API Docs for SAX and DOM                 DOMParser.hpp Go to the documentation of this file. 00001 /* 00002 * The Apache Software License, Version 1.1 00003 * 00004 * Copyright (c) 1999-2000 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) 1999, 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: DOMParser.hpp,v 1.15 2001/01/25 23:59:15 lehors Exp $ 00059 * 00060 */ 00061 00062 #if !defined(DOMPARSER_HPP) 00063 #define DOMPARSER_HPP 00064 00065 00066 #include <dom/DOM_Document.hpp> 00067 #include <framework/XMLDocumentHandler.hpp> 00068 #include <framework/XMLErrorReporter.hpp> 00069 #include <framework/XMLEntityHandler.hpp> 00070 #include <util/ValueStackOf.hpp> 00071 00072 #include <validators/DTD/DocTypeHandler.hpp> 00073 #include <dom/DOM_DocumentType.hpp> 00074 #include <validators/DTD/DTDElementDecl.hpp> 00075 #include <validators/DTD/DTDValidator.hpp> 00076 00077 class EntityResolver; 00078 class ErrorHandler; 00079 class XMLPScanToken; 00080 class XMLScanner; 00081 class XMLValidator; 00082 00083 00094 class DOMParser : 00095 00096 public XMLDocumentHandler 00097 , public XMLErrorReporter 00098 , public XMLEntityHandler 00099 , public DocTypeHandler 00100 { 00101 public : 00102 // ----------------------------------------------------------------------- 00103 // Class types 00104 // ----------------------------------------------------------------------- 00105 enum ValSchemes 00106 { 00107 Val_Never 00108 , Val_Always 00109 , Val_Auto 00110 }; 00111 00112 00113 // ----------------------------------------------------------------------- 00114 // Constructors and Detructor 00115 // ----------------------------------------------------------------------- 00116 00119 00128 DOMParser(XMLValidator* const valToAdopt = 0); 00129 00133 ~DOMParser(); 00134 00136 00142 void reset(); 00143 00144 00145 // ----------------------------------------------------------------------- 00146 // Getter methods 00147 // ----------------------------------------------------------------------- 00148 00151 00161 DOM_Document getDocument(); 00162 00170 ErrorHandler* getErrorHandler(); 00171 00179 const ErrorHandler* getErrorHandler() const; 00180 00188 EntityResolver* getEntityResolver(); 00189 00197 const EntityResolver* getEntityResolver() const; 00198 00206 const XMLScanner& getScanner() const; 00207 00215 const XMLValidator& getValidator() const; 00216 00224 ValSchemes getValidationScheme() const; 00225 00236 bool getDoNamespaces() const; 00237 00250 bool getExitOnFirstFatalError() const; 00251 00263 bool getExpandEntityReferences() const; 00264 00282 bool getCreateEntityReferenceNodes()const; 00283 00294 bool getIncludeIgnorableWhitespace() const; 00295 00305 bool getToCreateXMLDeclTypeNode() const; 00306 00307 00309 00310 00311 // ----------------------------------------------------------------------- 00312 // Setter methods 00313 // ----------------------------------------------------------------------- 00314 00317 00331 void setErrorHandler(ErrorHandler* const handler); 00332 00348 void setEntityResolver(EntityResolver* const handler); 00349 00368 void setDoNamespaces(const bool newState); 00369 00386 void setExitOnFirstFatalError(const bool newState); 00387 00402 void setExpandEntityReferences(const bool expand); 00403 00420 void setCreateEntityReferenceNodes(const bool create); 00421 00440 void setIncludeIgnorableWhitespace(const bool include); 00441 00458 void setValidationScheme(const ValSchemes newScheme); 00459 00470 void setToCreateXMLDeclTypeNode(const bool create); 00471 00473 00474 00475 // ----------------------------------------------------------------------- 00476 // Parsing methods 00477 // ----------------------------------------------------------------------- 00478 00481 00495 void parse(const InputSource& source, const bool reuseValidator = false); 00496 00511 void parse(const XMLCh* const systemId, const bool reuseValidator = false); 00512 00524 void parse(const char* const systemId, const bool reuseValidator = false); 00525 00555 bool parseFirst 00556 ( 00557 const XMLCh* const systemId 00558 , XMLPScanToken& toFill 00559 , const bool reuseValidator = false 00560 ); 00561 00592 bool parseFirst 00593 ( 00594 const char* const systemId 00595 , XMLPScanToken& toFill 00596 , const bool reuseValidator = false 00597 ); 00598 00629 bool parseFirst 00630 ( 00631 const InputSource& source 00632 , XMLPScanToken& toFill 00633 , const bool reuseValidator = false 00634 ); 00635 00658 bool parseNext(XMLPScanToken& token); 00659 00685 void parseReset(XMLPScanToken& token); 00686 00688 00689 00690 00691 // ----------------------------------------------------------------------- 00692 // Implementation of the XMLErrorReporter interface. 