xmlformatter.cpp

来自「IBM的解析xml的工具Xerces的源代码」· C++ 代码 · 共 706 行 · 第 1/2 页

    fIsXML11 = false;  // docVersion 1.0 is not 1.1!
}

XMLFormatter::~XMLFormatter()
{
    fMemoryManager->deallocate(fAposRef); //delete [] fAposRef;
    fMemoryManager->deallocate(fAmpRef); //delete [] fAmpRef;
    fMemoryManager->deallocate(fGTRef); //delete [] fGTRef;
    fMemoryManager->deallocate(fLTRef); //delete [] fLTRef;
    fMemoryManager->deallocate(fQuoteRef); //delete [] fQuoteRef;
    fMemoryManager->deallocate(fOutEncoding); //delete [] fOutEncoding;
    delete fXCoder;

    // We DO NOT own the target object!
}


// ---------------------------------------------------------------------------
//  XMLFormatter: Formatting methods
// ---------------------------------------------------------------------------
void
XMLFormatter::formatBuf(const   XMLCh* const    toFormat
                        , const unsigned int    count
                        , const EscapeFlags     escapeFlags
                        , const UnRepFlags      unrepFlags)
{
    //
    //  Figure out the actual escape flag value. If the parameter is not
    //  the default, then take it. Else take the current default.
    //
    const EscapeFlags actualEsc = (escapeFlags == DefaultEscape)
                                ? fEscapeFlags : escapeFlags;

    // And do the same for the unrep flags
    const UnRepFlags  actualUnRep = (unrepFlags == DefaultUnRep)
                                    ? fUnRepFlags : unrepFlags;

    //
    //  If the actual unrep action is that they want to provide char refs
    //  for unrepresentable chars, then this one is a much more difficult
    //  one to do cleanly, and we handle it separately.
    //
    if (actualUnRep == UnRep_CharRef)
    {
        specialFormat(toFormat, count, actualEsc);
        return;
    }

    //
    //  If we don't have any escape flags set, then we can do the most
    //  efficient loop, else we have to do it the hard way.
    //
    const XMLCh*    srcPtr = toFormat;
    const XMLCh*    endPtr = toFormat + count;
    if (actualEsc == NoEscapes)
    {
        //
        //  Just do a whole buffer at a time into the temp buffer, cap 
        //  it off, and send it to the target. 
        //
        if (srcPtr < endPtr)
           srcPtr += handleUnEscapedChars(srcPtr, endPtr - srcPtr, actualUnRep); 
    }
     else
    {
        //
        //  Escape chars that require it according tot he scale flags 
        //  we were given. For the others, try to accumulate them and 
        //  format them in as big as bulk as we can. 
        //
        while (srcPtr < endPtr)
        {
            //
            //  Run a temp pointer up until we hit a character that we have
            //  to escape. Then we can convert all the chars between our
            //  current source pointer and here all at once.
            //
            const XMLCh* tmpPtr = srcPtr;
            while ((tmpPtr < endPtr) && !inEscapeList(actualEsc, *tmpPtr))
                tmpPtr++;

            //
            //  If we got any chars, then lets convert them and write them
            //  out.
            //
            if (tmpPtr > srcPtr)
               srcPtr += handleUnEscapedChars(srcPtr, tmpPtr - srcPtr,  
                                              actualUnRep); 

             else if (tmpPtr < endPtr)
            {
                //
                //  Ok, so we've hit a char that must be escaped. So do
                //  this one specially.
                //
                const XMLByte * theChars;                
                switch (*srcPtr) { 
                    case chAmpersand :
                        theChars = getCharRef(fAmpLen, fAmpRef, gAmpRef); 
                        fTarget->writeChars(theChars, fAmpLen, this);
                        break;

                    case chSingleQuote :
                        theChars = getCharRef(fAposLen, fAposRef, gAposRef); 
                        fTarget->writeChars(theChars, fAposLen, this);
                        break;

