IGXMLScanner2.cpp

    return 0;
}



// ---------------------------------------------------------------------------
//  IGXMLScanner: Private parsing methods
// ---------------------------------------------------------------------------

//  This method is called to do a raw scan of an attribute value. It does not
//  do normalization (since we don't know their types yet.) It just scans the
//  value and does entity expansion.
//
//  End of entity's must be dealt with here. During DTD scan, they can come
//  from external entities. During content, they can come from any entity.
//  We just eat the end of entity and continue with our scan until we come
//  to the closing quote. If an unterminated value causes us to go through
//  subsequent entities, that will cause errors back in the calling code,
//  but there's little we can do about it here.
bool IGXMLScanner::basicAttrValueScan(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. It has to be in the same entity
                //  as where we started. Quotes in nested entities are ignored.
                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 . We ignore the empty flag in
                //  this one.
                escaped = false;
                if (nextCh == chAmpersand)
                {
                    // If it was not returned directly, then jump back up
                    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 it was escaped, then put in a 0xFFFF value. This will
                //  be used later during validation and normalization of the
                //  value to know that the following character was via an
                //  escape char.
                if (escaped)
                    toFill.append(0xFFFF);

                // 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;
}


bool IGXMLScanner::scanAttValue(  const   XMLAttDef* const    attDef
                                  , const XMLCh* const        attrName
                                  ,       XMLBuffer&          toFill)
{
    enum States
    {
        InWhitespace
        , InContent
    };

    // Get the type and name
    const XMLAttDef::AttTypes type = (attDef)
                ?attDef->getType()
                :XMLAttDef::CData;

    // 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();

    // Get attribute def - to check to see if it's declared externally or not
    bool  isAttExternal = (attDef)
                ?attDef->isExternal()
                :false;

    //  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;
    States  curState = InContent;
    bool    firstNonWS = false;
    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
                    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 && (nextCh == chOpenAngle))
                emitError(XMLErrs::BracketInAttrValue, attrName);

            //  If the attribute is a CDATA type we do simple replacement of
            //  tabs and new lines with spaces, if the character is not escaped
            //  by way of a char ref.
            //
            //  Otherwise, we do the standard non-CDATA normalization of
            //  compressing whitespace to single spaces and getting rid of leading
            //  and trailing whitespace.
            if (type == XMLAttDef::CData)
            {
                if (!escaped)
                {
                    if ((nextCh == 0x09) || (nextCh == 0x0A) || (nextCh == 0x0D))
                    {
                        // Check Validity Constraint for Standalone document declaration
                        // XML 1.0, Section 2.9
                        if (fStandalone && fValidate && isAttExternal)
                        {
                            // Can't have a standalone document declaration of "yes" if  attribute
                            // values are subject to normalisation
                            fValidator->emitError(XMLValid::NoAttNormForStandalone, attrName);
                        }
                        nextCh = chSpace;
                    }
                }
            }
            else
            {
                if (curState == InWhitespace)
                {
                    if ((escaped && nextCh != chSpace) || !fReaderMgr.getCurrentReader()->isWhitespace(nextCh))
                    {
                        if (firstNonWS)
                            toFill.append(chSpace);
                        curState = InContent;
                        firstNonWS = true;
                    }
                    else
                    {
                        continue;
                    }
                }
                else if (curState == InContent)
                {
                    if ((nextCh == chSpace) ||
                        (fReaderMgr.getCurrentReader()->isWhitespace(nextCh) && !escaped))
                    {
                        curState = InWhitespace;

                        // Check Validity Constraint for Standalone document declaration
                        // XML 1.0, Section 2.9
                        if (fStandalone && fValidate && isAttExternal)
                        {
                            if (!firstNonWS || (nextCh != chSpace) || (fReaderMgr.lookingAtSpace()))
                            {
                                 // Can't have a standalone document declaration of "yes" if  attribute
                                 // values are subject to normalisation
                                 fValidator->emitError(XMLValid::NoAttNormForStandalone, attrName);
                            }
                        }
                        continue;
                    }
                    firstNonWS = true;
                }
            }

            // 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 IGXMLScanner::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;
    const ElemStack::StackElem* topElem = fElemStack.topElement();

    // 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) 
                charOpts = XMLElementDecl::SpacesOk;
            else if(modelType == SchemaElementDecl::Empty) 
                charOpts = XMLElementDecl::NoCharData; 
        } 
    } else // DTD grammar
        charOpts = topElem->fThisElement->getCharDataOpts();

