dgxmlscanner.cpp

来自「IBM的解析xml的工具Xerces的源代码」· C++ 代码 · 共 1,902 行 · 第 1/5 页

//  This method handles the high level logic of scanning the DOCType
//  declaration. This calls the DTDScanner and kicks off both the scanning of
//  the internal subset and the scanning of the external subset, if any.
//
//  When we get here the '<!DOCTYPE' part has already been scanned, which is
//  what told us that we had a doc type decl to parse.
void DGXMLScanner::scanDocTypeDecl()
{
    if (fDocTypeHandler)
        fDocTypeHandler->resetDocType();

    // There must be some space after DOCTYPE
    if (!fReaderMgr.skipPastSpaces())
    {
        emitError(XMLErrs::ExpectedWhitespace);

        // Just skip the Doctype declaration and return
        fReaderMgr.skipPastChar(chCloseAngle);
        return;
    }

    // Get a buffer for the root element
    XMLBufBid bbRootName(&fBufMgr);

    //  Get a name from the input, which should be the name of the root
    //  element of the upcoming content.
    fReaderMgr.getName(bbRootName.getBuffer());
    if (bbRootName.isEmpty())
    {
        emitError(XMLErrs::NoRootElemInDOCTYPE);
        fReaderMgr.skipPastChar(chCloseAngle);
        return;
    }

    //  Store the root element name for later check
    setRootElemName(bbRootName.getRawBuffer());

    //  This element obviously is not going to exist in the element decl
    //  pool yet, but we need to call docTypeDecl. So force it into
    //  the element decl pool, marked as being there because it was in
    //  the DOCTYPE. Later, when its declared, the status will be updated.
    //
    //  Only do this if we are not reusing the validator! If we are reusing,
    //  then look it up instead. It has to exist!
    DTDElementDecl* rootDecl = new (fGrammarPoolMemoryManager) DTDElementDecl
    (
        bbRootName.getRawBuffer()
        , fEmptyNamespaceId
        , DTDElementDecl::Any
        , fGrammarPoolMemoryManager
    ); 

    Janitor<DTDElementDecl> rootDeclJanitor(rootDecl);    
    rootDecl->setCreateReason(DTDElementDecl::AsRootElem);
    rootDecl->setExternalElemDeclaration(true);
    if(!fUseCachedGrammar) 
    {
        // this will break getRootElemId on DTDGrammar when
        // cached grammars are in use, but 
        // why would one use this anyway???
        ((DTDGrammar*)fGrammar)->setRootElemId(fGrammar->putElemDecl(rootDecl));
        rootDeclJanitor.release();
    } else 
    {
        // put this in the undeclared pool so it gets deleted...
        XMLElementDecl* elemDecl = fDTDElemNonDeclPool->getByKey(bbRootName.getRawBuffer());
        if (elemDecl)
        {
            rootDecl->setId(elemDecl->getId());
        }
        else
        {
            rootDecl->setId(fDTDElemNonDeclPool->put((DTDElementDecl*)rootDecl));
            rootDeclJanitor.release();
        }
    }

    // Skip any spaces after the name
    fReaderMgr.skipPastSpaces();

    //  And now if we are looking at a >, then we are done. It is not
    //  required to have an internal or external subset, though why you
    //  would not escapes me.
    if (fReaderMgr.skippedChar(chCloseAngle)) {

        //  If we have a doc type handler and advanced callbacks are enabled,
        //  call the doctype event.
        if (fDocTypeHandler)
            fDocTypeHandler->doctypeDecl(*rootDecl, 0, 0, false);
        return;
    }

    // either internal/external subset
    if (fValScheme == Val_Auto && !fValidate)
        fValidate = true;

    bool    hasIntSubset = false;
    bool    hasExtSubset = false;
    XMLCh*  sysId = 0;
    XMLCh*  pubId = 0;

    DTDScanner dtdScanner
    (
        (DTDGrammar*) fGrammar
        , fDocTypeHandler
        , fGrammarPoolMemoryManager
        , fMemoryManager
    );
    dtdScanner.setScannerInfo(this, &fReaderMgr, &fBufMgr);

    //  If the next character is '[' then we have no external subset cause
    //  there is no system id, just the opening character of the internal
    //  subset. Else, has to be an id.
    //
    // Just look at the next char, don't eat it.
    if (fReaderMgr.peekNextChar() == chOpenSquare)
    {
        hasIntSubset = true;
    }
    else
    {
        // Indicate we have an external subset
        hasExtSubset = true;
        fHasNoDTD = false;