                    case chDoubleQuote :
                        theChars = getCharRef(fQuoteLen, fQuoteRef, gQuoteRef); 
                        fTarget->writeChars(theChars, fQuoteLen, this);
                        break;

                    case chCloseAngle :
                        theChars = getCharRef(fGTLen, fGTRef, gGTRef); 
                        fTarget->writeChars(theChars, fGTLen, this);
                        break;

                    case chOpenAngle :
                        theChars = getCharRef(fLTLen, fLTRef, gLTRef); 
                        fTarget->writeChars(theChars, fLTLen, this);
                        break;

                    default:
                        // control characters
                        writeCharRef(*srcPtr);
                        break;
                }
                srcPtr++;
            }
        }
    }
}


unsigned int 
XMLFormatter::handleUnEscapedChars(const XMLCh *                  srcPtr, 
                                   const unsigned int             oCount, 
                                   const UnRepFlags               actualUnRep) 
{ 
   //
   //  Use that to figure out what I should pass to the transcoder. If we
   //  are doing character references or failing for unrepresentable chars,
   //  then we just throw, since we should never get a call for something
   //  we cannot represent. Else, we tell it to just use the replacement
   //  char.
   //
   const XMLTranscoder::UnRepOpts unRepOpts = (actualUnRep == UnRep_Replace)
                                             ? XMLTranscoder::UnRep_RepChar
                                             : XMLTranscoder::UnRep_Throw;
                                             	
   unsigned int charsEaten; 
   unsigned int count = oCount; 
 
   while (count) { 
      const unsigned srcChars  
         = count > kTmpBufSize ? kTmpBufSize : count; 
 
      const unsigned int outBytes  
         = fXCoder->transcodeTo(srcPtr, srcChars,  
                                fTmpBuf, kTmpBufSize, 
                                charsEaten, unRepOpts); 
 
      if (outBytes) { 
         fTmpBuf[outBytes]     = 0; fTmpBuf[outBytes + 1] = 0; 
         fTmpBuf[outBytes + 2] = 0; fTmpBuf[outBytes + 3] = 0; 
         fTarget->writeChars(fTmpBuf, outBytes, this); 
      } 
 
      srcPtr += charsEaten; 
      count  -= charsEaten; 
   } 
    
   return oCount; // This should be an assertion that count == 0. 
} 
 
 
XMLFormatter& XMLFormatter::operator<<(const XMLCh* const toFormat)
{
    const unsigned int len = XMLString::stringLen(toFormat);
    formatBuf(toFormat, len);
    return *this;
}

XMLFormatter& XMLFormatter::operator<<(const XMLCh toFormat)
{
    // Make a temp string format that
    XMLCh szTmp[2];
    szTmp[0] = toFormat;
    szTmp[1] = 0;

    formatBuf(szTmp, 1);
    return *this;
}

/**
 * the parameter, count, is needed since stringLen()
 * does not work on a BOM like "0xFE0xFF0x000x00" or
 * "0x000x000xFF0xFE"
 **/
void XMLFormatter::writeBOM(const XMLByte* const toFormat
                          , const unsigned int   count)
{
    fTarget->writeChars(toFormat, count, this);    
}

// ---------------------------------------------------------------------------
//  XMLFormatter: Private helper methods
// ---------------------------------------------------------------------------
void XMLFormatter::writeCharRef(const XMLCh &toWrite)
{
    XMLCh tmpBuf[32];
    tmpBuf[0] = chAmpersand;
    tmpBuf[1] = chPound;
    tmpBuf[2] = chLatin_x;

    // Build a char ref for the current char
    XMLString::binToText(toWrite, &tmpBuf[3], 8, 16, fMemoryManager);
    const unsigned int bufLen = XMLString::stringLen(tmpBuf);
    tmpBuf[bufLen] = chSemiColon;
    tmpBuf[bufLen+1] = chNull;

    // write it out
    formatBuf(tmpBuf
            , bufLen + 1
            , XMLFormatter::NoEscapes
            , XMLFormatter::UnRep_Fail);

