wfxmlscanner.cpp

来自「IBM的解析xml的工具Xerces的源代码」· C++ 代码 · 共 1,883 行 · 第 1/5 页

    (
        elemDecl->getElementName()->getPrefix()
        , ElemStack::Mode_Element
    );

    // Now we can update the element stack
    fElemStack.setCurrentURI(uriId);

    // Tell the document handler about this start tag
    if (fDocHandler)
    {
        fDocHandler->startElement
        (
            *elemDecl
            , uriId
            , elemDecl->getElementName()->getPrefix()
            , *fAttrList
            , attCount
            , false
            , isRoot
        );
    }

    //  If empty, validate content right now if we are validating and then
    //  pop the element stack top. Else, we have to update the current stack
    //  top's namespace mapping elements.
    if (isEmpty)
    {
        // Pop the element stack back off since it'll never be used now
        fElemStack.popTop();

        // If we have a doc handler, tell it about the end tag
        if (fDocHandler)
        {
            fDocHandler->endElement
            (
                *elemDecl
                , uriId
                , isRoot
                , elemDecl->getElementName()->getPrefix()
            );
        }

        // If the elem stack is empty, then it was an empty root
        if (isRoot)
            gotData = false;
    }

    return true;
}

unsigned int
WFXMLScanner::resolveQName(const   XMLCh* const qName
                           ,       XMLBuffer&   prefixBuf
                           , const short        mode
                           ,       int&         prefixColonPos)
{
    //  Lets split out the qName into a URI and name buffer first. The URI
    //  can be empty.
    prefixColonPos = XMLString::indexOf(qName, chColon);
    if (prefixColonPos == -1)
    {
        //  Its all name with no prefix, so put the whole thing into the name
        //  buffer. Then map the empty string to a URI, since the empty string
        //  represents the default namespace. This will either return some
        //  explicit URI which the default namespace is mapped to, or the
        //  the default global namespace.
        bool unknown = false;

        prefixBuf.reset();
        return fElemStack.mapPrefixToURI(XMLUni::fgZeroLenString, (ElemStack::MapModes) mode, unknown);
    }
    else
    {
        //  Copy the chars up to but not including the colon into the prefix
        //  buffer.
        prefixBuf.set(qName, prefixColonPos);

        //  Watch for the special namespace prefixes. We always map these to
        //  special URIs. 'xml' gets mapped to the official URI that its defined
        //  to map to by the NS spec. xmlns gets mapped to a special place holder
        //  URI that we define (so that it maps to something checkable.)
        const XMLCh* prefixRawBuf = prefixBuf.getRawBuffer();
        if (XMLString::equals(prefixRawBuf, XMLUni::fgXMLNSString)) {

            // if this is an element, it is an error to have xmlns as prefix
            if (mode == ElemStack::Mode_Element)
                emitError(XMLErrs::NoXMLNSAsElementPrefix, qName);

            return fXMLNSNamespaceId;
        }
        else if (XMLString::equals(prefixRawBuf, XMLUni::fgXMLString)) {
            return  fXMLNamespaceId;
        }
        else
        {
            bool unknown = false;
            unsigned int uriId = fElemStack.mapPrefixToURI(prefixRawBuf, (ElemStack::MapModes) mode, unknown);

            if (unknown)
                emitError(XMLErrs::UnknownPrefix, prefixRawBuf);

            return uriId;
        }
    }
}

// ---------------------------------------------------------------------------
//  XMLScanner: Private parsing methods
// ---------------------------------------------------------------------------
bool WFXMLScanner::scanAttValue(const XMLCh* const attrName
                              ,     XMLBuffer&   toFill)
{
    // Reset the target buffer
    toFill.reset();

    // Get the next char which must be a single or double quote
    XMLCh quoteCh;
    if (!fReaderMgr.skipIfQuote(quoteCh))
        return false;

