[358] | 1 | /*
|
---|
| 2 | * Copyright 1999-2001,2004 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: DFAContentModel.hpp,v $
|
---|
| 19 | * Revision 1.8 2004/09/16 13:32:03 amassari
|
---|
| 20 | * Updated error message for UPA to also state the complex type that is failing the test
|
---|
| 21 | *
|
---|
| 22 | * Revision 1.7 2004/09/08 13:56:51 peiyongz
|
---|
| 23 | * Apache License Version 2.0
|
---|
| 24 | *
|
---|
| 25 | * Revision 1.6 2003/12/17 00:18:38 cargilld
|
---|
| 26 | * Update to memory management so that the static memory manager (one used to call Initialize) is only for static data.
|
---|
| 27 | *
|
---|
| 28 | * Revision 1.5 2003/05/16 21:43:20 knoaman
|
---|
| 29 | * Memory manager implementation: Modify constructors to pass in the memory manager.
|
---|
| 30 | *
|
---|
| 31 | * Revision 1.4 2003/05/15 18:48:27 knoaman
|
---|
| 32 | * Partial implementation of the configurable memory manager.
|
---|
| 33 | *
|
---|
| 34 | * Revision 1.3 2003/03/07 18:16:57 tng
|
---|
| 35 | * Return a reference instead of void for operator=
|
---|
| 36 | *
|
---|
| 37 | * Revision 1.2 2002/11/04 14:54:58 tng
|
---|
| 38 | * C++ Namespace Support.
|
---|
| 39 | *
|
---|
| 40 | * Revision 1.1.1.1 2002/02/01 22:22:38 peiyongz
|
---|
| 41 | * sane_include
|
---|
| 42 | *
|
---|
| 43 | * Revision 1.13 2001/11/21 14:30:13 knoaman
|
---|
| 44 | * Fix for UPA checking.
|
---|
| 45 | *
|
---|
| 46 | * Revision 1.12 2001/08/24 12:48:48 tng
|
---|
| 47 | * Schema: AllContentModel
|
---|
| 48 | *
|
---|
| 49 | * Revision 1.11 2001/08/21 16:06:11 tng
|
---|
| 50 | * Schema: Unique Particle Attribution Constraint Checking.
|
---|
| 51 | *
|
---|
| 52 | * Revision 1.10 2001/08/13 15:06:39 knoaman
|
---|
| 53 | * update <any> validation.
|
---|
| 54 | *
|
---|
| 55 | * Revision 1.9 2001/06/13 20:50:55 peiyongz
|
---|
| 56 | * fIsMixed: to handle mixed Content Model
|
---|
| 57 | *
|
---|
| 58 | * Revision 1.8 2001/05/11 13:27:18 tng
|
---|
| 59 | * Copyright update.
|
---|
| 60 | *
|
---|
| 61 | * Revision 1.7 2001/05/03 21:02:30 tng
|
---|
| 62 | * Schema: Add SubstitutionGroupComparator and update exception messages. By Pei Yong Zhang.
|
---|
| 63 | *
|
---|
| 64 | * Revision 1.6 2001/04/19 18:17:30 tng
|
---|
| 65 | * Schema: SchemaValidator update, and use QName in Content Model
|
---|
| 66 | *
|
---|
| 67 | * Revision 1.5 2001/03/21 21:56:27 tng
|
---|
| 68 | * Schema: Add Schema Grammar, Schema Validator, and split the DTDValidator into DTDValidator, DTDScanner, and DTDGrammar.
|
---|
| 69 | *
|
---|
| 70 | * Revision 1.4 2001/03/21 19:29:55 tng
|
---|
| 71 | * Schema: Content Model Updates, by Pei Yong Zhang.
|
---|
| 72 | *
|
---|
| 73 | * Revision 1.3 2001/02/27 18:32:32 tng
|
---|
| 74 | * Schema: Use XMLElementDecl instead of DTDElementDecl in Content Model.
|
---|
| 75 | *
|
---|
| 76 | * Revision 1.2 2001/02/27 14:48:52 tng
|
---|
| 77 | * Schema: Add CMAny and ContentLeafNameTypeVector, by Pei Yong Zhang
|
---|
| 78 | *
|
---|
| 79 | * Revision 1.1 2001/02/16 14:17:29 tng
|
---|
| 80 | * Schema: Move the common Content Model files that are shared by DTD
|
---|
| 81 | * and schema from 'DTD' folder to 'common' folder. By Pei Yong Zhang.
