Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 1 | /* |
| 2 | * regexp.c: generic and extensible Regular Expression engine |
| 3 | * |
| 4 | * Basically designed with the purpose of compiling regexps for |
| 5 | * the variety of validation/shemas mechanisms now available in |
| 6 | * XML related specifications thise includes: |
| 7 | * - XML-1.0 DTD validation |
| 8 | * - XML Schemas structure part 1 |
| 9 | * - XML Schemas Datatypes part 2 especially Appendix F |
| 10 | * - RELAX-NG/TREX i.e. the counter proposal |
| 11 | * |
| 12 | * See Copyright for the status of this software. |
| 13 | * |
| 14 | * Daniel Veillard <veillard@redhat.com> |
| 15 | */ |
| 16 | |
| 17 | #define IN_LIBXML |
| 18 | #include "libxml.h" |
| 19 | |
| 20 | #ifdef LIBXML_REGEXP_ENABLED |
| 21 | |
| 22 | #include <stdio.h> |
| 23 | #include <string.h> |
| 24 | #include <libxml/tree.h> |
| 25 | #include <libxml/parserInternals.h> |
| 26 | #include <libxml/xmlregexp.h> |
| 27 | #include <libxml/xmlautomata.h> |
| 28 | #include <libxml/xmlunicode.h> |
| 29 | |
| 30 | /* #define DEBUG_REGEXP_GRAPH */ |
| 31 | /* #define DEBUG_REGEXP_EXEC */ |
| 32 | /* #define DEBUG_PUSH */ |
Daniel Veillard | 23e7357 | 2002-09-19 19:56:43 +0000 | [diff] [blame] | 33 | /* #define DEBUG_COMPACTION */ |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 34 | |
| 35 | #define ERROR(str) ctxt->error = 1; \ |
| 36 | xmlGenericError(xmlGenericErrorContext, "Regexp: %s: %s\n", str, ctxt->cur) |
| 37 | #define NEXT ctxt->cur++ |
| 38 | #define CUR (*(ctxt->cur)) |
| 39 | #define NXT(index) (ctxt->cur[index]) |
| 40 | |
| 41 | #define CUR_SCHAR(s, l) xmlStringCurrentChar(NULL, s, &l) |
| 42 | #define NEXTL(l) ctxt->cur += l; |
| 43 | |
Daniel Veillard | e19fc23 | 2002-04-22 16:01:24 +0000 | [diff] [blame] | 44 | /** |
| 45 | * TODO: |
| 46 | * |
| 47 | * macro to flag unimplemented blocks |
| 48 | */ |
| 49 | #define TODO \ |
| 50 | xmlGenericError(xmlGenericErrorContext, \ |
| 51 | "Unimplemented block at %s:%d\n", \ |
| 52 | __FILE__, __LINE__); |
| 53 | |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 54 | |
| 55 | /************************************************************************ |
| 56 | * * |
| 57 | * Datatypes and structures * |
| 58 | * * |
| 59 | ************************************************************************/ |
| 60 | |
| 61 | typedef enum { |
| 62 | XML_REGEXP_EPSILON = 1, |
| 63 | XML_REGEXP_CHARVAL, |
| 64 | XML_REGEXP_RANGES, |
| 65 | XML_REGEXP_SUBREG, |
| 66 | XML_REGEXP_STRING, |
| 67 | XML_REGEXP_ANYCHAR, /* . */ |
| 68 | XML_REGEXP_ANYSPACE, /* \s */ |
| 69 | XML_REGEXP_NOTSPACE, /* \S */ |
| 70 | XML_REGEXP_INITNAME, /* \l */ |
| 71 | XML_REGEXP_NOTINITNAME, /* \l */ |
| 72 | XML_REGEXP_NAMECHAR, /* \c */ |
| 73 | XML_REGEXP_NOTNAMECHAR, /* \C */ |
| 74 | XML_REGEXP_DECIMAL, /* \d */ |
| 75 | XML_REGEXP_NOTDECIMAL, /* \d */ |
| 76 | XML_REGEXP_REALCHAR, /* \w */ |
| 77 | XML_REGEXP_NOTREALCHAR, /* \w */ |
| 78 | XML_REGEXP_LETTER, |
| 79 | XML_REGEXP_LETTER_UPPERCASE, |
| 80 | XML_REGEXP_LETTER_LOWERCASE, |
| 81 | XML_REGEXP_LETTER_TITLECASE, |
