[358] | 1 | /*
|
---|
| 2 | * Copyright 1999-2004 The Apache Software Foundation.
|
---|
| 3 | *
|
---|
| 4 | * Licensed under the Apache License, Version 2.0 (the "License");
|
---|
| 5 | * you may not use this file except in compliance with the License.
|
---|
| 6 | * You may obtain a copy of the License at
|
---|
| 7 | *
|
---|
| 8 | * http://www.apache.org/licenses/LICENSE-2.0
|
---|
| 9 | *
|
---|
| 10 | * Unless required by applicable law or agreed to in writing, software
|
---|
| 11 | * distributed under the License is distributed on an "AS IS" BASIS,
|
---|
| 12 | * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
---|
| 13 | * See the License for the specific language governing permissions and
|
---|
| 14 | * limitations under the License.
|
---|
| 15 | */
|
---|
| 16 |
|
---|
| 17 | /*
|
---|
| 18 | * $Log: DTDScanner.hpp,v $
|
---|
| 19 | * Revision 1.11 2004/09/08 13:56:50 peiyongz
|
---|
| 20 | * Apache License Version 2.0
|
---|
| 21 | *
|
---|
| 22 | * Revision 1.10 2004/01/29 11:52:30 cargilld
|
---|
| 23 | * Code cleanup changes to get rid of various compiler diagnostic messages.
|
---|
| 24 | *
|
---|
| 25 | * Revision 1.9 2003/07/10 19:50:12 peiyongz
|
---|
| 26 | * Stateless Grammar: create grammar components with grammarPool's memory Manager
|
---|
| 27 | *
|
---|
| 28 | * Revision 1.8 2003/05/16 21:43:19 knoaman
|
---|
| 29 | * Memory manager implementation: Modify constructors to pass in the memory manager.
|
---|
| 30 | *
|
---|
| 31 | * Revision 1.7 2003/05/15 18:54:51 knoaman
|
---|
| 32 | * Partial implementation of the configurable memory manager.
|
---|
| 33 | *
|
---|
| 34 | * Revision 1.6 2003/02/05 22:07:09 tng
|
---|
| 35 | * [Bug 3111] Problem with LexicalHandler::startDTD() and LexicalHandler::endDTD().
|
---|
| 36 | *
|
---|
| 37 | * Revision 1.5 2002/12/04 02:47:25 knoaman
|
---|
| 38 | * scanner re-organization.
|
---|
| 39 | *
|
---|
| 40 | * Revision 1.4 2002/11/04 14:50:40 tng
|
---|
| 41 | * C++ Namespace Support.
|
---|
| 42 | *
|
---|
| 43 | * Revision 1.3 2002/07/11 18:39:48 knoaman
|
---|
| 44 | * Access entities through the DTDGrammar instead of the scanner.
|
---|
| 45 | *
|
---|
| 46 | * Revision 1.2 2002/05/30 16:17:19 tng
|
---|
| 47 | * Add feature to optionally ignore external DTD.
|
---|
| 48 | *
|
---|
| 49 | * Revision 1.1.1.1 2002/02/01 22:22:44 peiyongz
|
---|
| 50 | * sane_include
|
---|
| 51 | *
|
---|
| 52 | * Revision 1.4 2001/06/21 14:25:56 knoaman
|
---|
| 53 | * Fix for bug 1946
|
---|
| 54 | *
|
---|
| 55 | * Revision 1.3 2001/05/11 13:27:10 tng
|
---|
| 56 | * Copyright update.
|
---|
| 57 | *
|
---|
| 58 | * Revision 1.2 2001/04/19 18:17:22 tng
|
---|
| 59 | * Schema: SchemaValidator update, and use QName in Content Model
|
---|
| 60 | *
|
---|
| 61 | * Revision 1.1 2001/03/21 21:56:20 tng
|
---|
| 62 | * Schema: Add Schema Grammar, Schema Validator, and split the DTDValidator into DTDValidator, DTDScanner, and DTDGrammar.
|
---|
| 63 | *
|
---|
| 64 | */
|
---|
| 65 |
|
---|
| 66 |
|
---|
| 67 |
|
---|
| 68 | #if !defined(DTDSCANNER_HPP)
|
---|
| 69 | #define DTDSCANNER_HPP
|
---|
| 70 |
|
---|
| 71 | #include <xercesc/validators/DTD/DTDGrammar.hpp>
|
---|
| 72 | #include <xercesc/validators/DTD/DTDEntityDecl.hpp>
|
---|
| 73 |
|
---|
| 74 | XERCES_CPP_NAMESPACE_BEGIN
|
---|
| 75 |
|
---|
| 76 | class XMLScanner;
|
---|
| 77 |
|
---|
| 78 | /*
|
---|
| 79 | * Default implementation of an XML DTD scanner.
