initial commit, 4.5 stable

thirdparty/icu4c/i18n/scriptset.cpp

@@ -0,0 +1,313 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+/*
+**********************************************************************
+*   Copyright (C) 2014, International Business Machines
+*   Corporation and others.  All Rights Reserved.
+**********************************************************************
+*
+* scriptset.cpp
+*
+* created on: 2013 Jan 7
+* created by: Andy Heninger
+*/
+
+#include "unicode/utypes.h"
+
+#include "unicode/uchar.h"
+#include "unicode/unistr.h"
+
+#include "scriptset.h"
+#include "uassert.h"
+#include "cmemory.h"
+
+U_NAMESPACE_BEGIN
+
+//----------------------------------------------------------------------------
+//
+//  ScriptSet implementation
+//
+//----------------------------------------------------------------------------
+ScriptSet::ScriptSet() {
+    uprv_memset(bits, 0, sizeof(bits));
+}
+
+ScriptSet::~ScriptSet() {
+}
+
+ScriptSet::ScriptSet(const ScriptSet &other) {
+    *this = other;
+}
+
+ScriptSet & ScriptSet::operator =(const ScriptSet &other) {
+    uprv_memcpy(bits, other.bits, sizeof(bits));
+    return *this;
+}
+
+bool ScriptSet::operator == (const ScriptSet &other) const {
+    for (uint32_t i=0; i<UPRV_LENGTHOF(bits); i++) {
+        if (bits[i] != other.bits[i]) {
+            return false;
+        }
+    }
+    return true;
+}
+
+UBool ScriptSet::test(UScriptCode script, UErrorCode &status) const {
+    if (U_FAILURE(status)) {
+        return false;
+    }
+    if (script < 0 || static_cast<int32_t>(script) >= SCRIPT_LIMIT) {
+        status = U_ILLEGAL_ARGUMENT_ERROR;
+        return false;
+    }
+    uint32_t index = script / 32;
+    uint32_t bit   = 1 << (script & 31);
+    return ((bits[index] & bit) != 0);
+}
+
+
+ScriptSet &ScriptSet::set(UScriptCode script, UErrorCode &status) {
+    if (U_FAILURE(status)) {
+        return *this;
+    }
+    if (script < 0 || static_cast<int32_t>(script) >= SCRIPT_LIMIT) {
+        status = U_ILLEGAL_ARGUMENT_ERROR;
+        return *this;
+    }
+    uint32_t index = script / 32;
+    uint32_t bit   = 1 << (script & 31);
+    bits[index] |= bit;
+    return *this;
+}
+
+ScriptSet &ScriptSet::reset(UScriptCode script, UErrorCode &status) {
+    if (U_FAILURE(status)) {
+        return *this;
+    }
+    if (script < 0 || static_cast<int32_t>(script) >= SCRIPT_LIMIT) {
+        status = U_ILLEGAL_ARGUMENT_ERROR;
+        return *this;
+    }
+    uint32_t index = script / 32;
+    uint32_t bit   = 1 << (script & 31);
+    bits[index] &= ~bit;
+    return *this;
+}
+
+
+
+ScriptSet &ScriptSet::Union(const ScriptSet &other) {
+    for (uint32_t i=0; i<UPRV_LENGTHOF(bits); i++) {
+        bits[i] |= other.bits[i];
+    }
+    return *this;
+}
+
+ScriptSet &ScriptSet::intersect(const ScriptSet &other) {
+    for (uint32_t i=0; i<UPRV_LENGTHOF(bits); i++) {
+        bits[i] &= other.bits[i];
+    }
+    return *this;
+}
+
+ScriptSet &ScriptSet::intersect(UScriptCode script, UErrorCode &status) {
+    ScriptSet t;
+    t.set(script, status);
+    if (U_SUCCESS(status)) {
+        this->intersect(t);
+    }
+    return *this;
+}
+
+UBool ScriptSet::intersects(const ScriptSet &other) const {
+    for (uint32_t i=0; i<UPRV_LENGTHOF(bits); i++) {
+        if ((bits[i] & other.bits[i]) != 0) {
+            return true;
+        }
+    }
+    return false;
+}
+
+UBool ScriptSet::contains(const ScriptSet &other) const {
+    ScriptSet t(*this);
+    t.intersect(other);
+    return (t == other);
+}
+
+
+ScriptSet &ScriptSet::setAll() {
+    for (uint32_t i=0; i<UPRV_LENGTHOF(bits); i++) {
+        bits[i] = 0xffffffffu;
+    }
+    return *this;
+}
+
+
+ScriptSet &ScriptSet::resetAll() {
+    uprv_memset(bits, 0, sizeof(bits));
+    return *this;
+}
+
+int32_t ScriptSet::countMembers() const {
+    // This bit counter is good for sparse numbers of '1's, which is
+    //  very much the case that we will usually have.
+    int32_t count = 0;
+    for (uint32_t i=0; i<UPRV_LENGTHOF(bits); i++) {
+        uint32_t x = bits[i];
+        while (x > 0) {
+            count++;
+            x &= (x - 1);    // and off the least significant one bit.
+        }
+    }
+    return count;
+}
+
+int32_t ScriptSet::hashCode() const {
+    int32_t hash = 0;
+    for (int32_t i=0; i<UPRV_LENGTHOF(bits); i++) {
+        hash ^= bits[i];
+    }
+    return hash;
+}
+
+int32_t ScriptSet::nextSetBit(int32_t fromIndex) const {
+    // TODO: Wants a better implementation.
+    if (fromIndex < 0) {
+        return -1;
+    }
+    UErrorCode status = U_ZERO_ERROR;
+    for (int32_t scriptIndex = fromIndex; scriptIndex < SCRIPT_LIMIT; scriptIndex++) {
+        if (test(static_cast<UScriptCode>(scriptIndex), status)) {
+            return scriptIndex;
+        }
+    }
+    return -1;
+}
+
+UBool ScriptSet::isEmpty() const {
+    for (uint32_t i=0; i<UPRV_LENGTHOF(bits); i++) {
+        if (bits[i] != 0) {
+            return false;
+        }
+    }
+    return true;
+}
+
+UnicodeString &ScriptSet::displayScripts(UnicodeString &dest) const {
+    UBool firstTime = true;
+    for (int32_t i = nextSetBit(0); i >= 0; i = nextSetBit(i + 1)) {
+        if (!firstTime) {
+            dest.append(static_cast<char16_t>(0x20));
+        }
+        firstTime = false;
+        const char* scriptName = uscript_getShortName(static_cast<UScriptCode>(i));
+        dest.append(UnicodeString(scriptName, -1, US_INV));
+    }
+    return dest;
+}
+
+ScriptSet &ScriptSet::parseScripts(const UnicodeString &scriptString, UErrorCode &status) {
+    resetAll();
+    if (U_FAILURE(status)) {
+        return *this;
+    }
+    UnicodeString oneScriptName;
+    for (int32_t i=0; i<scriptString.length();) {
+        UChar32 c = scriptString.char32At(i);
+        i = scriptString.moveIndex32(i, 1);
+        if (!u_isUWhiteSpace(c)) {
+            oneScriptName.append(c);
+            if (i < scriptString.length()) {
+                continue;
+            }
+        }
+        if (oneScriptName.length() > 0) {
+            char buf[40];
+            oneScriptName.extract(0, oneScriptName.length(), buf, sizeof(buf)-1, US_INV);
+            buf[sizeof(buf)-1] = 0;
+            int32_t sc = u_getPropertyValueEnum(UCHAR_SCRIPT, buf);
+            if (sc == UCHAR_INVALID_CODE) {
+                status = U_ILLEGAL_ARGUMENT_ERROR;
+            } else {
+                this->set(static_cast<UScriptCode>(sc), status);
+            }
+            if (U_FAILURE(status)) {
+                return *this;
+            }
+            oneScriptName.remove();
+        }
+    }
+    return *this;
+}
+
+void ScriptSet::setScriptExtensions(UChar32 codePoint, UErrorCode& status) {
+    if (U_FAILURE(status)) { return; }
+    static const int32_t FIRST_GUESS_SCRIPT_CAPACITY = 20;
+    MaybeStackArray<UScriptCode,FIRST_GUESS_SCRIPT_CAPACITY> scripts;
+    UErrorCode internalStatus = U_ZERO_ERROR;
+    int32_t script_count = -1;
+
+    while (true) {
+        script_count = uscript_getScriptExtensions(
+            codePoint, scripts.getAlias(), scripts.getCapacity(), &internalStatus);
+        if (internalStatus == U_BUFFER_OVERFLOW_ERROR) {
+            // Need to allocate more space
+            if (scripts.resize(script_count) == nullptr) {
+                status = U_MEMORY_ALLOCATION_ERROR;
+                return;
+            }
+            internalStatus = U_ZERO_ERROR;
+        } else {
+            break;
+        }
+    }
+
+    // Check if we failed for some reason other than buffer overflow
+    if (U_FAILURE(internalStatus)) {
+        status = internalStatus;
+        return;
+    }
+
+    // Load the scripts into the ScriptSet and return
+    for (int32_t i = 0; i < script_count; i++) {
+        this->set(scripts[i], status);
+        if (U_FAILURE(status)) { return; }
+    }
+}
+
+U_NAMESPACE_END
+
+U_CAPI UBool U_EXPORT2
+uhash_equalsScriptSet(const UElement key1, const UElement key2) {
+    icu::ScriptSet *s1 = static_cast<icu::ScriptSet *>(key1.pointer);
+    icu::ScriptSet *s2 = static_cast<icu::ScriptSet *>(key2.pointer);
+    return (*s1 == *s2);
+}
+
+U_CAPI int32_t U_EXPORT2
+uhash_compareScriptSet(UElement key0, UElement key1) {
+    icu::ScriptSet *s0 = static_cast<icu::ScriptSet *>(key0.pointer);
+    icu::ScriptSet *s1 = static_cast<icu::ScriptSet *>(key1.pointer);
+    int32_t diff = s0->countMembers() - s1->countMembers();
+    if (diff != 0) return diff;
+    int32_t i0 = s0->nextSetBit(0);
+    int32_t i1 = s1->nextSetBit(0);
+    while ((diff = i0-i1) == 0 && i0 > 0) {
+        i0 = s0->nextSetBit(i0+1);
+        i1 = s1->nextSetBit(i1+1);
+    }
+    return diff;
+}
+
+U_CAPI int32_t U_EXPORT2
+uhash_hashScriptSet(const UElement key) {
+    icu::ScriptSet *s = static_cast<icu::ScriptSet *>(key.pointer);
+    return s->hashCode();
+}
+
+U_CAPI void U_EXPORT2
+uhash_deleteScriptSet(void *obj) {
+    icu::ScriptSet *s = static_cast<icu::ScriptSet *>(obj);
+    delete s;
+}

thirdparty/icu4c/i18n/scriptset.h

@@ -0,0 +1,89 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+/*
+**********************************************************************
+*   Copyright (C) 2013, International Business Machines
+*   Corporation and others.  All Rights Reserved.
+**********************************************************************
+*
+* scriptset.h
+*
+* created on: 2013 Jan 7
+* created by: Andy Heninger
+*/
+
+#ifndef __SCRIPTSET_H__
+#define __SCRIPTSET_H__
+
+#include "unicode/utypes.h"
+#include "unicode/uobject.h"
+#include "unicode/uscript.h"
+
+#include "uelement.h"
+
+U_NAMESPACE_BEGIN
+
+//-------------------------------------------------------------------------------
+//
+//  ScriptSet - A bit set representing a set of scripts.
+//
+//              This class was originally used exclusively with script sets appearing
+//              as part of the spoof check whole script confusable binary data. Its
+//              use has since become more general, but the continued use to wrap
+//              prebuilt binary data does constrain the design.
+//
+//-------------------------------------------------------------------------------
+class U_I18N_API ScriptSet: public UMemory {
+  public:
+    static constexpr int32_t SCRIPT_LIMIT = 224;  // multiple of 32!
+
+    ScriptSet();
+    ScriptSet(const ScriptSet &other);
+    ~ScriptSet();
+
+    bool operator == (const ScriptSet &other) const;
+    bool operator != (const ScriptSet &other) const {return !(*this == other);}
+    ScriptSet & operator = (const ScriptSet &other);
+
+    UBool      test(UScriptCode script, UErrorCode &status) const;
+    ScriptSet &Union(const ScriptSet &other);
+    ScriptSet &set(UScriptCode script, UErrorCode &status);
+    ScriptSet &reset(UScriptCode script, UErrorCode &status);
+    ScriptSet &intersect(const ScriptSet &other);
+    ScriptSet &intersect(UScriptCode script, UErrorCode &status);
+    UBool      intersects(const ScriptSet &other) const;  // Sets contain at least one script in common.
+    UBool      contains(const ScriptSet &other) const;    // All set bits in other are also set in this.
+
+    ScriptSet &setAll();
+    ScriptSet &resetAll();
+    int32_t countMembers() const;
+    int32_t hashCode() const;
+    int32_t nextSetBit(int32_t script) const;
+
+    UBool isEmpty() const;
+
+    UnicodeString &displayScripts(UnicodeString &dest) const; // append script names to dest string.
+    ScriptSet & parseScripts(const UnicodeString &scriptsString, UErrorCode &status);  // Replaces ScriptSet contents.
+
+    // Wraps around UScript::getScriptExtensions() and adds the corresponding scripts to this instance.
