*;
+    using LocalPointerBase<T>::operator->;
+    /**
+     * Constructor takes ownership.
+     * @param p simple pointer to an array of T objects that is adopted
+     * @stable ICU 4.4
+     */
+    explicit LocalArray(T *p=nullptr) : LocalPointerBase<T>(p) {}
+    /**
+     * Constructor takes ownership and reports an error if nullptr.
+     *
+     * This constructor is intended to be used with other-class constructors
+     * that may report a failure UErrorCode,
+     * so that callers need to check only for U_FAILURE(errorCode)
+     * and not also separately for isNull().
+     *
+     * @param p simple pointer to an array of T objects that is adopted
+     * @param errorCode in/out UErrorCode, set to U_MEMORY_ALLOCATION_ERROR
+     *     if p==nullptr and no other failure code had been set
+     * @stable ICU 56
+     */
+    LocalArray(T *p, UErrorCode &errorCode) : LocalPointerBase<T>(p) {
+        if(p==nullptr && U_SUCCESS(errorCode)) {
+            errorCode=U_MEMORY_ALLOCATION_ERROR;
+        }
+    }
+    /**
+     * Move constructor, leaves src with isNull().
+     * @param src source smart pointer
+     * @stable ICU 56
+     */
+    LocalArray(LocalArray<T> &&src) noexcept : LocalPointerBase<T>(src.ptr) {
+        src.ptr=nullptr;
+    }
+
+    /**
+     * Constructs a LocalArray from a C++11 std::unique_ptr of an array type.
+     * The LocalPointer steals the array owned by the std::unique_ptr.
+     *
+     * This constructor works via move semantics. If your std::unique_ptr is
+     * in a local variable, you must use std::move.
+     *
+     * @param p The std::unique_ptr from which the array will be stolen.
+     * @stable ICU 64
+     */
+    explicit LocalArray(std::unique_ptr<T[]> &&p)
+        : LocalPointerBase<T>(p.release()) {}
+
+    /**
+     * Destructor deletes the array it owns.
+     * @stable ICU 4.4
+     */
+    ~LocalArray() {
+        delete[] LocalPointerBase<T>::ptr;
+    }
+    /**
+     * Move assignment operator, leaves src with isNull().
+     * The behavior is undefined if *this and src are the same object.
+     * @param src source smart pointer
+     * @return *this
+     * @stable ICU 56
+     */
+    LocalArray<T> &operator=(LocalArray<T> &&src) noexcept {
+        delete[] LocalPointerBase<T>::ptr;
+        LocalPointerBase<T>::ptr=src.ptr;
+        src.ptr=nullptr;
+        return *this;
+    }
+
+    /**
+     * Move-assign from an std::unique_ptr to this LocalPointer.
+     * Steals the array from the std::unique_ptr.
+     *
+     * @param p The std::unique_ptr from which the array will be stolen.
+     * @return *this
+     * @stable ICU 64
+     */
+    LocalArray<T> &operator=(std::unique_ptr<T[]> &&p) noexcept {
+        adoptInstead(p.release());
+        return *this;
+    }
+
+    /**
+     * Swap pointers.
+     * @param other other smart pointer
+     * @stable ICU 56
+     */
+    void swap(LocalArray<T> &other) noexcept {
+        T *temp=LocalPointerBase<T>::ptr;
+        LocalPointerBase<T>::ptr=other.ptr;
+        other.ptr=temp;
+    }
+    /**
+     * Non-member LocalArray swap function.
+     * @param p1 will get p2's pointer
+     * @param p2 will get p1's pointer
+     * @stable ICU 56
+     */
+    friend inline void swap(LocalArray<T> &p1, LocalArray<T> &p2) noexcept {
+        p1.swap(p2);
+    }
+    /**
+     * Deletes the array it owns,
+     * and adopts (takes ownership of) the one passed in.
+     * @param p simple pointer to an array of T objects that is adopted
+     * @stable ICU 4.4
+     */
+    void adoptInstead(T *p) {
+        delete[] LocalPointerBase<T>::ptr;
+        LocalPointerBase<T>::ptr=p;
+    }
+    /**
+     * Deletes the array it owns,
+     * and adopts (takes ownership of) the one passed in.
+     *
+     * If U_FAILURE(errorCode), then the current array is retained and the new one deleted.
+     *
+     * If U_SUCCESS(errorCode) but the input pointer is nullptr,
+     * then U_MEMORY_ALLOCATION_ERROR is set,
+     * the current array is deleted, and nullptr is set.
+     *
+     * @param p simple pointer to an array of T objects that is adopted
+     * @param errorCode in/out UErrorCode, set to U_MEMORY_ALLOCATION_ERROR
+     *     if p==nullptr and no other failure code had been set
+     * @stable ICU 56
+     */
+    void adoptInsteadAndCheckErrorCode(T *p, UErrorCode &errorCode) {
+        if(U_SUCCESS(errorCode)) {
+            delete[] LocalPointerBase<T>::ptr;
+            LocalPointerBase<T>::ptr=p;
+            if(p==nullptr) {
+                errorCode=U_MEMORY_ALLOCATION_ERROR;
+            }
+        } else {
+            delete[] p;
+        }
+    }
+    /**
+     * Array item access (writable).
+     * No index bounds check.
+     * @param i array index
+     * @return reference to the array item
+     * @stable ICU 4.4
+     */
+    T &operator[](ptrdiff_t i) const { return LocalPointerBase<T>::ptr[i]; }
+
+    /**
+     * Conversion operator to a C++11 std::unique_ptr.
+     * Disowns the object and gives it to the returned std::unique_ptr.
+     *
+     * This operator works via move semantics. If your LocalPointer is
+     * in a local variable, you must use std::move.
+     *
+     * @return An std::unique_ptr owning the pointer previously owned by this
+     *         icu::LocalPointer.
+     * @stable ICU 64
+     */
+    operator std::unique_ptr<T[]> () && {
+        return std::unique_ptr<T[]>(LocalPointerBase<T>::orphan());
+    }
+};
+
+/**
+ * \def U_DEFINE_LOCAL_OPEN_POINTER
+ * "Smart pointer" definition macro, deletes objects via the closeFunction.
+ * Defines a subclass of LocalPointerBase which works just
+ * like LocalPointer<Type> except that this subclass will use the closeFunction
+ * rather than the C++ delete operator.
+ *
+ * Usage example:
+ * \code
+ * LocalUCaseMapPointer csm(ucasemap_open(localeID, options, &errorCode));
+ * utf8OutLength=ucasemap_utf8ToLower(csm.getAlias(),
+ *     utf8Out, (int32_t)sizeof(utf8Out),
+ *     utf8In, utf8InLength, &errorCode);
+ * if(U_FAILURE(errorCode)) { return; }  // no need to explicitly delete the UCaseMap
+ * \endcode
+ *
+ * @see LocalPointerBase
+ * @see LocalPointer
+ * @stable ICU 4.4
+ */
+#define U_DEFINE_LOCAL_OPEN_POINTER(LocalPointerClassName, Type, closeFunction) \
+    using LocalPointerClassName = internal::LocalOpenPointer<Type, closeFunction>
+
+#ifndef U_IN_DOXYGEN
+namespace internal {
+/**
+ * Implementation, do not use directly: use U_DEFINE_LOCAL_OPEN_POINTER.
+ *
+ * @see U_DEFINE_LOCAL_OPEN_POINTER
+ * @internal
+ */
+template <typename Type, auto closeFunction>
+class LocalOpenPointer : public LocalPointerBase<Type> {
+    using LocalPointerBase<Type>::ptr;
+public:
+    using LocalPointerBase<Type>::operator*;
+    using LocalPointerBase<Type>::operator->;
+    explicit LocalOpenPointer(Type *p=nullptr) : LocalPointerBase<Type>(p) {}
+    LocalOpenPointer(LocalOpenPointer &&src) noexcept
+            : LocalPointerBase<Type>(src.ptr) {
+        src.ptr=nullptr;
+    }
+    /* TODO: Be agnostic of the deleter function signature from the user-provided std::unique_ptr? */
+    explicit LocalOpenPointer(std::unique_ptr<Type, decltype(closeFunction)> &&p)
+            : LocalPointerBase<Type>(p.release()) {}
+    ~LocalOpenPointer() { if (ptr != nullptr) { closeFunction(ptr); } }
+    LocalOpenPointer &operator=(LocalOpenPointer &&src) noexcept {
+        if (ptr != nullptr) { closeFunction(ptr); }
+        LocalPointerBase<Type>::ptr=src.ptr;
+        src.ptr=nullptr;
+        return *this;
+    }
+    /* TODO: Be agnostic of the deleter function signature from the user-provided std::unique_ptr? */
+    LocalOpenPointer &operator=(std::unique_ptr<Type, decltype(closeFunction)> &&p) {
+        adoptInstead(p.release());
+        return *this;
+    }
+    void swap(LocalOpenPointer &other) noexcept {
+        Type *temp=LocalPointerBase<Type>::ptr;
+        LocalPointerBase<Type>::ptr=other.ptr;
+        other.ptr=temp;
+    }
+    friend inline void swap(LocalOpenPointer &p1, LocalOpenPointer &p2) noexcept {
+        p1.swap(p2);
+    }
+    void adoptInstead(Type *p) {
+        if (ptr != nullptr) { closeFunction(ptr); }
+        ptr=p;
+    }
+    operator std::unique_ptr<Type, decltype(closeFunction)> () && {
+        return std::unique_ptr<Type, decltype(closeFunction)>(LocalPointerBase<Type>::orphan(), closeFunction);
+    }
+};
+}  // namespace internal
+#endif
+
+U_NAMESPACE_END
+
+#endif  /* U_SHOW_CPLUSPLUS_API */
+#endif  /* __LOCALPOINTER_H__ */

thirdparty/icu4c/common/unicode/locdspnm.h

@@ -0,0 +1,211 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+/*
+******************************************************************************
+* Copyright (C) 2010-2016, International Business Machines Corporation and
+* others. All Rights Reserved.
+******************************************************************************
+*/
+
+#ifndef LOCDSPNM_H
+#define LOCDSPNM_H
+
+#include "unicode/utypes.h"
+
+#if U_SHOW_CPLUSPLUS_API
+
+/**
+ * \file
+ * \brief C++ API: Provides display names of Locale and its components.
+ */
+
+#if !UCONFIG_NO_FORMATTING
+
+#include "unicode/locid.h"
+#include "unicode/strenum.h"
+#include "unicode/uscript.h"
+#include "unicode/uldnames.h"
+#include "unicode/udisplaycontext.h"
+
+U_NAMESPACE_BEGIN
+
+/**
+ * Returns display names of Locales and components of Locales. For
+ * more information on language, script, region, variant, key, and
+ * values, see Locale.
+ * @stable ICU 4.4
+ */
+class U_COMMON_API LocaleDisplayNames : public UObject {
+public:
+    /**
+     * Destructor.
+     * @stable ICU 4.4
+     */
+    virtual ~LocaleDisplayNames();
+
+    /**
+     * Convenience overload of
+     * {@link #createInstance(const Locale& locale, UDialectHandling dialectHandling)}
+     * that specifies STANDARD dialect handling.
+     * @param locale the display locale
+     * @return a LocaleDisplayNames instance
+     * @stable ICU 4.4
+     */
+    inline static LocaleDisplayNames* U_EXPORT2 createInstance(const Locale& locale);
+
+    /**
+     * Returns an instance of LocaleDisplayNames that returns names
+     * formatted for the provided locale, using the provided
+     * dialectHandling.
+     *
+     * @param locale the display locale
+     * @param dialectHandling how to select names for locales
+     * @return a LocaleDisplayNames instance
+     * @stable ICU 4.4
+     */
+    static LocaleDisplayNames* U_EXPORT2 createInstance(const Locale& locale,
+                            UDialectHandling dialectHandling);
+
+    /**
+     * Returns an instance of LocaleDisplayNames that returns names formatted
+     * for the provided locale, using the provided UDisplayContext settings.
+     *
+     * @param locale the display locale
+     * @param contexts List of one or more context settings (e.g. for dialect
+     *               handling, capitalization, etc.
+     * @param length Number of items in the contexts list
+     * @return a LocaleDisplayNames instance
+     * @stable ICU 51
+     */
+    static LocaleDisplayNames* U_EXPORT2 createInstance(const Locale& locale,
+                            UDisplayContext *contexts, int32_t length);
+
+    // getters for state
+    /**
+     * Returns the locale used to determine the display names. This is
+     * not necessarily the same locale passed to {@link #createInstance}.
+     * @return the display locale
+     * @stable ICU 4.4
+     */
+    virtual const Locale& getLocale() const = 0;
+
+    /**
+     * Returns the dialect handling used in the display names.
+     * @return the dialect handling enum
+     * @stable ICU 4.4
+     */
+    virtual UDialectHandling getDialectHandling() const = 0;
+
+    /**
+     * Returns the UDisplayContext value for the specified UDisplayContextType.
+     * @param type the UDisplayContextType whose value to return
+     * @return the UDisplayContext for the specified type.
+     * @stable ICU 51
+     */
+    virtual UDisplayContext getContext(UDisplayContextType type) const = 0;
+
+    // names for entire locales
+    /**
+     * Returns the display name of the provided locale.
+     * @param locale the locale whose display name to return
+     * @param result receives the locale's display name
+     * @return the display name of the provided locale
+     * @stable ICU 4.4
+     */
+    virtual UnicodeString& localeDisplayName(const Locale& locale,
+                         UnicodeString& result) const = 0;
+
+    /**
+     * Returns the display name of the provided locale id.
+     * @param localeId the id of the locale whose display name to return
+     * @param result receives the locale's display name
+     * @return the display name of the provided locale
+     * @stable ICU 4.4
+     */
+    virtual UnicodeString& localeDisplayName(const char* localeId,
+                         UnicodeString& result) const = 0;
+
+    // names for components of a locale id
+    /**
+     * Returns the display name of the provided language code.
+     * @param lang the language code
+     * @param result receives the language code's display name
+     * @return the display name of the provided language code
+     * @stable ICU 4.4
+     */
+    virtual UnicodeString& languageDisplayName(const char* lang,
+                           UnicodeString& result) const = 0;
+
+    /**
+     * Returns the display name of the provided script code.
+     * @param script the script code
+     * @param result receives the script code's display name
+     * @return the display name of the provided script code
+     * @stable ICU 4.4
+     */
+    virtual UnicodeString& scriptDisplayName(const char* script,
+                         UnicodeString& result) const = 0;
+
+    /**
+     * Returns the display name of the provided script code.
+     * @param scriptCode the script code number
+     * @param result receives the script code's display name
+     * @return the display name of the provided script code
+     * @stable ICU 4.4
+     */
+    virtual UnicodeString& scriptDisplayName(UScriptCode scriptCode,
+                         UnicodeString& result) const = 0;
+
+    /**
+     * Returns the display name of the provided region code.
+     * @param region the region code
+     * @param result receives the region code's display name
+     * @return the display name of the provided region code
+     * @stable ICU 4.4
+     */
+    virtual UnicodeString& regionDisplayName(const char* region,
+                         UnicodeString& result) const = 0;
+
+    /**
+     * Returns the display name of the provided variant.
+     * @param variant the variant string
+     * @param result receives the variant's display name
+     * @return the display name of the provided variant
+     * @stable ICU 4.4
+     */
+    virtual UnicodeString& variantDisplayName(const char* variant,
+                          UnicodeString& result) const = 0;
+
+    /**
+     * Returns the display name of the provided locale key.
+     * @param key the locale key name
+     * @param result receives the locale key's display name
+     * @return the display name of the provided locale key
+     * @stable ICU 4.4
+     */
+    virtual UnicodeString& keyDisplayName(const char* key,
+                      UnicodeString& result) const = 0;
+
+    /**
+     * Returns the display name of the provided value (used with the provided key).
+     * @param key the locale key name
+     * @param value the locale key's value
+     * @param result receives the value's display name
+     * @return the display name of the provided value
+     * @stable ICU 4.4
+     */
+    virtual UnicodeString& keyValueDisplayName(const char* key, const char* value,
+                           UnicodeString& result) const = 0;
+};
+
+inline LocaleDisplayNames* LocaleDisplayNames::createInstance(const Locale& locale) {
+  return LocaleDisplayNames::createInstance(locale, ULDN_STANDARD_NAMES);
+}
+
+U_NAMESPACE_END
+
+#endif
+
+#endif /* U_SHOW_CPLUSPLUS_API */
+
+#endif

thirdparty/icu4c/common/unicode/messagepattern.h

@@ -0,0 +1,950 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+/*
+*******************************************************************************
+*   Copyright (C) 2011-2013, International Business Machines
+*   Corporation and others.  All Rights Reserved.
+*******************************************************************************
+*   file name:  messagepattern.h
+*   encoding:   UTF-8
+*   tab size:   8 (not used)
+*   indentation:4
+*
+*   created on: 2011mar14
+*   created by: Markus W. Scherer
+*/
+
+#ifndef __MESSAGEPATTERN_H__
+#define __MESSAGEPATTERN_H__
+
+/**
+ * \file
+ * \brief C++ API: MessagePattern class: Parses and represents ICU MessageFormat patterns.
+ */
+
+#include "unicode/utypes.h"
+
+#if U_SHOW_CPLUSPLUS_API
+
+#if !UCONFIG_NO_FORMATTING
+
+#include "unicode/parseerr.h"
+#include "unicode/unistr.h"
+
+/**
+ * Mode for when an apostrophe starts quoted literal text for MessageFormat output.
+ * The default is DOUBLE_OPTIONAL unless overridden via uconfig.h
+ * (UCONFIG_MSGPAT_DEFAULT_APOSTROPHE_MODE).
+ * <p>
+ * A pair of adjacent apostrophes always results in a single apostrophe in the output,
+ * even when the pair is between two single, text-quoting apostrophes.
+ * <p>
+ * The following table shows examples of desired MessageFormat.format() output
+ * with the pattern strings that yield that output.
+ * <p>
+ * <table>
+ *   <tr>
+ *     <th>Desired output</th>
+ *     <th>DOUBLE_OPTIONAL</th>
+ *     <th>DOUBLE_REQUIRED</th>
+ *   </tr>
+ *   <tr>
+ *     <td>I see {many}</td>
+ *     <td>I see '{many}'</td>
+ *     <td>(same)</td>
+ *   </tr>
+ *   <tr>
+ *     <td>I said {'Wow!'}</td>
+ *     <td>I said '{''Wow!''}'</td>
+ *     <td>(same)</td>
+ *   </tr>
+ *   <tr>
+ *     <td>I don't know</td>
+ *     <td>I don't know OR<br> I don''t know</td>
+ *     <td>I don''t know</td>
+ *   </tr>
+ * </table>
+ * @stable ICU 4.8
+ * @see UCONFIG_MSGPAT_DEFAULT_APOSTROPHE_MODE
+ */
+enum UMessagePatternApostropheMode {
+    /**
+     * A literal apostrophe is represented by
+     * either a single or a double apostrophe pattern character.
+     * Within a MessageFormat pattern, a single apostrophe only starts quoted literal text
+     * if it immediately precedes a curly brace {},
+     * or a pipe symbol | if inside a choice format,
+     * or a pound symbol # if inside a plural format.
+     * <p>
+     * This is the default behavior starting with ICU 4.8.
+     * @stable ICU 4.8
+     */
+    UMSGPAT_APOS_DOUBLE_OPTIONAL,
+    /**
+     * A literal apostrophe must be represented by
+     * a double apostrophe pattern character.
+     * A single apostrophe always starts quoted literal text.
+     * <p>
+     * This is the behavior of ICU 4.6 and earlier, and of the JDK.
+     * @stable ICU 4.8
+     */
+    UMSGPAT_APOS_DOUBLE_REQUIRED
+};
+/**
+ * @stable ICU 4.8
+ */
+typedef enum UMessagePatternApostropheMode UMessagePatternApostropheMode;
+
+/**
+ * MessagePattern::Part type constants.
+ * @stable ICU 4.8
+ */
+enum UMessagePatternPartType {
+    /**
+     * Start of a message pattern (main or nested).
+     * The length is 0 for the top-level message
+     * and for a choice argument sub-message, otherwise 1 for the '{'.
+     * The value indicates the nesting level, starting with 0 for the main message.
+     * <p>
+     * There is always a later MSG_LIMIT part.
+     * @stable ICU 4.8
+     */
+    UMSGPAT_PART_TYPE_MSG_START,
+    /**
+     * End of a message pattern (main or nested).
+     * The length is 0 for the top-level message and
+     * the last sub-message of a choice argument,
+     * otherwise 1 for the '}' or (in a choice argument style) the '|'.
+     * The value indicates the nesting level, starting with 0 for the main message.
+     * @stable ICU 4.8
+     */
+    UMSGPAT_PART_TYPE_MSG_LIMIT,
+    /**
+     * Indicates a substring of the pattern string which is to be skipped when formatting.
+     * For example, an apostrophe that begins or ends quoted text
+     * would be indicated with such a part.
+     * The value is undefined and currently always 0.
+     * @stable ICU 4.8
+     */
+    UMSGPAT_PART_TYPE_SKIP_SYNTAX,
+    /**
+     * Indicates that a syntax character needs to be inserted for auto-quoting.
+     * The length is 0.
+     * The value is the character code of the insertion character. (U+0027=APOSTROPHE)
+     * @stable ICU 4.8
+     */
+    UMSGPAT_PART_TYPE_INSERT_CHAR,
+    /**
+     * Indicates a syntactic (non-escaped) # symbol in a plural variant.
+     * When formatting, replace this part's substring with the
+     * (value-offset) for the plural argument value.
+     * The value is undefined and currently always 0.
+     * @stable ICU 4.8
+     */
+    UMSGPAT_PART_TYPE_REPLACE_NUMBER,
+    /**
+     * Start of an argument.
+     * The length is 1 for the '{'.
+     * The value is the ordinal value of the ArgType. Use getArgType().
+     * <p>
+     * This part is followed by either an ARG_NUMBER or ARG_NAME,
+     * followed by optional argument sub-parts (see UMessagePatternArgType constants)
+     * and finally an ARG_LIMIT part.
+     * @stable ICU 4.8
+     */
+    UMSGPAT_PART_TYPE_ARG_START,
+    /**
+     * End of an argument.
+     * The length is 1 for the '}'.
+     * The value is the ordinal value of the ArgType. Use getArgType().
+     * @stable ICU 4.8
+     */
+    UMSGPAT_PART_TYPE_ARG_LIMIT,
+    /**
+     * The argument number, provided by the value.
+     * @stable ICU 4.8
+     */
+    UMSGPAT_PART_TYPE_ARG_NUMBER,
+    /**
+     * The argument name.
+     * The value is undefined and currently always 0.
+     * @stable ICU 4.8
+     */
+    UMSGPAT_PART_TYPE_ARG_NAME,
+    /**
+     * The argument type.
+     * The value is undefined and currently always 0.
+     * @stable ICU 4.8
+     */
+    UMSGPAT_PART_TYPE_ARG_TYPE,
+    /**
+     * The argument style text.
+     * The value is undefined and currently always 0.
+     * @stable ICU 4.8
+     */
+    UMSGPAT_PART_TYPE_ARG_STYLE,
+    /**
+     * A selector substring in a "complex" argument style.
+     * The value is undefined and currently always 0.
+     * @stable ICU 4.8
+     */
+    UMSGPAT_PART_TYPE_ARG_SELECTOR,
+    /**
+     * An integer value, for example the offset or an explicit selector value
+     * in a PluralFormat style.
+     * The part value is the integer value.
+     * @stable ICU 4.8
+     */
+    UMSGPAT_PART_TYPE_ARG_INT,
+    /**
+     * A numeric value, for example the offset or an explicit selector value
+     * in a PluralFormat style.
+     * The part value is an index into an internal array of numeric values;
+     * use getNumericValue().
+     * @stable ICU 4.8
+     */
+    UMSGPAT_PART_TYPE_ARG_DOUBLE
+};
+/**
+ * @stable ICU 4.8
+ */
+typedef enum UMessagePatternPartType UMessagePatternPartType;
+
+/**
+ * Argument type constants.
+ * Returned by Part.getArgType() for ARG_START and ARG_LIMIT parts.
+ *
+ * Messages nested inside an argument are each delimited by MSG_START and MSG_LIMIT,
+ * with a nesting level one greater than the surrounding message.
+ * @stable ICU 4.8
+ */
+enum UMessagePatternArgType {
+    /**
+     * The argument has no specified type.
+     * @stable ICU 4.8
+     */
+    UMSGPAT_ARG_TYPE_NONE,
+    /**
+     * The argument has a "simple" type which is provided by the ARG_TYPE part.
+     * An ARG_STYLE part might follow that.
+     * @stable ICU 4.8
+     */
+    UMSGPAT_ARG_TYPE_SIMPLE,
+    /**
+     * The argument is a ChoiceFormat with one or more
+     * ((ARG_INT | ARG_DOUBLE), ARG_SELECTOR, message) tuples.
+     * @stable ICU 4.8
+     */
+    UMSGPAT_ARG_TYPE_CHOICE,
+    /**
+     * The argument is a cardinal-number PluralFormat with an optional ARG_INT or ARG_DOUBLE offset
+     * (e.g., offset:1)
+     * and one or more (ARG_SELECTOR [explicit-value] message) tuples.
+     * If the selector has an explicit value (e.g., =2), then
+     * that value is provided by the ARG_INT or ARG_DOUBLE part preceding the message.
+     * Otherwise the message immediately follows the ARG_SELECTOR.
+     * @stable ICU 4.8
+     */
+    UMSGPAT_ARG_TYPE_PLURAL,
+    /**
+     * The argument is a SelectFormat with one or more (ARG_SELECTOR, message) pairs.
+     * @stable ICU 4.8
+     */
+    UMSGPAT_ARG_TYPE_SELECT,
+    /**
+     * The argument is an ordinal-number PluralFormat
+     * with the same style parts sequence and semantics as UMSGPAT_ARG_TYPE_PLURAL.
+     * @stable ICU 50
+     */
+    UMSGPAT_ARG_TYPE_SELECTORDINAL
+};
+/**
+ * @stable ICU 4.8
+ */
+typedef enum UMessagePatternArgType UMessagePatternArgType;
+
+/**
+ * \def UMSGPAT_ARG_TYPE_HAS_PLURAL_STYLE
+ * Returns true if the argument type has a plural style part sequence and semantics,
+ * for example UMSGPAT_ARG_TYPE_PLURAL and UMSGPAT_ARG_TYPE_SELECTORDINAL.
+ * @stable ICU 50
+ */
+#define UMSGPAT_ARG_TYPE_HAS_PLURAL_STYLE(argType) \
+    ((argType)==UMSGPAT_ARG_TYPE_PLURAL || (argType)==UMSGPAT_ARG_TYPE_SELECTORDINAL)
+
+enum {
+    /**
+     * Return value from MessagePattern.validateArgumentName() for when
+     * the string is a valid "pattern identifier" but not a number.
+     * @stable ICU 4.8
+     */
+    UMSGPAT_ARG_NAME_NOT_NUMBER=-1,
+
+    /**
+     * Return value from MessagePattern.validateArgumentName() for when
+     * the string is invalid.
+     * It might not be a valid "pattern identifier",
+     * or it have only ASCII digits but there is a leading zero or the number is too large.
+     * @stable ICU 4.8
+     */
+    UMSGPAT_ARG_NAME_NOT_VALID=-2
+};
+
+/**
+ * Special value that is returned by getNumericValue(Part) when no
+ * numeric value is defined for a part.
+ * @see MessagePattern.getNumericValue()
+ * @stable ICU 4.8
+ */
+#define UMSGPAT_NO_NUMERIC_VALUE ((double)(-123456789))
+
+U_NAMESPACE_BEGIN
+
+class MessagePatternDoubleList;
+class MessagePatternPartsList;
+
+/**
+ * Parses and represents ICU MessageFormat patterns.
+ * Also handles patterns for ChoiceFormat, PluralFormat and SelectFormat.
+ * Used in the implementations of those classes as well as in tools
+ * for message validation, translation and format conversion.
+ * <p>
+ * The parser handles all syntax relevant for identifying message arguments.
+ * This includes "complex" arguments whose style strings contain
+ * nested MessageFormat pattern substrings.
+ * For "simple" arguments (with no nested MessageFormat pattern substrings),
+ * the argument style is not parsed any further.
+ * <p>
+ * The parser handles named and numbered message arguments and allows both in one message.
+ * <p>
+ * Once a pattern has been parsed successfully, iterate through the parsed data
+ * with countParts(), getPart() and related methods.
+ * <p>
+ * The data logically represents a parse tree, but is stored and accessed
+ * as a list of "parts" for fast and simple parsing and to minimize object allocations.
+ * Arguments and nested messages are best handled via recursion.
+ * For every _START "part", MessagePattern.getLimitPartIndex() efficiently returns
+ * the index of the corresponding _LIMIT "part".
+ * <p>
+ * List of "parts":
+ * <pre>
+ * message = MSG_START (SKIP_SYNTAX | INSERT_CHAR | REPLACE_NUMBER | argument)* MSG_LIMIT
+ * argument = noneArg | simpleArg | complexArg
+ * complexArg = choiceArg | pluralArg | selectArg
+ *
+ * noneArg = ARG_START.NONE (ARG_NAME | ARG_NUMBER) ARG_LIMIT.NONE
+ * simpleArg = ARG_START.SIMPLE (ARG_NAME | ARG_NUMBER) ARG_TYPE [ARG_STYLE] ARG_LIMIT.SIMPLE
+ * choiceArg = ARG_START.CHOICE (ARG_NAME | ARG_NUMBER) choiceStyle ARG_LIMIT.CHOICE
+ * pluralArg = ARG_START.PLURAL (ARG_NAME | ARG_NUMBER) pluralStyle ARG_LIMIT.PLURAL
+ * selectArg = ARG_START.SELECT (ARG_NAME | ARG_NUMBER) selectStyle ARG_LIMIT.SELECT
+ *
+ * choiceStyle = ((ARG_INT | ARG_DOUBLE) ARG_SELECTOR message)+
+ * pluralStyle = [ARG_INT | ARG_DOUBLE] (ARG_SELECTOR [ARG_INT | ARG_DOUBLE] message)+
+ * selectStyle = (ARG_SELECTOR message)+
+ * </pre>
+ * <ul>
+ *   <li>Literal output text is not represented directly by "parts" but accessed
+ *       between parts of a message, from one part's getLimit() to the next part's getIndex().
+ *   <li><code>ARG_START.CHOICE</code> stands for an ARG_START Part with ArgType CHOICE.
+ *   <li>In the choiceStyle, the ARG_SELECTOR has the '<', the '#' or
+ *       the less-than-or-equal-to sign (U+2264).
+ *   <li>In the pluralStyle, the first, optional numeric Part has the "offset:" value.
+ *       The optional numeric Part between each (ARG_SELECTOR, message) pair
+ *       is the value of an explicit-number selector like "=2",
+ *       otherwise the selector is a non-numeric identifier.
+ *   <li>The REPLACE_NUMBER Part can occur only in an immediate sub-message of the pluralStyle.
+ * </ul>
+ * <p>
+ * This class is not intended for public subclassing.
+ *
+ * @stable ICU 4.8
+ */
+class U_COMMON_API MessagePattern : public UObject {
+public:
+    /**
+     * Constructs an empty MessagePattern with default UMessagePatternApostropheMode.
+     * @param errorCode Standard ICU error code. Its input value must
+     *                  pass the U_SUCCESS() test, or else the function returns
+     *                  immediately. Check for U_FAILURE() on output or use with
+     *                  function chaining. (See User Guide for details.)
+     * @stable ICU 4.8
+     */
+    MessagePattern(UErrorCode &errorCode);
+
+    /**
+     * Constructs an empty MessagePattern.
+     * @param mode Explicit UMessagePatternApostropheMode.
+     * @param errorCode Standard ICU error code. Its input value must
+     *                  pass the U_SUCCESS() test, or else the function returns
+     *                  immediately. Check for U_FAILURE() on output or use with
+     *                  function chaining. (See User Guide for details.)
+     * @stable ICU 4.8
+     */
+    MessagePattern(UMessagePatternApostropheMode mode, UErrorCode &errorCode);
+
+    /**
+     * Constructs a MessagePattern with default UMessagePatternApostropheMode and
+     * parses the MessageFormat pattern string.
+     * @param pattern a MessageFormat pattern string
+     * @param parseError Struct to receive information on the position
+     *                   of an error within the pattern.
+     *                   Can be nullptr.
+     * @param errorCode Standard ICU error code. Its input value must
+     *                  pass the U_SUCCESS() test, or else the function returns
+     *                  immediately. Check for U_FAILURE() on output or use with
+     *                  function chaining. (See User Guide for details.)
+     * TODO: turn @throws into UErrorCode specifics?
+     * @throws IllegalArgumentException for syntax errors in the pattern string
+     * @throws IndexOutOfBoundsException if certain limits are exceeded
+     *         (e.g., argument number too high, argument name too long, etc.)
+     * @throws NumberFormatException if a number could not be parsed
+     * @stable ICU 4.8
+     */
+    MessagePattern(const UnicodeString &pattern, UParseError *parseError, UErrorCode &errorCode);
+
+    /**
+     * Copy constructor.
+     * @param other Object to copy.
+     * @stable ICU 4.8
+     */
+    MessagePattern(const MessagePattern &other);
+
+    /**
+     * Assignment operator.
+     * @param other Object to copy.
+     * @return *this=other
+     * @stable ICU 4.8
+     */
+    MessagePattern &operator=(const MessagePattern &other);
+
+    /**
+     * Destructor.
+     * @stable ICU 4.8
+     */
+    virtual ~MessagePattern();
+
+    /**
+     * Parses a MessageFormat pattern string.
+     * @param pattern a MessageFormat pattern string
+     * @param parseError Struct to receive information on the position
+     *                   of an error within the pattern.
+     *                   Can be nullptr.
+     * @param errorCode Standard ICU error code. Its input value must
+     *                  pass the U_SUCCESS() test, or else the function returns
+     *                  immediately. Check for U_FAILURE() on output or use with
+     *                  function chaining. (See User Guide for details.)
+     * @return *this
+     * @throws IllegalArgumentException for syntax errors in the pattern string
+     * @throws IndexOutOfBoundsException if certain limits are exceeded
+     *         (e.g., argument number too high, argument name too long, etc.)
+     * @throws NumberFormatException if a number could not be parsed
+     * @stable ICU 4.8
+     */
+    MessagePattern &parse(const UnicodeString &pattern,
+                          UParseError *parseError, UErrorCode &errorCode);
+
+    /**
+     * Parses a ChoiceFormat pattern string.
+     * @param pattern a ChoiceFormat pattern string
+     * @param parseError Struct to receive information on the position
+     *                   of an error within the pattern.
+     *                   Can be nullptr.
+     * @param errorCode Standard ICU error code. Its input value must
+     *                  pass the U_SUCCESS() test, or else the function returns
+     *                  immediately. Check for U_FAILURE() on output or use with
+     *                  function chaining. (See User Guide for details.)
+     * @return *this
+     * @throws IllegalArgumentException for syntax errors in the pattern string
+     * @throws IndexOutOfBoundsException if certain limits are exceeded
+     *         (e.g., argument number too high, argument name too long, etc.)
+     * @throws NumberFormatException if a number could not be parsed
+     * @stable ICU 4.8
+     */
+    MessagePattern &parseChoiceStyle(const UnicodeString &pattern,
+                                     UParseError *parseError, UErrorCode &errorCode);
+
+    /**
+     * Parses a PluralFormat pattern string.
+     * @param pattern a PluralFormat pattern string
+     * @param parseError Struct to receive information on the position
+     *                   of an error within the pattern.
+     *                   Can be nullptr.
+     * @param errorCode Standard ICU error code. Its input value must
+     *                  pass the U_SUCCESS() test, or else the function returns
+     *                  immediately. Check for U_FAILURE() on output or use with
+     *                  function chaining. (See User Guide for details.)
+     * @return *this
+     * @throws IllegalArgumentException for syntax errors in the pattern string
+     * @throws IndexOutOfBoundsException if certain limits are exceeded
+     *         (e.g., argument number too high, argument name too long, etc.)
+     * @throws NumberFormatException if a number could not be parsed
+     * @stable ICU 4.8
+     */
+    MessagePattern &parsePluralStyle(const UnicodeString &pattern,
+                                     UParseError *parseError, UErrorCode &errorCode);
+
+    /**
+     * Parses a SelectFormat pattern string.
+     * @param pattern a SelectFormat pattern string
+     * @param parseError Struct to receive information on the position
+     *                   of an error within the pattern.
+     *                   Can be nullptr.
+     * @param errorCode Standard ICU error code. Its input value must
+     *                  pass the U_SUCCESS() test, or else the function returns
+     *                  immediately. Check for U_FAILURE() on output or use with
+     *                  function chaining. (See User Guide for details.)
+     * @return *this
+     * @throws IllegalArgumentException for syntax errors in the pattern string
+     * @throws IndexOutOfBoundsException if certain limits are exceeded
+     *         (e.g., argument number too high, argument name too long, etc.)
+     * @throws NumberFormatException if a number could not be parsed
+     * @stable ICU 4.8
+     */
+    MessagePattern &parseSelectStyle(const UnicodeString &pattern,
+                                     UParseError *parseError, UErrorCode &errorCode);
+
+    /**
+     * Clears this MessagePattern.
+     * countParts() will return 0.
+     * @stable ICU 4.8
+     */
+    void clear();
+
+    /**
+     * Clears this MessagePattern and sets the UMessagePatternApostropheMode.
+     * countParts() will return 0.
+     * @param mode The new UMessagePatternApostropheMode.
+     * @stable ICU 4.8
+     */
+    void clearPatternAndSetApostropheMode(UMessagePatternApostropheMode mode) {
+        clear();
+        aposMode=mode;
+    }
+
+    /**
+     * @param other another object to compare with.
+     * @return true if this object is equivalent to the other one.
+     * @stable ICU 4.8
+     */
+    bool operator==(const MessagePattern &other) const;
+
+    /**
+     * @param other another object to compare with.
+     * @return false if this object is equivalent to the other one.
+     * @stable ICU 4.8
+     */
+    inline bool operator!=(const MessagePattern &other) const {
+        return !operator==(other);
+    }
+
+    /**
+     * @return A hash code for this object.
+     * @stable ICU 4.8
+     */
+    int32_t hashCode() const;
+
+    /**
+     * @return this instance's UMessagePatternApostropheMode.
+     * @stable ICU 4.8
+     */
+    UMessagePatternApostropheMode getApostropheMode() const {
+        return aposMode;
+    }
+
+    // Java has package-private jdkAposMode() here.
+    // In C++, this is declared in the MessageImpl class.
+
+    /**
+     * @return the parsed pattern string (null if none was parsed).
+     * @stable ICU 4.8
+     */
+    const UnicodeString &getPatternString() const {
+        return msg;
+    }
+
+    /**
+     * Does the parsed pattern have named arguments like {first_name}?
+     * @return true if the parsed pattern has at least one named argument.
+     * @stable ICU 4.8
+     */
+    UBool hasNamedArguments() const {
+        return hasArgNames;
+    }
+
+    /**
+     * Does the parsed pattern have numbered arguments like {2}?
+     * @return true if the parsed pattern has at least one numbered argument.
+     * @stable ICU 4.8
+     */
+    UBool hasNumberedArguments() const {
+        return hasArgNumbers;
+    }
+
+    /**
+     * Validates and parses an argument name or argument number string.
+     * An argument name must be a "pattern identifier", that is, it must contain
+     * no Unicode Pattern_Syntax or Pattern_White_Space characters.
+     * If it only contains ASCII digits, then it must be a small integer with no leading zero.
+     * @param name Input string.
+     * @return &gt;=0 if the name is a valid number,
+     *         ARG_NAME_NOT_NUMBER (-1) if it is a "pattern identifier" but not all ASCII digits,
+     *         ARG_NAME_NOT_VALID (-2) if it is neither.
+     * @stable ICU 4.8
+     */
+    static int32_t validateArgumentName(const UnicodeString &name);
+
+    /**
+     * Returns a version of the parsed pattern string where each ASCII apostrophe
+     * is doubled (escaped) if it is not already, and if it is not interpreted as quoting syntax.
+     * <p>
+     * For example, this turns "I don't '{know}' {gender,select,female{h''er}other{h'im}}."
+     * into "I don''t '{know}' {gender,select,female{h''er}other{h''im}}."
+     * @return the deep-auto-quoted version of the parsed pattern string.
+     * @see MessageFormat.autoQuoteApostrophe()
+     * @stable ICU 4.8
+     */
+    UnicodeString autoQuoteApostropheDeep() const;
+
+    class Part;
+
+    /**
+     * Returns the number of "parts" created by parsing the pattern string.
+     * Returns 0 if no pattern has been parsed or clear() was called.
+     * @return the number of pattern parts.
+     * @stable ICU 4.8
+     */
+    int32_t countParts() const {
+        return partsLength;
+    }
+
+    /**
+     * Gets the i-th pattern "part".
+     * @param i The index of the Part data. (0..countParts()-1)
+     * @return the i-th pattern "part".
+     * @stable ICU 4.8
+     */
+    const Part &getPart(int32_t i) const {
+        return parts[i];
+    }
+
+    /**
+     * Returns the UMessagePatternPartType of the i-th pattern "part".
+     * Convenience method for getPart(i).getType().
+     * @param i The index of the Part data. (0..countParts()-1)
+     * @return The UMessagePatternPartType of the i-th Part.
+     * @stable ICU 4.8
+     */
+    UMessagePatternPartType getPartType(int32_t i) const {
+        return getPart(i).type;
+    }
+
+    /**
+     * Returns the pattern index of the specified pattern "part".
+     * Convenience method for getPart(partIndex).getIndex().
+     * @param partIndex The index of the Part data. (0..countParts()-1)
+     * @return The pattern index of this Part.
+     * @stable ICU 4.8
+     */
+    int32_t getPatternIndex(int32_t partIndex) const {
+        return getPart(partIndex).index;
+    }
+
+    /**
+     * Returns the substring of the pattern string indicated by the Part.
+     * Convenience method for getPatternString().substring(part.getIndex(), part.getLimit()).
+     * @param part a part of this MessagePattern.
+     * @return the substring associated with part.
+     * @stable ICU 4.8
+     */
+    UnicodeString getSubstring(const Part &part) const {
+        return msg.tempSubString(part.index, part.length);
+    }
+
+    /**
+     * Compares the part's substring with the input string s.
+     * @param part a part of this MessagePattern.
+     * @param s a string.
+     * @return true if getSubstring(part).equals(s).
+     * @stable ICU 4.8
+     */
+    UBool partSubstringMatches(const Part &part, const UnicodeString &s) const {
+        return 0==msg.compare(part.index, part.length, s);
+    }
+
+    /**
+     * Returns the numeric value associated with an ARG_INT or ARG_DOUBLE.
+     * @param part a part of this MessagePattern.
+     * @return the part's numeric value, or UMSGPAT_NO_NUMERIC_VALUE if this is not a numeric part.
+     * @stable ICU 4.8
+     */
+    double getNumericValue(const Part &part) const;
+
+    /**
+     * Returns the "offset:" value of a PluralFormat argument, or 0 if none is specified.
+     * @param pluralStart the index of the first PluralFormat argument style part. (0..countParts()-1)
+     * @return the "offset:" value.
+     * @stable ICU 4.8
+     */
+    double getPluralOffset(int32_t pluralStart) const;
+
+    /**
+     * Returns the index of the ARG|MSG_LIMIT part corresponding to the ARG|MSG_START at start.
+     * @param start The index of some Part data (0..countParts()-1);
+     *        this Part should be of Type ARG_START or MSG_START.
+     * @return The first i>start where getPart(i).getType()==ARG|MSG_LIMIT at the same nesting level,
+     *         or start itself if getPartType(msgStart)!=ARG|MSG_START.
+     * @stable ICU 4.8
+     */
+    int32_t getLimitPartIndex(int32_t start) const {
+        int32_t limit=getPart(start).limitPartIndex;
+        if(limit<start) {
+            return start;
+        }
+        return limit;
+    }
+
+    /**
+     * A message pattern "part", representing a pattern parsing event.
+     * There is a part for the start and end of a message or argument,
+     * for quoting and escaping of and with ASCII apostrophes,
+     * and for syntax elements of "complex" arguments.
+     * @stable ICU 4.8
+     */
+    class Part : public UMemory {
+    public:
+        /**
+         * Default constructor, do not use.
+         * @internal
+         */
+        Part() {}
+
+        /**
+         * Returns the type of this part.
+         * @return the part type.
+         * @stable ICU 4.8
+         */
+        UMessagePatternPartType getType() const {
+            return type;
+        }
+
+        /**
+         * Returns the pattern string index associated with this Part.
+         * @return this part's pattern string index.
+         * @stable ICU 4.8
+         */
+        int32_t getIndex() const {
+            return index;
+        }
+
+        /**
+         * Returns the length of the pattern substring associated with this Part.
+         * This is 0 for some parts.
+         * @return this part's pattern substring length.
+         * @stable ICU 4.8
+         */
+        int32_t getLength() const {
+            return length;
+        }
+
+        /**
+         * Returns the pattern string limit (exclusive-end) index associated with this Part.
+         * Convenience method for getIndex()+getLength().
+         * @return this part's pattern string limit index, same as getIndex()+getLength().
+         * @stable ICU 4.8
+         */
+        int32_t getLimit() const {
+            return index+length;
+        }
+
+        /**
+         * Returns a value associated with this part.
+         * See the documentation of each part type for details.
+         * @return the part value.
+         * @stable ICU 4.8
+         */
+        int32_t getValue() const {
+            return value;
+        }
+
+        /**
+         * Returns the argument type if this part is of type ARG_START or ARG_LIMIT,
+         * otherwise UMSGPAT_ARG_TYPE_NONE.
+         * @return the argument type for this part.
+         * @stable ICU 4.8
+         */
+        UMessagePatternArgType getArgType() const {
+            UMessagePatternPartType msgType=getType();
+            if(msgType ==UMSGPAT_PART_TYPE_ARG_START || msgType ==UMSGPAT_PART_TYPE_ARG_LIMIT) {
+                return static_cast<UMessagePatternArgType>(value);
+            } else {
+                return UMSGPAT_ARG_TYPE_NONE;
+            }
+        }
+
+        /**
+         * Indicates whether the Part type has a numeric value.
+         * If so, then that numeric value can be retrieved via MessagePattern.getNumericValue().
+         * @param type The Part type to be tested.
+         * @return true if the Part type has a numeric value.
+         * @stable ICU 4.8
+         */
+        static UBool hasNumericValue(UMessagePatternPartType type) {
+            return type==UMSGPAT_PART_TYPE_ARG_INT || type==UMSGPAT_PART_TYPE_ARG_DOUBLE;
+        }
+
+        /**
+         * @param other another object to compare with.
+         * @return true if this object is equivalent to the other one.
+         * @stable ICU 4.8
+         */
+        bool operator==(const Part &other) const;
+
+        /**
+         * @param other another object to compare with.
+         * @return false if this object is equivalent to the other one.
+         * @stable ICU 4.8
+         */
+        inline bool operator!=(const Part &other) const {
+            return !operator==(other);
+        }
+
+        /**
+         * @return A hash code for this object.
+         * @stable ICU 4.8
+         */
+        int32_t hashCode() const {
+            return ((type*37+index)*37+length)*37+value;
+        }
+
+    private:
+        friend class MessagePattern;
+
+        static const int32_t MAX_LENGTH=0xffff;
+        static const int32_t MAX_VALUE=0x7fff;
+        static const int32_t MAX_NESTED_LEVELS=0x03ff;
+
+        // Some fields are not final because they are modified during pattern parsing.
+        // After pattern parsing, the parts are effectively immutable.
+        UMessagePatternPartType type;
+        int32_t index;
+        uint16_t length;
+        int16_t value;
+        int32_t limitPartIndex;
+    };
+
+private:
+    void preParse(const UnicodeString &pattern, UParseError *parseError, UErrorCode &errorCode);
+
+    void postParse();
+
+    int32_t parseMessage(int32_t index, int32_t msgStartLength,
+                         int32_t nestingLevel, UMessagePatternArgType parentType,
+                         UParseError *parseError, UErrorCode &errorCode);
+
+    int32_t parseArg(int32_t index, int32_t argStartLength, int32_t nestingLevel,
+                     UParseError *parseError, UErrorCode &errorCode);
+
+    int32_t parseSimpleStyle(int32_t index, UParseError *parseError, UErrorCode &errorCode);
+
+    int32_t parseChoiceStyle(int32_t index, int32_t nestingLevel,
+                             UParseError *parseError, UErrorCode &errorCode);
+
+    int32_t parsePluralOrSelectStyle(UMessagePatternArgType argType, int32_t index, int32_t nestingLevel,
+                                     UParseError *parseError, UErrorCode &errorCode);
+
+    /**
+     * Validates and parses an argument name or argument number string.
+     * This internal method assumes that the input substring is a "pattern identifier".
+     * @return &gt;=0 if the name is a valid number,
+     *         ARG_NAME_NOT_NUMBER (-1) if it is a "pattern identifier" but not all ASCII digits,
+     *         ARG_NAME_NOT_VALID (-2) if it is neither.
+     * @see #validateArgumentName(String)
+     */
+    static int32_t parseArgNumber(const UnicodeString &s, int32_t start, int32_t limit);
+
+    int32_t parseArgNumber(int32_t start, int32_t limit) {
+        return parseArgNumber(msg, start, limit);
+    }
+
+    /**
+     * Parses a number from the specified message substring.
+     * @param start start index into the message string
+     * @param limit limit index into the message string, must be start<limit
+     * @param allowInfinity true if U+221E is allowed (for ChoiceFormat)
+     * @param parseError
+     * @param errorCode
+     */
+    void parseDouble(int32_t start, int32_t limit, UBool allowInfinity,
+                     UParseError *parseError, UErrorCode &errorCode);
+
+    // Java has package-private appendReducedApostrophes() here.
+    // In C++, this is declared in the MessageImpl class.
+
+    int32_t skipWhiteSpace(int32_t index);
+
+    int32_t skipIdentifier(int32_t index);
+
+    /**
+     * Skips a sequence of characters that could occur in a double value.
+     * Does not fully parse or validate the value.
+     */
+    int32_t skipDouble(int32_t index);
+
+    static UBool isArgTypeChar(UChar32 c);
+
+    UBool isChoice(int32_t index);
+
+    UBool isPlural(int32_t index);
+
+    UBool isSelect(int32_t index);
+
+    UBool isOrdinal(int32_t index);
+
+    /**
+     * @return true if we are inside a MessageFormat (sub-)pattern,
+     *         as opposed to inside a top-level choice/plural/select pattern.
+     */
+    UBool inMessageFormatPattern(int32_t nestingLevel);
+
+    /**
+     * @return true if we are in a MessageFormat sub-pattern
+     *         of a top-level ChoiceFormat pattern.
+     */
+    UBool inTopLevelChoiceMessage(int32_t nestingLevel, UMessagePatternArgType parentType);
+
+    void addPart(UMessagePatternPartType type, int32_t index, int32_t length,
+                 int32_t value, UErrorCode &errorCode);
+
+    void addLimitPart(int32_t start,
+                      UMessagePatternPartType type, int32_t index, int32_t length,
+                      int32_t value, UErrorCode &errorCode);
+
+    void addArgDoublePart(double numericValue, int32_t start, int32_t length, UErrorCode &errorCode);
+
+    void setParseError(UParseError *parseError, int32_t index);
+
+    UBool init(UErrorCode &errorCode);
+    UBool copyStorage(const MessagePattern &other, UErrorCode &errorCode);
+
+    UMessagePatternApostropheMode aposMode;
+    UnicodeString msg;
+    // ArrayList<Part> parts=new ArrayList<Part>();
+    MessagePatternPartsList *partsList;
+    Part *parts;
+    int32_t partsLength;
+    // ArrayList<Double> numericValues;
+    MessagePatternDoubleList *numericValuesList;
+    double *numericValues;
+    int32_t numericValuesLength;
+    UBool hasArgNames;
+    UBool hasArgNumbers;
+    UBool needsAutoQuoting;
+};
+
+U_NAMESPACE_END
+
+#endif  // !UCONFIG_NO_FORMATTING
+
+#endif /* U_SHOW_CPLUSPLUS_API */
+
+#endif  // __MESSAGEPATTERN_H__

thirdparty/icu4c/common/unicode/normalizer2.h

@@ -0,0 +1,791 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+/*
+*******************************************************************************
+*
+*   Copyright (C) 2009-2013, International Business Machines
+*   Corporation and others.  All Rights Reserved.
+*
+*******************************************************************************
+*   file name:  normalizer2.h
+*   encoding:   UTF-8
+*   tab size:   8 (not used)
+*   indentation:4
+*
+*   created on: 2009nov22
+*   created by: Markus W. Scherer
+*/
+
+#ifndef __NORMALIZER2_H__
+#define __NORMALIZER2_H__
+
+/**
+ * \file
+ * \brief C++ API: New API for Unicode Normalization.
+ */
+
+#include "unicode/utypes.h"
+
+#if U_SHOW_CPLUSPLUS_API
+
+#if !UCONFIG_NO_NORMALIZATION
+
+#include "unicode/stringpiece.h"
+#include "unicode/uniset.h"
+#include "unicode/unistr.h"
+#include "unicode/unorm2.h"
+
+U_NAMESPACE_BEGIN
+
+class ByteSink;
+
+/**
+ * Unicode normalization functionality for standard Unicode normalization or
+ * for using custom mapping tables.
+ * All instances of this class are unmodifiable/immutable.
+ * Instances returned by getInstance() are singletons that must not be deleted by the caller.
+ * The Normalizer2 class is not intended for public subclassing.
+ *
+ * The primary functions are to produce a normalized string and to detect whether
+ * a string is already normalized.
+ * The most commonly used normalization forms are those defined in
+ * http://www.unicode.org/unicode/reports/tr15/
+ * However, this API supports additional normalization forms for specialized purposes.
+ * For example, NFKC_Casefold is provided via getInstance("nfkc_cf", COMPOSE)
+ * and can be used in implementations of UTS #46.
+ *
+ * Not only are the standard compose and decompose modes supplied,
+ * but additional modes are provided as documented in the Mode enum.
+ *
+ * Some of the functions in this class identify normalization boundaries.
+ * At a normalization boundary, the portions of the string
+ * before it and starting from it do not interact and can be handled independently.
+ *
+ * The spanQuickCheckYes() stops at a normalization boundary.
+ * When the goal is a normalized string, then the text before the boundary
+ * can be copied, and the remainder can be processed with normalizeSecondAndAppend().
+ *
+ * The hasBoundaryBefore(), hasBoundaryAfter() and isInert() functions test whether
+ * a character is guaranteed to be at a normalization boundary,
+ * regardless of context.
+ * This is used for moving from one normalization boundary to the next
+ * or preceding boundary, and for performing iterative normalization.
+ *
+ * Iterative normalization is useful when only a small portion of a
+ * longer string needs to be processed.
+ * For example, in ICU, iterative normalization is used by the NormalizationTransliterator
+ * (to avoid replacing already-normalized text) and ucol_nextSortKeyPart()
+ * (to process only the substring for which sort key bytes are computed).
+ *
+ * The set of normalization boundaries returned by these functions may not be
+ * complete: There may be more boundaries that could be returned.
+ * Different functions may return different boundaries.
+ * @stable ICU 4.4
+ */
+class U_COMMON_API Normalizer2 : public UObject {
+public:
+    /**
+     * Destructor.
+     * @stable ICU 4.4
+     */
+    ~Normalizer2();
+
+    /**
+     * Returns a Normalizer2 instance for Unicode NFC normalization.
+     * Same as getInstance(nullptr, "nfc", UNORM2_COMPOSE, errorCode).
+     * Returns an unmodifiable singleton instance. Do not delete it.
+     * @param errorCode Standard ICU error code. Its input value must
+     *                  pass the U_SUCCESS() test, or else the function returns
+     *                  immediately. Check for U_FAILURE() on output or use with
+     *                  function chaining. (See User Guide for details.)
+     * @return the requested Normalizer2, if successful
+     * @stable ICU 49
+     */
+    static const Normalizer2 *
+    getNFCInstance(UErrorCode &errorCode);
+
+    /**
+     * Returns a Normalizer2 instance for Unicode NFD normalization.
+     * Same as getInstance(nullptr, "nfc", UNORM2_DECOMPOSE, errorCode).
+     * Returns an unmodifiable singleton instance. Do not delete it.
+     * @param errorCode Standard ICU error code. Its input value must
+     *                  pass the U_SUCCESS() test, or else the function returns
+     *                  immediately. Check for U_FAILURE() on output or use with
+     *                  function chaining. (See User Guide for details.)
+     * @return the requested Normalizer2, if successful
+     * @stable ICU 49
+     */
+    static const Normalizer2 *
+    getNFDInstance(UErrorCode &errorCode);
+
+    /**
+     * Returns a Normalizer2 instance for Unicode NFKC normalization.
+     * Same as getInstance(nullptr, "nfkc", UNORM2_COMPOSE, errorCode).
+     * Returns an unmodifiable singleton instance. Do not delete it.
+     * @param errorCode Standard ICU error code. Its input value must
+     *                  pass the U_SUCCESS() test, or else the function returns
+     *                  immediately. Check for U_FAILURE() on output or use with
+     *                  function chaining. (See User Guide for details.)
+     * @return the requested Normalizer2, if successful
+     * @stable ICU 49
+     */
+    static const Normalizer2 *
+    getNFKCInstance(UErrorCode &errorCode);
+
+    /**
+     * Returns a Normalizer2 instance for Unicode NFKD normalization.
+     * Same as getInstance(nullptr, "nfkc", UNORM2_DECOMPOSE, errorCode).
+     * Returns an unmodifiable singleton instance. Do not delete it.
+     * @param errorCode Standard ICU error code. Its input value must
+     *                  pass the U_SUCCESS() test, or else the function returns
+     *                  immediately. Check for U_FAILURE() on output or use with
+     *                  function chaining. (See User Guide for details.)
+     * @return the requested Normalizer2, if successful
+     * @stable ICU 49
+     */
+    static const Normalizer2 *
+    getNFKDInstance(UErrorCode &errorCode);
+
+    /**
+     * Returns a Normalizer2 instance for Unicode toNFKC_Casefold() normalization
+     * which is equivalent to applying the NFKC_Casefold mappings and then NFC.
+     * See https://www.unicode.org/reports/tr44/#NFKC_Casefold
+     *
+     * Same as getInstance(nullptr, "nfkc_cf", UNORM2_COMPOSE, errorCode).
+     * Returns an unmodifiable singleton instance. Do not delete it.
+     * @param errorCode Standard ICU error code. Its input value must
+     *                  pass the U_SUCCESS() test, or else the function returns
+     *                  immediately. Check for U_FAILURE() on output or use with
+     *                  function chaining. (See User Guide for details.)
+     * @return the requested Normalizer2, if successful
+     * @stable ICU 49
+     */
+    static const Normalizer2 *
+    getNFKCCasefoldInstance(UErrorCode &errorCode);
+
+    /**
+     * Returns a Normalizer2 instance for a variant of Unicode toNFKC_Casefold() normalization
+     * which is equivalent to applying the NFKC_Simple_Casefold mappings and then NFC.
+     * See https://www.unicode.org/reports/tr44/#NFKC_Simple_Casefold
+     *
+     * Same as getInstance(nullptr, "nfkc_scf", UNORM2_COMPOSE, errorCode).
+     * Returns an unmodifiable singleton instance. Do not delete it.
+     * @param errorCode Standard ICU error code. Its input value must
+     *                  pass the U_SUCCESS() test, or else the function returns
+     *                  immediately. Check for U_FAILURE() on output or use with
+     *                  function chaining. (See User Guide for details.)
+     * @return the requested Normalizer2, if successful
+     * @stable ICU 74
+     */
+    static const Normalizer2 *
+    getNFKCSimpleCasefoldInstance(UErrorCode &errorCode);
+
+    /**
+     * Returns a Normalizer2 instance which uses the specified data file
+     * (packageName/name similar to ucnv_openPackage() and ures_open()/ResourceBundle)
+     * and which composes or decomposes text according to the specified mode.
+     * Returns an unmodifiable singleton instance. Do not delete it.
+     *
+     * Use packageName=nullptr for data files that are part of ICU's own data.
+     * Use name="nfc" and UNORM2_COMPOSE/UNORM2_DECOMPOSE for Unicode standard NFC/NFD.
+     * Use name="nfkc" and UNORM2_COMPOSE/UNORM2_DECOMPOSE for Unicode standard NFKC/NFKD.
+     * Use name="nfkc_cf" and UNORM2_COMPOSE for Unicode standard NFKC_CF=NFKC_Casefold.
+     *
+     * @param packageName nullptr for ICU built-in data, otherwise application data package name
+     * @param name "nfc" or "nfkc" or "nfkc_cf" or "nfkc_scf" or name of custom data file
+     * @param mode normalization mode (compose or decompose etc.)
+     * @param errorCode Standard ICU error code. Its input value must
+     *                  pass the U_SUCCESS() test, or else the function returns
+     *                  immediately. Check for U_FAILURE() on output or use with
+     *                  function chaining. (See User Guide for details.)
+     * @return the requested Normalizer2, if successful
+     * @stable ICU 4.4
+     */
+    static const Normalizer2 *
+    getInstance(const char *packageName,
+                const char *name,
+                UNormalization2Mode mode,
+                UErrorCode &errorCode);
+
+    /**
+     * Returns the normalized form of the source string.
+     * @param src source string
+     * @param errorCode Standard ICU error code. Its input value must
+     *                  pass the U_SUCCESS() test, or else the function returns
+     *                  immediately. Check for U_FAILURE() on output or use with
+     *                  function chaining. (See User Guide for details.)
+     * @return normalized src
+     * @stable ICU 4.4
+     */
+    UnicodeString
+    normalize(const UnicodeString &src, UErrorCode &errorCode) const {
+        UnicodeString result;
+        normalize(src, result, errorCode);
+        return result;
+    }
+    /**
+     * Writes the normalized form of the source string to the destination string
+     * (replacing its contents) and returns the destination string.
+     * The source and destination strings must be different objects.
+     * @param src source string
+     * @param dest destination string; its contents is replaced with normalized src
+     * @param errorCode Standard ICU error code. Its input value must
+     *                  pass the U_SUCCESS() test, or else the function returns
+     *                  immediately. Check for U_FAILURE() on output or use with
+     *                  function chaining. (See User Guide for details.)
+     * @return dest
+     * @stable ICU 4.4
+     */
+    virtual UnicodeString &
+    normalize(const UnicodeString &src,
+              UnicodeString &dest,
+              UErrorCode &errorCode) const = 0;
+
+    /**
+     * Normalizes a UTF-8 string and optionally records how source substrings
+     * relate to changed and unchanged result substrings.
+     *
+     * Implemented completely for all built-in modes except for FCD.
+     * The base class implementation converts to & from UTF-16 and does not support edits.
+     *
+     * @param options   Options bit set, usually 0. See U_OMIT_UNCHANGED_TEXT and U_EDITS_NO_RESET.
+     * @param src       Source UTF-8 string.
+     * @param sink      A ByteSink to which the normalized UTF-8 result string is written.
+     *                  sink.Flush() is called at the end.
+     * @param edits     Records edits for index mapping, working with styled text,
+     *                  and getting only changes (if any).
+     *                  The Edits contents is undefined if any error occurs.
+     *                  This function calls edits->reset() first unless
+     *                  options includes U_EDITS_NO_RESET. edits can be nullptr.
+     * @param errorCode Standard ICU error code. Its input value must
+     *                  pass the U_SUCCESS() test, or else the function returns
+     *                  immediately. Check for U_FAILURE() on output or use with
+     *                  function chaining. (See User Guide for details.)
+     * @stable ICU 60
+     */
+    virtual void
+    normalizeUTF8(uint32_t options, StringPiece src, ByteSink &sink,
+                  Edits *edits, UErrorCode &errorCode) const;
+
+    /**
+     * Appends the normalized form of the second string to the first string
+     * (merging them at the boundary) and returns the first string.
+     * The result is normalized if the first string was normalized.
+     * The first and second strings must be different objects.
+     * @param first string, should be normalized
+     * @param second string, will be normalized
+     * @param errorCode Standard ICU error code. Its input value must
+     *                  pass the U_SUCCESS() test, or else the function returns
+     *                  immediately. Check for U_FAILURE() on output or use with
+     *                  function chaining. (See User Guide for details.)
+     * @return first
+     * @stable ICU 4.4
+     */
+    virtual UnicodeString &
+    normalizeSecondAndAppend(UnicodeString &first,
+                             const UnicodeString &second,
+                             UErrorCode &errorCode) const = 0;
+    /**
+     * Appends the second string to the first string
+     * (merging them at the boundary) and returns the first string.
+     * The result is normalized if both the strings were normalized.
+     * The first and second strings must be different objects.
+     * @param first string, should be normalized
+     * @param second string, should be normalized
+     * @param errorCode Standard ICU error code. Its input value must
+     *                  pass the U_SUCCESS() test, or else the function returns
+     *                  immediately. Check for U_FAILURE() on output or use with
+     *                  function chaining. (See User Guide for details.)
+     * @return first
+     * @stable ICU 4.4
+     */
+    virtual UnicodeString &
+    append(UnicodeString &first,
+           const UnicodeString &second,
+           UErrorCode &errorCode) const = 0;
+
+    /**
+     * Gets the decomposition mapping of c.
+     * Roughly equivalent to normalizing the String form of c
+     * on a UNORM2_DECOMPOSE Normalizer2 instance, but much faster, and except that this function
+     * returns false and does not write a string
+     * if c does not have a decomposition mapping in this instance's data.
+     * This function is independent of the mode of the Normalizer2.
+     * @param c code point
+     * @param decomposition String object which will be set to c's
+     *                      decomposition mapping, if there is one.
+     * @return true if c has a decomposition, otherwise false
+     * @stable ICU 4.6
+     */
+    virtual UBool
+    getDecomposition(UChar32 c, UnicodeString &decomposition) const = 0;
+
+    /**
+     * Gets the raw decomposition mapping of c.
+     *
+     * This is similar to the getDecomposition() method but returns the
+     * raw decomposition mapping as specified in UnicodeData.txt or
+     * (for custom data) in the mapping files processed by the gennorm2 tool.
+     * By contrast, getDecomposition() returns the processed,
+     * recursively-decomposed version of this mapping.
+     *
+     * When used on a standard NFKC Normalizer2 instance,
+     * getRawDecomposition() returns the Unicode Decomposition_Mapping (dm) property.
+     *
+     * When used on a standard NFC Normalizer2 instance,
+     * it returns the Decomposition_Mapping only if the Decomposition_Type (dt) is Canonical (Can);
+     * in this case, the result contains either one or two code points (=1..4 char16_ts).
+     *
+     * This function is independent of the mode of the Normalizer2.
+     * The default implementation returns false.
+     * @param c code point
+     * @param decomposition String object which will be set to c's
+     *                      raw decomposition mapping, if there is one.
+     * @return true if c has a decomposition, otherwise false
+     * @stable ICU 49
+     */
+    virtual UBool
+    getRawDecomposition(UChar32 c, UnicodeString &decomposition) const;
+
+    /**
+     * Performs pairwise composition of a & b and returns the composite if there is one.
+     *
+     * Returns a composite code point c only if c has a two-way mapping to a+b.
+     * In standard Unicode normalization, this means that
+     * c has a canonical decomposition to a+b
+     * and c does not have the Full_Composition_Exclusion property.
+     *
+     * This function is independent of the mode of the Normalizer2.
+     * The default implementation returns a negative value.
+     * @param a A (normalization starter) code point.
+     * @param b Another code point.
+     * @return The non-negative composite code point if there is one; otherwise a negative value.
+     * @stable ICU 49
+     */
+    virtual UChar32
+    composePair(UChar32 a, UChar32 b) const;
+
+    /**
+     * Gets the combining class of c.
+     * The default implementation returns 0
+     * but all standard implementations return the Unicode Canonical_Combining_Class value.
+     * @param c code point
+     * @return c's combining class
+     * @stable ICU 49
+     */
+    virtual uint8_t
+    getCombiningClass(UChar32 c) const;
+
+    /**
+     * Tests if the string is normalized.
+     * Internally, in cases where the quickCheck() method would return "maybe"
+     * (which is only possible for the two COMPOSE modes) this method
+     * resolves to "yes" or "no" to provide a definitive result,
+     * at the cost of doing more work in those cases.
+     * @param s input string
+     * @param errorCode Standard ICU error code. Its input value must
+     *                  pass the U_SUCCESS() test, or else the function returns
+     *                  immediately. Check for U_FAILURE() on output or use with
+     *                  function chaining. (See User Guide for details.)
+     * @return true if s is normalized
+     * @stable ICU 4.4
+     */
+    virtual UBool
+    isNormalized(const UnicodeString &s, UErrorCode &errorCode) const = 0;
+    /**
+     * Tests if the UTF-8 string is normalized.
+     * Internally, in cases where the quickCheck() method would return "maybe"
+     * (which is only possible for the two COMPOSE modes) this method
+     * resolves to "yes" or "no" to provide a definitive result,
+     * at the cost of doing more work in those cases.
+     *
+     * This works for all normalization modes.
+     * It is optimized for UTF-8 for all built-in modes except for FCD.
+     * The base class implementation converts to UTF-16 and calls isNormalized().
+     *
+     * @param s UTF-8 input string
+     * @param errorCode Standard ICU error code. Its input value must
+     *                  pass the U_SUCCESS() test, or else the function returns
+     *                  immediately. Check for U_FAILURE() on output or use with
+     *                  function chaining. (See User Guide for details.)
+     * @return true if s is normalized
+     * @stable ICU 60
+     */
+    virtual UBool
+    isNormalizedUTF8(StringPiece s, UErrorCode &errorCode) const;
+
+
+    /**
+     * Tests if the string is normalized.
+     * For the two COMPOSE modes, the result could be "maybe" in cases that
+     * would take a little more work to resolve definitively.
+     * Use spanQuickCheckYes() and normalizeSecondAndAppend() for a faster
+     * combination of quick check + normalization, to avoid
+     * re-checking the "yes" prefix.
+     * @param s input string
+     * @param errorCode Standard ICU error code. Its input value must
+     *                  pass the U_SUCCESS() test, or else the function returns
+     *                  immediately. Check for U_FAILURE() on output or use with
+     *                  function chaining. (See User Guide for details.)
+     * @return UNormalizationCheckResult
+     * @stable ICU 4.4
+     */
+    virtual UNormalizationCheckResult
+    quickCheck(const UnicodeString &s, UErrorCode &errorCode) const = 0;
+
+    /**
+     * Returns the end of the normalized substring of the input string.
+     * In other words, with <code>end=spanQuickCheckYes(s, ec);</code>
+     * the substring <code>UnicodeString(s, 0, end)</code>
+     * will pass the quick check with a "yes" result.
+     *
+     * The returned end index is usually one or more characters before the
+     * "no" or "maybe" character: The end index is at a normalization boundary.
+     * (See the class documentation for more about normalization boundaries.)
+     *
+     * When the goal is a normalized string and most input strings are expected
+     * to be normalized already, then call this method,
+     * and if it returns a prefix shorter than the input string,
+     * copy that prefix and use normalizeSecondAndAppend() for the remainder.
+     * @param s input string
+     * @param errorCode Standard ICU error code. Its input value must
+     *                  pass the U_SUCCESS() test, or else the function returns
+     *                  immediately. Check for U_FAILURE() on output or use with
+     *                  function chaining. (See User Guide for details.)
+     * @return "yes" span end index
+     * @stable ICU 4.4
+     */
+    virtual int32_t
+    spanQuickCheckYes(const UnicodeString &s, UErrorCode &errorCode) const = 0;
+
+    /**
+     * Tests if the character always has a normalization boundary before it,
+     * regardless of context.
+     * If true, then the character does not normalization-interact with
+     * preceding characters.
+     * In other words, a string containing this character can be normalized
+     * by processing portions before this character and starting from this
+     * character independently.
+     * This is used for iterative normalization. See the class documentation for details.
+     * @param c character to test
+     * @return true if c has a normalization boundary before it
+     * @stable ICU 4.4
+     */
+    virtual UBool hasBoundaryBefore(UChar32 c) const = 0;
+
+    /**
+     * Tests if the character always has a normalization boundary after it,
+     * regardless of context.
+     * If true, then the character does not normalization-interact with
+     * following characters.
+     * In other words, a string containing this character can be normalized
+     * by processing portions up to this character and after this
+     * character independently.
+     * This is used for iterative normalization. See the class documentation for details.
+     * Note that this operation may be significantly slower than hasBoundaryBefore().
+     * @param c character to test
+     * @return true if c has a normalization boundary after it
+     * @stable ICU 4.4
+     */
+    virtual UBool hasBoundaryAfter(UChar32 c) const = 0;
+
+    /**
+     * Tests if the character is normalization-inert.
+     * If true, then the character does not change, nor normalization-interact with
+     * preceding or following characters.
+     * In other words, a string containing this character can be normalized
+     * by processing portions before this character and after this
+     * character independently.
+     * This is used for iterative normalization. See the class documentation for details.
+     * Note that this operation may be significantly slower than hasBoundaryBefore().
+     * @param c character to test
+     * @return true if c is normalization-inert
+     * @stable ICU 4.4
+     */
+    virtual UBool isInert(UChar32 c) const = 0;
+};
+
+/**
+ * Normalization filtered by a UnicodeSet.
+ * Normalizes portions of the text contained in the filter set and leaves
+ * portions not contained in the filter set unchanged.
+ * Filtering is done via UnicodeSet::span(..., USET_SPAN_SIMPLE).
+ * Not-in-the-filter text is treated as "is normalized" and "quick check yes".
+ * This class implements all of (and only) the Normalizer2 API.
+ * An instance of this class is unmodifiable/immutable but is constructed and
+ * must be destructed by the owner.
+ * @stable ICU 4.4
+ */
+class U_COMMON_API FilteredNormalizer2 : public Normalizer2 {
+public:
+    /**
+     * Constructs a filtered normalizer wrapping any Normalizer2 instance
+     * and a filter set.
+     * Both are aliased and must not be modified or deleted while this object
+     * is used.
+     * The filter set should be frozen; otherwise the performance will suffer greatly.
+     * @param n2 wrapped Normalizer2 instance
+     * @param filterSet UnicodeSet which determines the characters to be normalized
+     * @stable ICU 4.4
+     */
+    FilteredNormalizer2(const Normalizer2 &n2, const UnicodeSet &filterSet) :
+            norm2(n2), set(filterSet) {}
+
+    /**
+     * Destructor.
+     * @stable ICU 4.4
+     */
+    ~FilteredNormalizer2();
+
+    /**
+     * Writes the normalized form of the source string to the destination string
+     * (replacing its contents) and returns the destination string.
+     * The source and destination strings must be different objects.
+     * @param src source string
+     * @param dest destination string; its contents is replaced with normalized src
+     * @param errorCode Standard ICU error code. Its input value must
+     *                  pass the U_SUCCESS() test, or else the function returns
+     *                  immediately. Check for U_FAILURE() on output or use with
+     *                  function chaining. (See User Guide for details.)
+     * @return dest
+     * @stable ICU 4.4
+     */
+    virtual UnicodeString &
+    normalize(const UnicodeString &src,
+              UnicodeString &dest,
+              UErrorCode &errorCode) const override;
+
+    /**
+     * Normalizes a UTF-8 string and optionally records how source substrings
+     * relate to changed and unchanged result substrings.
+     *
+     * Implemented completely for most built-in modes except for FCD.
+     * The base class implementation converts to & from UTF-16 and does not support edits.
+     *
+     * @param options   Options bit set, usually 0. See U_OMIT_UNCHANGED_TEXT and U_EDITS_NO_RESET.
+     * @param src       Source UTF-8 string.
+     * @param sink      A ByteSink to which the normalized UTF-8 result string is written.
+     *                  sink.Flush() is called at the end.
+     * @param edits     Records edits for index mapping, working with styled text,
+     *                  and getting only changes (if any).
+     *                  The Edits contents is undefined if any error occurs.
+     *                  This function calls edits->reset() first unless
+     *                  options includes U_EDITS_NO_RESET. edits can be nullptr.
+     * @param errorCode Standard ICU error code. Its input value must
+     *                  pass the U_SUCCESS() test, or else the function returns
+     *                  immediately. Check for U_FAILURE() on output or use with
+     *                  function chaining. (See User Guide for details.)
+     * @stable ICU 60
+     */
+    virtual void
+    normalizeUTF8(uint32_t options, StringPiece src, ByteSink &sink,
+                  Edits *edits, UErrorCode &errorCode) const override;
+
+    /**
+     * Appends the normalized form of the second string to the first string
+     * (merging them at the boundary) and returns the first string.
+     * The result is normalized if the first string was normalized.
+     * The first and second strings must be different objects.
+     * @param first string, should be normalized
+     * @param second string, will be normalized
+     * @param errorCode Standard ICU error code. Its input value must
+     *                  pass the U_SUCCESS() test, or else the function returns
+     *                  immediately. Check for U_FAILURE() on output or use with
+     *                  function chaining. (See User Guide for details.)
+     * @return first
+     * @stable ICU 4.4
+     */
+    virtual UnicodeString &
+    normalizeSecondAndAppend(UnicodeString &first,
+                             const UnicodeString &second,
+                             UErrorCode &errorCode) const override;
+    /**
+     * Appends the second string to the first string
+     * (merging them at the boundary) and returns the first string.
+     * The result is normalized if both the strings were normalized.
+     * The first and second strings must be different objects.
+     * @param first string, should be normalized
+     * @param second string, should be normalized
+     * @param errorCode Standard ICU error code. Its input value must
+     *                  pass the U_SUCCESS() test, or else the function returns
+     *                  immediately. Check for U_FAILURE() on output or use with
+     *                  function chaining. (See User Guide for details.)
+     * @return first
+     * @stable ICU 4.4
+     */
+    virtual UnicodeString &
+    append(UnicodeString &first,
+           const UnicodeString &second,
+           UErrorCode &errorCode) const override;
+
+    /**
+     * Gets the decomposition mapping of c.
+     * For details see the base class documentation.
+     *
+     * This function is independent of the mode of the Normalizer2.
+     * @param c code point
+     * @param decomposition String object which will be set to c's
+     *                      decomposition mapping, if there is one.
+     * @return true if c has a decomposition, otherwise false
+     * @stable ICU 4.6
+     */
+    virtual UBool
+    getDecomposition(UChar32 c, UnicodeString &decomposition) const override;
+
+    /**
+     * Gets the raw decomposition mapping of c.
+     * For details see the base class documentation.
+     *
+     * This function is independent of the mode of the Normalizer2.
+     * @param c code point
+     * @param decomposition String object which will be set to c's
+     *                      raw decomposition mapping, if there is one.
+     * @return true if c has a decomposition, otherwise false
+     * @stable ICU 49
+     */
+    virtual UBool
+    getRawDecomposition(UChar32 c, UnicodeString &decomposition) const override;
+
+    /**
+     * Performs pairwise composition of a & b and returns the composite if there is one.
+     * For details see the base class documentation.
+     *
+     * This function is independent of the mode of the Normalizer2.
+     * @param a A (normalization starter) code point.
+     * @param b Another code point.
+     * @return The non-negative composite code point if there is one; otherwise a negative value.
+     * @stable ICU 49
+     */
+    virtual UChar32
+    composePair(UChar32 a, UChar32 b) const override;
+
+    /**
+     * Gets the combining class of c.
+     * The default implementation returns 0
+     * but all standard implementations return the Unicode Canonical_Combining_Class value.
+     * @param c code point
+     * @return c's combining class
+     * @stable ICU 49
+     */
+    virtual uint8_t
+    getCombiningClass(UChar32 c) const override;
+
+    /**
+     * Tests if the string is normalized.
+     * For details see the Normalizer2 base class documentation.
+     * @param s input string
+     * @param errorCode Standard ICU error code. Its input value must
+     *                  pass the U_SUCCESS() test, or else the function returns
+     *                  immediately. Check for U_FAILURE() on output or use with
+     *                  function chaining. (See User Guide for details.)
+     * @return true if s is normalized
+     * @stable ICU 4.4
+     */
+    virtual UBool
+    isNormalized(const UnicodeString &s, UErrorCode &errorCode) const override;
+    /**
+     * Tests if the UTF-8 string is normalized.
+     * Internally, in cases where the quickCheck() method would return "maybe"
+     * (which is only possible for the two COMPOSE modes) this method
+     * resolves to "yes" or "no" to provide a definitive result,
+     * at the cost of doing more work in those cases.
+     *
+     * This works for all normalization modes.
+     * It is optimized for UTF-8 for all built-in modes except for FCD.
+     * The base class implementation converts to UTF-16 and calls isNormalized().
+     *
+     * @param s UTF-8 input string
+     * @param errorCode Standard ICU error code. Its input value must
+     *                  pass the U_SUCCESS() test, or else the function returns
+     *                  immediately. Check for U_FAILURE() on output or use with
+     *                  function chaining. (See User Guide for details.)
+     * @return true if s is normalized
+     * @stable ICU 60
+     */
+    virtual UBool
+    isNormalizedUTF8(StringPiece s, UErrorCode &errorCode) const override;
+    /**
+     * Tests if the string is normalized.
+     * For details see the Normalizer2 base class documentation.
+     * @param s input string
+     * @param errorCode Standard ICU error code. Its input value must
+     *                  pass the U_SUCCESS() test, or else the function returns
+     *                  immediately. Check for U_FAILURE() on output or use with
+     *                  function chaining. (See User Guide for details.)
+     * @return UNormalizationCheckResult
+     * @stable ICU 4.4
+     */
+    virtual UNormalizationCheckResult
+    quickCheck(const UnicodeString &s, UErrorCode &errorCode) const override;
+    /**
+     * Returns the end of the normalized substring of the input string.
+     * For details see the Normalizer2 base class documentation.
+     * @param s input string
+     * @param errorCode Standard ICU error code. Its input value must
+     *                  pass the U_SUCCESS() test, or else the function returns
+     *                  immediately. Check for U_FAILURE() on output or use with
+     *                  function chaining. (See User Guide for details.)
+     * @return "yes" span end index
+     * @stable ICU 4.4
+     */
+    virtual int32_t
+    spanQuickCheckYes(const UnicodeString &s, UErrorCode &errorCode) const override;
+
+    /**
+     * Tests if the character always has a normalization boundary before it,
+     * regardless of context.
+     * For details see the Normalizer2 base class documentation.
+     * @param c character to test
+     * @return true if c has a normalization boundary before it
+     * @stable ICU 4.4
+     */
+    virtual UBool hasBoundaryBefore(UChar32 c) const override;
+
+    /**
+     * Tests if the character always has a normalization boundary after it,
+     * regardless of context.
+     * For details see the Normalizer2 base class documentation.
+     * @param c character to test
+     * @return true if c has a normalization boundary after it
+     * @stable ICU 4.4
+     */
+    virtual UBool hasBoundaryAfter(UChar32 c) const override;
+
+    /**
+     * Tests if the character is normalization-inert.
+     * For details see the Normalizer2 base class documentation.
+     * @param c character to test
+     * @return true if c is normalization-inert
+     * @stable ICU 4.4
+     */
+    virtual UBool isInert(UChar32 c) const override;
+private:
+    UnicodeString &
+    normalize(const UnicodeString &src,
+              UnicodeString &dest,
+              USetSpanCondition spanCondition,
+              UErrorCode &errorCode) const;
+
+    void
+    normalizeUTF8(uint32_t options, const char *src, int32_t length,
+                  ByteSink &sink, Edits *edits,
+                  USetSpanCondition spanCondition,
+                  UErrorCode &errorCode) const;
+
+    UnicodeString &
+    normalizeSecondAndAppend(UnicodeString &first,
+                             const UnicodeString &second,
+                             UBool doNormalize,
+                             UErrorCode &errorCode) const;
+
+    const Normalizer2 &norm2;
+    const UnicodeSet &set;
+};
+
+U_NAMESPACE_END
+
+#endif  // !UCONFIG_NO_NORMALIZATION
+
+#endif /* U_SHOW_CPLUSPLUS_API */
+
+#endif  // __NORMALIZER2_H__

thirdparty/icu4c/common/unicode/normlzr.h

@@ -0,0 +1,816 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+/*
+ ********************************************************************
+ * COPYRIGHT:
+ * Copyright (c) 1996-2015, International Business Machines Corporation and
+ * others. All Rights Reserved.
+ ********************************************************************
+ */
+
+#ifndef NORMLZR_H
+#define NORMLZR_H
+
+#include "unicode/utypes.h"
+
+#if U_SHOW_CPLUSPLUS_API
+
+/**
+ * \file 
+ * \brief C++ API: Unicode Normalization
+ */
+ 
+#if !UCONFIG_NO_NORMALIZATION
+
+#include "unicode/chariter.h"
+#include "unicode/normalizer2.h"
+#include "unicode/unistr.h"
+#include "unicode/unorm.h"
+#include "unicode/uobject.h"
+
+U_NAMESPACE_BEGIN
+/**
+ * Old Unicode normalization API.
+ *
+ * This API has been replaced by the Normalizer2 class and is only available
+ * for backward compatibility. This class simply delegates to the Normalizer2 class.
+ * There is one exception: The new API does not provide a replacement for Normalizer::compare().
+ *
+ * The Normalizer class supports the standard normalization forms described in
+ * <a href="http://www.unicode.org/unicode/reports/tr15/" target="unicode">
+ * Unicode Standard Annex #15: Unicode Normalization Forms</a>.
+ *
+ * The Normalizer class consists of two parts:
+ * - static functions that normalize strings or test if strings are normalized
+ * - a Normalizer object is an iterator that takes any kind of text and
+ *   provides iteration over its normalized form
+ *
+ * The Normalizer class is not suitable for subclassing.
+ *
+ * For basic information about normalization forms and details about the C API
+ * please see the documentation in unorm.h.
+ *
+ * The iterator API with the Normalizer constructors and the non-static functions
+ * use a CharacterIterator as input. It is possible to pass a string which
+ * is then internally wrapped in a CharacterIterator.
+ * The input text is not normalized all at once, but incrementally where needed
+ * (providing efficient random access).
+ * This allows to pass in a large text but spend only a small amount of time
+ * normalizing a small part of that text.
+ * However, if the entire text is normalized, then the iterator will be
+ * slower than normalizing the entire text at once and iterating over the result.
+ * A possible use of the Normalizer iterator is also to report an index into the
+ * original text that is close to where the normalized characters come from.
+ *
+ * <em>Important:</em> The iterator API was cleaned up significantly for ICU 2.0.
+ * The earlier implementation reported the getIndex() inconsistently,
+ * and previous() could not be used after setIndex(), next(), first(), and current().
+ *
+ * Normalizer allows to start normalizing from anywhere in the input text by
+ * calling setIndexOnly(), first(), or last().
+ * Without calling any of these, the iterator will start at the beginning of the text.
+ *
+ * At any time, next() returns the next normalized code point (UChar32),
+ * with post-increment semantics (like CharacterIterator::next32PostInc()).
+ * previous() returns the previous normalized code point (UChar32),
+ * with pre-decrement semantics (like CharacterIterator::previous32()).
+ *
+ * current() returns the current code point
+ * (respectively the one at the newly set index) without moving
+ * the getIndex(). Note that if the text at the current position
+ * needs to be normalized, then these functions will do that.
+ * (This is why current() is not const.)
+ * It is more efficient to call setIndexOnly() instead, which does not
+ * normalize.
+ *
+ * getIndex() always refers to the position in the input text where the normalized
+ * code points are returned from. It does not always change with each returned
+ * code point.
+ * The code point that is returned from any of the functions
+ * corresponds to text at or after getIndex(), according to the
+ * function's iteration semantics (post-increment or pre-decrement).
+ *
+ * next() returns a code point from at or after the getIndex()
+ * from before the next() call. After the next() call, the getIndex()
+ * might have moved to where the next code point will be returned from
+ * (from a next() or current() call).
+ * This is semantically equivalent to array access with array[index++]
+ * (post-increment semantics).
+ *
+ * previous() returns a code point from at or after the getIndex()
+ * from after the previous() call.
+ * This is semantically equivalent to array access with array[--index]
+ * (pre-decrement semantics).
+ *
+ * Internally, the Normalizer iterator normalizes a small piece of text
+ * starting at the getIndex() and ending at a following "safe" index.
+ * The normalized results is stored in an internal string buffer, and
+ * the code points are iterated from there.
+ * With multiple iteration calls, this is repeated until the next piece
+ * of text needs to be normalized, and the getIndex() needs to be moved.
+ *
+ * The following "safe" index, the internal buffer, and the secondary
+ * iteration index into that buffer are not exposed on the API.
+ * This also means that it is currently not practical to return to
+ * a particular, arbitrary position in the text because one would need to
+ * know, and be able to set, in addition to the getIndex(), at least also the
+ * current index into the internal buffer.
+ * It is currently only possible to observe when getIndex() changes
+ * (with careful consideration of the iteration semantics),
+ * at which time the internal index will be 0.
+ * For example, if getIndex() is different after next() than before it,
+ * then the internal index is 0 and one can return to this getIndex()
+ * later with setIndexOnly().
+ *
+ * Note: While the setIndex() and getIndex() refer to indices in the
+ * underlying Unicode input text, the next() and previous() methods
+ * iterate through characters in the normalized output.
+ * This means that there is not necessarily a one-to-one correspondence
+ * between characters returned by next() and previous() and the indices
+ * passed to and returned from setIndex() and getIndex().
+ * It is for this reason that Normalizer does not implement the CharacterIterator interface.
+ *
+ * @author Laura Werner, Mark Davis, Markus Scherer
+ * @stable ICU 2.0
+ */
+class U_COMMON_API Normalizer : public UObject {
+public:
+#ifndef U_HIDE_DEPRECATED_API
+  /**
+   * If DONE is returned from an iteration function that returns a code point,
+   * then there are no more normalization results available.
+   * @deprecated ICU 56 Use Normalizer2 instead.
+   */
+  enum {
+      DONE=0xffff
+  };
+
+  // Constructors
+
+  /**
+   * Creates a new <code>Normalizer</code> object for iterating over the
+   * normalized form of a given string.
+   * <p>
+   * @param str   The string to be normalized.  The normalization
+   *              will start at the beginning of the string.
+   *
+   * @param mode  The normalization mode.
+   * @deprecated ICU 56 Use Normalizer2 instead.
+   */
+  Normalizer(const UnicodeString& str, UNormalizationMode mode);
+
+  /**
+   * Creates a new <code>Normalizer</code> object for iterating over the
+   * normalized form of a given string.
+   * <p>
+   * @param str   The string to be normalized.  The normalization
+   *              will start at the beginning of the string.
+   *
+   * @param length Length of the string, or -1 if NUL-terminated.
+   * @param mode  The normalization mode.
+   * @deprecated ICU 56 Use Normalizer2 instead.
+   */
+  Normalizer(ConstChar16Ptr str, int32_t length, UNormalizationMode mode);
+
+  /**
+   * Creates a new <code>Normalizer</code> object for iterating over the
+   * normalized form of the given text.
+   * <p>
+   * @param iter  The input text to be normalized.  The normalization
+   *              will start at the beginning of the string.
+   *
+   * @param mode  The normalization mode.
+   * @deprecated ICU 56 Use Normalizer2 instead.
+   */
+  Normalizer(const CharacterIterator& iter, UNormalizationMode mode);
+#endif  /* U_HIDE_DEPRECATED_API */
+
+#ifndef U_FORCE_HIDE_DEPRECATED_API
+  /**
+   * Copy constructor.
+   * @param copy The object to be copied.
+   * @deprecated ICU 56 Use Normalizer2 instead.
+   */
+  Normalizer(const Normalizer& copy);
+
+  /**
+   * Destructor
+   * @deprecated ICU 56 Use Normalizer2 instead.
+   */
+  virtual ~Normalizer();
+#endif  // U_FORCE_HIDE_DEPRECATED_API
+
+  //-------------------------------------------------------------------------
+  // Static utility methods
+  //-------------------------------------------------------------------------
+
+#ifndef U_HIDE_DEPRECATED_API
+  /**
+   * Normalizes a <code>UnicodeString</code> according to the specified normalization mode.
+   * This is a wrapper for unorm_normalize(), using UnicodeString's.
+   *
+   * The <code>options</code> parameter specifies which optional
+   * <code>Normalizer</code> features are to be enabled for this operation.
+   *
+   * @param source    the input string to be normalized.
+   * @param mode      the normalization mode
+   * @param options   the optional features to be enabled (0 for no options)
+   * @param result    The normalized string (on output).
+   * @param status    The error code.
+   * @deprecated ICU 56 Use Normalizer2 instead.
+   */
+  static void U_EXPORT2 normalize(const UnicodeString& source,
+                        UNormalizationMode mode, int32_t options,
+                        UnicodeString& result,
+                        UErrorCode &status);
+
+  /**
+   * Compose a <code>UnicodeString</code>.
+   * This is equivalent to normalize() with mode UNORM_NFC or UNORM_NFKC.
+   * This is a wrapper for unorm_normalize(), using UnicodeString's.
+   *
+   * The <code>options</code> parameter specifies which optional
+   * <code>Normalizer</code> features are to be enabled for this operation.
+   *
+   * @param source    the string to be composed.
+   * @param compat    Perform compatibility decomposition before composition.
+   *                  If this argument is <code>false</code>, only canonical
+   *                  decomposition will be performed.
+   * @param options   the optional features to be enabled (0 for no options)
+   * @param result    The composed string (on output).
+   * @param status    The error code.
+   * @deprecated ICU 56 Use Normalizer2 instead.
+   */
+  static void U_EXPORT2 compose(const UnicodeString& source,
+                      UBool compat, int32_t options,
+                      UnicodeString& result,
+                      UErrorCode &status);
+
+  /**
+   * Static method to decompose a <code>UnicodeString</code>.
+   * This is equivalent to normalize() with mode UNORM_NFD or UNORM_NFKD.
+   * This is a wrapper for unorm_normalize(), using UnicodeString's.
+   *
+   * The <code>options</code> parameter specifies which optional
+   * <code>Normalizer</code> features are to be enabled for this operation.
+   *
+   * @param source    the string to be decomposed.
+   * @param compat    Perform compatibility decomposition.
+   *                  If this argument is <code>false</code>, only canonical
+   *                  decomposition will be performed.
+   * @param options   the optional features to be enabled (0 for no options)
+   * @param result    The decomposed string (on output).
+   * @param status    The error code.
+   * @deprecated ICU 56 Use Normalizer2 instead.
+   */
+  static void U_EXPORT2 decompose(const UnicodeString& source,
+                        UBool compat, int32_t options,
+                        UnicodeString& result,
+                        UErrorCode &status);
+
+  /**
+   * Performing quick check on a string, to quickly determine if the string is
+   * in a particular normalization format.
+   * This is a wrapper for unorm_quickCheck(), using a UnicodeString.
+   *
+   * Three types of result can be returned UNORM_YES, UNORM_NO or
+   * UNORM_MAYBE. Result UNORM_YES indicates that the argument
+   * string is in the desired normalized format, UNORM_NO determines that
+   * argument string is not in the desired normalized format. A
+   * UNORM_MAYBE result indicates that a more thorough check is required,
+   * the user may have to put the string in its normalized form and compare the
+   * results.
+   * @param source       string for determining if it is in a normalized format
+   * @param mode         normalization format
+   * @param status A reference to a UErrorCode to receive any errors
+   * @return UNORM_YES, UNORM_NO or UNORM_MAYBE
+   *
+   * @see isNormalized
+   * @deprecated ICU 56 Use Normalizer2 instead.
+   */
+  static inline UNormalizationCheckResult
+  quickCheck(const UnicodeString &source, UNormalizationMode mode, UErrorCode &status);
+
+  /**
+   * Performing quick check on a string; same as the other version of quickCheck
+   * but takes an extra options parameter like most normalization functions.
+   *
+   * @param source       string for determining if it is in a normalized format
+   * @param mode         normalization format
+   * @param options      the optional features to be enabled (0 for no options)
+   * @param status A reference to a UErrorCode to receive any errors
+   * @return UNORM_YES, UNORM_NO or UNORM_MAYBE
+   *
+   * @see isNormalized
+   * @deprecated ICU 56 Use Normalizer2 instead.
+   */
+  static UNormalizationCheckResult
+  quickCheck(const UnicodeString &source, UNormalizationMode mode, int32_t options, UErrorCode &status);
+
+  /**
+   * Test if a string is in a given normalization form.
+   * This is semantically equivalent to source.equals(normalize(source, mode)) .
+   *
+   * Unlike unorm_quickCheck(), this function returns a definitive result,
+   * never a "maybe".
+   * For NFD, NFKD, and FCD, both functions work exactly the same.
+   * For NFC and NFKC where quickCheck may return "maybe", this function will
+   * perform further tests to arrive at a true/false result.
+   *
+   * @param src        String that is to be tested if it is in a normalization format.
+   * @param mode       Which normalization form to test for.
+   * @param errorCode  ICU error code in/out parameter.
+   *                   Must fulfill U_SUCCESS before the function call.
+   * @return Boolean value indicating whether the source string is in the
+   *         "mode" normalization form.
+   *
+   * @see quickCheck
+   * @deprecated ICU 56 Use Normalizer2 instead.
+   */
+  static inline UBool
+  isNormalized(const UnicodeString &src, UNormalizationMode mode, UErrorCode &errorCode);
+
+  /**
+   * Test if a string is in a given normalization form; same as the other version of isNormalized
+   * but takes an extra options parameter like most normalization functions.
+   *
+   * @param src        String that is to be tested if it is in a normalization format.
+   * @param mode       Which normalization form to test for.
+   * @param options      the optional features to be enabled (0 for no options)
+   * @param errorCode  ICU error code in/out parameter.
+   *                   Must fulfill U_SUCCESS before the function call.
+   * @return Boolean value indicating whether the source string is in the
+   *         "mode" normalization form.
+   *
+   * @see quickCheck
+   * @deprecated ICU 56 Use Normalizer2 instead.
+   */
+  static UBool
+  isNormalized(const UnicodeString &src, UNormalizationMode mode, int32_t options, UErrorCode &errorCode);
+
+  /**
+   * Concatenate normalized strings, making sure that the result is normalized as well.
+   *
+   * If both the left and the right strings are in
+   * the normalization form according to "mode/options",
+   * then the result will be
+   *
+   * \code
+   *     dest=normalize(left+right, mode, options)
+   * \endcode
+   *
+   * For details see unorm_concatenate in unorm.h.
+   *
+   * @param left Left source string.
+   * @param right Right source string.
+   * @param result The output string.
+   * @param mode The normalization mode.
+   * @param options A bit set of normalization options.
+   * @param errorCode ICU error code in/out parameter.
+   *                   Must fulfill U_SUCCESS before the function call.
+   * @return result
+   *
+   * @see unorm_concatenate
+   * @see normalize
+   * @see unorm_next
+   * @see unorm_previous
+   *
+   * @deprecated ICU 56 Use Normalizer2 instead.
+   */
+  static UnicodeString &
+  U_EXPORT2 concatenate(const UnicodeString &left, const UnicodeString &right,
+              UnicodeString &result,
+              UNormalizationMode mode, int32_t options,
+              UErrorCode &errorCode);
+#endif  /* U_HIDE_DEPRECATED_API */
+
+  /**
+   * Compare two strings for canonical equivalence.
+   * Further options include case-insensitive comparison and
+   * code point order (as opposed to code unit order).
+   *
+   * Canonical equivalence between two strings is defined as their normalized
+   * forms (NFD or NFC) being identical.
+   * This function compares strings incrementally instead of normalizing
+   * (and optionally case-folding) both strings entirely,
+   * improving performance significantly.
+   *
+   * Bulk normalization is only necessary if the strings do not fulfill the FCD
+   * conditions. Only in this case, and only if the strings are relatively long,
+   * is memory allocated temporarily.
+   * For FCD strings and short non-FCD strings there is no memory allocation.
+   *
+   * Semantically, this is equivalent to
+   *   strcmp[CodePointOrder](NFD(foldCase(s1)), NFD(foldCase(s2)))
+   * where code point order and foldCase are all optional.
+   *
+   * UAX 21 2.5 Caseless Matching specifies that for a canonical caseless match
+   * the case folding must be performed first, then the normalization.
+   *
+   * @param s1 First source string.
+   * @param s2 Second source string.
+   *
+   * @param options A bit set of options:
+   *   - U_FOLD_CASE_DEFAULT or 0 is used for default options:
+   *     Case-sensitive comparison in code unit order, and the input strings
+   *     are quick-checked for FCD.
+   *
+   *   - UNORM_INPUT_IS_FCD
+   *     Set if the caller knows that both s1 and s2 fulfill the FCD conditions.
+   *     If not set, the function will quickCheck for FCD
+   *     and normalize if necessary.
+   *
+   *   - U_COMPARE_CODE_POINT_ORDER
+   *     Set to choose code point order instead of code unit order
+   *     (see u_strCompare for details).
+   *
+   *   - U_COMPARE_IGNORE_CASE
+   *     Set to compare strings case-insensitively using case folding,
+   *     instead of case-sensitively.
+   *     If set, then the following case folding options are used.
+   *
+   *   - Options as used with case-insensitive comparisons, currently:
+   *
+   *   - U_FOLD_CASE_EXCLUDE_SPECIAL_I
+   *    (see u_strCaseCompare for details)
+   *
+   *   - regular normalization options shifted left by UNORM_COMPARE_NORM_OPTIONS_SHIFT
+   *
+   * @param errorCode ICU error code in/out parameter.
+   *                  Must fulfill U_SUCCESS before the function call.
+   * @return <0 or 0 or >0 as usual for string comparisons
+   *
+   * @see unorm_compare
+   * @see normalize
+   * @see UNORM_FCD
+   * @see u_strCompare
+   * @see u_strCaseCompare
+   *
+   * @stable ICU 2.2
+   */
+  static inline int32_t
+  compare(const UnicodeString &s1, const UnicodeString &s2,
+          uint32_t options,
+          UErrorCode &errorCode);
+
+#ifndef U_HIDE_DEPRECATED_API
+  //-------------------------------------------------------------------------
+  // Iteration API
+  //-------------------------------------------------------------------------
+
+  /**
+   * Return the current character in the normalized text.
+   * current() may need to normalize some text at getIndex().
+   * The getIndex() is not changed.
+   *
+   * @return the current normalized code point
+   * @deprecated ICU 56 Use Normalizer2 instead.
+   */
+  UChar32 current();
+
+  /**
+   * Return the first character in the normalized text.
+   * This is equivalent to setIndexOnly(startIndex()) followed by next().
+   * (Post-increment semantics.)
+   *
+   * @return the first normalized code point
+   * @deprecated ICU 56 Use Normalizer2 instead.
+   */
+  UChar32 first();
+
+  /**
+   * Return the last character in the normalized text.
+   * This is equivalent to setIndexOnly(endIndex()) followed by previous().
+   * (Pre-decrement semantics.)
+   *
+   * @return the last normalized code point
+   * @deprecated ICU 56 Use Normalizer2 instead.
+   */
+  UChar32 last();
+
+  /**
+   * Return the next character in the normalized text.
+   * (Post-increment semantics.)
+   * If the end of the text has already been reached, DONE is returned.
+   * The DONE value could be confused with a U+FFFF non-character code point
+   * in the text. If this is possible, you can test getIndex()<endIndex()
+   * before calling next(), or (getIndex()<endIndex() || last()!=DONE)
+   * after calling next(). (Calling last() will change the iterator state!)
+   *
+   * The C API unorm_next() is more efficient and does not have this ambiguity.
+   *
+   * @return the next normalized code point
+   * @deprecated ICU 56 Use Normalizer2 instead.
+   */
+  UChar32 next();
+
+  /**
+   * Return the previous character in the normalized text and decrement.
+   * (Pre-decrement semantics.)
+   * If the beginning of the text has already been reached, DONE is returned.
+   * The DONE value could be confused with a U+FFFF non-character code point
+   * in the text. If this is possible, you can test
+   * (getIndex()>startIndex() || first()!=DONE). (Calling first() will change
+   * the iterator state!)
+   *
+   * The C API unorm_previous() is more efficient and does not have this ambiguity.
+   *
+   * @return the previous normalized code point
+   * @deprecated ICU 56 Use Normalizer2 instead.
+   */
+  UChar32 previous();
+
+  /**
+   * Set the iteration position in the input text that is being normalized,
+   * without any immediate normalization.
+   * After setIndexOnly(), getIndex() will return the same index that is
+   * specified here.
+   *
+   * @param index the desired index in the input text.
+   * @deprecated ICU 56 Use Normalizer2 instead.
+   */
+  void                 setIndexOnly(int32_t index);
+
+  /**
+   * Reset the index to the beginning of the text.
+   * This is equivalent to setIndexOnly(startIndex)).
+   * @deprecated ICU 56 Use Normalizer2 instead.
+   */
+  void reset();
+
+  /**
+   * Retrieve the current iteration position in the input text that is
+   * being normalized.
+   *
+   * A following call to next() will return a normalized code point from
+   * the input text at or after this index.
+   *
+   * After a call to previous(), getIndex() will point at or before the
+   * position in the input text where the normalized code point
+   * was returned from with previous().
+   *
+   * @return the current index in the input text
+   * @deprecated ICU 56 Use Normalizer2 instead.
+   */
+  int32_t getIndex() const;
+
+  /**
+   * Retrieve the index of the start of the input text. This is the begin index
+   * of the <code>CharacterIterator</code> or the start (i.e. index 0) of the string
+   * over which this <code>Normalizer</code> is iterating.
+   *
+   * @return the smallest index in the input text where the Normalizer operates
+   * @deprecated ICU 56 Use Normalizer2 instead.
+   */
+  int32_t startIndex() const;
+
+  /**
+   * Retrieve the index of the end of the input text. This is the end index
+   * of the <code>CharacterIterator</code> or the length of the string
+   * over which this <code>Normalizer</code> is iterating.
+   * This end index is exclusive, i.e., the Normalizer operates only on characters
+   * before this index.
+   *
+   * @return the first index in the input text where the Normalizer does not operate
+   * @deprecated ICU 56 Use Normalizer2 instead.
+   */
+  int32_t endIndex() const;
+
+  /**
+   * Returns true when both iterators refer to the same character in the same
+   * input text.
+   *
+   * @param that a Normalizer object to compare this one to
+   * @return comparison result
+   * @deprecated ICU 56 Use Normalizer2 instead.
+   */
+  bool         operator==(const Normalizer& that) const;
+
+  /**
+   * Returns false when both iterators refer to the same character in the same
+   * input text.
+   *
+   * @param that a Normalizer object to compare this one to
+   * @return comparison result
+   * @deprecated ICU 56 Use Normalizer2 instead.
+   */
+  inline bool         operator!=(const Normalizer& that) const;
+
+  /**
+   * Returns a pointer to a new Normalizer that is a clone of this one.
+   * The caller is responsible for deleting the new clone.
+   * @return a pointer to a new Normalizer
+   * @deprecated ICU 56 Use Normalizer2 instead.
+   */
+  Normalizer*        clone() const;
+
+  /**
+   * Generates a hash code for this iterator.
+   *
+   * @return the hash code
+   * @deprecated ICU 56 Use Normalizer2 instead.
+   */
+  int32_t hashCode() const;
+
+  //-------------------------------------------------------------------------
+  // Property access methods
+  //-------------------------------------------------------------------------
+
+  /**
+   * Set the normalization mode for this object.
+   * <p>
+   * <b>Note:</b>If the normalization mode is changed while iterating
+   * over a string, calls to {@link #next() } and {@link #previous() } may
+   * return previously buffers characters in the old normalization mode
+   * until the iteration is able to re-sync at the next base character.
+   * It is safest to call {@link #setIndexOnly }, {@link #reset() },
+   * {@link #setText }, {@link #first() },
+   * {@link #last() }, etc. after calling <code>setMode</code>.
+   * <p>
+   * @param newMode the new mode for this <code>Normalizer</code>.
+   * @see #getUMode
+   * @deprecated ICU 56 Use Normalizer2 instead.
+   */
+  void setMode(UNormalizationMode newMode);
+
+  /**
+   * Return the normalization mode for this object.
+   *
+   * This is an unusual name because there used to be a getMode() that
+   * returned a different type.
+   *
+   * @return the mode for this <code>Normalizer</code>
+   * @see #setMode
+   * @deprecated ICU 56 Use Normalizer2 instead.
+   */
+  UNormalizationMode getUMode() const;
+
+  /**
+   * Set options that affect this <code>Normalizer</code>'s operation.
+   * Options do not change the basic composition or decomposition operation
+   * that is being performed, but they control whether
+   * certain optional portions of the operation are done.
+   * Currently the only available option is obsolete.
+   *
+   * It is possible to specify multiple options that are all turned on or off.
+   *
+   * @param   option  the option(s) whose value is/are to be set.
+   * @param   value   the new setting for the option.  Use <code>true</code> to
+   *                  turn the option(s) on and <code>false</code> to turn it/them off.
+   *
+   * @see #getOption
+   * @deprecated ICU 56 Use Normalizer2 instead.
+   */
+  void setOption(int32_t option,
+         UBool value);
+
+  /**
+   * Determine whether an option is turned on or off.
+   * If multiple options are specified, then the result is true if any
+   * of them are set.
+   * <p>
+   * @param option the option(s) that are to be checked
+   * @return true if any of the option(s) are set
+   * @see #setOption
+   * @deprecated ICU 56 Use Normalizer2 instead.
+   */
+  UBool getOption(int32_t option) const;
+
+  /**
+   * Set the input text over which this <code>Normalizer</code> will iterate.
+   * The iteration position is set to the beginning.
+   *
+   * @param newText a string that replaces the current input text
+   * @param status a UErrorCode
+   * @deprecated ICU 56 Use Normalizer2 instead.
+   */
+  void setText(const UnicodeString& newText,
+           UErrorCode &status);
+
+  /**
+   * Set the input text over which this <code>Normalizer</code> will iterate.
+   * The iteration position is set to the beginning.
+   *
+   * @param newText a CharacterIterator object that replaces the current input text
+   * @param status a UErrorCode
+   * @deprecated ICU 56 Use Normalizer2 instead.
+   */
+  void setText(const CharacterIterator& newText,
+           UErrorCode &status);
+
+  /**
+   * Set the input text over which this <code>Normalizer</code> will iterate.
+   * The iteration position is set to the beginning.
+   *
+   * @param newText a string that replaces the current input text
+   * @param length the length of the string, or -1 if NUL-terminated
+   * @param status a UErrorCode
+   * @deprecated ICU 56 Use Normalizer2 instead.
+   */
+  void setText(ConstChar16Ptr newText,
+                    int32_t length,
+            UErrorCode &status);
+  /**
+   * Copies the input text into the UnicodeString argument.
+   *
+   * @param result Receives a copy of the text under iteration.
+   * @deprecated ICU 56 Use Normalizer2 instead.
+   */
+  void            getText(UnicodeString&  result);
+
+  /**
+   * ICU "poor man's RTTI", returns a UClassID for this class.
+   * @returns a UClassID for this class.
+   * @deprecated ICU 56 Use Normalizer2 instead.
+   */
+  static UClassID U_EXPORT2 getStaticClassID();
+#endif  /* U_HIDE_DEPRECATED_API */
+
+#ifndef U_FORCE_HIDE_DEPRECATED_API
+  /**
+   * ICU "poor man's RTTI", returns a UClassID for the actual class.
+   * @return a UClassID for the actual class.
+   * @deprecated ICU 56 Use Normalizer2 instead.
+   */
+  virtual UClassID getDynamicClassID() const override;
+#endif  // U_FORCE_HIDE_DEPRECATED_API
+
+private:
+  //-------------------------------------------------------------------------
+  // Private functions
+  //-------------------------------------------------------------------------
+
+  Normalizer() = delete; // default constructor not implemented
+  Normalizer &operator=(const Normalizer &that) = delete; // assignment operator not implemented
+
+  // Private utility methods for iteration
+  // For documentation, see the source code
+  UBool nextNormalize();
+  UBool previousNormalize();
+
+  void    init();
+  void clearBuffer();
+
+  //-------------------------------------------------------------------------
+  // Private data
+  //-------------------------------------------------------------------------
+
+  FilteredNormalizer2*fFilteredNorm2;  // owned if not nullptr
+  const Normalizer2  *fNorm2;  // not owned; may be equal to fFilteredNorm2
+  UNormalizationMode  fUMode;  // deprecated
+  int32_t             fOptions;
+
+  // The input text and our position in it
+  CharacterIterator  *text;
+
+  // The normalization buffer is the result of normalization
+  // of the source in [currentIndex..nextIndex[ .
+  int32_t         currentIndex, nextIndex;
+
+  // A buffer for holding intermediate results
+  UnicodeString       buffer;
+  int32_t         bufferPos;
+};
+
+//-------------------------------------------------------------------------
+// Inline implementations
+//-------------------------------------------------------------------------
+
+#ifndef U_HIDE_DEPRECATED_API
+inline bool
+Normalizer::operator!= (const Normalizer& other) const
+{ return ! operator==(other); }
+
+inline UNormalizationCheckResult
+Normalizer::quickCheck(const UnicodeString& source,
+                       UNormalizationMode mode,
+                       UErrorCode &status) {
+    return quickCheck(source, mode, 0, status);
+}
+
+inline UBool
+Normalizer::isNormalized(const UnicodeString& source,
+                         UNormalizationMode mode,
+                         UErrorCode &status) {
+    return isNormalized(source, mode, 0, status);
+}
+#endif  /* U_HIDE_DEPRECATED_API */
+
+inline int32_t
+Normalizer::compare(const UnicodeString &s1, const UnicodeString &s2,
+                    uint32_t options,
+                    UErrorCode &errorCode) {
+  // all argument checking is done in unorm_compare
+  return unorm_compare(toUCharPtr(s1.getBuffer()), s1.length(),
+                       toUCharPtr(s2.getBuffer()), s2.length(),
+                       options,
+                       &errorCode);
+}
+
+U_NAMESPACE_END
+
+#endif /* #if !UCONFIG_NO_NORMALIZATION */
+
+#endif // NORMLZR_H
+
+#endif /* U_SHOW_CPLUSPLUS_API */

thirdparty/icu4c/common/unicode/parseerr.h

@@ -0,0 +1,94 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+/*
+**********************************************************************
+*   Copyright (C) 1999-2005, International Business Machines
+*   Corporation and others.  All Rights Reserved.
+**********************************************************************
+*   Date        Name        Description
+*   03/14/00    aliu        Creation.
+*   06/27/00    aliu        Change from C++ class to C struct
+**********************************************************************
+*/
+#ifndef PARSEERR_H
+#define PARSEERR_H
+
+#include "unicode/utypes.h"
+
+
+/**
+ * \file
+ * \brief C API: Parse Error Information
+ */
+/**
+ * The capacity of the context strings in UParseError.
+ * @stable ICU 2.0
+ */ 
+enum { U_PARSE_CONTEXT_LEN = 16 };
+
+/**
+ * A UParseError struct is used to returned detailed information about
+ * parsing errors.  It is used by ICU parsing engines that parse long
+ * rules, patterns, or programs, where the text being parsed is long
+ * enough that more information than a UErrorCode is needed to
+ * localize the error.
+ *
+ * <p>The line, offset, and context fields are optional; parsing
+ * engines may choose not to use to use them.
+ *
+ * <p>The preContext and postContext strings include some part of the
+ * context surrounding the error.  If the source text is "let for=7"
+ * and "for" is the error (e.g., because it is a reserved word), then
+ * some examples of what a parser might produce are the following:
+ *
+ * <pre>
+ * preContext   postContext
+ * ""           ""            The parser does not support context
+ * "let "       "=7"          Pre- and post-context only
+ * "let "       "for=7"       Pre- and post-context and error text
+ * ""           "for"         Error text only
+ * </pre>
+ *
+ * <p>Examples of engines which use UParseError (or may use it in the
+ * future) are Transliterator, RuleBasedBreakIterator, and
+ * RegexPattern.
+ * 
+ * @stable ICU 2.0
+ */
+typedef struct UParseError {
+
+    /**
+     * The line on which the error occurred.  If the parser uses this
+     * field, it sets it to the line number of the source text line on
+     * which the error appears, which will be a value >= 1.  If the
+     * parse does not support line numbers, the value will be <= 0.
+     * @stable ICU 2.0
+     */
+    int32_t        line;
+
+    /**
+     * The character offset to the error.  If the line field is >= 1,
+     * then this is the offset from the start of the line.  Otherwise,
+     * this is the offset from the start of the text.  If the parser
+     * does not support this field, it will have a value < 0.
+     * @stable ICU 2.0
+     */
+    int32_t        offset;
+
+    /**
+     * Textual context before the error.  Null-terminated.  The empty
+     * string if not supported by parser.
+     * @stable ICU 2.0   
+     */
+    UChar          preContext[U_PARSE_CONTEXT_LEN];
+
+    /**
+     * The error itself and/or textual context after the error.
+     * Null-terminated.  The empty string if not supported by parser.
+     * @stable ICU 2.0   
+     */
+    UChar          postContext[U_PARSE_CONTEXT_LEN];
+
+} UParseError;
+
+#endif

thirdparty/icu4c/common/unicode/parsepos.h

@@ -0,0 +1,237 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+/*
+* Copyright (C) 1997-2005, International Business Machines Corporation and others. All Rights Reserved.
+*******************************************************************************
+*
+* File PARSEPOS.H
+*
+* Modification History:
+*
+*   Date        Name        Description
+*   07/09/97    helena      Converted from java.
+*   07/17/98    stephen     Added errorIndex support.
+*   05/11/99    stephen     Cleaned up.
+*******************************************************************************
+*/
+
+#ifndef PARSEPOS_H
+#define PARSEPOS_H
+
+#include "unicode/utypes.h"
+
+#if U_SHOW_CPLUSPLUS_API
+
+#include "unicode/uobject.h"
+
+ 
+U_NAMESPACE_BEGIN
+
+/**
+ * \file
+ * \brief C++ API: Canonical Iterator
+ */
+/** 
+ * <code>ParsePosition</code> is a simple class used by <code>Format</code>
+ * and its subclasses to keep track of the current position during parsing.
+ * The <code>parseObject</code> method in the various <code>Format</code>
+ * classes requires a <code>ParsePosition</code> object as an argument.
+ *
+ * <p>
+ * By design, as you parse through a string with different formats,
+ * you can use the same <code>ParsePosition</code>, since the index parameter
+ * records the current position.
+ *
+ * The ParsePosition class is not suitable for subclassing.
+ *
+ * @version     1.3 10/30/97
+ * @author      Mark Davis, Helena Shih
+ * @see         java.text.Format
+ */
+
+class U_COMMON_API ParsePosition : public UObject {
+public:
+    /**
+     * Default constructor, the index starts with 0 as default.
+     * @stable ICU 2.0
+     */
+    ParsePosition()
+        : UObject(),
+        index(0),
+        errorIndex(-1)
+      {}
+
+    /**
+     * Create a new ParsePosition with the given initial index.
+     * @param newIndex the new text offset.
+     * @stable ICU 2.0
+     */
+    ParsePosition(int32_t newIndex)
+        : UObject(),
+        index(newIndex),
+        errorIndex(-1)
+      {}
+
+    /**
+     * Copy constructor
+     * @param copy the object to be copied from.
+     * @stable ICU 2.0
+     */
+    ParsePosition(const ParsePosition& copy)
+        : UObject(copy),
+        index(copy.index),
+        errorIndex(copy.errorIndex)
+      {}
+
+    /**
+     * Destructor
+     * @stable ICU 2.0
+     */
+    virtual ~ParsePosition();
+
+    /**
+     * Assignment operator
+     * @stable ICU 2.0
+     */
+    inline ParsePosition&      operator=(const ParsePosition& copy);
+
+    /**
+     * Equality operator.
+     * @return true if the two parse positions are equal, false otherwise.
+     * @stable ICU 2.0
+     */
+    inline bool               operator==(const ParsePosition& that) const;
+
+    /**
+     * Equality operator.
+     * @return true if the two parse positions are not equal, false otherwise.
+     * @stable ICU 2.0
+     */
+    inline bool               operator!=(const ParsePosition& that) const;
+
+    /**
+     * Clone this object.
+     * Clones can be used concurrently in multiple threads.
+     * If an error occurs, then nullptr is returned.
+     * The caller must delete the clone.
+     *
+     * @return a clone of this object
+     *
+     * @see getDynamicClassID
+     * @stable ICU 2.8
+     */
+    ParsePosition *clone() const;
+
+    /**
+     * Retrieve the current parse position.  On input to a parse method, this
+     * is the index of the character at which parsing will begin; on output, it
+     * is the index of the character following the last character parsed.
+     * @return the current index.
+     * @stable ICU 2.0
+     */
+    inline int32_t getIndex() const;
+
+    /**
+     * Set the current parse position.
+     * @param index the new index.
+     * @stable ICU 2.0
+     */
+    inline void setIndex(int32_t index);
+
+    /**
+     * Set the index at which a parse error occurred.  Formatters
+     * should set this before returning an error code from their
+     * parseObject method.  The default value is -1 if this is not
+     * set.
+     * @stable ICU 2.0
+     */
+    inline void setErrorIndex(int32_t ei);
+
+    /**
+     * Retrieve the index at which an error occurred, or -1 if the
+     * error index has not been set.
+     * @stable ICU 2.0
+     */
+    inline int32_t getErrorIndex() const;
+
+    /**
+     * ICU "poor man's RTTI", returns a UClassID for this class.
+     *
+     * @stable ICU 2.2
+     */
+    static UClassID U_EXPORT2 getStaticClassID();
+
+    /**
+     * ICU "poor man's RTTI", returns a UClassID for the actual class.
+     *
+     * @stable ICU 2.2
+     */
+    virtual UClassID getDynamicClassID() const override;
+
+private:
+    /**
+     * Input: the place you start parsing.
+     * <br>Output: position where the parse stopped.
+     * This is designed to be used serially,
+     * with each call setting index up for the next one.
+     */
+    int32_t index;
+
+    /**
+     * The index at which a parse error occurred.
+     */
+    int32_t errorIndex;
+
+};
+
+inline ParsePosition&
+ParsePosition::operator=(const ParsePosition& copy)
+{
+  index = copy.index;
+  errorIndex = copy.errorIndex;
+  return *this;
+}
+
+inline bool
+ParsePosition::operator==(const ParsePosition& copy) const
+{
+  if(index != copy.index || errorIndex != copy.errorIndex)
+  return false;
+  else
+  return true;
+}
+
+inline bool
+ParsePosition::operator!=(const ParsePosition& copy) const
+{
+  return !operator==(copy);
+}
+
+inline int32_t
+ParsePosition::getIndex() const
+{
+  return index;
+}
+
+inline void
+ParsePosition::setIndex(int32_t offset)
+{
+  this->index = offset;
+}
+
+inline int32_t
+ParsePosition::getErrorIndex() const
+{
+  return errorIndex;
+}
+
+inline void
+ParsePosition::setErrorIndex(int32_t ei)
+{
+  this->errorIndex = ei;
+}
+U_NAMESPACE_END
+
+#endif /* U_SHOW_CPLUSPLUS_API */
+
+#endif

