dtdscanner.cpp

来自「IBM的解析xml的工具Xerces的源代码」· C++ 代码 · 共 1,936 行 · 第 1/5 页

    //
    //  If we are in the internal subset and in markup, then this is
    //  an error but we go ahead and do it anyway.
    //
    if (fInternalSubset && inMarkup)
        fScanner->emitError(XMLErrs::PERefInMarkupInIntSubset);

    if (!fReaderMgr->getName(bbName.getBuffer()))
    {
        fScanner->emitError(XMLErrs::ExpectedPEName);

        // Skip the semicolon if that's what we ended up on
        fReaderMgr->skippedChar(chSemiColon);
        return false;
    }

    // If no terminating semicolon, emit an error but try to keep going
    if (!fReaderMgr->skippedChar(chSemiColon))
        fScanner->emitError(XMLErrs::UnterminatedEntityRef, bbName.getRawBuffer());

    //
    //  Look it up in the PE decl pool and see if it exists. If not, just
    //  emit an error and continue.
    //
    XMLEntityDecl* decl = fPEntityDeclPool->getByKey(bbName.getRawBuffer());
    if (!decl)
    {
        // XML 1.0 Section 4.1
        if (fScanner->getStandalone()) {
            // no need to check fScanner->fHasNoDTD which is for sure false
            // since we are in expandPERef already
            fScanner->emitError(XMLErrs::EntityNotFound, bbName.getRawBuffer());
        }
        else {
            if (fScanner->getDoValidation())
                fScanner->getValidator()->emitError(XMLValid::VC_EntityNotFound, bbName.getRawBuffer());
        }

        return false;
    }

    //
    // XML 1.0 Section 2.9
    //  If we are a standalone document, then it has to have been declared
    //  in the internal subset. Keep going though.
    //
    if (fScanner->getDoValidation() && fScanner->getStandalone() && !decl->getDeclaredInIntSubset())
        fScanner->getValidator()->emitError(XMLValid::VC_IllegalRefInStandalone, bbName.getRawBuffer());

    //
    //  Okee dokee, we found it. So create either a memory stream with
    //  the entity value contents, or a file stream if its an external
    //  entity.
    //
    if (decl->isExternal())
    {
        // And now create a reader to read this entity
        InputSource* srcUsed;
        XMLReader* reader = fReaderMgr->createReader
        (
            decl->getBaseURI()
            , decl->getSystemId()
            , decl->getPublicId()
            , false
            , inLiteral ? XMLReader::RefFrom_Literal : XMLReader::RefFrom_NonLiteral
            , XMLReader::Type_PE
            , XMLReader::Source_External
            , srcUsed
        );

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

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

        // Set the 'throw at end' flag, to the one we were given
        reader->setThrowAtEnd(throwEndOfExt);

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

        //
        //  If the caller wants us to scan the external entity, then lets
        //  do that now.
        //
        if (scanExternal)
        {
            XMLEntityHandler* entHandler = fScanner->getEntityHandler();

            // If we have an entity handler, tell it we are starting this entity
            if (entHandler)
                entHandler->startInputSource(*srcUsed);

            //
            //  Scan the external entity now. The parameter tells it that
            //  it is not in an include section. Get the current reader
            //  level so we can catch partial markup errors and be sure
            //  to get back to here if we get an exception out of the
            //  ext subset scan.
            //
            const unsigned int readerNum = fReaderMgr->getCurrentReaderNum();
            try
            {
                scanExtSubsetDecl(false, false);
            }
            catch(const OutOfMemoryException&)
            {
                throw;
            }
            catch(...)
            {
                // Pop the reader back to the original level
                fReaderMgr->cleanStackBackTo(readerNum);

                // End the input source, even though its not happy
                if (entHandler)
                    entHandler->endInputSource(*srcUsed);
                throw;
            }