+    void setScriptExtensions(UChar32 codePoint, UErrorCode& status);
+
+  private:
+    uint32_t  bits[SCRIPT_LIMIT / 32];
+};
+
+U_NAMESPACE_END
+
+U_CAPI int32_t U_EXPORT2
+uhash_compareScriptSet(const UElement key1, const UElement key2);
+
+U_CAPI int32_t U_EXPORT2
+uhash_hashScriptSet(const UElement key);
+
+U_CAPI void U_EXPORT2
+uhash_deleteScriptSet(void *obj);
+
+U_CAPI UBool U_EXPORT2
+uhash_equalsScriptSet(const UElement key1, const UElement key2);
+
+#endif // __SCRIPTSET_H_

thirdparty/icu4c/i18n/ucln_in.cpp

@@ -0,0 +1,65 @@
+// © 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:  ucln_in.cpp
+*   encoding:   UTF-8
+*   tab size:   8 (not used)
+*   indentation:4
+*
+*   created on: 2001July05
+*   created by: George Rhoten
+*/
+
+#include "ucln.h"
+#include "ucln_in.h"
+#include "mutex.h"
+#include "uassert.h"
+
+/**  Auto-client for UCLN_I18N **/
+#define UCLN_TYPE UCLN_I18N
+#include "ucln_imp.h"
+
+/* Leave this copyright notice here! It needs to go somewhere in this library. */
+static const char copyright[] = U_COPYRIGHT_STRING;
+
+static cleanupFunc *gCleanupFunctions[UCLN_I18N_COUNT];
+
+static UBool U_CALLCONV i18n_cleanup()
+{
+    int32_t libType = UCLN_I18N_START;
+    (void)copyright;   /* Suppress unused variable warning with clang. */
+
+    while (++libType<UCLN_I18N_COUNT) {
+        if (gCleanupFunctions[libType])
+        {
+            gCleanupFunctions[libType]();
+            gCleanupFunctions[libType] = nullptr;
+        }
+    }
+#if !UCLN_NO_AUTO_CLEANUP && (defined(UCLN_AUTO_ATEXIT) || defined(UCLN_AUTO_LOCAL))
+    ucln_unRegisterAutomaticCleanup();
+#endif
+    return true;
+}
+
+void ucln_i18n_registerCleanup(ECleanupI18NType type,
+                               cleanupFunc *func) {
+    U_ASSERT(UCLN_I18N_START < type && type < UCLN_I18N_COUNT);
+    {
+        icu::Mutex m;   // See ticket 10295 for discussion.
+        ucln_registerCleanup(UCLN_I18N, i18n_cleanup);
+        if (UCLN_I18N_START < type && type < UCLN_I18N_COUNT) {
+            gCleanupFunctions[type] = func;
+        }
+    }
+#if !UCLN_NO_AUTO_CLEANUP && (defined(UCLN_AUTO_ATEXIT) || defined(UCLN_AUTO_LOCAL))
+    ucln_registerAutomaticCleanup();
+#endif
+}
+

thirdparty/icu4c/i18n/ucln_in.h

@@ -0,0 +1,76 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+/*
+******************************************************************************
+* Copyright (C) 2001-2016, International Business Machines
+*                Corporation and others. All Rights Reserved.
+******************************************************************************
+*   file name:  ucln_in.h
+*   encoding:   UTF-8
+*   tab size:   8 (not used)
+*   indentation:4
+*
+*   created on: 2001July05
+*   created by: George Rhoten
+*/
+
+#ifndef __UCLN_IN_H__
+#define __UCLN_IN_H__
+
+#include "unicode/utypes.h"
+#include "ucln.h"
+
+/*
+Please keep the order of enums declared in same order
+as the functions are suppose to be called.
+It's usually best to have child dependencies called first. */
+typedef enum ECleanupI18NType {
+    UCLN_I18N_START = -1,
+    UCLN_I18N_UNIT_EXTRAS,
+    UCLN_I18N_NUMBER_SKELETONS,
+    UCLN_I18N_CURRENCY_SPACING,
+    UCLN_I18N_SPOOF,
+    UCLN_I18N_SPOOFDATA,
+    UCLN_I18N_TRANSLITERATOR,
+    UCLN_I18N_REGEX,
+    UCLN_I18N_JAPANESE_CALENDAR,
+    UCLN_I18N_ISLAMIC_CALENDAR,
+    UCLN_I18N_CHINESE_CALENDAR,
+    UCLN_I18N_HEBREW_CALENDAR,
+    UCLN_I18N_ASTRO_CALENDAR,
+    UCLN_I18N_DANGI_CALENDAR,
+    UCLN_I18N_PERSIAN_CALENDAR,
+    UCLN_I18N_CALENDAR,
+    UCLN_I18N_TIMEZONEFORMAT,
+    UCLN_I18N_TZDBTIMEZONENAMES,
+    UCLN_I18N_TIMEZONEGENERICNAMES,
+    UCLN_I18N_TIMEZONENAMES,
+    UCLN_I18N_ZONEMETA,
+    UCLN_I18N_TIMEZONE,
+    UCLN_I18N_DIGITLIST,
+    UCLN_I18N_DECFMT,
+    UCLN_I18N_NUMFMT,
+    UCLN_I18N_ALLOWED_HOUR_FORMATS,
+    UCLN_I18N_DAYPERIODRULES,
+    UCLN_I18N_SMPDTFMT,
+    UCLN_I18N_USEARCH,
+    UCLN_I18N_COLLATOR,
+    UCLN_I18N_UCOL_RES,
+    UCLN_I18N_CSDET,
+    UCLN_I18N_COLLATION_ROOT,
+    UCLN_I18N_GENDERINFO,
+    UCLN_I18N_CDFINFO,
+    UCLN_I18N_REGION,
+    UCLN_I18N_LIST_FORMATTER,
+    UCLN_I18N_NUMSYS,
+    UCLN_I18N_MF2_UNISETS,
+    UCLN_I18N_COUNT /* This must be last */
+} ECleanupI18NType;
+
+/* Main library cleanup registration function. */
+/* See common/ucln.h for details on adding a cleanup function. */
+/* Note: the global mutex must not be held when calling this function. */
+U_CFUNC void U_EXPORT2 ucln_i18n_registerCleanup(ECleanupI18NType type,
+                                                 cleanupFunc *func);
+
+#endif

thirdparty/icu4c/i18n/uspoof.cpp

@@ -0,0 +1,925 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+/*
+***************************************************************************
+* Copyright (C) 2008-2015, International Business Machines Corporation
+* and others. All Rights Reserved.
+***************************************************************************
+*   file name:  uspoof.cpp
+*   encoding:   UTF-8
+*   tab size:   8 (not used)
+*   indentation:4
+*
+*   created on: 2008Feb13
+*   created by: Andy Heninger
+*
+*   Unicode Spoof Detection
+*/
+#include "unicode/ubidi.h"
+#include "unicode/utypes.h"
+#include "unicode/normalizer2.h"
+#include "unicode/uspoof.h"
+#include "unicode/ustring.h"
+#include "unicode/utf16.h"
+#include "cmemory.h"
+#include "cstring.h"
+#include "mutex.h"
+#include "scriptset.h"
+#include "uassert.h"
+#include "ucln_in.h"
+#include "uspoof_impl.h"
+#include "umutex.h"
+
+
+#if !UCONFIG_NO_NORMALIZATION
+
+U_NAMESPACE_USE
+
+
+//
+// Static Objects used by the spoof impl, their thread safe initialization and their cleanup.
+//
+static UnicodeSet *gInclusionSet = nullptr;
+static UnicodeSet *gRecommendedSet = nullptr;
+static const Normalizer2 *gNfdNormalizer = nullptr;
+static UInitOnce gSpoofInitStaticsOnce {};
+
+namespace {
+
+UBool U_CALLCONV
+uspoof_cleanup() {
+    delete gInclusionSet;
+    gInclusionSet = nullptr;
+    delete gRecommendedSet;
+    gRecommendedSet = nullptr;
+    gNfdNormalizer = nullptr;
+    gSpoofInitStaticsOnce.reset();
+    return true;
+}
+
+void U_CALLCONV initializeStatics(UErrorCode &status) {
+    gInclusionSet = new UnicodeSet();
+    gRecommendedSet = new UnicodeSet();
+    if (gInclusionSet == nullptr || gRecommendedSet == nullptr) {
+        status = U_MEMORY_ALLOCATION_ERROR;
+        delete gInclusionSet;
+        gInclusionSet = nullptr;
+        delete gRecommendedSet;
+        gRecommendedSet = nullptr;
+        return;
+    }
+    gInclusionSet->applyIntPropertyValue(UCHAR_IDENTIFIER_TYPE, U_ID_TYPE_INCLUSION, status);
+    gRecommendedSet->applyIntPropertyValue(UCHAR_IDENTIFIER_TYPE, U_ID_TYPE_RECOMMENDED, status);
+    if (U_FAILURE(status)) {
+        delete gInclusionSet;
+        gInclusionSet = nullptr;
+        delete gRecommendedSet;
+        gRecommendedSet = nullptr;
+        return;
+    }
+    gInclusionSet->freeze();
+    gRecommendedSet->freeze();
+    gNfdNormalizer = Normalizer2::getNFDInstance(status);
+    ucln_i18n_registerCleanup(UCLN_I18N_SPOOF, uspoof_cleanup);
+}
+
+}  // namespace
+
+U_CFUNC void uspoof_internalInitStatics(UErrorCode *status) {
+    umtx_initOnce(gSpoofInitStaticsOnce, &initializeStatics, *status);
+}
+
+U_CAPI USpoofChecker * U_EXPORT2
+uspoof_open(UErrorCode *status) {
+    umtx_initOnce(gSpoofInitStaticsOnce, &initializeStatics, *status);
+    if (U_FAILURE(*status)) {
+        return nullptr;
+    }
+    SpoofImpl *si = new SpoofImpl(*status);
+    if (si == nullptr) {
+        *status = U_MEMORY_ALLOCATION_ERROR;
+        return nullptr;
+    }
+    if (U_FAILURE(*status)) {
+        delete si;
+        return nullptr;
+    }
+    return si->asUSpoofChecker();
+}
+
+
+U_CAPI USpoofChecker * U_EXPORT2
+uspoof_openFromSerialized(const void *data, int32_t length, int32_t *pActualLength,
+                          UErrorCode *status) {
+    if (U_FAILURE(*status)) {
+        return nullptr;
+    }
+
+    if (data == nullptr) {
+        *status = U_ILLEGAL_ARGUMENT_ERROR;
+        return nullptr;
+    }
+
+    umtx_initOnce(gSpoofInitStaticsOnce, &initializeStatics, *status);
+    if (U_FAILURE(*status))
+    {
+        return nullptr;
+    }
+
+    SpoofData *sd = new SpoofData(data, length, *status);
+    if (sd == nullptr) {
+        *status = U_MEMORY_ALLOCATION_ERROR;
+        return nullptr;
+    }
+
+    if (U_FAILURE(*status)) {
+        delete sd;
+        return nullptr;
+    }
+
+    SpoofImpl *si = new SpoofImpl(sd, *status);
+    if (si == nullptr) {
+        *status = U_MEMORY_ALLOCATION_ERROR;
+        delete sd; // explicit delete as the destructor for si won't be called.
+        return nullptr;
+    }
+
+    if (U_FAILURE(*status)) {
+        delete si; // no delete for sd, as the si destructor will delete it.
+        return nullptr;
+    }
+
+    if (pActualLength != nullptr) {
+        *pActualLength = sd->size();
+    }
+    return si->asUSpoofChecker();
+}
+
+
+U_CAPI USpoofChecker * U_EXPORT2
+uspoof_clone(const USpoofChecker *sc, UErrorCode *status) {
+    const SpoofImpl *src = SpoofImpl::validateThis(sc, *status);
+    if (src == nullptr) {
+        return nullptr;
+    }
+    SpoofImpl *result = new SpoofImpl(*src, *status);   // copy constructor
+    if (result == nullptr) {
+        *status = U_MEMORY_ALLOCATION_ERROR;
+        return nullptr;
+    }
+    if (U_FAILURE(*status)) {
+        delete result;
+        result = nullptr;
+    }
+    return result->asUSpoofChecker();
+}
+
+
+U_CAPI void U_EXPORT2
+uspoof_close(USpoofChecker *sc) {
+    UErrorCode status = U_ZERO_ERROR;
+    SpoofImpl *This = SpoofImpl::validateThis(sc, status);
+    delete This;
+}
+
+
+U_CAPI void U_EXPORT2
+uspoof_setChecks(USpoofChecker *sc, int32_t checks, UErrorCode *status) {
+    SpoofImpl *This = SpoofImpl::validateThis(sc, *status);
+    if (This == nullptr) {
+        return;
+    }
+
+    // Verify that the requested checks are all ones (bits) that 
+    //   are acceptable, known values.
