void IGXMLScanner::scanCharData()

in src/xercesc/internal/IGXMLScanner2.cpp [2842:3068]


void IGXMLScanner::scanCharData(XMLBuffer& toUse)
{
    //  We have to watch for the stupid ]]> sequence, which is illegal in
    //  character data. So this is a little state machine that handles that.
    enum States
    {
        State_Waiting
        , State_GotOne
        , State_GotTwo
    };

    // Reset the buffer before we start
    toUse.reset();

    // Turn on the 'throw at end' flag of the reader manager
    ThrowEOEJanitor jan(&fReaderMgr, true);

    //  In order to be more efficient we have to use kind of a deeply nested
    //  set of blocks here. The outer block puts on a try and catches end of
    //  entity exceptions. The inner loop is the per-character loop. If we
    //  put the try inside the inner loop, it would work but would require
    //  the exception handling code setup/teardown code to be invoked for
    //  each character.
    XMLCh   nextCh;
    XMLCh   secondCh = 0;
    States  curState = State_Waiting;
    bool    escaped = false;
    bool    gotLeadingSurrogate = false;
    bool    notDone = true;
    while (notDone)
    {
        try
        {
            while (true)
            {
                //  Eat through as many plain content characters as possible without
                //  needing special handling.  Moving most content characters here,
                //  in this one call, rather than running the overall loop once
                //  per content character, is a speed optimization.
                if (curState == State_Waiting  &&  !gotLeadingSurrogate)
                {
                     fReaderMgr.movePlainContentChars(toUse);
                }

                // Try to get another char from the source
                //   The code from here on down covers all contengencies,
                if (!fReaderMgr.getNextCharIfNot(chOpenAngle, nextCh))
                {
                    // If we were waiting for a trailing surrogate, its an error
                    if (gotLeadingSurrogate)
                        emitError(XMLErrs::Expected2ndSurrogateChar);

                    notDone = false;
                    break;
                }

                //  Watch for a reference. Note that the escapement mechanism
                //  is ignored in this content.
                escaped = false;
                if (nextCh == chAmpersand)
                {
                    sendCharData(toUse);

                    // Turn off the throwing at the end of entity during this
                    ThrowEOEJanitor jan(&fReaderMgr, false);

                    if (scanEntityRef(false, nextCh, secondCh, escaped) != EntityExp_Returned)
                    {
                        gotLeadingSurrogate = false;
                        continue;
                    }
                    else
                    {
                        if (escaped && !fElemStack.isEmpty())
                            fElemStack.setReferenceEscaped();
                    }
                }
                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);

                        // Make sure the returned char is a valid XML char
                        if (!fReaderMgr.getCurrentReader()->isXMLChar(nextCh))
                        {
                            XMLCh tmpBuf[9];
                            XMLString::binToText
                            (
                                nextCh
                                , tmpBuf
                                , 8
                                , 16
                                , fMemoryManager
                            );
                            emitError(XMLErrs::InvalidCharacter, tmpBuf);
                        }
                    }
                    gotLeadingSurrogate = false;
                }

                 // Keep the state machine up to date
                if (!escaped)
                {
                    if (nextCh == chCloseSquare)
                    {
                        if (curState == State_Waiting)
                            curState = State_GotOne;
                        else if (curState == State_GotOne)
                            curState = State_GotTwo;
                    }
                    else if (nextCh == chCloseAngle)
                    {
                        if (curState == State_GotTwo)
                            emitError(XMLErrs::BadSequenceInCharData);
                        curState = State_Waiting;
                    }
                    else
                    {
                        curState = State_Waiting;
                    }
                }
                else
                {
                    curState = State_Waiting;
                }

                // Add this char to the buffer
                toUse.append(nextCh);

                if (secondCh)
                {
                    toUse.append(secondCh);
                    secondCh=0;
                }
            }
        }
        catch(const EndOfEntityException& toCatch)
        {
            //  Some entity ended, so we have to send any accumulated
            //  chars and send an end of entity event.
            sendCharData(toUse);
            gotLeadingSurrogate = false;

            if (fDocHandler)
                fDocHandler->endEntityReference(toCatch.getEntity());
        }
    }

    // Check the validity constraints as per XML 1.0 Section 2.9
    if (fValidate && fStandalone)
    {
        // See if the text contains whitespace
        // Get the raw data we need for the callback
        const XMLCh* rawBuf = toUse.getRawBuffer();
        const XMLSize_t len = toUse.getLen();
        const bool isSpaces = fReaderMgr.getCurrentReader()->containsWhiteSpace(rawBuf, len);

        if (isSpaces)
        {
            // And see if the current element is a 'Children' style content model
            const ElemStack::StackElem* topElem = fElemStack.topElement();

            if (topElem->fThisElement->isExternal()) {

                // Get the character data opts for the current element
                XMLElementDecl::CharDataOpts charOpts = XMLElementDecl::AllCharData;
                if(fGrammar->getGrammarType() == Grammar::SchemaGrammarType)
                {
                    // And see if the current element is a 'Children' style content model
                    ComplexTypeInfo *currType = ((SchemaValidator*)fValidator)->getCurrentTypeInfo();
                    if(currType)
                    {
                        SchemaElementDecl::ModelTypes modelType = (SchemaElementDecl::ModelTypes) currType->getContentType();
                        if(modelType == SchemaElementDecl::Children ||
                           modelType == SchemaElementDecl::ElementOnlyEmpty)
                            charOpts = XMLElementDecl::SpacesOk;
                        else if(modelType == SchemaElementDecl::Empty)
                            charOpts = XMLElementDecl::NoCharData;
                    }
                } else // DTD grammar
                    charOpts = topElem->fThisElement->getCharDataOpts();

                if (charOpts == XMLElementDecl::SpacesOk)  // => Element Content
                {
                    // Error - standalone should have a value of "no" as whitespace detected in an
                    // element type with element content whose element declaration was external
                    //
                    fValidator->emitError(XMLValid::NoWSForStandalone);
                    if(fGrammarType == Grammar::SchemaGrammarType)
                    {
                        if (getPSVIHandler())
                        {
                            // REVISIT:
                            // PSVIElement->setValidity(PSVIItem::VALIDITY_INVALID);
                        }
                    }
                }
            }
        }
    }
    // Send any char data that we accumulated into the buffer
    sendCharData(toUse);
}