[2674] | 1 | /* |
---|
| 2 | * Licensed to the Apache Software Foundation (ASF) under one or more |
---|
| 3 | * contributor license agreements. See the NOTICE file distributed with |
---|
| 4 | * this work for additional information regarding copyright ownership. |
---|
| 5 | * The ASF licenses this file to You under the Apache License, Version 2.0 |
---|
| 6 | * (the "License"); you may not use this file except in compliance with |
---|
| 7 | * the License. You may obtain a copy of the License at |
---|
| 8 | * |
---|
| 9 | * http://www.apache.org/licenses/LICENSE-2.0 |
---|
| 10 | * |
---|
| 11 | * Unless required by applicable law or agreed to in writing, software |
---|
| 12 | * distributed under the License is distributed on an "AS IS" BASIS, |
---|
| 13 | * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. |
---|
| 14 | * See the License for the specific language governing permissions and |
---|
| 15 | * limitations under the License. |
---|
| 16 | */ |
---|
| 17 | |
---|
| 18 | /* |
---|
| 19 | * $Id: IGXMLScanner.hpp 568078 2007-08-21 11:43:25Z amassari $ |
---|
| 20 | */ |
---|
| 21 | |
---|
| 22 | |
---|
| 23 | #if !defined(IGXMLSCANNER_HPP) |
---|
| 24 | #define IGXMLSCANNER_HPP |
---|
| 25 | |
---|
| 26 | #include <xercesc/internal/XMLScanner.hpp> |
---|
| 27 | #include <xercesc/util/KVStringPair.hpp> |
---|
| 28 | #include <xercesc/util/NameIdPool.hpp> |
---|
| 29 | #include <xercesc/util/RefHash3KeysIdPool.hpp> |
---|
| 30 | #include <xercesc/validators/common/Grammar.hpp> |
---|
| 31 | #include <xercesc/validators/schema/SchemaElementDecl.hpp> |
---|
| 32 | |
---|
| 33 | XERCES_CPP_NAMESPACE_BEGIN |
---|
| 34 | |
---|
| 35 | class DTDElementDecl; |
---|
| 36 | class DTDGrammar; |
---|
| 37 | class DTDValidator; |
---|
| 38 | class SchemaValidator; |
---|
| 39 | class IdentityConstraintHandler; |
---|
| 40 | class IdentityConstraint; |
---|
| 41 | class ContentLeafNameTypeVector; |
---|
| 42 | class SchemaAttDef; |
---|
| 43 | class XMLContentModel; |
---|
| 44 | class XSModel; |
---|
| 45 | class PSVIAttributeList; |
---|
| 46 | class PSVIElement; |
---|
| 47 | |
---|
| 48 | // This is an integrated scanner class, which does DTD/XML Schema grammar |
---|
| 49 | // processing. |
---|
| 50 | class XMLPARSER_EXPORT IGXMLScanner : public XMLScanner |
---|
| 51 | { |
---|
| 52 | public : |
---|
| 53 | // ----------------------------------------------------------------------- |
---|
| 54 | // Constructors and Destructor |
---|
| 55 | // ----------------------------------------------------------------------- |
---|
| 56 | IGXMLScanner |
---|
| 57 | ( |
---|
| 58 | XMLValidator* const valToAdopt |
---|
| 59 | , GrammarResolver* const grammarResolver |
---|
| 60 | , MemoryManager* const manager = XMLPlatformUtils::fgMemoryManager |
---|
| 61 | ); |
---|
| 62 | IGXMLScanner |
---|
| 63 | ( |
---|
| 64 | XMLDocumentHandler* const docHandler |
---|
| 65 | , DocTypeHandler* const docTypeHandler |
---|
| 66 | , XMLEntityHandler* const entityHandler |
---|
| 67 | , XMLErrorReporter* const errReporter |
---|
| 68 | , XMLValidator* const valToAdopt |
---|
| 69 | , GrammarResolver* const grammarResolver |
---|
| 70 | , MemoryManager* const manager = XMLPlatformUtils::fgMemoryManager |
---|
| 71 | ); |
---|
| 72 | virtual ~IGXMLScanner(); |
---|
| 73 | |
---|
| 74 | // ----------------------------------------------------------------------- |
---|
| 75 | // XMLScanner public virtual methods |
---|
| 76 | // ----------------------------------------------------------------------- |
---|
| 77 | virtual const XMLCh* getName() const; |
---|
| 78 | virtual NameIdPool<DTDEntityDecl>* getEntityDeclPool(); |
