igxmlscanner2.cpp

来自「IBM的解析xml的工具Xerces的源代码」· C++ 代码 · 共 1,601 行 · 第 1/5 页

                         fValidator->emitError(XMLValid::NoAttNormForStandalone, attName);
                    }
                    nextCh = chSpace;
                }
            }
        }
        else
        {
            if (curState == InWhitespace)
            {
                if (!fReaderMgr.getCurrentReader()->isWhitespace(nextCh))
                {
                    if (firstNonWS)
                        toFill.append(chSpace);
                    curState = InContent;
                    firstNonWS = true;
                }
                else
                {
                    srcPtr++;
                    continue;
                }
            }
            else if (curState == InContent)
            {
                if (fReaderMgr.getCurrentReader()->isWhitespace(nextCh))
                {
                    curState = InWhitespace;
                    srcPtr++;

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

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

        // And move up to the next character in the source
        srcPtr++;
    }

    return retVal;
}

//  This method will just normalize the input value as CDATA without
//  any standalone checking.
bool IGXMLScanner::normalizeAttRawValue( const   XMLCh* const        attrName
                                      , const XMLCh* const        value
                                      ,       XMLBuffer&          toFill)
{
    // Assume its going to go fine, and empty the target buffer in preperation
    bool retVal = true;
    toFill.reset();

    //  Loop through the chars of the source value and normalize it according
    //  to the type.
    bool escaped;
    XMLCh nextCh;
    const XMLCh* srcPtr = value;
    while (*srcPtr)
    {
        //  Get the next character from the source. We have to watch for
        //  escaped characters (which are indicated by a 0xFFFF value followed
        //  by the char that was escaped.)
        nextCh = *srcPtr;
        escaped = (nextCh == 0xFFFF);
        if (escaped)
            nextCh = *++srcPtr;

        //  If its not escaped, then make sure its not a < character, which is
        //  not allowed in attribute values.
        if (!escaped && (*srcPtr == chOpenAngle))
        {
            emitError(XMLErrs::BracketInAttrValue, attrName);
            retVal = false;
        }

        if (!escaped)
        {
            //  NOTE: Yes this is a little redundant in that a 0x20 is
            //  replaced with an 0x20. But its faster to do this (I think)
            //  than checking for 9, A, and D separately.
            if (fReaderMgr.getCurrentReader()->isWhitespace(nextCh))
                nextCh = chSpace;
        }

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

        // And move up to the next character in the source
        srcPtr++;
    }
    return retVal;
}

unsigned int
IGXMLScanner::resolvePrefix(  const   XMLCh* const        prefix
                              , const ElemStack::MapModes mode)
{
    //  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.)
    if (XMLString::equals(prefix, XMLUni::fgXMLNSString))
        return fXMLNSNamespaceId;
    else if (XMLString::equals(prefix, XMLUni::fgXMLString))
        return fXMLNamespaceId;

    //  Ask the element stack to search up itself for a mapping for the
    //  passed prefix.
    bool unknown;
    unsigned int uriId = fElemStack.mapPrefixToURI(prefix, mode, unknown);

    // If it was unknown, then the URI was faked in but we have to issue an error
    if (unknown)
        emitError(XMLErrs::UnknownPrefix, prefix);

    return uriId;
}

unsigned int
IGXMLScanner::resolvePrefix(  const   XMLCh* const        prefix
                              ,       XMLBuffer&          bufToFill
                              , const ElemStack::MapModes mode)
{
    //  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.)
    if (XMLString::equals(prefix, XMLUni::fgXMLNSString))
        return fXMLNSNamespaceId;
    else if (XMLString::equals(prefix, XMLUni::fgXMLString))
        return fXMLNamespaceId;

    //  Ask the element stack to search up itself for a mapping for the
    //  passed prefix.
    bool unknown;
    unsigned int uriId = fElemStack.mapPrefixToURI(prefix, mode, unknown);

    // If it was unknown, then the URI was faked in but we have to issue an error
    if (unknown)
        emitError(XMLErrs::UnknownPrefix, prefix);

    getURIText(uriId,bufToFill);

    return uriId;
}


//  This method will reset the scanner data structures, and related plugged
//  in stuff, for a new scan session. We get the input source for the primary
//  XML entity, create the reader for it, and push it on the stack so that
//  upon successful return from here we are ready to go.
void IGXMLScanner::scanReset(const InputSource& src)
{

    //  This call implicitly tells us that we are going to reuse the scanner
    //  if it was previously used. So tell the validator to reset itself.
    //
    //  But, if the fUseCacheGrammar flag is set, then don't reset it.
    //
    //  NOTE:   The ReaderMgr is flushed on the way out, because that is
    //          required to insure that files are closed.
    fGrammarResolver->cacheGrammarFromParse(fToCacheGrammar);
    fGrammarResolver->useCachedGrammarInParse(fUseCachedGrammar);