thirdparty/icu4c/common/unicode/platform.h

@@ -0,0 +1,861 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+/*
+******************************************************************************
+*
+*   Copyright (C) 1997-2016, International Business Machines
+*   Corporation and others.  All Rights Reserved.
+*
+******************************************************************************
+*
+*  FILE NAME : platform.h
+*
+*   Date        Name        Description
+*   05/13/98    nos         Creation (content moved here from ptypes.h).
+*   03/02/99    stephen     Added AS400 support.
+*   03/30/99    stephen     Added Linux support.
+*   04/13/99    stephen     Reworked for autoconf.
+******************************************************************************
+*/
+
+#ifndef _PLATFORM_H
+#define _PLATFORM_H
+
+#include "unicode/uconfig.h"
+#include "unicode/uvernum.h"
+
+/**
+ * \file
+ * \brief Basic types for the platform.
+ *
+ * This file used to be generated by autoconf/configure.
+ * Starting with ICU 49, platform.h is a normal source file,
+ * to simplify cross-compiling and working with non-autoconf/make build systems.
+ *
+ * When a value in this file does not work on a platform, then please
+ * try to derive it from the U_PLATFORM value
+ * (for which we might need a new value constant in rare cases)
+ * and/or from other macros that are predefined by the compiler
+ * or defined in standard (POSIX or platform or compiler) headers.
+ *
+ * As a temporary workaround, you can add an explicit \#define for some macros
+ * before it is first tested, or add an equivalent -D macro definition
+ * to the compiler's command line.
+ *
+ * Note: Some compilers provide ways to show the predefined macros.
+ * For example, with gcc you can compile an empty .c file and have the compiler
+ * print the predefined macros with
+ * \code
+ * gcc -E -dM -x c /dev/null | sort
+ * \endcode
+ * (You can provide an actual empty .c file rather than /dev/null.
+ * <code>-x c++</code> is for C++.)
+ */
+
+/**
+ * Define some things so that they can be documented.
+ * @internal
+ */
+#ifdef U_IN_DOXYGEN
+/*
+ * Problem: "platform.h:335: warning: documentation for unknown define U_HAVE_STD_STRING found." means that U_HAVE_STD_STRING is not documented.
+ * Solution: #define any defines for non @internal API here, so that they are visible in the docs.  If you just set PREDEFINED in Doxyfile.in,  they won't be documented.
+ */
+
+/* None for now. */
+#endif
+
+/**
+ * \def U_PLATFORM
+ * The U_PLATFORM macro defines the platform we're on.
+ *
+ * We used to define one different, value-less macro per platform.
+ * That made it hard to know the set of relevant platforms and macros,
+ * and hard to deal with variants of platforms.
+ *
+ * Starting with ICU 49, we define platforms as numeric macros,
+ * with ranges of values for related platforms and their variants.
+ * The U_PLATFORM macro is set to one of these values.
+ *
+ * Historical note from the Solaris Wikipedia article:
+ * AT&T and Sun collaborated on a project to merge the most popular Unix variants
+ * on the market at that time: BSD, System V, and Xenix.
+ * This became Unix System V Release 4 (SVR4).
+ *
+ * @internal
+ */
+
+/** Unknown platform. @internal */
+#define U_PF_UNKNOWN 0
+/** Windows @internal */
+#define U_PF_WINDOWS 1000
+/** MinGW. Windows, calls to Win32 API, but using GNU gcc and binutils. @internal */
+#define U_PF_MINGW 1800
+/**
+ * Cygwin. Windows, calls to cygwin1.dll for Posix functions,
+ * using MSVC or GNU gcc and binutils.
+ * @internal
+ */
+#define U_PF_CYGWIN 1900
+/* Reserve 2000 for U_PF_UNIX? */
+/** HP-UX is based on UNIX System V. @internal */
+#define U_PF_HPUX 2100
+/** Solaris is a Unix operating system based on SVR4. @internal */
+#define U_PF_SOLARIS 2600
+/** BSD is a UNIX operating system derivative. @internal */
+#define U_PF_BSD 3000
+/** AIX is based on UNIX System V Releases and 4.3 BSD. @internal */
+#define U_PF_AIX 3100
+/** IRIX is based on UNIX System V with BSD extensions. @internal */
+#define U_PF_IRIX 3200
+/**
+ * Darwin is a POSIX-compliant operating system, composed of code developed by Apple,
+ * as well as code derived from NeXTSTEP, BSD, and other projects,
+ * built around the Mach kernel.
+ * Darwin forms the core set of components upon which Mac OS X, Apple TV, and iOS are based.
+ * (Original description modified from WikiPedia.)
+ * @internal
+ */
+#define U_PF_DARWIN 3500
+/** iPhone OS (iOS) is a derivative of Mac OS X. @internal */
+#define U_PF_IPHONE 3550
+/** QNX is a commercial Unix-like real-time operating system related to BSD. @internal */
+#define U_PF_QNX 3700
+/** Linux is a Unix-like operating system. @internal */
+#define U_PF_LINUX 4000
+/**
+ * Native Client is pretty close to Linux.
+ * See https://developer.chrome.com/native-client and
+ *  http://www.chromium.org/nativeclient
+ *  @internal
+ */
+#define U_PF_BROWSER_NATIVE_CLIENT 4020
+/** Android is based on Linux. @internal */
+#define U_PF_ANDROID 4050
+/** Haiku is a POSIX-ish platform. @internal */
+#define U_PF_HAIKU 4080
+/** Fuchsia is a POSIX-ish platform. @internal */
+#define U_PF_FUCHSIA 4100
+/* Maximum value for Linux-based platform is 4499 */
+/**
+ * Emscripten is a C++ transpiler for the Web that can target asm.js or
+ * WebAssembly. It provides some POSIX-compatible wrappers and stubs and
+ * some Linux-like functionality, but is not fully compatible with
+ * either.
+ * @internal
+ */
+#define U_PF_EMSCRIPTEN 5010
+/** z/OS is the successor to OS/390 which was the successor to MVS. @internal */
+#define U_PF_OS390 9000
+/** "IBM i" is the current name of what used to be i5/OS and earlier OS/400. @internal */
+#define U_PF_OS400 9400
+
+#ifdef U_PLATFORM
+    /* Use the predefined value. */
+#elif defined(__MINGW32__)
+#   define U_PLATFORM U_PF_MINGW
+#elif defined(__CYGWIN__)
+#   define U_PLATFORM U_PF_CYGWIN
+    /* Cygwin uchar.h doesn't exist until Cygwin 3.5. */
+#   include <cygwin/version.h>
+#elif defined(WIN32) || defined(_WIN32) || defined(WIN64) || defined(_WIN64)
+#   define U_PLATFORM U_PF_WINDOWS
+#elif defined(__ANDROID__)
+#   define U_PLATFORM U_PF_ANDROID
+    /* Android wchar_t support depends on the API level. */
+#   include <android/api-level.h>
+#elif defined(__pnacl__) || defined(__native_client__)
+#   define U_PLATFORM U_PF_BROWSER_NATIVE_CLIENT
+#elif defined(__Fuchsia__)
+#   define U_PLATFORM U_PF_FUCHSIA
+#elif defined(linux) || defined(__linux__) || defined(__linux)
+#   define U_PLATFORM U_PF_LINUX
+#elif defined(__APPLE__) && defined(__MACH__)
+#   include <TargetConditionals.h>
+#   if (defined(TARGET_OS_IPHONE) && TARGET_OS_IPHONE) && (defined(TARGET_OS_MACCATALYST) && !TARGET_OS_MACCATALYST)   /* variant of TARGET_OS_MAC */
+#       define U_PLATFORM U_PF_IPHONE
+#   else
+#       define U_PLATFORM U_PF_DARWIN
+#   endif
+#elif defined(BSD) || defined(__FreeBSD__) || defined(__FreeBSD_kernel__) || defined(__NetBSD__) || defined(__OpenBSD__) || defined(__MirBSD__)
+#   if defined(__FreeBSD__)
+#       include <sys/endian.h>
+#   endif
+#   define U_PLATFORM U_PF_BSD
+#elif defined(sun) || defined(__sun)
+    /* Check defined(__SVR4) || defined(__svr4__) to distinguish Solaris from SunOS? */
+#   define U_PLATFORM U_PF_SOLARIS
+#   if defined(__GNUC__)
+        /* Solaris/GCC needs this header file to get the proper endianness. Normally, this
+         * header file is included with stddef.h but on Solairs/GCC, the GCC version of stddef.h
+         *  is included which does not include this header file.
+         */
+#       include <sys/isa_defs.h>
+#   endif
+#elif defined(_AIX) || defined(__TOS_AIX__)
+#   define U_PLATFORM U_PF_AIX
+#elif defined(_hpux) || defined(hpux) || defined(__hpux)
+#   define U_PLATFORM U_PF_HPUX
+#elif defined(sgi) || defined(__sgi)
+#   define U_PLATFORM U_PF_IRIX
+#elif defined(__QNX__) || defined(__QNXNTO__)
+#   define U_PLATFORM U_PF_QNX
+#elif defined(__TOS_MVS__)
+#   define U_PLATFORM U_PF_OS390
+#elif defined(__OS400__) || defined(__TOS_OS400__)
+#   define U_PLATFORM U_PF_OS400
+#elif defined(__HAIKU__)
+#   define U_PLATFORM U_PF_HAIKU
+#elif defined(__EMSCRIPTEN__)
+#   define U_PLATFORM U_PF_EMSCRIPTEN
+#else
+#   define U_PLATFORM U_PF_UNKNOWN
+#endif
+
+/**
+ * \def U_REAL_MSVC
+ * Defined if the compiler is the real MSVC compiler (and not something like
+ * Clang setting _MSC_VER in order to compile Windows code that requires it).
+ * Otherwise undefined.
+ * @internal
+ */
+#if (defined(_MSC_VER) && !(defined(__clang__) && __clang__)) || defined(U_IN_DOXYGEN)
+#   define U_REAL_MSVC
+#endif
+
+/**
+ * \def CYGWINMSVC
+ * Defined if this is Windows with Cygwin, but using MSVC rather than gcc.
+ * Otherwise undefined.
+ * @internal
+ */
+/* Commented out because this is already set in mh-cygwin-msvc
+#if U_PLATFORM == U_PF_CYGWIN && defined(_MSC_VER)
+#   define CYGWINMSVC
+#endif
+*/
+#ifdef U_IN_DOXYGEN
+#   define CYGWINMSVC
+#endif
+
+/**
+ * \def U_PLATFORM_USES_ONLY_WIN32_API
+ * Defines whether the platform uses only the Win32 API.
+ * Set to 1 for Windows/MSVC, ClangCL and MinGW but not Cygwin.
+ * @internal
+ */
+#ifdef U_PLATFORM_USES_ONLY_WIN32_API
+    /* Use the predefined value. */
+#elif (U_PF_WINDOWS <= U_PLATFORM && U_PLATFORM <= U_PF_MINGW) || defined(CYGWINMSVC)
+#   define U_PLATFORM_USES_ONLY_WIN32_API 1
+#else
+    /* Cygwin implements POSIX. */
+#   define U_PLATFORM_USES_ONLY_WIN32_API 0
+#endif
+
+/**
+ * \def U_PLATFORM_HAS_WIN32_API
+ * Defines whether the Win32 API is available on the platform.
+ * Set to 1 for Windows/MSVC, ClangCL, MinGW and Cygwin.
+ * @internal
+ */
+#ifdef U_PLATFORM_HAS_WIN32_API
+    /* Use the predefined value. */
+#elif U_PF_WINDOWS <= U_PLATFORM && U_PLATFORM <= U_PF_CYGWIN
+#   define U_PLATFORM_HAS_WIN32_API 1
+#else
+#   define U_PLATFORM_HAS_WIN32_API 0
+#endif
+
+/**
+ * \def U_PLATFORM_HAS_WINUWP_API
+ * Defines whether target is intended for Universal Windows Platform API
+ * Set to 1 for Windows10 Release Solution Configuration
+ * @internal
+ */
+#ifdef U_PLATFORM_HAS_WINUWP_API
+    /* Use the predefined value. */
+#else
+#   define U_PLATFORM_HAS_WINUWP_API 0
+#endif
+
+/**
+ * \def U_PLATFORM_IMPLEMENTS_POSIX
+ * Defines whether the platform implements (most of) the POSIX API.
+ * Set to 1 for Cygwin and most other platforms.
+ * @internal
+ */
+#ifdef U_PLATFORM_IMPLEMENTS_POSIX
+    /* Use the predefined value. */
+#elif U_PLATFORM_USES_ONLY_WIN32_API
+#   define U_PLATFORM_IMPLEMENTS_POSIX 0
+#else
+#   define U_PLATFORM_IMPLEMENTS_POSIX 1
+#endif
+
+/**
+ * \def U_PLATFORM_IS_LINUX_BASED
+ * Defines whether the platform is Linux or one of its derivatives.
+ * @internal
+ */
+#ifdef U_PLATFORM_IS_LINUX_BASED
+    /* Use the predefined value. */
+#elif U_PF_LINUX <= U_PLATFORM && U_PLATFORM <= 4499
+#   define U_PLATFORM_IS_LINUX_BASED 1
+#else
+#   define U_PLATFORM_IS_LINUX_BASED 0
+#endif
+
+/**
+ * \def U_PLATFORM_IS_DARWIN_BASED
+ * Defines whether the platform is Darwin or one of its derivatives.
+ * @internal
+ */
+#ifdef U_PLATFORM_IS_DARWIN_BASED
+    /* Use the predefined value. */
+#elif U_PF_DARWIN <= U_PLATFORM && U_PLATFORM <= U_PF_IPHONE
+#   define U_PLATFORM_IS_DARWIN_BASED 1
+#else
+#   define U_PLATFORM_IS_DARWIN_BASED 0
+#endif
+
+/*===========================================================================*/
+/** @{ Compiler and environment features                                     */
+/*===========================================================================*/
+
+/**
+ * \def U_GCC_MAJOR_MINOR
+ * Indicates whether the compiler is gcc (test for != 0),
+ * and if so, contains its major (times 100) and minor version numbers.
+ * If the compiler is not gcc, then U_GCC_MAJOR_MINOR == 0.
+ *
+ * For example, for testing for whether we have gcc, and whether it's 4.6 or higher,
+ * use "#if U_GCC_MAJOR_MINOR >= 406".
+ * @internal
+ */
+#ifdef __GNUC__
+#   define U_GCC_MAJOR_MINOR (__GNUC__ * 100 + __GNUC_MINOR__)
+#else
+#   define U_GCC_MAJOR_MINOR 0
+#endif
+
+/**
+ * \def U_IS_BIG_ENDIAN
+ * Determines the endianness of the platform.
+ * @internal
+ */
+#ifdef U_IS_BIG_ENDIAN
+    /* Use the predefined value. */
+#elif defined(BYTE_ORDER) && defined(BIG_ENDIAN)
+#   define U_IS_BIG_ENDIAN (BYTE_ORDER == BIG_ENDIAN)
+#elif defined(__BYTE_ORDER__) && defined(__ORDER_BIG_ENDIAN__)
+    /* gcc */
+#   define U_IS_BIG_ENDIAN (__BYTE_ORDER__ == __ORDER_BIG_ENDIAN__)
+#elif defined(__BIG_ENDIAN__) || defined(_BIG_ENDIAN)
+#   define U_IS_BIG_ENDIAN 1
+#elif defined(__LITTLE_ENDIAN__) || defined(_LITTLE_ENDIAN)
+#   define U_IS_BIG_ENDIAN 0
+#elif U_PLATFORM == U_PF_OS390 || U_PLATFORM == U_PF_OS400 || defined(__s390__) || defined(__s390x__)
+    /* These platforms do not appear to predefine any endianness macros. */
+#   define U_IS_BIG_ENDIAN 1
+#elif defined(_PA_RISC1_0) || defined(_PA_RISC1_1) || defined(_PA_RISC2_0)
+    /* HPPA do not appear to predefine any endianness macros. */
+#   define U_IS_BIG_ENDIAN 1
+#elif defined(sparc) || defined(__sparc) || defined(__sparc__)
+    /* Some sparc based systems (e.g. Linux) do not predefine any endianness macros. */
+#   define U_IS_BIG_ENDIAN 1
+#else
+#   define U_IS_BIG_ENDIAN 0
+#endif
+
+/**
+ * \def U_HAVE_PLACEMENT_NEW
+ * Determines whether to override placement new and delete for STL.
+ * @stable ICU 2.6
+ */
+#ifdef U_HAVE_PLACEMENT_NEW
+    /* Use the predefined value. */
+#elif defined(__BORLANDC__)
+#   define U_HAVE_PLACEMENT_NEW 0
+#else
+#   define U_HAVE_PLACEMENT_NEW 1
+#endif
+
+/**
+ * \def U_HAVE_DEBUG_LOCATION_NEW 
+ * Define this to define the MFC debug version of the operator new.
+ *
+ * @stable ICU 3.4
+ */
+#ifdef U_HAVE_DEBUG_LOCATION_NEW
+    /* Use the predefined value. */
+#elif defined(_MSC_VER)
+#   define U_HAVE_DEBUG_LOCATION_NEW 1
+#else
+#   define U_HAVE_DEBUG_LOCATION_NEW 0
+#endif
+
+/* Compatibility with compilers other than clang: http://clang.llvm.org/docs/LanguageExtensions.html */
+#ifdef __has_attribute
+#   define UPRV_HAS_ATTRIBUTE(x) __has_attribute(x)
+#else
+#   define UPRV_HAS_ATTRIBUTE(x) 0
+#endif
+#ifdef __has_cpp_attribute
+#   define UPRV_HAS_CPP_ATTRIBUTE(x) __has_cpp_attribute(x)
+#else
+#   define UPRV_HAS_CPP_ATTRIBUTE(x) 0
+#endif
+#ifdef __has_declspec_attribute
+#   define UPRV_HAS_DECLSPEC_ATTRIBUTE(x) __has_declspec_attribute(x)
+#else
+#   define UPRV_HAS_DECLSPEC_ATTRIBUTE(x) 0
+#endif
+#ifdef __has_builtin
+#   define UPRV_HAS_BUILTIN(x) __has_builtin(x)
+#else
+#   define UPRV_HAS_BUILTIN(x) 0
+#endif
+#ifdef __has_feature
+#   define UPRV_HAS_FEATURE(x) __has_feature(x)
+#else
+#   define UPRV_HAS_FEATURE(x) 0
+#endif
+#ifdef __has_extension
+#   define UPRV_HAS_EXTENSION(x) __has_extension(x)
+#else
+#   define UPRV_HAS_EXTENSION(x) 0
+#endif
+#ifdef __has_warning
+#   define UPRV_HAS_WARNING(x) __has_warning(x)
+#else
+#   define UPRV_HAS_WARNING(x) 0
+#endif
+
+
+#if defined(__clang__)
+#define UPRV_NO_SANITIZE_UNDEFINED __attribute__((no_sanitize("undefined")))
+#else
+#define UPRV_NO_SANITIZE_UNDEFINED
+#endif
+
+/**
+ * \def U_MALLOC_ATTR
+ * Attribute to mark functions as malloc-like
+ * @internal
+ */
+#if defined(__GNUC__) && __GNUC__>=3
+#    define U_MALLOC_ATTR __attribute__ ((__malloc__))
+#else
+#    define U_MALLOC_ATTR
+#endif
+
+/**
+ * \def U_ALLOC_SIZE_ATTR
+ * Attribute to specify the size of the allocated buffer for malloc-like functions
+ * @internal
+ */
+#if (defined(__GNUC__) &&                                            \
+        (__GNUC__ > 4 || (__GNUC__ == 4 && __GNUC_MINOR__ >= 3))) || \
+        UPRV_HAS_ATTRIBUTE(alloc_size)
+#   define U_ALLOC_SIZE_ATTR(X) __attribute__ ((alloc_size(X)))
+#   define U_ALLOC_SIZE_ATTR2(X,Y) __attribute__ ((alloc_size(X,Y)))
+#else
+#   define U_ALLOC_SIZE_ATTR(X)
+#   define U_ALLOC_SIZE_ATTR2(X,Y)
+#endif
+
+/**
+ * \def U_CPLUSPLUS_VERSION
+ * 0 if no C++; 1, 11, 14, ... if C++.
+ * Support for specific features cannot always be determined by the C++ version alone.
+ * @internal
+ */
+#ifdef U_CPLUSPLUS_VERSION
+#   if U_CPLUSPLUS_VERSION != 0 && !defined(__cplusplus)
+#       undef U_CPLUSPLUS_VERSION
+#       define U_CPLUSPLUS_VERSION 0
+#   endif
+    /* Otherwise use the predefined value. */
+#elif !defined(__cplusplus)
+#   define U_CPLUSPLUS_VERSION 0
+#elif __cplusplus >= 201703L || (defined(_MSVC_LANG) && _MSVC_LANG >= 201703L)
+#   define U_CPLUSPLUS_VERSION 17
+#elif __cplusplus >= 201402L || (defined(_MSVC_LANG) && _MSVC_LANG >= 201402L)
+#   define U_CPLUSPLUS_VERSION 14
+#elif __cplusplus >= 201103L || (defined(_MSVC_LANG) && _MSVC_LANG >= 201103L)
+#   define U_CPLUSPLUS_VERSION 11
+#else
+    // C++98 or C++03
+#   define U_CPLUSPLUS_VERSION 1
+#endif
+
+/**
+ * \def U_FALLTHROUGH
+ * Annotate intentional fall-through between switch labels.
+ * http://clang.llvm.org/docs/AttributeReference.html#fallthrough-clang-fallthrough
+ * @internal
+ */
+#ifndef __cplusplus
+    // Not for C.
+#elif defined(U_FALLTHROUGH)
+    // Use the predefined value.
+#elif defined(__clang__)
+    // Test for compiler vs. feature separately.
+    // Other compilers might choke on the feature test.
+#    if UPRV_HAS_CPP_ATTRIBUTE(clang::fallthrough) || \
+             (UPRV_HAS_FEATURE(cxx_attributes) &&     \
+             UPRV_HAS_WARNING("-Wimplicit-fallthrough"))
+#       define U_FALLTHROUGH [[clang::fallthrough]]
+#   endif
+#elif defined(__GNUC__) && (__GNUC__ >= 7)
+#   define U_FALLTHROUGH __attribute__((fallthrough))
+#endif
+
+#ifndef U_FALLTHROUGH
+#   define U_FALLTHROUGH
+#endif
+
+/** @} */
+
+/*===========================================================================*/
+/** @{ Character data types                                                  */
+/*===========================================================================*/
+
+/**
+ * U_CHARSET_FAMILY is equal to this value when the platform is an ASCII based platform.
+ * @stable ICU 2.0
+ */
+#define U_ASCII_FAMILY 0
+
+/**
+ * U_CHARSET_FAMILY is equal to this value when the platform is an EBCDIC based platform.
+ * @stable ICU 2.0
+ */
+#define U_EBCDIC_FAMILY 1
+
+/**
+ * \def U_CHARSET_FAMILY
+ *
+ * <p>These definitions allow to specify the encoding of text
+ * in the char data type as defined by the platform and the compiler.
+ * It is enough to determine the code point values of "invariant characters",
+ * which are the ones shared by all encodings that are in use
+ * on a given platform.</p>
+ *
+ * <p>Those "invariant characters" should be all the uppercase and lowercase
+ * latin letters, the digits, the space, and "basic punctuation".
+ * Also, '\\n', '\\r', '\\t' should be available.</p>
+ *
+ * <p>The list of "invariant characters" is:<br>
+ * \code
+ *    A-Z  a-z  0-9  SPACE  "  %  &amp;  '  (  )  *  +  ,  -  .  /  :  ;  <  =  >  ?  _
+ * \endcode
+ * <br>
+ * (52 letters + 10 numbers + 20 punc/sym/space = 82 total)</p>
+ *
+ * <p>This matches the IBM Syntactic Character Set (CS 640).</p>
+ *
+ * <p>In other words, all the graphic characters in 7-bit ASCII should
+ * be safely accessible except the following:</p>
+ *
+ * \code
+ *    '\' <backslash>
+ *    '[' <left bracket>
+ *    ']' <right bracket>
+ *    '{' <left brace>
+ *    '}' <right brace>
+ *    '^' <circumflex>
+ *    '~' <tilde>
+ *    '!' <exclamation mark>
+ *    '#' <number sign>
+ *    '|' <vertical line>
+ *    '$' <dollar sign>
+ *    '@' <commercial at>
+ *    '`' <grave accent>
+ * \endcode
+ * @stable ICU 2.0
+ */
+#ifdef U_CHARSET_FAMILY
+    /* Use the predefined value. */
+#elif U_PLATFORM == U_PF_OS390 && (!defined(__CHARSET_LIB) || !__CHARSET_LIB)
+#   define U_CHARSET_FAMILY U_EBCDIC_FAMILY
+#elif U_PLATFORM == U_PF_OS400 && !defined(__UTF32__)
+#   define U_CHARSET_FAMILY U_EBCDIC_FAMILY
+#else
+#   define U_CHARSET_FAMILY U_ASCII_FAMILY
+#endif
+
+/**
+ * \def U_CHARSET_IS_UTF8
+ *
+ * Hardcode the default charset to UTF-8.
+ *
+ * If this is set to 1, then
+ * - ICU will assume that all non-invariant char*, StringPiece, std::string etc.
+ *   contain UTF-8 text, regardless of what the system API uses
+ * - some ICU code will use fast functions like u_strFromUTF8()
+ *   rather than the more general and more heavy-weight conversion API (ucnv.h)
+ * - ucnv_getDefaultName() always returns "UTF-8"
+ * - ucnv_setDefaultName() is disabled and will not change the default charset
+ * - static builds of ICU are smaller
+ * - more functionality is available with the UCONFIG_NO_CONVERSION build-time
+ *   configuration option (see unicode/uconfig.h)
+ * - the UCONFIG_NO_CONVERSION build option in uconfig.h is more usable
+ *
+ * @stable ICU 4.2
+ * @see UCONFIG_NO_CONVERSION
+ */
+#ifdef U_CHARSET_IS_UTF8
+    /* Use the predefined value. */
+#elif U_PLATFORM_IS_LINUX_BASED || U_PLATFORM_IS_DARWIN_BASED || \
+        U_PLATFORM == U_PF_EMSCRIPTEN
+#   define U_CHARSET_IS_UTF8 1
+#else
+#   define U_CHARSET_IS_UTF8 0
+#endif
+
+/** @} */
+
+/*===========================================================================*/
+/** @{ Information about wchar support                                       */
+/*===========================================================================*/
+
+/**
+ * \def U_HAVE_WCHAR_H
+ * Indicates whether <wchar.h> is available (1) or not (0). Set to 1 by default.
+ *
+ * @stable ICU 2.0
+ */
+#ifdef U_HAVE_WCHAR_H
+    /* Use the predefined value. */
+#elif U_PLATFORM == U_PF_ANDROID && __ANDROID_API__ < 9
+    /*
+     * Android before Gingerbread (Android 2.3, API level 9) did not support wchar_t.
+     * The type and header existed, but the library functions did not work as expected.
+     * The size of wchar_t was 1 but L"xyz" string literals had 32-bit units anyway.
+     */
+#   define U_HAVE_WCHAR_H 0
+#else
+#   define U_HAVE_WCHAR_H 1
+#endif
+
+/**
+ * \def U_SIZEOF_WCHAR_T
+ * U_SIZEOF_WCHAR_T==sizeof(wchar_t)
+ *
+ * @stable ICU 2.0
+ */
+#ifdef U_SIZEOF_WCHAR_T
+    /* Use the predefined value. */
+#elif (U_PLATFORM == U_PF_ANDROID && __ANDROID_API__ < 9)
+    /*
+     * Classic Mac OS and Mac OS X before 10.3 (Panther) did not support wchar_t or wstring.
+     * Newer Mac OS X has size 4.
+     */
+#   define U_SIZEOF_WCHAR_T 1
+#elif U_PLATFORM_HAS_WIN32_API || U_PLATFORM == U_PF_CYGWIN
+#   define U_SIZEOF_WCHAR_T 2
+#elif U_PLATFORM == U_PF_AIX
+    /*
+     * AIX 6.1 information, section "Wide character data representation":
+     * "... the wchar_t datatype is 32-bit in the 64-bit environment and
+     * 16-bit in the 32-bit environment."
+     * and
+     * "All locales use Unicode for their wide character code values (process code),
+     * except the IBM-eucTW codeset."
+     */
+#   ifdef __64BIT__
+#       define U_SIZEOF_WCHAR_T 4
+#   else
+#       define U_SIZEOF_WCHAR_T 2
+#   endif
+#elif U_PLATFORM == U_PF_OS390
+    /*
+     * z/OS V1R11 information center, section "LP64 | ILP32":
+     * "In 31-bit mode, the size of long and pointers is 4 bytes and the size of wchar_t is 2 bytes.
+     * Under LP64, the size of long and pointer is 8 bytes and the size of wchar_t is 4 bytes."
+     */
+#   ifdef _LP64
+#       define U_SIZEOF_WCHAR_T 4
+#   else
+#       define U_SIZEOF_WCHAR_T 2
+#   endif
+#elif U_PLATFORM == U_PF_OS400
+#   if defined(__UTF32__)
+        /*
+         * LOCALETYPE(*LOCALEUTF) is specified.
+         * Wide-character strings are in UTF-32,
+         * narrow-character strings are in UTF-8.
+         */
+#       define U_SIZEOF_WCHAR_T 4
+#   elif defined(__UCS2__)
+        /*
+         * LOCALETYPE(*LOCALEUCS2) is specified.
+         * Wide-character strings are in UCS-2,
+         * narrow-character strings are in EBCDIC.
+         */
+#       define U_SIZEOF_WCHAR_T 2
+#   else
+        /*
+         * LOCALETYPE(*CLD) or LOCALETYPE(*LOCALE) is specified.
+         * Wide-character strings are in 16-bit EBCDIC,
+         * narrow-character strings are in EBCDIC.
+         */
+#       define U_SIZEOF_WCHAR_T 2
+#   endif
+#else
+#   define U_SIZEOF_WCHAR_T 4
+#endif
+
+#ifndef U_HAVE_WCSCPY
+#define U_HAVE_WCSCPY U_HAVE_WCHAR_H
+#endif
+
+/** @} */
+
+/**
+ * \def U_HAVE_CHAR16_T
+ * Defines whether the char16_t type is available for UTF-16
+ * and u"abc" UTF-16 string literals are supported.
+ * This is a new standard type and standard string literal syntax in C++11
+ * but has been available in some compilers before.
+ * @internal
+ */
+#ifdef U_HAVE_CHAR16_T
+    /* Use the predefined value. */
+#else
+    /*
+     * Notes:
+     * C++11 and C11 require support for UTF-16 literals
+     * Doesn't work on Mac C11 (see workaround in ptypes.h)
+     * or Cygwin less than 3.5.
+     */
+#   if defined(__cplusplus)
+#       define U_HAVE_CHAR16_T 1
+#   elif U_PLATFORM_IS_DARWIN_BASED || (U_PLATFORM == U_PF_CYGWIN && CYGWIN_VERSION_DLL_MAJOR < 3005)
+#       define U_HAVE_CHAR16_T 0
+#   else
+        // conformant C11
+#       define U_HAVE_CHAR16_T 1
+#   endif
+#endif
+
+/**
+ * @{
+ * \def U_DECLARE_UTF16
+ * Do not use this macro because it is not defined on all platforms.
+ * In C++, use std::u16string_view literals, see the UNICODE_STRING docs.
+ * In C, use u"UTF-16 literals".
+ * See also the public U_STRING_DECL macro.
+ * @internal
+ */
+#ifdef U_DECLARE_UTF16
+    /* Use the predefined value. */
+#elif U_HAVE_CHAR16_T \
+    || (defined(__xlC__) && defined(__IBM_UTF_LITERAL) && U_SIZEOF_WCHAR_T != 2) \
+    || (defined(__HP_aCC) && __HP_aCC >= 035000) \
+    || (defined(__HP_cc) && __HP_cc >= 111106) \
+    || (defined(U_IN_DOXYGEN))
+#   define U_DECLARE_UTF16(string) u ## string
+#elif U_SIZEOF_WCHAR_T == 2 \
+    && (U_CHARSET_FAMILY == 0 || (U_PF_OS390 <= U_PLATFORM && U_PLATFORM <= U_PF_OS400 && defined(__UCS2__)))
+#   define U_DECLARE_UTF16(string) L ## string
+#else
+    /* Leave U_DECLARE_UTF16 undefined. See unistr.h. */
+#endif
+
+/** @} */
+
+/*===========================================================================*/
+/** @{ Symbol import-export control                                          */
+/*===========================================================================*/
+
+#ifdef U_EXPORT
+    /* Use the predefined value. */
+#elif defined(U_STATIC_IMPLEMENTATION)
+#   define U_EXPORT
+#elif defined(_MSC_VER) || (UPRV_HAS_DECLSPEC_ATTRIBUTE(__dllexport__) && \
+                            UPRV_HAS_DECLSPEC_ATTRIBUTE(__dllimport__))
+#   define U_EXPORT __declspec(dllexport)
+#elif defined(__GNUC__) || defined(__open_xl__)
+#   define U_EXPORT __attribute__((visibility("default")))
+#elif (defined(__SUNPRO_CC) && __SUNPRO_CC >= 0x550) \
+   || (defined(__SUNPRO_C) && __SUNPRO_C >= 0x550) 
+#   define U_EXPORT __global
+/*#elif defined(__HP_aCC) || defined(__HP_cc)
+#   define U_EXPORT __declspec(dllexport)*/
+#else
+#   define U_EXPORT
+#endif
+
+/* U_CALLCONV is related to U_EXPORT2 */
+#ifdef U_EXPORT2
+    /* Use the predefined value. */
+#elif defined(_MSC_VER)
+#   define U_EXPORT2 __cdecl
+#else
+#   define U_EXPORT2
+#endif
+
+#ifdef U_IMPORT
+    /* Use the predefined value. */
+#elif defined(_MSC_VER) || (UPRV_HAS_DECLSPEC_ATTRIBUTE(__dllexport__) && \
+                            UPRV_HAS_DECLSPEC_ATTRIBUTE(__dllimport__))
+    /* Windows needs to export/import data. */
+#   define U_IMPORT __declspec(dllimport)
+#else
+#   define U_IMPORT 
+#endif
+
+/**
+ * \def U_HIDDEN
+ * This is used to mark internal structs declared within external classes,
+ * to prevent the internal structs from having the same visibility as the
+ * class within which they are declared. 
+ * @internal
+ */
+#ifdef U_HIDDEN
+    /* Use the predefined value. */
+#elif defined(__GNUC__) || defined(__open_xl__)
+#   define U_HIDDEN __attribute__((visibility("hidden")))
+#else
+#   define U_HIDDEN 
+#endif
+
+/**
+ * \def U_CALLCONV
+ * Similar to U_CDECL_BEGIN/U_CDECL_END, this qualifier is necessary
+ * in callback function typedefs to make sure that the calling convention
+ * is compatible.
+ *
+ * This is only used for non-ICU-API functions.
+ * When a function is a public ICU API,
+ * you must use the U_CAPI and U_EXPORT2 qualifiers.
+ *
+ * Please note, you need to use U_CALLCONV after the *.
+ *
+ * NO : "static const char U_CALLCONV *func( . . . )"
+ * YES: "static const char* U_CALLCONV func( . . . )"
+ *
+ * @stable ICU 2.0
+ */
+#if U_PLATFORM == U_PF_OS390 && defined(__cplusplus)
+#    define U_CALLCONV __cdecl
+#else
+#    define U_CALLCONV U_EXPORT2
+#endif
+
+/**
+ * \def U_CALLCONV_FPTR
+ * Similar to U_CALLCONV, but only used on function pointers.
+ * @internal
+ */
+#if U_PLATFORM == U_PF_OS390 && defined(__cplusplus)
+#    define U_CALLCONV_FPTR U_CALLCONV
+#else
+#    define U_CALLCONV_FPTR
+#endif
+/** @} */
+
+#endif  // _PLATFORM_H

thirdparty/icu4c/common/unicode/ptypes.h

@@ -0,0 +1,66 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+/*
+******************************************************************************
+*
+*   Copyright (C) 1997-2012, International Business Machines
+*   Corporation and others.  All Rights Reserved.
+*
+******************************************************************************
+*
+*  FILE NAME : ptypes.h
+*
+*   Date        Name        Description
+*   05/13/98    nos         Creation (content moved here from ptypes.h).
+*   03/02/99    stephen     Added AS400 support.
+*   03/30/99    stephen     Added Linux support.
+*   04/13/99    stephen     Reworked for autoconf.
+*   09/18/08    srl         Moved basic types back to ptypes.h from platform.h
+******************************************************************************
+*/
+
+/**
+ * \file
+ * \brief C API: Definitions of integer types of various widths
+ */
+
+#ifndef _PTYPES_H
+#define _PTYPES_H
+
+/**
+ * \def __STDC_LIMIT_MACROS
+ * According to the Linux stdint.h, the ISO C99 standard specifies that in C++ implementations
+ * macros like INT32_MIN and UINTPTR_MAX should only be defined if explicitly requested.
+ * We need to define __STDC_LIMIT_MACROS before including stdint.h in C++ code
+ * that uses such limit macros.
+ * @internal
+ */
+#ifndef __STDC_LIMIT_MACROS
+#define __STDC_LIMIT_MACROS
+#endif
+
+/* NULL, size_t, wchar_t */
+#include <stddef.h>
+
+/* More platform-specific definitions. */
+#include "unicode/platform.h"
+
+/*===========================================================================*/
+/* Generic data types                                                        */
+/*===========================================================================*/
+
+#include <stdint.h>
+
+// C++11 and C11 both specify that the data type char16_t should exist, C++11
+// as a keyword and C11 as a typedef in the uchar.h header file, but not all
+// implementations (looking at you, Apple, spring 2024) actually do this, so
+// ICU4C must detect and deal with that.
+#if !defined(__cplusplus) && !defined(U_IN_DOXYGEN)
+#   if U_HAVE_CHAR16_T
+#       include <uchar.h>
+#   else
+        typedef uint16_t char16_t;
+#   endif
+#endif
+
+#endif /* _PTYPES_H */

thirdparty/icu4c/common/unicode/putil.h

@@ -0,0 +1,183 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+/*
+******************************************************************************
+*
+*   Copyright (C) 1997-2014, International Business Machines
+*   Corporation and others.  All Rights Reserved.
+*
+******************************************************************************
+*
+*  FILE NAME : putil.h
+*
+*   Date        Name        Description
+*   05/14/98    nos         Creation (content moved here from utypes.h).
+*   06/17/99    erm         Added IEEE_754
+*   07/22/98    stephen     Added IEEEremainder, max, min, trunc
+*   08/13/98    stephen     Added isNegativeInfinity, isPositiveInfinity
+*   08/24/98    stephen     Added longBitsFromDouble
+*   03/02/99    stephen     Removed openFile().  Added AS400 support.
+*   04/15/99    stephen     Converted to C
+*   11/15/99    helena      Integrated S/390 changes for IEEE support.
+*   01/11/00    helena      Added u_getVersion.
+******************************************************************************
+*/
+
+#ifndef PUTIL_H
+#define PUTIL_H
+
+#include "unicode/utypes.h"
+ /**
+  * \file
+  * \brief C API: Platform Utilities
+  */
+
+/*==========================================================================*/
+/* Platform utilities                                                       */
+/*==========================================================================*/
+
+/**
+ * Platform utilities isolates the platform dependencies of the
+ * library.  For each platform which this code is ported to, these
+ * functions may have to be re-implemented.
+ */
+
+/**
+ * Return the ICU data directory. 
+ * The data directory is where common format ICU data files (.dat files)
+ *   are loaded from.  Note that normal use of the built-in ICU
+ *   facilities does not require loading of an external data file;
+ *   unless you are adding custom data to ICU, the data directory
+ *   does not need to be set.
+ *
+ * The data directory is determined as follows:
+ *    If u_setDataDirectory() has been called, that is it, otherwise
+ *    if the ICU_DATA environment variable is set, use that, otherwise
+ *    If a data directory was specified at ICU build time
+ *      <code>
+ * \code
+ *        #define ICU_DATA_DIR "path" 
+ * \endcode
+ * </code> use that,
+ *    otherwise no data directory is available.
+ *
+ * @return the data directory, or an empty string ("") if no data directory has
+ *         been specified.
+ *   
+ * @stable ICU 2.0
+ */
+U_CAPI const char* U_EXPORT2 u_getDataDirectory(void);
+
+
+/** 
+ * Set the ICU data directory. 
+ * The data directory is where common format ICU data files (.dat files)
+ *   are loaded from.  Note that normal use of the built-in ICU
+ *   facilities does not require loading of an external data file;
+ *   unless you are adding custom data to ICU, the data directory
+ *   does not need to be set.
+ *
+ * This function should be called at most once in a process, before the
+ * first ICU operation (e.g., u_init()) that will require the loading of an
+ * ICU data file.
+ * This function is not thread-safe. Use it before calling ICU APIs from
+ * multiple threads.
+ *
+ * @param directory The directory to be set.
+ *
+ * @see u_init
+ * @stable ICU 2.0
+ */
+U_CAPI void U_EXPORT2 u_setDataDirectory(const char *directory);
+
+#ifndef U_HIDE_INTERNAL_API
+/**
+  * Return the time zone files override directory, or an empty string if
+  * no directory was specified. Certain time zone resources will be preferentially
+  * loaded from individual files in this directory.
+  *
+  * @return the time zone data override directory.
+  * @internal
+  */ 
+U_CAPI const char * U_EXPORT2 u_getTimeZoneFilesDirectory(UErrorCode *status);
+
+/**
+  * Set the time zone files override directory.
+  * This function is not thread safe; it must not be called concurrently with
+  *   u_getTimeZoneFilesDirectory() or any other use of ICU time zone functions.
+  * This function should only be called before using any ICU service that
+  *   will access the time zone data.
+  * @internal
+  */
+U_CAPI void U_EXPORT2 u_setTimeZoneFilesDirectory(const char *path, UErrorCode *status);
+#endif  /* U_HIDE_INTERNAL_API */
+
+
+/**
+ * @{
+ * Filesystem file and path separator characters.
+ * Example: '/' and ':' on Unix, '\\' and ';' on Windows.
+ * @stable ICU 2.0
+ */
+#if U_PLATFORM_USES_ONLY_WIN32_API
+#   define U_FILE_SEP_CHAR '\\'
+#   define U_FILE_ALT_SEP_CHAR '/'
+#   define U_PATH_SEP_CHAR ';'
+#   define U_FILE_SEP_STRING "\\"
+#   define U_FILE_ALT_SEP_STRING "/"
+#   define U_PATH_SEP_STRING ";"
+#else
+#   define U_FILE_SEP_CHAR '/'
+#   define U_FILE_ALT_SEP_CHAR '/'
+#   define U_PATH_SEP_CHAR ':'
+#   define U_FILE_SEP_STRING "/"
+#   define U_FILE_ALT_SEP_STRING "/"
+#   define U_PATH_SEP_STRING ":"
+#endif
+
+/** @} */
+
+/**
+ * Convert char characters to UChar characters.
+ * This utility function is useful only for "invariant characters"
+ * that are encoded in the platform default encoding.
+ * They are a small, constant subset of the encoding and include
+ * just the latin letters, digits, and some punctuation.
+ * For details, see U_CHARSET_FAMILY.
+ *
+ * @param cs Input string, points to <code>length</code>
+ *           character bytes from a subset of the platform encoding.
+ * @param us Output string, points to memory for <code>length</code>
+ *           Unicode characters.
+ * @param length The number of characters to convert; this may
+ *               include the terminating <code>NUL</code>.
+ *
+ * @see U_CHARSET_FAMILY
+ * @stable ICU 2.0
+ */
+U_CAPI void U_EXPORT2
+u_charsToUChars(const char *cs, UChar *us, int32_t length);
+
+/**
+ * Convert UChar characters to char characters.
+ * This utility function is useful only for "invariant characters"
+ * that can be encoded in the platform default encoding.
+ * They are a small, constant subset of the encoding and include
+ * just the latin letters, digits, and some punctuation.
+ * For details, see U_CHARSET_FAMILY.
+ *
+ * @param us Input string, points to <code>length</code>
+ *           Unicode characters that can be encoded with the
+ *           codepage-invariant subset of the platform encoding.
+ * @param cs Output string, points to memory for <code>length</code>
+ *           character bytes.
+ * @param length The number of characters to convert; this may
+ *               include the terminating <code>NUL</code>.
+ *
+ * @see U_CHARSET_FAMILY
+ * @stable ICU 2.0
+ */
+U_CAPI void U_EXPORT2
+u_UCharsToChars(const UChar *us, char *cs, int32_t length);
+
+#endif

thirdparty/icu4c/common/unicode/rbbi.h

@@ -0,0 +1,823 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+/*
+***************************************************************************
+*   Copyright (C) 1999-2016 International Business Machines Corporation   *
+*   and others. All rights reserved.                                      *
+***************************************************************************
+
+**********************************************************************
+*   Date        Name        Description
+*   10/22/99    alan        Creation.
+*   11/11/99    rgillam     Complete port from Java.
+**********************************************************************
+*/
+
+#ifndef RBBI_H
+#define RBBI_H
+
+#include "unicode/utypes.h"
+
+#if U_SHOW_CPLUSPLUS_API
+
+/**
+ * \file
+ * \brief C++ API: Rule Based Break Iterator
+ */
+
+#if !UCONFIG_NO_BREAK_ITERATION
+
+#include "unicode/brkiter.h"
+#include "unicode/udata.h"
+#include "unicode/parseerr.h"
+#include "unicode/schriter.h"
+
+struct UCPTrie;
+
+U_NAMESPACE_BEGIN
+
+/** @internal */
+class  LanguageBreakEngine;
+struct RBBIDataHeader;
+class  RBBIDataWrapper;
+class  UnhandledEngine;
+class  UStack;
+
+
+#ifndef U_HIDE_INTERNAL_API
+/**
+ * The ExternalBreakEngine class define an abstract interface for the host environment
+ * to provide a low level facility to break text for unicode text in script that the text boundary
+ * cannot be handled by upper level rule based logic, for example, for Chinese and Japanese
+ * word breaking, Thai, Khmer, Burmese, Lao and other Southeast Asian scripts.
+ * The host environment implement one or more subclass of ExternalBreakEngine and
+ * register them in the initialization time by calling
+ * RuleBasedBreakIterator::registerExternalBreakEngine(). ICU adopt and own the engine and will
+ * delete the registered external engine in proper time during the clean up
+ * event.
+ * @internal ICU 74 technology preview
+ */
+class ExternalBreakEngine : public UObject {
+  public:
+    /**
+     * destructor
+     * @internal ICU 74 technology preview
+     */
+    virtual ~ExternalBreakEngine() {}
+
+    /**
+     * <p>Indicate whether this engine handles a particular character when
+     * the RuleBasedBreakIterator is used for a particular locale. This method is used
+     * by the RuleBasedBreakIterator to find a break engine.</p>
+     * @param c A character which begins a run that the engine might handle.
+     * @param locale    The locale.
+     * @return true if this engine handles the particular character for that locale.
+     * @internal ICU 74 technology preview
+     */
+    virtual bool isFor(UChar32 c, const char* locale) const = 0;
+
+    /**
+     * <p>Indicate whether this engine handles a particular character.This method is
+     * used by the RuleBasedBreakIterator after it already find a break engine to see which
+     * characters after the first one can be handled by this break engine.</p>
+     * @param c A character that the engine might handle.
+     * @return true if this engine handles the particular character.
+     * @internal ICU 74 technology preview
+     */
+    virtual bool handles(UChar32 c) const = 0;
+
+    /**
+     * <p>Divide up a range of text handled by this break engine.</p>
+     *
+     * @param text A UText representing the text
+     * @param start The start of the range of known characters
+     * @param end The end of the range of known characters
+     * @param foundBreaks Output of C array of int32_t break positions, or
+     * nullptr
+     * @param foundBreaksCapacity The capacity of foundBreaks
+     * @param status Information on any errors encountered.
+     * @return The number of breaks found
+     * @internal ICU 74 technology preview
+     */
+     virtual int32_t fillBreaks(UText* text,  int32_t start, int32_t end,
+                               int32_t* foundBreaks, int32_t foundBreaksCapacity,
+                               UErrorCode& status) const = 0;
+};
+#endif  /* U_HIDE_INTERNAL_API */
+
+
+/**
+ *
+ * A subclass of BreakIterator whose behavior is specified using a list of rules.
+ * <p>Instances of this class are most commonly created by the factory methods of
+ *  BreakIterator::createWordInstance(), BreakIterator::createLineInstance(), etc.,
+ *  and then used via the abstract API in class BreakIterator</p>
+ *
+ * <p>See the ICU User Guide for information on Break Iterator Rules.</p>
+ *
+ * <p>This class is not intended to be subclassed.</p>
+ */
+class U_COMMON_API RuleBasedBreakIterator /*final*/ : public BreakIterator {
+
+private:
+    /**
+     * The UText through which this BreakIterator accesses the text
+     * @internal (private)
+     */
+    UText  fText = UTEXT_INITIALIZER;
+
+#ifndef U_HIDE_INTERNAL_API
+public:
+#endif /* U_HIDE_INTERNAL_API */
+    /**
+     * The rule data for this BreakIterator instance.
+     * Not for general use; Public only for testing purposes.
+     * @internal
+     */
+    RBBIDataWrapper    *fData = nullptr;
+
+private:
+    /**
+      * The saved error code associated with this break iterator.
+      * This is the value to be returned by copyErrorTo().
+      */
+    UErrorCode      fErrorCode = U_ZERO_ERROR;
+
+    /**
+      * The current  position of the iterator. Pinned, 0 < fPosition <= text.length.
+      * Never has the value UBRK_DONE (-1).
+      */
+    int32_t         fPosition = 0;
+
+    /**
+      * TODO:
+      */
+    int32_t         fRuleStatusIndex = 0;
+
+    /**
+     *   Cache of previously determined boundary positions.
+     */
+    class BreakCache;
+    BreakCache         *fBreakCache = nullptr;
+
+    /**
+     *  Cache of boundary positions within a region of text that has been
+     *  sub-divided by dictionary based breaking.
+     */
+    class DictionaryCache;
+    DictionaryCache *fDictionaryCache = nullptr;
+
+    /**
+     *
+     * If present, UStack of LanguageBreakEngine objects that might handle
+     * dictionary characters. Searched from top to bottom to find an object to
+     * handle a given character.
+     * @internal (private)
+     */
+    UStack              *fLanguageBreakEngines = nullptr;
+
+    /**
+     *
+     * If present, the special LanguageBreakEngine used for handling
+     * characters that are in the dictionary set, but not handled by any
+     * LanguageBreakEngine.
+     * @internal (private)
+     */
+    UnhandledEngine     *fUnhandledBreakEngine = nullptr;
+
+    /**
+     * Counter for the number of characters encountered with the "dictionary"
+     *   flag set.
+     * @internal (private)
+     */
+    uint32_t            fDictionaryCharCount = 0;
+
+    /**
+     *   A character iterator that refers to the same text as the UText, above.
+     *   Only included for compatibility with old API, which was based on CharacterIterators.
+     *   Value may be adopted from outside, or one of fSCharIter or fDCharIter, below.
+     */
+    CharacterIterator  *fCharIter = &fSCharIter;
+
+    /**
+     *   When the input text is provided by a UnicodeString, this will point to
+     *    a characterIterator that wraps that data.  Needed only for the
+     *    implementation of getText(), a backwards compatibility issue.
+     */
+    UCharCharacterIterator fSCharIter {u"", 0};
+
+    /**
+      * True when iteration has run off the end, and iterator functions should return UBRK_DONE.
+      */
+    bool           fDone = false;
+
+    /**
+     *  Array of look-ahead tentative results.
+     */
+    int32_t *fLookAheadMatches = nullptr;
+
+    /**
+     *  A flag to indicate if phrase based breaking is enabled.
+     */
+    UBool fIsPhraseBreaking = false;
+
+    //=======================================================================
+    // constructors
+    //=======================================================================
+
+    /**
+     * Constructor from a flattened set of RBBI data in malloced memory.
+     *             RulesBasedBreakIterators built from a custom set of rules
+     *             are created via this constructor; the rules are compiled
+     *             into memory, then the break iterator is constructed here.
+     *
+     *             The break iterator adopts the memory, and will
+     *             free it when done.
+     * @internal (private)
+     */
+    RuleBasedBreakIterator(RBBIDataHeader* data, UErrorCode &status);
+
+    /**
+     * This constructor uses the udata interface to create a BreakIterator
+     * whose internal tables live in a memory-mapped file.  "image" is an
+     * ICU UDataMemory handle for the pre-compiled break iterator tables.
+     * @param image handle to the memory image for the break iterator data.
+     *        Ownership of the UDataMemory handle passes to the Break Iterator,
+     *        which will be responsible for closing it when it is no longer needed.
+     * @param status Information on any errors encountered.
+     * @param isPhraseBreaking true if phrase based breaking is required, otherwise false.
+     * @see udata_open
+     * @see #getBinaryRules
+     * @internal (private)
+     */
+    RuleBasedBreakIterator(UDataMemory* image, UBool isPhraseBreaking, UErrorCode &status);
+
+    /** @internal */
+    friend class RBBIRuleBuilder;
+    /** @internal */
+    friend class BreakIterator;
+
+    /**
+     * Default constructor with an error code parameter.
+     * Aside from error handling, otherwise identical to the default constructor.
+     * Internally, handles common initialization for other constructors.
+     * @internal (private)
+     */
+    RuleBasedBreakIterator(UErrorCode *status);
+
+public:
+
+    /** Default constructor.  Creates an empty shell of an iterator, with no
+     *  rules or text to iterate over.   Object can subsequently be assigned to,
+     *  but is otherwise unusable.
+     *  @stable ICU 2.2
+     */
+    RuleBasedBreakIterator();
+
+    /**
+     * Copy constructor.  Will produce a break iterator with the same behavior,
+     * and which iterates over the same text, as the one passed in.
+     * @param that The RuleBasedBreakIterator passed to be copied
+     * @stable ICU 2.0
+     */
+    RuleBasedBreakIterator(const RuleBasedBreakIterator& that);
+
+    /**
+     * Construct a RuleBasedBreakIterator from a set of rules supplied as a string.
+     * @param rules The break rules to be used.
+     * @param parseError  In the event of a syntax error in the rules, provides the location
+     *                    within the rules of the problem.
+     * @param status Information on any errors encountered.
+     * @stable ICU 2.2
+     */
+    RuleBasedBreakIterator( const UnicodeString    &rules,
+                             UParseError           &parseError,
+                             UErrorCode            &status);
+
+    /**
+     * Construct a RuleBasedBreakIterator from a set of precompiled binary rules.
+     * Binary rules are obtained from RulesBasedBreakIterator::getBinaryRules().
+     * Construction of a break iterator in this way is substantially faster than
+     * construction from source rules.
+     *
+     * Ownership of the storage containing the compiled rules remains with the
+     * caller of this function.  The compiled rules must not be  modified or
+     * deleted during the life of the break iterator.
+     *
+     * The compiled rules are not compatible across different major versions of ICU.
+     * The compiled rules are compatible only between machines with the same
+     * byte ordering (little or big endian) and the same base character set family
+     * (ASCII or EBCDIC).
+     *
+     * @see #getBinaryRules
+     * @param compiledRules A pointer to the compiled break rules to be used.
+     * @param ruleLength The length of the compiled break rules, in bytes.  This
+     *   corresponds to the length value produced by getBinaryRules().
+     * @param status Information on any errors encountered, including invalid
+     *   binary rules.
+     * @stable ICU 4.8
+     */
+    RuleBasedBreakIterator(const uint8_t *compiledRules,
+                           uint32_t       ruleLength,
+                           UErrorCode    &status);
+
+    /**
+     * This constructor uses the udata interface to create a BreakIterator
+     * whose internal tables live in a memory-mapped file.  "image" is an
+     * ICU UDataMemory handle for the pre-compiled break iterator tables.
+     * @param image handle to the memory image for the break iterator data.
+     *        Ownership of the UDataMemory handle passes to the Break Iterator,
+     *        which will be responsible for closing it when it is no longer needed.
+     * @param status Information on any errors encountered.
+     * @see udata_open
+     * @see #getBinaryRules
+     * @stable ICU 2.8
+     */
+    RuleBasedBreakIterator(UDataMemory* image, UErrorCode &status);
+
+    /**
+     * Destructor
+     *  @stable ICU 2.0
+     */
+    virtual ~RuleBasedBreakIterator();
+
+    /**
+     * Assignment operator.  Sets this iterator to have the same behavior,
+     * and iterate over the same text, as the one passed in.
+     * @param that The RuleBasedBreakItertor passed in
+     * @return the newly created RuleBasedBreakIterator
+     *  @stable ICU 2.0
+     */
+    RuleBasedBreakIterator& operator=(const RuleBasedBreakIterator& that);
+
+    /**
+     * Equality operator.  Returns true if both BreakIterators are of the
+     * same class, have the same behavior, and iterate over the same text.
+     * @param that The BreakIterator to be compared for equality
+     * @return true if both BreakIterators are of the
+     * same class, have the same behavior, and iterate over the same text.
+     *  @stable ICU 2.0
+     */
+    virtual bool operator==(const BreakIterator& that) const override;
+
+    /**
+     * Not-equal operator.  If operator== returns true, this returns false,
+     * and vice versa.
+     * @param that The BreakIterator to be compared for inequality
+     * @return true if both BreakIterators are not same.
+     *  @stable ICU 2.0
+     */
+    inline bool operator!=(const BreakIterator& that) const {
+        return !operator==(that);
+    }
+
+    /**
+     * Returns a newly-constructed RuleBasedBreakIterator with the same
+     * behavior, and iterating over the same text, as this one.
+     * Differs from the copy constructor in that it is polymorphic, and
+     * will correctly clone (copy) a derived class.
+     * clone() is thread safe.  Multiple threads may simultaneously
+     * clone the same source break iterator.
+     * @return a newly-constructed RuleBasedBreakIterator
+     * @stable ICU 2.0
+     */
+    virtual RuleBasedBreakIterator* clone() const override;
+
+    /**
+     * Compute a hash code for this BreakIterator
+     * @return A hash code
+     *  @stable ICU 2.0
+     */
+    virtual int32_t hashCode() const;
+
+    /**
+     * Returns the description used to create this iterator
+     * @return the description used to create this iterator
+     *  @stable ICU 2.0
+     */
+    virtual const UnicodeString& getRules() const;
+
+    //=======================================================================
+    // BreakIterator overrides
+    //=======================================================================
+
+    /**
+     * <p>
+     * Return a CharacterIterator over the text being analyzed.
+     * The returned character iterator is owned by the break iterator, and must
+     * not be deleted by the caller.  Repeated calls to this function may
+     * return the same CharacterIterator.
+     * </p>
+     * <p>
+     * The returned character iterator must not be used concurrently with
+     * the break iterator.  If concurrent operation is needed, clone the
+     * returned character iterator first and operate on the clone.
+     * </p>
+     * <p>
+     * When the break iterator is operating on text supplied via a UText,
+     * this function will fail, returning a CharacterIterator containing no text.
+     * The function getUText() provides similar functionality,
+     * is reliable, and is more efficient.
+     * </p>
+     *
+     * TODO:  deprecate this function?
+     *
+     * @return An iterator over the text being analyzed.
+     * @stable ICU 2.0
+     */
+    virtual CharacterIterator& getText() const override;
+
+    /**
+      *  Get a UText for the text being analyzed.
+      *  The returned UText is a shallow clone of the UText used internally
+      *  by the break iterator implementation.  It can safely be used to
+      *  access the text without impacting any break iterator operations,
+      *  but the underlying text itself must not be altered.
+      *
+      * @param fillIn A UText to be filled in.  If nullptr, a new UText will be
+      *           allocated to hold the result.
+      * @param status receives any error codes.
+      * @return   The current UText for this break iterator.  If an input
+      *           UText was provided, it will always be returned.
+      * @stable ICU 3.4
+      */
+     virtual UText *getUText(UText *fillIn, UErrorCode &status) const override;
+
+    /**
+     * Set the iterator to analyze a new piece of text.  This function resets
+     * the current iteration position to the beginning of the text.
+     * @param newText An iterator over the text to analyze.  The BreakIterator
+     * takes ownership of the character iterator.  The caller MUST NOT delete it!
+     *  @stable ICU 2.0
+     */
+    virtual void adoptText(CharacterIterator* newText) override;
+
+    /**
+     * Set the iterator to analyze a new piece of text.  This function resets
+     * the current iteration position to the beginning of the text.
+     *
+     * The BreakIterator will retain a reference to the supplied string.
+     * The caller must not modify or delete the text while the BreakIterator
+     * retains the reference.
+     *
+     * @param newText The text to analyze.
+     *  @stable ICU 2.0
+     */
+    virtual void setText(const UnicodeString& newText) override;
+
+    /**
+     * Reset the break iterator to operate over the text represented by
+     * the UText.  The iterator position is reset to the start.
+     *
+     * This function makes a shallow clone of the supplied UText.  This means
+     * that the caller is free to immediately close or otherwise reuse the
+     * Utext that was passed as a parameter, but that the underlying text itself
+     * must not be altered while being referenced by the break iterator.
+     *
+     * @param text    The UText used to change the text.
+     * @param status  Receives any error codes.
+     * @stable ICU 3.4
+     */
+    virtual void  setText(UText *text, UErrorCode &status) override;
+
+    /**
+     * Sets the current iteration position to the beginning of the text, position zero.
+     * @return The offset of the beginning of the text, zero.
+     *  @stable ICU 2.0
+     */
+    virtual int32_t first() override;
+
+    /**
+     * Sets the current iteration position to the end of the text.
+     * @return The text's past-the-end offset.
+     *  @stable ICU 2.0
+     */
+    virtual int32_t last() override;
+
+    /**
+     * Advances the iterator either forward or backward the specified number of steps.
+     * Negative values move backward, and positive values move forward.  This is
+     * equivalent to repeatedly calling next() or previous().
+     * @param n The number of steps to move.  The sign indicates the direction
+     * (negative is backwards, and positive is forwards).
+     * @return The character offset of the boundary position n boundaries away from
+     * the current one.
+     *  @stable ICU 2.0
+     */
+    virtual int32_t next(int32_t n) override;
+
+    /**
+     * Advances the iterator to the next boundary position.
+     * @return The position of the first boundary after this one.
+     *  @stable ICU 2.0
+     */
+    virtual int32_t next() override;
+
+    /**
+     * Moves the iterator backwards, to the last boundary preceding this one.
+     * @return The position of the last boundary position preceding this one.
+     *  @stable ICU 2.0
+     */
+    virtual int32_t previous() override;
+
+    /**
+     * Sets the iterator to refer to the first boundary position following
+     * the specified position.
+     * @param offset The position from which to begin searching for a break position.
+     * @return The position of the first break after the current position.
+     *  @stable ICU 2.0
+     */
+    virtual int32_t following(int32_t offset) override;
+
+    /**
+     * Sets the iterator to refer to the last boundary position before the
+     * specified position.
+     * @param offset The position to begin searching for a break from.
+     * @return The position of the last boundary before the starting position.
+     *  @stable ICU 2.0
+     */
+    virtual int32_t preceding(int32_t offset) override;
+
+    /**
+     * Returns true if the specified position is a boundary position.  As a side
+     * effect, leaves the iterator pointing to the first boundary position at
+     * or after "offset".
+     * @param offset the offset to check.
+     * @return True if "offset" is a boundary position.
+     *  @stable ICU 2.0
+     */
+    virtual UBool isBoundary(int32_t offset) override;
+
+    /**
+     * Returns the current iteration position. Note that UBRK_DONE is never
+     * returned from this function; if iteration has run to the end of a
+     * string, current() will return the length of the string while
+     * next() will return UBRK_DONE).
+     * @return The current iteration position.
+     * @stable ICU 2.0
+     */
+    virtual int32_t current() const override;
+
+    /**
+     * Return the status tag from the break rule that determined the boundary at
+     * the current iteration position.  For break rules that do not specify a
+     * status, a default value of 0 is returned.  If more than one break rule
+     * would cause a boundary to be located at some position in the text,
+     * the numerically largest of the applicable status values is returned.
+     * <p>
+     * Of the standard types of ICU break iterators, only word break and
+     * line break provide status values.  The values are defined in
+     * the header file ubrk.h.  For Word breaks, the status allows distinguishing between words
+     * that contain alphabetic letters, "words" that appear to be numbers,
+     * punctuation and spaces, words containing ideographic characters, and
+     * more.  For Line Break, the status distinguishes between hard (mandatory) breaks
+     * and soft (potential) break positions.
+     * <p>
+     * <code>getRuleStatus()</code> can be called after obtaining a boundary
+     * position from <code>next()</code>, <code>previous()</code>, or
+     * any other break iterator functions that returns a boundary position.
+     * <p>
+     * Note that <code>getRuleStatus()</code> returns the value corresponding to
+     * <code>current()</code> index even after <code>next()</code> has returned DONE.
+     * <p>
+     * When creating custom break rules, one is free to define whatever
+     * status values may be convenient for the application.
+     * <p>
+     * @return the status from the break rule that determined the boundary
+     * at the current iteration position.
+     *
+     * @see UWordBreak
+     * @stable ICU 2.2
+     */
+    virtual int32_t getRuleStatus() const override;
+
+   /**
+    * Get the status (tag) values from the break rule(s) that determined the boundary
+    * at the current iteration position.
+    * <p>
+    * The returned status value(s) are stored into an array provided by the caller.
+    * The values are stored in sorted (ascending) order.
+    * If the capacity of the output array is insufficient to hold the data,
+    *  the output will be truncated to the available length, and a
+    *  U_BUFFER_OVERFLOW_ERROR will be signaled.
+    *
+    * @param fillInVec an array to be filled in with the status values.
+    * @param capacity  the length of the supplied vector.  A length of zero causes
+    *                  the function to return the number of status values, in the
+    *                  normal way, without attempting to store any values.
+    * @param status    receives error codes.
+    * @return          The number of rule status values from the rules that determined
+    *                  the boundary at the current iteration position.
+    *                  In the event of a U_BUFFER_OVERFLOW_ERROR, the return value
+    *                  is the total number of status values that were available,
+    *                  not the reduced number that were actually returned.
+    * @see getRuleStatus
+    * @stable ICU 3.0
+    */
+    virtual int32_t getRuleStatusVec(int32_t *fillInVec, int32_t capacity, UErrorCode &status) override;
+
+    /**
+     * Returns a unique class ID POLYMORPHICALLY.  Pure virtual override.
+     * This method is to implement a simple version of RTTI, since not all
+     * C++ compilers support genuine RTTI.  Polymorphic operator==() and
+     * clone() methods call this method.
+     *
+     * @return          The class ID for this object. All objects of a
+     *                  given class have the same class ID.  Objects of
+     *                  other classes have different class IDs.
+     * @stable ICU 2.0
+     */
+    virtual UClassID getDynamicClassID() const override;
+
+    /**
+     * Returns the class ID for this class.  This is useful only for
+     * comparing to a return value from getDynamicClassID().  For example:
+     *
+     *      Base* polymorphic_pointer = createPolymorphicObject();
+     *      if (polymorphic_pointer->getDynamicClassID() ==
+     *          Derived::getStaticClassID()) ...
+     *
+     * @return          The class ID for all objects of this class.
+     * @stable ICU 2.0
+     */
+    static UClassID U_EXPORT2 getStaticClassID();
+
+#ifndef U_FORCE_HIDE_DEPRECATED_API
+    /**
+     * Deprecated functionality. Use clone() instead.
+     *
+     * Create a clone (copy) of this break iterator in memory provided
+     *  by the caller.  The idea is to increase performance by avoiding
+     *  a storage allocation.  Use of this function is NOT RECOMMENDED.
+     *  Performance gains are minimal, and correct buffer management is
+     *  tricky.  Use clone() instead.
+     *
+     * @param stackBuffer  The pointer to the memory into which the cloned object
+     *                     should be placed.  If nullptr,  allocate heap memory
+     *                     for the cloned object.
+     * @param BufferSize   The size of the buffer.  If zero, return the required
+     *                     buffer size, but do not clone the object.  If the
+     *                     size was too small (but not zero), allocate heap
+     *                     storage for the cloned object.
+     *
+     * @param status       Error status.  U_SAFECLONE_ALLOCATED_WARNING will be
+     *                     returned if the provided buffer was too small, and
+     *                     the clone was therefore put on the heap.
+     *
+     * @return  Pointer to the clone object.  This may differ from the stackBuffer
+     *          address if the byte alignment of the stack buffer was not suitable
+     *          or if the stackBuffer was too small to hold the clone.
+     * @deprecated ICU 52. Use clone() instead.
+     */
+    virtual RuleBasedBreakIterator *createBufferClone(void *stackBuffer,
+                                                      int32_t &BufferSize,
+                                                      UErrorCode &status) override;
+#endif  // U_FORCE_HIDE_DEPRECATED_API
+
+    /**
+     * Return the binary form of compiled break rules,
+     * which can then be used to create a new break iterator at some
+     * time in the future.  Creating a break iterator from pre-compiled rules
+     * is much faster than building one from the source form of the
+     * break rules.
+     *
+     * The binary data can only be used with the same version of ICU
+     *  and on the same platform type (processor endian-ness)
+     *
+     * @param length Returns the length of the binary data.  (Out parameter.)
+     *
+     * @return   A pointer to the binary (compiled) rule data.  The storage
+     *           belongs to the RulesBasedBreakIterator object, not the
+     *           caller, and must not be modified or deleted.
+     * @stable ICU 4.8
+     */
+    virtual const uint8_t *getBinaryRules(uint32_t &length);
+
+    /**
+     *  Set the subject text string upon which the break iterator is operating
+     *  without changing any other aspect of the matching state.
+     *  The new and previous text strings must have the same content.
+     *
+     *  This function is intended for use in environments where ICU is operating on
+     *  strings that may move around in memory.  It provides a mechanism for notifying
+     *  ICU that the string has been relocated, and providing a new UText to access the
+     *  string in its new position.
+     *
+     *  Note that the break iterator implementation never copies the underlying text
+     *  of a string being processed, but always operates directly on the original text
+     *  provided by the user. Refreshing simply drops the references to the old text
+     *  and replaces them with references to the new.
+     *
+     *  Caution:  this function is normally used only by very specialized,
+     *  system-level code.  One example use case is with garbage collection that moves
+     *  the text in memory.
+     *
+     * @param input      The new (moved) text string.
+     * @param status     Receives errors detected by this function.
+     * @return           *this
+     *
+     * @stable ICU 49
+     */
+    virtual RuleBasedBreakIterator &refreshInputText(UText *input, UErrorCode &status) override;
+
+
+private:
+    //=======================================================================
+    // implementation
+    //=======================================================================
+    /**
+     * Iterate backwards from an arbitrary position in the input text using the
+     * synthesized Safe Reverse rules.
+     * This locates a "Safe Position" from which the forward break rules
+     * will operate correctly. A Safe Position is not necessarily a boundary itself.
+     *
+     * @param fromPosition the position in the input text to begin the iteration.
+     * @internal (private)
+     */
+    int32_t handleSafePrevious(int32_t fromPosition);
+
+    /**
+     * Find a rule-based boundary by running the state machine.
+     * Input
+     *    fPosition, the position in the text to begin from.
+     * Output
+     *    fPosition:           the boundary following the starting position.
+     *    fDictionaryCharCount the number of dictionary characters encountered.
+     *                         If > 0, the segment will be further subdivided
+     *    fRuleStatusIndex     Info from the state table indicating which rules caused the boundary.
+     *
+     * @internal (private)
+     */
+    int32_t handleNext();
+
+    /*
+     * Templatized version of handleNext() and handleSafePrevious().
+     *
+     * There will be exactly four instantiations, two each for 8 and 16 bit tables,
+     * two each for 8 and 16 bit trie.
+     * Having separate instantiations for the table types keeps conditional tests of
+     * the table type out of the inner loops, at the expense of replicated code.
+     *
+     * The template parameter for the Trie access function is a value, not a type.
+     * Doing it this way, the compiler will inline the Trie function in the
+     * expanded functions. (Both the 8 and 16 bit access functions have the same type
+     * signature)
+     */
+
+    typedef uint16_t (*PTrieFunc)(const UCPTrie *, UChar32);
+
+    template<typename RowType, PTrieFunc trieFunc>
+    int32_t handleSafePrevious(int32_t fromPosition);
+
+    template<typename RowType, PTrieFunc trieFunc>
+    int32_t handleNext();
+
+
+    /**
+     * This function returns the appropriate LanguageBreakEngine for a
+     * given character c.
+     * @param c         A character in the dictionary set
+     * @param locale    The locale.
+     * @internal (private)
+     */
+    const LanguageBreakEngine *getLanguageBreakEngine(UChar32 c, const char* locale);
+
+  public:
+#ifndef U_HIDE_INTERNAL_API
+    /**
+     *   Debugging function only.
+     *   @internal
+     */
+     void dumpCache();
+
+    /**
+     * Debugging function only.
+     * @internal
+     */
+    void dumpTables();
+#endif  /* U_HIDE_INTERNAL_API */
+
+#ifndef U_HIDE_INTERNAL_API
+    /**
+     * Register a new external break engine. The external break engine will be adopted.
+     * Because ICU may choose to cache break engine internally, this must
+     * be called at application startup, prior to any calls to
+     * object methods of RuleBasedBreakIterator to avoid undefined behavior.
+     * @param toAdopt the ExternalBreakEngine instance to be adopted
+     * @param status the in/out status code, no special meanings are assigned
+     * @internal ICU 74 technology preview
+     */
+    static void U_EXPORT2 registerExternalBreakEngine(
+                  ExternalBreakEngine* toAdopt, UErrorCode& status);
+#endif  /* U_HIDE_INTERNAL_API */
+
+};
+
+
+U_NAMESPACE_END
+
+#endif /* #if !UCONFIG_NO_BREAK_ITERATION */
+
+#endif /* U_SHOW_CPLUSPLUS_API */
+
+#endif

thirdparty/icu4c/common/unicode/rep.h

@@ -0,0 +1,266 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+/*
+**************************************************************************
+* Copyright (C) 1999-2012, International Business Machines Corporation and
+* others. All Rights Reserved.
+**************************************************************************
+*   Date        Name        Description
+*   11/17/99    aliu        Creation.  Ported from java.  Modified to
+*                           match current UnicodeString API.  Forced
+*                           to use name "handleReplaceBetween" because
+*                           of existing methods in UnicodeString.
+**************************************************************************
+*/
+
+#ifndef REP_H
+#define REP_H
+
+#include "unicode/utypes.h"
+
+#if U_SHOW_CPLUSPLUS_API
+
+#include "unicode/uobject.h"
+
+/**
+ * \file 
+ * \brief C++ API: Replaceable String
+ */
+ 
+U_NAMESPACE_BEGIN
+
+class UnicodeString;
+
+/**
+ * <code>Replaceable</code> is an abstract base class representing a
+ * string of characters that supports the replacement of a range of
+ * itself with a new string of characters.  It is used by APIs that
+ * change a piece of text while retaining metadata.  Metadata is data
+ * other than the Unicode characters returned by char32At().  One
+ * example of metadata is style attributes; another is an edit
+ * history, marking each character with an author and revision number.
+ *
+ * <p>An implicit aspect of the <code>Replaceable</code> API is that
+ * during a replace operation, new characters take on the metadata of
+ * the old characters.  For example, if the string "the <b>bold</b>
+ * font" has range (4, 8) replaced with "strong", then it becomes "the
+ * <b>strong</b> font".
+ *
+ * <p><code>Replaceable</code> specifies ranges using a start
+ * offset and a limit offset.  The range of characters thus specified
+ * includes the characters at offset start..limit-1.  That is, the
+ * start offset is inclusive, and the limit offset is exclusive.
+ *
+ * <p><code>Replaceable</code> also includes API to access characters
+ * in the string: <code>length()</code>, <code>charAt()</code>,
+ * <code>char32At()</code>, and <code>extractBetween()</code>.
+ *
+ * <p>For a subclass to support metadata, typical behavior of
+ * <code>replace()</code> is the following:
+ * <ul>
+ *   <li>Set the metadata of the new text to the metadata of the first
+ *   character replaced</li>
+ *   <li>If no characters are replaced, use the metadata of the
+ *   previous character</li>
+ *   <li>If there is no previous character (i.e. start == 0), use the
+ *   following character</li>
+ *   <li>If there is no following character (i.e. the replaceable was
+ *   empty), use default metadata.<br>
+ *   <li>If the code point U+FFFF is seen, it should be interpreted as
+ *   a special marker having no metadata<li>
+ *   </li>
+ * </ul>
+ * If this is not the behavior, the subclass should document any differences.
+ * @author Alan Liu
+ * @stable ICU 2.0
+ */
+class U_COMMON_API Replaceable : public UObject {
+
+public:
+    /**
+     * Destructor.
+     * @stable ICU 2.0
+     */
+    virtual ~Replaceable();
+
+    /**
+     * Returns the number of 16-bit code units in the text.
+     * @return number of 16-bit code units in text
+     * @stable ICU 1.8
+     */ 
+    inline int32_t length() const;
+
+    /**
+     * Returns the 16-bit code unit at the given offset into the text.
+     * @param offset an integer between 0 and <code>length()</code>-1
+     * inclusive
+     * @return 16-bit code unit of text at given offset
+     * @stable ICU 1.8
+     */
+    inline char16_t charAt(int32_t offset) const;
+
+    /**
+     * Returns the 32-bit code point at the given 16-bit offset into
+     * the text.  This assumes the text is stored as 16-bit code units
+     * with surrogate pairs intermixed.  If the offset of a leading or
+     * trailing code unit of a surrogate pair is given, return the
+     * code point of the surrogate pair.
+     *
+     * @param offset an integer between 0 and <code>length()</code>-1
+     * inclusive
+     * @return 32-bit code point of text at given offset
+     * @stable ICU 1.8
+     */
+    inline UChar32 char32At(int32_t offset) const;
+
+    /**
+     * Copies characters in the range [<tt>start</tt>, <tt>limit</tt>) 
+     * into the UnicodeString <tt>target</tt>.
+     * @param start offset of first character which will be copied
+     * @param limit offset immediately following the last character to
+     * be copied
+     * @param target UnicodeString into which to copy characters.
+     * @return A reference to <TT>target</TT>
+     * @stable ICU 2.1
+     */
+    virtual void extractBetween(int32_t start,
+                                int32_t limit,
+                                UnicodeString& target) const = 0;
+
+    /**
+     * Replaces a substring of this object with the given text.  If the
+     * characters being replaced have metadata, the new characters
+     * that replace them should be given the same metadata.
+     *
+     * <p>Subclasses must ensure that if the text between start and
+     * limit is equal to the replacement text, that replace has no
+     * effect. That is, any metadata
+     * should be unaffected. In addition, subclasses are encouraged to
+     * check for initial and trailing identical characters, and make a
+     * smaller replacement if possible. This will preserve as much
+     * metadata as possible.
+     * @param start the beginning index, inclusive; <code>0 <= start
+     * <= limit</code>.
+     * @param limit the ending index, exclusive; <code>start <= limit
+     * <= length()</code>.
+     * @param text the text to replace characters <code>start</code>
+     * to <code>limit - 1</code> 
+     * @stable ICU 2.0
+     */
+    virtual void handleReplaceBetween(int32_t start,
+                                      int32_t limit,
+                                      const UnicodeString& text) = 0;
+    // Note: All other methods in this class take the names of
+    // existing UnicodeString methods.  This method is the exception.
+    // It is named differently because all replace methods of
+    // UnicodeString return a UnicodeString&.  The 'between' is
+    // required in order to conform to the UnicodeString naming
+    // convention; API taking start/length are named <operation>, and
+    // those taking start/limit are named <operationBetween>.  The
+    // 'handle' is added because 'replaceBetween' and
+    // 'doReplaceBetween' are already taken.
+
+    /**
+     * Copies a substring of this object, retaining metadata.
+     * This method is used to duplicate or reorder substrings.
+     * The destination index must not overlap the source range.
+     * 
+     * @param start the beginning index, inclusive; <code>0 <= start <=
+     * limit</code>.
+     * @param limit the ending index, exclusive; <code>start <= limit <=
+     * length()</code>.
+     * @param dest the destination index.  The characters from
+     * <code>start..limit-1</code> will be copied to <code>dest</code>.
+     * Implementations of this method may assume that <code>dest <= start ||
+     * dest >= limit</code>.
+     * @stable ICU 2.0
+     */
+    virtual void copy(int32_t start, int32_t limit, int32_t dest) = 0;
+
+    /**
+     * Returns true if this object contains metadata.  If a
+     * Replaceable object has metadata, calls to the Replaceable API
+     * must be made so as to preserve metadata.  If it does not, calls
+     * to the Replaceable API may be optimized to improve performance.
+     * The default implementation returns true.
+     * @return true if this object contains metadata
+     * @stable ICU 2.2
+     */
+    virtual UBool hasMetaData() const;
+
+    /**
+     * Clone this object, an instance of a subclass of Replaceable.
+     * Clones can be used concurrently in multiple threads.
+     * If a subclass does not implement clone(), or if an error occurs,
+     * then nullptr is returned.
+     * The caller must delete the clone.
+     *
+     * @return a clone of this object
+     *
+     * @see getDynamicClassID
+     * @stable ICU 2.6
+     */
+    virtual Replaceable *clone() const;
+
+protected:
+
+    /**
+     * Default constructor.
+     * @stable ICU 2.4
+     */
+    inline Replaceable();
+
+    /*
+     * Assignment operator not declared. The compiler will provide one
+     * which does nothing since this class does not contain any data members.
+     * API/code coverage may show the assignment operator as present and
+     * untested - ignore.
+     * Subclasses need this assignment operator if they use compiler-provided
+     * assignment operators of their own. An alternative to not declaring one
+     * here would be to declare and empty-implement a protected or public one.
+    Replaceable &Replaceable::operator=(const Replaceable &);
+     */
+
+    /**
+     * Virtual version of length().
+     * @stable ICU 2.4
+     */ 
+    virtual int32_t getLength() const = 0;
+
+    /**
+     * Virtual version of charAt().
+     * @stable ICU 2.4
+     */
+    virtual char16_t getCharAt(int32_t offset) const = 0;
+
+    /**
+     * Virtual version of char32At().
+     * @stable ICU 2.4
+     */
+    virtual UChar32 getChar32At(int32_t offset) const = 0;
+};
+
+inline Replaceable::Replaceable() {}
+
+inline int32_t
+Replaceable::length() const {
+    return getLength();
+}
+
+inline char16_t
+Replaceable::charAt(int32_t offset) const {
+    return getCharAt(offset);
+}
+
+inline UChar32
+Replaceable::char32At(int32_t offset) const {
+    return getChar32At(offset);
+}
+
+// There is no rep.cpp, see unistr.cpp for Replaceable function implementations.
+
+U_NAMESPACE_END
+
+#endif /* U_SHOW_CPLUSPLUS_API */
+
+#endif

thirdparty/icu4c/common/unicode/resbund.h

@@ -0,0 +1,489 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+/*
+******************************************************************************
+*
+*   Copyright (C) 1996-2013, International Business Machines Corporation
+*   and others.  All Rights Reserved.
+*
+******************************************************************************
+*
+* File resbund.h
+*
+*   CREATED BY
+*       Richard Gillam
+*
+* Modification History:
+*
+*   Date        Name        Description
+*   2/5/97      aliu        Added scanForLocaleInFile.  Added
+*                           constructor which attempts to read resource bundle
+*                           from a specific file, without searching other files.
+*   2/11/97     aliu        Added UErrorCode return values to constructors.  Fixed
+*                           infinite loops in scanForFile and scanForLocale.
+*                           Modified getRawResourceData to not delete storage
+*                           in localeData and resourceData which it doesn't own.
+*                           Added Mac compatibility #ifdefs for tellp() and
+*                           ios::nocreate.
+*   2/18/97     helena      Updated with 100% documentation coverage.
+*   3/13/97     aliu        Rewrote to load in entire resource bundle and store
+*                           it as a Hashtable of ResourceBundleData objects.
+*                           Added state table to govern parsing of files.
+*                           Modified to load locale index out of new file
+*                           distinct from default.txt.
+*   3/25/97     aliu        Modified to support 2-d arrays, needed for timezone
+*                           data. Added support for custom file suffixes.  Again,
+*                           needed to support timezone data.
+*   4/7/97      aliu        Cleaned up.
+* 03/02/99      stephen     Removed dependency on FILE*.
+* 03/29/99      helena      Merged Bertrand and Stephen's changes.
+* 06/11/99      stephen     Removed parsing of .txt files.
+*                           Reworked to use new binary format.
+*                           Cleaned up.
+* 06/14/99      stephen     Removed methods taking a filename suffix.
+* 11/09/99      weiv        Added getLocale(), fRealLocale, removed fRealLocaleID
+******************************************************************************
+*/
+
+#ifndef RESBUND_H
+#define RESBUND_H
+
+#include "unicode/utypes.h"
+
+#if U_SHOW_CPLUSPLUS_API
+
+#include "unicode/uobject.h"
+#include "unicode/ures.h"
+#include "unicode/unistr.h"
+#include "unicode/locid.h"
+
+/**
+ * \file 
+ * \brief C++ API: Resource Bundle
+ */
+ 
+U_NAMESPACE_BEGIN
+
+/**
+ * A class representing a collection of resource information pertaining to a given
+ * locale. A resource bundle provides a way of accessing locale- specific information in
+ * a data file. You create a resource bundle that manages the resources for a given
+ * locale and then ask it for individual resources.
+ * <P>
+ * Resource bundles in ICU4C are currently defined using text files which conform to the following
+ * <a href="https://github.com/unicode-org/icu-docs/blob/main/design/bnf_rb.txt">BNF definition</a>.
+ * More on resource bundle concepts and syntax can be found in the
+ * <a href="https://unicode-org.github.io/icu/userguide/locale/resources">Users Guide</a>.
+ * <P>
+ *
+ * The ResourceBundle class is not suitable for subclassing.
+ *
+ * @stable ICU 2.0
+ */
+class U_COMMON_API ResourceBundle : public UObject {
+public:
+    /**
+     * Constructor
+     *
+     * @param packageName   The packageName and locale together point to an ICU udata object, 
+     *                      as defined by <code> udata_open( packageName, "res", locale, err) </code> 
+     *                      or equivalent.  Typically, packageName will refer to a (.dat) file, or to
+     *                      a package registered with udata_setAppData(). Using a full file or directory
+     *                      pathname for packageName is deprecated.
+     * @param locale  This is the locale this resource bundle is for. To get resources
+     *                for the French locale, for example, you would create a
+     *                ResourceBundle passing Locale::FRENCH for the "locale" parameter,
+     *                and all subsequent calls to that resource bundle will return
+     *                resources that pertain to the French locale. If the caller doesn't
+     *                pass a locale parameter, the default locale for the system (as
+     *                returned by Locale::getDefault()) will be used.
+     * @param err     The Error Code.
+     * The UErrorCode& err parameter is used to return status information to the user. To
+     * check whether the construction succeeded or not, you should check the value of
+     * U_SUCCESS(err). If you wish more detailed information, you can check for
+     * informational error results which still indicate success. U_USING_FALLBACK_WARNING
+     * indicates that a fall back locale was used. For example, 'de_CH' was requested,
+     * but nothing was found there, so 'de' was used. U_USING_DEFAULT_WARNING indicates that
+     * the default locale data was used; neither the requested locale nor any of its
+     * fall back locales could be found.
+     * @stable ICU 2.0
+     */
+    ResourceBundle(const UnicodeString&    packageName,
+                   const Locale&           locale,
+                   UErrorCode&              err);
+
+    /**
+     * Construct a resource bundle for the default bundle in the specified package.
+     *
+     * @param packageName   The packageName and locale together point to an ICU udata object, 
+     *                      as defined by <code> udata_open( packageName, "res", locale, err) </code> 
+     *                      or equivalent.  Typically, packageName will refer to a (.dat) file, or to
+     *                      a package registered with udata_setAppData(). Using a full file or directory
+     *                      pathname for packageName is deprecated.
+     * @param err A UErrorCode value
+     * @stable ICU 2.0
+     */
+    ResourceBundle(const UnicodeString&    packageName,
+                   UErrorCode&              err);
+
+    /**
+     * Construct a resource bundle for the ICU default bundle.
+     *
+     * @param err A UErrorCode value
+     * @stable ICU 2.0
+     */
+    ResourceBundle(UErrorCode &err);
+
+    /**
+     * Standard constructor, constructs a resource bundle for the locale-specific
+     * bundle in the specified package.
+     *
+     * @param packageName   The packageName and locale together point to an ICU udata object, 
+     *                      as defined by <code> udata_open( packageName, "res", locale, err) </code> 
+     *                      or equivalent.  Typically, packageName will refer to a (.dat) file, or to
+     *                      a package registered with udata_setAppData(). Using a full file or directory
+     *                      pathname for packageName is deprecated.
+     *                      nullptr is used to refer to ICU data.
+     * @param locale The locale for which to open a resource bundle.
+     * @param err A UErrorCode value
+     * @stable ICU 2.0
+     */
+    ResourceBundle(const char* packageName,
+                   const Locale& locale,
+                   UErrorCode& err);
+
+    /**
+     * Copy constructor.
+     *
+     * @param original The resource bundle to copy.
+     * @stable ICU 2.0
+     */
+    ResourceBundle(const ResourceBundle &original);
+
+    /**
+     * Constructor from a C UResourceBundle. The resource bundle is
+     * copied and not adopted. ures_close will still need to be used on the
+     * original resource bundle.
+     *
+     * @param res A pointer to the C resource bundle.
+     * @param status A UErrorCode value.
+     * @stable ICU 2.0
+     */
+    ResourceBundle(UResourceBundle *res,
+                   UErrorCode &status);
+
+    /**
+     * Assignment operator.
+     *
+     * @param other The resource bundle to copy.
+     * @stable ICU 2.0
+     */
+    ResourceBundle&
+      operator=(const ResourceBundle& other);
+
+    /** Destructor.
+     * @stable ICU 2.0
+     */
+    virtual ~ResourceBundle();
+
+    /**
+     * Clone this object.
+     * Clones can be used concurrently in multiple threads.
+     * If an error occurs, then nullptr is returned.
+     * The caller must delete the clone.
+     *
+     * @return a clone of this object
+     *
+     * @see getDynamicClassID
+     * @stable ICU 2.8
+     */
+    ResourceBundle *clone() const;
+
+    /**
+     * Returns the size of a resource. Size for scalar types is always 1, and for vector/table types is
+     * the number of child resources.
+     * @warning Integer array is treated as a scalar type. There are no
+     *          APIs to access individual members of an integer array. It
+     *          is always returned as a whole.
+     *
+     * @return number of resources in a given resource.
+     * @stable ICU 2.0
+     */
+    int32_t getSize() const;
+
+    /**
+     * returns a string from a string resource type
+     *
+     * @param status  fills in the outgoing error code
+     *                could be <TT>U_MISSING_RESOURCE_ERROR</TT> if the key is not found
+     *                could be a warning
+     *                e.g.: <TT>U_USING_FALLBACK_WARNING</TT>,<TT>U_USING_DEFAULT_WARNING </TT>
+     * @return a pointer to a zero-terminated char16_t array which lives in a memory mapped/DLL file.
+     * @stable ICU 2.0
+     */
+    UnicodeString
+      getString(UErrorCode& status) const;
+
+    /**
+     * returns a binary data from a resource. Can be used at most primitive resource types (binaries,
+     * strings, ints)
+     *
+     * @param len     fills in the length of resulting byte chunk
+     * @param status  fills in the outgoing error code
+     *                could be <TT>U_MISSING_RESOURCE_ERROR</TT> if the key is not found
+     *                could be a warning
+     *                e.g.: <TT>U_USING_FALLBACK_WARNING</TT>,<TT>U_USING_DEFAULT_WARNING </TT>
+     * @return a pointer to a chunk of unsigned bytes which live in a memory mapped/DLL file.
+     * @stable ICU 2.0
+     */
+    const uint8_t*
+      getBinary(int32_t& len, UErrorCode& status) const;
+
+
+    /**
+     * returns an integer vector from a resource.
+     *
+     * @param len     fills in the length of resulting integer vector
+     * @param status  fills in the outgoing error code
+     *                could be <TT>U_MISSING_RESOURCE_ERROR</TT> if the key is not found
+     *                could be a warning
+     *                e.g.: <TT>U_USING_FALLBACK_WARNING</TT>,<TT>U_USING_DEFAULT_WARNING </TT>
+     * @return a pointer to a vector of integers that lives in a memory mapped/DLL file.
+     * @stable ICU 2.0
+     */
+    const int32_t*
+      getIntVector(int32_t& len, UErrorCode& status) const;
+
+    /**
+     * returns an unsigned integer from a resource.
+     * This integer is originally 28 bits.
+     *
+     * @param status  fills in the outgoing error code
+     *                could be <TT>U_MISSING_RESOURCE_ERROR</TT> if the key is not found
+     *                could be a warning
+     *                e.g.: <TT>U_USING_FALLBACK_WARNING</TT>,<TT>U_USING_DEFAULT_WARNING </TT>
+     * @return an unsigned integer value
+     * @stable ICU 2.0
+     */
+    uint32_t
+      getUInt(UErrorCode& status) const;
+
+    /**
+     * returns a signed integer from a resource.
+     * This integer is originally 28 bit and the sign gets propagated.
+     *
+     * @param status  fills in the outgoing error code
+     *                could be <TT>U_MISSING_RESOURCE_ERROR</TT> if the key is not found
+     *                could be a warning
+     *                e.g.: <TT>U_USING_FALLBACK_WARNING</TT>,<TT>U_USING_DEFAULT_WARNING </TT>
+     * @return a signed integer value
+     * @stable ICU 2.0
+     */
+    int32_t
+      getInt(UErrorCode& status) const;
+
+    /**
+     * Checks whether the resource has another element to iterate over.
+     *
+     * @return true if there are more elements, false if there is no more elements
+     * @stable ICU 2.0
+     */
+    UBool hasNext() const;
+
+    /**
+     * Resets the internal context of a resource so that iteration starts from the first element.
+     *
+     * @stable ICU 2.0
+     */
+    void resetIterator();
+
+    /**
+     * Returns the key associated with this resource. Not all the resources have a key - only
+     * those that are members of a table.
+     *
+     * @return a key associated to this resource, or nullptr if it doesn't have a key
+     * @stable ICU 2.0
+     */
+    const char* getKey() const;
+
+    /**
+     * Gets the locale ID of the resource bundle as a string.
+     * Same as getLocale().getName() .
+     *
+     * @return the locale ID of the resource bundle as a string
+     * @stable ICU 2.0
+     */
+    const char* getName() const;
+
+    /**
+     * Returns the type of a resource. Available types are defined in enum UResType
+     *
+     * @return type of the given resource.
+     * @stable ICU 2.0
+     */
+    UResType getType() const;
+
+    /**
+     * Returns the next resource in a given resource or nullptr if there are no more resources
+     *
+     * @param status            fills in the outgoing error code
+     * @return                  ResourceBundle object.
+     * @stable ICU 2.0
+     */
+    ResourceBundle
+      getNext(UErrorCode& status);
+
+    /**
+     * Returns the next string in a resource or nullptr if there are no more resources
+     * to iterate over.
+     *
+     * @param status            fills in the outgoing error code
+     * @return an UnicodeString object.
+     * @stable ICU 2.0
+     */
+    UnicodeString
+      getNextString(UErrorCode& status);
+
+    /**
+     * Returns the next string in a resource or nullptr if there are no more resources
+     * to iterate over.
+     *
+     * @param key               fill in for key associated with this string
+     * @param status            fills in the outgoing error code
+     * @return an UnicodeString object.
+     * @stable ICU 2.0
+     */
+    UnicodeString
+      getNextString(const char ** key,
+                    UErrorCode& status);
+
+    /**
+     * Returns the resource in a resource at the specified index.
+     *
+     * @param index             an index to the wanted resource.
+     * @param status            fills in the outgoing error code
+     * @return                  ResourceBundle object. If there is an error, resource is invalid.
+     * @stable ICU 2.0
+     */
+    ResourceBundle
+      get(int32_t index,
+          UErrorCode& status) const;
+
+    /**
+     * Returns the string in a given resource at the specified index.
+     *
+     * @param index             an index to the wanted string.
+     * @param status            fills in the outgoing error code
+     * @return                  an UnicodeString object. If there is an error, string is bogus
+     * @stable ICU 2.0
+     */
+    UnicodeString
+      getStringEx(int32_t index,
+                  UErrorCode& status) const;
+
+    /**
+     * Returns a resource in a resource that has a given key. This procedure works only with table
+     * resources.
+     *
+     * @param key               a key associated with the wanted resource
+     * @param status            fills in the outgoing error code.
+     * @return                  ResourceBundle object. If there is an error, resource is invalid.
+     * @stable ICU 2.0
+     */
+    ResourceBundle
+      get(const char* key,
+          UErrorCode& status) const;
+
+    /**
+     * Returns a string in a resource that has a given key. This procedure works only with table
+     * resources.
+     *
+     * @param key               a key associated with the wanted string
+     * @param status            fills in the outgoing error code
+     * @return                  an UnicodeString object. If there is an error, string is bogus
+     * @stable ICU 2.0
+     */
+    UnicodeString
+      getStringEx(const char* key,
+                  UErrorCode& status) const;
+
+#ifndef U_HIDE_DEPRECATED_API
+    /**
+     * Return the version number associated with this ResourceBundle as a string. Please
+     * use getVersion, as this method is going to be deprecated.
+     *
+     * @return  A version number string as specified in the resource bundle or its parent.
+     *          The caller does not own this string.
+     * @see getVersion
+     * @deprecated ICU 2.8 Use getVersion instead.
+     */
+    const char* getVersionNumber() const;
+#endif  /* U_HIDE_DEPRECATED_API */
+
+    /**
+     * Return the version number associated with this ResourceBundle as a UVersionInfo array.
+     *
+     * @param versionInfo A UVersionInfo array that is filled with the version number
+     *                    as specified in the resource bundle or its parent.
+     * @stable ICU 2.0
+     */
+    void
+      getVersion(UVersionInfo versionInfo) const;
+
+#ifndef U_HIDE_DEPRECATED_API
+    /**
+     * Return the Locale associated with this ResourceBundle.
+     *
+     * @return a Locale object
+     * @deprecated ICU 2.8 Use getLocale(ULocDataLocaleType type, UErrorCode &status) overload instead.
+     */
+    const Locale& getLocale() const;
+#endif  /* U_HIDE_DEPRECATED_API */
+
+    /**
+     * Return the Locale associated with this ResourceBundle.
+     * @param type You can choose between requested, valid and actual
+     *             locale. For description see the definition of
+     *             ULocDataLocaleType in uloc.h
+     * @param status just for catching illegal arguments
+     *
+     * @return a Locale object
+     * @stable ICU 2.8
+     */
+    Locale
+      getLocale(ULocDataLocaleType type, UErrorCode &status) const;
+#ifndef U_HIDE_INTERNAL_API
+    /**
+     * This API implements multilevel fallback
+     * @internal
+     */
+    ResourceBundle
+        getWithFallback(const char* key, UErrorCode& status);
+#endif  /* U_HIDE_INTERNAL_API */
+    /**
+     * ICU "poor man's RTTI", returns a UClassID for the actual class.
+     *
+     * @stable ICU 2.2
+     */
+    virtual UClassID getDynamicClassID() const override;
+
+    /**
+     * ICU "poor man's RTTI", returns a UClassID for this class.
+     *
+     * @stable ICU 2.2
+     */
+    static UClassID U_EXPORT2 getStaticClassID();
+
+private:
+    ResourceBundle() = delete; // default constructor not implemented
+
+    UResourceBundle *fResource;
+    void constructForLocale(const UnicodeString& path, const Locale& locale, UErrorCode& error);
+    Locale *fLocale;
+};
+
+U_NAMESPACE_END
+
+#endif /* U_SHOW_CPLUSPLUS_API */
+
+#endif

thirdparty/icu4c/common/unicode/schriter.h

@@ -0,0 +1,187 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+/*
+******************************************************************************
+*
+*   Copyright (C) 1998-2005, International Business Machines
+*   Corporation and others.  All Rights Reserved.
+*
+******************************************************************************
+*
+* File schriter.h
+*
+* Modification History:
+*
+*   Date        Name        Description
+*  05/05/99     stephen     Cleaned up.
+******************************************************************************
+*/
+
+#ifndef SCHRITER_H
+#define SCHRITER_H
+
+#include "unicode/utypes.h"
+
+#if U_SHOW_CPLUSPLUS_API
+
+#include "unicode/chariter.h"
+#include "unicode/uchriter.h"
+
+/**
+ * \file 
+ * \brief C++ API: String Character Iterator
+ */
+ 
+U_NAMESPACE_BEGIN
+/**
+ * A concrete subclass of CharacterIterator that iterates over the
+ * characters (code units or code points) in a UnicodeString.
+ * It's possible not only to create an
+ * iterator that iterates over an entire UnicodeString, but also to
+ * create one that iterates over only a subrange of a UnicodeString
+ * (iterators over different subranges of the same UnicodeString don't
+ * compare equal).
+ * @see CharacterIterator
+ * @see ForwardCharacterIterator
+ * @stable ICU 2.0
+ */
+class U_COMMON_API StringCharacterIterator : public UCharCharacterIterator {
+public:
+  /**
+   * Create an iterator over the UnicodeString referred to by "textStr".
+   * The UnicodeString object is copied.
+   * The iteration range is the whole string, and the starting position is 0.
+   * @param textStr The unicode string used to create an iterator
+   * @stable ICU 2.0
+   */
+  StringCharacterIterator(const UnicodeString& textStr);
+
+  /**
+   * Create an iterator over the UnicodeString referred to by "textStr".
+   * The iteration range is the whole string, and the starting
+   * position is specified by "textPos".  If "textPos" is outside the valid
+   * iteration range, the behavior of this object is undefined.
+   * @param textStr The unicode string used to create an iterator
+   * @param textPos The starting position of the iteration
+   * @stable ICU 2.0
+   */
+  StringCharacterIterator(const UnicodeString&    textStr,
+              int32_t              textPos);
+
+  /**
+   * Create an iterator over the UnicodeString referred to by "textStr".
+   * The UnicodeString object is copied.
+   * The iteration range begins with the code unit specified by
+   * "textBegin" and ends with the code unit BEFORE the code unit specified
+   * by "textEnd".  The starting position is specified by "textPos".  If
+   * "textBegin" and "textEnd" don't form a valid range on "text" (i.e.,
+   * textBegin >= textEnd or either is negative or greater than text.size()),
+   * or "textPos" is outside the range defined by "textBegin" and "textEnd",
+   * the behavior of this iterator is undefined.
+   * @param textStr    The unicode string used to create the StringCharacterIterator
+   * @param textBegin  The begin position of the iteration range
+   * @param textEnd    The end position of the iteration range
+   * @param textPos    The starting position of the iteration
+   * @stable ICU 2.0
+   */
+  StringCharacterIterator(const UnicodeString&    textStr,
+              int32_t              textBegin,
+              int32_t              textEnd,
+              int32_t              textPos);
+
+  /**
+   * Copy constructor.  The new iterator iterates over the same range
+   * of the same string as "that", and its initial position is the
+   * same as "that"'s current position.
+   * The UnicodeString object in "that" is copied.
+   * @param that The StringCharacterIterator to be copied
+   * @stable ICU 2.0
+   */
+  StringCharacterIterator(const StringCharacterIterator&  that);
+
+  /**
+   * Destructor.
+   * @stable ICU 2.0
+   */
+  virtual ~StringCharacterIterator();
+
+  /**
+   * Assignment operator.  *this is altered to iterate over the same
+   * range of the same string as "that", and refers to the same
+   * character within that string as "that" does.
+   * @param that The object to be copied.
+   * @return the newly created object.
+   * @stable ICU 2.0
+   */
+  StringCharacterIterator&
+  operator=(const StringCharacterIterator&    that);
+
+  /**
+   * Returns true if the iterators iterate over the same range of the
+   * same string and are pointing at the same character.
+   * @param that The ForwardCharacterIterator to be compared for equality
+   * @return true if the iterators iterate over the same range of the
+   * same string and are pointing at the same character.
+   * @stable ICU 2.0
+   */
+  virtual bool           operator==(const ForwardCharacterIterator& that) const override;
+
+  /**
+   * Returns a new StringCharacterIterator referring to the same
+   * character in the same range of the same string as this one.  The
+   * caller must delete the new iterator.
+   * @return the newly cloned object.
+   * @stable ICU 2.0
+   */
+  virtual StringCharacterIterator* clone() const override;
+
+  /**
+   * Sets the iterator to iterate over the provided string.
+   * @param newText The string to be iterated over
+   * @stable ICU 2.0
+   */
+  void setText(const UnicodeString& newText);
+
+  /**
+   * Copies the UnicodeString under iteration into the UnicodeString
+   * referred to by "result".  Even if this iterator iterates across
+   * only a part of this string, the whole string is copied.
+   * @param result Receives a copy of the text under iteration.
+   * @stable ICU 2.0
+   */
+  virtual void            getText(UnicodeString& result) override;
+
+  /**
+   * Return a class ID for this object (not really public)
+   * @return a class ID for this object.
+   * @stable ICU 2.0
+   */
+  virtual UClassID getDynamicClassID() const override;
+
+  /**
+   * Return a class ID for this class (not really public)
+   * @return a class ID for this class
+   * @stable ICU 2.0
+   */
+  static UClassID U_EXPORT2 getStaticClassID();
+
+protected:
+  /**
+   * Default constructor, iteration over empty string.
+   * @stable ICU 2.0
+   */
+  StringCharacterIterator();
+
+  /**
+   * Copy of the iterated string object.
+   * @stable ICU 2.0
+   */
+  UnicodeString            text;
+
+};
+
+U_NAMESPACE_END
+
+#endif /* U_SHOW_CPLUSPLUS_API */
+
+#endif

thirdparty/icu4c/common/unicode/simpleformatter.h

@@ -0,0 +1,339 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+/*
+******************************************************************************
+* Copyright (C) 2014-2016, International Business Machines
+* Corporation and others.  All Rights Reserved.
+******************************************************************************
+* simpleformatter.h
+*/
+
+#ifndef __SIMPLEFORMATTER_H__
+#define __SIMPLEFORMATTER_H__
+
+/**
+ * \file
+ * \brief C++ API: Simple formatter, minimal subset of MessageFormat.
+ */
+
+#include "unicode/utypes.h"
+
+#if U_SHOW_CPLUSPLUS_API
+
+#include "unicode/unistr.h"
+
+U_NAMESPACE_BEGIN
+
+// Forward declaration:
+namespace number::impl {
+class SimpleModifier;
+}
+
+/**
+ * Formats simple patterns like "{1} was born in {0}".
+ * Minimal subset of MessageFormat; fast, simple, minimal dependencies.
+ * Supports only numbered arguments with no type nor style parameters,
+ * and formats only string values.
+ * Quoting via ASCII apostrophe compatible with ICU MessageFormat default behavior.
+ *
+ * Factory methods set error codes for syntax errors
+ * and for too few or too many arguments/placeholders.
+ *
+ * SimpleFormatter objects are thread-safe except for assignment and applying new patterns.
+ *
+ * Example:
+ * <pre>
+ * UErrorCode errorCode = U_ZERO_ERROR;
+ * SimpleFormatter fmt("{1} '{born}' in {0}", errorCode);
+ * UnicodeString result;
+ *
+ * // Output: "paul {born} in england"
+ * fmt.format("england", "paul", result, errorCode);
+ * </pre>
+ *
+ * This class is not intended for public subclassing.
+ *
+ * @see MessageFormat
+ * @see UMessagePatternApostropheMode
+ * @stable ICU 57
+ */
+class U_COMMON_API SimpleFormatter final : public UMemory {
+public:
+    /**
+     * Default constructor.
+     * @stable ICU 57
+     */
+    SimpleFormatter() : compiledPattern(static_cast<char16_t>(0)) {}
+
+    /**
+     * Constructs a formatter from the pattern string.
+     *
+     * @param pattern The pattern string.
+     * @param errorCode ICU error code in/out parameter.
+     *                  Must fulfill U_SUCCESS before the function call.
+     *                  Set to U_ILLEGAL_ARGUMENT_ERROR for bad argument syntax.
+     * @stable ICU 57
+     */
+    SimpleFormatter(const UnicodeString& pattern, UErrorCode &errorCode) {
+        applyPattern(pattern, errorCode);
+    }
+
+    /**
+     * Constructs a formatter from the pattern string.
+     * The number of arguments checked against the given limits is the
+     * highest argument number plus one, not the number of occurrences of arguments.
+     *
+     * @param pattern The pattern string.
+     * @param min The pattern must have at least this many arguments.
+     * @param max The pattern must have at most this many arguments.
+     * @param errorCode ICU error code in/out parameter.
+     *                  Must fulfill U_SUCCESS before the function call.
+     *                  Set to U_ILLEGAL_ARGUMENT_ERROR for bad argument syntax and
+     *                  too few or too many arguments.
+     * @stable ICU 57
+     */
+    SimpleFormatter(const UnicodeString& pattern, int32_t min, int32_t max,
+                    UErrorCode &errorCode) {
+        applyPatternMinMaxArguments(pattern, min, max, errorCode);
+    }
+
+    /**
+     * Copy constructor.
+     * @stable ICU 57
+     */
+    SimpleFormatter(const SimpleFormatter& other)
+            : compiledPattern(other.compiledPattern) {}
+
+    /**
+     * Assignment operator.
+     * @stable ICU 57
+     */
+    SimpleFormatter &operator=(const SimpleFormatter& other);
+
+    /**
+     * Destructor.
+     * @stable ICU 57
+     */
+    ~SimpleFormatter();
+
+    /**
+     * Changes this object according to the new pattern.
+     *
+     * @param pattern The pattern string.
+     * @param errorCode ICU error code in/out parameter.
+     *                  Must fulfill U_SUCCESS before the function call.
+     *                  Set to U_ILLEGAL_ARGUMENT_ERROR for bad argument syntax.
+     * @return true if U_SUCCESS(errorCode).
+     * @stable ICU 57
+     */
+    UBool applyPattern(const UnicodeString &pattern, UErrorCode &errorCode) {
+        return applyPatternMinMaxArguments(pattern, 0, INT32_MAX, errorCode);
+    }
+
+    /**
+     * Changes this object according to the new pattern.
+     * The number of arguments checked against the given limits is the
+     * highest argument number plus one, not the number of occurrences of arguments.
+     *
+     * @param pattern The pattern string.
+     * @param min The pattern must have at least this many arguments.
+     * @param max The pattern must have at most this many arguments.
+     * @param errorCode ICU error code in/out parameter.
+     *                  Must fulfill U_SUCCESS before the function call.
+     *                  Set to U_ILLEGAL_ARGUMENT_ERROR for bad argument syntax and
+     *                  too few or too many arguments.
+     * @return true if U_SUCCESS(errorCode).
+     * @stable ICU 57
+     */
+    UBool applyPatternMinMaxArguments(const UnicodeString &pattern,
+                                      int32_t min, int32_t max, UErrorCode &errorCode);
+
+    /**
+     * @return The max argument number + 1.
+     * @stable ICU 57
+     */
+    int32_t getArgumentLimit() const {
+        return getArgumentLimit(compiledPattern.getBuffer(), compiledPattern.length());
+    }
+
+    /**
+     * Formats the given value, appending to the appendTo builder.
+     * The argument value must not be the same object as appendTo.
+     * getArgumentLimit() must be at most 1.
+     *
+     * @param value0 Value for argument {0}.
+     * @param appendTo Gets the formatted pattern and value appended.
+     * @param errorCode ICU error code in/out parameter.
+     *                  Must fulfill U_SUCCESS before the function call.
+     * @return appendTo
+     * @stable ICU 57
+     */
+    UnicodeString &format(
+            const UnicodeString &value0,
+            UnicodeString &appendTo, UErrorCode &errorCode) const;
+
+    /**
+     * Formats the given values, appending to the appendTo builder.
+     * An argument value must not be the same object as appendTo.
+     * getArgumentLimit() must be at most 2.
+     *
+     * @param value0 Value for argument {0}.
+     * @param value1 Value for argument {1}.
+     * @param appendTo Gets the formatted pattern and values appended.
+     * @param errorCode ICU error code in/out parameter.
+     *                  Must fulfill U_SUCCESS before the function call.
+     * @return appendTo
+     * @stable ICU 57
+     */
+    UnicodeString &format(
+            const UnicodeString &value0,
+            const UnicodeString &value1,
+            UnicodeString &appendTo, UErrorCode &errorCode) const;
+
+    /**
+     * Formats the given values, appending to the appendTo builder.
+     * An argument value must not be the same object as appendTo.
+     * getArgumentLimit() must be at most 3.
+     *
+     * @param value0 Value for argument {0}.
+     * @param value1 Value for argument {1}.
+     * @param value2 Value for argument {2}.
+     * @param appendTo Gets the formatted pattern and values appended.
+     * @param errorCode ICU error code in/out parameter.
+     *                  Must fulfill U_SUCCESS before the function call.
+     * @return appendTo
+     * @stable ICU 57
+     */
+    UnicodeString &format(
+            const UnicodeString &value0,
+            const UnicodeString &value1,
+            const UnicodeString &value2,
+            UnicodeString &appendTo, UErrorCode &errorCode) const;
+
+    /**
+     * Formats the given values, appending to the appendTo string.
+     *
+     * @param values The argument values.
+     *               An argument value must not be the same object as appendTo.
+     *               Can be nullptr if valuesLength==getArgumentLimit()==0.
+     * @param valuesLength The length of the values array.
+     *                     Must be at least getArgumentLimit().
+     * @param appendTo Gets the formatted pattern and values appended.
+     * @param offsets offsets[i] receives the offset of where
+     *                values[i] replaced pattern argument {i}.
+     *                Can be shorter or longer than values. Can be nullptr if offsetsLength==0.
+     *                If there is no {i} in the pattern, then offsets[i] is set to -1.
+     * @param offsetsLength The length of the offsets array.
+     * @param errorCode ICU error code in/out parameter.
+     *                  Must fulfill U_SUCCESS before the function call.
+     * @return appendTo
+     * @stable ICU 57
+     */
+    UnicodeString &formatAndAppend(
+            const UnicodeString *const *values, int32_t valuesLength,
+            UnicodeString &appendTo,
+            int32_t *offsets, int32_t offsetsLength, UErrorCode &errorCode) const;
+
+    /**
+     * Formats the given values, replacing the contents of the result string.
+     * May optimize by actually appending to the result if it is the same object
+     * as the value corresponding to the initial argument in the pattern.
+     *
+     * @param values The argument values.
+     *               An argument value may be the same object as result.
+     *               Can be nullptr if valuesLength==getArgumentLimit()==0.
+     * @param valuesLength The length of the values array.
+     *                     Must be at least getArgumentLimit().
+     * @param result Gets its contents replaced by the formatted pattern and values.
+     * @param offsets offsets[i] receives the offset of where
+     *                values[i] replaced pattern argument {i}.
+     *                Can be shorter or longer than values. Can be nullptr if offsetsLength==0.
+     *                If there is no {i} in the pattern, then offsets[i] is set to -1.
+     * @param offsetsLength The length of the offsets array.
+     * @param errorCode ICU error code in/out parameter.
+     *                  Must fulfill U_SUCCESS before the function call.
+     * @return result
+     * @stable ICU 57
+     */
+    UnicodeString &formatAndReplace(
+            const UnicodeString *const *values, int32_t valuesLength,
+            UnicodeString &result,
+            int32_t *offsets, int32_t offsetsLength, UErrorCode &errorCode) const;
+
+    /**
+     * Returns the pattern text with none of the arguments.
+     * Like formatting with all-empty string values.
+     * @stable ICU 57
+     */
+    UnicodeString getTextWithNoArguments() const {
+        return getTextWithNoArguments(
+            compiledPattern.getBuffer(),
+            compiledPattern.length(),
+            nullptr,
+            0);
+    }
+
+#ifndef U_HIDE_INTERNAL_API
+    /**
+     * Returns the pattern text with none of the arguments.
+     * Like formatting with all-empty string values.
+     *
+     * TODO(ICU-20406): Replace this with an Iterator interface.
+     *
+     * @param offsets offsets[i] receives the offset of where {i} was located
+     *                before it was replaced by an empty string.
+     *                For example, "a{0}b{1}" produces offset 1 for i=0 and 2 for i=1.
+     *                Can be nullptr if offsetsLength==0.
+     *                If there is no {i} in the pattern, then offsets[i] is set to -1.
+     * @param offsetsLength The length of the offsets array.
+     *
+     * @internal
+     */
+    UnicodeString getTextWithNoArguments(int32_t *offsets, int32_t offsetsLength) const {
+        return getTextWithNoArguments(
+            compiledPattern.getBuffer(),
+            compiledPattern.length(),
+            offsets,
+            offsetsLength);
+    }
+#endif // U_HIDE_INTERNAL_API
+
+private:
+    /**
+     * Binary representation of the compiled pattern.
+     * Index 0: One more than the highest argument number.
+     * Followed by zero or more arguments or literal-text segments.
+     *
+     * An argument is stored as its number, less than ARG_NUM_LIMIT.
+     * A literal-text segment is stored as its length (at least 1) offset by ARG_NUM_LIMIT,
+     * followed by that many chars.
+     */
+    UnicodeString compiledPattern;
+
+    static inline int32_t getArgumentLimit(const char16_t *compiledPattern,
+                                              int32_t compiledPatternLength) {
+        return compiledPatternLength == 0 ? 0 : compiledPattern[0];
+    }
+
+    static UnicodeString getTextWithNoArguments(
+        const char16_t *compiledPattern,
+        int32_t compiledPatternLength,
+        int32_t *offsets,
+        int32_t offsetsLength);
+
+    static UnicodeString &format(
+            const char16_t *compiledPattern, int32_t compiledPatternLength,
+            const UnicodeString *const *values,
+            UnicodeString &result, const UnicodeString *resultCopy, UBool forbidResultAsValue,
+            int32_t *offsets, int32_t offsetsLength,
+            UErrorCode &errorCode);
+
+    // Give access to internals to SimpleModifier for number formatting
+    friend class number::impl::SimpleModifier;
+};
+
+U_NAMESPACE_END
+
+#endif /* U_SHOW_CPLUSPLUS_API */
+
+#endif  // __SIMPLEFORMATTER_H__

thirdparty/icu4c/common/unicode/std_string.h

@@ -0,0 +1,41 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+/*
+*******************************************************************************
+*
+*   Copyright (C) 2009-2014, International Business Machines
+*   Corporation and others.  All Rights Reserved.
+*
+*******************************************************************************
+*   file name:  std_string.h
+*   encoding:   UTF-8
+*   tab size:   8 (not used)
+*   indentation:4
+*
+*   created on: 2009feb19
+*   created by: Markus W. Scherer
+*/
+
+#ifndef __STD_STRING_H__
+#define __STD_STRING_H__
+
+/**
+ * \file
+ * \brief C++ API: Central ICU header for including the C++ standard <string>
+ *                 header and for related definitions.
+ */
+
+#include "unicode/utypes.h"
+
+#if U_SHOW_CPLUSPLUS_API
+
+// Workaround for a libstdc++ bug before libstdc++4.6 (2011).
+// https://bugs.llvm.org/show_bug.cgi?id=13364
+#if defined(__GLIBCXX__)
+namespace std { class type_info; }
+#endif
+#include <string>
+
+#endif /* U_SHOW_CPLUSPLUS_API */
+
+#endif  // __STD_STRING_H__

thirdparty/icu4c/common/unicode/strenum.h

@@ -0,0 +1,281 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+/*
+*******************************************************************************
+*
+*   Copyright (C) 2002-2012, International Business Machines
+*   Corporation and others.  All Rights Reserved.
+*
+*******************************************************************************
+*/
+
+#ifndef STRENUM_H
+#define STRENUM_H
+
+#include "unicode/utypes.h"
+
+#if U_SHOW_CPLUSPLUS_API
+
+#include "unicode/uobject.h"
+#include "unicode/unistr.h"
+
+/**
+ * \file 
+ * \brief C++ API: String Enumeration
+ */
+ 
+U_NAMESPACE_BEGIN
+
+/**
+ * Base class for 'pure' C++ implementations of uenum api.  Adds a
+ * method that returns the next UnicodeString since in C++ this can
+ * be a common storage format for strings.
+ *
+ * <p>The model is that the enumeration is over strings maintained by
+ * a 'service.'  At any point, the service might change, invalidating
+ * the enumerator (though this is expected to be rare).  The iterator
+ * returns an error if this has occurred.  Lack of the error is no
+ * guarantee that the service didn't change immediately after the
+ * call, so the returned string still might not be 'valid' on
+ * subsequent use.</p>
+ *
+ * <p>Strings may take the form of const char*, const char16_t*, or const
+ * UnicodeString*.  The type you get is determine by the variant of
+ * 'next' that you call.  In general the StringEnumeration is
+ * optimized for one of these types, but all StringEnumerations can
+ * return all types.  Returned strings are each terminated with a NUL.
+ * Depending on the service data, they might also include embedded NUL
+ * characters, so API is provided to optionally return the true
+ * length, counting the embedded NULs but not counting the terminating
+ * NUL.</p>
+ *
+ * <p>The pointers returned by next, unext, and snext become invalid
+ * upon any subsequent call to the enumeration's destructor, next,
+ * unext, snext, or reset.</p>
+ *
+ * ICU 2.8 adds some default implementations and helper functions
+ * for subclasses.
+ *
+ * @stable ICU 2.4 
+ */
+class U_COMMON_API StringEnumeration : public UObject { 
+public:
+    /**
+     * Destructor.
+     * @stable ICU 2.4
+     */
+    virtual ~StringEnumeration();
+
+    /**
+     * Clone this object, an instance of a subclass of StringEnumeration.
+     * Clones can be used concurrently in multiple threads.
+     * If a subclass does not implement clone(), or if an error occurs,
+     * then nullptr is returned.
+     * The caller must delete the clone.
+     *
+     * @return a clone of this object
+     *
+     * @see getDynamicClassID
+     * @stable ICU 2.8
+     */
+    virtual StringEnumeration *clone() const;
+
+    /**
+     * <p>Return the number of elements that the iterator traverses.  If
+     * the iterator is out of sync with its service, status is set to
+     * U_ENUM_OUT_OF_SYNC_ERROR, and the return value is zero.</p>
+     *
+     * <p>The return value will not change except possibly as a result of
+     * a subsequent call to reset, or if the iterator becomes out of sync.</p>
+     *
+     * <p>This is a convenience function. It can end up being very
+     * expensive as all the items might have to be pre-fetched
+     * (depending on the storage format of the data being
+     * traversed).</p>
+     *
+     * @param status the error code.
+     * @return number of elements in the iterator.
+     *
+     * @stable ICU 2.4 */
+    virtual int32_t count(UErrorCode& status) const = 0;
+
+    /**
+     * <p>Returns the next element as a NUL-terminated char*.  If there
+     * are no more elements, returns nullptr.  If the resultLength pointer
+     * is not nullptr, the length of the string (not counting the
+     * terminating NUL) is returned at that address.  If an error
+     * status is returned, the value at resultLength is undefined.</p>
+     *
+     * <p>The returned pointer is owned by this iterator and must not be
+     * deleted by the caller.  The pointer is valid until the next call
+     * to next, unext, snext, reset, or the enumerator's destructor.</p>
+     *
+     * <p>If the iterator is out of sync with its service, status is set
+     * to U_ENUM_OUT_OF_SYNC_ERROR and nullptr is returned.</p>
+     *
+     * <p>If the native service string is a char16_t* string, it is
+     * converted to char* with the invariant converter.  If the
+     * conversion fails (because a character cannot be converted) then
+     * status is set to U_INVARIANT_CONVERSION_ERROR and the return
+     * value is undefined (though not nullptr).</p>
+     *
+     * Starting with ICU 2.8, the default implementation calls snext()
+     * and handles the conversion.
+     * Either next() or snext() must be implemented differently by a subclass.
+     *
+     * @param status the error code.
+     * @param resultLength a pointer to receive the length, can be nullptr.
+     * @return a pointer to the string, or nullptr.
+     *
+     * @stable ICU 2.4 
+     */
+    virtual const char* next(int32_t *resultLength, UErrorCode& status);
+
+    /**
+     * <p>Returns the next element as a NUL-terminated char16_t*.  If there
+     * are no more elements, returns nullptr.  If the resultLength pointer
+     * is not nullptr, the length of the string (not counting the
+     * terminating NUL) is returned at that address.  If an error
+     * status is returned, the value at resultLength is undefined.</p>
+     *
+     * <p>The returned pointer is owned by this iterator and must not be
+     * deleted by the caller.  The pointer is valid until the next call
+     * to next, unext, snext, reset, or the enumerator's destructor.</p>
+     *
+     * <p>If the iterator is out of sync with its service, status is set
+     * to U_ENUM_OUT_OF_SYNC_ERROR and nullptr is returned.</p>
+     *
+     * Starting with ICU 2.8, the default implementation calls snext()
+     * and handles the conversion.
+     *
+     * @param status the error code.
+     * @param resultLength a pointer to receive the length, can be nullptr.
+     * @return a pointer to the string, or nullptr.
+     *
+     * @stable ICU 2.4 
+     */
+    virtual const char16_t* unext(int32_t *resultLength, UErrorCode& status);
+
+    /**
+     * <p>Returns the next element a UnicodeString*.  If there are no
+     * more elements, returns nullptr.</p>
+     *
+     * <p>The returned pointer is owned by this iterator and must not be
+     * deleted by the caller.  The pointer is valid until the next call
+     * to next, unext, snext, reset, or the enumerator's destructor.</p>
+     *
+     * <p>If the iterator is out of sync with its service, status is set
+     * to U_ENUM_OUT_OF_SYNC_ERROR and nullptr is returned.</p>
+     *
+     * Starting with ICU 2.8, the default implementation calls next()
+     * and handles the conversion.
+     * Either next() or snext() must be implemented differently by a subclass.
+     *
+     * @param status the error code.
+     * @return a pointer to the string, or nullptr.
+     *
+     * @stable ICU 2.4 
+     */
+    virtual const UnicodeString* snext(UErrorCode& status);
+
+    /**
+     * <p>Resets the iterator.  This re-establishes sync with the
+     * service and rewinds the iterator to start at the first
+     * element.</p>
+     *
+     * <p>Previous pointers returned by next, unext, or snext become
+     * invalid, and the value returned by count might change.</p>
+     *
+     * @param status the error code.
+     *
+     * @stable ICU 2.4 
+     */
+    virtual void reset(UErrorCode& status) = 0;
+
+    /**
+     * Compares this enumeration to other to check if both are equal
+     *
+     * @param that The other string enumeration to compare this object to
+     * @return true if the enumerations are equal. false if not.
+     * @stable ICU 3.6 
+     */
+    virtual bool operator==(const StringEnumeration& that)const;
+    /**
+     * Compares this enumeration to other to check if both are not equal
+     *
+     * @param that The other string enumeration to compare this object to
+     * @return true if the enumerations are equal. false if not.
+     * @stable ICU 3.6 
+     */
+    virtual bool operator!=(const StringEnumeration& that)const;
+
+protected:
+    /**
+     * UnicodeString field for use with default implementations and subclasses.
+     * @stable ICU 2.8
+     */
+    UnicodeString unistr;
+    /**
+     * char * default buffer for use with default implementations and subclasses.
+     * @stable ICU 2.8
+     */
+    char charsBuffer[32];
+    /**
+     * char * buffer for use with default implementations and subclasses.
+     * Allocated in constructor and in ensureCharsCapacity().
+     * @stable ICU 2.8
+     */
+    char *chars;
+    /**
+     * Capacity of chars, for use with default implementations and subclasses.
+     * @stable ICU 2.8
+     */
+    int32_t charsCapacity;
+
+    /**
+     * Default constructor for use with default implementations and subclasses.
+     * @stable ICU 2.8
+     */
+    StringEnumeration();
+
+    /**
+     * Ensures that chars is at least as large as the requested capacity.
+     * For use with default implementations and subclasses.
+     *
+     * @param capacity Requested capacity.
+     * @param status ICU in/out error code.
+     * @stable ICU 2.8
+     */
+    void ensureCharsCapacity(int32_t capacity, UErrorCode &status);
+
+    /**
+     * Converts s to Unicode and sets unistr to the result.
+     * For use with default implementations and subclasses,
+     * especially for implementations of snext() in terms of next().
+     * This is provided with a helper function instead of a default implementation
+     * of snext() to avoid potential infinite loops between next() and snext().
+     *
+     * For example:
+     * \code
+     * const UnicodeString* snext(UErrorCode& status) {
+     *   int32_t resultLength=0;
+     *   const char *s=next(&resultLength, status);
+     *   return setChars(s, resultLength, status);
+     * }
+     * \endcode
+     *
+     * @param s String to be converted to Unicode.
+     * @param length Length of the string.
+     * @param status ICU in/out error code.
+     * @return A pointer to unistr.
+     * @stable ICU 2.8
+     */
+    UnicodeString *setChars(const char *s, int32_t length, UErrorCode &status);
+};
+
+U_NAMESPACE_END
+
+#endif /* U_SHOW_CPLUSPLUS_API */
+
+/* STRENUM_H */
+#endif