| 82 | XML_REGEXP_LETTER_MODIFIER, |
| 83 | XML_REGEXP_LETTER_OTHERS, |
| 84 | XML_REGEXP_MARK, |
| 85 | XML_REGEXP_MARK_NONSPACING, |
| 86 | XML_REGEXP_MARK_SPACECOMBINING, |
| 87 | XML_REGEXP_MARK_ENCLOSING, |
| 88 | XML_REGEXP_NUMBER, |
| 89 | XML_REGEXP_NUMBER_DECIMAL, |
| 90 | XML_REGEXP_NUMBER_LETTER, |
| 91 | XML_REGEXP_NUMBER_OTHERS, |
| 92 | XML_REGEXP_PUNCT, |
| 93 | XML_REGEXP_PUNCT_CONNECTOR, |
| 94 | XML_REGEXP_PUNCT_DASH, |
| 95 | XML_REGEXP_PUNCT_OPEN, |
| 96 | XML_REGEXP_PUNCT_CLOSE, |
| 97 | XML_REGEXP_PUNCT_INITQUOTE, |
| 98 | XML_REGEXP_PUNCT_FINQUOTE, |
| 99 | XML_REGEXP_PUNCT_OTHERS, |
| 100 | XML_REGEXP_SEPAR, |
| 101 | XML_REGEXP_SEPAR_SPACE, |
| 102 | XML_REGEXP_SEPAR_LINE, |
| 103 | XML_REGEXP_SEPAR_PARA, |
| 104 | XML_REGEXP_SYMBOL, |
| 105 | XML_REGEXP_SYMBOL_MATH, |
| 106 | XML_REGEXP_SYMBOL_CURRENCY, |
| 107 | XML_REGEXP_SYMBOL_MODIFIER, |
| 108 | XML_REGEXP_SYMBOL_OTHERS, |
| 109 | XML_REGEXP_OTHER, |
| 110 | XML_REGEXP_OTHER_CONTROL, |
| 111 | XML_REGEXP_OTHER_FORMAT, |
| 112 | XML_REGEXP_OTHER_PRIVATE, |
| 113 | XML_REGEXP_OTHER_NA, |
| 114 | XML_REGEXP_BLOCK_NAME |
| 115 | } xmlRegAtomType; |
| 116 | |
| 117 | typedef enum { |
| 118 | XML_REGEXP_QUANT_EPSILON = 1, |
| 119 | XML_REGEXP_QUANT_ONCE, |
| 120 | XML_REGEXP_QUANT_OPT, |
| 121 | XML_REGEXP_QUANT_MULT, |
| 122 | XML_REGEXP_QUANT_PLUS, |
Daniel Veillard | 7646b18 | 2002-04-20 06:41:40 +0000 | [diff] [blame] | 123 | XML_REGEXP_QUANT_ONCEONLY, |
| 124 | XML_REGEXP_QUANT_ALL, |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 125 | XML_REGEXP_QUANT_RANGE |
| 126 | } xmlRegQuantType; |
| 127 | |
| 128 | typedef enum { |
| 129 | XML_REGEXP_START_STATE = 1, |
| 130 | XML_REGEXP_FINAL_STATE, |
| 131 | XML_REGEXP_TRANS_STATE |
| 132 | } xmlRegStateType; |
| 133 | |
| 134 | typedef enum { |
| 135 | XML_REGEXP_MARK_NORMAL = 0, |
| 136 | XML_REGEXP_MARK_START, |
| 137 | XML_REGEXP_MARK_VISITED |
| 138 | } xmlRegMarkedType; |
| 139 | |
| 140 | typedef struct _xmlRegRange xmlRegRange; |
| 141 | typedef xmlRegRange *xmlRegRangePtr; |
| 142 | |
| 143 | struct _xmlRegRange { |
| 144 | int neg; |
| 145 | xmlRegAtomType type; |
| 146 | int start; |
| 147 | int end; |
| 148 | xmlChar *blockName; |
| 149 | }; |
| 150 | |
| 151 | typedef struct _xmlRegAtom xmlRegAtom; |
| 152 | typedef xmlRegAtom *xmlRegAtomPtr; |
| 153 | |
| 154 | typedef struct _xmlAutomataState xmlRegState; |
| 155 | typedef xmlRegState *xmlRegStatePtr; |
| 156 | |
| 157 | struct _xmlRegAtom { |
| 158 | int no; |
| 159 | xmlRegAtomType type; |
| 160 | xmlRegQuantType quant; |
| 161 | int min; |
| 162 | int max; |
| 163 | |
| 164 | void *valuep; |
Daniel Veillard | a646cfd | 2002-09-17 21:50:03 +0000 | [diff] [blame] | 165 | void *valuep2; |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 166 | int neg; |
| 167 | int codepoint; |
| 168 | xmlRegStatePtr start; |
| 169 | xmlRegStatePtr stop; |
| 170 | int maxRanges; |
| 171 | int nbRanges; |
| 172 | xmlRegRangePtr *ranges; |
| 173 | void *data; |
| 174 | }; |
| 175 | |
| 176 | typedef struct _xmlRegCounter xmlRegCounter; |
| 177 | typedef xmlRegCounter *xmlRegCounterPtr; |