        // Get buffers for the ids
        XMLBufBid bbPubId(&fBufMgr);
        XMLBufBid bbSysId(&fBufMgr);

        // Get the external subset id
        if (!dtdScanner.scanId(bbPubId.getBuffer(), bbSysId.getBuffer(), DTDScanner::IDType_External))
        {
            fReaderMgr.skipPastChar(chCloseAngle);
            return;
        }

        // Get copies of the ids we got
        pubId = XMLString::replicate(bbPubId.getRawBuffer(), fMemoryManager);
        sysId = XMLString::replicate(bbSysId.getRawBuffer(), fMemoryManager);

        // Skip spaces and check again for the opening of an internal subset
        fReaderMgr.skipPastSpaces();

        // Just look at the next char, don't eat it.
        if (fReaderMgr.peekNextChar() == chOpenSquare) {
            hasIntSubset = true;
        }
    }

    // Insure that the ids get cleaned up, if they got allocated
    ArrayJanitor<XMLCh> janSysId(sysId, fMemoryManager);
    ArrayJanitor<XMLCh> janPubId(pubId, fMemoryManager);

    //  If we have a doc type handler and advanced callbacks are enabled,
    //  call the doctype event.
    if (fDocTypeHandler)
        fDocTypeHandler->doctypeDecl(*rootDecl, pubId, sysId, hasIntSubset, hasExtSubset);

    //  Ok, if we had an internal subset, we are just past the [ character
    //  and need to parse that first.
    if (hasIntSubset)
    {
        // Eat the opening square bracket
        fReaderMgr.getNextChar();

        checkInternalDTD(hasExtSubset, sysId);

        //  And try to scan the internal subset. If we fail, try to recover
        //  by skipping forward tot he close angle and returning.
        if (!dtdScanner.scanInternalSubset())
        {
            fReaderMgr.skipPastChar(chCloseAngle);
            return;
        }

        //  Do a sanity check that some expanded PE did not propogate out of
        //  the doctype. This could happen if it was terminated early by bad
        //  syntax.
        if (fReaderMgr.getReaderDepth() > 1)
        {
            emitError(XMLErrs::PEPropogated);

            // Ask the reader manager to pop back down to the main level
            fReaderMgr.cleanStackBackTo(1);
        }

        fReaderMgr.skipPastSpaces();
    }

    // And that should leave us at the closing > of the DOCTYPE line
    if (!fReaderMgr.skippedChar(chCloseAngle))
    {
        //  Do a special check for the common scenario of an extra ] char at
        //  the end. This is easy to recover from.
        if (fReaderMgr.skippedChar(chCloseSquare)
        &&  fReaderMgr.skippedChar(chCloseAngle))
        {
            emitError(XMLErrs::ExtraCloseSquare);
        }
         else
        {
            emitError(XMLErrs::UnterminatedDOCTYPE);
            fReaderMgr.skipPastChar(chCloseAngle);
        }
    }

    //  If we had an external subset, then we need to deal with that one
    //  next. If we are reusing the validator, then don't scan it.
    if (hasExtSubset) {

        if (fUseCachedGrammar)
        {
            InputSource* sysIdSrc = resolveSystemId(sysId);
            Janitor<InputSource> janSysIdSrc(sysIdSrc);
            Grammar* grammar = fGrammarResolver->getGrammar(sysIdSrc->getSystemId());

            if (grammar && grammar->getGrammarType() == Grammar::DTDGrammarType) {

                fDTDGrammar = (DTDGrammar*) grammar;
                fGrammar = fDTDGrammar;
                fValidator->setGrammar(fGrammar);
                // we *cannot* identify the root element on 
                // cached grammars; else we risk breaking multithreaded
                // applications.  - NG
                /*******
                rootDecl = (DTDElementDecl*) fGrammar->getElemDecl(fEmptyNamespaceId, 0, bbRootName.getRawBuffer(), Grammar::TOP_LEVEL_SCOPE);

                if (rootDecl)
                    ((DTDGrammar*)fGrammar)->setRootElemId(rootDecl->getId());
                else {
                    rootDecl = new (fGrammarPoolMemoryManager) DTDElementDecl
                    (
                        bbRootName.getRawBuffer()
                        , fEmptyNamespaceId
                        , DTDElementDecl::Any
                        , fGrammarPoolMemoryManager
                    );
                    rootDecl->setCreateReason(DTDElementDecl::AsRootElem);
                    rootDecl->setExternalElemDeclaration(true);
                    ((DTDGrammar*)fGrammar)->setRootElemId(fGrammar->putElemDecl(rootDecl));
                }
                *********/

                return;
            }
        }

        if (fLoadExternalDTD || fValidate)
        {
            // And now create a reader to read this entity
            InputSource* srcUsed;
            XMLReader* reader = fReaderMgr.createReader
            (
                sysId
                , pubId
                , false
                , XMLReader::RefFrom_NonLiteral
                , XMLReader::Type_General
                , XMLReader::Source_External
                , srcUsed
                , fCalculateSrcOfs
            );

