/ATF2/control-software/epics-3.14.8/extensions/src/ChannelArchiver/ThirdParty/xerces-c-src2_4_0/src/xercesc/internal/IGXMLScanner.cpp

Large files files are truncated, but you can click here to view the full file

/*
 * The Apache Software License, Version 1.1
 *
 * Copyright (c) 2002 The Apache Software Foundation.  All rights
 * reserved.
 *
 * Redistribution and use in source and binary forms, with or without
 * modification, are permitted provided that the following conditions
 * are met:
 *
 * 1. Redistributions of source code must retain the above copyright
 *    notice, this list of conditions and the following disclaimer.
 *
 * 2. Redistributions in binary form must reproduce the above copyright
 *    notice, this list of conditions and the following disclaimer in
 *    the documentation and/or other materials provided with the
 *    distribution.
 *
 * 3. The end-user documentation included with the redistribution,
 *    if any, must include the following acknowledgment:
 *       "This product includes software developed by the
 *        Apache Software Foundation (http://www.apache.org/)."
 *    Alternately, this acknowledgment may appear in the software itself,
 *    if and wherever such third-party acknowledgments normally appear.
 *
 * 4. The names "Xerces" and "Apache Software Foundation" must
 *    not be used to endorse or promote products derived from this
 *    software without prior written permission. For written
 *    permission, please contact apache\@apache.org.
 *
 * 5. Products derived from this software may not be called "Apache",
 *    nor may "Apache" appear in their name, without prior written
 *    permission of the Apache Software Foundation.
 *
 * THIS SOFTWARE IS PROVIDED ``AS IS'' AND ANY EXPRESSED OR IMPLIED
 * WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
 * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
 * DISCLAIMED.  IN NO EVENT SHALL THE APACHE SOFTWARE FOUNDATION OR
 * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
 * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
 * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF
 * USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
 * ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY,
 * OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT
 * OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
 * SUCH DAMAGE.
 * ====================================================================
 *
 * This software consists of voluntary contributions made by many
 * individuals on behalf of the Apache Software Foundation, and was
 * originally based on software copyright (c) 1999, International
 * Business Machines, Inc., http://www.ibm.com .  For more information
 * on the Apache Software Foundation, please see
 * <http://www.apache.org/>.
 */

/*
 * $Id: IGXMLScanner.cpp,v 1.1.1.1 2009/03/14 06:42:17 whitegr Exp $
 */


// ---------------------------------------------------------------------------
//  Includes
// ---------------------------------------------------------------------------
#include <xercesc/internal/IGXMLScanner.hpp>
#include <xercesc/util/RuntimeException.hpp>
#include <xercesc/util/HashPtr.hpp>
#include <xercesc/util/UnexpectedEOFException.hpp>
#include <xercesc/sax/InputSource.hpp>
#include <xercesc/framework/XMLDocumentHandler.hpp>
#include <xercesc/framework/XMLEntityHandler.hpp>
#include <xercesc/framework/XMLPScanToken.hpp>
#include <xercesc/internal/EndOfEntityException.hpp>
#include <xercesc/framework/MemoryManager.hpp>
#include <xercesc/framework/XMLGrammarPool.hpp>
#include <xercesc/framework/XMLDTDDescription.hpp>
#include <xercesc/framework/psvi/PSVIHandler.hpp>
#include <xercesc/framework/psvi/PSVIAttributeList.hpp>
#include <xercesc/validators/common/GrammarResolver.hpp>
#include <xercesc/validators/DTD/DocTypeHandler.hpp>
#include <xercesc/validators/DTD/DTDScanner.hpp>
#include <xercesc/validators/DTD/DTDValidator.hpp>
#include <xercesc/validators/schema/SchemaValidator.hpp>
#include <xercesc/validators/schema/identity/FieldActivator.hpp>
#include <xercesc/validators/schema/identity/XPathMatcherStack.hpp>
#include <xercesc/validators/schema/identity/ValueStoreCache.hpp>
#include <xercesc/validators/schema/identity/IC_Selector.hpp>
#include <xercesc/validators/schema/identity/ValueStore.hpp>
#include <xercesc/util/OutOfMemoryException.hpp>

XERCES_CPP_NAMESPACE_BEGIN

// ---------------------------------------------------------------------------
//  IGXMLScanner: Constructors and Destructor
// ---------------------------------------------------------------------------
IGXMLScanner::IGXMLScanner( XMLValidator* const  valToAdopt
                          , GrammarResolver* const grammarResolver
                          , MemoryManager* const manager) :

    XMLScanner(valToAdopt, grammarResolver, manager)
    , fSeeXsi(false)
    , fElemStateSize(16)
    , fElemState(0)
    , fContent(1023, manager)
    , fRawAttrList(0)
    , fDTDValidator(0)
    , fSchemaValidator(0)
    , fDTDGrammar(0)
    , fMatcherStack(0)
    , fValueStoreCache(0)
    , fFieldActivator(0)
    , fDTDElemNonDeclPool(0)
    , fSchemaElemNonDeclPool(0)
    , fElemCount(0)
    , fAttDefRegistry(0)
    , fUndeclaredAttrRegistry(0)
    , fUndeclaredAttrRegistryNS(0)
    , fPSVIAttrList(0)
    , fModel(0)
    , fPSVIElement(0)
    , fErrorStack(0)
{
    try
    {
         commonInit();