    //  We have to get the current reader because we have to ignore closing
    //  quotes until we hit the same reader again.
    const unsigned int curReader = fReaderMgr.getCurrentReaderNum();

    //  Loop until we get the attribute value. Note that we use a double
    //  loop here to avoid the setup/teardown overhead of the exception
    //  handler on every round.
    XMLCh   nextCh;
    XMLCh   secondCh = 0;
    bool    gotLeadingSurrogate = false;
    bool    escaped;
    while (true)
    {
    try
    {
        while(true)
        {
            nextCh = fReaderMgr.getNextChar();

            if (!nextCh)
                ThrowXMLwithMemMgr(UnexpectedEOFException, XMLExcepts::Gen_UnexpectedEOF, fMemoryManager);

            // Check for our ending quote in the same entity
            if (nextCh == quoteCh)
            {
                if (curReader == fReaderMgr.getCurrentReaderNum())
                    return true;

                // Watch for spillover into a previous entity
                if (curReader > fReaderMgr.getCurrentReaderNum())
                {
                    emitError(XMLErrs::PartialMarkupInEntity);
                    return false;
                }
            }

            //  Check for an entity ref now, before we let it affect our
            //  whitespace normalization logic below. We ignore the empty flag
            //  in this one.
            escaped = false;
            if (nextCh == chAmpersand)
            {
                if (scanEntityRef(true, nextCh, secondCh, escaped) != EntityExp_Returned)
                {
                    gotLeadingSurrogate = false;
                    continue;
                }
            }
            else if ((nextCh >= 0xD800) && (nextCh <= 0xDBFF))
            {
                // Deal with surrogate pairs
                //  Its a leading surrogate. If we already got one, then
                //  issue an error, else set leading flag to make sure that
                //  we look for a trailing next time.
                if (gotLeadingSurrogate)
                {
                    emitError(XMLErrs::Expected2ndSurrogateChar);
                }
                else
                    gotLeadingSurrogate = true;
            }
            else
            {
                //  If its a trailing surrogate, make sure that we are
                //  prepared for that. Else, its just a regular char so make
                //  sure that we were not expected a trailing surrogate.
                if ((nextCh >= 0xDC00) && (nextCh <= 0xDFFF))
                {
                    // Its trailing, so make sure we were expecting it
                    if (!gotLeadingSurrogate)
                        emitError(XMLErrs::Unexpected2ndSurrogateChar);
                }
                else
                {
                    //  Its just a char, so make sure we were not expecting a
                    //  trailing surrogate.
                    if (gotLeadingSurrogate) {
                        emitError(XMLErrs::Expected2ndSurrogateChar);
                    }
                    // Its got to at least be a valid XML character
                    else if (!fReaderMgr.getCurrentReader()->isXMLChar(nextCh))
                    {
                        XMLCh tmpBuf[9];
                        XMLString::binToText
                        (
                            nextCh
                            , tmpBuf
                            , 8
                            , 16
                            , fMemoryManager
                        );
                        emitError(XMLErrs::InvalidCharacterInAttrValue, attrName, tmpBuf);
                    }
                }
                gotLeadingSurrogate = false;
            }

            //  If its not escaped, then make sure its not a < character, which
            //  is not allowed in attribute values.
            if (!escaped) {
                if (nextCh == chOpenAngle)
                    emitError(XMLErrs::BracketInAttrValue, attrName);
                else if (fReaderMgr.getCurrentReader()->isWhitespace(nextCh))
                    nextCh = chSpace;
            }

            // Else add it to the buffer
            toFill.append(nextCh);

            if (secondCh)
            {
                toFill.append(secondCh);
                secondCh=0;
            }
        }
    }
    catch(const EndOfEntityException&)
    {
        // Just eat it and continue.
        gotLeadingSurrogate = false;
        escaped = false;
    }
    }
    return true;
}