            // Put a janitor on the input source
            Janitor<InputSource> janSrc(srcUsed);

            //  If it failed then throw an exception
            if (!reader)
                ThrowXMLwithMemMgr1(RuntimeException, XMLExcepts::Gen_CouldNotOpenDTD, srcUsed->getSystemId(), fMemoryManager);

            if (fToCacheGrammar) {

                unsigned int stringId = fGrammarResolver->getStringPool()->addOrFind(srcUsed->getSystemId());
                const XMLCh* sysIdStr = fGrammarResolver->getStringPool()->getValueForId(stringId);

                fGrammarResolver->orphanGrammar(XMLUni::fgDTDEntityString);
                ((XMLDTDDescription*) (fGrammar->getGrammarDescription()))->setSystemId(sysIdStr);
                fGrammarResolver->putGrammar(fGrammar);
            }

            //  In order to make the processing work consistently, we have to
            //  make this look like an external entity. So create an entity
            //  decl and fill it in and push it with the reader, as happens
            //  with an external entity. Put a janitor on it to insure it gets
            //  cleaned up. The reader manager does not adopt them.
            const XMLCh gDTDStr[] = { chLatin_D, chLatin_T, chLatin_D , chNull };
            DTDEntityDecl* declDTD = new (fGrammarPoolMemoryManager) DTDEntityDecl(gDTDStr, false, fGrammarPoolMemoryManager);
            declDTD->setSystemId(sysId);
            Janitor<DTDEntityDecl> janDecl(declDTD);

            // Mark this one as a throw at end
            reader->setThrowAtEnd(true);

            // And push it onto the stack, with its pseudo name
            fReaderMgr.pushReader(reader, declDTD);

            // Tell it its not in an include section
            dtdScanner.scanExtSubsetDecl(false, true);
        }
    }
}

bool DGXMLScanner::scanStartTag(bool& gotData)
{
    //  Assume we will still have data until proven otherwise. It will only
    //  ever be false if this is the root and its empty.
    gotData = true;

    //  Get the QName. In this case, we are not doing namespaces, so we just
    //  use it as is and don't have to break it into parts.
    if (!fReaderMgr.getName(fQNameBuf))
    {
        emitError(XMLErrs::ExpectedElementName);
        fReaderMgr.skipToChar(chOpenAngle);
        return false;
    }

    // Assume it won't be an empty tag
    bool isEmpty = false;

    // See if its the root element
    const bool isRoot = fElemStack.isEmpty();

    //  Lets try to look up the element in the validator's element decl pool
    //  We can pass bogus values for the URI id and the base name. We know that
    //  this can only be called if we are doing a DTD style validator and that
    //  he will only look at the QName.
    //
    //  We *do not* tell him to fault in a decl if he does not find one - NG.
    bool wasAdded = false;
    const XMLCh* qnameRawBuf = fQNameBuf.getRawBuffer(); 

    XMLElementDecl* elemDecl = fGrammar->getElemDecl
    (
        fEmptyNamespaceId
        , 0
        , qnameRawBuf
        , Grammar::TOP_LEVEL_SCOPE
    );
    // look in the undeclared pool:
    if(!elemDecl) 
    {
        elemDecl = fDTDElemNonDeclPool->getByKey(qnameRawBuf);
    }
    if(!elemDecl) 
    {
        wasAdded = true;
        elemDecl = new (fMemoryManager) DTDElementDecl 
        (
            qnameRawBuf
            , fEmptyNamespaceId
            , DTDElementDecl::Any
            , fMemoryManager
        );
        elemDecl->setId(fDTDElemNonDeclPool->put((DTDElementDecl*)elemDecl));
    }

    if (fValidate) {

        if (wasAdded)
        {
            // This is to tell the reuse Validator that this element was
            // faulted-in, was not an element in the validator pool originally
            elemDecl->setCreateReason(XMLElementDecl::JustFaultIn);

            fValidator->emitError
            (
                XMLValid::ElementNotDefined
                , qnameRawBuf
            );
        }
        // If its not marked declared, then emit an error
        else if (!elemDecl->isDeclared())
        {
            fValidator->emitError
            (
                XMLValid::ElementNotDefined
                , qnameRawBuf
            );
        }