         // use fDTDValidator as the default validator
         if (!valToAdopt)
             fValidator = fDTDValidator;
    }
    catch(const OutOfMemoryException&)
    {
        throw;
    }
    catch(...)
    {
        cleanUp();
        throw;
    }
}

IGXMLScanner::IGXMLScanner( XMLDocumentHandler* const docHandler
                          , DocTypeHandler* const     docTypeHandler
                          , XMLEntityHandler* const   entityHandler
                          , XMLErrorReporter* const   errHandler
                          , XMLValidator* const       valToAdopt
                          , GrammarResolver* const    grammarResolver
                          , MemoryManager* const      manager) :

    XMLScanner(docHandler, docTypeHandler, entityHandler, errHandler, valToAdopt, grammarResolver, manager)
    , fSeeXsi(false)
    , fElemStateSize(16)
    , fElemState(0)
    , fContent(1023, manager)
    , fRawAttrList(0)
    , fDTDValidator(0)
    , fSchemaValidator(0)
    , fDTDGrammar(0)
    , fMatcherStack(0)
    , fValueStoreCache(0)
    , fFieldActivator(0)
    , fDTDElemNonDeclPool(0)
    , fSchemaElemNonDeclPool(0)
    , fElemCount(0)
    , fAttDefRegistry(0)
    , fUndeclaredAttrRegistry(0)
    , fUndeclaredAttrRegistryNS(0)
    , fPSVIAttrList(0)
    , fModel(0)
    , fPSVIElement(0)
    , fErrorStack(0)
{
    try
    {	
        commonInit();

        //use fDTDValidator as the default validator
        if (!valToAdopt)
            fValidator = fDTDValidator;
    }
    catch(const OutOfMemoryException&)
    {
        throw;
    }
    catch(...)
    {
        cleanUp();
        throw;
    }
}

IGXMLScanner::~IGXMLScanner()
{
    cleanUp();
}

// ---------------------------------------------------------------------------
//  XMLScanner: Getter methods
// ---------------------------------------------------------------------------
NameIdPool<DTDEntityDecl>* IGXMLScanner::getEntityDeclPool()
{
    if(!fDTDGrammar)
        return 0;
    return fDTDGrammar->getEntityDeclPool();
}

const NameIdPool<DTDEntityDecl>* IGXMLScanner::getEntityDeclPool() const
{
    if(!fDTDGrammar)
        return 0;
    return fDTDGrammar->getEntityDeclPool();
}

// ---------------------------------------------------------------------------
//  IGXMLScanner: Main entry point to scan a document
// ---------------------------------------------------------------------------
void IGXMLScanner::scanDocument(const InputSource& src)
{
    //  Bump up the sequence id for this parser instance. This will invalidate
    //  any previous progressive scan tokens.
    fSequenceId++;

    try
    {
        //  Reset the scanner and its plugged in stuff for a new run. This
        //  resets all the data structures, creates the initial reader and
        //  pushes it on the stack, and sets up the base document path.
        scanReset(src);

        // If we have a document handler, then call the start document
        if (fDocHandler)
            fDocHandler->startDocument();

        //  Scan the prolog part, which is everything before the root element
        //  including the DTD subsets.
        scanProlog();

        //  If we got to the end of input, then its not a valid XML file.
        //  Else, go on to scan the content.
        if (fReaderMgr.atEOF())
        {
            emitError(XMLErrs::EmptyMainEntity);
        }
        else
        {
            // Scan content, and tell it its not an external entity
            if (scanContent(false))
            {
                // Do post-parse validation if required
                if (fValidate)
                {
                    //  We handle ID reference semantics at this level since
                    //  its required by XML 1.0.
                    checkIDRefs();

                    // Then allow the validator to do any extra stuff it wants
//                    fValidator->postParseValidation();
                }

                // That went ok, so scan for any miscellaneous stuff
                if (!fReaderMgr.atEOF())
                    scanMiscellaneous();
            }
        }

        // If we have a document handler, then call the end document
        if (fDocHandler)
            fDocHandler->endDocument();