|
---|
| 82 | *
|
---|
| 83 | * Revision 1.4 2000/03/02 19:55:38 roddey
|
---|
| 84 | * This checkin includes many changes done while waiting for the
|
---|
| 85 | * 1.1.0 code to be finished. I can't list them all here, but a list is
|
---|
| 86 | * available elsewhere.
|
---|
| 87 | *
|
---|
| 88 | * Revision 1.3 2000/02/24 20:16:48 abagchi
|
---|
| 89 | * Swat for removing Log from API docs
|
---|
| 90 | *
|
---|
| 91 | * Revision 1.2 2000/02/09 21:42:37 abagchi
|
---|
| 92 | * Copyright swat
|
---|
| 93 | *
|
---|
| 94 | * Revision 1.1.1.1 1999/11/09 01:03:19 twl
|
---|
| 95 | * Initial checkin
|
---|
| 96 | *
|
---|
| 97 | * Revision 1.2 1999/11/08 20:45:38 rahul
|
---|
| 98 | * Swat for adding in Product name and CVS comment log variable.
|
---|
| 99 | *
|
---|
| 100 | */
|
---|
| 101 |
|
---|
| 102 | #if !defined(DFACONTENTMODEL_HPP)
|
---|
| 103 | #define DFACONTENTMODEL_HPP
|
---|
| 104 |
|
---|
| 105 | #include <xercesc/util/XercesDefs.hpp>
|
---|
| 106 | #include <xercesc/util/ArrayIndexOutOfBoundsException.hpp>
|
---|
| 107 | #include <xercesc/framework/XMLContentModel.hpp>
|
---|
| 108 | #include <xercesc/validators/common/ContentLeafNameTypeVector.hpp>
|
---|
| 109 |
|
---|
| 110 | XERCES_CPP_NAMESPACE_BEGIN
|
---|
| 111 |
|
---|
| 112 | class ContentSpecNode;
|
---|
| 113 | class CMLeaf;
|
---|
| 114 | class CMNode;
|
---|
| 115 | class CMStateSet;
|
---|
| 116 |
|
---|
| 117 | //
|
---|
| 118 | // DFAContentModel is the heavy weight derivative of ContentModel that does
|
---|
| 119 | // all of the non-trivial element content validation. This guy does the full
|
---|
| 120 | // bore regular expression to DFA conversion to create a DFA that it then
|
---|
| 121 | // uses in its validation algorithm.
|
---|
| 122 | //
|
---|
| 123 | // NOTE: Upstream work insures that this guy will never see a content model
|
---|
| 124 | // with PCDATA in it. Any model with PCDATA is 'mixed' and is handled
|
---|
| 125 | // via the MixedContentModel class, since mixed models are very
|
---|
| 126 | // constrained in form and easily handled via a special case. This
|
---|
| 127 | // also makes our life much easier here.
|
---|
| 128 | //
|
---|
| 129 | class DFAContentModel : public XMLContentModel
|
---|
| 130 | {
|
---|
| 131 | public:
|
---|
| 132 | // -----------------------------------------------------------------------
|
---|
| 133 | // Constructors and Destructor
|
---|
| 134 | // -----------------------------------------------------------------------
|
---|
| 135 | DFAContentModel
|
---|
| 136 | (
|
---|
| 137 | const bool dtd
|
---|
| 138 | , ContentSpecNode* const elemContentSpec
|
---|
| 139 | , MemoryManager* const manager = XMLPlatformUtils::fgMemoryManager
|
---|
| 140 | );
|
---|
| 141 | DFAContentModel
|
---|
| 142 | (
|
---|
| 143 | const bool dtd
|
---|
| 144 | , ContentSpecNode* const elemContentSpec
|
---|
| 145 | , const bool isMixed
|
---|
| 146 | , MemoryManager* const manager
|
---|
| 147 | );
|
---|
| 148 |
|
---|
| 149 | virtual ~DFAContentModel();
|
---|
| 150 |
|
---|
| 151 |
|
---|
| 152 | // -----------------------------------------------------------------------
|
---|
| 153 | // Implementation of the virtual content model interface
|
---|
| 154 | // -----------------------------------------------------------------------
|
---|
| 155 | virtual int validateContent
|
---|
| 156 | (
|
---|
| 157 | QName** const children
|
---|
| 158 | , const unsigned int childCount
|
---|
| 159 | , const unsigned int emptyNamespaceId
|
---|
| 160 | ) const;
|
---|
| 161 |
|
---|
| 162 | virtual int validateContentSpecial
|
---|
| 163 | (
|
---|
| 164 | QName** const children
|
---|
| 165 | , const unsigned int childCount
|
---|
| 166 | , const unsigned int emptyNamespaceId
|
---|
