http://xml.apache.org/http://www.apache.org/http://www.w3.org/

Home

Readme
Release Info

Installation
Download
Build

FAQs
Samples
API Docs

DOM C++ Binding
Programming
Migration Guide

Feedback
Bug-Reporting
PDF Document

CVS Repository
Mail Archive

API Docs for SAX and DOM
 

Main Page   Class Hierarchy   Alphabetical List   Compound List   File List   Compound Members   File Members  

AbstractDOMParser.hpp

Go to the documentation of this file.
00001 /*
00002  * The Apache Software License, Version 1.1
00003  *
00004  * Copyright (c) 2002 The Apache Software Foundation.  All rights
00005  * reserved.
00006  *
00007  * Redistribution and use in source and binary forms, with or without
00008  * modification, are permitted provided that the following conditions
00009  * are met:
00010  *
00011  * 1. Redistributions of source code must retain the above copyright
00012  *    notice, this list of conditions and the following disclaimer.
00013  *
00014  * 2. Redistributions in binary form must reproduce the above copyright
00015  *    notice, this list of conditions and the following disclaimer in
00016  *    the documentation and/or other materials provided with the
00017  *    distribution.
00018  *
00019  * 3. The end-user documentation included with the redistribution,
00020  *    if any, must include the following acknowledgment:
00021  *       "This product includes software developed by the
00022  *        Apache Software Foundation (http://www.apache.org/)."
00023  *    Alternately, this acknowledgment may appear in the software itself,
00024  *    if and wherever such third-party acknowledgments normally appear.
00025  *
00026  * 4. The names "Xerces" and "Apache Software Foundation" must
00027  *    not be used to endorse or promote products derived from this
00028  *    software without prior written permission. For written
00029  *    permission, please contact apache\@apache.org.
00030  *
00031  * 5. Products derived from this software may not be called "Apache",
00032  *    nor may "Apache" appear in their name, without prior written
00033  *    permission of the Apache Software Foundation.
00034  *
00035  * THIS SOFTWARE IS PROVIDED ``AS IS'' AND ANY EXPRESSED OR IMPLIED
00036  * WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
00037  * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
00038  * DISCLAIMED.  IN NO EVENT SHALL THE APACHE SOFTWARE FOUNDATION OR
00039  * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
00040  * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
00041  * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF
00042  * USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
00043  * ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY,
00044  * OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT
00045  * OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
00046  * SUCH DAMAGE.
00047  * ====================================================================
00048  *
00049  * This software consists of voluntary contributions made by many
00050  * individuals on behalf of the Apache Software Foundation, and was
00051  * originally based on software copyright (c) 2001, International
00052  * Business Machines, Inc., http://www.ibm.com .  For more information
00053  * on the Apache Software Foundation, please see
00054  * <http://www.apache.org/>.
00055  */
00056 
00057 /*
00058  * $Id: AbstractDOMParser.hpp,v 1.16 2003/01/03 20:09:36 tng Exp $
00059  *
00060  */
00061 #if !defined(ABSTRACTDOMPARSER_HPP)
00062 #define ABSTRACTDOMPARSER_HPP
00063 
00064 #include <xercesc/dom/DOMDocument.hpp>
00065 #include <xercesc/framework/XMLDocumentHandler.hpp>
00066 #include <xercesc/framework/XMLErrorReporter.hpp>
00067 #include <xercesc/framework/XMLEntityHandler.hpp>
00068 #include <xercesc/util/ValueStackOf.hpp>
00069 #include <xercesc/validators/DTD/DocTypeHandler.hpp>
00070 #include <xercesc/dom/DOMDocumentType.hpp>
00071 #include <xercesc/validators/DTD/DTDElementDecl.hpp>
00072 #include <xercesc/framework/XMLBufferMgr.hpp>
00073 
00074 
00075 XERCES_CPP_NAMESPACE_BEGIN
00076 
00077 class XMLPScanToken;
00078 class XMLScanner;
00079 class XMLValidator;
00080 class DOMDocumentImpl;
00081 class DOMDocumentTypeImpl;
00082 class DOMElement;
00083 class GrammarResolver;
00084 
00085 
00090 class  AbstractDOMParser :
00091 
00092     public XMLDocumentHandler
00093     , public XMLErrorReporter
00094     , public XMLEntityHandler
00095     , public DocTypeHandler
00096 {
00097 public :
00098     // -----------------------------------------------------------------------
