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