    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 (fValidate && fStandalone && (fReaderMgr.getCurrentReader()->isWhitespace(nextCh)))
        {
            // This document is standalone; this ignorable CDATA whitespace is forbidden.
            // XML 1.0, Section 2.9
            // And see if the current element is a 'Children' style content model
            if (topElem->fThisElement->isExternal()) {

                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);
                        }
                    }
                }
            }
        }

        //  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 (fGrammarType == Grammar::SchemaGrammarType) {

                unsigned int xsLen = bbCData.getLen();
                const XMLCh* xsNormalized = bbCData.getRawBuffer();
                DatatypeValidator* tempDV = ((SchemaValidator*) fValidator)->getCurrentDatatypeValidator();
                if (tempDV && tempDV->getWSFacet() != DatatypeValidator::PRESERVE)
                {
                    // normalize the character according to schema whitespace facet
                    ((SchemaValidator*) fValidator)->normalizeWhiteSpace(tempDV, xsNormalized, fWSNormalizeBuf);
                    xsNormalized = fWSNormalizeBuf.getRawBuffer();
                    xsLen = fWSNormalizeBuf.getLen();
                    if (fNormalizeData && fValidate) {
                        bbCData.set(xsNormalized);
                    }
                }

                if (fValidate) {

                    // tell the schema validation about the character data for checkContent later
                    ((SchemaValidator*)fValidator)->setDatatypeBuffer(xsNormalized);

                    if (charOpts != XMLElementDecl::AllCharData)
                    {
                        // They definitely cannot handle any type of char data
                        fValidator->emitError(XMLValid::NoCharDataInCM);
                        if (getPSVIHandler())
                        {
                            // REVISIT:                                   
                            // PSVIElement->setValidity(PSVIItem::VALIDITY_INVALID);
                        }
                    }
                }

                // call all active identity constraints
                if (toCheckIdentityConstraint() && fICHandler->getMatcherCount()) {
                    fContent.append(xsNormalized, xsLen);
                }
            }
            else {
                if (fValidate) {

                    if (charOpts != XMLElementDecl::AllCharData)
                    {
                        // They definitely cannot handle any type of char data
                        fValidator->emitError(XMLValid::NoCharDataInCM);
                    }
                }
            }

            // 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 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 unsigned int 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) 
                            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);
}


//  This method will scan a general/character entity ref. It will either
//  expand a char ref and return it directly, or push a reader for a general
//  entity.
//
//  The return value indicates whether the char parameters hold the value
//  or whether the value was pushed as a reader, or that it failed.
//
//  The escaped flag tells the caller whether the returned parameter resulted
//  from a character reference, which escapes the character in some cases. It
//  only makes any difference if the return value indicates the value was
//  returned directly.
IGXMLScanner::EntityExpRes
IGXMLScanner::scanEntityRef(  const   bool    inAttVal
                            ,       XMLCh&  firstCh
                            ,       XMLCh&  secondCh
                            ,       bool&   escaped)
{
    // Assume no escape
    secondCh = 0;
    escaped = false;

    // We have to insure that its all in one entity
    const unsigned int curReader = fReaderMgr.getCurrentReaderNum();

    //  If the next char is a pound, then its a character reference and we
    //  need to expand it always.
    if (fReaderMgr.skippedChar(chPound))
    {
        //  Its a character reference, so scan it and get back the numeric
        //  value it represents.
        if (!scanCharRef(firstCh, secondCh))
            return EntityExp_Failed;

        escaped = true;

        if (curReader != fReaderMgr.getCurrentReaderNum())
            emitError(XMLErrs::PartialMarkupInEntity);

        return EntityExp_Returned;
    }

    // Expand it since its a normal entity ref
    XMLBufBid bbName(&fBufMgr);
    if (!fReaderMgr.getName(bbName.getBuffer()))
    {
        emitError(XMLErrs::ExpectedEntityRefName);
        return EntityExp_Failed;
    }

    //  Next char must be a semi-colon. But if its not, just emit
    //  an error and try to continue.
    if (!fReaderMgr.skippedChar(chSemiColon))
        emitError(XMLErrs::UnterminatedEntityRef, bbName.getRawBuffer());

    // Make sure we ended up on the same entity reader as the & char
    if (curReader != fReaderMgr.getCurrentReaderNum())
        emitError(XMLErrs::PartialMarkupInEntity);