        // Reset the reader manager to close all files, sockets, etc...
        fReaderMgr.reset();
    }
    //  NOTE:
    //
    //  In all of the error processing below, the emitError() call MUST come
    //  before the flush of the reader mgr, or it will fail because it tries
    //  to find out the position in the XML source of the error.
    catch(const XMLErrs::Codes)
    {
        // This is a 'first fatal error' type exit, so reset and fall through
        fReaderMgr.reset();
    }
    catch(const XMLValid::Codes)
    {
        // This is a 'first fatal error' type exit, so reset and fall through
        fReaderMgr.reset();
    }
    catch(const XMLException& excToCatch)
    {
        //  Emit the error and catch any user exception thrown from here. Make
        //  sure in all cases we flush the reader manager.
        fInException = true;
        try
        {
            if (excToCatch.getErrorType() == XMLErrorReporter::ErrType_Warning)
                emitError
                (
                    XMLErrs::XMLException_Warning
                    , excToCatch.getType()
                    , excToCatch.getMessage()
                );
            else if (excToCatch.getErrorType() >= XMLErrorReporter::ErrType_Fatal)
                emitError
                (
                    XMLErrs::XMLException_Fatal
                    , excToCatch.getType()
                    , excToCatch.getMessage()
                );
            else
                emitError
                (
                    XMLErrs::XMLException_Error
                    , excToCatch.getType()
                    , excToCatch.getMessage()
                );
        }
        catch(const OutOfMemoryException&)
        {
            throw;
        }
        catch(...)
        {
            // Flush the reader manager and rethrow user's error
            fReaderMgr.reset();
            throw;
        }

        // If it returned, then reset the reader manager and fall through
        fReaderMgr.reset();
    }
    catch(const OutOfMemoryException&)
    {
        throw;
    }
    catch(...)
    {
        // Reset and rethrow
        fReaderMgr.reset();
        throw;
    }
}


bool IGXMLScanner::scanNext(XMLPScanToken& token)
{
    // Make sure this token is still legal
    if (!isLegalToken(token))
        ThrowXML(RuntimeException, XMLExcepts::Scan_BadPScanToken);

    // Find the next token and remember the reader id
    unsigned int orgReader;
    XMLTokens curToken;

    bool retVal = true;

    try
    {
        while (true)
        {
            //  We have to handle any end of entity exceptions that happen here.
            //  We could be at the end of X nested entities, each of which will
            //  generate an end of entity exception as we try to move forward.
            try
            {
                curToken = senseNextToken(orgReader);
                break;
            }
            catch(const EndOfEntityException& toCatch)
            {
                // Send an end of entity reference event
                if (fDocHandler)
                    fDocHandler->endEntityReference(toCatch.getEntity());
            }
        }

        if (curToken == Token_CharData)
        {
            scanCharData(fCDataBuf);
        }
        else if (curToken == Token_EOF)
        {
            if (!fElemStack.isEmpty())
            {
                const ElemStack::StackElem* topElem = fElemStack.popTop();
                emitError
                (
                    XMLErrs::EndedWithTagsOnStack
                    , topElem->fThisElement->getFullName()
                );
            }

            retVal = false;
        }
        else
        {
            // Its some sort of markup
            bool gotData = true;
            switch(curToken)
            {
                case Token_CData :
                    // Make sure we are within content
                    if (fElemStack.isEmpty())
                        emitError(XMLErrs::CDATAOutsideOfContent);
                    scanCDSection();
                    break;

                case Token_Comment :
                    scanComment();
                    break;

                case Token_EndTag :
                    scanEndTag(gotData);
                    break;

                case Token_PI :
                    scanPI();
                    break;

                case Token_StartTag :
                    if (fDoNamespaces)
                        scanStartTagNS(gotData);
                    else
                        scanStartTag(gotData);
                    break;

                default :
                    fReaderMgr.skipToChar(chOpenAngle);
                    break;
            }

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

            // If we hit the end, then do the miscellaneous part
            if (!gotData)
            {
                // Do post-parse validation if required
                if (fValidate)
                {
                    //  We handle ID reference semantics at this level since
                    //  its required by XML 1.0.
                    checkIDRefs();

                    // Then allow the validator to do any extra stuff it wants
//                    fValidator->postParseValidation();
                }

                // That went ok, so scan for any miscellaneous stuff
                scanMiscellaneous();

                if (fValidate)
                    fValueStoreCache->endDocument();

                if (fDocHandler)
                    fDocHandler->endDocument();
            }
        }
    }
    //  NOTE:
    //
    //  In all of the error processing below, the emitError() call MUST come
    //  before the flush of the reader mgr, or it will fail because it tries
    //  to find out the position in the XML source of the error.
    catch(const XMLErrs::Codes)
    {
        // This is a 'first failure' exception, so reset and return failure
        fReaderMgr.reset();
        return false;
    }
    catch(const XMLValid::Codes)
    {
        // This is a 'first fatal error' type exit, so reset and reuturn failure
        fReaderMgr.reset();
        return false;
    }
    catch(const XMLException& excToCatch)
    {
        //  Emit the error and catch any user exception thrown from here. Make
        //  sure in all cases we flush the reader manager.
        fInException = true;
        try
        {
            if (excToCatch.getErrorType() == XMLErrorReporter::ErrType_Warning)
                emitError
                (
                    XMLErrs::XMLException_Warning
                    , excToCatch.getType()
                    , excToCatch.getMessage()
                );
            else if (excToCatch.getErrorType() >= XMLErrorReporter::ErrType_Fatal)
                emitError
                (
                    XMLErrs::XMLException_Fatal
                    , excToCatch.getType()
                    , excToCatch.getMessage()
                );
            else
                emitError
                (
                    XMLErrs::XMLException_Error
                    , excToCatch.getType()
                    , excToCatch.getMessage()
                );
        }
        catch(const OutOfMemoryException&)
        {
            throw;
        }
        catch(...)
        {
            // Reset and rethrow user error
            fReaderMgr.reset();
            throw;
        }