+    if (checks & ~(USPOOF_ALL_CHECKS | USPOOF_AUX_INFO)) {
+        *status = U_ILLEGAL_ARGUMENT_ERROR; 
+        return;
+    }
+
+    This->fChecks = checks;
+}
+
+
+U_CAPI int32_t U_EXPORT2
+uspoof_getChecks(const USpoofChecker *sc, UErrorCode *status) {
+    const SpoofImpl *This = SpoofImpl::validateThis(sc, *status);
+    if (This == nullptr) {
+        return 0;
+    }
+    return This->fChecks;
+}
+
+U_CAPI void U_EXPORT2
+uspoof_setRestrictionLevel(USpoofChecker *sc, URestrictionLevel restrictionLevel) {
+    UErrorCode status = U_ZERO_ERROR;
+    SpoofImpl *This = SpoofImpl::validateThis(sc, status);
+    if (This != nullptr) {
+        This->fRestrictionLevel = restrictionLevel;
+        This->fChecks |= USPOOF_RESTRICTION_LEVEL;
+    }
+}
+
+U_CAPI URestrictionLevel U_EXPORT2
+uspoof_getRestrictionLevel(const USpoofChecker *sc) {
+    UErrorCode status = U_ZERO_ERROR;
+    const SpoofImpl *This = SpoofImpl::validateThis(sc, status);
+    if (This == nullptr) {
+        return USPOOF_UNRESTRICTIVE;
+    }
+    return This->fRestrictionLevel;
+}
+
+U_CAPI void U_EXPORT2
+uspoof_setAllowedLocales(USpoofChecker *sc, const char *localesList, UErrorCode *status) {
+    SpoofImpl *This = SpoofImpl::validateThis(sc, *status);
+    if (This == nullptr) {
+        return;
+    }
+    This->setAllowedLocales(localesList, *status);
+}
+
+U_CAPI const char * U_EXPORT2
+uspoof_getAllowedLocales(USpoofChecker *sc, UErrorCode *status) {
+    SpoofImpl *This = SpoofImpl::validateThis(sc, *status);
+    if (This == nullptr) {
+        return nullptr;
+    }
+    return This->getAllowedLocales(*status);
+}
+
+
+U_CAPI const USet * U_EXPORT2
+uspoof_getAllowedChars(const USpoofChecker *sc, UErrorCode *status) {
+    const UnicodeSet *result = uspoof_getAllowedUnicodeSet(sc, status);
+    return result->toUSet();
+}
+
+U_CAPI const UnicodeSet * U_EXPORT2
+uspoof_getAllowedUnicodeSet(const USpoofChecker *sc, UErrorCode *status) {
+    const SpoofImpl *This = SpoofImpl::validateThis(sc, *status);
+    if (This == nullptr) {
+        return nullptr;
+    }
+    return This->fAllowedCharsSet;
+}
+
+
+U_CAPI void U_EXPORT2
+uspoof_setAllowedChars(USpoofChecker *sc, const USet *chars, UErrorCode *status) {
+    const UnicodeSet *set = UnicodeSet::fromUSet(chars);
+    uspoof_setAllowedUnicodeSet(sc, set, status);
+}
+
+
+U_CAPI void U_EXPORT2
+uspoof_setAllowedUnicodeSet(USpoofChecker *sc, const UnicodeSet *chars, UErrorCode *status) {
+    SpoofImpl *This = SpoofImpl::validateThis(sc, *status);
+    if (This == nullptr) {
+        return;
+    }
+    if (chars->isBogus()) {
+        *status = U_ILLEGAL_ARGUMENT_ERROR;
+        return;
+    }
+    UnicodeSet *clonedSet = chars->clone();
+    if (clonedSet == nullptr || clonedSet->isBogus()) {
+        *status = U_MEMORY_ALLOCATION_ERROR;
+        return;
+    }
+    clonedSet->freeze();
+    delete This->fAllowedCharsSet;
+    This->fAllowedCharsSet = clonedSet;
+    This->fChecks |= USPOOF_CHAR_LIMIT;
+}
+
+
+U_CAPI int32_t U_EXPORT2
+uspoof_check(const USpoofChecker *sc,
+             const char16_t *id, int32_t length,
+             int32_t *position,
+             UErrorCode *status) {
+
+    // Backwards compatibility:
+    if (position != nullptr) {
+        *position = 0;
+    }
+
+    // Delegate to uspoof_check2
+    return uspoof_check2(sc, id, length, nullptr, status);
+}
+
+
+U_CAPI int32_t U_EXPORT2
+uspoof_check2(const USpoofChecker *sc,
+    const char16_t* id, int32_t length,
+    USpoofCheckResult* checkResult,
+    UErrorCode *status) {
+
+    const SpoofImpl *This = SpoofImpl::validateThis(sc, *status);
+    if (This == nullptr) {
+        return 0;
+    }
+    if (length < -1) {
+        *status = U_ILLEGAL_ARGUMENT_ERROR;
+        return 0;
+    }
+    UnicodeString idStr((length == -1), id, length);  // Aliasing constructor.
+    int32_t result = uspoof_check2UnicodeString(sc, idStr, checkResult, status);
+    return result;
+}
+
+
+U_CAPI int32_t U_EXPORT2
+uspoof_checkUTF8(const USpoofChecker *sc,
+                 const char *id, int32_t length,
+                 int32_t *position,
+                 UErrorCode *status) {
+
+    // Backwards compatibility:
+    if (position != nullptr) {
+        *position = 0;
+    }
+
+    // Delegate to uspoof_check2
+    return uspoof_check2UTF8(sc, id, length, nullptr, status);
+}
+
+
+U_CAPI int32_t U_EXPORT2
+uspoof_check2UTF8(const USpoofChecker *sc,
+    const char *id, int32_t length,
+    USpoofCheckResult* checkResult,
+    UErrorCode *status) {
+
+    if (U_FAILURE(*status)) {
+        return 0;
+    }
+    UnicodeString idStr = UnicodeString::fromUTF8(StringPiece(id, length>=0 ? length : static_cast<int32_t>(uprv_strlen(id))));
+    int32_t result = uspoof_check2UnicodeString(sc, idStr, checkResult, status);
+    return result;
+}
+
+
+U_CAPI int32_t U_EXPORT2
+uspoof_areConfusable(const USpoofChecker *sc,
+                     const char16_t *id1, int32_t length1,
+                     const char16_t *id2, int32_t length2,
+                     UErrorCode *status) {
+    SpoofImpl::validateThis(sc, *status);
+    if (U_FAILURE(*status)) {
+        return 0;
+    }
+    if (length1 < -1 || length2 < -1) {
+        *status = U_ILLEGAL_ARGUMENT_ERROR;
+        return 0;
+    }
+        
+    UnicodeString id1Str((length1==-1), id1, length1);  // Aliasing constructor
+    UnicodeString id2Str((length2==-1), id2, length2);  // Aliasing constructor
+    return uspoof_areConfusableUnicodeString(sc, id1Str, id2Str, status);
+}
+
+
+U_CAPI int32_t U_EXPORT2
+uspoof_areConfusableUTF8(const USpoofChecker *sc,
+                         const char *id1, int32_t length1,
+                         const char *id2, int32_t length2,
+                         UErrorCode *status) {
+    SpoofImpl::validateThis(sc, *status);
+    if (U_FAILURE(*status)) {
+        return 0;
+    }
+    if (length1 < -1 || length2 < -1) {
+        *status = U_ILLEGAL_ARGUMENT_ERROR;
+        return 0;
+    }
+    UnicodeString id1Str = UnicodeString::fromUTF8(StringPiece(id1, length1>=0? length1 : static_cast<int32_t>(uprv_strlen(id1))));
+    UnicodeString id2Str = UnicodeString::fromUTF8(StringPiece(id2, length2>=0? length2 : static_cast<int32_t>(uprv_strlen(id2))));
+    int32_t results = uspoof_areConfusableUnicodeString(sc, id1Str, id2Str, status);
+    return results;
+}
+ 
+
+U_CAPI int32_t U_EXPORT2
+uspoof_areConfusableUnicodeString(const USpoofChecker *sc,
+                                  const icu::UnicodeString &id1,
+                                  const icu::UnicodeString &id2,
+                                  UErrorCode *status) {
+    const SpoofImpl *This = SpoofImpl::validateThis(sc, *status);
+    if (U_FAILURE(*status)) {
+        return 0;
+    }
+    //
+    // See section 4 of UAX 39 for the algorithm for checking whether two strings are confusable,
+    //   and for definitions of the types (single, whole, mixed-script) of confusables.
+    
+    // We only care about a few of the check flags.  Ignore the others.
+    // If no tests relevant to this function have been specified, return an error.
+    // TODO:  is this really the right thing to do?  It's probably an error on the caller's part,
+    //        but logically we would just return 0 (no error).
+    if ((This->fChecks & USPOOF_CONFUSABLE) == 0) {
+        *status = U_INVALID_STATE_ERROR;
+        return 0;
+    }
+
+    // Compute the skeletons and check for confusability.
+    UnicodeString id1Skeleton;
+    uspoof_getSkeletonUnicodeString(sc, 0 /* deprecated */, id1, id1Skeleton, status);
+    UnicodeString id2Skeleton;
+    uspoof_getSkeletonUnicodeString(sc, 0 /* deprecated */, id2, id2Skeleton, status);
+    if (U_FAILURE(*status)) { return 0; }
+    if (id1Skeleton != id2Skeleton) {
+        return 0;
+    }
+
+    // If we get here, the strings are confusable.  Now we just need to set the flags for the appropriate classes
+    // of confusables according to UTS 39 section 4.
+    // Start by computing the resolved script sets of id1 and id2.
+    ScriptSet id1RSS;
+    This->getResolvedScriptSet(id1, id1RSS, *status);
+    ScriptSet id2RSS;
+    This->getResolvedScriptSet(id2, id2RSS, *status);
+
+    // Turn on all applicable flags
+    int32_t result = 0;
+    if (id1RSS.intersects(id2RSS)) {
+        result |= USPOOF_SINGLE_SCRIPT_CONFUSABLE;
+    } else {
+        result |= USPOOF_MIXED_SCRIPT_CONFUSABLE;
+        if (!id1RSS.isEmpty() && !id2RSS.isEmpty()) {
+            result |= USPOOF_WHOLE_SCRIPT_CONFUSABLE;
+        }
+    }
+
+    // Turn off flags that the user doesn't want
+    if ((This->fChecks & USPOOF_SINGLE_SCRIPT_CONFUSABLE) == 0) {
+        result &= ~USPOOF_SINGLE_SCRIPT_CONFUSABLE;
+    }
+    if ((This->fChecks & USPOOF_MIXED_SCRIPT_CONFUSABLE) == 0) {
+        result &= ~USPOOF_MIXED_SCRIPT_CONFUSABLE;
+    }
+    if ((This->fChecks & USPOOF_WHOLE_SCRIPT_CONFUSABLE) == 0) {
+        result &= ~USPOOF_WHOLE_SCRIPT_CONFUSABLE;
+    }
+
+    return result;
+}
+
+U_CAPI uint32_t U_EXPORT2 uspoof_areBidiConfusable(const USpoofChecker *sc, UBiDiDirection direction,
+                                                  const char16_t *id1, int32_t length1,
+                                                  const char16_t *id2, int32_t length2,
+                                                   UErrorCode *status) {
+    UnicodeString id1Str((length1 == -1), id1, length1); // Aliasing constructor
+    UnicodeString id2Str((length2 == -1), id2, length2); // Aliasing constructor
+    if (id1Str.isBogus() || id2Str.isBogus()) {
+        *status = U_ILLEGAL_ARGUMENT_ERROR;
+        return 0;
+    }
+    return uspoof_areBidiConfusableUnicodeString(sc, direction, id1Str, id2Str, status);
+}
+
+U_CAPI uint32_t U_EXPORT2 uspoof_areBidiConfusableUTF8(const USpoofChecker *sc, UBiDiDirection direction,
+                                                      const char *id1, int32_t length1, const char *id2,
+                                                      int32_t length2, UErrorCode *status) {
+    if (length1 < -1 || length2 < -1) {
+        *status = U_ILLEGAL_ARGUMENT_ERROR;
+        return 0;
+    }
+    UnicodeString id1Str = UnicodeString::fromUTF8(
+        StringPiece(id1, length1 >= 0 ? length1 : static_cast<int32_t>(uprv_strlen(id1))));
+    UnicodeString id2Str = UnicodeString::fromUTF8(
+        StringPiece(id2, length2 >= 0 ? length2 : static_cast<int32_t>(uprv_strlen(id2))));
+    return uspoof_areBidiConfusableUnicodeString(sc, direction, id1Str, id2Str, status);
+}
+
+U_CAPI uint32_t U_EXPORT2 uspoof_areBidiConfusableUnicodeString(const USpoofChecker *sc,
+                                                               UBiDiDirection direction,
+                                                               const icu::UnicodeString &id1,
+                                                               const icu::UnicodeString &id2,
+                                                               UErrorCode *status) {
+    const SpoofImpl *This = SpoofImpl::validateThis(sc, *status);
+    if (U_FAILURE(*status)) {
+        return 0;
+    }
+    //
+    // See section 4 of UTS 39 for the algorithm for checking whether two strings are confusable,
+    //   and for definitions of the types (single, whole, mixed-script) of confusables.
+
+    // We only care about a few of the check flags.  Ignore the others.
+    // If no tests relevant to this function have been specified, return an error.
+    // TODO:  is this really the right thing to do?  It's probably an error on the caller's part,
+    //        but logically we would just return 0 (no error).
+    if ((This->fChecks & USPOOF_CONFUSABLE) == 0) {
+        *status = U_INVALID_STATE_ERROR;
+        return 0;
+    }
+
+    // Compute the skeletons and check for confusability.
+    UnicodeString id1Skeleton;
+    uspoof_getBidiSkeletonUnicodeString(sc, direction, id1, id1Skeleton, status);
+    UnicodeString id2Skeleton;
+    uspoof_getBidiSkeletonUnicodeString(sc, direction, id2, id2Skeleton, status);
+    if (U_FAILURE(*status)) {
+        return 0;
+    }
+    if (id1Skeleton != id2Skeleton) {
+        return 0;
+    }
+
+    // If we get here, the strings are confusable.  Now we just need to set the flags for the appropriate
+    // classes of confusables according to UTS 39 section 4. Start by computing the resolved script sets
+    // of id1 and id2.