| 178 | |
| 179 | struct _xmlRegCounter { |
| 180 | int min; |
| 181 | int max; |
| 182 | }; |
| 183 | |
| 184 | typedef struct _xmlRegTrans xmlRegTrans; |
| 185 | typedef xmlRegTrans *xmlRegTransPtr; |
| 186 | |
| 187 | struct _xmlRegTrans { |
| 188 | xmlRegAtomPtr atom; |
| 189 | int to; |
| 190 | int counter; |
| 191 | int count; |
| 192 | }; |
| 193 | |
| 194 | struct _xmlAutomataState { |
| 195 | xmlRegStateType type; |
| 196 | xmlRegMarkedType mark; |
Daniel Veillard | 23e7357 | 2002-09-19 19:56:43 +0000 | [diff] [blame] | 197 | xmlRegMarkedType reached; |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 198 | int no; |
| 199 | |
| 200 | int maxTrans; |
| 201 | int nbTrans; |
| 202 | xmlRegTrans *trans; |
| 203 | }; |
| 204 | |
| 205 | typedef struct _xmlAutomata xmlRegParserCtxt; |
| 206 | typedef xmlRegParserCtxt *xmlRegParserCtxtPtr; |
| 207 | |
| 208 | struct _xmlAutomata { |
| 209 | xmlChar *string; |
| 210 | xmlChar *cur; |
| 211 | |
| 212 | int error; |
| 213 | int neg; |
| 214 | |
| 215 | xmlRegStatePtr start; |
| 216 | xmlRegStatePtr end; |
| 217 | xmlRegStatePtr state; |
| 218 | |
| 219 | xmlRegAtomPtr atom; |
| 220 | |
| 221 | int maxAtoms; |
| 222 | int nbAtoms; |
| 223 | xmlRegAtomPtr *atoms; |
| 224 | |
| 225 | int maxStates; |
| 226 | int nbStates; |
| 227 | xmlRegStatePtr *states; |
| 228 | |
| 229 | int maxCounters; |
| 230 | int nbCounters; |
| 231 | xmlRegCounter *counters; |
Daniel Veillard | e19fc23 | 2002-04-22 16:01:24 +0000 | [diff] [blame] | 232 | |
| 233 | int determinist; |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 234 | }; |
| 235 | |
| 236 | struct _xmlRegexp { |
| 237 | xmlChar *string; |
| 238 | int nbStates; |
| 239 | xmlRegStatePtr *states; |
| 240 | int nbAtoms; |
| 241 | xmlRegAtomPtr *atoms; |
| 242 | int nbCounters; |
| 243 | xmlRegCounter *counters; |
Daniel Veillard | e19fc23 | 2002-04-22 16:01:24 +0000 | [diff] [blame] | 244 | int determinist; |
Daniel Veillard | 23e7357 | 2002-09-19 19:56:43 +0000 | [diff] [blame] | 245 | /* |
| 246 | * That's the compact form for determinists automatas |
| 247 | */ |
| 248 | int nbstates; |
| 249 | int *compact; |
Daniel Veillard | 118aed7 | 2002-09-24 14:13:13 +0000 | [diff] [blame] | 250 | void **transdata; |
Daniel Veillard | 23e7357 | 2002-09-19 19:56:43 +0000 | [diff] [blame] | 251 | int nbstrings; |
| 252 | xmlChar **stringMap; |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 253 | }; |
| 254 | |
| 255 | typedef struct _xmlRegExecRollback xmlRegExecRollback; |
| 256 | typedef xmlRegExecRollback *xmlRegExecRollbackPtr; |
| 257 | |
| 258 | struct _xmlRegExecRollback { |
| 259 | xmlRegStatePtr state;/* the current state */ |
| 260 | int index; /* the index in the input stack */ |
| 261 | int nextbranch; /* the next transition to explore in that state */ |
| 262 | int *counts; /* save the automate state if it has some */ |
| 263 | }; |
| 264 | |
| 265 | typedef struct _xmlRegInputToken xmlRegInputToken; |
| 266 | typedef xmlRegInputToken *xmlRegInputTokenPtr; |
| 267 | |
| 268 | struct _xmlRegInputToken { |
| 269 | xmlChar *value; |
| 270 | void *data; |