thirdparty/icu4c/common/unicode/stringoptions.h

@@ -0,0 +1,190 @@
+// © 2017 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+
+// stringoptions.h
+// created: 2017jun08 Markus W. Scherer
+
+#ifndef __STRINGOPTIONS_H__
+#define __STRINGOPTIONS_H__
+
+#include "unicode/utypes.h"
+
+/**
+ * \file
+ * \brief C API: Bit set option bit constants for various string and character processing functions.
+ */
+
+/**
+ * Option value for case folding: Use default mappings defined in CaseFolding.txt.
+ *
+ * @stable ICU 2.0
+ */
+#define U_FOLD_CASE_DEFAULT 0
+
+/**
+ * Option value for case folding:
+ *
+ * Use the modified set of mappings provided in CaseFolding.txt to handle dotted I
+ * and dotless i appropriately for Turkic languages (tr, az).
+ *
+ * Before Unicode 3.2, CaseFolding.txt contains mappings marked with 'I' that
+ * are to be included for default mappings and
+ * excluded for the Turkic-specific mappings.
+ *
+ * Unicode 3.2 CaseFolding.txt instead contains mappings marked with 'T' that
+ * are to be excluded for default mappings and
+ * included for the Turkic-specific mappings.
+ *
+ * @stable ICU 2.0
+ */
+#define U_FOLD_CASE_EXCLUDE_SPECIAL_I 1
+
+/**
+ * Titlecase the string as a whole rather than each word.
+ * (Titlecase only the character at index 0, possibly adjusted.)
+ * Option bits value for titlecasing APIs that take an options bit set.
+ *
+ * It is an error to specify multiple titlecasing iterator options together,
+ * including both an options bit and an explicit BreakIterator.
+ *
+ * @see U_TITLECASE_ADJUST_TO_CASED
+ * @stable ICU 60
+ */
+#define U_TITLECASE_WHOLE_STRING 0x20
+
+/**
+ * Titlecase sentences rather than words.
+ * (Titlecase only the first character of each sentence, possibly adjusted.)
+ * Option bits value for titlecasing APIs that take an options bit set.
+ *
+ * It is an error to specify multiple titlecasing iterator options together,
+ * including both an options bit and an explicit BreakIterator.
+ *
+ * @see U_TITLECASE_ADJUST_TO_CASED
+ * @stable ICU 60
+ */
+#define U_TITLECASE_SENTENCES 0x40
+
+/**
+ * Do not lowercase non-initial parts of words when titlecasing.
+ * Option bit for titlecasing APIs that take an options bit set.
+ *
+ * By default, titlecasing will titlecase the character at each
+ * (possibly adjusted) BreakIterator index and
+ * lowercase all other characters up to the next iterator index.
+ * With this option, the other characters will not be modified.
+ *
+ * @see U_TITLECASE_ADJUST_TO_CASED
+ * @see UnicodeString::toTitle
+ * @see CaseMap::toTitle
+ * @see ucasemap_setOptions
+ * @see ucasemap_toTitle
+ * @see ucasemap_utf8ToTitle
+ * @stable ICU 3.8
+ */
+#define U_TITLECASE_NO_LOWERCASE 0x100
+
+/**
+ * Do not adjust the titlecasing BreakIterator indexes;
+ * titlecase exactly the characters at breaks from the iterator.
+ * Option bit for titlecasing APIs that take an options bit set.
+ *
+ * By default, titlecasing will take each break iterator index,
+ * adjust it to the next relevant character (see U_TITLECASE_ADJUST_TO_CASED),
+ * and titlecase that one.
+ *
+ * Other characters are lowercased.
+ *
+ * It is an error to specify multiple titlecasing adjustment options together.
+ *
+ * @see U_TITLECASE_ADJUST_TO_CASED
+ * @see U_TITLECASE_NO_LOWERCASE
+ * @see UnicodeString::toTitle
+ * @see CaseMap::toTitle
+ * @see ucasemap_setOptions
+ * @see ucasemap_toTitle
+ * @see ucasemap_utf8ToTitle
+ * @stable ICU 3.8
+ */
+#define U_TITLECASE_NO_BREAK_ADJUSTMENT 0x200
+
+/**
+ * Adjust each titlecasing BreakIterator index to the next cased character.
+ * (See the Unicode Standard, chapter 3, Default Case Conversion, R3 toTitlecase(X).)
+ * Option bit for titlecasing APIs that take an options bit set.
+ *
+ * This used to be the default index adjustment in ICU.
+ * Since ICU 60, the default index adjustment is to the next character that is
+ * a letter, number, symbol, or private use code point.
+ * (Uncased modifier letters are skipped.)
+ * The difference in behavior is small for word titlecasing,
+ * but the new adjustment is much better for whole-string and sentence titlecasing:
+ * It yields "49ers" and "«丰(abc)»" instead of "49Ers" and "«丰(Abc)»".
+ *
+ * It is an error to specify multiple titlecasing adjustment options together.
+ *
+ * @see U_TITLECASE_NO_BREAK_ADJUSTMENT
+ * @stable ICU 60
+ */
+#define U_TITLECASE_ADJUST_TO_CASED 0x400
+
+/**
+ * Option for string transformation functions to not first reset the Edits object.
+ * Used for example in some case-mapping and normalization functions.
+ *
+ * @see CaseMap
+ * @see Edits
+ * @see Normalizer2
+ * @stable ICU 60
+ */
+#define U_EDITS_NO_RESET 0x2000
+
+/**
+ * Omit unchanged text when recording how source substrings
+ * relate to changed and unchanged result substrings.
+ * Used for example in some case-mapping and normalization functions.
+ *
+ * @see CaseMap
+ * @see Edits
+ * @see Normalizer2
+ * @stable ICU 60
+ */
+#define U_OMIT_UNCHANGED_TEXT 0x4000
+
+/**
+ * Option bit for u_strCaseCompare, u_strcasecmp, unorm_compare, etc:
+ * Compare strings in code point order instead of code unit order.
+ * @stable ICU 2.2
+ */
+#define U_COMPARE_CODE_POINT_ORDER  0x8000
+
+/**
+ * Option bit for unorm_compare:
+ * Perform case-insensitive comparison.
+ * @stable ICU 2.2
+ */
+#define U_COMPARE_IGNORE_CASE       0x10000
+
+/**
+ * Option bit for unorm_compare:
+ * Both input strings are assumed to fulfill FCD conditions.
+ * @stable ICU 2.2
+ */
+#define UNORM_INPUT_IS_FCD          0x20000
+
+// Related definitions elsewhere.
+// Options that are not meaningful in the same functions
+// can share the same bits.
+//
+// Public:
+// unicode/unorm.h #define UNORM_COMPARE_NORM_OPTIONS_SHIFT 20
+//
+// Internal: (may change or be removed)
+// ucase.h #define _STRCASECMP_OPTIONS_MASK 0xffff
+// ucase.h #define _FOLD_CASE_OPTIONS_MASK 7
+// ucasemap_imp.h #define U_TITLECASE_ITERATOR_MASK 0xe0
+// ucasemap_imp.h #define U_TITLECASE_ADJUSTMENT_MASK 0x600
+// ustr_imp.h #define _STRNCMP_STYLE 0x1000
+// unormcmp.cpp #define _COMPARE_EQUIV 0x80000
+
+#endif  // __STRINGOPTIONS_H__

thirdparty/icu4c/common/unicode/stringpiece.h

@@ -0,0 +1,354 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+// Copyright (C) 2009-2013, International Business Machines
+// Corporation and others. All Rights Reserved.
+//
+// Copyright 2001 and onwards Google Inc.
+// Author: Sanjay Ghemawat
+
+// This code is a contribution of Google code, and the style used here is
+// a compromise between the original Google code and the ICU coding guidelines.
+// For example, data types are ICU-ified (size_t,int->int32_t),
+// and API comments doxygen-ified, but function names and behavior are
+// as in the original, if possible.
+// Assertion-style error handling, not available in ICU, was changed to
+// parameter "pinning" similar to UnicodeString.
+//
+// In addition, this is only a partial port of the original Google code,
+// limited to what was needed so far. The (nearly) complete original code
+// is in the ICU svn repository at icuhtml/trunk/design/strings/contrib
+// (see ICU ticket 6765, r25517).
+
+#ifndef __STRINGPIECE_H__
+#define __STRINGPIECE_H__
+
+/**
+ * \file 
+ * \brief C++ API: StringPiece: Read-only byte string wrapper class.
+ */
+
+#include "unicode/utypes.h"
+
+#if U_SHOW_CPLUSPLUS_API
+
+#include <cstddef>
+#include <string_view>
+#include <type_traits>
+
+#include "unicode/uobject.h"
+#include "unicode/std_string.h"
+
+// Arghh!  I wish C++ literals were "string".
+
+U_NAMESPACE_BEGIN
+
+/**
+ * A string-like object that points to a sized piece of memory.
+ *
+ * We provide non-explicit singleton constructors so users can pass
+ * in a "const char*" or a "string" wherever a "StringPiece" is
+ * expected.
+ *
+ * Functions or methods may use StringPiece parameters to accept either a
+ * "const char*" or a "string" value that will be implicitly converted to a
+ * StringPiece.
+ *
+ * Systematic usage of StringPiece is encouraged as it will reduce unnecessary
+ * conversions from "const char*" to "string" and back again.
+ *
+ * @stable ICU 4.2
+ */
+class U_COMMON_API StringPiece : public UMemory {
+ private:
+  const char*   ptr_;
+  int32_t       length_;
+
+ public:
+  /**
+   * Default constructor, creates an empty StringPiece.
+   * @stable ICU 4.2
+   */
+  StringPiece() : ptr_(nullptr), length_(0) { }
+
+  /**
+   * Constructs from a NUL-terminated const char * pointer.
+   * @param str a NUL-terminated const char * pointer
+   * @stable ICU 4.2
+   */
+  StringPiece(const char* str);
+#if defined(__cpp_char8_t) || defined(U_IN_DOXYGEN)
+  /**
+   * Constructs from a NUL-terminated const char8_t * pointer.
+   * @param str a NUL-terminated const char8_t * pointer
+   * @stable ICU 67
+   */
+  StringPiece(const char8_t* str) : StringPiece(reinterpret_cast<const char*>(str)) {}
+#endif
+  /**
+   * Constructs an empty StringPiece.
+   * Needed for type disambiguation from multiple other overloads.
+   * @param p nullptr
+   * @stable ICU 67
+   */
+  StringPiece(std::nullptr_t p) : ptr_(p), length_(0) {}
+
+  /**
+   * Constructs from a std::string.
+   * @stable ICU 4.2
+   */
+  StringPiece(const std::string& str)
+    : ptr_(str.data()), length_(static_cast<int32_t>(str.size())) { }
+#if defined(__cpp_lib_char8_t) || defined(U_IN_DOXYGEN)
+  /**
+   * Constructs from a std::u8string.
+   * @stable ICU 67
+   */
+  StringPiece(const std::u8string& str)
+    : ptr_(reinterpret_cast<const char*>(str.data())),
+      length_(static_cast<int32_t>(str.size())) { }
+#endif
+
+  /**
+   * Constructs from some other implementation of a string piece class, from any
+   * C++ record type that has these two methods:
+   *
+   * \code{.cpp}
+   *
+   *   struct OtherStringPieceClass {
+   *     const char* data();  // or const char8_t*
+   *     size_t size();
+   *   };
+   *
+   * \endcode
+   *
+   * The other string piece class will typically be std::string_view from C++17
+   * or absl::string_view from Abseil.
+   *
+   * Starting with C++20, data() may also return a const char8_t* pointer,
+   * as from std::u8string_view.
+   *
+   * @param str the other string piece
+   * @stable ICU 65
+   */
+  template <typename T,
+            typename = std::enable_if_t<
+                (std::is_same_v<decltype(T().data()), const char*>
+#if defined(__cpp_char8_t)
+                    || std::is_same_v<decltype(T().data()), const char8_t*>
+#endif
+                ) &&
+                std::is_same_v<decltype(T().size()), size_t>>>
+  StringPiece(T str)
+      : ptr_(reinterpret_cast<const char*>(str.data())),
+        length_(static_cast<int32_t>(str.size())) {}
+
+  /**
+   * Constructs from a const char * pointer and a specified length.
+   * @param offset a const char * pointer (need not be terminated)
+   * @param len the length of the string; must be non-negative
+   * @stable ICU 4.2
+   */
+  StringPiece(const char* offset, int32_t len) : ptr_(offset), length_(len) { }
+#if defined(__cpp_char8_t) || defined(U_IN_DOXYGEN)
+  /**
+   * Constructs from a const char8_t * pointer and a specified length.
+   * @param str a const char8_t * pointer (need not be terminated)
+   * @param len the length of the string; must be non-negative
+   * @stable ICU 67
+   */
+  StringPiece(const char8_t* str, int32_t len) :
+      StringPiece(reinterpret_cast<const char*>(str), len) {}
+#endif
+
+  /**
+   * Substring of another StringPiece.
+   * @param x the other StringPiece
+   * @param pos start position in x; must be non-negative and <= x.length().
+   * @stable ICU 4.2
+   */
+  StringPiece(const StringPiece& x, int32_t pos);
+  /**
+   * Substring of another StringPiece.
+   * @param x the other StringPiece
+   * @param pos start position in x; must be non-negative and <= x.length().
+   * @param len length of the substring;
+   *            must be non-negative and will be pinned to at most x.length() - pos.
+   * @stable ICU 4.2
+   */
+  StringPiece(const StringPiece& x, int32_t pos, int32_t len);
+
+#ifndef U_HIDE_INTERNAL_API
+  /**
+   * Converts to a std::string_view().
+   * @internal
+   */
+  inline operator std::string_view() const {
+    return {data(), static_cast<std::string_view::size_type>(size())};
+  }
+#endif  // U_HIDE_INTERNAL_API
+
+  /**
+   * Returns the string pointer. May be nullptr if it is empty.
+   *
+   * data() may return a pointer to a buffer with embedded NULs, and the
+   * returned buffer may or may not be null terminated.  Therefore it is
+   * typically a mistake to pass data() to a routine that expects a NUL
+   * terminated string.
+   * @return the string pointer
+   * @stable ICU 4.2
+   */
+  const char* data() const { return ptr_; }
+  /**
+   * Returns the string length. Same as length().
+   * @return the string length
+   * @stable ICU 4.2
+   */
+  int32_t size() const { return length_; }
+  /**
+   * Returns the string length. Same as size().
+   * @return the string length
+   * @stable ICU 4.2
+   */
+  int32_t length() const { return length_; }
+  /**
+   * Returns whether the string is empty.
+   * @return true if the string is empty
+   * @stable ICU 4.2
+   */
+  UBool empty() const { return length_ == 0; }
+
+  /**
+   * Sets to an empty string.
+   * @stable ICU 4.2
+   */
+  void clear() { ptr_ = nullptr; length_ = 0; }
+
+  /**
+   * Reset the stringpiece to refer to new data.
+   * @param xdata pointer the new string data.  Need not be nul terminated.
+   * @param len the length of the new data
+   * @stable ICU 4.8
+   */
+  void set(const char* xdata, int32_t len) { ptr_ = xdata; length_ = len; }
+
+  /**
+   * Reset the stringpiece to refer to new data.
+   * @param str a pointer to a NUL-terminated string. 
+   * @stable ICU 4.8
+   */
+  void set(const char* str);
+
+#if defined(__cpp_char8_t) || defined(U_IN_DOXYGEN)
+  /**
+   * Resets the stringpiece to refer to new data.
+   * @param xdata pointer the new string data. Need not be NUL-terminated.
+   * @param len the length of the new data
+   * @stable ICU 67
+   */
+  inline void set(const char8_t* xdata, int32_t len) {
+      set(reinterpret_cast<const char*>(xdata), len);
+  }
+
+  /**
+   * Resets the stringpiece to refer to new data.
+   * @param str a pointer to a NUL-terminated string.
+   * @stable ICU 67
+   */
+  inline void set(const char8_t* str) {
+      set(reinterpret_cast<const char*>(str));
+  }
+#endif
+
+  /**
+   * Removes the first n string units.
+   * @param n prefix length, must be non-negative and <=length()
+   * @stable ICU 4.2
+   */
+  void remove_prefix(int32_t n) {
+    if (n >= 0) {
+      if (n > length_) {
+        n = length_;
+      }
+      ptr_ += n;
+      length_ -= n;
+    }
+  }
+
+  /**
+   * Removes the last n string units.
+   * @param n suffix length, must be non-negative and <=length()
+   * @stable ICU 4.2
+   */
+  void remove_suffix(int32_t n) {
+    if (n >= 0) {
+      if (n <= length_) {
+        length_ -= n;
+      } else {
+        length_ = 0;
+      }
+    }
+  }
+
+  /**
+   * Searches the StringPiece for the given search string (needle);
+   * @param needle The string for which to search.
+   * @param offset Where to start searching within this string (haystack).
+   * @return The offset of needle in haystack, or -1 if not found.
+   * @stable ICU 67
+   */
+  int32_t find(StringPiece needle, int32_t offset);
+
+  /**
+   * Compares this StringPiece with the other StringPiece, with semantics
+   * similar to std::string::compare().
+   * @param other The string to compare to.
+   * @return below zero if this < other; above zero if this > other; 0 if this == other.
+   * @stable ICU 67
+   */
+  int32_t compare(StringPiece other);
+
+  /**
+   * Maximum integer, used as a default value for substring methods.
+   * @stable ICU 4.2
+   */
+  static const int32_t npos; // = 0x7fffffff;
+
+  /**
+   * Returns a substring of this StringPiece.
+   * @param pos start position; must be non-negative and <= length().
+   * @param len length of the substring;
+   *            must be non-negative and will be pinned to at most length() - pos.
+   * @return the substring StringPiece
+   * @stable ICU 4.2
+   */
+  StringPiece substr(int32_t pos, int32_t len = npos) const {
+    return StringPiece(*this, pos, len);
+  }
+};
+
+/**
+ * Global operator == for StringPiece
+ * @param x The first StringPiece to compare.
+ * @param y The second StringPiece to compare.
+ * @return true if the string data is equal
+ * @stable ICU 4.8
+ */
+U_EXPORT UBool U_EXPORT2 
+operator==(const StringPiece& x, const StringPiece& y);
+
+/**
+ * Global operator != for StringPiece
+ * @param x The first StringPiece to compare.
+ * @param y The second StringPiece to compare.
+ * @return true if the string data is not equal
+ * @stable ICU 4.8
+ */
+inline bool operator!=(const StringPiece& x, const StringPiece& y) {
+  return !(x == y);
+}
+
+U_NAMESPACE_END
+
+#endif /* U_SHOW_CPLUSPLUS_API */
+
+#endif  // __STRINGPIECE_H__

thirdparty/icu4c/common/unicode/stringtriebuilder.h

@@ -0,0 +1,426 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+/*
+*******************************************************************************
+*   Copyright (C) 2010-2012,2014, International Business Machines
+*   Corporation and others.  All Rights Reserved.
+*******************************************************************************
+*   file name:  stringtriebuilder.h
+*   encoding:   UTF-8
+*   tab size:   8 (not used)
+*   indentation:4
+*
+*   created on: 2010dec24
+*   created by: Markus W. Scherer
+*/
+
+#ifndef __STRINGTRIEBUILDER_H__
+#define __STRINGTRIEBUILDER_H__
+
+#include "unicode/utypes.h"
+
+#if U_SHOW_CPLUSPLUS_API
+
+#include "unicode/uobject.h"
+
+/**
+ * \file
+ * \brief C++ API: Builder API for trie builders
+ */
+
+// Forward declaration.
+/// \cond
+struct UHashtable;
+typedef struct UHashtable UHashtable;
+/// \endcond
+
+/**
+ * Build options for BytesTrieBuilder and CharsTrieBuilder.
+ * @stable ICU 4.8
+ */
+enum UStringTrieBuildOption {
+    /**
+     * Builds a trie quickly.
+     * @stable ICU 4.8
+     */
+    USTRINGTRIE_BUILD_FAST,
+    /**
+     * Builds a trie more slowly, attempting to generate
+     * a shorter but equivalent serialization.
+     * This build option also uses more memory.
+     *
+     * This option can be effective when many integer values are the same
+     * and string/byte sequence suffixes can be shared.
+     * Runtime speed is not expected to improve.
+     * @stable ICU 4.8
+     */
+    USTRINGTRIE_BUILD_SMALL
+};
+
+U_NAMESPACE_BEGIN
+
+/**
+ * Base class for string trie builder classes.
+ *
+ * This class is not intended for public subclassing.
+ * @stable ICU 4.8
+ */
+class U_COMMON_API StringTrieBuilder : public UObject {
+public:
+#ifndef U_HIDE_INTERNAL_API
+    /** @internal */
+    static int32_t hashNode(const void *node);
+    /** @internal */
+    static UBool equalNodes(const void *left, const void *right);
+#endif  /* U_HIDE_INTERNAL_API */
+
+protected:
+    // Do not enclose the protected default constructor with #ifndef U_HIDE_INTERNAL_API
+    // or else the compiler will create a public default constructor.
+    /** @internal */
+    StringTrieBuilder();
+    /** @internal */
+    virtual ~StringTrieBuilder();
+
+#ifndef U_HIDE_INTERNAL_API
+    /** @internal */
+    void createCompactBuilder(int32_t sizeGuess, UErrorCode &errorCode);
+    /** @internal */
+    void deleteCompactBuilder();
+
+    /** @internal */
+    void build(UStringTrieBuildOption buildOption, int32_t elementsLength, UErrorCode &errorCode);
+
+    /** @internal */
+    int32_t writeNode(int32_t start, int32_t limit, int32_t unitIndex);
+    /** @internal */
+    int32_t writeBranchSubNode(int32_t start, int32_t limit, int32_t unitIndex, int32_t length);
+#endif  /* U_HIDE_INTERNAL_API */
+
+    class Node;
+
+#ifndef U_HIDE_INTERNAL_API
+    /** @internal */
+    Node *makeNode(int32_t start, int32_t limit, int32_t unitIndex, UErrorCode &errorCode);
+    /** @internal */
+    Node *makeBranchSubNode(int32_t start, int32_t limit, int32_t unitIndex,
+                            int32_t length, UErrorCode &errorCode);
+#endif  /* U_HIDE_INTERNAL_API */
+
+    /** @internal */
+    virtual int32_t getElementStringLength(int32_t i) const = 0;
+    /** @internal */
+    virtual char16_t getElementUnit(int32_t i, int32_t unitIndex) const = 0;
+    /** @internal */
+    virtual int32_t getElementValue(int32_t i) const = 0;
+
+    // Finds the first unit index after this one where
+    // the first and last element have different units again.
+    /** @internal */
+    virtual int32_t getLimitOfLinearMatch(int32_t first, int32_t last, int32_t unitIndex) const = 0;
+
+    // Number of different units at unitIndex.
+    /** @internal */
+    virtual int32_t countElementUnits(int32_t start, int32_t limit, int32_t unitIndex) const = 0;
+    /** @internal */
+    virtual int32_t skipElementsBySomeUnits(int32_t i, int32_t unitIndex, int32_t count) const = 0;
+    /** @internal */
+    virtual int32_t indexOfElementWithNextUnit(int32_t i, int32_t unitIndex, char16_t unit) const = 0;
+
+    /** @internal */
+    virtual UBool matchNodesCanHaveValues() const = 0;
+
+    /** @internal */
+    virtual int32_t getMaxBranchLinearSubNodeLength() const = 0;
+    /** @internal */
+    virtual int32_t getMinLinearMatch() const = 0;
+    /** @internal */
+    virtual int32_t getMaxLinearMatchLength() const = 0;
+
+#ifndef U_HIDE_INTERNAL_API
+    // max(BytesTrie::kMaxBranchLinearSubNodeLength, UCharsTrie::kMaxBranchLinearSubNodeLength).
+    /** @internal */
+    static const int32_t kMaxBranchLinearSubNodeLength=5;
+
+    // Maximum number of nested split-branch levels for a branch on all 2^16 possible char16_t units.
+    // log2(2^16/kMaxBranchLinearSubNodeLength) rounded up.
+    /** @internal */
+    static const int32_t kMaxSplitBranchLevels=14;
+
+    /**
+     * Makes sure that there is only one unique node registered that is
+     * equivalent to newNode.
+     * @param newNode Input node. The builder takes ownership.
+     * @param errorCode ICU in/out UErrorCode.
+                        Set to U_MEMORY_ALLOCATION_ERROR if it was success but newNode==nullptr.
+     * @return newNode if it is the first of its kind, or
+     *         an equivalent node if newNode is a duplicate.
+     * @internal
+     */
+    Node *registerNode(Node *newNode, UErrorCode &errorCode);
+    /**
+     * Makes sure that there is only one unique FinalValueNode registered
+     * with this value.
+     * Avoids creating a node if the value is a duplicate.
+     * @param value A final value.
+     * @param errorCode ICU in/out UErrorCode.
+                        Set to U_MEMORY_ALLOCATION_ERROR if it was success but newNode==nullptr.
+     * @return A FinalValueNode with the given value.
+     * @internal
+     */
+    Node *registerFinalValue(int32_t value, UErrorCode &errorCode);
+#endif  /* U_HIDE_INTERNAL_API */
+
+    /*
+     * C++ note:
+     * registerNode() and registerFinalValue() take ownership of their input nodes,
+     * and only return owned nodes.
+     * If they see a failure UErrorCode, they will delete the input node.
+     * If they get a nullptr pointer, they will record a U_MEMORY_ALLOCATION_ERROR.
+     * If there is a failure, they return nullptr.
+     *
+     * nullptr Node pointers can be safely passed into other Nodes because
+     * they call the static Node::hashCode() which checks for a nullptr pointer first.
+     *
+     * Therefore, as long as builder functions register a new node,
+     * they need to check for failures only before explicitly dereferencing
+     * a Node pointer, or before setting a new UErrorCode.
+     */
+
+    // Hash set of nodes, maps from nodes to integer 1.
+    /** @internal */
+    UHashtable *nodes;
+
+    // Do not conditionalize the following with #ifndef U_HIDE_INTERNAL_API,
+    // it is needed for layout of other objects.
+    /**
+     * @internal
+     * \cond
+     */
+    class Node : public UObject {
+    public:
+        Node(int32_t initialHash) : hash(initialHash), offset(0) {}
+        inline int32_t hashCode() const { return hash; }
+        // Handles node==nullptr.
+        static inline int32_t hashCode(const Node *node) { return node==nullptr ? 0 : node->hashCode(); }
+        // Base class operator==() compares the actual class types.
+        virtual bool operator==(const Node &other) const;
+        inline bool operator!=(const Node &other) const { return !operator==(other); }
+        /**
+         * Traverses the Node graph and numbers branch edges, with rightmost edges first.
+         * This is to avoid writing a duplicate node twice.
+         *
+         * Branch nodes in this trie data structure are not symmetric.
+         * Most branch edges "jump" to other nodes but the rightmost branch edges
+         * just continue without a jump.
+         * Therefore, write() must write the rightmost branch edge last
+         * (trie units are written backwards), and must write it at that point even if
+         * it is a duplicate of a node previously written elsewhere.
+         *
+         * This function visits and marks right branch edges first.
+         * Edges are numbered with increasingly negative values because we share the
+         * offset field which gets positive values when nodes are written.
+         * A branch edge also remembers the first number for any of its edges.
+         *
+         * When a further-left branch edge has a number in the range of the rightmost
+         * edge's numbers, then it will be written as part of the required right edge
+         * and we can avoid writing it first.
+         *
+         * After root.markRightEdgesFirst(-1) the offsets of all nodes are negative
+         * edge numbers.
+         *
+         * @param edgeNumber The first edge number for this node and its sub-nodes.
+         * @return An edge number that is at least the maximum-negative
+         *         of the input edge number and the numbers of this node and all of its sub-nodes.
+         */
+        virtual int32_t markRightEdgesFirst(int32_t edgeNumber);
+        // write() must set the offset to a positive value.
+        virtual void write(StringTrieBuilder &builder) = 0;
+        // See markRightEdgesFirst.
+        inline void writeUnlessInsideRightEdge(int32_t firstRight, int32_t lastRight,
+                                               StringTrieBuilder &builder) {
+            // Note: Edge numbers are negative, lastRight<=firstRight.
+            // If offset>0 then this node and its sub-nodes have been written already
+            // and we need not write them again.
+            // If this node is part of the unwritten right branch edge,
+            // then we wait until that is written.
+            if(offset<0 && (offset<lastRight || firstRight<offset)) {
+                write(builder);
+            }
+        }
+        inline int32_t getOffset() const { return offset; }
+    protected:
+        int32_t hash;
+        int32_t offset;
+    };
+
+#ifndef U_HIDE_INTERNAL_API
+    // This class should not be overridden because
+    // registerFinalValue() compares a stack-allocated FinalValueNode
+    // (stack-allocated so that we don't unnecessarily create lots of duplicate nodes)
+    // with the input node, and the
+    // !Node::operator==(other) used inside FinalValueNode::operator==(other)
+    // will be false if the typeid's are different.
+    /** @internal */
+    class FinalValueNode : public Node {
+    public:
+        FinalValueNode(int32_t v) : Node(0x111111u*37u+v), value(v) {}
+        virtual bool operator==(const Node &other) const override;
+        virtual void write(StringTrieBuilder &builder) override;
+    protected:
+        int32_t value;
+    };
+#endif  /* U_HIDE_INTERNAL_API */
+
+    // Do not conditionalize the following with #ifndef U_HIDE_INTERNAL_API,
+    // it is needed for layout of other objects.
+    /**
+     * @internal 
+     */
+    class ValueNode : public Node {
+    public:
+        ValueNode(int32_t initialHash) : Node(initialHash), hasValue(false), value(0) {}
+        virtual bool operator==(const Node &other) const override;
+        void setValue(int32_t v) {
+            hasValue=true;
+            value=v;
+            hash=hash*37u+v;
+        }
+    protected:
+        UBool hasValue;
+        int32_t value;
+    };
+
+#ifndef U_HIDE_INTERNAL_API
+    /** 
+     * @internal 
+     */
+    class IntermediateValueNode : public ValueNode {
+    public:
+        IntermediateValueNode(int32_t v, Node *nextNode)
+                : ValueNode(0x222222u*37u+hashCode(nextNode)), next(nextNode) { setValue(v); }
+        virtual bool operator==(const Node &other) const override;
+        virtual int32_t markRightEdgesFirst(int32_t edgeNumber) override;
+        virtual void write(StringTrieBuilder &builder) override;
+    protected:
+        Node *next;
+    };
+#endif  /* U_HIDE_INTERNAL_API */
+
+    // Do not conditionalize the following with #ifndef U_HIDE_INTERNAL_API,
+    // it is needed for layout of other objects.
+    /**
+     * @internal 
+     */
+    class LinearMatchNode : public ValueNode {
+    public:
+        LinearMatchNode(int32_t len, Node *nextNode)
+                : ValueNode((0x333333u*37u+len)*37u+hashCode(nextNode)),
+                  length(len), next(nextNode) {}
+        virtual bool operator==(const Node &other) const override;
+        virtual int32_t markRightEdgesFirst(int32_t edgeNumber) override;
+    protected:
+        int32_t length;
+        Node *next;
+    };
+
+#ifndef U_HIDE_INTERNAL_API
+    /**
+     * @internal 
+     */
+    class BranchNode : public Node {
+    public:
+        BranchNode(int32_t initialHash) : Node(initialHash) {}
+    protected:
+        int32_t firstEdgeNumber;
+    };
+
+    /**
+     * @internal 
+     */
+    class ListBranchNode : public BranchNode {
+    public:
+        ListBranchNode() : BranchNode(0x444444), length(0) {}
+        virtual bool operator==(const Node &other) const override;
+        virtual int32_t markRightEdgesFirst(int32_t edgeNumber) override;
+        virtual void write(StringTrieBuilder &builder) override;
+        // Adds a unit with a final value.
+        void add(int32_t c, int32_t value) {
+            units[length] = static_cast<char16_t>(c);
+            equal[length]=nullptr;
+            values[length]=value;
+            ++length;
+            hash=(hash*37u+c)*37u+value;
+        }
+        // Adds a unit which leads to another match node.
+        void add(int32_t c, Node *node) {
+            units[length] = static_cast<char16_t>(c);
+            equal[length]=node;
+            values[length]=0;
+            ++length;
+            hash=(hash*37u+c)*37u+hashCode(node);
+        }
+    protected:
+        Node *equal[kMaxBranchLinearSubNodeLength];  // nullptr means "has final value".
+        int32_t length;
+        int32_t values[kMaxBranchLinearSubNodeLength];
+        char16_t units[kMaxBranchLinearSubNodeLength];
+    };
+
+    /**
+     * @internal 
+     */
+    class SplitBranchNode : public BranchNode {
+    public:
+        SplitBranchNode(char16_t middleUnit, Node *lessThanNode, Node *greaterOrEqualNode)
+                : BranchNode(((0x555555u*37u+middleUnit)*37u+
+                              hashCode(lessThanNode))*37u+hashCode(greaterOrEqualNode)),
+                  unit(middleUnit), lessThan(lessThanNode), greaterOrEqual(greaterOrEqualNode) {}
+        virtual bool operator==(const Node &other) const override;
+        virtual int32_t markRightEdgesFirst(int32_t edgeNumber) override;
+        virtual void write(StringTrieBuilder &builder) override;
+    protected:
+        char16_t unit;
+        Node *lessThan;
+        Node *greaterOrEqual;
+    };
+
+    // Branch head node, for writing the actual node lead unit.
+    /** @internal */
+    class BranchHeadNode : public ValueNode {
+    public:
+        BranchHeadNode(int32_t len, Node *subNode)
+                : ValueNode((0x666666u*37u+len)*37u+hashCode(subNode)),
+                  length(len), next(subNode) {}
+        virtual bool operator==(const Node &other) const override;
+        virtual int32_t markRightEdgesFirst(int32_t edgeNumber) override;
+        virtual void write(StringTrieBuilder &builder) override;
+    protected:
+        int32_t length;
+        Node *next;  // A branch sub-node.
+    };
+
+#endif  /* U_HIDE_INTERNAL_API */
+    /// \endcond
+
+    /** @internal */
+    virtual Node *createLinearMatchNode(int32_t i, int32_t unitIndex, int32_t length,
+                                        Node *nextNode) const = 0;
+
+    /** @internal */
+    virtual int32_t write(int32_t unit) = 0;
+    /** @internal */
+    virtual int32_t writeElementUnits(int32_t i, int32_t unitIndex, int32_t length) = 0;
+    /** @internal */
+    virtual int32_t writeValueAndFinal(int32_t i, UBool isFinal) = 0;
+    /** @internal */
+    virtual int32_t writeValueAndType(UBool hasValue, int32_t value, int32_t node) = 0;
+    /** @internal */
+    virtual int32_t writeDeltaTo(int32_t jumpTarget) = 0;
+};
+
+U_NAMESPACE_END
+
+#endif /* U_SHOW_CPLUSPLUS_API */
+
+#endif  // __STRINGTRIEBUILDER_H__

thirdparty/icu4c/common/unicode/symtable.h

@@ -0,0 +1,119 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+/*
+**********************************************************************
+*   Copyright (c) 2000-2005, International Business Machines
+*   Corporation and others.  All Rights Reserved.
+**********************************************************************
+*   Date        Name        Description
+*   02/04/00    aliu        Creation.
+**********************************************************************
+*/
+#ifndef SYMTABLE_H
+#define SYMTABLE_H
+
+#include "unicode/utypes.h"
+
+#if U_SHOW_CPLUSPLUS_API
+
+#include "unicode/uobject.h"
+
+/**
+ * \file 
+ * \brief C++ API: An interface that defines both lookup protocol and parsing of
+ * symbolic names.
+ */
+ 
+U_NAMESPACE_BEGIN
+
+class ParsePosition;
+class UnicodeFunctor;
+class UnicodeSet;
+class UnicodeString;
+
+/**
+ * An interface that defines both lookup protocol and parsing of
+ * symbolic names.
+ *
+ * <p>A symbol table maintains two kinds of mappings.  The first is
+ * between symbolic names and their values.  For example, if the
+ * variable with the name "start" is set to the value "alpha"
+ * (perhaps, though not necessarily, through an expression such as
+ * "$start=alpha"), then the call lookup("start") will return the
+ * char[] array ['a', 'l', 'p', 'h', 'a'].
+ *
+ * <p>The second kind of mapping is between character values and
+ * UnicodeMatcher objects.  This is used by RuleBasedTransliterator,
+ * which uses characters in the private use area to represent objects
+ * such as UnicodeSets.  If U+E015 is mapped to the UnicodeSet [a-z],
+ * then lookupMatcher(0xE015) will return the UnicodeSet [a-z].
+ *
+ * <p>Finally, a symbol table defines parsing behavior for symbolic
+ * names.  All symbolic names start with the SYMBOL_REF character.
+ * When a parser encounters this character, it calls parseReference()
+ * with the position immediately following the SYMBOL_REF.  The symbol
+ * table parses the name, if there is one, and returns it.
+ *
+ * @stable ICU 2.8
+ */
+class U_COMMON_API SymbolTable /* not : public UObject because this is an interface/mixin class */ {
+public:
+
+    /**
+     * The character preceding a symbol reference name.
+     * @stable ICU 2.8
+     */
+    enum { SYMBOL_REF = 0x0024 /*$*/ };
+
+    /**
+     * Destructor.
+     * @stable ICU 2.8
+     */
+    virtual ~SymbolTable();
+
+    /**
+     * Lookup the characters associated with this string and return it.
+     * Return <tt>nullptr</tt> if no such name exists.  The resultant
+     * string may have length zero.
+     * @param s the symbolic name to lookup
+     * @return a string containing the name's value, or <tt>nullptr</tt> if
+     * there is no mapping for s.
+     * @stable ICU 2.8
+     */
+    virtual const UnicodeString* lookup(const UnicodeString& s) const = 0;
+
+    /**
+     * Lookup the UnicodeMatcher associated with the given character, and
+     * return it.  Return <tt>nullptr</tt> if not found.
+     * @param ch a 32-bit code point from 0 to 0x10FFFF inclusive.
+     * @return the UnicodeMatcher object represented by the given
+     * character, or nullptr if there is no mapping for ch.
+     * @stable ICU 2.8
+     */
+    virtual const UnicodeFunctor* lookupMatcher(UChar32 ch) const = 0;
+
+    /**
+     * Parse a symbol reference name from the given string, starting
+     * at the given position.  If no valid symbol reference name is
+     * found, return the empty string and leave pos unchanged.  That is, if the
+     * character at pos cannot start a name, or if pos is at or after
+     * text.length(), then return an empty string.  This indicates an
+     * isolated SYMBOL_REF character.
+     * @param text the text to parse for the name
+     * @param pos on entry, the index of the first character to parse.
+     * This is the character following the SYMBOL_REF character.  On
+     * exit, the index after the last parsed character.  If the parse
+     * failed, pos is unchanged on exit.
+     * @param limit the index after the last character to be parsed.
+     * @return the parsed name, or an empty string if there is no
+     * valid symbolic name at the given position.
+     * @stable ICU 2.8
+     */
+    virtual UnicodeString parseReference(const UnicodeString& text,
+                                         ParsePosition& pos, int32_t limit) const = 0;
+};
+U_NAMESPACE_END
+
+#endif /* U_SHOW_CPLUSPLUS_API */
+
+#endif

thirdparty/icu4c/common/unicode/ubiditransform.h

@@ -0,0 +1,326 @@
+/*
+******************************************************************************
+*
+* © 2016 and later: Unicode, Inc. and others.
+* License & terms of use: http://www.unicode.org/copyright.html
+*
+******************************************************************************
+*   file name:  ubiditransform.h
+*   encoding:   UTF-8
+*   tab size:   8 (not used)
+*   indentation:4
+*
+*   created on: 2016jul24
+*   created by: Lina Kemmel
+*
+*/
+
+#ifndef UBIDITRANSFORM_H
+#define UBIDITRANSFORM_H
+
+#include "unicode/utypes.h"
+#include "unicode/ubidi.h"
+#include "unicode/uchar.h"
+
+#if U_SHOW_CPLUSPLUS_API
+#include "unicode/localpointer.h"
+#endif   // U_SHOW_CPLUSPLUS_API
+
+/**
+ * \file
+ * \brief C API: Bidi Transformations
+ */
+
+/**
+ * `UBiDiOrder` indicates the order of text.
+ *
+ * This bidi transformation engine supports all possible combinations (4 in
+ * total) of input and output text order:
+ *
+ *   - <logical input, visual output>: unless the output direction is RTL, this
+ *     corresponds to a normal operation of the Bidi algorithm as described in the
+ *     Unicode Technical Report and implemented by `UBiDi` when the
+ *     reordering mode is set to `UBIDI_REORDER_DEFAULT`. Visual RTL
+ *     mode is not supported by `UBiDi` and is accomplished through
+ *     reversing a visual LTR string,
+ *
+ *   - <visual input, logical output>: unless the input direction is RTL, this
+ *     corresponds to an "inverse bidi algorithm" in `UBiDi` with the
+ *     reordering mode set to `UBIDI_REORDER_INVERSE_LIKE_DIRECT`.
+ *     Visual RTL mode is not not supported by `UBiDi` and is
+ *     accomplished through reversing a visual LTR string,
+ *
+ *   - <logical input, logical output>: if the input and output base directions
+ *     mismatch, this corresponds to the `UBiDi` implementation with the
+ *     reordering mode set to `UBIDI_REORDER_RUNS_ONLY`; and if the
+ *     input and output base directions are identical, the transformation engine
+ *     will only handle character mirroring and Arabic shaping operations without
+ *     reordering,
+ *
+ *   - <visual input, visual output>: this reordering mode is not supported by
+ *     the `UBiDi` engine; it implies character mirroring, Arabic
+ *     shaping, and - if the input/output base directions mismatch -  string
+ *     reverse operations.
+ * @see ubidi_setInverse
+ * @see ubidi_setReorderingMode
+ * @see UBIDI_REORDER_DEFAULT
+ * @see UBIDI_REORDER_INVERSE_LIKE_DIRECT
+ * @see UBIDI_REORDER_RUNS_ONLY
+ * @stable ICU 58
+ */
+typedef enum {
+    /** 0: Constant indicating a logical order.
+      * This is the default for input text.
+      * @stable ICU 58
+      */
+    UBIDI_LOGICAL = 0,
+    /** 1: Constant indicating a visual order.
+      * This is a default for output text.
+      * @stable ICU 58
+      */
+    UBIDI_VISUAL
+} UBiDiOrder;
+
+/**
+ * <code>UBiDiMirroring</code> indicates whether or not characters with the
+ * "mirrored" property in RTL runs should be replaced with their mirror-image
+ * counterparts.
+ * @see UBIDI_DO_MIRRORING
+ * @see ubidi_setReorderingOptions
+ * @see ubidi_writeReordered
+ * @see ubidi_writeReverse
+ * @stable ICU 58
+ */
+typedef enum {
+    /** 0: Constant indicating that character mirroring should not be
+      * performed.
+      * This is the default.
+      * @stable ICU 58
+      */
+    UBIDI_MIRRORING_OFF = 0,
+    /** 1: Constant indicating that character mirroring should be performed.
+      * This corresponds to calling <code>ubidi_writeReordered</code> or
+      * <code>ubidi_writeReverse</code> with the
+      * <code>UBIDI_DO_MIRRORING</code> option bit set.
+      * @stable ICU 58
+      */
+    UBIDI_MIRRORING_ON
+} UBiDiMirroring;
+
+/**
+ * Forward declaration of the <code>UBiDiTransform</code> structure that stores
+ * information used by the layout transformation engine.
+ * @stable ICU 58
+ */
+typedef struct UBiDiTransform UBiDiTransform;
+
+/**
+ * Performs transformation of text from the bidi layout defined by the input
+ * ordering scheme to the bidi layout defined by the output ordering scheme,
+ * and applies character mirroring and Arabic shaping operations.<p>
+ * In terms of <code>UBiDi</code>, such a transformation implies:
+ * <ul>
+ * <li>calling <code>ubidi_setReorderingMode</code> as needed (when the
+ * reordering mode is other than normal),</li>
+ * <li>calling <code>ubidi_setInverse</code> as needed (when text should be
+ * transformed from a visual to a logical form),</li>
+ * <li>resolving embedding levels of each character in the input text by
+ * calling <code>ubidi_setPara</code>,</li>
+ * <li>reordering the characters based on the computed embedding levels, also
+ * performing character mirroring as needed, and streaming the result to the
+ * output, by calling <code>ubidi_writeReordered</code>,</li>
+ * <li>performing Arabic digit and letter shaping on the output text by calling
+ * <code>u_shapeArabic</code>.</li>
+ * </ul>
+ * An "ordering scheme" encompasses the base direction and the order of text,
+ * and these characteristics must be defined by the caller for both input and
+ * output explicitly .<p>
+ * There are 36 possible combinations of <input, output> ordering schemes,
+ * which are partially supported by <code>UBiDi</code> already. Examples of the
+ * currently supported combinations:
+ * <ul>
+ * <li><Logical LTR, Visual LTR>: this is equivalent to calling
+ * <code>ubidi_setPara</code> with <code>paraLevel == UBIDI_LTR</code>,</li>
+ * <li><Logical RTL, Visual LTR>: this is equivalent to calling
+ * <code>ubidi_setPara</code> with <code>paraLevel == UBIDI_RTL</code>,</li>
+ * <li><Logical Default ("Auto") LTR, Visual LTR>: this is equivalent to
+ * calling <code>ubidi_setPara</code> with
+ * <code>paraLevel == UBIDI_DEFAULT_LTR</code>,</li>
+ * <li><Logical Default ("Auto") RTL, Visual LTR>: this is equivalent to
+ * calling <code>ubidi_setPara</code> with
+ * <code>paraLevel == UBIDI_DEFAULT_RTL</code>,</li>
+ * <li><Visual LTR, Logical LTR>: this is equivalent to
+ * calling <code>ubidi_setInverse(UBiDi*, true)</code> and then
+ * <code>ubidi_setPara</code> with <code>paraLevel == UBIDI_LTR</code>,</li>
+ * <li><Visual LTR, Logical RTL>: this is equivalent to
+ * calling <code>ubidi_setInverse(UBiDi*, true)</code> and then
+ * <code>ubidi_setPara</code> with <code>paraLevel == UBIDI_RTL</code>.</li>
+ * </ul>
+ * All combinations that involve the Visual RTL scheme are unsupported by
+ * <code>UBiDi</code>, for instance:
+ * <ul>
+ * <li><Logical LTR, Visual RTL>,</li>
+ * <li><Visual RTL, Logical RTL>.</li>
+ * </ul>
+ * <p>Example of usage of the transformation engine:<br>
+ * <pre>
+ * \code
+ * UChar text1[] = {'a', 'b', 'c', 0x0625, '1', 0};
+ * UChar text2[] = {'a', 'b', 'c', 0x0625, '1', 0};
+ * UErrorCode errorCode = U_ZERO_ERROR;
+ * // Run a transformation.
+ * ubiditransform_transform(pBidiTransform,
+ *          text1, -1, text2, -1,
+ *          UBIDI_LTR, UBIDI_VISUAL,
+ *          UBIDI_RTL, UBIDI_LOGICAL,
+ *          UBIDI_MIRRORING_OFF,
+ *          U_SHAPE_DIGITS_AN2EN | U_SHAPE_DIGIT_TYPE_AN_EXTENDED,
+ *          &errorCode);
+ * // Do something with text2.
+ *  text2[4] = '2';
+ * // Run a reverse transformation.
+ * ubiditransform_transform(pBidiTransform,
+ *          text2, -1, text1, -1,
+ *          UBIDI_RTL, UBIDI_LOGICAL,
+ *          UBIDI_LTR, UBIDI_VISUAL,
+ *          UBIDI_MIRRORING_OFF,
+ *          U_SHAPE_DIGITS_EN2AN | U_SHAPE_DIGIT_TYPE_AN_EXTENDED,
+ *          &errorCode);
+ *\endcode
+ * </pre>
+ * </p>
+ *
+ * @param pBiDiTransform A pointer to a <code>UBiDiTransform</code> object
+ *        allocated with <code>ubiditransform_open()</code> or
+ *        <code>NULL</code>.<p>
+ *        This object serves for one-time setup to amortize initialization
+ *        overheads. Use of this object is not thread-safe. All other threads
+ *        should allocate a new <code>UBiDiTransform</code> object by calling
+ *        <code>ubiditransform_open()</code> before using it. Alternatively,
+ *        a caller can set this parameter to <code>NULL</code>, in which case
+ *        the object will be allocated by the engine on the fly.</p>
+ * @param src A pointer to the text that the Bidi layout transformations will
+ *        be performed on.
+ *        <p><strong>Note:</strong> the text must be (at least)
+ *        <code>srcLength</code> long.</p>
+ * @param srcLength The length of the text, in number of UChars. If
+ *        <code>length == -1</code> then the text must be zero-terminated.
+ * @param dest A pointer to where the processed text is to be copied.
+ * @param destSize The size of the <code>dest</code> buffer, in number of
+ *        UChars. If the <code>U_SHAPE_LETTERS_UNSHAPE</code> option is set,
+ *        then the destination length could be as large as
+ *        <code>srcLength * 2</code>. Otherwise, the destination length will
+ *        not exceed <code>srcLength</code>. If the caller reserves the last
+ *        position for zero-termination, it should be excluded from
+ *        <code>destSize</code>.
+ *        <p><code>destSize == -1</code> is allowed and makes sense when
+ *        <code>dest</code> was holds some meaningful value, e.g. that of
+ *        <code>src</code>. In this case <code>dest</code> must be
+ *        zero-terminated.</p>
+ * @param inParaLevel A base embedding level of the input as defined in
+ *        <code>ubidi_setPara</code> documentation for the
+ *        <code>paraLevel</code> parameter.
+ * @param inOrder An order of the input, which can be one of the
+ *        <code>UBiDiOrder</code> values.
+ * @param outParaLevel A base embedding level of the output as defined in
+ *        <code>ubidi_setPara</code> documentation for the
+ *        <code>paraLevel</code> parameter.
+ * @param outOrder An order of the output, which can be one of the
+ *        <code>UBiDiOrder</code> values.
+ * @param doMirroring Indicates whether or not to perform character mirroring,
+ *        and can accept one of the <code>UBiDiMirroring</code> values.
+ * @param shapingOptions Arabic digit and letter shaping options defined in the
+ *        ushape.h documentation.
+ *        <p><strong>Note:</strong> Direction indicator options are computed by
+ *        the transformation engine based on the effective ordering schemes, so
+ *        user-defined direction indicators will be ignored.</p>
+ * @param pErrorCode A pointer to an error code value.
+ *
+ * @return The destination length, i.e. the number of UChars written to
+ *         <code>dest</code>. If the transformation fails, the return value
+ *         will be 0 (and the error code will be written to
+ *         <code>pErrorCode</code>).
+ *
+ * @see UBiDiLevel
+ * @see UBiDiOrder
+ * @see UBiDiMirroring
+ * @see ubidi_setPara
+ * @see u_shapeArabic
+ * @stable ICU 58
+ */
+U_CAPI uint32_t U_EXPORT2
+ubiditransform_transform(UBiDiTransform *pBiDiTransform,
+            const UChar *src, int32_t srcLength,
+            UChar *dest, int32_t destSize,
+            UBiDiLevel inParaLevel, UBiDiOrder inOrder,
+            UBiDiLevel outParaLevel, UBiDiOrder outOrder,
+            UBiDiMirroring doMirroring, uint32_t shapingOptions,
+            UErrorCode *pErrorCode);
+
+/**
+ * Allocates a <code>UBiDiTransform</code> object. This object can be reused,
+ * e.g. with different ordering schemes, mirroring or shaping options.<p>
+ * <strong>Note:</strong>The object can only be reused in the same thread.
+ * All other threads should allocate a new <code>UBiDiTransform</code> object
+ * before using it.<p>
+ * Example of usage:<p>
+ * <pre>
+ * \code
+ * UErrorCode errorCode = U_ZERO_ERROR;
+ * // Open a new UBiDiTransform.
+ * UBiDiTransform* transform = ubiditransform_open(&errorCode);
+ * // Run a transformation.
+ * ubiditransform_transform(transform,
+ *          text1, -1, text2, -1,
+ *          UBIDI_RTL, UBIDI_LOGICAL,
+ *          UBIDI_LTR, UBIDI_VISUAL,
+ *          UBIDI_MIRRORING_ON,
+ *          U_SHAPE_DIGITS_EN2AN,
+ *          &errorCode);
+ * // Do something with the output text and invoke another transformation using
+ * //   that text as input.
+ * ubiditransform_transform(transform,
+ *          text2, -1, text3, -1,
+ *          UBIDI_LTR, UBIDI_VISUAL,
+ *          UBIDI_RTL, UBIDI_VISUAL,
+ *          UBIDI_MIRRORING_ON,
+ *          0, &errorCode);
+ *\endcode
+ * </pre>
+ * <p>
+ * The <code>UBiDiTransform</code> object must be deallocated by calling
+ * <code>ubiditransform_close()</code>.
+ *
+ * @return An empty <code>UBiDiTransform</code> object.
+ * @stable ICU 58
+ */
+U_CAPI UBiDiTransform* U_EXPORT2
+ubiditransform_open(UErrorCode *pErrorCode);
+
+/**
+ * Deallocates the given <code>UBiDiTransform</code> object.
+ * @stable ICU 58
+ */
+U_CAPI void U_EXPORT2
+ubiditransform_close(UBiDiTransform *pBidiTransform);
+
+#if U_SHOW_CPLUSPLUS_API
+
+U_NAMESPACE_BEGIN
+
+/**
+ * \class LocalUBiDiTransformPointer
+ * "Smart pointer" class, closes a UBiDiTransform via ubiditransform_close().
+ * For most methods see the LocalPointerBase base class.
+ *
+ * @see LocalPointerBase
+ * @see LocalPointer
+ * @stable ICU 58
+ */
+U_DEFINE_LOCAL_OPEN_POINTER(LocalUBiDiTransformPointer, UBiDiTransform, ubiditransform_close);
+
+U_NAMESPACE_END
+
+#endif
+
+#endif

thirdparty/icu4c/common/unicode/ubrk.h

@@ -0,0 +1,647 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+/*
+******************************************************************************
+* Copyright (C) 1996-2015, International Business Machines Corporation and others.
+* All Rights Reserved.
+******************************************************************************
+*/
+
+#ifndef UBRK_H
+#define UBRK_H
+
+#include "unicode/utypes.h"
+#include "unicode/uloc.h"
+#include "unicode/utext.h"
+
+#if U_SHOW_CPLUSPLUS_API
+#include "unicode/localpointer.h"
+#endif   // U_SHOW_CPLUSPLUS_API
+
+/**
+ * A text-break iterator.
+ *  For usage in C programs.
+ */
+#ifndef UBRK_TYPEDEF_UBREAK_ITERATOR
+#   define UBRK_TYPEDEF_UBREAK_ITERATOR
+    /**
+     *  Opaque type representing an ICU Break iterator object.
+     *  @stable ICU 2.0
+     */
+    typedef struct UBreakIterator UBreakIterator;
+#endif
+
+#if !UCONFIG_NO_BREAK_ITERATION
+
+#include "unicode/parseerr.h"
+
+/**
+ * \file
+ * \brief C API: BreakIterator
+ *
+ * <h2> BreakIterator C API </h2>
+ *
+ * The BreakIterator C API defines  methods for finding the location
+ * of boundaries in text. Pointer to a UBreakIterator maintain a
+ * current position and scan over text returning the index of characters
+ * where boundaries occur.
+ * <p>
+ * Line boundary analysis determines where a text string can be broken
+ * when line-wrapping. The mechanism correctly handles punctuation and
+ * hyphenated words.
+ * <p>
+ * Note: The locale keyword "lb" can be used to modify line break
+ * behavior according to the CSS level 3 line-break options, see
+ * <http://dev.w3.org/csswg/css-text/#line-breaking>. For example:
+ * "ja@lb=strict", "zh@lb=loose".
+ * <p>
+ * Sentence boundary analysis allows selection with correct
+ * interpretation of periods within numbers and abbreviations, and
+ * trailing punctuation marks such as quotation marks and parentheses.
+ * <p>
+ * Note: The locale keyword "ss" can be used to enable use of
+ * segmentation suppression data (preventing breaks in English after
+ * abbreviations such as "Mr." or "Est.", for example), as follows:
+ * "en@ss=standard".
+ * <p>
+ * Word boundary analysis is used by search and replace functions, as
+ * well as within text editing applications that allow the user to
+ * select words with a double click. Word selection provides correct
+ * interpretation of punctuation marks within and following
+ * words. Characters that are not part of a word, such as symbols or
+ * punctuation marks, have word-breaks on both sides.
+ * <p>
+ * Character boundary analysis identifies the boundaries of
+ * "Extended Grapheme Clusters", which are groupings of codepoints
+ * that should be treated as character-like units for many text operations.
+ * Please see Unicode Standard Annex #29, Unicode Text Segmentation,
+ * http://www.unicode.org/reports/tr29/ for additional information
+ * on grapheme clusters and guidelines on their use.
+ * <p>
+ * Title boundary analysis locates all positions,
+ * typically starts of words, that should be set to Title Case
+ * when title casing the text.
+ * <p>
+ * The text boundary positions are found according to the rules
+ * described in Unicode Standard Annex #29, Text Boundaries, and
+ * Unicode Standard Annex #14, Line Breaking Properties.  These
+ * are available at http://www.unicode.org/reports/tr14/ and
+ * http://www.unicode.org/reports/tr29/.
+ * <p>
+ * In addition to the plain C API defined in this header file, an
+ * object oriented C++ API with equivalent functionality is defined in the
+ * file brkiter.h.
+ * <p>
+ * Code snippets illustrating the use of the Break Iterator APIs
+ * are available in the ICU User Guide,
+ * https://unicode-org.github.io/icu/userguide/boundaryanalysis/
+ * and in the sample program icu/source/samples/break/break.cpp
+ */
+
+/** The possible types of text boundaries.  @stable ICU 2.0 */
+typedef enum UBreakIteratorType {
+  /** Character breaks  @stable ICU 2.0 */
+  UBRK_CHARACTER = 0,
+  /** Word breaks @stable ICU 2.0 */
+  UBRK_WORD = 1,
+  /** Line breaks @stable ICU 2.0 */
+  UBRK_LINE = 2,
+  /** Sentence breaks @stable ICU 2.0 */
+  UBRK_SENTENCE = 3,
+
+#ifndef U_HIDE_DEPRECATED_API
+  /**
+   * Title Case breaks
+   * The iterator created using this type locates title boundaries as described for
+   * Unicode 3.2 only. For Unicode 4.0 and above title boundary iteration,
+   * please use Word Boundary iterator.
+   *
+   * @deprecated ICU 2.8 Use the word break iterator for titlecasing for Unicode 4 and later.
+   */
+  UBRK_TITLE = 4,
+    /**
+     * One more than the highest normal UBreakIteratorType value.
+     * @deprecated ICU 58 The numeric value may change over time, see ICU ticket #12420.
+     */
+    UBRK_COUNT = 5
+#endif  // U_HIDE_DEPRECATED_API
+} UBreakIteratorType;
+
+/** Value indicating all text boundaries have been returned.
+ *  @stable ICU 2.0
+ */
+#define UBRK_DONE ((int32_t) -1)
+
+
+/**
+ *  Enum constants for the word break tags returned by
+ *  getRuleStatus().  A range of values is defined for each category of
+ *  word, to allow for further subdivisions of a category in future releases.
+ *  Applications should check for tag values falling within the range, rather
+ *  than for single individual values.
+ *
+ * The numeric values of all of these constants are stable (will not change).
+ *
+ * @stable ICU 2.2
+*/
+typedef enum UWordBreak {
+    /** Tag value for "words" that do not fit into any of other categories.
+     *  Includes spaces and most punctuation. */
+    UBRK_WORD_NONE           = 0,
+    /** Upper bound for tags for uncategorized words. */
+    UBRK_WORD_NONE_LIMIT     = 100,
+    /** Tag value for words that appear to be numbers, lower limit.    */
+    UBRK_WORD_NUMBER         = 100,
+    /** Tag value for words that appear to be numbers, upper limit.    */
+    UBRK_WORD_NUMBER_LIMIT   = 200,
+    /** Tag value for words that contain letters, excluding
+     *  hiragana, katakana or ideographic characters, lower limit.    */
+    UBRK_WORD_LETTER         = 200,
+    /** Tag value for words containing letters, upper limit  */
+    UBRK_WORD_LETTER_LIMIT   = 300,
+    /** Tag value for words containing kana characters, lower limit */
+    UBRK_WORD_KANA           = 300,
+    /** Tag value for words containing kana characters, upper limit */
+    UBRK_WORD_KANA_LIMIT     = 400,
+    /** Tag value for words containing ideographic characters, lower limit */
+    UBRK_WORD_IDEO           = 400,
+    /** Tag value for words containing ideographic characters, upper limit */
+    UBRK_WORD_IDEO_LIMIT     = 500
+} UWordBreak;
+
+/**
+ *  Enum constants for the line break tags returned by getRuleStatus().
+ *  A range of values is defined for each category of
+ *  word, to allow for further subdivisions of a category in future releases.
+ *  Applications should check for tag values falling within the range, rather
+ *  than for single individual values.
+ *
+ * The numeric values of all of these constants are stable (will not change).
+ *
+ * @stable ICU 2.8
+*/
+typedef enum ULineBreakTag {
+    /** Tag value for soft line breaks, positions at which a line break
+      *  is acceptable but not required                */
+    UBRK_LINE_SOFT            = 0,
+    /** Upper bound for soft line breaks.              */
+    UBRK_LINE_SOFT_LIMIT      = 100,
+    /** Tag value for a hard, or mandatory line break  */
+    UBRK_LINE_HARD            = 100,
+    /** Upper bound for hard line breaks.              */
+    UBRK_LINE_HARD_LIMIT      = 200
+} ULineBreakTag;
+
+
+
+/**
+ *  Enum constants for the sentence break tags returned by getRuleStatus().
+ *  A range of values is defined for each category of
+ *  sentence, to allow for further subdivisions of a category in future releases.
+ *  Applications should check for tag values falling within the range, rather
+ *  than for single individual values.
+ *
+ * The numeric values of all of these constants are stable (will not change).
+ *
+ * @stable ICU 2.8
+*/
+typedef enum USentenceBreakTag {
+    /** Tag value for for sentences  ending with a sentence terminator
+      * ('.', '?', '!', etc.) character, possibly followed by a
+      * hard separator (CR, LF, PS, etc.)
+      */
+    UBRK_SENTENCE_TERM       = 0,
+    /** Upper bound for tags for sentences ended by sentence terminators.    */
+    UBRK_SENTENCE_TERM_LIMIT = 100,
+    /** Tag value for for sentences that do not contain an ending
+      * sentence terminator ('.', '?', '!', etc.) character, but
+      * are ended only by a hard separator (CR, LF, PS, etc.) or end of input.
+      */
+    UBRK_SENTENCE_SEP        = 100,
+    /** Upper bound for tags for sentences ended by a separator.              */
+    UBRK_SENTENCE_SEP_LIMIT  = 200
+    /** Tag value for a hard, or mandatory line break  */
+} USentenceBreakTag;
+
+
+/**
+ * Open a new UBreakIterator for locating text boundaries for a specified locale.
+ * A UBreakIterator may be used for detecting character, line, word,
+ * and sentence breaks in text.
+ * @param type The type of UBreakIterator to open: one of UBRK_CHARACTER, UBRK_WORD,
+ * UBRK_LINE, UBRK_SENTENCE
+ * @param locale The locale specifying the text-breaking conventions. Note that
+ * locale keys such as "lb" and "ss" may be used to modify text break behavior,
+ * see general discussion of BreakIterator C API.
+ * @param text The text to be iterated over. May be null, in which case ubrk_setText() is
+ *        used to specify the text to be iterated.
+ * @param textLength The number of characters in text, or -1 if null-terminated.
+ * @param status A UErrorCode to receive any errors.
+ * @return A UBreakIterator for the specified locale.
+ * @see ubrk_openRules
+ * @stable ICU 2.0
+ */
+U_CAPI UBreakIterator* U_EXPORT2
+ubrk_open(UBreakIteratorType type,
+      const char *locale,
+      const UChar *text,
+      int32_t textLength,
+      UErrorCode *status);
+
+/**
+ * Open a new UBreakIterator for locating text boundaries using specified breaking rules.
+ * The rule syntax is ... (TBD)
+ * @param rules A set of rules specifying the text breaking conventions.
+ * @param rulesLength The number of characters in rules, or -1 if null-terminated.
+ * @param text The text to be iterated over.  May be null, in which case ubrk_setText() is
+ *        used to specify the text to be iterated.
+ * @param textLength The number of characters in text, or -1 if null-terminated.
+ * @param parseErr   Receives position and context information for any syntax errors
+ *                   detected while parsing the rules.
+ * @param status A UErrorCode to receive any errors.
+ * @return A UBreakIterator for the specified rules.
+ * @see ubrk_open
+ * @stable ICU 2.2
+ */
+U_CAPI UBreakIterator* U_EXPORT2
+ubrk_openRules(const UChar     *rules,
+               int32_t         rulesLength,
+               const UChar     *text,
+               int32_t          textLength,
+               UParseError     *parseErr,
+               UErrorCode      *status);
+
+/**
+ * Open a new UBreakIterator for locating text boundaries using precompiled binary rules.
+ * Opening a UBreakIterator this way is substantially faster than using ubrk_openRules.
+ * Binary rules may be obtained using ubrk_getBinaryRules. The compiled rules are not
+ * compatible across different major versions of ICU, nor across platforms of different
+ * endianness or different base character set family (ASCII vs EBCDIC).
+ * @param binaryRules A set of compiled binary rules specifying the text breaking
+ *                    conventions. Ownership of the storage containing the compiled
+ *                    rules remains with the caller of this function. The compiled
+ *                    rules must not be modified or deleted during the life of the
+ *                    break iterator.
+ * @param rulesLength The length of binaryRules in bytes; must be >= 0.
+ * @param text        The text to be iterated over.  May be null, in which case
+ *                    ubrk_setText() is used to specify the text to be iterated.
+ * @param textLength  The number of characters in text, or -1 if null-terminated.
+ * @param status      Pointer to UErrorCode to receive any errors.
+ * @return            UBreakIterator for the specified rules.
+ * @see ubrk_getBinaryRules
+ * @stable ICU 59
+ */
+U_CAPI UBreakIterator* U_EXPORT2
+ubrk_openBinaryRules(const uint8_t *binaryRules, int32_t rulesLength,
+                     const UChar *  text, int32_t textLength,
+                     UErrorCode *   status);
+
+#ifndef U_HIDE_DEPRECATED_API
+
+/**
+ * Thread safe cloning operation
+ * @param bi iterator to be cloned
+ * @param stackBuffer <em>Deprecated functionality as of ICU 52, use NULL.</em><br>
+ *  user allocated space for the new clone. If NULL new memory will be allocated.
+ *  If buffer is not large enough, new memory will be allocated.
+ *  Clients can use the U_BRK_SAFECLONE_BUFFERSIZE.
+ * @param pBufferSize <em>Deprecated functionality as of ICU 52, use NULL or 1.</em><br>
+ *  pointer to size of allocated space.
+ *  If *pBufferSize == 0, a sufficient size for use in cloning will
+ *  be returned ('pre-flighting')
+ *  If *pBufferSize is not enough for a stack-based safe clone,
+ *  new memory will be allocated.
+ * @param status to indicate whether the operation went on smoothly or there were errors
+ *  An informational status value, U_SAFECLONE_ALLOCATED_ERROR, is used
+ * if pBufferSize != NULL and any allocations were necessary
+ * @return pointer to the new clone
+ * @deprecated ICU 69 Use ubrk_clone() instead.
+ */
+U_DEPRECATED UBreakIterator * U_EXPORT2
+ubrk_safeClone(
+          const UBreakIterator *bi,
+          void *stackBuffer,
+          int32_t *pBufferSize,
+          UErrorCode *status);
+
+#endif /* U_HIDE_DEPRECATED_API */
+
+/**
+ * Thread safe cloning operation.
+ * @param bi iterator to be cloned
+ * @param status to indicate whether the operation went on smoothly or there were errors
+ * @return pointer to the new clone
+ * @stable ICU 69
+ */
+U_CAPI UBreakIterator * U_EXPORT2
+ubrk_clone(const UBreakIterator *bi,
+           UErrorCode *status);
+
+#ifndef U_HIDE_DEPRECATED_API
+
+/**
+  * A recommended size (in bytes) for the memory buffer to be passed to ubrk_saveClone().
+  * @deprecated ICU 52. Do not rely on ubrk_safeClone() cloning into any provided buffer.
+  */
+#define U_BRK_SAFECLONE_BUFFERSIZE 1
+
+#endif /* U_HIDE_DEPRECATED_API */
+
+/**
+* Close a UBreakIterator.
+* Once closed, a UBreakIterator may no longer be used.
+* @param bi The break iterator to close.
+ * @stable ICU 2.0
+*/
+U_CAPI void U_EXPORT2
+ubrk_close(UBreakIterator *bi);
+
+#if U_SHOW_CPLUSPLUS_API
+
+U_NAMESPACE_BEGIN
+
+/**
+ * \class LocalUBreakIteratorPointer
+ * "Smart pointer" class, closes a UBreakIterator via ubrk_close().
+ * For most methods see the LocalPointerBase base class.
+ *
+ * @see LocalPointerBase
+ * @see LocalPointer
+ * @stable ICU 4.4
+ */
+U_DEFINE_LOCAL_OPEN_POINTER(LocalUBreakIteratorPointer, UBreakIterator, ubrk_close);
+
+U_NAMESPACE_END
+
+#endif
+
+/**
+ * Sets an existing iterator to point to a new piece of text.
+ * The break iterator retains a pointer to the supplied text.
+ * The caller must not modify or delete the text while the BreakIterator
+ * retains the reference.
+ *
+ * @param bi The iterator to use
+ * @param text The text to be set
+ * @param textLength The length of the text
+ * @param status The error code
+ * @stable ICU 2.0
+ */
+U_CAPI void U_EXPORT2
+ubrk_setText(UBreakIterator* bi,
+             const UChar*    text,
+             int32_t         textLength,
+             UErrorCode*     status);
+
+
+/**
+ * Sets an existing iterator to point to a new piece of text.
+ *
+ * All index positions returned by break iterator functions are
+ * native indices from the UText. For example, when breaking UTF-8
+ * encoded text, the break positions returned by \ref ubrk_next, \ref ubrk_previous, etc.
+ * will be UTF-8 string indices, not UTF-16 positions.
+ *
+ * @param bi The iterator to use
+ * @param text The text to be set.
+ *             This function makes a shallow clone of the supplied UText.  This means
+ *             that the caller is free to immediately close or otherwise reuse the
+ *             UText that was passed as a parameter, but that the underlying text itself
+ *             must not be altered while being referenced by the break iterator.
+ * @param status The error code
+ * @stable ICU 3.4
+ */
+U_CAPI void U_EXPORT2
+ubrk_setUText(UBreakIterator* bi,
+             UText*          text,
+             UErrorCode*     status);
+
+
+
+/**
+ * Determine the most recently-returned text boundary.
+ *
+ * @param bi The break iterator to use.
+ * @return The character index most recently returned by \ref ubrk_next, \ref ubrk_previous,
+ * \ref ubrk_first, or \ref ubrk_last.
+ * @stable ICU 2.0
+ */
+U_CAPI int32_t U_EXPORT2
+ubrk_current(const UBreakIterator *bi);
+
+/**
+ * Advance the iterator to the boundary following the current boundary.
+ *
+ * @param bi The break iterator to use.
+ * @return The character index of the next text boundary, or UBRK_DONE
+ * if all text boundaries have been returned.
+ * @see ubrk_previous
+ * @stable ICU 2.0
+ */
+U_CAPI int32_t U_EXPORT2
+ubrk_next(UBreakIterator *bi);
+
+/**
+ * Set the iterator position to the boundary preceding the current boundary.
+ *
+ * @param bi The break iterator to use.
+ * @return The character index of the preceding text boundary, or UBRK_DONE
+ * if all text boundaries have been returned.
+ * @see ubrk_next
+ * @stable ICU 2.0
+ */
+U_CAPI int32_t U_EXPORT2
+ubrk_previous(UBreakIterator *bi);
+
+/**
+ * Set the iterator position to zero, the start of the text being scanned.
+ * @param bi The break iterator to use.
+ * @return The new iterator position (zero).
+ * @see ubrk_last
+ * @stable ICU 2.0
+ */
+U_CAPI int32_t U_EXPORT2
+ubrk_first(UBreakIterator *bi);
+
+/**
+ * Set the iterator position to the index immediately <EM>beyond</EM> the last character in the text being scanned.
+ * This is not the same as the last character.
+ * @param bi The break iterator to use.
+ * @return The character offset immediately <EM>beyond</EM> the last character in the
+ * text being scanned.
+ * @see ubrk_first
+ * @stable ICU 2.0
+ */
+U_CAPI int32_t U_EXPORT2
+ubrk_last(UBreakIterator *bi);
+
+/**
+ * Set the iterator position to the first boundary preceding the specified offset.
+ * The new position is always smaller than offset, or UBRK_DONE.
+ * @param bi The break iterator to use.
+ * @param offset The offset to begin scanning.
+ * @return The text boundary preceding offset, or UBRK_DONE.
+ * @see ubrk_following
+ * @stable ICU 2.0
+ */
+U_CAPI int32_t U_EXPORT2
+ubrk_preceding(UBreakIterator *bi,
+           int32_t offset);
+
+/**
+ * Advance the iterator to the first boundary following the specified offset.
+ * The value returned is always greater than offset, or UBRK_DONE.
+ * @param bi The break iterator to use.
+ * @param offset The offset to begin scanning.
+ * @return The text boundary following offset, or UBRK_DONE.
+ * @see ubrk_preceding
+ * @stable ICU 2.0
+ */
+U_CAPI int32_t U_EXPORT2
+ubrk_following(UBreakIterator *bi,
+           int32_t offset);
+
+/**
+* Get a locale for which text breaking information is available.
+* A UBreakIterator in a locale returned by this function will perform the correct
+* text breaking for the locale.
+* @param index The index of the desired locale.
+* @return A locale for which number text breaking information is available, or 0 if none.
+* @see ubrk_countAvailable
+* @stable ICU 2.0
+*/
+U_CAPI const char* U_EXPORT2
+ubrk_getAvailable(int32_t index);
+
+/**
+* Determine how many locales have text breaking information available.
+* This function is most useful as determining the loop ending condition for
+* calls to \ref ubrk_getAvailable.
+* @return The number of locales for which text breaking information is available.
+* @see ubrk_getAvailable
+* @stable ICU 2.0
+*/
+U_CAPI int32_t U_EXPORT2
+ubrk_countAvailable(void);
+
+
+/**
+* Returns true if the specified position is a boundary position.  As a side
+* effect, leaves the iterator pointing to the first boundary position at
+* or after "offset".
+* @param bi The break iterator to use.
+* @param offset the offset to check.
+* @return True if "offset" is a boundary position.
+* @stable ICU 2.0
+*/
+U_CAPI  UBool U_EXPORT2
+ubrk_isBoundary(UBreakIterator *bi, int32_t offset);
+
+/**
+ * Return the status from the break rule that determined the most recently
+ * returned break position.  The values appear in the rule source
+ * within brackets, {123}, for example.  For rules that do not specify a
+ * status, a default value of 0 is returned.
+ * <p>
+ * For word break iterators, the possible values are defined in enum UWordBreak.
+ * @stable ICU 2.2
+ */
+U_CAPI  int32_t U_EXPORT2
+ubrk_getRuleStatus(UBreakIterator *bi);
+
+/**
+ * Get the statuses from the break rules that determined the most recently
+ * returned break position.  The values appear in the rule source
+ * within brackets, {123}, for example.  The default status value for rules
+ * that do not explicitly provide one is zero.
+ * <p>
+ * For word break iterators, the possible values are defined in enum UWordBreak.
+ * @param bi        The break iterator to use
+ * @param fillInVec an array to be filled in with the status values.
+ * @param capacity  the length of the supplied vector.  A length of zero causes
+ *                  the function to return the number of status values, in the
+ *                  normal way, without attempting to store any values.
+ * @param status    receives error codes.
+ * @return          The number of rule status values from rules that determined
+ *                  the most recent boundary returned by the break iterator.
+ * @stable ICU 3.0
+ */
+U_CAPI  int32_t U_EXPORT2
+ubrk_getRuleStatusVec(UBreakIterator *bi, int32_t *fillInVec, int32_t capacity, UErrorCode *status);
+
+/**
+ * Return the locale of the break iterator. You can choose between the valid and
+ * the actual locale.
+ * @param bi break iterator
+ * @param type locale type (valid or actual)
+ * @param status error code
+ * @return locale string
+ * @stable ICU 2.8
+ */
+U_CAPI const char* U_EXPORT2
+ubrk_getLocaleByType(const UBreakIterator *bi, ULocDataLocaleType type, UErrorCode* status);
+
+/**
+  *  Set the subject text string upon which the break iterator is operating
+  *  without changing any other aspect of the state.
+  *  The new and previous text strings must have the same content.
+  *
+  *  This function is intended for use in environments where ICU is operating on
+  *  strings that may move around in memory.  It provides a mechanism for notifying
+  *  ICU that the string has been relocated, and providing a new UText to access the
+  *  string in its new position.
+  *
+  *  Note that the break iterator never copies the underlying text
+  *  of a string being processed, but always operates directly on the original text
+  *  provided by the user. Refreshing simply drops the references to the old text
+  *  and replaces them with references to the new.
+  *
+  *  Caution:  this function is normally used only by very specialized
+  *            system-level code.   One example use case is with garbage collection
+  *            that moves the text in memory.
+  *
+  * @param bi         The break iterator.
+  * @param text       The new (moved) text string.
+  * @param status     Receives errors detected by this function.
+  *
+  * @stable ICU 49
+  */
+U_CAPI void U_EXPORT2
+ubrk_refreshUText(UBreakIterator *bi,
+                       UText          *text,
+                       UErrorCode     *status);
+
+
+/**
+ * Get a compiled binary version of the rules specifying the behavior of a UBreakIterator.
+ * The binary rules may be used with ubrk_openBinaryRules to open a new UBreakIterator
+ * more quickly than using ubrk_openRules. The compiled rules are not compatible across
+ * different major versions of ICU, nor across platforms of different endianness or
+ * different base character set family (ASCII vs EBCDIC). Supports preflighting (with
+ * binaryRules=NULL and rulesCapacity=0) to get the rules length without copying them to
+ * the binaryRules buffer. However, whether preflighting or not, if the actual length
+ * is greater than INT32_MAX, then the function returns 0 and sets *status to
+ * U_INDEX_OUTOFBOUNDS_ERROR.
+
+ * @param bi            The break iterator to use.
+ * @param binaryRules   Buffer to receive the compiled binary rules; set to NULL for
+ *                      preflighting.
+ * @param rulesCapacity Capacity (in bytes) of the binaryRules buffer; set to 0 for
+ *                      preflighting. Must be >= 0.
+ * @param status        Pointer to UErrorCode to receive any errors, such as
+ *                      U_BUFFER_OVERFLOW_ERROR, U_INDEX_OUTOFBOUNDS_ERROR, or
+ *                      U_ILLEGAL_ARGUMENT_ERROR.
+ * @return              The actual byte length of the binary rules, if <= INT32_MAX;
+ *                      otherwise 0. If not preflighting and this is larger than
+ *                      rulesCapacity, *status will be set to an error.
+ * @see ubrk_openBinaryRules
+ * @stable ICU 59
+ */
+U_CAPI int32_t U_EXPORT2
+ubrk_getBinaryRules(UBreakIterator *bi,
+                    uint8_t *       binaryRules, int32_t rulesCapacity,
+                    UErrorCode *    status);
+
+#endif /* #if !UCONFIG_NO_BREAK_ITERATION */
+
+#endif

thirdparty/icu4c/common/unicode/ucasemap.h

@@ -0,0 +1,388 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+/*
+*******************************************************************************
+*
+*   Copyright (C) 2005-2012, International Business Machines
+*   Corporation and others.  All Rights Reserved.
+*
+*******************************************************************************
+*   file name:  ucasemap.h
+*   encoding:   UTF-8
+*   tab size:   8 (not used)
+*   indentation:4
+*
+*   created on: 2005may06
+*   created by: Markus W. Scherer
+*
+*   Case mapping service object and functions using it.
+*/
+
+#ifndef __UCASEMAP_H__
+#define __UCASEMAP_H__
+
+#include "unicode/utypes.h"
+#include "unicode/stringoptions.h"
+#include "unicode/ustring.h"
+
+#if U_SHOW_CPLUSPLUS_API
+#include "unicode/localpointer.h"
+#endif   // U_SHOW_CPLUSPLUS_API
+
+/**
+ * \file
+ * \brief C API: Unicode case mapping functions using a UCaseMap service object.
+ *
+ * The service object takes care of memory allocations, data loading, and setup
+ * for the attributes, as usual.
+ *
+ * Currently, the functionality provided here does not overlap with uchar.h
+ * and ustring.h, except for ucasemap_toTitle().
+ *
+ * ucasemap_utf8XYZ() functions operate directly on UTF-8 strings.
+ */
+
+/**
+ * UCaseMap is an opaque service object for newer ICU case mapping functions.
+ * Older functions did not use a service object.
+ * @stable ICU 3.4
+ */
+struct UCaseMap;
+typedef struct UCaseMap UCaseMap; /**< C typedef for struct UCaseMap. @stable ICU 3.4 */
+
+/**
+ * Open a UCaseMap service object for a locale and a set of options.
+ * The locale ID and options are preprocessed so that functions using the
+ * service object need not process them in each call.
+ *
+ * @param locale ICU locale ID, used for language-dependent
+ *               upper-/lower-/title-casing according to the Unicode standard.
+ *               Usual semantics: ""=root, NULL=default locale, etc.
+ * @param options Options bit set, used for case folding and string comparisons.
+ *                Same flags as for u_foldCase(), u_strFoldCase(),
+ *                u_strCaseCompare(), etc.
+ *                Use 0 or U_FOLD_CASE_DEFAULT for default behavior.
+ * @param pErrorCode Must be a valid pointer to an error code value,
+ *                   which must not indicate a failure before the function call.
+ * @return Pointer to a UCaseMap service object, if successful.
+ *
+ * @see U_FOLD_CASE_DEFAULT
+ * @see U_FOLD_CASE_EXCLUDE_SPECIAL_I
+ * @see U_TITLECASE_NO_LOWERCASE
+ * @see U_TITLECASE_NO_BREAK_ADJUSTMENT
+ * @stable ICU 3.4
+ */
+U_CAPI UCaseMap * U_EXPORT2
+ucasemap_open(const char *locale, uint32_t options, UErrorCode *pErrorCode);
+
+/**
+ * Close a UCaseMap service object.
+ * @param csm Object to be closed.
+ * @stable ICU 3.4
+ */
+U_CAPI void U_EXPORT2
+ucasemap_close(UCaseMap *csm);
+
+#if U_SHOW_CPLUSPLUS_API
+
+U_NAMESPACE_BEGIN
+
+/**
+ * \class LocalUCaseMapPointer
+ * "Smart pointer" class, closes a UCaseMap via ucasemap_close().
+ * For most methods see the LocalPointerBase base class.
+ *
+ * @see LocalPointerBase
+ * @see LocalPointer
+ * @stable ICU 4.4
+ */
+U_DEFINE_LOCAL_OPEN_POINTER(LocalUCaseMapPointer, UCaseMap, ucasemap_close);
+
+U_NAMESPACE_END
+
+#endif
+
+/**
+ * Get the locale ID that is used for language-dependent case mappings.
+ * @param csm UCaseMap service object.
+ * @return locale ID
+ * @stable ICU 3.4
+ */
+U_CAPI const char * U_EXPORT2
+ucasemap_getLocale(const UCaseMap *csm);
+
+/**
+ * Get the options bit set that is used for case folding and string comparisons.
+ * @param csm UCaseMap service object.
+ * @return options bit set
+ * @stable ICU 3.4
+ */
+U_CAPI uint32_t U_EXPORT2
+ucasemap_getOptions(const UCaseMap *csm);
+
+/**
+ * Set the locale ID that is used for language-dependent case mappings.
+ *
+ * @param csm UCaseMap service object.
+ * @param locale Locale ID, see ucasemap_open().
+ * @param pErrorCode Must be a valid pointer to an error code value,
+ *                   which must not indicate a failure before the function call.
+ *
+ * @see ucasemap_open
+ * @stable ICU 3.4
+ */
+U_CAPI void U_EXPORT2
+ucasemap_setLocale(UCaseMap *csm, const char *locale, UErrorCode *pErrorCode);
+
+/**
+ * Set the options bit set that is used for case folding and string comparisons.
+ *
+ * @param csm UCaseMap service object.
+ * @param options Options bit set, see ucasemap_open().
+ * @param pErrorCode Must be a valid pointer to an error code value,
+ *                   which must not indicate a failure before the function call.
+ *
+ * @see ucasemap_open
+ * @stable ICU 3.4
+ */
+U_CAPI void U_EXPORT2
+ucasemap_setOptions(UCaseMap *csm, uint32_t options, UErrorCode *pErrorCode);
+
+#if !UCONFIG_NO_BREAK_ITERATION
+
+/**
+ * Get the break iterator that is used for titlecasing.
+ * Do not modify the returned break iterator.
+ * @param csm UCaseMap service object.
+ * @return titlecasing break iterator
+ * @stable ICU 3.8
+ */
+U_CAPI const UBreakIterator * U_EXPORT2
+ucasemap_getBreakIterator(const UCaseMap *csm);
+
+/**
+ * Set the break iterator that is used for titlecasing.
+ * The UCaseMap service object releases a previously set break iterator
+ * and "adopts" this new one, taking ownership of it.
+ * It will be released in a subsequent call to ucasemap_setBreakIterator()
+ * or ucasemap_close().
+ *
+ * Break iterator operations are not thread-safe. Therefore, titlecasing
+ * functions use non-const UCaseMap objects. It is not possible to titlecase
+ * strings concurrently using the same UCaseMap.
+ *
+ * @param csm UCaseMap service object.
+ * @param iterToAdopt Break iterator to be adopted for titlecasing.
+ * @param pErrorCode Must be a valid pointer to an error code value,
+ *                   which must not indicate a failure before the function call.
+ *
+ * @see ucasemap_toTitle
+ * @see ucasemap_utf8ToTitle
+ * @stable ICU 3.8
+ */
+U_CAPI void U_EXPORT2
+ucasemap_setBreakIterator(UCaseMap *csm, UBreakIterator *iterToAdopt, UErrorCode *pErrorCode);
+
+/**
+ * Titlecase a UTF-16 string. This function is almost a duplicate of u_strToTitle(),
+ * except that it takes ucasemap_setOptions() into account and has performance
+ * advantages from being able to use a UCaseMap object for multiple case mapping
+ * operations, saving setup time.
+ *
+ * Casing is locale-dependent and context-sensitive.
+ * Titlecasing uses a break iterator to find the first characters of words
+ * that are to be titlecased. It titlecases those characters and lowercases
+ * all others. (This can be modified with ucasemap_setOptions().)
+ *
+ * Note: This function takes a non-const UCaseMap pointer because it will
+ * open a default break iterator if no break iterator was set yet,
+ * and effectively call ucasemap_setBreakIterator();
+ * also because the break iterator is stateful and will be modified during
+ * the iteration.
+ *
+ * The titlecase break iterator can be provided to customize for arbitrary
+ * styles, using rules and dictionaries beyond the standard iterators.
+ * If the break iterator passed in is null, the default Unicode algorithm
+ * will be used to determine the titlecase positions.
+ *
+ * This function uses only the setText(), first() and next() methods of the
+ * provided break iterator.
+ *
+ * The result may be longer or shorter than the original.
+ * The source string and the destination buffer must not overlap.
+ *
+ * @param csm       UCaseMap service object. This pointer is non-const!
+ *                  See the note above for details.
+ * @param dest      A buffer for the result string. The result will be NUL-terminated if
+ *                  the buffer is large enough.
+ *                  The contents is undefined in case of failure.
+ * @param destCapacity The size of the buffer (number of UChars). If it is 0, then
+ *                  dest may be NULL and the function will only return the length of the result
+ *                  without writing any of the result string.
+ * @param src       The original string.
+ * @param srcLength The length of the original string. If -1, then src must be NUL-terminated.
+ * @param pErrorCode Must be a valid pointer to an error code value,
+ *                  which must not indicate a failure before the function call.
+ * @return The length of the result string, if successful - or in case of a buffer overflow,
+ *         in which case it will be greater than destCapacity.
+ *
+ * @see u_strToTitle
+ * @stable ICU 3.8
+ */
+U_CAPI int32_t U_EXPORT2
+ucasemap_toTitle(UCaseMap *csm,
+                 UChar *dest, int32_t destCapacity,
+                 const UChar *src, int32_t srcLength,
+                 UErrorCode *pErrorCode);
+
+#endif  // UCONFIG_NO_BREAK_ITERATION
+
+/**
+ * Lowercase the characters in a UTF-8 string.
+ * Casing is locale-dependent and context-sensitive.
+ * The result may be longer or shorter than the original.
+ * The source string and the destination buffer must not overlap.
+ *
+ * @param csm       UCaseMap service object.
+ * @param dest      A buffer for the result string. The result will be NUL-terminated if
+ *                  the buffer is large enough.
+ *                  The contents is undefined in case of failure.
+ * @param destCapacity The size of the buffer (number of bytes). If it is 0, then
+ *                  dest may be NULL and the function will only return the length of the result
+ *                  without writing any of the result string.
+ * @param src       The original string.
+ * @param srcLength The length of the original string. If -1, then src must be NUL-terminated.
+ * @param pErrorCode Must be a valid pointer to an error code value,
+ *                  which must not indicate a failure before the function call.
+ * @return The length of the result string, if successful - or in case of a buffer overflow,
+ *         in which case it will be greater than destCapacity.
+ *
+ * @see u_strToLower
+ * @stable ICU 3.4
+ */
+U_CAPI int32_t U_EXPORT2
+ucasemap_utf8ToLower(const UCaseMap *csm,
+                     char *dest, int32_t destCapacity,
+                     const char *src, int32_t srcLength,
+                     UErrorCode *pErrorCode);
+
+/**
+ * Uppercase the characters in a UTF-8 string.
+ * Casing is locale-dependent and context-sensitive.
+ * The result may be longer or shorter than the original.
+ * The source string and the destination buffer must not overlap.
+ *
+ * @param csm       UCaseMap service object.
+ * @param dest      A buffer for the result string. The result will be NUL-terminated if
+ *                  the buffer is large enough.
+ *                  The contents is undefined in case of failure.
+ * @param destCapacity The size of the buffer (number of bytes). If it is 0, then
+ *                  dest may be NULL and the function will only return the length of the result
+ *                  without writing any of the result string.
+ * @param src       The original string.
+ * @param srcLength The length of the original string. If -1, then src must be NUL-terminated.
+ * @param pErrorCode Must be a valid pointer to an error code value,
+ *                  which must not indicate a failure before the function call.
+ * @return The length of the result string, if successful - or in case of a buffer overflow,
+ *         in which case it will be greater than destCapacity.
+ *
+ * @see u_strToUpper
+ * @stable ICU 3.4
+ */
+U_CAPI int32_t U_EXPORT2
+ucasemap_utf8ToUpper(const UCaseMap *csm,
+                     char *dest, int32_t destCapacity,
+                     const char *src, int32_t srcLength,
+                     UErrorCode *pErrorCode);
+
+#if !UCONFIG_NO_BREAK_ITERATION
+
+/**
+ * Titlecase a UTF-8 string.
+ * Casing is locale-dependent and context-sensitive.
+ * Titlecasing uses a break iterator to find the first characters of words
+ * that are to be titlecased. It titlecases those characters and lowercases
+ * all others. (This can be modified with ucasemap_setOptions().)
+ *
+ * Note: This function takes a non-const UCaseMap pointer because it will
+ * open a default break iterator if no break iterator was set yet,
+ * and effectively call ucasemap_setBreakIterator();
+ * also because the break iterator is stateful and will be modified during
+ * the iteration.
+ *
+ * The titlecase break iterator can be provided to customize for arbitrary
+ * styles, using rules and dictionaries beyond the standard iterators.
+ * If the break iterator passed in is null, the default Unicode algorithm
+ * will be used to determine the titlecase positions.
+ *
+ * This function uses only the setUText(), first(), next() and close() methods of the
+ * provided break iterator.
+ *
+ * The result may be longer or shorter than the original.
+ * The source string and the destination buffer must not overlap.
+ *
+ * @param csm       UCaseMap service object. This pointer is non-const!
+ *                  See the note above for details.
+ * @param dest      A buffer for the result string. The result will be NUL-terminated if
+ *                  the buffer is large enough.
+ *                  The contents is undefined in case of failure.
+ * @param destCapacity The size of the buffer (number of bytes). If it is 0, then
+ *                  dest may be NULL and the function will only return the length of the result
+ *                  without writing any of the result string.
+ * @param src       The original string.
+ * @param srcLength The length of the original string. If -1, then src must be NUL-terminated.
+ * @param pErrorCode Must be a valid pointer to an error code value,
+ *                  which must not indicate a failure before the function call.
+ * @return The length of the result string, if successful - or in case of a buffer overflow,
+ *         in which case it will be greater than destCapacity.
+ *
+ * @see u_strToTitle
+ * @see U_TITLECASE_NO_LOWERCASE
+ * @see U_TITLECASE_NO_BREAK_ADJUSTMENT
+ * @stable ICU 3.8
+ */
+U_CAPI int32_t U_EXPORT2
+ucasemap_utf8ToTitle(UCaseMap *csm,
+                    char *dest, int32_t destCapacity,
+                    const char *src, int32_t srcLength,
+                    UErrorCode *pErrorCode);
+
+#endif
+
+/**
+ * Case-folds the characters in a UTF-8 string.
+ *
+ * Case-folding is locale-independent and not context-sensitive,
+ * but there is an option for whether to include or exclude mappings for dotted I
+ * and dotless i that are marked with 'T' in CaseFolding.txt.
+ *
+ * The result may be longer or shorter than the original.
+ * The source string and the destination buffer must not overlap.
+ *
+ * @param csm       UCaseMap service object.
+ * @param dest      A buffer for the result string. The result will be NUL-terminated if
+ *                  the buffer is large enough.
+ *                  The contents is undefined in case of failure.
+ * @param destCapacity The size of the buffer (number of bytes). If it is 0, then
+ *                  dest may be NULL and the function will only return the length of the result
+ *                  without writing any of the result string.
+ * @param src       The original string.
+ * @param srcLength The length of the original string. If -1, then src must be NUL-terminated.
+ * @param pErrorCode Must be a valid pointer to an error code value,
+ *                  which must not indicate a failure before the function call.
+ * @return The length of the result string, if successful - or in case of a buffer overflow,
+ *         in which case it will be greater than destCapacity.
+ *
+ * @see u_strFoldCase
+ * @see ucasemap_setOptions
+ * @see U_FOLD_CASE_DEFAULT
+ * @see U_FOLD_CASE_EXCLUDE_SPECIAL_I
+ * @stable ICU 3.8
+ */
+U_CAPI int32_t U_EXPORT2
+ucasemap_utf8FoldCase(const UCaseMap *csm,
+                      char *dest, int32_t destCapacity,
+                      const char *src, int32_t srcLength,
+                      UErrorCode *pErrorCode);
+
+#endif

thirdparty/icu4c/common/unicode/ucat.h

@@ -0,0 +1,160 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+/*
+**********************************************************************
+* Copyright (c) 2003-2004, International Business Machines
+* Corporation and others.  All Rights Reserved.
+**********************************************************************
+* Author: Alan Liu
+* Created: March 19 2003
+* Since: ICU 2.6
+**********************************************************************
+*/
+#ifndef UCAT_H
+#define UCAT_H
+
+#include "unicode/utypes.h"
+#include "unicode/ures.h"
+
+/**
+ * \file
+ * \brief C API: Message Catalog Wrappers
+ *
+ * This C API provides look-alike functions that deliberately resemble
+ * the POSIX catopen, catclose, and catgets functions.  The underlying
+ * implementation is in terms of ICU resource bundles, rather than
+ * POSIX message catalogs.
+ *
+ * The ICU resource bundles obey standard ICU inheritance policies.
+ * To facilitate this, sets and messages are flattened into one tier.
+ * This is done by creating resource bundle keys of the form
+ * <set_num>%<msg_num> where set_num is the set number and msg_num is
+ * the message number, formatted as decimal strings.
+ *
+ * Example:  Consider a message catalog containing two sets:
+ *
+ * Set 1: Message 4  = "Good morning."
+ *        Message 5  = "Good afternoon."
+ *        Message 7  = "Good evening."
+ *        Message 8  = "Good night."
+ * Set 4: Message 14 = "Please "
+ *        Message 19 = "Thank you."
+ *        Message 20 = "Sincerely,"
+ *
+ * The ICU resource bundle source file would, assuming it is named
+ * "greet.txt", would look like this:
+ *
+ * greet
+ * {
+ *     1%4  { "Good morning." }
+ *     1%5  { "Good afternoon." }
+ *     1%7  { "Good evening." }
+ *     1%8  { "Good night." }
+ * 
+ *     4%14 { "Please " }
+ *     4%19 { "Thank you." }
+ *     4%20 { "Sincerely," }
+ * }
+ *
+ * The catgets function is commonly used in combination with functions
+ * like printf and strftime.  ICU components like message format can
+ * be used instead, although they use a different format syntax.
+ * There is an ICU package, icuio, that provides some of
+ * the POSIX-style formatting API.
+ */
+
+U_CDECL_BEGIN
+
+/**
+ * An ICU message catalog descriptor, analogous to nl_catd.
+ * 
+ * @stable ICU 2.6
+ */
+typedef UResourceBundle* u_nl_catd;
+
+/**
+ * Open and return an ICU message catalog descriptor. The descriptor
+ * may be passed to u_catgets() to retrieve localized strings.
+ *
+ * @param name string containing the full path pointing to the
+ * directory where the resources reside followed by the package name
+ * e.g. "/usr/resource/my_app/resources/guimessages" on a Unix system.
+ * If NULL, ICU default data files will be used.
+ *
+ * Unlike POSIX, environment variables are not interpolated within the
+ * name.
+ *
+ * @param locale the locale for which we want to open the resource. If
+ * NULL, the default ICU locale will be used (see uloc_getDefault). If
+ * strlen(locale) == 0, the root locale will be used.
+ *
+ * @param ec input/output error code. Upon output,
+ * U_USING_FALLBACK_WARNING indicates that a fallback locale was
+ * used. For example, 'de_CH' was requested, but nothing was found
+ * there, so 'de' was used. U_USING_DEFAULT_WARNING indicates that the
+ * default locale data or root locale data was used; neither the
+ * requested locale nor any of its fallback locales were found.
+ *
+ * @return a message catalog descriptor that may be passed to
+ * u_catgets(). If the ec parameter indicates success, then the caller
+ * is responsible for calling u_catclose() to close the message
+ * catalog. If the ec parameter indicates failure, then NULL will be
+ * returned.
+ * 
+ * @stable ICU 2.6
+ */
+U_CAPI u_nl_catd U_EXPORT2
+u_catopen(const char* name, const char* locale, UErrorCode* ec);
+
+/**
+ * Close an ICU message catalog, given its descriptor.
+ *
+ * @param catd a message catalog descriptor to be closed. May be NULL,
+ * in which case no action is taken.
+ * 
+ * @stable ICU 2.6
+ */
+U_CAPI void U_EXPORT2
+u_catclose(u_nl_catd catd);
+
+/**
+ * Retrieve a localized string from an ICU message catalog.
+ *
+ * @param catd a message catalog descriptor returned by u_catopen.
+ *
+ * @param set_num the message catalog set number. Sets need not be
+ * numbered consecutively.
+ *
+ * @param msg_num the message catalog message number within the
+ * set. Messages need not be numbered consecutively.
+ *
+ * @param s the default string. This is returned if the string
+ * specified by the set_num and msg_num is not found. It must be
+ * zero-terminated.
+ *
+ * @param len fill-in parameter to receive the length of the result.
+ * May be NULL, in which case it is ignored.
+ *
+ * @param ec input/output error code. May be U_USING_FALLBACK_WARNING
+ * or U_USING_DEFAULT_WARNING. U_MISSING_RESOURCE_ERROR indicates that
+ * the set_num/msg_num tuple does not specify a valid message string
+ * in this catalog.
+ *
+ * @return a pointer to a zero-terminated UChar array which lives in
+ * an internal buffer area, typically a memory mapped/DLL file. The
+ * caller must NOT delete this pointer. If the call is unsuccessful
+ * for any reason, then s is returned.  This includes the situation in
+ * which ec indicates a failing error code upon entry to this
+ * function.
+ * 
+ * @stable ICU 2.6
+ */
+U_CAPI const UChar* U_EXPORT2
+u_catgets(u_nl_catd catd, int32_t set_num, int32_t msg_num,
+          const UChar* s,
+          int32_t* len, UErrorCode* ec);
+
+U_CDECL_END
+
+#endif /*UCAT_H*/
+/*eof*/

thirdparty/icu4c/common/unicode/ucharstrie.h

@@ -0,0 +1,623 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+/*
+*******************************************************************************
+*   Copyright (C) 2010-2012, International Business Machines
+*   Corporation and others.  All Rights Reserved.
+*******************************************************************************
+*   file name:  ucharstrie.h
+*   encoding:   UTF-8
+*   tab size:   8 (not used)
+*   indentation:4
+*
+*   created on: 2010nov14
+*   created by: Markus W. Scherer
+*/
+
+#ifndef __UCHARSTRIE_H__
+#define __UCHARSTRIE_H__
+
+/**
+ * \file
+ * \brief C++ API: Trie for mapping Unicode strings (or 16-bit-unit sequences)
+ *                 to integer values.
+ */
+
+#include "unicode/utypes.h"
+
+#if U_SHOW_CPLUSPLUS_API
+
+#include "unicode/unistr.h"
+#include "unicode/uobject.h"
+#include "unicode/ustringtrie.h"
+
+U_NAMESPACE_BEGIN
+
+class Appendable;
+class UCharsTrieBuilder;
+class UVector32;
+
+/**
+ * Light-weight, non-const reader class for a UCharsTrie.
+ * Traverses a char16_t-serialized data structure with minimal state,
+ * for mapping strings (16-bit-unit sequences) to non-negative integer values.
+ *
+ * This class owns the serialized trie data only if it was constructed by
+ * the builder's build() method.
+ * The public constructor and the copy constructor only alias the data (only copy the pointer).
+ * There is no assignment operator.
+ *
+ * This class is not intended for public subclassing.
+ * @stable ICU 4.8
+ */
+class U_COMMON_API UCharsTrie : public UMemory {
+public:
+    /**
+     * Constructs a UCharsTrie reader instance.
+     *
+     * The trieUChars must contain a copy of a char16_t sequence from the UCharsTrieBuilder,
+     * starting with the first char16_t of that sequence.
+     * The UCharsTrie object will not read more char16_ts than
+     * the UCharsTrieBuilder generated in the corresponding build() call.
+     *
+     * The array is not copied/cloned and must not be modified while
+     * the UCharsTrie object is in use.
+     *
+     * @param trieUChars The char16_t array that contains the serialized trie.
+     * @stable ICU 4.8
+     */
+    UCharsTrie(ConstChar16Ptr trieUChars)
+            : ownedArray_(nullptr), uchars_(trieUChars),
+              pos_(uchars_), remainingMatchLength_(-1) {}
+
+    /**
+     * Destructor.
+     * @stable ICU 4.8
+     */
+    ~UCharsTrie();
+
+    /**
+     * Copy constructor, copies the other trie reader object and its state,
+     * but not the char16_t array which will be shared. (Shallow copy.)
+     * @param other Another UCharsTrie object.
+     * @stable ICU 4.8
+     */
+    UCharsTrie(const UCharsTrie &other)
+            : ownedArray_(nullptr), uchars_(other.uchars_),
+              pos_(other.pos_), remainingMatchLength_(other.remainingMatchLength_) {}
+
+    /**
+     * Resets this trie to its initial state.
+     * @return *this
+     * @stable ICU 4.8
+     */
+    UCharsTrie &reset() {
+        pos_=uchars_;
+        remainingMatchLength_=-1;
+        return *this;
+    }
+
+    /**
+     * Returns the state of this trie as a 64-bit integer.
+     * The state value is never 0.
+     *
+     * @return opaque state value
+     * @see resetToState64
+     * @stable ICU 65
+     */
+    uint64_t getState64() const {
+        return (static_cast<uint64_t>(remainingMatchLength_ + 2) << kState64RemainingShift) |
+            static_cast<uint64_t>(pos_ - uchars_);
+    }
+
+    /**
+     * Resets this trie to the saved state.
+     * Unlike resetToState(State), the 64-bit state value
+     * must be from getState64() from the same trie object or
+     * from one initialized the exact same way.
+     * Because of no validation, this method is faster.
+     *
+     * @param state The opaque trie state value from getState64().
+     * @return *this
+     * @see getState64
+     * @see resetToState
+     * @see reset
+     * @stable ICU 65
+     */
+    UCharsTrie &resetToState64(uint64_t state) {
+        remainingMatchLength_ = static_cast<int32_t>(state >> kState64RemainingShift) - 2;
+        pos_ = uchars_ + (state & kState64PosMask);
+        return *this;
+    }
+
+    /**
+     * UCharsTrie state object, for saving a trie's current state
+     * and resetting the trie back to this state later.
+     * @stable ICU 4.8
+     */
+    class State : public UMemory {
+    public:
+        /**
+         * Constructs an empty State.
+         * @stable ICU 4.8
+         */
+        State() { uchars=nullptr; }
+    private:
+        friend class UCharsTrie;
+
+        const char16_t *uchars;
+        const char16_t *pos;
+        int32_t remainingMatchLength;
+    };
+
+    /**
+     * Saves the state of this trie.
+     * @param state The State object to hold the trie's state.
+     * @return *this
+     * @see resetToState
+     * @stable ICU 4.8
+     */
+    const UCharsTrie &saveState(State &state) const {
+        state.uchars=uchars_;
+        state.pos=pos_;
+        state.remainingMatchLength=remainingMatchLength_;
+        return *this;
+    }
+
+    /**
+     * Resets this trie to the saved state.
+     * If the state object contains no state, or the state of a different trie,
+     * then this trie remains unchanged.
+     * @param state The State object which holds a saved trie state.
+     * @return *this
+     * @see saveState
+     * @see reset
+     * @stable ICU 4.8
+     */
+    UCharsTrie &resetToState(const State &state) {
+        if(uchars_==state.uchars && uchars_!=nullptr) {
+            pos_=state.pos;
+            remainingMatchLength_=state.remainingMatchLength;
+        }
+        return *this;
+    }
+
+    /**
+     * Determines whether the string so far matches, whether it has a value,
+     * and whether another input char16_t can continue a matching string.
+     * @return The match/value Result.
+     * @stable ICU 4.8
+     */
+    UStringTrieResult current() const;
+
+    /**
+     * Traverses the trie from the initial state for this input char16_t.
+     * Equivalent to reset().next(uchar).
+     * @param uchar Input char value. Values below 0 and above 0xffff will never match.
+     * @return The match/value Result.
+     * @stable ICU 4.8
+     */
+    inline UStringTrieResult first(int32_t uchar) {
+        remainingMatchLength_=-1;
+        return nextImpl(uchars_, uchar);
+    }
+
+    /**
+     * Traverses the trie from the initial state for the
+     * one or two UTF-16 code units for this input code point.
+     * Equivalent to reset().nextForCodePoint(cp).
+     * @param cp A Unicode code point 0..0x10ffff.
+     * @return The match/value Result.
+     * @stable ICU 4.8
+     */
+    UStringTrieResult firstForCodePoint(UChar32 cp);
+
+    /**
+     * Traverses the trie from the current state for this input char16_t.
+     * @param uchar Input char value. Values below 0 and above 0xffff will never match.
+     * @return The match/value Result.
+     * @stable ICU 4.8
+     */
+    UStringTrieResult next(int32_t uchar);
+
+    /**
+     * Traverses the trie from the current state for the
+     * one or two UTF-16 code units for this input code point.
+     * @param cp A Unicode code point 0..0x10ffff.
+     * @return The match/value Result.
+     * @stable ICU 4.8
+     */
+    UStringTrieResult nextForCodePoint(UChar32 cp);
+
+    /**
+     * Traverses the trie from the current state for this string.
+     * Equivalent to
+     * \code
+     * Result result=current();
+     * for(each c in s)
+     *   if(!USTRINGTRIE_HAS_NEXT(result)) return USTRINGTRIE_NO_MATCH;
+     *   result=next(c);
+     * return result;
+     * \endcode
+     * @param s A string. Can be nullptr if length is 0.
+     * @param length The length of the string. Can be -1 if NUL-terminated.
+     * @return The match/value Result.
+     * @stable ICU 4.8
+     */
+    UStringTrieResult next(ConstChar16Ptr s, int32_t length);
+
+    /**
+     * Returns a matching string's value if called immediately after
+     * current()/first()/next() returned USTRINGTRIE_INTERMEDIATE_VALUE or USTRINGTRIE_FINAL_VALUE.
+     * getValue() can be called multiple times.
+     *
+     * Do not call getValue() after USTRINGTRIE_NO_MATCH or USTRINGTRIE_NO_VALUE!
+     * @return The value for the string so far.
+     * @stable ICU 4.8
+     */
+    inline int32_t getValue() const {
+        const char16_t *pos=pos_;
+        int32_t leadUnit=*pos++;
+        // U_ASSERT(leadUnit>=kMinValueLead);
+        return leadUnit&kValueIsFinal ?
+            readValue(pos, leadUnit&0x7fff) : readNodeValue(pos, leadUnit);
+    }
+
+    /**
+     * Determines whether all strings reachable from the current state
+     * map to the same value.
+     * @param uniqueValue Receives the unique value, if this function returns true.
+     *                    (output-only)
+     * @return true if all strings reachable from the current state
+     *         map to the same value.
+     * @stable ICU 4.8
+     */
+    inline UBool hasUniqueValue(int32_t &uniqueValue) const {
+        const char16_t *pos=pos_;
+        // Skip the rest of a pending linear-match node.
+        return pos!=nullptr && findUniqueValue(pos+remainingMatchLength_+1, false, uniqueValue);
+    }
+
+    /**
+     * Finds each char16_t which continues the string from the current state.
+     * That is, each char16_t c for which it would be next(c)!=USTRINGTRIE_NO_MATCH now.
+     * @param out Each next char16_t is appended to this object.
+     * @return the number of char16_ts which continue the string from here
+     * @stable ICU 4.8
+     */
+    int32_t getNextUChars(Appendable &out) const;
+
+    /**
+     * Iterator for all of the (string, value) pairs in a UCharsTrie.
+     * @stable ICU 4.8
+     */
+    class U_COMMON_API Iterator : public UMemory {
+    public:
+        /**
+         * Iterates from the root of a char16_t-serialized UCharsTrie.
+         * @param trieUChars The trie char16_ts.
+         * @param maxStringLength If 0, the iterator returns full strings.
+         *                        Otherwise, the iterator returns strings with this maximum length.
+         * @param errorCode Standard ICU error code. Its input value must
+         *                  pass the U_SUCCESS() test, or else the function returns
+         *                  immediately. Check for U_FAILURE() on output or use with
+         *                  function chaining. (See User Guide for details.)
+         * @stable ICU 4.8
+         */
+        Iterator(ConstChar16Ptr trieUChars, int32_t maxStringLength, UErrorCode &errorCode);
+
+        /**
+         * Iterates from the current state of the specified UCharsTrie.
+         * @param trie The trie whose state will be copied for iteration.
+         * @param maxStringLength If 0, the iterator returns full strings.
+         *                        Otherwise, the iterator returns strings with this maximum length.
+         * @param errorCode Standard ICU error code. Its input value must
+         *                  pass the U_SUCCESS() test, or else the function returns
+         *                  immediately. Check for U_FAILURE() on output or use with
+         *                  function chaining. (See User Guide for details.)
+         * @stable ICU 4.8
+         */
+        Iterator(const UCharsTrie &trie, int32_t maxStringLength, UErrorCode &errorCode);
+
+        /**
+         * Destructor.
+         * @stable ICU 4.8
+         */
+        ~Iterator();
+
+        /**
+         * Resets this iterator to its initial state.
+         * @return *this
+         * @stable ICU 4.8
+         */
+        Iterator &reset();
+
+        /**
+         * @return true if there are more elements.
+         * @stable ICU 4.8
+         */
+        UBool hasNext() const;
+
+        /**
+         * Finds the next (string, value) pair if there is one.
+         *
+         * If the string is truncated to the maximum length and does not
+         * have a real value, then the value is set to -1.
+         * In this case, this "not a real value" is indistinguishable from
+         * a real value of -1.
+         * @param errorCode Standard ICU error code. Its input value must
+         *                  pass the U_SUCCESS() test, or else the function returns
+         *                  immediately. Check for U_FAILURE() on output or use with
+         *                  function chaining. (See User Guide for details.)
+         * @return true if there is another element.
+         * @stable ICU 4.8
+         */
+        UBool next(UErrorCode &errorCode);
+
+        /**
+         * @return The string for the last successful next().
+         * @stable ICU 4.8
+         */
+        const UnicodeString &getString() const { return str_; }
+        /**
+         * @return The value for the last successful next().
+         * @stable ICU 4.8
+         */
+        int32_t getValue() const { return value_; }
+
+    private:
+        UBool truncateAndStop() {
+            pos_=nullptr;
+            value_=-1;  // no real value for str
+            return true;
+        }
+
+        const char16_t *branchNext(const char16_t *pos, int32_t length, UErrorCode &errorCode);
+
+        const char16_t *uchars_;
+        const char16_t *pos_;
+        const char16_t *initialPos_;
+        int32_t remainingMatchLength_;
+        int32_t initialRemainingMatchLength_;
+        UBool skipValue_;  // Skip intermediate value which was already delivered.
+
+        UnicodeString str_;
+        int32_t maxLength_;
+        int32_t value_;
+
+        // The stack stores pairs of integers for backtracking to another
+        // outbound edge of a branch node.
+        // The first integer is an offset from uchars_.
+        // The second integer has the str_.length() from before the node in bits 15..0,
+        // and the remaining branch length in bits 31..16.
+        // (We could store the remaining branch length minus 1 in bits 30..16 and not use the sign bit,
+        // but the code looks more confusing that way.)
+        UVector32 *stack_;
+    };
+
+private:
+    friend class UCharsTrieBuilder;
+
+    /**
+     * Constructs a UCharsTrie reader instance.
+     * Unlike the public constructor which just aliases an array,
+     * this constructor adopts the builder's array.
+     * This constructor is only called by the builder.
+     */
+    UCharsTrie(char16_t *adoptUChars, const char16_t *trieUChars)
+            : ownedArray_(adoptUChars), uchars_(trieUChars),
+              pos_(uchars_), remainingMatchLength_(-1) {}
+
+    // No assignment operator.
+    UCharsTrie &operator=(const UCharsTrie &other) = delete;
+
+    inline void stop() {
+        pos_=nullptr;
+    }
+
+    // Reads a compact 32-bit integer.
+    // pos is already after the leadUnit, and the lead unit has bit 15 reset.
+    static inline int32_t readValue(const char16_t *pos, int32_t leadUnit) {
+        int32_t value;
+        if(leadUnit<kMinTwoUnitValueLead) {
+            value=leadUnit;
+        } else if(leadUnit<kThreeUnitValueLead) {
+            value=((leadUnit-kMinTwoUnitValueLead)<<16)|*pos;
+        } else {
+            value=(pos[0]<<16)|pos[1];
+        }
+        return value;
+    }
+    static inline const char16_t *skipValue(const char16_t *pos, int32_t leadUnit) {
+        if(leadUnit>=kMinTwoUnitValueLead) {
+            if(leadUnit<kThreeUnitValueLead) {
+                ++pos;
+            } else {
+                pos+=2;
+            }
+        }
+        return pos;
+    }
+    static inline const char16_t *skipValue(const char16_t *pos) {
+        int32_t leadUnit=*pos++;
+        return skipValue(pos, leadUnit&0x7fff);
+    }
+
+    static inline int32_t readNodeValue(const char16_t *pos, int32_t leadUnit) {
+        // U_ASSERT(kMinValueLead<=leadUnit && leadUnit<kValueIsFinal);
+        int32_t value;
+        if(leadUnit<kMinTwoUnitNodeValueLead) {
+            value=(leadUnit>>6)-1;
+        } else if(leadUnit<kThreeUnitNodeValueLead) {
+            value=(((leadUnit&0x7fc0)-kMinTwoUnitNodeValueLead)<<10)|*pos;
+        } else {
+            value=(pos[0]<<16)|pos[1];
+        }
+        return value;
+    }
+    static inline const char16_t *skipNodeValue(const char16_t *pos, int32_t leadUnit) {
+        // U_ASSERT(kMinValueLead<=leadUnit && leadUnit<kValueIsFinal);
+        if(leadUnit>=kMinTwoUnitNodeValueLead) {
+            if(leadUnit<kThreeUnitNodeValueLead) {
+                ++pos;
+            } else {
+                pos+=2;
+            }
+        }
+        return pos;
+    }
+
+    static inline const char16_t *jumpByDelta(const char16_t *pos) {
+        int32_t delta=*pos++;
+        if(delta>=kMinTwoUnitDeltaLead) {
+            if(delta==kThreeUnitDeltaLead) {
+                delta=(pos[0]<<16)|pos[1];
+                pos+=2;
+            } else {
+                delta=((delta-kMinTwoUnitDeltaLead)<<16)|*pos++;
+            }
+        }
+        return pos+delta;
+    }
+
+    static const char16_t *skipDelta(const char16_t *pos) {
+        int32_t delta=*pos++;
+        if(delta>=kMinTwoUnitDeltaLead) {
+            if(delta==kThreeUnitDeltaLead) {
+                pos+=2;
+            } else {
+                ++pos;
+            }
+        }
+        return pos;
+    }
+
+    static inline UStringTrieResult valueResult(int32_t node) {
+        return static_cast<UStringTrieResult>(USTRINGTRIE_INTERMEDIATE_VALUE - (node >> 15));
+    }
+
+    // Handles a branch node for both next(uchar) and next(string).
+    UStringTrieResult branchNext(const char16_t *pos, int32_t length, int32_t uchar);
+
+    // Requires remainingLength_<0.
+    UStringTrieResult nextImpl(const char16_t *pos, int32_t uchar);
+
+    // Helper functions for hasUniqueValue().
+    // Recursively finds a unique value (or whether there is not a unique one)
+    // from a branch.
+    static const char16_t *findUniqueValueFromBranch(const char16_t *pos, int32_t length,
+                                                  UBool haveUniqueValue, int32_t &uniqueValue);
+    // Recursively finds a unique value (or whether there is not a unique one)
+    // starting from a position on a node lead unit.
+    static UBool findUniqueValue(const char16_t *pos, UBool haveUniqueValue, int32_t &uniqueValue);
+
+    // Helper functions for getNextUChars().
+    // getNextUChars() when pos is on a branch node.
+    static void getNextBranchUChars(const char16_t *pos, int32_t length, Appendable &out);
+
+    // UCharsTrie data structure
+    //
+    // The trie consists of a series of char16_t-serialized nodes for incremental
+    // Unicode string/char16_t sequence matching. (char16_t=16-bit unsigned integer)
+    // The root node is at the beginning of the trie data.
+    //
+    // Types of nodes are distinguished by their node lead unit ranges.
+    // After each node, except a final-value node, another node follows to
+    // encode match values or continue matching further units.
+    //
+    // Node types:
+    //  - Final-value node: Stores a 32-bit integer in a compact, variable-length format.
+    //    The value is for the string/char16_t sequence so far.
+    //  - Match node, optionally with an intermediate value in a different compact format.
+    //    The value, if present, is for the string/char16_t sequence so far.
+    //
+    //  Aside from the value, which uses the node lead unit's high bits:
+    //
+    //  - Linear-match node: Matches a number of units.
+    //  - Branch node: Branches to other nodes according to the current input unit.
+    //    The node unit is the length of the branch (number of units to select from)
+    //    minus 1. It is followed by a sub-node:
+    //    - If the length is at most kMaxBranchLinearSubNodeLength, then
+    //      there are length-1 (key, value) pairs and then one more comparison unit.
+    //      If one of the key units matches, then the value is either a final value for
+    //      the string so far, or a "jump" delta to the next node.
+    //      If the last unit matches, then matching continues with the next node.
+    //      (Values have the same encoding as final-value nodes.)
+    //    - If the length is greater than kMaxBranchLinearSubNodeLength, then
+    //      there is one unit and one "jump" delta.
+    //      If the input unit is less than the sub-node unit, then "jump" by delta to
+    //      the next sub-node which will have a length of length/2.
+    //      (The delta has its own compact encoding.)
+    //      Otherwise, skip the "jump" delta to the next sub-node
+    //      which will have a length of length-length/2.
+
+    // Match-node lead unit values, after masking off intermediate-value bits:
+
+    // 0000..002f: Branch node. If node!=0 then the length is node+1, otherwise
+    // the length is one more than the next unit.
+
+    // For a branch sub-node with at most this many entries, we drop down
+    // to a linear search.
+    static const int32_t kMaxBranchLinearSubNodeLength=5;
+
+    // 0030..003f: Linear-match node, match 1..16 units and continue reading the next node.
+    static const int32_t kMinLinearMatch=0x30;
+    static const int32_t kMaxLinearMatchLength=0x10;
+
+    // Match-node lead unit bits 14..6 for the optional intermediate value.
+    // If these bits are 0, then there is no intermediate value.
+    // Otherwise, see the *NodeValue* constants below.
+    static const int32_t kMinValueLead=kMinLinearMatch+kMaxLinearMatchLength;  // 0x0040
+    static const int32_t kNodeTypeMask=kMinValueLead-1;  // 0x003f
+
+    // A final-value node has bit 15 set.
+    static const int32_t kValueIsFinal=0x8000;
+
+    // Compact value: After testing and masking off bit 15, use the following thresholds.
+    static const int32_t kMaxOneUnitValue=0x3fff;
+
+    static const int32_t kMinTwoUnitValueLead=kMaxOneUnitValue+1;  // 0x4000
+    static const int32_t kThreeUnitValueLead=0x7fff;
+
+    static const int32_t kMaxTwoUnitValue=((kThreeUnitValueLead-kMinTwoUnitValueLead)<<16)-1;  // 0x3ffeffff
+
+    // Compact intermediate-value integer, lead unit shared with a branch or linear-match node.
+    static const int32_t kMaxOneUnitNodeValue=0xff;
+    static const int32_t kMinTwoUnitNodeValueLead=kMinValueLead+((kMaxOneUnitNodeValue+1)<<6);  // 0x4040
+    static const int32_t kThreeUnitNodeValueLead=0x7fc0;
+
+    static const int32_t kMaxTwoUnitNodeValue=
+        ((kThreeUnitNodeValueLead-kMinTwoUnitNodeValueLead)<<10)-1;  // 0xfdffff
+
+    // Compact delta integers.
+    static const int32_t kMaxOneUnitDelta=0xfbff;
+    static const int32_t kMinTwoUnitDeltaLead=kMaxOneUnitDelta+1;  // 0xfc00
+    static const int32_t kThreeUnitDeltaLead=0xffff;
+
+    static const int32_t kMaxTwoUnitDelta=((kThreeUnitDeltaLead-kMinTwoUnitDeltaLead)<<16)-1;  // 0x03feffff
+
+    // For getState64():
+    // The remainingMatchLength_ is -1..14=(kMaxLinearMatchLength=0x10)-2
+    // so we need at least 5 bits for that.
+    // We add 2 to store it as a positive value 1..16=kMaxLinearMatchLength.
+    static constexpr int32_t kState64RemainingShift = 59;
+    static constexpr uint64_t kState64PosMask = (UINT64_C(1) << kState64RemainingShift) - 1;
+
+    char16_t *ownedArray_;
+
+    // Fixed value referencing the UCharsTrie words.
+    const char16_t *uchars_;
+
+    // Iterator variables.
+
+    // Pointer to next trie unit to read. nullptr if no more matches.
+    const char16_t *pos_;
+    // Remaining length of a linear-match node, minus 1. Negative if not in such a node.
+    int32_t remainingMatchLength_;
+};
+
+U_NAMESPACE_END
+
+#endif /* U_SHOW_CPLUSPLUS_API */
+
+#endif  // __UCHARSTRIE_H__