| 167 | , GrammarResolver* const pGrammarResolver
|
---|
| 168 | , XMLStringPool* const pStringPool
|
---|
| 169 | ) const;
|
---|
| 170 |
|
---|
| 171 | virtual void checkUniqueParticleAttribution
|
---|
| 172 | (
|
---|
| 173 | SchemaGrammar* const pGrammar
|
---|
| 174 | , GrammarResolver* const pGrammarResolver
|
---|
| 175 | , XMLStringPool* const pStringPool
|
---|
| 176 | , XMLValidator* const pValidator
|
---|
| 177 | , unsigned int* const pContentSpecOrgURI
|
---|
| 178 | , const XMLCh* pComplexTypeName = 0
|
---|
| 179 | ) ;
|
---|
| 180 |
|
---|
| 181 | virtual ContentLeafNameTypeVector* getContentLeafNameTypeVector() const ;
|
---|
| 182 |
|
---|
| 183 | virtual unsigned int getNextState(const unsigned int currentState,
|
---|
| 184 | const unsigned int elementIndex) const;
|
---|
| 185 |
|
---|
| 186 | private :
|
---|
| 187 | // -----------------------------------------------------------------------
|
---|
| 188 | // Unimplemented constructors and operators
|
---|
| 189 | // -----------------------------------------------------------------------
|
---|
| 190 | DFAContentModel();
|
---|
| 191 | DFAContentModel(const DFAContentModel&);
|
---|
| 192 | DFAContentModel& operator=(const DFAContentModel&);
|
---|
| 193 |
|
---|
| 194 |
|
---|
| 195 | // -----------------------------------------------------------------------
|
---|
| 196 | // Private helper methods
|
---|
| 197 | // -----------------------------------------------------------------------
|
---|
| 198 | void buildDFA(ContentSpecNode* const curNode);
|
---|
| 199 | CMNode* buildSyntaxTree(ContentSpecNode* const curNode);
|
---|
| 200 | void calcFollowList(CMNode* const curNode);
|
---|
| 201 | unsigned int* makeDefStateList() const;
|
---|
| 202 | int postTreeBuildInit
|
---|
| 203 | (
|
---|
| 204 | CMNode* const nodeCur
|
---|
| 205 | , const unsigned int curIndex
|
---|
| 206 | );
|
---|
| 207 |
|
---|
| 208 |
|
---|
| 209 | // -----------------------------------------------------------------------
|
---|
| 210 | // Private data members
|
---|
| 211 | //
|
---|
| 212 | // fElemMap
|
---|
| 213 | // fElemMapSize
|
---|
| 214 | // This is the map of unique input symbol elements to indices into
|
---|
| 215 | // each state's per-input symbol transition table entry. This is part
|
---|
| 216 | // of the built DFA information that must be kept around to do the
|
---|
| 217 | // actual validation.
|
---|
| 218 | //
|
---|
| 219 | // fElemMapType
|
---|
| 220 | // This is a map of whether the element map contains information
|
---|
| 221 | // related to ANY models.
|
---|
| 222 | //
|
---|
| 223 | // fEmptyOk
|
---|
| 224 | // This is an optimization. While building the transition table we
|
---|
| 225 | // can see whether this content model would approve of an empty
|
---|
| 226 | // content (which could happen if everything was optional.) So we
|
---|
| 227 | // set this flag and short circuit that check, which would otherwise
|
---|
| 228 | // be ugly and time consuming if we tried to determine it at each
|
---|
| 229 | // validation call.
|
---|
| 230 | //
|
---|
| 231 | // fEOCPos
|
---|
| 232 | // The NFA position of the special EOC (end of content) node. This
|
---|
| 233 | // is saved away since its used during the DFA build.
|
---|
| 234 | //
|
---|
| 235 | // fFinalStateFlags
|
---|
| 236 | // This is an array of booleans, one per state (there are
|
---|
| 237 | // fTransTableSize states in the DFA) that indicates whether that
|
---|
| 238 | // state is a final state.
|
---|
| 239 | //
|
---|
| 240 | // fFollowList
|
---|
| 241 | // The list of follow positions for each NFA position (i.e. for each
|
---|
| 242 | // non-epsilon leaf node.) This is only used during the building of
|
---|
| 243 | // the DFA, and is let go afterwards.