| 271 | }; |
| 272 | |
| 273 | struct _xmlRegExecCtxt { |
| 274 | int status; /* execution status != 0 indicate an error */ |
| 275 | int determinist; /* did we found an inderterministic behaviour */ |
| 276 | xmlRegexpPtr comp; /* the compiled regexp */ |
| 277 | xmlRegExecCallbacks callback; |
| 278 | void *data; |
| 279 | |
| 280 | xmlRegStatePtr state;/* the current state */ |
| 281 | int transno; /* the current transition on that state */ |
| 282 | int transcount; /* the number of char in char counted transitions */ |
| 283 | |
| 284 | /* |
| 285 | * A stack of rollback states |
| 286 | */ |
| 287 | int maxRollbacks; |
| 288 | int nbRollbacks; |
| 289 | xmlRegExecRollback *rollbacks; |
| 290 | |
| 291 | /* |
| 292 | * The state of the automata if any |
| 293 | */ |
| 294 | int *counts; |
| 295 | |
| 296 | /* |
| 297 | * The input stack |
| 298 | */ |
| 299 | int inputStackMax; |
| 300 | int inputStackNr; |
| 301 | int index; |
| 302 | int *charStack; |
| 303 | const xmlChar *inputString; /* when operating on characters */ |
| 304 | xmlRegInputTokenPtr inputStack;/* when operating on strings */ |
| 305 | |
| 306 | }; |
| 307 | |
Daniel Veillard | 441bc32 | 2002-04-20 17:38:48 +0000 | [diff] [blame] | 308 | #define REGEXP_ALL_COUNTER 0x123456 |
| 309 | #define REGEXP_ALL_LAX_COUNTER 0x123457 |
Daniel Veillard | 7646b18 | 2002-04-20 06:41:40 +0000 | [diff] [blame] | 310 | |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 311 | static void xmlFAParseRegExp(xmlRegParserCtxtPtr ctxt, int top); |
Daniel Veillard | 23e7357 | 2002-09-19 19:56:43 +0000 | [diff] [blame] | 312 | static void xmlRegFreeState(xmlRegStatePtr state); |
| 313 | static void xmlRegFreeAtom(xmlRegAtomPtr atom); |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 314 | |
| 315 | /************************************************************************ |
| 316 | * * |
| 317 | * Allocation/Deallocation * |
| 318 | * * |
| 319 | ************************************************************************/ |
| 320 | |
Daniel Veillard | 23e7357 | 2002-09-19 19:56:43 +0000 | [diff] [blame] | 321 | static int xmlFAComputesDeterminism(xmlRegParserCtxtPtr ctxt); |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 322 | /** |
| 323 | * xmlRegEpxFromParse: |
| 324 | * @ctxt: the parser context used to build it |
| 325 | * |
| 326 | * Allocate a new regexp and fill it with the reult from the parser |
| 327 | * |
| 328 | * Returns the new regexp or NULL in case of error |
| 329 | */ |
| 330 | static xmlRegexpPtr |
| 331 | xmlRegEpxFromParse(xmlRegParserCtxtPtr ctxt) { |
| 332 | xmlRegexpPtr ret; |
| 333 | |
| 334 | ret = (xmlRegexpPtr) xmlMalloc(sizeof(xmlRegexp)); |
Daniel Veillard | a76fe5c | 2003-04-24 16:06:47 +0000 | [diff] [blame] | 335 | if (ret == NULL) { |
| 336 | xmlGenericError(xmlGenericErrorContext, |
| 337 | "out of memory compiling regexp\n"); |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 338 | return(NULL); |
Daniel Veillard | a76fe5c | 2003-04-24 16:06:47 +0000 | [diff] [blame] | 339 | } |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 340 | memset(ret, 0, sizeof(xmlRegexp)); |