        // Reset and return failure
        fReaderMgr.reset();
        return false;
    }
    catch(const OutOfMemoryException&)
    {
        throw;
    }
    catch(...)
    {
        // Reset and rethrow original error
        fReaderMgr.reset();
        throw;
    }

    // If we hit the end, then flush the reader manager
    if (!retVal)
        fReaderMgr.reset();

    return retVal;
}



// ---------------------------------------------------------------------------
//  IGXMLScanner: Private helper methods. Most of these are implemented in
//  IGXMLScanner2.Cpp.
// ---------------------------------------------------------------------------

//  This method handles the common initialization, to avoid having to do
//  it redundantly in multiple constructors.
void IGXMLScanner::commonInit()
{

    //  Create the element state array
    fElemState = (unsigned int*) fMemoryManager->allocate
    (
        fElemStateSize * sizeof(unsigned int)
    ); //new unsigned int[fElemStateSize];

    //  And we need one for the raw attribute scan. This just stores key/
    //  value string pairs (prior to any processing.)
    fRawAttrList = new (fMemoryManager) RefVectorOf<KVStringPair>(32, true, fMemoryManager);

    //  Create the Validator and init them
    fDTDValidator = new (fMemoryManager) DTDValidator();
    initValidator(fDTDValidator);
    fSchemaValidator = new (fMemoryManager) SchemaValidator(0, fMemoryManager);
    initValidator(fSchemaValidator);

    // Create IdentityConstraint info
    fMatcherStack = new (fMemoryManager) XPathMatcherStack(fMemoryManager);
    fValueStoreCache = new (fMemoryManager) ValueStoreCache(fMemoryManager);
    fFieldActivator = new (fMemoryManager) FieldActivator(fValueStoreCache, fMatcherStack, fMemoryManager);
    fValueStoreCache->setScanner(this);

    // Create schemaLocation pair info
    fLocationPairs = new (fMemoryManager) ValueVectorOf<XMLCh*>(8, fMemoryManager);
    // create pools for undeclared elements
    fDTDElemNonDeclPool = new (fMemoryManager) NameIdPool<DTDElementDecl>(29, 128, fMemoryManager);
    fSchemaElemNonDeclPool = new (fMemoryManager) RefHash3KeysIdPool<SchemaElementDecl>(29, true, 128, fMemoryManager); 
    fAttDefRegistry = new (fMemoryManager) RefHashTableOf<unsigned int>
    (
        509, false, new (fMemoryManager)HashPtr(), fMemoryManager
    );
    fUndeclaredAttrRegistry = new (fMemoryManager) RefHashTableOf<unsigned int>
    (
        509, false, new (fMemoryManager)HashXMLCh(), fMemoryManager
    );
    fUndeclaredAttrRegistryNS = new (fMemoryManager) RefHash2KeysTableOf<unsigned int>
    (
        509, false, new (fMemoryManager)HashXMLCh(), fMemoryManager
    );
    fPSVIAttrList = new (fMemoryManager) PSVIAttributeList(fMemoryManager);
}

void IGXMLScanner::cleanUp()
{
    fMemoryManager->deallocate(fElemState); //delete [] fElemState;
    delete fRawAttrList;
    delete fDTDValidator;
    delete fSchemaValidator;
    delete fFieldActivator;
    delete fMatcherStack;
    delete fValueStoreCache;
    delete fLocationPairs;
    delete fDTDElemNonDeclPool;
    delete fSchemaElemNonDeclPool;
    delete fAttDefRegistry;
    delete fUndeclaredAttrRegistry;
    delete fUndeclaredAttrRegistryNS;
    delete fPSVIAttrList;
    delete fPSVIElement;
    delete fErrorStack;
}

// ---------------------------------------------------------------------------
//  IGXMLScanner: Private scanning methods
// ---------------------------------------------------------------------------

//  This method is called from scanStartTag() to handle the very raw initial
//  scan of the attributes. It just fills in the passed collection with
//  key/value pairs for each attribute. No processing is done on them at all.
unsigned int
IGXMLScanner::rawAttrScan(const   XMLCh* const                elemName
                          ,       RefVectorOf<KVStringPair>&  toFill
                          ,       bool&                       isEmpty)
{
    //  Keep up with how many attributes we've seen so far, and how many
    //  elements are available in the vector. This way we can reuse old
    //  elements until we run out and then expand it.
    unsigned int attCount = 0;
    unsigned int curVecSize = toFill.size();

    // Assume it is not empty
    isEmpty = false;