thirdparty/icu4c/common/unicode/ucharstriebuilder.h

@@ -0,0 +1,193 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+/*
+*******************************************************************************
+*   Copyright (C) 2010-2016, International Business Machines
+*   Corporation and others.  All Rights Reserved.
+*******************************************************************************
+*   file name:  ucharstriebuilder.h
+*   encoding:   UTF-8
+*   tab size:   8 (not used)
+*   indentation:4
+*
+*   created on: 2010nov14
+*   created by: Markus W. Scherer
+*/
+
+#ifndef __UCHARSTRIEBUILDER_H__
+#define __UCHARSTRIEBUILDER_H__
+
+#include "unicode/utypes.h"
+
+#if U_SHOW_CPLUSPLUS_API
+
+#include "unicode/stringtriebuilder.h"
+#include "unicode/ucharstrie.h"
+#include "unicode/unistr.h"
+
+/**
+ * \file
+ * \brief C++ API: Builder for icu::UCharsTrie
+ */
+
+U_NAMESPACE_BEGIN
+
+class UCharsTrieElement;
+
+/**
+ * Builder class for UCharsTrie.
+ *
+ * This class is not intended for public subclassing.
+ * @stable ICU 4.8
+ */
+class U_COMMON_API UCharsTrieBuilder : public StringTrieBuilder {
+public:
+    /**
+     * Constructs an empty builder.
+     * @param errorCode Standard ICU error code.
+     * @stable ICU 4.8
+     */
+    UCharsTrieBuilder(UErrorCode &errorCode);
+
+    /**
+     * Destructor.
+     * @stable ICU 4.8
+     */
+    virtual ~UCharsTrieBuilder();
+
+    /**
+     * Adds a (string, value) pair.
+     * The string must be unique.
+     * The string contents will be copied; the builder does not keep
+     * a reference to the input UnicodeString or its buffer.
+     * @param s The input string.
+     * @param value The value associated with this string.
+     * @param errorCode Standard ICU error code. Its input value must
+     *                  pass the U_SUCCESS() test, or else the function returns
+     *                  immediately. Check for U_FAILURE() on output or use with
+     *                  function chaining. (See User Guide for details.)
+     * @return *this
+     * @stable ICU 4.8
+     */
+    UCharsTrieBuilder &add(const UnicodeString &s, int32_t value, UErrorCode &errorCode);
+
+    /**
+     * Builds a UCharsTrie for the add()ed data.
+     * Once built, no further data can be add()ed until clear() is called.
+     *
+     * A UCharsTrie cannot be empty. At least one (string, value) pair
+     * must have been add()ed.
+     *
+     * This method passes ownership of the builder's internal result array to the new trie object.
+     * Another call to any build() variant will re-serialize the trie.
+     * After clear() has been called, a new array will be used as well.
+     * @param buildOption Build option, see UStringTrieBuildOption.
+     * @param errorCode Standard ICU error code. Its input value must
+     *                  pass the U_SUCCESS() test, or else the function returns
+     *                  immediately. Check for U_FAILURE() on output or use with
+     *                  function chaining. (See User Guide for details.)
+     * @return A new UCharsTrie for the add()ed data.
+     * @stable ICU 4.8
+     */
+    UCharsTrie *build(UStringTrieBuildOption buildOption, UErrorCode &errorCode);
+
+    /**
+     * Builds a UCharsTrie for the add()ed data and char16_t-serializes it.
+     * Once built, no further data can be add()ed until clear() is called.
+     *
+     * A UCharsTrie cannot be empty. At least one (string, value) pair
+     * must have been add()ed.
+     *
+     * Multiple calls to buildUnicodeString() set the UnicodeStrings to the
+     * builder's same char16_t array, without rebuilding.
+     * If buildUnicodeString() is called after build(), the trie will be
+     * re-serialized into a new array (because build() passes on ownership).
+     * If build() is called after buildUnicodeString(), the trie object returned
+     * by build() will become the owner of the underlying data for the
+     * previously returned UnicodeString.
+     * After clear() has been called, a new array will be used as well.
+     * @param buildOption Build option, see UStringTrieBuildOption.
+     * @param result A UnicodeString which will be set to the char16_t-serialized
+     *               UCharsTrie for the add()ed data.
+     * @param errorCode Standard ICU error code. Its input value must
+     *                  pass the U_SUCCESS() test, or else the function returns
+     *                  immediately. Check for U_FAILURE() on output or use with
+     *                  function chaining. (See User Guide for details.)
+     * @return result
+     * @stable ICU 4.8
+     */
+    UnicodeString &buildUnicodeString(UStringTrieBuildOption buildOption, UnicodeString &result,
+                                      UErrorCode &errorCode);
+
+    /**
+     * Removes all (string, value) pairs.
+     * New data can then be add()ed and a new trie can be built.
+     * @return *this
+     * @stable ICU 4.8
+     */
+    UCharsTrieBuilder &clear() {
+        strings.remove();
+        elementsLength=0;
+        ucharsLength=0;
+        return *this;
+    }
+
+private:
+    UCharsTrieBuilder(const UCharsTrieBuilder &other) = delete;  // no copy constructor
+    UCharsTrieBuilder &operator=(const UCharsTrieBuilder &other) = delete;  // no assignment operator
+
+    void buildUChars(UStringTrieBuildOption buildOption, UErrorCode &errorCode);
+
+    virtual int32_t getElementStringLength(int32_t i) const override;
+    virtual char16_t getElementUnit(int32_t i, int32_t unitIndex) const override;
+    virtual int32_t getElementValue(int32_t i) const override;
+
+    virtual int32_t getLimitOfLinearMatch(int32_t first, int32_t last, int32_t unitIndex) const override;
+
+    virtual int32_t countElementUnits(int32_t start, int32_t limit, int32_t unitIndex) const override;
+    virtual int32_t skipElementsBySomeUnits(int32_t i, int32_t unitIndex, int32_t count) const override;
+    virtual int32_t indexOfElementWithNextUnit(int32_t i, int32_t unitIndex, char16_t unit) const override;
+
+    virtual UBool matchNodesCanHaveValues() const override { return true; }
+
+    virtual int32_t getMaxBranchLinearSubNodeLength() const override { return UCharsTrie::kMaxBranchLinearSubNodeLength; }
+    virtual int32_t getMinLinearMatch() const override { return UCharsTrie::kMinLinearMatch; }
+    virtual int32_t getMaxLinearMatchLength() const override { return UCharsTrie::kMaxLinearMatchLength; }
+
+    class UCTLinearMatchNode : public LinearMatchNode {
+    public:
+        UCTLinearMatchNode(const char16_t *units, int32_t len, Node *nextNode);
+        virtual bool operator==(const Node &other) const override;
+        virtual void write(StringTrieBuilder &builder) override;
+    private:
+        const char16_t *s;
+    };
+
+    virtual Node *createLinearMatchNode(int32_t i, int32_t unitIndex, int32_t length,
+                                        Node *nextNode) const override;
+
+    UBool ensureCapacity(int32_t length);
+    virtual int32_t write(int32_t unit) override;
+    int32_t write(const char16_t *s, int32_t length);
+    virtual int32_t writeElementUnits(int32_t i, int32_t unitIndex, int32_t length) override;
+    virtual int32_t writeValueAndFinal(int32_t i, UBool isFinal) override;
+    virtual int32_t writeValueAndType(UBool hasValue, int32_t value, int32_t node) override;
+    virtual int32_t writeDeltaTo(int32_t jumpTarget) override;
+
+    UnicodeString strings;
+    UCharsTrieElement *elements;
+    int32_t elementsCapacity;
+    int32_t elementsLength;
+
+    // char16_t serialization of the trie.
+    // Grows from the back: ucharsLength measures from the end of the buffer!
+    char16_t *uchars;
+    int32_t ucharsCapacity;
+    int32_t ucharsLength;
+};
+
+U_NAMESPACE_END
+
+#endif /* U_SHOW_CPLUSPLUS_API */
+
+#endif  // __UCHARSTRIEBUILDER_H__

thirdparty/icu4c/common/unicode/uchriter.h

@@ -0,0 +1,393 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+/*
+**********************************************************************
+*   Copyright (C) 1998-2005, International Business Machines
+*   Corporation and others.  All Rights Reserved.
+**********************************************************************
+*/
+
+#ifndef UCHRITER_H
+#define UCHRITER_H
+
+#include "unicode/utypes.h"
+
+#if U_SHOW_CPLUSPLUS_API
+
+#include "unicode/chariter.h"
+
+/**
+ * \file 
+ * \brief C++ API: char16_t Character Iterator
+ */
+ 
+U_NAMESPACE_BEGIN
+
+/**
+ * A concrete subclass of CharacterIterator that iterates over the
+ * characters (code units or code points) in a char16_t array.
+ * It's possible not only to create an
+ * iterator that iterates over an entire char16_t array, but also to
+ * create one that iterates over only a subrange of a char16_t array
+ * (iterators over different subranges of the same char16_t array don't
+ * compare equal).
+ * @see CharacterIterator
+ * @see ForwardCharacterIterator
+ * @stable ICU 2.0
+ */
+class U_COMMON_API UCharCharacterIterator : public CharacterIterator {
+public:
+  /**
+   * Create an iterator over the char16_t array referred to by "textPtr".
+   * The iteration range is 0 to <code>length-1</code>.
+   * text is only aliased, not adopted (the
+   * destructor will not delete it).
+   * @param textPtr The char16_t array to be iterated over
+   * @param length The length of the char16_t array
+   * @stable ICU 2.0
+   */
+  UCharCharacterIterator(ConstChar16Ptr textPtr, int32_t length);
+
+  /**
+   * Create an iterator over the char16_t array referred to by "textPtr".
+   * The iteration range is 0 to <code>length-1</code>.
+   * text is only aliased, not adopted (the
+   * destructor will not delete it).
+   * The starting
+   * position is specified by "position". If "position" is outside the valid
+   * iteration range, the behavior of this object is undefined.
+   * @param textPtr The char16_t array to be iterated over
+   * @param length The length of the char16_t array
+   * @param position The starting position of the iteration
+   * @stable ICU 2.0
+   */
+  UCharCharacterIterator(ConstChar16Ptr textPtr, int32_t length,
+                         int32_t position);
+
+  /**
+   * Create an iterator over the char16_t array referred to by "textPtr".
+   * The iteration range is 0 to <code>end-1</code>.
+   * text is only aliased, not adopted (the
+   * destructor will not delete it).
+   * The starting
+   * position is specified by "position". If begin and end do not
+   * form a valid iteration range or "position" is outside the valid
+   * iteration range, the behavior of this object is undefined.
+   * @param textPtr The char16_t array to be iterated over
+   * @param length The length of the char16_t array
+   * @param textBegin  The begin position of the iteration range
+   * @param textEnd    The end position of the iteration range
+   * @param position    The starting position of the iteration
+   * @stable ICU 2.0
+   */
+  UCharCharacterIterator(ConstChar16Ptr textPtr, int32_t length,
+                         int32_t textBegin,
+                         int32_t textEnd,
+                         int32_t position);
+
+  /**
+   * Copy constructor.  The new iterator iterates over the same range
+   * of the same string as "that", and its initial position is the
+   * same as "that"'s current position.
+   * @param that The UCharCharacterIterator to be copied
+   * @stable ICU 2.0
+   */
+  UCharCharacterIterator(const UCharCharacterIterator&  that);
+
+  /**
+   * Destructor.
+   * @stable ICU 2.0
+   */
+  virtual ~UCharCharacterIterator();
+
+  /**
+   * Assignment operator.  *this is altered to iterate over the sane
+   * range of the same string as "that", and refers to the same
+   * character within that string as "that" does.
+   * @param that The object to be copied
+   * @return the newly created object
+   * @stable ICU 2.0
+   */
+  UCharCharacterIterator&
+  operator=(const UCharCharacterIterator&    that);
+
+  /**
+   * Returns true if the iterators iterate over the same range of the
+   * same string and are pointing at the same character.
+   * @param that The ForwardCharacterIterator used to be compared for equality
+   * @return true if the iterators iterate over the same range of the
+   * same string and are pointing at the same character.
+   * @stable ICU 2.0
+   */
+  virtual bool           operator==(const ForwardCharacterIterator& that) const override;
+
+  /**
+   * Generates a hash code for this iterator.
+   * @return the hash code.
+   * @stable ICU 2.0
+   */
+  virtual int32_t hashCode() const override;
+
+  /**
+   * Returns a new UCharCharacterIterator referring to the same
+   * character in the same range of the same string as this one.  The
+   * caller must delete the new iterator.
+   * @return the CharacterIterator newly created
+   * @stable ICU 2.0
+   */
+  virtual UCharCharacterIterator* clone() const override;
+
+  /**
+   * Sets the iterator to refer to the first code unit in its
+   * iteration range, and returns that code unit.
+   * This can be used to begin an iteration with next().
+   * @return the first code unit in its iteration range.
+   * @stable ICU 2.0
+   */
+  virtual char16_t first() override;
+
+  /**
+   * Sets the iterator to refer to the first code unit in its
+   * iteration range, returns that code unit, and moves the position
+   * to the second code unit. This is an alternative to setToStart()
+   * for forward iteration with nextPostInc().
+   * @return the first code unit in its iteration range
+   * @stable ICU 2.0
+   */
+  virtual char16_t firstPostInc() override;
+
+  /**
+   * Sets the iterator to refer to the first code point in its
+   * iteration range, and returns that code unit,
+   * This can be used to begin an iteration with next32().
+   * Note that an iteration with next32PostInc(), beginning with,
+   * e.g., setToStart() or firstPostInc(), is more efficient.
+   * @return the first code point in its iteration range
+   * @stable ICU 2.0
+   */
+  virtual UChar32 first32() override;
+
+  /**
+   * Sets the iterator to refer to the first code point in its
+   * iteration range, returns that code point, and moves the position
+   * to the second code point. This is an alternative to setToStart()
+   * for forward iteration with next32PostInc().
+   * @return the first code point in its iteration range.
+   * @stable ICU 2.0
+   */
+  virtual UChar32 first32PostInc() override;
+
+  /**
+   * Sets the iterator to refer to the last code unit in its
+   * iteration range, and returns that code unit.
+   * This can be used to begin an iteration with previous().
+   * @return the last code unit in its iteration range.
+   * @stable ICU 2.0
+   */
+  virtual char16_t last() override;
+
+  /**
+   * Sets the iterator to refer to the last code point in its
+   * iteration range, and returns that code unit.
+   * This can be used to begin an iteration with previous32().
+   * @return the last code point in its iteration range.
+   * @stable ICU 2.0
+   */
+  virtual UChar32 last32() override;
+
+  /**
+   * Sets the iterator to refer to the "position"-th code unit
+   * in the text-storage object the iterator refers to, and
+   * returns that code unit.
+   * @param position the position within the text-storage object
+   * @return the code unit
+   * @stable ICU 2.0
+   */
+  virtual char16_t         setIndex(int32_t position) override;
+
+  /**
+   * Sets the iterator to refer to the beginning of the code point
+   * that contains the "position"-th code unit
+   * in the text-storage object the iterator refers to, and
+   * returns that code point.
+   * The current position is adjusted to the beginning of the code point
+   * (its first code unit).
+   * @param position the position within the text-storage object
+   * @return the code unit
+   * @stable ICU 2.0
+   */
+  virtual UChar32       setIndex32(int32_t position) override;
+
+  /**
+   * Returns the code unit the iterator currently refers to.
+   * @return the code unit the iterator currently refers to.
+   * @stable ICU 2.0
+   */
+  virtual char16_t current() const override;
+
+  /**
+   * Returns the code point the iterator currently refers to.
+   * @return the code point the iterator currently refers to.
+   * @stable ICU 2.0
+   */
+  virtual UChar32 current32() const override;
+
+  /**
+   * Advances to the next code unit in the iteration range (toward
+   * endIndex()), and returns that code unit.  If there are no more
+   * code units to return, returns DONE.
+   * @return the next code unit in the iteration range.
+   * @stable ICU 2.0
+   */
+  virtual char16_t next() override;
+
+  /**
+   * Gets the current code unit for returning and advances to the next code unit
+   * in the iteration range
+   * (toward endIndex()).  If there are
+   * no more code units to return, returns DONE.
+   * @return the current code unit.
+   * @stable ICU 2.0
+   */
+  virtual char16_t nextPostInc() override;
+
+  /**
+   * Advances to the next code point in the iteration range (toward
+   * endIndex()), and returns that code point.  If there are no more
+   * code points to return, returns DONE.
+   * Note that iteration with "pre-increment" semantics is less
+   * efficient than iteration with "post-increment" semantics
+   * that is provided by next32PostInc().
+   * @return the next code point in the iteration range.
+   * @stable ICU 2.0
+   */
+  virtual UChar32 next32() override;
+
+  /**
+   * Gets the current code point for returning and advances to the next code point
+   * in the iteration range
+   * (toward endIndex()).  If there are
+   * no more code points to return, returns DONE.
+   * @return the current point.
+   * @stable ICU 2.0
+   */
+  virtual UChar32 next32PostInc() override;
+
+  /**
+   * Returns false if there are no more code units or code points
+   * at or after the current position in the iteration range.
+   * This is used with nextPostInc() or next32PostInc() in forward
+   * iteration.
+   * @return false if there are no more code units or code points
+   * at or after the current position in the iteration range.
+   * @stable ICU 2.0
+   */
+  virtual UBool        hasNext() override;
+
+  /**
+   * Advances to the previous code unit in the iteration range (toward
+   * startIndex()), and returns that code unit.  If there are no more
+   * code units to return, returns DONE.
+   * @return the previous code unit in the iteration range.
+   * @stable ICU 2.0
+   */
+  virtual char16_t previous() override;
+
+  /**
+   * Advances to the previous code point in the iteration range (toward
+   * startIndex()), and returns that code point.  If there are no more
+   * code points to return, returns DONE.
+   * @return the previous code point in the iteration range.
+   * @stable ICU 2.0
+   */
+  virtual UChar32 previous32() override;
+
+  /**
+   * Returns false if there are no more code units or code points
+   * before the current position in the iteration range.
+   * This is used with previous() or previous32() in backward
+   * iteration.
+   * @return false if there are no more code units or code points
+   * before the current position in the iteration range.
+   * @stable ICU 2.0
+   */
+  virtual UBool        hasPrevious() override;
+
+  /**
+   * Moves the current position relative to the start or end of the
+   * iteration range, or relative to the current position itself.
+   * The movement is expressed in numbers of code units forward
+   * or backward by specifying a positive or negative delta.
+   * @param delta the position relative to origin. A positive delta means forward;
+   * a negative delta means backward.
+   * @param origin Origin enumeration {kStart, kCurrent, kEnd}
+   * @return the new position
+   * @stable ICU 2.0
+   */
+  virtual int32_t      move(int32_t delta, EOrigin origin) override;
+
+  /**
+   * Moves the current position relative to the start or end of the
+   * iteration range, or relative to the current position itself.
+   * The movement is expressed in numbers of code points forward
+   * or backward by specifying a positive or negative delta.
+   * @param delta the position relative to origin. A positive delta means forward;
+   * a negative delta means backward.
+   * @param origin Origin enumeration {kStart, kCurrent, kEnd}
+   * @return the new position
+   * @stable ICU 2.0
+   */
+#ifdef move32
+   // One of the system headers right now is sometimes defining a conflicting macro we don't use
+#undef move32
+#endif
+  virtual int32_t      move32(int32_t delta, EOrigin origin) override;
+
+  /**
+   * Sets the iterator to iterate over a new range of text
+   * @stable ICU 2.0
+   */
+  void setText(ConstChar16Ptr newText, int32_t newTextLength);
+
+  /**
+   * Copies the char16_t array under iteration into the UnicodeString
+   * referred to by "result".  Even if this iterator iterates across
+   * only a part of this string, the whole string is copied.
+   * @param result Receives a copy of the text under iteration.
+   * @stable ICU 2.0
+   */
+  virtual void            getText(UnicodeString& result) override;
+
+  /**
+   * Return a class ID for this class (not really public)
+   * @return a class ID for this class
+   * @stable ICU 2.0
+   */
+  static UClassID U_EXPORT2 getStaticClassID();
+
+  /**
+   * Return a class ID for this object (not really public)
+   * @return a class ID for this object.
+   * @stable ICU 2.0
+   */
+  virtual UClassID getDynamicClassID() const override;
+
+protected:
+  /**
+   * Protected constructor
+   * @stable ICU 2.0
+   */
+  UCharCharacterIterator();
+  /**
+   * Protected member text
+   * @stable ICU 2.0
+   */
+  const char16_t*            text;
+
+};
+
+U_NAMESPACE_END
+
+#endif /* U_SHOW_CPLUSPLUS_API */
+
+#endif

thirdparty/icu4c/common/unicode/uclean.h

@@ -0,0 +1,262 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+/*
+******************************************************************************
+* Copyright (C) 2001-2014, International Business Machines
+*                Corporation and others. All Rights Reserved.
+******************************************************************************
+*   file name:  uclean.h
+*   encoding:   UTF-8
+*   tab size:   8 (not used)
+*   indentation:4
+*
+*   created on: 2001July05
+*   created by: George Rhoten
+*/
+
+#ifndef __UCLEAN_H__
+#define __UCLEAN_H__
+
+#include "unicode/utypes.h"
+/**
+ * \file
+ * \brief C API: Initialize and clean up ICU
+ */
+ 
+/**
+ *  Initialize ICU.
+ *
+ *  Use of this function is optional.  It is OK to simply use ICU
+ *  services and functions without first having initialized
+ *  ICU by calling u_init().
+ *
+ *  u_init() will attempt to load some part of ICU's data, and is
+ *  useful as a test for configuration or installation problems that
+ *  leave the ICU data inaccessible.  A successful invocation of u_init()
+ *  does not, however, guarantee that all ICU data is accessible.
+ *
+ *  Multiple calls to u_init() cause no harm, aside from the small amount
+ *  of time required.
+ *
+ *  In old versions of ICU, u_init() was required in multi-threaded applications
+ *  to ensure the thread safety of ICU.  u_init() is no longer needed for this purpose.
+ *
+ * @param status An ICU UErrorCode parameter. It must not be <code>NULL</code>.
+ *    An Error will be returned if some required part of ICU data can not
+ *    be loaded or initialized.
+ *    The function returns immediately if the input error code indicates a
+ *    failure, as usual.
+ *
+ * @stable ICU 2.6
+ */  
+U_CAPI void U_EXPORT2 
+u_init(UErrorCode *status);
+
+#ifndef U_HIDE_SYSTEM_API
+/**
+ * Clean up the system resources, such as allocated memory or open files,
+ * used in all ICU libraries. This will free/delete all memory owned by the
+ * ICU libraries, and return them to their original load state. All open ICU
+ * items (collators, resource bundles, converters, etc.) must be closed before
+ * calling this function, otherwise ICU may not free its allocated memory
+ * (e.g. close your converters and resource bundles before calling this
+ * function). Generally, this function should be called once just before
+ * an application exits. For applications that dynamically load and unload
+ * the ICU libraries (relatively uncommon), u_cleanup() should be called
+ * just before the library unload.
+ * <p>
+ * u_cleanup() also clears any ICU heap functions, mutex functions or
+ * trace functions that may have been set for the process.  
+ * This has the effect of restoring ICU to its initial condition, before
+ * any of these override functions were installed.  Refer to
+ * u_setMemoryFunctions(), u_setMutexFunctions and 
+ * utrace_setFunctions().  If ICU is to be reinitialized after
+ * calling u_cleanup(), these runtime override functions will need to
+ * be set up again if they are still required.
+ * <p>
+ * u_cleanup() is not thread safe.  All other threads should stop using ICU
+ * before calling this function.
+ * <p>
+ * Any open ICU items will be left in an undefined state by u_cleanup(),
+ * and any subsequent attempt to use such an item will give unpredictable
+ * results.
+ * <p>
+ * After calling u_cleanup(), an application may continue to use ICU by
+ * calling u_init().  An application must invoke u_init() first from one single
+ * thread before allowing other threads call u_init().  All threads existing
+ * at the time of the first thread's call to u_init() must also call
+ * u_init() themselves before continuing with other ICU operations.  
+ * <p>
+ * The use of u_cleanup() just before an application terminates is optional,
+ * but it should be called only once for performance reasons. The primary
+ * benefit is to eliminate reports of memory or resource leaks originating
+ * in ICU code from the results generated by heap analysis tools.
+ * <p>
+ * <strong>Use this function with great care!</strong>
+ * </p>
+ *
+ * @stable ICU 2.0
+ * @system
+ */
+U_CAPI void U_EXPORT2 
+u_cleanup(void);
+
+U_CDECL_BEGIN
+/**
+  *  Pointer type for a user supplied memory allocation function.
+  *  @param context user supplied value, obtained from u_setMemoryFunctions().
+  *  @param size    The number of bytes to be allocated
+  *  @return        Pointer to the newly allocated memory, or NULL if the allocation failed.
+  *  @stable ICU 2.8
+  *  @system
+  */
+typedef void *U_CALLCONV UMemAllocFn(const void *context, size_t size);
+/**
+  *  Pointer type for a user supplied memory re-allocation function.
+  *  @param context user supplied value, obtained from u_setMemoryFunctions().
+  *  @param mem     Pointer to the memory block to be resized.
+  *  @param size    The new size for the block.
+  *  @return        Pointer to the newly allocated memory, or NULL if the allocation failed.
+  *  @stable ICU 2.8
+  *  @system
+  */
+typedef void *U_CALLCONV UMemReallocFn(const void *context, void *mem, size_t size);
+/**
+  *  Pointer type for a user supplied memory free  function.  Behavior should be
+  *  similar the standard C library free().
+  *  @param context user supplied value, obtained from u_setMemoryFunctions().
+  *  @param mem     Pointer to the memory block to be freed.
+  *  @return        Pointer to the resized memory block, or NULL if the resizing failed.
+  *  @stable ICU 2.8
+  *  @system
+  */
+typedef void  U_CALLCONV UMemFreeFn (const void *context, void *mem);
+
+/**
+ *  Set the functions that ICU will use for memory allocation.
+ *  Use of this function is optional; by default (without this function), ICU will
+ *  use the standard C library malloc() and free() functions.
+ *  This function can only be used when ICU is in an initial, unused state, before
+ *  u_init() has been called.
+ *  @param context This pointer value will be saved, and then (later) passed as
+ *                 a parameter to the memory functions each time they
+ *                 are called.
+ *  @param a       Pointer to a user-supplied malloc function.
+ *  @param r       Pointer to a user-supplied realloc function.
+ *  @param f       Pointer to a user-supplied free function.
+ *  @param status  Receives error values.
+ *  @stable ICU 2.8
+ *  @system
+ */  
+U_CAPI void U_EXPORT2 
+u_setMemoryFunctions(const void *context, UMemAllocFn * U_CALLCONV_FPTR a, UMemReallocFn * U_CALLCONV_FPTR r, UMemFreeFn * U_CALLCONV_FPTR f, 
+                    UErrorCode *status);
+
+U_CDECL_END
+
+#ifndef U_HIDE_DEPRECATED_API
+/*********************************************************************************
+ *
+ * Deprecated Functions
+ *
+ *    The following functions for user supplied mutexes are no longer supported.
+ *    Any attempt to use them will return a U_UNSUPPORTED_ERROR.
+ *
+ **********************************************************************************/
+
+/**
+  * An opaque pointer type that represents an ICU mutex.
+  * For user-implemented mutexes, the value will typically point to a
+  *  struct or object that implements the mutex.
+  * @deprecated ICU 52. This type is no longer supported.
+  * @system
+  */
+typedef void *UMTX;
+
+U_CDECL_BEGIN
+/**
+  *  Function Pointer type for a user supplied mutex initialization function.
+  *  The user-supplied function will be called by ICU whenever ICU needs to create a
+  *  new mutex.  The function implementation should create a mutex, and store a pointer
+  *  to something that uniquely identifies the mutex into the UMTX that is supplied
+  *  as a parameter.
+  *  @param context user supplied value, obtained from u_setMutexFunctions().
+  *  @param mutex   Receives a pointer that identifies the new mutex.
+  *                 The mutex init function must set the UMTX to a non-null value.   
+  *                 Subsequent calls by ICU to lock, unlock, or destroy a mutex will 
+  *                 identify the mutex by the UMTX value.
+  *  @param status  Error status.  Report errors back to ICU by setting this variable
+  *                 with an error code.
+  *  @deprecated ICU 52. This function is no longer supported.
+  *  @system
+  */
+typedef void U_CALLCONV UMtxInitFn (const void *context, UMTX  *mutex, UErrorCode* status);
+
+
+/**
+  *  Function Pointer type for a user supplied mutex functions.
+  *  One of the  user-supplied functions with this signature will be called by ICU
+  *  whenever ICU needs to lock, unlock, or destroy a mutex.
+  *  @param context user supplied value, obtained from u_setMutexFunctions().
+  *  @param mutex   specify the mutex on which to operate.
+  *  @deprecated ICU 52. This function is no longer supported.
+  *  @system
+  */
+typedef void U_CALLCONV UMtxFn   (const void *context, UMTX  *mutex);
+U_CDECL_END
+
+/**
+  *  Set the functions that ICU will use for mutex operations
+  *  Use of this function is optional; by default (without this function), ICU will
+  *  directly access system functions for mutex operations
+  *  This function can only be used when ICU is in an initial, unused state, before
+  *  u_init() has been called.
+  *  @param context This pointer value will be saved, and then (later) passed as
+  *                 a parameter to the user-supplied mutex functions each time they
+  *                 are called. 
+  *  @param init    Pointer to a mutex initialization function.  Must be non-null.
+  *  @param destroy Pointer to the mutex destroy function.  Must be non-null.
+  *  @param lock    pointer to the mutex lock function.  Must be non-null.
+  *  @param unlock  Pointer to the mutex unlock function.  Must be non-null.
+  *  @param status  Receives error values.
+  *  @deprecated ICU 52. This function is no longer supported.
+  *  @system
+  */  
+U_DEPRECATED void U_EXPORT2 
+u_setMutexFunctions(const void *context, UMtxInitFn *init, UMtxFn *destroy, UMtxFn *lock, UMtxFn *unlock,
+                    UErrorCode *status);
+
+
+/**
+  *  Pointer type for a user supplied atomic increment or decrement function.
+  *  @param context user supplied value, obtained from u_setAtomicIncDecFunctions().
+  *  @param p   Pointer to a 32 bit int to be incremented or decremented
+  *  @return    The value of the variable after the inc or dec operation.
+  *  @deprecated ICU 52. This function is no longer supported.
+  *  @system
+  */
+typedef int32_t U_CALLCONV UMtxAtomicFn(const void *context, int32_t *p);
+
+/**
+ *  Set the functions that ICU will use for atomic increment and decrement of int32_t values.
+ *  Use of this function is optional; by default (without this function), ICU will
+ *  use its own internal implementation of atomic increment/decrement.
+ *  This function can only be used when ICU is in an initial, unused state, before
+ *  u_init() has been called.
+ *  @param context This pointer value will be saved, and then (later) passed as
+ *                 a parameter to the increment and decrement functions each time they
+ *                 are called.  This function can only be called 
+ *  @param inc     Pointer to a function to do an atomic increment operation.  Must be non-null.
+ *  @param dec     Pointer to a function to do an atomic decrement operation.  Must be non-null.
+ *  @param status  Receives error values.
+ *  @deprecated ICU 52. This function is no longer supported.
+ *  @system
+ */  
+U_DEPRECATED void U_EXPORT2 
+u_setAtomicIncDecFunctions(const void *context, UMtxAtomicFn *inc, UMtxAtomicFn *dec,
+                    UErrorCode *status);
+
+#endif  /* U_HIDE_DEPRECATED_API */
+#endif  /* U_HIDE_SYSTEM_API */
+
+#endif

thirdparty/icu4c/common/unicode/ucnv_cb.h

@@ -0,0 +1,164 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+/*
+**********************************************************************
+*   Copyright (C) 2000-2004, International Business Machines
+*   Corporation and others.  All Rights Reserved.
+**********************************************************************
+ *  ucnv_cb.h:
+ *  External APIs for the ICU's codeset conversion library
+ *  Helena Shih
+ * 
+ * Modification History:
+ *
+ *   Date        Name        Description
+ */
+
+/**
+ * \file
+ * \brief C API: UConverter functions to aid the writers of callbacks
+ *
+ * <h2> Callback API for UConverter </h2>
+ * 
+ * These functions are provided here for the convenience of the callback
+ * writer. If you are just looking for callback functions to use, please
+ * see ucnv_err.h.  DO NOT call these functions directly when you are 
+ * working with converters, unless your code has been called as a callback
+ * via ucnv_setFromUCallback or ucnv_setToUCallback !!
+ * 
+ * A note about error codes and overflow.  Unlike other ICU functions,
+ * these functions do not expect the error status to be U_ZERO_ERROR.
+ * Callbacks must be much more careful about their error codes.
+ * The error codes used here are in/out parameters, which should be passed
+ * back in the callback's error parameter.
+ * 
+ * For example, if you call ucnv_cbfromUWriteBytes to write data out 
+ * to the output codepage, it may return U_BUFFER_OVERFLOW_ERROR if 
+ * the data did not fit in the target. But this isn't a failing error, 
+ * in fact, ucnv_cbfromUWriteBytes may be called AGAIN with the error
+ * status still U_BUFFER_OVERFLOW_ERROR to attempt to write further bytes,
+ * which will also go into the internal overflow buffers.
+ * 
+ * Concerning offsets, the 'offset' parameters here are relative to the start
+ * of SOURCE.  For example, Suppose the string "ABCD" was being converted 
+ * from Unicode into a codepage which doesn't have a mapping for 'B'.
+ * 'A' will be written out correctly, but
+ * The FromU Callback will be called on an unassigned character for 'B'.
+ * At this point, this is the state of the world:
+ *    Target:    A [..]     [points after A]
+ *    Source:  A B [C] D    [points to C - B has been consumed]
+ *             0 1  2  3 
+ *    codePoint = "B"       [the unassigned codepoint] 
+ * 
+ * Now, suppose a callback wants to write the substitution character '?' to
+ * the target. It calls ucnv_cbFromUWriteBytes() to write the ?. 
+ * It should pass ZERO as the offset, because the offset as far as the 
+ * callback is concerned is relative to the SOURCE pointer [which points 
+ * before 'C'.]  If the callback goes into the args and consumes 'C' also,
+ * it would call FromUWriteBytes with an offset of 1 (and advance the source
+ * pointer).
+ *
+ */
+
+#ifndef UCNV_CB_H
+#define UCNV_CB_H
+
+#include "unicode/utypes.h"
+
+#if !UCONFIG_NO_CONVERSION
+
+#include "unicode/ucnv.h"
+#include "unicode/ucnv_err.h"
+
+/**
+ * ONLY used by FromU callback functions.
+ * Writes out the specified byte output bytes to the target byte buffer or to converter internal buffers.
+ *
+ * @param args callback fromUnicode arguments
+ * @param source source bytes to write
+ * @param length length of bytes to write
+ * @param offsetIndex the relative offset index from callback.
+ * @param err error status. If <TT>U_BUFFER_OVERFLOW</TT> is returned, then U_BUFFER_OVERFLOW <STRONG>must</STRONG> 
+ * be returned to the user, because it means that not all data could be written into the target buffer, and some is 
+ * in the converter error buffer.
+ * @see ucnv_cbFromUWriteSub
+ * @stable ICU 2.0
+ */
+U_CAPI void U_EXPORT2
+ucnv_cbFromUWriteBytes (UConverterFromUnicodeArgs *args,
+                        const char* source,
+                        int32_t length,
+                        int32_t offsetIndex,
+                        UErrorCode * err);
+
+/**
+ * ONLY used by FromU callback functions.  
+ * This function will write out the correct substitution character sequence 
+ * to the target.
+ *
+ * @param args callback fromUnicode arguments
+ * @param offsetIndex the relative offset index from the current source pointer to be used
+ * @param err error status. If <TT>U_BUFFER_OVERFLOW</TT> is returned, then U_BUFFER_OVERFLOW <STRONG>must</STRONG> 
+ * be returned to the user, because it means that not all data could be written into the target buffer, and some is 
+ * in the converter error buffer.
+ * @see ucnv_cbFromUWriteBytes
+ * @stable ICU 2.0
+ */
+U_CAPI void U_EXPORT2 
+ucnv_cbFromUWriteSub (UConverterFromUnicodeArgs *args,
+                      int32_t offsetIndex,
+                      UErrorCode * err);
+
+/**
+ * ONLY used by fromU callback functions.  
+ * This function will write out the error character(s) to the target UChar buffer.
+ *
+ * @param args callback fromUnicode arguments
+ * @param source pointer to pointer to first UChar to write [on exit: 1 after last UChar processed]
+ * @param sourceLimit pointer after last UChar to write
+ * @param offsetIndex the relative offset index from callback which will be set
+ * @param err error status <TT>U_BUFFER_OVERFLOW</TT>
+ * @see ucnv_cbToUWriteSub
+ * @stable ICU 2.0
+ */
+U_CAPI void U_EXPORT2 ucnv_cbFromUWriteUChars(UConverterFromUnicodeArgs *args,
+                             const UChar** source,
+                             const UChar*  sourceLimit,
+                             int32_t offsetIndex,
+                             UErrorCode * err);
+
+/**
+ * ONLY used by ToU callback functions.
+ *  This function will write out the specified characters to the target 
+ * UChar buffer.
+ *
+ * @param args callback toUnicode arguments
+ * @param source source string to write
+ * @param length the length of source string
+ * @param offsetIndex the relative offset index which will be written.
+ * @param err error status <TT>U_BUFFER_OVERFLOW</TT>
+ * @see ucnv_cbToUWriteSub
+ * @stable ICU 2.0
+ */
+U_CAPI void U_EXPORT2 ucnv_cbToUWriteUChars (UConverterToUnicodeArgs *args,
+                                             const UChar* source,
+                                             int32_t length,
+                                             int32_t offsetIndex,
+                                             UErrorCode * err);
+
+/**
+ * ONLY used by ToU  callback functions.  
+ * This function will write out the Unicode substitution character (U+FFFD).
+ *
+ * @param args callback fromUnicode arguments
+ * @param offsetIndex the relative offset index from callback.
+ * @param err error status <TT>U_BUFFER_OVERFLOW</TT>
+ * @see ucnv_cbToUWriteUChars
+ * @stable ICU 2.0
+ */
+U_CAPI void U_EXPORT2 ucnv_cbToUWriteSub (UConverterToUnicodeArgs *args,
+                       int32_t offsetIndex,
+                       UErrorCode * err);
+#endif
+
+#endif

thirdparty/icu4c/common/unicode/ucnv_err.h

@@ -0,0 +1,465 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+/*
+**********************************************************************
+*   Copyright (C) 1999-2009, International Business Machines
+*   Corporation and others.  All Rights Reserved.
+**********************************************************************
+ *
+ *
+ *   ucnv_err.h:
+ */
+
+/**
+ * \file
+ * \brief C API: UConverter predefined error callbacks
+ *
+ *  <h2>Error Behaviour Functions</h2>
+ *  Defines some error behaviour functions called by ucnv_{from,to}Unicode
+ *  These are provided as part of ICU and many are stable, but they
+ *  can also be considered only as an example of what can be done with
+ *  callbacks.  You may of course write your own.
+ *
+ *  If you want to write your own, you may also find the functions from
+ *  ucnv_cb.h useful when writing your own callbacks.
+ *
+ *  These functions, although public, should NEVER be called directly.
+ *  They should be used as parameters to the ucnv_setFromUCallback
+ *  and ucnv_setToUCallback functions, to set the behaviour of a converter
+ *  when it encounters ILLEGAL/UNMAPPED/INVALID sequences.
+ *
+ *  usage example:  'STOP' doesn't need any context, but newContext
+ *    could be set to something other than 'NULL' if needed. The available
+ *    contexts in this header can modify the default behavior of the callback.
+ *
+ *  \code
+ *  UErrorCode err = U_ZERO_ERROR;
+ *  UConverter *myConverter = ucnv_open("ibm-949", &err);
+ *  const void *oldContext;
+ *  UConverterFromUCallback oldAction;
+ *
+ *
+ *  if (U_SUCCESS(err))
+ *  {
+ *      ucnv_setFromUCallBack(myConverter,
+ *                       UCNV_FROM_U_CALLBACK_STOP,
+ *                       NULL,
+ *                       &oldAction,
+ *                       &oldContext,
+ *                       &status);
+ *  }
+ *  \endcode
+ *
+ *  The code above tells "myConverter" to stop when it encounters an
+ *  ILLEGAL/TRUNCATED/INVALID sequences when it is used to convert from
+ *  Unicode -> Codepage. The behavior from Codepage to Unicode is not changed,
+ *  and ucnv_setToUCallBack would need to be called in order to change
+ *  that behavior too.
+ *
+ *  Here is an example with a context:
+ *
+ *  \code
+ *  UErrorCode err = U_ZERO_ERROR;
+ *  UConverter *myConverter = ucnv_open("ibm-949", &err);
+ *  const void *oldContext;
+ *  UConverterFromUCallback oldAction;
+ *
+ *
+ *  if (U_SUCCESS(err))
+ *  {
+ *      ucnv_setToUCallBack(myConverter,
+ *                       UCNV_TO_U_CALLBACK_SUBSTITUTE,
+ *                       UCNV_SUB_STOP_ON_ILLEGAL,
+ *                       &oldAction,
+ *                       &oldContext,
+ *                       &status);
+ *  }
+ *  \endcode
+ *
+ *  The code above tells "myConverter" to stop when it encounters an
+ *  ILLEGAL/TRUNCATED/INVALID sequences when it is used to convert from
+ *  Codepage -> Unicode. Any unmapped and legal characters will be
+ *  substituted to be the default substitution character.
+ */
+
+#ifndef UCNV_ERR_H
+#define UCNV_ERR_H
+
+#include "unicode/utypes.h"
+
+#if !UCONFIG_NO_CONVERSION
+
+/** Forward declaring the UConverter structure. @stable ICU 2.0 */
+struct UConverter;
+
+/** @stable ICU 2.0 */
+typedef struct UConverter UConverter;
+
+/**
+ * FROM_U, TO_U context options for sub callback
+ * @stable ICU 2.0
+ */
+#define UCNV_SUB_STOP_ON_ILLEGAL "i"
+
+/**
+ * FROM_U, TO_U context options for skip callback
+ * @stable ICU 2.0
+ */
+#define UCNV_SKIP_STOP_ON_ILLEGAL "i"
+
+/**
+ * FROM_U_CALLBACK_ESCAPE context option to escape the code unit according to ICU (%UXXXX) 
+ * @stable ICU 2.0
+ */
+#define UCNV_ESCAPE_ICU       NULL
+/**
+ * FROM_U_CALLBACK_ESCAPE context option to escape the code unit according to JAVA (\\uXXXX)
+ * @stable ICU 2.0
+ */
+#define UCNV_ESCAPE_JAVA      "J"
+/**
+ * FROM_U_CALLBACK_ESCAPE context option to escape the code unit according to C (\\uXXXX \\UXXXXXXXX)
+ * TO_U_CALLBACK_ESCAPE option to escape the character value according to C (\\xXXXX)
+ * @stable ICU 2.0
+ */
+#define UCNV_ESCAPE_C         "C"
+/**
+ * FROM_U_CALLBACK_ESCAPE context option to escape the code unit according to XML Decimal escape \htmlonly(&amp;#DDDD;)\endhtmlonly
+ * TO_U_CALLBACK_ESCAPE context option to escape the character value according to XML Decimal escape \htmlonly(&amp;#DDDD;)\endhtmlonly
+ * @stable ICU 2.0
+ */
+#define UCNV_ESCAPE_XML_DEC   "D"
+/**
+ * FROM_U_CALLBACK_ESCAPE context option to escape the code unit according to XML Hex escape \htmlonly(&amp;#xXXXX;)\endhtmlonly
+ * TO_U_CALLBACK_ESCAPE context option to escape the character value according to XML Hex escape \htmlonly(&amp;#xXXXX;)\endhtmlonly
+ * @stable ICU 2.0
+ */
+#define UCNV_ESCAPE_XML_HEX   "X"
+/**
+ * FROM_U_CALLBACK_ESCAPE context option to escape the code unit according to Unicode (U+XXXXX)
+ * @stable ICU 2.0
+ */
+#define UCNV_ESCAPE_UNICODE   "U"
+
+/**
+ * FROM_U_CALLBACK_ESCAPE context option to escape the code unit according to CSS2 conventions (\\HH..H<space>, that is,
+ * a backslash, 1..6 hex digits, and a space)
+ * @stable ICU 4.0
+ */
+#define UCNV_ESCAPE_CSS2   "S"
+
+/** 
+ * The process condition code to be used with the callbacks.  
+ * Codes which are greater than UCNV_IRREGULAR should be 
+ * passed on to any chained callbacks.
+ * @stable ICU 2.0
+ */
+typedef enum {
+    UCNV_UNASSIGNED = 0,  /**< The code point is unassigned.
+                             The error code U_INVALID_CHAR_FOUND will be set. */
+    UCNV_ILLEGAL = 1,     /**< The code point is illegal. For example, 
+                             \\x81\\x2E is illegal in SJIS because \\x2E
+                             is not a valid trail byte for the \\x81 
+                             lead byte.
+                             Also, starting with Unicode 3.0.1, non-shortest byte sequences
+                             in UTF-8 (like \\xC1\\xA1 instead of \\x61 for U+0061)
+                             are also illegal, not just irregular.
+                             The error code U_ILLEGAL_CHAR_FOUND will be set. */
+    UCNV_IRREGULAR = 2,   /**< The codepoint is not a regular sequence in 
+                             the encoding. For example, \\xED\\xA0\\x80..\\xED\\xBF\\xBF
+                             are irregular UTF-8 byte sequences for single surrogate
+                             code points.
+                             The error code U_INVALID_CHAR_FOUND will be set. */
+    UCNV_RESET = 3,       /**< The callback is called with this reason when a
+                             'reset' has occurred. Callback should reset all
+                             state. */
+    UCNV_CLOSE = 4,        /**< Called when the converter is closed. The
+                             callback should release any allocated memory.*/
+    UCNV_CLONE = 5         /**< Called when ucnv_safeClone() is called on the
+                              converter. the pointer available as the
+                              'context' is an alias to the original converters'
+                              context pointer. If the context must be owned
+                              by the new converter, the callback must clone 
+                              the data and call ucnv_setFromUCallback 
+                              (or setToUCallback) with the correct pointer.
+                              @stable ICU 2.2
+                           */
+} UConverterCallbackReason;
+
+
+/**
+ * The structure for the fromUnicode callback function parameter.
+ * @stable ICU 2.0
+ */
+typedef struct {
+    uint16_t size;              /**< The size of this struct. @stable ICU 2.0 */
+    UBool flush;                /**< The internal state of converter will be reset and data flushed if set to true. @stable ICU 2.0    */
+    UConverter *converter;      /**< Pointer to the converter that is opened and to which this struct is passed as an argument. @stable ICU 2.0  */
+    const UChar *source;        /**< Pointer to the source source buffer. @stable ICU 2.0    */
+    const UChar *sourceLimit;   /**< Pointer to the limit (end + 1) of source buffer. @stable ICU 2.0    */
+    char *target;               /**< Pointer to the target buffer. @stable ICU 2.0    */
+    const char *targetLimit;    /**< Pointer to the limit (end + 1) of target buffer. @stable ICU 2.0     */
+    int32_t *offsets;           /**< Pointer to the buffer that receives the offsets. *offset = blah ; offset++;. @stable ICU 2.0  */
+} UConverterFromUnicodeArgs;
+
+
+/**
+ * The structure for the toUnicode callback function parameter.
+ * @stable ICU 2.0
+ */
+typedef struct {
+    uint16_t size;              /**< The size of this struct   @stable ICU 2.0 */
+    UBool flush;                /**< The internal state of converter will be reset and data flushed if set to true. @stable ICU 2.0   */
+    UConverter *converter;      /**< Pointer to the converter that is opened and to which this struct is passed as an argument. @stable ICU 2.0 */
+    const char *source;         /**< Pointer to the source source buffer. @stable ICU 2.0    */
+    const char *sourceLimit;    /**< Pointer to the limit (end + 1) of source buffer. @stable ICU 2.0    */
+    UChar *target;              /**< Pointer to the target buffer. @stable ICU 2.0    */
+    const UChar *targetLimit;   /**< Pointer to the limit (end + 1) of target buffer. @stable ICU 2.0     */
+    int32_t *offsets;           /**< Pointer to the buffer that receives the offsets. *offset = blah ; offset++;. @stable ICU 2.0  */
+} UConverterToUnicodeArgs;
+
+
+/**
+ * DO NOT CALL THIS FUNCTION DIRECTLY!
+ * This From Unicode callback STOPS at the ILLEGAL_SEQUENCE,
+ * returning the error code back to the caller immediately.
+ *
+ * @param context Pointer to the callback's private data
+ * @param fromUArgs Information about the conversion in progress
+ * @param codeUnits Points to 'length' UChars of the concerned Unicode sequence
+ * @param length Size (in bytes) of the concerned codepage sequence
+ * @param codePoint Single UChar32 (UTF-32) containing the concerend Unicode codepoint.
+ * @param reason Defines the reason the callback was invoked
+ * @param err This should always be set to a failure status prior to calling.
+ * @stable ICU 2.0
+ */
+U_CAPI void U_EXPORT2 UCNV_FROM_U_CALLBACK_STOP (
+                  const void *context,
+                  UConverterFromUnicodeArgs *fromUArgs,
+                  const UChar* codeUnits,
+                  int32_t length,
+                  UChar32 codePoint,
+                  UConverterCallbackReason reason,
+                  UErrorCode * err);
+
+
+
+/**
+ * DO NOT CALL THIS FUNCTION DIRECTLY!
+ * This To Unicode callback STOPS at the ILLEGAL_SEQUENCE,
+ * returning the error code back to the caller immediately.
+ *
+ * @param context Pointer to the callback's private data
+ * @param toUArgs Information about the conversion in progress
+ * @param codeUnits Points to 'length' bytes of the concerned codepage sequence
+ * @param length Size (in bytes) of the concerned codepage sequence
+ * @param reason Defines the reason the callback was invoked
+ * @param err This should always be set to a failure status prior to calling.
+ * @stable ICU 2.0
+ */
+U_CAPI void U_EXPORT2 UCNV_TO_U_CALLBACK_STOP (
+                  const void *context,
+                  UConverterToUnicodeArgs *toUArgs,
+                  const char* codeUnits,
+                  int32_t length,
+                  UConverterCallbackReason reason,
+                  UErrorCode * err);
+
+/**
+ * DO NOT CALL THIS FUNCTION DIRECTLY!
+ * This From Unicode callback skips any ILLEGAL_SEQUENCE, or
+ * skips only UNASSIGNED_SEQUENCE depending on the context parameter
+ * simply ignoring those characters. 
+ *
+ * @param context  The function currently recognizes the callback options:
+ *                 UCNV_SKIP_STOP_ON_ILLEGAL: STOPS at the ILLEGAL_SEQUENCE,
+ *                      returning the error code back to the caller immediately.
+ *                 NULL: Skips any ILLEGAL_SEQUENCE
+ * @param fromUArgs Information about the conversion in progress
+ * @param codeUnits Points to 'length' UChars of the concerned Unicode sequence
+ * @param length Size (in bytes) of the concerned codepage sequence
+ * @param codePoint Single UChar32 (UTF-32) containing the concerend Unicode codepoint.
+ * @param reason Defines the reason the callback was invoked
+ * @param err Return value will be set to success if the callback was handled,
+ *      otherwise this value will be set to a failure status.
+ * @stable ICU 2.0
+ */
+U_CAPI void U_EXPORT2 UCNV_FROM_U_CALLBACK_SKIP (
+                  const void *context,
+                  UConverterFromUnicodeArgs *fromUArgs,
+                  const UChar* codeUnits,
+                  int32_t length,
+                  UChar32 codePoint,
+                  UConverterCallbackReason reason,
+                  UErrorCode * err);
+
+/**
+ * DO NOT CALL THIS FUNCTION DIRECTLY!
+ * This From Unicode callback will Substitute the ILLEGAL SEQUENCE, or 
+ * UNASSIGNED_SEQUENCE depending on context parameter, with the
+ * current substitution string for the converter. This is the default
+ * callback.
+ *
+ * @param context The function currently recognizes the callback options:
+ *                 UCNV_SUB_STOP_ON_ILLEGAL: STOPS at the ILLEGAL_SEQUENCE,
+ *                      returning the error code back to the caller immediately.
+ *                 NULL: Substitutes any ILLEGAL_SEQUENCE
+ * @param fromUArgs Information about the conversion in progress
+ * @param codeUnits Points to 'length' UChars of the concerned Unicode sequence
+ * @param length Size (in bytes) of the concerned codepage sequence
+ * @param codePoint Single UChar32 (UTF-32) containing the concerend Unicode codepoint.
+ * @param reason Defines the reason the callback was invoked
+ * @param err Return value will be set to success if the callback was handled,
+ *      otherwise this value will be set to a failure status.
+ * @see ucnv_setSubstChars
+ * @stable ICU 2.0
+ */
+U_CAPI void U_EXPORT2 UCNV_FROM_U_CALLBACK_SUBSTITUTE (
+                  const void *context,
+                  UConverterFromUnicodeArgs *fromUArgs,
+                  const UChar* codeUnits,
+                  int32_t length,
+                  UChar32 codePoint,
+                  UConverterCallbackReason reason,
+                  UErrorCode * err);
+
+/**
+ * DO NOT CALL THIS FUNCTION DIRECTLY!
+ * This From Unicode callback will Substitute the ILLEGAL SEQUENCE with the
+ * hexadecimal representation of the illegal codepoints
+ *
+ * @param context The function currently recognizes the callback options:
+ *        <ul>
+ *        <li>UCNV_ESCAPE_ICU: Substitutes the  ILLEGAL SEQUENCE with the hexadecimal 
+ *          representation in the format  %UXXXX, e.g. "%uFFFE%u00AC%uC8FE"). 
+ *          In the Event the converter doesn't support the characters {%,U}[A-F][0-9], 
+ *          it will  substitute  the illegal sequence with the substitution characters.
+ *          Note that  codeUnit(32bit int eg: unit of a surrogate pair) is represented as
+ *          %UD84D%UDC56</li>
+ *        <li>UCNV_ESCAPE_JAVA: Substitutes the  ILLEGAL SEQUENCE with the hexadecimal 
+ *          representation in the format  \\uXXXX, e.g. "\\uFFFE\\u00AC\\uC8FE"). 
+ *          In the Event the converter doesn't support the characters {\,u}[A-F][0-9], 
+ *          it will  substitute  the illegal sequence with the substitution characters.
+ *          Note that  codeUnit(32bit int eg: unit of a surrogate pair) is represented as
+ *          \\uD84D\\uDC56</li>
+ *        <li>UCNV_ESCAPE_C: Substitutes the  ILLEGAL SEQUENCE with the hexadecimal 
+ *          representation in the format  \\uXXXX, e.g. "\\uFFFE\\u00AC\\uC8FE"). 
+ *          In the Event the converter doesn't support the characters {\,u,U}[A-F][0-9], 
+ *          it will  substitute  the illegal sequence with the substitution characters.
+ *          Note that  codeUnit(32bit int eg: unit of a surrogate pair) is represented as
+ *          \\U00023456</li>
+ *        <li>UCNV_ESCAPE_XML_DEC: Substitutes the  ILLEGAL SEQUENCE with the decimal 
+ *          representation in the format \htmlonly&amp;#DDDDDDDD;, e.g. "&amp;#65534;&amp;#172;&amp;#51454;")\endhtmlonly. 
+ *          In the Event the converter doesn't support the characters {&amp;,#}[0-9], 
+ *          it will  substitute  the illegal sequence with the substitution characters.
+ *          Note that  codeUnit(32bit int eg: unit of a surrogate pair) is represented as
+ *          &amp;#144470; and Zero padding is ignored.</li>
+ *        <li>UCNV_ESCAPE_XML_HEX:Substitutes the  ILLEGAL SEQUENCE with the decimal 
+ *          representation in the format \htmlonly&amp;#xXXXX; e.g. "&amp;#xFFFE;&amp;#x00AC;&amp;#xC8FE;")\endhtmlonly. 
+ *          In the Event the converter doesn't support the characters {&,#,x}[0-9], 
+ *          it will  substitute  the illegal sequence with the substitution characters.
+ *          Note that  codeUnit(32bit int eg: unit of a surrogate pair) is represented as
+ *          \htmlonly&amp;#x23456;\endhtmlonly</li>
+ *        </ul>
+ * @param fromUArgs Information about the conversion in progress
+ * @param codeUnits Points to 'length' UChars of the concerned Unicode sequence
+ * @param length Size (in bytes) of the concerned codepage sequence
+ * @param codePoint Single UChar32 (UTF-32) containing the concerend Unicode codepoint.
+ * @param reason Defines the reason the callback was invoked
+ * @param err Return value will be set to success if the callback was handled,
+ *      otherwise this value will be set to a failure status.
+ * @stable ICU 2.0
+ */
+U_CAPI void U_EXPORT2 UCNV_FROM_U_CALLBACK_ESCAPE (
+                  const void *context,
+                  UConverterFromUnicodeArgs *fromUArgs,
+                  const UChar* codeUnits,
+                  int32_t length,
+                  UChar32 codePoint,
+                  UConverterCallbackReason reason,
+                  UErrorCode * err);
+
+
+/**
+ * DO NOT CALL THIS FUNCTION DIRECTLY!
+ * This To Unicode callback skips any ILLEGAL_SEQUENCE, or
+ * skips only UNASSIGNED_SEQUENCE depending on the context parameter
+ * simply ignoring those characters. 
+ *
+ * @param context  The function currently recognizes the callback options:
+ *                 UCNV_SKIP_STOP_ON_ILLEGAL: STOPS at the ILLEGAL_SEQUENCE,
+ *                      returning the error code back to the caller immediately.
+ *                 NULL: Skips any ILLEGAL_SEQUENCE
+ * @param toUArgs Information about the conversion in progress
+ * @param codeUnits Points to 'length' bytes of the concerned codepage sequence
+ * @param length Size (in bytes) of the concerned codepage sequence
+ * @param reason Defines the reason the callback was invoked
+ * @param err Return value will be set to success if the callback was handled,
+ *      otherwise this value will be set to a failure status.
+ * @stable ICU 2.0
+ */
+U_CAPI void U_EXPORT2 UCNV_TO_U_CALLBACK_SKIP (
+                  const void *context,
+                  UConverterToUnicodeArgs *toUArgs,
+                  const char* codeUnits,
+                  int32_t length,
+                  UConverterCallbackReason reason,
+                  UErrorCode * err);
+
+/**
+ * DO NOT CALL THIS FUNCTION DIRECTLY!
+ * This To Unicode callback will Substitute the ILLEGAL SEQUENCE,or 
+ * UNASSIGNED_SEQUENCE depending on context parameter,  with the
+ * Unicode substitution character, U+FFFD.
+ *
+ * @param context  The function currently recognizes the callback options:
+ *                 UCNV_SUB_STOP_ON_ILLEGAL: STOPS at the ILLEGAL_SEQUENCE,
+ *                      returning the error code back to the caller immediately.
+ *                 NULL: Substitutes any ILLEGAL_SEQUENCE
+ * @param toUArgs Information about the conversion in progress
+ * @param codeUnits Points to 'length' bytes of the concerned codepage sequence
+ * @param length Size (in bytes) of the concerned codepage sequence
+ * @param reason Defines the reason the callback was invoked
+ * @param err Return value will be set to success if the callback was handled,
+ *      otherwise this value will be set to a failure status.
+ * @stable ICU 2.0
+ */
+U_CAPI void U_EXPORT2 UCNV_TO_U_CALLBACK_SUBSTITUTE (
+                  const void *context,
+                  UConverterToUnicodeArgs *toUArgs,
+                  const char* codeUnits,
+                  int32_t length,
+                  UConverterCallbackReason reason,
+                  UErrorCode * err);
+
+/**
+ * DO NOT CALL THIS FUNCTION DIRECTLY!
+ * This To Unicode callback will Substitute the ILLEGAL SEQUENCE with the
+ * hexadecimal representation of the illegal bytes
+ *  (in the format  %XNN, e.g. "%XFF%X0A%XC8%X03").
+ *
+ * @param context This function currently recognizes the callback options:
+ *      UCNV_ESCAPE_ICU, UCNV_ESCAPE_JAVA, UCNV_ESCAPE_C, UCNV_ESCAPE_XML_DEC,
+ *      UCNV_ESCAPE_XML_HEX and UCNV_ESCAPE_UNICODE.
+ * @param toUArgs Information about the conversion in progress
+ * @param codeUnits Points to 'length' bytes of the concerned codepage sequence
+ * @param length Size (in bytes) of the concerned codepage sequence
+ * @param reason Defines the reason the callback was invoked
+ * @param err Return value will be set to success if the callback was handled,
+ *      otherwise this value will be set to a failure status.
+ * @stable ICU 2.0
+ */
+
+U_CAPI void U_EXPORT2 UCNV_TO_U_CALLBACK_ESCAPE (
+                  const void *context,
+                  UConverterToUnicodeArgs *toUArgs,
+                  const char* codeUnits,
+                  int32_t length,
+                  UConverterCallbackReason reason,
+                  UErrorCode * err);
+
+#endif
+
+#endif
+
+/*UCNV_ERR_H*/ 

thirdparty/icu4c/common/unicode/ucnvsel.h

@@ -0,0 +1,193 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+/*
+*******************************************************************************
+*
+*   Copyright (C) 2008-2011, International Business Machines
+*   Corporation, Google and others.  All Rights Reserved.
+*
+*******************************************************************************
+*/
+/*
+ * Author : eldawy@google.com (Mohamed Eldawy)
+ * ucnvsel.h
+ *
+ * Purpose: To generate a list of encodings capable of handling
+ * a given Unicode text
+ *
+ * Started 09-April-2008
+ */
+
+#ifndef __ICU_UCNV_SEL_H__
+#define __ICU_UCNV_SEL_H__
+
+#include "unicode/utypes.h"
+
+#if !UCONFIG_NO_CONVERSION
+
+#include "unicode/uset.h"
+#include "unicode/utf16.h"
+#include "unicode/uenum.h"
+#include "unicode/ucnv.h"
+
+#if U_SHOW_CPLUSPLUS_API
+#include "unicode/localpointer.h"
+#endif   // U_SHOW_CPLUSPLUS_API
+
+/**
+ * \file
+ * \brief C API: Encoding/charset encoding selector
+ *
+ * A converter selector is built with a set of encoding/charset names
+ * and given an input string returns the set of names of the
+ * corresponding converters which can convert the string.
+ *
+ * A converter selector can be serialized into a buffer and reopened
+ * from the serialized form.
+ */
+
+struct UConverterSelector;
+/**
+ * @{
+ * Typedef for selector data structure.
+ */
+typedef struct UConverterSelector UConverterSelector;
+/** @} */
+
+/**
+ * Open a selector.
+ * If converterListSize is 0, build for all available converters.
+ * If excludedCodePoints is NULL, don't exclude any code points.
+ *
+ * @param converterList a pointer to encoding names needed to be involved. 
+ *                      Can be NULL if converterListSize==0.
+ *                      The list and the names will be cloned, and the caller
+ *                      retains ownership of the original.
+ * @param converterListSize number of encodings in above list.
+ *                          If 0, builds a selector for all available converters.
+ * @param excludedCodePoints a set of code points to be excluded from consideration.
+ *                           That is, excluded code points in a string do not change
+ *                           the selection result. (They might be handled by a callback.)
+ *                           Use NULL to exclude nothing.
+ * @param whichSet what converter set to use? Use this to determine whether
+ *                 to consider only roundtrip mappings or also fallbacks.
+ * @param status an in/out ICU UErrorCode
+ * @return the new selector
+ *
+ * @stable ICU 4.2
+ */
+U_CAPI UConverterSelector* U_EXPORT2
+ucnvsel_open(const char* const*  converterList, int32_t converterListSize,
+             const USet* excludedCodePoints,
+             const UConverterUnicodeSet whichSet, UErrorCode* status);
+
+/**
+ * Closes a selector.
+ * If any Enumerations were returned by ucnv_select*, they become invalid.
+ * They can be closed before or after calling ucnv_closeSelector,
+ * but should never be used after the selector is closed.
+ *
+ * @see ucnv_selectForString
+ * @see ucnv_selectForUTF8
+ *
+ * @param sel selector to close
+ *
+ * @stable ICU 4.2
+ */
+U_CAPI void U_EXPORT2
+ucnvsel_close(UConverterSelector *sel);
+
+#if U_SHOW_CPLUSPLUS_API
+
+U_NAMESPACE_BEGIN
+
+/**
+ * \class LocalUConverterSelectorPointer
+ * "Smart pointer" class, closes a UConverterSelector via ucnvsel_close().
+ * For most methods see the LocalPointerBase base class.
+ *
+ * @see LocalPointerBase
+ * @see LocalPointer
+ * @stable ICU 4.4
+ */
+U_DEFINE_LOCAL_OPEN_POINTER(LocalUConverterSelectorPointer, UConverterSelector, ucnvsel_close);
+
+U_NAMESPACE_END
+
+#endif
+
+/**
+ * Open a selector from its serialized form.
+ * The buffer must remain valid and unchanged for the lifetime of the selector.
+ * This is much faster than creating a selector from scratch.
+ * Using a serialized form from a different machine (endianness/charset) is supported.
+ *
+ * @param buffer pointer to the serialized form of a converter selector;
+ *               must be 32-bit-aligned
+ * @param length the capacity of this buffer (can be equal to or larger than
+ *               the actual data length)
+ * @param status an in/out ICU UErrorCode
+ * @return the new selector
+ *
+ * @stable ICU 4.2
+ */
+U_CAPI UConverterSelector* U_EXPORT2
+ucnvsel_openFromSerialized(const void* buffer, int32_t length, UErrorCode* status);
+
+/**
+ * Serialize a selector into a linear buffer.
+ * The serialized form is portable to different machines.
+ *
+ * @param sel selector to consider
+ * @param buffer pointer to 32-bit-aligned memory to be filled with the
+ *               serialized form of this converter selector
+ * @param bufferCapacity the capacity of this buffer
+ * @param status an in/out ICU UErrorCode
+ * @return the required buffer capacity to hold serialize data (even if the call fails
+ *         with a U_BUFFER_OVERFLOW_ERROR, it will return the required capacity)
+ *
+ * @stable ICU 4.2
+ */
+U_CAPI int32_t U_EXPORT2
+ucnvsel_serialize(const UConverterSelector* sel,
+                  void* buffer, int32_t bufferCapacity, UErrorCode* status);
+
+/**
+ * Select converters that can map all characters in a UTF-16 string,
+ * ignoring the excluded code points.
+ *
+ * @param sel a selector
+ * @param s UTF-16 string
+ * @param length length of the string, or -1 if NUL-terminated
+ * @param status an in/out ICU UErrorCode
+ * @return an enumeration containing encoding names.
+ *         The returned encoding names and their order will be the same as
+ *         supplied when building the selector.
+ *
+ * @stable ICU 4.2
+ */
+U_CAPI UEnumeration * U_EXPORT2
+ucnvsel_selectForString(const UConverterSelector* sel,
+                        const UChar *s, int32_t length, UErrorCode *status);
+
+/**
+ * Select converters that can map all characters in a UTF-8 string,
+ * ignoring the excluded code points.
+ *
+ * @param sel a selector
+ * @param s UTF-8 string
+ * @param length length of the string, or -1 if NUL-terminated
+ * @param status an in/out ICU UErrorCode
+ * @return an enumeration containing encoding names.
+ *         The returned encoding names and their order will be the same as
+ *         supplied when building the selector.
+ *
+ * @stable ICU 4.2
+ */
+U_CAPI UEnumeration * U_EXPORT2
+ucnvsel_selectForUTF8(const UConverterSelector* sel,
+                      const char *s, int32_t length, UErrorCode *status);
+
+#endif  /* !UCONFIG_NO_CONVERSION */
+
+#endif  /* __ICU_UCNV_SEL_H__ */

thirdparty/icu4c/common/unicode/uconfig.h

@@ -0,0 +1,477 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+/*  
+**********************************************************************
+*   Copyright (C) 2002-2016, International Business Machines
+*   Corporation and others.  All Rights Reserved.
+**********************************************************************
+*   file name:  uconfig.h
+*   encoding:   UTF-8
+*   tab size:   8 (not used)
+*   indentation:4
+*
+*   created on: 2002sep19
+*   created by: Markus W. Scherer
+*/
+
+#ifndef __UCONFIG_H__
+#define __UCONFIG_H__
+
+
+/*!
+ * \file
+ * \brief User-configurable settings
+ *
+ * Miscellaneous switches:
+ *
+ * A number of macros affect a variety of minor aspects of ICU.
+ * Most of them used to be defined elsewhere (e.g., in utypes.h or platform.h)
+ * and moved here to make them easier to find.
+ *
+ * Switches for excluding parts of ICU library code modules:
+ *
+ * Changing these macros allows building partial, smaller libraries for special purposes.
+ * By default, all modules are built.
+ * The switches are fairly coarse, controlling large modules.
+ * Basic services cannot be turned off.
+ *
+ * Building with any of these options does not guarantee that the
+ * ICU build process will completely work. It is recommended that
+ * the ICU libraries and data be built using the normal build.
+ * At that time you should remove the data used by those services.
+ * After building the ICU data library, you should rebuild the ICU
+ * libraries with these switches customized to your needs.
+ *
+ * @stable ICU 2.4
+ */
+
+/**
+ * If this switch is defined, ICU will attempt to load a header file named "uconfig_local.h"
+ * prior to determining default settings for uconfig variables.
+ *
+ * @internal ICU 4.0
+ */
+#if defined(UCONFIG_USE_LOCAL)
+#include "uconfig_local.h"
+#endif
+
+/**
+ * \def U_DEBUG
+ * Determines whether to include debugging code.
+ * Automatically set on Windows, but most compilers do not have
+ * related predefined macros.
+ * @internal
+ */
+#ifdef U_DEBUG
+    /* Use the predefined value. */
+#elif defined(_DEBUG)
+    /*
+     * _DEBUG is defined by Visual Studio debug compilation.
+     * Do *not* test for its NDEBUG macro: It is an orthogonal macro
+     * which disables assert().
+     */
+#   define U_DEBUG 1
+# else
+#   define U_DEBUG 0
+#endif
+
+/**
+ * Determines whether to enable auto cleanup of libraries. 
+ * @internal
+ */
+#ifndef UCLN_NO_AUTO_CLEANUP
+#define UCLN_NO_AUTO_CLEANUP 1
+#endif
+
+/**
+ * \def U_DISABLE_RENAMING
+ * Determines whether to disable renaming or not.
+ * @internal
+ */
+#ifndef U_DISABLE_RENAMING
+#define U_DISABLE_RENAMING 0
+#endif
+
+/**
+ * \def U_NO_DEFAULT_INCLUDE_UTF_HEADERS
+ * Determines whether utypes.h includes utf.h, utf8.h, utf16.h and utf_old.h.
+ * utypes.h includes those headers if this macro is defined to 0.
+ * Otherwise, each those headers must be included explicitly when using one of their macros.
+ * Defaults to 0 for backward compatibility, except inside ICU.
+ * @stable ICU 49
+ */
+#ifdef U_NO_DEFAULT_INCLUDE_UTF_HEADERS
+    /* Use the predefined value. */
+#elif defined(U_COMBINED_IMPLEMENTATION) || defined(U_COMMON_IMPLEMENTATION) || defined(U_I18N_IMPLEMENTATION) || \
+      defined(U_IO_IMPLEMENTATION) || defined(U_LAYOUT_IMPLEMENTATION) || defined(U_LAYOUTEX_IMPLEMENTATION) || \
+      defined(U_TOOLUTIL_IMPLEMENTATION)
+#   define U_NO_DEFAULT_INCLUDE_UTF_HEADERS 1
+#else
+#   define U_NO_DEFAULT_INCLUDE_UTF_HEADERS 0
+#endif
+
+/**
+ * \def U_OVERRIDE_CXX_ALLOCATION
+ * Determines whether to override new and delete.
+ * ICU is normally built such that all of its C++ classes, via their UMemory base,
+ * override operators new and delete to use its internal, customizable,
+ * non-exception-throwing memory allocation functions. (Default value 1 for this macro.)
+ *
+ * This is especially important when the application and its libraries use multiple heaps.
+ * For example, on Windows, this allows the ICU DLL to be used by
+ * applications that statically link the C Runtime library.
+ *
+ * @stable ICU 2.2
+ */
+#ifndef U_OVERRIDE_CXX_ALLOCATION
+#define U_OVERRIDE_CXX_ALLOCATION 1
+#endif
+
+/**
+ * \def U_ENABLE_TRACING
+ * Determines whether to enable tracing.
+ * @internal
+ */
+#ifndef U_ENABLE_TRACING
+#define U_ENABLE_TRACING 0
+#endif
+
+/**
+ * \def UCONFIG_ENABLE_PLUGINS
+ * Determines whether to enable ICU plugins.
+ * @internal
+ */
+#ifndef UCONFIG_ENABLE_PLUGINS
+#define UCONFIG_ENABLE_PLUGINS 0
+#endif
+
+/**
+ * \def U_ENABLE_DYLOAD
+ * Whether to enable Dynamic loading in ICU.
+ * @internal
+ */
+#ifndef U_ENABLE_DYLOAD
+#define U_ENABLE_DYLOAD 1
+#endif
+
+/**
+ * \def U_CHECK_DYLOAD
+ * Whether to test Dynamic loading as an OS capability.
+ * @internal
+ */
+#ifndef U_CHECK_DYLOAD
+#define U_CHECK_DYLOAD 1
+#endif
+
+/**
+ * \def U_DEFAULT_SHOW_DRAFT
+ * Do we allow ICU users to use the draft APIs by default?
+ * @internal
+ */
+#ifndef U_DEFAULT_SHOW_DRAFT
+#define U_DEFAULT_SHOW_DRAFT 1
+#endif
+
+/*===========================================================================*/
+/* Custom icu entry point renaming                                           */
+/*===========================================================================*/
+
+/**
+ * \def U_HAVE_LIB_SUFFIX
+ * 1 if a custom library suffix is set.
+ * @internal
+ */
+#ifdef U_HAVE_LIB_SUFFIX
+    /* Use the predefined value. */
+#elif defined(U_LIB_SUFFIX_C_NAME) || defined(U_IN_DOXYGEN)
+#   define U_HAVE_LIB_SUFFIX 1
+#endif
+
+/**
+ * \def U_LIB_SUFFIX_C_NAME_STRING
+ * Defines the library suffix as a string with C syntax.
+ * @internal
+ */
+#ifdef U_LIB_SUFFIX_C_NAME_STRING
+    /* Use the predefined value. */
+#elif defined(U_LIB_SUFFIX_C_NAME)
+#   define CONVERT_TO_STRING(s) #s
+#   define U_LIB_SUFFIX_C_NAME_STRING CONVERT_TO_STRING(U_LIB_SUFFIX_C_NAME)
+#else
+#   define U_LIB_SUFFIX_C_NAME_STRING ""
+#endif
+
+/* common/i18n library switches --------------------------------------------- */
+
+/**
+ * \def UCONFIG_ONLY_COLLATION
+ * This switch turns off modules that are not needed for collation.
+ *
+ * It does not turn off legacy conversion because that is necessary
+ * for ICU to work on EBCDIC platforms (for the default converter).
+ * If you want "only collation" and do not build for EBCDIC,
+ * then you can define UCONFIG_NO_CONVERSION or UCONFIG_NO_LEGACY_CONVERSION to 1 as well.
+ *
+ * @stable ICU 2.4
+ */
+#ifndef UCONFIG_ONLY_COLLATION
+#   define UCONFIG_ONLY_COLLATION 0
+#endif
+
+#if UCONFIG_ONLY_COLLATION
+    /* common library */
+#   define UCONFIG_NO_BREAK_ITERATION 1
+#   define UCONFIG_NO_IDNA 1
+
+    /* i18n library */
+#   if UCONFIG_NO_COLLATION
+#       error Contradictory collation switches in uconfig.h.
+#   endif
+#   define UCONFIG_NO_FORMATTING 1
+#   define UCONFIG_NO_TRANSLITERATION 1
+#   define UCONFIG_NO_REGULAR_EXPRESSIONS 1
+#endif
+
+/* common library switches -------------------------------------------------- */
+
+/**
+ * \def UCONFIG_NO_FILE_IO
+ * This switch turns off all file access in the common library
+ * where file access is only used for data loading.
+ * ICU data must then be provided in the form of a data DLL (or with an
+ * equivalent way to link to the data residing in an executable,
+ * as in building a combined library with both the common library's code and
+ * the data), or via udata_setCommonData().
+ * Application data must be provided via udata_setAppData() or by using
+ * "open" functions that take pointers to data, for example ucol_openBinary().
+ *
+ * File access is not used at all in the i18n library.
+ *
+ * File access cannot be turned off for the icuio library or for the ICU
+ * test suites and ICU tools.
+ *
+ * @stable ICU 3.6
+ */
+#ifndef UCONFIG_NO_FILE_IO
+#   define UCONFIG_NO_FILE_IO 0
+#endif
+
+#if UCONFIG_NO_FILE_IO && defined(U_TIMEZONE_FILES_DIR) 
+#   error Contradictory file io switches in uconfig.h.
+#endif
+
+/**
+ * \def UCONFIG_NO_CONVERSION
+ * ICU will not completely build (compiling the tools fails) with this
+ * switch turned on.
+ * This switch turns off all converters.
+ *
+ * You may want to use this together with U_CHARSET_IS_UTF8 defined to 1
+ * in utypes.h if char* strings in your environment are always in UTF-8.
+ *
+ * @stable ICU 3.2
+ * @see U_CHARSET_IS_UTF8
+ */
+#ifndef UCONFIG_NO_CONVERSION
+#   define UCONFIG_NO_CONVERSION 0
+#endif
+
+#if UCONFIG_NO_CONVERSION
+#   define UCONFIG_NO_LEGACY_CONVERSION 1
+#endif
+
+/**
+ * \def UCONFIG_ONLY_HTML_CONVERSION
+ * This switch turns off all of the converters NOT listed in
+ * the HTML encoding standard:
+ * http://www.w3.org/TR/encoding/#names-and-labels
+ *
+ * This is not possible on EBCDIC platforms
+ * because they need ibm-37 or ibm-1047 default converters.
+ *
+ * @stable ICU 55
+ */
+#ifndef UCONFIG_ONLY_HTML_CONVERSION
+#   define UCONFIG_ONLY_HTML_CONVERSION 0
+#endif
+
+/**
+ * \def UCONFIG_NO_LEGACY_CONVERSION
+ * This switch turns off all converters except for
+ * - Unicode charsets (UTF-7/8/16/32, CESU-8, SCSU, BOCU-1)
+ * - US-ASCII
+ * - ISO-8859-1
+ *
+ * Turning off legacy conversion is not possible on EBCDIC platforms
+ * because they need ibm-37 or ibm-1047 default converters.
+ *
+ * @stable ICU 2.4
+ */
+#ifndef UCONFIG_NO_LEGACY_CONVERSION
+#   define UCONFIG_NO_LEGACY_CONVERSION 0
+#endif
+
+/**
+ * \def UCONFIG_NO_NORMALIZATION
+ * This switch turns off normalization.
+ * It implies turning off several other services as well, for example
+ * collation and IDNA.
+ *
+ * @stable ICU 2.6
+ */
+#ifndef UCONFIG_NO_NORMALIZATION
+#   define UCONFIG_NO_NORMALIZATION 0
+#endif
+
+/**
+ * \def UCONFIG_USE_ML_PHRASE_BREAKING
+ * This switch turns on BudouX ML phrase-based line breaking, rather than using the dictionary.
+ *
+ * @internal
+ */
+#ifndef UCONFIG_USE_ML_PHRASE_BREAKING
+#   define UCONFIG_USE_ML_PHRASE_BREAKING 0
+#endif
+
+#if UCONFIG_NO_NORMALIZATION
+    /* common library */
+    /* ICU 50 CJK dictionary BreakIterator uses normalization */
+#   define UCONFIG_NO_BREAK_ITERATION 1
+    /* IDNA (UTS #46) is implemented via normalization */
+#   define UCONFIG_NO_IDNA 1
+
+    /* i18n library */
+#   if UCONFIG_ONLY_COLLATION
+#       error Contradictory collation switches in uconfig.h.
+#   endif
+#   define UCONFIG_NO_COLLATION 1
+#   define UCONFIG_NO_TRANSLITERATION 1
+#endif
+
+/**
+ * \def UCONFIG_NO_BREAK_ITERATION
+ * This switch turns off break iteration.
+ *
+ * @stable ICU 2.4
+ */
+#ifndef UCONFIG_NO_BREAK_ITERATION
+#   define UCONFIG_NO_BREAK_ITERATION 0
+#endif
+
+/**
+ * \def UCONFIG_NO_IDNA
+ * This switch turns off IDNA.
+ *
+ * @stable ICU 2.6
+ */
+#ifndef UCONFIG_NO_IDNA
+#   define UCONFIG_NO_IDNA 0
+#endif
+
+/**
+ * \def UCONFIG_MSGPAT_DEFAULT_APOSTROPHE_MODE
+ * Determines the default UMessagePatternApostropheMode.
+ * See the documentation for that enum.
+ *
+ * @stable ICU 4.8
+ */
+#ifndef UCONFIG_MSGPAT_DEFAULT_APOSTROPHE_MODE
+#   define UCONFIG_MSGPAT_DEFAULT_APOSTROPHE_MODE UMSGPAT_APOS_DOUBLE_OPTIONAL
+#endif
+
+/**
+ * \def UCONFIG_USE_WINDOWS_LCID_MAPPING_API
+ * On platforms where U_PLATFORM_HAS_WIN32_API is true, this switch determines
+ * if the Windows platform APIs are used for LCID<->Locale Name conversions.
+ * Otherwise, only the built-in ICU tables are used.
+ * 
+ * @internal ICU 64
+ */
+#ifndef UCONFIG_USE_WINDOWS_LCID_MAPPING_API
+#   define UCONFIG_USE_WINDOWS_LCID_MAPPING_API 1
+#endif
+
+/* i18n library switches ---------------------------------------------------- */
+
+/**
+ * \def UCONFIG_NO_COLLATION
+ * This switch turns off collation and collation-based string search.
+ *
+ * @stable ICU 2.4
+ */
+#ifndef UCONFIG_NO_COLLATION
+#   define UCONFIG_NO_COLLATION 0
+#endif
+
+/**
+ * \def UCONFIG_NO_FORMATTING
+ * This switch turns off formatting and calendar/timezone services.
+ *
+ * @stable ICU 2.4
+ */
+#ifndef UCONFIG_NO_FORMATTING
+#   define UCONFIG_NO_FORMATTING 0
+#endif
+
+/**
+ * \def UCONFIG_NO_MF2
+ * This switch turns off the experimental MessageFormat 2.0 API.
+ *
+ * @internal ICU 75 technology preview
+ * @deprecated This API is for technology preview only.
+ */
+#ifndef UCONFIG_NO_MF2
+#   define UCONFIG_NO_MF2 0
+#endif
+
+/**
+ * \def UCONFIG_NO_TRANSLITERATION
+ * This switch turns off transliteration.
+ *
+ * @stable ICU 2.4
+ */
+#ifndef UCONFIG_NO_TRANSLITERATION
+#   define UCONFIG_NO_TRANSLITERATION 0
+#endif
+
+/**
+ * \def UCONFIG_NO_REGULAR_EXPRESSIONS
+ * This switch turns off regular expressions.
+ *
+ * @stable ICU 2.4
+ */
+#ifndef UCONFIG_NO_REGULAR_EXPRESSIONS
+#   define UCONFIG_NO_REGULAR_EXPRESSIONS 0
+#endif
+
+/**
+ * \def UCONFIG_NO_SERVICE
+ * This switch turns off service registration.
+ *
+ * @stable ICU 3.2
+ */
+#ifndef UCONFIG_NO_SERVICE
+#   define UCONFIG_NO_SERVICE 0
+#endif
+
+/**
+ * \def UCONFIG_HAVE_PARSEALLINPUT
+ * This switch turns on the "parse all input" attribute. Binary incompatible.
+ *
+ * @internal
+ */
+#ifndef UCONFIG_HAVE_PARSEALLINPUT
+#   define UCONFIG_HAVE_PARSEALLINPUT 1
+#endif
+
+/**
+ * \def UCONFIG_NO_FILTERED_BREAK_ITERATION
+ * This switch turns off filtered break iteration code.
+ *
+ * @internal
+ */
+#ifndef UCONFIG_NO_FILTERED_BREAK_ITERATION
+#   define UCONFIG_NO_FILTERED_BREAK_ITERATION 0
+#endif
+
+#endif  // __UCONFIG_H__

thirdparty/icu4c/common/unicode/ucpmap.h

@@ -0,0 +1,158 @@
+// © 2018 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+
+// ucpmap.h
+// created: 2018sep03 Markus W. Scherer
+
+#ifndef __UCPMAP_H__
+#define __UCPMAP_H__
+
+#include "unicode/utypes.h"
+
+U_CDECL_BEGIN
+
+/**
+ * \file
+ * \brief C API: This file defines an abstract map from Unicode code points to integer values.
+ *
+ * @see UCPMap
+ * @see UCPTrie
+ * @see UMutableCPTrie
+ */
+
+/**
+ * Abstract map from Unicode code points (U+0000..U+10FFFF) to integer values.
+ *
+ * @see UCPTrie
+ * @see UMutableCPTrie
+ * @stable ICU 63
+ */
+typedef struct UCPMap UCPMap;
+
+/**
+ * Selectors for how ucpmap_getRange() etc. should report value ranges overlapping with surrogates.
+ * Most users should use UCPMAP_RANGE_NORMAL.
+ *
+ * @see ucpmap_getRange
+ * @see ucptrie_getRange
+ * @see umutablecptrie_getRange
+ * @stable ICU 63
+ */
+enum UCPMapRangeOption {
+    /**
+     * ucpmap_getRange() enumerates all same-value ranges as stored in the map.
+     * Most users should use this option.
+     * @stable ICU 63
+     */
+    UCPMAP_RANGE_NORMAL,
+    /**
+     * ucpmap_getRange() enumerates all same-value ranges as stored in the map,
+     * except that lead surrogates (U+D800..U+DBFF) are treated as having the
+     * surrogateValue, which is passed to getRange() as a separate parameter.
+     * The surrogateValue is not transformed via filter().
+     * See U_IS_LEAD(c).
+     *
+     * Most users should use UCPMAP_RANGE_NORMAL instead.
+     *
+     * This option is useful for maps that map surrogate code *units* to
+     * special values optimized for UTF-16 string processing
+     * or for special error behavior for unpaired surrogates,
+     * but those values are not to be associated with the lead surrogate code *points*.
+     * @stable ICU 63
+     */
+    UCPMAP_RANGE_FIXED_LEAD_SURROGATES,
+    /**
+     * ucpmap_getRange() enumerates all same-value ranges as stored in the map,
+     * except that all surrogates (U+D800..U+DFFF) are treated as having the
+     * surrogateValue, which is passed to getRange() as a separate parameter.
+     * The surrogateValue is not transformed via filter().
+     * See U_IS_SURROGATE(c).
+     *
+     * Most users should use UCPMAP_RANGE_NORMAL instead.
+     *
+     * This option is useful for maps that map surrogate code *units* to
+     * special values optimized for UTF-16 string processing
+     * or for special error behavior for unpaired surrogates,
+     * but those values are not to be associated with the lead surrogate code *points*.
+     * @stable ICU 63
+     */
+    UCPMAP_RANGE_FIXED_ALL_SURROGATES
+};
+#ifndef U_IN_DOXYGEN
+typedef enum UCPMapRangeOption UCPMapRangeOption;
+#endif
+
+/**
+ * Returns the value for a code point as stored in the map, with range checking.
+ * Returns an implementation-defined error value if c is not in the range 0..U+10FFFF.
+ *
+ * @param map the map
+ * @param c the code point
+ * @return the map value,
+ *         or an implementation-defined error value if the code point is not in the range 0..U+10FFFF
+ * @stable ICU 63
+ */
+U_CAPI uint32_t U_EXPORT2
+ucpmap_get(const UCPMap *map, UChar32 c);
+
+/**
+ * Callback function type: Modifies a map value.
+ * Optionally called by ucpmap_getRange()/ucptrie_getRange()/umutablecptrie_getRange().
+ * The modified value will be returned by the getRange function.
+ *
+ * Can be used to ignore some of the value bits,
+ * make a filter for one of several values,
+ * return a value index computed from the map value, etc.
+ *
+ * @param context an opaque pointer, as passed into the getRange function
+ * @param value a value from the map
+ * @return the modified value
+ * @stable ICU 63
+ */
+typedef uint32_t U_CALLCONV
+UCPMapValueFilter(const void *context, uint32_t value);
+
+/**
+ * Returns the last code point such that all those from start to there have the same value.
+ * Can be used to efficiently iterate over all same-value ranges in a map.
+ * (This is normally faster than iterating over code points and get()ting each value,
+ * but much slower than a data structure that stores ranges directly.)
+ *
+ * If the UCPMapValueFilter function pointer is not NULL, then
+ * the value to be delivered is passed through that function, and the return value is the end
+ * of the range where all values are modified to the same actual value.
+ * The value is unchanged if that function pointer is NULL.
+ *
+ * Example:
+ * \code
+ * UChar32 start = 0, end;
+ * uint32_t value;
+ * while ((end = ucpmap_getRange(map, start, UCPMAP_RANGE_NORMAL, 0,
+ *                               NULL, NULL, &value)) >= 0) {
+ *     // Work with the range start..end and its value.
+ *     start = end + 1;
+ * }
+ * \endcode
+ *
+ * @param map the map
+ * @param start range start
+ * @param option defines whether surrogates are treated normally,
+ *               or as having the surrogateValue; usually UCPMAP_RANGE_NORMAL
+ * @param surrogateValue value for surrogates; ignored if option==UCPMAP_RANGE_NORMAL
+ * @param filter a pointer to a function that may modify the map data value,
+ *     or NULL if the values from the map are to be used unmodified
+ * @param context an opaque pointer that is passed on to the filter function
+ * @param pValue if not NULL, receives the value that every code point start..end has;
+ *     may have been modified by filter(context, map value)
+ *     if that function pointer is not NULL
+ * @return the range end code point, or -1 if start is not a valid code point
+ * @stable ICU 63
+ */
+U_CAPI UChar32 U_EXPORT2
+ucpmap_getRange(const UCPMap *map, UChar32 start,
+                UCPMapRangeOption option, uint32_t surrogateValue,
+                UCPMapValueFilter *filter, const void *context, uint32_t *pValue);
+
+U_CDECL_END
+
+#endif

thirdparty/icu4c/common/unicode/ucptrie.h

@@ -0,0 +1,645 @@
+// © 2017 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+
+// ucptrie.h (modified from utrie2.h)
+// created: 2017dec29 Markus W. Scherer
+
+#ifndef __UCPTRIE_H__
+#define __UCPTRIE_H__
+
+#include "unicode/utypes.h"
+#include "unicode/ucpmap.h"
+#include "unicode/utf8.h"
+
+#if U_SHOW_CPLUSPLUS_API
+#include "unicode/localpointer.h"
+#endif   // U_SHOW_CPLUSPLUS_API
+
+U_CDECL_BEGIN
+
+/**
+ * \file
+ * \brief C API: This file defines an immutable Unicode code point trie.
+ *
+ * @see UCPTrie
+ * @see UMutableCPTrie
+ */
+
+#ifndef U_IN_DOXYGEN
+/** @internal */
+typedef union UCPTrieData {
+    /** @internal */
+    const void *ptr0;
+    /** @internal */
+    const uint16_t *ptr16;
+    /** @internal */
+    const uint32_t *ptr32;
+    /** @internal */
+    const uint8_t *ptr8;
+} UCPTrieData;
+#endif
+
+/**
+ * Immutable Unicode code point trie structure.
+ * Fast, reasonably compact, map from Unicode code points (U+0000..U+10FFFF) to integer values.
+ * For details see https://icu.unicode.org/design/struct/utrie
+ *
+ * Do not access UCPTrie fields directly; use public functions and macros.
+ * Functions are easy to use: They support all trie types and value widths.
+ *
+ * When performance is really important, macros provide faster access.
+ * Most macros are specific to either "fast" or "small" tries, see UCPTrieType.
+ * There are "fast" macros for special optimized use cases.
+ *
+ * The macros will return bogus values, or may crash, if used on the wrong type or value width.
+ *
+ * @see UMutableCPTrie
+ * @stable ICU 63
+ */
+struct UCPTrie {
+#ifndef U_IN_DOXYGEN
+    /** @internal */
+    const uint16_t *index;
+    /** @internal */
+    UCPTrieData data;
+
+    /** @internal */
+    int32_t indexLength;
+    /** @internal */
+    int32_t dataLength;
+    /** Start of the last range which ends at U+10FFFF. @internal */
+    UChar32 highStart;
+    /** highStart>>12 @internal */
+    uint16_t shifted12HighStart;
+
+    /** @internal */
+    int8_t type;  // UCPTrieType
+    /** @internal */
+    int8_t valueWidth;  // UCPTrieValueWidth
+
+    /** padding/reserved @internal */
+    uint32_t reserved32;
+    /** padding/reserved @internal */
+    uint16_t reserved16;
+
+    /**
+     * Internal index-3 null block offset.
+     * Set to an impossibly high value (e.g., 0xffff) if there is no dedicated index-3 null block.
+     * @internal
+     */
+    uint16_t index3NullOffset;
+    /**
+     * Internal data null block offset, not shifted.
+     * Set to an impossibly high value (e.g., 0xfffff) if there is no dedicated data null block.
+     * @internal
+     */
+    int32_t dataNullOffset;
+    /** @internal */
+    uint32_t nullValue;
+
+#ifdef UCPTRIE_DEBUG
+    /** @internal */
+    const char *name;
+#endif
+#endif
+};
+#ifndef U_IN_DOXYGEN
+typedef struct UCPTrie UCPTrie;
+#endif
+
+/**
+ * Selectors for the type of a UCPTrie.
+ * Different trade-offs for size vs. speed.
+ *
+ * @see umutablecptrie_buildImmutable
+ * @see ucptrie_openFromBinary
+ * @see ucptrie_getType
+ * @stable ICU 63
+ */
+enum UCPTrieType {
+    /**
+     * For ucptrie_openFromBinary() to accept any type.
+     * ucptrie_getType() will return the actual type.
+     * @stable ICU 63
+     */
+    UCPTRIE_TYPE_ANY = -1,
+    /**
+     * Fast/simple/larger BMP data structure. Use functions and "fast" macros.
+     * @stable ICU 63
+     */
+    UCPTRIE_TYPE_FAST,
+    /**
+     * Small/slower BMP data structure. Use functions and "small" macros.
+     * @stable ICU 63
+     */
+    UCPTRIE_TYPE_SMALL
+};
+#ifndef U_IN_DOXYGEN
+typedef enum UCPTrieType UCPTrieType;
+#endif
+
+/**
+ * Selectors for the number of bits in a UCPTrie data value.
+ *
+ * @see umutablecptrie_buildImmutable
+ * @see ucptrie_openFromBinary
+ * @see ucptrie_getValueWidth
+ * @stable ICU 63
+ */
+enum UCPTrieValueWidth {
+    /**
+     * For ucptrie_openFromBinary() to accept any data value width.
+     * ucptrie_getValueWidth() will return the actual data value width.
+     * @stable ICU 63
+     */
+    UCPTRIE_VALUE_BITS_ANY = -1,
+    /**
+     * The trie stores 16 bits per data value.
+     * It returns them as unsigned values 0..0xffff=65535.
+     * @stable ICU 63
+     */
+    UCPTRIE_VALUE_BITS_16,
+    /**
+     * The trie stores 32 bits per data value.
+     * @stable ICU 63
+     */
+    UCPTRIE_VALUE_BITS_32,
+    /**
+     * The trie stores 8 bits per data value.
+     * It returns them as unsigned values 0..0xff=255.
+     * @stable ICU 63
+     */
+    UCPTRIE_VALUE_BITS_8
+};
+#ifndef U_IN_DOXYGEN
+typedef enum UCPTrieValueWidth UCPTrieValueWidth;
+#endif
+
+/**
+ * Opens a trie from its binary form, stored in 32-bit-aligned memory.
+ * Inverse of ucptrie_toBinary().
+ *
+ * The memory must remain valid and unchanged as long as the trie is used.
+ * You must ucptrie_close() the trie once you are done using it.
+ *
+ * @param type selects the trie type; results in an
+ *             U_INVALID_FORMAT_ERROR if it does not match the binary data;
+ *             use UCPTRIE_TYPE_ANY to accept any type
+ * @param valueWidth selects the number of bits in a data value; results in an
+ *                  U_INVALID_FORMAT_ERROR if it does not match the binary data;
+ *                  use UCPTRIE_VALUE_BITS_ANY to accept any data value width
+ * @param data a pointer to 32-bit-aligned memory containing the binary data of a UCPTrie
+ * @param length the number of bytes available at data;
+ *               can be more than necessary
+ * @param pActualLength receives the actual number of bytes at data taken up by the trie data;
+ *                      can be NULL
+ * @param pErrorCode an in/out ICU UErrorCode
+ * @return the trie
+ *
+ * @see umutablecptrie_open
+ * @see umutablecptrie_buildImmutable
+ * @see ucptrie_toBinary
+ * @stable ICU 63
+ */
+U_CAPI UCPTrie * U_EXPORT2
+ucptrie_openFromBinary(UCPTrieType type, UCPTrieValueWidth valueWidth,
+                       const void *data, int32_t length, int32_t *pActualLength,
+                       UErrorCode *pErrorCode);
+
+/**
+ * Closes a trie and releases associated memory.
+ *
+ * @param trie the trie
+ * @stable ICU 63
+ */
+U_CAPI void U_EXPORT2
+ucptrie_close(UCPTrie *trie);
+
+/**
+ * Returns the trie type.
+ *
+ * @param trie the trie
+ * @return the trie type
+ * @see ucptrie_openFromBinary
+ * @see UCPTRIE_TYPE_ANY
+ * @stable ICU 63
+ */
+U_CAPI UCPTrieType U_EXPORT2
+ucptrie_getType(const UCPTrie *trie);
+
+/**
+ * Returns the number of bits in a trie data value.
+ *
+ * @param trie the trie
+ * @return the number of bits in a trie data value
+ * @see ucptrie_openFromBinary
+ * @see UCPTRIE_VALUE_BITS_ANY
+ * @stable ICU 63
+ */
+U_CAPI UCPTrieValueWidth U_EXPORT2
+ucptrie_getValueWidth(const UCPTrie *trie);
+
+/**
+ * Returns the value for a code point as stored in the trie, with range checking.
+ * Returns the trie error value if c is not in the range 0..U+10FFFF.
+ *
+ * Easier to use than UCPTRIE_FAST_GET() and similar macros but slower.
+ * Easier to use because, unlike the macros, this function works on all UCPTrie
+ * objects, for all types and value widths.
+ *
+ * @param trie the trie
+ * @param c the code point
+ * @return the trie value,
+ *         or the trie error value if the code point is not in the range 0..U+10FFFF
+ * @stable ICU 63
+ */
+U_CAPI uint32_t U_EXPORT2
+ucptrie_get(const UCPTrie *trie, UChar32 c);
+
+/**
+ * Returns the last code point such that all those from start to there have the same value.
+ * Can be used to efficiently iterate over all same-value ranges in a trie.
+ * (This is normally faster than iterating over code points and get()ting each value,
+ * but much slower than a data structure that stores ranges directly.)
+ *
+ * If the UCPMapValueFilter function pointer is not NULL, then
+ * the value to be delivered is passed through that function, and the return value is the end
+ * of the range where all values are modified to the same actual value.
+ * The value is unchanged if that function pointer is NULL.
+ *
+ * Example:
+ * \code
+ * UChar32 start = 0, end;
+ * uint32_t value;
+ * while ((end = ucptrie_getRange(trie, start, UCPMAP_RANGE_NORMAL, 0,
+ *                                NULL, NULL, &value)) >= 0) {
+ *     // Work with the range start..end and its value.
+ *     start = end + 1;
+ * }
+ * \endcode
+ *
+ * @param trie the trie
+ * @param start range start
+ * @param option defines whether surrogates are treated normally,
+ *               or as having the surrogateValue; usually UCPMAP_RANGE_NORMAL
+ * @param surrogateValue value for surrogates; ignored if option==UCPMAP_RANGE_NORMAL
+ * @param filter a pointer to a function that may modify the trie data value,
+ *     or NULL if the values from the trie are to be used unmodified
+ * @param context an opaque pointer that is passed on to the filter function
+ * @param pValue if not NULL, receives the value that every code point start..end has;
+ *     may have been modified by filter(context, trie value)
+ *     if that function pointer is not NULL
+ * @return the range end code point, or -1 if start is not a valid code point
+ * @stable ICU 63
+ */
+U_CAPI UChar32 U_EXPORT2
+ucptrie_getRange(const UCPTrie *trie, UChar32 start,
+                 UCPMapRangeOption option, uint32_t surrogateValue,
+                 UCPMapValueFilter *filter, const void *context, uint32_t *pValue);
+
+/**
+ * Writes a memory-mappable form of the trie into 32-bit aligned memory.
+ * Inverse of ucptrie_openFromBinary().
+ *
+ * @param trie the trie
+ * @param data a pointer to 32-bit-aligned memory to be filled with the trie data;
+ *             can be NULL if capacity==0
+ * @param capacity the number of bytes available at data, or 0 for pure preflighting
+ * @param pErrorCode an in/out ICU UErrorCode;
+ *                   U_BUFFER_OVERFLOW_ERROR if the capacity is too small
+ * @return the number of bytes written or (if buffer overflow) needed for the trie
+ *
+ * @see ucptrie_openFromBinary()
+ * @stable ICU 63
+ */
+U_CAPI int32_t U_EXPORT2
+ucptrie_toBinary(const UCPTrie *trie, void *data, int32_t capacity, UErrorCode *pErrorCode);
+
+/**
+ * Macro parameter value for a trie with 16-bit data values.
+ * Use the name of this macro as a "dataAccess" parameter in other macros.
+ * Do not use this macro in any other way.
+ *
+ * @see UCPTRIE_VALUE_BITS_16
+ * @stable ICU 63
+ */
+#define UCPTRIE_16(trie, i) ((trie)->data.ptr16[i])
+
+/**
+ * Macro parameter value for a trie with 32-bit data values.
+ * Use the name of this macro as a "dataAccess" parameter in other macros.
+ * Do not use this macro in any other way.
+ *
+ * @see UCPTRIE_VALUE_BITS_32
+ * @stable ICU 63
+ */
+#define UCPTRIE_32(trie, i) ((trie)->data.ptr32[i])
+
+/**
+ * Macro parameter value for a trie with 8-bit data values.
+ * Use the name of this macro as a "dataAccess" parameter in other macros.
+ * Do not use this macro in any other way.
+ *
+ * @see UCPTRIE_VALUE_BITS_8
+ * @stable ICU 63
+ */
+#define UCPTRIE_8(trie, i) ((trie)->data.ptr8[i])
+
+/**
+ * Returns a trie value for a code point, with range checking.
+ * Returns the trie error value if c is not in the range 0..U+10FFFF.
+ *
+ * @param trie (const UCPTrie *, in) the trie; must have type UCPTRIE_TYPE_FAST
+ * @param dataAccess UCPTRIE_16, UCPTRIE_32, or UCPTRIE_8 according to the trie’s value width
+ * @param c (UChar32, in) the input code point
+ * @return The code point's trie value.
+ * @stable ICU 63
+ */
+#define UCPTRIE_FAST_GET(trie, dataAccess, c) dataAccess(trie, _UCPTRIE_CP_INDEX(trie, 0xffff, c))
+
+/**
+ * Returns a 16-bit trie value for a code point, with range checking.
+ * Returns the trie error value if c is not in the range U+0000..U+10FFFF.
+ *
+ * @param trie (const UCPTrie *, in) the trie; must have type UCPTRIE_TYPE_SMALL
+ * @param dataAccess UCPTRIE_16, UCPTRIE_32, or UCPTRIE_8 according to the trie’s value width
+ * @param c (UChar32, in) the input code point
+ * @return The code point's trie value.
+ * @stable ICU 63
+ */
+#define UCPTRIE_SMALL_GET(trie, dataAccess, c) \
+    dataAccess(trie, _UCPTRIE_CP_INDEX(trie, UCPTRIE_SMALL_MAX, c))
+
+/**
+ * UTF-16: Reads the next code point (UChar32 c, out), post-increments src,
+ * and gets a value from the trie.
+ * Sets the trie error value if c is an unpaired surrogate.
+ *
+ * @param trie (const UCPTrie *, in) the trie; must have type UCPTRIE_TYPE_FAST
+ * @param dataAccess UCPTRIE_16, UCPTRIE_32, or UCPTRIE_8 according to the trie’s value width
+ * @param src (const UChar *, in/out) the source text pointer
+ * @param limit (const UChar *, in) the limit pointer for the text, or NULL if NUL-terminated
+ * @param c (UChar32, out) variable for the code point
+ * @param result (out) variable for the trie lookup result
+ * @stable ICU 63
+ */
+#define UCPTRIE_FAST_U16_NEXT(trie, dataAccess, src, limit, c, result) UPRV_BLOCK_MACRO_BEGIN { \
+    (c) = *(src)++; \
+    int32_t __index; \
+    if (!U16_IS_SURROGATE(c)) { \
+        __index = _UCPTRIE_FAST_INDEX(trie, c); \
+    } else { \
+        uint16_t __c2; \
+        if (U16_IS_SURROGATE_LEAD(c) && (src) != (limit) && U16_IS_TRAIL(__c2 = *(src))) { \
+            ++(src); \
+            (c) = U16_GET_SUPPLEMENTARY((c), __c2); \
+            __index = _UCPTRIE_SMALL_INDEX(trie, c); \
+        } else { \
+            __index = (trie)->dataLength - UCPTRIE_ERROR_VALUE_NEG_DATA_OFFSET; \
+        } \
+    } \
+    (result) = dataAccess(trie, __index); \
+} UPRV_BLOCK_MACRO_END
+
+/**
+ * UTF-16: Reads the previous code point (UChar32 c, out), pre-decrements src,
+ * and gets a value from the trie.
+ * Sets the trie error value if c is an unpaired surrogate.
+ *
+ * @param trie (const UCPTrie *, in) the trie; must have type UCPTRIE_TYPE_FAST
+ * @param dataAccess UCPTRIE_16, UCPTRIE_32, or UCPTRIE_8 according to the trie’s value width
+ * @param start (const UChar *, in) the start pointer for the text
+ * @param src (const UChar *, in/out) the source text pointer
+ * @param c (UChar32, out) variable for the code point
+ * @param result (out) variable for the trie lookup result
+ * @stable ICU 63
+ */
+#define UCPTRIE_FAST_U16_PREV(trie, dataAccess, start, src, c, result) UPRV_BLOCK_MACRO_BEGIN { \
+    (c) = *--(src); \
+    int32_t __index; \
+    if (!U16_IS_SURROGATE(c)) { \
+        __index = _UCPTRIE_FAST_INDEX(trie, c); \
+    } else { \
+        uint16_t __c2; \
+        if (U16_IS_SURROGATE_TRAIL(c) && (src) != (start) && U16_IS_LEAD(__c2 = *((src) - 1))) { \
+            --(src); \
+            (c) = U16_GET_SUPPLEMENTARY(__c2, (c)); \
+            __index = _UCPTRIE_SMALL_INDEX(trie, c); \
+        } else { \
+            __index = (trie)->dataLength - UCPTRIE_ERROR_VALUE_NEG_DATA_OFFSET; \
+        } \
+    } \
+    (result) = dataAccess(trie, __index); \
+} UPRV_BLOCK_MACRO_END
+
+/**
+ * UTF-8: Post-increments src and gets a value from the trie.
+ * Sets the trie error value for an ill-formed byte sequence.
+ *
+ * Unlike UCPTRIE_FAST_U16_NEXT() this UTF-8 macro does not provide the code point
+ * because it would be more work to do so and is often not needed.
+ * If the trie value differs from the error value, then the byte sequence is well-formed,
+ * and the code point can be assembled without revalidation.
+ *
+ * @param trie (const UCPTrie *, in) the trie; must have type UCPTRIE_TYPE_FAST
+ * @param dataAccess UCPTRIE_16, UCPTRIE_32, or UCPTRIE_8 according to the trie’s value width
+ * @param src (const char *, in/out) the source text pointer
+ * @param limit (const char *, in) the limit pointer for the text (must not be NULL)
+ * @param result (out) variable for the trie lookup result
+ * @stable ICU 63
+ */
+#define UCPTRIE_FAST_U8_NEXT(trie, dataAccess, src, limit, result) UPRV_BLOCK_MACRO_BEGIN { \
+    int32_t __lead = (uint8_t)*(src)++; \
+    if (!U8_IS_SINGLE(__lead)) { \
+        uint8_t __t1, __t2, __t3; \
+        if ((src) != (limit) && \
+            (__lead >= 0xe0 ? \
+                __lead < 0xf0 ?  /* U+0800..U+FFFF except surrogates */ \
+                    U8_LEAD3_T1_BITS[__lead &= 0xf] & (1 << ((__t1 = *(src)) >> 5)) && \
+                    ++(src) != (limit) && (__t2 = *(src) - 0x80) <= 0x3f && \
+                    (__lead = ((int32_t)(trie)->index[(__lead << 6) + (__t1 & 0x3f)]) + __t2, 1) \
+                :  /* U+10000..U+10FFFF */ \
+                    (__lead -= 0xf0) <= 4 && \
+                    U8_LEAD4_T1_BITS[(__t1 = *(src)) >> 4] & (1 << __lead) && \
+                    (__lead = (__lead << 6) | (__t1 & 0x3f), ++(src) != (limit)) && \
+                    (__t2 = *(src) - 0x80) <= 0x3f && \
+                    ++(src) != (limit) && (__t3 = *(src) - 0x80) <= 0x3f && \
+                    (__lead = __lead >= (trie)->shifted12HighStart ? \
+                        (trie)->dataLength - UCPTRIE_HIGH_VALUE_NEG_DATA_OFFSET : \
+                        ucptrie_internalSmallU8Index((trie), __lead, __t2, __t3), 1) \
+            :  /* U+0080..U+07FF */ \
+                __lead >= 0xc2 && (__t1 = *(src) - 0x80) <= 0x3f && \
+                (__lead = (int32_t)(trie)->index[__lead & 0x1f] + __t1, 1))) { \
+            ++(src); \
+        } else { \
+            __lead = (trie)->dataLength - UCPTRIE_ERROR_VALUE_NEG_DATA_OFFSET;  /* ill-formed*/ \
+        } \
+    } \
+    (result) = dataAccess(trie, __lead); \
+} UPRV_BLOCK_MACRO_END
+
+/**
+ * UTF-8: Pre-decrements src and gets a value from the trie.
+ * Sets the trie error value for an ill-formed byte sequence.
+ *
+ * Unlike UCPTRIE_FAST_U16_PREV() this UTF-8 macro does not provide the code point
+ * because it would be more work to do so and is often not needed.
+ * If the trie value differs from the error value, then the byte sequence is well-formed,
+ * and the code point can be assembled without revalidation.
+ *
+ * @param trie (const UCPTrie *, in) the trie; must have type UCPTRIE_TYPE_FAST
+ * @param dataAccess UCPTRIE_16, UCPTRIE_32, or UCPTRIE_8 according to the trie’s value width
+ * @param start (const char *, in) the start pointer for the text
+ * @param src (const char *, in/out) the source text pointer
+ * @param result (out) variable for the trie lookup result
+ * @stable ICU 63
+ */
+#define UCPTRIE_FAST_U8_PREV(trie, dataAccess, start, src, result) UPRV_BLOCK_MACRO_BEGIN { \
+    int32_t __index = (uint8_t)*--(src); \
+    if (!U8_IS_SINGLE(__index)) { \
+        __index = ucptrie_internalU8PrevIndex((trie), __index, (const uint8_t *)(start), \
+                                              (const uint8_t *)(src)); \
+        (src) -= __index & 7; \
+        __index >>= 3; \
+    } \
+    (result) = dataAccess(trie, __index); \
+} UPRV_BLOCK_MACRO_END
+
+/**
+ * Returns a trie value for an ASCII code point, without range checking.
+ *
+ * @param trie (const UCPTrie *, in) the trie (of either fast or small type)
+ * @param dataAccess UCPTRIE_16, UCPTRIE_32, or UCPTRIE_8 according to the trie’s value width
+ * @param c (UChar32, in) the input code point; must be U+0000..U+007F
+ * @return The ASCII code point's trie value.
+ * @stable ICU 63
+ */
+#define UCPTRIE_ASCII_GET(trie, dataAccess, c) dataAccess(trie, c)
+
+/**
+ * Returns a trie value for a BMP code point (U+0000..U+FFFF), without range checking.
+ * Can be used to look up a value for a UTF-16 code unit if other parts of
+ * the string processing check for surrogates.
+ *
+ * @param trie (const UCPTrie *, in) the trie; must have type UCPTRIE_TYPE_FAST
+ * @param dataAccess UCPTRIE_16, UCPTRIE_32, or UCPTRIE_8 according to the trie’s value width
+ * @param c (UChar32, in) the input code point, must be U+0000..U+FFFF
+ * @return The BMP code point's trie value.
+ * @stable ICU 63
+ */
+#define UCPTRIE_FAST_BMP_GET(trie, dataAccess, c) dataAccess(trie, _UCPTRIE_FAST_INDEX(trie, c))
+
+/**
+ * Returns a trie value for a supplementary code point (U+10000..U+10FFFF),
+ * without range checking.
+ *
+ * @param trie (const UCPTrie *, in) the trie; must have type UCPTRIE_TYPE_FAST
+ * @param dataAccess UCPTRIE_16, UCPTRIE_32, or UCPTRIE_8 according to the trie’s value width
+ * @param c (UChar32, in) the input code point, must be U+10000..U+10FFFF
+ * @return The supplementary code point's trie value.
+ * @stable ICU 63
+ */
+#define UCPTRIE_FAST_SUPP_GET(trie, dataAccess, c) dataAccess(trie, _UCPTRIE_SMALL_INDEX(trie, c))
+
+/* Internal definitions ----------------------------------------------------- */
+
+#ifndef U_IN_DOXYGEN
+
+/**
+ * Internal implementation constants.
+ * These are needed for the API macros, but users should not use these directly.
+ * @internal
+ */
+enum {
+    /** @internal */
+    UCPTRIE_FAST_SHIFT = 6,
+
+    /** Number of entries in a data block for code points below the fast limit. 64=0x40 @internal */
+    UCPTRIE_FAST_DATA_BLOCK_LENGTH = 1 << UCPTRIE_FAST_SHIFT,
+
+    /** Mask for getting the lower bits for the in-fast-data-block offset. @internal */
+    UCPTRIE_FAST_DATA_MASK = UCPTRIE_FAST_DATA_BLOCK_LENGTH - 1,
+
+    /** @internal */
+    UCPTRIE_SMALL_MAX = 0xfff,
+
+    /**
+     * Offset from dataLength (to be subtracted) for fetching the
+     * value returned for out-of-range code points and ill-formed UTF-8/16.
+     * @internal
+     */
+    UCPTRIE_ERROR_VALUE_NEG_DATA_OFFSET = 1,
+    /**
+     * Offset from dataLength (to be subtracted) for fetching the
+     * value returned for code points highStart..U+10FFFF.
+     * @internal
+     */
+    UCPTRIE_HIGH_VALUE_NEG_DATA_OFFSET = 2
+};
+
+/* Internal functions and macros -------------------------------------------- */
+// Do not conditionalize with #ifndef U_HIDE_INTERNAL_API, needed for public API
+
+/** @internal */
+U_CAPI int32_t U_EXPORT2
+ucptrie_internalSmallIndex(const UCPTrie *trie, UChar32 c);
+
+/** @internal */
+U_CAPI int32_t U_EXPORT2
+ucptrie_internalSmallU8Index(const UCPTrie *trie, int32_t lt1, uint8_t t2, uint8_t t3);
+
+/**
+ * Internal function for part of the UCPTRIE_FAST_U8_PREVxx() macro implementations.
+ * Do not call directly.
+ * @internal
+ */
+U_CAPI int32_t U_EXPORT2
+ucptrie_internalU8PrevIndex(const UCPTrie *trie, UChar32 c,
+                            const uint8_t *start, const uint8_t *src);
+
+/** Internal trie getter for a code point below the fast limit. Returns the data index. @internal */
+#define _UCPTRIE_FAST_INDEX(trie, c) \
+    ((int32_t)(trie)->index[(c) >> UCPTRIE_FAST_SHIFT] + ((c) & UCPTRIE_FAST_DATA_MASK))
+
+/** Internal trie getter for a code point at or above the fast limit. Returns the data index. @internal */
+#define _UCPTRIE_SMALL_INDEX(trie, c) \
+    ((c) >= (trie)->highStart ? \
+        (trie)->dataLength - UCPTRIE_HIGH_VALUE_NEG_DATA_OFFSET : \
+        ucptrie_internalSmallIndex(trie, c))
+
+/**
+ * Internal trie getter for a code point, with checking that c is in U+0000..10FFFF.
+ * Returns the data index.
+ * @internal
+ */
+#define _UCPTRIE_CP_INDEX(trie, fastMax, c) \
+    ((uint32_t)(c) <= (uint32_t)(fastMax) ? \
+        _UCPTRIE_FAST_INDEX(trie, c) : \
+        (uint32_t)(c) <= 0x10ffff ? \
+            _UCPTRIE_SMALL_INDEX(trie, c) : \
+            (trie)->dataLength - UCPTRIE_ERROR_VALUE_NEG_DATA_OFFSET)
+
+U_CDECL_END
+
+#endif  // U_IN_DOXYGEN
+
+#if U_SHOW_CPLUSPLUS_API
+
+U_NAMESPACE_BEGIN
+
+/**
+ * \class LocalUCPTriePointer
+ * "Smart pointer" class, closes a UCPTrie via ucptrie_close().
+ * For most methods see the LocalPointerBase base class.
+ *
+ * @see LocalPointerBase
+ * @see LocalPointer
+ * @stable ICU 63
+ */
+U_DEFINE_LOCAL_OPEN_POINTER(LocalUCPTriePointer, UCPTrie, ucptrie_close);
+
+U_NAMESPACE_END
+
+#endif  // U_SHOW_CPLUSPLUS_API
+
+#endif

thirdparty/icu4c/common/unicode/ucurr.h

@@ -0,0 +1,466 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+/*
+**********************************************************************
+* Copyright (c) 2002-2016, International Business Machines
+* Corporation and others.  All Rights Reserved.
+**********************************************************************
+*/
+#ifndef _UCURR_H_
+#define _UCURR_H_
+
+#include "unicode/utypes.h"
+#include "unicode/uenum.h"
+
+/**
+ * \file 
+ * \brief C API: Encapsulates information about a currency.
+ *
+ * The ucurr API encapsulates information about a currency, as defined by
+ * ISO 4217.  A currency is represented by a 3-character string
+ * containing its ISO 4217 code.  This API can return various data
+ * necessary the proper display of a currency:
+ *
+ * <ul><li>A display symbol, for a specific locale
+ * <li>The number of fraction digits to display
+ * <li>A rounding increment
+ * </ul>
+ *
+ * The <tt>DecimalFormat</tt> class uses these data to display
+ * currencies.
+ * @author Alan Liu
+ * @since ICU 2.2
+ */
+
+#if !UCONFIG_NO_FORMATTING
+
+/**
+ * Currency Usage used for Decimal Format
+ * @stable ICU 54
+ */
+enum UCurrencyUsage {
+    /**
+     * a setting to specify currency usage which determines currency digit
+     * and rounding for standard usage, for example: "50.00 NT$"
+     * used as DEFAULT value
+     * @stable ICU 54
+     */
+    UCURR_USAGE_STANDARD=0,
+    /**
+     * a setting to specify currency usage which determines currency digit
+     * and rounding for cash usage, for example: "50 NT$"
+     * @stable ICU 54
+     */
+    UCURR_USAGE_CASH=1,
+#ifndef U_HIDE_DEPRECATED_API
+    /**
+     * One higher than the last enum UCurrencyUsage constant.
+     * @deprecated ICU 58 The numeric value may change over time, see ICU ticket #12420.
+     */
+    UCURR_USAGE_COUNT=2
+#endif  // U_HIDE_DEPRECATED_API
+};
+/** Currency Usage used for Decimal Format */
+typedef enum UCurrencyUsage UCurrencyUsage; 
+
+/**
+ * Finds a currency code for the given locale.
+ * @param locale the locale for which to retrieve a currency code. 
+ *               Currency can be specified by the "currency" keyword
+ *               in which case it overrides the default currency code
+ * @param buff   fill in buffer. Can be NULL for preflighting.
+ * @param buffCapacity capacity of the fill in buffer. Can be 0 for
+ *               preflighting. If it is non-zero, the buff parameter
+ *               must not be NULL.
+ * @param ec error code
+ * @return length of the currency string. It should always be 3. If 0,
+ *                currency couldn't be found or the input values are 
+ *                invalid. 
+ * @stable ICU 2.8
+ */
+U_CAPI int32_t U_EXPORT2
+ucurr_forLocale(const char* locale,
+                UChar* buff,
+                int32_t buffCapacity,
+                UErrorCode* ec);
+
+/**
+ * Selector constants for ucurr_getName().
+ *
+ * @see ucurr_getName
+ * @stable ICU 2.6
+ */
+typedef enum UCurrNameStyle {
+    /**
+     * Selector for ucurr_getName indicating a symbolic name for a
+     * currency, such as "$" for USD.
+     * @stable ICU 2.6
+     */
+    UCURR_SYMBOL_NAME,
+
+    /**
+     * Selector for ucurr_getName indicating the long name for a
+     * currency, such as "US Dollar" for USD.
+     * @stable ICU 2.6
+     */
+    UCURR_LONG_NAME,
+
+    /**
+     * Selector for getName() indicating the narrow currency symbol.
+     * The narrow currency symbol is similar to the regular currency
+     * symbol, but it always takes the shortest form: for example,
+     * "$" instead of "US$" for USD in en-CA.
+     *
+     * @stable ICU 61
+     */
+    UCURR_NARROW_SYMBOL_NAME,
+
+    /**
+     * Selector for getName() indicating the formal currency symbol.
+     * The formal currency symbol is similar to the regular currency
+     * symbol, but it always takes the form used in formal settings
+     * such as banking; for example, "NT$" instead of "$" for TWD in zh-TW.
+     *
+     * @stable ICU 68
+     */
+    UCURR_FORMAL_SYMBOL_NAME,
+
+    /**
+     * Selector for getName() indicating the variant currency symbol.
+     * The variant symbol for a currency is an alternative symbol
+     * that is not necessarily as widely used as the regular symbol.
+     *
+     * @stable ICU 68
+     */
+    UCURR_VARIANT_SYMBOL_NAME
+    
+} UCurrNameStyle;
+
+#if !UCONFIG_NO_SERVICE
+/**
+ * @stable ICU 2.6
+ */
+typedef const void* UCurrRegistryKey;
+
+/**
+ * Register an (existing) ISO 4217 currency code for the given locale.
+ * Only the country code and the two variants EURO and PRE_EURO are
+ * recognized.
+ * @param isoCode the three-letter ISO 4217 currency code
+ * @param locale  the locale for which to register this currency code
+ * @param status the in/out status code
+ * @return a registry key that can be used to unregister this currency code, or NULL
+ * if there was an error.
+ * @stable ICU 2.6
+ */
+U_CAPI UCurrRegistryKey U_EXPORT2
+ucurr_register(const UChar* isoCode, 
+                   const char* locale,  
+                   UErrorCode* status);
+/**
+ * Unregister the previously-registered currency definitions using the
+ * URegistryKey returned from ucurr_register.  Key becomes invalid after
+ * a successful call and should not be used again.  Any currency 
+ * that might have been hidden by the original ucurr_register call is 
+ * restored.
+ * @param key the registry key returned by a previous call to ucurr_register
+ * @param status the in/out status code, no special meanings are assigned
+ * @return true if the currency for this key was successfully unregistered
+ * @stable ICU 2.6
+ */
+U_CAPI UBool U_EXPORT2
+ucurr_unregister(UCurrRegistryKey key, UErrorCode* status);
+#endif /* UCONFIG_NO_SERVICE */
+
+/**
+ * Returns the display name for the given currency in the
+ * given locale.  For example, the display name for the USD
+ * currency object in the en_US locale is "$".
+ * @param currency null-terminated 3-letter ISO 4217 code
+ * @param locale locale in which to display currency
+ * @param nameStyle selector for which kind of name to return
+ * @param isChoiceFormat always set to false, or can be NULL;
+ *     display names are static strings;
+ *     since ICU 4.4, ChoiceFormat patterns are no longer supported
+ * @param len fill-in parameter to receive length of result
+ * @param ec error code
+ * @return pointer to display string of 'len' UChars.  If the resource
+ * data contains no entry for 'currency', then 'currency' itself is
+ * returned.
+ * @stable ICU 2.6
+ */
+U_CAPI const UChar* U_EXPORT2
+ucurr_getName(const UChar* currency,
+              const char* locale,
+              UCurrNameStyle nameStyle,
+              UBool* isChoiceFormat,
+              int32_t* len,
+              UErrorCode* ec);
+
+/**
+ * Returns the plural name for the given currency in the
+ * given locale.  For example, the plural name for the USD
+ * currency object in the en_US locale is "US dollar" or "US dollars".
+ * @param currency null-terminated 3-letter ISO 4217 code
+ * @param locale locale in which to display currency
+ * @param isChoiceFormat always set to false, or can be NULL;
+ *     display names are static strings;
+ *     since ICU 4.4, ChoiceFormat patterns are no longer supported
+ * @param pluralCount plural count
+ * @param len fill-in parameter to receive length of result
+ * @param ec error code
+ * @return pointer to display string of 'len' UChars.  If the resource
+ * data contains no entry for 'currency', then 'currency' itself is
+ * returned.
+ * @stable ICU 4.2
+ */
+U_CAPI const UChar* U_EXPORT2
+ucurr_getPluralName(const UChar* currency,
+                    const char* locale,
+                    UBool* isChoiceFormat,
+                    const char* pluralCount,
+                    int32_t* len,
+                    UErrorCode* ec);
+
+/**
+ * Returns the number of the number of fraction digits that should
+ * be displayed for the given currency.
+ * This is equivalent to ucurr_getDefaultFractionDigitsForUsage(currency,UCURR_USAGE_STANDARD,ec);
+ *
+ * Important: The number of fraction digits for a given currency is NOT
+ * guaranteed to be constant across versions of ICU or CLDR. For example,
+ * do NOT use this value as a mechanism for deciding the magnitude used
+ * to store currency values in a database. You should use this value for
+ * display purposes only.
+ *
+ * @param currency null-terminated 3-letter ISO 4217 code
+ * @param ec input-output error code
+ * @return a non-negative number of fraction digits to be
+ * displayed, or 0 if there is an error
+ * @stable ICU 3.0
+ */
+U_CAPI int32_t U_EXPORT2
+ucurr_getDefaultFractionDigits(const UChar* currency,
+                               UErrorCode* ec);
+
+/**
+ * Returns the number of the number of fraction digits that should
+ * be displayed for the given currency with usage.
+ *
+ * Important: The number of fraction digits for a given currency is NOT
+ * guaranteed to be constant across versions of ICU or CLDR. For example,
+ * do NOT use this value as a mechanism for deciding the magnitude used
+ * to store currency values in a database. You should use this value for
+ * display purposes only.
+ *
+ * @param currency null-terminated 3-letter ISO 4217 code
+ * @param usage enum usage for the currency
+ * @param ec input-output error code
+ * @return a non-negative number of fraction digits to be
+ * displayed, or 0 if there is an error
+ * @stable ICU 54
+ */
+U_CAPI int32_t U_EXPORT2
+ucurr_getDefaultFractionDigitsForUsage(const UChar* currency, 
+                                       const UCurrencyUsage usage,
+                                       UErrorCode* ec);
+
+/**
+ * Returns the rounding increment for the given currency, or 0.0 if no
+ * rounding is done by the currency.
+ * This is equivalent to ucurr_getRoundingIncrementForUsage(currency,UCURR_USAGE_STANDARD,ec);
+ * @param currency null-terminated 3-letter ISO 4217 code
+ * @param ec input-output error code
+ * @return the non-negative rounding increment, or 0.0 if none,
+ * or 0.0 if there is an error
+ * @stable ICU 3.0
+ */
+U_CAPI double U_EXPORT2
+ucurr_getRoundingIncrement(const UChar* currency,
+                           UErrorCode* ec);
+
+/**
+ * Returns the rounding increment for the given currency, or 0.0 if no
+ * rounding is done by the currency given usage.
+ * @param currency null-terminated 3-letter ISO 4217 code
+ * @param usage enum usage for the currency
+ * @param ec input-output error code
+ * @return the non-negative rounding increment, or 0.0 if none,
+ * or 0.0 if there is an error
+ * @stable ICU 54
+ */
+U_CAPI double U_EXPORT2
+ucurr_getRoundingIncrementForUsage(const UChar* currency,
+                                   const UCurrencyUsage usage,
+                                   UErrorCode* ec);
+
+/**
+ * Selector constants for ucurr_openCurrencies().
+ *
+ * @see ucurr_openCurrencies
+ * @stable ICU 3.2
+ */
+typedef enum UCurrCurrencyType {
+    /**
+     * Select all ISO-4217 currency codes.
+     * @stable ICU 3.2
+     */
+    UCURR_ALL = INT32_MAX,
+    /**
+     * Select only ISO-4217 commonly used currency codes.
+     * These currencies can be found in common use, and they usually have
+     * bank notes or coins associated with the currency code.
+     * This does not include fund codes, precious metals and other
+     * various ISO-4217 codes limited to special financial products.
+     * @stable ICU 3.2
+     */
+    UCURR_COMMON = 1,
+    /**
+     * Select ISO-4217 uncommon currency codes.
+     * These codes respresent fund codes, precious metals and other
+     * various ISO-4217 codes limited to special financial products.
+     * A fund code is a monetary resource associated with a currency.
+     * @stable ICU 3.2
+     */
+    UCURR_UNCOMMON = 2,
+    /**
+     * Select only deprecated ISO-4217 codes.
+     * These codes are no longer in general public use.
+     * @stable ICU 3.2
+     */
+    UCURR_DEPRECATED = 4,
+    /**
+     * Select only non-deprecated ISO-4217 codes.
+     * These codes are in general public use.
+     * @stable ICU 3.2
+     */
+    UCURR_NON_DEPRECATED = 8
+} UCurrCurrencyType;
+
+/**
+ * Provides a UEnumeration object for listing ISO-4217 codes.
+ * @param currType You can use one of several UCurrCurrencyType values for this
+ *      variable. You can also | (or) them together to get a specific list of
+ *      currencies. Most people will want to use the (UCURR_COMMON|UCURR_NON_DEPRECATED) value to
+ *      get a list of current currencies.
+ * @param pErrorCode Error code
+ * @stable ICU 3.2
+ */
+U_CAPI UEnumeration * U_EXPORT2
+ucurr_openISOCurrencies(uint32_t currType, UErrorCode *pErrorCode);
+
+/**
+  * Queries if the given ISO 4217 3-letter code is available on the specified date range. 
+  * 
+  * Note: For checking availability of a currency on a specific date, specify the date on both 'from' and 'to' 
+  * 
+  * When 'from' is U_DATE_MIN and 'to' is U_DATE_MAX, this method checks if the specified currency is available any time. 
+  * If 'from' and 'to' are same UDate value, this method checks if the specified currency is available on that date.
+  * 
+  * @param isoCode 
+  *            The ISO 4217 3-letter code. 
+  * 
+  * @param from 
+  *            The lower bound of the date range, inclusive. When 'from' is U_DATE_MIN, check the availability 
+  *            of the currency any date before 'to' 
+  * 
+  * @param to 
+  *            The upper bound of the date range, inclusive. When 'to' is U_DATE_MAX, check the availability of 
+  *            the currency any date after 'from' 
+  * 
+  * @param errorCode 
+  *            ICU error code 
+   * 
+  * @return true if the given ISO 4217 3-letter code is supported on the specified date range. 
+  * 
+  * @stable ICU 4.8 
+  */ 
+U_CAPI UBool U_EXPORT2
+ucurr_isAvailable(const UChar* isoCode, 
+             UDate from, 
+             UDate to, 
+             UErrorCode* errorCode);
+
+/** 
+ * Finds the number of valid currency codes for the
+ * given locale and date.
+ * @param locale the locale for which to retrieve the
+ *               currency count.
+ * @param date   the date for which to retrieve the
+ *               currency count for the given locale.
+ * @param ec     error code
+ * @return       the number of currency codes for the
+ *               given locale and date.  If 0, currency
+ *               codes couldn't be found for the input
+ *               values are invalid.
+ * @stable ICU 4.0
+ */
+U_CAPI int32_t U_EXPORT2
+ucurr_countCurrencies(const char* locale, 
+                 UDate date, 
+                 UErrorCode* ec); 
+
+/** 
+ * Finds a currency code for the given locale and date 
+ * @param locale the locale for which to retrieve a currency code.  
+ *               Currency can be specified by the "currency" keyword 
+ *               in which case it overrides the default currency code 
+ * @param date   the date for which to retrieve a currency code for 
+ *               the given locale. 
+ * @param index  the index within the available list of currency codes
+ *               for the given locale on the given date.
+ * @param buff   fill in buffer. Can be NULL for preflighting. 
+ * @param buffCapacity capacity of the fill in buffer. Can be 0 for 
+ *               preflighting. If it is non-zero, the buff parameter 
+ *               must not be NULL. 
+ * @param ec     error code 
+ * @return       length of the currency string. It should always be 3. 
+ *               If 0, currency couldn't be found or the input values are  
+ *               invalid.  
+ * @stable ICU 4.0 
+ */ 
+U_CAPI int32_t U_EXPORT2 
+ucurr_forLocaleAndDate(const char* locale, 
+                UDate date, 
+                int32_t index,
+                UChar* buff, 
+                int32_t buffCapacity, 
+                UErrorCode* ec); 
+
+/**
+ * Given a key and a locale, returns an array of string values in a preferred
+ * order that would make a difference. These are all and only those values where
+ * the open (creation) of the service with the locale formed from the input locale
+ * plus input keyword and that value has different behavior than creation with the
+ * input locale alone.
+ * @param key           one of the keys supported by this service.  For now, only
+ *                      "currency" is supported.
+ * @param locale        the locale
+ * @param commonlyUsed  if set to true it will return only commonly used values
+ *                      with the given locale in preferred order.  Otherwise,
+ *                      it will return all the available values for the locale.
+ * @param status error status
+ * @return a string enumeration over keyword values for the given key and the locale.
+ * @stable ICU 4.2
+ */
+U_CAPI UEnumeration* U_EXPORT2
+ucurr_getKeywordValuesForLocale(const char* key,
+                                const char* locale,
+                                UBool commonlyUsed,
+                                UErrorCode* status);
+
+/**
+ * Returns the ISO 4217 numeric code for the currency.
+ * <p>Note: If the ISO 4217 numeric code is not assigned for the currency or
+ * the currency is unknown, this function returns 0.
+ *
+ * @param currency null-terminated 3-letter ISO 4217 code
+ * @return The ISO 4217 numeric code of the currency
+ * @stable ICU 49
+ */
+U_CAPI int32_t U_EXPORT2
+ucurr_getNumericCode(const UChar* currency);
+
+#endif /* #if !UCONFIG_NO_FORMATTING */
+
+#endif