| 341 | ret->string = ctxt->string; |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 342 | ret->nbStates = ctxt->nbStates; |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 343 | ret->states = ctxt->states; |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 344 | ret->nbAtoms = ctxt->nbAtoms; |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 345 | ret->atoms = ctxt->atoms; |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 346 | ret->nbCounters = ctxt->nbCounters; |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 347 | ret->counters = ctxt->counters; |
Daniel Veillard | e19fc23 | 2002-04-22 16:01:24 +0000 | [diff] [blame] | 348 | ret->determinist = ctxt->determinist; |
Daniel Veillard | 23e7357 | 2002-09-19 19:56:43 +0000 | [diff] [blame] | 349 | |
| 350 | if ((ret->determinist != 0) && |
| 351 | (ret->nbCounters == 0) && |
Daniel Veillard | 118aed7 | 2002-09-24 14:13:13 +0000 | [diff] [blame] | 352 | (ret->atoms != NULL) && |
Daniel Veillard | 23e7357 | 2002-09-19 19:56:43 +0000 | [diff] [blame] | 353 | (ret->atoms[0] != NULL) && |
| 354 | (ret->atoms[0]->type == XML_REGEXP_STRING)) { |
| 355 | int i, j, nbstates = 0, nbatoms = 0; |
| 356 | int *stateRemap; |
| 357 | int *stringRemap; |
| 358 | int *transitions; |
Daniel Veillard | 118aed7 | 2002-09-24 14:13:13 +0000 | [diff] [blame] | 359 | void **transdata; |
Daniel Veillard | 23e7357 | 2002-09-19 19:56:43 +0000 | [diff] [blame] | 360 | xmlChar **stringMap; |
| 361 | xmlChar *value; |
| 362 | |
| 363 | /* |
| 364 | * Switch to a compact representation |
| 365 | * 1/ counting the effective number of states left |
| 366 | * 2/ conting the unique number of atoms, and check that |
| 367 | * they are all of the string type |
| 368 | * 3/ build a table state x atom for the transitions |
| 369 | */ |
| 370 | |
| 371 | stateRemap = xmlMalloc(ret->nbStates * sizeof(int)); |
Daniel Veillard | a76fe5c | 2003-04-24 16:06:47 +0000 | [diff] [blame] | 372 | if (stateRemap == NULL) { |
| 373 | xmlGenericError(xmlGenericErrorContext, |
| 374 | "out of memory compiling regexp\n"); |
| 375 | xmlFree(ret); |
| 376 | return(NULL); |
| 377 | } |
Daniel Veillard | 23e7357 | 2002-09-19 19:56:43 +0000 | [diff] [blame] | 378 | for (i = 0;i < ret->nbStates;i++) { |
| 379 | if (ret->states[i] != NULL) { |
| 380 | stateRemap[i] = nbstates; |
| 381 | nbstates++; |
| 382 | } else { |
| 383 | stateRemap[i] = -1; |
| 384 | } |
| 385 | } |
| 386 | #ifdef DEBUG_COMPACTION |
| 387 | printf("Final: %d states\n", nbstates); |
| 388 | #endif |
| 389 | stringMap = xmlMalloc(ret->nbAtoms * sizeof(char *)); |
Daniel Veillard | a76fe5c | 2003-04-24 16:06:47 +0000 | [diff] [blame] | 390 | if (stringMap == NULL) { |
| 391 | xmlGenericError(xmlGenericErrorContext, |
| 392 | "out of memory compiling regexp\n"); |
| 393 | xmlFree(stateRemap); |
| 394 | xmlFree(ret); |
| 395 | return(NULL); |
| 396 | } |
Daniel Veillard | 23e7357 | 2002-09-19 19:56:43 +0000 | [diff] [blame] | 397 | stringRemap = xmlMalloc(ret->nbAtoms * sizeof(int)); |
Daniel Veillard | a76fe5c | 2003-04-24 16:06:47 +0000 | [diff] [blame] | 398 | if (stringRemap == NULL) { |