00099     //  Class types
00100     // -----------------------------------------------------------------------
00103 
00111     enum ValSchemes
00112     {
00113         Val_Never
00114         , Val_Always
00115         , Val_Auto
00116     };
00118 
00119 
00120     // -----------------------------------------------------------------------
00121     //  Constructors and Detructor
00122     // -----------------------------------------------------------------------
00125 
00129     virtual ~AbstractDOMParser();
00130 
00132 
00133     // -----------------------------------------------------------------------
00134     //  Utility methods
00135     // -----------------------------------------------------------------------
00136 
00144     void reset();
00145 
00159     DOMDocument* adoptDocument();
00160 
00162 
00163 
00164     // -----------------------------------------------------------------------
00165     //  Getter methods
00166     // -----------------------------------------------------------------------
00167 
00170 
00182     DOMDocument* getDocument();
00183 
00191     const XMLValidator& getValidator() const;
00192 
00200     ValSchemes getValidationScheme() const;
00201 
00212     bool getDoSchema() const;
00213 
00224     bool getValidationSchemaFullChecking() const;
00225 
00237     int getErrorCount() const;
00238 
00249     bool getDoNamespaces() const;
00250 
00263     bool getExitOnFirstFatalError() const;
00264 
00275     bool getValidationConstraintFatal() const;
00276 
00286     bool  getCreateEntityReferenceNodes()const;
00287 
00298     bool getIncludeIgnorableWhitespace() const;
00299 
00319     XMLCh* getExternalSchemaLocation() const;
00320 
00340     XMLCh* getExternalNoNamespaceSchemaLocation() const;
00341 
00353     bool getLoadExternalDTD() const;
00354 
00363     bool  getCreateCommentNodes()const;
00364 
00376     bool getCalculateSrcOfs() const;
00377 
00388     bool getStandardUriConformant() const;
00389 
00391 
00392 
00393     // -----------------------------------------------------------------------
00394     //  Setter methods
00395     // -----------------------------------------------------------------------
00396 
00399 
00414     void setDoNamespaces(const bool newState);
00415 
00432     void setExitOnFirstFatalError(const bool newState);
00433 
00453     void setValidationConstraintFatal(const bool newState);
00454 
00473     void setCreateEntityReferenceNodes(const bool create);
00474 
00496     void setIncludeIgnorableWhitespace(const bool include);
00497 
00514     void setValidationScheme(const ValSchemes newScheme);
00515 
00531     void setDoSchema(const bool newState);
00532 
00549     void setValidationSchemaFullChecking(const bool schemaFullChecking);
00550 
00571     void setExternalSchemaLocation(const XMLCh* const schemaLocation);
00572 
00581     void setExternalSchemaLocation(const char* const schemaLocation);
00582 
00597     void setExternalNoNamespaceSchemaLocation(const XMLCh* const noNamespaceSchemaLocation);
00598 
00607     void setExternalNoNamespaceSchemaLocation(const char* const noNamespaceSchemaLocation);
00608 
00625     void setLoadExternalDTD(const bool newState);
00626 
00637     void setCreateCommentNodes(const bool create);
00638 
00651     void setCalculateSrcOfs(const bool newState);
00652 
00663     void setStandardUriConformant(const bool newState);
00664 
00672     void useScanner(const XMLCh* const scannerName);
00673 
00675 
00676 
00677     // -----------------------------------------------------------------------
00678     //  Parsing methods
00679     // -----------------------------------------------------------------------
00680 
00683 
00699     void parse(const InputSource& source);
00700 
00717     void parse(const XMLCh* const systemId);
00718 
00734     void parse(const char* const systemId);
00735 
00762     bool parseFirst
00763     (
00764         const   XMLCh* const    systemId
00765         ,       XMLPScanToken&  toFill
00766     );
00767 
00795     bool parseFirst
00796     (
00797         const   char* const     systemId
00798         ,       XMLPScanToken&  toFill
00799     );
00800 
00828     bool parseFirst
00829     (
00830         const   InputSource&    source
00831         ,       XMLPScanToken&  toFill
00832     );
00833 
00856     bool parseNext(XMLPScanToken& token);
00857 
00883     void parseReset(XMLPScanToken& token);
00884 
00886 
00887     // -----------------------------------------------------------------------
00888     //  Implementation of the XMLDocumentHandler interface.
