xmlreader.cpp

来自「IBM的解析xml的工具Xerces的源代码」· C++ 代码 · 共 1,501 行 · 第 1/4 页

        fCharSizeBuf[fCharsAvail] = 0;
        fCharOfsBuf[fCharsAvail] = 0;
        fCharBuf[fCharsAvail++] = chSpace;
    }
}


XMLReader::~XMLReader()
{
    fMemoryManager->deallocate(fEncodingStr);
    fMemoryManager->deallocate(fPublicId);
    fMemoryManager->deallocate(fSystemId);
    delete fStream;
    delete fTranscoder;
}


// ---------------------------------------------------------------------------
//  XMLReader: Character buffer management methods
// ---------------------------------------------------------------------------
unsigned int XMLReader::getSrcOffset() const
{
    if (!fSrcOfsSupported || !fCalculateSrcOfs)
        ThrowXMLwithMemMgr(RuntimeException, XMLExcepts::Reader_SrcOfsNotSupported, fMemoryManager);

    //
    //  Take the current source offset and add in the sizes that we've
    //  eaten from the source so far.
    //
    if( fCharIndex == 0 ) {
        return fSrcOfsBase;
    }

    if( fCharIndex < fCharsAvail ) {
        return (fSrcOfsBase + fCharOfsBuf[fCharIndex]);
    }

    return (fSrcOfsBase + fCharOfsBuf[fCharIndex-1] + fCharSizeBuf[fCharIndex-1]);
}


bool XMLReader::refreshCharBuffer()
{
    // If the no more flag is set, then don't both doing anything
    if (fNoMore)
        return false;

    unsigned int startInd;

    // See if we have any existing chars.
    const unsigned int spareChars = fCharsAvail - fCharIndex;

    // If we are full, then don't do anything.
    if (spareChars == kCharBufSize)
        return true;

    //
    //  If no transcoder has been created yet, then we never saw the
    //  any encoding="" string and the encoding was not forced, so lets
    //  create one now. We know that it won't change now.
    //
    //  However, note that if we autosensed EBCDIC, then we have to
    //  consider it an error if we never got an encoding since we don't
    //  know what variant of EBCDIC it is.
    //
    if (!fTranscoder)
    {
        if (fEncoding == XMLRecognizer::EBCDIC)
            ThrowXMLwithMemMgr(RuntimeException, XMLExcepts::Reader_EncodingStrRequired, fMemoryManager);

        // Ask the transcoding service to make use a transcoder
        XMLTransService::Codes failReason;
        fTranscoder = XMLPlatformUtils::fgTransService->makeNewTranscoderFor
        (
            fEncodingStr
            , failReason
            , kCharBufSize
            , fMemoryManager
        );

        if (!fTranscoder)
        {
            ThrowXMLwithMemMgr1
            (
                TranscodingException
                , XMLExcepts::Trans_CantCreateCvtrFor
                , fEncodingStr
                , fMemoryManager
            );
        }
    }

    //
    //  Add the number of source bytes eaten so far to the base src
    //  offset member.
    //
    if (fCalculateSrcOfs) {
        for (startInd = 0; startInd < fCharIndex; startInd++)
            fSrcOfsBase += fCharSizeBuf[startInd];
    }

    //
    //  If there are spare chars, then move then down to the bottom. We
    //  have to move the char sizes down also.
    //
    startInd = 0;
    if (spareChars)
    {
        for (unsigned int index = fCharIndex; index < fCharsAvail; index++)
        {
            fCharBuf[startInd] = fCharBuf[index];
            fCharSizeBuf[startInd] = fCharSizeBuf[index];
            startInd++;
        }
    }

    //
    //  And then get more chars, starting after any spare chars that were
    //  left over from the last time.
    //
    fCharsAvail = xcodeMoreChars
    (
        &fCharBuf[startInd]
        , &fCharSizeBuf[startInd]
        , kCharBufSize - spareChars
    );

    // Add back in the spare chars
    fCharsAvail += spareChars;

    // Reset the buffer index to zero, so we start from the 0th char again
    fCharIndex = 0;

    //
    //  If no chars available, then we have to check for one last thing. If
    //  this is reader for a PE and its not being expanded inside a literal,
    //  then unget a trailing space. We use a boolean to avoid triggering
    //  this more than once.
    //
    if (!fCharsAvail
    &&  (fType == Type_PE)
    &&  (fRefFrom == RefFrom_NonLiteral)
    &&  !fSentTrailingSpace)
    {
        fCharBuf[0] = chSpace;
        fCharsAvail = 1;
        fSentTrailingSpace = true;
    }

    //
    //  If we are on our first block of chars and the encoding is one of the
    //  UTF-16 formats, then check the first char for the BOM and skip over
    //  it manually.
    //
    if (fCharsAvail)
    {
        if ((fCurLine == 1) && (fCurCol == 1))
        {
            if (((fEncoding == XMLRecognizer::UTF_16L)
            ||   (fEncoding == XMLRecognizer::UTF_16B))
            &&  !startInd)
            {
                if ((fCharBuf[startInd] == chUnicodeMarker)
                ||  (fCharBuf[startInd] == chSwappedUnicodeMarker))
                {
                    fCharIndex++;
                }
            }
            // If there's a utf-8 BOM  (0xEF 0xBB 0xBF), skip past it.
            else {
                const char* asChars = (const char*)fRawByteBuf;
                if ((fRawBytesAvail > XMLRecognizer::fgUTF8BOMLen )&&
                    (XMLString::compareNString(  asChars
                    , XMLRecognizer::fgUTF8BOM
                    , XMLRecognizer::fgUTF8BOMLen) == 0) && !startInd)
                {
                    fCharIndex += XMLRecognizer::fgUTF8BOMLen;
                }
            }
        }
    }