00693 // ----------------------------------------------------------------------- 00694 00697 00722 virtual void error 00723 ( 00724 const unsigned int errCode 00725 , const XMLCh* const msgDomain 00726 , const XMLErrorReporter::ErrTypes errType 00727 , const XMLCh* const errorText 00728 , const XMLCh* const systemId 00729 , const XMLCh* const publicId 00730 , const unsigned int lineNum 00731 , const unsigned int colNum 00732 ); 00733 00742 virtual void resetErrors(); 00744 00745 00746 // ----------------------------------------------------------------------- 00747 // Implementation of the XMLEntityHandler interface. 00748 // ----------------------------------------------------------------------- 00749 00752 00765 virtual void endInputSource(const InputSource& inputSource); 00766 00782 virtual bool expandSystemId 00783 ( 00784 const XMLCh* const systemId 00785 , XMLBuffer& toFill 00786 ); 00787 00796 virtual void resetEntities(); 00797 00813 virtual InputSource* resolveEntity 00814 ( 00815 const XMLCh* const publicId 00816 , const XMLCh* const systemId 00817 ); 00818 00831 virtual void startInputSource(const InputSource& inputSource); 00832 00834 00835 00836 00837 // ----------------------------------------------------------------------- 00838 // Implementation of the XMLDocumentHandler interface. 00839 // ----------------------------------------------------------------------- 00840 00843 00856 virtual void docCharacters 00857 ( 00858 const XMLCh* const chars 00859 , const unsigned int length 00860 , const bool cdataSection 00861 ); 00862 00871 virtual void docComment 00872 ( 00873 const XMLCh* const comment 00874 ); 00875 00888 virtual void docPI 00889 ( 00890 const XMLCh* const target 00891 , const XMLCh* const data 00892 ); 00893 00898 virtual void endDocument(); 00899 00913 virtual void endElement 00914 ( 00915 const XMLElementDecl& elemDecl 00916 , const unsigned int urlId 00917 , const bool isRoot 00918 ); 00919 00928 virtual void endEntityReference 00929 ( 00930 const XMLEntityDecl& entDecl 00931 ); 00932 00951 virtual void ignorableWhitespace 00952 ( 00953 const XMLCh* const chars 00954 , const unsigned int length 00955 , const bool cdataSection 00956 ); 00957 00964 virtual void resetDocument(); 00965 00970 virtual void startDocument(); 00971 00999 virtual void startElement 01000 ( 01001 const XMLElementDecl& elemDecl 01002 , const unsigned int urlId 01003 , const XMLCh* const elemPrefix 01004 , const RefVectorOf<XMLAttr>& attrList 01005 , const unsigned int attrCount 01006 , const bool isEmpty 01007 , const bool isRoot 01008 ); 01009 01019 virtual void startEntityReference 01020 ( 01021 const XMLEntityDecl& entDecl 01022 ); 01023 01042 virtual void XMLDecl 01043 ( 01044 const XMLCh* const versionStr 01045 , const XMLCh* const encodingStr 01046 , const XMLCh* const standaloneStr 01047 , const XMLCh* const actualEncStr 01048 ); 01050 01051 01054 01064 bool getDoValidation() const; 01065 01079 void setDoValidation(const bool newState); 01080 //doctypehandler interfaces 01081 virtual void attDef 01082 ( 01083 const DTDElementDecl& elemDecl 01084 , const DTDAttDef& attDef 01085 , const bool ignoring 01086 ); 01087 01088 virtual void doctypeComment 01089 ( 01090 const XMLCh* const comment 01091 ); 01092 01093 virtual void doctypeDecl 01094 ( 01095 const DTDElementDecl& elemDecl 01096 , const XMLCh* const publicId 01097 , const XMLCh* const systemId 01098 , const bool hasIntSubset 01099 ); 01100 01101 virtual void doctypePI 01102 ( 01103 const XMLCh* const target 01104 , const XMLCh* const data 01105 ); 01106 01107 virtual void