//  This method scans a CDATA section. It collects the character into one
//  of the temp buffers and calls the document handler, if any, with the
//  characters. It assumes that the <![CDATA string has been scanned before
//  this call.
void WFXMLScanner::scanCDSection()
{
    static const XMLCh CDataClose[] =
    {
            chCloseSquare, chCloseAngle, chNull
    };

    //  The next character should be the opening square bracket. If not
    //  issue an error, but then try to recover by skipping any whitespace
    //  and checking again.
    if (!fReaderMgr.skippedChar(chOpenSquare))
    {
        emitError(XMLErrs::ExpectedOpenSquareBracket);
        fReaderMgr.skipPastSpaces();

        // If we still don't find it, then give up, else keep going
        if (!fReaderMgr.skippedChar(chOpenSquare))
            return;
    }

    // Get a buffer for this
    XMLBufBid bbCData(&fBufMgr);

    //  We just scan forward until we hit the end of CDATA section sequence.
    //  CDATA is effectively a big escape mechanism so we don't treat markup
    //  characters specially here.
    bool            emittedError = false;
    bool    gotLeadingSurrogate = false;
    while (true)
    {
        const XMLCh nextCh = fReaderMgr.getNextChar();

        // Watch for unexpected end of file
        if (!nextCh)
        {
            emitError(XMLErrs::UnterminatedCDATASection);
            ThrowXMLwithMemMgr(UnexpectedEOFException, XMLExcepts::Gen_UnexpectedEOF, fMemoryManager);
        }

        //  If this is a close square bracket it could be our closing
        //  sequence.
        if (nextCh == chCloseSquare && fReaderMgr.skippedString(CDataClose))
        {
            //  make sure we were not expecting a trailing surrogate.
            if (gotLeadingSurrogate)
                emitError(XMLErrs::Expected2ndSurrogateChar);

            // If we have a doc handler, call it
            if (fDocHandler)
            {
                fDocHandler->docCharacters
                (
                    bbCData.getRawBuffer()
                    , bbCData.getLen()
                    , true
                );
            }

            // And we are done
            break;
        }

        //  Make sure its a valid character. But if we've emitted an error
        //  already, don't bother with the overhead since we've already told
        //  them about it.
        if (!emittedError)
        {
            // Deal with surrogate pairs
            if ((nextCh >= 0xD800) && (nextCh <= 0xDBFF))
            {
                //  Its a leading surrogate. If we already got one, then
                //  issue an error, else set leading flag to make sure that
                //  we look for a trailing next time.
                if (gotLeadingSurrogate)
                    emitError(XMLErrs::Expected2ndSurrogateChar);
                else
                    gotLeadingSurrogate = true;
            }
            else
            {
                //  If its a trailing surrogate, make sure that we are
                //  prepared for that. Else, its just a regular char so make
                //  sure that we were not expected a trailing surrogate.
                if ((nextCh >= 0xDC00) && (nextCh <= 0xDFFF))
                {
                    // Its trailing, so make sure we were expecting it
                    if (!gotLeadingSurrogate)
                        emitError(XMLErrs::Unexpected2ndSurrogateChar);
                }
                else
                {
                    //  Its just a char, so make sure we were not expecting a
                    //  trailing surrogate.
                    if (gotLeadingSurrogate)
                        emitError(XMLErrs::Expected2ndSurrogateChar);

                    // Its got to at least be a valid XML character
                    else if (!fReaderMgr.getCurrentReader()->isXMLChar(nextCh))
                    {
                        XMLCh tmpBuf[9];
                        XMLString::binToText
                        (
                            nextCh
                            , tmpBuf
                            , 8
                            , 16
                            , fMemoryManager
                        );
                        emitError(XMLErrs::InvalidCharacter, tmpBuf);
                        emittedError = true;
                    }
                }
                gotLeadingSurrogate = false;
            }
        }

        // Add it to the buffer
        bbCData.append(nextCh);
    }
}


void WFXMLScanner::scanCharData(XML