}

void XMLFormatter::writeCharRef(unsigned long toWrite)
{
    XMLCh tmpBuf[32];
    tmpBuf[0] = chAmpersand;
    tmpBuf[1] = chPound;
    tmpBuf[2] = chLatin_x;

    // Build a char ref for the current char
    XMLString::binToText(toWrite, &tmpBuf[3], 8, 16, fMemoryManager);
    const unsigned int bufLen = XMLString::stringLen(tmpBuf);
    tmpBuf[bufLen] = chSemiColon;
    tmpBuf[bufLen+1] = chNull;

    // write it out
    formatBuf(tmpBuf
            , bufLen + 1
            , XMLFormatter::NoEscapes
            , XMLFormatter::UnRep_Fail);

}


const XMLByte* XMLFormatter::getCharRef(unsigned int & count, 
                                        XMLByte*       &ref, 
                                        const XMLCh *  stdRef) 
{
   if (!ref) { 

       unsigned int charsEaten;
       const unsigned int outBytes = 
           fXCoder->transcodeTo(stdRef, XMLString::stringLen(stdRef), 
                                fTmpBuf, kTmpBufSize, charsEaten, 
                                XMLTranscoder::UnRep_Throw); 

       fTmpBuf[outBytes] = 0; 
       fTmpBuf[outBytes + 1] = 0;
       fTmpBuf[outBytes + 2] = 0; 
       fTmpBuf[outBytes + 3] = 0;

       ref = (XMLByte*) fMemoryManager->allocate
       (
           (outBytes + 4) * sizeof(XMLByte)
       );//new XMLByte[outBytes + 4]; 
       memcpy(ref, fTmpBuf, outBytes + 4); 
       count = outBytes; 
   }

   return ref; 
}

void XMLFormatter::specialFormat(const  XMLCh* const    toFormat
                                , const unsigned int    count
                                , const EscapeFlags     escapeFlags)
{
    //
    //  We have to check each character and see if it could be represented.
    //  As long as it can, we just keep up with where we started and how
    //  many chars we've checked. When we hit an unrepresentable one, we
    //  stop, transcode everything we've collected, then start handling
    //  the unrepresentables via char refs. We repeat this until we get all
    //  the chars done.
    //
    const XMLCh*    srcPtr = toFormat;
    const XMLCh*    endPtr = toFormat + count;

    while (srcPtr < endPtr)
    {
        const XMLCh* tmpPtr = srcPtr;
        while (tmpPtr < endPtr)
        {
            if (fXCoder->canTranscodeTo(*tmpPtr))
                tmpPtr++;
            else
                break;
        }

        if (tmpPtr > srcPtr)
        {
            // We got at least some chars that can be done normally
            formatBuf
            (
                srcPtr
                , tmpPtr - srcPtr
                , escapeFlags
                , XMLFormatter::UnRep_Fail
            );

            // Update the source pointer to our new spot
            srcPtr = tmpPtr;
        }
         else
        {
 
            //  We hit something unrepresentable. So continue forward doing
            //  char refs until we hit something representable again or the
            //  end of input.
            //
            while (srcPtr < endPtr)
            {
                if ((*srcPtr & 0xFC00) == 0xD800) {
                    // we have encountered a surrogate, need to recombine before printing out					  
                    // use writeCharRef that takes unsigned long to get values larger than
                    // hex 0xFFFF printed.
                    tmpPtr = srcPtr;
                    tmpPtr++; // point at low surrogate
                    writeCharRef((unsigned long) (0x10000+((*srcPtr-0xD800)<<10)+*tmpPtr-0xDC00));
                    srcPtr++; // advance to low surrogate (will advance again below)
                }
                else {
                    writeCharRef(*srcPtr);
                }

                // Move up the source pointer and break out if needed
                srcPtr++;
                if (fXCoder->canTranscodeTo(*srcPtr))
                    break;
            }
        }
    }
}

XERCES_CPP_NAMESPACE_END