|
---|
| 80 | */
|
---|
| 81 | class DocTypeHandler;
|
---|
| 82 |
|
---|
| 83 | class VALIDATORS_EXPORT DTDScanner : public XMemory
|
---|
| 84 | {
|
---|
| 85 | public:
|
---|
| 86 | // -----------------------------------------------------------------------
|
---|
| 87 | // Class specific types
|
---|
| 88 | //
|
---|
| 89 | // EntityExpRes
|
---|
| 90 | // Returned from scanEntityRef() to indicate how the expanded text
|
---|
| 91 | // was treated.
|
---|
| 92 | //
|
---|
| 93 | // IDTypes
|
---|
| 94 | // Type of the ID
|
---|
| 95 | // -----------------------------------------------------------------------
|
---|
| 96 | enum EntityExpRes
|
---|
| 97 | {
|
---|
| 98 | EntityExp_Failed
|
---|
| 99 | , EntityExp_Pushed
|
---|
| 100 | , EntityExp_Returned
|
---|
| 101 | };
|
---|
| 102 |
|
---|
| 103 | enum IDTypes
|
---|
| 104 | {
|
---|
| 105 | IDType_Public
|
---|
| 106 | , IDType_External
|
---|
| 107 | , IDType_Either
|
---|
| 108 | };
|
---|
| 109 |
|
---|
| 110 |
|
---|
| 111 |
|
---|
| 112 | // -----------------------------------------------------------------------
|
---|
| 113 | // Constructors and Destructor
|
---|
| 114 | // -----------------------------------------------------------------------
|
---|
| 115 | DTDScanner
|
---|
| 116 | (
|
---|
| 117 | DTDGrammar* dtdGrammar
|
---|
| 118 | , DocTypeHandler* const docTypeHandler
|
---|
| 119 | , MemoryManager* const grammarPoolMemoryManager
|
---|
| 120 | , MemoryManager* const manager = XMLPlatformUtils::fgMemoryManager
|
---|
| 121 | );
|
---|
| 122 | virtual ~DTDScanner();
|
---|
| 123 |
|
---|
| 124 | // -----------------------------------------------------------------------
|
---|
| 125 | // Getter methods
|
---|
| 126 | // -----------------------------------------------------------------------
|
---|
| 127 | DocTypeHandler* getDocTypeHandler();
|
---|
| 128 | const DocTypeHandler* getDocTypeHandler() const;
|
---|
| 129 |
|
---|
| 130 | // -----------------------------------------------------------------------
|
---|
| 131 | // Setter methods
|
---|
| 132 | //
|
---|
| 133 | // setScannerInfo() is called by the scanner to tell the DTDScanner
|
---|
| 134 | // about the stuff it needs to have access to.
|
---|
| 135 | // -----------------------------------------------------------------------
|
---|
| 136 | void setScannerInfo
|
---|
| 137 | (
|
---|
| 138 | XMLScanner* const owningScanner
|
---|
| 139 | , ReaderMgr* const readerMgr
|
---|
| 140 | , XMLBufferMgr* const bufMgr
|
---|
| 141 | );
|
---|
| 142 |
|
---|
| 143 | void setDocTypeHandler
|
---|
| 144 | (
|
---|
| 145 | DocTypeHandler* const handlerToSet
|
---|
| 146 | );
|
---|
| 147 |
|
---|
| 148 | void scanExtSubsetDecl(const bool inIncludeSect, const bool isDTD);
|
---|
| 149 | bool scanInternalSubset();
|
---|
| 150 | bool scanId
|
---|
| 151 | (
|
---|
| 152 | XMLBuffer& pubIdToFill
|
---|
| 153 | , XMLBuffer& sysIdToFill
|
---|
| 154 | , const IDTypes whatKind
|
---|
| 155 | );
|
---|
| 156 |
|
---|
| 157 | private:
|
---|
| 158 | // -----------------------------------------------------------------------
|
---|
| 159 | // Unimplemented constructors and operators
|
---|
| 160 | // -----------------------------------------------------------------------
|
---|
| 161 | DTDScanner(const DTDScanner &);
|
---|
| 162 | DTDScanner& operator = (const DTDScanner&);
|
---|
| 163 |
|
---|
| 164 | // -----------------------------------------------------------------------
|
---|
| 165 | // Private DTD scanning methods. These are all in XMLValidator2.cpp
|
---|