00889     // -----------------------------------------------------------------------
00890 
00893 
00906     virtual void docCharacters
00907     (
00908         const   XMLCh* const    chars
00909         , const unsigned int    length
00910         , const bool            cdataSection
00911     );
00912 
00921     virtual void docComment
00922     (
00923         const   XMLCh* const    comment
00924     );
00925 
00938     virtual void docPI
00939     (
00940         const   XMLCh* const    target
00941         , const XMLCh* const    data
00942     );
00943 
00948     virtual void endDocument();
00949 
00966     virtual void endElement
00967     (
00968         const   XMLElementDecl& elemDecl
00969         , const unsigned int    urlId
00970         , const bool            isRoot
00971         , const XMLCh* const    elemPrefix
00972     );
00973 
00982     virtual void endEntityReference
00983     (
00984         const   XMLEntityDecl&  entDecl
00985     );
00986 
01005     virtual void ignorableWhitespace
01006     (
01007         const   XMLCh* const    chars
01008         , const unsigned int    length
01009         , const bool            cdataSection
01010     );
01011 
01018     virtual void resetDocument();
01019 
01024     virtual void startDocument();
01025 
01053     virtual void startElement
01054     (
01055         const   XMLElementDecl&         elemDecl
01056         , const unsigned int            urlId
01057         , const XMLCh* const            elemPrefix
01058         , const RefVectorOf<XMLAttr>&   attrList
01059         , const unsigned int            attrCount
01060         , const bool                    isEmpty
01061         , const bool                    isRoot
01062     );
01063 
01073     virtual void startEntityReference
01074     (
01075         const   XMLEntityDecl&  entDecl
01076     );
01077 
01096     virtual void XMLDecl
01097     (
01098         const   XMLCh* const    versionStr
01099         , const XMLCh* const    encodingStr
01100         , const XMLCh* const    standaloneStr
01101         , const XMLCh* const    actualEncStr
01102     );
01104 
01105 
01106     // -----------------------------------------------------------------------
01107     //  Implementation of the deprecated DocTypeHandler interface.
01108     // -----------------------------------------------------------------------
01111     virtual void attDef
01112     (
01113         const   DTDElementDecl&     elemDecl
01114         , const DTDAttDef&          attDef
01115         , const bool                ignoring
01116     );
01117 
01118     virtual void doctypeComment
01119     (
01120         const   XMLCh* const    comment
01121     );
01122 
01123     virtual void doctypeDecl
01124     (
01125         const   DTDElementDecl& elemDecl
01126         , const XMLCh* const    publicId
01127         , const XMLCh* const    systemId
01128         , const bool            hasIntSubset
01129         , const bool            hasExtSubset = false
01130     );
01131 
01132     virtual void doctypePI
01133     (
01134         const   XMLCh* const    target
01135         , const XMLCh* const    data
01136     );
01137 
01138     virtual void doctypeWhitespace
01139     (
01140         const   XMLCh* const    chars
01141         , const unsigned int    length
01142     );
01143 
01144     virtual void elementDecl
01145     (
01146         const   DTDElementDecl& decl
01147         , const bool            isIgnored
01148     );
01149 
01150     virtual void endAttList
01151     (
01152         const   DTDElementDecl& elemDecl
01153     );
01154 
01155     virtual void endIntSubset();
01156 
01157     virtual void endExtSubset();
01158 
01159     virtual void entityDecl
01160     (
01161         const   DTDEntityDecl&  entityDecl
01162         , const bool            isPEDecl
01163         , const bool            isIgnored
01164     );
01165 
01166     virtual void resetDocType();
01167 
01168     virtual void notationDecl
01169     (
01170         const   XMLNotationDecl&    notDecl
01171         , const bool                isIgnored
01172     );
01173 
01174     virtual void startAttList
01175     (
01176         const   DTDElementDecl& elemDecl
01177     );
01178 
01179     virtual void startIntSubset();
01180 
01181     virtual void startExtSubset();
01182 
01183     virtual void TextDecl
01184     (
01185         const   XMLCh* const    versionStr
01186         , const XMLCh* const    encodingStr
01187     );
01188 
01189 
01191 
01192 
01193     // -----------------------------------------------------------------------
