Core: Drop custom copymem/zeromem defines

We've been using standard C library functions `memcpy`/`memset` for these since
2016 with 67f65f66391327b2967a20a89c3627e1dd6e84eb.

There was still the possibility for third-party platform ports to override the
definitions with a custom header, but this doesn't seem useful anymore.

modules/glslang/register_types.cpp

@@ -173,7 +173,7 @@ static Vector<uint8_t> _compile_shader_glsl(RenderingDevice::ShaderStage p_stage
 	ret.resize(SpirV.size() * sizeof(uint32_t));
 	{
 		uint8_t *w = ret.ptrw();
-		copymem(w, &SpirV[0], SpirV.size() * sizeof(uint32_t));
+		memcpy(w, &SpirV[0], SpirV.size() * sizeof(uint32_t));
 	}
 
 	return ret;