thirdparty/icu4c/common/unicode/udata.h

@@ -0,0 +1,440 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+/*
+******************************************************************************
+*
+*   Copyright (C) 1999-2014, International Business Machines
+*   Corporation and others.  All Rights Reserved.
+*
+******************************************************************************
+*   file name:  udata.h
+*   encoding:   UTF-8
+*   tab size:   8 (not used)
+*   indentation:4
+*
+*   created on: 1999oct25
+*   created by: Markus W. Scherer
+*/
+
+#ifndef __UDATA_H__
+#define __UDATA_H__
+
+#include "unicode/utypes.h"
+
+#if U_SHOW_CPLUSPLUS_API
+#include "unicode/localpointer.h"
+#endif   // U_SHOW_CPLUSPLUS_API
+
+U_CDECL_BEGIN
+
+/**
+ * \file
+ * \brief C API: Data loading interface
+ *
+ * <h2>Information about data loading interface</h2>
+ *
+ * This API is used to find and efficiently load data for ICU and applications
+ * using ICU. It provides an abstract interface that specifies a data type and
+ * name to find and load the data. Normally this API is used by other ICU APIs
+ * to load required data out of the ICU data library, but it can be used to
+ * load data out of other places.
+ *
+ * See the User Guide Data Management chapter.
+ */
+ 
+#ifndef U_HIDE_INTERNAL_API
+/**
+ * Character used to separate package names from tree names 
+ * @internal ICU 3.0
+ */
+#define U_TREE_SEPARATOR '-'
+
+/**
+ * String used to separate package names from tree names 
+ * @internal ICU 3.0
+ */
+#define U_TREE_SEPARATOR_STRING "-"
+
+/**
+ * Character used to separate parts of entry names
+ * @internal ICU 3.0
+ */
+#define U_TREE_ENTRY_SEP_CHAR '/'
+
+/**
+ * String used to separate parts of entry names
+ * @internal ICU 3.0
+ */
+#define U_TREE_ENTRY_SEP_STRING "/"
+
+/**
+ * Alias for standard ICU data 
+ * @internal ICU 3.0
+ */
+#define U_ICUDATA_ALIAS "ICUDATA"
+
+#endif /* U_HIDE_INTERNAL_API */
+
+/**
+ * UDataInfo contains the properties about the requested data.
+ * This is meta data.
+ *
+ * <p>This structure may grow in the future, indicated by the
+ * <code>size</code> field.</p>
+ *
+ * <p>ICU data must be at least 8-aligned, and should be 16-aligned.
+ * The UDataInfo struct begins 4 bytes after the start of the data item,
+ * so it is 4-aligned.
+ *
+ * <p>The platform data property fields help determine if a data
+ * file can be efficiently used on a given machine.
+ * The particular fields are of importance only if the data
+ * is affected by the properties - if there is integer data
+ * with word sizes > 1 byte, char* text, or UChar* text.</p>
+ *
+ * <p>The implementation for the <code>udata_open[Choice]()</code>
+ * functions may reject data based on the value in <code>isBigEndian</code>.
+ * No other field is used by the <code>udata</code> API implementation.</p>
+ *
+ * <p>The <code>dataFormat</code> may be used to identify
+ * the kind of data, e.g. a converter table.</p>
+ *
+ * <p>The <code>formatVersion</code> field should be used to
+ * make sure that the format can be interpreted.
+ * It may be a good idea to check only for the one or two highest
+ * of the version elements to allow the data memory to
+ * get more or somewhat rearranged contents, for as long
+ * as the using code can still interpret the older contents.</p>
+ *
+ * <p>The <code>dataVersion</code> field is intended to be a
+ * common place to store the source version of the data;
+ * for data from the Unicode character database, this could
+ * reflect the Unicode version.</p>
+ *
+ * @stable ICU 2.0
+ */
+typedef struct {
+    /** sizeof(UDataInfo)
+     *  @stable ICU 2.0 */
+    uint16_t size;
+
+    /** unused, set to 0 
+     *  @stable ICU 2.0*/
+    uint16_t reservedWord;
+
+    /* platform data properties */
+    /** 0 for little-endian machine, 1 for big-endian
+     *  @stable ICU 2.0 */
+    uint8_t isBigEndian;
+
+    /** see U_CHARSET_FAMILY values in utypes.h 
+     *  @stable ICU 2.0*/
+    uint8_t charsetFamily;
+
+    /** sizeof(UChar), one of { 1, 2, 4 } 
+     *  @stable ICU 2.0*/
+    uint8_t sizeofUChar;
+
+    /** unused, set to 0 
+     *  @stable ICU 2.0*/
+    uint8_t reservedByte;
+
+    /** data format identifier 
+     *  @stable ICU 2.0*/
+    uint8_t dataFormat[4];
+
+    /** versions: [0] major [1] minor [2] milli [3] micro 
+     *  @stable ICU 2.0*/
+    uint8_t formatVersion[4];
+
+    /** versions: [0] major [1] minor [2] milli [3] micro 
+     *  @stable ICU 2.0*/
+    uint8_t dataVersion[4];
+} UDataInfo;
+
+/* API for reading data -----------------------------------------------------*/
+
+/**
+ * Forward declaration of the data memory type.
+ * @stable ICU 2.0
+ */
+typedef struct UDataMemory UDataMemory;
+
+/**
+ * Callback function for udata_openChoice().
+ * @param context parameter passed into <code>udata_openChoice()</code>.
+ * @param type The type of the data as passed into <code>udata_openChoice()</code>.
+ *             It may be <code>NULL</code>.
+ * @param name The name of the data as passed into <code>udata_openChoice()</code>.
+ * @param pInfo A pointer to the <code>UDataInfo</code> structure
+ *              of data that has been loaded and will be returned
+ *              by <code>udata_openChoice()</code> if this function
+ *              returns <code>true</code>.
+ * @return true if the current data memory is acceptable
+ * @stable ICU 2.0
+ */
+typedef UBool U_CALLCONV
+UDataMemoryIsAcceptable(void *context,
+                        const char *type, const char *name,
+                        const UDataInfo *pInfo);
+
+
+/**
+ * Convenience function.
+ * This function works the same as <code>udata_openChoice</code>
+ * except that any data that matches the type and name
+ * is assumed to be acceptable.
+ * @param path Specifies an absolute path and/or a basename for the
+ *             finding of the data in the file system.
+ *             <code>NULL</code> for ICU data.
+ * @param type A string that specifies the type of data to be loaded.
+ *             For example, resource bundles are loaded with type "res",
+ *             conversion tables with type "cnv".
+ *             This may be <code>NULL</code> or empty.
+ * @param name A string that specifies the name of the data.
+ * @param pErrorCode An ICU UErrorCode parameter. It must not be <code>NULL</code>.
+ * @return A pointer (handle) to a data memory object, or <code>NULL</code>
+ *         if an error occurs. Call <code>udata_getMemory()</code>
+ *         to get a pointer to the actual data.
+ *
+ * @see udata_openChoice
+ * @stable ICU 2.0
+ */
+U_CAPI UDataMemory * U_EXPORT2
+udata_open(const char *path, const char *type, const char *name,
+           UErrorCode *pErrorCode);
+
+/**
+ * Data loading function.
+ * This function is used to find and load efficiently data for
+ * ICU and applications using ICU.
+ * It provides an abstract interface that allows to specify a data
+ * type and name to find and load the data.
+ *
+ * <p>The implementation depends on platform properties and user preferences
+ * and may involve loading shared libraries (DLLs), mapping
+ * files into memory, or fopen()/fread() files.
+ * It may also involve using static memory or database queries etc.
+ * Several or all data items may be combined into one entity
+ * (DLL, memory-mappable file).</p>
+ *
+ * <p>The data is always preceded by a header that includes
+ * a <code>UDataInfo</code> structure.
+ * The caller's <code>isAcceptable()</code> function is called to make
+ * sure that the data is useful. It may be called several times if it
+ * rejects the data and there is more than one location with data
+ * matching the type and name.</p>
+ *
+ * <p>If <code>path==NULL</code>, then ICU data is loaded.
+ * Otherwise, it is separated into a basename and a basename-less directory string.
+ * The basename is used as the data package name, and the directory is
+ * logically prepended to the ICU data directory string.</p>
+ *
+ * <p>For details about ICU data loading see the User Guide
+ * Data Management chapter. (https://unicode-org.github.io/icu/userguide/icu_data/)</p>
+ *
+ * @param path Specifies an absolute path and/or a basename for the
+ *             finding of the data in the file system.
+ *             <code>NULL</code> for ICU data.
+ * @param type A string that specifies the type of data to be loaded.
+ *             For example, resource bundles are loaded with type "res",
+ *             conversion tables with type "cnv".
+ *             This may be <code>NULL</code> or empty.
+ * @param name A string that specifies the name of the data.
+ * @param isAcceptable This function is called to verify that loaded data
+ *                     is useful for the client code. If it returns false
+ *                     for all data items, then <code>udata_openChoice()</code>
+ *                     will return with an error.
+ * @param context Arbitrary parameter to be passed into isAcceptable.
+ * @param pErrorCode An ICU UErrorCode parameter. It must not be <code>NULL</code>.
+ * @return A pointer (handle) to a data memory object, or <code>NULL</code>
+ *         if an error occurs. Call <code>udata_getMemory()</code>
+ *         to get a pointer to the actual data.
+ * @stable ICU 2.0
+ */
+U_CAPI UDataMemory * U_EXPORT2
+udata_openChoice(const char *path, const char *type, const char *name,
+                 UDataMemoryIsAcceptable *isAcceptable, void *context,
+                 UErrorCode *pErrorCode);
+
+/**
+ * Close the data memory.
+ * This function must be called to allow the system to
+ * release resources associated with this data memory.
+ * @param pData The pointer to data memory object
+ * @stable ICU 2.0
+ */
+U_CAPI void U_EXPORT2
+udata_close(UDataMemory *pData);
+
+/**
+ * Get the pointer to the actual data inside the data memory.
+ * The data is read-only.
+ *
+ * ICU data must be at least 8-aligned, and should be 16-aligned.
+ *
+ * @param pData The pointer to data memory object
+ * @stable ICU 2.0
+ */
+U_CAPI const void * U_EXPORT2
+udata_getMemory(UDataMemory *pData);
+
+/**
+ * Get the information from the data memory header.
+ * This allows to get access to the header containing
+ * platform data properties etc. which is not part of
+ * the data itself and can therefore not be accessed
+ * via the pointer that <code>udata_getMemory()</code> returns.
+ *
+ * @param pData pointer to the data memory object
+ * @param pInfo pointer to a UDataInfo object;
+ *              its <code>size</code> field must be set correctly,
+ *              typically to <code>sizeof(UDataInfo)</code>.
+ *
+ * <code>*pInfo</code> will be filled with the UDataInfo structure
+ * in the data memory object. If this structure is smaller than
+ * <code>pInfo->size</code>, then the <code>size</code> will be
+ * adjusted and only part of the structure will be filled.
+ * @stable ICU 2.0
+ */
+U_CAPI void U_EXPORT2
+udata_getInfo(UDataMemory *pData, UDataInfo *pInfo);
+
+/**
+ * This function bypasses the normal ICU data loading process and
+ * allows you to force ICU's system data to come out of a user-specified
+ * area in memory.
+ *
+ * ICU data must be at least 8-aligned, and should be 16-aligned.
+ * See https://unicode-org.github.io/icu/userguide/icu_data
+ *
+ * The format of this data is that of the icu common data file, as is
+ * generated by the pkgdata tool with mode=common or mode=dll.
+ * You can read in a whole common mode file and pass the address to the start of the
+ * data, or (with the appropriate link options) pass in the pointer to
+ * the data that has been loaded from a dll by the operating system,
+ * as shown in this code:
+ *
+ *       extern const char U_IMPORT U_ICUDATA_ENTRY_POINT [];
+ *        // U_ICUDATA_ENTRY_POINT is same as entry point specified to pkgdata tool
+ *       UErrorCode  status = U_ZERO_ERROR;
+ *
+ *       udata_setCommonData(&U_ICUDATA_ENTRY_POINT, &status);
+ *
+ * It is important that the declaration be as above. The entry point
+ * must not be declared as an extern void*.
+ *
+ * Starting with ICU 4.4, it is possible to set several data packages,
+ * one per call to this function.
+ * udata_open() will look for data in the multiple data packages in the order
+ * in which they were set.
+ * The position of the linked-in or default-name ICU .data package in the
+ * search list depends on when the first data item is loaded that is not contained
+ * in the already explicitly set packages.
+ * If data was loaded implicitly before the first call to this function
+ * (for example, via opening a converter, constructing a UnicodeString
+ * from default-codepage data, using formatting or collation APIs, etc.),
+ * then the default data will be first in the list.
+ *
+ * This function has no effect on application (non ICU) data.  See udata_setAppData()
+ * for similar functionality for application data.
+ *
+ * @param data pointer to ICU common data
+ * @param err outgoing error status <code>U_USING_DEFAULT_WARNING, U_UNSUPPORTED_ERROR</code>
+ * @stable ICU 2.0
+ */
+U_CAPI void U_EXPORT2
+udata_setCommonData(const void *data, UErrorCode *err);
+
+
+/**
+ * This function bypasses the normal ICU data loading process for application-specific
+ * data and allows you to force the it to come out of a user-specified
+ * pointer.
+ *
+ * ICU data must be at least 8-aligned, and should be 16-aligned.
+ * See https://unicode-org.github.io/icu/userguide/icu_data
+ *
+ * The format of this data is that of the icu common data file, like 'icudt26l.dat'
+ * or the corresponding shared library (DLL) file.
+ * The application must read in or otherwise construct an image of the data and then
+ * pass the address of it to this function.
+ *
+ *
+ * Warning:  setAppData will set a U_USING_DEFAULT_WARNING code if
+ *           data with the specified path that has already been opened, or
+ *           if setAppData with the same path has already been called.
+ *           Any such calls to setAppData will have no effect.
+ *
+ *
+ * @param packageName the package name by which the application will refer
+ *             to (open) this data
+ * @param data pointer to the data
+ * @param err outgoing error status <code>U_USING_DEFAULT_WARNING, U_UNSUPPORTED_ERROR</code>
+ * @see udata_setCommonData
+ * @stable ICU 2.0
+ */
+U_CAPI void U_EXPORT2
+udata_setAppData(const char *packageName, const void *data, UErrorCode *err);
+
+/**
+ * Possible settings for udata_setFileAccess()
+ * @see udata_setFileAccess
+ * @stable ICU 3.4
+ */
+typedef enum UDataFileAccess {
+    /** ICU looks for data in single files first, then in packages. (default) @stable ICU 3.4 */
+    UDATA_FILES_FIRST,
+    /** An alias for the default access mode. @stable ICU 3.4 */
+    UDATA_DEFAULT_ACCESS = UDATA_FILES_FIRST,
+    /** ICU only loads data from packages, not from single files. @stable ICU 3.4 */
+    UDATA_ONLY_PACKAGES,
+    /** ICU loads data from packages first, and only from single files
+        if the data cannot be found in a package. @stable ICU 3.4 */
+    UDATA_PACKAGES_FIRST,
+    /** ICU does not access the file system for data loading. @stable ICU 3.4 */
+    UDATA_NO_FILES,
+#ifndef U_HIDE_DEPRECATED_API
+    /**
+     * Number of real UDataFileAccess values.
+     * @deprecated ICU 58 The numeric value may change over time, see ICU ticket #12420.
+     */
+    UDATA_FILE_ACCESS_COUNT
+#endif  // U_HIDE_DEPRECATED_API
+} UDataFileAccess;
+
+/**
+ * This function may be called to control how ICU loads data. It must be called
+ * before any ICU data is loaded, including application data loaded with 
+ * ures/ResourceBundle or udata APIs. This function is not multithread safe.  
+ * The results of calling it while other threads are loading data are undefined.
+ * @param access The type of file access to be used
+ * @param status Error code.
+ * @see UDataFileAccess
+ * @stable ICU 3.4 
+ */
+U_CAPI void U_EXPORT2
+udata_setFileAccess(UDataFileAccess access, UErrorCode *status);
+
+U_CDECL_END
+
+#if U_SHOW_CPLUSPLUS_API
+
+U_NAMESPACE_BEGIN
+
+/**
+ * \class LocalUDataMemoryPointer
+ * "Smart pointer" class, closes a UDataMemory via udata_close().
+ * For most methods see the LocalPointerBase base class.
+ *
+ * @see LocalPointerBase
+ * @see LocalPointer
+ * @stable ICU 4.4
+ */
+U_DEFINE_LOCAL_OPEN_POINTER(LocalUDataMemoryPointer, UDataMemory, udata_close);
+
+U_NAMESPACE_END
+
+#endif  // U_SHOW_CPLUSPLUS_API
+
+#endif

thirdparty/icu4c/common/unicode/udisplaycontext.h

@@ -0,0 +1,173 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+/*
+*****************************************************************************************
+* Copyright (C) 2014-2016, International Business Machines
+* Corporation and others. All Rights Reserved.
+*****************************************************************************************
+*/
+
+#ifndef UDISPLAYCONTEXT_H
+#define UDISPLAYCONTEXT_H
+
+#include "unicode/utypes.h"
+
+#if !UCONFIG_NO_FORMATTING
+
+/**
+ * \file
+ * \brief C API: Display context types (enum values)
+ */
+
+/**
+ * Display context types, for getting values of a particular setting.
+ * Note, the specific numeric values are internal and may change.
+ * @stable ICU 51
+ */
+enum UDisplayContextType {
+    /**
+     * Type to retrieve the dialect handling setting, e.g.
+     * UDISPCTX_STANDARD_NAMES or UDISPCTX_DIALECT_NAMES.
+     * @stable ICU 51
+     */
+    UDISPCTX_TYPE_DIALECT_HANDLING = 0,
+    /**
+     * Type to retrieve the capitalization context setting, e.g.
+     * UDISPCTX_CAPITALIZATION_NONE, UDISPCTX_CAPITALIZATION_FOR_MIDDLE_OF_SENTENCE,
+     * UDISPCTX_CAPITALIZATION_FOR_BEGINNING_OF_SENTENCE, etc.
+     * @stable ICU 51
+     */
+    UDISPCTX_TYPE_CAPITALIZATION = 1,
+    /**
+     * Type to retrieve the display length setting, e.g.
+     * UDISPCTX_LENGTH_FULL, UDISPCTX_LENGTH_SHORT.
+     * @stable ICU 54
+     */
+    UDISPCTX_TYPE_DISPLAY_LENGTH = 2,
+    /**
+     * Type to retrieve the substitute handling setting, e.g.
+     * UDISPCTX_SUBSTITUTE, UDISPCTX_NO_SUBSTITUTE.
+     * @stable ICU 58
+     */
+    UDISPCTX_TYPE_SUBSTITUTE_HANDLING = 3
+};
+/**
+*  @stable ICU 51
+*/
+typedef enum UDisplayContextType UDisplayContextType;
+
+/**
+ * Display context settings.
+ * Note, the specific numeric values are internal and may change.
+ * @stable ICU 51
+ */
+enum UDisplayContext {
+    /**
+     * ================================
+     * DIALECT_HANDLING can be set to one of UDISPCTX_STANDARD_NAMES or
+     * UDISPCTX_DIALECT_NAMES. Use UDisplayContextType UDISPCTX_TYPE_DIALECT_HANDLING
+     * to get the value.
+     */
+    /**
+     * A possible setting for DIALECT_HANDLING:
+     * use standard names when generating a locale name,
+     * e.g. en_GB displays as 'English (United Kingdom)'.
+     * @stable ICU 51
+     */
+    UDISPCTX_STANDARD_NAMES = (UDISPCTX_TYPE_DIALECT_HANDLING<<8) + 0,
+    /**
+     * A possible setting for DIALECT_HANDLING:
+     * use dialect names, when generating a locale name,
+     * e.g. en_GB displays as 'British English'.
+     * @stable ICU 51
+     */
+    UDISPCTX_DIALECT_NAMES = (UDISPCTX_TYPE_DIALECT_HANDLING<<8) + 1,
+    /**
+     * ================================
+     * CAPITALIZATION can be set to one of UDISPCTX_CAPITALIZATION_NONE,
+     * UDISPCTX_CAPITALIZATION_FOR_MIDDLE_OF_SENTENCE,
+     * UDISPCTX_CAPITALIZATION_FOR_BEGINNING_OF_SENTENCE,
+     * UDISPCTX_CAPITALIZATION_FOR_UI_LIST_OR_MENU, or
+     * UDISPCTX_CAPITALIZATION_FOR_STANDALONE.
+     * Use UDisplayContextType UDISPCTX_TYPE_CAPITALIZATION to get the value.
+     */
+    /**
+     * The capitalization context to be used is unknown (this is the default value).
+     * @stable ICU 51
+     */
+    UDISPCTX_CAPITALIZATION_NONE = (UDISPCTX_TYPE_CAPITALIZATION<<8) + 0,
+    /**
+     * The capitalization context if a date, date symbol or display name is to be
+     * formatted with capitalization appropriate for the middle of a sentence.
+     * @stable ICU 51
+     */
+    UDISPCTX_CAPITALIZATION_FOR_MIDDLE_OF_SENTENCE = (UDISPCTX_TYPE_CAPITALIZATION<<8) + 1,
+    /**
+     * The capitalization context if a date, date symbol or display name is to be
+     * formatted with capitalization appropriate for the beginning of a sentence.
+     * @stable ICU 51
+     */
+    UDISPCTX_CAPITALIZATION_FOR_BEGINNING_OF_SENTENCE = (UDISPCTX_TYPE_CAPITALIZATION<<8) + 2,
+    /**
+     * The capitalization context if a date, date symbol or display name is to be
+     * formatted with capitalization appropriate for a user-interface list or menu item.
+     * @stable ICU 51
+     */
+    UDISPCTX_CAPITALIZATION_FOR_UI_LIST_OR_MENU = (UDISPCTX_TYPE_CAPITALIZATION<<8) + 3,
+    /**
+     * The capitalization context if a date, date symbol or display name is to be
+     * formatted with capitalization appropriate for stand-alone usage such as an
+     * isolated name on a calendar page.
+     * @stable ICU 51
+     */
+    UDISPCTX_CAPITALIZATION_FOR_STANDALONE = (UDISPCTX_TYPE_CAPITALIZATION<<8) + 4,
+    /**
+     * ================================
+     * DISPLAY_LENGTH can be set to one of UDISPCTX_LENGTH_FULL or
+     * UDISPCTX_LENGTH_SHORT. Use UDisplayContextType UDISPCTX_TYPE_DISPLAY_LENGTH
+     * to get the value.
+     */
+    /**
+     * A possible setting for DISPLAY_LENGTH:
+     * use full names when generating a locale name,
+     * e.g. "United States" for US.
+     * @stable ICU 54
+     */
+    UDISPCTX_LENGTH_FULL = (UDISPCTX_TYPE_DISPLAY_LENGTH<<8) + 0,
+    /**
+     * A possible setting for DISPLAY_LENGTH:
+     * use short names when generating a locale name,
+     * e.g. "U.S." for US.
+     * @stable ICU 54
+     */
+    UDISPCTX_LENGTH_SHORT = (UDISPCTX_TYPE_DISPLAY_LENGTH<<8) + 1,
+    /**
+     * ================================
+     * SUBSTITUTE_HANDLING can be set to one of UDISPCTX_SUBSTITUTE or
+     * UDISPCTX_NO_SUBSTITUTE. Use UDisplayContextType UDISPCTX_TYPE_SUBSTITUTE_HANDLING
+     * to get the value.
+     */
+    /**
+     * A possible setting for SUBSTITUTE_HANDLING:
+     * Returns a fallback value (e.g., the input code) when no data is available.
+     * This is the default value.
+     * @stable ICU 58
+     */
+    UDISPCTX_SUBSTITUTE = (UDISPCTX_TYPE_SUBSTITUTE_HANDLING<<8) + 0,
+    /**
+     * A possible setting for SUBSTITUTE_HANDLING:
+     * Returns a null value with error code set to U_ILLEGAL_ARGUMENT_ERROR when no
+     * data is available.
+     * @stable ICU 58
+     */
+    UDISPCTX_NO_SUBSTITUTE = (UDISPCTX_TYPE_SUBSTITUTE_HANDLING<<8) + 1
+
+};
+/**
+*  @stable ICU 51
+*/
+typedef enum UDisplayContext UDisplayContext;
+
+#endif /* #if !UCONFIG_NO_FORMATTING */
+
+#endif

thirdparty/icu4c/common/unicode/uenum.h

@@ -0,0 +1,209 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+/*
+*******************************************************************************
+*
+*   Copyright (C) 2002-2013, International Business Machines
+*   Corporation and others.  All Rights Reserved.
+*
+*******************************************************************************
+*   file name:  uenum.h
+*   encoding:   UTF-8
+*   tab size:   8 (not used)
+*   indentation:2
+*
+*   created on: 2002jul08
+*   created by: Vladimir Weinstein
+*/
+
+#ifndef __UENUM_H
+#define __UENUM_H
+
+#include "unicode/utypes.h"
+
+#if U_SHOW_CPLUSPLUS_API
+#include "unicode/localpointer.h"
+
+U_NAMESPACE_BEGIN
+class StringEnumeration;
+U_NAMESPACE_END
+#endif   // U_SHOW_CPLUSPLUS_API
+
+/**
+ * \file
+ * \brief C API: String Enumeration 
+ */
+ 
+/**
+ * An enumeration object.
+ * For usage in C programs.
+ * @stable ICU 2.2
+ */
+struct UEnumeration;
+/** structure representing an enumeration object instance @stable ICU 2.2 */
+typedef struct UEnumeration UEnumeration;
+
+/**
+ * Disposes of resources in use by the iterator.  If en is NULL,
+ * does nothing.  After this call, any char* or UChar* pointer
+ * returned by uenum_unext() or uenum_next() is invalid.
+ * @param en UEnumeration structure pointer
+ * @stable ICU 2.2
+ */
+U_CAPI void U_EXPORT2
+uenum_close(UEnumeration* en);
+
+#if U_SHOW_CPLUSPLUS_API
+
+U_NAMESPACE_BEGIN
+
+/**
+ * \class LocalUEnumerationPointer
+ * "Smart pointer" class, closes a UEnumeration via uenum_close().
+ * For most methods see the LocalPointerBase base class.
+ *
+ * @see LocalPointerBase
+ * @see LocalPointer
+ * @stable ICU 4.4
+ */
+U_DEFINE_LOCAL_OPEN_POINTER(LocalUEnumerationPointer, UEnumeration, uenum_close);
+
+U_NAMESPACE_END
+
+#endif
+
+/**
+ * Returns the number of elements that the iterator traverses.  If
+ * the iterator is out-of-sync with its service, status is set to
+ * U_ENUM_OUT_OF_SYNC_ERROR.
+ * This is a convenience function. It can end up being very
+ * expensive as all the items might have to be pre-fetched (depending
+ * on the type of data being traversed). Use with caution and only 
+ * when necessary.
+ * @param en UEnumeration structure pointer
+ * @param status error code, can be U_ENUM_OUT_OF_SYNC_ERROR if the
+ *               iterator is out of sync.
+ * @return number of elements in the iterator
+ * @stable ICU 2.2
+ */
+U_CAPI int32_t U_EXPORT2
+uenum_count(UEnumeration* en, UErrorCode* status);
+
+/**
+ * Returns the next element in the iterator's list.  If there are
+ * no more elements, returns NULL.  If the iterator is out-of-sync
+ * with its service, status is set to U_ENUM_OUT_OF_SYNC_ERROR and
+ * NULL is returned.  If the native service string is a char* string,
+ * it is converted to UChar* with the invariant converter.
+ * The result is terminated by (UChar)0.
+ * @param en the iterator object
+ * @param resultLength pointer to receive the length of the result
+ *                     (not including the terminating \\0).
+ *                     If the pointer is NULL it is ignored.
+ * @param status the error code, set to U_ENUM_OUT_OF_SYNC_ERROR if
+ *               the iterator is out of sync with its service.
+ * @return a pointer to the string.  The string will be
+ *         zero-terminated.  The return pointer is owned by this iterator
+ *         and must not be deleted by the caller.  The pointer is valid
+ *         until the next call to any uenum_... method, including
+ *         uenum_next() or uenum_unext().  When all strings have been
+ *         traversed, returns NULL.
+ * @stable ICU 2.2
+ */
+U_CAPI const UChar* U_EXPORT2
+uenum_unext(UEnumeration* en,
+            int32_t* resultLength,
+            UErrorCode* status);
+
+/**
+ * Returns the next element in the iterator's list.  If there are
+ * no more elements, returns NULL.  If the iterator is out-of-sync
+ * with its service, status is set to U_ENUM_OUT_OF_SYNC_ERROR and
+ * NULL is returned.  If the native service string is a UChar*
+ * string, it is converted to char* with the invariant converter.
+ * The result is terminated by (char)0.  If the conversion fails
+ * (because a character cannot be converted) then status is set to
+ * U_INVARIANT_CONVERSION_ERROR and the return value is undefined
+ * (but non-NULL).
+ * @param en the iterator object
+ * @param resultLength pointer to receive the length of the result
+ *                     (not including the terminating \\0).
+ *                     If the pointer is NULL it is ignored.
+ * @param status the error code, set to U_ENUM_OUT_OF_SYNC_ERROR if
+ *               the iterator is out of sync with its service.  Set to
+ *               U_INVARIANT_CONVERSION_ERROR if the underlying native string is
+ *               UChar* and conversion to char* with the invariant converter
+ *               fails. This error pertains only to current string, so iteration
+ *               might be able to continue successfully.
+ * @return a pointer to the string.  The string will be
+ *         zero-terminated.  The return pointer is owned by this iterator
+ *         and must not be deleted by the caller.  The pointer is valid
+ *         until the next call to any uenum_... method, including
+ *         uenum_next() or uenum_unext().  When all strings have been
+ *         traversed, returns NULL.
+ * @stable ICU 2.2
+ */
+U_CAPI const char* U_EXPORT2
+uenum_next(UEnumeration* en,
+           int32_t* resultLength,
+           UErrorCode* status);
+
+/**
+ * Resets the iterator to the current list of service IDs.  This
+ * re-establishes sync with the service and rewinds the iterator
+ * to start at the first element.
+ * @param en the iterator object
+ * @param status the error code, set to U_ENUM_OUT_OF_SYNC_ERROR if
+ *               the iterator is out of sync with its service.  
+ * @stable ICU 2.2
+ */
+U_CAPI void U_EXPORT2
+uenum_reset(UEnumeration* en, UErrorCode* status);
+
+#if U_SHOW_CPLUSPLUS_API
+
+/**
+ * Given a StringEnumeration, wrap it in a UEnumeration.  The
+ * StringEnumeration is adopted; after this call, the caller must not
+ * delete it (regardless of error status).
+ * @param adopted the C++ StringEnumeration to be wrapped in a UEnumeration.
+ * @param ec the error code.
+ * @return a UEnumeration wrapping the adopted StringEnumeration.
+ * @stable ICU 4.2
+ */
+U_CAPI UEnumeration* U_EXPORT2
+uenum_openFromStringEnumeration(icu::StringEnumeration* adopted, UErrorCode* ec);
+
+#endif
+
+/**
+ * Given an array of const UChar* strings, return a UEnumeration.  String pointers from 0..count-1 must not be null.
+ * Do not free or modify either the string array or the characters it points to until this object has been destroyed with uenum_close.
+ * \snippet test/cintltst/uenumtst.c uenum_openUCharStringsEnumeration
+ * @param strings array of const UChar* strings (each null terminated). All storage is owned by the caller.
+ * @param count length of the array
+ * @param ec error code
+ * @return the new UEnumeration object. Caller is responsible for calling uenum_close to free memory.
+ * @see uenum_close
+ * @stable ICU 50
+ */
+U_CAPI UEnumeration* U_EXPORT2
+uenum_openUCharStringsEnumeration(const UChar* const strings[], int32_t count,
+                                 UErrorCode* ec);
+
+/**
+ * Given an array of const char* strings (invariant chars only), return a UEnumeration.  String pointers from 0..count-1 must not be null.
+ * Do not free or modify either the string array or the characters it points to until this object has been destroyed with uenum_close.
+ * \snippet test/cintltst/uenumtst.c uenum_openCharStringsEnumeration
+ * @param strings array of char* strings (each null terminated).  All storage is owned by the caller.
+ * @param count length of the array
+ * @param ec error code
+ * @return the new UEnumeration object. Caller is responsible for calling uenum_close to free memory
+ * @see uenum_close
+ * @stable ICU 50
+ */
+U_CAPI UEnumeration* U_EXPORT2
+uenum_openCharStringsEnumeration(const char* const strings[], int32_t count,
+                                 UErrorCode* ec);
+
+#endif

thirdparty/icu4c/common/unicode/uidna.h

@@ -0,0 +1,794 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+/*
+ *******************************************************************************
+ *
+ *   Copyright (C) 2003-2014, International Business Machines
+ *   Corporation and others.  All Rights Reserved.
+ *
+ *******************************************************************************
+ *   file name:  uidna.h
+ *   encoding:   UTF-8
+ *   tab size:   8 (not used)
+ *   indentation:4
+ *
+ *   created on: 2003feb1
+ *   created by: Ram Viswanadha
+ */
+
+#ifndef __UIDNA_H__
+#define __UIDNA_H__
+
+#include "unicode/utypes.h"
+
+#if !UCONFIG_NO_IDNA
+
+#include <stdbool.h>
+#include "unicode/parseerr.h"
+
+#if U_SHOW_CPLUSPLUS_API
+#include "unicode/localpointer.h"
+#endif   // U_SHOW_CPLUSPLUS_API
+
+/**
+ * \file
+ * \brief C API: Internationalizing Domain Names in Applications (IDNA)
+ *
+ * IDNA2008 is implemented according to UTS #46, see the IDNA C++ class in idna.h.
+ *
+ * The C API functions which do take a UIDNA * service object pointer
+ * implement UTS #46 and IDNA2008.
+ *
+ * IDNA2003 is obsolete.
+ * The C API functions which do not take a service object pointer
+ * implement IDNA2003. They are all deprecated.
+ */
+
+/*
+ * IDNA option bit set values.
+ */
+enum {
+    /**
+     * Default options value: UTS #46 nontransitional processing.
+     * For use in static worker and factory methods.
+     *
+     * Since ICU 76, this is the same as
+     * UIDNA_NONTRANSITIONAL_TO_ASCII | UIDNA_NONTRANSITIONAL_TO_UNICODE,
+     * corresponding to Unicode 15.1 UTS #46 deprecating transitional processing.
+     * (These options are ignored by the IDNA2003 implementation.)
+     *
+     * Before ICU 76, this constant did not set any of the options.
+     *
+     * @stable ICU 2.6
+     */
+    UIDNA_DEFAULT=0x30,
+#ifndef U_HIDE_DEPRECATED_API
+    /**
+     * Option to allow unassigned code points in domain names and labels.
+     * For use in static worker and factory methods.
+     * <p>This option is ignored by the UTS46 implementation.
+     * (UTS #46 disallows unassigned code points.)
+     * @deprecated ICU 55 Use UTS #46 instead via uidna_openUTS46() or class IDNA.
+     */
+    UIDNA_ALLOW_UNASSIGNED=1,
+#endif  /* U_HIDE_DEPRECATED_API */
+    /**
+     * Option to check whether the input conforms to the STD3 ASCII rules,
+     * for example the restriction of labels to LDH characters
+     * (ASCII Letters, Digits and Hyphen-Minus).
+     * For use in static worker and factory methods.
+     * @stable ICU 2.6
+     */
+    UIDNA_USE_STD3_RULES=2,
+    /**
+     * IDNA option to check for whether the input conforms to the BiDi rules.
+     * For use in static worker and factory methods.
+     * <p>This option is ignored by the IDNA2003 implementation.
+     * (IDNA2003 always performs a BiDi check.)
+     * @stable ICU 4.6
+     */
+    UIDNA_CHECK_BIDI=4,
+    /**
+     * IDNA option to check for whether the input conforms to the CONTEXTJ rules.
+     * For use in static worker and factory methods.
+     * <p>This option is ignored by the IDNA2003 implementation.
+     * (The CONTEXTJ check is new in IDNA2008.)
+     * @stable ICU 4.6
+     */
+    UIDNA_CHECK_CONTEXTJ=8,
+    /**
+     * IDNA option for nontransitional processing in ToASCII().
+     * For use in static worker and factory methods.
+     *
+     * <p>By default, ToASCII() uses transitional processing.
+     * Unicode 15.1 UTS #46 deprecated transitional processing.
+     *
+     * <p>This option is ignored by the IDNA2003 implementation.
+     * (This is only relevant for compatibility of newer IDNA implementations with IDNA2003.)
+     * @stable ICU 4.6
+     * @see UIDNA_DEFAULT
+     */
+    UIDNA_NONTRANSITIONAL_TO_ASCII=0x10,
+    /**
+     * IDNA option for nontransitional processing in ToUnicode().
+     * For use in static worker and factory methods.
+     *
+     * <p>By default, ToUnicode() uses transitional processing.
+     * Unicode 15.1 UTS #46 deprecated transitional processing.
+     *
+     * <p>This option is ignored by the IDNA2003 implementation.
+     * (This is only relevant for compatibility of newer IDNA implementations with IDNA2003.)
+     * @stable ICU 4.6
+     * @see UIDNA_DEFAULT
+     */
+    UIDNA_NONTRANSITIONAL_TO_UNICODE=0x20,
+    /**
+     * IDNA option to check for whether the input conforms to the CONTEXTO rules.
+     * For use in static worker and factory methods.
+     * <p>This option is ignored by the IDNA2003 implementation.
+     * (The CONTEXTO check is new in IDNA2008.)
+     * <p>This is for use by registries for IDNA2008 conformance.
+     * UTS #46 does not require the CONTEXTO check.
+     * @stable ICU 49
+     */
+    UIDNA_CHECK_CONTEXTO=0x40
+};
+
+/**
+ * Opaque C service object type for the new IDNA API.
+ * @stable ICU 4.6
+ */
+struct UIDNA;
+typedef struct UIDNA UIDNA;  /**< C typedef for struct UIDNA. @stable ICU 4.6 */
+
+/**
+ * Returns a UIDNA instance which implements UTS #46.
+ * Returns an unmodifiable instance, owned by the caller.
+ * Cache it for multiple operations, and uidna_close() it when done.
+ * The instance is thread-safe, that is, it can be used concurrently.
+ *
+ * For details about the UTS #46 implementation see the IDNA C++ class in idna.h.
+ *
+ * @param options Bit set to modify the processing and error checking.
+ *                These should include UIDNA_DEFAULT, or
+ *                UIDNA_NONTRANSITIONAL_TO_ASCII | UIDNA_NONTRANSITIONAL_TO_UNICODE.
+ *                See option bit set values in uidna.h.
+ * @param pErrorCode Standard ICU error code. Its input value must
+ *                  pass the U_SUCCESS() test, or else the function returns
+ *                  immediately. Check for U_FAILURE() on output or use with
+ *                  function chaining. (See User Guide for details.)
+ * @return the UTS #46 UIDNA instance, if successful
+ * @stable ICU 4.6
+ */
+U_CAPI UIDNA * U_EXPORT2
+uidna_openUTS46(uint32_t options, UErrorCode *pErrorCode);
+
+/**
+ * Closes a UIDNA instance.
+ * @param idna UIDNA instance to be closed
+ * @stable ICU 4.6
+ */
+U_CAPI void U_EXPORT2
+uidna_close(UIDNA *idna);
+
+#if U_SHOW_CPLUSPLUS_API
+
+U_NAMESPACE_BEGIN
+
+/**
+ * \class LocalUIDNAPointer
+ * "Smart pointer" class, closes a UIDNA via uidna_close().
+ * For most methods see the LocalPointerBase base class.
+ *
+ * @see LocalPointerBase
+ * @see LocalPointer
+ * @stable ICU 4.6
+ */
+U_DEFINE_LOCAL_OPEN_POINTER(LocalUIDNAPointer, UIDNA, uidna_close);
+
+U_NAMESPACE_END
+
+#endif
+
+/**
+ * Output container for IDNA processing errors.
+ * Initialize with UIDNA_INFO_INITIALIZER:
+ * \code
+ * UIDNAInfo info = UIDNA_INFO_INITIALIZER;
+ * int32_t length = uidna_nameToASCII(..., &info, &errorCode);
+ * if(U_SUCCESS(errorCode) && info.errors!=0) { ... }
+ * \endcode
+ * @stable ICU 4.6
+ */
+typedef struct UIDNAInfo {
+    /** sizeof(UIDNAInfo) @stable ICU 4.6 */
+    int16_t size;
+    /**
+     * Set to true if transitional and nontransitional processing produce different results.
+     * For details see C++ IDNAInfo::isTransitionalDifferent().
+     * @stable ICU 4.6
+     */
+    UBool isTransitionalDifferent;
+    UBool reservedB3;  /**< Reserved field, do not use. @internal */
+    /**
+     * Bit set indicating IDNA processing errors. 0 if no errors.
+     * See UIDNA_ERROR_... constants.
+     * @stable ICU 4.6
+     */
+    uint32_t errors;
+    int32_t reservedI2;  /**< Reserved field, do not use. @internal */
+    int32_t reservedI3;  /**< Reserved field, do not use. @internal */
+} UIDNAInfo;
+
+/**
+ * Static initializer for a UIDNAInfo struct.
+ * @stable ICU 4.6
+ */
+#define UIDNA_INFO_INITIALIZER { \
+    (int16_t)sizeof(UIDNAInfo), \
+    false, false, \
+    0, 0, 0 }
+
+/**
+ * Converts a single domain name label into its ASCII form for DNS lookup.
+ * If any processing step fails, then pInfo->errors will be non-zero and
+ * the result might not be an ASCII string.
+ * The label might be modified according to the types of errors.
+ * Labels with severe errors will be left in (or turned into) their Unicode form.
+ *
+ * The UErrorCode indicates an error only in exceptional cases,
+ * such as a U_MEMORY_ALLOCATION_ERROR.
+ *
+ * @param idna UIDNA instance
+ * @param label Input domain name label
+ * @param length Label length, or -1 if NUL-terminated
+ * @param dest Destination string buffer
+ * @param capacity Destination buffer capacity
+ * @param pInfo Output container of IDNA processing details.
+ * @param pErrorCode Standard ICU error code. Its input value must
+ *                  pass the U_SUCCESS() test, or else the function returns
+ *                  immediately. Check for U_FAILURE() on output or use with
+ *                  function chaining. (See User Guide for details.)
+ * @return destination string length
+ * @stable ICU 4.6
+ */
+U_CAPI int32_t U_EXPORT2
+uidna_labelToASCII(const UIDNA *idna,
+                   const UChar *label, int32_t length,
+                   UChar *dest, int32_t capacity,
+                   UIDNAInfo *pInfo, UErrorCode *pErrorCode);
+
+/**
+ * Converts a single domain name label into its Unicode form for human-readable display.
+ * If any processing step fails, then pInfo->errors will be non-zero.
+ * The label might be modified according to the types of errors.
+ *
+ * The UErrorCode indicates an error only in exceptional cases,
+ * such as a U_MEMORY_ALLOCATION_ERROR.
+ *
+ * @param idna UIDNA instance
+ * @param label Input domain name label
+ * @param length Label length, or -1 if NUL-terminated
+ * @param dest Destination string buffer
+ * @param capacity Destination buffer capacity
+ * @param pInfo Output container of IDNA processing details.
+ * @param pErrorCode Standard ICU error code. Its input value must
+ *                  pass the U_SUCCESS() test, or else the function returns
+ *                  immediately. Check for U_FAILURE() on output or use with
+ *                  function chaining. (See User Guide for details.)
+ * @return destination string length
+ * @stable ICU 4.6
+ */
+U_CAPI int32_t U_EXPORT2
+uidna_labelToUnicode(const UIDNA *idna,
+                     const UChar *label, int32_t length,
+                     UChar *dest, int32_t capacity,
+                     UIDNAInfo *pInfo, UErrorCode *pErrorCode);
+
+/**
+ * Converts a whole domain name into its ASCII form for DNS lookup.
+ * If any processing step fails, then pInfo->errors will be non-zero and
+ * the result might not be an ASCII string.
+ * The domain name might be modified according to the types of errors.
+ * Labels with severe errors will be left in (or turned into) their Unicode form.
+ *
+ * The UErrorCode indicates an error only in exceptional cases,
+ * such as a U_MEMORY_ALLOCATION_ERROR.
+ *
+ * @param idna UIDNA instance
+ * @param name Input domain name
+ * @param length Domain name length, or -1 if NUL-terminated
+ * @param dest Destination string buffer
+ * @param capacity Destination buffer capacity
+ * @param pInfo Output container of IDNA processing details.
+ * @param pErrorCode Standard ICU error code. Its input value must
+ *                  pass the U_SUCCESS() test, or else the function returns
+ *                  immediately. Check for U_FAILURE() on output or use with
+ *                  function chaining. (See User Guide for details.)
+ * @return destination string length
+ * @stable ICU 4.6
+ */
+U_CAPI int32_t U_EXPORT2
+uidna_nameToASCII(const UIDNA *idna,
+                  const UChar *name, int32_t length,
+                  UChar *dest, int32_t capacity,
+                  UIDNAInfo *pInfo, UErrorCode *pErrorCode);
+
+/**
+ * Converts a whole domain name into its Unicode form for human-readable display.
+ * If any processing step fails, then pInfo->errors will be non-zero.
+ * The domain name might be modified according to the types of errors.
+ *
+ * The UErrorCode indicates an error only in exceptional cases,
+ * such as a U_MEMORY_ALLOCATION_ERROR.
+ *
+ * @param idna UIDNA instance
+ * @param name Input domain name
+ * @param length Domain name length, or -1 if NUL-terminated
+ * @param dest Destination string buffer
+ * @param capacity Destination buffer capacity
+ * @param pInfo Output container of IDNA processing details.
+ * @param pErrorCode Standard ICU error code. Its input value must
+ *                  pass the U_SUCCESS() test, or else the function returns
+ *                  immediately. Check for U_FAILURE() on output or use with
+ *                  function chaining. (See User Guide for details.)
+ * @return destination string length
+ * @stable ICU 4.6
+ */
+U_CAPI int32_t U_EXPORT2
+uidna_nameToUnicode(const UIDNA *idna,
+                    const UChar *name, int32_t length,
+                    UChar *dest, int32_t capacity,
+                    UIDNAInfo *pInfo, UErrorCode *pErrorCode);
+
+/* UTF-8 versions of the processing methods --------------------------------- */
+
+/**
+ * Converts a single domain name label into its ASCII form for DNS lookup.
+ * UTF-8 version of uidna_labelToASCII(), same behavior.
+ *
+ * @param idna UIDNA instance
+ * @param label Input domain name label
+ * @param length Label length, or -1 if NUL-terminated
+ * @param dest Destination string buffer
+ * @param capacity Destination buffer capacity
+ * @param pInfo Output container of IDNA processing details.
+ * @param pErrorCode Standard ICU error code. Its input value must
+ *                  pass the U_SUCCESS() test, or else the function returns
+ *                  immediately. Check for U_FAILURE() on output or use with
+ *                  function chaining. (See User Guide for details.)
+ * @return destination string length
+ * @stable ICU 4.6
+ */
+U_CAPI int32_t U_EXPORT2
+uidna_labelToASCII_UTF8(const UIDNA *idna,
+                        const char *label, int32_t length,
+                        char *dest, int32_t capacity,
+                        UIDNAInfo *pInfo, UErrorCode *pErrorCode);
+
+/**
+ * Converts a single domain name label into its Unicode form for human-readable display.
+ * UTF-8 version of uidna_labelToUnicode(), same behavior.
+ *
+ * @param idna UIDNA instance
+ * @param label Input domain name label
+ * @param length Label length, or -1 if NUL-terminated
+ * @param dest Destination string buffer
+ * @param capacity Destination buffer capacity
+ * @param pInfo Output container of IDNA processing details.
+ * @param pErrorCode Standard ICU error code. Its input value must
+ *                  pass the U_SUCCESS() test, or else the function returns
+ *                  immediately. Check for U_FAILURE() on output or use with
+ *                  function chaining. (See User Guide for details.)
+ * @return destination string length
+ * @stable ICU 4.6
+ */
+U_CAPI int32_t U_EXPORT2
+uidna_labelToUnicodeUTF8(const UIDNA *idna,
+                         const char *label, int32_t length,
+                         char *dest, int32_t capacity,
+                         UIDNAInfo *pInfo, UErrorCode *pErrorCode);
+
+/**
+ * Converts a whole domain name into its ASCII form for DNS lookup.
+ * UTF-8 version of uidna_nameToASCII(), same behavior.
+ *
+ * @param idna UIDNA instance
+ * @param name Input domain name
+ * @param length Domain name length, or -1 if NUL-terminated
+ * @param dest Destination string buffer
+ * @param capacity Destination buffer capacity
+ * @param pInfo Output container of IDNA processing details.
+ * @param pErrorCode Standard ICU error code. Its input value must
+ *                  pass the U_SUCCESS() test, or else the function returns
+ *                  immediately. Check for U_FAILURE() on output or use with
+ *                  function chaining. (See User Guide for details.)
+ * @return destination string length
+ * @stable ICU 4.6
+ */
+U_CAPI int32_t U_EXPORT2
+uidna_nameToASCII_UTF8(const UIDNA *idna,
+                       const char *name, int32_t length,
+                       char *dest, int32_t capacity,
+                       UIDNAInfo *pInfo, UErrorCode *pErrorCode);
+
+/**
+ * Converts a whole domain name into its Unicode form for human-readable display.
+ * UTF-8 version of uidna_nameToUnicode(), same behavior.
+ *
+ * @param idna UIDNA instance
+ * @param name Input domain name
+ * @param length Domain name length, or -1 if NUL-terminated
+ * @param dest Destination string buffer
+ * @param capacity Destination buffer capacity
+ * @param pInfo Output container of IDNA processing details.
+ * @param pErrorCode Standard ICU error code. Its input value must
+ *                  pass the U_SUCCESS() test, or else the function returns
+ *                  immediately. Check for U_FAILURE() on output or use with
+ *                  function chaining. (See User Guide for details.)
+ * @return destination string length
+ * @stable ICU 4.6
+ */
+U_CAPI int32_t U_EXPORT2
+uidna_nameToUnicodeUTF8(const UIDNA *idna,
+                        const char *name, int32_t length,
+                        char *dest, int32_t capacity,
+                        UIDNAInfo *pInfo, UErrorCode *pErrorCode);
+
+/*
+ * IDNA error bit set values.
+ * When a domain name or label fails a processing step or does not meet the
+ * validity criteria, then one or more of these error bits are set.
+ */
+enum {
+    /**
+     * A non-final domain name label (or the whole domain name) is empty.
+     * @stable ICU 4.6
+     */
+    UIDNA_ERROR_EMPTY_LABEL=1,
+    /**
+     * A domain name label is longer than 63 bytes.
+     * (See STD13/RFC1034 3.1. Name space specifications and terminology.)
+     * This is only checked in ToASCII operations, and only if the output label is all-ASCII.
+     * @stable ICU 4.6
+     */
+    UIDNA_ERROR_LABEL_TOO_LONG=2,
+    /**
+     * A domain name is longer than 255 bytes in its storage form.
+     * (See STD13/RFC1034 3.1. Name space specifications and terminology.)
+     * This is only checked in ToASCII operations, and only if the output domain name is all-ASCII.
+     * @stable ICU 4.6
+     */
+    UIDNA_ERROR_DOMAIN_NAME_TOO_LONG=4,
+    /**
+     * A label starts with a hyphen-minus ('-').
+     * @stable ICU 4.6
+     */
+    UIDNA_ERROR_LEADING_HYPHEN=8,
+    /**
+     * A label ends with a hyphen-minus ('-').
+     * @stable ICU 4.6
+     */
+    UIDNA_ERROR_TRAILING_HYPHEN=0x10,
+    /**
+     * A label contains hyphen-minus ('-') in the third and fourth positions.
+     * @stable ICU 4.6
+     */
+    UIDNA_ERROR_HYPHEN_3_4=0x20,
+    /**
+     * A label starts with a combining mark.
+     * @stable ICU 4.6
+     */
+    UIDNA_ERROR_LEADING_COMBINING_MARK=0x40,
+    /**
+     * A label or domain name contains disallowed characters.
+     * @stable ICU 4.6
+     */
+    UIDNA_ERROR_DISALLOWED=0x80,
+    /**
+     * A label starts with "xn--" but does not contain valid Punycode.
+     * That is, an xn-- label failed Punycode decoding.
+     * @stable ICU 4.6
+     */
+    UIDNA_ERROR_PUNYCODE=0x100,
+    /**
+     * A label contains a dot=full stop.
+     * This can occur in an input string for a single-label function.
+     * @stable ICU 4.6
+     */
+    UIDNA_ERROR_LABEL_HAS_DOT=0x200,
+    /**
+     * An ACE label does not contain a valid label string.
+     * The label was successfully ACE (Punycode) decoded but the resulting
+     * string had severe validation errors. For example,
+     * it might contain characters that are not allowed in ACE labels,
+     * or it might not be normalized.
+     * @stable ICU 4.6
+     */
+    UIDNA_ERROR_INVALID_ACE_LABEL=0x400,
+    /**
+     * A label does not meet the IDNA BiDi requirements (for right-to-left characters).
+     * @stable ICU 4.6
+     */
+    UIDNA_ERROR_BIDI=0x800,
+    /**
+     * A label does not meet the IDNA CONTEXTJ requirements.
+     * @stable ICU 4.6
+     */
+    UIDNA_ERROR_CONTEXTJ=0x1000,
+    /**
+     * A label does not meet the IDNA CONTEXTO requirements for punctuation characters.
+     * Some punctuation characters "Would otherwise have been DISALLOWED"
+     * but are allowed in certain contexts. (RFC 5892)
+     * @stable ICU 49
+     */
+    UIDNA_ERROR_CONTEXTO_PUNCTUATION=0x2000,
+    /**
+     * A label does not meet the IDNA CONTEXTO requirements for digits.
+     * Arabic-Indic Digits (U+066x) must not be mixed with Extended Arabic-Indic Digits (U+06Fx).
+     * @stable ICU 49
+     */
+    UIDNA_ERROR_CONTEXTO_DIGITS=0x4000
+};
+
+#ifndef U_HIDE_DEPRECATED_API
+
+/* IDNA2003 API ------------------------------------------------------------- */
+
+/**
+ * IDNA2003: This function implements the ToASCII operation as defined in the IDNA RFC.
+ * This operation is done on <b>single labels</b> before sending it to something that expects
+ * ASCII names. A label is an individual part of a domain name. Labels are usually
+ * separated by dots; e.g. "www.example.com" is composed of 3 labels "www","example", and "com".
+ *
+ * IDNA2003 API Overview:
+ *
+ * The uidna_ API implements the IDNA protocol as defined in the IDNA RFC
+ * (http://www.ietf.org/rfc/rfc3490.txt).
+ * The RFC defines 2 operations: ToASCII and ToUnicode. Domain name labels
+ * containing non-ASCII code points are processed by the
+ * ToASCII operation before passing it to resolver libraries. Domain names
+ * that are obtained from resolver libraries are processed by the
+ * ToUnicode operation before displaying the domain name to the user.
+ * IDNA requires that implementations process input strings with Nameprep
+ * (http://www.ietf.org/rfc/rfc3491.txt),
+ * which is a profile of Stringprep (http://www.ietf.org/rfc/rfc3454.txt),
+ * and then with Punycode (http://www.ietf.org/rfc/rfc3492.txt).
+ * Implementations of IDNA MUST fully implement Nameprep and Punycode;
+ * neither Nameprep nor Punycode are optional.
+ * The input and output of ToASCII and ToUnicode operations are Unicode
+ * and are designed to be chainable, i.e., applying ToASCII or ToUnicode operations
+ * multiple times to an input string will yield the same result as applying the operation
+ * once.
+ * ToUnicode(ToUnicode(ToUnicode...(ToUnicode(string)))) == ToUnicode(string) 
+ * ToASCII(ToASCII(ToASCII...(ToASCII(string))) == ToASCII(string).
+ *
+ * @param src               Input UChar array containing label in Unicode.
+ * @param srcLength         Number of UChars in src, or -1 if NUL-terminated.
+ * @param dest              Output UChar array with ASCII (ACE encoded) label.
+ * @param destCapacity      Size of dest.
+ * @param options           A bit set of options:
+ *
+ *  - UIDNA_DEFAULT             Use default options, i.e., do not process unassigned code points
+ *                              and do not use STD3 ASCII rules
+ *                              If unassigned code points are found the operation fails with 
+ *                              U_UNASSIGNED_ERROR error code.
+ *
+ *  - UIDNA_ALLOW_UNASSIGNED    Unassigned values can be converted to ASCII for query operations
+ *                              If this option is set, the unassigned code points are in the input 
+ *                              are treated as normal Unicode code points.
+ *
+ *  - UIDNA_USE_STD3_RULES      Use STD3 ASCII rules for host name syntax restrictions
+ *                              If this option is set and the input does not satisfy STD3 rules,  
+ *                              the operation will fail with U_IDNA_STD3_ASCII_RULES_ERROR
+ *
+ * @param parseError        Pointer to UParseError struct to receive information on position 
+ *                          of error if an error is encountered. Can be NULL.
+ * @param status            ICU in/out error code parameter.
+ *                          U_INVALID_CHAR_FOUND if src contains
+ *                          unmatched single surrogates.
+ *                          U_INDEX_OUTOFBOUNDS_ERROR if src contains
+ *                          too many code points.
+ *                          U_BUFFER_OVERFLOW_ERROR if destCapacity is not enough
+ * @return The length of the result string, if successful - or in case of a buffer overflow,
+ *         in which case it will be greater than destCapacity.
+ * @deprecated ICU 55 Use UTS #46 instead via uidna_openUTS46() or class IDNA.
+ */
+U_DEPRECATED int32_t U_EXPORT2
+uidna_toASCII(const UChar* src, int32_t srcLength, 
+              UChar* dest, int32_t destCapacity,
+              int32_t options,
+              UParseError* parseError,
+              UErrorCode* status);
+
+
+/**
+ * IDNA2003: This function implements the ToUnicode operation as defined in the IDNA RFC.
+ * This operation is done on <b>single labels</b> before sending it to something that expects
+ * Unicode names. A label is an individual part of a domain name. Labels are usually
+ * separated by dots; for e.g. "www.example.com" is composed of 3 labels "www","example", and "com".
+ *
+ * @param src               Input UChar array containing ASCII (ACE encoded) label.
+ * @param srcLength         Number of UChars in src, or -1 if NUL-terminated.
+ * @param dest Output       Converted UChar array containing Unicode equivalent of label.
+ * @param destCapacity      Size of dest.
+ * @param options           A bit set of options:
+ *
+ *  - UIDNA_DEFAULT             Use default options, i.e., do not process unassigned code points
+ *                              and do not use STD3 ASCII rules
+ *                              If unassigned code points are found the operation fails with 
+ *                              U_UNASSIGNED_ERROR error code.
+ *
+ *  - UIDNA_ALLOW_UNASSIGNED      Unassigned values can be converted to ASCII for query operations
+ *                              If this option is set, the unassigned code points are in the input 
+ *                              are treated as normal Unicode code points. <b> Note: </b> This option is 
+ *                              required on toUnicode operation because the RFC mandates 
+ *                              verification of decoded ACE input by applying toASCII and comparing
+ *                              its output with source
+ *
+ *  - UIDNA_USE_STD3_RULES      Use STD3 ASCII rules for host name syntax restrictions
+ *                              If this option is set and the input does not satisfy STD3 rules,  
+ *                              the operation will fail with U_IDNA_STD3_ASCII_RULES_ERROR
+ *
+ * @param parseError        Pointer to UParseError struct to receive information on position 
+ *                          of error if an error is encountered. Can be NULL.
+ * @param status            ICU in/out error code parameter.
+ *                          U_INVALID_CHAR_FOUND if src contains
+ *                          unmatched single surrogates.
+ *                          U_INDEX_OUTOFBOUNDS_ERROR if src contains
+ *                          too many code points.
+ *                          U_BUFFER_OVERFLOW_ERROR if destCapacity is not enough
+ * @return The length of the result string, if successful - or in case of a buffer overflow,
+ *         in which case it will be greater than destCapacity.
+ * @deprecated ICU 55 Use UTS #46 instead via uidna_openUTS46() or class IDNA.
+ */
+U_DEPRECATED int32_t U_EXPORT2
+uidna_toUnicode(const UChar* src, int32_t srcLength,
+                UChar* dest, int32_t destCapacity,
+                int32_t options,
+                UParseError* parseError,
+                UErrorCode* status);
+
+
+/**
+ * IDNA2003: Convenience function that implements the IDNToASCII operation as defined in the IDNA RFC.
+ * This operation is done on complete domain names, e.g: "www.example.com". 
+ * It is important to note that this operation can fail. If it fails, then the input 
+ * domain name cannot be used as an Internationalized Domain Name and the application
+ * should have methods defined to deal with the failure.
+ *
+ * <b>Note:</b> IDNA RFC specifies that a conformant application should divide a domain name
+ * into separate labels, decide whether to apply allowUnassigned and useSTD3ASCIIRules on each, 
+ * and then convert. This function does not offer that level of granularity. The options once  
+ * set will apply to all labels in the domain name
+ *
+ * @param src               Input UChar array containing IDN in Unicode.
+ * @param srcLength         Number of UChars in src, or -1 if NUL-terminated.
+ * @param dest              Output UChar array with ASCII (ACE encoded) IDN.
+ * @param destCapacity      Size of dest.
+ * @param options           A bit set of options:
+ *
+ *  - UIDNA_DEFAULT             Use default options, i.e., do not process unassigned code points
+ *                              and do not use STD3 ASCII rules
+ *                              If unassigned code points are found the operation fails with 
+ *                              U_UNASSIGNED_CODE_POINT_FOUND error code.
+ *
+ *  - UIDNA_ALLOW_UNASSIGNED    Unassigned values can be converted to ASCII for query operations
+ *                              If this option is set, the unassigned code points are in the input 
+ *                              are treated as normal Unicode code points.
+ *
+ *  - UIDNA_USE_STD3_RULES      Use STD3 ASCII rules for host name syntax restrictions
+ *                              If this option is set and the input does not satisfy STD3 rules,  
+ *                              the operation will fail with U_IDNA_STD3_ASCII_RULES_ERROR
+ *
+ * @param parseError        Pointer to UParseError struct to receive information on position 
+ *                          of error if an error is encountered. Can be NULL.
+ * @param status            ICU in/out error code parameter.
+ *                          U_INVALID_CHAR_FOUND if src contains
+ *                          unmatched single surrogates.
+ *                          U_INDEX_OUTOFBOUNDS_ERROR if src contains
+ *                          too many code points.
+ *                          U_BUFFER_OVERFLOW_ERROR if destCapacity is not enough
+ * @return The length of the result string, if successful - or in case of a buffer overflow,
+ *         in which case it will be greater than destCapacity.
+ * @deprecated ICU 55 Use UTS #46 instead via uidna_openUTS46() or class IDNA.
+ */
+U_DEPRECATED int32_t U_EXPORT2
+uidna_IDNToASCII(  const UChar* src, int32_t srcLength,
+                   UChar* dest, int32_t destCapacity,
+                   int32_t options,
+                   UParseError* parseError,
+                   UErrorCode* status);
+
+/**
+ * IDNA2003: Convenience function that implements the IDNToUnicode operation as defined in the IDNA RFC.
+ * This operation is done on complete domain names, e.g: "www.example.com". 
+ *
+ * <b>Note:</b> IDNA RFC specifies that a conformant application should divide a domain name
+ * into separate labels, decide whether to apply allowUnassigned and useSTD3ASCIIRules on each, 
+ * and then convert. This function does not offer that level of granularity. The options once  
+ * set will apply to all labels in the domain name
+ *
+ * @param src               Input UChar array containing IDN in ASCII (ACE encoded) form.
+ * @param srcLength         Number of UChars in src, or -1 if NUL-terminated.
+ * @param dest Output       UChar array containing Unicode equivalent of source IDN.
+ * @param destCapacity      Size of dest.
+ * @param options           A bit set of options:
+ *
+ *  - UIDNA_DEFAULT             Use default options, i.e., do not process unassigned code points
+ *                              and do not use STD3 ASCII rules
+ *                              If unassigned code points are found the operation fails with 
+ *                              U_UNASSIGNED_CODE_POINT_FOUND error code.
+ *
+ *  - UIDNA_ALLOW_UNASSIGNED    Unassigned values can be converted to ASCII for query operations
+ *                              If this option is set, the unassigned code points are in the input 
+ *                              are treated as normal Unicode code points.
+ *
+ *  - UIDNA_USE_STD3_RULES      Use STD3 ASCII rules for host name syntax restrictions
+ *                              If this option is set and the input does not satisfy STD3 rules,  
+ *                              the operation will fail with U_IDNA_STD3_ASCII_RULES_ERROR
+ *
+ * @param parseError        Pointer to UParseError struct to receive information on position 
+ *                          of error if an error is encountered. Can be NULL.
+ * @param status            ICU in/out error code parameter.
+ *                          U_INVALID_CHAR_FOUND if src contains
+ *                          unmatched single surrogates.
+ *                          U_INDEX_OUTOFBOUNDS_ERROR if src contains
+ *                          too many code points.
+ *                          U_BUFFER_OVERFLOW_ERROR if destCapacity is not enough
+ * @return The length of the result string, if successful - or in case of a buffer overflow,
+ *         in which case it will be greater than destCapacity.
+ * @deprecated ICU 55 Use UTS #46 instead via uidna_openUTS46() or class IDNA.
+ */
+U_DEPRECATED int32_t U_EXPORT2
+uidna_IDNToUnicode(  const UChar* src, int32_t srcLength,
+                     UChar* dest, int32_t destCapacity,
+                     int32_t options,
+                     UParseError* parseError,
+                     UErrorCode* status);
+
+/**
+ * IDNA2003: Compare two IDN strings for equivalence.
+ * This function splits the domain names into labels and compares them.
+ * According to IDN RFC, whenever two labels are compared, they are 
+ * considered equal if and only if their ASCII forms (obtained by 
+ * applying toASCII) match using an case-insensitive ASCII comparison.
+ * Two domain names are considered a match if and only if all labels 
+ * match regardless of whether label separators match.
+ *
+ * @param s1                First source string.
+ * @param length1           Length of first source string, or -1 if NUL-terminated.
+ *
+ * @param s2                Second source string.
+ * @param length2           Length of second source string, or -1 if NUL-terminated.
+ * @param options           A bit set of options:
+ *
+ *  - UIDNA_DEFAULT             Use default options, i.e., do not process unassigned code points
+ *                              and do not use STD3 ASCII rules
+ *                              If unassigned code points are found the operation fails with 
+ *                              U_UNASSIGNED_CODE_POINT_FOUND error code.
+ *
+ *  - UIDNA_ALLOW_UNASSIGNED    Unassigned values can be converted to ASCII for query operations
+ *                              If this option is set, the unassigned code points are in the input 
+ *                              are treated as normal Unicode code points.
+ *
+ *  - UIDNA_USE_STD3_RULES      Use STD3 ASCII rules for host name syntax restrictions
+ *                              If this option is set and the input does not satisfy STD3 rules,  
+ *                              the operation will fail with U_IDNA_STD3_ASCII_RULES_ERROR
+ *
+ * @param status            ICU error code in/out parameter.
+ *                          Must fulfill U_SUCCESS before the function call.
+ * @return <0 or 0 or >0 as usual for string comparisons
+ * @deprecated ICU 55 Use UTS #46 instead via uidna_openUTS46() or class IDNA.
+ */
+U_DEPRECATED int32_t U_EXPORT2
+uidna_compare(  const UChar *s1, int32_t length1,
+                const UChar *s2, int32_t length2,
+                int32_t options,
+                UErrorCode* status);
+
+#endif  /* U_HIDE_DEPRECATED_API */
+
+#endif /* #if !UCONFIG_NO_IDNA */
+
+#endif

thirdparty/icu4c/common/unicode/uiter.h

@@ -0,0 +1,709 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+/*
+*******************************************************************************
+*
+*   Copyright (C) 2002-2011 International Business Machines
+*   Corporation and others.  All Rights Reserved.
+*
+*******************************************************************************
+*   file name:  uiter.h
+*   encoding:   UTF-8
+*   tab size:   8 (not used)
+*   indentation:4
+*
+*   created on: 2002jan18
+*   created by: Markus W. Scherer
+*/
+
+#ifndef __UITER_H__
+#define __UITER_H__
+
+/**
+ * \file
+ * \brief C API: Unicode Character Iteration
+ *
+ * @see UCharIterator
+ */
+
+#include "unicode/utypes.h"
+
+#if U_SHOW_CPLUSPLUS_API
+    U_NAMESPACE_BEGIN
+
+    class CharacterIterator;
+    class Replaceable;
+
+    U_NAMESPACE_END
+#endif
+
+U_CDECL_BEGIN
+
+struct UCharIterator;
+typedef struct UCharIterator UCharIterator; /**< C typedef for struct UCharIterator. @stable ICU 2.1 */
+
+/**
+ * Origin constants for UCharIterator.getIndex() and UCharIterator.move().
+ * @see UCharIteratorMove
+ * @see UCharIterator
+ * @stable ICU 2.1
+ */
+typedef enum UCharIteratorOrigin {
+    UITER_START, UITER_CURRENT, UITER_LIMIT, UITER_ZERO, UITER_LENGTH
+} UCharIteratorOrigin;
+
+/** Constants for UCharIterator. @stable ICU 2.6 */
+enum {
+    /**
+     * Constant value that may be returned by UCharIteratorMove
+     * indicating that the final UTF-16 index is not known, but that the move succeeded.
+     * This can occur when moving relative to limit or length, or
+     * when moving relative to the current index after a setState()
+     * when the current UTF-16 index is not known.
+     *
+     * It would be very inefficient to have to count from the beginning of the text
+     * just to get the current/limit/length index after moving relative to it.
+     * The actual index can be determined with getIndex(UITER_CURRENT)
+     * which will count the UChars if necessary.
+     *
+     * @stable ICU 2.6
+     */
+    UITER_UNKNOWN_INDEX=-2
+};
+
+
+/**
+ * Constant for UCharIterator getState() indicating an error or
+ * an unknown state.
+ * Returned by uiter_getState()/UCharIteratorGetState
+ * when an error occurs.
+ * Also, some UCharIterator implementations may not be able to return
+ * a valid state for each position. This will be clearly documented
+ * for each such iterator (none of the public ones here).
+ *
+ * @stable ICU 2.6
+ */
+#define UITER_NO_STATE ((uint32_t)0xffffffff)
+
+/**
+ * Function type declaration for UCharIterator.getIndex().
+ *
+ * Gets the current position, or the start or limit of the
+ * iteration range.
+ *
+ * This function may perform slowly for UITER_CURRENT after setState() was called,
+ * or for UITER_LENGTH, because an iterator implementation may have to count
+ * UChars if the underlying storage is not UTF-16.
+ *
+ * @param iter the UCharIterator structure ("this pointer")
+ * @param origin get the 0, start, limit, length, or current index
+ * @return the requested index, or U_SENTINEL in an error condition
+ *
+ * @see UCharIteratorOrigin
+ * @see UCharIterator
+ * @stable ICU 2.1
+ */
+typedef int32_t U_CALLCONV
+UCharIteratorGetIndex(UCharIterator *iter, UCharIteratorOrigin origin);
+
+/**
+ * Function type declaration for UCharIterator.move().
+ *
+ * Use iter->move(iter, index, UITER_ZERO) like CharacterIterator::setIndex(index).
+ *
+ * Moves the current position relative to the start or limit of the
+ * iteration range, or relative to the current position itself.
+ * The movement is expressed in numbers of code units forward
+ * or backward by specifying a positive or negative delta.
+ * Out of bounds movement will be pinned to the start or limit.
+ *
+ * This function may perform slowly for moving relative to UITER_LENGTH
+ * because an iterator implementation may have to count the rest of the
+ * UChars if the native storage is not UTF-16.
+ *
+ * When moving relative to the limit or length, or
+ * relative to the current position after setState() was called,
+ * move() may return UITER_UNKNOWN_INDEX (-2) to avoid an inefficient
+ * determination of the actual UTF-16 index.
+ * The actual index can be determined with getIndex(UITER_CURRENT)
+ * which will count the UChars if necessary.
+ * See UITER_UNKNOWN_INDEX for details.
+ *
+ * @param iter the UCharIterator structure ("this pointer")
+ * @param delta can be positive, zero, or negative
+ * @param origin move relative to the 0, start, limit, length, or current index
+ * @return the new index, or U_SENTINEL on an error condition,
+ *         or UITER_UNKNOWN_INDEX when the index is not known.
+ *
+ * @see UCharIteratorOrigin
+ * @see UCharIterator
+ * @see UITER_UNKNOWN_INDEX
+ * @stable ICU 2.1
+ */
+typedef int32_t U_CALLCONV
+UCharIteratorMove(UCharIterator *iter, int32_t delta, UCharIteratorOrigin origin);
+
+/**
+ * Function type declaration for UCharIterator.hasNext().
+ *
+ * Check if current() and next() can still
+ * return another code unit.
+ *
+ * @param iter the UCharIterator structure ("this pointer")
+ * @return boolean value for whether current() and next() can still return another code unit
+ *
+ * @see UCharIterator
+ * @stable ICU 2.1
+ */
+typedef UBool U_CALLCONV
+UCharIteratorHasNext(UCharIterator *iter);
+
+/**
+ * Function type declaration for UCharIterator.hasPrevious().
+ *
+ * Check if previous() can still return another code unit.
+ *
+ * @param iter the UCharIterator structure ("this pointer")
+ * @return boolean value for whether previous() can still return another code unit
+ *
+ * @see UCharIterator
+ * @stable ICU 2.1
+ */
+typedef UBool U_CALLCONV
+UCharIteratorHasPrevious(UCharIterator *iter);
+ 
+/**
+ * Function type declaration for UCharIterator.current().
+ *
+ * Return the code unit at the current position,
+ * or U_SENTINEL if there is none (index is at the limit).
+ *
+ * @param iter the UCharIterator structure ("this pointer")
+ * @return the current code unit
+ *
+ * @see UCharIterator
+ * @stable ICU 2.1
+ */
+typedef UChar32 U_CALLCONV
+UCharIteratorCurrent(UCharIterator *iter);
+
+/**
+ * Function type declaration for UCharIterator.next().
+ *
+ * Return the code unit at the current index and increment
+ * the index (post-increment, like s[i++]),
+ * or return U_SENTINEL if there is none (index is at the limit).
+ *
+ * @param iter the UCharIterator structure ("this pointer")
+ * @return the current code unit (and post-increment the current index)
+ *
+ * @see UCharIterator
+ * @stable ICU 2.1
+ */
+typedef UChar32 U_CALLCONV
+UCharIteratorNext(UCharIterator *iter);
+
+/**
+ * Function type declaration for UCharIterator.previous().
+ *
+ * Decrement the index and return the code unit from there
+ * (pre-decrement, like s[--i]),
+ * or return U_SENTINEL if there is none (index is at the start).
+ *
+ * @param iter the UCharIterator structure ("this pointer")
+ * @return the previous code unit (after pre-decrementing the current index)
+ *
+ * @see UCharIterator
+ * @stable ICU 2.1
+ */
+typedef UChar32 U_CALLCONV
+UCharIteratorPrevious(UCharIterator *iter);
+
+/**
+ * Function type declaration for UCharIterator.reservedFn().
+ * Reserved for future use.
+ *
+ * @param iter the UCharIterator structure ("this pointer")
+ * @param something some integer argument
+ * @return some integer
+ *
+ * @see UCharIterator
+ * @stable ICU 2.1
+ */
+typedef int32_t U_CALLCONV
+UCharIteratorReserved(UCharIterator *iter, int32_t something);
+
+/**
+ * Function type declaration for UCharIterator.getState().
+ *
+ * Get the "state" of the iterator in the form of a single 32-bit word.
+ * It is recommended that the state value be calculated to be as small as
+ * is feasible. For strings with limited lengths, fewer than 32 bits may
+ * be sufficient.
+ *
+ * This is used together with setState()/UCharIteratorSetState
+ * to save and restore the iterator position more efficiently than with
+ * getIndex()/move().
+ *
+ * The iterator state is defined as a uint32_t value because it is designed
+ * for use in ucol_nextSortKeyPart() which provides 32 bits to store the state
+ * of the character iterator.
+ *
+ * With some UCharIterator implementations (e.g., UTF-8),
+ * getting and setting the UTF-16 index with existing functions
+ * (getIndex(UITER_CURRENT) followed by move(pos, UITER_ZERO)) is possible but
+ * relatively slow because the iterator has to "walk" from a known index
+ * to the requested one.
+ * This takes more time the farther it needs to go.
+ *
+ * An opaque state value allows an iterator implementation to provide
+ * an internal index (UTF-8: the source byte array index) for
+ * fast, constant-time restoration.
+ *
+ * After calling setState(), a getIndex(UITER_CURRENT) may be slow because
+ * the UTF-16 index may not be restored as well, but the iterator can deliver
+ * the correct text contents and move relative to the current position
+ * without performance degradation.
+ *
+ * Some UCharIterator implementations may not be able to return
+ * a valid state for each position, in which case they return UITER_NO_STATE instead.
+ * This will be clearly documented for each such iterator (none of the public ones here).
+ *
+ * @param iter the UCharIterator structure ("this pointer")
+ * @return the state word
+ *
+ * @see UCharIterator
+ * @see UCharIteratorSetState
+ * @see UITER_NO_STATE
+ * @stable ICU 2.6
+ */
+typedef uint32_t U_CALLCONV
+UCharIteratorGetState(const UCharIterator *iter);
+
+/**
+ * Function type declaration for UCharIterator.setState().
+ *
+ * Restore the "state" of the iterator using a state word from a getState() call.
+ * The iterator object need not be the same one as for which getState() was called,
+ * but it must be of the same type (set up using the same uiter_setXYZ function)
+ * and it must iterate over the same string
+ * (binary identical regardless of memory address).
+ * For more about the state word see UCharIteratorGetState.
+ *
+ * After calling setState(), a getIndex(UITER_CURRENT) may be slow because
+ * the UTF-16 index may not be restored as well, but the iterator can deliver
+ * the correct text contents and move relative to the current position
+ * without performance degradation.
+ *
+ * @param iter the UCharIterator structure ("this pointer")
+ * @param state the state word from a getState() call
+ *              on a same-type, same-string iterator
+ * @param pErrorCode Must be a valid pointer to an error code value,
+ *                   which must not indicate a failure before the function call.
+ *
+ * @see UCharIterator
+ * @see UCharIteratorGetState
+ * @stable ICU 2.6
+ */
+typedef void U_CALLCONV
+UCharIteratorSetState(UCharIterator *iter, uint32_t state, UErrorCode *pErrorCode);
+
+
+/**
+ * C API for code unit iteration.
+ * This can be used as a C wrapper around
+ * CharacterIterator, Replaceable, or implemented using simple strings, etc.
+ *
+ * There are two roles for using UCharIterator:
+ *
+ * A "provider" sets the necessary function pointers and controls the "protected"
+ * fields of the UCharIterator structure. A "provider" passes a UCharIterator
+ * into C APIs that need a UCharIterator as an abstract, flexible string interface.
+ *
+ * Implementations of such C APIs are "callers" of UCharIterator functions;
+ * they only use the "public" function pointers and never access the "protected"
+ * fields directly.
+ *
+ * The current() and next() functions only check the current index against the
+ * limit, and previous() only checks the current index against the start,
+ * to see if the iterator already reached the end of the iteration range.
+ *
+ * The assumption - in all iterators - is that the index is moved via the API,
+ * which means it won't go out of bounds, or the index is modified by
+ * user code that knows enough about the iterator implementation to set valid
+ * index values.
+ *
+ * UCharIterator functions return code unit values 0..0xffff,
+ * or U_SENTINEL if the iteration bounds are reached.
+ *
+ * @stable ICU 2.1
+ */
+struct UCharIterator {
+    /**
+     * (protected) Pointer to string or wrapped object or similar.
+     * Not used by caller.
+     * @stable ICU 2.1
+     */
+    const void *context;
+
+    /**
+     * (protected) Length of string or similar.
+     * Not used by caller.
+     * @stable ICU 2.1
+     */
+    int32_t length;
+
+    /**
+     * (protected) Start index or similar.
+     * Not used by caller.
+     * @stable ICU 2.1
+     */
+    int32_t start;
+
+    /**
+     * (protected) Current index or similar.
+     * Not used by caller.
+     * @stable ICU 2.1
+     */
+    int32_t index;
+
+    /**
+     * (protected) Limit index or similar.
+     * Not used by caller.
+     * @stable ICU 2.1
+     */
+    int32_t limit;
+
+    /**
+     * (protected) Used by UTF-8 iterators and possibly others.
+     * @stable ICU 2.1
+     */
+    int32_t reservedField;
+
+    /**
+     * (public) Returns the current position or the
+     * start or limit index of the iteration range.
+     *
+     * @see UCharIteratorGetIndex
+     * @stable ICU 2.1
+     */
+    UCharIteratorGetIndex *getIndex;
+
+    /**
+     * (public) Moves the current position relative to the start or limit of the
+     * iteration range, or relative to the current position itself.
+     * The movement is expressed in numbers of code units forward
+     * or backward by specifying a positive or negative delta.
+     *
+     * @see UCharIteratorMove
+     * @stable ICU 2.1
+     */
+    UCharIteratorMove *move;
+
+    /**
+     * (public) Check if current() and next() can still
+     * return another code unit.
+     *
+     * @see UCharIteratorHasNext
+     * @stable ICU 2.1
+     */
+    UCharIteratorHasNext *hasNext;
+
+    /**
+     * (public) Check if previous() can still return another code unit.
+     *
+     * @see UCharIteratorHasPrevious
+     * @stable ICU 2.1
+     */
+    UCharIteratorHasPrevious *hasPrevious;
+
+    /**
+     * (public) Return the code unit at the current position,
+     * or U_SENTINEL if there is none (index is at the limit).
+     *
+     * @see UCharIteratorCurrent
+     * @stable ICU 2.1
+     */
+    UCharIteratorCurrent *current;
+
+    /**
+     * (public) Return the code unit at the current index and increment
+     * the index (post-increment, like s[i++]),
+     * or return U_SENTINEL if there is none (index is at the limit).
+     *
+     * @see UCharIteratorNext
+     * @stable ICU 2.1
+     */
+    UCharIteratorNext *next;
+
+    /**
+     * (public) Decrement the index and return the code unit from there
+     * (pre-decrement, like s[--i]),
+     * or return U_SENTINEL if there is none (index is at the start).
+     *
+     * @see UCharIteratorPrevious
+     * @stable ICU 2.1
+     */
+    UCharIteratorPrevious *previous;
+
+    /**
+     * (public) Reserved for future use. Currently NULL.
+     *
+     * @see UCharIteratorReserved
+     * @stable ICU 2.1
+     */
+    UCharIteratorReserved *reservedFn;
+
+    /**
+     * (public) Return the state of the iterator, to be restored later with setState().
+     * This function pointer is NULL if the iterator does not implement it.
+     *
+     * @see UCharIteratorGet
+     * @stable ICU 2.6
+     */
+    UCharIteratorGetState *getState;
+
+    /**
+     * (public) Restore the iterator state from the state word from a call
+     * to getState().
+     * This function pointer is NULL if the iterator does not implement it.
+     *
+     * @see UCharIteratorSet
+     * @stable ICU 2.6
+     */
+    UCharIteratorSetState *setState;
+};
+
+/**
+ * Helper function for UCharIterator to get the code point
+ * at the current index.
+ *
+ * Return the code point that includes the code unit at the current position,
+ * or U_SENTINEL if there is none (index is at the limit).
+ * If the current code unit is a lead or trail surrogate,
+ * then the following or preceding surrogate is used to form
+ * the code point value.
+ *
+ * @param iter the UCharIterator structure ("this pointer")
+ * @return the current code point
+ *
+ * @see UCharIterator
+ * @see U16_GET
+ * @see UnicodeString::char32At()
+ * @stable ICU 2.1
+ */
+U_CAPI UChar32 U_EXPORT2
+uiter_current32(UCharIterator *iter);
+
+/**
+ * Helper function for UCharIterator to get the next code point.
+ *
+ * Return the code point at the current index and increment
+ * the index (post-increment, like s[i++]),
+ * or return U_SENTINEL if there is none (index is at the limit).
+ *
+ * @param iter the UCharIterator structure ("this pointer")
+ * @return the current code point (and post-increment the current index)
+ *
+ * @see UCharIterator
+ * @see U16_NEXT
+ * @stable ICU 2.1
+ */
+U_CAPI UChar32 U_EXPORT2
+uiter_next32(UCharIterator *iter);
+
+/**
+ * Helper function for UCharIterator to get the previous code point.
+ *
+ * Decrement the index and return the code point from there
+ * (pre-decrement, like s[--i]),
+ * or return U_SENTINEL if there is none (index is at the start).
+ *
+ * @param iter the UCharIterator structure ("this pointer")
+ * @return the previous code point (after pre-decrementing the current index)
+ *
+ * @see UCharIterator
+ * @see U16_PREV
+ * @stable ICU 2.1
+ */
+U_CAPI UChar32 U_EXPORT2
+uiter_previous32(UCharIterator *iter);
+
+/**
+ * Get the "state" of the iterator in the form of a single 32-bit word.
+ * This is a convenience function that calls iter->getState(iter)
+ * if iter->getState is not NULL;
+ * if it is NULL or any other error occurs, then UITER_NO_STATE is returned.
+ *
+ * Some UCharIterator implementations may not be able to return
+ * a valid state for each position, in which case they return UITER_NO_STATE instead.
+ * This will be clearly documented for each such iterator (none of the public ones here).
+ *
+ * @param iter the UCharIterator structure ("this pointer")
+ * @return the state word
+ *
+ * @see UCharIterator
+ * @see UCharIteratorGetState
+ * @see UITER_NO_STATE
+ * @stable ICU 2.6
+ */
+U_CAPI uint32_t U_EXPORT2
+uiter_getState(const UCharIterator *iter);
+
+/**
+ * Restore the "state" of the iterator using a state word from a getState() call.
+ * This is a convenience function that calls iter->setState(iter, state, pErrorCode)
+ * if iter->setState is not NULL; if it is NULL, then U_UNSUPPORTED_ERROR is set.
+ *
+ * @param iter the UCharIterator structure ("this pointer")
+ * @param state the state word from a getState() call
+ *              on a same-type, same-string iterator
+ * @param pErrorCode Must be a valid pointer to an error code value,
+ *                   which must not indicate a failure before the function call.
+ *
+ * @see UCharIterator
+ * @see UCharIteratorSetState
+ * @stable ICU 2.6
+ */
+U_CAPI void U_EXPORT2
+uiter_setState(UCharIterator *iter, uint32_t state, UErrorCode *pErrorCode);
+
+/**
+ * Set up a UCharIterator to iterate over a string.
+ *
+ * Sets the UCharIterator function pointers for iteration over the string s
+ * with iteration boundaries start=index=0 and length=limit=string length.
+ * The "provider" may set the start, index, and limit values at any time
+ * within the range 0..length.
+ * The length field will be ignored.
+ *
+ * The string pointer s is set into UCharIterator.context without copying
+ * or reallocating the string contents.
+ *
+ * getState() simply returns the current index.
+ * move() will always return the final index.
+ *
+ * @param iter UCharIterator structure to be set for iteration
+ * @param s String to iterate over
+ * @param length Length of s, or -1 if NUL-terminated
+ *
+ * @see UCharIterator
+ * @stable ICU 2.1
+ */
+U_CAPI void U_EXPORT2
+uiter_setString(UCharIterator *iter, const UChar *s, int32_t length);
+
+/**
+ * Set up a UCharIterator to iterate over a UTF-16BE string
+ * (byte vector with a big-endian pair of bytes per UChar).
+ *
+ * Everything works just like with a normal UChar iterator (uiter_setString),
+ * except that UChars are assembled from byte pairs,
+ * and that the length argument here indicates an even number of bytes.
+ *
+ * getState() simply returns the current index.
+ * move() will always return the final index.
+ *
+ * @param iter UCharIterator structure to be set for iteration
+ * @param s UTF-16BE string to iterate over
+ * @param length Length of s as an even number of bytes, or -1 if NUL-terminated
+ *               (NUL means pair of 0 bytes at even index from s)
+ *
+ * @see UCharIterator
+ * @see uiter_setString
+ * @stable ICU 2.6
+ */
+U_CAPI void U_EXPORT2
+uiter_setUTF16BE(UCharIterator *iter, const char *s, int32_t length);
+
+/**
+ * Set up a UCharIterator to iterate over a UTF-8 string.
+ *
+ * Sets the UCharIterator function pointers for iteration over the UTF-8 string s
+ * with UTF-8 iteration boundaries 0 and length.
+ * The implementation counts the UTF-16 index on the fly and
+ * lazily evaluates the UTF-16 length of the text.
+ *
+ * The start field is used as the UTF-8 offset, the limit field as the UTF-8 length.
+ * When the reservedField is not 0, then it contains a supplementary code point
+ * and the UTF-16 index is between the two corresponding surrogates.
+ * At that point, the UTF-8 index is behind that code point.
+ *
+ * The UTF-8 string pointer s is set into UCharIterator.context without copying
+ * or reallocating the string contents.
+ *
+ * getState() returns a state value consisting of
+ * - the current UTF-8 source byte index (bits 31..1)
+ * - a flag (bit 0) that indicates whether the UChar position is in the middle
+ *   of a surrogate pair
+ *   (from a 4-byte UTF-8 sequence for the corresponding supplementary code point)
+ *
+ * getState() cannot also encode the UTF-16 index in the state value.
+ * move(relative to limit or length), or
+ * move(relative to current) after setState(), may return UITER_UNKNOWN_INDEX.
+ *
+ * @param iter UCharIterator structure to be set for iteration
+ * @param s UTF-8 string to iterate over
+ * @param length Length of s in bytes, or -1 if NUL-terminated
+ *
+ * @see UCharIterator
+ * @stable ICU 2.6
+ */
+U_CAPI void U_EXPORT2
+uiter_setUTF8(UCharIterator *iter, const char *s, int32_t length);
+
+#if U_SHOW_CPLUSPLUS_API
+
+/**
+ * Set up a UCharIterator to wrap around a C++ CharacterIterator.
+ *
+ * Sets the UCharIterator function pointers for iteration using the
+ * CharacterIterator charIter.
+ *
+ * The CharacterIterator pointer charIter is set into UCharIterator.context
+ * without copying or cloning the CharacterIterator object.
+ * The other "protected" UCharIterator fields are set to 0 and will be ignored.
+ * The iteration index and boundaries are controlled by the CharacterIterator.
+ *
+ * getState() simply returns the current index.
+ * move() will always return the final index.
+ *
+ * @param iter UCharIterator structure to be set for iteration
+ * @param charIter CharacterIterator to wrap
+ *
+ * @see UCharIterator
+ * @stable ICU 2.1
+ */
+U_CAPI void U_EXPORT2
+uiter_setCharacterIterator(UCharIterator *iter, icu::CharacterIterator *charIter);
+
+/**
+ * Set up a UCharIterator to iterate over a C++ Replaceable.
+ *
+ * Sets the UCharIterator function pointers for iteration over the
+ * Replaceable rep with iteration boundaries start=index=0 and
+ * length=limit=rep->length().
+ * The "provider" may set the start, index, and limit values at any time
+ * within the range 0..length=rep->length().
+ * The length field will be ignored.
+ *
+ * The Replaceable pointer rep is set into UCharIterator.context without copying
+ * or cloning/reallocating the Replaceable object.
+ *
+ * getState() simply returns the current index.
+ * move() will always return the final index.
+ *
+ * @param iter UCharIterator structure to be set for iteration
+ * @param rep Replaceable to iterate over
+ *
+ * @see UCharIterator
+ * @stable ICU 2.1
+ */
+U_CAPI void U_EXPORT2
+uiter_setReplaceable(UCharIterator *iter, const icu::Replaceable *rep);
+
+#endif
+
+U_CDECL_END
+
+#endif

thirdparty/icu4c/common/unicode/uldnames.h

@@ -0,0 +1,307 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+/*
+*******************************************************************************
+*   Copyright (C) 2010-2016, International Business Machines Corporation and
+*   others.  All Rights Reserved.
+*******************************************************************************
+*/
+
+#ifndef __ULDNAMES_H__
+#define __ULDNAMES_H__
+
+/**
+ * \file
+ * \brief C API: Provides display names of Locale ids and their components.
+ */
+
+#include "unicode/utypes.h"
+#include "unicode/uscript.h"
+#include "unicode/udisplaycontext.h"
+
+#if U_SHOW_CPLUSPLUS_API
+#include "unicode/localpointer.h"
+#endif   // U_SHOW_CPLUSPLUS_API
+
+/**
+ * Enum used in LocaleDisplayNames::createInstance.
+ * @stable ICU 4.4
+ */
+typedef enum {
+    /**
+     * Use standard names when generating a locale name,
+     * e.g. en_GB displays as 'English (United Kingdom)'.
+     * @stable ICU 4.4
+     */
+    ULDN_STANDARD_NAMES = 0,
+    /**
+     * Use dialect names, when generating a locale name,
+     * e.g. en_GB displays as 'British English'.
+     * @stable ICU 4.4
+     */
+    ULDN_DIALECT_NAMES
+} UDialectHandling;
+
+/**
+ * Opaque C service object type for the locale display names API
+ * @stable ICU 4.4
+ */
+struct ULocaleDisplayNames;
+
+/** 
+ * C typedef for struct ULocaleDisplayNames. 
+ * @stable ICU 4.4 
+ */
+typedef struct ULocaleDisplayNames ULocaleDisplayNames;  
+
+#if !UCONFIG_NO_FORMATTING
+
+/**
+ * Returns an instance of LocaleDisplayNames that returns names
+ * formatted for the provided locale, using the provided
+ * dialectHandling.  The usual value for dialectHandling is
+ * ULOC_STANDARD_NAMES.
+ *
+ * @param locale the display locale 
+ * @param dialectHandling how to select names for locales 
+ * @return a ULocaleDisplayNames instance 
+ * @param pErrorCode the status code
+ * @stable ICU 4.4
+ */
+U_CAPI ULocaleDisplayNames * U_EXPORT2
+uldn_open(const char * locale,
+          UDialectHandling dialectHandling,
+          UErrorCode *pErrorCode);
+
+/**
+ * Closes a ULocaleDisplayNames instance obtained from uldn_open().
+ * @param ldn the ULocaleDisplayNames instance to be closed
+ * @stable ICU 4.4
+ */
+U_CAPI void U_EXPORT2
+uldn_close(ULocaleDisplayNames *ldn);
+
+#if U_SHOW_CPLUSPLUS_API
+
+U_NAMESPACE_BEGIN
+
+/**
+ * \class LocalULocaleDisplayNamesPointer
+ * "Smart pointer" class, closes a ULocaleDisplayNames via uldn_close().
+ * For most methods see the LocalPointerBase base class.
+ *
+ * @see LocalPointerBase
+ * @see LocalPointer
+ * @stable ICU 4.4
+ */
+U_DEFINE_LOCAL_OPEN_POINTER(LocalULocaleDisplayNamesPointer, ULocaleDisplayNames, uldn_close);
+
+U_NAMESPACE_END
+
+#endif
+
+/* getters for state */
+
+/**
+ * Returns the locale used to determine the display names. This is
+ * not necessarily the same locale passed to {@link #uldn_open}.
+ * @param ldn the LocaleDisplayNames instance
+ * @return the display locale 
+ * @stable ICU 4.4
+ */
+U_CAPI const char * U_EXPORT2
+uldn_getLocale(const ULocaleDisplayNames *ldn);
+
+/**
+ * Returns the dialect handling used in the display names.
+ * @param ldn the LocaleDisplayNames instance
+ * @return the dialect handling enum
+ * @stable ICU 4.4
+ */
+U_CAPI UDialectHandling U_EXPORT2
+uldn_getDialectHandling(const ULocaleDisplayNames *ldn);
+
+/* names for entire locales */
+
+/**
+ * Returns the display name of the provided locale.
+ * @param ldn the LocaleDisplayNames instance
+ * @param locale the locale whose display name to return
+ * @param result receives the display name
+ * @param maxResultSize the size of the result buffer
+ * @param pErrorCode the status code
+ * @return the actual buffer size needed for the display name.  If it's
+ * greater than maxResultSize, the returned name will be truncated.
+ * @stable ICU 4.4
+ */
+U_CAPI int32_t U_EXPORT2
+uldn_localeDisplayName(const ULocaleDisplayNames *ldn,
+                       const char *locale,
+                       UChar *result,
+                       int32_t maxResultSize,
+                       UErrorCode *pErrorCode);
+
+/* names for components of a locale */
+
+/**
+ * Returns the display name of the provided language code.
+ * @param ldn the LocaleDisplayNames instance
+ * @param lang the language code whose display name to return
+ * @param result receives the display name
+ * @param maxResultSize the size of the result buffer
+ * @param pErrorCode the status code
+ * @return the actual buffer size needed for the display name.  If it's
+ * greater than maxResultSize, the returned name will be truncated.
+ * @stable ICU 4.4
+ */
+U_CAPI int32_t U_EXPORT2
+uldn_languageDisplayName(const ULocaleDisplayNames *ldn,
+                         const char *lang,
+                         UChar *result,
+                         int32_t maxResultSize,
+                         UErrorCode *pErrorCode);
+
+/**
+ * Returns the display name of the provided script.
+ * @param ldn the LocaleDisplayNames instance
+ * @param script the script whose display name to return
+ * @param result receives the display name
+ * @param maxResultSize the size of the result buffer
+ * @param pErrorCode the status code
+ * @return the actual buffer size needed for the display name.  If it's
+ * greater than maxResultSize, the returned name will be truncated.
+ * @stable ICU 4.4
+ */
+U_CAPI int32_t U_EXPORT2
+uldn_scriptDisplayName(const ULocaleDisplayNames *ldn,
+                       const char *script,
+                       UChar *result,
+                       int32_t maxResultSize,
+                       UErrorCode *pErrorCode);
+
+/**
+ * Returns the display name of the provided script code.
+ * @param ldn the LocaleDisplayNames instance
+ * @param scriptCode the script code whose display name to return
+ * @param result receives the display name
+ * @param maxResultSize the size of the result buffer
+ * @param pErrorCode the status code
+ * @return the actual buffer size needed for the display name.  If it's
+ * greater than maxResultSize, the returned name will be truncated.
+ * @stable ICU 4.4
+ */
+U_CAPI int32_t U_EXPORT2
+uldn_scriptCodeDisplayName(const ULocaleDisplayNames *ldn,
+                           UScriptCode scriptCode,
+                           UChar *result,
+                           int32_t maxResultSize,
+                           UErrorCode *pErrorCode);
+
+/**
+ * Returns the display name of the provided region code.
+ * @param ldn the LocaleDisplayNames instance
+ * @param region the region code whose display name to return
+ * @param result receives the display name
+ * @param maxResultSize the size of the result buffer
+ * @param pErrorCode the status code
+ * @return the actual buffer size needed for the display name.  If it's
+ * greater than maxResultSize, the returned name will be truncated.
+ * @stable ICU 4.4
+ */
+U_CAPI int32_t U_EXPORT2
+uldn_regionDisplayName(const ULocaleDisplayNames *ldn,
+                       const char *region,
+                       UChar *result,
+                       int32_t maxResultSize,
+                       UErrorCode *pErrorCode);
+
+/**
+ * Returns the display name of the provided variant
+ * @param ldn the LocaleDisplayNames instance
+ * @param variant the variant whose display name to return
+ * @param result receives the display name
+ * @param maxResultSize the size of the result buffer
+ * @param pErrorCode the status code
+ * @return the actual buffer size needed for the display name.  If it's
+ * greater than maxResultSize, the returned name will be truncated.
+ * @stable ICU 4.4
+ */
+U_CAPI int32_t U_EXPORT2
+uldn_variantDisplayName(const ULocaleDisplayNames *ldn,
+                        const char *variant,
+                        UChar *result,
+                        int32_t maxResultSize,
+                        UErrorCode *pErrorCode);
+
+/**
+ * Returns the display name of the provided locale key
+ * @param ldn the LocaleDisplayNames instance
+ * @param key the locale key whose display name to return
+ * @param result receives the display name
+ * @param maxResultSize the size of the result buffer
+ * @param pErrorCode the status code
+ * @return the actual buffer size needed for the display name.  If it's
+ * greater than maxResultSize, the returned name will be truncated.
+ * @stable ICU 4.4
+ */
+U_CAPI int32_t U_EXPORT2
+uldn_keyDisplayName(const ULocaleDisplayNames *ldn,
+                    const char *key,
+                    UChar *result,
+                    int32_t maxResultSize,
+                    UErrorCode *pErrorCode);
+
+/**
+ * Returns the display name of the provided value (used with the provided key).
+ * @param ldn the LocaleDisplayNames instance
+ * @param key the locale key
+ * @param value the locale key's value
+ * @param result receives the display name
+ * @param maxResultSize the size of the result buffer
+ * @param pErrorCode the status code
+ * @return the actual buffer size needed for the display name.  If it's
+ * greater than maxResultSize, the returned name will be truncated.
+ * @stable ICU 4.4
+ */
+U_CAPI int32_t U_EXPORT2
+uldn_keyValueDisplayName(const ULocaleDisplayNames *ldn,
+                         const char *key,
+                         const char *value,
+                         UChar *result,
+                         int32_t maxResultSize,
+                         UErrorCode *pErrorCode);
+
+/**
+* Returns an instance of LocaleDisplayNames that returns names formatted
+* for the provided locale, using the provided UDisplayContext settings.
+*
+* @param locale The display locale 
+* @param contexts List of one or more context settings (e.g. for dialect
+*               handling, capitalization, etc.
+* @param length Number of items in the contexts list
+* @param pErrorCode Pointer to UErrorCode input/output status. If at entry this indicates
+*               a failure status, the function will do nothing; otherwise this will be
+*               updated with any new status from the function. 
+* @return a ULocaleDisplayNames instance 
+* @stable ICU 51
+*/
+U_CAPI ULocaleDisplayNames * U_EXPORT2
+uldn_openForContext(const char * locale, UDisplayContext *contexts,
+                    int32_t length, UErrorCode *pErrorCode);
+
+/**
+* Returns the UDisplayContext value for the specified UDisplayContextType.
+* @param ldn the ULocaleDisplayNames instance
+* @param type the UDisplayContextType whose value to return
+* @param pErrorCode Pointer to UErrorCode input/output status. If at entry this indicates
+*               a failure status, the function will do nothing; otherwise this will be
+*               updated with any new status from the function. 
+* @return the UDisplayContextValue for the specified type.
+* @stable ICU 51
+*/
+U_CAPI UDisplayContext U_EXPORT2
+uldn_getContext(const ULocaleDisplayNames *ldn, UDisplayContextType type,
+                UErrorCode *pErrorCode);
+
+#endif  /* !UCONFIG_NO_FORMATTING */
+#endif  /* __ULDNAMES_H__ */

thirdparty/icu4c/common/unicode/ulocale.h

@@ -0,0 +1,226 @@
+// © 2023 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+
+#ifndef ULOCALE_H
+#define ULOCALE_H
+
+#include "unicode/localpointer.h"
+#include "unicode/uenum.h"
+#include "unicode/utypes.h"
+
+/**
+ * \file
+ * \brief C API: Locale ID functionality similar to C++ class Locale
+ */
+
+/**
+ * Opaque C service object type for the locale API
+ * @stable ICU 74
+ */
+struct ULocale;
+
+/**
+ * C typedef for struct ULocale.
+ * @stable ICU 74
+ */
+typedef struct ULocale ULocale;
+
+/**
+ * Constructs an ULocale from the locale ID.
+ * The created ULocale should be destroyed by calling
+ * ulocale_close();
+ * @param localeID the locale, a const char * pointer (need not be terminated when
+ *               the length is non-negative)
+ * @param length the length of the locale; if negative, then the locale need to be
+ *               null terminated.
+ * @param err the error code
+ * @return the locale.
+ *
+ * @stable ICU 74
+ */
+U_CAPI ULocale* U_EXPORT2
+ulocale_openForLocaleID(const char* localeID, int32_t length, UErrorCode* err);
+
+/**
+ * Constructs an ULocale from the provided IETF BCP 47 language tag.
+ * The created ULocale should be destroyed by calling
+ * ulocale_close();
+ * @param tag the language tag, defined as IETF BCP 47 language tag, const
+ *            char* pointer (need not be terminated when the length is non-negative)
+ * @param length the length of the tag; if negative, then the tag need to be
+ *               null terminated.
+ * @param err the error code
+ * @return the locale.
+ *
+ * @stable ICU 74
+ */
+U_CAPI ULocale* U_EXPORT2
+ulocale_openForLanguageTag(const char* tag, int32_t length, UErrorCode* err);
+
+/**
+ * Close the locale and destroy it's internal states.
+ *
+ * @param locale the locale
+ * @stable ICU 74
+ */
+U_CAPI void U_EXPORT2
+ulocale_close(ULocale* locale);
+
+/**
+ * Returns the locale's ISO-639 language code.
+ *
+ * @param locale the locale
+ * @return      the language code of the locale.
+ * @stable ICU 74
+ */
+U_CAPI const char* U_EXPORT2
+ulocale_getLanguage(const ULocale* locale);
+
+/**
+ * Returns the locale's ISO-15924 abbreviation script code.
+ *
+ * @param locale the locale
+ * @return      A pointer to the script.
+ * @stable ICU 74
+ */
+U_CAPI const char* U_EXPORT2
+ulocale_getScript(const ULocale* locale);
+
+/**
+ * Returns the locale's ISO-3166 region code.
+ *
+ * @param locale the locale
+ * @return      A pointer to the region.
+ * @stable ICU 74
+ */
+U_CAPI const char* U_EXPORT2
+ulocale_getRegion(const ULocale* locale);
+
+/**
+ * Returns the locale's variant code.
+ *
+ * @param locale the locale
+ * @return      A pointer to the variant.
+ * @stable ICU 74
+ */
+U_CAPI const char* U_EXPORT2
+ulocale_getVariant(const ULocale* locale);
+
+/**
+ * Returns the programmatic name of the entire locale, with the language,
+ * country and variant separated by underbars. If a field is missing, up
+ * to two leading underbars will occur. Example: "en", "de_DE", "en_US_WIN",
+ * "de__POSIX", "fr__MAC", "__MAC", "_MT", "_FR_EURO"
+ *
+ * @param locale the locale
+ * @return      A pointer to "name".
+ * @stable ICU 74
+ */
+U_CAPI const char* U_EXPORT2
+ulocale_getLocaleID(const ULocale* locale);
+
+/**
+ * Returns the programmatic name of the entire locale as ulocale_getLocaleID()
+ * would return, but without keywords.
+ *
+ * @param locale the locale
+ * @return      A pointer to "base name".
+ * @stable ICU 74
+ */
+U_CAPI const char* U_EXPORT2
+ulocale_getBaseName(const ULocale* locale);
+
+/**
+ * Gets the bogus state. Locale object can be bogus if it doesn't exist
+ *
+ * @param locale the locale
+ * @return false if it is a real locale, true if it is a bogus locale
+ * @stable ICU 74
+ */
+U_CAPI bool U_EXPORT2
+ulocale_isBogus(const ULocale* locale);
+
+/**
+ * Gets the list of keywords for the specified locale.
+ *
+ * @param locale the locale
+ * @param err the error code
+ * @return pointer to UEnumeration, or nullptr if there are no keywords.
+ * Client must call uenum_close() to dispose the returned value.
+ * @stable ICU 74
+ */
+U_CAPI UEnumeration* U_EXPORT2
+ulocale_getKeywords(const ULocale* locale, UErrorCode *err);
+
+/**
+ * Gets the list of unicode keywords for the specified locale.
+ *
+ * @param locale the locale
+ * @param err the error code
+ * @return pointer to UEnumeration, or nullptr if there are no keywords.
+ * Client must call uenum_close() to dispose the returned value.
+ * @stable ICU 74
+ */
+U_CAPI UEnumeration* U_EXPORT2
+ulocale_getUnicodeKeywords(const ULocale* locale, UErrorCode *err);
+
+/**
+ * Gets the value for a keyword.
+ *
+ * This uses legacy keyword=value pairs, like "collation=phonebook".
+ *
+ * @param locale the locale
+ * @param keyword the keyword, a const char * pointer (need not be
+ *                terminated when the length is non-negative)
+ * @param keywordLength the length of the keyword; if negative, then the
+ *                      keyword need to be null terminated.
+ * @param valueBuffer The buffer to receive the value.
+ * @param valueBufferCapacity The capacity of receiving valueBuffer.
+ * @param err the error code
+ * @stable ICU 74
+ */
+U_CAPI int32_t U_EXPORT2
+ulocale_getKeywordValue(
+    const ULocale* locale, const char* keyword, int32_t keywordLength,
+    char* valueBuffer, int32_t valueBufferCapacity, UErrorCode *err);
+
+/**
+ * Gets the Unicode value for a Unicode keyword.
+ *
+ * This uses Unicode key-value pairs, like "co-phonebk".
+ *
+ * @param locale the locale
+ * @param keyword the Unicode keyword, a const char * pointer (need not be
+ *                terminated when the length is non-negative)
+ * @param keywordLength the length of the Unicode keyword; if negative,
+ *                      then the keyword need to be null terminated.
+ * @param valueBuffer The buffer to receive the Unicode value.
+ * @param valueBufferCapacity The capacity of receiving valueBuffer.
+ * @param err the error code
+ * @stable ICU 74
+ */
+U_CAPI int32_t U_EXPORT2
+ulocale_getUnicodeKeywordValue(
+    const ULocale* locale, const char* keyword, int32_t keywordLength,
+    char* valueBuffer, int32_t valueBufferCapacity, UErrorCode *err);
+
+#if U_SHOW_CPLUSPLUS_API
+
+U_NAMESPACE_BEGIN
+
+/**
+ * \class LocalULocalePointer
+ * "Smart pointer" class, closes a ULocale via ulocale_close().
+ * For most methods see the LocalPointerBase base class.
+ *
+ * @see LocalPointerBase
+ * @see LocalPointer
+ * @stable ICU 74
+ */
+U_DEFINE_LOCAL_OPEN_POINTER(LocalULocalePointer, ULocale, ulocale_close);
+
+U_NAMESPACE_END
+
+#endif  /* U_SHOW_CPLUSPLUS_API */
+
+#endif /*_ULOCALE */