            // If we have an entity handler, tell it we are ending this entity
            if (entHandler)
                entHandler->endInputSource(*srcUsed);
        }
        else {
            // If it starts with the XML string, then parse a text decl
            if (fScanner->checkXMLDecl(true))
                scanTextDecl();
        }
    }
     else
    {
        // Create a reader over a memory stream over the entity value
        XMLReader* valueReader = fReaderMgr->createIntEntReader
        (
            decl->getName()
            , inLiteral ? XMLReader::RefFrom_Literal : XMLReader::RefFrom_NonLiteral
            , XMLReader::Type_PE
            , decl->getValue()
            , decl->getValueLen()
            , false
        );

        //
        //  Trt 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
        //  will have just been discarded, but we just keep going.
        //
        if (!fReaderMgr->pushReader(valueReader, decl))
            fScanner->emitError(XMLErrs::RecursiveEntity, decl->getName());
    }

    return true;
}


bool DTDScanner::getQuotedString(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;

    while (true)
    {
        // Get another char
        const XMLCh nextCh = fReaderMgr->getNextChar();

        // See if it matches the starting quote char
        if (nextCh == quoteCh)
            break;

        //
        //  We should never get either an end of file null char here. If we
        //  do, just fail. It will be handled more gracefully in the higher
        //  level code that called us.
        //
        if (!nextCh)
            return false;

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


XMLAttDef*
DTDScanner::scanAttDef(DTDElementDecl& parentElem, XMLBuffer& bufToUse)
{
    // Check for PE ref or optional whitespace
    checkForPERef(false, true);

    // Get the name of the attribute
    if (!fReaderMgr->getName(bufToUse))
    {
        fScanner->emitError(XMLErrs::ExpectedAttrName);
        return 0;
    }

    //
    //  Look up this attribute in the parent element's attribute list. If
    //  it already exists, then use the dummy.
    //
    DTDAttDef* decl = parentElem.getAttDef(bufToUse.getRawBuffer());
    if (decl)
    {
        // It already exists, so put out a warning
        fScanner->emitError
        (
            XMLErrs::AttListAlreadyExists
            , bufToUse.getRawBuffer()
            , parentElem.getFullName()
        );

        // Use the dummy decl to parse into and set its name to the name we got
        if (!fDumAttDef)
        {
            fDumAttDef = new (fMemoryManager) DTDAttDef(fMemoryManager);
            fDumAttDef->setId(fNextAttrId++);
        }
        fDumAttDef->setName(bufToUse.getRawBuffer());
        decl = fDumAttDef;
    }
     else
    {
        //
        //  It does not already exist so create a new one, give it the next
        //  available unique id, and add it
        //
        decl = new (fGrammarPoolMemoryManager) DTDAttDef
        (
            bufToUse.getRawBuffer()
            , XMLAttDef::CData
            , XMLAttDef::Implied
            , fGrammarPoolMemoryManager
        );
        decl->setId(fNextAttrId++);
        decl->setExternalAttDeclaration(isReadingExternalEntity());
        parentElem.addAttDef(decl);
    }

    // Set a flag to indicate whether we are doing a dummy parse
    const bool isIgnored = (decl == fDumAttDef);

    // Space is required here, so check for PE ref, and require space
    if (!checkForPERef(false, true))
        fScanner->emitError(XMLErrs::ExpectedWhitespace);

    //
    //  Next has to be one of the attribute type strings. This tells us what
    //  is to follow.
    //
    if (fReaderMgr->skippedString(XMLUni::fgCDATAString))
    {
        decl->setType(XMLAttDef::CData);
    }
     else if (fReaderMgr->skippedString(XMLUni::fgIDString))
    {
        if (!fReaderMgr->skippedString(XMLUni::fgRefString))
            decl->setType(XMLAttDef::ID);
        else if (!fReaderMgr->skippedChar(chLatin_S))
            decl->setType(XMLAttDef::IDRef);
        else
            decl->setType(XMLAttDef::IDRefs);
    }
     else if (fReaderMgr->skippedString(XMLUni::fgEntitString))
    {
        if (fReaderMgr->skippedChar(chLatin_Y))
        {
            decl->setType(XMLAttDef::Entity);
        }
         else if (fReaderMgr->skippedString(XMLUni::fgIESString))
        {
            decl->setType(XMLAttDef::Entities);
        }
         else
        {
            fScanner->emitError
            (
                XMLErrs::ExpectedAttributeType
                , decl->getFullName()
                , parentElem.getFullName()
            );
            return 0;
        }
    }
     else if (fReaderMgr->skippedString(XMLUni::fgNmTokenString))
    {
        if (fReaderMgr->skippedChar(chLatin_S))
            decl->setType(XMLAttDef::NmTokens);
        else
            decl->setType(XMLAttDef::NmToken);
    }
     else if (fReaderMgr->skippedString(XMLUni::fgNotationString))
    {
        // Check for PE ref and require space
        if (!checkForPERef(false, true))
            fScanner->emitError(XMLErrs::ExpectedWhitespace);