+    ScriptSet id1RSS;
+    This->getResolvedScriptSet(id1, id1RSS, *status);
+    ScriptSet id2RSS;
+    This->getResolvedScriptSet(id2, id2RSS, *status);
+
+    // Turn on all applicable flags
+    uint32_t result = 0;
+    if (id1RSS.intersects(id2RSS)) {
+        result |= USPOOF_SINGLE_SCRIPT_CONFUSABLE;
+    } else {
+        result |= USPOOF_MIXED_SCRIPT_CONFUSABLE;
+        if (!id1RSS.isEmpty() && !id2RSS.isEmpty()) {
+            result |= USPOOF_WHOLE_SCRIPT_CONFUSABLE;
+        }
+    }
+
+    // Turn off flags that the user doesn't want
+    return result & This->fChecks;
+}
+
+
+U_CAPI int32_t U_EXPORT2
+uspoof_checkUnicodeString(const USpoofChecker *sc,
+                          const icu::UnicodeString &id,
+                          int32_t *position,
+                          UErrorCode *status) {
+
+    // Backwards compatibility:
+    if (position != nullptr) {
+        *position = 0;
+    }
+
+    // Delegate to uspoof_check2
+    return uspoof_check2UnicodeString(sc, id, nullptr, status);
+}
+
+namespace {
+
+int32_t checkImpl(const SpoofImpl* This, const UnicodeString& id, CheckResult* checkResult, UErrorCode* status) {
+    U_ASSERT(This != nullptr);
+    U_ASSERT(checkResult != nullptr);
+    checkResult->clear();
+    int32_t result = 0;
+
+    if (0 != (This->fChecks & USPOOF_RESTRICTION_LEVEL)) {
+        URestrictionLevel idRestrictionLevel = This->getRestrictionLevel(id, *status);
+        if (idRestrictionLevel > This->fRestrictionLevel) {
+            result |= USPOOF_RESTRICTION_LEVEL;
+        }
+        checkResult->fRestrictionLevel = idRestrictionLevel;
+    }
+
+    if (0 != (This->fChecks & USPOOF_MIXED_NUMBERS)) {
+        UnicodeSet numerics;
+        This->getNumerics(id, numerics, *status);
+        if (numerics.size() > 1) {
+            result |= USPOOF_MIXED_NUMBERS;
+        }
+        checkResult->fNumerics = numerics;  // UnicodeSet::operator=
+    }
+
+    if (0 != (This->fChecks & USPOOF_HIDDEN_OVERLAY)) {
+        int32_t index = This->findHiddenOverlay(id, *status);
+        if (index != -1) {
+            result |= USPOOF_HIDDEN_OVERLAY;
+        }
+    }
+
+
+    if (0 != (This->fChecks & USPOOF_CHAR_LIMIT)) {
+        int32_t i;
+        UChar32 c;
+        int32_t length = id.length();
+        for (i=0; i<length ;) {
+            c = id.char32At(i);
+            i += U16_LENGTH(c);
+            if (!This->fAllowedCharsSet->contains(c)) {
+                result |= USPOOF_CHAR_LIMIT;
+                break;
+            }
+        }
+    }
+
+    if (0 != (This->fChecks & USPOOF_INVISIBLE)) {
+        // This check needs to be done on NFD input
+        UnicodeString nfdText;
+        gNfdNormalizer->normalize(id, nfdText, *status);
+        int32_t nfdLength = nfdText.length();
+
+        // scan for more than one occurrence of the same non-spacing mark
+        // in a sequence of non-spacing marks.
+        int32_t     i;
+        UChar32     c;
+        UChar32     firstNonspacingMark = 0;
+        UBool       haveMultipleMarks = false;  
+        UnicodeSet  marksSeenSoFar;   // Set of combining marks in a single combining sequence.
+        
+        for (i=0; i<nfdLength ;) {
+            c = nfdText.char32At(i);
+            i += U16_LENGTH(c);
+            if (u_charType(c) != U_NON_SPACING_MARK) {
+                firstNonspacingMark = 0;
+                if (haveMultipleMarks) {
+                    marksSeenSoFar.clear();
+                    haveMultipleMarks = false;
+                }
+                continue;
+            }
+            if (firstNonspacingMark == 0) {
+                firstNonspacingMark = c;
+                continue;
+            }
+            if (!haveMultipleMarks) {
+                marksSeenSoFar.add(firstNonspacingMark);
+                haveMultipleMarks = true;
+            }
+            if (marksSeenSoFar.contains(c)) {
+                // report the error, and stop scanning.
+                // No need to find more than the first failure.
+                result |= USPOOF_INVISIBLE;
+                break;
+            }
+            marksSeenSoFar.add(c);
+        }
+    }
+
+    checkResult->fChecks = result;
+    return checkResult->toCombinedBitmask(This->fChecks);
+}
+
+}  // namespace
+
+U_CAPI int32_t U_EXPORT2
+uspoof_check2UnicodeString(const USpoofChecker *sc,
+                          const icu::UnicodeString &id,
+                          USpoofCheckResult* checkResult,
+                          UErrorCode *status) {
+    const SpoofImpl *This = SpoofImpl::validateThis(sc, *status);
+    if (This == nullptr) {
+        return false;
+    }
+
+    if (checkResult != nullptr) {
+        CheckResult* ThisCheckResult = CheckResult::validateThis(checkResult, *status);
+        if (ThisCheckResult == nullptr) {
+            return false;
+        }
+        return checkImpl(This, id, ThisCheckResult, status);
+    } else {
+        // Stack-allocate the checkResult since this method doesn't return it
+        CheckResult stackCheckResult;
+        return checkImpl(This, id, &stackCheckResult, status);
+    }
+}
+
+
+U_CAPI int32_t U_EXPORT2
+uspoof_getSkeleton(const USpoofChecker *sc,
+                   uint32_t type,
+                   const char16_t *id,  int32_t length,
+                   char16_t *dest, int32_t destCapacity,
+                   UErrorCode *status) {
+
+    SpoofImpl::validateThis(sc, *status);
+    if (U_FAILURE(*status)) {
+        return 0;
+    }
+    if (length<-1 || destCapacity<0 || (destCapacity==0 && dest!=nullptr)) {
+        *status = U_ILLEGAL_ARGUMENT_ERROR;
+        return 0;
+    }
+
+    UnicodeString idStr((length==-1), id, length);  // Aliasing constructor
+    UnicodeString destStr;
+    uspoof_getSkeletonUnicodeString(sc, type, idStr, destStr, status);
+    destStr.extract(dest, destCapacity, *status);
+    return destStr.length();
+}
+
+U_CAPI int32_t U_EXPORT2 uspoof_getBidiSkeleton(const USpoofChecker *sc, UBiDiDirection direction,
+                                                const UChar *id, int32_t length, UChar *dest,
+                                                int32_t destCapacity, UErrorCode *status) {
+    UnicodeString idStr((length == -1), id, length); // Aliasing constructor
+    if (idStr.isBogus()) {
+        *status = U_ILLEGAL_ARGUMENT_ERROR;
+        return 0;
+    }
+    UnicodeString destStr;
+    uspoof_getBidiSkeletonUnicodeString(sc, direction, idStr, destStr, status);
+    return destStr.extract(dest, destCapacity, *status);
+}
+
+
+
+U_I18N_API UnicodeString &U_EXPORT2 uspoof_getBidiSkeletonUnicodeString(const USpoofChecker *sc,
+                                                                        UBiDiDirection direction,
+                                                                        const UnicodeString &id,
+                                                                        UnicodeString &dest,
+                                                                        UErrorCode *status) {
+    dest.remove();
+    if (direction != UBIDI_LTR && direction != UBIDI_RTL) {
+      *status = U_ILLEGAL_ARGUMENT_ERROR;
+      return dest;
+    }
+    UBiDi *bidi = ubidi_open();
+    ubidi_setPara(bidi, id.getBuffer(), id.length(), direction,
+                  /*embeddingLevels*/ nullptr, status);
+    if (U_FAILURE(*status)) {
+        ubidi_close(bidi);
+        return dest;
+    }
+    UnicodeString reordered;
+    int32_t const size = ubidi_getProcessedLength(bidi);
+    UChar* const reorderedBuffer = reordered.getBuffer(size);
+    if (reorderedBuffer == nullptr) {
+        *status = U_MEMORY_ALLOCATION_ERROR;
+        ubidi_close(bidi);
+        return dest;
+    }
+    ubidi_writeReordered(bidi, reorderedBuffer, size,
+                         UBIDI_KEEP_BASE_COMBINING | UBIDI_DO_MIRRORING, status);
+    reordered.releaseBuffer(size);
+    ubidi_close(bidi);
+
+    if (U_FAILURE(*status)) {
+        return dest;
+    }
+
+    // The type parameter is deprecated since ICU 58; any number may be passed.
+    constexpr uint32_t deprecatedType = 58;
+    return uspoof_getSkeletonUnicodeString(sc, deprecatedType, reordered, dest, status);
+}
+
+
+
+U_I18N_API UnicodeString &  U_EXPORT2
+uspoof_getSkeletonUnicodeString(const USpoofChecker *sc,
+                                uint32_t /*type*/,
+                                const UnicodeString &id,
+                                UnicodeString &dest,
+                                UErrorCode *status) {
+    const SpoofImpl *This = SpoofImpl::validateThis(sc, *status);
+    if (U_FAILURE(*status)) {
+        return dest;
+    }
+
+    UnicodeString nfdId;
+    gNfdNormalizer->normalize(id, nfdId, *status);
+
+    // Apply the skeleton mapping to the NFD normalized input string
+    // Accumulate the skeleton, possibly unnormalized, in a UnicodeString.
+    int32_t inputIndex = 0;
+    UnicodeString skelStr;
+    int32_t normalizedLen = nfdId.length();
+    for (inputIndex=0; inputIndex < normalizedLen; ) {
+        UChar32 c = nfdId.char32At(inputIndex);
+        inputIndex += U16_LENGTH(c);
+        if (!u_hasBinaryProperty(c, UCHAR_DEFAULT_IGNORABLE_CODE_POINT)) {
+            This->fSpoofData->confusableLookup(c, skelStr);
+        }
+    }
+
+    gNfdNormalizer->normalize(skelStr, dest, *status);
+    return dest;
+}
+
+U_CAPI int32_t U_EXPORT2 uspoof_getSkeletonUTF8(const USpoofChecker *sc, uint32_t type, const char *id,
+                                                int32_t length, char *dest, int32_t destCapacity,
+                       UErrorCode *status) {
+    SpoofImpl::validateThis(sc, *status);
+    if (U_FAILURE(*status)) {
+        return 0;
+    }
+    if (length<-1 || destCapacity<0 || (destCapacity==0 && dest!=nullptr)) {
+        *status = U_ILLEGAL_ARGUMENT_ERROR;
+        return 0;
+    }
+
+    UnicodeString srcStr = UnicodeString::fromUTF8(
+        StringPiece(id, length >= 0 ? length : static_cast<int32_t>(uprv_strlen(id))));
+    UnicodeString destStr;
+    uspoof_getSkeletonUnicodeString(sc, type, srcStr, destStr, status);
+    if (U_FAILURE(*status)) {
+        return 0;
+    }
+
+    int32_t lengthInUTF8 = 0;
+    u_strToUTF8(dest, destCapacity, &lengthInUTF8, destStr.getBuffer(), destStr.length(), status);
+    return lengthInUTF8;
+}
+
+U_CAPI int32_t U_EXPORT2 uspoof_getBidiSkeletonUTF8(const USpoofChecker *sc, UBiDiDirection direction,
+                                                    const char *id, int32_t length, char *dest,
+                                                    int32_t destCapacity, UErrorCode *status) {
+    if (length < -1) {
+        *status = U_ILLEGAL_ARGUMENT_ERROR;
+        return 0;
+    }
+
+    UnicodeString srcStr = UnicodeString::fromUTF8(
+        StringPiece(id, length >= 0 ? length : static_cast<int32_t>(uprv_strlen(id))));
+    UnicodeString destStr;
+    uspoof_getBidiSkeletonUnicodeString(sc, direction, srcStr, destStr, status);
+    if (U_FAILURE(*status)) {
+        return 0;
+    }
+
+    int32_t lengthInUTF8 = 0;
+    u_strToUTF8(dest, destCapacity, &lengthInUTF8, destStr.getBuffer(), destStr.length(), status);
+    return lengthInUTF8;
+}
+
+
+U_CAPI int32_t U_EXPORT2
+uspoof_serialize(USpoofChecker *sc,void *buf, int32_t capacity, UErrorCode *status) {
+    SpoofImpl *This = SpoofImpl::validateThis(sc, *status);
+    if (This == nullptr) {
+        U_ASSERT(U_FAILURE(*status));
+        return 0;
+    }
+
+    return This->fSpoofData->serialize(buf, capacity, *status);
+}
+
+U_CAPI const USet * U_EXPORT2
+uspoof_getInclusionSet(UErrorCode *status) {
+    umtx_initOnce(gSpoofInitStaticsOnce, &initializeStatics, *status);
+    return gInclusionSet->toUSet();
+}
+
+U_CAPI const USet * U_EXPORT2
+uspoof_getRecommendedSet(UErrorCode *status) {
+    umtx_initOnce(gSpoofInitStaticsOnce, &initializeStatics, *status);
+    return gRecommendedSet->toUSet();
+}
+
+U_I18N_API const UnicodeSet * U_EXPORT2
+uspoof_getInclusionUnicodeSet(UErrorCode *status) {
+    umtx_initOnce(gSpoofInitStaticsOnce, &initializeStatics, *status);
+    return gInclusionSet;
+}
+
+U_I18N_API const UnicodeSet * U_EXPORT2
+uspoof_getRecommendedUnicodeSet(UErrorCode *status) {
+    umtx_initOnce(gSpoofInitStaticsOnce, &initializeStatics, *status);
+    return gRecommendedSet;
+}
+
+//------------------
+// CheckResult APIs
+//------------------
+
+U_CAPI USpoofCheckResult* U_EXPORT2
+uspoof_openCheckResult(UErrorCode *status) {
+    CheckResult* checkResult = new CheckResult();
+    if (checkResult == nullptr) {
+        *status = U_MEMORY_ALLOCATION_ERROR;
+        return nullptr;
+    }
+    return checkResult->asUSpoofCheckResult();
+}
+
+U_CAPI void U_EXPORT2
+uspoof_closeCheckResult(USpoofCheckResult* checkResult) {
+    UErrorCode status = U_ZERO_ERROR;
+    CheckResult* This = CheckResult::validateThis(checkResult, status);
+    delete This;
+}
+
+U_CAPI int32_t U_EXPORT2
+uspoof_getCheckResultChecks(const USpoofCheckResult *checkResult, UErrorCode *status) {
+    const CheckResult* This = CheckResult::validateThis(checkResult, *status);
+    if (U_FAILURE(*status)) { return 0; }
+    return This->fChecks;
+}
+
+U_CAPI URestrictionLevel U_EXPORT2
+uspoof_getCheckResultRestrictionLevel(const USpoofCheckResult *checkResult, UErrorCode *status) {
+    const CheckResult* This = CheckResult::validateThis(checkResult, *status);
+    if (U_FAILURE(*status)) { return USPOOF_UNRESTRICTIVE; }
+    return This->fRestrictionLevel;
+}
+
+U_CAPI const USet* U_EXPORT2
+uspoof_getCheckResultNumerics(const USpoofCheckResult *checkResult, UErrorCode *status) {
+    const CheckResult* This = CheckResult::validateThis(checkResult, *status);
+    if (U_FAILURE(*status)) { return nullptr; }
+    return This->fNumerics.toUSet();
+}
+
+
+
+#endif // !UCONFIG_NO_NORMALIZATION

thirdparty/icu4c/i18n/uspoof_impl.cpp

@@ -0,0 +1,959 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+/*
+**********************************************************************
+*   Copyright (C) 2008-2016, International Business Machines
+*   Corporation and others.  All Rights Reserved.
