macosunicodeconverter.cpp

来自「IBM的解析xml的工具Xerces的源代码」· C++ 代码 · 共 1,049 行 · 第 1/3 页

MacOSLCPTranscoder::~MacOSLCPTranscoder()
{
	//	Dispose the XMLTranscoder we're using
    delete mTranscoder;
}


// ---------------------------------------------------------------------------
//  MacOSLCPTranscoder: Implementation of the virtual transcoder interface
// ---------------------------------------------------------------------------

// ---------------------------------------------------------------------------
//	In order to implement calcRequiredSize we have to go ahead and do the
//	conversion, which seems quite painful. The Mac Unicode converter has
//	no way of saying "don't actually do the conversion." So we end up
//	converting twice. It would be nice if the calling code could do some
//	extra buffering to avoid this result.
// ---------------------------------------------------------------------------
unsigned int
MacOSLCPTranscoder::calcRequiredSize(const char* const srcText
                                     , MemoryManager* const manager)
{
	if (!srcText)
		return 0;
		
	//  Lock our mutex to gain exclusive access to the transcoder
	//  since the lcp transcoders are used globally.
	XMLMutexLock lock(&mMutex);

	std::size_t totalCharsProduced = 0;

	const char* src = srcText;
	unsigned int srcCnt = std::strlen(src);
    
    //  Iterate over the characters, converting into a temporary buffer which we'll discard.
    //  All this to get the size required.
	while (srcCnt > 0)
    {
        TempXMLBuf tmpBuf;
        unsigned int bytesConsumed = 0;
		unsigned int charsProduced = mTranscoder->transcodeFrom((XMLByte*)src, srcCnt,
														tmpBuf, kTempBufCount,
														bytesConsumed,
														NULL);
		
        src     += bytesConsumed;
        srcCnt  -= bytesConsumed;

        totalCharsProduced += charsProduced;
        
        //  Bail out if nothing more was produced
        if (charsProduced == 0)
            break;
    }

	//	Return number of XMLCh characters required (not counting terminating NULL!)
	return totalCharsProduced;
}


// ---------------------------------------------------------------------------
//	In order to implement calcRequiredSize we have to go ahead and do the
//	conversion, which seems quite painful. The Mac Unicode converter has
//	no way of saying "don't actually do the conversion." So we end up
//	converting twice. It would be nice if the calling code could do some
//	extra buffering to avoid this result.
// ---------------------------------------------------------------------------
unsigned int
MacOSLCPTranscoder::calcRequiredSize(const XMLCh* const srcText
                                     , MemoryManager* const manager)
{
	if (!srcText)
		return 0;

	//  Lock our mutex to gain exclusive access to the transcoder
	//  since the lcp transcoders are used globally.
	XMLMutexLock lock(&mMutex);
	std::size_t     totalBytesProduced = 0;

	const XMLCh*    src     = srcText;
	unsigned int    srcCnt  = XMLString::stringLen(src);
    
    //  Iterate over the characters, converting into a temporary buffer which we'll discard.
    //  All this to get the size required.
   while (srcCnt > 0)
    {
        TempCharBuf tmpBuf;
        unsigned int charsConsumed = 0;
		unsigned int bytesProduced = mTranscoder->transcodeTo(src, srcCnt,
                                            (XMLByte*)tmpBuf, kTempBufCount,
                                            charsConsumed,
                                            XMLTranscoder::UnRep_RepChar);
        
        src     += charsConsumed;
        srcCnt  -= charsConsumed;

        totalBytesProduced += bytesProduced;
        
        //  Bail out if nothing more was produced
        if (bytesProduced == 0)
            break;
    }

	//	Return number of characters required (not counting terminating NULL!)
	return totalBytesProduced;
}


char*
MacOSLCPTranscoder::transcode(const XMLCh* const srcText)
{
	//	Transcode using a memory manager that allocates
	//	memory using new[].
	return transcode(srcText, XMLPlatformUtils::fgArrayMemoryManager);
}


char*
MacOSLCPTranscoder::transcode(const XMLCh* const srcText,
                              MemoryManager* const manager)
{
	if (!srcText)
		return NULL;

	//  Lock our mutex to gain exclusive access to the transcoder
	//  since the lcp transcoders are used globally.
	XMLMutexLock lock(&mMutex);

	ArrayJanitor<char> result(0);
	const XMLCh* src		= srcText;
	unsigned int srcCnt		= XMLString::stringLen(src);
	std::size_t resultCnt	= 0;

    //  Iterate over the characters, buffering into a local temporary
    //  buffer, which we dump into an allocated (and reallocated, as necessary)
    //  string for return.
    while (srcCnt > 0)
    {
		//  Transcode some characters
        TempCharBuf tmpBuf;
        unsigned int charsConsumed = 0;
        unsigned int bytesProduced = mTranscoder->transcodeTo(src, srcCnt,
                                            (XMLByte*)tmpBuf, kTempBufCount,
                                            charsConsumed,
                                            XMLTranscoder::UnRep_RepChar);
        src     += charsConsumed;
        srcCnt  -= charsConsumed;

		//	Move the data to result buffer, reallocating as needed
		if (bytesProduced > 0)
		{
			//	Allocate space for result
			std::size_t newCnt = resultCnt + bytesProduced;
			ArrayJanitor<char> newResult
            (
                (char*) manager->allocate((newCnt + 1) * sizeof(char)) //new char[newCnt + 1]
                , manager
            );
			if (newResult.get() != NULL)
			{
				//	Incorporate previous result
				if (result.get() != NULL)
					std::memcpy(newResult.get(), result.get(), resultCnt);
				result.reset(newResult.release());

