GrabBag/VrUtils/Src/VrCodeFormatUtils.cpp

#include "VrCodeFormatUtils.h"
#ifdef _WIN32
#include "windows.h"
#endif
//GB2312<31><32>UTF-8<><38>ת<EFBFBD><D7AA>
int CVrCodeFormatUtils::GB2312ToUtf8(const char* gb2312, char* utf8)
{
	int len = 0;
#ifdef _WIN32
	len = MultiByteToWideChar(CP_ACP, 0, gb2312, -1, NULL, 0);
	wchar_t* wstr = new wchar_t[len + 1];
	memset(wstr, 0, len + 1);
	MultiByteToWideChar(CP_ACP, 0, gb2312, -1, wstr, len);
	len = WideCharToMultiByte(CP_UTF8, 0, wstr, -1, NULL, 0, NULL, NULL);
	WideCharToMultiByte(CP_UTF8, 0, wstr, -1, utf8, len, NULL, NULL);
	if (wstr) delete[] wstr;
#endif
	return len;
}

//<2F>ж<EFBFBD><D0B6>Ƿ<EFBFBD><C7B7><EFBFBD>utf8
bool CVrCodeFormatUtils::IsTextUTF8(const char* str, long length)
{
	unsigned char chr;

	int nBytes = 0; // UFT8<54><38><EFBFBD><EFBFBD>1-6<><36><EFBFBD>ֽڱ<D6BD><DAB1><EFBFBD>,ASCII<49><49>һ<EFBFBD><D2BB><EFBFBD>ֽ<EFBFBD>
	bool bAllAscii = true; // <20><><EFBFBD><EFBFBD>ȫ<EFBFBD><C8AB><EFBFBD><EFBFBD><EFBFBD><EFBFBD>ASCII, ˵<><CBB5><EFBFBD><EFBFBD><EFBFBD><EFBFBD>UTF-8

	for (int i = 0; i < length; i++)
	{
		chr = *(str + i);

		if ((chr & 0x80) != 0) // <20>ж<EFBFBD><D0B6>Ƿ<EFBFBD>ASCII<49><49><EFBFBD><EFBFBD>,<2C><><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD>,˵<><CBB5><EFBFBD>п<EFBFBD><D0BF><EFBFBD><EFBFBD><EFBFBD>UTF-8, ASCII<49><49>7λ<37><CEBB><EFBFBD><EFBFBD>,<2C><><EFBFBD><EFBFBD>һ<EFBFBD><D2BB><EFBFBD>ֽڴ<D6BD>,<2C><><EFBFBD><EFBFBD>λ<EFBFBD><CEBB><EFBFBD><EFBFBD>Ϊ0,o0xxxxxxx
		{
			bAllAscii = false;
		}

		if (nBytes == 0) // <20><><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD>ASCII<49><49>,Ӧ<><D3A6><EFBFBD>Ƕ<EFBFBD><C7B6>ֽڷ<D6BD>,<2C><><EFBFBD><EFBFBD><EFBFBD>ֽ<EFBFBD><D6BD><EFBFBD>
		{
			if (chr >= 0x80)
			{
				if (chr >= 0xFC && chr <= 0xFD)
					nBytes = 6;
				else if (chr >= 0xF8)
					nBytes = 5;
				else if (chr >= 0xF0)
					nBytes = 4;
				else if (chr >= 0xE0)
					nBytes = 3;
				else if (chr >= 0xC0)
					nBytes = 2;
				else
					return false;

				nBytes--;
			}
		}
		else // every char of ascii buffer looks like 10xxxxxx, except the first char
		{
			if ((chr & 0xC0) != 0x80)
			{
				return false;
			}
			nBytes--;
		}
	}

	if (nBytes > 0) // format error
	{
		return false;
	}

	if (bAllAscii) // if all chars are ascii, the buffer is not utf-8
	{
		return false;
	}

	return true;
}


//UTF-8<><38>GB2312<31><32>ת<EFBFBD><D7AA>
int CVrCodeFormatUtils::Utf8ToGB2312(const char* utf8, char* gb2312)
{
	int len = 0;
#ifdef _WIN32
	len = MultiByteToWideChar(CP_UTF8, 0, utf8, -1, NULL, 0);
	wchar_t* wstr = new wchar_t[len + 1];
	memset(wstr, 0, len + 1);
	MultiByteToWideChar(CP_UTF8, 0, utf8, -1, wstr, len);
	len = WideCharToMultiByte(CP_ACP, 0, wstr, -1, NULL, 0, NULL, NULL);
	WideCharToMultiByte(CP_ACP, 0, wstr, -1, gb2312, len, NULL, NULL);
	if (wstr) delete[] wstr;
#endif
	return len;
}


//GB2312<31><32>Unicode<64><65>ת<EFBFBD><D7AA>
int CVrCodeFormatUtils::GB2312ToUnicode(const char* gb2312, char* unicode)
{
	int len = 0;
#ifdef _WIN32
	UINT nCodePage = 936; //GB2312
	len = MultiByteToWideChar(nCodePage, 0, gb2312, -1, NULL, 0);
	wchar_t* wstr = new wchar_t[len + 1];
	memset(wstr, 0, len + 1);
	MultiByteToWideChar(nCodePage, 0, gb2312, -1, wstr, len);
	len = len * sizeof(wchar_t);
	memcpy(unicode, wstr, len);
	if (wstr) delete[] wstr;
#endif
	return len;
}


//Unicode<64><65>GB2312<31><32>ת<EFBFBD><D7AA>
int CVrCodeFormatUtils::UnicodeToGB2312(const char* unicode, int size, char*gb2312)
{
	int len = 0;
#ifdef _WIN32
	UINT nCodePage = 936; //GB2312
	wchar_t* wstr = new wchar_t[size / 2 + 1];
	memcpy(wstr, unicode, size);
	len = WideCharToMultiByte(nCodePage, 0, wstr, -1, NULL, 0, NULL, NULL);
	WideCharToMultiByte(nCodePage, 0, wstr, -1, gb2312, len, NULL, NULL);
	if (wstr) delete[] wstr;
#endif
	return len;
}


//UTF-8<><38>Unicode<64><65>ת<EFBFBD><D7AA>
int CVrCodeFormatUtils::Utf8ToUnicode(const char* utf8, char*unicode)
{
	int len = 0;
#ifdef _WIN32
	MultiByteToWideChar(CP_UTF8, 0, utf8, -1, NULL, 0);
	wchar_t* wstr = new wchar_t[len + 1];
	memset(wstr, 0, len + 1);
	MultiByteToWideChar(CP_UTF8, 0, utf8, -1, wstr, len);
	memcpy(unicode, wstr, len);
	if (wstr) delete[] wstr;
#endif
	return len;
}


//Unicode<64><65>UTF-8<><38>ת<EFBFBD><D7AA>
int CVrCodeFormatUtils::UnicodeToUtf8(const char* unicode, int size, char* utf8)
{
	int len = 0;
#ifdef _WIN32
	wchar_t* wstr = new wchar_t[size / 2 + 1];
	memcpy(wstr, unicode, size);
	len = WideCharToMultiByte(CP_UTF8, 0, wstr, -1, NULL, 0, NULL, NULL);
	WideCharToMultiByte(CP_UTF8, 0, wstr, -1, utf8, len, NULL, NULL);
	if (wstr) delete[] wstr;
#endif
	return len;
}