+**********************************************************************
+*/
+
+#include "unicode/utypes.h"
+#include "unicode/uspoof.h"
+#include "unicode/uchar.h"
+#include "unicode/uniset.h"
+#include "unicode/utf16.h"
+#include "utrie2.h"
+#include "cmemory.h"
+#include "cstring.h"
+#include "scriptset.h"
+#include "umutex.h"
+#include "udataswp.h"
+#include "uassert.h"
+#include "ucln_in.h"
+#include "uspoof_impl.h"
+
+#if !UCONFIG_NO_NORMALIZATION
+
+
+U_NAMESPACE_BEGIN
+
+UOBJECT_DEFINE_RTTI_IMPLEMENTATION(SpoofImpl)
+
+SpoofImpl::SpoofImpl(SpoofData *data, UErrorCode& status) {
+    construct(status);
+    fSpoofData = data;
+}
+
+SpoofImpl::SpoofImpl(UErrorCode& status) {
+    construct(status);
+
+    // TODO: Call this method where it is actually needed, instead of in the
+    // constructor, to allow for lazy data loading.  See #12696.
+    fSpoofData = SpoofData::getDefault(status);
+}
+
+SpoofImpl::SpoofImpl() {
+    UErrorCode status = U_ZERO_ERROR;
+    construct(status);
+
+    // TODO: Call this method where it is actually needed, instead of in the
+    // constructor, to allow for lazy data loading.  See #12696.
+    fSpoofData = SpoofData::getDefault(status);
+}
+
+void SpoofImpl::construct(UErrorCode& status) {
+    fChecks = USPOOF_ALL_CHECKS;
+    fSpoofData = nullptr;
+    fAllowedCharsSet = nullptr;
+    fAllowedLocales = nullptr;
+    fRestrictionLevel = USPOOF_HIGHLY_RESTRICTIVE;
+
+    if (U_FAILURE(status)) { return; }
+
+    UnicodeSet *allowedCharsSet = new UnicodeSet(0, 0x10ffff);
+    fAllowedCharsSet = allowedCharsSet;
+    fAllowedLocales  = uprv_strdup("");
+    if (fAllowedCharsSet == nullptr || fAllowedLocales == nullptr) {
+        status = U_MEMORY_ALLOCATION_ERROR;
+        return;
+    }
+    allowedCharsSet->freeze();
+}
+
+
+// Copy Constructor, used by the user level clone() function.
+SpoofImpl::SpoofImpl(const SpoofImpl &src, UErrorCode &status)  :
+        fChecks(USPOOF_ALL_CHECKS), fSpoofData(nullptr), fAllowedCharsSet(nullptr) , 
+        fAllowedLocales(nullptr) {
+    if (U_FAILURE(status)) {
+        return;
+    }
+    fChecks = src.fChecks;
+    if (src.fSpoofData != nullptr) {
+        fSpoofData = src.fSpoofData->addReference();
+    }
+    fAllowedCharsSet = src.fAllowedCharsSet->clone();
+    fAllowedLocales = uprv_strdup(src.fAllowedLocales);
+    if (fAllowedCharsSet == nullptr || fAllowedLocales == nullptr) {
+        status = U_MEMORY_ALLOCATION_ERROR;
+    }
+    fRestrictionLevel = src.fRestrictionLevel;
+}
+
+SpoofImpl::~SpoofImpl() {
+    if (fSpoofData != nullptr) {
+        fSpoofData->removeReference();   // Will delete if refCount goes to zero.
+    }
+    delete fAllowedCharsSet;
+    uprv_free((void *)fAllowedLocales);
+}
+
+//  Cast this instance as a USpoofChecker for the C API.
+USpoofChecker *SpoofImpl::asUSpoofChecker() {
+    return exportForC();
+}
+
+//
+//  Incoming parameter check on Status and the SpoofChecker object
+//    received from the C API.
+//
+const SpoofImpl *SpoofImpl::validateThis(const USpoofChecker *sc, UErrorCode &status) {
+    const auto* This = validate(sc, status);
+    if (U_FAILURE(status)) {
+        return nullptr;
+    }
+    if (This->fSpoofData != nullptr && !This->fSpoofData->validateDataVersion(status)) {
+        return nullptr;
+    }
+    return This;
+}
+
+SpoofImpl *SpoofImpl::validateThis(USpoofChecker *sc, UErrorCode &status) {
+    return const_cast<SpoofImpl *>
+        (SpoofImpl::validateThis(const_cast<const USpoofChecker *>(sc), status));
+}
+
+
+void SpoofImpl::setAllowedLocales(const char *localesList, UErrorCode &status) {
+    UnicodeSet    allowedChars;
+    UnicodeSet    *tmpSet = nullptr;
+    const char    *locStart = localesList;
+    const char    *locEnd = nullptr;
+    const char    *localesListEnd = localesList + uprv_strlen(localesList);
+    int32_t        localeListCount = 0;   // Number of locales provided by caller.
+
+    // Loop runs once per locale from the localesList, a comma separated list of locales.
+    do {
+        locEnd = uprv_strchr(locStart, ',');
+        if (locEnd == nullptr) {
+            locEnd = localesListEnd;
+        }
+        while (*locStart == ' ') {
+            locStart++;
+        }
+        const char *trimmedEnd = locEnd-1;
+        while (trimmedEnd > locStart && *trimmedEnd == ' ') {
+            trimmedEnd--;
+        }
+        if (trimmedEnd <= locStart) {
+            break;
+        }
+        const char* locale = uprv_strndup(locStart, static_cast<int32_t>(trimmedEnd + 1 - locStart));
+        localeListCount++;
+
+        // We have one locale from the locales list.
+        // Add the script chars for this locale to the accumulating set of allowed chars.
+        // If the locale is no good, we will be notified back via status.
+        addScriptChars(locale, &allowedChars, status);
+        uprv_free((void *)locale);
+        if (U_FAILURE(status)) {
+            break;
+        }
+        locStart = locEnd + 1;
+    } while (locStart < localesListEnd);
+
+    // If our caller provided an empty list of locales, we disable the allowed characters checking
+    if (localeListCount == 0) {
+        uprv_free((void *)fAllowedLocales);
+        fAllowedLocales = uprv_strdup("");
+        tmpSet = new UnicodeSet(0, 0x10ffff);
+        if (fAllowedLocales == nullptr || tmpSet == nullptr) {
+            status = U_MEMORY_ALLOCATION_ERROR;
+            return;
+        } 
+        tmpSet->freeze();
+        delete fAllowedCharsSet;
+        fAllowedCharsSet = tmpSet;
+        fChecks &= ~USPOOF_CHAR_LIMIT;
+        return;
+    }
+
+        
+    // Add all common and inherited characters to the set of allowed chars.
+    UnicodeSet tempSet;
+    tempSet.applyIntPropertyValue(UCHAR_SCRIPT, USCRIPT_COMMON, status);
+    allowedChars.addAll(tempSet);
+    tempSet.applyIntPropertyValue(UCHAR_SCRIPT, USCRIPT_INHERITED, status);
+    allowedChars.addAll(tempSet);
+    
+    // If anything went wrong, we bail out without changing
+    // the state of the spoof checker.
+    if (U_FAILURE(status)) {
+        return;
+    }
+
+    // Store the updated spoof checker state.
+    tmpSet = allowedChars.clone();
+    const char *tmpLocalesList = uprv_strdup(localesList);
+    if (tmpSet == nullptr || tmpLocalesList == nullptr) {
+        status = U_MEMORY_ALLOCATION_ERROR;
+        return;
+    }
+    uprv_free((void *)fAllowedLocales);
+    fAllowedLocales = tmpLocalesList;
+    tmpSet->freeze();
+    delete fAllowedCharsSet;
+    fAllowedCharsSet = tmpSet;
+    fChecks |= USPOOF_CHAR_LIMIT;
+}
+
+
+const char * SpoofImpl::getAllowedLocales(UErrorCode &/*status*/) {
+    return fAllowedLocales;
+}
+
+
+// Given a locale (a language), add all the characters from all of the scripts used with that language
+// to the allowedChars UnicodeSet
+
+void SpoofImpl::addScriptChars(const char *locale, UnicodeSet *allowedChars, UErrorCode &status) {
+    UScriptCode scripts[30];
+
+    int32_t numScripts = uscript_getCode(locale, scripts, UPRV_LENGTHOF(scripts), &status);
+    if (U_FAILURE(status)) {
+        return;
+    }
+    if (status == U_USING_DEFAULT_WARNING) {
+        status = U_ILLEGAL_ARGUMENT_ERROR;
+        return;
+    }
+    UnicodeSet tmpSet;
+    int32_t    i;
+    for (i=0; i<numScripts; i++) {
+        tmpSet.applyIntPropertyValue(UCHAR_SCRIPT, scripts[i], status);
+        allowedChars->addAll(tmpSet);
+    }
+}
+
+// Computes the augmented script set for a code point, according to UTS 39 section 5.1.
+void SpoofImpl::getAugmentedScriptSet(UChar32 codePoint, ScriptSet& result, UErrorCode& status) {
+    result.resetAll();
+    result.setScriptExtensions(codePoint, status);
+    if (U_FAILURE(status)) { return; }
+
+    // Section 5.1 step 1
+    if (result.test(USCRIPT_HAN, status)) {
+        result.set(USCRIPT_HAN_WITH_BOPOMOFO, status);
+        result.set(USCRIPT_JAPANESE, status);
+        result.set(USCRIPT_KOREAN, status);
+    }
+    if (result.test(USCRIPT_HIRAGANA, status)) {
+        result.set(USCRIPT_JAPANESE, status);
+    }
+    if (result.test(USCRIPT_KATAKANA, status)) {
+        result.set(USCRIPT_JAPANESE, status);
+    }
+    if (result.test(USCRIPT_HANGUL, status)) {
+        result.set(USCRIPT_KOREAN, status);
+    }
+    if (result.test(USCRIPT_BOPOMOFO, status)) {
+        result.set(USCRIPT_HAN_WITH_BOPOMOFO, status);
+    }
+
+    // Section 5.1 step 2
+    if (result.test(USCRIPT_COMMON, status) || result.test(USCRIPT_INHERITED, status)) {
+        result.setAll();
+    }
+}
+
+// Computes the resolved script set for a string, according to UTS 39 section 5.1.
+void SpoofImpl::getResolvedScriptSet(const UnicodeString& input, ScriptSet& result, UErrorCode& status) const {
+    getResolvedScriptSetWithout(input, USCRIPT_CODE_LIMIT, result, status);
+}
+
+// Computes the resolved script set for a string, omitting characters having the specified script.
+// If USCRIPT_CODE_LIMIT is passed as the second argument, all characters are included.