doctypeWhitespace 01108 ( 01109 const XMLCh* const chars 01110 , const unsigned int length 01111 ); 01112 01113 virtual void elementDecl 01114 ( 01115 const DTDElementDecl& decl 01116 , const bool isIgnored 01117 ); 01118 01119 virtual void endAttList 01120 ( 01121 const DTDElementDecl& elemDecl 01122 ); 01123 01124 virtual void endIntSubset(); 01125 01126 virtual void endExtSubset(); 01127 01128 virtual void entityDecl 01129 ( 01130 const DTDEntityDecl& entityDecl 01131 , const bool isPEDecl 01132 , const bool isIgnored 01133 ); 01134 01135 virtual void resetDocType(); 01136 01137 virtual void notationDecl 01138 ( 01139 const XMLNotationDecl& notDecl 01140 , const bool isIgnored 01141 ); 01142 01143 virtual void startAttList 01144 ( 01145 const DTDElementDecl& elemDecl 01146 ); 01147 01148 virtual void startIntSubset(); 01149 01150 virtual void startExtSubset(); 01151 01152 virtual void TextDecl 01153 ( 01154 const XMLCh* const versionStr 01155 , const XMLCh* const encodingStr 01156 ); 01157 01158 01160 01161 01162 protected : 01163 // ----------------------------------------------------------------------- 01164 // Protected getter methods 01165 // ----------------------------------------------------------------------- 01166 01169 01174 DOM_Node getCurrentNode(); 01175 01177 01178 01179 // ----------------------------------------------------------------------- 01180 // Protected setter methods 01181 // ----------------------------------------------------------------------- 01182 01185 01193 void setCurrentNode(DOM_Node toSet); 01194 01201 void setDocument(DOM_Document toSet); 01203 01204 01205 private : 01206 //local private function to populate the doctype data 01207 virtual void populateDocumentType(); 01208 01209 // ----------------------------------------------------------------------- 01210 // Private data members 01211 // 01212 // fCurrentNode 01213 // fCurrentParent 01214 // Used to track the current node during nested element events. Since 01215 // the tree must be built from a set of disjoint callbacks, we need 01216 // these to keep up with where we currently are. 01217 // 01218 // fDocument 01219 // The root document object, filled with the document contents. 01220 // 01221 // fEntityResolver 01222 // The installed SAX entity resolver, if any. Null if none. 01223 // 01224 // fErrorHandler 01225 // The installed SAX error handler, if any. Null if none. 01226 // 01227 // fCreateEntityReferenceNode 01228 // Indicates whether entity reference nodes should be created. 01229 // 01230 // fIncludeIgnorableWhitespace 01231 // Indicates whether ignorable whiltespace should be added to 01232 // the DOM tree for validating parsers. 01233 // 01234 // fNodeStack 01235 // Used to track previous parent nodes during nested element events. 01236 // 01237 // fParseInProgress 01238 // Used to prevent multiple entrance to the parser while its doing 01239 // a parse. 01240 // 01241 // fScanner 01242 // The scanner used for this parser. This is created during the 01243 // constructor. 01244 // 01245 // fValidator 01246 // The validator that is installed. If none is provided, we will 01247 // create and install a DTD validator. We install this on the 01248 // scanner we create, which it will use to do validation. We set 01249 // ourself on it as the error reporter for validity errors. 01250 // 01251 // fWithinElement 01252 // A flag to indicate that the parser is within at least one level 01253 // of element processing. 01254 // 01255 // fDocumentType 01256 // Used to store and update the documentType variable information 01257 // in fDocument 01258 // 01259 // fOldDocTypeHandler 01260 // Used to chain the old documentType node if the user has set it 01261 // from outside 01262 // 01263 // fToCreateXMLDecTypeNode 01264 // A flag to create a DOM_XMLDecl node in the ODM tree if it exists 01265 // This is an extension to xerces implementation 01266 // 01267 // ----------------------------------------------------------------------- 01268 DOM_Node fCurrentParent; 01269 DOM_Node fCurrentNode; 01270 DOM_Document fDocument; 01271 EntityResolver* fEntityResolver; 01272 ErrorHandler* fErrorHandler; 