    //
    //  If we get here with no more chars, then set the fNoMore flag which
    //  lets us optimize and know without checking that no more chars are
    //  available.
    //
    if (!fCharsAvail)
        fNoMore = true;

    //  Calculate fCharOfsBuf using the elements from fCharBufSize
    if (fCalculateSrcOfs)
    {
        fCharOfsBuf[0] = 0;
        for (unsigned int index = 1; index < fCharsAvail; ++index) {
            fCharOfsBuf[index] = fCharOfsBuf[index-1]+fCharSizeBuf[index-1];
        }
    }

    return (fCharsAvail != 0);
}



// ---------------------------------------------------------------------------
//  XMLReader: Scanning methods
// ---------------------------------------------------------------------------
bool XMLReader::getName(XMLBuffer& toFill, const bool token)
{
    //  Ok, first lets see if we have chars in the buffer. If not, then lets
    //  reload.
    if (fCharIndex == fCharsAvail)
    {
        if (!refreshCharBuffer())
            return false;
    }

    //  Lets check the first char for being a first name char. If not, then
    //  what's the point in living mannnn? Just give up now. We only do this
    //  if its a name and not a name token that they want.
    if (!token)
    {
        if (fXMLVersion == XMLV1_1 && ((fCharBuf[fCharIndex] >= 0xD800) && (fCharBuf[fCharIndex] <= 0xDB7F))) {
           if ((fCharBuf[fCharIndex+1] < 0xDC00) || (fCharBuf[fCharIndex+1] > 0xDFFF))
               return false;

            // Looks ok, so lets eat it and put it in our buffer
            toFill.append(fCharBuf[fCharIndex++]);
            fCurCol++;
            toFill.append(fCharBuf[fCharIndex++]);
            fCurCol++;
        }
        else {
            if (!isFirstNameChar(fCharBuf[fCharIndex]))
                return false;

            // Looks ok, so lets eat it and put it in our buffer.  Update column also!
            toFill.append(fCharBuf[fCharIndex++]);
            fCurCol++;
        }

    }

    //  And now we loop until we run out of data in this reader or we hit
    //  a non-name char.
    do {

        while (fCharIndex < fCharsAvail)
        {
            //  Check the current char and take it if its a name char. Else
            //  break out.
            if (fXMLVersion == XMLV1_1 && ((fCharBuf[fCharIndex] >= 0xD800) && (fCharBuf[fCharIndex] <= 0xDB7F))) {
               if ((fCharBuf[fCharIndex+1] < 0xDC00) || (fCharBuf[fCharIndex+1] > 0xDFFF))
                   return !toFill.isEmpty();

                toFill.append(fCharBuf[fCharIndex++]);
                fCurCol++;
                toFill.append(fCharBuf[fCharIndex++]);
                fCurCol++;
            }
            else {
                if (!isNameChar(fCharBuf[fCharIndex]))
                {
                    return !toFill.isEmpty();
                }

                toFill.append(fCharBuf[fCharIndex++]);
                fCurCol++;
            }
        }

    // If we don't get no more, then break out.
    } while (refreshCharBuffer());

    return !toFill.isEmpty();
}


bool XMLReader::getSpaces(XMLBuffer& toFill)
{
    //
    //  We just loop until we either hit a non-space or the end of this
    //  entity. We return true if we returned because of a non-space and
    //  false if because of end of entity.
    //
    //  NOTE:   We have to maintain line/col info here and we have to do
    //          whitespace normalization if we are not already internalized.
    //
    while (true)
    {
        // Loop through the current chars in the buffer
        while (fCharIndex < fCharsAvail)
        {
            // Get the current char out of the buffer
            XMLCh curCh = fCharBuf[fCharIndex];

            //
            //  See if its a white space char. If so, then process it. Else
            //  we've hit a non-space and need to return.
            //
            if (isWhitespace(curCh))
            {
                // Eat this char
                fCharIndex++;

                handleEOL(curCh, false);

                // Ok we can add this guy to our buffer
                toFill.append(curCh);
            }
             else
            {
                // Return true to indicate we broke out due to a whitespace
                return true;
            }
        }

        //
        //  We've eaten up the current buffer, so lets try to reload it. If
        //  we don't get anything new, then break out. If we do, then we go
        //  back to the top to keep getting spaces.
        //
        if (!refreshCharBuffer())
            break;
    }
    return false;
}


bool XMLReader::getUpToCharOrWS(XMLBuffer& toFill, const XMLCh toCheck)
{
    while (true)
    {
        // Loop through the current chars in the buffer
        while (fCharIndex < fCharsAvail)
        {
            // Get the current char out of the buffer
            XMLCh curCh = fCharBuf[fCharIndex];

            //
            //  See if its not a white space or our target char, then process
            //  it. Else, we need to return.
            //
            if (!isWhitespace(curCh) && (curCh != toCheck))
            {
                // Eat this char
                fCharIndex++;

                handleEOL(curCh, false);

                // Add it to our buffer
                toFill.append(curCh);
            }
             else
            {
                return true;
            }
        }

        //
        //  We've eaten up the current buffer, so lets try to reload it. If
        //  we don't get anything new, then break out. If we do, then we go
        //  back to the top to keep getting spaces.
        //
        if (!refreshCharBuffer())
            break;
    }

    // We never hit any non-space and ate up the whole reader
    return false;

}

bool XMLReader::skipIfQuote(XMLCh& chGotten)
{
    if (fCharIndex == fCharsAvail)