thirdparty/icu4c/common/unicode/ulocbuilder.h

@@ -0,0 +1,437 @@
+// © 2023 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+#ifndef __ULOCBUILDER_H__
+#define __ULOCBUILDER_H__
+
+#include "unicode/localpointer.h"
+#include "unicode/ulocale.h"
+#include "unicode/utypes.h"
+
+/**
+ * \file
+ * \brief C API: Builder API for Locale
+ */
+
+/**
+ * Opaque C service object type for the locale builder API
+ * @stable ICU 74
+ */
+struct ULocaleBuilder;
+
+/**
+ * C typedef for struct ULocaleBuilder.
+ * @stable ICU 74
+ */
+typedef struct ULocaleBuilder ULocaleBuilder;
+
+/**
+ * <code>ULocaleBuilder</code> is used to build valid <code>locale</code> id
+ * string or IETF BCP 47 language tag from values configured by the setters.
+ * The <code>ULocaleBuilder</code> checks if a value configured by a
+ * setter satisfies the syntax requirements defined by the <code>Locale</code>
+ * class.  A string of Locale created by a <code>ULocaleBuilder</code> is
+ * well-formed and can be transformed to a well-formed IETF BCP 47 language tag
+ * without losing information.
+ *
+ * <p>The following example shows how to create a <code>locale</code> string
+ * with the <code>ULocaleBuilder</code>.
+ * <blockquote>
+ * <pre>
+ *     UErrorCode err = U_ZERO_ERROR;
+ *     char buffer[ULOC_FULLNAME_CAPACITY];
+ *     ULocaleBuilder* builder = ulocbld_open();
+ *     ulocbld_setLanguage(builder, "sr", -1);
+ *     ulocbld_setScript(builder, "Latn", -1);
+ *     ulocbld_setRegion(builder, "RS", -1);
+ *     int32_t length = ulocbld_buildLocaleID(
+ *         builder, buffer, ULOC_FULLNAME_CAPACITY, &error);
+ *     ulocbld_close(builder);
+ * </pre>
+ * </blockquote>
+ *
+ * <p>ULocaleBuilders can be reused; <code>ulocbld_clear()</code> resets all
+ * fields to their default values.
+ *
+ * <p>ULocaleBuilder tracks errors in an internal UErrorCode. For all setters,
+ * except ulocbld_setLanguageTag and ulocbld_setLocale, ULocaleBuilder will return immediately
+ * if the internal UErrorCode is in error state.
+ * To reset internal state and error code, call clear method.
+ * The ulocbld_setLanguageTag and setLocale method will first clear the internal
+ * UErrorCode, then track the error of the validation of the input parameter
+ * into the internal UErrorCode.
+ *
+ * @stable ICU 74
+ */
+
+/**
+ * Constructs an empty ULocaleBuilder. The default value of all
+ * fields, extensions, and private use information is the
+ * empty string. The created builder should be destroyed by calling
+ * ulocbld_close();
+ *
+ * @stable ICU 74
+ */
+U_CAPI ULocaleBuilder* U_EXPORT2
+ulocbld_open(void);
+
+/**
+ * Close the builder and destroy it's internal states.
+ * @param builder the builder
+ * @stable ICU 74
+ */
+U_CAPI void U_EXPORT2
+ulocbld_close(ULocaleBuilder* builder);
+
+/**
+ * Resets the <code>ULocaleBuilder</code> to match the provided
+ * <code>locale</code>.  Existing state is discarded.
+ *
+ * <p>All fields of the locale must be well-formed.
+ * <p>This method clears the internal UErrorCode.
+ *
+ * @param builder the builder
+ * @param locale the locale, a const char * pointer (need not be terminated when
+ *               the length is non-negative)
+ * @param length the length of the locale; if negative, then the locale need to be
+ *               null terminated,
+ *
+ * @stable ICU 74
+ */
+U_CAPI void U_EXPORT2
+ulocbld_setLocale(ULocaleBuilder* builder, const char* locale, int32_t length);
+
+/**
+ * Resets the <code>ULocaleBuilder</code> to match the provided
+ * <code>ULocale</code>. Existing state is discarded.
+ *
+ * <p>The locale must be not bogus.
+ * <p>This method clears the internal UErrorCode.
+ *
+ * @param builder the builder.
+ * @param locale the locale, a ULocale* pointer. The builder adopts the locale
+ *               after the call and the client must not delete it.
+ *
+ * @stable ICU 74
+ */
+U_CAPI void U_EXPORT2
+ulocbld_adoptULocale(ULocaleBuilder* builder, ULocale* locale);
+
+/**
+ * Resets the ULocaleBuilder to match the provided IETF BCP 47 language tag.
+ * Discards the existing state.
+ * The empty string causes the builder to be reset, like {@link #ulocbld_clear}.
+ * Legacy language tags (marked as “Type: grandfathered” in BCP 47)
+ * are converted to their canonical form before being processed.
+ * Otherwise, the <code>language tag</code> must be well-formed,
+ * or else the ulocbld_buildLocaleID() and ulocbld_buildLanguageTag() methods
+ * will later report an U_ILLEGAL_ARGUMENT_ERROR.
+ *
+ * <p>This method clears the internal UErrorCode.
+ *
+ * @param builder the builder
+ * @param tag the language tag, defined as IETF BCP 47 language tag, a
+ *               const char * pointer (need not be terminated when
+ *               the length is non-negative)
+ * @param length the length of the tag; if negative, then the tag need to be
+ *               null terminated,
+ * @stable ICU 74
+ */
+U_CAPI void U_EXPORT2
+ulocbld_setLanguageTag(ULocaleBuilder* builder, const char* tag, int32_t length);
+
+/**
+ * Sets the language.  If <code>language</code> is the empty string, the
+ * language in this <code>ULocaleBuilder</code> is removed. Otherwise, the
+ * <code>language</code> must be well-formed, or else the ulocbld_buildLocaleID()
+ * and ulocbld_buildLanguageTag() methods will
+ * later report an U_ILLEGAL_ARGUMENT_ERROR.
+ *
+ * <p>The syntax of language value is defined as
+ * [unicode_language_subtag](http://www.unicode.org/reports/tr35/tr35.html#unicode_language_subtag).
+ *
+ * @param builder the builder
+ * @param language the language, a const char * pointer (need not be terminated when
+ *               the length is non-negative)
+ * @param length the length of the language; if negative, then the language need to be
+ *               null terminated,
+ * @stable ICU 74
+ */
+U_CAPI void U_EXPORT2
+ulocbld_setLanguage(ULocaleBuilder* builder, const char* language, int32_t length);
+
+/**
+ * Sets the script. If <code>script</code> is the empty string, the script in
+ * this <code>ULocaleBuilder</code> is removed.
+ * Otherwise, the <code>script</code> must be well-formed, or else the
+ * ulocbld_buildLocaleID() and ulocbld_buildLanguageTag() methods will later
+ * report an U_ILLEGAL_ARGUMENT_ERROR.
+ *
+ * <p>The script value is a four-letter script code as
+ * [unicode_script_subtag](http://www.unicode.org/reports/tr35/tr35.html#unicode_script_subtag)
+ * defined by ISO 15924
+ *
+ * @param builder the builder
+ * @param script the script, a const char * pointer (need not be terminated when
+ *               the length is non-negative)
+ * @param length the length of the script; if negative, then the script need to be
+ *               null terminated,
+ * @stable ICU 74
+ */
+U_CAPI void U_EXPORT2
+ulocbld_setScript(ULocaleBuilder* builder, const char* script, int32_t length);
+
+/**
+ * Sets the region.  If region is the empty string, the region in this
+ * <code>ULocaleBuilder</code> is removed. Otherwise, the <code>region</code>
+ * must be well-formed, or else the ulocbld_buildLocaleID() and
+ * ulocbld_buildLanguageTag() methods will later report an
+ * U_ILLEGAL_ARGUMENT_ERROR.
+ *
+ * <p>The region value is defined by
+ *  [unicode_region_subtag](http://www.unicode.org/reports/tr35/tr35.html#unicode_region_subtag)
+ * as a two-letter ISO 3166 code or a three-digit UN M.49 area code.
+ *
+ * <p>The region value in the <code>Locale</code> created by the
+ * <code>ULocaleBuilder</code> is always normalized to upper case.
+ *
+ * @param builder the builder
+ * @param region the region, a const char * pointer (need not be terminated when
+ *               the length is non-negative)
+ * @param length the length of the region; if negative, then the region need to be
+ *               null terminated,
+ * @stable ICU 74
+ */
+U_CAPI void U_EXPORT2
+ulocbld_setRegion(ULocaleBuilder* builder, const char* region, int32_t length);
+
+/**
+ * Sets the variant.  If variant is the empty string, the variant in this
+ * <code>ULocaleBuilder</code> is removed.  Otherwise, the <code>variant</code>
+ * must be well-formed, or else the ulocbld_buildLocaleID() and
+ * ulocbld_buildLanguageTag() methods will later report an
+ * U_ILLEGAL_ARGUMENT_ERROR.
+ *
+ * <p><b>Note:</b> This method checks if <code>variant</code>
+ * satisfies the
+ * [unicode_variant_subtag](http://www.unicode.org/reports/tr35/tr35.html#unicode_variant_subtag)
+ * syntax requirements, and normalizes the value to lowercase letters. However,
+ * the <code>Locale</code> class does not impose any syntactic
+ * restriction on variant. To set an ill-formed variant, use a Locale constructor.
+ * If there are multiple unicode_variant_subtag, the caller must concatenate
+ * them with '-' as separator (ex: "foobar-fibar").
+ *
+ * @param builder the builder
+ * @param variant the variant, a const char * pointer (need not be terminated when
+ *               the length is non-negative)
+ * @param length the length of the variant; if negative, then the variant need to be
+ *               null terminated,
+ * @stable ICU 74
+ */
+U_CAPI void U_EXPORT2
+ulocbld_setVariant(ULocaleBuilder* builder, const char* variant, int32_t length);
+
+/**
+ * Sets the extension for the given key. If the value is the empty string,
+ * the extension is removed.  Otherwise, the <code>key</code> and
+ * <code>value</code> must be well-formed, or else the ulocbld_buildLocaleID()
+ * and ulocbld_buildLanguageTag() methods will
+ * later report an U_ILLEGAL_ARGUMENT_ERROR.
+ *
+ * <p><b>Note:</b> The key ('u') is used for the Unicode locale extension.
+ * Setting a value for this key replaces any existing Unicode locale key/type
+ * pairs with those defined in the extension.
+ *
+ * <p><b>Note:</b> The key ('x') is used for the private use code. To be
+ * well-formed, the value for this key needs only to have subtags of one to
+ * eight alphanumeric characters, not two to eight as in the general case.
+ *
+ * @param builder the builder
+ * @param key the extension key
+ * @param value the value, a const char * pointer (need not be terminated when
+ *               the length is non-negative)
+ * @param length the length of the value; if negative, then the value need to be
+ *               null terminated,
+ * @stable ICU 74
+ */
+U_CAPI void U_EXPORT2
+ulocbld_setExtension(ULocaleBuilder* builder, char key, const char* value, int32_t length);
+
+/**
+ * Sets the Unicode locale keyword type for the given key. If the type
+ * StringPiece is constructed with a nullptr, the keyword is removed.
+ * If the type is the empty string, the keyword is set without type subtags.
+ * Otherwise, the key and type must be well-formed, or else the
+ * ulocbld_buildLocaleID() and ulocbld_buildLanguageTag() methods will later
+ * report an U_ILLEGAL_ARGUMENT_ERROR.
+ *
+ * <p>Keys and types are converted to lower case.
+ *
+ * <p><b>Note</b>:Setting the 'u' extension via {@link #ulocbld_setExtension}
+ * replaces all Unicode locale keywords with those defined in the
+ * extension.
+ *
+ * @param builder the builder
+ * @param key the Unicode locale key, a const char * pointer (need not be
+ *               terminated when the length is non-negative)
+ * @param keyLength the length of the key; if negative, then the key need to be
+ *               null terminated,
+ * @param type the Unicode locale type, a const char * pointer (need not be
+ *               terminated when the length is non-negative)
+ * @param typeLength the length of the type; if negative, then the type need to
+ *               be null terminated,
+ * @return This builder.
+ * @stable ICU 74
+ */
+U_CAPI void U_EXPORT2
+ulocbld_setUnicodeLocaleKeyword(ULocaleBuilder* builder,
+        const char* key, int32_t keyLength, const char* type, int32_t typeLength);
+
+/**
+ * Adds a unicode locale attribute, if not already present, otherwise
+ * has no effect.  The attribute must not be empty string and must be
+ * well-formed or U_ILLEGAL_ARGUMENT_ERROR will be set to status
+ * during the ulocbld_buildLocaleID() and ulocbld_buildLanguageTag() calls.
+ *
+ * @param builder the builder
+ * @param attribute the attribute, a const char * pointer (need not be
+ *               terminated when the length is non-negative)
+ * @param length the length of the attribute; if negative, then the attribute
+ *               need to be null terminated,
+ * @stable ICU 74
+ */
+U_CAPI void U_EXPORT2
+ulocbld_addUnicodeLocaleAttribute(
+    ULocaleBuilder* builder, const char* attribute, int32_t length);
+
+/**
+ * Removes a unicode locale attribute, if present, otherwise has no
+ * effect.  The attribute must not be empty string and must be well-formed
+ * or U_ILLEGAL_ARGUMENT_ERROR will be set to status during the ulocbld_buildLocaleID()
+ * and ulocbld_buildLanguageTag() calls.
+ *
+ * <p>Attribute comparison for removal is case-insensitive.
+ *
+ * @param builder the builder
+ * @param attribute the attribute, a const char * pointer (need not be
+ *               terminated when the length is non-negative)
+ * @param length the length of the attribute; if negative, then the attribute
+ *               need to be null terminated,
+ * @stable ICU 74
+ */
+U_CAPI void U_EXPORT2
+ulocbld_removeUnicodeLocaleAttribute(
+    ULocaleBuilder* builder, const char* attribute, int32_t length);
+
+/**
+ * Resets the builder to its initial, empty state.
+ * <p>This method clears the internal UErrorCode.
+ *
+ * @param builder the builder
+ * @stable ICU 74
+ */
+U_CAPI void U_EXPORT2
+ulocbld_clear(ULocaleBuilder* builder);
+
+/**
+ * Resets the extensions to their initial, empty state.
+ * Language, script, region and variant are unchanged.
+ *
+ * @param builder the builder
+ * @stable ICU 74
+ */
+U_CAPI void U_EXPORT2
+ulocbld_clearExtensions(ULocaleBuilder* builder);
+
+/**
+ * Build the LocaleID string from the fields set on this builder.
+ * If any set methods or during the ulocbld_buildLocaleID() call require memory
+ * allocation but fail U_MEMORY_ALLOCATION_ERROR will be set to status.
+ * If any of the fields set by the setters are not well-formed, the status
+ * will be set to U_ILLEGAL_ARGUMENT_ERROR. The state of the builder will
+ * not change after the ulocbld_buildLocaleID() call and the caller is
+ * free to keep using the same builder to build more locales.
+ *
+ * @param builder the builder
+ * @param locale the locale id
+ * @param localeCapacity the size of the locale buffer to store the locale id
+ * @param err the error code
+ * @return the length of the locale id in buffer
+ * @stable ICU 74
+ */
+U_CAPI int32_t U_EXPORT2
+ulocbld_buildLocaleID(ULocaleBuilder* builder, char* locale,
+                      int32_t localeCapacity, UErrorCode* err);
+
+/**
+ * Build the ULocale object from the fields set on this builder.
+ * If any set methods or during the ulocbld_buildULocale() call require memory
+ * allocation but fail U_MEMORY_ALLOCATION_ERROR will be set to status.
+ * If any of the fields set by the setters are not well-formed, the status
+ * will be set to U_ILLEGAL_ARGUMENT_ERROR. The state of the builder will
+ * not change after the ulocbld_buildULocale() call and the caller is
+ * free to keep using the same builder to build more locales.
+ *
+ * @param builder the builder.
+ * @param err the error code.
+ * @return the locale, a ULocale* pointer. The created ULocale must be
+ *          destroyed by calling {@link ulocale_close}.
+ * @stable ICU 74
+ */
+U_CAPI ULocale* U_EXPORT2
+ulocbld_buildULocale(ULocaleBuilder* builder, UErrorCode* err);
+
+/**
+ * Build the IETF BCP 47 language tag string from the fields set on this builder.
+ * If any set methods or during the ulocbld_buildLanguageTag() call require memory
+ * allocation but fail U_MEMORY_ALLOCATION_ERROR will be set to status.
+ * If any of the fields set by the setters are not well-formed, the status
+ * will be set to U_ILLEGAL_ARGUMENT_ERROR. The state of the builder will
+ * not change after the ulocbld_buildLanguageTag() call and the caller is free
+ * to keep using the same builder to build more locales.
+ *
+ * @param builder the builder
+ * @param language the language tag
+ * @param languageCapacity the size of the language buffer to store the language
+ * tag
+ * @param err the error code
+ * @return the length of the language tag in buffer
+ * @stable ICU 74
+ */
+U_CAPI int32_t U_EXPORT2
+ulocbld_buildLanguageTag(ULocaleBuilder* builder, char* language,
+                      int32_t languageCapacity, UErrorCode* err);
+
+/**
+ * Sets the UErrorCode if an error occurred while recording sets.
+ * Preserves older error codes in the outErrorCode.
+ *
+ * @param builder the builder
+ * @param outErrorCode Set to an error code that occurred while setting subtags.
+ *                  Unchanged if there is no such error or if outErrorCode
+ *                  already contained an error.
+ * @return true if U_FAILURE(*outErrorCode)
+ * @stable ICU 74
+ */
+U_CAPI UBool U_EXPORT2
+ulocbld_copyErrorTo(const ULocaleBuilder* builder, UErrorCode *outErrorCode);
+
+#if U_SHOW_CPLUSPLUS_API
+
+U_NAMESPACE_BEGIN
+
+/**
+ * \class LocalULocaleBuilderPointer
+ * "Smart pointer" class, closes a ULocaleBuilder via ulocbld_close().
+ * For most methods see the LocalPointerBase base class.
+ *
+ * @see LocalPointerBase
+ * @see LocalPointer
+ * @stable ICU 74
+ */
+U_DEFINE_LOCAL_OPEN_POINTER(LocalULocaleBuilderPointer, ULocaleBuilder, ulocbld_close);
+
+U_NAMESPACE_END
+
+#endif  /* U_SHOW_CPLUSPLUS_API */
+
+#endif  // __ULOCBUILDER_H__

thirdparty/icu4c/common/unicode/umachine.h

@@ -0,0 +1,451 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+/*
+******************************************************************************
+*
+*   Copyright (C) 1999-2015, International Business Machines
+*   Corporation and others.  All Rights Reserved.
+*
+******************************************************************************
+*   file name:  umachine.h
+*   encoding:   UTF-8
+*   tab size:   8 (not used)
+*   indentation:4
+*
+*   created on: 1999sep13
+*   created by: Markus W. Scherer
+*
+*   This file defines basic types and constants for ICU to be
+*   platform-independent. umachine.h and utf.h are included into
+*   utypes.h to provide all the general definitions for ICU.
+*   All of these definitions used to be in utypes.h before
+*   the UTF-handling macros made this unmaintainable.
+*/
+
+#ifndef __UMACHINE_H__
+#define __UMACHINE_H__
+
+
+/**
+ * \file
+ * \brief Basic types and constants for UTF
+ *
+ * <h2> Basic types and constants for UTF </h2>
+ *   This file defines basic types and constants for utf.h to be
+ *   platform-independent. umachine.h and utf.h are included into
+ *   utypes.h to provide all the general definitions for ICU.
+ *   All of these definitions used to be in utypes.h before
+ *   the UTF-handling macros made this unmaintainable.
+ *
+ */
+/*==========================================================================*/
+/* Include platform-dependent definitions                                   */
+/* which are contained in the platform-specific file platform.h             */
+/*==========================================================================*/
+
+#include "unicode/ptypes.h" /* platform.h is included in ptypes.h */
+
+/*
+ * ANSI C headers:
+ * stddef.h defines wchar_t
+ */
+#include <stdbool.h>
+#include <stddef.h>
+
+/*==========================================================================*/
+/* For C wrappers, we use the symbol U_CAPI.                                */
+/* This works properly if the includer is C or C++.                         */
+/* Functions are declared   U_CAPI return-type U_EXPORT2 function-name()... */
+/*==========================================================================*/
+
+/**
+ * \def U_CFUNC
+ * This is used in a declaration of a library private ICU C function.
+ * @stable ICU 2.4
+ */
+
+/**
+ * \def U_CDECL_BEGIN
+ * This is used to begin a declaration of a library private ICU C API.
+ * @stable ICU 2.4
+ */
+
+/**
+ * \def U_CDECL_END
+ * This is used to end a declaration of a library private ICU C API
+ * @stable ICU 2.4
+ */
+
+#ifdef __cplusplus
+#   define U_CFUNC extern "C"
+#   define U_CDECL_BEGIN extern "C" {
+#   define U_CDECL_END   }
+#else
+#   define U_CFUNC extern
+#   define U_CDECL_BEGIN
+#   define U_CDECL_END
+#endif
+
+#ifndef U_ATTRIBUTE_DEPRECATED
+/**
+ * \def U_ATTRIBUTE_DEPRECATED
+ *  This is used for GCC specific attributes
+ * @internal
+ */
+#if U_GCC_MAJOR_MINOR >= 302
+#    define U_ATTRIBUTE_DEPRECATED __attribute__ ((deprecated))
+/**
+ * \def U_ATTRIBUTE_DEPRECATED
+ * This is used for Visual C++ specific attributes 
+ * @internal
+ */
+#elif defined(_MSC_VER) && (_MSC_VER >= 1400)
+#    define U_ATTRIBUTE_DEPRECATED __declspec(deprecated)
+#else
+#    define U_ATTRIBUTE_DEPRECATED
+#endif
+#endif
+
+/** This is used to declare a function as a public ICU C API @stable ICU 2.0*/
+#define U_CAPI U_CFUNC U_EXPORT
+/** Obsolete/same as U_CAPI; was used to declare a function as a stable public ICU C API*/
+#define U_STABLE U_CAPI
+/** Obsolete/same as U_CAPI; was used to declare a function as a draft public ICU C API  */
+#define U_DRAFT  U_CAPI
+/** This is used to declare a function as a deprecated public ICU C API  */
+#define U_DEPRECATED U_CAPI U_ATTRIBUTE_DEPRECATED
+/** Obsolete/same as U_CAPI; was used to declare a function as an obsolete public ICU C API  */
+#define U_OBSOLETE U_CAPI
+/** Obsolete/same as U_CAPI; was used to declare a function as an internal ICU C API  */
+#define U_INTERNAL U_CAPI
+
+// Before ICU 65, function-like, multi-statement ICU macros were just defined as
+// series of statements wrapped in { } blocks and the caller could choose to
+// either treat them as if they were actual functions and end the invocation
+// with a trailing ; creating an empty statement after the block or else omit
+// this trailing ; using the knowledge that the macro would expand to { }.
+//
+// But doing so doesn't work well with macros that look like functions and
+// compiler warnings about empty statements (ICU-20601) and ICU 65 therefore
+// switches to the standard solution of wrapping such macros in do { } while.
+//
+// This will however break existing code that depends on being able to invoke
+// these macros without a trailing ; so to be able to remain compatible with
+// such code the wrapper is itself defined as macros so that it's possible to
+// build ICU 65 and later with the old macro behaviour, like this:
+//
+// export CPPFLAGS='-DUPRV_BLOCK_MACRO_BEGIN="" -DUPRV_BLOCK_MACRO_END=""'
+// runConfigureICU ...
+//
+
+/**
+ * \def UPRV_BLOCK_MACRO_BEGIN
+ * Defined as the "do" keyword by default.
+ * @internal
+ */
+#ifndef UPRV_BLOCK_MACRO_BEGIN
+#define UPRV_BLOCK_MACRO_BEGIN do
+#endif
+
+/**
+ * \def UPRV_BLOCK_MACRO_END
+ * Defined as "while (false)" by default.
+ * @internal
+ */
+#ifndef UPRV_BLOCK_MACRO_END
+#define UPRV_BLOCK_MACRO_END while (false)
+#endif
+
+/*==========================================================================*/
+/* limits for int32_t etc., like in POSIX inttypes.h                        */
+/*==========================================================================*/
+
+#ifndef INT8_MIN
+/** The smallest value an 8 bit signed integer can hold @stable ICU 2.0 */
+#   define INT8_MIN        ((int8_t)(-128))
+#endif
+#ifndef INT16_MIN
+/** The smallest value a 16 bit signed integer can hold @stable ICU 2.0 */
+#   define INT16_MIN       ((int16_t)(-32767-1))
+#endif
+#ifndef INT32_MIN
+/** The smallest value a 32 bit signed integer can hold @stable ICU 2.0 */
+#   define INT32_MIN       ((int32_t)(-2147483647-1))
+#endif
+
+#ifndef INT8_MAX
+/** The largest value an 8 bit signed integer can hold @stable ICU 2.0 */
+#   define INT8_MAX        ((int8_t)(127))
+#endif
+#ifndef INT16_MAX
+/** The largest value a 16 bit signed integer can hold @stable ICU 2.0 */
+#   define INT16_MAX       ((int16_t)(32767))
+#endif
+#ifndef INT32_MAX
+/** The largest value a 32 bit signed integer can hold @stable ICU 2.0 */
+#   define INT32_MAX       ((int32_t)(2147483647))
+#endif
+
+#ifndef UINT8_MAX
+/** The largest value an 8 bit unsigned integer can hold @stable ICU 2.0 */
+#   define UINT8_MAX       ((uint8_t)(255U))
+#endif
+#ifndef UINT16_MAX
+/** The largest value a 16 bit unsigned integer can hold @stable ICU 2.0 */
+#   define UINT16_MAX      ((uint16_t)(65535U))
+#endif
+#ifndef UINT32_MAX
+/** The largest value a 32 bit unsigned integer can hold @stable ICU 2.0 */
+#   define UINT32_MAX      ((uint32_t)(4294967295U))
+#endif
+
+#if defined(U_INT64_T_UNAVAILABLE)
+# error int64_t is required for decimal format and rule-based number format.
+#else
+# ifndef INT64_C
+/**
+ * Provides a platform independent way to specify a signed 64-bit integer constant.
+ * note: may be wrong for some 64 bit platforms - ensure your compiler provides INT64_C
+ * @stable ICU 2.8
+ */
+#   define INT64_C(c) c ## LL
+# endif
+# ifndef UINT64_C
+/**
+ * Provides a platform independent way to specify an unsigned 64-bit integer constant.
+ * note: may be wrong for some 64 bit platforms - ensure your compiler provides UINT64_C
+ * @stable ICU 2.8
+ */
+#   define UINT64_C(c) c ## ULL
+# endif
+# ifndef U_INT64_MIN
+/** The smallest value a 64 bit signed integer can hold @stable ICU 2.8 */
+#     define U_INT64_MIN       ((int64_t)(INT64_C(-9223372036854775807)-1))
+# endif
+# ifndef U_INT64_MAX
+/** The largest value a 64 bit signed integer can hold @stable ICU 2.8 */
+#     define U_INT64_MAX       ((int64_t)(INT64_C(9223372036854775807)))
+# endif
+# ifndef U_UINT64_MAX
+/** The largest value a 64 bit unsigned integer can hold @stable ICU 2.8 */
+#     define U_UINT64_MAX      ((uint64_t)(UINT64_C(18446744073709551615)))
+# endif
+#endif
+
+/*==========================================================================*/
+/* Boolean data type                                                        */
+/*==========================================================================*/
+
+/**
+ * The ICU boolean type, a signed-byte integer.
+ * ICU-specific for historical reasons: The C and C++ standards used to not define type bool.
+ * Also provides a fixed type definition, as opposed to
+ * type bool whose details (e.g., sizeof) may vary by compiler and between C and C++.
+ *
+ * @stable ICU 2.0
+ */
+typedef int8_t UBool;
+
+/**
+ * \def U_DEFINE_FALSE_AND_TRUE
+ * Normally turns off defining macros FALSE=0 & TRUE=1 in public ICU headers.
+ * These obsolete macros sometimes break compilation of other code that
+ * defines enum constants or similar with these names.
+ * C++ has long defined bool/false/true.
+ * C99 also added definitions for these, although as macros; see stdbool.h.
+ *
+ * You may transitionally define U_DEFINE_FALSE_AND_TRUE=1 if you need time to migrate code.
+ *
+ * @internal ICU 68
+ */
+#ifdef U_DEFINE_FALSE_AND_TRUE
+    // Use the predefined value.
+#else
+    // Default to avoiding collision with non-macro definitions of FALSE & TRUE.
+#   define U_DEFINE_FALSE_AND_TRUE 0
+#endif
+
+#if U_DEFINE_FALSE_AND_TRUE || defined(U_IN_DOXYGEN)
+#ifndef TRUE
+/**
+ * The TRUE value of a UBool.
+ *
+ * @deprecated ICU 68 Use standard "true" instead.
+ */
+#   define TRUE  1
+#endif
+#ifndef FALSE
+/**
+ * The FALSE value of a UBool.
+ *
+ * @deprecated ICU 68 Use standard "false" instead.
+ */
+#   define FALSE 0
+#endif
+#endif  // U_DEFINE_FALSE_AND_TRUE
+
+/*==========================================================================*/
+/* Unicode data types                                                       */
+/*==========================================================================*/
+
+/* wchar_t-related definitions -------------------------------------------- */
+
+/*
+ * \def U_WCHAR_IS_UTF16
+ * Defined if wchar_t uses UTF-16.
+ *
+ * @stable ICU 2.0
+ */
+/*
+ * \def U_WCHAR_IS_UTF32
+ * Defined if wchar_t uses UTF-32.
+ *
+ * @stable ICU 2.0
+ */
+#if !defined(U_WCHAR_IS_UTF16) && !defined(U_WCHAR_IS_UTF32)
+#   ifdef __STDC_ISO_10646__
+#       if (U_SIZEOF_WCHAR_T==2)
+#           define U_WCHAR_IS_UTF16
+#       elif (U_SIZEOF_WCHAR_T==4)
+#           define  U_WCHAR_IS_UTF32
+#       endif
+#   elif defined __UCS2__
+#       if (U_PF_OS390 <= U_PLATFORM && U_PLATFORM <= U_PF_OS400) && (U_SIZEOF_WCHAR_T==2)
+#           define U_WCHAR_IS_UTF16
+#       endif
+#   elif defined(__UCS4__) || (U_PLATFORM == U_PF_OS400 && defined(__UTF32__))
+#       if (U_SIZEOF_WCHAR_T==4)
+#           define U_WCHAR_IS_UTF32
+#       endif
+#   elif U_PLATFORM_IS_DARWIN_BASED || (U_SIZEOF_WCHAR_T==4 && U_PLATFORM_IS_LINUX_BASED)
+#       define U_WCHAR_IS_UTF32
+#   elif U_PLATFORM_HAS_WIN32_API
+#       define U_WCHAR_IS_UTF16
+#   endif
+#endif
+
+/* UChar and UChar32 definitions -------------------------------------------- */
+
+/** Number of bytes in a UChar (always 2). @stable ICU 2.0 */
+#define U_SIZEOF_UCHAR 2
+
+/**
+ * \def U_CHAR16_IS_TYPEDEF
+ * If 1, then char16_t is a typedef and not a real type (yet)
+ * @internal
+ */
+#if defined(_MSC_VER) && (_MSC_VER < 1900)
+// Versions of Visual Studio/MSVC below 2015 do not support char16_t as a real type,
+// and instead use a typedef.  https://msdn.microsoft.com/library/bb531344.aspx
+# define U_CHAR16_IS_TYPEDEF 1
+#else
+# define U_CHAR16_IS_TYPEDEF 0
+#endif
+
+
+/**
+ * \var UChar
+ *
+ * The base type for UTF-16 code units and pointers.
+ * Unsigned 16-bit integer.
+ * Starting with ICU 59, C++ API uses char16_t directly, while C API continues to use UChar.
+ *
+ * UChar is configurable by defining the macro UCHAR_TYPE
+ * on the preprocessor or compiler command line:
+ * -DUCHAR_TYPE=uint16_t or -DUCHAR_TYPE=wchar_t (if U_SIZEOF_WCHAR_T==2) etc.
+ * (The UCHAR_TYPE can also be \#defined earlier in this file, for outside the ICU library code.)
+ * This is for transitional use from application code that uses uint16_t or wchar_t for UTF-16.
+ *
+ * The default is UChar=char16_t.
+ *
+ * C++11 defines char16_t as bit-compatible with uint16_t, but as a distinct type.
+ *
+ * In C, char16_t is a simple typedef of uint_least16_t.
+ * ICU requires uint_least16_t=uint16_t for data memory mapping.
+ * On macOS, char16_t is not available because the uchar.h standard header is missing.
+ *
+ * @stable ICU 4.4
+ */
+
+#if 1
+    // #if 1 is normal. UChar defaults to char16_t in C++.
+    // For configuration testing of UChar=uint16_t temporarily change this to #if 0.
+#else
+#   define UCHAR_TYPE uint16_t
+#endif
+
+#if defined(U_ALL_IMPLEMENTATION) || !defined(UCHAR_TYPE)
+    typedef char16_t UChar;
+#else
+    typedef UCHAR_TYPE UChar;
+#endif
+
+/**
+ * \var OldUChar
+ * Default ICU 58 definition of UChar.
+ * A base type for UTF-16 code units and pointers.
+ * Unsigned 16-bit integer.
+ *
+ * Define OldUChar to be wchar_t if that is 16 bits wide.
+ * If wchar_t is not 16 bits wide, then define UChar to be uint16_t.
+ *
+ * This makes the definition of OldUChar platform-dependent
+ * but allows direct string type compatibility with platforms with
+ * 16-bit wchar_t types.
+ *
+ * This is how UChar was defined in ICU 58, for transition convenience.
+ * Exception: ICU 58 UChar was defined to UCHAR_TYPE if that macro was defined.
+ * The current UChar responds to UCHAR_TYPE but OldUChar does not.
+ *
+ * @stable ICU 59
+ */
+#if U_SIZEOF_WCHAR_T==2
+    typedef wchar_t OldUChar;
+#elif defined(__CHAR16_TYPE__)
+    typedef __CHAR16_TYPE__ OldUChar;
+#else
+    typedef uint16_t OldUChar;
+#endif
+
+/**
+ * Define UChar32 as a type for single Unicode code points.
+ * UChar32 is a signed 32-bit integer (same as int32_t).
+ *
+ * The Unicode code point range is 0..0x10ffff.
+ * All other values (negative or >=0x110000) are illegal as Unicode code points.
+ * They may be used as sentinel values to indicate "done", "error"
+ * or similar non-code point conditions.
+ *
+ * Before ICU 2.4 (Jitterbug 2146), UChar32 was defined
+ * to be wchar_t if that is 32 bits wide (wchar_t may be signed or unsigned)
+ * or else to be uint32_t.
+ * That is, the definition of UChar32 was platform-dependent.
+ *
+ * @see U_SENTINEL
+ * @stable ICU 2.4
+ */
+typedef int32_t UChar32;
+
+/**
+ * This value is intended for sentinel values for APIs that
+ * (take or) return single code points (UChar32).
+ * It is outside of the Unicode code point range 0..0x10ffff.
+ * 
+ * For example, a "done" or "error" value in a new API
+ * could be indicated with U_SENTINEL.
+ *
+ * ICU APIs designed before ICU 2.4 usually define service-specific "done"
+ * values, mostly 0xffff.
+ * Those may need to be distinguished from
+ * actual U+ffff text contents by calling functions like
+ * CharacterIterator::hasNext() or UnicodeString::length().
+ *
+ * @return -1
+ * @see UChar32
+ * @stable ICU 2.4
+ */
+#define U_SENTINEL (-1)
+
+#include "unicode/urename.h"
+
+#endif

thirdparty/icu4c/common/unicode/umisc.h

@@ -0,0 +1,62 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+/*
+**********************************************************************
+*   Copyright (C) 1999-2006, International Business Machines
+*   Corporation and others.  All Rights Reserved.
+**********************************************************************
+*   file name:  umisc.h
+*   encoding:   UTF-8
+*   tab size:   8 (not used)
+*   indentation:4
+*
+*   created on: 1999oct15
+*   created by: Markus W. Scherer
+*/
+
+#ifndef UMISC_H
+#define UMISC_H
+
+#include "unicode/utypes.h"
+
+/**
+ * \file
+ * \brief C API: Miscellaneous definitions
+ *
+ *  This file contains miscellaneous definitions for the C APIs.
+ */
+
+U_CDECL_BEGIN
+
+/** A struct representing a range of text containing a specific field 
+ *  @stable ICU 2.0
+ */
+typedef struct UFieldPosition {
+  /**
+   * The field 
+   * @stable ICU 2.0
+   */
+  int32_t field;
+  /**
+   * The start of the text range containing field 
+   * @stable ICU 2.0
+   */
+  int32_t beginIndex;
+  /** 
+   * The limit of the text range containing field 
+   * @stable ICU 2.0
+   */
+  int32_t endIndex;
+} UFieldPosition;
+
+#if !UCONFIG_NO_SERVICE
+/**
+ * Opaque type returned by registerInstance, registerFactory and unregister for service registration.
+ * @stable ICU 2.6
+ */
+typedef const void* URegistryKey;
+#endif
+
+U_CDECL_END
+
+#endif

thirdparty/icu4c/common/unicode/umutablecptrie.h

@@ -0,0 +1,240 @@
+// © 2017 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+
+// umutablecptrie.h (split out of ucptrie.h)
+// created: 2018jan24 Markus W. Scherer
+
+#ifndef __UMUTABLECPTRIE_H__
+#define __UMUTABLECPTRIE_H__
+
+#include "unicode/utypes.h"
+
+#include "unicode/ucpmap.h"
+#include "unicode/ucptrie.h"
+#include "unicode/utf8.h"
+
+#if U_SHOW_CPLUSPLUS_API
+#include "unicode/localpointer.h"
+#endif   // U_SHOW_CPLUSPLUS_API
+
+U_CDECL_BEGIN
+
+/**
+ * \file
+ * \brief C API: This file defines a mutable Unicode code point trie.
+ *
+ * @see UCPTrie
+ * @see UMutableCPTrie
+ */
+
+/**
+ * Mutable Unicode code point trie.
+ * Fast map from Unicode code points (U+0000..U+10FFFF) to 32-bit integer values.
+ * For details see https://icu.unicode.org/design/struct/utrie
+ *
+ * Setting values (especially ranges) and lookup is fast.
+ * The mutable trie is only somewhat space-efficient.
+ * It builds a compacted, immutable UCPTrie.
+ *
+ * This trie can be modified while iterating over its contents.
+ * For example, it is possible to merge its values with those from another
+ * set of ranges (e.g., another mutable or immutable trie):
+ * Iterate over those source ranges; for each of them iterate over this trie;
+ * add the source value into the value of each trie range.
+ *
+ * @see UCPTrie
+ * @see umutablecptrie_buildImmutable
+ * @stable ICU 63
+ */
+typedef struct UMutableCPTrie UMutableCPTrie;
+
+/**
+ * Creates a mutable trie that initially maps each Unicode code point to the same value.
+ * It uses 32-bit data values until umutablecptrie_buildImmutable() is called.
+ * umutablecptrie_buildImmutable() takes a valueWidth parameter which
+ * determines the number of bits in the data value in the resulting UCPTrie.
+ * You must umutablecptrie_close() the trie once you are done using it.
+ *
+ * @param initialValue the initial value that is set for all code points
+ * @param errorValue the value for out-of-range code points and ill-formed UTF-8/16
+ * @param pErrorCode an in/out ICU UErrorCode
+ * @return the trie
+ * @stable ICU 63
+ */
+U_CAPI UMutableCPTrie * U_EXPORT2
+umutablecptrie_open(uint32_t initialValue, uint32_t errorValue, UErrorCode *pErrorCode);
+
+/**
+ * Clones a mutable trie.
+ * You must umutablecptrie_close() the clone once you are done using it.
+ *
+ * @param other the trie to clone
+ * @param pErrorCode an in/out ICU UErrorCode
+ * @return the trie clone
+ * @stable ICU 63
+ */
+U_CAPI UMutableCPTrie * U_EXPORT2
+umutablecptrie_clone(const UMutableCPTrie *other, UErrorCode *pErrorCode);
+
+/**
+ * Closes a mutable trie and releases associated memory.
+ *
+ * @param trie the trie
+ * @stable ICU 63
+ */
+U_CAPI void U_EXPORT2
+umutablecptrie_close(UMutableCPTrie *trie);
+
+/**
+ * Creates a mutable trie with the same contents as the UCPMap.
+ * You must umutablecptrie_close() the mutable trie once you are done using it.
+ *
+ * @param map the source map
+ * @param pErrorCode an in/out ICU UErrorCode
+ * @return the mutable trie
+ * @stable ICU 63
+ */
+U_CAPI UMutableCPTrie * U_EXPORT2
+umutablecptrie_fromUCPMap(const UCPMap *map, UErrorCode *pErrorCode);
+
+/**
+ * Creates a mutable trie with the same contents as the immutable one.
+ * You must umutablecptrie_close() the mutable trie once you are done using it.
+ *
+ * @param trie the immutable trie
+ * @param pErrorCode an in/out ICU UErrorCode
+ * @return the mutable trie
+ * @stable ICU 63
+ */
+U_CAPI UMutableCPTrie * U_EXPORT2
+umutablecptrie_fromUCPTrie(const UCPTrie *trie, UErrorCode *pErrorCode);
+
+/**
+ * Returns the value for a code point as stored in the trie.
+ *
+ * @param trie the trie
+ * @param c the code point
+ * @return the value
+ * @stable ICU 63
+ */
+U_CAPI uint32_t U_EXPORT2
+umutablecptrie_get(const UMutableCPTrie *trie, UChar32 c);
+
+/**
+ * Returns the last code point such that all those from start to there have the same value.
+ * Can be used to efficiently iterate over all same-value ranges in a trie.
+ * (This is normally faster than iterating over code points and get()ting each value,
+ * but much slower than a data structure that stores ranges directly.)
+ *
+ * The trie can be modified between calls to this function.
+ *
+ * If the UCPMapValueFilter function pointer is not NULL, then
+ * the value to be delivered is passed through that function, and the return value is the end
+ * of the range where all values are modified to the same actual value.
+ * The value is unchanged if that function pointer is NULL.
+ *
+ * See the same-signature ucptrie_getRange() for a code sample.
+ *
+ * @param trie the trie
+ * @param start range start
+ * @param option defines whether surrogates are treated normally,
+ *               or as having the surrogateValue; usually UCPMAP_RANGE_NORMAL
+ * @param surrogateValue value for surrogates; ignored if option==UCPMAP_RANGE_NORMAL
+ * @param filter a pointer to a function that may modify the trie data value,
+ *     or NULL if the values from the trie are to be used unmodified
+ * @param context an opaque pointer that is passed on to the filter function
+ * @param pValue if not NULL, receives the value that every code point start..end has;
+ *     may have been modified by filter(context, trie value)
+ *     if that function pointer is not NULL
+ * @return the range end code point, or -1 if start is not a valid code point
+ * @stable ICU 63
+ */
+U_CAPI UChar32 U_EXPORT2
+umutablecptrie_getRange(const UMutableCPTrie *trie, UChar32 start,
+                        UCPMapRangeOption option, uint32_t surrogateValue,
+                        UCPMapValueFilter *filter, const void *context, uint32_t *pValue);
+
+/**
+ * Sets a value for a code point.
+ *
+ * @param trie the trie
+ * @param c the code point
+ * @param value the value
+ * @param pErrorCode an in/out ICU UErrorCode
+ * @stable ICU 63
+ */
+U_CAPI void U_EXPORT2
+umutablecptrie_set(UMutableCPTrie *trie, UChar32 c, uint32_t value, UErrorCode *pErrorCode);
+
+/**
+ * Sets a value for each code point [start..end].
+ * Faster and more space-efficient than setting the value for each code point separately.
+ *
+ * @param trie the trie
+ * @param start the first code point to get the value
+ * @param end the last code point to get the value (inclusive)
+ * @param value the value
+ * @param pErrorCode an in/out ICU UErrorCode
+ * @stable ICU 63
+ */
+U_CAPI void U_EXPORT2
+umutablecptrie_setRange(UMutableCPTrie *trie,
+                        UChar32 start, UChar32 end,
+                        uint32_t value, UErrorCode *pErrorCode);
+
+/**
+ * Compacts the data and builds an immutable UCPTrie according to the parameters.
+ * After this, the mutable trie will be empty.
+ *
+ * The mutable trie stores 32-bit values until buildImmutable() is called.
+ * If values shorter than 32 bits are to be stored in the immutable trie,
+ * then the upper bits are discarded.
+ * For example, when the mutable trie contains values 0x81, -0x7f, and 0xa581,
+ * and the value width is 8 bits, then each of these is stored as 0x81
+ * and the immutable trie will return that as an unsigned value.
+ * (Some implementations may want to make productive temporary use of the upper bits
+ * until buildImmutable() discards them.)
+ *
+ * Not every possible set of mappings can be built into a UCPTrie,
+ * because of limitations resulting from speed and space optimizations.
+ * Every Unicode assigned character can be mapped to a unique value.
+ * Typical data yields data structures far smaller than the limitations.
+ *
+ * It is possible to construct extremely unusual mappings that exceed the data structure limits.
+ * In such a case this function will fail with a U_INDEX_OUTOFBOUNDS_ERROR.
+ *
+ * @param trie the trie trie
+ * @param type selects the trie type
+ * @param valueWidth selects the number of bits in a trie data value; if smaller than 32 bits,
+ *                   then the values stored in the trie will be truncated first
+ * @param pErrorCode an in/out ICU UErrorCode
+ *
+ * @see umutablecptrie_fromUCPTrie
+ * @stable ICU 63
+ */
+U_CAPI UCPTrie * U_EXPORT2
+umutablecptrie_buildImmutable(UMutableCPTrie *trie, UCPTrieType type, UCPTrieValueWidth valueWidth,
+                              UErrorCode *pErrorCode);
+
+U_CDECL_END
+
+#if U_SHOW_CPLUSPLUS_API
+
+U_NAMESPACE_BEGIN
+
+/**
+ * \class LocalUMutableCPTriePointer
+ * "Smart pointer" class, closes a UMutableCPTrie via umutablecptrie_close().
+ * For most methods see the LocalPointerBase base class.
+ *
+ * @see LocalPointerBase
+ * @see LocalPointer
+ * @stable ICU 63
+ */
+U_DEFINE_LOCAL_OPEN_POINTER(LocalUMutableCPTriePointer, UMutableCPTrie, umutablecptrie_close);
+
+U_NAMESPACE_END
+
+#endif
+
+#endif

thirdparty/icu4c/common/unicode/unifilt.h

@@ -0,0 +1,136 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+/*
+**********************************************************************
+* Copyright (C) 1999-2010, International Business Machines Corporation and others.
+* All Rights Reserved.
+**********************************************************************
+*   Date        Name        Description
+*   11/17/99    aliu        Creation.
+**********************************************************************
+*/
+#ifndef UNIFILT_H
+#define UNIFILT_H
+
+#include "unicode/utypes.h"
+
+#if U_SHOW_CPLUSPLUS_API
+
+#include "unicode/unifunct.h"
+#include "unicode/unimatch.h"
+
+/**
+ * \file 
+ * \brief C++ API: Unicode Filter
+ */
+
+U_NAMESPACE_BEGIN
+
+/**
+ * U_ETHER is used to represent character values for positions outside
+ * a range.  For example, transliterator uses this to represent
+ * characters outside the range contextStart..contextLimit-1.  This
+ * allows explicit matching by rules and UnicodeSets of text outside a
+ * defined range.
+ * @stable ICU 3.0
+ */
+#define U_ETHER ((char16_t)0xFFFF)
+
+/**
+ *
+ * <code>UnicodeFilter</code> defines a protocol for selecting a
+ * subset of the full range (U+0000 to U+10FFFF) of Unicode characters.
+ * Currently, filters are used in conjunction with classes like
+ * {@link Transliterator} to only process selected characters through a
+ * transformation.
+ *
+ * <p>Note: UnicodeFilter currently stubs out two pure virtual methods
+ * of its base class, UnicodeMatcher.  These methods are toPattern()
+ * and matchesIndexValue().  This is done so that filter classes that
+ * are not actually used as matchers -- specifically, those in the
+ * UnicodeFilterLogic component, and those in tests -- can continue to
+ * work without defining these methods.  As long as a filter is not
+ * used in an RBT during real transliteration, these methods will not
+ * be called.  However, this breaks the UnicodeMatcher base class
+ * protocol, and it is not a correct solution.
+ *
+ * <p>In the future we may revisit the UnicodeMatcher / UnicodeFilter
+ * hierarchy and either redesign it, or simply remove the stubs in
+ * UnicodeFilter and force subclasses to implement the full
+ * UnicodeMatcher protocol.
+ *
+ * @see UnicodeFilterLogic
+ * @stable ICU 2.0
+ */
+class U_COMMON_API UnicodeFilter : public UnicodeFunctor, public UnicodeMatcher {
+
+public:
+    /**
+     * Destructor
+     * @stable ICU 2.0
+     */
+    virtual ~UnicodeFilter();
+
+    /**
+     * Clones this object polymorphically.
+     * The caller owns the result and should delete it when done.
+     * @return clone, or nullptr if an error occurred
+     * @stable ICU 2.4
+     */
+    virtual UnicodeFilter* clone() const override = 0;
+
+    /**
+     * Returns <tt>true</tt> for characters that are in the selected
+     * subset.  In other words, if a character is <b>to be
+     * filtered</b>, then <tt>contains()</tt> returns
+     * <b><tt>false</tt></b>.
+     * @stable ICU 2.0
+     */
+    virtual UBool contains(UChar32 c) const = 0;
+
+    /**
+     * UnicodeFunctor API.  Cast 'this' to a UnicodeMatcher* pointer
+     * and return the pointer.
+     * @stable ICU 2.4
+     */
+    virtual UnicodeMatcher* toMatcher() const override;
+
+    /**
+     * Implement UnicodeMatcher API.
+     * @stable ICU 2.4
+     */
+    virtual UMatchDegree matches(const Replaceable& text,
+                                 int32_t& offset,
+                                 int32_t limit,
+                                 UBool incremental) override;
+
+    /**
+     * UnicodeFunctor API.  Nothing to do.
+     * @stable ICU 2.4
+     */
+    virtual void setData(const TransliterationRuleData*) override;
+
+    /**
+     * ICU "poor man's RTTI", returns a UClassID for this class.
+     *
+     * @stable ICU 2.2
+     */
+    static UClassID U_EXPORT2 getStaticClassID();
+
+protected:
+
+    /*
+     * Since this class has pure virtual functions,
+     * a constructor can't be used.
+     * @stable ICU 2.0
+     */
+/*    UnicodeFilter();*/
+};
+
+/*inline UnicodeFilter::UnicodeFilter() {}*/
+
+U_NAMESPACE_END
+
+#endif /* U_SHOW_CPLUSPLUS_API */
+
+#endif

thirdparty/icu4c/common/unicode/unifunct.h

@@ -0,0 +1,132 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+/*
+**********************************************************************
+*   Copyright (c) 2002-2005, International Business Machines Corporation
+*   and others.  All Rights Reserved.
+**********************************************************************
+*   Date        Name        Description
+*   01/14/2002  aliu        Creation.
+**********************************************************************
+*/
+#ifndef UNIFUNCT_H
+#define UNIFUNCT_H
+
+#include "unicode/utypes.h"
+
+#if U_SHOW_CPLUSPLUS_API
+
+#include "unicode/uobject.h"
+
+/**
+ * \file 
+ * \brief C++ API: Unicode Functor
+ */
+ 
+U_NAMESPACE_BEGIN
+
+class UnicodeMatcher;
+class UnicodeReplacer;
+class TransliterationRuleData;
+
+/**
+ * <code>UnicodeFunctor</code> is an abstract base class for objects
+ * that perform match and/or replace operations on Unicode strings.
+ * @author Alan Liu
+ * @stable ICU 2.4
+ */
+class U_COMMON_API UnicodeFunctor : public UObject {
+
+public:
+
+    /**
+     * Destructor
+     * @stable ICU 2.4
+     */
+    virtual ~UnicodeFunctor();
+
+    /**
+     * Return a copy of this object.  All UnicodeFunctor objects
+     * have to support cloning in order to allow classes using
+     * UnicodeFunctor to implement cloning.
+     * @stable ICU 2.4
+     */
+    virtual UnicodeFunctor* clone() const = 0;
+
+    /**
+     * Cast 'this' to a UnicodeMatcher* pointer and return the
+     * pointer, or null if this is not a UnicodeMatcher*.  Subclasses
+     * that mix in UnicodeMatcher as a base class must override this.
+     * This protocol is required because a pointer to a UnicodeFunctor
+     * cannot be cast to a pointer to a UnicodeMatcher, since
+     * UnicodeMatcher is a mixin that does not derive from
+     * UnicodeFunctor.
+     * @stable ICU 2.4
+     */
+    virtual UnicodeMatcher* toMatcher() const;
+
+    /**
+     * Cast 'this' to a UnicodeReplacer* pointer and return the
+     * pointer, or null if this is not a UnicodeReplacer*.  Subclasses
+     * that mix in UnicodeReplacer as a base class must override this.
+     * This protocol is required because a pointer to a UnicodeFunctor
+     * cannot be cast to a pointer to a UnicodeReplacer, since
+     * UnicodeReplacer is a mixin that does not derive from
+     * UnicodeFunctor.
+     * @stable ICU 2.4
+     */
+    virtual UnicodeReplacer* toReplacer() const;
+
+    /**
+     * Return the class ID for this class.  This is useful only for
+     * comparing to a return value from getDynamicClassID().
+     * @return          The class ID for all objects of this class.
+     * @stable ICU 2.0
+     */
+    static UClassID U_EXPORT2 getStaticClassID();
+
+    /**
+     * Returns a unique class ID <b>polymorphically</b>.  This method
+     * is to implement a simple version of RTTI, since not all C++
+     * compilers support genuine RTTI.  Polymorphic operator==() and
+     * clone() methods call this method.
+     *
+     * <p>Concrete subclasses of UnicodeFunctor should use the macro
+     *    UOBJECT_DEFINE_RTTI_IMPLEMENTATION from uobject.h to
+     *    provide definitions getStaticClassID and getDynamicClassID.
+     *
+     * @return The class ID for this object. All objects of a given
+     * class have the same class ID.  Objects of other classes have
+     * different class IDs.
+     * @stable ICU 2.4
+     */
+    virtual UClassID getDynamicClassID() const override = 0;
+
+    /**
+     * Set the data object associated with this functor.  The data
+     * object provides context for functor-to-standin mapping.  This
+     * method is required when assigning a functor to a different data
+     * object.  This function MAY GO AWAY later if the architecture is
+     * changed to pass data object pointers through the API.
+     * @internal ICU 2.1
+     */
+    virtual void setData(const TransliterationRuleData*) = 0;
+
+protected:
+
+    /**
+     * Since this class has pure virtual functions,
+     * a constructor can't be used.
+     * @stable ICU 2.0
+     */
+    /*UnicodeFunctor();*/
+
+};
+
+/*inline UnicodeFunctor::UnicodeFunctor() {}*/
+
+U_NAMESPACE_END
+
+#endif /* U_SHOW_CPLUSPLUS_API */
+
+#endif

thirdparty/icu4c/common/unicode/unimatch.h

@@ -0,0 +1,168 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+/*
+* Copyright (C) 2001-2005, International Business Machines Corporation and others. All Rights Reserved.
+**********************************************************************
+*   Date        Name        Description
+*   07/18/01    aliu        Creation.
+**********************************************************************
+*/
+#ifndef UNIMATCH_H
+#define UNIMATCH_H
+
+#include "unicode/utypes.h"
+
+/**
+ * \file 
+ * \brief C++ API: Unicode Matcher
+ */
+
+#if U_SHOW_CPLUSPLUS_API
+
+U_NAMESPACE_BEGIN
+
+class Replaceable;
+class UnicodeString;
+class UnicodeSet;
+
+/**
+ * Constants returned by <code>UnicodeMatcher::matches()</code>
+ * indicating the degree of match.
+ * @stable ICU 2.4
+ */
+enum UMatchDegree {
+    /**
+     * Constant returned by <code>matches()</code> indicating a
+     * mismatch between the text and this matcher.  The text contains
+     * a character which does not match, or the text does not contain
+     * all desired characters for a non-incremental match.
+     * @stable ICU 2.4
+     */
+    U_MISMATCH,
+    
+    /**
+     * Constant returned by <code>matches()</code> indicating a
+     * partial match between the text and this matcher.  This value is
+     * only returned for incremental match operations.  All characters
+     * of the text match, but more characters are required for a
+     * complete match.  Alternatively, for variable-length matchers,
+     * all characters of the text match, and if more characters were
+     * supplied at limit, they might also match.
+     * @stable ICU 2.4
+     */
+    U_PARTIAL_MATCH,
+    
+    /**
+     * Constant returned by <code>matches()</code> indicating a
+     * complete match between the text and this matcher.  For an
+     * incremental variable-length match, this value is returned if
+     * the given text matches, and it is known that additional
+     * characters would not alter the extent of the match.
+     * @stable ICU 2.4
+     */
+    U_MATCH
+};
+
+/**
+ * <code>UnicodeMatcher</code> defines a protocol for objects that can
+ * match a range of characters in a Replaceable string.
+ * @stable ICU 2.4
+ */
+class U_COMMON_API UnicodeMatcher /* not : public UObject because this is an interface/mixin class */ {
+
+public:
+    /**
+     * Destructor.
+     * @stable ICU 2.4
+     */
+    virtual ~UnicodeMatcher();
+
+    /**
+     * Return a UMatchDegree value indicating the degree of match for
+     * the given text at the given offset.  Zero, one, or more
+     * characters may be matched.
+     *
+     * Matching in the forward direction is indicated by limit >
+     * offset.  Characters from offset forwards to limit-1 will be
+     * considered for matching.
+     * 
+     * Matching in the reverse direction is indicated by limit <
+     * offset.  Characters from offset backwards to limit+1 will be
+     * considered for matching.
+     *
+     * If limit == offset then the only match possible is a zero
+     * character match (which subclasses may implement if desired).
+     *
+     * As a side effect, advance the offset parameter to the limit of
+     * the matched substring.  In the forward direction, this will be
+     * the index of the last matched character plus one.  In the
+     * reverse direction, this will be the index of the last matched
+     * character minus one.
+     *
+     * <p>Note:  This method is not const because some classes may
+     * modify their state as the result of a match.
+     *
+     * @param text the text to be matched
+     * @param offset on input, the index into text at which to begin
+     * matching.  On output, the limit of the matched text.  The
+     * number of matched characters is the output value of offset
+     * minus the input value.  Offset should always point to the
+     * HIGH SURROGATE (leading code unit) of a pair of surrogates,
+     * both on entry and upon return.
+     * @param limit the limit index of text to be matched.  Greater
+     * than offset for a forward direction match, less than offset for
+     * a backward direction match.  The last character to be
+     * considered for matching will be text.charAt(limit-1) in the
+     * forward direction or text.charAt(limit+1) in the backward
+     * direction.
+     * @param incremental if true, then assume further characters may
+     * be inserted at limit and check for partial matching.  Otherwise
+     * assume the text as given is complete.
+     * @return a match degree value indicating a full match, a partial
+     * match, or a mismatch.  If incremental is false then
+     * U_PARTIAL_MATCH should never be returned.
+     * @stable ICU 2.4
+     */
+    virtual UMatchDegree matches(const Replaceable& text,
+                                 int32_t& offset,
+                                 int32_t limit,
+                                 UBool incremental) = 0;
+
+    /**
+     * Returns a string representation of this matcher.  If the result of
+     * calling this function is passed to the appropriate parser, it
+     * will produce another matcher that is equal to this one.
+     * @param result the string to receive the pattern.  Previous
+     * contents will be deleted.
+     * @param escapeUnprintable if true then convert unprintable
+     * character to their hex escape representations, \\uxxxx or
+     * \\Uxxxxxxxx.  Unprintable characters are those other than
+     * U+000A, U+0020..U+007E.
+     * @stable ICU 2.4
+     */
+    virtual UnicodeString& toPattern(UnicodeString& result,
+                                     UBool escapeUnprintable = false) const = 0;
+
+    /**
+     * Returns true if this matcher will match a character c, where c
+     * & 0xFF == v, at offset, in the forward direction (with limit >
+     * offset).  This is used by <tt>RuleBasedTransliterator</tt> for
+     * indexing.
+     * @stable ICU 2.4
+     */
+    virtual UBool matchesIndexValue(uint8_t v) const = 0;
+
+    /**
+     * Union the set of all characters that may be matched by this object
+     * into the given set.
+     * @param toUnionTo the set into which to union the source characters
+     * @stable ICU 2.4
+     */
+    virtual void addMatchSetTo(UnicodeSet& toUnionTo) const = 0;
+};
+
+U_NAMESPACE_END
+
+#endif /* U_SHOW_CPLUSPLUS_API */
+
+#endif

thirdparty/icu4c/common/unicode/unorm.h

@@ -0,0 +1,476 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+/*
+*******************************************************************************
+* Copyright (c) 1996-2016, International Business Machines Corporation
+*               and others. All Rights Reserved.
+*******************************************************************************
+* File unorm.h
+*
+* Created by: Vladimir Weinstein 12052000
+*
+* Modification history :
+*
+* Date        Name        Description
+* 02/01/01    synwee      Added normalization quickcheck enum and method.
+*/
+#ifndef UNORM_H
+#define UNORM_H
+
+#include "unicode/utypes.h"
+
+#if !UCONFIG_NO_NORMALIZATION
+
+#include "unicode/uiter.h"
+#include "unicode/unorm2.h"
+
+/**
+ * \file
+ * \brief C API: Unicode Normalization
+ *
+ * Old Unicode normalization API.
+ *
+ * This API has been replaced by the unorm2.h API and is only available
+ * for backward compatibility. The functions here simply delegate to the
+ * unorm2.h functions, for example unorm2_getInstance() and unorm2_normalize().
+ * There is one exception: The new API does not provide a replacement for unorm_compare().
+ * Its declaration has been moved to unorm2.h.
+ *
+ * <code>unorm_normalize</code> transforms Unicode text into an equivalent composed or
+ * decomposed form, allowing for easier sorting and searching of text.
+ * <code>unorm_normalize</code> supports the standard normalization forms described in
+ * <a href="http://www.unicode.org/unicode/reports/tr15/" target="unicode">
+ * Unicode Standard Annex #15: Unicode Normalization Forms</a>.
+ *
+ * Characters with accents or other adornments can be encoded in
+ * several different ways in Unicode.  For example, take the character A-acute.
+ * In Unicode, this can be encoded as a single character (the
+ * "composed" form):
+ *
+ * \code
+ *      00C1    LATIN CAPITAL LETTER A WITH ACUTE
+ * \endcode
+ *
+ * or as two separate characters (the "decomposed" form):
+ *
+ * \code
+ *      0041    LATIN CAPITAL LETTER A
+ *      0301    COMBINING ACUTE ACCENT
+ * \endcode
+ *
+ * To a user of your program, however, both of these sequences should be
+ * treated as the same "user-level" character "A with acute accent".  When you are searching or
+ * comparing text, you must ensure that these two sequences are treated
+ * equivalently.  In addition, you must handle characters with more than one
+ * accent.  Sometimes the order of a character's combining accents is
+ * significant, while in other cases accent sequences in different orders are
+ * really equivalent.
+ *
+ * Similarly, the string "ffi" can be encoded as three separate letters:
+ *
+ * \code
+ *      0066    LATIN SMALL LETTER F
+ *      0066    LATIN SMALL LETTER F
+ *      0069    LATIN SMALL LETTER I
+ * \endcode
+ *
+ * or as the single character
+ *
+ * \code
+ *      FB03    LATIN SMALL LIGATURE FFI
+ * \endcode
+ *
+ * The ffi ligature is not a distinct semantic character, and strictly speaking
+ * it shouldn't be in Unicode at all, but it was included for compatibility
+ * with existing character sets that already provided it.  The Unicode standard
+ * identifies such characters by giving them "compatibility" decompositions
+ * into the corresponding semantic characters.  When sorting and searching, you
+ * will often want to use these mappings.
+ *
+ * <code>unorm_normalize</code> helps solve these problems by transforming text into the
+ * canonical composed and decomposed forms as shown in the first example above.
+ * In addition, you can have it perform compatibility decompositions so that
+ * you can treat compatibility characters the same as their equivalents.
+ * Finally, <code>unorm_normalize</code> rearranges accents into the proper canonical
+ * order, so that you do not have to worry about accent rearrangement on your
+ * own.
+ *
+ * Form FCD, "Fast C or D", is also designed for collation.
+ * It allows to work on strings that are not necessarily normalized
+ * with an algorithm (like in collation) that works under "canonical closure", i.e., it treats precomposed
+ * characters and their decomposed equivalents the same.
+ *
+ * It is not a normalization form because it does not provide for uniqueness of representation. Multiple strings
+ * may be canonically equivalent (their NFDs are identical) and may all conform to FCD without being identical
+ * themselves.
+ *
+ * The form is defined such that the "raw decomposition", the recursive canonical decomposition of each character,
+ * results in a string that is canonically ordered. This means that precomposed characters are allowed for as long
+ * as their decompositions do not need canonical reordering.
+ *
+ * Its advantage for a process like collation is that all NFD and most NFC texts - and many unnormalized texts -
+ * already conform to FCD and do not need to be normalized (NFD) for such a process. The FCD quick check will
+ * return UNORM_YES for most strings in practice.
+ *
+ * unorm_normalize(UNORM_FCD) may be implemented with UNORM_NFD.
+ *
+ * For more details on FCD see the collation design document:
+ * https://htmlpreview.github.io/?https://github.com/unicode-org/icu-docs/blob/main/design/collation/ICU_collation_design.htm
+ *
+ * ICU collation performs either NFD or FCD normalization automatically if normalization
+ * is turned on for the collator object.
+ * Beyond collation and string search, normalized strings may be useful for string equivalence comparisons,
+ * transliteration/transcription, unique representations, etc.
+ *
+ * The W3C generally recommends to exchange texts in NFC.
+ * Note also that most legacy character encodings use only precomposed forms and often do not
+ * encode any combining marks by themselves. For conversion to such character encodings the
+ * Unicode text needs to be normalized to NFC.
+ * For more usage examples, see the Unicode Standard Annex.
+ */
+
+// Do not conditionalize the following enum with #ifndef U_HIDE_DEPRECATED_API,
+// it is needed for layout of Normalizer object.
+#ifndef U_FORCE_HIDE_DEPRECATED_API
+
+/**
+ * Constants for normalization modes.
+ * @deprecated ICU 56 Use unorm2.h instead.
+ */
+typedef enum {
+  /** No decomposition/composition. @deprecated ICU 56 Use unorm2.h instead. */
+  UNORM_NONE = 1,
+  /** Canonical decomposition. @deprecated ICU 56 Use unorm2.h instead. */
+  UNORM_NFD = 2,
+  /** Compatibility decomposition. @deprecated ICU 56 Use unorm2.h instead. */
+  UNORM_NFKD = 3,
+  /** Canonical decomposition followed by canonical composition. @deprecated ICU 56 Use unorm2.h instead. */
+  UNORM_NFC = 4,
+  /** Default normalization. @deprecated ICU 56 Use unorm2.h instead. */
+  UNORM_DEFAULT = UNORM_NFC,
+  /** Compatibility decomposition followed by canonical composition. @deprecated ICU 56 Use unorm2.h instead. */
+  UNORM_NFKC =5,
+  /** "Fast C or D" form. @deprecated ICU 56 Use unorm2.h instead. */
+  UNORM_FCD = 6,
+
+  /** One more than the highest normalization mode constant. @deprecated ICU 56 Use unorm2.h instead. */
+  UNORM_MODE_COUNT
+} UNormalizationMode;
+
+#endif  // U_FORCE_HIDE_DEPRECATED_API
+
+#ifndef U_HIDE_DEPRECATED_API
+
+/**
+ * Constants for options flags for normalization.
+ * Use 0 for default options,
+ * including normalization according to the Unicode version
+ * that is currently supported by ICU (see u_getUnicodeVersion).
+ * @deprecated ICU 56 Use unorm2.h instead.
+ */
+enum {
+    /**
+     * Options bit set value to select Unicode 3.2 normalization
+     * (except NormalizationCorrections).
+     * At most one Unicode version can be selected at a time.
+     * @deprecated ICU 56 Use unorm2.h instead.
+     */
+    UNORM_UNICODE_3_2=0x20
+};
+
+/**
+ * Lowest-order bit number of unorm_compare() options bits corresponding to
+ * normalization options bits.
+ *
+ * The options parameter for unorm_compare() uses most bits for
+ * itself and for various comparison and folding flags.
+ * The most significant bits, however, are shifted down and passed on
+ * to the normalization implementation.
+ * (That is, from unorm_compare(..., options, ...),
+ * options>>UNORM_COMPARE_NORM_OPTIONS_SHIFT will be passed on to the
+ * internal normalization functions.)
+ *
+ * @see unorm_compare
+ * @deprecated ICU 56 Use unorm2.h instead.
+ */
+#define UNORM_COMPARE_NORM_OPTIONS_SHIFT 20
+
+/**
+ * Normalize a string.
+ * The string will be normalized according the specified normalization mode
+ * and options.
+ * The source and result buffers must not be the same, nor overlap.
+ *
+ * @param source The string to normalize.
+ * @param sourceLength The length of source, or -1 if NUL-terminated.
+ * @param mode The normalization mode; one of UNORM_NONE,
+ *             UNORM_NFD, UNORM_NFC, UNORM_NFKC, UNORM_NFKD, UNORM_DEFAULT.
+ * @param options The normalization options, ORed together (0 for no options).
+ * @param result A pointer to a buffer to receive the result string.
+ *               The result string is NUL-terminated if possible.
+ * @param resultLength The maximum size of result.
+ * @param status A pointer to a UErrorCode to receive any errors.
+ * @return The total buffer size needed; if greater than resultLength,
+ *         the output was truncated, and the error code is set to U_BUFFER_OVERFLOW_ERROR.
+ * @deprecated ICU 56 Use unorm2.h instead.
+ */
+U_DEPRECATED int32_t U_EXPORT2
+unorm_normalize(const UChar *source, int32_t sourceLength,
+                UNormalizationMode mode, int32_t options,
+                UChar *result, int32_t resultLength,
+                UErrorCode *status);
+
+/**
+ * Performing quick check on a string, to quickly determine if the string is
+ * in a particular normalization format.
+ * Three types of result can be returned UNORM_YES, UNORM_NO or
+ * UNORM_MAYBE. Result UNORM_YES indicates that the argument
+ * string is in the desired normalized format, UNORM_NO determines that
+ * argument string is not in the desired normalized format. A
+ * UNORM_MAYBE result indicates that a more thorough check is required,
+ * the user may have to put the string in its normalized form and compare the
+ * results.
+ *
+ * @param source       string for determining if it is in a normalized format
+ * @param sourcelength length of source to test, or -1 if NUL-terminated
+ * @param mode         which normalization form to test for
+ * @param status       a pointer to a UErrorCode to receive any errors
+ * @return UNORM_YES, UNORM_NO or UNORM_MAYBE
+ *
+ * @see unorm_isNormalized
+ * @deprecated ICU 56 Use unorm2.h instead.
+ */
+U_DEPRECATED UNormalizationCheckResult U_EXPORT2
+unorm_quickCheck(const UChar *source, int32_t sourcelength,
+                 UNormalizationMode mode,
+                 UErrorCode *status);
+
+/**
+ * Performing quick check on a string; same as unorm_quickCheck but
+ * takes an extra options parameter like most normalization functions.
+ *
+ * @param src        String that is to be tested if it is in a normalization format.
+ * @param srcLength  Length of source to test, or -1 if NUL-terminated.
+ * @param mode       Which normalization form to test for.
+ * @param options    The normalization options, ORed together (0 for no options).
+ * @param pErrorCode ICU error code in/out parameter.
+ *                   Must fulfill U_SUCCESS before the function call.
+ * @return UNORM_YES, UNORM_NO or UNORM_MAYBE
+ *
+ * @see unorm_quickCheck
+ * @see unorm_isNormalized
+ * @deprecated ICU 56 Use unorm2.h instead.
+ */
+U_DEPRECATED UNormalizationCheckResult U_EXPORT2
+unorm_quickCheckWithOptions(const UChar *src, int32_t srcLength,
+                            UNormalizationMode mode, int32_t options,
+                            UErrorCode *pErrorCode);
+
+/**
+ * Test if a string is in a given normalization form.
+ * This is semantically equivalent to source.equals(normalize(source, mode)) .
+ *
+ * Unlike unorm_quickCheck(), this function returns a definitive result,
+ * never a "maybe".
+ * For NFD, NFKD, and FCD, both functions work exactly the same.
+ * For NFC and NFKC where quickCheck may return "maybe", this function will
+ * perform further tests to arrive at a true/false result.
+ *
+ * @param src        String that is to be tested if it is in a normalization format.
+ * @param srcLength  Length of source to test, or -1 if NUL-terminated.
+ * @param mode       Which normalization form to test for.
+ * @param pErrorCode ICU error code in/out parameter.
+ *                   Must fulfill U_SUCCESS before the function call.
+ * @return Boolean value indicating whether the source string is in the
+ *         "mode" normalization form.
+ *
+ * @see unorm_quickCheck
+ * @deprecated ICU 56 Use unorm2.h instead.
+ */
+U_DEPRECATED UBool U_EXPORT2
+unorm_isNormalized(const UChar *src, int32_t srcLength,
+                   UNormalizationMode mode,
+                   UErrorCode *pErrorCode);
+
+/**
+ * Test if a string is in a given normalization form; same as unorm_isNormalized but
+ * takes an extra options parameter like most normalization functions.
+ *
+ * @param src        String that is to be tested if it is in a normalization format.
+ * @param srcLength  Length of source to test, or -1 if NUL-terminated.
+ * @param mode       Which normalization form to test for.
+ * @param options    The normalization options, ORed together (0 for no options).
+ * @param pErrorCode ICU error code in/out parameter.
+ *                   Must fulfill U_SUCCESS before the function call.
+ * @return Boolean value indicating whether the source string is in the
+ *         "mode/options" normalization form.
+ *
+ * @see unorm_quickCheck
+ * @see unorm_isNormalized
+ * @deprecated ICU 56 Use unorm2.h instead.
+ */
+U_DEPRECATED UBool U_EXPORT2
+unorm_isNormalizedWithOptions(const UChar *src, int32_t srcLength,
+                              UNormalizationMode mode, int32_t options,
+                              UErrorCode *pErrorCode);
+
+/**
+ * Iterative normalization forward.
+ * This function (together with unorm_previous) is somewhat
+ * similar to the C++ Normalizer class (see its non-static functions).
+ *
+ * Iterative normalization is useful when only a small portion of a longer
+ * string/text needs to be processed.
+ *
+ * For example, the likelihood may be high that processing the first 10% of some
+ * text will be sufficient to find certain data.
+ * Another example: When one wants to concatenate two normalized strings and get a
+ * normalized result, it is much more efficient to normalize just a small part of
+ * the result around the concatenation place instead of re-normalizing everything.
+ *
+ * The input text is an instance of the C character iteration API UCharIterator.
+ * It may wrap around a simple string, a CharacterIterator, a Replaceable, or any
+ * other kind of text object.
+ *
+ * If a buffer overflow occurs, then the caller needs to reset the iterator to the
+ * old index and call the function again with a larger buffer - if the caller cares
+ * for the actual output.
+ * Regardless of the output buffer, the iterator will always be moved to the next
+ * normalization boundary.
+ *
+ * This function (like unorm_previous) serves two purposes:
+ *
+ * 1) To find the next boundary so that the normalization of the part of the text
+ * from the current position to that boundary does not affect and is not affected
+ * by the part of the text beyond that boundary.
+ *
+ * 2) To normalize the text up to the boundary.
+ *
+ * The second step is optional, per the doNormalize parameter.
+ * It is omitted for operations like string concatenation, where the two adjacent
+ * string ends need to be normalized together.
+ * In such a case, the output buffer will just contain a copy of the text up to the
+ * boundary.
+ *
+ * pNeededToNormalize is an output-only parameter. Its output value is only defined
+ * if normalization was requested (doNormalize) and successful (especially, no
+ * buffer overflow).
+ * It is useful for operations like a normalizing transliterator, where one would
+ * not want to replace a piece of text if it is not modified.
+ *
+ * If doNormalize==true and pNeededToNormalize!=NULL then *pNeeded... is set true
+ * if the normalization was necessary.
+ *
+ * If doNormalize==false then *pNeededToNormalize will be set to false.
+ *
+ * If the buffer overflows, then *pNeededToNormalize will be undefined;
+ * essentially, whenever U_FAILURE is true (like in buffer overflows), this result
+ * will be undefined.
+ *
+ * @param src The input text in the form of a C character iterator.
+ * @param dest The output buffer; can be NULL if destCapacity==0 for pure preflighting.
+ * @param destCapacity The number of UChars that fit into dest.
+ * @param mode The normalization mode.
+ * @param options The normalization options, ORed together (0 for no options).
+ * @param doNormalize Indicates if the source text up to the next boundary
+ *                    is to be normalized (true) or just copied (false).
+ * @param pNeededToNormalize Output flag indicating if the normalization resulted in
+ *                           different text from the input.
+ *                           Not defined if an error occurs including buffer overflow.
+ *                           Always false if !doNormalize.
+ * @param pErrorCode ICU error code in/out parameter.
+ *                   Must fulfill U_SUCCESS before the function call.
+ * @return Length of output (number of UChars) when successful or buffer overflow.
+ *
+ * @see unorm_previous
+ * @see unorm_normalize
+ *
+ * @deprecated ICU 56 Use unorm2.h instead.
+ */
+U_DEPRECATED int32_t U_EXPORT2
+unorm_next(UCharIterator *src,
+           UChar *dest, int32_t destCapacity,
+           UNormalizationMode mode, int32_t options,
+           UBool doNormalize, UBool *pNeededToNormalize,
+           UErrorCode *pErrorCode);
+
+/**
+ * Iterative normalization backward.
+ * This function (together with unorm_next) is somewhat
+ * similar to the C++ Normalizer class (see its non-static functions).
+ * For all details see unorm_next.
+ *
+ * @param src The input text in the form of a C character iterator.
+ * @param dest The output buffer; can be NULL if destCapacity==0 for pure preflighting.
+ * @param destCapacity The number of UChars that fit into dest.
+ * @param mode The normalization mode.
+ * @param options The normalization options, ORed together (0 for no options).
+ * @param doNormalize Indicates if the source text up to the next boundary
+ *                    is to be normalized (true) or just copied (false).
+ * @param pNeededToNormalize Output flag indicating if the normalization resulted in
+ *                           different text from the input.
+ *                           Not defined if an error occurs including buffer overflow.
+ *                           Always false if !doNormalize.
+ * @param pErrorCode ICU error code in/out parameter.
+ *                   Must fulfill U_SUCCESS before the function call.
+ * @return Length of output (number of UChars) when successful or buffer overflow.
+ *
+ * @see unorm_next
+ * @see unorm_normalize
+ *
+ * @deprecated ICU 56 Use unorm2.h instead.
+ */
+U_DEPRECATED int32_t U_EXPORT2
+unorm_previous(UCharIterator *src,
+               UChar *dest, int32_t destCapacity,
+               UNormalizationMode mode, int32_t options,
+               UBool doNormalize, UBool *pNeededToNormalize,
+               UErrorCode *pErrorCode);
+
+/**
+ * Concatenate normalized strings, making sure that the result is normalized as well.
+ *
+ * If both the left and the right strings are in
+ * the normalization form according to "mode/options",
+ * then the result will be
+ *
+ * \code
+ *     dest=normalize(left+right, mode, options)
+ * \endcode
+ *
+ * With the input strings already being normalized,
+ * this function will use unorm_next() and unorm_previous()
+ * to find the adjacent end pieces of the input strings.
+ * Only the concatenation of these end pieces will be normalized and
+ * then concatenated with the remaining parts of the input strings.
+ *
+ * It is allowed to have dest==left to avoid copying the entire left string.
+ *
+ * @param left Left source string, may be same as dest.
+ * @param leftLength Length of left source string, or -1 if NUL-terminated.
+ * @param right Right source string. Must not be the same as dest, nor overlap.
+ * @param rightLength Length of right source string, or -1 if NUL-terminated.
+ * @param dest The output buffer; can be NULL if destCapacity==0 for pure preflighting.
+ * @param destCapacity The number of UChars that fit into dest.
+ * @param mode The normalization mode.
+ * @param options The normalization options, ORed together (0 for no options).
+ * @param pErrorCode ICU error code in/out parameter.
+ *                   Must fulfill U_SUCCESS before the function call.
+ * @return Length of output (number of UChars) when successful or buffer overflow.
+ *
+ * @see unorm_normalize
+ * @see unorm_next
+ * @see unorm_previous
+ *
+ * @deprecated ICU 56 Use unorm2.h instead.
+ */
+U_DEPRECATED int32_t U_EXPORT2
+unorm_concatenate(const UChar *left, int32_t leftLength,
+                  const UChar *right, int32_t rightLength,
+                  UChar *dest, int32_t destCapacity,
+                  UNormalizationMode mode, int32_t options,
+                  UErrorCode *pErrorCode);
+
+#endif  /* U_HIDE_DEPRECATED_API */
+#endif /* #if !UCONFIG_NO_NORMALIZATION */
+#endif

thirdparty/icu4c/common/unicode/unorm2.h

@@ -0,0 +1,626 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+/*
+*******************************************************************************
+*
+*   Copyright (C) 2009-2015, International Business Machines
+*   Corporation and others.  All Rights Reserved.
+*
+*******************************************************************************
+*   file name:  unorm2.h
+*   encoding:   UTF-8
+*   tab size:   8 (not used)
+*   indentation:4
+*
+*   created on: 2009dec15
+*   created by: Markus W. Scherer
+*/
+
+#ifndef __UNORM2_H__
+#define __UNORM2_H__
+
+/**
+ * \file
+ * \brief C API: New API for Unicode Normalization.
+ *
+ * Unicode normalization functionality for standard Unicode normalization or
+ * for using custom mapping tables.
+ * All instances of UNormalizer2 are unmodifiable/immutable.
+ * Instances returned by unorm2_getInstance() are singletons that must not be deleted by the caller.
+ * For more details see the Normalizer2 C++ class.
+ */
+
+#include "unicode/utypes.h"
+#include "unicode/stringoptions.h"
+#include "unicode/uset.h"
+
+#if U_SHOW_CPLUSPLUS_API
+#include "unicode/localpointer.h"
+#endif   // U_SHOW_CPLUSPLUS_API
+
+/**
+ * Constants for normalization modes.
+ * For details about standard Unicode normalization forms
+ * and about the algorithms which are also used with custom mapping tables
+ * see http://www.unicode.org/unicode/reports/tr15/
+ * @stable ICU 4.4
+ */
+typedef enum {
+    /**
+     * Decomposition followed by composition.
+     * Same as standard NFC when using an "nfc" instance.
+     * Same as standard NFKC when using an "nfkc" instance.
+     * For details about standard Unicode normalization forms
+     * see http://www.unicode.org/unicode/reports/tr15/
+     * @stable ICU 4.4
+     */
+    UNORM2_COMPOSE,
+    /**
+     * Map, and reorder canonically.
+     * Same as standard NFD when using an "nfc" instance.
+     * Same as standard NFKD when using an "nfkc" instance.
+     * For details about standard Unicode normalization forms
+     * see http://www.unicode.org/unicode/reports/tr15/
+     * @stable ICU 4.4
+     */
+    UNORM2_DECOMPOSE,
+    /**
+     * "Fast C or D" form.
+     * If a string is in this form, then further decomposition <i>without reordering</i>
+     * would yield the same form as DECOMPOSE.
+     * Text in "Fast C or D" form can be processed efficiently with data tables
+     * that are "canonically closed", that is, that provide equivalent data for
+     * equivalent text, without having to be fully normalized.
+     * Not a standard Unicode normalization form.
+     * Not a unique form: Different FCD strings can be canonically equivalent.
+     * For details see http://www.unicode.org/notes/tn5/#FCD
+     * @stable ICU 4.4
+     */
+    UNORM2_FCD,
+    /**
+     * Compose only contiguously.
+     * Also known as "FCC" or "Fast C Contiguous".
+     * The result will often but not always be in NFC.
+     * The result will conform to FCD which is useful for processing.
+     * Not a standard Unicode normalization form.
+     * For details see http://www.unicode.org/notes/tn5/#FCC
+     * @stable ICU 4.4
+     */
+    UNORM2_COMPOSE_CONTIGUOUS
+} UNormalization2Mode;
+
+/**
+ * Result values for normalization quick check functions.
+ * For details see http://www.unicode.org/reports/tr15/#Detecting_Normalization_Forms
+ * @stable ICU 2.0
+ */
+typedef enum UNormalizationCheckResult {
+  /**
+   * The input string is not in the normalization form.
+   * @stable ICU 2.0
+   */
+  UNORM_NO,
+  /**
+   * The input string is in the normalization form.
+   * @stable ICU 2.0
+   */
+  UNORM_YES,
+  /**
+   * The input string may or may not be in the normalization form.
+   * This value is only returned for composition forms like NFC and FCC,
+   * when a backward-combining character is found for which the surrounding text
+   * would have to be analyzed further.
+   * @stable ICU 2.0
+   */
+  UNORM_MAYBE
+} UNormalizationCheckResult;
+
+/**
+ * Opaque C service object type for the new normalization API.
+ * @stable ICU 4.4
+ */
+struct UNormalizer2;
+typedef struct UNormalizer2 UNormalizer2;  /**< C typedef for struct UNormalizer2. @stable ICU 4.4 */
+
+#if !UCONFIG_NO_NORMALIZATION
+
+/**
+ * Returns a UNormalizer2 instance for Unicode NFC normalization.
+ * Same as unorm2_getInstance(NULL, "nfc", UNORM2_COMPOSE, pErrorCode).
+ * Returns an unmodifiable singleton instance. Do not delete it.
+ * @param pErrorCode Standard ICU error code. Its input value must
+ *                  pass the U_SUCCESS() test, or else the function returns
+ *                  immediately. Check for U_FAILURE() on output or use with
+ *                  function chaining. (See User Guide for details.)
+ * @return the requested Normalizer2, if successful
+ * @stable ICU 49
+ */
+U_CAPI const UNormalizer2 * U_EXPORT2
+unorm2_getNFCInstance(UErrorCode *pErrorCode);
+
+/**
+ * Returns a UNormalizer2 instance for Unicode NFD normalization.
+ * Same as unorm2_getInstance(NULL, "nfc", UNORM2_DECOMPOSE, pErrorCode).
+ * Returns an unmodifiable singleton instance. Do not delete it.
+ * @param pErrorCode Standard ICU error code. Its input value must
+ *                  pass the U_SUCCESS() test, or else the function returns
+ *                  immediately. Check for U_FAILURE() on output or use with
+ *                  function chaining. (See User Guide for details.)
+ * @return the requested Normalizer2, if successful
+ * @stable ICU 49
+ */
+U_CAPI const UNormalizer2 * U_EXPORT2
+unorm2_getNFDInstance(UErrorCode *pErrorCode);
+
+/**
+ * Returns a UNormalizer2 instance for Unicode NFKC normalization.
+ * Same as unorm2_getInstance(NULL, "nfkc", UNORM2_COMPOSE, pErrorCode).
+ * Returns an unmodifiable singleton instance. Do not delete it.
+ * @param pErrorCode Standard ICU error code. Its input value must
+ *                  pass the U_SUCCESS() test, or else the function returns
+ *                  immediately. Check for U_FAILURE() on output or use with
+ *                  function chaining. (See User Guide for details.)
+ * @return the requested Normalizer2, if successful
+ * @stable ICU 49
+ */
+U_CAPI const UNormalizer2 * U_EXPORT2
+unorm2_getNFKCInstance(UErrorCode *pErrorCode);
+
+/**
+ * Returns a UNormalizer2 instance for Unicode NFKD normalization.
+ * Same as unorm2_getInstance(NULL, "nfkc", UNORM2_DECOMPOSE, pErrorCode).
+ * Returns an unmodifiable singleton instance. Do not delete it.
+ * @param pErrorCode Standard ICU error code. Its input value must
+ *                  pass the U_SUCCESS() test, or else the function returns
+ *                  immediately. Check for U_FAILURE() on output or use with
+ *                  function chaining. (See User Guide for details.)
+ * @return the requested Normalizer2, if successful
+ * @stable ICU 49
+ */
+U_CAPI const UNormalizer2 * U_EXPORT2
+unorm2_getNFKDInstance(UErrorCode *pErrorCode);
+
+/**
+ * Returns a UNormalizer2 instance for Unicode toNFKC_Casefold() normalization
+ * which is equivalent to applying the NFKC_Casefold mappings and then NFC.
+ * See https://www.unicode.org/reports/tr44/#NFKC_Casefold
+ *
+ * Same as unorm2_getInstance(NULL, "nfkc_cf", UNORM2_COMPOSE, pErrorCode).
+ * Returns an unmodifiable singleton instance. Do not delete it.
+ * @param pErrorCode Standard ICU error code. Its input value must
+ *                  pass the U_SUCCESS() test, or else the function returns
+ *                  immediately. Check for U_FAILURE() on output or use with
+ *                  function chaining. (See User Guide for details.)
+ * @return the requested Normalizer2, if successful
+ * @stable ICU 49
+ */
+U_CAPI const UNormalizer2 * U_EXPORT2
+unorm2_getNFKCCasefoldInstance(UErrorCode *pErrorCode);
+
+/**
+ * Returns a UNormalizer2 instance for a variant of Unicode toNFKC_Casefold() normalization
+ * which is equivalent to applying the NFKC_Simple_Casefold mappings and then NFC.
+ * See https://www.unicode.org/reports/tr44/#NFKC_Simple_Casefold
+ *
+ * Same as unorm2_getInstance(NULL, "nfkc_scf", UNORM2_COMPOSE, pErrorCode).
+ * Returns an unmodifiable singleton instance. Do not delete it.
+ * @param pErrorCode Standard ICU error code. Its input value must
+ *                  pass the U_SUCCESS() test, or else the function returns
+ *                  immediately. Check for U_FAILURE() on output or use with
+ *                  function chaining. (See User Guide for details.)
+ * @return the requested Normalizer2, if successful
+ * @stable ICU 74
+ */
+U_CAPI const UNormalizer2 * U_EXPORT2
+unorm2_getNFKCSimpleCasefoldInstance(UErrorCode *pErrorCode);
+
+/**
+ * Returns a UNormalizer2 instance which uses the specified data file
+ * (packageName/name similar to ucnv_openPackage() and ures_open()/ResourceBundle)
+ * and which composes or decomposes text according to the specified mode.
+ * Returns an unmodifiable singleton instance. Do not delete it.
+ *
+ * Use packageName=NULL for data files that are part of ICU's own data.
+ * Use name="nfc" and UNORM2_COMPOSE/UNORM2_DECOMPOSE for Unicode standard NFC/NFD.
+ * Use name="nfkc" and UNORM2_COMPOSE/UNORM2_DECOMPOSE for Unicode standard NFKC/NFKD.
+ * Use name="nfkc_cf" and UNORM2_COMPOSE for Unicode standard NFKC_CF=NFKC_Casefold.
+ *
+ * @param packageName NULL for ICU built-in data, otherwise application data package name
+ * @param name "nfc" or "nfkc" or "nfkc_cf" or "nfkc_scf" or name of custom data file
+ * @param mode normalization mode (compose or decompose etc.)
+ * @param pErrorCode Standard ICU error code. Its input value must
+ *                  pass the U_SUCCESS() test, or else the function returns
+ *                  immediately. Check for U_FAILURE() on output or use with
+ *                  function chaining. (See User Guide for details.)
+ * @return the requested UNormalizer2, if successful
+ * @stable ICU 4.4
+ */
+U_CAPI const UNormalizer2 * U_EXPORT2
+unorm2_getInstance(const char *packageName,
+                   const char *name,
+                   UNormalization2Mode mode,
+                   UErrorCode *pErrorCode);
+
+/**
+ * Constructs a filtered normalizer wrapping any UNormalizer2 instance
+ * and a filter set.
+ * Both are aliased and must not be modified or deleted while this object
+ * is used.
+ * The filter set should be frozen; otherwise the performance will suffer greatly.
+ * @param norm2 wrapped UNormalizer2 instance
+ * @param filterSet USet which determines the characters to be normalized
+ * @param pErrorCode Standard ICU error code. Its input value must
+ *                   pass the U_SUCCESS() test, or else the function returns
+ *                   immediately. Check for U_FAILURE() on output or use with
+ *                   function chaining. (See User Guide for details.)
+ * @return the requested UNormalizer2, if successful
+ * @stable ICU 4.4
+ */
+U_CAPI UNormalizer2 * U_EXPORT2
+unorm2_openFiltered(const UNormalizer2 *norm2, const USet *filterSet, UErrorCode *pErrorCode);
+
+/**
+ * Closes a UNormalizer2 instance from unorm2_openFiltered().
+ * Do not close instances from unorm2_getInstance()!
+ * @param norm2 UNormalizer2 instance to be closed
+ * @stable ICU 4.4
+ */
+U_CAPI void U_EXPORT2
+unorm2_close(UNormalizer2 *norm2);
+
+#if U_SHOW_CPLUSPLUS_API
+
+U_NAMESPACE_BEGIN
+
+/**
+ * \class LocalUNormalizer2Pointer
+ * "Smart pointer" class, closes a UNormalizer2 via unorm2_close().
+ * For most methods see the LocalPointerBase base class.
+ *
+ * @see LocalPointerBase
+ * @see LocalPointer
+ * @stable ICU 4.4
+ */
+U_DEFINE_LOCAL_OPEN_POINTER(LocalUNormalizer2Pointer, UNormalizer2, unorm2_close);
+
+U_NAMESPACE_END
+
+#endif
+
+/**
+ * Writes the normalized form of the source string to the destination string
+ * (replacing its contents) and returns the length of the destination string.
+ * The source and destination strings must be different buffers.
+ * @param norm2 UNormalizer2 instance
+ * @param src source string
+ * @param length length of the source string, or -1 if NUL-terminated
+ * @param dest destination string; its contents is replaced with normalized src
+ * @param capacity number of UChars that can be written to dest
+ * @param pErrorCode Standard ICU error code. Its input value must
+ *                   pass the U_SUCCESS() test, or else the function returns
+ *                   immediately. Check for U_FAILURE() on output or use with
+ *                   function chaining. (See User Guide for details.)
+ * @return dest
+ * @stable ICU 4.4
+ */
+U_CAPI int32_t U_EXPORT2
+unorm2_normalize(const UNormalizer2 *norm2,
+                 const UChar *src, int32_t length,
+                 UChar *dest, int32_t capacity,
+                 UErrorCode *pErrorCode);
+/**
+ * Appends the normalized form of the second string to the first string
+ * (merging them at the boundary) and returns the length of the first string.
+ * The result is normalized if the first string was normalized.
+ * The first and second strings must be different buffers.
+ * @param norm2 UNormalizer2 instance
+ * @param first string, should be normalized
+ * @param firstLength length of the first string, or -1 if NUL-terminated
+ * @param firstCapacity number of UChars that can be written to first
+ * @param second string, will be normalized
+ * @param secondLength length of the source string, or -1 if NUL-terminated
+ * @param pErrorCode Standard ICU error code. Its input value must
+ *                   pass the U_SUCCESS() test, or else the function returns
+ *                   immediately. Check for U_FAILURE() on output or use with
+ *                   function chaining. (See User Guide for details.)
+ * @return first
+ * @stable ICU 4.4
+ */
+U_CAPI int32_t U_EXPORT2
+unorm2_normalizeSecondAndAppend(const UNormalizer2 *norm2,
+                                UChar *first, int32_t firstLength, int32_t firstCapacity,
+                                const UChar *second, int32_t secondLength,
+                                UErrorCode *pErrorCode);
+/**
+ * Appends the second string to the first string
+ * (merging them at the boundary) and returns the length of the first string.
+ * The result is normalized if both the strings were normalized.
+ * The first and second strings must be different buffers.
+ * @param norm2 UNormalizer2 instance
+ * @param first string, should be normalized
+ * @param firstLength length of the first string, or -1 if NUL-terminated
+ * @param firstCapacity number of UChars that can be written to first
+ * @param second string, should be normalized
+ * @param secondLength length of the source string, or -1 if NUL-terminated
+ * @param pErrorCode Standard ICU error code. Its input value must
+ *                   pass the U_SUCCESS() test, or else the function returns
+ *                   immediately. Check for U_FAILURE() on output or use with
+ *                   function chaining. (See User Guide for details.)
+ * @return first
+ * @stable ICU 4.4
+ */
+U_CAPI int32_t U_EXPORT2
+unorm2_append(const UNormalizer2 *norm2,
+              UChar *first, int32_t firstLength, int32_t firstCapacity,
+              const UChar *second, int32_t secondLength,
+              UErrorCode *pErrorCode);
+
+/**
+ * Gets the decomposition mapping of c.
+ * Roughly equivalent to normalizing the String form of c
+ * on a UNORM2_DECOMPOSE UNormalizer2 instance, but much faster, and except that this function
+ * returns a negative value and does not write a string
+ * if c does not have a decomposition mapping in this instance's data.
+ * This function is independent of the mode of the UNormalizer2.
+ * @param norm2 UNormalizer2 instance
+ * @param c code point
+ * @param decomposition String buffer which will be set to c's
+ *                      decomposition mapping, if there is one.
+ * @param capacity number of UChars that can be written to decomposition
+ * @param pErrorCode Standard ICU error code. Its input value must
+ *                   pass the U_SUCCESS() test, or else the function returns
+ *                   immediately. Check for U_FAILURE() on output or use with
+ *                   function chaining. (See User Guide for details.)
+ * @return the non-negative length of c's decomposition, if there is one; otherwise a negative value
+ * @stable ICU 4.6
+ */
+U_CAPI int32_t U_EXPORT2
+unorm2_getDecomposition(const UNormalizer2 *norm2,
+                        UChar32 c, UChar *decomposition, int32_t capacity,
+                        UErrorCode *pErrorCode);
+
+/**
+ * Gets the raw decomposition mapping of c.
+ *
+ * This is similar to the unorm2_getDecomposition() function but returns the
+ * raw decomposition mapping as specified in UnicodeData.txt or
+ * (for custom data) in the mapping files processed by the gennorm2 tool.
+ * By contrast, unorm2_getDecomposition() returns the processed,
+ * recursively-decomposed version of this mapping.
+ *
+ * When used on a standard NFKC Normalizer2 instance,
+ * unorm2_getRawDecomposition() returns the Unicode Decomposition_Mapping (dm) property.
+ *
+ * When used on a standard NFC Normalizer2 instance,
+ * it returns the Decomposition_Mapping only if the Decomposition_Type (dt) is Canonical (Can);
+ * in this case, the result contains either one or two code points (=1..4 UChars).
+ *
+ * This function is independent of the mode of the UNormalizer2.
+ * @param norm2 UNormalizer2 instance
+ * @param c code point
+ * @param decomposition String buffer which will be set to c's
+ *                      raw decomposition mapping, if there is one.
+ * @param capacity number of UChars that can be written to decomposition
+ * @param pErrorCode Standard ICU error code. Its input value must
+ *                   pass the U_SUCCESS() test, or else the function returns
+ *                   immediately. Check for U_FAILURE() on output or use with
+ *                   function chaining. (See User Guide for details.)
+ * @return the non-negative length of c's raw decomposition, if there is one; otherwise a negative value
+ * @stable ICU 49
+ */
+U_CAPI int32_t U_EXPORT2
+unorm2_getRawDecomposition(const UNormalizer2 *norm2,
+                           UChar32 c, UChar *decomposition, int32_t capacity,
+                           UErrorCode *pErrorCode);
+
+/**
+ * Performs pairwise composition of a & b and returns the composite if there is one.
+ *
+ * Returns a composite code point c only if c has a two-way mapping to a+b.
+ * In standard Unicode normalization, this means that
+ * c has a canonical decomposition to a+b
+ * and c does not have the Full_Composition_Exclusion property.
+ *
+ * This function is independent of the mode of the UNormalizer2.
+ * @param norm2 UNormalizer2 instance
+ * @param a A (normalization starter) code point.
+ * @param b Another code point.
+ * @return The non-negative composite code point if there is one; otherwise a negative value.
+ * @stable ICU 49
+ */
+U_CAPI UChar32 U_EXPORT2
+unorm2_composePair(const UNormalizer2 *norm2, UChar32 a, UChar32 b);
+
+/**
+ * Gets the combining class of c.
+ * The default implementation returns 0
+ * but all standard implementations return the Unicode Canonical_Combining_Class value.
+ * @param norm2 UNormalizer2 instance
+ * @param c code point
+ * @return c's combining class
+ * @stable ICU 49
+ */
+U_CAPI uint8_t U_EXPORT2
+unorm2_getCombiningClass(const UNormalizer2 *norm2, UChar32 c);
+
+/**
+ * Tests if the string is normalized.
+ * Internally, in cases where the quickCheck() method would return "maybe"
+ * (which is only possible for the two COMPOSE modes) this method
+ * resolves to "yes" or "no" to provide a definitive result,
+ * at the cost of doing more work in those cases.
+ * @param norm2 UNormalizer2 instance
+ * @param s input string
+ * @param length length of the string, or -1 if NUL-terminated
+ * @param pErrorCode Standard ICU error code. Its input value must
+ *                   pass the U_SUCCESS() test, or else the function returns
+ *                   immediately. Check for U_FAILURE() on output or use with
+ *                   function chaining. (See User Guide for details.)
+ * @return true if s is normalized
+ * @stable ICU 4.4
+ */
+U_CAPI UBool U_EXPORT2
+unorm2_isNormalized(const UNormalizer2 *norm2,
+                    const UChar *s, int32_t length,
+                    UErrorCode *pErrorCode);
+
+/**
+ * Tests if the string is normalized.
+ * For the two COMPOSE modes, the result could be "maybe" in cases that
+ * would take a little more work to resolve definitively.
+ * Use spanQuickCheckYes() and normalizeSecondAndAppend() for a faster
+ * combination of quick check + normalization, to avoid
+ * re-checking the "yes" prefix.
+ * @param norm2 UNormalizer2 instance
+ * @param s input string
+ * @param length length of the string, or -1 if NUL-terminated
+ * @param pErrorCode Standard ICU error code. Its input value must
+ *                   pass the U_SUCCESS() test, or else the function returns
+ *                   immediately. Check for U_FAILURE() on output or use with
+ *                   function chaining. (See User Guide for details.)
+ * @return UNormalizationCheckResult
+ * @stable ICU 4.4
+ */
+U_CAPI UNormalizationCheckResult U_EXPORT2
+unorm2_quickCheck(const UNormalizer2 *norm2,
+                  const UChar *s, int32_t length,
+                  UErrorCode *pErrorCode);
+
+/**
+ * Returns the end of the normalized substring of the input string.
+ * In other words, with <code>end=spanQuickCheckYes(s, ec);</code>
+ * the substring <code>UnicodeString(s, 0, end)</code>
+ * will pass the quick check with a "yes" result.
+ *
+ * The returned end index is usually one or more characters before the
+ * "no" or "maybe" character: The end index is at a normalization boundary.
+ * (See the class documentation for more about normalization boundaries.)
+ *
+ * When the goal is a normalized string and most input strings are expected
+ * to be normalized already, then call this method,
+ * and if it returns a prefix shorter than the input string,
+ * copy that prefix and use normalizeSecondAndAppend() for the remainder.
+ * @param norm2 UNormalizer2 instance
+ * @param s input string
+ * @param length length of the string, or -1 if NUL-terminated
+ * @param pErrorCode Standard ICU error code. Its input value must
+ *                   pass the U_SUCCESS() test, or else the function returns
+ *                   immediately. Check for U_FAILURE() on output or use with
+ *                   function chaining. (See User Guide for details.)
+ * @return "yes" span end index
+ * @stable ICU 4.4
+ */
+U_CAPI int32_t U_EXPORT2
+unorm2_spanQuickCheckYes(const UNormalizer2 *norm2,
+                         const UChar *s, int32_t length,
+                         UErrorCode *pErrorCode);
+
+/**
+ * Tests if the character always has a normalization boundary before it,
+ * regardless of context.
+ * For details see the Normalizer2 base class documentation.
+ * @param norm2 UNormalizer2 instance
+ * @param c character to test
+ * @return true if c has a normalization boundary before it
+ * @stable ICU 4.4
+ */
+U_CAPI UBool U_EXPORT2
+unorm2_hasBoundaryBefore(const UNormalizer2 *norm2, UChar32 c);
+
+/**
+ * Tests if the character always has a normalization boundary after it,
+ * regardless of context.
+ * For details see the Normalizer2 base class documentation.
+ * @param norm2 UNormalizer2 instance
+ * @param c character to test
+ * @return true if c has a normalization boundary after it
+ * @stable ICU 4.4
+ */
+U_CAPI UBool U_EXPORT2
+unorm2_hasBoundaryAfter(const UNormalizer2 *norm2, UChar32 c);
+
+/**
+ * Tests if the character is normalization-inert.
+ * For details see the Normalizer2 base class documentation.
+ * @param norm2 UNormalizer2 instance
+ * @param c character to test
+ * @return true if c is normalization-inert
+ * @stable ICU 4.4
+ */
+U_CAPI UBool U_EXPORT2
+unorm2_isInert(const UNormalizer2 *norm2, UChar32 c);
+
+/**
+ * Compares two strings for canonical equivalence.
+ * Further options include case-insensitive comparison and
+ * code point order (as opposed to code unit order).
+ *
+ * Canonical equivalence between two strings is defined as their normalized
+ * forms (NFD or NFC) being identical.
+ * This function compares strings incrementally instead of normalizing
+ * (and optionally case-folding) both strings entirely,
+ * improving performance significantly.
+ *
+ * Bulk normalization is only necessary if the strings do not fulfill the FCD
+ * conditions. Only in this case, and only if the strings are relatively long,
+ * is memory allocated temporarily.
+ * For FCD strings and short non-FCD strings there is no memory allocation.
+ *
+ * Semantically, this is equivalent to
+ *   strcmp[CodePointOrder](NFD(foldCase(NFD(s1))), NFD(foldCase(NFD(s2))))
+ * where code point order and foldCase are all optional.
+ *
+ * UAX 21 2.5 Caseless Matching specifies that for a canonical caseless match
+ * the case folding must be performed first, then the normalization.
+ *
+ * @param s1 First source string.
+ * @param length1 Length of first source string, or -1 if NUL-terminated.
+ *
+ * @param s2 Second source string.
+ * @param length2 Length of second source string, or -1 if NUL-terminated.
+ *
+ * @param options A bit set of options:
+ *   - U_FOLD_CASE_DEFAULT or 0 is used for default options:
+ *     Case-sensitive comparison in code unit order, and the input strings
+ *     are quick-checked for FCD.
+ *
+ *   - UNORM_INPUT_IS_FCD
+ *     Set if the caller knows that both s1 and s2 fulfill the FCD conditions.
+ *     If not set, the function will quickCheck for FCD
+ *     and normalize if necessary.
+ *
+ *   - U_COMPARE_CODE_POINT_ORDER
+ *     Set to choose code point order instead of code unit order
+ *     (see u_strCompare for details).
+ *
+ *   - U_COMPARE_IGNORE_CASE
+ *     Set to compare strings case-insensitively using case folding,
+ *     instead of case-sensitively.
+ *     If set, then the following case folding options are used.
+ *
+ *   - Options as used with case-insensitive comparisons, currently:
+ *
+ *   - U_FOLD_CASE_EXCLUDE_SPECIAL_I
+ *    (see u_strCaseCompare for details)
+ *
+ *   - regular normalization options shifted left by UNORM_COMPARE_NORM_OPTIONS_SHIFT
+ *
+ * @param pErrorCode ICU error code in/out parameter.
+ *                   Must fulfill U_SUCCESS before the function call.
+ * @return <0 or 0 or >0 as usual for string comparisons
+ *
+ * @see unorm_normalize
+ * @see UNORM_FCD
+ * @see u_strCompare
+ * @see u_strCaseCompare
+ *
+ * @stable ICU 2.2
+ */
+U_CAPI int32_t U_EXPORT2
+unorm_compare(const UChar *s1, int32_t length1,
+              const UChar *s2, int32_t length2,
+              uint32_t options,
+              UErrorCode *pErrorCode);
+
+#endif  /* !UCONFIG_NO_NORMALIZATION */
+#endif  /* __UNORM2_H__ */

thirdparty/icu4c/common/unicode/uobject.h

@@ -0,0 +1,324 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+/*
+******************************************************************************
+*
+*   Copyright (C) 2002-2012, International Business Machines
+*   Corporation and others.  All Rights Reserved.
+*
+******************************************************************************
+*   file name:  uobject.h
+*   encoding:   UTF-8
+*   tab size:   8 (not used)
+*   indentation:4
+*
+*   created on: 2002jun26
+*   created by: Markus W. Scherer
+*/
+
+#ifndef __UOBJECT_H__
+#define __UOBJECT_H__
+
+#include "unicode/utypes.h"
+
+#if U_SHOW_CPLUSPLUS_API
+
+#include "unicode/platform.h"
+
+/**
+ * \file
+ * \brief C++ API: Common ICU base class UObject.
+ */
+
+/**
+ * \def U_NO_THROW
+ *         Since ICU 64, use noexcept instead.
+ *
+ *         Previously, define this to define the throw() specification so
+ *                 certain functions do not throw any exceptions
+ *
+ *         UMemory operator new methods should have the throw() specification 
+ *         appended to them, so that the compiler adds the additional nullptr check 
+ *         before calling constructors. Without, if <code>operator new</code> returns nullptr the
+ *         constructor is still called, and if the constructor references member 
+ *         data, (which it typically does), the result is a segmentation violation.
+ *
+ * @stable ICU 4.2. Since ICU 64, Use noexcept instead. See ICU-20422.
+ */
+#ifndef U_NO_THROW
+#define U_NO_THROW noexcept
+#endif
+
+/*===========================================================================*/
+/* UClassID-based RTTI */
+/*===========================================================================*/
+
+/**
+ * UClassID is used to identify classes without using the compiler's RTTI.
+ * This was used before C++ compilers consistently supported RTTI.
+ * ICU 4.6 requires compiler RTTI to be turned on.
+ *
+ * Each class hierarchy which needs
+ * to implement polymorphic clone() or operator==() defines two methods,
+ * described in detail below.  UClassID values can be compared using
+ * operator==(). Nothing else should be done with them.
+ *
+ * \par
+ * In class hierarchies that implement "poor man's RTTI",
+ * each concrete subclass implements getDynamicClassID() in the same way:
+ *
+ * \code
+ *      class Derived {
+ *      public:
+ *          virtual UClassID getDynamicClassID() const
+ *            { return Derived::getStaticClassID(); }
+ *      }
+ * \endcode
+ *
+ * Each concrete class implements getStaticClassID() as well, which allows
+ * clients to test for a specific type.
+ *
+ * \code
+ *      class Derived {
+ *      public:
+ *          static UClassID U_EXPORT2 getStaticClassID();
+ *      private:
+ *          static char fgClassID;
+ *      }
+ *
+ *      // In Derived.cpp:
+ *      UClassID Derived::getStaticClassID()
+ *        { return (UClassID)&Derived::fgClassID; }
+ *      char Derived::fgClassID = 0; // Value is irrelevant
+ * \endcode
+ * @stable ICU 2.0
+ */
+typedef void* UClassID;
+
+U_NAMESPACE_BEGIN
+
+/**
+ * UMemory is the common ICU base class.
+ * All other ICU C++ classes are derived from UMemory (starting with ICU 2.4).
+ *
+ * This is primarily to make it possible and simple to override the
+ * C++ memory management by adding new/delete operators to this base class.
+ *
+ * To override ALL ICU memory management, including that from plain C code,
+ * replace the allocation functions declared in cmemory.h
+ *
+ * UMemory does not contain any virtual functions.
+ * Common "boilerplate" functions are defined in UObject.
+ *
+ * @stable ICU 2.4
+ */
+class U_COMMON_API UMemory {
+public:
+
+/* test versions for debugging shaper heap memory problems */
+#ifdef SHAPER_MEMORY_DEBUG  
+    static void * NewArray(int size, int count);
+    static void * GrowArray(void * array, int newSize );
+    static void   FreeArray(void * array );
+#endif
+
+#if U_OVERRIDE_CXX_ALLOCATION
+    /**
+     * Override for ICU4C C++ memory management.
+     * simple, non-class types are allocated using the macros in common/cmemory.h
+     * (uprv_malloc(), uprv_free(), uprv_realloc());
+     * they or something else could be used here to implement C++ new/delete
+     * for ICU4C C++ classes
+     * @stable ICU 2.4
+     */
+    static void * U_EXPORT2 operator new(size_t size) noexcept;
+
+    /**
+     * Override for ICU4C C++ memory management.
+     * See new().
+     * @stable ICU 2.4
+     */
+    static void * U_EXPORT2 operator new[](size_t size) noexcept;
+
+    /**
+     * Override for ICU4C C++ memory management.
+     * simple, non-class types are allocated using the macros in common/cmemory.h
+     * (uprv_malloc(), uprv_free(), uprv_realloc());
+     * they or something else could be used here to implement C++ new/delete
+     * for ICU4C C++ classes
+     * @stable ICU 2.4
+     */
+    static void U_EXPORT2 operator delete(void *p) noexcept;
+
+    /**
+     * Override for ICU4C C++ memory management.
+     * See delete().
+     * @stable ICU 2.4
+     */
+    static void U_EXPORT2 operator delete[](void *p) noexcept;
+
+#if U_HAVE_PLACEMENT_NEW
+    /**
+     * Override for ICU4C C++ memory management for STL.
+     * See new().
+     * @stable ICU 2.6
+     */
+    static inline void * U_EXPORT2 operator new(size_t, void *ptr) noexcept { return ptr; }
+
+    /**
+     * Override for ICU4C C++ memory management for STL.
+     * See delete().
+     * @stable ICU 2.6
+     */
+    static inline void U_EXPORT2 operator delete(void *, void *) noexcept {}
+#endif /* U_HAVE_PLACEMENT_NEW */
+#if U_HAVE_DEBUG_LOCATION_NEW
+    /**
+      * This method overrides the MFC debug version of the operator new
+      * 
+      * @param size   The requested memory size
+      * @param file   The file where the allocation was requested
+      * @param line   The line where the allocation was requested 
+      */ 
+    static void * U_EXPORT2 operator new(size_t size, const char* file, int line) noexcept;
+    /**
+      * This method provides a matching delete for the MFC debug new
+      * 
+      * @param p      The pointer to the allocated memory
+      * @param file   The file where the allocation was requested
+      * @param line   The line where the allocation was requested 
+      */ 
+    static void U_EXPORT2 operator delete(void* p, const char* file, int line) noexcept;
+#endif /* U_HAVE_DEBUG_LOCATION_NEW */
+#endif /* U_OVERRIDE_CXX_ALLOCATION */
+
+    /*
+     * Assignment operator not declared. The compiler will provide one
+     * which does nothing since this class does not contain any data members.
+     * API/code coverage may show the assignment operator as present and
+     * untested - ignore.
+     * Subclasses need this assignment operator if they use compiler-provided
+     * assignment operators of their own. An alternative to not declaring one
+     * here would be to declare and empty-implement a protected or public one.
+    UMemory &UMemory::operator=(const UMemory &);
+     */
+};
+
+/**
+ * UObject is the common ICU "boilerplate" class.
+ * UObject inherits UMemory (starting with ICU 2.4),
+ * and all other public ICU C++ classes
+ * are derived from UObject (starting with ICU 2.2).
+ *
+ * UObject contains common virtual functions, in particular a virtual destructor.
+ *
+ * The clone() function is not available in UObject because it is not
+ * implemented by all ICU classes.
+ * Many ICU services provide a clone() function for their class trees,
+ * defined on the service's C++ base class
+ * (which itself is a subclass of UObject).
+ *
+ * @stable ICU 2.2
+ */
+class U_COMMON_API UObject : public UMemory {
+public:
+    /**
+     * Destructor.
+     *
+     * @stable ICU 2.2
+     */
+    virtual ~UObject();
+
+    /**
+     * ICU4C "poor man's RTTI", returns a UClassID for the actual ICU class.
+     * The base class implementation returns a dummy value.
+     *
+     * Use compiler RTTI rather than ICU's "poor man's RTTI".
+     * Since ICU 4.6, new ICU C++ class hierarchies do not implement "poor man's RTTI".
+     *
+     * @stable ICU 2.2
+     */
+    virtual UClassID getDynamicClassID() const;
+
+protected:
+    // the following functions are protected to prevent instantiation and
+    // direct use of UObject itself
+
+    // default constructor
+    // inline UObject() {}
+
+    // copy constructor
+    // inline UObject(const UObject &other) {}
+
+#if 0
+    // TODO Sometime in the future. Implement operator==().
+    // (This comment inserted in 2.2)
+    // some or all of the following "boilerplate" functions may be made public
+    // in a future ICU4C release when all subclasses implement them
+
+    // assignment operator
+    // (not virtual, see "Taligent's Guide to Designing Programs" pp.73..74)
+    // commented out because the implementation is the same as a compiler's default
+    // UObject &operator=(const UObject &other) { return *this; }
+
+    // comparison operators
+    virtual inline bool operator==(const UObject &other) const { return this==&other; }
+    inline bool operator!=(const UObject &other) const { return !operator==(other); }
+
+    // clone() commented out from the base class:
+    // some compilers do not support co-variant return types
+    // (i.e., subclasses would have to return UObject * as well, instead of SubClass *)
+    // see also UObject class documentation.
+    // virtual UObject *clone() const;
+#endif
+
+    /*
+     * Assignment operator not declared. The compiler will provide one
+     * which does nothing since this class does not contain any data members.
+     * API/code coverage may show the assignment operator as present and
+     * untested - ignore.
+     * Subclasses need this assignment operator if they use compiler-provided
+     * assignment operators of their own. An alternative to not declaring one
+     * here would be to declare and empty-implement a protected or public one.
+    UObject &UObject::operator=(const UObject &);
+     */
+};
+
+#ifndef U_HIDE_INTERNAL_API
+/**
+ * This is a simple macro to add ICU RTTI to an ICU object implementation.
+ * This does not go into the header. This should only be used in *.cpp files.
+ *
+ * @param myClass The name of the class that needs RTTI defined.
+ * @internal
+ */
+#define UOBJECT_DEFINE_RTTI_IMPLEMENTATION(myClass) \
+    UClassID U_EXPORT2 myClass::getStaticClassID() { \
+        static char classID = 0; \
+        return (UClassID)&classID; \
+    } \
+    UClassID myClass::getDynamicClassID() const \
+    { return myClass::getStaticClassID(); }
+
+
+/**
+ * This macro adds ICU RTTI to an ICU abstract class implementation.
+ * This macro should be invoked in *.cpp files.  The corresponding
+ * header should declare getStaticClassID.
+ *
+ * @param myClass The name of the class that needs RTTI defined.
+ * @internal
+ */
+#define UOBJECT_DEFINE_ABSTRACT_RTTI_IMPLEMENTATION(myClass) \
+    UClassID U_EXPORT2 myClass::getStaticClassID() { \
+        static char classID = 0; \
+        return (UClassID)&classID; \
+    }
+
+#endif  /* U_HIDE_INTERNAL_API */
+
+U_NAMESPACE_END
+
+#endif /* U_SHOW_CPLUSPLUS_API */
+
+#endif

thirdparty/icu4c/common/unicode/urep.h

@@ -0,0 +1,157 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+/*
+******************************************************************************
+*   Copyright (C) 1997-2010, International Business Machines
+*   Corporation and others.  All Rights Reserved.
+******************************************************************************
+*   Date        Name        Description
+*   06/23/00    aliu        Creation.
+******************************************************************************
+*/
+
+#ifndef __UREP_H
+#define __UREP_H
+
+#include "unicode/utypes.h"
+
+U_CDECL_BEGIN
+
+/********************************************************************
+ * General Notes
+ ********************************************************************
+ * TODO
+ * Add usage scenario
+ * Add test code
+ * Talk about pinning
+ * Talk about "can truncate result if out of memory"
+ */
+
+/********************************************************************
+ * Data Structures
+ ********************************************************************/
+/**
+ * \file
+ * \brief C API: Callbacks for UReplaceable
+ */
+/**
+ * An opaque replaceable text object.  This will be manipulated only
+ * through the caller-supplied UReplaceableFunctor struct.  Related
+ * to the C++ class Replaceable.
+ * This is currently only used in the Transliterator C API, see utrans.h .
+ * @stable ICU 2.0
+ */
+typedef void* UReplaceable;
+
+/**
+ * A set of function pointers that transliterators use to manipulate a
+ * UReplaceable.  The caller should supply the required functions to
+ * manipulate their text appropriately.  Related to the C++ class
+ * Replaceable.
+ * @stable ICU 2.0
+ */
+typedef struct UReplaceableCallbacks {
+
+    /**
+     * Function pointer that returns the number of UChar code units in
+     * this text.
+     *
+     * @param rep A pointer to "this" UReplaceable object.
+     * @return The length of the text.
+     * @stable ICU 2.0
+     */
+    int32_t (*length)(const UReplaceable* rep);
+
+    /**
+     * Function pointer that returns a UChar code units at the given
+     * offset into this text; 0 <= offset < n, where n is the value
+     * returned by (*length)(rep).  See unistr.h for a description of
+     * charAt() vs. char32At().
+     *
+     * @param rep A pointer to "this" UReplaceable object.
+     * @param offset The index at which to fetch the UChar (code unit).
+     * @return The UChar (code unit) at offset, or U+FFFF if the offset is out of bounds.
+     * @stable ICU 2.0
+     */
+    UChar   (*charAt)(const UReplaceable* rep,
+                      int32_t offset);
+
+    /**
+     * Function pointer that returns a UChar32 code point at the given
+     * offset into this text.  See unistr.h for a description of
+     * charAt() vs. char32At().
+     *
+     * @param rep A pointer to "this" UReplaceable object.
+     * @param offset The index at which to fetch the UChar32 (code point).
+     * @return The UChar32 (code point) at offset, or U+FFFF if the offset is out of bounds.
+     * @stable ICU 2.0
+     */
+    UChar32 (*char32At)(const UReplaceable* rep,
+                        int32_t offset);
+    
+    /**
+     * Function pointer that replaces text between start and limit in
+     * this text with the given text.  Attributes (out of band info)
+     * should be retained.
+     *
+     * @param rep A pointer to "this" UReplaceable object.
+     * @param start the starting index of the text to be replaced,
+     * inclusive.
+     * @param limit the ending index of the text to be replaced,
+     * exclusive.
+     * @param text the new text to replace the UChars from
+     * start..limit-1.
+     * @param textLength the number of UChars at text, or -1 if text
+     * is null-terminated.
+     * @stable ICU 2.0
+     */
+    void    (*replace)(UReplaceable* rep,
+                       int32_t start,
+                       int32_t limit,
+                       const UChar* text,
+                       int32_t textLength);
+    
+    /**
+     * Function pointer that copies the characters in the range
+     * [<tt>start</tt>, <tt>limit</tt>) into the array <tt>dst</tt>.
+     *
+     * @param rep A pointer to "this" UReplaceable object.
+     * @param start offset of first character which will be copied
+     * into the array
+     * @param limit offset immediately following the last character to
+     * be copied
+     * @param dst array in which to copy characters.  The length of
+     * <tt>dst</tt> must be at least <tt>(limit - start)</tt>.
+     * @stable ICU 2.1
+     */
+    void    (*extract)(UReplaceable* rep,
+                       int32_t start,
+                       int32_t limit,
+                       UChar* dst);
+
+    /**
+     * Function pointer that copies text between start and limit in
+     * this text to another index in the text.  Attributes (out of
+     * band info) should be retained.  After this call, there will be
+     * (at least) two copies of the characters originally located at
+     * start..limit-1.
+     *
+     * @param rep A pointer to "this" UReplaceable object.
+     * @param start the starting index of the text to be copied,
+     * inclusive.
+     * @param limit the ending index of the text to be copied,
+     * exclusive.
+     * @param dest the index at which the copy of the UChars should be
+     * inserted.
+     * @stable ICU 2.0
+     */
+    void    (*copy)(UReplaceable* rep,
+                    int32_t start,
+                    int32_t limit,
+                    int32_t dest);    
+
+} UReplaceableCallbacks;
+
+U_CDECL_END
+
+#endif