    // Look up the name in the general entity pool
    XMLEntityDecl* decl = fDTDGrammar->getEntityDecl(bbName.getRawBuffer());

    // If it does not exist, then obviously an error
    if (!decl)
    {
        // XML 1.0 Section 4.1
        // Well-formedness Constraint for entity not found:
        //   In a document without any DTD, a document with only an internal DTD subset which contains no parameter entity references,
        //      or a document with "standalone='yes'", for an entity reference that does not occur within the external subset
        //      or a parameter entity
        //
        // Else it's Validity Constraint
        if (fStandalone || fHasNoDTD)
            emitError(XMLErrs::EntityNotFound, bbName.getRawBuffer());
        else {
            if (fValidate)
                fValidator->emitError(XMLValid::VC_EntityNotFound, bbName.getRawBuffer());
        }

        return EntityExp_Failed;
    }

    // XML 1.0 Section 4.1
    //  If we are a standalone document, then it has to have been declared
    //  in the internal subset.
    if (fStandalone && !decl->getDeclaredInIntSubset())
        emitError(XMLErrs::IllegalRefInStandalone, bbName.getRawBuffer());

    if (decl->isExternal())
    {
        // If its unparsed, then its not valid here
        if (decl->isUnparsed())
        {
            emitError(XMLErrs::NoUnparsedEntityRefs, bbName.getRawBuffer());
            return EntityExp_Failed;
        }

        // If we are in an attribute value, then not valid but keep going
        if (inAttVal)
            emitError(XMLErrs::NoExtRefsInAttValue);

        // And now create a reader to read this entity
        InputSource* srcUsed;
        XMLReader* reader = fReaderMgr.createReader
        (
            decl->getBaseURI()
            , decl->getSystemId()
            , decl->getPublicId()
            , false
            , XMLReader::RefFrom_NonLiteral
            , XMLReader::Type_General
            , XMLReader::Source_External
            , srcUsed
            , fCalculateSrcOfs
        );

        // Put a janitor on the source so it gets cleaned up on exit
        Janitor<InputSource> janSrc(srcUsed);

        //  If the creation failed, and its not because the source was empty,
        //  then emit an error and return.
        if (!reader)
            ThrowXMLwithMemMgr1(RuntimeException, XMLExcepts::Gen_CouldNotOpenExtEntity, srcUsed->getSystemId(), fMemoryManager);

        //  Push the reader. If its a recursive expansion, then emit an error
        //  and return an failure.
        if (!fReaderMgr.pushReader(reader, decl))
        {
            emitError(XMLErrs::RecursiveEntity, decl->getName());
            return EntityExp_Failed;
        }

        // here's where we need to check if there's a SecurityManager,
        // how many entity references we've had
        if(fSecurityManager != 0 && ++fEntityExpansionCount > fEntityExpansionLimit) {
            XMLCh expLimStr[16];
            XMLString::binToText(fEntityExpansionLimit, expLimStr, 15, 10, fMemoryManager);
            emitError
            ( 
                XMLErrs::EntityExpansionLimitExceeded
                , expLimStr
            );
            // there seems nothing  better to be done than to reset the entity expansion counter
            fEntityExpansionCount = 0;
        }

        //  Do a start entity reference event.
        //
        //  <TBD> For now, we supress them in att values. Later, when
        //  the stuff is in place to correctly allow DOM to handle them
        //  we'll turn this back on.
        if (fDocHandler && !inAttVal)
            fDocHandler->startEntityReference(*decl);

        // If it starts with the XML string, then parse a text decl
        if (checkXMLDecl(true))
            scanXMLDecl(Decl_Text);
    }
    else
    {
        //  If its one of the special char references, then we can return
        //  it as a character, and its considered escaped.
        if (decl->getIsSpecialChar())
        {
            firstCh = decl->getValue()[0];
            escaped = true;
            return EntityExp_Returned;
        }

        //  Create a reader over a memory stream over the entity value
        //  We force it to assume UTF-16 by passing in an encoding
        //  string. This way it won't both trying to predecode the
        //  first line, looking for an XML/TextDecl.
        XMLReader* valueReader = fReaderMgr.createIntEntReader
        (
            decl->getName()
            , XMLReader::RefFrom_NonLiteral
            , XMLReader::Type_General
            , decl->getValue()
            , decl->getValueLen()
            , false
        );

        //  Try to push the entity reader onto the reader manager stack,
        //  where it will become the subsequent input. If it fails, that
        //  means the entity is recursive, so issue an error. The reader