| 399 | xmlGenericError(xmlGenericErrorContext, |
| 400 | "out of memory compiling regexp\n"); |
| 401 | xmlFree(stringMap); |
| 402 | xmlFree(stateRemap); |
| 403 | xmlFree(ret); |
| 404 | return(NULL); |
| 405 | } |
Daniel Veillard | 23e7357 | 2002-09-19 19:56:43 +0000 | [diff] [blame] | 406 | for (i = 0;i < ret->nbAtoms;i++) { |
| 407 | if ((ret->atoms[i]->type == XML_REGEXP_STRING) && |
| 408 | (ret->atoms[i]->quant == XML_REGEXP_QUANT_ONCE)) { |
| 409 | value = ret->atoms[i]->valuep; |
| 410 | for (j = 0;j < nbatoms;j++) { |
| 411 | if (xmlStrEqual(stringMap[j], value)) { |
| 412 | stringRemap[i] = j; |
| 413 | break; |
| 414 | } |
| 415 | } |
| 416 | if (j >= nbatoms) { |
| 417 | stringRemap[i] = nbatoms; |
| 418 | stringMap[nbatoms] = xmlStrdup(value); |
Daniel Veillard | a76fe5c | 2003-04-24 16:06:47 +0000 | [diff] [blame] | 419 | if (stringMap[nbatoms] == NULL) { |
| 420 | for (i = 0;i < nbatoms;i++) |
| 421 | xmlFree(stringMap[i]); |
| 422 | xmlFree(stringRemap); |
| 423 | xmlFree(stringMap); |
| 424 | xmlFree(stateRemap); |
| 425 | xmlFree(ret); |
| 426 | return(NULL); |
| 427 | } |
Daniel Veillard | 23e7357 | 2002-09-19 19:56:43 +0000 | [diff] [blame] | 428 | nbatoms++; |
| 429 | } |
| 430 | } else { |
| 431 | xmlFree(stateRemap); |
| 432 | xmlFree(stringRemap); |
| 433 | for (i = 0;i < nbatoms;i++) |
| 434 | xmlFree(stringMap[i]); |
| 435 | xmlFree(stringMap); |
Daniel Veillard | a76fe5c | 2003-04-24 16:06:47 +0000 | [diff] [blame] | 436 | xmlFree(ret); |
| 437 | return(NULL); |
Daniel Veillard | 23e7357 | 2002-09-19 19:56:43 +0000 | [diff] [blame] | 438 | } |
| 439 | } |
| 440 | #ifdef DEBUG_COMPACTION |
| 441 | printf("Final: %d atoms\n", nbatoms); |
| 442 | #endif |
Daniel Veillard | a76fe5c | 2003-04-24 16:06:47 +0000 | [diff] [blame] | 443 | transitions = (int *) xmlMalloc((nbstates + 1) * |
| 444 | (nbatoms + 1) * sizeof(int)); |
| 445 | if (transitions == NULL) { |
| 446 | xmlFree(stateRemap); |
| 447 | xmlFree(stringRemap); |
| 448 | xmlFree(stringMap); |
| 449 | xmlFree(ret); |
| 450 | return(NULL); |
| 451 | } |
| 452 | memset(transitions, 0, (nbstates + 1) * (nbatoms + 1) * sizeof(int)); |
Daniel Veillard | 23e7357 | 2002-09-19 19:56:43 +0000 | [diff] [blame] | 453 | |
| 454 | /* |
| 455 | * Allocate the transition table. The first entry for each |
| 456 | * state correspond to the state type. |
| 457 | */ |
Daniel Veillard | 118aed7 | 2002-09-24 14:13:13 +0000 | [diff] [blame] | 458 | transdata = NULL; |
Daniel Veillard | 23e7357 | 2002-09-19 19:56:43 +0000 | [diff] [blame] | 459 | |
| 460 | for (i = 0;i < ret->nbStates;i++) { |
| 461 | int stateno, atomno, targetno, prev; |
| 462 | xmlRegStatePtr state; |
| 463 | xmlRegTransPtr trans; |
| 464 | |
| 465 | stateno = stateRemap[i]; |
| 466 | if (stateno == -1) |
| 467 | continue; |
| 468 | state = ret->states[i]; |
| 469 | |
| 470 | transitions[stateno * (nbatoms + 1)] = state->type; |
| 471 | |
| 472 | for (j = 0;j < state->nbTrans;j++) { |
| 473 | trans = &(state->trans[j]); |
| 474 | if ((trans->to == -1) || (trans->atom == NULL)) |
| 475 | continue; |