| 166 | // -----------------------------------------------------------------------
|
---|
| 167 | bool checkForPERef
|
---|
| 168 | (
|
---|
| 169 | const bool inLiteral
|
---|
| 170 | , const bool inMarkup
|
---|
| 171 | );
|
---|
| 172 | bool expandPERef
|
---|
| 173 | (
|
---|
| 174 | const bool scanExternal
|
---|
| 175 | , const bool inLiteral
|
---|
| 176 | , const bool inMarkup
|
---|
| 177 | , const bool throwEndOfExt = false
|
---|
| 178 | );
|
---|
| 179 | bool getQuotedString(XMLBuffer& toFill);
|
---|
| 180 | XMLAttDef* scanAttDef(DTDElementDecl& elemDecl, XMLBuffer& bufToUse);
|
---|
| 181 | bool scanAttValue
|
---|
| 182 | (
|
---|
| 183 | const XMLCh* const attrName
|
---|
| 184 | , XMLBuffer& toFill
|
---|
| 185 | , const XMLAttDef::AttTypes type
|
---|
| 186 | );
|
---|
| 187 | void scanAttListDecl();
|
---|
| 188 | ContentSpecNode* scanChildren
|
---|
| 189 | (
|
---|
| 190 | const DTDElementDecl& elemDecl
|
---|
| 191 | , XMLBuffer& bufToUse
|
---|
| 192 | );
|
---|
| 193 | bool scanCharRef(XMLCh& toFill, XMLCh& second);
|
---|
| 194 | void scanComment();
|
---|
| 195 | bool scanContentSpec(DTDElementDecl& toFill);
|
---|
| 196 | void scanDefaultDecl(DTDAttDef& toFill);
|
---|
| 197 | void scanElementDecl();
|
---|
| 198 | void scanEntityDecl();
|
---|
| 199 | bool scanEntityDef();
|
---|
| 200 | bool scanEntityLiteral(XMLBuffer& toFill);
|
---|
| 201 | bool scanEntityDef(DTDEntityDecl& decl, const bool isPEDecl);
|
---|
| 202 | EntityExpRes scanEntityRef(XMLCh& firstCh, XMLCh& secondCh, bool& escaped);
|
---|
| 203 | bool scanEnumeration
|
---|
| 204 | (
|
---|
| 205 | const DTDAttDef& attDef
|
---|
| 206 | , XMLBuffer& toFill
|
---|
| 207 | , const bool notation
|
---|
| 208 | );
|
---|
| 209 | bool scanEq();
|
---|
| 210 | void scanIgnoredSection();
|
---|
| 211 | void scanMarkupDecl(const bool parseTextDecl);
|
---|
| 212 | bool scanMixed(DTDElementDecl& toFill);
|
---|
| 213 | void scanNotationDecl();
|
---|
| 214 | void scanPI();
|
---|
| 215 | bool scanPublicLiteral(XMLBuffer& toFill);
|
---|
| 216 | bool scanSystemLiteral(XMLBuffer& toFill);
|
---|
| 217 | void scanTextDecl();
|
---|
| 218 | bool isReadingExternalEntity();
|
---|
| 219 |
|
---|
| 220 |
|
---|
| 221 | // -----------------------------------------------------------------------
|
---|
| 222 | // Private data members
|
---|
| 223 | //
|
---|
| 224 | // fDocTypeHandler
|
---|
| 225 | // This holds the optional doc type handler that can be installed
|
---|
| 226 | // and used to call back for all markup events. It is DTD specific.
|
---|
| 227 | //
|
---|
| 228 | // fDumAttDef
|
---|
| 229 | // fDumElemDecl
|
---|
| 230 | // fDumEntityDecl
|
---|
| 231 | // These are dummy objects into which mark decls are parsed when
|
---|
| 232 | // they are just overrides of previously declared markup decls. In
|
---|
| 233 | // such situations, the first one wins but we need to have somewhere
|
---|
| 234 | // to parse them into. So these are lazily created and used as needed
|
---|
| 235 | // when such markup decls are seen.
|
---|
| 236 | //
|
---|
| 237 | // fInternalSubset
|
---|
| 238 | // This is used to track whether we are in the internal subset or not,
|
---|
| 239 | // in which case we are in the external subset.
|
---|
| 240 | //
|
---|
| 241 | // fNextAttrId
|
---|
| 242 | // Since att defs are per-element, we don't have a validator wide
|
---|
| 243 | // attribute def pool. So we use a simpler data structure in each
|
---|
| 244 | // element decl to store its att defs, and we use this simple counter
|
---|
| 245 | // to apply a unique id to each new attribute.