thirdparty/icu4c/common/unicode/ures.h

@@ -0,0 +1,912 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+/*
+**********************************************************************
+*   Copyright (C) 1997-2016, International Business Machines
+*   Corporation and others.  All Rights Reserved.
+**********************************************************************
+*
+* File URES.H (formerly CRESBUND.H)
+*
+* Modification History:
+*
+*   Date        Name        Description
+*   04/01/97    aliu        Creation.
+*   02/22/99    damiba      overhaul.
+*   04/04/99    helena      Fixed internal header inclusion.
+*   04/15/99    Madhu       Updated Javadoc
+*   06/14/99    stephen     Removed functions taking a filename suffix.
+*   07/20/99    stephen     Language-independent typedef to void*
+*   11/09/99    weiv        Added ures_getLocale()
+*   06/24/02    weiv        Added support for resource sharing
+******************************************************************************
+*/
+
+#ifndef URES_H
+#define URES_H
+
+#include "unicode/char16ptr.h"
+#include "unicode/utypes.h"
+#include "unicode/uloc.h"
+
+#if U_SHOW_CPLUSPLUS_API
+#include "unicode/localpointer.h"
+#endif   // U_SHOW_CPLUSPLUS_API
+
+/**
+ * \file
+ * \brief C API: Resource Bundle
+ *
+ * <h2>C API: Resource Bundle</h2>
+ *
+ * C API representing a collection of resource information pertaining to a given
+ * locale. A resource bundle provides a way of accessing locale- specific information in
+ * a data file. You create a resource bundle that manages the resources for a given
+ * locale and then ask it for individual resources.
+ * <P>
+ * Resource bundles in ICU4C are currently defined using text files which conform to the following
+ * <a href="https://github.com/unicode-org/icu-docs/blob/main/design/bnf_rb.txt">BNF definition</a>.
+ * More on resource bundle concepts and syntax can be found in the
+ * <a href="https://unicode-org.github.io/icu/userguide/locale/resources">Users Guide</a>.
+ * <P>
+ */
+
+/**
+ * UResourceBundle is an opaque type for handles for resource bundles in C APIs.
+ * @stable ICU 2.0
+ */
+struct UResourceBundle;
+
+/**
+ * @stable ICU 2.0
+ */
+typedef struct UResourceBundle UResourceBundle;
+
+/**
+ * Numeric constants for types of resource items.
+ * @see ures_getType
+ * @stable ICU 2.0
+ */
+typedef enum {
+    /** Resource type constant for "no resource". @stable ICU 2.6 */
+    URES_NONE=-1,
+
+    /** Resource type constant for 16-bit Unicode strings. @stable ICU 2.6 */
+    URES_STRING=0,
+
+    /** Resource type constant for binary data. @stable ICU 2.6 */
+    URES_BINARY=1,
+
+    /** Resource type constant for tables of key-value pairs. @stable ICU 2.6 */
+    URES_TABLE=2,
+
+    /**
+     * Resource type constant for aliases;
+     * internally stores a string which identifies the actual resource
+     * storing the data (can be in a different resource bundle).
+     * Resolved internally before delivering the actual resource through the API.
+     * @stable ICU 2.6
+     */
+    URES_ALIAS=3,
+
+    /**
+     * Resource type constant for a single 28-bit integer, interpreted as
+     * signed or unsigned by the ures_getInt() or ures_getUInt() function.
+     * @see ures_getInt
+     * @see ures_getUInt
+     * @stable ICU 2.6
+     */
+    URES_INT=7,
+
+    /** Resource type constant for arrays of resources. @stable ICU 2.6 */
+    URES_ARRAY=8,
+
+    /**
+     * Resource type constant for vectors of 32-bit integers.
+     * @see ures_getIntVector
+     * @stable ICU 2.6
+     */
+    URES_INT_VECTOR = 14,
+#ifndef U_HIDE_DEPRECATED_API
+    /** @deprecated ICU 2.6 Use the URES_ constant instead. */
+    RES_NONE=URES_NONE,
+    /** @deprecated ICU 2.6 Use the URES_ constant instead. */
+    RES_STRING=URES_STRING,
+    /** @deprecated ICU 2.6 Use the URES_ constant instead. */
+    RES_BINARY=URES_BINARY,
+    /** @deprecated ICU 2.6 Use the URES_ constant instead. */
+    RES_TABLE=URES_TABLE,
+    /** @deprecated ICU 2.6 Use the URES_ constant instead. */
+    RES_ALIAS=URES_ALIAS,
+    /** @deprecated ICU 2.6 Use the URES_ constant instead. */
+    RES_INT=URES_INT,
+    /** @deprecated ICU 2.6 Use the URES_ constant instead. */
+    RES_ARRAY=URES_ARRAY,
+    /** @deprecated ICU 2.6 Use the URES_ constant instead. */
+    RES_INT_VECTOR=URES_INT_VECTOR,
+    /** @deprecated ICU 2.6 Not used. */
+    RES_RESERVED=15,
+
+    /**
+     * One more than the highest normal UResType value.
+     * @deprecated ICU 58 The numeric value may change over time, see ICU ticket #12420.
+     */
+    URES_LIMIT = 16
+#endif  // U_HIDE_DEPRECATED_API
+} UResType;
+
+/*
+ * Functions to create and destroy resource bundles.
+ */
+
+/**
+ * Opens a UResourceBundle, from which users can extract strings by using
+ * their corresponding keys.
+ * Note that the caller is responsible of calling <TT>ures_close</TT> on each successfully
+ * opened resource bundle.
+ * @param packageName   The packageName and locale together point to an ICU udata object,
+ *                      as defined by <code> udata_open( packageName, "res", locale, err) </code>
+ *                      or equivalent.  Typically, packageName will refer to a (.dat) file, or to
+ *                      a package registered with udata_setAppData(). Using a full file or directory
+ *                      pathname for packageName is deprecated. If NULL, ICU data will be used.
+ * @param locale  specifies the locale for which we want to open the resource
+ *                if NULL, the default locale will be used. If strlen(locale) == 0
+ *                root locale will be used.
+ *
+ * @param status  fills in the outgoing error code.
+ * The UErrorCode err parameter is used to return status information to the user. To
+ * check whether the construction succeeded or not, you should check the value of
+ * U_SUCCESS(err). If you wish more detailed information, you can check for
+ * informational status results which still indicate success. U_USING_FALLBACK_WARNING
+ * indicates that a fall back locale was used. For example, 'de_CH' was requested,
+ * but nothing was found there, so 'de' was used. U_USING_DEFAULT_WARNING indicates that
+ * the default locale data or root locale data was used; neither the requested locale
+ * nor any of its fall back locales could be found. Please see the users guide for more
+ * information on this topic.
+ * @return      a newly allocated resource bundle.
+ * @see ures_close
+ * @stable ICU 2.0
+ */
+U_CAPI UResourceBundle*  U_EXPORT2
+ures_open(const char*    packageName,
+          const char*  locale,
+          UErrorCode*     status);
+
+
+/** This function does not care what kind of localeID is passed in. It simply opens a bundle with
+ *  that name. Fallback mechanism is disabled for the new bundle. If the requested bundle contains
+ *  an %%ALIAS directive, the results are undefined.
+ * @param packageName   The packageName and locale together point to an ICU udata object,
+ *                      as defined by <code> udata_open( packageName, "res", locale, err) </code>
+ *                      or equivalent.  Typically, packageName will refer to a (.dat) file, or to
+ *                      a package registered with udata_setAppData(). Using a full file or directory
+ *                      pathname for packageName is deprecated. If NULL, ICU data will be used.
+ * @param locale  specifies the locale for which we want to open the resource
+ *                if NULL, the default locale will be used. If strlen(locale) == 0
+ *                root locale will be used.
+ *
+ * @param status fills in the outgoing error code. Either U_ZERO_ERROR or U_MISSING_RESOURCE_ERROR
+ * @return      a newly allocated resource bundle or NULL if it doesn't exist.
+ * @see ures_close
+ * @stable ICU 2.0
+ */
+U_CAPI UResourceBundle* U_EXPORT2
+ures_openDirect(const char* packageName,
+                const char* locale,
+                UErrorCode* status);
+
+/**
+ * Same as ures_open() but takes a const UChar *path.
+ * This path will be converted to char * using the default converter,
+ * then ures_open() is called.
+ *
+ * @param packageName   The packageName and locale together point to an ICU udata object,
+ *                      as defined by <code> udata_open( packageName, "res", locale, err) </code>
+ *                      or equivalent.  Typically, packageName will refer to a (.dat) file, or to
+ *                      a package registered with udata_setAppData(). Using a full file or directory
+ *                      pathname for packageName is deprecated. If NULL, ICU data will be used.
+ * @param locale  specifies the locale for which we want to open the resource
+ *                if NULL, the default locale will be used. If strlen(locale) == 0
+ *                root locale will be used.
+ * @param status  fills in the outgoing error code.
+ * @return      a newly allocated resource bundle.
+ * @see ures_open
+ * @stable ICU 2.0
+ */
+U_CAPI UResourceBundle* U_EXPORT2
+ures_openU(const UChar* packageName,
+           const char* locale,
+           UErrorCode* status);
+
+#ifndef U_HIDE_DEPRECATED_API
+/**
+ * Returns the number of strings/arrays in resource bundles.
+ * Better to use ures_getSize, as this function will be deprecated.
+ *
+ *@param resourceBundle resource bundle containing the desired strings
+ *@param resourceKey key tagging the resource
+ *@param err fills in the outgoing error code
+ *                could be <TT>U_MISSING_RESOURCE_ERROR</TT> if the key is not found
+ *                could be a non-failing error
+ *                e.g.: <TT>U_USING_FALLBACK_WARNING</TT>,<TT>U_USING_FALLBACK_WARNING </TT>
+ *@return: for    <STRONG>Arrays</STRONG>: returns the number of resources in the array
+ *                <STRONG>Tables</STRONG>: returns the number of resources in the table
+ *                <STRONG>single string</STRONG>: returns 1
+ *@see ures_getSize
+ * @deprecated ICU 2.8 User ures_getSize instead
+ */
+U_DEPRECATED int32_t U_EXPORT2
+ures_countArrayItems(const UResourceBundle* resourceBundle,
+                     const char* resourceKey,
+                     UErrorCode* err);
+#endif  /* U_HIDE_DEPRECATED_API */
+
+/**
+ * Close a resource bundle, all pointers returned from the various ures_getXXX calls
+ * on this particular bundle should be considered invalid henceforth.
+ *
+ * @param resourceBundle a pointer to a resourceBundle struct. Can be NULL.
+ * @see ures_open
+ * @stable ICU 2.0
+ */
+U_CAPI void U_EXPORT2
+ures_close(UResourceBundle* resourceBundle);
+
+#if U_SHOW_CPLUSPLUS_API
+
+U_NAMESPACE_BEGIN
+
+/**
+ * \class LocalUResourceBundlePointer
+ * "Smart pointer" class, closes a UResourceBundle via ures_close().
+ * For most methods see the LocalPointerBase base class.
+ *
+ * @see LocalPointerBase
+ * @see LocalPointer
+ * @stable ICU 4.4
+ */
+U_DEFINE_LOCAL_OPEN_POINTER(LocalUResourceBundlePointer, UResourceBundle, ures_close);
+
+U_NAMESPACE_END
+
+#endif
+
+#ifndef U_HIDE_DEPRECATED_API
+/**
+ * Return the version number associated with this ResourceBundle as a string. Please
+ * use ures_getVersion as this function is going to be deprecated.
+ *
+ * @param resourceBundle The resource bundle for which the version is checked.
+ * @return  A version number string as specified in the resource bundle or its parent.
+ *          The caller does not own this string.
+ * @see ures_getVersion
+ * @deprecated ICU 2.8 Use ures_getVersion instead.
+ */
+U_DEPRECATED const char* U_EXPORT2
+ures_getVersionNumber(const UResourceBundle*   resourceBundle);
+#endif  /* U_HIDE_DEPRECATED_API */
+
+/**
+ * Return the version number associated with this ResourceBundle as an
+ * UVersionInfo array.
+ *
+ * @param resB The resource bundle for which the version is checked.
+ * @param versionInfo A UVersionInfo array that is filled with the version number
+ *                    as specified in the resource bundle or its parent.
+ * @stable ICU 2.0
+ */
+U_CAPI void U_EXPORT2
+ures_getVersion(const UResourceBundle* resB,
+                UVersionInfo versionInfo);
+
+#ifndef U_HIDE_DEPRECATED_API
+/**
+ * Return the name of the Locale associated with this ResourceBundle. This API allows
+ * you to query for the real locale of the resource. For example, if you requested
+ * "en_US_CALIFORNIA" and only "en_US" bundle exists, "en_US" will be returned.
+ * For subresources, the locale where this resource comes from will be returned.
+ * If fallback has occurred, getLocale will reflect this.
+ *
+ * @param resourceBundle resource bundle in question
+ * @param status just for catching illegal arguments
+ * @return  A Locale name
+ * @deprecated ICU 2.8 Use ures_getLocaleByType instead.
+ */
+U_DEPRECATED const char* U_EXPORT2
+ures_getLocale(const UResourceBundle* resourceBundle,
+               UErrorCode* status);
+#endif  /* U_HIDE_DEPRECATED_API */
+
+/**
+ * Return the name of the Locale associated with this ResourceBundle.
+ * You can choose between requested, valid and real locale.
+ *
+ * @param resourceBundle resource bundle in question
+ * @param type You can choose between requested, valid and actual
+ *             locale. For description see the definition of
+ *             ULocDataLocaleType in uloc.h
+ * @param status just for catching illegal arguments
+ * @return  A Locale name
+ * @stable ICU 2.8
+ */
+U_CAPI const char* U_EXPORT2
+ures_getLocaleByType(const UResourceBundle* resourceBundle,
+                     ULocDataLocaleType type,
+                     UErrorCode* status);
+
+
+#ifndef U_HIDE_INTERNAL_API
+/**
+ * Same as ures_open() but uses the fill-in parameter instead of allocating a new bundle.
+ *
+ * TODO need to revisit usefulness of this function
+ *      and usage model for fillIn parameters without knowing sizeof(UResourceBundle)
+ * @param r The existing UResourceBundle to fill in. If NULL then status will be
+ *               set to U_ILLEGAL_ARGUMENT_ERROR.
+ * @param packageName   The packageName and locale together point to an ICU udata object,
+ *                      as defined by <code> udata_open( packageName, "res", locale, err) </code>
+ *                      or equivalent.  Typically, packageName will refer to a (.dat) file, or to
+ *                      a package registered with udata_setAppData(). Using a full file or directory
+ *                      pathname for packageName is deprecated. If NULL, ICU data will be used.
+ * @param localeID specifies the locale for which we want to open the resource
+ * @param status The error code.
+ * @internal
+ */
+U_CAPI void U_EXPORT2
+ures_openFillIn(UResourceBundle *r,
+                const char* packageName,
+                const char* localeID,
+                UErrorCode* status);
+#endif  /* U_HIDE_INTERNAL_API */
+
+/**
+ * Returns a string from a string resource type
+ *
+ * @param resourceBundle a string resource
+ * @param len    fills in the length of resulting string
+ * @param status fills in the outgoing error code
+ *                could be <TT>U_MISSING_RESOURCE_ERROR</TT> if the key is not found
+ *                Always check the value of status. Don't count on returning NULL.
+ *                could be a non-failing error
+ *                e.g.: <TT>U_USING_FALLBACK_WARNING</TT>,<TT>U_USING_DEFAULT_WARNING </TT>
+ * @return a pointer to a zero-terminated UChar array which lives in a memory mapped/DLL file.
+ * @see ures_getBinary
+ * @see ures_getIntVector
+ * @see ures_getInt
+ * @see ures_getUInt
+ * @stable ICU 2.0
+ */
+U_CAPI const UChar* U_EXPORT2
+ures_getString(const UResourceBundle* resourceBundle,
+               int32_t* len,
+               UErrorCode* status);
+
+/**
+ * Returns a UTF-8 string from a string resource.
+ * The UTF-8 string may be returnable directly as a pointer, or
+ * it may need to be copied, or transformed from UTF-16 using u_strToUTF8()
+ * or equivalent.
+ *
+ * If forceCopy==true, then the string is always written to the dest buffer
+ * and dest is returned.
+ *
+ * If forceCopy==false, then the string is returned as a pointer if possible,
+ * without needing a dest buffer (it can be NULL). If the string needs to be
+ * copied or transformed, then it may be placed into dest at an arbitrary offset.
+ *
+ * If the string is to be written to dest, then U_BUFFER_OVERFLOW_ERROR and
+ * U_STRING_NOT_TERMINATED_WARNING are set if appropriate, as usual.
+ *
+ * If the string is transformed from UTF-16, then a conversion error may occur
+ * if an unpaired surrogate is encountered. If the function is successful, then
+ * the output UTF-8 string is always well-formed.
+ *
+ * @param resB Resource bundle.
+ * @param dest Destination buffer. Can be NULL only if capacity=*length==0.
+ * @param length Input: Capacity of destination buffer.
+ *               Output: Actual length of the UTF-8 string, not counting the
+ *               terminating NUL, even in case of U_BUFFER_OVERFLOW_ERROR.
+ *               Can be NULL, meaning capacity=0 and the string length is not
+ *               returned to the caller.
+ * @param forceCopy If true, then the output string will always be written to
+ *                  dest, with U_BUFFER_OVERFLOW_ERROR and
+ *                  U_STRING_NOT_TERMINATED_WARNING set if appropriate.
+ *                  If false, then the dest buffer may or may not contain a
+ *                  copy of the string. dest may or may not be modified.
+ *                  If a copy needs to be written, then the UErrorCode parameter
+ *                  indicates overflow etc. as usual.
+ * @param status Pointer to a standard ICU error code. Its input value must
+ *               pass the U_SUCCESS() test, or else the function returns
+ *               immediately. Check for U_FAILURE() on output or use with
+ *               function chaining. (See User Guide for details.)
+ * @return The pointer to the UTF-8 string. It may be dest, or at some offset
+ *         from dest (only if !forceCopy), or in unrelated memory.
+ *         Always NUL-terminated unless the string was written to dest and
+ *         length==capacity (in which case U_STRING_NOT_TERMINATED_WARNING is set).
+ *
+ * @see ures_getString
+ * @see u_strToUTF8
+ * @stable ICU 3.6
+ */
+U_CAPI const char * U_EXPORT2
+ures_getUTF8String(const UResourceBundle *resB,
+                   char *dest, int32_t *length,
+                   UBool forceCopy,
+                   UErrorCode *status);
+
+/**
+ * Returns a binary data from a binary resource.
+ *
+ * @param resourceBundle a string resource
+ * @param len    fills in the length of resulting byte chunk
+ * @param status fills in the outgoing error code
+ *                could be <TT>U_MISSING_RESOURCE_ERROR</TT> if the key is not found
+ *                Always check the value of status. Don't count on returning NULL.
+ *                could be a non-failing error
+ *                e.g.: <TT>U_USING_FALLBACK_WARNING</TT>,<TT>U_USING_DEFAULT_WARNING </TT>
+ * @return a pointer to a chunk of unsigned bytes which live in a memory mapped/DLL file.
+ * @see ures_getString
+ * @see ures_getIntVector
+ * @see ures_getInt
+ * @see ures_getUInt
+ * @stable ICU 2.0
+ */
+U_CAPI const uint8_t* U_EXPORT2
+ures_getBinary(const UResourceBundle* resourceBundle,
+               int32_t* len,
+               UErrorCode* status);
+
+/**
+ * Returns a 32 bit integer array from a resource.
+ *
+ * @param resourceBundle an int vector resource
+ * @param len    fills in the length of resulting byte chunk
+ * @param status fills in the outgoing error code
+ *                could be <TT>U_MISSING_RESOURCE_ERROR</TT> if the key is not found
+ *                Always check the value of status. Don't count on returning NULL.
+ *                could be a non-failing error
+ *                e.g.: <TT>U_USING_FALLBACK_WARNING</TT>,<TT>U_USING_DEFAULT_WARNING </TT>
+ * @return a pointer to a chunk of integers which live in a memory mapped/DLL file.
+ * @see ures_getBinary
+ * @see ures_getString
+ * @see ures_getInt
+ * @see ures_getUInt
+ * @stable ICU 2.0
+ */
+U_CAPI const int32_t* U_EXPORT2
+ures_getIntVector(const UResourceBundle* resourceBundle,
+                  int32_t* len,
+                  UErrorCode* status);
+
+/**
+ * Returns an unsigned integer from a resource.
+ * This integer is originally 28 bits.
+ *
+ * @param resourceBundle a string resource
+ * @param status fills in the outgoing error code
+ *                could be <TT>U_MISSING_RESOURCE_ERROR</TT> if the key is not found
+ *                could be a non-failing error
+ *                e.g.: <TT>U_USING_FALLBACK_WARNING</TT>,<TT>U_USING_DEFAULT_WARNING </TT>
+ * @return an integer value
+ * @see ures_getInt
+ * @see ures_getIntVector
+ * @see ures_getBinary
+ * @see ures_getString
+ * @stable ICU 2.0
+ */
+U_CAPI uint32_t U_EXPORT2
+ures_getUInt(const UResourceBundle* resourceBundle,
+             UErrorCode *status);
+
+/**
+ * Returns a signed integer from a resource.
+ * This integer is originally 28 bit and the sign gets propagated.
+ *
+ * @param resourceBundle a string resource
+ * @param status  fills in the outgoing error code
+ *                could be <TT>U_MISSING_RESOURCE_ERROR</TT> if the key is not found
+ *                could be a non-failing error
+ *                e.g.: <TT>U_USING_FALLBACK_WARNING</TT>,<TT>U_USING_DEFAULT_WARNING </TT>
+ * @return an integer value
+ * @see ures_getUInt
+ * @see ures_getIntVector
+ * @see ures_getBinary
+ * @see ures_getString
+ * @stable ICU 2.0
+ */
+U_CAPI int32_t U_EXPORT2
+ures_getInt(const UResourceBundle* resourceBundle,
+            UErrorCode *status);
+
+/**
+ * Returns the size of a resource. Size for scalar types is always 1,
+ * and for vector/table types is the number of child resources.
+ * @warning Integer array is treated as a scalar type. There are no
+ *          APIs to access individual members of an integer array. It
+ *          is always returned as a whole.
+ * @param resourceBundle a resource
+ * @return number of resources in a given resource.
+ * @stable ICU 2.0
+ */
+U_CAPI int32_t U_EXPORT2
+ures_getSize(const UResourceBundle *resourceBundle);
+
+/**
+ * Returns the type of a resource. Available types are defined in enum UResType
+ *
+ * @param resourceBundle a resource
+ * @return type of the given resource.
+ * @see UResType
+ * @stable ICU 2.0
+ */
+U_CAPI UResType U_EXPORT2
+ures_getType(const UResourceBundle *resourceBundle);
+
+/**
+ * Returns the key associated with a given resource. Not all the resources have a key - only
+ * those that are members of a table.
+ *
+ * @param resourceBundle a resource
+ * @return a key associated to this resource, or NULL if it doesn't have a key
+ * @stable ICU 2.0
+ */
+U_CAPI const char * U_EXPORT2
+ures_getKey(const UResourceBundle *resourceBundle);
+
+/* ITERATION API
+    This API provides means for iterating through a resource
+*/
+
+/**
+ * Resets the internal context of a resource so that iteration starts from the first element.
+ *
+ * @param resourceBundle a resource
+ * @stable ICU 2.0
+ */
+U_CAPI void U_EXPORT2
+ures_resetIterator(UResourceBundle *resourceBundle);
+
+/**
+ * Checks whether the given resource has another element to iterate over.
+ *
+ * @param resourceBundle a resource
+ * @return true if there are more elements, false if there is no more elements
+ * @stable ICU 2.0
+ */
+U_CAPI UBool U_EXPORT2
+ures_hasNext(const UResourceBundle *resourceBundle);
+
+/**
+ * Returns the next resource in a given resource or NULL if there are no more resources
+ * to iterate over. Features a fill-in parameter.
+ *
+ * @param resourceBundle    a resource
+ * @param fillIn            if NULL a new UResourceBundle struct is allocated and must be closed by the caller.
+ *                          Alternatively, you can supply a struct to be filled by this function.
+ * @param status            fills in the outgoing error code. You may still get a non NULL result even if an
+ *                          error occurred. Check status instead.
+ * @return                  a pointer to a UResourceBundle struct. If fill in param was NULL, caller must close it
+ * @stable ICU 2.0
+ */
+U_CAPI UResourceBundle* U_EXPORT2
+ures_getNextResource(UResourceBundle *resourceBundle,
+                     UResourceBundle *fillIn,
+                     UErrorCode *status);
+
+/**
+ * Returns the next string in a given resource or NULL if there are no more resources
+ * to iterate over.
+ *
+ * @param resourceBundle    a resource
+ * @param len               fill in length of the string
+ * @param key               fill in for key associated with this string. NULL if no key
+ * @param status            fills in the outgoing error code. If an error occurred, we may return NULL, but don't
+ *                          count on it. Check status instead!
+ * @return a pointer to a zero-terminated UChar array which lives in a memory mapped/DLL file.
+ * @stable ICU 2.0
+ */
+U_CAPI const UChar* U_EXPORT2
+ures_getNextString(UResourceBundle *resourceBundle,
+                   int32_t* len,
+                   const char ** key,
+                   UErrorCode *status);
+
+/**
+ * Returns the resource in a given resource at the specified index. Features a fill-in parameter.
+ *
+ * @param resourceBundle    the resource bundle from which to get a sub-resource
+ * @param indexR            an index to the wanted resource.
+ * @param fillIn            if NULL a new UResourceBundle struct is allocated and must be closed by the caller.
+ *                          Alternatively, you can supply a struct to be filled by this function.
+ * @param status            fills in the outgoing error code. Don't count on NULL being returned if an error has
+ *                          occurred. Check status instead.
+ * @return                  a pointer to a UResourceBundle struct. If fill in param was NULL, caller must close it
+ * @stable ICU 2.0
+ */
+U_CAPI UResourceBundle* U_EXPORT2
+ures_getByIndex(const UResourceBundle *resourceBundle,
+                int32_t indexR,
+                UResourceBundle *fillIn,
+                UErrorCode *status);
+
+/**
+ * Returns the string in a given resource at the specified index.
+ *
+ * @param resourceBundle    a resource
+ * @param indexS            an index to the wanted string.
+ * @param len               fill in length of the string
+ * @param status            fills in the outgoing error code. If an error occurred, we may return NULL, but don't
+ *                          count on it. Check status instead!
+ * @return                  a pointer to a zero-terminated UChar array which lives in a memory mapped/DLL file.
+ * @stable ICU 2.0
+ */
+U_CAPI const UChar* U_EXPORT2
+ures_getStringByIndex(const UResourceBundle *resourceBundle,
+                      int32_t indexS,
+                      int32_t* len,
+                      UErrorCode *status);
+
+/**
+ * Returns a UTF-8 string from a resource at the specified index.
+ * The UTF-8 string may be returnable directly as a pointer, or
+ * it may need to be copied, or transformed from UTF-16 using u_strToUTF8()
+ * or equivalent.
+ *
+ * If forceCopy==true, then the string is always written to the dest buffer
+ * and dest is returned.
+ *
+ * If forceCopy==false, then the string is returned as a pointer if possible,
+ * without needing a dest buffer (it can be NULL). If the string needs to be
+ * copied or transformed, then it may be placed into dest at an arbitrary offset.
+ *
+ * If the string is to be written to dest, then U_BUFFER_OVERFLOW_ERROR and
+ * U_STRING_NOT_TERMINATED_WARNING are set if appropriate, as usual.
+ *
+ * If the string is transformed from UTF-16, then a conversion error may occur
+ * if an unpaired surrogate is encountered. If the function is successful, then
+ * the output UTF-8 string is always well-formed.
+ *
+ * @param resB Resource bundle.
+ * @param stringIndex An index to the wanted string.
+ * @param dest Destination buffer. Can be NULL only if capacity=*length==0.
+ * @param pLength Input: Capacity of destination buffer.
+ *               Output: Actual length of the UTF-8 string, not counting the
+ *               terminating NUL, even in case of U_BUFFER_OVERFLOW_ERROR.
+ *               Can be NULL, meaning capacity=0 and the string length is not
+ *               returned to the caller.
+ * @param forceCopy If true, then the output string will always be written to
+ *                  dest, with U_BUFFER_OVERFLOW_ERROR and
+ *                  U_STRING_NOT_TERMINATED_WARNING set if appropriate.
+ *                  If false, then the dest buffer may or may not contain a
+ *                  copy of the string. dest may or may not be modified.
+ *                  If a copy needs to be written, then the UErrorCode parameter
+ *                  indicates overflow etc. as usual.
+ * @param status Pointer to a standard ICU error code. Its input value must
+ *               pass the U_SUCCESS() test, or else the function returns
+ *               immediately. Check for U_FAILURE() on output or use with
+ *               function chaining. (See User Guide for details.)
+ * @return The pointer to the UTF-8 string. It may be dest, or at some offset
+ *         from dest (only if !forceCopy), or in unrelated memory.
+ *         Always NUL-terminated unless the string was written to dest and
+ *         length==capacity (in which case U_STRING_NOT_TERMINATED_WARNING is set).
+ *
+ * @see ures_getStringByIndex
+ * @see u_strToUTF8
+ * @stable ICU 3.6
+ */
+U_CAPI const char * U_EXPORT2
+ures_getUTF8StringByIndex(const UResourceBundle *resB,
+                          int32_t stringIndex,
+                          char *dest, int32_t *pLength,
+                          UBool forceCopy,
+                          UErrorCode *status);
+
+/**
+ * Returns a resource in a given resource that has a given key. This procedure works only with table
+ * resources. Features a fill-in parameter.
+ *
+ * @param resourceBundle    a resource
+ * @param key               a key associated with the wanted resource
+ * @param fillIn            if NULL a new UResourceBundle struct is allocated and must be closed by the caller.
+ *                          Alternatively, you can supply a struct to be filled by this function.
+ * @param status            fills in the outgoing error code.
+ * @return                  a pointer to a UResourceBundle struct. If fill in param was NULL, caller must close it
+ * @stable ICU 2.0
+ */
+U_CAPI UResourceBundle* U_EXPORT2
+ures_getByKey(const UResourceBundle *resourceBundle,
+              const char* key,
+              UResourceBundle *fillIn,
+              UErrorCode *status);
+
+/**
+ * Returns a string in a given resource that has a given key. This procedure works only with table
+ * resources.
+ *
+ * @param resB              a resource
+ * @param key               a key associated with the wanted string
+ * @param len               fill in length of the string
+ * @param status            fills in the outgoing error code. If an error occurred, we may return NULL, but don't
+ *                          count on it. Check status instead!
+ * @return                  a pointer to a zero-terminated UChar array which lives in a memory mapped/DLL file.
+ * @stable ICU 2.0
+ */
+U_CAPI const UChar* U_EXPORT2
+ures_getStringByKey(const UResourceBundle *resB,
+                    const char* key,
+                    int32_t* len,
+                    UErrorCode *status);
+
+/**
+ * Returns a UTF-8 string from a resource and a key.
+ * This function works only with table resources.
+ *
+ * The UTF-8 string may be returnable directly as a pointer, or
+ * it may need to be copied, or transformed from UTF-16 using u_strToUTF8()
+ * or equivalent.
+ *
+ * If forceCopy==true, then the string is always written to the dest buffer
+ * and dest is returned.
+ *
+ * If forceCopy==false, then the string is returned as a pointer if possible,
+ * without needing a dest buffer (it can be NULL). If the string needs to be
+ * copied or transformed, then it may be placed into dest at an arbitrary offset.
+ *
+ * If the string is to be written to dest, then U_BUFFER_OVERFLOW_ERROR and
+ * U_STRING_NOT_TERMINATED_WARNING are set if appropriate, as usual.
+ *
+ * If the string is transformed from UTF-16, then a conversion error may occur
+ * if an unpaired surrogate is encountered. If the function is successful, then
+ * the output UTF-8 string is always well-formed.
+ *
+ * @param resB Resource bundle.
+ * @param key  A key associated with the wanted resource
+ * @param dest Destination buffer. Can be NULL only if capacity=*length==0.
+ * @param pLength Input: Capacity of destination buffer.
+ *               Output: Actual length of the UTF-8 string, not counting the
+ *               terminating NUL, even in case of U_BUFFER_OVERFLOW_ERROR.
+ *               Can be NULL, meaning capacity=0 and the string length is not
+ *               returned to the caller.
+ * @param forceCopy If true, then the output string will always be written to
+ *                  dest, with U_BUFFER_OVERFLOW_ERROR and
+ *                  U_STRING_NOT_TERMINATED_WARNING set if appropriate.
+ *                  If false, then the dest buffer may or may not contain a
+ *                  copy of the string. dest may or may not be modified.
+ *                  If a copy needs to be written, then the UErrorCode parameter
+ *                  indicates overflow etc. as usual.
+ * @param status Pointer to a standard ICU error code. Its input value must
+ *               pass the U_SUCCESS() test, or else the function returns
+ *               immediately. Check for U_FAILURE() on output or use with
+ *               function chaining. (See User Guide for details.)
+ * @return The pointer to the UTF-8 string. It may be dest, or at some offset
+ *         from dest (only if !forceCopy), or in unrelated memory.
+ *         Always NUL-terminated unless the string was written to dest and
+ *         length==capacity (in which case U_STRING_NOT_TERMINATED_WARNING is set).
+ *
+ * @see ures_getStringByKey
+ * @see u_strToUTF8
+ * @stable ICU 3.6
+ */
+U_CAPI const char * U_EXPORT2
+ures_getUTF8StringByKey(const UResourceBundle *resB,
+                        const char *key,
+                        char *dest, int32_t *pLength,
+                        UBool forceCopy,
+                        UErrorCode *status);
+
+#if U_SHOW_CPLUSPLUS_API
+#include "unicode/unistr.h"
+
+U_NAMESPACE_BEGIN
+/**
+ * Returns the string value from a string resource bundle.
+ *
+ * @param resB    a resource, should have type URES_STRING
+ * @param status: fills in the outgoing error code
+ *                could be <TT>U_MISSING_RESOURCE_ERROR</TT> if the key is not found
+ *                could be a non-failing error
+ *                e.g.: <TT>U_USING_FALLBACK_WARNING</TT>,<TT>U_USING_DEFAULT_WARNING </TT>
+ * @return The string value, or a bogus string if there is a failure UErrorCode.
+ * @stable ICU 2.0
+ */
+inline UnicodeString
+ures_getUnicodeString(const UResourceBundle *resB, UErrorCode* status) {
+    UnicodeString result;
+    int32_t len = 0;
+    const char16_t *r = ConstChar16Ptr(ures_getString(resB, &len, status));
+    if(U_SUCCESS(*status)) {
+        result.setTo(true, r, len);
+    } else {
+        result.setToBogus();
+    }
+    return result;
+}
+
+/**
+ * Returns the next string in a resource, or an empty string if there are no more resources
+ * to iterate over.
+ * Use ures_getNextString() instead to distinguish between
+ * the end of the iteration and a real empty string value.
+ *
+ * @param resB              a resource
+ * @param key               fill in for key associated with this string
+ * @param status            fills in the outgoing error code
+ * @return The string value, or a bogus string if there is a failure UErrorCode.
+ * @stable ICU 2.0
+ */
+inline UnicodeString
+ures_getNextUnicodeString(UResourceBundle *resB, const char ** key, UErrorCode* status) {
+    UnicodeString result;
+    int32_t len = 0;
+    const char16_t* r = ConstChar16Ptr(ures_getNextString(resB, &len, key, status));
+    if(U_SUCCESS(*status)) {
+        result.setTo(true, r, len);
+    } else {
+        result.setToBogus();
+    }
+    return result;
+}
+
+/**
+ * Returns the string in a given resource array or table at the specified index.
+ *
+ * @param resB              a resource
+ * @param indexS            an index to the wanted string.
+ * @param status            fills in the outgoing error code
+ * @return The string value, or a bogus string if there is a failure UErrorCode.
+ * @stable ICU 2.0
+ */
+inline UnicodeString
+ures_getUnicodeStringByIndex(const UResourceBundle *resB, int32_t indexS, UErrorCode* status) {
+    UnicodeString result;
+    int32_t len = 0;
+    const char16_t* r = ConstChar16Ptr(ures_getStringByIndex(resB, indexS, &len, status));
+    if(U_SUCCESS(*status)) {
+        result.setTo(true, r, len);
+    } else {
+        result.setToBogus();
+    }
+    return result;
+}
+
+/**
+ * Returns a string in a resource that has a given key.
+ * This procedure works only with table resources.
+ *
+ * @param resB              a resource
+ * @param key               a key associated with the wanted string
+ * @param status            fills in the outgoing error code
+ * @return The string value, or a bogus string if there is a failure UErrorCode.
+ * @stable ICU 2.0
+ */
+inline UnicodeString
+ures_getUnicodeStringByKey(const UResourceBundle *resB, const char* key, UErrorCode* status) {
+    UnicodeString result;
+    int32_t len = 0;
+    const char16_t* r = ConstChar16Ptr(ures_getStringByKey(resB, key, &len, status));
+    if(U_SUCCESS(*status)) {
+        result.setTo(true, r, len);
+    } else {
+        result.setToBogus();
+    }
+    return result;
+}
+
+U_NAMESPACE_END
+
+#endif
+
+/**
+ * Create a string enumerator, owned by the caller, of all locales located within
+ * the specified resource tree.
+ * @param packageName name of the tree, such as (NULL) or U_ICUDATA_ALIAS or  or "ICUDATA-coll"
+ * This call is similar to uloc_getAvailable().
+ * @param status error code
+ * @stable ICU 3.2
+ */
+U_CAPI UEnumeration* U_EXPORT2
+ures_openAvailableLocales(const char *packageName, UErrorCode *status);
+
+
+#endif /*_URES*/
+/*eof*/

thirdparty/icu4c/common/unicode/uscript.h

@@ -0,0 +1,742 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+/*
+ **********************************************************************
+ *   Copyright (C) 1997-2016, International Business Machines
+ *   Corporation and others.  All Rights Reserved.
+ **********************************************************************
+ *
+ * File USCRIPT.H
+ *
+ * Modification History:
+ *
+ *   Date        Name        Description
+ *   07/06/2001    Ram         Creation.
+ ******************************************************************************
+ */
+
+#ifndef USCRIPT_H
+#define USCRIPT_H
+#include "unicode/utypes.h"
+
+/**
+ * \file
+ * \brief C API: Unicode Script Information
+ */
+
+/**
+ * Constants for ISO 15924 script codes.
+ *
+ * The current set of script code constants supports at least all scripts
+ * that are encoded in the version of Unicode which ICU currently supports.
+ * The names of the constants are usually derived from the
+ * Unicode script property value aliases.
+ * See UAX #24 Unicode Script Property (http://www.unicode.org/reports/tr24/)
+ * and http://www.unicode.org/Public/UCD/latest/ucd/PropertyValueAliases.txt .
+ *
+ * In addition, constants for many ISO 15924 script codes
+ * are included, for use with language tags, CLDR data, and similar.
+ * Some of those codes are not used in the Unicode Character Database (UCD).
+ * For example, there are no characters that have a UCD script property value of
+ * Hans or Hant. All Han ideographs have the Hani script property value in Unicode.
+ *
+ * Private-use codes Qaaa..Qabx are not included, except as used in the UCD or in CLDR.
+ *
+ * Starting with ICU 55, script codes are only added when their scripts
+ * have been or will certainly be encoded in Unicode,
+ * and have been assigned Unicode script property value aliases,
+ * to ensure that their script names are stable and match the names of the constants.
+ * Script codes like Latf and Aran that are not subject to separate encoding
+ * may be added at any time.
+ *
+ * @stable ICU 2.2
+ */
+typedef enum UScriptCode {
+    /*
+     * Note: UScriptCode constants and their ISO script code comments
+     * are parsed by preparseucd.py.
+     * It matches lines like
+     *     USCRIPT_<Unicode Script value name> = <integer>,  / * <ISO script code> * /
+     */
+
+      /** @stable ICU 2.2 */
+      USCRIPT_INVALID_CODE = -1,
+      /** @stable ICU 2.2 */
+      USCRIPT_COMMON       =  0,  /* Zyyy */
+      /** @stable ICU 2.2 */
+      USCRIPT_INHERITED    =  1,  /* Zinh */ /* "Code for inherited script", for non-spacing combining marks; also Qaai */
+      /** @stable ICU 2.2 */
+      USCRIPT_ARABIC       =  2,  /* Arab */
+      /** @stable ICU 2.2 */
+      USCRIPT_ARMENIAN     =  3,  /* Armn */
+      /** @stable ICU 2.2 */
+      USCRIPT_BENGALI      =  4,  /* Beng */
+      /** @stable ICU 2.2 */
+      USCRIPT_BOPOMOFO     =  5,  /* Bopo */
+      /** @stable ICU 2.2 */
+      USCRIPT_CHEROKEE     =  6,  /* Cher */
+      /** @stable ICU 2.2 */
+      USCRIPT_COPTIC       =  7,  /* Copt */
+      /** @stable ICU 2.2 */
+      USCRIPT_CYRILLIC     =  8,  /* Cyrl */
+      /** @stable ICU 2.2 */
+      USCRIPT_DESERET      =  9,  /* Dsrt */
+      /** @stable ICU 2.2 */
+      USCRIPT_DEVANAGARI   = 10,  /* Deva */
+      /** @stable ICU 2.2 */
+      USCRIPT_ETHIOPIC     = 11,  /* Ethi */
+      /** @stable ICU 2.2 */
+      USCRIPT_GEORGIAN     = 12,  /* Geor */
+      /** @stable ICU 2.2 */
+      USCRIPT_GOTHIC       = 13,  /* Goth */
+      /** @stable ICU 2.2 */
+      USCRIPT_GREEK        = 14,  /* Grek */
+      /** @stable ICU 2.2 */
+      USCRIPT_GUJARATI     = 15,  /* Gujr */
+      /** @stable ICU 2.2 */
+      USCRIPT_GURMUKHI     = 16,  /* Guru */
+      /** @stable ICU 2.2 */
+      USCRIPT_HAN          = 17,  /* Hani */
+      /** @stable ICU 2.2 */
+      USCRIPT_HANGUL       = 18,  /* Hang */
+      /** @stable ICU 2.2 */
+      USCRIPT_HEBREW       = 19,  /* Hebr */
+      /** @stable ICU 2.2 */
+      USCRIPT_HIRAGANA     = 20,  /* Hira */
+      /** @stable ICU 2.2 */
+      USCRIPT_KANNADA      = 21,  /* Knda */
+      /** @stable ICU 2.2 */
+      USCRIPT_KATAKANA     = 22,  /* Kana */
+      /** @stable ICU 2.2 */
+      USCRIPT_KHMER        = 23,  /* Khmr */
+      /** @stable ICU 2.2 */
+      USCRIPT_LAO          = 24,  /* Laoo */
+      /** @stable ICU 2.2 */
+      USCRIPT_LATIN        = 25,  /* Latn */
+      /** @stable ICU 2.2 */
+      USCRIPT_MALAYALAM    = 26,  /* Mlym */
+      /** @stable ICU 2.2 */
+      USCRIPT_MONGOLIAN    = 27,  /* Mong */
+      /** @stable ICU 2.2 */
+      USCRIPT_MYANMAR      = 28,  /* Mymr */
+      /** @stable ICU 2.2 */
+      USCRIPT_OGHAM        = 29,  /* Ogam */
+      /** @stable ICU 2.2 */
+      USCRIPT_OLD_ITALIC   = 30,  /* Ital */
+      /** @stable ICU 2.2 */
+      USCRIPT_ORIYA        = 31,  /* Orya */
+      /** @stable ICU 2.2 */
+      USCRIPT_RUNIC        = 32,  /* Runr */
+      /** @stable ICU 2.2 */
+      USCRIPT_SINHALA      = 33,  /* Sinh */
+      /** @stable ICU 2.2 */
+      USCRIPT_SYRIAC       = 34,  /* Syrc */
+      /** @stable ICU 2.2 */
+      USCRIPT_TAMIL        = 35,  /* Taml */
+      /** @stable ICU 2.2 */
+      USCRIPT_TELUGU       = 36,  /* Telu */
+      /** @stable ICU 2.2 */
+      USCRIPT_THAANA       = 37,  /* Thaa */
+      /** @stable ICU 2.2 */
+      USCRIPT_THAI         = 38,  /* Thai */
+      /** @stable ICU 2.2 */
+      USCRIPT_TIBETAN      = 39,  /* Tibt */
+      /** Canadian_Aboriginal script. @stable ICU 2.6 */
+      USCRIPT_CANADIAN_ABORIGINAL = 40,  /* Cans */
+      /** Canadian_Aboriginal script (alias). @stable ICU 2.2 */
+      USCRIPT_UCAS         = USCRIPT_CANADIAN_ABORIGINAL,
+      /** @stable ICU 2.2 */
+      USCRIPT_YI           = 41,  /* Yiii */
+      /* New scripts in Unicode 3.2 */
+      /** @stable ICU 2.2 */
+      USCRIPT_TAGALOG      = 42,  /* Tglg */
+      /** @stable ICU 2.2 */
+      USCRIPT_HANUNOO      = 43,  /* Hano */
+      /** @stable ICU 2.2 */
+      USCRIPT_BUHID        = 44,  /* Buhd */
+      /** @stable ICU 2.2 */
+      USCRIPT_TAGBANWA     = 45,  /* Tagb */
+
+      /* New scripts in Unicode 4 */
+      /** @stable ICU 2.6 */
+      USCRIPT_BRAILLE      = 46,  /* Brai */
+      /** @stable ICU 2.6 */
+      USCRIPT_CYPRIOT      = 47,  /* Cprt */
+      /** @stable ICU 2.6 */
+      USCRIPT_LIMBU        = 48,  /* Limb */
+      /** @stable ICU 2.6 */
+      USCRIPT_LINEAR_B     = 49,  /* Linb */
+      /** @stable ICU 2.6 */
+      USCRIPT_OSMANYA      = 50,  /* Osma */
+      /** @stable ICU 2.6 */
+      USCRIPT_SHAVIAN      = 51,  /* Shaw */
+      /** @stable ICU 2.6 */
+      USCRIPT_TAI_LE       = 52,  /* Tale */
+      /** @stable ICU 2.6 */
+      USCRIPT_UGARITIC     = 53,  /* Ugar */
+
+      /** New script code in Unicode 4.0.1 @stable ICU 3.0 */
+      USCRIPT_KATAKANA_OR_HIRAGANA = 54,/*Hrkt */
+
+      /* New scripts in Unicode 4.1 */
+      /** @stable ICU 3.4 */
+      USCRIPT_BUGINESE      = 55, /* Bugi */
+      /** @stable ICU 3.4 */
+      USCRIPT_GLAGOLITIC    = 56, /* Glag */
+      /** @stable ICU 3.4 */
+      USCRIPT_KHAROSHTHI    = 57, /* Khar */
+      /** @stable ICU 3.4 */
+      USCRIPT_SYLOTI_NAGRI  = 58, /* Sylo */
+      /** @stable ICU 3.4 */
+      USCRIPT_NEW_TAI_LUE   = 59, /* Talu */
+      /** @stable ICU 3.4 */
+      USCRIPT_TIFINAGH      = 60, /* Tfng */
+      /** @stable ICU 3.4 */
+      USCRIPT_OLD_PERSIAN   = 61, /* Xpeo */
+
+      /* New script codes from Unicode and ISO 15924 */
+      /** @stable ICU 3.6 */
+      USCRIPT_BALINESE                      = 62, /* Bali */
+      /** @stable ICU 3.6 */
+      USCRIPT_BATAK                         = 63, /* Batk */
+      /** @stable ICU 3.6 */
+      USCRIPT_BLISSYMBOLS                   = 64, /* Blis */
+      /** @stable ICU 3.6 */
+      USCRIPT_BRAHMI                        = 65, /* Brah */
+      /** @stable ICU 3.6 */
+      USCRIPT_CHAM                          = 66, /* Cham */
+      /** @stable ICU 3.6 */
+      USCRIPT_CIRTH                         = 67, /* Cirt */
+      /** @stable ICU 3.6 */
+      USCRIPT_OLD_CHURCH_SLAVONIC_CYRILLIC  = 68, /* Cyrs */
+      /** @stable ICU 3.6 */
+      USCRIPT_DEMOTIC_EGYPTIAN              = 69, /* Egyd */
+      /** @stable ICU 3.6 */
+      USCRIPT_HIERATIC_EGYPTIAN             = 70, /* Egyh */
+      /** @stable ICU 3.6 */
+      USCRIPT_EGYPTIAN_HIEROGLYPHS          = 71, /* Egyp */
+      /** @stable ICU 3.6 */
+      USCRIPT_KHUTSURI                      = 72, /* Geok */
+      /** @stable ICU 3.6 */
+      USCRIPT_SIMPLIFIED_HAN                = 73, /* Hans */
+      /** @stable ICU 3.6 */
+      USCRIPT_TRADITIONAL_HAN               = 74, /* Hant */
+      /** @stable ICU 3.6 */
+      USCRIPT_PAHAWH_HMONG                  = 75, /* Hmng */
+      /** @stable ICU 3.6 */
+      USCRIPT_OLD_HUNGARIAN                 = 76, /* Hung */
+      /** @stable ICU 3.6 */
+      USCRIPT_HARAPPAN_INDUS                = 77, /* Inds */
+      /** @stable ICU 3.6 */
+      USCRIPT_JAVANESE                      = 78, /* Java */
+      /** @stable ICU 3.6 */
+      USCRIPT_KAYAH_LI                      = 79, /* Kali */
+      /** @stable ICU 3.6 */
+      USCRIPT_LATIN_FRAKTUR                 = 80, /* Latf */
+      /** @stable ICU 3.6 */
+      USCRIPT_LATIN_GAELIC                  = 81, /* Latg */
+      /** @stable ICU 3.6 */
+      USCRIPT_LEPCHA                        = 82, /* Lepc */
+      /** @stable ICU 3.6 */
+      USCRIPT_LINEAR_A                      = 83, /* Lina */
+      /** @stable ICU 4.6 */
+      USCRIPT_MANDAIC                       = 84, /* Mand */
+      /** @stable ICU 3.6 */
+      USCRIPT_MANDAEAN                      = USCRIPT_MANDAIC,
+      /** @stable ICU 3.6 */
+      USCRIPT_MAYAN_HIEROGLYPHS             = 85, /* Maya */
+      /** @stable ICU 4.6 */
+      USCRIPT_MEROITIC_HIEROGLYPHS          = 86, /* Mero */
+      /** @stable ICU 3.6 */
+      USCRIPT_MEROITIC                      = USCRIPT_MEROITIC_HIEROGLYPHS,
+      /** @stable ICU 3.6 */
+      USCRIPT_NKO                           = 87, /* Nkoo */
+      /** @stable ICU 3.6 */
+      USCRIPT_ORKHON                        = 88, /* Orkh */
+      /** @stable ICU 3.6 */
+      USCRIPT_OLD_PERMIC                    = 89, /* Perm */
+      /** @stable ICU 3.6 */
+      USCRIPT_PHAGS_PA                      = 90, /* Phag */
+      /** @stable ICU 3.6 */
+      USCRIPT_PHOENICIAN                    = 91, /* Phnx */
+      /** @stable ICU 52 */
+      USCRIPT_MIAO                          = 92, /* Plrd */
+      /** @stable ICU 3.6 */
+      USCRIPT_PHONETIC_POLLARD              = USCRIPT_MIAO,
+      /** @stable ICU 3.6 */
+      USCRIPT_RONGORONGO                    = 93, /* Roro */
+      /** @stable ICU 3.6 */
+      USCRIPT_SARATI                        = 94, /* Sara */
+      /** @stable ICU 3.6 */
+      USCRIPT_ESTRANGELO_SYRIAC             = 95, /* Syre */
+      /** @stable ICU 3.6 */
+      USCRIPT_WESTERN_SYRIAC                = 96, /* Syrj */
+      /** @stable ICU 3.6 */
+      USCRIPT_EASTERN_SYRIAC                = 97, /* Syrn */
+      /** @stable ICU 3.6 */
+      USCRIPT_TENGWAR                       = 98, /* Teng */
+      /** @stable ICU 3.6 */
+      USCRIPT_VAI                           = 99, /* Vaii */
+      /** @stable ICU 3.6 */
+      USCRIPT_VISIBLE_SPEECH                = 100,/* Visp */
+      /** @stable ICU 3.6 */
+      USCRIPT_CUNEIFORM                     = 101,/* Xsux */
+      /** @stable ICU 3.6 */
+      USCRIPT_UNWRITTEN_LANGUAGES           = 102,/* Zxxx */
+      /** @stable ICU 3.6 */
+      USCRIPT_UNKNOWN                       = 103,/* Zzzz */ /* Unknown="Code for uncoded script", for unassigned code points */
+
+      /** @stable ICU 3.8 */
+      USCRIPT_CARIAN                        = 104,/* Cari */
+      /** @stable ICU 3.8 */
+      USCRIPT_JAPANESE                      = 105,/* Jpan */
+      /** @stable ICU 3.8 */
+      USCRIPT_LANNA                         = 106,/* Lana */
+      /** @stable ICU 3.8 */
+      USCRIPT_LYCIAN                        = 107,/* Lyci */
+      /** @stable ICU 3.8 */
+      USCRIPT_LYDIAN                        = 108,/* Lydi */
+      /** @stable ICU 3.8 */
+      USCRIPT_OL_CHIKI                      = 109,/* Olck */
+      /** @stable ICU 3.8 */
+      USCRIPT_REJANG                        = 110,/* Rjng */
+      /** @stable ICU 3.8 */
+      USCRIPT_SAURASHTRA                    = 111,/* Saur */
+      /** Sutton SignWriting @stable ICU 3.8 */
+      USCRIPT_SIGN_WRITING                  = 112,/* Sgnw */
+      /** @stable ICU 3.8 */
+      USCRIPT_SUNDANESE                     = 113,/* Sund */
+      /** @stable ICU 3.8 */
+      USCRIPT_MOON                          = 114,/* Moon */
+      /** @stable ICU 3.8 */
+      USCRIPT_MEITEI_MAYEK                  = 115,/* Mtei */
+
+      /** @stable ICU 4.0 */
+      USCRIPT_IMPERIAL_ARAMAIC              = 116,/* Armi */
+      /** @stable ICU 4.0 */
+      USCRIPT_AVESTAN                       = 117,/* Avst */
+      /** @stable ICU 4.0 */
+      USCRIPT_CHAKMA                        = 118,/* Cakm */
+      /** @stable ICU 4.0 */
+      USCRIPT_KOREAN                        = 119,/* Kore */
+      /** @stable ICU 4.0 */
+      USCRIPT_KAITHI                        = 120,/* Kthi */
+      /** @stable ICU 4.0 */
+      USCRIPT_MANICHAEAN                    = 121,/* Mani */
+      /** @stable ICU 4.0 */
+      USCRIPT_INSCRIPTIONAL_PAHLAVI         = 122,/* Phli */
+      /** @stable ICU 4.0 */
+      USCRIPT_PSALTER_PAHLAVI               = 123,/* Phlp */
+      /** @stable ICU 4.0 */
+      USCRIPT_BOOK_PAHLAVI                  = 124,/* Phlv */
+      /** @stable ICU 4.0 */
+      USCRIPT_INSCRIPTIONAL_PARTHIAN        = 125,/* Prti */
+      /** @stable ICU 4.0 */
+      USCRIPT_SAMARITAN                     = 126,/* Samr */
+      /** @stable ICU 4.0 */
+      USCRIPT_TAI_VIET                      = 127,/* Tavt */
+      /** @stable ICU 4.0 */
+      USCRIPT_MATHEMATICAL_NOTATION         = 128,/* Zmth */
+      /** @stable ICU 4.0 */
+      USCRIPT_SYMBOLS                       = 129,/* Zsym */
+
+      /** @stable ICU 4.4 */
+      USCRIPT_BAMUM                         = 130,/* Bamu */
+      /** @stable ICU 4.4 */
+      USCRIPT_LISU                          = 131,/* Lisu */
+      /** @stable ICU 4.4 */
+      USCRIPT_NAKHI_GEBA                    = 132,/* Nkgb */
+      /** @stable ICU 4.4 */
+      USCRIPT_OLD_SOUTH_ARABIAN             = 133,/* Sarb */
+
+      /** @stable ICU 4.6 */
+      USCRIPT_BASSA_VAH                     = 134,/* Bass */
+      /** @stable ICU 54 */
+      USCRIPT_DUPLOYAN                      = 135,/* Dupl */
+#ifndef U_HIDE_DEPRECATED_API
+      /** @deprecated ICU 54 Typo, use USCRIPT_DUPLOYAN */
+      USCRIPT_DUPLOYAN_SHORTAND             = USCRIPT_DUPLOYAN,
+#endif  /* U_HIDE_DEPRECATED_API */
+      /** @stable ICU 4.6 */
+      USCRIPT_ELBASAN                       = 136,/* Elba */
+      /** @stable ICU 4.6 */
+      USCRIPT_GRANTHA                       = 137,/* Gran */
+      /** @stable ICU 4.6 */
+      USCRIPT_KPELLE                        = 138,/* Kpel */
+      /** @stable ICU 4.6 */
+      USCRIPT_LOMA                          = 139,/* Loma */
+      /** Mende Kikakui @stable ICU 4.6 */
+      USCRIPT_MENDE                         = 140,/* Mend */
+      /** @stable ICU 4.6 */
+      USCRIPT_MEROITIC_CURSIVE              = 141,/* Merc */
+      /** @stable ICU 4.6 */
+      USCRIPT_OLD_NORTH_ARABIAN             = 142,/* Narb */
+      /** @stable ICU 4.6 */
+      USCRIPT_NABATAEAN                     = 143,/* Nbat */
+      /** @stable ICU 4.6 */
+      USCRIPT_PALMYRENE                     = 144,/* Palm */
+      /** @stable ICU 54 */
+      USCRIPT_KHUDAWADI                     = 145,/* Sind */
+      /** @stable ICU 4.6 */
+      USCRIPT_SINDHI                        = USCRIPT_KHUDAWADI,
+      /** @stable ICU 4.6 */
+      USCRIPT_WARANG_CITI                   = 146,/* Wara */
+
+      /** @stable ICU 4.8 */
+      USCRIPT_AFAKA                         = 147,/* Afak */
+      /** @stable ICU 4.8 */
+      USCRIPT_JURCHEN                       = 148,/* Jurc */
+      /** @stable ICU 4.8 */
+      USCRIPT_MRO                           = 149,/* Mroo */
+      /** @stable ICU 4.8 */
+      USCRIPT_NUSHU                         = 150,/* Nshu */
+      /** @stable ICU 4.8 */
+      USCRIPT_SHARADA                       = 151,/* Shrd */
+      /** @stable ICU 4.8 */
+      USCRIPT_SORA_SOMPENG                  = 152,/* Sora */
+      /** @stable ICU 4.8 */
+      USCRIPT_TAKRI                         = 153,/* Takr */
+      /** @stable ICU 4.8 */
+      USCRIPT_TANGUT                        = 154,/* Tang */
+      /** @stable ICU 4.8 */
+      USCRIPT_WOLEAI                        = 155,/* Wole */
+
+      /** @stable ICU 49 */
+      USCRIPT_ANATOLIAN_HIEROGLYPHS         = 156,/* Hluw */
+      /** @stable ICU 49 */
+      USCRIPT_KHOJKI                        = 157,/* Khoj */
+      /** @stable ICU 49 */
+      USCRIPT_TIRHUTA                       = 158,/* Tirh */
+
+      /** @stable ICU 52 */
+      USCRIPT_CAUCASIAN_ALBANIAN            = 159,/* Aghb */
+      /** @stable ICU 52 */
+      USCRIPT_MAHAJANI                      = 160,/* Mahj */
+
+      /** @stable ICU 54 */
+      USCRIPT_AHOM                          = 161,/* Ahom */
+      /** @stable ICU 54 */
+      USCRIPT_HATRAN                        = 162,/* Hatr */
+      /** @stable ICU 54 */
+      USCRIPT_MODI                          = 163,/* Modi */
+      /** @stable ICU 54 */
+      USCRIPT_MULTANI                       = 164,/* Mult */
+      /** @stable ICU 54 */
+      USCRIPT_PAU_CIN_HAU                   = 165,/* Pauc */
+      /** @stable ICU 54 */
+      USCRIPT_SIDDHAM                       = 166,/* Sidd */
+
+      /** @stable ICU 58 */
+      USCRIPT_ADLAM                         = 167,/* Adlm */
+      /** @stable ICU 58 */
+      USCRIPT_BHAIKSUKI                     = 168,/* Bhks */
+      /** @stable ICU 58 */
+      USCRIPT_MARCHEN                       = 169,/* Marc */
+      /** @stable ICU 58 */
+      USCRIPT_NEWA                          = 170,/* Newa */
+      /** @stable ICU 58 */
+      USCRIPT_OSAGE                         = 171,/* Osge */
+
+      /** @stable ICU 58 */
+      USCRIPT_HAN_WITH_BOPOMOFO             = 172,/* Hanb */
+      /** @stable ICU 58 */
+      USCRIPT_JAMO                          = 173,/* Jamo */
+      /** @stable ICU 58 */
+      USCRIPT_SYMBOLS_EMOJI                 = 174,/* Zsye */
+
+      /** @stable ICU 60 */
+      USCRIPT_MASARAM_GONDI                 = 175,/* Gonm */
+      /** @stable ICU 60 */
+      USCRIPT_SOYOMBO                       = 176,/* Soyo */
+      /** @stable ICU 60 */
+      USCRIPT_ZANABAZAR_SQUARE              = 177,/* Zanb */
+
+      /** @stable ICU 62 */
+      USCRIPT_DOGRA                         = 178,/* Dogr */
+      /** @stable ICU 62 */
+      USCRIPT_GUNJALA_GONDI                 = 179,/* Gong */
+      /** @stable ICU 62 */
+      USCRIPT_MAKASAR                       = 180,/* Maka */
+      /** @stable ICU 62 */
+      USCRIPT_MEDEFAIDRIN                   = 181,/* Medf */
+      /** @stable ICU 62 */
+      USCRIPT_HANIFI_ROHINGYA               = 182,/* Rohg */
+      /** @stable ICU 62 */
+      USCRIPT_SOGDIAN                       = 183,/* Sogd */
+      /** @stable ICU 62 */
+      USCRIPT_OLD_SOGDIAN                   = 184,/* Sogo */
+
+      /** @stable ICU 64 */
+      USCRIPT_ELYMAIC                       = 185,/* Elym */
+      /** @stable ICU 64 */
+      USCRIPT_NYIAKENG_PUACHUE_HMONG        = 186,/* Hmnp */
+      /** @stable ICU 64 */
+      USCRIPT_NANDINAGARI                   = 187,/* Nand */
+      /** @stable ICU 64 */
+      USCRIPT_WANCHO                        = 188,/* Wcho */
+
+      /** @stable ICU 66 */
+      USCRIPT_CHORASMIAN                    = 189,/* Chrs */
+      /** @stable ICU 66 */
+      USCRIPT_DIVES_AKURU                   = 190,/* Diak */
+      /** @stable ICU 66 */
+      USCRIPT_KHITAN_SMALL_SCRIPT           = 191,/* Kits */
+      /** @stable ICU 66 */
+      USCRIPT_YEZIDI                        = 192,/* Yezi */
+
+      /** @stable ICU 70 */
+      USCRIPT_CYPRO_MINOAN                  = 193,/* Cpmn */
+      /** @stable ICU 70 */
+      USCRIPT_OLD_UYGHUR                    = 194,/* Ougr */
+      /** @stable ICU 70 */
+      USCRIPT_TANGSA                        = 195,/* Tnsa */
+      /** @stable ICU 70 */
+      USCRIPT_TOTO                          = 196,/* Toto */
+      /** @stable ICU 70 */
+      USCRIPT_VITHKUQI                      = 197,/* Vith */
+
+      /** @stable ICU 72 */
+      USCRIPT_KAWI                          = 198,/* Kawi */
+      /** @stable ICU 72 */
+      USCRIPT_NAG_MUNDARI                   = 199,/* Nagm */
+
+      /** @stable ICU 75 */
+      USCRIPT_ARABIC_NASTALIQ               = 200, /* Aran */
+
+      /** @stable ICU 76 */
+      USCRIPT_GARAY                         = 201, /* Gara */
+      /** @stable ICU 76 */
+      USCRIPT_GURUNG_KHEMA                  = 202, /* Gukh */
+      /** @stable ICU 76 */
+      USCRIPT_KIRAT_RAI                     = 203, /* Krai */
+      /** @stable ICU 76 */
+      USCRIPT_OL_ONAL                       = 204, /* Onao */
+      /** @stable ICU 76 */
+      USCRIPT_SUNUWAR                       = 205, /* Sunu */
+      /** @stable ICU 76 */
+      USCRIPT_TODHRI                        = 206, /* Todr */
+      /** @stable ICU 76 */
+      USCRIPT_TULU_TIGALARI                 = 207, /* Tutg */
+
+#ifndef U_HIDE_DEPRECATED_API
+    /**
+     * One more than the highest normal UScriptCode value.
+     * The highest value is available via u_getIntPropertyMaxValue(UCHAR_SCRIPT).
+     *
+     * @deprecated ICU 58 The numeric value may change over time, see ICU ticket #12420.
+     */
+    USCRIPT_CODE_LIMIT    = 208
+#endif  // U_HIDE_DEPRECATED_API
+} UScriptCode;
+
+/**
+ * Gets the script codes associated with the given locale or ISO 15924 abbreviation or name.
+ * Fills in USCRIPT_MALAYALAM given "Malayam" OR "Mlym".
+ * Fills in USCRIPT_LATIN given "en" OR "en_US"
+ * If the required capacity is greater than the capacity of the destination buffer,
+ * then the error code is set to U_BUFFER_OVERFLOW_ERROR and the required capacity is returned.
+ *
+ * <p>Note: To search by short or long script alias only, use
+ * u_getPropertyValueEnum(UCHAR_SCRIPT, alias) instead.  That does
+ * a fast lookup with no access of the locale data.
+ *
+ * @param nameOrAbbrOrLocale name of the script, as given in
+ * PropertyValueAliases.txt, or ISO 15924 code or locale
+ * @param fillIn the UScriptCode buffer to fill in the script code
+ * @param capacity the capacity (size) of UScriptCode buffer passed in.
+ * @param err the error status code.
+ * @return The number of script codes filled in the buffer passed in
+ * @stable ICU 2.4
+ */
+U_CAPI int32_t  U_EXPORT2
+uscript_getCode(const char* nameOrAbbrOrLocale,UScriptCode* fillIn,int32_t capacity,UErrorCode *err);
+
+/**
+ * Returns the long Unicode script name, if there is one.
+ * Otherwise returns the 4-letter ISO 15924 script code.
+ * Returns "Malayam" given USCRIPT_MALAYALAM.
+ *
+ * @param scriptCode UScriptCode enum
+ * @return long script name as given in PropertyValueAliases.txt, or the 4-letter code,
+ * or NULL if scriptCode is invalid
+ * @stable ICU 2.4
+ */
+U_CAPI const char*  U_EXPORT2
+uscript_getName(UScriptCode scriptCode);
+
+/**
+ * Returns the 4-letter ISO 15924 script code,
+ * which is the same as the short Unicode script name if Unicode has names for the script.
+ * Returns "Mlym" given USCRIPT_MALAYALAM.
+ *
+ * @param scriptCode UScriptCode enum
+ * @return short script name (4-letter code), or NULL if scriptCode is invalid
+ * @stable ICU 2.4
+ */
+U_CAPI const char*  U_EXPORT2
+uscript_getShortName(UScriptCode scriptCode);
+
+/**
+ * Gets the script code associated with the given codepoint.
+ * Returns USCRIPT_MALAYALAM given 0x0D02
+ * @param codepoint UChar32 codepoint
+ * @param err the error status code.
+ * @return The UScriptCode, or 0 if codepoint is invalid
+ * @stable ICU 2.4
+ */
+U_CAPI UScriptCode  U_EXPORT2
+uscript_getScript(UChar32 codepoint, UErrorCode *err);
+
+/**
+ * Do the Script_Extensions of code point c contain script sc?
+ * If c does not have explicit Script_Extensions, then this tests whether
+ * c has the Script property value sc.
+ *
+ * Some characters are commonly used in multiple scripts.
+ * For more information, see UAX #24: http://www.unicode.org/reports/tr24/.
+ * @param c code point
+ * @param sc script code
+ * @return true if sc is in Script_Extensions(c)
+ * @stable ICU 49
+ */
+U_CAPI UBool U_EXPORT2
+uscript_hasScript(UChar32 c, UScriptCode sc);
+
+/**
+ * Writes code point c's Script_Extensions as a list of UScriptCode values
+ * to the output scripts array and returns the number of script codes.
+ * - If c does have Script_Extensions, then the Script property value
+ *   (normally Common or Inherited) is not included.
+ * - If c does not have Script_Extensions, then the one Script code is written to the output array.
+ * - If c is not a valid code point, then the one USCRIPT_UNKNOWN code is written.
+ * In other words, if the return value is 1,
+ * then the output array contains exactly c's single Script code.
+ * If the return value is n>=2, then the output array contains c's n Script_Extensions script codes.
+ *
+ * Some characters are commonly used in multiple scripts.
+ * For more information, see UAX #24: http://www.unicode.org/reports/tr24/.
+ *
+ * If there are more than capacity script codes to be written, then
+ * U_BUFFER_OVERFLOW_ERROR is set and the number of Script_Extensions is returned.
+ * (Usual ICU buffer handling behavior.)
+ *
+ * @param c code point
+ * @param scripts output script code array
+ * @param capacity capacity of the scripts array
+ * @param errorCode Standard ICU error code. Its input value must
+ *                  pass the U_SUCCESS() test, or else the function returns
+ *                  immediately. Check for U_FAILURE() on output or use with
+ *                  function chaining. (See User Guide for details.)
+ * @return number of script codes in c's Script_Extensions, or 1 for the single Script value,
+ *         written to scripts unless U_BUFFER_OVERFLOW_ERROR indicates insufficient capacity
+ * @stable ICU 49
+ */
+U_CAPI int32_t U_EXPORT2
+uscript_getScriptExtensions(UChar32 c,
+                            UScriptCode *scripts, int32_t capacity,
+                            UErrorCode *errorCode);
+
+/**
+ * Script usage constants.
+ * See UAX #31 Unicode Identifier and Pattern Syntax.
+ * http://www.unicode.org/reports/tr31/#Table_Candidate_Characters_for_Exclusion_from_Identifiers
+ *
+ * @stable ICU 51
+ */
+typedef enum UScriptUsage {
+    /** Not encoded in Unicode. @stable ICU 51 */
+    USCRIPT_USAGE_NOT_ENCODED,
+    /** Unknown script usage. @stable ICU 51 */
+    USCRIPT_USAGE_UNKNOWN,
+    /** Candidate for Exclusion from Identifiers. @stable ICU 51 */
+    USCRIPT_USAGE_EXCLUDED,
+    /** Limited Use script. @stable ICU 51 */
+    USCRIPT_USAGE_LIMITED_USE,
+    /** Aspirational Use script. @stable ICU 51 */
+    USCRIPT_USAGE_ASPIRATIONAL,
+    /** Recommended script. @stable ICU 51 */
+    USCRIPT_USAGE_RECOMMENDED
+} UScriptUsage;
+
+/**
+ * Writes the script sample character string.
+ * This string normally consists of one code point but might be longer.
+ * The string is empty if the script is not encoded.
+ *
+ * @param script script code
+ * @param dest output string array
+ * @param capacity number of UChars in the dest array
+ * @param pErrorCode standard ICU in/out error code, must pass U_SUCCESS() on input
+ * @return the string length, even if U_BUFFER_OVERFLOW_ERROR
+ * @stable ICU 51
+ */
+U_CAPI int32_t U_EXPORT2
+uscript_getSampleString(UScriptCode script, UChar *dest, int32_t capacity, UErrorCode *pErrorCode);
+
+#if U_SHOW_CPLUSPLUS_API
+
+U_NAMESPACE_BEGIN
+class UnicodeString;
+U_NAMESPACE_END
+
+/**
+ * Returns the script sample character string.
+ * This string normally consists of one code point but might be longer.
+ * The string is empty if the script is not encoded.
+ *
+ * @param script script code
+ * @return the sample character string
+ * @stable ICU 51
+ */
+U_COMMON_API icu::UnicodeString U_EXPORT2
+uscript_getSampleUnicodeString(UScriptCode script);
+
+#endif
+
+/**
+ * Returns the script usage according to UAX #31 Unicode Identifier and Pattern Syntax.
+ * Returns USCRIPT_USAGE_NOT_ENCODED if the script is not encoded in Unicode.
+ *
+ * @param script script code
+ * @return script usage
+ * @see UScriptUsage
+ * @stable ICU 51
+ */
+U_CAPI UScriptUsage U_EXPORT2
+uscript_getUsage(UScriptCode script);
+
+/**
+ * Returns true if the script is written right-to-left.
+ * For example, Arab and Hebr.
+ *
+ * @param script script code
+ * @return true if the script is right-to-left
+ * @stable ICU 51
+ */
+U_CAPI UBool U_EXPORT2
+uscript_isRightToLeft(UScriptCode script);
+
+/**
+ * Returns true if the script allows line breaks between letters (excluding hyphenation).
+ * Such a script typically requires dictionary-based line breaking.
+ * For example, Hani and Thai.
+ *
+ * @param script script code
+ * @return true if the script allows line breaks between letters
+ * @stable ICU 51
+ */
+U_CAPI UBool U_EXPORT2
+uscript_breaksBetweenLetters(UScriptCode script);
+
+/**
+ * Returns true if in modern (or most recent) usage of the script case distinctions are customary.
+ * For example, Latn and Cyrl.
+ *
+ * @param script script code
+ * @return true if the script is cased
+ * @stable ICU 51
+ */
+U_CAPI UBool U_EXPORT2
+uscript_isCased(UScriptCode script);
+
+#endif

thirdparty/icu4c/common/unicode/usetiter.h

@@ -0,0 +1,323 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+/*
+**********************************************************************
+* Copyright (c) 2002-2014, International Business Machines
+* Corporation and others.  All Rights Reserved.
+**********************************************************************
+*/
+#ifndef USETITER_H
+#define USETITER_H
+
+#include "unicode/utypes.h"
+
+#if U_SHOW_CPLUSPLUS_API
+
+#include "unicode/uobject.h"
+#include "unicode/unistr.h"
+
+/**
+ * \file 
+ * \brief C++ API: UnicodeSetIterator iterates over the contents of a UnicodeSet.
+ */
+
+U_NAMESPACE_BEGIN
+
+class UnicodeSet;
+class UnicodeString;
+
+/**
+ *
+ * UnicodeSetIterator iterates over the contents of a UnicodeSet.  It
+ * iterates over either code points or code point ranges.  After all
+ * code points or ranges have been returned, it returns the
+ * multicharacter strings of the UnicodeSet, if any.
+ *
+ * This class is not intended for public subclassing.
+ *
+ * <p>To iterate over code points and strings, use a loop like this:
+ * <pre>
+ * UnicodeSetIterator it(set);
+ * while (it.next()) {
+ *     processItem(it.getString());
+ * }
+ * </pre>
+ * <p>Each item in the set is accessed as a string.  Set elements
+ *    consisting of single code points are returned as strings containing
+ *    just the one code point.
+ *
+ * <p>To iterate over code point ranges, instead of individual code points,
+ *    use a loop like this:
+ * <pre>
+ * UnicodeSetIterator it(set);
+ * while (it.nextRange()) {
+ *   if (it.isString()) {
+ *     processString(it.getString());
+ *   } else {
+ *     processCodepointRange(it.getCodepoint(), it.getCodepointEnd());
+ *   }
+ * }
+ * </pre>
+ *
+ * To iterate over only the strings, start with <code>skipToStrings()</code>.
+ *
+ * @author M. Davis
+ * @stable ICU 2.4
+ */
+class U_COMMON_API UnicodeSetIterator final : public UObject {
+    /**
+     * Value of <tt>codepoint</tt> if the iterator points to a string.
+     * If <tt>codepoint == IS_STRING</tt>, then examine
+     * <tt>string</tt> for the current iteration result.
+     */
+    enum { IS_STRING = -1 };
+
+    /**
+     * Current code point, or the special value <tt>IS_STRING</tt>, if
+     * the iterator points to a string.
+     */
+    UChar32 codepoint;
+
+    /**
+     * When iterating over ranges using <tt>nextRange()</tt>,
+     * <tt>codepointEnd</tt> contains the inclusive end of the
+     * iteration range, if <tt>codepoint != IS_STRING</tt>.  If
+     * iterating over code points using <tt>next()</tt>, or if
+     * <tt>codepoint == IS_STRING</tt>, then the value of
+     * <tt>codepointEnd</tt> is undefined.
+     */
+    UChar32 codepointEnd;
+
+    /**
+     * If <tt>codepoint == IS_STRING</tt>, then <tt>string</tt> points
+     * to the current string.  If <tt>codepoint != IS_STRING</tt>, the
+     * value of <tt>string</tt> is undefined.
+     */
+    const UnicodeString* string;
+
+ public:
+
+    /**
+     * Create an iterator over the given set.  The iterator is valid
+     * only so long as <tt>set</tt> is valid.
+     * @param set set to iterate over
+     * @stable ICU 2.4
+     */
+    UnicodeSetIterator(const UnicodeSet& set);
+
+    /**
+     * Create an iterator over nothing.  <tt>next()</tt> and
+     * <tt>nextRange()</tt> return false. This is a convenience
+     * constructor allowing the target to be set later.
+     * @stable ICU 2.4
+     */
+    UnicodeSetIterator();
+
+    /**
+     * Destructor.
+     * @stable ICU 2.4
+     */
+    virtual ~UnicodeSetIterator();
+
+    /**
+     * Returns true if the current element is a string.  If so, the
+     * caller can retrieve it with <tt>getString()</tt>.  If this
+     * method returns false, the current element is a code point or
+     * code point range, depending on whether <tt>next()</tt> or
+     * <tt>nextRange()</tt> was called.
+     * Elements of types string and codepoint can both be retrieved
+     * with the function <tt>getString()</tt>.
+     * Elements of type codepoint can also be retrieved with
+     * <tt>getCodepoint()</tt>.
+     * For ranges, <tt>getCodepoint()</tt> returns the starting codepoint
+     * of the range, and <tt>getCodepointEnd()</tt> returns the end
+     * of the range.
+     * @stable ICU 2.4
+     */
+    inline UBool isString() const;
+
+    /**
+     * Returns the current code point, if <tt>isString()</tt> returned
+     * false.  Otherwise returns an undefined result.
+     * @stable ICU 2.4
+     */
+    inline UChar32 getCodepoint() const;
+
+    /**
+     * Returns the end of the current code point range, if
+     * <tt>isString()</tt> returned false and <tt>nextRange()</tt> was
+     * called.  Otherwise returns an undefined result.
+     * @stable ICU 2.4
+     */
+    inline UChar32 getCodepointEnd() const;
+
+    /**
+     * Returns the current string, if <tt>isString()</tt> returned
+     * true.  If the current iteration item is a code point, a UnicodeString
+     * containing that single code point is returned.
+     *
+     * Ownership of the returned string remains with the iterator.
+     * The string is guaranteed to remain valid only until the iterator is
+     *   advanced to the next item, or until the iterator is deleted.
+     * 
+     * @stable ICU 2.4
+     */
+    const UnicodeString& getString();
+
+    /**
+     * Skips over the remaining code points/ranges, if any.
+     * A following call to next() or nextRange() will yield a string, if there is one.
+     * No-op if next() would return false, or if it would yield a string anyway.
+     *
+     * @return *this
+     * @stable ICU 70
+     * @see UnicodeSet#strings()
+     */
+    inline UnicodeSetIterator &skipToStrings() {
+        // Finish code point/range iteration.
+        range = endRange;
+        endElement = -1;
+        nextElement = 0;
+        return *this;
+    }
+
+    /**
+     * Advances the iteration position to the next element in the set, 
+     * which can be either a single code point or a string.  
+     * If there are no more elements in the set, return false.
+     *
+     * <p>
+     * If <tt>isString() == true</tt>, the value is a
+     * string, otherwise the value is a
+     * single code point.  Elements of either type can be retrieved
+     * with the function <tt>getString()</tt>, while elements of
+     * consisting of a single code point can be retrieved with
+     * <tt>getCodepoint()</tt>
+     *
+     * <p>The order of iteration is all code points in sorted order,
+     * followed by all strings sorted order.    Do not mix
+     * calls to <tt>next()</tt> and <tt>nextRange()</tt> without
+     * calling <tt>reset()</tt> between them.  The results of doing so
+     * are undefined.
+     *
+     * @return true if there was another element in the set.
+     * @stable ICU 2.4
+     */
+    UBool next();
+
+    /**
+     * Returns the next element in the set, either a code point range
+     * or a string.  If there are no more elements in the set, return
+     * false.  If <tt>isString() == true</tt>, the value is a
+     * string and can be accessed with <tt>getString()</tt>.  Otherwise the value is a
+     * range of one or more code points from <tt>getCodepoint()</tt> to
+     * <tt>getCodepointeEnd()</tt> inclusive.
+     *
+     * <p>The order of iteration is all code points ranges in sorted
+     * order, followed by all strings sorted order.  Ranges are
+     * disjoint and non-contiguous.  The value returned from <tt>getString()</tt>
+     * is undefined unless <tt>isString() == true</tt>.  Do not mix calls to
+     * <tt>next()</tt> and <tt>nextRange()</tt> without calling
+     * <tt>reset()</tt> between them.  The results of doing so are
+     * undefined.
+     *
+     * @return true if there was another element in the set.
+     * @stable ICU 2.4
+     */
+    UBool nextRange();
+
+    /**
+     * Sets this iterator to visit the elements of the given set and
+     * resets it to the start of that set.  The iterator is valid only
+     * so long as <tt>set</tt> is valid.
+     * @param set the set to iterate over.
+     * @stable ICU 2.4
+     */
+    void reset(const UnicodeSet& set);
+
+    /**
+     * Resets this iterator to the start of the set.
+     * @stable ICU 2.4
+     */
+    void reset();
+
+    /**
+     * ICU "poor man's RTTI", returns a UClassID for this class.
+     *
+     * @stable ICU 2.4
+     */
+    static UClassID U_EXPORT2 getStaticClassID();
+
+    /**
+     * ICU "poor man's RTTI", returns a UClassID for the actual class.
+     *
+     * @stable ICU 2.4
+     */
+    virtual UClassID getDynamicClassID() const override;
+
+    // ======================= PRIVATES ===========================
+
+private:
+
+    // endElement and nextElements are really UChar32's, but we keep
+    // them as signed int32_t's so we can do comparisons with
+    // endElement set to -1.  Leave them as int32_t's.
+    /** The set
+     */
+    const UnicodeSet* set;
+    /** End range
+     */
+    int32_t endRange;
+    /** Range
+     */
+    int32_t range;
+    /** End element
+     */
+    int32_t endElement;
+    /** Next element
+     */
+    int32_t nextElement;
+    /** Next string
+     */
+    int32_t nextString;
+    /** String count
+     */
+    int32_t stringCount;
+
+    /**
+     *  Points to the string to use when the caller asks for a
+     *  string and the current iteration item is a code point, not a string.
+     */
+    UnicodeString *cpString;
+
+    /** Copy constructor. Disallowed.
+     */
+    UnicodeSetIterator(const UnicodeSetIterator&) = delete;
+
+    /** Assignment operator. Disallowed.
+     */
+    UnicodeSetIterator& operator=(const UnicodeSetIterator&) = delete;
+
+    /** Load range
+     */
+    void loadRange(int32_t range);
+};
+
+inline UBool UnicodeSetIterator::isString() const {
+    return codepoint < 0;
+}
+
+inline UChar32 UnicodeSetIterator::getCodepoint() const {
+    return codepoint;
+}
+
+inline UChar32 UnicodeSetIterator::getCodepointEnd() const {
+    return codepointEnd;
+}
+
+
+U_NAMESPACE_END
+
+#endif /* U_SHOW_CPLUSPLUS_API */
+
+#endif

thirdparty/icu4c/common/unicode/ushape.h

@@ -0,0 +1,476 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+/*
+******************************************************************************
+*
+*   Copyright (C) 2000-2012, International Business Machines
+*   Corporation and others.  All Rights Reserved.
+*
+******************************************************************************
+*   file name:  ushape.h
+*   encoding:   UTF-8
+*   tab size:   8 (not used)
+*   indentation:4
+*
+*   created on: 2000jun29
+*   created by: Markus W. Scherer
+*/
+
+#ifndef __USHAPE_H__
+#define __USHAPE_H__
+
+#include "unicode/utypes.h"
+
+/**
+ * \file
+ * \brief C API:  Arabic shaping
+ * 
+ */
+
+/**
+ * Shape Arabic text on a character basis.
+ *
+ * <p>This function performs basic operations for "shaping" Arabic text. It is most
+ * useful for use with legacy data formats and legacy display technology
+ * (simple terminals). All operations are performed on Unicode characters.</p>
+ *
+ * <p>Text-based shaping means that some character code points in the text are
+ * replaced by others depending on the context. It transforms one kind of text
+ * into another. In comparison, modern displays for Arabic text select
+ * appropriate, context-dependent font glyphs for each text element, which means
+ * that they transform text into a glyph vector.</p>
+ *
+ * <p>Text transformations are necessary when modern display technology is not
+ * available or when text needs to be transformed to or from legacy formats that
+ * use "shaped" characters. Since the Arabic script is cursive, connecting
+ * adjacent letters to each other, computers select images for each letter based
+ * on the surrounding letters. This usually results in four images per Arabic
+ * letter: initial, middle, final, and isolated forms. In Unicode, on the other
+ * hand, letters are normally stored abstract, and a display system is expected
+ * to select the necessary glyphs. (This makes searching and other text
+ * processing easier because the same letter has only one code.) It is possible
+ * to mimic this with text transformations because there are characters in
+ * Unicode that are rendered as letters with a specific shape
+ * (or cursive connectivity). They were included for interoperability with
+ * legacy systems and codepages, and for unsophisticated display systems.</p>
+ *
+ * <p>A second kind of text transformations is supported for Arabic digits:
+ * For compatibility with legacy codepages that only include European digits,
+ * it is possible to replace one set of digits by another, changing the
+ * character code points. These operations can be performed for either
+ * Arabic-Indic Digits (U+0660...U+0669) or Eastern (Extended) Arabic-Indic
+ * digits (U+06f0...U+06f9).</p>
+ *
+ * <p>Some replacements may result in more or fewer characters (code points).
+ * By default, this means that the destination buffer may receive text with a
+ * length different from the source length. Some legacy systems rely on the
+ * length of the text to be constant. They expect extra spaces to be added
+ * or consumed either next to the affected character or at the end of the
+ * text.</p>
+ *
+ * <p>For details about the available operations, see the description of the
+ * <code>U_SHAPE_...</code> options.</p>
+ *
+ * @param source The input text.
+ *
+ * @param sourceLength The number of UChars in <code>source</code>.
+ *
+ * @param dest The destination buffer that will receive the results of the
+ *             requested operations. It may be <code>NULL</code> only if
+ *             <code>destSize</code> is 0. The source and destination must not
+ *             overlap.
+ *
+ * @param destSize The size (capacity) of the destination buffer in UChars.
+ *                 If <code>destSize</code> is 0, then no output is produced,
+ *                 but the necessary buffer size is returned ("preflighting").
+ *
+ * @param options This is a 32-bit set of flags that specify the operations
+ *                that are performed on the input text. If no error occurs,
+ *                then the result will always be written to the destination
+ *                buffer.
+ *
+ * @param pErrorCode must be a valid pointer to an error code value,
+ *        which must not indicate a failure before the function call.
+ *
+ * @return The number of UChars written to the destination buffer.
+ *         If an error occurred, then no output was written, or it may be
+ *         incomplete. If <code>U_BUFFER_OVERFLOW_ERROR</code> is set, then
+ *         the return value indicates the necessary destination buffer size.
+ * @stable ICU 2.0
+ */
+U_CAPI int32_t U_EXPORT2
+u_shapeArabic(const UChar *source, int32_t sourceLength,
+              UChar *dest, int32_t destSize,
+              uint32_t options,
+              UErrorCode *pErrorCode);
+
+/**
+ * Memory option: allow the result to have a different length than the source.
+ * Affects: LamAlef options
+ * @stable ICU 2.0
+ */
+#define U_SHAPE_LENGTH_GROW_SHRINK              0
+
+/**
+ * Memory option: allow the result to have a different length than the source.
+ * Affects: LamAlef options
+ * This option is an alias to U_SHAPE_LENGTH_GROW_SHRINK
+ * @stable ICU 4.2
+ */
+#define U_SHAPE_LAMALEF_RESIZE                  0 
+
+/**
+ * Memory option: the result must have the same length as the source.
+ * If more room is necessary, then try to consume spaces next to modified characters.
+ * @stable ICU 2.0
+ */
+#define U_SHAPE_LENGTH_FIXED_SPACES_NEAR        1
+
+/**
+ * Memory option: the result must have the same length as the source.
+ * If more room is necessary, then try to consume spaces next to modified characters.
+ * Affects: LamAlef options
+ * This option is an alias to U_SHAPE_LENGTH_FIXED_SPACES_NEAR
+ * @stable ICU 4.2
+ */
+#define U_SHAPE_LAMALEF_NEAR                    1 
+
+/**
+ * Memory option: the result must have the same length as the source.
+ * If more room is necessary, then try to consume spaces at the end of the text.
+ * @stable ICU 2.0
+ */
+#define U_SHAPE_LENGTH_FIXED_SPACES_AT_END      2
+
+/**
+ * Memory option: the result must have the same length as the source.
+ * If more room is necessary, then try to consume spaces at the end of the text.
+ * Affects: LamAlef options
+ * This option is an alias to U_SHAPE_LENGTH_FIXED_SPACES_AT_END
+ * @stable ICU 4.2
+ */
+#define U_SHAPE_LAMALEF_END                     2 
+
+/**
+ * Memory option: the result must have the same length as the source.
+ * If more room is necessary, then try to consume spaces at the beginning of the text.
+ * @stable ICU 2.0
+ */
+#define U_SHAPE_LENGTH_FIXED_SPACES_AT_BEGINNING 3
+
+/**
+ * Memory option: the result must have the same length as the source.
+ * If more room is necessary, then try to consume spaces at the beginning of the text.
+ * Affects: LamAlef options
+ * This option is an alias to U_SHAPE_LENGTH_FIXED_SPACES_AT_BEGINNING
+ * @stable ICU 4.2
+ */
+#define U_SHAPE_LAMALEF_BEGIN                    3 
+
+
+/**
+ * Memory option: the result must have the same length as the source.
+ * Shaping Mode: For each LAMALEF character found, expand LAMALEF using space at end.
+ *               If there is no space at end, use spaces at beginning of the buffer. If there
+ *               is no space at beginning of the buffer, use spaces at the near (i.e. the space
+ *               after the LAMALEF character).
+ *               If there are no spaces found, an error U_NO_SPACE_AVAILABLE (as defined in utypes.h) 
+ *               will be set in pErrorCode
+ *
+ * Deshaping Mode: Perform the same function as the flag equals U_SHAPE_LAMALEF_END. 
+ * Affects: LamAlef options
+ * @stable ICU 4.2
+ */
+#define U_SHAPE_LAMALEF_AUTO                     0x10000 
+
+/** Bit mask for memory options. @stable ICU 2.0 */
+#define U_SHAPE_LENGTH_MASK                      0x10003 /* Changed old value 3 */
+
+
+/**
+ * Bit mask for LamAlef memory options.
+ * @stable ICU 4.2
+ */
+#define U_SHAPE_LAMALEF_MASK                     0x10003 /* updated */
+
+/** Direction indicator: the source is in logical (keyboard) order. @stable ICU 2.0 */
+#define U_SHAPE_TEXT_DIRECTION_LOGICAL          0
+
+/**
+ * Direction indicator:
+ * the source is in visual RTL order,
+ * the rightmost displayed character stored first.
+ * This option is an alias to U_SHAPE_TEXT_DIRECTION_LOGICAL
+ * @stable ICU 4.2
+ */
+#define U_SHAPE_TEXT_DIRECTION_VISUAL_RTL       0
+
+/**
+ * Direction indicator:
+ * the source is in visual LTR order,
+ * the leftmost displayed character stored first.
+ * @stable ICU 2.0
+ */
+#define U_SHAPE_TEXT_DIRECTION_VISUAL_LTR       4
+
+/** Bit mask for direction indicators. @stable ICU 2.0 */
+#define U_SHAPE_TEXT_DIRECTION_MASK             4
+
+
+/** Letter shaping option: do not perform letter shaping. @stable ICU 2.0 */
+#define U_SHAPE_LETTERS_NOOP                    0
+
+/** Letter shaping option: replace abstract letter characters by "shaped" ones. @stable ICU 2.0 */
+#define U_SHAPE_LETTERS_SHAPE                   8
+
+/** Letter shaping option: replace "shaped" letter characters by abstract ones. @stable ICU 2.0 */
+#define U_SHAPE_LETTERS_UNSHAPE                 0x10
+
+/**
+ * Letter shaping option: replace abstract letter characters by "shaped" ones.
+ * The only difference with U_SHAPE_LETTERS_SHAPE is that Tashkeel letters
+ * are always "shaped" into the isolated form instead of the medial form
+ * (selecting code points from the Arabic Presentation Forms-B block).
+ * @stable ICU 2.0
+ */
+#define U_SHAPE_LETTERS_SHAPE_TASHKEEL_ISOLATED 0x18
+
+
+/** Bit mask for letter shaping options. @stable ICU 2.0 */
+#define U_SHAPE_LETTERS_MASK                        0x18
+
+
+/** Digit shaping option: do not perform digit shaping. @stable ICU 2.0 */
+#define U_SHAPE_DIGITS_NOOP                     0
+
+/**
+ * Digit shaping option:
+ * Replace European digits (U+0030...) by Arabic-Indic digits.
+ * @stable ICU 2.0
+ */
+#define U_SHAPE_DIGITS_EN2AN                    0x20
+
+/**
+ * Digit shaping option:
+ * Replace Arabic-Indic digits by European digits (U+0030...).
+ * @stable ICU 2.0
+ */
+#define U_SHAPE_DIGITS_AN2EN                    0x40
+
+/**
+ * Digit shaping option:
+ * Replace European digits (U+0030...) by Arabic-Indic digits if the most recent
+ * strongly directional character is an Arabic letter
+ * (<code>u_charDirection()</code> result <code>U_RIGHT_TO_LEFT_ARABIC</code> [AL]).<br>
+ * The direction of "preceding" depends on the direction indicator option.
+ * For the first characters, the preceding strongly directional character
+ * (initial state) is assumed to be not an Arabic letter
+ * (it is <code>U_LEFT_TO_RIGHT</code> [L] or <code>U_RIGHT_TO_LEFT</code> [R]).
+ * @stable ICU 2.0
+ */
+#define U_SHAPE_DIGITS_ALEN2AN_INIT_LR          0x60
+
+/**
+ * Digit shaping option:
+ * Replace European digits (U+0030...) by Arabic-Indic digits if the most recent
+ * strongly directional character is an Arabic letter
+ * (<code>u_charDirection()</code> result <code>U_RIGHT_TO_LEFT_ARABIC</code> [AL]).<br>
+ * The direction of "preceding" depends on the direction indicator option.
+ * For the first characters, the preceding strongly directional character
+ * (initial state) is assumed to be an Arabic letter.
+ * @stable ICU 2.0
+ */
+#define U_SHAPE_DIGITS_ALEN2AN_INIT_AL          0x80
+
+/** Not a valid option value. May be replaced by a new option. @stable ICU 2.0 */
+#define U_SHAPE_DIGITS_RESERVED                 0xa0
+
+/** Bit mask for digit shaping options. @stable ICU 2.0 */
+#define U_SHAPE_DIGITS_MASK                     0xe0
+
+
+/** Digit type option: Use Arabic-Indic digits (U+0660...U+0669). @stable ICU 2.0 */
+#define U_SHAPE_DIGIT_TYPE_AN                   0
+
+/** Digit type option: Use Eastern (Extended) Arabic-Indic digits (U+06f0...U+06f9). @stable ICU 2.0 */
+#define U_SHAPE_DIGIT_TYPE_AN_EXTENDED          0x100
+
+/** Not a valid option value. May be replaced by a new option. @stable ICU 2.0 */
+#define U_SHAPE_DIGIT_TYPE_RESERVED             0x200
+
+/** Bit mask for digit type options. @stable ICU 2.0 */
+#define U_SHAPE_DIGIT_TYPE_MASK                 0x300 /* I need to change this from 0x3f00 to 0x300 */
+
+/** 
+ * Tashkeel aggregation option:
+ * Replaces any combination of U+0651 with one of
+ * U+064C, U+064D, U+064E, U+064F, U+0650 with
+ * U+FC5E, U+FC5F, U+FC60, U+FC61, U+FC62 consecutively.
+ * @stable ICU 3.6
+ */
+#define U_SHAPE_AGGREGATE_TASHKEEL              0x4000
+/** Tashkeel aggregation option: do not aggregate tashkeels. @stable ICU 3.6 */
+#define U_SHAPE_AGGREGATE_TASHKEEL_NOOP         0
+/** Bit mask for tashkeel aggregation. @stable ICU 3.6 */
+#define U_SHAPE_AGGREGATE_TASHKEEL_MASK         0x4000
+
+/** 
+ * Presentation form option:
+ * Don't replace Arabic Presentation Forms-A and Arabic Presentation Forms-B
+ * characters with 0+06xx characters, before shaping.
+ * @stable ICU 3.6
+ */
+#define U_SHAPE_PRESERVE_PRESENTATION           0x8000
+/** Presentation form option: 
+ * Replace Arabic Presentation Forms-A and Arabic Presentationo Forms-B with 
+ * their unshaped correspondents in range 0+06xx, before shaping.
+ * @stable ICU 3.6 
+ */
+#define U_SHAPE_PRESERVE_PRESENTATION_NOOP      0
+/** Bit mask for preserve presentation form. @stable ICU 3.6 */
+#define U_SHAPE_PRESERVE_PRESENTATION_MASK      0x8000
+
+/* Seen Tail option */ 
+/**
+ * Memory option: the result must have the same length as the source.
+ * Shaping mode: The SEEN family character will expand into two characters using space near 
+ *               the SEEN family character(i.e. the space after the character).
+ *               If there are no spaces found, an error U_NO_SPACE_AVAILABLE (as defined in utypes.h) 
+ *               will be set in pErrorCode
+ *
+ * De-shaping mode: Any Seen character followed by Tail character will be
+ *                  replaced by one cell Seen and a space will replace the Tail.
+ * Affects: Seen options
+ * @stable ICU 4.2
+ */
+#define U_SHAPE_SEEN_TWOCELL_NEAR     0x200000
+
+/**
+ * Bit mask for Seen memory options. 
+ * @stable ICU 4.2
+ */
+#define U_SHAPE_SEEN_MASK             0x700000
+
+/* YehHamza option */ 
+/**
+ * Memory option: the result must have the same length as the source.
+ * Shaping mode: The YEHHAMZA character will expand into two characters using space near it 
+ *              (i.e. the space after the character
+ *               If there are no spaces found, an error U_NO_SPACE_AVAILABLE (as defined in utypes.h) 
+ *               will be set in pErrorCode
+ *
+ * De-shaping mode: Any Yeh (final or isolated) character followed by Hamza character will be
+ *                  replaced by one cell YehHamza and space will replace the Hamza.
+ * Affects: YehHamza options
+ * @stable ICU 4.2
+ */
+#define U_SHAPE_YEHHAMZA_TWOCELL_NEAR      0x1000000
+
+
+/**
+ * Bit mask for YehHamza memory options. 
+ * @stable ICU 4.2
+ */
+#define U_SHAPE_YEHHAMZA_MASK              0x3800000
+
+/* New Tashkeel options */ 
+/**
+ * Memory option: the result must have the same length as the source.
+ * Shaping mode: Tashkeel characters will be replaced by spaces. 
+ *               Spaces will be placed at beginning of the buffer
+ *
+ * De-shaping mode: N/A
+ * Affects: Tashkeel options
+ * @stable ICU 4.2
+ */
+#define U_SHAPE_TASHKEEL_BEGIN                      0x40000
+
+/**
+ * Memory option: the result must have the same length as the source.
+ * Shaping mode: Tashkeel characters will be replaced by spaces. 
+ *               Spaces will be placed at end of the buffer
+ *
+ * De-shaping mode: N/A
+ * Affects: Tashkeel options
+ * @stable ICU 4.2
+ */
+#define U_SHAPE_TASHKEEL_END                        0x60000
+
+/**
+ * Memory option: allow the result to have a different length than the source.
+ * Shaping mode: Tashkeel characters will be removed, buffer length will shrink. 
+ * De-shaping mode: N/A 
+ *
+ * Affect: Tashkeel options
+ * @stable ICU 4.2
+ */
+#define U_SHAPE_TASHKEEL_RESIZE                     0x80000
+
+/**
+ * Memory option: the result must have the same length as the source.
+ * Shaping mode: Tashkeel characters will be replaced by Tatweel if it is connected to adjacent
+ *               characters (i.e. shaped on Tatweel) or replaced by space if it is not connected.
+ *
+ * De-shaping mode: N/A
+ * Affects: YehHamza options
+ * @stable ICU 4.2
+ */
+#define U_SHAPE_TASHKEEL_REPLACE_BY_TATWEEL         0xC0000
+
+/** 
+ * Bit mask for Tashkeel replacement with Space or Tatweel memory options. 
+ * @stable ICU 4.2
+ */
+#define U_SHAPE_TASHKEEL_MASK                       0xE0000
+
+
+/* Space location Control options */ 
+/**
+ * This option affect the meaning of BEGIN and END options. if this option is not used the default
+ * for BEGIN and END will be as following: 
+ * The Default (for both Visual LTR, Visual RTL and Logical Text)
+ *           1. BEGIN always refers to the start address of physical memory.
+ *           2. END always refers to the end address of physical memory.
+ *
+ * If this option is used it will swap the meaning of BEGIN and END only for Visual LTR text. 
+ *
+ * The effect on BEGIN and END Memory Options will be as following:
+ *    A. BEGIN For Visual LTR text: This will be the beginning (right side) of the visual text(
+ *       corresponding to the physical memory address end for Visual LTR text, Same as END in 
+ *       default behavior)
+ *    B. BEGIN For Logical text: Same as BEGIN in default behavior. 
+ *    C. END For Visual LTR text: This will be the end (left side) of the visual text (corresponding
+ *       to the physical memory address beginning for Visual LTR text, Same as BEGIN in default behavior.
+ *    D. END For Logical text: Same as END in default behavior). 
+ * Affects: All LamAlef BEGIN, END and AUTO options.
+ * @stable ICU 4.2
+ */
+#define U_SHAPE_SPACES_RELATIVE_TO_TEXT_BEGIN_END 0x4000000
+
+/**
+ * Bit mask for swapping BEGIN and END for Visual LTR text 
+ * @stable ICU 4.2
+ */
+#define U_SHAPE_SPACES_RELATIVE_TO_TEXT_MASK      0x4000000
+
+/**
+ * If this option is used, shaping will use the new Unicode code point for TAIL (i.e. 0xFE73). 
+ * If this option is not specified (Default), old unofficial Unicode TAIL code point is used (i.e. 0x200B)
+ * De-shaping will not use this option as it will always search for both the new Unicode code point for the 
+ * TAIL (i.e. 0xFE73) or the old unofficial Unicode TAIL code point (i.e. 0x200B) and de-shape the
+ * Seen-Family letter accordingly.
+ *
+ * Shaping Mode: Only shaping.
+ * De-shaping Mode: N/A.
+ * Affects: All Seen options
+ * @stable ICU 4.8
+ */
+#define U_SHAPE_TAIL_NEW_UNICODE        0x8000000
+
+/**
+ * Bit mask for new Unicode Tail option 
+ * @stable ICU 4.8
+ */
+#define U_SHAPE_TAIL_TYPE_MASK          0x8000000
+
+#endif

thirdparty/icu4c/common/unicode/usprep.h

@@ -0,0 +1,274 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+/*
+ *******************************************************************************
+ *
+ *   Copyright (C) 2003-2014, International Business Machines
+ *   Corporation and others.  All Rights Reserved.
+ *
+ *******************************************************************************
+ *   file name:  usprep.h
+ *   encoding:   UTF-8
+ *   tab size:   8 (not used)
+ *   indentation:4
+ *
+ *   created on: 2003jul2
+ *   created by: Ram Viswanadha
+ */
+
+#ifndef __USPREP_H__
+#define __USPREP_H__
+
+/**
+ * \file 
+ * \brief C API: Implements the StringPrep algorithm.
+ */
+
+#include "unicode/utypes.h"
+
+#if U_SHOW_CPLUSPLUS_API
+#include "unicode/localpointer.h"
+#endif   // U_SHOW_CPLUSPLUS_API
+
+/**
+ *
+ * StringPrep API implements the StingPrep framework as described by RFC 3454.
+ * StringPrep prepares Unicode strings for use in network protocols.
+ * Profiles of StingPrep are set of rules and data according to with the
+ * Unicode Strings are prepared. Each profiles contains tables which describe
+ * how a code point should be treated. The tables are broadly classified into
+ * <ul>
+ *     <li> Unassigned Table: Contains code points that are unassigned 
+ *          in the Unicode Version supported by StringPrep. Currently 
+ *          RFC 3454 supports Unicode 3.2. </li>
+ *     <li> Prohibited Table: Contains code points that are prohibited from
+ *          the output of the StringPrep processing function. </li>
+ *     <li> Mapping Table: Contains code points that are deleted from the output or case mapped. </li>
+ * </ul>
+ * 
+ * The procedure for preparing Unicode strings:
+ * <ol>
+ *      <li> Map: For each character in the input, check if it has a mapping
+ *           and, if so, replace it with its mapping. </li>
+ *      <li> Normalize: Possibly normalize the result of step 1 using Unicode
+ *           normalization. </li>
+ *      <li> Prohibit: Check for any characters that are not allowed in the
+ *        output.  If any are found, return an error.</li>
+ *      <li> Check bidi: Possibly check for right-to-left characters, and if
+ *           any are found, make sure that the whole string satisfies the
+ *           requirements for bidirectional strings.  If the string does not
+ *           satisfy the requirements for bidirectional strings, return an
+ *           error.  </li>
+ * </ol>
+ * @author Ram Viswanadha
+ */
+#if !UCONFIG_NO_IDNA
+
+#include "unicode/parseerr.h"
+
+/**
+ * The StringPrep profile
+ * @stable ICU 2.8
+ */
+typedef struct UStringPrepProfile UStringPrepProfile;
+
+
+/** 
+ * Option to prohibit processing of unassigned code points in the input
+ * 
+ * @see  usprep_prepare
+ * @stable ICU 2.8
+ */
+#define USPREP_DEFAULT 0x0000
+
+/** 
+ * Option to allow processing of unassigned code points in the input
+ * 
+ * @see  usprep_prepare
+ * @stable ICU 2.8
+ */
+#define USPREP_ALLOW_UNASSIGNED 0x0001
+
+/**
+ * enums for the standard stringprep profile types
+ * supported by usprep_openByType.
+ * @see usprep_openByType
+ * @stable ICU 4.2
+ */
+typedef enum UStringPrepProfileType {
+    /**
+     * RFC3491 Nameprep
+     * @stable ICU 4.2
+     */
+    USPREP_RFC3491_NAMEPREP,
+    /**
+     * RFC3530 nfs4_cs_prep
+     * @stable ICU 4.2
+     */
+	USPREP_RFC3530_NFS4_CS_PREP,
+    /**
+     * RFC3530 nfs4_cs_prep with case insensitive option
+     * @stable ICU 4.2
+     */
+	USPREP_RFC3530_NFS4_CS_PREP_CI,
+    /**
+     * RFC3530 nfs4_cis_prep
+     * @stable ICU 4.2
+     */
+	USPREP_RFC3530_NFS4_CIS_PREP,
+    /**
+     * RFC3530 nfs4_mixed_prep for prefix
+     * @stable ICU 4.2
+     */
+	USPREP_RFC3530_NFS4_MIXED_PREP_PREFIX,
+    /**
+     * RFC3530 nfs4_mixed_prep for suffix
+     * @stable ICU 4.2
+     */
+	USPREP_RFC3530_NFS4_MIXED_PREP_SUFFIX,
+    /**
+     * RFC3722 iSCSI
+     * @stable ICU 4.2
+     */
+	USPREP_RFC3722_ISCSI,
+    /**
+     * RFC3920 XMPP Nodeprep
+     * @stable ICU 4.2
+     */
+	USPREP_RFC3920_NODEPREP,
+    /**
+     * RFC3920 XMPP Resourceprep
+     * @stable ICU 4.2
+     */
+	USPREP_RFC3920_RESOURCEPREP,
+    /**
+     * RFC4011 Policy MIB Stringprep
+     * @stable ICU 4.2
+     */
+	USPREP_RFC4011_MIB,
+    /**
+     * RFC4013 SASLprep
+     * @stable ICU 4.2
+     */
+    USPREP_RFC4013_SASLPREP,
+    /**
+     * RFC4505 trace
+     * @stable ICU 4.2
+     */
+	USPREP_RFC4505_TRACE,
+    /**
+     * RFC4518 LDAP
+     * @stable ICU 4.2
+     */
+	USPREP_RFC4518_LDAP,
+    /**
+     * RFC4518 LDAP for case ignore, numeric and stored prefix
+     * matching rules
+     * @stable ICU 4.2
+     */
+	USPREP_RFC4518_LDAP_CI
+} UStringPrepProfileType;
+
+/**
+ * Creates a StringPrep profile from the data file.
+ *
+ * @param path      string containing the full path pointing to the directory
+ *                  where the profile reside followed by the package name
+ *                  e.g. "/usr/resource/my_app/profiles/mydata" on a Unix system.
+ *                  if NULL, ICU default data files will be used.
+ * @param fileName  name of the profile file to be opened
+ * @param status    ICU error code in/out parameter. Must not be NULL.
+ *                  Must fulfill U_SUCCESS before the function call.
+ * @return Pointer to UStringPrepProfile that is opened. Should be closed by
+ * calling usprep_close()
+ * @see usprep_close()
+ * @stable ICU 2.8
+ */
+U_CAPI UStringPrepProfile* U_EXPORT2
+usprep_open(const char* path, 
+            const char* fileName,
+            UErrorCode* status);
+
+/**
+ * Creates a StringPrep profile for the specified profile type.
+ *
+ * @param type		The profile type
+ * @param status    ICU error code in/out parameter. Must not be NULL.
+ *                  Must fulfill U_SUCCESS before the function call.
+ * @return          Pointer to UStringPrepProfile that is opened. Should be closed by
+ *                  calling usprep_close()
+ * @see usprep_close()
+ * @stable ICU 4.2
+ */
+U_CAPI UStringPrepProfile* U_EXPORT2
+usprep_openByType(UStringPrepProfileType type,
+				  UErrorCode* status);
+
+/**
+ * Closes the profile
+ * @param profile The profile to close
+ * @stable ICU 2.8
+ */
+U_CAPI void U_EXPORT2
+usprep_close(UStringPrepProfile* profile);
+
+#if U_SHOW_CPLUSPLUS_API
+
+U_NAMESPACE_BEGIN
+
+/**
+ * \class LocalUStringPrepProfilePointer
+ * "Smart pointer" class, closes a UStringPrepProfile via usprep_close().
+ * For most methods see the LocalPointerBase base class.
+ *
+ * @see LocalPointerBase
+ * @see LocalPointer
+ * @stable ICU 4.4
+ */
+U_DEFINE_LOCAL_OPEN_POINTER(LocalUStringPrepProfilePointer, UStringPrepProfile, usprep_close);
+
+U_NAMESPACE_END
+
+#endif
+
+/**
+ * Prepare the input buffer for use in applications with the given profile. This operation maps, normalizes(NFKC),
+ * checks for prohibited and BiDi characters in the order defined by RFC 3454
+ * depending on the options specified in the profile.
+ *
+ * @param prep          The profile to use 
+ * @param src           Pointer to UChar buffer containing the string to prepare
+ * @param srcLength     Number of characters in the source string
+ * @param dest          Pointer to the destination buffer to receive the output
+ * @param destCapacity  The capacity of destination array
+ * @param options       A bit set of options:
+ *
+ *  - USPREP_DEFAULT            Prohibit processing of unassigned code points in the input
+ *
+ *  - USPREP_ALLOW_UNASSIGNED   Treat the unassigned code points are in the input 
+ *                              as normal Unicode code points.
+ *
+ * @param parseError        Pointer to UParseError struct to receive information on position 
+ *                          of error if an error is encountered. Can be NULL.
+ * @param status            ICU in/out error code parameter.
+ *                          U_INVALID_CHAR_FOUND if src contains
+ *                          unmatched single surrogates.
+ *                          U_INDEX_OUTOFBOUNDS_ERROR if src contains
+ *                          too many code points.
+ *                          U_BUFFER_OVERFLOW_ERROR if destCapacity is not enough
+ * @return The number of UChars in the destination buffer
+ * @stable ICU 2.8
+ */
+
+U_CAPI int32_t U_EXPORT2
+usprep_prepare(   const UStringPrepProfile* prep,
+                  const UChar* src, int32_t srcLength, 
+                  UChar* dest, int32_t destCapacity,
+                  int32_t options,
+                  UParseError* parseError,
+                  UErrorCode* status );
+
+
+#endif /* #if !UCONFIG_NO_IDNA */
+
+#endif

thirdparty/icu4c/common/unicode/ustringtrie.h

@@ -0,0 +1,97 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+/*
+*******************************************************************************
+*   Copyright (C) 2010-2012, International Business Machines
+*   Corporation and others.  All Rights Reserved.
+*******************************************************************************
+*   file name:  udicttrie.h
+*   encoding:   UTF-8
+*   tab size:   8 (not used)
+*   indentation:4
+*
+*   created on: 2010dec17
+*   created by: Markus W. Scherer
+*/
+
+#ifndef __USTRINGTRIE_H__
+#define __USTRINGTRIE_H__
+
+/**
+ * \file
+ * \brief C API: Helper definitions for dictionary trie APIs.
+ */
+
+#include "unicode/utypes.h"
+
+
+/**
+ * Return values for BytesTrie::next(), UCharsTrie::next() and similar methods.
+ * @see USTRINGTRIE_MATCHES
+ * @see USTRINGTRIE_HAS_VALUE
+ * @see USTRINGTRIE_HAS_NEXT
+ * @stable ICU 4.8
+ */
+enum UStringTrieResult {
+    /**
+     * The input unit(s) did not continue a matching string.
+     * Once current()/next() return USTRINGTRIE_NO_MATCH,
+     * all further calls to current()/next() will also return USTRINGTRIE_NO_MATCH,
+     * until the trie is reset to its original state or to a saved state.
+     * @stable ICU 4.8
+     */
+    USTRINGTRIE_NO_MATCH,
+    /**
+     * The input unit(s) continued a matching string
+     * but there is no value for the string so far.
+     * (It is a prefix of a longer string.)
+     * @stable ICU 4.8
+     */
+    USTRINGTRIE_NO_VALUE,
+    /**
+     * The input unit(s) continued a matching string
+     * and there is a value for the string so far.
+     * This value will be returned by getValue().
+     * No further input byte/unit can continue a matching string.
+     * @stable ICU 4.8
+     */
+    USTRINGTRIE_FINAL_VALUE,
+    /**
+     * The input unit(s) continued a matching string
+     * and there is a value for the string so far.
+     * This value will be returned by getValue().
+     * Another input byte/unit can continue a matching string.
+     * @stable ICU 4.8
+     */
+    USTRINGTRIE_INTERMEDIATE_VALUE
+};
+
+/**
+ * Same as (result!=USTRINGTRIE_NO_MATCH).
+ * @param result A result from BytesTrie::first(), UCharsTrie::next() etc.
+ * @return true if the input bytes/units so far are part of a matching string/byte sequence.
+ * @stable ICU 4.8
+ */
+#define USTRINGTRIE_MATCHES(result) ((result)!=USTRINGTRIE_NO_MATCH)
+
+/**
+ * Equivalent to (result==USTRINGTRIE_INTERMEDIATE_VALUE || result==USTRINGTRIE_FINAL_VALUE) but
+ * this macro evaluates result exactly once.
+ * @param result A result from BytesTrie::first(), UCharsTrie::next() etc.
+ * @return true if there is a value for the input bytes/units so far.
+ * @see BytesTrie::getValue
+ * @see UCharsTrie::getValue
+ * @stable ICU 4.8
+ */
+#define USTRINGTRIE_HAS_VALUE(result) ((result)>=USTRINGTRIE_FINAL_VALUE)
+
+/**
+ * Equivalent to (result==USTRINGTRIE_NO_VALUE || result==USTRINGTRIE_INTERMEDIATE_VALUE) but
+ * this macro evaluates result exactly once.
+ * @param result A result from BytesTrie::first(), UCharsTrie::next() etc.
+ * @return true if another input byte/unit can continue a matching string.
+ * @stable ICU 4.8
+ */
+#define USTRINGTRIE_HAS_NEXT(result) ((result)&1)
+
+#endif  /* __USTRINGTRIE_H__ */

thirdparty/icu4c/common/unicode/utf.h

@@ -0,0 +1,225 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+/*
+*******************************************************************************
+*
+*   Copyright (C) 1999-2011, International Business Machines
+*   Corporation and others.  All Rights Reserved.
+*
+*******************************************************************************
+*   file name:  utf.h
+*   encoding:   UTF-8
+*   tab size:   8 (not used)
+*   indentation:4
+*
+*   created on: 1999sep09
+*   created by: Markus W. Scherer
+*/
+
+/**
+ * \file
+ * \brief C API: Code point macros
+ *
+ * This file defines macros for checking whether a code point is
+ * a surrogate or a non-character etc.
+ *
+ * If U_NO_DEFAULT_INCLUDE_UTF_HEADERS is 0 then utf.h is included by utypes.h
+ * and itself includes utf8.h and utf16.h after some
+ * common definitions.
+ * If U_NO_DEFAULT_INCLUDE_UTF_HEADERS is 1 then each of these headers must be
+ * included explicitly if their definitions are used.
+ *
+ * utf8.h and utf16.h define macros for efficiently getting code points
+ * in and out of UTF-8/16 strings.
+ * utf16.h macros have "U16_" prefixes.
+ * utf8.h defines similar macros with "U8_" prefixes for UTF-8 string handling.
+ *
+ * ICU mostly processes 16-bit Unicode strings.
+ * Most of the time, such strings are well-formed UTF-16.
+ * Single, unpaired surrogates must be handled as well, and are treated in ICU
+ * like regular code points where possible.
+ * (Pairs of surrogate code points are indistinguishable from supplementary
+ * code points encoded as pairs of supplementary code units.)
+ *
+ * In fact, almost all Unicode code points in normal text (>99%)
+ * are on the BMP (<=U+ffff) and even <=U+d7ff.
+ * ICU functions handle supplementary code points (U+10000..U+10ffff)
+ * but are optimized for the much more frequently occurring BMP code points.
+ *
+ * umachine.h defines UChar to be an unsigned 16-bit integer.
+ * Since ICU 59, ICU uses char16_t in C++, UChar only in C,
+ * and defines UChar=char16_t by default. See the UChar API docs for details.
+ *
+ * UChar32 is defined to be a signed 32-bit integer (int32_t), large enough for a 21-bit
+ * Unicode code point (Unicode scalar value, 0..0x10ffff) and U_SENTINEL (-1).
+ * Before ICU 2.4, the definition of UChar32 was similarly platform-dependent as
+ * the definition of UChar. For details see the documentation for UChar32 itself.
+ *
+ * utf.h defines a small number of C macros for single Unicode code points.
+ * These are simple checks for surrogates and non-characters.
+ * For actual Unicode character properties see uchar.h.
+ *
+ * By default, string operations must be done with error checking in case
+ * a string is not well-formed UTF-16 or UTF-8.
+ *
+ * The U16_ macros detect if a surrogate code unit is unpaired
+ * (lead unit without trail unit or vice versa) and just return the unit itself
+ * as the code point.
+ *
+ * The U8_ macros detect illegal byte sequences and return a negative value.
+ * Starting with ICU 60, the observable length of a single illegal byte sequence
+ * skipped by one of these macros follows the Unicode 6+ recommendation
+ * which is consistent with the W3C Encoding Standard.
+ *
+ * There are ..._OR_FFFD versions of both U16_ and U8_ macros
+ * that return U+FFFD for illegal code unit sequences.
+ *
+ * The regular "safe" macros require that the initial, passed-in string index
+ * is within bounds. They only check the index when they read more than one
+ * code unit. This is usually done with code similar to the following loop:
+ * <pre>while(i<length) {
+ *   U16_NEXT(s, i, length, c);
+ *   // use c
+ * }</pre>
+ *
+ * When it is safe to assume that text is well-formed UTF-16
+ * (does not contain single, unpaired surrogates), then one can use
+ * U16_..._UNSAFE macros.
+ * These do not check for proper code unit sequences or truncated text and may
+ * yield wrong results or even cause a crash if they are used with "malformed"
+ * text.
+ * In practice, U16_..._UNSAFE macros will produce slightly less code but
+ * should not be faster because the processing is only different when a
+ * surrogate code unit is detected, which will be rare.
+ *
+ * Similarly for UTF-8, there are "safe" macros without a suffix,
+ * and U8_..._UNSAFE versions.
+ * The performance differences are much larger here because UTF-8 provides so
+ * many opportunities for malformed sequences.
+ * The unsafe UTF-8 macros are entirely implemented inside the macro definitions
+ * and are fast, while the safe UTF-8 macros call functions for some complicated cases.
+ *
+ * Unlike with UTF-16, malformed sequences cannot be expressed with distinct
+ * code point values (0..U+10ffff). They are indicated with negative values instead.
+ *
+ * For more information see the ICU User Guide Strings chapter
+ * (https://unicode-org.github.io/icu/userguide/strings).
+ *
+ * <em>Usage:</em>
+ * ICU coding guidelines for if() statements should be followed when using these macros.
+ * Compound statements (curly braces {}) must be used  for if-else-while... 
+ * bodies and all macro statements should be terminated with semicolon.
+ *
+ * @stable ICU 2.4
+ */
+
+#ifndef __UTF_H__
+#define __UTF_H__
+
+#include "unicode/umachine.h"
+/* include the utfXX.h after the following definitions */
+
+/* single-code point definitions -------------------------------------------- */
+
+/**
+ * Is this code point a Unicode noncharacter?
+ * @param c 32-bit code point
+ * @return true or false
+ * @stable ICU 2.4
+ */
+#define U_IS_UNICODE_NONCHAR(c) \
+    ((c)>=0xfdd0 && \
+     ((c)<=0xfdef || ((c)&0xfffe)==0xfffe) && (c)<=0x10ffff)
+
+/**
+ * Is c a Unicode code point value (0..U+10ffff)
+ * that can be assigned a character?
+ *
+ * Code points that are not characters include:
+ * - single surrogate code points (U+d800..U+dfff, 2048 code points)
+ * - the last two code points on each plane (U+__fffe and U+__ffff, 34 code points)
+ * - U+fdd0..U+fdef (new with Unicode 3.1, 32 code points)
+ * - the highest Unicode code point value is U+10ffff
+ *
+ * This means that all code points below U+d800 are character code points,
+ * and that boundary is tested first for performance.
+ *
+ * @param c 32-bit code point
+ * @return true or false
+ * @stable ICU 2.4
+ */
+#define U_IS_UNICODE_CHAR(c) \
+    ((uint32_t)(c)<0xd800 || \
+        (0xdfff<(c) && (c)<=0x10ffff && !U_IS_UNICODE_NONCHAR(c)))
+
+/**
+ * Is this code point a BMP code point (U+0000..U+ffff)?
+ * @param c 32-bit code point
+ * @return true or false
+ * @stable ICU 2.8
+ */
+#define U_IS_BMP(c) ((uint32_t)(c)<=0xffff)
+
+/**
+ * Is this code point a supplementary code point (U+10000..U+10ffff)?
+ * @param c 32-bit code point
+ * @return true or false
+ * @stable ICU 2.8
+ */
+#define U_IS_SUPPLEMENTARY(c) ((uint32_t)((c)-0x10000)<=0xfffff)
+ 
+/**
+ * Is this code point a lead surrogate (U+d800..U+dbff)?
+ * @param c 32-bit code point
+ * @return true or false
+ * @stable ICU 2.4
+ */
+#define U_IS_LEAD(c) (((c)&0xfffffc00)==0xd800)
+
+/**
+ * Is this code point a trail surrogate (U+dc00..U+dfff)?
+ * @param c 32-bit code point
+ * @return true or false
+ * @stable ICU 2.4
+ */
+#define U_IS_TRAIL(c) (((c)&0xfffffc00)==0xdc00)
+
+/**
+ * Is this code point a surrogate (U+d800..U+dfff)?
+ * @param c 32-bit code point
+ * @return true or false
+ * @stable ICU 2.4
+ */
+#define U_IS_SURROGATE(c) (((c)&0xfffff800)==0xd800)
+
+/**
+ * Assuming c is a surrogate code point (U_IS_SURROGATE(c)),
+ * is it a lead surrogate?
+ * @param c 32-bit code point
+ * @return true or false
+ * @stable ICU 2.4
+ */
+#define U_IS_SURROGATE_LEAD(c) (((c)&0x400)==0)
+
+/**
+ * Assuming c is a surrogate code point (U_IS_SURROGATE(c)),
+ * is it a trail surrogate?
+ * @param c 32-bit code point
+ * @return true or false
+ * @stable ICU 4.2
+ */
+#define U_IS_SURROGATE_TRAIL(c) (((c)&0x400)!=0)
+
+/* include the utfXX.h ------------------------------------------------------ */
+
+#if !U_NO_DEFAULT_INCLUDE_UTF_HEADERS
+
+#include "unicode/utf8.h"
+#include "unicode/utf16.h"
+
+/* utf_old.h contains deprecated, pre-ICU 2.4 definitions */
+#include "unicode/utf_old.h"
+
+#endif  /* !U_NO_DEFAULT_INCLUDE_UTF_HEADERS */
+
+#endif  /* __UTF_H__ */