+void SpoofImpl::getResolvedScriptSetWithout(const UnicodeString& input, UScriptCode script, ScriptSet& result, UErrorCode& status) const {
+    result.setAll();
+
+    ScriptSet temp;
+    UChar32 codePoint;
+    for (int32_t i = 0; i < input.length(); i += U16_LENGTH(codePoint)) {
+        codePoint = input.char32At(i);
+
+        // Compute the augmented script set for the character
+        getAugmentedScriptSet(codePoint, temp, status);
+        if (U_FAILURE(status)) { return; }
+
+        // Intersect the augmented script set with the resolved script set, but only if the character doesn't
+        // have the script specified in the function call
+        if (script == USCRIPT_CODE_LIMIT || !temp.test(script, status)) {
+            result.intersect(temp);
+        }
+    }
+}
+
+// Computes the set of numerics for a string, according to UTS 39 section 5.3.
+void SpoofImpl::getNumerics(const UnicodeString& input, UnicodeSet& result, UErrorCode& /*status*/) const {
+    result.clear();
+
+    UChar32 codePoint;
+    for (int32_t i = 0; i < input.length(); i += U16_LENGTH(codePoint)) {
+        codePoint = input.char32At(i);
+
+        // Store a representative character for each kind of decimal digit
+        if (u_charType(codePoint) == U_DECIMAL_DIGIT_NUMBER) {
+            // Store the zero character as a representative for comparison.
+            // Unicode guarantees it is codePoint - value
+            result.add(codePoint - static_cast<UChar32>(u_getNumericValue(codePoint)));
+        }
+    }
+}
+
+// Computes the restriction level of a string, according to UTS 39 section 5.2.
+URestrictionLevel SpoofImpl::getRestrictionLevel(const UnicodeString& input, UErrorCode& status) const {
+    // Section 5.2 step 1:
+    if (!fAllowedCharsSet->containsAll(input)) {
+        return USPOOF_UNRESTRICTIVE;
+    }
+
+    // Section 5.2 step 2
+    // Java use a static UnicodeSet for this test.  In C++, avoid the static variable
+    // and just do a simple for loop.
+    UBool allASCII = true;
+    for (int32_t i=0, length=input.length(); i<length; i++) {
+        if (input.charAt(i) > 0x7f) {
+            allASCII = false;
+            break;
+        }
+    }
+    if (allASCII) {
+        return USPOOF_ASCII;
+    }
+
+    // Section 5.2 steps 3:
+    ScriptSet resolvedScriptSet;
+    getResolvedScriptSet(input, resolvedScriptSet, status);
+    if (U_FAILURE(status)) { return USPOOF_UNRESTRICTIVE; }
+
+    // Section 5.2 step 4:
+    if (!resolvedScriptSet.isEmpty()) {
+        return USPOOF_SINGLE_SCRIPT_RESTRICTIVE;
+    }
+
+    // Section 5.2 step 5:
+    ScriptSet resolvedNoLatn;
+    getResolvedScriptSetWithout(input, USCRIPT_LATIN, resolvedNoLatn, status);
+    if (U_FAILURE(status)) { return USPOOF_UNRESTRICTIVE; }
+
+    // Section 5.2 step 6:
+    if (resolvedNoLatn.test(USCRIPT_HAN_WITH_BOPOMOFO, status)
+            || resolvedNoLatn.test(USCRIPT_JAPANESE, status)
+            || resolvedNoLatn.test(USCRIPT_KOREAN, status)) {
+        return USPOOF_HIGHLY_RESTRICTIVE;
+    }
+
+    // Section 5.2 step 7:
+    if (!resolvedNoLatn.isEmpty()
+            && !resolvedNoLatn.test(USCRIPT_CYRILLIC, status)
+            && !resolvedNoLatn.test(USCRIPT_GREEK, status)
+            && !resolvedNoLatn.test(USCRIPT_CHEROKEE, status)) {
+        return USPOOF_MODERATELY_RESTRICTIVE;
+    }
+
+    // Section 5.2 step 8:
+    return USPOOF_MINIMALLY_RESTRICTIVE;
+}
+
+int32_t SpoofImpl::findHiddenOverlay(const UnicodeString& input, UErrorCode&) const {
+    bool sawLeadCharacter = false;
+    for (int32_t i=0; i<input.length();) {
+        UChar32 cp = input.char32At(i);
+        if (sawLeadCharacter && cp == 0x0307) {
+            return i;
+        }
+        uint8_t combiningClass = u_getCombiningClass(cp);
+        // Skip over characters except for those with combining class 0 (non-combining characters) or with
+        // combining class 230 (same class as U+0307)
+        U_ASSERT(u_getCombiningClass(0x0307) == 230);
+        if (combiningClass == 0 || combiningClass == 230) {
+            sawLeadCharacter = isIllegalCombiningDotLeadCharacter(cp);
+        }
+        i += U16_LENGTH(cp);
+    }
+    return -1;
+}
+
+static inline bool isIllegalCombiningDotLeadCharacterNoLookup(UChar32 cp) {
+    return cp == u'i' || cp == u'j' || cp == u'ı' || cp == u'ȷ' || cp == u'l' ||
+           u_hasBinaryProperty(cp, UCHAR_SOFT_DOTTED);
+}
+
+bool SpoofImpl::isIllegalCombiningDotLeadCharacter(UChar32 cp) const {
+    if (isIllegalCombiningDotLeadCharacterNoLookup(cp)) {
+        return true;
+    }
+    UnicodeString skelStr;
+    fSpoofData->confusableLookup(cp, skelStr);
+    UChar32 finalCp = skelStr.char32At(skelStr.moveIndex32(skelStr.length(), -1));
+    if (finalCp != cp && isIllegalCombiningDotLeadCharacterNoLookup(finalCp)) {
+        return true;
+    }
+    return false;
+}
+
+
+
+// Convert a text format hex number.  Utility function used by builder code.  Static.
+// Input: char16_t *string text.  Output: a UChar32
+// Input has been pre-checked, and will have no non-hex chars.
+// The number must fall in the code point range of 0..0x10ffff
+// Static Function.
+UChar32 SpoofImpl::ScanHex(const char16_t *s, int32_t start, int32_t limit, UErrorCode &status) {
+    if (U_FAILURE(status)) {
+        return 0;
+    }
+    U_ASSERT(limit-start > 0);
+    uint32_t val = 0;
+    int i;
+    for (i=start; i<limit; i++) {
+        int digitVal = s[i] - 0x30;
+        if (digitVal>9) {
+            digitVal = 0xa + (s[i] - 0x41);  // Upper Case 'A'
+        }
+        if (digitVal>15) {
+            digitVal = 0xa + (s[i] - 0x61);  // Lower Case 'a'
+        }
+        U_ASSERT(digitVal <= 0xf);
+        val <<= 4;
+        val += digitVal;
+    }
+    if (val > 0x10ffff) {
+        status = U_PARSE_ERROR;
+        val = 0;
+    }
+    return static_cast<UChar32>(val);
+}
+
+
+//-----------------------------------------
+//
+//   class CheckResult Implementation
+//
+//-----------------------------------------
+
+CheckResult::CheckResult() {
+    clear();
+}
+
+USpoofCheckResult* CheckResult::asUSpoofCheckResult() {
+    return exportForC();
+}
+
+//
+//  Incoming parameter check on Status and the CheckResult object
+//    received from the C API.
+//
+const CheckResult* CheckResult::validateThis(const USpoofCheckResult *ptr, UErrorCode &status) {
+    return validate(ptr, status);
+}
+
+CheckResult* CheckResult::validateThis(USpoofCheckResult *ptr, UErrorCode &status) {
+    return validate(ptr, status);
+}
+
+void CheckResult::clear() {
+    fChecks = 0;
+    fNumerics.clear();
+    fRestrictionLevel = USPOOF_UNDEFINED_RESTRICTIVE;
+}
+
+int32_t CheckResult::toCombinedBitmask(int32_t enabledChecks) {
+    if ((enabledChecks & USPOOF_AUX_INFO) != 0 && fRestrictionLevel != USPOOF_UNDEFINED_RESTRICTIVE) {
+        return fChecks | fRestrictionLevel;
+    } else {
+        return fChecks;
+    }
+}
+
+CheckResult::~CheckResult() {
+}
+
+//----------------------------------------------------------------------------------------------
+//
+//   class SpoofData Implementation
+//
+//----------------------------------------------------------------------------------------------
+
+
+UBool SpoofData::validateDataVersion(UErrorCode &status) const {
+    if (U_FAILURE(status) ||
+        fRawData == nullptr ||
+        fRawData->fMagic != USPOOF_MAGIC ||
+        fRawData->fFormatVersion[0] != USPOOF_CONFUSABLE_DATA_FORMAT_VERSION ||
+        fRawData->fFormatVersion[1] != 0 ||
+        fRawData->fFormatVersion[2] != 0 ||
+        fRawData->fFormatVersion[3] != 0) {
+            status = U_INVALID_FORMAT_ERROR;
+            return false;
+    }
+    return true;
+}
+
+static UBool U_CALLCONV
+spoofDataIsAcceptable(void *context,
+                        const char * /* type */, const char * /*name*/,
+                        const UDataInfo *pInfo) {
+    if(
+        pInfo->size >= 20 &&
+        pInfo->isBigEndian == U_IS_BIG_ENDIAN &&
+        pInfo->charsetFamily == U_CHARSET_FAMILY &&
+        pInfo->dataFormat[0] == 0x43 &&  // dataFormat="Cfu "
+        pInfo->dataFormat[1] == 0x66 &&
+        pInfo->dataFormat[2] == 0x75 &&
+        pInfo->dataFormat[3] == 0x20 &&
+        pInfo->formatVersion[0] == USPOOF_CONFUSABLE_DATA_FORMAT_VERSION
+    ) {
+        UVersionInfo *version = static_cast<UVersionInfo *>(context);
+        if(version != nullptr) {
+            uprv_memcpy(version, pInfo->dataVersion, 4);
+        }
+        return true;
+    } else {
+        return false;
+    }
+}
+
+//  Methods for the loading of the default confusables data file.  The confusable
+//  data is loaded only when it is needed.
+//
+//  SpoofData::getDefault() - Return the default confusables data, and call the
+//                            initOnce() if it is not available.  Adds a reference
+//                            to the SpoofData that the caller is responsible for
+//                            decrementing when they are done with the data.
+//
+//  uspoof_loadDefaultData - Called once, from initOnce().  The resulting SpoofData
+//                           is shared by all spoof checkers using the default data.
+//
+//  uspoof_cleanupDefaultData - Called during cleanup.
+//
+
+static UInitOnce gSpoofInitDefaultOnce {};
+static SpoofData* gDefaultSpoofData;
+
+static UBool U_CALLCONV
+uspoof_cleanupDefaultData() {
+    if (gDefaultSpoofData) {
+        // Will delete, assuming all user-level spoof checkers were closed.
+        gDefaultSpoofData->removeReference();
+        gDefaultSpoofData = nullptr;
+        gSpoofInitDefaultOnce.reset();
+    }
+    return true;
+}
+
+static void U_CALLCONV uspoof_loadDefaultData(UErrorCode& status) {
+    UDataMemory *udm = udata_openChoice(nullptr, "cfu", "confusables",
+                                        spoofDataIsAcceptable, 
+                                        nullptr,       // context, would receive dataVersion if supplied.
+                                        &status);
+    if (U_FAILURE(status)) { return; }
+    gDefaultSpoofData = new SpoofData(udm, status);
+    if (U_FAILURE(status)) {
+        delete gDefaultSpoofData;
+        gDefaultSpoofData = nullptr;
+        return;
+    }
+    if (gDefaultSpoofData == nullptr) {
+        status = U_MEMORY_ALLOCATION_ERROR;
+        return;
+    }
+    ucln_i18n_registerCleanup(UCLN_I18N_SPOOFDATA, uspoof_cleanupDefaultData);
+}
+
+SpoofData* SpoofData::getDefault(UErrorCode& status) {
+    umtx_initOnce(gSpoofInitDefaultOnce, &uspoof_loadDefaultData, status);
+    if (U_FAILURE(status)) { return nullptr; }
+    gDefaultSpoofData->addReference();
+    return gDefaultSpoofData;
+}
+
+
+
+SpoofData::SpoofData(UDataMemory *udm, UErrorCode &status)
+{
+    reset();
+    if (U_FAILURE(status)) {
+        return;
+    }
+    fUDM = udm;
+    // fRawData is non-const because it may be constructed by the data builder.
+    fRawData = reinterpret_cast<SpoofDataHeader *>(
+            const_cast<void *>(udata_getMemory(udm)));
+    validateDataVersion(status);
+    initPtrs(status);
+}
+
+
+SpoofData::SpoofData(const void *data, int32_t length, UErrorCode &status)
+{
+    reset();
+    if (U_FAILURE(status)) {
+        return;
+    }
+    if (static_cast<size_t>(length) < sizeof(SpoofDataHeader)) {
+        status = U_INVALID_FORMAT_ERROR;
+        return;
+    }
+    if (data == nullptr) {
+        status = U_ILLEGAL_ARGUMENT_ERROR;
+        return;
+    }
+    void *ncData = const_cast<void *>(data);
+    fRawData = static_cast<SpoofDataHeader *>(ncData);
+    if (length < fRawData->fLength) {
+        status = U_INVALID_FORMAT_ERROR;
+        return;
+    }
+    validateDataVersion(status);
+    initPtrs(status);
+}
+
+
+// Spoof Data constructor for use from data builder.