        decl->setType(XMLAttDef::Notation);
        if (!scanEnumeration(*decl, bufToUse, true))
            return 0;

        // Set the value as the enumeration for this decl
        decl->setEnumeration(bufToUse.getRawBuffer());
    }
     else if (fReaderMgr->skippedChar(chOpenParen))
    {
        decl->setType(XMLAttDef::Enumeration);
        if (!scanEnumeration(*decl, bufToUse, false))
            return 0;

        // Set the value as the enumeration for this decl
        decl->setEnumeration(bufToUse.getRawBuffer());
    }
     else
    {
        fScanner->emitError
        (
            XMLErrs::ExpectedAttributeType
            , decl->getFullName()
            , parentElem.getFullName()
        );
        return 0;
    }

    // Space is required here, so check for PE ref, and require space
    if (!checkForPERef(false, true))
        fScanner->emitError(XMLErrs::ExpectedWhitespace);

    // And then scan for the optional default value declaration
    scanDefaultDecl(*decl);

    // If validating, then do a couple of validation constraints
    if (fScanner->getDoValidation())
    {
        if (decl->getType() == XMLAttDef::ID)
        {
            if ((decl->getDefaultType() != XMLAttDef::Implied)
            &&  (decl->getDefaultType() != XMLAttDef::Required))
            {
                fScanner->getValidator()->emitError(XMLValid::BadIDAttrDefType, decl->getFullName());
            }
        }

        // if attdef is xml:space, check correct enumeration (default|preserve)
        const XMLCh fgXMLSpace[] = { chLatin_x, chLatin_m, chLatin_l, chColon, chLatin_s, chLatin_p, chLatin_a, chLatin_c, chLatin_e, chNull };

        if (XMLString::equals(decl->getFullName(),fgXMLSpace)) {
            const XMLCh fgPreserve[] = { chLatin_p, chLatin_r, chLatin_e, chLatin_s, chLatin_e, chLatin_r, chLatin_v, chLatin_e, chNull };
            const XMLCh fgDefault[] = { chLatin_d, chLatin_e, chLatin_f, chLatin_a, chLatin_u, chLatin_l, chLatin_t, chNull };
            bool ok = false;
            if (decl->getType() == XMLAttDef::Enumeration) {
                BaseRefVectorOf<XMLCh>* enumVector = XMLString::tokenizeString(decl->getEnumeration(), fMemoryManager);
                int size = enumVector->size();
                ok = (size == 1 &&
                     (XMLString::equals(enumVector->elementAt(0), fgDefault) ||
                      XMLString::equals(enumVector->elementAt(0), fgPreserve))) ||
                     (size == 2 &&
                     (XMLString::equals(enumVector->elementAt(0), fgDefault) &&
                      XMLString::equals(enumVector->elementAt(1), fgPreserve))) ||
                     (size == 2 &&
                     (XMLString::equals(enumVector->elementAt(1), fgDefault) &&
                      XMLString::equals(enumVector->elementAt(0), fgPreserve)));
                delete enumVector;
            }
            if (!ok)
                fScanner->getValidator()->emitError(XMLValid::IllegalXMLSpace);
        }
    }

    // If we have a doc type handler, tell it about this attdef.
    if (fDocTypeHandler)