01194     //  Deprecated Methods
01195     // -----------------------------------------------------------------------
01208     bool getDoValidation() const;
01209 
01223     void setDoValidation(const bool newState);
01224 
01238     bool getExpandEntityReferences() const;
01239 
01256     void setExpandEntityReferences(const bool expand);
01257 
01259 
01260 protected :
01261     // -----------------------------------------------------------------------
01262     //  Protected Constructor Methods
01263     // -----------------------------------------------------------------------
01275     AbstractDOMParser(XMLValidator* const valToAdopt = 0);
01276 
01278 
01279     // -----------------------------------------------------------------------
01280     //  Protected getter methods
01281     // -----------------------------------------------------------------------
01289     DOMNode* getCurrentNode();
01290 
01295     XMLScanner* getScanner() const;
01296 
01301     GrammarResolver* getGrammarResolver() const;
01302 
01308     bool getParseInProgress() const;
01309 
01311 
01312 
01313     // -----------------------------------------------------------------------
01314     //  Protected setter methods
01315     // -----------------------------------------------------------------------
01316 
01319 
01327     void setCurrentNode(DOMNode* toSet);
01328 
01335     void setDocument(DOMDocument* toSet);
01336 
01343     void setParseInProgress(const bool toSet);
01345 
01346     // -----------------------------------------------------------------------
01347     //  Protected Helper methods
01348     // -----------------------------------------------------------------------
01351     virtual DOMElement* createElementNSNode(const XMLCh *fNamespaceURI,
01352                                               const XMLCh *qualifiedName);
01353 
01354     void resetPool();
01355 
01359     bool isDocumentAdopted() const;
01360 
01362 
01363 
01364 private :
01365     // -----------------------------------------------------------------------
01366     //  Initialize/Cleanup methods
01367     // -----------------------------------------------------------------------
01368     void initialize();
01369     void cleanUp();
01370 
01371     // -----------------------------------------------------------------------
01372     //  Private data members
01373     //
01374     //  fCurrentNode
01375     //  fCurrentParent
01376     //      Used to track the current node during nested element events. Since
01377     //      the tree must be built from a set of disjoint callbacks, we need
01378     //      these to keep up with where we currently are.
01379     //
01380     //  fCurrentEntity
01381     //      Used to track the current entity decl.  If a text decl is seen later on,
01382     //      it is used to update the encoding and version information.
01383     //
01384     //  fDocument
01385     //      The root document object, filled with the document contents.
01386     //
01387     //  fCreateEntityReferenceNodes
01388     //      Indicates whether entity reference nodes should be created.
01389     //
01390     //  fIncludeIgnorableWhitespace
01391     //      Indicates whether ignorable whiltespace should be added to
01392     //      the DOM tree for validating parsers.
01393     //
01394     //  fScanner
01395     //      The scanner used for this parser. This is created during the
01396     //      constructor.
01397     //
01398     //  fNodeStack
01399     //      Used to track previous parent nodes during nested element events.
01400     //
01401     //  fParseInProgress
01402     //      Used to prevent multiple entrance to the parser while its doing
01403     //      a parse.
01404     //
01405     //  fWithinElement
01406     //      A flag to indicate that the parser is within at least one level
01407     //      of element processing.
01408     //
01409     //  fDocumentType
01410     //      Used to store and update the documentType variable information
01411     //      in fDocument
01412     //
01413     //  fDocumentVector
01414     //      Store all the previous fDocument(s) (thus not the current fDocument)
01415     //      created in this parser.  It is destroyed when the parser is destructed.
01416     //
01417     //  fCreateCommentNodes
01418     //      Indicates whether comment nodes should be created.
01419     //
01420     //  fDocumentAdoptedByUser
01421     //      The DOMDocument ownership has been transferred to application
01422     //      If set to true, the parser does not own the document anymore