|
---|
| 244 | //
|
---|
| 245 | // fHeadNode
|
---|
| 246 | // This is the head node of our intermediate representation. It is
|
---|
| 247 | // only non-null during the building of the DFA (just so that it
|
---|
| 248 | // does not have to be passed all around.) Once the DFA is built,
|
---|
| 249 | // this is no longer required so its deleted.
|
---|
| 250 | //
|
---|
| 251 | // fLeafCount
|
---|
| 252 | // The count of leaf nodes. This is an important number that set some
|
---|
| 253 | // limits on the sizes of data structures in the DFA process.
|
---|
| 254 | //
|
---|
| 255 | // fLeafList
|
---|
| 256 | // An array of non-epsilon leaf nodes, which is used during the DFA
|
---|
| 257 | // build operation, then dropped. These are just references to nodes
|
---|
| 258 | // pointed to by fHeadNode, so we don't have to clean them up, just
|
---|
| 259 | // the actually leaf list array itself needs cleanup.
|
---|
| 260 | //
|
---|
| 261 | // fLeafListType
|
---|
| 262 | // Array mapping ANY types to the leaf list.
|
---|
| 263 | //
|
---|
| 264 | // fTransTable
|
---|
| 265 | // fTransTableSize
|
---|
| 266 | // This is the transition table that is the main by product of all
|
---|
| 267 | // of the effort here. It is an array of arrays of ints. The first
|
---|
| 268 | // dimension is the number of states we end up with in the DFA. The
|
---|
| 269 | // second dimensions is the number of unique elements in the content
|
---|
| 270 | // model (fElemMapSize). Each entry in the second dimension indicates
|
---|
| 271 | // the new state given that input for the first dimension's start
|
---|
| 272 | // state.
|
---|
| 273 | //
|
---|
| 274 | // The fElemMap array handles mapping from element indexes to
|
---|
| 275 | // positions in the second dimension of the transition table.
|
---|
| 276 | //
|
---|
| 277 | // fTransTableSize is the number of valid entries in the transition
|
---|
| 278 | // table, and in the other related tables such as fFinalStateFlags.
|
---|
| 279 | //
|
---|
| 280 | // fDTD
|
---|
| 281 | // Boolean to allow DTDs to validate even with namespace support.
|
---|
| 282 | //
|
---|
| 283 | // fIsMixed
|
---|
| 284 | // DFA ContentModel with mixed PCDATA.
|
---|
| 285 | // -----------------------------------------------------------------------
|
---|
| 286 | QName** fElemMap;
|
---|
| 287 | ContentSpecNode::NodeTypes *fElemMapType;
|
---|
| 288 | unsigned int fElemMapSize;
|
---|
| 289 | bool fEmptyOk;
|
---|
| 290 | unsigned int fEOCPos;
|
---|
| 291 | bool* fFinalStateFlags;
|
---|
| 292 | CMStateSet** fFollowList;
|
---|
| 293 | CMNode* fHeadNode;
|
---|
| 294 | unsigned int fLeafCount;
|
---|
| 295 | CMLeaf** fLeafList;
|
---|
| 296 | ContentSpecNode::NodeTypes *fLeafListType;
|
---|
| 297 | unsigned int** fTransTable;
|
---|
| 298 | unsigned int fTransTableSize;
|
---|
| 299 | bool fDTD;
|
---|
| 300 | bool fIsMixed;
|
---|
| 301 | ContentLeafNameTypeVector *fLeafNameTypeVector;
|
---|
| 302 | MemoryManager* fMemoryManager;
|
---|
| 303 | };
|
---|
| 304 |
|
---|
| 305 |
|
---|
| 306 | inline unsigned int
|
---|
| 307 | DFAContentModel::getNextState(const unsigned int currentState,
|
---|
| 308 | const unsigned int elementIndex) const {
|
---|
| 309 |
|
---|
| 310 | if (currentState == XMLContentModel::gInvalidTrans) {
|
---|
| 311 | return XMLContentModel::gInvalidTrans;
|
---|
| 312 | }
|
---|
| 313 |
|
---|
| 314 | if (currentState >= fTransTableSize || elementIndex >= fElemMapSize) {
|
---|
| 315 | ThrowXMLwithMemMgr(ArrayIndexOutOfBoundsException, XMLExcepts::Array_BadIndex, fMemoryManager);
|
---|
| 316 | }
|
---|
| 317 |
|
---|
| 318 | return fTransTable[currentState][elementIndex];
|
---|
| 319 | }
|
---|
| 320 |
|
---|
| 321 | XERCES_CPP_NAMESPACE_END
|
---|
| 322 |
|
---|
| 323 | #endif
|
---|
| 324 |
|
---|