    //  We loop until we either see a /> or >, handling key/value pairs util
    //  we get there. We place them in the passed vector, which we will expand
    //  as required to hold them.
    while (true)
    {
        // Get the next character, which should be non-space
        XMLCh nextCh = fReaderMgr.peekNextChar();

        //  If the next character is not a slash or closed angle bracket,
        //  then it must be whitespace, since whitespace is required
        //  between the end of the last attribute and the name of the next
        //  one.
        //
        if (attCount)
        {
            if ((nextCh != chForwardSlash) && (nextCh != chCloseAngle))
            {
                if (fReaderMgr.getCurrentReader()->isWhitespace(nextCh))
                {
                    // Ok, skip by them and get another char
                    fReaderMgr.getNextChar();
                    fReaderMgr.skipPastSpaces();
                    nextCh = fReaderMgr.peekNextChar();
                }
                 else
                {
                    // Emit the error but keep on going
                    emitError(XMLErrs::ExpectedWhitespace);
                }
            }
        }

        //  Ok, here we first check for any of the special case characters.
        //  If its not one, then we do the normal case processing, which
        //  assumes that we've hit an attribute value, Otherwise, we do all
        //  the special case checks.
        if (!fReaderMgr.getCurrentReader()->isSpecialStartTagChar(nextCh))
        {
            //  Assume its going to be an attribute, so get a name from
            //  the input.
            if (!fReaderMgr.getName(fAttNameBuf))
            {
                emitError(XMLErrs::ExpectedAttrName);
                fReaderMgr.skipPastChar(chCloseAngle);
                return attCount;
            }

            // And next must be an equal sign
            if (!scanEq())
            {
                static const XMLCh tmpList[] =
                {
                    chSingleQuote, chDoubleQuote, chCloseAngle
                    , chOpenAngle, chForwardSlash, chNull
                };

                emitError(XMLErrs::ExpectedEqSign);

                //  Try to sync back up by skipping forward until we either
                //  hit something meaningful.
                const XMLCh chFound = fReaderMgr.skipUntilInOrWS(tmpList);

                if ((chFound == chCloseAngle) || (chFound == chForwardSlash))
                {
                    // Jump back to top for normal processing of these
                    continue;
                }
                else if ((chFound == chSingleQuote)
                      ||  (chFound == chDoubleQuote)
                      ||  fReaderMgr.getCurrentReader()->isWhitespace(chFound))
                {
                    // Just fall through assuming that the value is to follow
                }
                else if (chFound == chOpenAngle)
                {
                    // Assume a malformed tag and that new one is starting
                    emitError(XMLErrs::UnterminatedStartTag, elemName);
                    return attCount;
                }
                else
                {
                    // Something went really wrong
                    return attCount;
                }
            }

            //  Next should be the quoted attribute value. We just do a simple
            //  and stupid scan of this value. The only thing we do here
            //  is to expand entity references.
            if (!basicAttrValueScan(fAttNameBuf.getRawBuffer(), fAttValueBuf))
            {
                static const XMLCh tmpList[] =
                {
                    chCloseAngle, chOpenAngle, chForwardSlash, chNull
                };

                emitError(XMLErrs::ExpectedAttrValue);

                //  It failed, so lets try to get synced back up. We skip
                //  forward until we find some whitespace or one of the
                //  chars in our list.
                const XMLCh chFound = fReaderMgr.skipUntilInOrWS(tmpList);

                if ((chFound == chCloseAngle)
                ||  (chFound == chForwardSlash)
                ||  fReaderMgr.getCurrentReader()->isWhitespace(chFound))
                {
                    //  Just fall through and process this attribute, though
                    //  the value will be "".
                }
                else if (chFound == chOpenAngle)
                {
                    // Assume a malformed tag and that new one is starting
                    emitError(XMLErrs::UnterminatedStartTag, elemName);
                    return attCount;
                }
                else
                {
                    // Something went really wrong
                    return attCount;
                }
            }

            //  Make sure that the name is basically well formed for namespace
            //  enabled rules. It either has no colons, or it has one which
            //  is neither the first or last char.
            const int colonFirst = XMLString::indexOf(fAttNameBuf.getRawBuffer(), chColon);
            if (colonFirst != -1)
            {
                const int colonLast = XMLString::lastIndexOf(fAttNameBuf.getRawBuffer(), chColon);

                if (colonFirst != colonLast)
                {
                    emitError(XMLErrs::TooManyColonsInName);
                    continue;
                }
                else if ((colonFirst == 0)
                      ||  (colonLast == (int)fAttNameBuf.getLen() - 1))
                {
                    emitError(XMLErrs::InvalidColonPos);
                    continue;
                }
            }

            //  And now lets add it to the passed collection. If we have not
            //  filled it up yet, then we use the next element. Else we add
            //  a new one.
            KVStringPair* curPair = 0;
            if (attCount >= curVecSize)
            {
                curPair = new (fMemoryManager) KVStringPair
                (
                    fAttNameBuf.getRawBuffer()
                    , fAttValueBuf.getRawBuffer()
                    , fMemoryManager
                );
                toFill.addElement(curPair);
            }
             else
            {
                curPair = toFill.elementAt(attCount);
                curPair->set(fAttNameBuf.getRawBuffer(), fAttValueBuf.getRawBuffer());
            }