+//   Initializes a new, empty data area that will be populated later.
+SpoofData::SpoofData(UErrorCode &status) {
+    reset();
+    if (U_FAILURE(status)) {
+        return;
+    }
+    fDataOwned = true;
+
+    // The spoof header should already be sized to be a multiple of 16 bytes.
+    // Just in case it's not, round it up.
+    uint32_t initialSize = (sizeof(SpoofDataHeader) + 15) & ~15;
+    U_ASSERT(initialSize == sizeof(SpoofDataHeader));
+    
+    fRawData = static_cast<SpoofDataHeader *>(uprv_malloc(initialSize));
+    fMemLimit = initialSize;
+    if (fRawData == nullptr) {
+        status = U_MEMORY_ALLOCATION_ERROR;
+        return;
+    }
+    uprv_memset(fRawData, 0, initialSize);
+
+    fRawData->fMagic = USPOOF_MAGIC;
+    fRawData->fFormatVersion[0] = USPOOF_CONFUSABLE_DATA_FORMAT_VERSION;
+    fRawData->fFormatVersion[1] = 0;
+    fRawData->fFormatVersion[2] = 0;
+    fRawData->fFormatVersion[3] = 0;
+    initPtrs(status);
+}
+
+// reset() - initialize all fields.
+//           Should be updated if any new fields are added.
+//           Called by constructors to put things in a known initial state.
+void SpoofData::reset() {
+   fRawData = nullptr;
+   fDataOwned = false;
+   fUDM      = nullptr;
+   fMemLimit = 0;
+   fRefCount = 1;
+   fCFUKeys = nullptr;
+   fCFUValues = nullptr;
+   fCFUStrings = nullptr;
+}
+
+
+//  SpoofData::initPtrs()
+//            Initialize the pointers to the various sections of the raw data.
+//
+//            This function is used both during the Trie building process (multiple
+//            times, as the individual data sections are added), and
+//            during the opening of a Spoof Checker from prebuilt data.
+//
+//            The pointers for non-existent data sections (identified by an offset of 0)
+//            are set to nullptr.
+//
+//            Note:  During building the data, adding each new data section
+//            reallocs the raw data area, which likely relocates it, which
+//            in turn requires reinitializing all of the pointers into it, hence
+//            multiple calls to this function during building.
+//
+void SpoofData::initPtrs(UErrorCode &status) {
+    fCFUKeys = nullptr;
+    fCFUValues = nullptr;
+    fCFUStrings = nullptr;
+    if (U_FAILURE(status)) {
+        return;
+    }
+    if (fRawData->fCFUKeys != 0) {
+        fCFUKeys = reinterpret_cast<int32_t*>(reinterpret_cast<char*>(fRawData) + fRawData->fCFUKeys);
+    }
+    if (fRawData->fCFUStringIndex != 0) {
+        fCFUValues = reinterpret_cast<uint16_t*>(reinterpret_cast<char*>(fRawData) + fRawData->fCFUStringIndex);
+    }
+    if (fRawData->fCFUStringTable != 0) {
+        fCFUStrings = reinterpret_cast<char16_t*>(reinterpret_cast<char*>(fRawData) + fRawData->fCFUStringTable);
+    }
+}
+
+
+SpoofData::~SpoofData() {
+    if (fDataOwned) {
+        uprv_free(fRawData);
+    }
+    fRawData = nullptr;
+    if (fUDM != nullptr) {
+        udata_close(fUDM);
+    }
+    fUDM = nullptr;
+}
+
+
+void SpoofData::removeReference() {
+    if (umtx_atomic_dec(&fRefCount) == 0) {
+        delete this;
+    }
+}
+
+
+SpoofData *SpoofData::addReference() {
+    umtx_atomic_inc(&fRefCount);
+    return this;
+}
+
+
+void *SpoofData::reserveSpace(int32_t numBytes,  UErrorCode &status) {
+    if (U_FAILURE(status)) {
+        return nullptr;
+    }
+    if (!fDataOwned) {
+        UPRV_UNREACHABLE_EXIT;
+    }
+
+    numBytes = (numBytes + 15) & ~15;   // Round up to a multiple of 16
+    uint32_t returnOffset = fMemLimit;
+    fMemLimit += numBytes;
+    fRawData = static_cast<SpoofDataHeader *>(uprv_realloc(fRawData, fMemLimit));
+    fRawData->fLength = fMemLimit;
+    uprv_memset((char *)fRawData + returnOffset, 0, numBytes);
+    initPtrs(status);
+    return reinterpret_cast<char*>(fRawData) + returnOffset;
+}
+
+int32_t SpoofData::serialize(void *buf, int32_t capacity, UErrorCode &status) const {
+    int32_t dataSize = fRawData->fLength;
+    if (capacity < dataSize) {
+        status = U_BUFFER_OVERFLOW_ERROR;
+        return dataSize;
+    }
+    uprv_memcpy(buf, fRawData, dataSize);
+    return dataSize;
+}
+
+int32_t SpoofData::size() const {
+    return fRawData->fLength;
+}
+
+//-------------------------------
+//
+// Front-end APIs for SpoofData
+//
+//-------------------------------
+
+int32_t SpoofData::confusableLookup(UChar32 inChar, UnicodeString &dest) const {
+    // Perform a binary search.
+    // [lo, hi), i.e lo is inclusive, hi is exclusive.
+    // The result after the loop will be in lo.
+    int32_t lo = 0;
+    int32_t hi = length();
+    do {
+        int32_t mid = (lo + hi) / 2;
+        if (codePointAt(mid) > inChar) {
+            hi = mid;
+        } else if (codePointAt(mid) < inChar) {
+            lo = mid;
+        } else {
+            // Found result.  Break early.
+            lo = mid;
+            break;
+        }
+    } while (hi - lo > 1);
+
+    // Did we find an entry?  If not, the char maps to itself.
+    if (codePointAt(lo) != inChar) {
+        dest.append(inChar);
+        return 1;
+    }
+
+    // Add the element to the string builder and return.
+    return appendValueTo(lo, dest);
+}
+
+int32_t SpoofData::length() const {
+    return fRawData->fCFUKeysSize;
+}
+
+UChar32 SpoofData::codePointAt(int32_t index) const {
+    return ConfusableDataUtils::keyToCodePoint(fCFUKeys[index]);
+}
+
+int32_t SpoofData::appendValueTo(int32_t index, UnicodeString& dest) const {
+    int32_t stringLength = ConfusableDataUtils::keyToLength(fCFUKeys[index]);
+
+    // Value is either a char (for strings of length 1) or
+    // an index into the string table (for longer strings)
+    uint16_t value = fCFUValues[index];
+    if (stringLength == 1) {
+        dest.append(static_cast<char16_t>(value));
+    } else {
+        dest.append(fCFUStrings + value, stringLength);
+    }
+
+    return stringLength;
+}
+
+
+U_NAMESPACE_END
+
+U_NAMESPACE_USE
+
+//-----------------------------------------------------------------------------
+//
+//  uspoof_swap   -  byte swap and char encoding swap of spoof data
+//
+//-----------------------------------------------------------------------------
+U_CAPI int32_t U_EXPORT2
+uspoof_swap(const UDataSwapper *ds, const void *inData, int32_t length, void *outData,
+           UErrorCode *status) {
+
+    if (status == nullptr || U_FAILURE(*status)) {
+        return 0;
+    }
+    if(ds==nullptr || inData==nullptr || length<-1 || (length>0 && outData==nullptr)) {
+        *status=U_ILLEGAL_ARGUMENT_ERROR;
+        return 0;
+    }
+
+    //
+    //  Check that the data header is for spoof data.
+    //    (Header contents are defined in gencfu.cpp)
+    //
+    const UDataInfo *pInfo = (const UDataInfo *)((const char *)inData+4);
+    if(!(  pInfo->dataFormat[0]==0x43 &&   /* dataFormat="Cfu " */
+           pInfo->dataFormat[1]==0x66 &&
+           pInfo->dataFormat[2]==0x75 &&
+           pInfo->dataFormat[3]==0x20 &&
+           pInfo->formatVersion[0]==USPOOF_CONFUSABLE_DATA_FORMAT_VERSION &&
+           pInfo->formatVersion[1]==0 &&
+           pInfo->formatVersion[2]==0 &&
+           pInfo->formatVersion[3]==0  )) {
+        udata_printError(ds, "uspoof_swap(): data format %02x.%02x.%02x.%02x "
+                             "(format version %02x %02x %02x %02x) is not recognized\n",
+                         pInfo->dataFormat[0], pInfo->dataFormat[1],
+                         pInfo->dataFormat[2], pInfo->dataFormat[3],
+                         pInfo->formatVersion[0], pInfo->formatVersion[1],
+                         pInfo->formatVersion[2], pInfo->formatVersion[3]);
+        *status=U_UNSUPPORTED_ERROR;
+        return 0;
+    }
+
+    //
+    // Swap the data header.  (This is the generic ICU Data Header, not the uspoof Specific
+    //                         header).  This swap also conveniently gets us
+    //                         the size of the ICU d.h., which lets us locate the start
+    //                         of the uspoof specific data.
+    //
+    int32_t headerSize=udata_swapDataHeader(ds, inData, length, outData, status);
+
+
+    //
+    // Get the Spoof Data Header, and check that it appears to be OK.
+    //
+    //
+    const uint8_t   *inBytes =(const uint8_t *)inData+headerSize;
+    SpoofDataHeader *spoofDH = (SpoofDataHeader *)inBytes;
+    if (ds->readUInt32(spoofDH->fMagic)   != USPOOF_MAGIC ||
+        ds->readUInt32(spoofDH->fLength)  <  sizeof(SpoofDataHeader)) 
+    {
+        udata_printError(ds, "uspoof_swap(): Spoof Data header is invalid.\n");
+        *status=U_UNSUPPORTED_ERROR;
+        return 0;
+    }
+
+    //
+    // Prefight operation?  Just return the size
+    //
+    int32_t spoofDataLength = ds->readUInt32(spoofDH->fLength);
+    int32_t totalSize = headerSize + spoofDataLength;
+    if (length < 0) {
+        return totalSize;
+    }
+
+    //
+    // Check that length passed in is consistent with length from Spoof data header.
+    //
+    if (length < totalSize) {
+        udata_printError(ds, "uspoof_swap(): too few bytes (%d after ICU Data header) for spoof data.\n",
+                            spoofDataLength);
+        *status=U_INDEX_OUTOFBOUNDS_ERROR;
+        return 0;
+        }
+
+
+    //
+    // Swap the Data.  Do the data itself first, then the Spoof Data Header, because
+    //                 we need to reference the header to locate the data, and an
+    //                 inplace swap of the header leaves it unusable.
+    //
+    uint8_t          *outBytes = (uint8_t *)outData + headerSize;
+    SpoofDataHeader  *outputDH = (SpoofDataHeader *)outBytes;
+
+    int32_t   sectionStart;
+    int32_t   sectionLength;
+
+    //
+    // If not swapping in place, zero out the output buffer before starting.
+    //    Gaps may exist between the individual sections, and these must be zeroed in
+    //    the output buffer.  The simplest way to do that is to just zero the whole thing.
+    //
+    if (inBytes != outBytes) {
+        uprv_memset(outBytes, 0, spoofDataLength);
+    }
+
+    // Confusables Keys Section   (fCFUKeys)
+    sectionStart  = ds->readUInt32(spoofDH->fCFUKeys);
+    sectionLength = ds->readUInt32(spoofDH->fCFUKeysSize) * 4;
+    ds->swapArray32(ds, inBytes+sectionStart, sectionLength, outBytes+sectionStart, status);
+
+    // String Index Section
+    sectionStart  = ds->readUInt32(spoofDH->fCFUStringIndex);
+    sectionLength = ds->readUInt32(spoofDH->fCFUStringIndexSize) * 2;
+    ds->swapArray16(ds, inBytes+sectionStart, sectionLength, outBytes+sectionStart, status);
+
+    // String Table Section
+    sectionStart  = ds->readUInt32(spoofDH->fCFUStringTable);
+    sectionLength = ds->readUInt32(spoofDH->fCFUStringTableLen) * 2;
+    ds->swapArray16(ds, inBytes+sectionStart, sectionLength, outBytes+sectionStart, status);
+
+    // And, last, swap the header itself.
+    //   int32_t   fMagic             // swap this
+    //   uint8_t   fFormatVersion[4]  // Do not swap this, just copy
+    //   int32_t   fLength and all the rest       // Swap the rest, all is 32 bit stuff.
+    //
+    uint32_t magic = ds->readUInt32(spoofDH->fMagic);
+    ds->writeUInt32((uint32_t *)&outputDH->fMagic, magic);
+
+    if (inBytes != outBytes) {
+        uprv_memcpy(outputDH->fFormatVersion, spoofDH->fFormatVersion, sizeof(spoofDH->fFormatVersion));
+    }
+    // swap starting at fLength
+    ds->swapArray32(ds, &spoofDH->fLength, sizeof(SpoofDataHeader)-8 /* minus magic and fFormatVersion[4] */, &outputDH->fLength, status);
+
+    return totalSize;
+}
+
+#endif
+
+

thirdparty/icu4c/i18n/uspoof_impl.h

@@ -0,0 +1,344 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+/*
+***************************************************************************
+* Copyright (C) 2008-2013, International Business Machines Corporation
+* and others. All Rights Reserved.