---|
| 79 | virtual const NameIdPool<DTDEntityDecl>* getEntityDeclPool() const; |
---|
| 80 | virtual unsigned int resolveQName |
---|
| 81 | ( |
---|
| 82 | const XMLCh* const qName |
---|
| 83 | , XMLBuffer& prefixBufToFill |
---|
| 84 | , const short mode |
---|
| 85 | , int& prefixColonPos |
---|
| 86 | ); |
---|
| 87 | virtual void scanDocument |
---|
| 88 | ( |
---|
| 89 | const InputSource& src |
---|
| 90 | ); |
---|
| 91 | virtual bool scanNext(XMLPScanToken& toFill); |
---|
| 92 | virtual Grammar* loadGrammar |
---|
| 93 | ( |
---|
| 94 | const InputSource& src |
---|
| 95 | , const short grammarType |
---|
| 96 | , const bool toCache = false |
---|
| 97 | ); |
---|
| 98 | |
---|
| 99 | virtual Grammar::GrammarType getCurrentGrammarType() const; |
---|
| 100 | |
---|
| 101 | private : |
---|
| 102 | // ----------------------------------------------------------------------- |
---|
| 103 | // Unimplemented constructors and operators |
---|
| 104 | // ----------------------------------------------------------------------- |
---|
| 105 | IGXMLScanner(); |
---|
| 106 | IGXMLScanner(const IGXMLScanner&); |
---|
| 107 | IGXMLScanner& operator=(const IGXMLScanner&); |
---|
| 108 | |
---|
| 109 | // ----------------------------------------------------------------------- |
---|
| 110 | // XMLScanner virtual methods |
---|
| 111 | // ----------------------------------------------------------------------- |
---|
| 112 | virtual void scanCDSection(); |
---|
| 113 | virtual void scanCharData(XMLBuffer& toToUse); |
---|
| 114 | virtual EntityExpRes scanEntityRef |
---|
| 115 | ( |
---|
| 116 | const bool inAttVal |
---|
| 117 | , XMLCh& firstCh |
---|
| 118 | , XMLCh& secondCh |
---|
| 119 | , bool& escaped |
---|
| 120 | ); |
---|
| 121 | virtual void scanDocTypeDecl(); |
---|
| 122 | virtual void scanReset(const InputSource& src); |
---|
| 123 | virtual void sendCharData(XMLBuffer& toSend); |
---|
| 124 | virtual InputSource* resolveSystemId(const XMLCh* const sysId |
---|
| 125 | ,const XMLCh* const pubId); |
---|
| 126 | |
---|
| 127 | // ----------------------------------------------------------------------- |
---|
| 128 | // Private helper methods |
---|
| 129 | // ----------------------------------------------------------------------- |
---|
| 130 | void commonInit(); |
---|
| 131 | void cleanUp(); |
---|
| 132 | |
---|
| 133 | unsigned int buildAttList |
---|
| 134 | ( |
---|
| 135 | const RefVectorOf<KVStringPair>& providedAttrs |
---|
| 136 | , const unsigned int attCount |
---|
| 137 | , XMLElementDecl* elemDecl |
---|
| 138 | , RefVectorOf<XMLAttr>& toFill |
---|
| 139 | ); |
---|
| 140 | bool normalizeAttValue |
---|
| 141 | ( |
---|
| 142 | const XMLAttDef* const attDef |
---|
| 143 | , const XMLCh* const name |
---|
| 144 | , const XMLCh* const value |
---|
| 145 | , XMLBuffer& toFill |
---|
| 146 | ); |
---|
| 147 | bool normalizeAttRawValue |
---|
| 148 | ( |
---|
| 149 | const XMLCh* const attrName |
---|
| 150 | , const XMLCh* const value |
---|
| 151 | , XMLBuffer& toFill |
---|
| 152 | ); |
---|
| 153 | unsigned int resolvePrefix |
---|
| 154 | ( |
---|
| 155 | const XMLCh* const prefix |
---|
| 156 | , const ElemStack::MapModes mode |
---|
| 157 | ); |
---|
| 158 | unsigned int resolvePrefix |
---|
| 159 | ( |
---|
| 160 | const XMLCh* const prefix |
---|
| 161 | , XMLBuffer& uriBufToFill |
---|
| 162 | , const ElemStack::MapModes mode |
---|
| 163 | ); |
---|
| 164 | void updateNSMap |
---|
| 165 | ( |
---|
| 166 | const XMLCh* const attrName |
---|
| 167 | , const XMLCh* const attrValue |
---|
| 168 | ); |
---|
| 169 | void updateNSMap |
---|