            // And bump the count of attributes we've gotten
            attCount++;

            // And go to the top again for another attribute
            continue;
        }

        //  It was some special case character so do all of the checks and
        //  deal with it.
        if (!nextCh)
            ThrowXML(UnexpectedEOFException, XMLExcepts::Gen_UnexpectedEOF);

        if (nextCh == chForwardSlash)
        {
            fReaderMgr.getNextChar();
            isEmpty = true;
            if (!fReaderMgr.skippedChar(chCloseAngle))
                emitError(XMLErrs::UnterminatedStartTag, elemName);
            break;
        }
        else if (nextCh == chCloseAngle)
        {
            fReaderMgr.getNextChar();
            break;
        }
        else if (nextCh == chOpenAngle)
        {
            //  Check for this one specially, since its going to be common
            //  and it is kind of auto-recovering since we've already hit the
            //  next open bracket, which is what we would have seeked to (and
            //  skipped this whole tag.)
            emitError(XMLErrs::UnterminatedStartTag, elemName);
            break;
        }
        else if ((nextCh == chSingleQuote) || (nextCh == chDoubleQuote))
        {
            //  Check for this one specially, which is probably a missing
            //  attribute name, e.g. ="value". Just issue expected name
            //  error and eat the quoted string, then jump back to the
            //  top again.
            emitError(XMLErrs::ExpectedAttrName);
            fReaderMgr.getNextChar();
            fReaderMgr.skipQuotedString(nextCh);
            fReaderMgr.skipPastSpaces();
            continue;
        }
    }

    return attCount;
}


//  This method will kick off the scanning of the primary content of the
//  document, i.e. the elements.
bool IGXMLScanner::scanContent(const bool extEntity)
{
    //  Go into a loop until we hit the end of the root element, or we fall
    //  out because there is no root element.
    //
    //  We have to do kind of a deeply nested double loop here in order to
    //  avoid doing the setup/teardown of the exception handler on each
    //  round. Doing it this way we only do it when an exception actually
    //  occurs.
    bool gotData = true;
    bool inMarkup = false;
    while (gotData)
    {
        try
        {
            while (gotData)
            {
                //  Sense what the next top level token is. According to what
                //  this tells us, we will call something to handle that kind
                //  of thing.
                unsigned int orgReader;
                const XMLTokens curToken = senseNextToken(orgReader);

                //  Handle character data and end of file specially. Char data
                //  is not markup so we don't want to handle it in the loop
                //  below.
                if (curToken == Token_CharData)
                {
                    //  Scan the character data and call appropriate events. Let
                    //  him use our local character data buffer for efficiency.
                    scanCharData(fCDataBuf);
                    continue;
                }
                else if (curToken == Token_EOF)
                {
                    //  The element stack better be empty at this point or we
                    //  ended prematurely before all elements were closed.
                    if (!fElemStack.isEmpty())
                    {
                        const ElemStack::StackElem* topElem = fElemStack.popTop();
                        emitError
                        (
                            XMLErrs::EndedWithTagsOnStack
                            , topElem->fThisElement->getFullName()
                        );
                    }

                    // Its the end of file, so clear the got data flag
                    gotData = false;
                    continue;
                }

                // We are in some sort of markup now
                inMarkup = true;

                //  According to the token we got, call the appropriate
                //  scanning method.
                switch(curToken)
                {
                    case Token_CData :
                        // Make sure we are within content
                        if (fElemStack.isEmpty())
                            emitError(XMLErrs::CDATAOutsideOfContent);
                        scanCDSection();
                        break;

                    case Token_Comment :
                        scanComment();
                        break;

                    case Token_EndTag :
                        scanEndTag(gotData);
                        break;

                    case Token_PI :
                        scanPI();
                        break;

                    case Token_StartTag :
                        if (fDoNamespaces)
                            scanStartTagNS(gotData);
                        else
                            scanStartTag(gotData);
                        break;

                    default :
                        fReaderMgr.skipToChar(chOpenAngle);
                        break;
                }

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

                // And we are back out of markup again
                inMarkup = false;
            }
        }
        catch(const EndOfEntityException& toCatch)
        {
            //  If we were in some markup when this happened, then its a
            //  partial markup error.
            if (inMarkup)
                emitError(XMLErrs::PartialMarkupInEntity);

            // Send an end of entity reference event
            if (fDocHandler)
                fDocHandler->endEntityReference(toCatch.getEntity());

            inMarkup = false;
        }
    }

    // It went ok, so return success
    return true;
}


void IGXMLScanner::scanEndTag(bool& gotData)
{
    //  Assume we will still have data until proven otherwise. It will only
    //  ever be false if this is the end of the root element.
    gotData = true;