|
---|
| 246 | //
|
---|
| 247 | // fDTDGrammar
|
---|
| 248 | // The DTD information we scanned like element decl, attribute decl
|
---|
| 249 | // are stored in this Grammar.
|
---|
| 250 | //
|
---|
| 251 | // fBufMgr
|
---|
| 252 | // This is the buffer manager of the scanner. This is provided as a
|
---|
| 253 | // convenience so that the DTDScanner doesn't have to create its own
|
---|
| 254 | // buffer manager during the parse process.
|
---|
| 255 | //
|
---|
| 256 | // fReaderMgr
|
---|
| 257 | // This is a pointer to the reader manager that is being used by the scanner.
|
---|
| 258 | //
|
---|
| 259 | // fScanner
|
---|
| 260 | // The pointer to the scanner to which this DTDScanner belongs
|
---|
| 261 | //
|
---|
| 262 | // fPEntityDeclPool
|
---|
| 263 | // This is a pool of EntityDecl objects, which contains all of the
|
---|
| 264 | // parameter entities that are declared in the DTD subsets.
|
---|
| 265 | //
|
---|
| 266 | // fEmptyNamespaceId
|
---|
| 267 | // The uri for all DTD decls
|
---|
| 268 | //
|
---|
| 269 | // fDocTypeReaderId
|
---|
| 270 | // The original reader in the fReaderMgr - to be compared against the
|
---|
| 271 | // current reader to decide whether we are processing an external/internal
|
---|
| 272 | // declaration
|
---|
| 273 | // -----------------------------------------------------------------------
|
---|
| 274 | MemoryManager* fMemoryManager;
|
---|
| 275 | MemoryManager* fGrammarPoolMemoryManager;
|
---|
| 276 | DocTypeHandler* fDocTypeHandler;
|
---|
| 277 | DTDAttDef* fDumAttDef;
|
---|
| 278 | DTDElementDecl* fDumElemDecl;
|
---|
| 279 | DTDEntityDecl* fDumEntityDecl;
|
---|
| 280 | bool fInternalSubset;
|
---|
| 281 | unsigned int fNextAttrId;
|
---|
| 282 | DTDGrammar* fDTDGrammar;
|
---|
| 283 | XMLBufferMgr* fBufMgr;
|
---|
| 284 | ReaderMgr* fReaderMgr;
|
---|
| 285 | XMLScanner* fScanner;
|
---|
| 286 | NameIdPool<DTDEntityDecl>* fPEntityDeclPool;
|
---|
| 287 | unsigned int fEmptyNamespaceId;
|
---|
| 288 | unsigned int fDocTypeReaderId;
|
---|
| 289 | };
|
---|
| 290 |
|
---|
| 291 |
|
---|
| 292 | // ---------------------------------------------------------------------------
|
---|
| 293 | // DTDScanner: Getter methods
|
---|
| 294 | // ---------------------------------------------------------------------------
|
---|
| 295 | inline DocTypeHandler* DTDScanner::getDocTypeHandler()
|
---|
| 296 | {
|
---|
| 297 | return fDocTypeHandler;
|
---|
| 298 | }
|
---|
| 299 |
|
---|
| 300 | inline const DocTypeHandler* DTDScanner::getDocTypeHandler() const
|
---|
| 301 | {
|
---|
| 302 | return fDocTypeHandler;
|
---|
| 303 | }
|
---|
| 304 |
|
---|
| 305 |
|
---|
| 306 | // ---------------------------------------------------------------------------
|
---|
| 307 | // DTDScanner: Setter methods
|
---|
| 308 | // ---------------------------------------------------------------------------
|
---|
| 309 | inline void DTDScanner::setDocTypeHandler(DocTypeHandler* const handlerToSet)
|
---|
| 310 | {
|
---|
| 311 | fDocTypeHandler = handlerToSet;
|
---|
| 312 | }
|
---|
| 313 |
|
---|
| 314 | // -----------------------------------------------------------------------
|
---|
| 315 | // Helper methods
|
---|
| 316 | // -----------------------------------------------------------------------
|
---|
| 317 | inline bool DTDScanner::isReadingExternalEntity() {
|
---|
| 318 | return (fDocTypeReaderId != fReaderMgr->getCurrentReaderNum());
|
---|
| 319 | }
|
---|
| 320 |
|
---|
| 321 | XERCES_CPP_NAMESPACE_END
|
---|
| 322 |
|
---|
| 323 | #endif
|
---|