thirdparty/icu4c/common/unicode/utf16.h

@@ -0,0 +1,734 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+/*
+*******************************************************************************
+*
+*   Copyright (C) 1999-2012, International Business Machines
+*   Corporation and others.  All Rights Reserved.
+*
+*******************************************************************************
+*   file name:  utf16.h
+*   encoding:   UTF-8
+*   tab size:   8 (not used)
+*   indentation:4
+*
+*   created on: 1999sep09
+*   created by: Markus W. Scherer
+*/
+
+/**
+ * \file
+ * \brief C API: 16-bit Unicode handling macros
+ * 
+ * This file defines macros to deal with 16-bit Unicode (UTF-16) code units and strings.
+ *
+ * For more information see utf.h and the ICU User Guide Strings chapter
+ * (https://unicode-org.github.io/icu/userguide/strings).
+ *
+ * <em>Usage:</em>
+ * ICU coding guidelines for if() statements should be followed when using these macros.
+ * Compound statements (curly braces {}) must be used  for if-else-while... 
+ * bodies and all macro statements should be terminated with semicolon.
+ */
+
+#ifndef __UTF16_H__
+#define __UTF16_H__
+
+#include <stdbool.h>
+#include "unicode/umachine.h"
+#ifndef __UTF_H__
+#   include "unicode/utf.h"
+#endif
+
+/* single-code point definitions -------------------------------------------- */
+
+/**
+ * Does this code unit alone encode a code point (BMP, not a surrogate)?
+ * @param c 16-bit code unit
+ * @return true or false
+ * @stable ICU 2.4
+ */
+#define U16_IS_SINGLE(c) !U_IS_SURROGATE(c)
+
+/**
+ * Is this code unit a lead surrogate (U+d800..U+dbff)?
+ * @param c 16-bit code unit
+ * @return true or false
+ * @stable ICU 2.4
+ */
+#define U16_IS_LEAD(c) (((c)&0xfffffc00)==0xd800)
+
+/**
+ * Is this code unit a trail surrogate (U+dc00..U+dfff)?
+ * @param c 16-bit code unit
+ * @return true or false
+ * @stable ICU 2.4
+ */
+#define U16_IS_TRAIL(c) (((c)&0xfffffc00)==0xdc00)
+
+/**
+ * Is this code unit a surrogate (U+d800..U+dfff)?
+ * @param c 16-bit code unit
+ * @return true or false
+ * @stable ICU 2.4
+ */
+#define U16_IS_SURROGATE(c) U_IS_SURROGATE(c)
+
+/**
+ * Assuming c is a surrogate code point (U16_IS_SURROGATE(c)),
+ * is it a lead surrogate?
+ * @param c 16-bit code unit
+ * @return true or false
+ * @stable ICU 2.4
+ */
+#define U16_IS_SURROGATE_LEAD(c) (((c)&0x400)==0)
+
+/**
+ * Assuming c is a surrogate code point (U16_IS_SURROGATE(c)),
+ * is it a trail surrogate?
+ * @param c 16-bit code unit
+ * @return true or false
+ * @stable ICU 4.2
+ */
+#define U16_IS_SURROGATE_TRAIL(c) (((c)&0x400)!=0)
+
+/**
+ * Helper constant for U16_GET_SUPPLEMENTARY.
+ * @internal
+ */
+#define U16_SURROGATE_OFFSET ((0xd800<<10UL)+0xdc00-0x10000)
+
+/**
+ * Get a supplementary code point value (U+10000..U+10ffff)
+ * from its lead and trail surrogates.
+ * The result is undefined if the input values are not
+ * lead and trail surrogates.
+ *
+ * @param lead lead surrogate (U+d800..U+dbff)
+ * @param trail trail surrogate (U+dc00..U+dfff)
+ * @return supplementary code point (U+10000..U+10ffff)
+ * @stable ICU 2.4
+ */
+#define U16_GET_SUPPLEMENTARY(lead, trail) \
+    (((UChar32)(lead)<<10UL)+(UChar32)(trail)-U16_SURROGATE_OFFSET)
+
+
+/**
+ * Get the lead surrogate (0xd800..0xdbff) for a
+ * supplementary code point (0x10000..0x10ffff).
+ * @param supplementary 32-bit code point (U+10000..U+10ffff)
+ * @return lead surrogate (U+d800..U+dbff) for supplementary
+ * @stable ICU 2.4
+ */
+#define U16_LEAD(supplementary) (UChar)(((supplementary)>>10)+0xd7c0)
+
+/**
+ * Get the trail surrogate (0xdc00..0xdfff) for a
+ * supplementary code point (0x10000..0x10ffff).
+ * @param supplementary 32-bit code point (U+10000..U+10ffff)
+ * @return trail surrogate (U+dc00..U+dfff) for supplementary
+ * @stable ICU 2.4
+ */
+#define U16_TRAIL(supplementary) (UChar)(((supplementary)&0x3ff)|0xdc00)
+
+/**
+ * How many 16-bit code units are used to encode this Unicode code point? (1 or 2)
+ * The result is not defined if c is not a Unicode code point (U+0000..U+10ffff).
+ * @param c 32-bit code point
+ * @return 1 or 2
+ * @stable ICU 2.4
+ */
+#define U16_LENGTH(c) ((uint32_t)(c)<=0xffff ? 1 : 2)
+
+/**
+ * The maximum number of 16-bit code units per Unicode code point (U+0000..U+10ffff).
+ * @return 2
+ * @stable ICU 2.4
+ */
+#define U16_MAX_LENGTH 2
+
+/**
+ * Get a code point from a string at a random-access offset,
+ * without changing the offset.
+ * "Unsafe" macro, assumes well-formed UTF-16.
+ *
+ * The offset may point to either the lead or trail surrogate unit
+ * for a supplementary code point, in which case the macro will read
+ * the adjacent matching surrogate as well.
+ * The result is undefined if the offset points to a single, unpaired surrogate.
+ * Iteration through a string is more efficient with U16_NEXT_UNSAFE or U16_NEXT.
+ *
+ * @param s const UChar * string
+ * @param i string offset
+ * @param c output UChar32 variable
+ * @see U16_GET
+ * @stable ICU 2.4
+ */
+#define U16_GET_UNSAFE(s, i, c) UPRV_BLOCK_MACRO_BEGIN { \
+    (c)=(s)[i]; \
+    if(U16_IS_SURROGATE(c)) { \
+        if(U16_IS_SURROGATE_LEAD(c)) { \
+            (c)=U16_GET_SUPPLEMENTARY((c), (s)[(i)+1]); \
+        } else { \
+            (c)=U16_GET_SUPPLEMENTARY((s)[(i)-1], (c)); \
+        } \
+    } \
+} UPRV_BLOCK_MACRO_END
+
+/**
+ * Get a code point from a string at a random-access offset,
+ * without changing the offset.
+ * "Safe" macro, handles unpaired surrogates and checks for string boundaries.
+ *
+ * The offset may point to either the lead or trail surrogate unit
+ * for a supplementary code point, in which case the macro will read
+ * the adjacent matching surrogate as well.
+ *
+ * The length can be negative for a NUL-terminated string.
+ *
+ * If the offset points to a single, unpaired surrogate, then
+ * c is set to that unpaired surrogate.
+ * Iteration through a string is more efficient with U16_NEXT_UNSAFE or U16_NEXT.
+ *
+ * @param s const UChar * string
+ * @param start starting string offset (usually 0)
+ * @param i string offset, must be start<=i<length
+ * @param length string length
+ * @param c output UChar32 variable
+ * @see U16_GET_UNSAFE
+ * @stable ICU 2.4
+ */
+#define U16_GET(s, start, i, length, c) UPRV_BLOCK_MACRO_BEGIN { \
+    (c)=(s)[i]; \
+    if(U16_IS_SURROGATE(c)) { \
+        uint16_t __c2; \
+        if(U16_IS_SURROGATE_LEAD(c)) { \
+            if((i)+1!=(length) && U16_IS_TRAIL(__c2=(s)[(i)+1])) { \
+                (c)=U16_GET_SUPPLEMENTARY((c), __c2); \
+            } \
+        } else { \
+            if((i)>(start) && U16_IS_LEAD(__c2=(s)[(i)-1])) { \
+                (c)=U16_GET_SUPPLEMENTARY(__c2, (c)); \
+            } \
+        } \
+    } \
+} UPRV_BLOCK_MACRO_END
+
+/**
+ * Get a code point from a string at a random-access offset,
+ * without changing the offset.
+ * "Safe" macro, handles unpaired surrogates and checks for string boundaries.
+ *
+ * The offset may point to either the lead or trail surrogate unit
+ * for a supplementary code point, in which case the macro will read
+ * the adjacent matching surrogate as well.
+ *
+ * The length can be negative for a NUL-terminated string.
+ *
+ * If the offset points to a single, unpaired surrogate, then
+ * c is set to U+FFFD.
+ * Iteration through a string is more efficient with U16_NEXT_UNSAFE or U16_NEXT_OR_FFFD.
+ *
+ * @param s const UChar * string
+ * @param start starting string offset (usually 0)
+ * @param i string offset, must be start<=i<length
+ * @param length string length
+ * @param c output UChar32 variable
+ * @see U16_GET_UNSAFE
+ * @stable ICU 60
+ */
+#define U16_GET_OR_FFFD(s, start, i, length, c) UPRV_BLOCK_MACRO_BEGIN { \
+    (c)=(s)[i]; \
+    if(U16_IS_SURROGATE(c)) { \
+        uint16_t __c2; \
+        if(U16_IS_SURROGATE_LEAD(c)) { \
+            if((i)+1!=(length) && U16_IS_TRAIL(__c2=(s)[(i)+1])) { \
+                (c)=U16_GET_SUPPLEMENTARY((c), __c2); \
+            } else { \
+                (c)=0xfffd; \
+            } \
+        } else { \
+            if((i)>(start) && U16_IS_LEAD(__c2=(s)[(i)-1])) { \
+                (c)=U16_GET_SUPPLEMENTARY(__c2, (c)); \
+            } else { \
+                (c)=0xfffd; \
+            } \
+        } \
+    } \
+} UPRV_BLOCK_MACRO_END
+
+/* definitions with forward iteration --------------------------------------- */
+
+/**
+ * Get a code point from a string at a code point boundary offset,
+ * and advance the offset to the next code point boundary.
+ * (Post-incrementing forward iteration.)
+ * "Unsafe" macro, assumes well-formed UTF-16.
+ *
+ * The offset may point to the lead surrogate unit
+ * for a supplementary code point, in which case the macro will read
+ * the following trail surrogate as well.
+ * If the offset points to a trail surrogate, then that itself
+ * will be returned as the code point.
+ * The result is undefined if the offset points to a single, unpaired lead surrogate.
+ *
+ * @param s const UChar * string
+ * @param i string offset
+ * @param c output UChar32 variable
+ * @see U16_NEXT
+ * @stable ICU 2.4
+ */
+#define U16_NEXT_UNSAFE(s, i, c) UPRV_BLOCK_MACRO_BEGIN { \
+    (c)=(s)[(i)++]; \
+    if(U16_IS_LEAD(c)) { \
+        (c)=U16_GET_SUPPLEMENTARY((c), (s)[(i)++]); \
+    } \
+} UPRV_BLOCK_MACRO_END
+
+/**
+ * Get a code point from a string at a code point boundary offset,
+ * and advance the offset to the next code point boundary.
+ * (Post-incrementing forward iteration.)
+ * "Safe" macro, handles unpaired surrogates and checks for string boundaries.
+ *
+ * The length can be negative for a NUL-terminated string.
+ *
+ * The offset may point to the lead surrogate unit
+ * for a supplementary code point, in which case the macro will read
+ * the following trail surrogate as well.
+ * If the offset points to a trail surrogate or
+ * to a single, unpaired lead surrogate, then c is set to that unpaired surrogate.
+ *
+ * @param s const UChar * string
+ * @param i string offset, must be i<length
+ * @param length string length
+ * @param c output UChar32 variable
+ * @see U16_NEXT_UNSAFE
+ * @stable ICU 2.4
+ */
+#define U16_NEXT(s, i, length, c) UPRV_BLOCK_MACRO_BEGIN { \
+    (c)=(s)[(i)++]; \
+    if(U16_IS_LEAD(c)) { \
+        uint16_t __c2; \
+        if((i)!=(length) && U16_IS_TRAIL(__c2=(s)[(i)])) { \
+            ++(i); \
+            (c)=U16_GET_SUPPLEMENTARY((c), __c2); \
+        } \
+    } \
+} UPRV_BLOCK_MACRO_END
+
+/**
+ * Get a code point from a string at a code point boundary offset,
+ * and advance the offset to the next code point boundary.
+ * (Post-incrementing forward iteration.)
+ * "Safe" macro, handles unpaired surrogates and checks for string boundaries.
+ *
+ * The length can be negative for a NUL-terminated string.
+ *
+ * The offset may point to the lead surrogate unit
+ * for a supplementary code point, in which case the macro will read
+ * the following trail surrogate as well.
+ * If the offset points to a trail surrogate or
+ * to a single, unpaired lead surrogate, then c is set to U+FFFD.
+ *
+ * @param s const UChar * string
+ * @param i string offset, must be i<length
+ * @param length string length
+ * @param c output UChar32 variable
+ * @see U16_NEXT_UNSAFE
+ * @stable ICU 60
+ */
+#define U16_NEXT_OR_FFFD(s, i, length, c) UPRV_BLOCK_MACRO_BEGIN { \
+    (c)=(s)[(i)++]; \
+    if(U16_IS_SURROGATE(c)) { \
+        uint16_t __c2; \
+        if(U16_IS_SURROGATE_LEAD(c) && (i)!=(length) && U16_IS_TRAIL(__c2=(s)[(i)])) { \
+            ++(i); \
+            (c)=U16_GET_SUPPLEMENTARY((c), __c2); \
+        } else { \
+            (c)=0xfffd; \
+        } \
+    } \
+} UPRV_BLOCK_MACRO_END
+
+/**
+ * Append a code point to a string, overwriting 1 or 2 code units.
+ * The offset points to the current end of the string contents
+ * and is advanced (post-increment).
+ * "Unsafe" macro, assumes a valid code point and sufficient space in the string.
+ * Otherwise, the result is undefined.
+ *
+ * @param s const UChar * string buffer
+ * @param i string offset
+ * @param c code point to append
+ * @see U16_APPEND
+ * @stable ICU 2.4
+ */
+#define U16_APPEND_UNSAFE(s, i, c) UPRV_BLOCK_MACRO_BEGIN { \
+    if((uint32_t)(c)<=0xffff) { \
+        (s)[(i)++]=(uint16_t)(c); \
+    } else { \
+        (s)[(i)++]=(uint16_t)(((c)>>10)+0xd7c0); \
+        (s)[(i)++]=(uint16_t)(((c)&0x3ff)|0xdc00); \
+    } \
+} UPRV_BLOCK_MACRO_END
+
+/**
+ * Append a code point to a string, overwriting 1 or 2 code units.
+ * The offset points to the current end of the string contents
+ * and is advanced (post-increment).
+ * "Safe" macro, checks for a valid code point.
+ * If a surrogate pair is written, checks for sufficient space in the string.
+ * If the code point is not valid or a trail surrogate does not fit,
+ * then isError is set to true.
+ *
+ * @param s const UChar * string buffer
+ * @param i string offset, must be i<capacity
+ * @param capacity size of the string buffer
+ * @param c code point to append
+ * @param isError output UBool set to true if an error occurs, otherwise not modified
+ * @see U16_APPEND_UNSAFE
+ * @stable ICU 2.4
+ */
+#define U16_APPEND(s, i, capacity, c, isError) UPRV_BLOCK_MACRO_BEGIN { \
+    if((uint32_t)(c)<=0xffff) { \
+        (s)[(i)++]=(uint16_t)(c); \
+    } else if((uint32_t)(c)<=0x10ffff && (i)+1<(capacity)) { \
+        (s)[(i)++]=(uint16_t)(((c)>>10)+0xd7c0); \
+        (s)[(i)++]=(uint16_t)(((c)&0x3ff)|0xdc00); \
+    } else /* c>0x10ffff or not enough space */ { \
+        (isError)=true; \
+    } \
+} UPRV_BLOCK_MACRO_END
+
+/**
+ * Advance the string offset from one code point boundary to the next.
+ * (Post-incrementing iteration.)
+ * "Unsafe" macro, assumes well-formed UTF-16.
+ *
+ * @param s const UChar * string
+ * @param i string offset
+ * @see U16_FWD_1
+ * @stable ICU 2.4
+ */
+#define U16_FWD_1_UNSAFE(s, i) UPRV_BLOCK_MACRO_BEGIN { \
+    if(U16_IS_LEAD((s)[(i)++])) { \
+        ++(i); \
+    } \
+} UPRV_BLOCK_MACRO_END
+
+/**
+ * Advance the string offset from one code point boundary to the next.
+ * (Post-incrementing iteration.)
+ * "Safe" macro, handles unpaired surrogates and checks for string boundaries.
+ *
+ * The length can be negative for a NUL-terminated string.
+ *
+ * @param s const UChar * string
+ * @param i string offset, must be i<length
+ * @param length string length
+ * @see U16_FWD_1_UNSAFE
+ * @stable ICU 2.4
+ */
+#define U16_FWD_1(s, i, length) UPRV_BLOCK_MACRO_BEGIN { \
+    if(U16_IS_LEAD((s)[(i)++]) && (i)!=(length) && U16_IS_TRAIL((s)[i])) { \
+        ++(i); \
+    } \
+} UPRV_BLOCK_MACRO_END
+
+/**
+ * Advance the string offset from one code point boundary to the n-th next one,
+ * i.e., move forward by n code points.
+ * (Post-incrementing iteration.)
+ * "Unsafe" macro, assumes well-formed UTF-16.
+ *
+ * @param s const UChar * string
+ * @param i string offset
+ * @param n number of code points to skip
+ * @see U16_FWD_N
+ * @stable ICU 2.4
+ */
+#define U16_FWD_N_UNSAFE(s, i, n) UPRV_BLOCK_MACRO_BEGIN { \
+    int32_t __N=(n); \
+    while(__N>0) { \
+        U16_FWD_1_UNSAFE(s, i); \
+        --__N; \
+    } \
+} UPRV_BLOCK_MACRO_END
+
+/**
+ * Advance the string offset from one code point boundary to the n-th next one,
+ * i.e., move forward by n code points.
+ * (Post-incrementing iteration.)
+ * "Safe" macro, handles unpaired surrogates and checks for string boundaries.
+ *
+ * The length can be negative for a NUL-terminated string.
+ *
+ * @param s const UChar * string
+ * @param i int32_t string offset, must be i<length
+ * @param length int32_t string length
+ * @param n number of code points to skip
+ * @see U16_FWD_N_UNSAFE
+ * @stable ICU 2.4
+ */
+#define U16_FWD_N(s, i, length, n) UPRV_BLOCK_MACRO_BEGIN { \
+    int32_t __N=(n); \
+    while(__N>0 && ((i)<(length) || ((length)<0 && (s)[i]!=0))) { \
+        U16_FWD_1(s, i, length); \
+        --__N; \
+    } \
+} UPRV_BLOCK_MACRO_END
+
+/**
+ * Adjust a random-access offset to a code point boundary
+ * at the start of a code point.
+ * If the offset points to the trail surrogate of a surrogate pair,
+ * then the offset is decremented.
+ * Otherwise, it is not modified.
+ * "Unsafe" macro, assumes well-formed UTF-16.
+ *
+ * @param s const UChar * string
+ * @param i string offset
+ * @see U16_SET_CP_START
+ * @stable ICU 2.4
+ */
+#define U16_SET_CP_START_UNSAFE(s, i) UPRV_BLOCK_MACRO_BEGIN { \
+    if(U16_IS_TRAIL((s)[i])) { \
+        --(i); \
+    } \
+} UPRV_BLOCK_MACRO_END
+
+/**
+ * Adjust a random-access offset to a code point boundary
+ * at the start of a code point.
+ * If the offset points to the trail surrogate of a surrogate pair,
+ * then the offset is decremented.
+ * Otherwise, it is not modified.
+ * "Safe" macro, handles unpaired surrogates and checks for string boundaries.
+ *
+ * @param s const UChar * string
+ * @param start starting string offset (usually 0)
+ * @param i string offset, must be start<=i
+ * @see U16_SET_CP_START_UNSAFE
+ * @stable ICU 2.4
+ */
+#define U16_SET_CP_START(s, start, i) UPRV_BLOCK_MACRO_BEGIN { \
+    if(U16_IS_TRAIL((s)[i]) && (i)>(start) && U16_IS_LEAD((s)[(i)-1])) { \
+        --(i); \
+    } \
+} UPRV_BLOCK_MACRO_END
+
+/* definitions with backward iteration -------------------------------------- */
+
+/**
+ * Move the string offset from one code point boundary to the previous one
+ * and get the code point between them.
+ * (Pre-decrementing backward iteration.)
+ * "Unsafe" macro, assumes well-formed UTF-16.
+ *
+ * The input offset may be the same as the string length.
+ * If the offset is behind a trail surrogate unit
+ * for a supplementary code point, then the macro will read
+ * the preceding lead surrogate as well.
+ * If the offset is behind a lead surrogate, then that itself
+ * will be returned as the code point.
+ * The result is undefined if the offset is behind a single, unpaired trail surrogate.
+ *
+ * @param s const UChar * string
+ * @param i string offset
+ * @param c output UChar32 variable
+ * @see U16_PREV
+ * @stable ICU 2.4
+ */
+#define U16_PREV_UNSAFE(s, i, c) UPRV_BLOCK_MACRO_BEGIN { \
+    (c)=(s)[--(i)]; \
+    if(U16_IS_TRAIL(c)) { \
+        (c)=U16_GET_SUPPLEMENTARY((s)[--(i)], (c)); \
+    } \
+} UPRV_BLOCK_MACRO_END
+
+/**
+ * Move the string offset from one code point boundary to the previous one
+ * and get the code point between them.
+ * (Pre-decrementing backward iteration.)
+ * "Safe" macro, handles unpaired surrogates and checks for string boundaries.
+ *
+ * The input offset may be the same as the string length.
+ * If the offset is behind a trail surrogate unit
+ * for a supplementary code point, then the macro will read
+ * the preceding lead surrogate as well.
+ * If the offset is behind a lead surrogate or behind a single, unpaired
+ * trail surrogate, then c is set to that unpaired surrogate.
+ *
+ * @param s const UChar * string
+ * @param start starting string offset (usually 0)
+ * @param i string offset, must be start<i
+ * @param c output UChar32 variable
+ * @see U16_PREV_UNSAFE
+ * @stable ICU 2.4
+ */
+#define U16_PREV(s, start, i, c) UPRV_BLOCK_MACRO_BEGIN { \
+    (c)=(s)[--(i)]; \
+    if(U16_IS_TRAIL(c)) { \
+        uint16_t __c2; \
+        if((i)>(start) && U16_IS_LEAD(__c2=(s)[(i)-1])) { \
+            --(i); \
+            (c)=U16_GET_SUPPLEMENTARY(__c2, (c)); \
+        } \
+    } \
+} UPRV_BLOCK_MACRO_END
+
+/**
+ * Move the string offset from one code point boundary to the previous one
+ * and get the code point between them.
+ * (Pre-decrementing backward iteration.)
+ * "Safe" macro, handles unpaired surrogates and checks for string boundaries.
+ *
+ * The input offset may be the same as the string length.
+ * If the offset is behind a trail surrogate unit
+ * for a supplementary code point, then the macro will read
+ * the preceding lead surrogate as well.
+ * If the offset is behind a lead surrogate or behind a single, unpaired
+ * trail surrogate, then c is set to U+FFFD.
+ *
+ * @param s const UChar * string
+ * @param start starting string offset (usually 0)
+ * @param i string offset, must be start<i
+ * @param c output UChar32 variable
+ * @see U16_PREV_UNSAFE
+ * @stable ICU 60
+ */
+#define U16_PREV_OR_FFFD(s, start, i, c) UPRV_BLOCK_MACRO_BEGIN { \
+    (c)=(s)[--(i)]; \
+    if(U16_IS_SURROGATE(c)) { \
+        uint16_t __c2; \
+        if(U16_IS_SURROGATE_TRAIL(c) && (i)>(start) && U16_IS_LEAD(__c2=(s)[(i)-1])) { \
+            --(i); \
+            (c)=U16_GET_SUPPLEMENTARY(__c2, (c)); \
+        } else { \
+            (c)=0xfffd; \
+        } \
+    } \
+} UPRV_BLOCK_MACRO_END
+
+/**
+ * Move the string offset from one code point boundary to the previous one.
+ * (Pre-decrementing backward iteration.)
+ * The input offset may be the same as the string length.
+ * "Unsafe" macro, assumes well-formed UTF-16.
+ *
+ * @param s const UChar * string
+ * @param i string offset
+ * @see U16_BACK_1
+ * @stable ICU 2.4
+ */
+#define U16_BACK_1_UNSAFE(s, i) UPRV_BLOCK_MACRO_BEGIN { \
+    if(U16_IS_TRAIL((s)[--(i)])) { \
+        --(i); \
+    } \
+} UPRV_BLOCK_MACRO_END
+
+/**
+ * Move the string offset from one code point boundary to the previous one.
+ * (Pre-decrementing backward iteration.)
+ * The input offset may be the same as the string length.
+ * "Safe" macro, handles unpaired surrogates and checks for string boundaries.
+ *
+ * @param s const UChar * string
+ * @param start starting string offset (usually 0)
+ * @param i string offset, must be start<i
+ * @see U16_BACK_1_UNSAFE
+ * @stable ICU 2.4
+ */
+#define U16_BACK_1(s, start, i) UPRV_BLOCK_MACRO_BEGIN { \
+    if(U16_IS_TRAIL((s)[--(i)]) && (i)>(start) && U16_IS_LEAD((s)[(i)-1])) { \
+        --(i); \
+    } \
+} UPRV_BLOCK_MACRO_END
+
+/**
+ * Move the string offset from one code point boundary to the n-th one before it,
+ * i.e., move backward by n code points.
+ * (Pre-decrementing backward iteration.)
+ * The input offset may be the same as the string length.
+ * "Unsafe" macro, assumes well-formed UTF-16.
+ *
+ * @param s const UChar * string
+ * @param i string offset
+ * @param n number of code points to skip
+ * @see U16_BACK_N
+ * @stable ICU 2.4
+ */
+#define U16_BACK_N_UNSAFE(s, i, n) UPRV_BLOCK_MACRO_BEGIN { \
+    int32_t __N=(n); \
+    while(__N>0) { \
+        U16_BACK_1_UNSAFE(s, i); \
+        --__N; \
+    } \
+} UPRV_BLOCK_MACRO_END
+
+/**
+ * Move the string offset from one code point boundary to the n-th one before it,
+ * i.e., move backward by n code points.
+ * (Pre-decrementing backward iteration.)
+ * The input offset may be the same as the string length.
+ * "Safe" macro, handles unpaired surrogates and checks for string boundaries.
+ *
+ * @param s const UChar * string
+ * @param start start of string
+ * @param i string offset, must be start<i
+ * @param n number of code points to skip
+ * @see U16_BACK_N_UNSAFE
+ * @stable ICU 2.4
+ */
+#define U16_BACK_N(s, start, i, n) UPRV_BLOCK_MACRO_BEGIN { \
+    int32_t __N=(n); \
+    while(__N>0 && (i)>(start)) { \
+        U16_BACK_1(s, start, i); \
+        --__N; \
+    } \
+} UPRV_BLOCK_MACRO_END
+
+/**
+ * Adjust a random-access offset to a code point boundary after a code point.
+ * If the offset is behind the lead surrogate of a surrogate pair,
+ * then the offset is incremented.
+ * Otherwise, it is not modified.
+ * The input offset may be the same as the string length.
+ * "Unsafe" macro, assumes well-formed UTF-16.
+ *
+ * @param s const UChar * string
+ * @param i string offset
+ * @see U16_SET_CP_LIMIT
+ * @stable ICU 2.4
+ */
+#define U16_SET_CP_LIMIT_UNSAFE(s, i) UPRV_BLOCK_MACRO_BEGIN { \
+    if(U16_IS_LEAD((s)[(i)-1])) { \
+        ++(i); \
+    } \
+} UPRV_BLOCK_MACRO_END
+
+/**
+ * Adjust a random-access offset to a code point boundary after a code point.
+ * If the offset is behind the lead surrogate of a surrogate pair,
+ * then the offset is incremented.
+ * Otherwise, it is not modified.
+ * The input offset may be the same as the string length.
+ * "Safe" macro, handles unpaired surrogates and checks for string boundaries.
+ *
+ * The length can be negative for a NUL-terminated string.
+ *
+ * @param s const UChar * string
+ * @param start int32_t starting string offset (usually 0)
+ * @param i int32_t string offset, start<=i<=length
+ * @param length int32_t string length
+ * @see U16_SET_CP_LIMIT_UNSAFE
+ * @stable ICU 2.4
+ */
+#define U16_SET_CP_LIMIT(s, start, i, length) UPRV_BLOCK_MACRO_BEGIN { \
+    if((start)<(i) && ((i)<(length) || (length)<0) && U16_IS_LEAD((s)[(i)-1]) && U16_IS_TRAIL((s)[i])) { \
+        ++(i); \
+    } \
+} UPRV_BLOCK_MACRO_END
+
+#endif

thirdparty/icu4c/common/unicode/utf32.h

@@ -0,0 +1,25 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+/*
+*******************************************************************************
+*
+*   Copyright (C) 1999-2001, International Business Machines
+*   Corporation and others.  All Rights Reserved.
+*
+*******************************************************************************
+*   file name:  utf32.h
+*   encoding:   UTF-8
+*   tab size:   8 (not used)
+*   indentation:4
+*
+*   created on: 1999sep20
+*   created by: Markus W. Scherer
+*/
+/**
+ * \file
+ * \brief C API: UTF-32 macros
+ *
+ * This file is obsolete and its contents moved to utf_old.h.
+ * See utf_old.h and Jitterbug 2150 and its discussion on the ICU mailing list
+ * in September 2002.
+ */

thirdparty/icu4c/common/unicode/utf8.h

@@ -0,0 +1,882 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+/*
+*******************************************************************************
+*
+*   Copyright (C) 1999-2015, International Business Machines
+*   Corporation and others.  All Rights Reserved.
+*
+*******************************************************************************
+*   file name:  utf8.h
+*   encoding:   UTF-8
+*   tab size:   8 (not used)
+*   indentation:4
+*
+*   created on: 1999sep13
+*   created by: Markus W. Scherer
+*/
+
+/**
+ * \file
+ * \brief C API: 8-bit Unicode handling macros
+ * 
+ * This file defines macros to deal with 8-bit Unicode (UTF-8) code units (bytes) and strings.
+ *
+ * For more information see utf.h and the ICU User Guide Strings chapter
+ * (https://unicode-org.github.io/icu/userguide/strings).
+ *
+ * <em>Usage:</em>
+ * ICU coding guidelines for if() statements should be followed when using these macros.
+ * Compound statements (curly braces {}) must be used  for if-else-while... 
+ * bodies and all macro statements should be terminated with semicolon.
+ */
+
+#ifndef __UTF8_H__
+#define __UTF8_H__
+
+#include <stdbool.h>
+#include "unicode/umachine.h"
+#ifndef __UTF_H__
+#   include "unicode/utf.h"
+#endif
+
+/* internal definitions ----------------------------------------------------- */
+
+/**
+ * Counts the trail bytes for a UTF-8 lead byte.
+ * Returns 0 for 0..0xc1 as well as for 0xf5..0xff.
+ * leadByte might be evaluated multiple times.
+ *
+ * This is internal since it is not meant to be called directly by external clients;
+ * however it is called by public macros in this file and thus must remain stable.
+ *
+ * @param leadByte The first byte of a UTF-8 sequence. Must be 0..0xff.
+ * @internal
+ */
+#define U8_COUNT_TRAIL_BYTES(leadByte) \
+    (U8_IS_LEAD(leadByte) ? \
+        ((uint8_t)(leadByte)>=0xe0)+((uint8_t)(leadByte)>=0xf0)+1 : 0)
+
+/**
+ * Counts the trail bytes for a UTF-8 lead byte of a valid UTF-8 sequence.
+ * Returns 0 for 0..0xc1. Undefined for 0xf5..0xff.
+ * leadByte might be evaluated multiple times.
+ *
+ * This is internal since it is not meant to be called directly by external clients;
+ * however it is called by public macros in this file and thus must remain stable.
+ *
+ * @param leadByte The first byte of a UTF-8 sequence. Must be 0..0xff.
+ * @internal
+ */
+#define U8_COUNT_TRAIL_BYTES_UNSAFE(leadByte) \
+    (((uint8_t)(leadByte)>=0xc2)+((uint8_t)(leadByte)>=0xe0)+((uint8_t)(leadByte)>=0xf0))
+
+/**
+ * Mask a UTF-8 lead byte, leave only the lower bits that form part of the code point value.
+ *
+ * This is internal since it is not meant to be called directly by external clients;
+ * however it is called by public macros in this file and thus must remain stable.
+ * @internal
+ */
+#define U8_MASK_LEAD_BYTE(leadByte, countTrailBytes) ((leadByte)&=(1<<(6-(countTrailBytes)))-1)
+
+/**
+ * Internal bit vector for 3-byte UTF-8 validity check, for use in U8_IS_VALID_LEAD3_AND_T1.
+ * Each bit indicates whether one lead byte + first trail byte pair starts a valid sequence.
+ * Lead byte E0..EF bits 3..0 are used as byte index,
+ * first trail byte bits 7..5 are used as bit index into that byte.
+ * @see U8_IS_VALID_LEAD3_AND_T1
+ * @internal
+ */
+#define U8_LEAD3_T1_BITS "\x20\x30\x30\x30\x30\x30\x30\x30\x30\x30\x30\x30\x30\x10\x30\x30"
+
+/**
+ * Internal 3-byte UTF-8 validity check.
+ * Non-zero if lead byte E0..EF and first trail byte 00..FF start a valid sequence.
+ * @internal
+ */
+#define U8_IS_VALID_LEAD3_AND_T1(lead, t1) (U8_LEAD3_T1_BITS[(lead)&0xf]&(1<<((uint8_t)(t1)>>5)))
+
+/**
+ * Internal bit vector for 4-byte UTF-8 validity check, for use in U8_IS_VALID_LEAD4_AND_T1.
+ * Each bit indicates whether one lead byte + first trail byte pair starts a valid sequence.
+ * First trail byte bits 7..4 are used as byte index,
+ * lead byte F0..F4 bits 2..0 are used as bit index into that byte.
+ * @see U8_IS_VALID_LEAD4_AND_T1
+ * @internal
+ */
+#define U8_LEAD4_T1_BITS "\x00\x00\x00\x00\x00\x00\x00\x00\x1E\x0F\x0F\x0F\x00\x00\x00\x00"
+
+/**
+ * Internal 4-byte UTF-8 validity check.
+ * Non-zero if lead byte F0..F4 and first trail byte 00..FF start a valid sequence.
+ * @internal
+ */
+#define U8_IS_VALID_LEAD4_AND_T1(lead, t1) (U8_LEAD4_T1_BITS[(uint8_t)(t1)>>4]&(1<<((lead)&7)))
+
+/**
+ * Function for handling "next code point" with error-checking.
+ *
+ * This is internal since it is not meant to be called directly by external clients;
+ * however it is called by public macros in this
+ * file and thus must remain stable, and should not be hidden when other internal
+ * functions are hidden (otherwise public macros would fail to compile).
+ * @internal
+ */
+U_CAPI UChar32 U_EXPORT2
+utf8_nextCharSafeBody(const uint8_t *s, int32_t *pi, int32_t length, UChar32 c, int8_t strict);
+
+/**
+ * Function for handling "append code point" with error-checking.
+ *
+ * This is internal since it is not meant to be called directly by external clients;
+ * however it is called by public macros in this
+ * file and thus must remain stable, and should not be hidden when other internal
+ * functions are hidden (otherwise public macros would fail to compile).
+ * @internal
+ */
+U_CAPI int32_t U_EXPORT2
+utf8_appendCharSafeBody(uint8_t *s, int32_t i, int32_t length, UChar32 c, UBool *pIsError);
+
+/**
+ * Function for handling "previous code point" with error-checking.
+ *
+ * This is internal since it is not meant to be called directly by external clients;
+ * however it is called by public macros in this
+ * file and thus must remain stable, and should not be hidden when other internal
+ * functions are hidden (otherwise public macros would fail to compile).
+ * @internal
+ */
+U_CAPI UChar32 U_EXPORT2
+utf8_prevCharSafeBody(const uint8_t *s, int32_t start, int32_t *pi, UChar32 c, int8_t strict);
+
+/**
+ * Function for handling "skip backward one code point" with error-checking.
+ *
+ * This is internal since it is not meant to be called directly by external clients;
+ * however it is called by public macros in this
+ * file and thus must remain stable, and should not be hidden when other internal
+ * functions are hidden (otherwise public macros would fail to compile).
+ * @internal
+ */
+U_CAPI int32_t U_EXPORT2
+utf8_back1SafeBody(const uint8_t *s, int32_t start, int32_t i);
+
+/* single-code point definitions -------------------------------------------- */
+
+/**
+ * Does this code unit (byte) encode a code point by itself (US-ASCII 0..0x7f)?
+ * @param c 8-bit code unit (byte)
+ * @return true or false
+ * @stable ICU 2.4
+ */
+#define U8_IS_SINGLE(c) (((c)&0x80)==0)
+
+/**
+ * Is this code unit (byte) a UTF-8 lead byte? (0xC2..0xF4)
+ * @param c 8-bit code unit (byte)
+ * @return true or false
+ * @stable ICU 2.4
+ */
+#define U8_IS_LEAD(c) ((uint8_t)((c)-0xc2)<=0x32)
+// 0x32=0xf4-0xc2
+
+/**
+ * Is this code unit (byte) a UTF-8 trail byte? (0x80..0xBF)
+ * @param c 8-bit code unit (byte)
+ * @return true or false
+ * @stable ICU 2.4
+ */
+#define U8_IS_TRAIL(c) ((int8_t)(c)<-0x40)
+
+/**
+ * How many code units (bytes) are used for the UTF-8 encoding
+ * of this Unicode code point?
+ * @param c 32-bit code point
+ * @return 1..4, or 0 if c is a surrogate or not a Unicode code point
+ * @stable ICU 2.4
+ */
+#define U8_LENGTH(c) \
+    ((uint32_t)(c)<=0x7f ? 1 : \
+        ((uint32_t)(c)<=0x7ff ? 2 : \
+            ((uint32_t)(c)<=0xd7ff ? 3 : \
+                ((uint32_t)(c)<=0xdfff || (uint32_t)(c)>0x10ffff ? 0 : \
+                    ((uint32_t)(c)<=0xffff ? 3 : 4)\
+                ) \
+            ) \
+        ) \
+    )
+
+/**
+ * The maximum number of UTF-8 code units (bytes) per Unicode code point (U+0000..U+10ffff).
+ * @return 4
+ * @stable ICU 2.4
+ */
+#define U8_MAX_LENGTH 4
+
+/**
+ * Get a code point from a string at a random-access offset,
+ * without changing the offset.
+ * The offset may point to either the lead byte or one of the trail bytes
+ * for a code point, in which case the macro will read all of the bytes
+ * for the code point.
+ * The result is undefined if the offset points to an illegal UTF-8
+ * byte sequence.
+ * Iteration through a string is more efficient with U8_NEXT_UNSAFE or U8_NEXT.
+ *
+ * @param s const uint8_t * string
+ * @param i string offset
+ * @param c output UChar32 variable
+ * @see U8_GET
+ * @stable ICU 2.4
+ */
+#define U8_GET_UNSAFE(s, i, c) UPRV_BLOCK_MACRO_BEGIN { \
+    int32_t _u8_get_unsafe_index=(int32_t)(i); \
+    U8_SET_CP_START_UNSAFE(s, _u8_get_unsafe_index); \
+    U8_NEXT_UNSAFE(s, _u8_get_unsafe_index, c); \
+} UPRV_BLOCK_MACRO_END
+
+/**
+ * Get a code point from a string at a random-access offset,
+ * without changing the offset.
+ * The offset may point to either the lead byte or one of the trail bytes
+ * for a code point, in which case the macro will read all of the bytes
+ * for the code point.
+ *
+ * The length can be negative for a NUL-terminated string.
+ *
+ * If the offset points to an illegal UTF-8 byte sequence, then
+ * c is set to a negative value.
+ * Iteration through a string is more efficient with U8_NEXT_UNSAFE or U8_NEXT.
+ *
+ * @param s const uint8_t * string
+ * @param start int32_t starting string offset
+ * @param i int32_t string offset, must be start<=i<length
+ * @param length int32_t string length
+ * @param c output UChar32 variable, set to <0 in case of an error
+ * @see U8_GET_UNSAFE
+ * @stable ICU 2.4
+ */
+#define U8_GET(s, start, i, length, c) UPRV_BLOCK_MACRO_BEGIN { \
+    int32_t _u8_get_index=(i); \
+    U8_SET_CP_START(s, start, _u8_get_index); \
+    U8_NEXT(s, _u8_get_index, length, c); \
+} UPRV_BLOCK_MACRO_END
+
+/**
+ * Get a code point from a string at a random-access offset,
+ * without changing the offset.
+ * The offset may point to either the lead byte or one of the trail bytes
+ * for a code point, in which case the macro will read all of the bytes
+ * for the code point.
+ *
+ * The length can be negative for a NUL-terminated string.
+ *
+ * If the offset points to an illegal UTF-8 byte sequence, then
+ * c is set to U+FFFD.
+ * Iteration through a string is more efficient with U8_NEXT_UNSAFE or U8_NEXT_OR_FFFD.
+ *
+ * This macro does not distinguish between a real U+FFFD in the text
+ * and U+FFFD returned for an ill-formed sequence.
+ * Use U8_GET() if that distinction is important.
+ *
+ * @param s const uint8_t * string
+ * @param start int32_t starting string offset
+ * @param i int32_t string offset, must be start<=i<length
+ * @param length int32_t string length
+ * @param c output UChar32 variable, set to U+FFFD in case of an error
+ * @see U8_GET
+ * @stable ICU 51
+ */
+#define U8_GET_OR_FFFD(s, start, i, length, c) UPRV_BLOCK_MACRO_BEGIN { \
+    int32_t _u8_get_index=(i); \
+    U8_SET_CP_START(s, start, _u8_get_index); \
+    U8_NEXT_OR_FFFD(s, _u8_get_index, length, c); \
+} UPRV_BLOCK_MACRO_END
+
+/* definitions with forward iteration --------------------------------------- */
+
+/**
+ * Get a code point from a string at a code point boundary offset,
+ * and advance the offset to the next code point boundary.
+ * (Post-incrementing forward iteration.)
+ * "Unsafe" macro, assumes well-formed UTF-8.
+ *
+ * The offset may point to the lead byte of a multi-byte sequence,
+ * in which case the macro will read the whole sequence.
+ * The result is undefined if the offset points to a trail byte
+ * or an illegal UTF-8 sequence.
+ *
+ * @param s const uint8_t * string
+ * @param i string offset
+ * @param c output UChar32 variable
+ * @see U8_NEXT
+ * @stable ICU 2.4
+ */
+#define U8_NEXT_UNSAFE(s, i, c) UPRV_BLOCK_MACRO_BEGIN { \
+    (c)=(uint8_t)(s)[(i)++]; \
+    if(!U8_IS_SINGLE(c)) { \
+        if((c)<0xe0) { \
+            (c)=(((c)&0x1f)<<6)|((s)[(i)++]&0x3f); \
+        } else if((c)<0xf0) { \
+            /* no need for (c&0xf) because the upper bits are truncated after <<12 in the cast to (UChar) */ \
+            (c)=(UChar)(((c)<<12)|(((s)[i]&0x3f)<<6)|((s)[(i)+1]&0x3f)); \
+            (i)+=2; \
+        } else { \
+            (c)=(((c)&7)<<18)|(((s)[i]&0x3f)<<12)|(((s)[(i)+1]&0x3f)<<6)|((s)[(i)+2]&0x3f); \
+            (i)+=3; \
+        } \
+    } \
+} UPRV_BLOCK_MACRO_END
+
+/**
+ * Get a code point from a string at a code point boundary offset,
+ * and advance the offset to the next code point boundary.
+ * (Post-incrementing forward iteration.)
+ * "Safe" macro, checks for illegal sequences and for string boundaries.
+ *
+ * The length can be negative for a NUL-terminated string.
+ *
+ * The offset may point to the lead byte of a multi-byte sequence,
+ * in which case the macro will read the whole sequence.
+ * If the offset points to a trail byte or an illegal UTF-8 sequence, then
+ * c is set to a negative value.
+ *
+ * @param s const uint8_t * string
+ * @param i int32_t string offset, must be i<length
+ * @param length int32_t string length
+ * @param c output UChar32 variable, set to <0 in case of an error
+ * @see U8_NEXT_UNSAFE
+ * @stable ICU 2.4
+ */
+#define U8_NEXT(s, i, length, c) U8_INTERNAL_NEXT_OR_SUB(s, i, length, c, U_SENTINEL)
+
+/**
+ * Get a code point from a string at a code point boundary offset,
+ * and advance the offset to the next code point boundary.
+ * (Post-incrementing forward iteration.)
+ * "Safe" macro, checks for illegal sequences and for string boundaries.
+ *
+ * The length can be negative for a NUL-terminated string.
+ *
+ * The offset may point to the lead byte of a multi-byte sequence,
+ * in which case the macro will read the whole sequence.
+ * If the offset points to a trail byte or an illegal UTF-8 sequence, then
+ * c is set to U+FFFD.
+ *
+ * This macro does not distinguish between a real U+FFFD in the text
+ * and U+FFFD returned for an ill-formed sequence.
+ * Use U8_NEXT() if that distinction is important.
+ *
+ * @param s const uint8_t * string
+ * @param i int32_t string offset, must be i<length
+ * @param length int32_t string length
+ * @param c output UChar32 variable, set to U+FFFD in case of an error
+ * @see U8_NEXT
+ * @stable ICU 51
+ */
+#define U8_NEXT_OR_FFFD(s, i, length, c) U8_INTERNAL_NEXT_OR_SUB(s, i, length, c, 0xfffd)
+
+/** @internal */
+#define U8_INTERNAL_NEXT_OR_SUB(s, i, length, c, sub) UPRV_BLOCK_MACRO_BEGIN { \
+    (c)=(uint8_t)(s)[(i)++]; \
+    if(!U8_IS_SINGLE(c)) { \
+        uint8_t __t = 0; \
+        if((i)!=(length) && \
+            /* fetch/validate/assemble all but last trail byte */ \
+            ((c)>=0xe0 ? \
+                ((c)<0xf0 ?  /* U+0800..U+FFFF except surrogates */ \
+                    U8_LEAD3_T1_BITS[(c)&=0xf]&(1<<((__t=(s)[i])>>5)) && \
+                    (__t&=0x3f, 1) \
+                :  /* U+10000..U+10FFFF */ \
+                    ((c)-=0xf0)<=4 && \
+                    U8_LEAD4_T1_BITS[(__t=(s)[i])>>4]&(1<<(c)) && \
+                    ((c)=((c)<<6)|(__t&0x3f), ++(i)!=(length)) && \
+                    (__t=(s)[i]-0x80)<=0x3f) && \
+                /* valid second-to-last trail byte */ \
+                ((c)=((c)<<6)|__t, ++(i)!=(length)) \
+            :  /* U+0080..U+07FF */ \
+                (c)>=0xc2 && ((c)&=0x1f, 1)) && \
+            /* last trail byte */ \
+            (__t=(s)[i]-0x80)<=0x3f && \
+            ((c)=((c)<<6)|__t, ++(i), 1)) { \
+        } else { \
+            (c)=(sub);  /* ill-formed*/ \
+        } \
+    } \
+} UPRV_BLOCK_MACRO_END
+
+/**
+ * Append a code point to a string, overwriting 1 to 4 bytes.
+ * The offset points to the current end of the string contents
+ * and is advanced (post-increment).
+ * "Unsafe" macro, assumes a valid code point and sufficient space in the string.
+ * Otherwise, the result is undefined.
+ *
+ * @param s const uint8_t * string buffer
+ * @param i string offset
+ * @param c code point to append
+ * @see U8_APPEND
+ * @stable ICU 2.4
+ */
+#define U8_APPEND_UNSAFE(s, i, c) UPRV_BLOCK_MACRO_BEGIN { \
+    uint32_t __uc=(c); \
+    if(__uc<=0x7f) { \
+        (s)[(i)++]=(uint8_t)__uc; \
+    } else { \
+        if(__uc<=0x7ff) { \
+            (s)[(i)++]=(uint8_t)((__uc>>6)|0xc0); \
+        } else { \
+            if(__uc<=0xffff) { \
+                (s)[(i)++]=(uint8_t)((__uc>>12)|0xe0); \
+            } else { \
+                (s)[(i)++]=(uint8_t)((__uc>>18)|0xf0); \
+                (s)[(i)++]=(uint8_t)(((__uc>>12)&0x3f)|0x80); \
+            } \
+            (s)[(i)++]=(uint8_t)(((__uc>>6)&0x3f)|0x80); \
+        } \
+        (s)[(i)++]=(uint8_t)((__uc&0x3f)|0x80); \
+    } \
+} UPRV_BLOCK_MACRO_END
+
+/**
+ * Append a code point to a string, overwriting 1 to 4 bytes.
+ * The offset points to the current end of the string contents
+ * and is advanced (post-increment).
+ * "Safe" macro, checks for a valid code point.
+ * If a non-ASCII code point is written, checks for sufficient space in the string.
+ * If the code point is not valid or trail bytes do not fit,
+ * then isError is set to true.
+ *
+ * @param s const uint8_t * string buffer
+ * @param i int32_t string offset, must be i<capacity
+ * @param capacity int32_t size of the string buffer
+ * @param c UChar32 code point to append
+ * @param isError output UBool set to true if an error occurs, otherwise not modified
+ * @see U8_APPEND_UNSAFE
+ * @stable ICU 2.4
+ */
+#define U8_APPEND(s, i, capacity, c, isError) UPRV_BLOCK_MACRO_BEGIN { \
+    uint32_t __uc=(c); \
+    if(__uc<=0x7f) { \
+        (s)[(i)++]=(uint8_t)__uc; \
+    } else if(__uc<=0x7ff && (i)+1<(capacity)) { \
+        (s)[(i)++]=(uint8_t)((__uc>>6)|0xc0); \
+        (s)[(i)++]=(uint8_t)((__uc&0x3f)|0x80); \
+    } else if((__uc<=0xd7ff || (0xe000<=__uc && __uc<=0xffff)) && (i)+2<(capacity)) { \
+        (s)[(i)++]=(uint8_t)((__uc>>12)|0xe0); \
+        (s)[(i)++]=(uint8_t)(((__uc>>6)&0x3f)|0x80); \
+        (s)[(i)++]=(uint8_t)((__uc&0x3f)|0x80); \
+    } else if(0xffff<__uc && __uc<=0x10ffff && (i)+3<(capacity)) { \
+        (s)[(i)++]=(uint8_t)((__uc>>18)|0xf0); \
+        (s)[(i)++]=(uint8_t)(((__uc>>12)&0x3f)|0x80); \
+        (s)[(i)++]=(uint8_t)(((__uc>>6)&0x3f)|0x80); \
+        (s)[(i)++]=(uint8_t)((__uc&0x3f)|0x80); \
+    } else { \
+        (isError)=true; \
+    } \
+} UPRV_BLOCK_MACRO_END
+
+/**
+ * Advance the string offset from one code point boundary to the next.
+ * (Post-incrementing iteration.)
+ * "Unsafe" macro, assumes well-formed UTF-8.
+ *
+ * @param s const uint8_t * string
+ * @param i string offset
+ * @see U8_FWD_1
+ * @stable ICU 2.4
+ */
+#define U8_FWD_1_UNSAFE(s, i) UPRV_BLOCK_MACRO_BEGIN { \
+    (i)+=1+U8_COUNT_TRAIL_BYTES_UNSAFE((s)[i]); \
+} UPRV_BLOCK_MACRO_END
+
+/**
+ * Advance the string offset from one code point boundary to the next.
+ * (Post-incrementing iteration.)
+ * "Safe" macro, checks for illegal sequences and for string boundaries.
+ *
+ * The length can be negative for a NUL-terminated string.
+ *
+ * @param s const uint8_t * string
+ * @param i int32_t string offset, must be i<length
+ * @param length int32_t string length
+ * @see U8_FWD_1_UNSAFE
+ * @stable ICU 2.4
+ */
+#define U8_FWD_1(s, i, length) UPRV_BLOCK_MACRO_BEGIN { \
+    uint8_t __b=(s)[(i)++]; \
+    if(U8_IS_LEAD(__b) && (i)!=(length)) { \
+        uint8_t __t1=(s)[i]; \
+        if((0xe0<=__b && __b<0xf0)) { \
+            if(U8_IS_VALID_LEAD3_AND_T1(__b, __t1) && \
+                    ++(i)!=(length) && U8_IS_TRAIL((s)[i])) { \
+                ++(i); \
+            } \
+        } else if(__b<0xe0) { \
+            if(U8_IS_TRAIL(__t1)) { \
+                ++(i); \
+            } \
+        } else /* c>=0xf0 */ { \
+            if(U8_IS_VALID_LEAD4_AND_T1(__b, __t1) && \
+                    ++(i)!=(length) && U8_IS_TRAIL((s)[i]) && \
+                    ++(i)!=(length) && U8_IS_TRAIL((s)[i])) { \
+                ++(i); \
+            } \
+        } \
+    } \
+} UPRV_BLOCK_MACRO_END
+
+/**
+ * Advance the string offset from one code point boundary to the n-th next one,
+ * i.e., move forward by n code points.
+ * (Post-incrementing iteration.)
+ * "Unsafe" macro, assumes well-formed UTF-8.
+ *
+ * @param s const uint8_t * string
+ * @param i string offset
+ * @param n number of code points to skip
+ * @see U8_FWD_N
+ * @stable ICU 2.4
+ */
+#define U8_FWD_N_UNSAFE(s, i, n) UPRV_BLOCK_MACRO_BEGIN { \
+    int32_t __N=(n); \
+    while(__N>0) { \
+        U8_FWD_1_UNSAFE(s, i); \
+        --__N; \
+    } \
+} UPRV_BLOCK_MACRO_END
+
+/**
+ * Advance the string offset from one code point boundary to the n-th next one,
+ * i.e., move forward by n code points.
+ * (Post-incrementing iteration.)
+ * "Safe" macro, checks for illegal sequences and for string boundaries.
+ *
+ * The length can be negative for a NUL-terminated string.
+ *
+ * @param s const uint8_t * string
+ * @param i int32_t string offset, must be i<length
+ * @param length int32_t string length
+ * @param n number of code points to skip
+ * @see U8_FWD_N_UNSAFE
+ * @stable ICU 2.4
+ */
+#define U8_FWD_N(s, i, length, n) UPRV_BLOCK_MACRO_BEGIN { \
+    int32_t __N=(n); \
+    while(__N>0 && ((i)<(length) || ((length)<0 && (s)[i]!=0))) { \
+        U8_FWD_1(s, i, length); \
+        --__N; \
+    } \
+} UPRV_BLOCK_MACRO_END
+
+/**
+ * Adjust a random-access offset to a code point boundary
+ * at the start of a code point.
+ * If the offset points to a UTF-8 trail byte,
+ * then the offset is moved backward to the corresponding lead byte.
+ * Otherwise, it is not modified.
+ * "Unsafe" macro, assumes well-formed UTF-8.
+ *
+ * @param s const uint8_t * string
+ * @param i string offset
+ * @see U8_SET_CP_START
+ * @stable ICU 2.4
+ */
+#define U8_SET_CP_START_UNSAFE(s, i) UPRV_BLOCK_MACRO_BEGIN { \
+    while(U8_IS_TRAIL((s)[i])) { --(i); } \
+} UPRV_BLOCK_MACRO_END
+
+/**
+ * Adjust a random-access offset to a code point boundary
+ * at the start of a code point.
+ * If the offset points to a UTF-8 trail byte,
+ * then the offset is moved backward to the corresponding lead byte.
+ * Otherwise, it is not modified.
+ *
+ * "Safe" macro, checks for illegal sequences and for string boundaries.
+ * Unlike U8_TRUNCATE_IF_INCOMPLETE(), this macro always reads s[i].
+ *
+ * @param s const uint8_t * string
+ * @param start int32_t starting string offset (usually 0)
+ * @param i int32_t string offset, must be start<=i
+ * @see U8_SET_CP_START_UNSAFE
+ * @see U8_TRUNCATE_IF_INCOMPLETE
+ * @stable ICU 2.4
+ */
+#define U8_SET_CP_START(s, start, i) UPRV_BLOCK_MACRO_BEGIN { \
+    if(U8_IS_TRAIL((s)[(i)])) { \
+        (i)=utf8_back1SafeBody(s, start, (i)); \
+    } \
+} UPRV_BLOCK_MACRO_END
+
+/**
+ * If the string ends with a UTF-8 byte sequence that is valid so far
+ * but incomplete, then reduce the length of the string to end before
+ * the lead byte of that incomplete sequence.
+ * For example, if the string ends with E1 80, the length is reduced by 2.
+ *
+ * In all other cases (the string ends with a complete sequence, or it is not
+ * possible for any further trail byte to extend the trailing sequence)
+ * the length remains unchanged.
+ *
+ * Useful for processing text split across multiple buffers
+ * (save the incomplete sequence for later)
+ * and for optimizing iteration
+ * (check for string length only once per character).
+ *
+ * "Safe" macro, checks for illegal sequences and for string boundaries.
+ * Unlike U8_SET_CP_START(), this macro never reads s[length].
+ *
+ * (In UTF-16, simply check for U16_IS_LEAD(last code unit).)
+ *
+ * @param s const uint8_t * string
+ * @param start int32_t starting string offset (usually 0)
+ * @param length int32_t string length (usually start<=length)
+ * @see U8_SET_CP_START
+ * @stable ICU 61
+ */
+#define U8_TRUNCATE_IF_INCOMPLETE(s, start, length) UPRV_BLOCK_MACRO_BEGIN { \
+    if((length)>(start)) { \
+        uint8_t __b1=s[(length)-1]; \
+        if(U8_IS_SINGLE(__b1)) { \
+            /* common ASCII character */ \
+        } else if(U8_IS_LEAD(__b1)) { \
+            --(length); \
+        } else if(U8_IS_TRAIL(__b1) && ((length)-2)>=(start)) { \
+            uint8_t __b2=s[(length)-2]; \
+            if(0xe0<=__b2 && __b2<=0xf4) { \
+                if(__b2<0xf0 ? U8_IS_VALID_LEAD3_AND_T1(__b2, __b1) : \
+                        U8_IS_VALID_LEAD4_AND_T1(__b2, __b1)) { \
+                    (length)-=2; \
+                } \
+            } else if(U8_IS_TRAIL(__b2) && ((length)-3)>=(start)) { \
+                uint8_t __b3=s[(length)-3]; \
+                if(0xf0<=__b3 && __b3<=0xf4 && U8_IS_VALID_LEAD4_AND_T1(__b3, __b2)) { \
+                    (length)-=3; \
+                } \
+            } \
+        } \
+    } \
+} UPRV_BLOCK_MACRO_END
+
+/* definitions with backward iteration -------------------------------------- */
+
+/**
+ * Move the string offset from one code point boundary to the previous one
+ * and get the code point between them.
+ * (Pre-decrementing backward iteration.)
+ * "Unsafe" macro, assumes well-formed UTF-8.
+ *
+ * The input offset may be the same as the string length.
+ * If the offset is behind a multi-byte sequence, then the macro will read
+ * the whole sequence.
+ * If the offset is behind a lead byte, then that itself
+ * will be returned as the code point.
+ * The result is undefined if the offset is behind an illegal UTF-8 sequence.
+ *
+ * @param s const uint8_t * string
+ * @param i string offset
+ * @param c output UChar32 variable
+ * @see U8_PREV
+ * @stable ICU 2.4
+ */
+#define U8_PREV_UNSAFE(s, i, c) UPRV_BLOCK_MACRO_BEGIN { \
+    (c)=(uint8_t)(s)[--(i)]; \
+    if(U8_IS_TRAIL(c)) { \
+        uint8_t __b, __count=1, __shift=6; \
+\
+        /* c is a trail byte */ \
+        (c)&=0x3f; \
+        for(;;) { \
+            __b=(s)[--(i)]; \
+            if(__b>=0xc0) { \
+                U8_MASK_LEAD_BYTE(__b, __count); \
+                (c)|=(UChar32)__b<<__shift; \
+                break; \
+            } else { \
+                (c)|=(UChar32)(__b&0x3f)<<__shift; \
+                ++__count; \
+                __shift+=6; \
+            } \
+        } \
+    } \
+} UPRV_BLOCK_MACRO_END
+
+/**
+ * Move the string offset from one code point boundary to the previous one
+ * and get the code point between them.
+ * (Pre-decrementing backward iteration.)
+ * "Safe" macro, checks for illegal sequences and for string boundaries.
+ *
+ * The input offset may be the same as the string length.
+ * If the offset is behind a multi-byte sequence, then the macro will read
+ * the whole sequence.
+ * If the offset is behind a lead byte, then that itself
+ * will be returned as the code point.
+ * If the offset is behind an illegal UTF-8 sequence, then c is set to a negative value.
+ *
+ * @param s const uint8_t * string
+ * @param start int32_t starting string offset (usually 0)
+ * @param i int32_t string offset, must be start<i
+ * @param c output UChar32 variable, set to <0 in case of an error
+ * @see U8_PREV_UNSAFE
+ * @stable ICU 2.4
+ */
+#define U8_PREV(s, start, i, c) UPRV_BLOCK_MACRO_BEGIN { \
+    (c)=(uint8_t)(s)[--(i)]; \
+    if(!U8_IS_SINGLE(c)) { \
+        (c)=utf8_prevCharSafeBody((const uint8_t *)s, start, &(i), c, -1); \
+    } \
+} UPRV_BLOCK_MACRO_END
+
+/**
+ * Move the string offset from one code point boundary to the previous one
+ * and get the code point between them.
+ * (Pre-decrementing backward iteration.)
+ * "Safe" macro, checks for illegal sequences and for string boundaries.
+ *
+ * The input offset may be the same as the string length.
+ * If the offset is behind a multi-byte sequence, then the macro will read
+ * the whole sequence.
+ * If the offset is behind a lead byte, then that itself
+ * will be returned as the code point.
+ * If the offset is behind an illegal UTF-8 sequence, then c is set to U+FFFD.
+ *
+ * This macro does not distinguish between a real U+FFFD in the text
+ * and U+FFFD returned for an ill-formed sequence.
+ * Use U8_PREV() if that distinction is important.
+ *
+ * @param s const uint8_t * string
+ * @param start int32_t starting string offset (usually 0)
+ * @param i int32_t string offset, must be start<i
+ * @param c output UChar32 variable, set to U+FFFD in case of an error
+ * @see U8_PREV
+ * @stable ICU 51
+ */
+#define U8_PREV_OR_FFFD(s, start, i, c) UPRV_BLOCK_MACRO_BEGIN { \
+    (c)=(uint8_t)(s)[--(i)]; \
+    if(!U8_IS_SINGLE(c)) { \
+        (c)=utf8_prevCharSafeBody((const uint8_t *)s, start, &(i), c, -3); \
+    } \
+} UPRV_BLOCK_MACRO_END
+
+/**
+ * Move the string offset from one code point boundary to the previous one.
+ * (Pre-decrementing backward iteration.)
+ * The input offset may be the same as the string length.
+ * "Unsafe" macro, assumes well-formed UTF-8.
+ *
+ * @param s const uint8_t * string
+ * @param i string offset
+ * @see U8_BACK_1
+ * @stable ICU 2.4
+ */
+#define U8_BACK_1_UNSAFE(s, i) UPRV_BLOCK_MACRO_BEGIN { \
+    while(U8_IS_TRAIL((s)[--(i)])) {} \
+} UPRV_BLOCK_MACRO_END
+
+/**
+ * Move the string offset from one code point boundary to the previous one.
+ * (Pre-decrementing backward iteration.)
+ * The input offset may be the same as the string length.
+ * "Safe" macro, checks for illegal sequences and for string boundaries.
+ *
+ * @param s const uint8_t * string
+ * @param start int32_t starting string offset (usually 0)
+ * @param i int32_t string offset, must be start<i
+ * @see U8_BACK_1_UNSAFE
+ * @stable ICU 2.4
+ */
+#define U8_BACK_1(s, start, i) UPRV_BLOCK_MACRO_BEGIN { \
+    if(U8_IS_TRAIL((s)[--(i)])) { \
+        (i)=utf8_back1SafeBody(s, start, (i)); \
+    } \
+} UPRV_BLOCK_MACRO_END
+
+/**
+ * Move the string offset from one code point boundary to the n-th one before it,
+ * i.e., move backward by n code points.
+ * (Pre-decrementing backward iteration.)
+ * The input offset may be the same as the string length.
+ * "Unsafe" macro, assumes well-formed UTF-8.
+ *
+ * @param s const uint8_t * string
+ * @param i string offset
+ * @param n number of code points to skip
+ * @see U8_BACK_N
+ * @stable ICU 2.4
+ */
+#define U8_BACK_N_UNSAFE(s, i, n) UPRV_BLOCK_MACRO_BEGIN { \
+    int32_t __N=(n); \
+    while(__N>0) { \
+        U8_BACK_1_UNSAFE(s, i); \
+        --__N; \
+    } \
+} UPRV_BLOCK_MACRO_END
+
+/**
+ * Move the string offset from one code point boundary to the n-th one before it,
+ * i.e., move backward by n code points.
+ * (Pre-decrementing backward iteration.)
+ * The input offset may be the same as the string length.
+ * "Safe" macro, checks for illegal sequences and for string boundaries.
+ *
+ * @param s const uint8_t * string
+ * @param start int32_t index of the start of the string
+ * @param i int32_t string offset, must be start<i
+ * @param n number of code points to skip
+ * @see U8_BACK_N_UNSAFE
+ * @stable ICU 2.4
+ */
+#define U8_BACK_N(s, start, i, n) UPRV_BLOCK_MACRO_BEGIN { \
+    int32_t __N=(n); \
+    while(__N>0 && (i)>(start)) { \
+        U8_BACK_1(s, start, i); \
+        --__N; \
+    } \
+} UPRV_BLOCK_MACRO_END
+
+/**
+ * Adjust a random-access offset to a code point boundary after a code point.
+ * If the offset is behind a partial multi-byte sequence,
+ * then the offset is incremented to behind the whole sequence.
+ * Otherwise, it is not modified.
+ * The input offset may be the same as the string length.
+ * "Unsafe" macro, assumes well-formed UTF-8.
+ *
+ * @param s const uint8_t * string
+ * @param i string offset
+ * @see U8_SET_CP_LIMIT
+ * @stable ICU 2.4
+ */
+#define U8_SET_CP_LIMIT_UNSAFE(s, i) UPRV_BLOCK_MACRO_BEGIN { \
+    U8_BACK_1_UNSAFE(s, i); \
+    U8_FWD_1_UNSAFE(s, i); \
+} UPRV_BLOCK_MACRO_END
+
+/**
+ * Adjust a random-access offset to a code point boundary after a code point.
+ * If the offset is behind a partial multi-byte sequence,
+ * then the offset is incremented to behind the whole sequence.
+ * Otherwise, it is not modified.
+ * The input offset may be the same as the string length.
+ * "Safe" macro, checks for illegal sequences and for string boundaries.
+ *
+ * The length can be negative for a NUL-terminated string.
+ *
+ * @param s const uint8_t * string
+ * @param start int32_t starting string offset (usually 0)
+ * @param i int32_t string offset, must be start<=i<=length
+ * @param length int32_t string length
+ * @see U8_SET_CP_LIMIT_UNSAFE
+ * @stable ICU 2.4
+ */
+#define U8_SET_CP_LIMIT(s, start, i, length) UPRV_BLOCK_MACRO_BEGIN { \
+    if((start)<(i) && ((i)<(length) || (length)<0)) { \
+        U8_BACK_1(s, start, i); \
+        U8_FWD_1(s, i, length); \
+    } \
+} UPRV_BLOCK_MACRO_END
+
+#endif

thirdparty/icu4c/common/unicode/utrace.h

@@ -0,0 +1,506 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+/*
+*******************************************************************************
+*
+*   Copyright (C) 2003-2013, International Business Machines
+*   Corporation and others.  All Rights Reserved.
+*
+*******************************************************************************
+*   file name:  utrace.h
+*   encoding:   UTF-8
+*   tab size:   8 (not used)
+*   indentation:4
+*
+*   created on: 2003aug06
+*   created by: Markus W. Scherer
+*
+*   Definitions for ICU tracing/logging.
+*
+*/
+
+#ifndef __UTRACE_H__
+#define __UTRACE_H__
+
+#include <stdarg.h>
+#include "unicode/utypes.h"
+
+/**
+ * \file
+ * \brief C API:  Definitions for ICU tracing/logging. 
+ *
+ * This provides API for debugging the internals of ICU without the use of
+ * a traditional debugger.
+ *
+ * By default, tracing is disabled in ICU. If you need to debug ICU with 
+ * tracing, please compile ICU with the --enable-tracing configure option.
+ */
+ 
+U_CDECL_BEGIN
+
+/**
+ * Trace severity levels.  Higher levels increase the verbosity of the trace output.
+ * @see utrace_setLevel
+ * @stable ICU 2.8
+ */
+typedef enum UTraceLevel {
+    /** Disable all tracing  @stable ICU 2.8*/
+    UTRACE_OFF=-1,
+    /** Trace error conditions only  @stable ICU 2.8*/
+    UTRACE_ERROR=0,
+    /** Trace errors and warnings  @stable ICU 2.8*/
+    UTRACE_WARNING=3,
+    /** Trace opens and closes of ICU services  @stable ICU 2.8*/
+    UTRACE_OPEN_CLOSE=5,
+    /** Trace an intermediate number of ICU operations  @stable ICU 2.8*/
+    UTRACE_INFO=7,
+    /** Trace the maximum number of ICU operations  @stable ICU 2.8*/
+    UTRACE_VERBOSE=9
+} UTraceLevel;
+
+/**
+ *  These are the ICU functions that will be traced when tracing is enabled.
+ *  @stable ICU 2.8
+ */
+typedef enum UTraceFunctionNumber {
+    UTRACE_FUNCTION_START=0,
+    UTRACE_U_INIT=UTRACE_FUNCTION_START,
+    UTRACE_U_CLEANUP,
+
+#ifndef U_HIDE_DEPRECATED_API
+    /**
+     * One more than the highest normal collation trace location.
+     * @deprecated ICU 58 The numeric value may change over time, see ICU ticket #12420.
+     */
+    UTRACE_FUNCTION_LIMIT,
+#endif  // U_HIDE_DEPRECATED_API
+
+    UTRACE_CONVERSION_START=0x1000,
+    UTRACE_UCNV_OPEN=UTRACE_CONVERSION_START,
+    UTRACE_UCNV_OPEN_PACKAGE,
+    UTRACE_UCNV_OPEN_ALGORITHMIC,
+    UTRACE_UCNV_CLONE,
+    UTRACE_UCNV_CLOSE,
+    UTRACE_UCNV_FLUSH_CACHE,
+    UTRACE_UCNV_LOAD,
+    UTRACE_UCNV_UNLOAD,
+
+#ifndef U_HIDE_DEPRECATED_API
+    /**
+     * One more than the highest normal collation trace location.
+     * @deprecated ICU 58 The numeric value may change over time, see ICU ticket #12420.
+     */
+    UTRACE_CONVERSION_LIMIT,
+#endif  // U_HIDE_DEPRECATED_API
+
+    UTRACE_COLLATION_START=0x2000,
+    UTRACE_UCOL_OPEN=UTRACE_COLLATION_START,
+    UTRACE_UCOL_CLOSE,
+    UTRACE_UCOL_STRCOLL,
+    UTRACE_UCOL_GET_SORTKEY,
+    UTRACE_UCOL_GETLOCALE,
+    UTRACE_UCOL_NEXTSORTKEYPART,
+    UTRACE_UCOL_STRCOLLITER,
+    UTRACE_UCOL_OPEN_FROM_SHORT_STRING,
+    UTRACE_UCOL_STRCOLLUTF8, /**< @stable ICU 50 */
+
+#ifndef U_HIDE_DEPRECATED_API
+    /**
+     * One more than the highest normal collation trace location.
+     * @deprecated ICU 58 The numeric value may change over time, see ICU ticket #12420.
+     */
+    UTRACE_COLLATION_LIMIT,
+#endif  // U_HIDE_DEPRECATED_API
+
+    /**
+     * The lowest resource/data location.
+     * @stable ICU 65
+     */
+    UTRACE_UDATA_START=0x3000,
+
+    /**
+     * Indicates that a value was read from a resource bundle. Provides three
+     * C-style strings to UTraceData: type, file name, and resource path. The
+     * possible types are:
+     *
+     * - "string" (a string value was accessed)
+     * - "binary" (a binary value was accessed)
+     * - "intvector" (a integer vector value was accessed)
+     * - "int" (a signed integer value was accessed)
+     * - "uint" (a unsigned integer value was accessed)
+     * - "get" (a path was loaded, but the value was not accessed)
+     * - "getalias" (a path was loaded, and an alias was resolved)
+     *
+     * @stable ICU 65
+     */
+    UTRACE_UDATA_RESOURCE=UTRACE_UDATA_START,
+
+    /**
+     * Indicates that a resource bundle was opened.
+     *
+     * Provides one C-style string to UTraceData: file name.
+     * @stable ICU 65
+     */
+    UTRACE_UDATA_BUNDLE,
+
+    /**
+     * Indicates that a data file was opened, but not *.res files.
+     *
+     * Provides one C-style string to UTraceData: file name.
+     *
+     * @stable ICU 65
+     */
+    UTRACE_UDATA_DATA_FILE,
+
+    /**
+     * Indicates that a *.res file was opened.
+     *
+     * This differs from UTRACE_UDATA_BUNDLE because a res file is typically
+     * opened only once per application runtime, but the bundle corresponding
+     * to that res file may be opened many times.
+     *
+     * Provides one C-style string to UTraceData: file name.
+     *
+     * @stable ICU 65
+     */
+    UTRACE_UDATA_RES_FILE,
+
+#ifndef U_HIDE_INTERNAL_API
+    /**
+     * One more than the highest normal resource/data trace location.
+     * @internal The numeric value may change over time, see ICU ticket #12420.
+     */
+    UTRACE_RES_DATA_LIMIT,
+#endif  // U_HIDE_INTERNAL_API
+
+    /**
+     * The lowest break iterator location.
+     * @stable ICU 67
+     */
+    UTRACE_UBRK_START=0x4000,
+
+    /**
+     * Indicates that a character instance of break iterator was created.
+     *
+     * @stable ICU 67
+     */
+    UTRACE_UBRK_CREATE_CHARACTER = UTRACE_UBRK_START,
+
+    /**
+     * Indicates that a word instance of break iterator was created.
+     *
+     * @stable ICU 67
+     */
+    UTRACE_UBRK_CREATE_WORD,
+
+    /**
+     * Indicates that a line instance of break iterator was created.
+     *
+     * Provides one C-style string to UTraceData: the lb value ("",
+     * "loose", "strict", or "normal").
+     *
+     * @stable ICU 67
+     */
+    UTRACE_UBRK_CREATE_LINE,
+
+    /**
+     * Indicates that a sentence instance of break iterator was created.
+     *
+     * @stable ICU 67
+     */
+    UTRACE_UBRK_CREATE_SENTENCE,
+
+    /**
+     * Indicates that a title instance of break iterator was created.
+     *
+     * @stable ICU 67
+     */
+    UTRACE_UBRK_CREATE_TITLE,
+
+    /**
+     * Indicates that an internal dictionary break engine was created.
+     *
+     * Provides one C-style string to UTraceData: the script code of what
+     * the break engine cover ("Hani", "Khmr", "Laoo", "Mymr", or "Thai").
+     *
+     * @stable ICU 67
+     */
+    UTRACE_UBRK_CREATE_BREAK_ENGINE,
+
+#ifndef U_HIDE_INTERNAL_API
+    /**
+     * One more than the highest normal break iterator trace location.
+     * @internal The numeric value may change over time, see ICU ticket #12420.
+     */
+    UTRACE_UBRK_LIMIT,
+#endif  // U_HIDE_INTERNAL_API
+
+} UTraceFunctionNumber;
+
+/**
+ * Setter for the trace level.
+ * @param traceLevel A UTraceLevel value.
+ * @stable ICU 2.8
+ */
+U_CAPI void U_EXPORT2
+utrace_setLevel(int32_t traceLevel);
+
+/**
+ * Getter for the trace level.
+ * @return The UTraceLevel value being used by ICU.
+ * @stable ICU 2.8
+ */
+U_CAPI int32_t U_EXPORT2
+utrace_getLevel(void);
+
+/* Trace function pointers types  ----------------------------- */
+
+/**
+  *  Type signature for the trace function to be called when entering a function.
+  *  @param context value supplied at the time the trace functions are set.
+  *  @param fnNumber Enum value indicating the ICU function being entered.
+  *  @stable ICU 2.8
+  */
+typedef void U_CALLCONV
+UTraceEntry(const void *context, int32_t fnNumber);
+
+/**
+  *  Type signature for the trace function to be called when exiting from a function.
+  *  @param context value supplied at the time the trace functions are set.
+  *  @param fnNumber Enum value indicating the ICU function being exited.
+  *  @param fmt     A formatting string that describes the number and types
+  *                 of arguments included with the variable args.  The fmt
+  *                 string has the same form as the utrace_vformat format
+  *                 string.
+  *  @param args    A variable arguments list.  Contents are described by
+  *                 the fmt parameter.
+  *  @see   utrace_vformat
+  *  @stable ICU 2.8
+  */
+typedef void U_CALLCONV
+UTraceExit(const void *context, int32_t fnNumber, 
+           const char *fmt, va_list args);
+
+/**
+  *  Type signature for the trace function to be called from within an ICU function
+  *  to display data or messages.
+  *  @param context  value supplied at the time the trace functions are set.
+  *  @param fnNumber Enum value indicating the ICU function being exited.
+  *  @param level    The current tracing level
+  *  @param fmt      A format string describing the tracing data that is supplied
+  *                  as variable args
+  *  @param args     The data being traced, passed as variable args.
+  *  @stable ICU 2.8
+  */
+typedef void U_CALLCONV
+UTraceData(const void *context, int32_t fnNumber, int32_t level,
+           const char *fmt, va_list args);
+
+/**
+  *  Set ICU Tracing functions.  Installs application-provided tracing
+  *  functions into ICU.  After doing this, subsequent ICU operations
+  *  will call back to the installed functions, providing a trace
+  *  of the use of ICU.  Passing a NULL pointer for a tracing function
+  *  is allowed, and inhibits tracing action at points where that function
+  *  would be called.
+  *  <p>
+  *  Tracing and Threads:  Tracing functions are global to a process, and
+  *  will be called in response to ICU operations performed by any
+  *  thread.  If tracing of an individual thread is desired, the
+  *  tracing functions must themselves filter by checking that the
+  *  current thread is the desired thread.
+  *
+  *  @param context an uninterpreted pointer.  Whatever is passed in
+  *                 here will in turn be passed to each of the tracing
+  *                 functions UTraceEntry, UTraceExit and UTraceData.
+  *                 ICU does not use or alter this pointer.
+  *  @param e       Callback function to be called on entry to a 
+  *                 a traced ICU function.
+  *  @param x       Callback function to be called on exit from a
+  *                 traced ICU function.
+  *  @param d       Callback function to be called from within a 
+  *                 traced ICU function, for the purpose of providing
+  *                 data to the trace.
+  *
+  *  @stable ICU 2.8
+  */
+U_CAPI void U_EXPORT2
+utrace_setFunctions(const void *context,
+                    UTraceEntry *e, UTraceExit *x, UTraceData *d);
+
+/**
+  * Get the currently installed ICU tracing functions.   Note that a null function
+  *   pointer will be returned if no trace function has been set.
+  *
+  * @param context  The currently installed tracing context.
+  * @param e        The currently installed UTraceEntry function.
+  * @param x        The currently installed UTraceExit function.
+  * @param d        The currently installed UTraceData function.
+  * @stable ICU 2.8
+  */
+U_CAPI void U_EXPORT2
+utrace_getFunctions(const void **context,
+                    UTraceEntry **e, UTraceExit **x, UTraceData **d);
+
+
+
+/*
+ *
+ * ICU trace format string syntax
+ *
+ * Format Strings are passed to UTraceData functions, and define the
+ * number and types of the trace data being passed on each call.
+ *
+ * The UTraceData function, which is supplied by the application,
+ * not by ICU, can either forward the trace data (passed via
+ * varargs) and the format string back to ICU for formatting into
+ * a displayable string, or it can interpret the format itself,
+ * and do as it wishes with the trace data.
+ *
+ *
+ * Goals for the format string
+ * - basic data output
+ * - easy to use for trace programmer
+ * - sufficient provision for data types for trace output readability
+ * - well-defined types and binary portable APIs
+ *
+ * Non-goals
+ * - printf compatibility
+ * - fancy formatting
+ * - argument reordering and other internationalization features
+ *
+ * ICU trace format strings contain plain text with argument inserts,
+ * much like standard printf format strings.
+ * Each insert begins with a '%', then optionally contains a 'v',
+ * then exactly one type character.
+ * Two '%' in a row represent a '%' instead of an insert.
+ * The trace format strings need not have \n at the end.
+ *
+ *
+ * Types
+ * -----
+ *
+ * Type characters:
+ * - c A char character in the default codepage.
+ * - s A NUL-terminated char * string in the default codepage.
+ * - S A UChar * string.  Requires two params, (ptr, length).  Length=-1 for nul term.
+ * - b A byte (8-bit integer).
+ * - h A 16-bit integer.  Also a 16 bit Unicode code unit.
+ * - d A 32-bit integer.  Also a 20 bit Unicode code point value. 
+ * - l A 64-bit integer.
+ * - p A data pointer.
+ *
+ * Vectors
+ * -------
+ *
+ * If the 'v' is not specified, then one item of the specified type
+ * is passed in.
+ * If the 'v' (for "vector") is specified, then a vector of items of the
+ * specified type is passed in, via a pointer to the first item
+ * and an int32_t value for the length of the vector.
+ * Length==-1 means zero or NUL termination.  Works for vectors of all types.
+ *
+ * Note:  %vS is a vector of (UChar *) strings.  The strings must
+ *        be nul terminated as there is no way to provide a
+ *        separate length parameter for each string.  The length
+ *        parameter (required for all vectors) is the number of
+ *        strings, not the length of the strings.
+ *
+ * Examples
+ * --------
+ *
+ * These examples show the parameters that will be passed to an application's
+ *   UTraceData() function for various formats.
+ *
+ * - the precise formatting is up to the application!
+ * - the examples use type casts for arguments only to _show_ the types of
+ *   arguments without needing variable declarations in the examples;
+ *   the type casts will not be necessary in actual code
+ *
+ * UTraceDataFunc(context, fnNumber, level,
+ *              "There is a character %c in the string %s.",   // Format String 
+ *              (char)c, (const char *)s);                     // varargs parameters
+ * ->   There is a character 0x42 'B' in the string "Bravo".
+ *
+ * UTraceDataFunc(context, fnNumber, level,
+ *              "Vector of bytes %vb vector of chars %vc",
+ *              (const uint8_t *)bytes, (int32_t)bytesLength,
+ *              (const char *)chars, (int32_t)charsLength);
+ * ->  Vector of bytes
+ *      42 63 64 3f [4]
+ *     vector of chars
+ *      "Bcd?"[4]
+ *
+ * UTraceDataFunc(context, fnNumber, level,
+ *              "An int32_t %d and a whole bunch of them %vd",
+ *              (int32_t)-5, (const int32_t *)ints, (int32_t)intsLength);
+ * ->   An int32_t 0xfffffffb and a whole bunch of them
+ *      fffffffb 00000005 0000010a [3]
+ *
+ */
+
+
+
+/**
+  *  Trace output Formatter.  An application's UTraceData tracing functions may call
+  *                 back to this function to format the trace output in a
+  *                 human readable form.  Note that a UTraceData function may choose
+  *                 to not format the data;  it could, for example, save it in
+  *                 in the raw form it was received (more compact), leaving
+  *                 formatting for a later trace analysis tool.
+  *  @param outBuf  pointer to a buffer to receive the formatted output.  Output
+  *                 will be nul terminated if there is space in the buffer -
+  *                 if the length of the requested output < the output buffer size.
+  *  @param capacity  Length of the output buffer.
+  *  @param indent  Number of spaces to indent the output.  Intended to allow
+  *                 data displayed from nested functions to be indented for readability.
+  *  @param fmt     Format specification for the data to output
+  *  @param args    Data to be formatted.
+  *  @return        Length of formatted output, including the terminating NUL.
+  *                 If buffer capacity is insufficient, the required capacity is returned. 
+  *  @stable ICU 2.8
+  */
+U_CAPI int32_t U_EXPORT2
+utrace_vformat(char *outBuf, int32_t capacity,
+              int32_t indent, const char *fmt,  va_list args);
+
+/**
+  *  Trace output Formatter.  An application's UTraceData tracing functions may call
+  *                 this function to format any additional trace data, beyond that
+  *                 provided by default, in human readable form with the same
+  *                 formatting conventions used by utrace_vformat().
+  *  @param outBuf  pointer to a buffer to receive the formatted output.  Output
+  *                 will be nul terminated if there is space in the buffer -
+  *                 if the length of the requested output < the output buffer size.
+  *  @param capacity  Length of the output buffer.
+  *  @param indent  Number of spaces to indent the output.  Intended to allow
+  *                 data displayed from nested functions to be indented for readability.
+  *  @param fmt     Format specification for the data to output
+  *  @param ...     Data to be formatted.
+  *  @return        Length of formatted output, including the terminating NUL.
+  *                 If buffer capacity is insufficient, the required capacity is returned. 
+  *  @stable ICU 2.8
+  */
+U_CAPI int32_t U_EXPORT2
+utrace_format(char *outBuf, int32_t capacity,
+              int32_t indent, const char *fmt,  ...);
+
+
+
+/* Trace function numbers --------------------------------------------------- */
+
+/**
+ * Get the name of a function from its trace function number.
+ *
+ * @param fnNumber The trace number for an ICU function.
+ * @return The name string for the function.
+ *
+ * @see UTraceFunctionNumber
+ * @stable ICU 2.8
+ */
+U_CAPI const char * U_EXPORT2
+utrace_functionName(int32_t fnNumber);
+
+U_CDECL_END
+
+#endif

thirdparty/icu4c/common/unicode/utypes.h

@@ -0,0 +1,761 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+/*
+**********************************************************************
+*   Copyright (C) 1996-2016, International Business Machines
+*   Corporation and others.  All Rights Reserved.
+**********************************************************************
+*
+*  FILE NAME : UTYPES.H (formerly ptypes.h)
+*
+*   Date        Name        Description
+*   12/11/96    helena      Creation.
+*   02/27/97    aliu        Added typedefs for UClassID, int8, int16, int32,
+*                           uint8, uint16, and uint32.
+*   04/01/97    aliu        Added XP_CPLUSPLUS and modified to work under C as
+*                            well as C++.
+*                           Modified to use memcpy() for uprv_arrayCopy() fns.
+*   04/14/97    aliu        Added TPlatformUtilities.
+*   05/07/97    aliu        Added import/export specifiers (replacing the old
+*                           broken EXT_CLASS).  Added version number for our
+*                           code.  Cleaned up header.
+*    6/20/97    helena      Java class name change.
+*   08/11/98    stephen     UErrorCode changed from typedef to enum
+*   08/12/98    erm         Changed T_ANALYTIC_PACKAGE_VERSION to 3
+*   08/14/98    stephen     Added uprv_arrayCopy() for int8_t, int16_t, int32_t
+*   12/09/98    jfitz       Added BUFFER_OVERFLOW_ERROR (bug 1100066)
+*   04/20/99    stephen     Cleaned up & reworked for autoconf.
+*                           Renamed to utypes.h.
+*   05/05/99    stephen     Changed to use <inttypes.h>
+*   12/07/99    helena      Moved copyright notice string from ucnv_bld.h here.
+*******************************************************************************
+*/
+
+#ifndef UTYPES_H
+#define UTYPES_H
+
+
+#include "unicode/umachine.h"
+#include "unicode/uversion.h"
+#include "unicode/uconfig.h"
+#include <float.h>
+
+#if !U_NO_DEFAULT_INCLUDE_UTF_HEADERS
+#   include "unicode/utf.h"
+#endif
+
+/*!
+ * \file
+ * \brief Basic definitions for ICU, for both C and C++ APIs
+ *
+ * This file defines basic types, constants, and enumerations directly or
+ * indirectly by including other header files, especially utf.h for the
+ * basic character and string definitions and umachine.h for consistent
+ * integer and other types.
+ */
+
+/** @{ API visibility control */
+
+/**
+ * \def U_SHOW_CPLUSPLUS_API
+ * When defined to 1 (=default) and compiled with a C++ compiler, both C and C++ APIs are visible.
+ * Otherwise, only C APIs are visible; this is for C++ users who want to
+ * restrict their usage to binary stable C APIs exported by ICU DLLs.
+ * @internal
+ */
+/**
+ * \def U_SHOW_CPLUSPLUS_HEADER_API
+ * When defined to 1 (=default) and compiled with a C++ compiler, C++ header-only APIs are visible.
+ * This is for C++ users who restrict their usage to binary stable C APIs exported by ICU DLLs
+ * (U_SHOW_CPLUSPLUS_API=0)
+ * but who still want to use C++ header-only APIs which do not rely on ICU DLL exports.
+ * @internal
+ */
+#ifdef __cplusplus
+#   ifndef U_SHOW_CPLUSPLUS_API
+#       define U_SHOW_CPLUSPLUS_API 1
+#   endif
+#   ifndef U_SHOW_CPLUSPLUS_HEADER_API
+#       define U_SHOW_CPLUSPLUS_HEADER_API 1
+#   endif
+#else
+#   undef U_SHOW_CPLUSPLUS_API
+#   define U_SHOW_CPLUSPLUS_API 0
+#   undef U_SHOW_CPLUSPLUS_HEADER_API
+#   define U_SHOW_CPLUSPLUS_HEADER_API 0
+#endif
+
+/**
+ * \def U_HIDE_DRAFT_API
+ * Define this to 1 to request that draft API be "hidden"
+ * @internal
+ */
+/**
+ * \def U_HIDE_INTERNAL_API
+ * Define this to 1 to request that internal API be "hidden"
+ * @internal
+ */
+#if !U_DEFAULT_SHOW_DRAFT && !defined(U_SHOW_DRAFT_API)
+#define U_HIDE_DRAFT_API 1
+#endif
+#if !U_DEFAULT_SHOW_DRAFT && !defined(U_SHOW_INTERNAL_API)
+#define U_HIDE_INTERNAL_API 1
+#endif
+
+/** @} */
+
+/*===========================================================================*/
+/* ICUDATA naming scheme                                                     */
+/*===========================================================================*/
+
+/**
+ * \def U_ICUDATA_TYPE_LETTER
+ *
+ * This is a platform-dependent string containing one letter:
+ * - b for big-endian, ASCII-family platforms
+ * - l for little-endian, ASCII-family platforms
+ * - e for big-endian, EBCDIC-family platforms
+ * This letter is part of the common data file name.
+ * @stable ICU 2.0
+ */
+
+/**
+ * \def U_ICUDATA_TYPE_LITLETTER
+ * The non-string form of U_ICUDATA_TYPE_LETTER
+ * @stable ICU 2.0
+ */
+#if U_CHARSET_FAMILY
+#   if U_IS_BIG_ENDIAN
+   /* EBCDIC - should always be BE */
+#     define U_ICUDATA_TYPE_LETTER "e"
+#     define U_ICUDATA_TYPE_LITLETTER e
+#   else
+#     error "Don't know what to do with little endian EBCDIC!"
+#     define U_ICUDATA_TYPE_LETTER "x"
+#     define U_ICUDATA_TYPE_LITLETTER x
+#   endif
+#else
+#   if U_IS_BIG_ENDIAN
+      /* Big-endian ASCII */
+#     define U_ICUDATA_TYPE_LETTER "b"
+#     define U_ICUDATA_TYPE_LITLETTER b
+#   else
+      /* Little-endian ASCII */
+#     define U_ICUDATA_TYPE_LETTER "l"
+#     define U_ICUDATA_TYPE_LITLETTER l
+#   endif
+#endif
+
+/**
+ * A single string literal containing the icudata stub name. i.e. 'icudt18e' for
+ * ICU 1.8.x on EBCDIC, etc..
+ * @stable ICU 2.0
+ */
+#define U_ICUDATA_NAME    "icudt" U_ICU_VERSION_SHORT U_ICUDATA_TYPE_LETTER
+#ifndef U_HIDE_INTERNAL_API
+#define U_USRDATA_NAME    "usrdt" U_ICU_VERSION_SHORT U_ICUDATA_TYPE_LETTER  /**< @internal */
+#define U_USE_USRDATA     0  /**< @internal */
+#endif  /* U_HIDE_INTERNAL_API */
+
+/**
+ *  U_ICU_ENTRY_POINT is the name of the DLL entry point to the ICU data library.
+ *    Defined as a literal, not a string.
+ *    Tricky Preprocessor use - ## operator replaces macro parameters with the literal string
+ *                              from the corresponding macro invocation, _before_ other macro substitutions.
+ *                              Need a nested \#defines to get the actual version numbers rather than
+ *                              the literal text U_ICU_VERSION_MAJOR_NUM into the name.
+ *                              The net result will be something of the form
+ *                                  \#define U_ICU_ENTRY_POINT icudt19_dat
+ * @stable ICU 2.4
+ */
+#define U_ICUDATA_ENTRY_POINT  U_DEF2_ICUDATA_ENTRY_POINT(U_ICU_VERSION_MAJOR_NUM,U_LIB_SUFFIX_C_NAME)
+
+#ifndef U_HIDE_INTERNAL_API
+/**
+ * Do not use. Note that it's OK for the 2nd argument to be undefined (literal).
+ * @internal
+ */
+#define U_DEF2_ICUDATA_ENTRY_POINT(major,suff) U_DEF_ICUDATA_ENTRY_POINT(major,suff)
+
+/**
+ * Do not use.
+ * @internal
+ */
+#ifndef U_DEF_ICUDATA_ENTRY_POINT
+/* affected by symbol renaming. See platform.h */
+#ifndef U_LIB_SUFFIX_C_NAME
+#define U_DEF_ICUDATA_ENTRY_POINT(major, suff) icudt##major##_dat
+#else
+#define U_DEF_ICUDATA_ENTRY_POINT(major, suff) icudt##suff ## major##_dat
+#endif
+#endif
+#endif  /* U_HIDE_INTERNAL_API */
+
+/**
+ * \def NULL
+ * Define NULL if necessary, to nullptr for C++ and to ((void *)0) for C.
+ * @stable ICU 2.0
+ */
+#ifndef NULL
+#ifdef __cplusplus
+#define NULL    nullptr
+#else
+#define NULL    ((void *)0)
+#endif
+#endif
+
+/*===========================================================================*/
+/* Calendar/TimeZone data types                                              */
+/*===========================================================================*/
+
+/**
+ * Date and Time data type.
+ * This is a primitive data type that holds the date and time
+ * as the number of milliseconds since 1970-jan-01, 00:00 UTC.
+ * UTC leap seconds are ignored.
+ * @stable ICU 2.0
+ */
+typedef double UDate;
+
+/** The number of milliseconds per second @stable ICU 2.0 */
+#define U_MILLIS_PER_SECOND        (1000)
+/** The number of milliseconds per minute @stable ICU 2.0 */
+#define U_MILLIS_PER_MINUTE       (60000)
+/** The number of milliseconds per hour @stable ICU 2.0 */
+#define U_MILLIS_PER_HOUR       (3600000)
+/** The number of milliseconds per day @stable ICU 2.0 */
+#define U_MILLIS_PER_DAY       (86400000)
+
+/** 
+ * Maximum UDate value 
+ * @stable ICU 4.8 
+ */ 
+#define U_DATE_MAX DBL_MAX
+
+/**
+ * Minimum UDate value 
+ * @stable ICU 4.8 
+ */ 
+#define U_DATE_MIN -U_DATE_MAX
+
+/*===========================================================================*/
+/* Shared library/DLL import-export API control                              */
+/*===========================================================================*/
+
+/*
+ * Control of symbol import/export.
+ * ICU is separated into three libraries.
+ */
+
+/**
+ * \def U_COMBINED_IMPLEMENTATION
+ * Set to export library symbols from inside the ICU library
+ * when all of ICU is in a single library.
+ * This can be set as a compiler option while building ICU, and it
+ * needs to be the first one tested to override U_COMMON_API, U_I18N_API, etc.
+ * @stable ICU 2.0
+ */
+
+/**
+ * \def U_DATA_API
+ * Set to export library symbols from inside the stubdata library,
+ * and to import them from outside.
+ * @stable ICU 3.0
+ */
+
+/**
+ * \def U_COMMON_API
+ * Set to export library symbols from inside the common library,
+ * and to import them from outside.
+ * @stable ICU 2.0
+ */
+
+/**
+ * \def U_I18N_API
+ * Set to export library symbols from inside the i18n library,
+ * and to import them from outside.
+ * @stable ICU 2.0
+ */
+
+/**
+ * \def U_LAYOUT_API
+ * Set to export library symbols from inside the layout engine library,
+ * and to import them from outside.
+ * @stable ICU 2.0
+ */
+
+/**
+ * \def U_LAYOUTEX_API
+ * Set to export library symbols from inside the layout extensions library,
+ * and to import them from outside.
+ * @stable ICU 2.6
+ */
+
+/**
+ * \def U_IO_API
+ * Set to export library symbols from inside the ustdio library,
+ * and to import them from outside.
+ * @stable ICU 2.0
+ */
+
+/**
+ * \def U_TOOLUTIL_API
+ * Set to export library symbols from inside the toolutil library,
+ * and to import them from outside.
+ * @stable ICU 3.4
+ */
+
+#ifdef U_IN_DOXYGEN
+// This definition is required when generating the API docs.
+#define U_COMBINED_IMPLEMENTATION 1
+#endif
+
+#if defined(U_COMBINED_IMPLEMENTATION)
+#define U_DATA_API     U_EXPORT
+#define U_COMMON_API   U_EXPORT
+#define U_I18N_API     U_EXPORT
+#define U_LAYOUT_API   U_EXPORT
+#define U_LAYOUTEX_API U_EXPORT
+#define U_IO_API       U_EXPORT
+#define U_TOOLUTIL_API U_EXPORT
+#elif defined(U_STATIC_IMPLEMENTATION)
+#define U_DATA_API
+#define U_COMMON_API
+#define U_I18N_API
+#define U_LAYOUT_API
+#define U_LAYOUTEX_API
+#define U_IO_API
+#define U_TOOLUTIL_API
+#elif defined(U_COMMON_IMPLEMENTATION)
+#define U_DATA_API     U_IMPORT
+#define U_COMMON_API   U_EXPORT
+#define U_I18N_API     U_IMPORT
+#define U_LAYOUT_API   U_IMPORT
+#define U_LAYOUTEX_API U_IMPORT
+#define U_IO_API       U_IMPORT
+#define U_TOOLUTIL_API U_IMPORT
+#elif defined(U_I18N_IMPLEMENTATION)
+#define U_DATA_API     U_IMPORT
+#define U_COMMON_API   U_IMPORT
+#define U_I18N_API     U_EXPORT
+#define U_LAYOUT_API   U_IMPORT
+#define U_LAYOUTEX_API U_IMPORT
+#define U_IO_API       U_IMPORT
+#define U_TOOLUTIL_API U_IMPORT
+#elif defined(U_LAYOUT_IMPLEMENTATION)
+#define U_DATA_API     U_IMPORT
+#define U_COMMON_API   U_IMPORT
+#define U_I18N_API     U_IMPORT
+#define U_LAYOUT_API   U_EXPORT
+#define U_LAYOUTEX_API U_IMPORT
+#define U_IO_API       U_IMPORT
+#define U_TOOLUTIL_API U_IMPORT
+#elif defined(U_LAYOUTEX_IMPLEMENTATION)
+#define U_DATA_API     U_IMPORT
+#define U_COMMON_API   U_IMPORT
+#define U_I18N_API     U_IMPORT
+#define U_LAYOUT_API   U_IMPORT
+#define U_LAYOUTEX_API U_EXPORT
+#define U_IO_API       U_IMPORT
+#define U_TOOLUTIL_API U_IMPORT
+#elif defined(U_IO_IMPLEMENTATION)
+#define U_DATA_API     U_IMPORT
+#define U_COMMON_API   U_IMPORT
+#define U_I18N_API     U_IMPORT
+#define U_LAYOUT_API   U_IMPORT
+#define U_LAYOUTEX_API U_IMPORT
+#define U_IO_API       U_EXPORT
+#define U_TOOLUTIL_API U_IMPORT
+#elif defined(U_TOOLUTIL_IMPLEMENTATION)
+#define U_DATA_API     U_IMPORT
+#define U_COMMON_API   U_IMPORT
+#define U_I18N_API     U_IMPORT
+#define U_LAYOUT_API   U_IMPORT
+#define U_LAYOUTEX_API U_IMPORT
+#define U_IO_API       U_IMPORT
+#define U_TOOLUTIL_API U_EXPORT
+#else
+#define U_DATA_API     U_IMPORT
+#define U_COMMON_API   U_IMPORT
+#define U_I18N_API     U_IMPORT
+#define U_LAYOUT_API   U_IMPORT
+#define U_LAYOUTEX_API U_IMPORT
+#define U_IO_API       U_IMPORT
+#define U_TOOLUTIL_API U_IMPORT
+#endif
+
+/**
+ * \def U_STANDARD_CPP_NAMESPACE
+ * Control of C++ Namespace
+ * @stable ICU 2.0
+ */
+#ifdef __cplusplus
+#define U_STANDARD_CPP_NAMESPACE        ::
+#else
+#define U_STANDARD_CPP_NAMESPACE
+#endif
+
+/*===========================================================================*/
+/* UErrorCode */
+/*===========================================================================*/
+
+/**
+ * Standard ICU4C error code type, a substitute for exceptions.
+ *
+ * Initialize the UErrorCode with U_ZERO_ERROR, and check for success or
+ * failure using U_SUCCESS() or U_FAILURE():
+ *
+ *     UErrorCode errorCode = U_ZERO_ERROR;
+ *     // call ICU API that needs an error code parameter.
+ *     if (U_FAILURE(errorCode)) {
+ *         // An error occurred. Handle it here.
+ *     }
+ *
+ * C++ code should use icu::ErrorCode, available in unicode/errorcode.h, or a
+ * suitable subclass.
+ *
+ * For more information, see:
+ * https://unicode-org.github.io/icu/userguide/dev/codingguidelines#details-about-icu-error-codes
+ *
+ * Note: By convention, ICU functions that take a reference (C++) or a pointer
+ * (C) to a UErrorCode first test:
+ *
+ *     if (U_FAILURE(errorCode)) { return immediately; }
+ *
+ * so that in a chain of such functions the first one that sets an error code
+ * causes the following ones to not perform any operations.
+ *
+ * @stable ICU 2.0
+ */
+typedef enum UErrorCode {
+    /* The ordering of U_ERROR_INFO_START Vs U_USING_FALLBACK_WARNING looks weird
+     * and is that way because VC++ debugger displays first encountered constant,
+     * which is not the what the code is used for
+     */
+
+    U_USING_FALLBACK_WARNING  = -128,   /**< A resource bundle lookup returned a fallback result (not an error) */
+
+    U_ERROR_WARNING_START     = -128,   /**< Start of information results (semantically successful) */
+
+    U_USING_DEFAULT_WARNING   = -127,   /**< A resource bundle lookup returned a result from the root locale (not an error) */
+
+    U_SAFECLONE_ALLOCATED_WARNING = -126, /**< A SafeClone operation required allocating memory (informational only) */
+
+    U_STATE_OLD_WARNING       = -125,   /**< ICU has to use compatibility layer to construct the service. Expect performance/memory usage degradation. Consider upgrading */
+
+    U_STRING_NOT_TERMINATED_WARNING = -124,/**< An output string could not be NUL-terminated because output length==destCapacity. */
+
+    U_SORT_KEY_TOO_SHORT_WARNING = -123, /**< Number of levels requested in getBound is higher than the number of levels in the sort key */
+
+    U_AMBIGUOUS_ALIAS_WARNING = -122,   /**< This converter alias can go to different converter implementations */
+
+    U_DIFFERENT_UCA_VERSION = -121,     /**< ucol_open encountered a mismatch between UCA version and collator image version, so the collator was constructed from rules. No impact to further function */
+    
+    U_PLUGIN_CHANGED_LEVEL_WARNING = -120, /**< A plugin caused a level change. May not be an error, but later plugins may not load. */
+
+
+#ifndef U_HIDE_DEPRECATED_API
+    /**
+     * One more than the highest normal UErrorCode warning value.
+     * @deprecated ICU 58 The numeric value may change over time, see ICU ticket #12420.
+     */
+    U_ERROR_WARNING_LIMIT,
+#endif  // U_HIDE_DEPRECATED_API
+
+    U_ZERO_ERROR              =  0,     /**< No error, no warning. */
+
+    U_ILLEGAL_ARGUMENT_ERROR  =  1,     /**< Start of codes indicating failure */
+    U_MISSING_RESOURCE_ERROR  =  2,     /**< The requested resource cannot be found */
+    U_INVALID_FORMAT_ERROR    =  3,     /**< Data format is not what is expected */
+    U_FILE_ACCESS_ERROR       =  4,     /**< The requested file cannot be found */
+    U_INTERNAL_PROGRAM_ERROR  =  5,     /**< Indicates a bug in the library code */
+    U_MESSAGE_PARSE_ERROR     =  6,     /**< Unable to parse a message (message format) */
+    U_MEMORY_ALLOCATION_ERROR =  7,     /**< Memory allocation error */
+    U_INDEX_OUTOFBOUNDS_ERROR =  8,     /**< Trying to access the index that is out of bounds */
+    U_PARSE_ERROR             =  9,     /**< Equivalent to Java ParseException */
+    U_INVALID_CHAR_FOUND      = 10,     /**< Character conversion: Unmappable input sequence. In other APIs: Invalid character. */
+    U_TRUNCATED_CHAR_FOUND    = 11,     /**< Character conversion: Incomplete input sequence. */
+    U_ILLEGAL_CHAR_FOUND      = 12,     /**< Character conversion: Illegal input sequence/combination of input units. */
+    U_INVALID_TABLE_FORMAT    = 13,     /**< Conversion table file found, but corrupted */
+    U_INVALID_TABLE_FILE      = 14,     /**< Conversion table file not found */
+    U_BUFFER_OVERFLOW_ERROR   = 15,     /**< A result would not fit in the supplied buffer */
+    U_UNSUPPORTED_ERROR       = 16,     /**< Requested operation not supported in current context */
+    U_RESOURCE_TYPE_MISMATCH  = 17,     /**< an operation is requested over a resource that does not support it */
+    U_ILLEGAL_ESCAPE_SEQUENCE = 18,     /**< ISO-2022 illegal escape sequence */
+    U_UNSUPPORTED_ESCAPE_SEQUENCE = 19, /**< ISO-2022 unsupported escape sequence */
+    U_NO_SPACE_AVAILABLE      = 20,     /**< No space available for in-buffer expansion for Arabic shaping */
+    U_CE_NOT_FOUND_ERROR      = 21,     /**< Currently used only while setting variable top, but can be used generally */
+    U_PRIMARY_TOO_LONG_ERROR  = 22,     /**< User tried to set variable top to a primary that is longer than two bytes */
+    U_STATE_TOO_OLD_ERROR     = 23,     /**< ICU cannot construct a service from this state, as it is no longer supported */
+    U_TOO_MANY_ALIASES_ERROR  = 24,     /**< There are too many aliases in the path to the requested resource.
+                                             It is very possible that a circular alias definition has occurred */
+    U_ENUM_OUT_OF_SYNC_ERROR  = 25,     /**< UEnumeration out of sync with underlying collection */
+    U_INVARIANT_CONVERSION_ERROR = 26,  /**< Unable to convert a UChar* string to char* with the invariant converter. */
+    U_INVALID_STATE_ERROR     = 27,     /**< Requested operation can not be completed with ICU in its current state */
+    U_COLLATOR_VERSION_MISMATCH = 28,   /**< Collator version is not compatible with the base version */
+    U_USELESS_COLLATOR_ERROR  = 29,     /**< Collator is options only and no base is specified */
+    U_NO_WRITE_PERMISSION     = 30,     /**< Attempt to modify read-only or constant data. */
+    /**
+     * The input is impractically long for an operation.
+     * It is rejected because it may lead to problems such as excessive
+     * processing time, stack depth, or heap memory requirements.
+     *
+     * @stable ICU 68
+     */
+    U_INPUT_TOO_LONG_ERROR = 31,
+
+#ifndef U_HIDE_DEPRECATED_API
+    /**
+     * One more than the highest standard error code.
+     * @deprecated ICU 58 The numeric value may change over time, see ICU ticket #12420.
+     */
+    U_STANDARD_ERROR_LIMIT = 32,
+#endif  // U_HIDE_DEPRECATED_API
+
+    /*
+     * Error codes in the range 0x10000 0x10100 are reserved for Transliterator.
+     */
+    U_BAD_VARIABLE_DEFINITION=0x10000,/**< Missing '$' or duplicate variable name */
+    U_PARSE_ERROR_START = 0x10000,    /**< Start of Transliterator errors */
+    U_MALFORMED_RULE,                 /**< Elements of a rule are misplaced */
+    U_MALFORMED_SET,                  /**< A UnicodeSet pattern is invalid*/
+    U_MALFORMED_SYMBOL_REFERENCE,     /**< UNUSED as of ICU 2.4 */
+    U_MALFORMED_UNICODE_ESCAPE,       /**< A Unicode escape pattern is invalid*/
+    U_MALFORMED_VARIABLE_DEFINITION,  /**< A variable definition is invalid */
+    U_MALFORMED_VARIABLE_REFERENCE,   /**< A variable reference is invalid */
+    U_MISMATCHED_SEGMENT_DELIMITERS,  /**< UNUSED as of ICU 2.4 */
+    U_MISPLACED_ANCHOR_START,         /**< A start anchor appears at an illegal position */
+    U_MISPLACED_CURSOR_OFFSET,        /**< A cursor offset occurs at an illegal position */
+    U_MISPLACED_QUANTIFIER,           /**< A quantifier appears after a segment close delimiter */
+    U_MISSING_OPERATOR,               /**< A rule contains no operator */
+    U_MISSING_SEGMENT_CLOSE,          /**< UNUSED as of ICU 2.4 */
+    U_MULTIPLE_ANTE_CONTEXTS,         /**< More than one ante context */
+    U_MULTIPLE_CURSORS,               /**< More than one cursor */
+    U_MULTIPLE_POST_CONTEXTS,         /**< More than one post context */
+    U_TRAILING_BACKSLASH,             /**< A dangling backslash */
+    U_UNDEFINED_SEGMENT_REFERENCE,    /**< A segment reference does not correspond to a defined segment */
+    U_UNDEFINED_VARIABLE,             /**< A variable reference does not correspond to a defined variable */
+    U_UNQUOTED_SPECIAL,               /**< A special character was not quoted or escaped */
+    U_UNTERMINATED_QUOTE,             /**< A closing single quote is missing */
+    U_RULE_MASK_ERROR,                /**< A rule is hidden by an earlier more general rule */
+    U_MISPLACED_COMPOUND_FILTER,      /**< A compound filter is in an invalid location */
+    U_MULTIPLE_COMPOUND_FILTERS,      /**< More than one compound filter */
+    U_INVALID_RBT_SYNTAX,             /**< A "::id" rule was passed to the RuleBasedTransliterator parser */
+    U_INVALID_PROPERTY_PATTERN,       /**< UNUSED as of ICU 2.4 */
+    U_MALFORMED_PRAGMA,               /**< A 'use' pragma is invalid */
+    U_UNCLOSED_SEGMENT,               /**< A closing ')' is missing */
+    U_ILLEGAL_CHAR_IN_SEGMENT,        /**< UNUSED as of ICU 2.4 */
+    U_VARIABLE_RANGE_EXHAUSTED,       /**< Too many stand-ins generated for the given variable range */
+    U_VARIABLE_RANGE_OVERLAP,         /**< The variable range overlaps characters used in rules */
+    U_ILLEGAL_CHARACTER,              /**< A special character is outside its allowed context */
+    U_INTERNAL_TRANSLITERATOR_ERROR,  /**< Internal transliterator system error */
+    U_INVALID_ID,                     /**< A "::id" rule specifies an unknown transliterator */
+    U_INVALID_FUNCTION,               /**< A "&fn()" rule specifies an unknown transliterator */
+#ifndef U_HIDE_DEPRECATED_API
+    /**
+     * One more than the highest normal Transliterator error code.
+     * @deprecated ICU 58 The numeric value may change over time, see ICU ticket #12420.
+     */
+    U_PARSE_ERROR_LIMIT,
+#endif  // U_HIDE_DEPRECATED_API
+
+    /*
+     * Error codes in the range 0x10100 0x10200 are reserved for the formatting API.
+     */
+    U_UNEXPECTED_TOKEN=0x10100,       /**< Syntax error in format pattern */
+    U_FMT_PARSE_ERROR_START=0x10100,  /**< Start of format library errors */
+    U_MULTIPLE_DECIMAL_SEPARATORS,    /**< More than one decimal separator in number pattern */
+    U_MULTIPLE_DECIMAL_SEPERATORS = U_MULTIPLE_DECIMAL_SEPARATORS, /**< Typo: kept for backward compatibility. Use U_MULTIPLE_DECIMAL_SEPARATORS */
+    U_MULTIPLE_EXPONENTIAL_SYMBOLS,   /**< More than one exponent symbol in number pattern */
+    U_MALFORMED_EXPONENTIAL_PATTERN,  /**< Grouping symbol in exponent pattern */
+    U_MULTIPLE_PERCENT_SYMBOLS,       /**< More than one percent symbol in number pattern */
+    U_MULTIPLE_PERMILL_SYMBOLS,       /**< More than one permill symbol in number pattern */
+    U_MULTIPLE_PAD_SPECIFIERS,        /**< More than one pad symbol in number pattern */
+    U_PATTERN_SYNTAX_ERROR,           /**< Syntax error in format pattern */
+    U_ILLEGAL_PAD_POSITION,           /**< Pad symbol misplaced in number pattern */
+    U_UNMATCHED_BRACES,               /**< Braces do not match in message pattern */
+    U_UNSUPPORTED_PROPERTY,           /**< UNUSED as of ICU 2.4 */
+    U_UNSUPPORTED_ATTRIBUTE,          /**< UNUSED as of ICU 2.4 */
+    U_ARGUMENT_TYPE_MISMATCH,         /**< Argument name and argument index mismatch in MessageFormat functions */
+    U_DUPLICATE_KEYWORD,              /**< Duplicate keyword in PluralFormat */
+    U_UNDEFINED_KEYWORD,              /**< Undefined Plural keyword */
+    U_DEFAULT_KEYWORD_MISSING,        /**< Missing DEFAULT rule in plural rules */
+    U_DECIMAL_NUMBER_SYNTAX_ERROR,    /**< Decimal number syntax error */
+    U_FORMAT_INEXACT_ERROR,           /**< Cannot format a number exactly and rounding mode is ROUND_UNNECESSARY @stable ICU 4.8 */
+    U_NUMBER_ARG_OUTOFBOUNDS_ERROR,   /**< The argument to a NumberFormatter helper method was out of bounds; the bounds are usually 0 to 999. @stable ICU 61 */
+    U_NUMBER_SKELETON_SYNTAX_ERROR,   /**< The number skeleton passed to C++ NumberFormatter or C UNumberFormatter was invalid or contained a syntax error. @stable ICU 62 */
+
+    /* MessageFormat 2.0 errors */
+    U_MF_UNRESOLVED_VARIABLE_ERROR,    /**< A variable is referred to but not bound by any definition @internal ICU 75 technology preview @deprecated This API is for technology preview only. */
+    U_MF_SYNTAX_ERROR,                 /**< Includes all syntax errors @internal ICU 75 technology preview @deprecated This API is for technology preview only. */
+    U_MF_UNKNOWN_FUNCTION_ERROR,       /**< An annotation refers to a function not defined by the standard or custom function registry @internal ICU 75 technology preview @deprecated This API is for technology preview only. */
+    U_MF_VARIANT_KEY_MISMATCH_ERROR,   /**< In a match-construct, one or more variants had a different number of keys from the number of selectors @internal ICU 75 technology preview @deprecated This API is for technology preview only. */
+    U_MF_FORMATTING_ERROR,             /**< Covers all runtime errors: for example, an internally inconsistent set of options. @internal ICU 75 technology preview @deprecated This API is for technology preview only. */
+    U_MF_NONEXHAUSTIVE_PATTERN_ERROR,  /**< In a match-construct, the variants do not cover all possible values @internal ICU 75 technology preview @deprecated This API is for technology preview only. */
+    U_MF_DUPLICATE_OPTION_NAME_ERROR,  /**< In an annotation, the same option name appears more than once @internal ICU 75 technology preview @deprecated This API is for technology preview only. */
+    U_MF_SELECTOR_ERROR,               /**< A selector function is applied to an operand of the wrong type @internal ICU 75 technology preview @deprecated This API is for technology preview only. */
+    U_MF_MISSING_SELECTOR_ANNOTATION_ERROR,  /**< A selector expression evaluates to an unannotated operand. @internal ICU 75 technology preview @deprecated This API is for technology preview only. */
+    U_MF_DUPLICATE_DECLARATION_ERROR, /**< The same variable is declared in more than one .local or .input declaration. @internal ICU 75 technology preview @deprecated This API is for technology preview only. */
+    U_MF_OPERAND_MISMATCH_ERROR,     /**< An operand provided to a function does not have the required form for that function @internal ICU 75 technology preview @deprecated This API is for technology preview only. */
+    U_MF_DUPLICATE_VARIANT_ERROR, /**< A message includes a variant with the same key list as another variant. @internal ICU 76 technology preview @deprecated This API is for technology preview only. */
+    U_MF_BAD_OPTION,             /**< An option value provided to a function does not have the required form for that option. @internal ICU 77 technology preview @deprecated This API is for technology preview only. */
+#ifndef U_HIDE_DEPRECATED_API
+    /**
+     * One more than the highest normal formatting API error code.
+     * @deprecated ICU 58 The numeric value may change over time, see ICU ticket #12420.
+     */
+    U_FMT_PARSE_ERROR_LIMIT = 0x10121,
+#endif  // U_HIDE_DEPRECATED_API
+
+    /*
+     * Error codes in the range 0x10200 0x102ff are reserved for BreakIterator.
+     */
+    U_BRK_INTERNAL_ERROR=0x10200,          /**< An internal error (bug) was detected.             */
+    U_BRK_ERROR_START=0x10200,             /**< Start of codes indicating Break Iterator failures */
+    U_BRK_HEX_DIGITS_EXPECTED,             /**< Hex digits expected as part of a escaped char in a rule. */
+    U_BRK_SEMICOLON_EXPECTED,              /**< Missing ';' at the end of a RBBI rule.            */
+    U_BRK_RULE_SYNTAX,                     /**< Syntax error in RBBI rule.                        */
+    U_BRK_UNCLOSED_SET,                    /**< UnicodeSet writing an RBBI rule missing a closing ']'. */
+    U_BRK_ASSIGN_ERROR,                    /**< Syntax error in RBBI rule assignment statement.   */
+    U_BRK_VARIABLE_REDFINITION,            /**< RBBI rule $Variable redefined.                    */
+    U_BRK_MISMATCHED_PAREN,                /**< Mis-matched parentheses in an RBBI rule.          */
+    U_BRK_NEW_LINE_IN_QUOTED_STRING,       /**< Missing closing quote in an RBBI rule.            */
+    U_BRK_UNDEFINED_VARIABLE,              /**< Use of an undefined $Variable in an RBBI rule.    */
+    U_BRK_INIT_ERROR,                      /**< Initialization failure.  Probable missing ICU Data. */
+    U_BRK_RULE_EMPTY_SET,                  /**< Rule contains an empty Unicode Set.               */
+    U_BRK_UNRECOGNIZED_OPTION,             /**< !!option in RBBI rules not recognized.            */
+    U_BRK_MALFORMED_RULE_TAG,              /**< The {nnn} tag on a rule is malformed              */
+#ifndef U_HIDE_DEPRECATED_API
+    /**
+     * One more than the highest normal BreakIterator error code.
+     * @deprecated ICU 58 The numeric value may change over time, see ICU ticket #12420.
+     */
+    U_BRK_ERROR_LIMIT,
+#endif  // U_HIDE_DEPRECATED_API
+
+    /*
+     * Error codes in the range 0x10300-0x103ff are reserved for regular expression related errors.
+     */
+    U_REGEX_INTERNAL_ERROR=0x10300,       /**< An internal error (bug) was detected.              */
+    U_REGEX_ERROR_START=0x10300,          /**< Start of codes indicating Regexp failures          */
+    U_REGEX_RULE_SYNTAX,                  /**< Syntax error in regexp pattern.                    */
+    U_REGEX_INVALID_STATE,                /**< RegexMatcher in invalid state for requested operation */
+    U_REGEX_BAD_ESCAPE_SEQUENCE,          /**< Unrecognized backslash escape sequence in pattern  */
+    U_REGEX_PROPERTY_SYNTAX,              /**< Incorrect Unicode property                         */
+    U_REGEX_UNIMPLEMENTED,                /**< Use of regexp feature that is not yet implemented. */
+    U_REGEX_MISMATCHED_PAREN,             /**< Incorrectly nested parentheses in regexp pattern.  */
+    U_REGEX_NUMBER_TOO_BIG,               /**< Decimal number is too large.                       */
+    U_REGEX_BAD_INTERVAL,                 /**< Error in {min,max} interval                        */
+    U_REGEX_MAX_LT_MIN,                   /**< In {min,max}, max is less than min.                */
+    U_REGEX_INVALID_BACK_REF,             /**< Back-reference to a non-existent capture group.    */
+    U_REGEX_INVALID_FLAG,                 /**< Invalid value for match mode flags.                */
+    U_REGEX_LOOK_BEHIND_LIMIT,            /**< Look-Behind pattern matches must have a bounded maximum length.    */
+    U_REGEX_SET_CONTAINS_STRING,          /**< Regexps cannot have UnicodeSets containing strings.*/
+#ifndef U_HIDE_DEPRECATED_API
+    U_REGEX_OCTAL_TOO_BIG,                /**< Octal character constants must be <= 0377. @deprecated ICU 54. This error cannot occur. */
+#endif  /* U_HIDE_DEPRECATED_API */
+    U_REGEX_MISSING_CLOSE_BRACKET=U_REGEX_SET_CONTAINS_STRING+2, /**< Missing closing bracket on a bracket expression. */
+    U_REGEX_INVALID_RANGE,                /**< In a character range [x-y], x is greater than y.   */
+    U_REGEX_STACK_OVERFLOW,               /**< Regular expression backtrack stack overflow.       */
+    U_REGEX_TIME_OUT,                     /**< Maximum allowed match time exceeded                */
+    U_REGEX_STOPPED_BY_CALLER,            /**< Matching operation aborted by user callback fn.    */
+    U_REGEX_PATTERN_TOO_BIG,              /**< Pattern exceeds limits on size or complexity. @stable ICU 55 */
+    U_REGEX_INVALID_CAPTURE_GROUP_NAME,   /**< Invalid capture group name. @stable ICU 55 */
+#ifndef U_HIDE_DEPRECATED_API
+    /**
+     * One more than the highest normal regular expression error code.
+     * @deprecated ICU 58 The numeric value may change over time, see ICU ticket #12420.
+     */
+    U_REGEX_ERROR_LIMIT=U_REGEX_STOPPED_BY_CALLER+3,
+#endif  // U_HIDE_DEPRECATED_API
+
+    /*
+     * Error codes in the range 0x10400-0x104ff are reserved for IDNA related error codes.
+     */
+    U_IDNA_PROHIBITED_ERROR=0x10400,
+    U_IDNA_ERROR_START=0x10400,
+    U_IDNA_UNASSIGNED_ERROR,
+    U_IDNA_CHECK_BIDI_ERROR,
+    U_IDNA_STD3_ASCII_RULES_ERROR,
+    U_IDNA_ACE_PREFIX_ERROR,
+    U_IDNA_VERIFICATION_ERROR,
+    U_IDNA_LABEL_TOO_LONG_ERROR,
+    U_IDNA_ZERO_LENGTH_LABEL_ERROR,
+    U_IDNA_DOMAIN_NAME_TOO_LONG_ERROR,
+#ifndef U_HIDE_DEPRECATED_API
+    /**
+     * One more than the highest normal IDNA error code.
+     * @deprecated ICU 58 The numeric value may change over time, see ICU ticket #12420.
+     */
+    U_IDNA_ERROR_LIMIT,
+#endif  // U_HIDE_DEPRECATED_API
+    /*
+     * Aliases for StringPrep
+     */
+    U_STRINGPREP_PROHIBITED_ERROR = U_IDNA_PROHIBITED_ERROR,
+    U_STRINGPREP_UNASSIGNED_ERROR = U_IDNA_UNASSIGNED_ERROR,
+    U_STRINGPREP_CHECK_BIDI_ERROR = U_IDNA_CHECK_BIDI_ERROR,
+    
+    /*
+     * Error codes in the range 0x10500-0x105ff are reserved for Plugin related error codes.
+     */
+    U_PLUGIN_ERROR_START=0x10500,         /**< Start of codes indicating plugin failures */
+    U_PLUGIN_TOO_HIGH=0x10500,            /**< The plugin's level is too high to be loaded right now. */
+    U_PLUGIN_DIDNT_SET_LEVEL,             /**< The plugin didn't call uplug_setPlugLevel in response to a QUERY */
+#ifndef U_HIDE_DEPRECATED_API
+    /**
+     * One more than the highest normal plug-in error code.
+     * @deprecated ICU 58 The numeric value may change over time, see ICU ticket #12420.
+     */
+    U_PLUGIN_ERROR_LIMIT,
+#endif  // U_HIDE_DEPRECATED_API
+
+#ifndef U_HIDE_DEPRECATED_API
+    /**
+     * One more than the highest normal error code.
+     * @deprecated ICU 58 The numeric value may change over time, see ICU ticket #12420.
+     */
+    U_ERROR_LIMIT=U_PLUGIN_ERROR_LIMIT
+#endif  // U_HIDE_DEPRECATED_API
+} UErrorCode;
+
+/* Use the following to determine if an UErrorCode represents */
+/* operational success or failure. */
+
+#ifdef __cplusplus
+    /**
+     * Does the error code indicate success?
+     * @stable ICU 2.0
+     */
+    static
+    inline UBool U_SUCCESS(UErrorCode code) { return code <= U_ZERO_ERROR; }
+    /**
+     * Does the error code indicate a failure?
+     * @stable ICU 2.0
+     */
+    static
+    inline UBool U_FAILURE(UErrorCode code) { return code > U_ZERO_ERROR; }
+#else
+    /**
+     * Does the error code indicate success?
+     * @stable ICU 2.0
+     */
+#   define U_SUCCESS(x) ((x)<=U_ZERO_ERROR)
+    /**
+     * Does the error code indicate a failure?
+     * @stable ICU 2.0
+     */
+#   define U_FAILURE(x) ((x)>U_ZERO_ERROR)
+#endif
+
+/**
+ * Return a string for a UErrorCode value.
+ * The string will be the same as the name of the error code constant
+ * in the UErrorCode enum above.
+ * @stable ICU 2.0
+ */
+U_CAPI const char * U_EXPORT2
+u_errorName(UErrorCode code);
+
+
+#endif /* _UTYPES */

thirdparty/icu4c/common/unicode/uvernum.h

@@ -0,0 +1,191 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+/*
+*******************************************************************************
+*   Copyright (C) 2000-2016, International Business Machines
+*   Corporation and others.  All Rights Reserved.
+*******************************************************************************
+*
+*   file name:  uvernum.h
+*   encoding:   UTF-8
+*   tab size:   8 (not used)
+*   indentation:4
+*
+*   Created by: Vladimir Weinstein
+*   Updated by: Steven R. Loomis
+*
+*/
+
+/**
+ * \file
+ * \brief C API: definitions of ICU version numbers
+ *
+ * This file is included by uversion.h and other files. This file contains only
+ * macros and definitions. The actual version numbers are defined here.
+ */
+
+ /*
+  * IMPORTANT: When updating version, the following things need to be done:
+  * source/common/unicode/uvernum.h - this file: update major, minor,
+  *        patchlevel, suffix, version, short version constants, namespace,
+  *                    renaming macro, and copyright
+  *
+  * The following files need to be updated as well, which can be done
+  *  by running the UNIX makefile target 'update-windows-makefiles' in icu4c/source.
+  *
+  * source/allinone/Build.Windows.IcuVersion.props - Update the IcuMajorVersion
+  * source/data/makedata.mak - change U_ICUDATA_NAME so that it contains
+  *                            the new major/minor combination, and UNICODE_VERSION
+  *                            for the Unicode version.
+  */
+
+#ifndef UVERNUM_H
+#define UVERNUM_H
+
+/** The standard copyright notice that gets compiled into each library.
+ *  This value will change in the subsequent releases of ICU
+ *  @stable ICU 2.4
+ */
+#define U_COPYRIGHT_STRING \
+  " Copyright (C) 2016 and later: Unicode, Inc. and others. License & terms of use: http://www.unicode.org/copyright.html "
+
+/** The current ICU major version as an integer.
+ *  This value will change in the subsequent releases of ICU
+ *  @stable ICU 2.4
+ */
+#define U_ICU_VERSION_MAJOR_NUM 77
+
+/** The current ICU minor version as an integer.
+ *  This value will change in the subsequent releases of ICU
+ *  @stable ICU 2.6
+ */
+#define U_ICU_VERSION_MINOR_NUM 1
+
+/** The current ICU patchlevel version as an integer.
+ *  This value will change in the subsequent releases of ICU
+ *  @stable ICU 2.4
+ */
+#define U_ICU_VERSION_PATCHLEVEL_NUM 0
+
+/** The current ICU build level version as an integer.
+ *  This value is for use by ICU clients. It defaults to 0.
+ *  @stable ICU 4.0
+ */
+#ifndef U_ICU_VERSION_BUILDLEVEL_NUM
+#define U_ICU_VERSION_BUILDLEVEL_NUM 0
+#endif
+
+/** Glued version suffix for renamers
+ *  This value will change in the subsequent releases of ICU
+ *  @stable ICU 2.6
+ */
+#define U_ICU_VERSION_SUFFIX _77
+
+/**
+ * \def U_DEF2_ICU_ENTRY_POINT_RENAME
+ * @internal
+ */
+/**
+ * \def U_DEF_ICU_ENTRY_POINT_RENAME
+ * @internal
+ */
+/** Glued version suffix function for renamers
+ *  This value will change in the subsequent releases of ICU.
+ *  If a custom suffix (such as matching library suffixes) is desired, this can be modified.
+ *  Note that if present, platform.h may contain an earlier definition of this macro.
+ *  \def U_ICU_ENTRY_POINT_RENAME
+ *  @stable ICU 4.2
+ */
+/**
+ * Disable the version suffix. Use the custom suffix if exists.
+ * \def U_DISABLE_VERSION_SUFFIX
+ * @internal
+ */
+#ifndef U_DISABLE_VERSION_SUFFIX
+#define U_DISABLE_VERSION_SUFFIX 0
+#endif
+
+#ifndef U_ICU_ENTRY_POINT_RENAME
+#ifdef U_HAVE_LIB_SUFFIX
+#   if !U_DISABLE_VERSION_SUFFIX
+#       define U_DEF_ICU_ENTRY_POINT_RENAME(x,y,z) x ## y ##  z
+#       define U_DEF2_ICU_ENTRY_POINT_RENAME(x,y,z) U_DEF_ICU_ENTRY_POINT_RENAME(x,y,z)
+#       define U_ICU_ENTRY_POINT_RENAME(x)    U_DEF2_ICU_ENTRY_POINT_RENAME(x,U_ICU_VERSION_SUFFIX,U_LIB_SUFFIX_C_NAME)
+#   else
+#       define U_DEF_ICU_ENTRY_POINT_RENAME(x,y) x ## y
+#       define U_DEF2_ICU_ENTRY_POINT_RENAME(x,y) U_DEF_ICU_ENTRY_POINT_RENAME(x,y)
+#       define U_ICU_ENTRY_POINT_RENAME(x)    U_DEF2_ICU_ENTRY_POINT_RENAME(x,U_LIB_SUFFIX_C_NAME)
+#   endif
+#else
+#   if !U_DISABLE_VERSION_SUFFIX
+#       define U_DEF_ICU_ENTRY_POINT_RENAME(x,y) x ## y
+#       define U_DEF2_ICU_ENTRY_POINT_RENAME(x,y) U_DEF_ICU_ENTRY_POINT_RENAME(x,y)
+#       define U_ICU_ENTRY_POINT_RENAME(x)    U_DEF2_ICU_ENTRY_POINT_RENAME(x,U_ICU_VERSION_SUFFIX)
+#   else
+#       define U_ICU_ENTRY_POINT_RENAME(x)    x
+#   endif
+#endif
+#endif
+
+/** The current ICU library version as a dotted-decimal string. The patchlevel
+ *  only appears in this string if it non-zero.
+ *  This value will change in the subsequent releases of ICU
+ *  @stable ICU 2.4
+ */
+#define U_ICU_VERSION "77.1"
+
+/**
+ * The current ICU library major version number as a string, for library name suffixes.
+ * This value will change in subsequent releases of ICU.
+ *
+ * Until ICU 4.8, this was the combination of the single-digit major and minor ICU version numbers
+ * into one string without dots ("48").
+ * Since ICU 49, it is the double-digit major ICU version number.
+ * See https://unicode-org.github.io/icu/userguide/design#version-numbers-in-icu
+ *
+ * @stable ICU 2.6
+ */
+#define U_ICU_VERSION_SHORT "77"
+
+#ifndef U_HIDE_INTERNAL_API
+/** Data version in ICU4C.
+ * @internal ICU 4.4 Internal Use Only
+ **/
+#define U_ICU_DATA_VERSION "77.1"
+#endif  /* U_HIDE_INTERNAL_API */
+
+/*===========================================================================
+ * ICU collation framework version information
+ * Version info that can be obtained from a collator is affected by these
+ * numbers in a secret and magic way. Please use collator version as whole
+ *===========================================================================
+ */
+
+/**
+ * Collation runtime version (sort key generator, strcoll).
+ * If the version is different, sort keys for the same string could be different.
+ * This value may change in subsequent releases of ICU.
+ * @stable ICU 2.4
+ */
+#define UCOL_RUNTIME_VERSION 9
+
+/**
+ * Collation builder code version.
+ * When this is different, the same tailoring might result
+ * in assigning different collation elements to code points.
+ * This value may change in subsequent releases of ICU.
+ * @stable ICU 2.4
+ */
+#define UCOL_BUILDER_VERSION 9
+
+#ifndef U_HIDE_DEPRECATED_API
+/**
+ * Constant 1.
+ * This was intended to be the version of collation tailorings,
+ * but instead the tailoring data carries a version number.
+ * @deprecated ICU 54
+ */
+#define UCOL_TAILORINGS_VERSION 1
+#endif  /* U_HIDE_DEPRECATED_API */
+
+#endif