+***************************************************************************
+*
+*  uspoof_impl.h
+*
+*    Implementation header for spoof detection
+*
+*/
+
+#ifndef USPOOFIM_H
+#define USPOOFIM_H
+
+#include "uassert.h"
+#include "unicode/utypes.h"
+#include "unicode/uspoof.h"
+#include "unicode/uscript.h"
+#include "unicode/udata.h"
+#include "udataswp.h"
+#include "utrie2.h"
+
+#if !UCONFIG_NO_NORMALIZATION
+
+#ifdef __cplusplus
+
+#include "capi_helper.h"
+#include "umutex.h"
+
+U_NAMESPACE_BEGIN
+
+// The maximum length (in UTF-16 UChars) of the skeleton replacement string resulting from
+//   a single input code point.  This is function of the unicode.org data.
+#define USPOOF_MAX_SKELETON_EXPANSION 20
+
+// The default stack buffer size for copies or conversions or normalizations
+// of input strings being checked.  (Used in multiple places.)
+#define USPOOF_STACK_BUFFER_SIZE 100
+
+// Magic number for sanity checking spoof data.
+#define USPOOF_MAGIC 0x3845fdef
+
+// Magic number for sanity checking spoof checkers.
+#define USPOOF_CHECK_MAGIC 0x2734ecde
+
+class ScriptSet;
+class SpoofData;
+struct SpoofDataHeader;
+class ConfusableDataUtils;
+
+/**
+  *  Class SpoofImpl corresponds directly to the plain C API opaque type
+  *  USpoofChecker.  One can be cast to the other.
+  */
+class SpoofImpl : public UObject,
+        public IcuCApiHelper<USpoofChecker, SpoofImpl, USPOOF_MAGIC> {
+public:
+    SpoofImpl(SpoofData *data, UErrorCode& status);
+    SpoofImpl(UErrorCode& status);
+    SpoofImpl();
+    void construct(UErrorCode& status);
+    virtual ~SpoofImpl();
+
+    /** Copy constructor, used by the user level uspoof_clone() function.
+     */
+    SpoofImpl(const SpoofImpl &src, UErrorCode &status);
+    
+    USpoofChecker *asUSpoofChecker();
+    static SpoofImpl *validateThis(USpoofChecker *sc, UErrorCode &status);
+    static const SpoofImpl *validateThis(const USpoofChecker *sc, UErrorCode &status);
+
+    /** Set and Get AllowedLocales, implementations of the corresponding API */
+    void setAllowedLocales(const char *localesList, UErrorCode &status);
+    const char * getAllowedLocales(UErrorCode &status);
+
+    // Add (union) to the UnicodeSet all of the characters for the scripts used for
+    // the specified locale.  Part of the implementation of setAllowedLocales.
+    void addScriptChars(const char *locale, UnicodeSet *allowedChars, UErrorCode &status);
+
+    // Functions implementing the features of UTS 39 section 5.
+    static void getAugmentedScriptSet(UChar32 codePoint, ScriptSet& result, UErrorCode& status);
+    void getResolvedScriptSet(const UnicodeString& input, ScriptSet& result, UErrorCode& status) const;
+    void getResolvedScriptSetWithout(const UnicodeString& input, UScriptCode script, ScriptSet& result, UErrorCode& status) const;
+    void getNumerics(const UnicodeString& input, UnicodeSet& result, UErrorCode& status) const;
+    URestrictionLevel getRestrictionLevel(const UnicodeString& input, UErrorCode& status) const;
+
+    int32_t findHiddenOverlay(const UnicodeString& input, UErrorCode& status) const;
+    bool isIllegalCombiningDotLeadCharacter(UChar32 cp) const;
+
+    /** parse a hex number.  Untility used by the builders.   */
+    static UChar32 ScanHex(const char16_t *s, int32_t start, int32_t limit, UErrorCode &status);
+
+    static UClassID U_EXPORT2 getStaticClassID();
+    virtual UClassID getDynamicClassID() const override;
+
+    //
+    // Data Members
+    //
+
+    int32_t           fChecks;            // Bit vector of checks to perform.
+
+    SpoofData        *fSpoofData;
+    
+    const UnicodeSet *fAllowedCharsSet;   // The UnicodeSet of allowed characters.
+                                          //   for this Spoof Checker.  Defaults to all chars.
+
+    const char       *fAllowedLocales;    // The list of allowed locales.
+    URestrictionLevel fRestrictionLevel;  // The maximum restriction level for an acceptable identifier.
+};
+
+/**
+ *  Class CheckResult corresponds directly to the plain C API opaque type
+ *  USpoofCheckResult.  One can be cast to the other.
+ */
+class CheckResult : public UObject,
+        public IcuCApiHelper<USpoofCheckResult, CheckResult, USPOOF_CHECK_MAGIC> {
+public:
+    CheckResult();
+    virtual ~CheckResult();
+
+    USpoofCheckResult *asUSpoofCheckResult();
+    static CheckResult *validateThis(USpoofCheckResult *ptr, UErrorCode &status);
+    static const CheckResult *validateThis(const USpoofCheckResult *ptr, UErrorCode &status);
+
+    void clear();
+
+    // Used to convert this CheckResult to the older int32_t return value API
+    int32_t toCombinedBitmask(int32_t expectedChecks);
+
+    // Data Members
+    int32_t fChecks;                       // Bit vector of checks that were failed.
+    UnicodeSet fNumerics;                  // Set of numerics found in the string.
+    URestrictionLevel fRestrictionLevel;   // The restriction level of the string.
+};
+
+
+//
+//  Confusable Mappings Data Structures, version 2.0
+//
+//    For the confusable data, we are essentially implementing a map,
+//       key:    a code point
+//       value:  a string.  Most commonly one char in length, but can be more.
+//
+//    The keys are stored as a sorted array of 32 bit ints.
+//             bits 0-23    a code point value
+//             bits 24-31   length of value string, in UChars (between 1 and 256 UChars).
+//        The key table is sorted in ascending code point order.  (not on the
+//        32 bit int value, the flag bits do not participate in the sorting.)
+//
+//        Lookup is done by means of a binary search in the key table.
+//
+//    The corresponding values are kept in a parallel array of 16 bit ints.
+//        If the value string is of length 1, it is literally in the value array.
+//        For longer strings, the value array contains an index into the strings table.
+//
+//    String Table:
+//       The strings table contains all of the value strings (those of length two or greater)
+//       concatenated together into one long char16_t (UTF-16) array.
+//
+//       There is no nul character or other mark between adjacent strings.
+//
+//----------------------------------------------------------------------------
+//
+//  Changes from format version 1 to format version 2:
+//      1) Removal of the whole-script confusable data tables.
+//      2) Removal of the SL/SA/ML/MA and multi-table flags in the key bitmask.
+//      3) Expansion of string length value in the key bitmask from 2 bits to 8 bits.
+//      4) Removal of the string lengths table since 8 bits is sufficient for the
+//         lengths of all entries in confusables.txt.
+
+
+
+// Internal functions for manipulating confusable data table keys
+#define USPOOF_CONFUSABLE_DATA_FORMAT_VERSION 2  // version for ICU 58
+class ConfusableDataUtils {
+public:
+    inline static UChar32 keyToCodePoint(int32_t key) {
+        return key & 0x00ffffff;
+    }
+    inline static int32_t keyToLength(int32_t key) {
+        return ((key & 0xff000000) >> 24) + 1;
+    }
+    inline static int32_t codePointAndLengthToKey(UChar32 codePoint, int32_t length) {
+        U_ASSERT((codePoint & 0x00ffffff) == codePoint);
+        U_ASSERT(length <= 256);
+        return codePoint | ((length - 1) << 24);
+    }
+};
+
+
+//-------------------------------------------------------------------------------------
+//
+//  SpoofData
+//
+//    A small class that wraps the raw (usually memory mapped) spoof data.
+//    Serves two primary functions:
+//      1.  Convenience.  Contains real pointers to the data, to avoid dealing with
+//          the offsets in the raw data.
+//      2.  Reference counting.  When a spoof checker is cloned, the raw data is shared
+//          and must be retained until all checkers using the data are closed.
+//    Nothing in this struct includes state that is specific to any particular
+//    USpoofDetector object.
+//
+//---------------------------------------------------------------------------------------
+class SpoofData: public UMemory {
+  public:
+    static SpoofData* getDefault(UErrorCode &status);   // Get standard ICU spoof data.
+    static void releaseDefault();   // Cleanup reference to default spoof data.
+
+    SpoofData(UErrorCode &status);   // Create new spoof data wrapper.
+                                     // Only used when building new data from rules.
+    
+    // Constructor for use when creating from prebuilt default data.
+    //   A UDataMemory is what the ICU internal data loading functions provide.
+    //   The udm is adopted by the SpoofData.
+    SpoofData(UDataMemory *udm, UErrorCode &status);
+
+    // Constructor for use when creating from serialized data.
+    //
+    SpoofData(const void *serializedData, int32_t length, UErrorCode &status);
+
+    //  Check raw Spoof Data Version compatibility.
+    //  Return true it looks good.
+    UBool validateDataVersion(UErrorCode &status) const;
+
+    ~SpoofData();                    // Destructor not normally used.
+                                     // Use removeReference() instead.
+    // Reference Counting functions.
+    //    Clone of a user-level spoof detector increments the ref count on the data.
+    //    Close of a user-level spoof detector decrements the ref count.
+    //    If the data is owned by us, it will be deleted when count goes to zero.
+    SpoofData *addReference(); 
+    void removeReference();
+
+    // Reset all fields to an initial state.
+    // Called from the top of all constructors.
+    void reset();
+
+    // Copy this instance's raw data buffer to the specified address.
+    int32_t serialize(void *buf, int32_t capacity, UErrorCode &status) const;
+
+    // Get the total number of bytes of data backed by this SpoofData.
+    // Not to be confused with length, which returns the number of confusable entries.
+    int32_t size() const;
+
+    // Get the confusable skeleton transform for a single code point.
+    // The result is a string with a length between 1 and 18 as of Unicode 9.
+    // This is the main public endpoint for this class.
+    // @return   The length in UTF-16 code units of the substitution string.
+    int32_t confusableLookup(UChar32 inChar, UnicodeString &dest) const;
+
+    // Get the number of confusable entries in this SpoofData.
+    int32_t length() const;
+
+    // Get the code point (key) at the specified index.
+    UChar32 codePointAt(int32_t index) const;
+
+    // Get the confusable skeleton (value) at the specified index.
+    // Append it to the specified UnicodeString&.
+    // @return   The length in UTF-16 code units of the skeleton string.
+    int32_t appendValueTo(int32_t index, UnicodeString& dest) const;
+
+  private:
+    // Reserve space in the raw data.  For use by builder when putting together a
+    //   new set of data.  Init the new storage to zero, to prevent inconsistent
+    //   results if it is not all otherwise set by the requester.
+    //  Return:
+    //    pointer to the new space that was added by this function.
+    void *reserveSpace(int32_t numBytes, UErrorCode &status);
+
+    // initialize the pointers from this object to the raw data.
+    void initPtrs(UErrorCode &status);
+
+    SpoofDataHeader             *fRawData;          // Ptr to the raw memory-mapped data
+    UBool                       fDataOwned;         // True if the raw data is owned, and needs
+                                                    //  to be deleted when refcount goes to zero.
+    UDataMemory                 *fUDM;              // If not nullptr, our data came from a
+                                                    //   UDataMemory, which we must close when
+                                                    //   we are done.
+
+    uint32_t                    fMemLimit;          // Limit of available raw data space
+    u_atomic_int32_t            fRefCount;
+
+    // Confusable data
+    int32_t                     *fCFUKeys;
+    uint16_t                    *fCFUValues;
+    char16_t                    *fCFUStrings;
+
+    friend class ConfusabledataBuilder;
+};
+
+//---------------------------------------------------------------------------------------
+//
+//  Raw Binary Data Formats, as loaded from the ICU data file,
+//    or as built by the builder.
+//
+//---------------------------------------------------------------------------------------
+struct SpoofDataHeader {
+    int32_t       fMagic;                // (0x3845fdef)
+    uint8_t       fFormatVersion[4];     // Data Format. Same as the value in struct UDataInfo
+                                         //   if there is one associated with this data.
+    int32_t       fLength;               // Total length in bytes of this spoof data,
+                                         //   including all sections, not just the header.
+
+    // The following four sections refer to data representing the confusable data
+    //   from the Unicode.org data from "confusables.txt"
+
+    int32_t       fCFUKeys;               // byte offset to Keys table (from SpoofDataHeader *)
+    int32_t       fCFUKeysSize;           // number of entries in keys table  (32 bits each)
+
+    // TODO: change name to fCFUValues, for consistency.
+    int32_t       fCFUStringIndex;        // byte offset to String Indexes table
+    int32_t       fCFUStringIndexSize;    // number of entries in String Indexes table (16 bits each)
+                                          //     (number of entries must be same as in Keys table
+
+    int32_t       fCFUStringTable;        // byte offset of String table
+    int32_t       fCFUStringTableLen;     // length of string table (in 16 bit UChars)
+
+    // The following sections are for data from xidmodifications.txt
+
+    int32_t       unused[15];              // Padding, Room for Expansion
+};
+
+
+
+U_NAMESPACE_END
+#endif /* __cplusplus */
+
+/**
+  * Endianness swap function for binary spoof data.
+  * @internal
+  */
+U_CAPI int32_t U_EXPORT2
+uspoof_swap(const UDataSwapper *ds, const void *inData, int32_t length, void *outData,
+            UErrorCode *status);
+
+
+#endif
+
+#endif  /* USPOOFIM_H */
+