| 170 | ( |
---|
| 171 | const XMLCh* const attrName |
---|
| 172 | , const XMLCh* const attrValue |
---|
| 173 | , const int colonPosition |
---|
| 174 | ); |
---|
| 175 | void scanRawAttrListforNameSpaces(int attCount); |
---|
| 176 | void parseSchemaLocation(const XMLCh* const schemaLocationStr); |
---|
| 177 | void resolveSchemaGrammar(const XMLCh* const loc, const XMLCh* const uri); |
---|
| 178 | bool switchGrammar(const XMLCh* const newGrammarNameSpace); |
---|
| 179 | bool laxElementValidation(QName* element, ContentLeafNameTypeVector* cv, |
---|
| 180 | const XMLContentModel* const cm, |
---|
| 181 | const unsigned int parentElemDepth); |
---|
| 182 | bool anyAttributeValidation(SchemaAttDef* attWildCard, |
---|
| 183 | unsigned int uriId, |
---|
| 184 | bool& skipThisOne, |
---|
| 185 | bool& laxThisOne); |
---|
| 186 | void resizeElemState(); |
---|
| 187 | void processSchemaLocation(XMLCh* const schemaLoc); |
---|
| 188 | |
---|
| 189 | void resizeRawAttrColonList(); |
---|
| 190 | |
---|
| 191 | unsigned int resolveQNameWithColon |
---|
| 192 | ( |
---|
| 193 | const XMLCh* const qName |
---|
| 194 | , XMLBuffer& prefixBufToFill |
---|
| 195 | , const short mode |
---|
| 196 | , const int prefixColonPos |
---|
| 197 | ); |
---|
| 198 | // ----------------------------------------------------------------------- |
---|
| 199 | // Private scanning methods |
---|
| 200 | // ----------------------------------------------------------------------- |
---|
| 201 | bool basicAttrValueScan |
---|
| 202 | ( |
---|
| 203 | const XMLCh* const attrName |
---|
| 204 | , XMLBuffer& toFill |
---|
| 205 | ); |
---|
| 206 | unsigned int rawAttrScan |
---|
| 207 | ( |
---|
| 208 | const XMLCh* const elemName |
---|
| 209 | , RefVectorOf<KVStringPair>& toFill |
---|
| 210 | , bool& isEmpty |
---|
| 211 | ); |
---|
| 212 | bool scanAttValue |
---|
| 213 | ( |
---|
| 214 | const XMLAttDef* const attDef |
---|
| 215 | , const XMLCh* const attrName |
---|
| 216 | , XMLBuffer& toFill |
---|
| 217 | ); |
---|
| 218 | bool scanContent(); |
---|
| 219 | void scanEndTag(bool& gotData); |
---|
| 220 | bool scanStartTag(bool& gotData); |
---|
| 221 | bool scanStartTagNS(bool& gotData); |
---|
| 222 | |
---|
| 223 | // ----------------------------------------------------------------------- |
---|
| 224 | // IdentityConstraints Activation methods |
---|
| 225 | // ----------------------------------------------------------------------- |
---|
| 226 | inline bool toCheckIdentityConstraint() const; |
---|
| 227 | |
---|
| 228 | // ----------------------------------------------------------------------- |
---|
| 229 | // Grammar preparsing methods |
---|
| 230 | // ----------------------------------------------------------------------- |
---|
| 231 | Grammar* loadXMLSchemaGrammar(const InputSource& src, const bool toCache = false); |
---|
| 232 | Grammar* loadDTDGrammar(const InputSource& src, const bool toCache = false); |
---|
| 233 | |
---|
| 234 | // ----------------------------------------------------------------------- |
---|
| 235 | // PSVI handling methods |
---|
| 236 | // ----------------------------------------------------------------------- |
---|
| 237 | void endElementPSVI(SchemaElementDecl* const elemDecl, |
---|
| 238 | DatatypeValidator* const memberDV); |
---|
| 239 | void resetPSVIElemContext(); |
---|
| 240 | |
---|
| 241 | // ----------------------------------------------------------------------- |
---|
| 242 | // Data members |
---|
| 243 | // |
---|
| 244 | // fRawAttrList |
---|
| 245 | // During the initial scan of the attributes we can only do a raw |
---|