01273 bool fCreateEntityReferenceNodes; 01274 bool fIncludeIgnorableWhitespace; 01275 ValueStackOf<DOM_Node>* fNodeStack; 01276 bool fParseInProgress; 01277 XMLScanner* fScanner; 01278 XMLValidator* fValidator; 01279 bool fWithinElement; 01280 DocumentTypeImpl* fDocumentType; 01281 DocTypeHandler* fOldDocTypeHandler; 01282 bool fToCreateXMLDeclTypeNode; 01283 }; 01284 01285 01286 01287 // --------------------------------------------------------------------------- 01288 // DOMParser: Handlers for the XMLEntityHandler interface 01289 // --------------------------------------------------------------------------- 01290 inline void DOMParser::endInputSource(const InputSource&) 01291 { 01292 // The DOM entity resolver doesn't handle this 01293 } 01294 01295 inline bool DOMParser::expandSystemId(const XMLCh* const, XMLBuffer&) 01296 { 01297 // The DOM entity resolver doesn't handle this 01298 return false; 01299 } 01300 01301 inline void DOMParser::resetEntities() 01302 { 01303 // Nothing to do on this one 01304 } 01305 01306 inline void DOMParser::startInputSource(const InputSource&) 01307 { 01308 // The DOM entity resolver doesn't handle this 01309 } 01310 01311 01312 // --------------------------------------------------------------------------- 01313 // DOMParser: Getter methods 01314 // --------------------------------------------------------------------------- 01315 inline DOM_Document DOMParser::getDocument() 01316 { 01317 return fDocument; 01318 } 01319 01320 inline ErrorHandler* DOMParser::getErrorHandler() 01321 { 01322 return fErrorHandler; 01323 } 01324 01325 inline const ErrorHandler* DOMParser::getErrorHandler() const 01326 { 01327 return fErrorHandler; 01328 } 01329 01330 inline EntityResolver* DOMParser::getEntityResolver() 01331 { 01332 return fEntityResolver; 01333 } 01334 01335 inline const EntityResolver* DOMParser::getEntityResolver() const 01336 { 01337 return fEntityResolver; 01338 } 01339 01340 inline bool DOMParser::getExpandEntityReferences() const 01341 { 01342 return fCreateEntityReferenceNodes; 01343 } 01344 inline bool DOMParser::getCreateEntityReferenceNodes() const 01345 { 01346 return fCreateEntityReferenceNodes; 01347 } 01348 01349 inline bool DOMParser::getIncludeIgnorableWhitespace() const 01350 { 01351 return fIncludeIgnorableWhitespace; 01352 } 01353 01354 inline const XMLScanner& DOMParser::getScanner() const 01355 { 01356 return *fScanner; 01357 } 01358 01359 inline bool DOMParser::getToCreateXMLDeclTypeNode() const 01360 { 01361 return fToCreateXMLDeclTypeNode; 01362 } 01363 01364 01365 // --------------------------------------------------------------------------- 01366 // DOMParser: Setter methods 01367 // --------------------------------------------------------------------------- 01368 inline void DOMParser::setExpandEntityReferences(const bool expand) 01369 { 01370 fCreateEntityReferenceNodes = expand; 01371 } 01372 01373 inline void DOMParser::setCreateEntityReferenceNodes(const bool create) 01374 { 01375 fCreateEntityReferenceNodes = create; 01376 } 01377 01378 inline void DOMParser::setIncludeIgnorableWhitespace(const bool include) 01379 { 01380 fIncludeIgnorableWhitespace = include; 01381 } 01382 01383 inline void DOMParser::setToCreateXMLDeclTypeNode(const bool create) 01384 { 01385 fToCreateXMLDeclTypeNode = create; 01386 } 01387 01388 01389 // --------------------------------------------------------------------------- 01390 // DOMParser: Protected getter methods 01391 // --------------------------------------------------------------------------- 01392 inline DOM_Node DOMParser::getCurrentNode() 01393 { 01394 return fCurrentNode; 01395 } 01396 01397 01398 // --------------------------------------------------------------------------- 01399 // DOMParser: Protected setter methods 01400 // --------------------------------------------------------------------------- 01401 inline void DOMParser::setCurrentNode(DOM_Node toSet) 01402 { 01403 fCurrentNode = toSet; 01404 } 01405 01406 inline void DOMParser::setDocument(DOM_Document toSet) 01407 { 01408 fDocument = toSet; 01409 } 01410 01411 #endif

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