    //  Check if the element stack is empty. If so, then this is an unbalanced
    //  element (i.e. more ends than starts, perhaps because of bad text
    //  causing one to be skipped.)
    if (fElemStack.isEmpty())
    {
        emitError(XMLErrs::MoreEndThanStartTags);
        fReaderMgr.skipPastChar(chCloseAngle);
        ThrowXML(RuntimeException, XMLExcepts::Scan_UnbalancedStartEnd);
    }

    // After the </ is the element QName, so get a name from the input
    if (!fReaderMgr.getName(fQNameBuf))
    {
        // It failed so we can't really do anything with it
        emitError(XMLErrs::ExpectedElementName);
        fReaderMgr.skipPastChar(chCloseAngle);
        //REVISIT: Do we restore PSVI information?
        return;
    }

    unsigned int uriId = fEmptyNamespaceId;
    int prefixColonPos = -1;
    if (fDoNamespaces)
    {
        uriId = resolveQName
        (
            fQNameBuf.getRawBuffer()
            , fPrefixBuf
            , ElemStack::Mode_Element
            , prefixColonPos
        );
    }

    //  Pop the stack of the element we are supposed to be ending. Remember
    //  that we don't own this. The stack just keeps them and reuses them.
    //
    //  NOTE: We CANNOT do this until we've resolved the element name because
    //  the element stack top contains the prefix to URI mappings for this
    //  element.
    unsigned int topUri = fElemStack.getCurrentURI();
    const ElemStack::StackElem* topElem = fElemStack.popTop();

    // See if it was the root element, to avoid multiple calls below
    const bool isRoot = fElemStack.isEmpty();

    // Make sure that its the end of the element that we expect
    XMLElementDecl* tempElement = topElem->fThisElement;
    if (fDoNamespaces && fGrammarType == Grammar::SchemaGrammarType) {

        const XMLCh* rawNameBuf = fQNameBuf.getRawBuffer();
        if ((topUri != uriId) || (!XMLString::equals(tempElement->getBaseName(), &rawNameBuf[prefixColonPos + 1])))
        {
            emitError
            (
                XMLErrs::ExpectedEndOfTagX
                , topElem->fThisElement->getFullName()
            );
        }
    }
    else {
        if (!XMLString::equals(tempElement->getFullName(), fQNameBuf.getRawBuffer()))
        {
            emitError
            (
                XMLErrs::ExpectedEndOfTagX
                , topElem->fThisElement->getFullName()
            );
        }
    }

    // Make sure we are back on the same reader as where we started
    if (topElem->fReaderNum != fReaderMgr.getCurrentReaderNum())
        emitError(XMLErrs::PartialTagMarkupError);

    // Skip optional whitespace
    fReaderMgr.skipPastSpaces();

    // Make sure we find the closing bracket
    if (!fReaderMgr.skippedChar(chCloseAngle))
    {
        emitError
        (
            XMLErrs::UnterminatedEndTag
            , topElem->fThisElement->getFullName()
        );
    }

    if (fPSVIHandler && fGrammarType == Grammar::SchemaGrammarType)
    {
        if (fValidate && topElem->fThisElement->isDeclared())
        {
            fPSVIElemContext.fCurrentDV = ((SchemaValidator*) fValidator)->getCurrentDatatypeValidator();
            fPSVIElemContext.fCurrentTypeInfo = ((SchemaValidator*) fValidator)->getCurrentTypeInfo();
            fPSVIElemContext.fNormalizedValue = ((SchemaValidator*) fValidator)->getNormalizedValue();

            if (XMLString::equals(fPSVIElemContext.fNormalizedValue, XMLUni::fgZeroLenString))
                fPSVIElemContext.fNormalizedValue = 0;
        }
        else
        {
            fPSVIElemContext.fCurrentDV = 0;
            fPSVIElemContext.fCurrentTypeInfo = 0;
            fPSVIElemContext.fNormalizedValue = 0;
        }
    }

    //  If validation is enabled, then lets pass him the list of children and
    //  this element and let him validate it.
    DatatypeValidator* psviMemberType = 0;
    if (fValidate)
    {
        int res = fValidator->checkContent
        (
            topElem->fThisElement
            , topElem->fChildren
            , topElem->fChildCount
        );

        if (res >= 0)
        {
            //  One of the elements is not valid for the content. NOTE that
            //  if no children were provided but the content model requires
            //  them, it comes back with a zero value. But we cannot use that
            //  to index the child array in this case, and have to put out a
            //  special message.
            if (!topElem->fChildCount)
            {
                fValidator->emitError
                (
                    XMLValid::EmptyNotValidForContent
                    , topElem->fThisElement->getFormattedContentModel()
                );
            }
            else if ((unsigned int)res >= topElem->fChildCount)
            {
                fValidator->emitError
                (
                    XMLValid::NotEnoughElemsForCM
                    , topElem->fThisElement->getFormattedContentModel()
                );
            }
            else
            {
                fValidator->emitError
                (
                    XMLValid::ElementNotValidForContent
                    , topElem->fChildren[res]->getRawName()
                    , topElem->fThisElement->getFormattedContentModel()
                );
            }
        }


        if (fGrammarType == Grammar::SchemaGrammarType) {

            if (fPSVIHandler)
            {
                fPSVIElemContext.fIsSpecified = ((SchemaValidator*) fValidator)->getIsElemSpecified();
                if (((SchemaValidator*) fValidator)->getErrorOccurred())
                    fPSVIElemContext.fErrorOccurred = true;
                else if (fPSVIElemContext.fCurrentDV->getType() == DatatypeValidator::Union)
                        psviMemberType = fValidationContext->getValidatingMemberType();
            }