    // fModel may need updating, as fGrammarResolver could have cleaned it
    if(fModel && getPSVIHandler())
        fModel = fGrammarResolver->getXSModel();

    {
        XMLDTDDescriptionImpl   theDTDDescription(XMLUni::fgDTDEntityString, fMemoryManager);
        fDTDGrammar = (DTDGrammar*) fGrammarResolver->getGrammar(&theDTDDescription);
    }

    if (!fDTDGrammar) {

        fDTDGrammar = new (fGrammarPoolMemoryManager) DTDGrammar(fGrammarPoolMemoryManager);
        fGrammarResolver->putGrammar(fDTDGrammar);
    }
    else
        fDTDGrammar->reset();

    fGrammar = fDTDGrammar;
    fGrammarType = fGrammar->getGrammarType();
    fRootGrammar = 0;

    if (fValidatorFromUser) {
        if (fValidator->handlesDTD())
            fValidator->setGrammar(fGrammar);
        else if (fValidator->handlesSchema()) {

            ((SchemaValidator*) fValidator)->setErrorReporter(fErrorReporter);
            ((SchemaValidator*) fValidator)->setGrammarResolver(fGrammarResolver);
            ((SchemaValidator*) fValidator)->setExitOnFirstFatal(fExitOnFirstFatal);
        }
    }
    else {
        // set fValidator as fDTDValidator
        fValidator = fDTDValidator;
        fValidator->setGrammar(fGrammar);
    }

    // Reset validation
    fValidate = (fValScheme == Val_Always) ? true : false;

    //  And for all installed handlers, send reset events. This gives them
    //  a chance to flush any cached data.
    if (fDocHandler)
        fDocHandler->resetDocument();
    if (fEntityHandler)
        fEntityHandler->resetEntities();
    if (fErrorReporter)
        fErrorReporter->resetErrors();

    // Clear out the id reference list
    resetValidationContext();

    // Reset the Root Element Name
    fMemoryManager->deallocate(fRootElemName);//delete [] fRootElemName;
    fRootElemName = 0;

    // Reset IdentityConstraints
    if (fICHandler)
        fICHandler->reset();

    //  Reset the element stack, and give it the latest ids for the special
    //  URIs it has to know about.
    fElemStack.reset
    (
        fEmptyNamespaceId
        , fUnknownNamespaceId
        , fXMLNamespaceId
        , fXMLNSNamespaceId
    );

    if (!fSchemaNamespaceId)
        fSchemaNamespaceId  = fURIStringPool->addOrFind(SchemaSymbols::fgURI_XSI);

    // Reset some status flags
    fInException = false;
    fStandalone = false;
    fErrorCount = 0;
    fHasNoDTD = true;
    fSeeXsi = false;

    // Reset PSVI context
    // note that we always need this around for DOMTypeInfo
    if (!fPSVIElement)
        fPSVIElement = new (fMemoryManager) PSVIElement(fMemoryManager);

    if (!fErrorStack)
    {
        fErrorStack = new (fMemoryManager) ValueStackOf<bool>(8, fMemoryManager);
    }
    else
    {
        fErrorStack->removeAllElements();
    }

    resetPSVIElemContext();

    // Reset the validators
    fDTDValidator->reset();
    fDTDValidator->setErrorReporter(fErrorReporter);
    fSchemaValidator->reset();
    fSchemaValidator->setErrorReporter(fErrorReporter);
    fSchemaValidator->setExitOnFirstFatal(fExitOnFirstFatal);
    fSchemaValidator->setGrammarResolver(fGrammarResolver);
    if (fValidatorFromUser)
        fValidator->reset();

    //  Handle the creation of the XML reader object for this input source.
    //  This will provide us with transcoding and basic lexing services.
    XMLReader* newReader = fReaderMgr.createReader
    (
        src
        , true
        , XMLReader::RefFrom_NonLiteral
        , XMLReader::Type_General
        , XMLReader::Source_External
        , fCalculateSrcOfs
    );

    if (!newReader) {
        if (src.getIssueFatalErrorIfNotFound())
            ThrowXMLwithMemMgr1(RuntimeException, XMLExcepts::Scan_CouldNotOpenSource, src.getSystemId(), fMemoryManager);
        else
            ThrowXMLwithMemMgr1(RuntimeException, XMLExcepts::Scan_CouldNotOpenSource_Warning, src.getSystemId(), fMemoryManager);
    }

    // Push this read onto the reader manager
    fReaderMgr.pushReader(newReader, 0);

    // and reset security-related things if necessary:
    if(fSecurityManager != 0) 
    {
        fEntityExpansionLimit = fSecurityManager->getEntityExpansionLimit();
        fEntityExpansionCount = 0;
    }
    fElemCount = 0;
    if(fUIntPoolRowTotal >= 32) 
    { // 8 KB tied up with validating attributes...
        fAttDefRegistry->removeAll();
        fUndeclaredAttrRegistry->removeAll();
        fUndeclaredAttrRegistryNS->removeAll();
        recreateUIntPool();
    }