				//	Copy in new data
				std::memcpy(result.get() + resultCnt, tmpBuf, bytesProduced);
				resultCnt = newCnt;
				
                //  Terminate the result
				result[resultCnt] = '\0';					
			}
		}
        else
            break;
    }

    if (!result.get())
	{
		//	No error, and no result: we probably processed a zero length
		//	input, in which case we want a valid zero length output.
		result.reset
        (
            (char*) manager->allocate(sizeof(char))//new char[1]
            , manager
        );
		result[0] = '\0';
	}

	return result.release();
}


XMLCh*
MacOSLCPTranscoder::transcode(const char* const srcText)
{
	//	Transcode using a memory manager that allocates
	//	memory using new[].
	return transcode(srcText, XMLPlatformUtils::fgArrayMemoryManager);
}


XMLCh*
MacOSLCPTranscoder::transcode(const char* const srcText,
                              MemoryManager* const manager)
{
	if (!srcText)
		return NULL;

	//  Lock our mutex to gain exclusive access to the transcoder
	//  since the lcp transcoders are used globally.
	XMLMutexLock lock(&mMutex);

	ArrayJanitor<XMLCh> result(0);
	const char* src			= srcText;
	std::size_t srcCnt		= std::strlen(src);
	std::size_t resultCnt	= 0;

    //  Iterate over the characters, buffering into a local temporary
    //  buffer, which we dump into an allocated (and reallocated, as necessary)
    //  string for return.
    while (srcCnt > 0)
    {
        //  Transcode some characters
		TempXMLBuf tmpBuf;
        unsigned int bytesConsumed = 0;
		unsigned int charsProduced = mTranscoder->transcodeFrom((XMLByte*)src, srcCnt,
												tmpBuf, kTempBufCount,
												bytesConsumed,
												NULL);
        src     += bytesConsumed;
        srcCnt  -= bytesConsumed;

		//	Move the data to result buffer, reallocating as needed
		if (charsProduced > 0)
		{
			//	Allocate space for result
			std::size_t newCnt = resultCnt + charsProduced;
			ArrayJanitor<XMLCh> newResult
            (
                (XMLCh*) manager->allocate((newCnt + 1) * sizeof(XMLCh)) //new XMLCh[newCnt + 1]
                , manager
            );
			if (newResult.get() != NULL)
			{
				//	Incorporate previous result
				if (result.get() != NULL)
					std::memcpy(newResult.get(), result.get(), resultCnt * sizeof(XMLCh));
				result.reset(newResult.release());

				//	Copy in new data
				std::memcpy(result.get() + resultCnt, tmpBuf, charsProduced * sizeof(XMLCh));
				resultCnt = newCnt;
				
				result[resultCnt] = 0;			
			}
		}
        else
            break;
    }

    if (!result.get())
	{
		//	No error, and no result: we probably processed a zero length
		//	input, in which case we want a valid zero length output.
		result.reset
        (
            (XMLCh*) manager->allocate(sizeof(XMLCh))//new XMLCh[1]
            , manager
        );
		result[0] = '\0';
	}
	
	return result.release();
}


bool
MacOSLCPTranscoder::transcode( 		 const   char* const	toTranscode
                                    ,       XMLCh* const    toFill
                                    , const unsigned int    maxChars
                                    , MemoryManager* const  manager)
{
    // toFill must contain space for maxChars XMLCh characters + 1 (for terminating NULL).

    // Check for a couple of psycho corner cases
    if (!toTranscode || !maxChars || !*toTranscode)
    {
        toFill[0] = 0;
        return true;
    }

	//  Lock our mutex to gain exclusive access to the transcoder
	//  since the lcp transcoders are used globally.
	XMLMutexLock lock(&mMutex);

    //  Call the transcoder to do the work
    unsigned int srcLen = std::strlen(toTranscode);
    unsigned int bytesConsumed = 0;
    unsigned int charsProduced = mTranscoder->transcodeFrom((XMLByte*)toTranscode, srcLen,
                                            toFill, maxChars,
											bytesConsumed,
											NULL);

    //	Zero terminate the output string
    toFill[charsProduced] = L'\0';
    
    //  Return true if we consumed all of the characters
    return (bytesConsumed == srcLen);
}


bool
MacOSLCPTranscoder::transcode( 		const   XMLCh* const    toTranscode
                                    ,       char* const     toFill
                                    , const unsigned int    maxChars
                                    , MemoryManager* const  manager)
{
    //	toFill must contain space for maxChars bytes + 1 (for terminating NULL).

    // Check for a couple of psycho corner cases
    if (!toTranscode || !maxChars || !*toTranscode)
    {
        toFill[0] = 0;
        return true;
    }

	//  Lock our mutex to gain exclusive access to the transcoder
	//  since the lcp transcoders are used globally.
	XMLMutexLock lock(&mMutex);

    //  Call the transcoder to do the work
    unsigned int srcLen = XMLString::stringLen(toTranscode);
    unsigned int charsConsumed = 0;
    unsigned int bytesProduced = mTranscoder->transcodeTo(toTranscode, srcLen,
                                            (XMLByte*)toFill, maxChars,
                                            charsConsumed,
                                            XMLTranscoder::UnRep_RepChar);

    //	Zero terminate the output string
    toFill[bytesProduced] = '\0';
    
    //  Return true if we consumed all of the characters
    return (charsConsumed == srcLen);
}


XERCES_CPP_NAMESPACE_END