            // call matchers and de-activate context
            int oldCount = fMatcherStack->getMatcherCount();

            if (oldCount ||
                ((SchemaElementDecl*)topElem->fThisElement)->getIdentityConstraintCount()) {

                for (int i = oldCount - 1; i >= 0; i--) {

                    XPathMatcher* matcher = fMatcherStack->getMatcherAt(i);
                    matcher->endElement(*(topElem->fThisElement), fContent.getRawBuffer());
                }

                if (fMatcherStack->size() > 0) {
                    fMatcherStack->popContext();
                }

                // handle everything *but* keyref's.
                int newCount = fMatcherStack->getMatcherCount();

                for (int j = oldCount - 1; j >= newCount; j--) {

                    XPathMatcher* matcher = fMatcherStack->getMatcherAt(j);
                    IdentityConstraint* ic = matcher->getIdentityConstraint();

                    if (ic  && (ic->getType() != IdentityConstraint::KEYREF))
                        fValueStoreCache->transplant(ic, matcher->getInitialDepth());
                }

                // now handle keyref's...
                for (int k = oldCount - 1; k >= newCount; k--) {

                    XPathMatcher* matcher = fMatcherStack->getMatcherAt(k);
                    IdentityConstraint* ic = matcher->getIdentityConstraint();

                    if (ic && (ic->getType() == IdentityConstraint::KEYREF)) {

                        ValueStore* values = fValueStoreCache->getValueStoreFor(ic, matcher->getInitialDepth());

                        if (values) { // nothing to do if nothing matched!
                            values->endDcocumentFragment(fValueStoreCache);
                        }
                    }
                }

                fValueStoreCache->endElement();
            }
        }
    }

    if (fGrammarType == Grammar::SchemaGrammarType)
    {
        if(!isRoot)
            ((SchemaElementDecl *)fElemStack.topElement()->fThisElement)->updateValidityFromElement(topElem->fThisElement, fGrammarType);

        if (fPSVIHandler)
        {
            endElementPSVI(
                (SchemaElementDecl*)topElem->fThisElement, psviMemberType);
        }
    }

    // If we have a doc handler, tell it about the end tag
    if (fDocHandler)
    {
        fDocHandler->endElement
        (
            *topElem->fThisElement
            , uriId
            , isRoot
            , fPrefixBuf.getRawBuffer()
        );
    }

    // reset xsi:type ComplexTypeInfo
    if (fGrammarType == Grammar::SchemaGrammarType) {        
        ((SchemaElementDecl*)topElem->fThisElement)->reset();
        if (!isRoot)
            ((SchemaElementDecl*)(fElemStack.topElement()->fThisElement))->
                setXsiComplexTypeInfo(((SchemaValidator*)fValidator)->getCurrentTypeInfo());
    }

    // If this was the root, then done with content
    gotData = !isRoot;

    if (gotData) {
        if (fDoNamespaces) {
            // Restore the grammar
            fGrammar = fElemStack.getCurrentGrammar();
            fGrammarType = fGrammar->getGrammarType();
            if (fGrammarType == Grammar::SchemaGrammarType && !fValidator->handlesSchema()) {
                if (fValidatorFromUser)
                    ThrowXML(RuntimeException, XMLExcepts::Gen_NoSchemaValidator);
                else {
                    fValidator = fSchemaValidator;
                }
            }
            else if (fGrammarType == Grammar::DTDGrammarType && !fValidator->handlesDTD()) {
                if (fValidatorFromUser)
                    ThrowXML(RuntimeException, XMLExcepts::Gen_NoDTDValidator);
                else {
                    fValidator = fDTDValidator;
                }
            }

            fValidator->setGrammar(fGrammar);
        }

        // Restore the validation flag
        fValidate = fElemStack.getValidationFlag();
    }
}


//  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 IGXMLScanner::scanDocTypeDecl()
{
    //  We have a doc type. So, switch the Grammar.
    switchGrammar(XMLUni::fgDTDEntityString);

    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
    ); 
    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));
    } else
    {
        // attach this to the undeclared element pool so that it gets deleted
        rootDecl->setId(fDTDElemNonDeclPool->put((DTDElementDecl*)rootDecl));
    }

    // 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();

        // We can't have any internal subset if we are reusing the validator
        if (fUseCachedGrammar || fToCacheGrammar)
            ThrowXML(RuntimeException, XMLExcepts::Val_CantHaveIntSS);

        //  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…

Large files files are truncated, but you can click here to view the full file