| 246 | // scan for key/value pairs. So this vector is used to store them |
---|
| 247 | // until they can be processed (and put into fAttrList.) |
---|
| 248 | // |
---|
| 249 | // fDTDValidator |
---|
| 250 | // The DTD validator instance. |
---|
| 251 | // |
---|
| 252 | // fSchemaValidator |
---|
| 253 | // The Schema validator instance. |
---|
| 254 | // |
---|
| 255 | // fSeeXsi |
---|
| 256 | // This flag indicates a schema has been seen. |
---|
| 257 | // |
---|
| 258 | // fElemState |
---|
| 259 | // fElemStateSize |
---|
| 260 | // Stores an element next state from DFA content model - used for |
---|
| 261 | // wildcard validation |
---|
| 262 | // |
---|
| 263 | // fDTDElemNonDeclPool |
---|
| 264 | // registry of "faulted-in" DTD element decls |
---|
| 265 | // fSchemaElemNonDeclPool |
---|
| 266 | // registry for elements without decls in the grammar |
---|
| 267 | // fElemCount |
---|
| 268 | // count of the number of start tags seen so far (starts at 1). |
---|
| 269 | // Used for duplicate attribute detection/processing of required/defaulted attributes |
---|
| 270 | // fAttDefRegistry |
---|
| 271 | // mapping from XMLAttDef instances to the count of the last |
---|
| 272 | // start tag where they were utilized. |
---|
| 273 | // fUndeclaredAttrRegistry |
---|
| 274 | // mapping of attr QNames to the count of the last start tag in which they occurred |
---|
| 275 | // fUndeclaredAttrRegistryNS |
---|
| 276 | // mapping of namespaceId/localName pairs to the count of the last |
---|
| 277 | // start tag in which they occurred. |
---|
| 278 | // fPSVIAttrList |
---|
| 279 | // PSVI attribute list implementation that needs to be |
---|
| 280 | // filled when a PSVIHandler is registered |
---|
| 281 | // |
---|
| 282 | // ----------------------------------------------------------------------- |
---|
| 283 | bool fSeeXsi; |
---|
| 284 | Grammar::GrammarType fGrammarType; |
---|
| 285 | unsigned int fElemStateSize; |
---|
| 286 | unsigned int* fElemState; |
---|
| 287 | XMLBuffer fContent; |
---|
| 288 | RefVectorOf<KVStringPair>* fRawAttrList; |
---|
| 289 | unsigned int fRawAttrColonListSize; |
---|
| 290 | int* fRawAttrColonList; |
---|
| 291 | DTDValidator* fDTDValidator; |
---|
| 292 | SchemaValidator* fSchemaValidator; |
---|
| 293 | DTDGrammar* fDTDGrammar; |
---|
| 294 | IdentityConstraintHandler* fICHandler; |
---|
| 295 | ValueVectorOf<XMLCh*>* fLocationPairs; |
---|
| 296 | NameIdPool<DTDElementDecl>* fDTDElemNonDeclPool; |
---|
| 297 | RefHash3KeysIdPool<SchemaElementDecl>* fSchemaElemNonDeclPool; |
---|
| 298 | unsigned int fElemCount; |
---|
| 299 | RefHashTableOf<unsigned int>* fAttDefRegistry; |
---|
| 300 | RefHashTableOf<unsigned int>* fUndeclaredAttrRegistry; |
---|
| 301 | RefHash2KeysTableOf<unsigned int>* fUndeclaredAttrRegistryNS; |
---|
| 302 | PSVIAttributeList * fPSVIAttrList; |
---|
| 303 | XSModel* fModel; |
---|
| 304 | PSVIElement* fPSVIElement; |
---|
| 305 | ValueStackOf<bool>* fErrorStack; |
---|
| 306 | PSVIElemContext fPSVIElemContext; |
---|
| 307 | }; |
---|
| 308 | |
---|
| 309 | inline const XMLCh* IGXMLScanner::getName() const |
---|
| 310 | { |
---|
| 311 | return XMLUni::fgIGXMLScanner; |
---|
| 312 | } |
---|
| 313 | |
---|
| 314 | inline bool IGXMLScanner::toCheckIdentityConstraint() const |
---|
| 315 | { |
---|
| 316 | return fValidate && fIdentityConstraintChecking && fICHandler; |
---|
| 317 | } |
---|
| 318 | |
---|
| 319 | inline Grammar::GrammarType IGXMLScanner::getCurrentGrammarType() const |
---|
| 320 | { |
---|
| 321 | return fGrammarType; |
---|
| 322 | } |
---|
| 323 | |
---|
| 324 | XERCES_CPP_NAMESPACE_END |
---|
| 325 | |
---|
| 326 | #endif |
---|