01423     //      and thus will not release its memory.
01424     //
01425     //  fInternalSubset
01426     //      Buffer for storing the internal subset information.
01427     //      Once complete (after DOCTYPE is finished scanning), send
01428     //      it to DocumentType Node
01429     // -----------------------------------------------------------------------
01430     bool                          fCreateEntityReferenceNodes;
01431     bool                          fIncludeIgnorableWhitespace;
01432     bool                          fWithinElement;
01433     bool                          fParseInProgress;
01434     bool                          fCreateCommentNodes;
01435     bool                          fDocumentAdoptedByUser;
01436     XMLBufferMgr                  fBufMgr;
01437     XMLScanner*                   fScanner;
01438     DOMNode*                      fCurrentParent;
01439     DOMNode*                      fCurrentNode;
01440     DOMEntity*                    fCurrentEntity;
01441     DOMDocumentImpl*              fDocument;
01442     ValueStackOf<DOMNode*>*       fNodeStack;
01443     DOMDocumentTypeImpl*          fDocumentType;
01444     RefVectorOf<DOMDocumentImpl>* fDocumentVector;
01445     GrammarResolver*              fGrammarResolver;
01446     XMLStringPool*                fURIStringPool;
01447     XMLValidator*                 fValidator;
01448     XMLBuffer&                    fInternalSubset;
01449 };
01450 
01451 
01452 
01453 // ---------------------------------------------------------------------------
01454 //  AbstractDOMParser: Getter methods
01455 // ---------------------------------------------------------------------------
01456 inline bool AbstractDOMParser::getExpandEntityReferences() const
01457 {
01458     return !fCreateEntityReferenceNodes;
01459 }
01460 inline bool AbstractDOMParser::getCreateEntityReferenceNodes() const
01461 {
01462     return fCreateEntityReferenceNodes;
01463 }
01464 
01465 inline bool AbstractDOMParser::getIncludeIgnorableWhitespace() const
01466 {
01467     return fIncludeIgnorableWhitespace;
01468 }
01469 
01470 inline bool AbstractDOMParser::getParseInProgress() const
01471 {
01472     return fParseInProgress;
01473 }
01474 
01475 inline XMLScanner* AbstractDOMParser::getScanner() const
01476 {
01477     return fScanner;
01478 }
01479 
01480 inline GrammarResolver* AbstractDOMParser::getGrammarResolver() const
01481 {
01482     return fGrammarResolver;
01483 }
01484 
01485 inline bool AbstractDOMParser::getCreateCommentNodes() const
01486 {
01487     return fCreateCommentNodes;
01488 }
01489 
01490 // ---------------------------------------------------------------------------
01491 //  AbstractDOMParser: Setter methods
01492 // ---------------------------------------------------------------------------
01493 inline void AbstractDOMParser::setExpandEntityReferences(const bool expand)
01494 {
01495     fCreateEntityReferenceNodes = !expand;
01496 }
01497 
01498 inline void AbstractDOMParser::setCreateEntityReferenceNodes(const bool create)
01499 {
01500     fCreateEntityReferenceNodes = create;
01501 }
01502 
01503 inline void AbstractDOMParser::setIncludeIgnorableWhitespace(const bool include)
01504 {
01505     fIncludeIgnorableWhitespace = include;
01506 }
01507 
01508 inline void AbstractDOMParser::setCreateCommentNodes(const bool create)
01509 {
01510     fCreateCommentNodes = create;
01511 }
01512 
01513 
01514 // ---------------------------------------------------------------------------
01515 //  AbstractDOMParser: Protected getter methods
01516 // ---------------------------------------------------------------------------
01517 inline DOMNode* AbstractDOMParser::getCurrentNode()
01518 {
01519     return fCurrentNode;
01520 }
01521 
01522 
01523 // ---------------------------------------------------------------------------
01524 //  AbstractDOMParser: Protected setter methods
01525 // ---------------------------------------------------------------------------
01526 inline void AbstractDOMParser::setCurrentNode(DOMNode* toSet)
01527 {
01528     fCurrentNode = toSet;
01529 }
01530 
01531 inline void AbstractDOMParser::setDocument(DOMDocument* toSet)
01532 {
01533     fDocument = (DOMDocumentImpl *)toSet;
01534 }
01535 
01536 inline void AbstractDOMParser::setParseInProgress(const bool toSet)
01537 {
01538     fParseInProgress = toSet;
01539 }
01540 
01541 XERCES_CPP_NAMESPACE_END
01542 
01543 #endif
01544 
01545 
01546 


Copyright © 2000 The Apache Software Foundation. All Rights Reserved.