| 476 | atomno = stringRemap[trans->atom->no]; |
Daniel Veillard | 118aed7 | 2002-09-24 14:13:13 +0000 | [diff] [blame] | 477 | if ((trans->atom->data != NULL) && (transdata == NULL)) { |
| 478 | transdata = (void **) xmlMalloc(nbstates * nbatoms * |
| 479 | sizeof(void *)); |
| 480 | if (transdata != NULL) |
| 481 | memset(transdata, 0, |
| 482 | nbstates * nbatoms * sizeof(void *)); |
Daniel Veillard | a76fe5c | 2003-04-24 16:06:47 +0000 | [diff] [blame] | 483 | else { |
| 484 | xmlGenericError(xmlGenericErrorContext, |
| 485 | "out of memory compiling regexp\n"); |
| 486 | break; |
| 487 | } |
Daniel Veillard | 118aed7 | 2002-09-24 14:13:13 +0000 | [diff] [blame] | 488 | } |
Daniel Veillard | 23e7357 | 2002-09-19 19:56:43 +0000 | [diff] [blame] | 489 | targetno = stateRemap[trans->to]; |
| 490 | /* |
| 491 | * if the same atome can generate transition to 2 different |
| 492 | * states then it means the automata is not determinist and |
| 493 | * the compact form can't be used ! |
| 494 | */ |
| 495 | prev = transitions[stateno * (nbatoms + 1) + atomno + 1]; |
| 496 | if (prev != 0) { |
| 497 | if (prev != targetno + 1) { |
Daniel Veillard | 23e7357 | 2002-09-19 19:56:43 +0000 | [diff] [blame] | 498 | ret->determinist = 0; |
| 499 | #ifdef DEBUG_COMPACTION |
| 500 | printf("Indet: state %d trans %d, atom %d to %d : %d to %d\n", |
| 501 | i, j, trans->atom->no, trans->to, atomno, targetno); |
| 502 | printf(" previous to is %d\n", prev); |
| 503 | #endif |
| 504 | ret->determinist = 0; |
Daniel Veillard | 118aed7 | 2002-09-24 14:13:13 +0000 | [diff] [blame] | 505 | if (transdata != NULL) |
| 506 | xmlFree(transdata); |
Daniel Veillard | 23e7357 | 2002-09-19 19:56:43 +0000 | [diff] [blame] | 507 | xmlFree(transitions); |
| 508 | xmlFree(stateRemap); |
| 509 | xmlFree(stringRemap); |
| 510 | for (i = 0;i < nbatoms;i++) |
| 511 | xmlFree(stringMap[i]); |
| 512 | xmlFree(stringMap); |
Daniel Veillard | a76fe5c | 2003-04-24 16:06:47 +0000 | [diff] [blame] | 513 | goto not_determ; |
Daniel Veillard | 23e7357 | 2002-09-19 19:56:43 +0000 | [diff] [blame] | 514 | } |
| 515 | } else { |
| 516 | #if 0 |
| 517 | printf("State %d trans %d: atom %d to %d : %d to %d\n", |
| 518 | i, j, trans->atom->no, trans->to, atomno, targetno); |
| 519 | #endif |
| 520 | transitions[stateno * (nbatoms + 1) + atomno + 1] = |
Daniel Veillard | 118aed7 | 2002-09-24 14:13:13 +0000 | [diff] [blame] | 521 | targetno + 1; /* to avoid 0 */ |
| 522 | if (transdata != NULL) |
| 523 | transdata[stateno * nbatoms + atomno] = |
| 524 | trans->atom->data; |
Daniel Veillard | 23e7357 | 2002-09-19 19:56:43 +0000 | [diff] [blame] | 525 | } |
| 526 | } |
| 527 | } |
| 528 | ret->determinist = 1; |
| 529 | #ifdef DEBUG_COMPACTION |
| 530 | /* |
| 531 | * Debug |
| 532 | */ |
| 533 | for (i = 0;i < nbstates;i++) { |
| 534 | for (j = 0;j < nbatoms + 1;j++) { |
| 535 | printf("%02d ", transitions[i * (nbatoms + 1) + j]); |
| 536 | } |
| 537 | printf("\n"); |
| 538 | } |
| 539 | printf("\n"); |
| 540 | #endif |
| 541 | /* |
| 542 | * Cleanup of the old data |
| 543 | */ |
| 544 | if (ret->states != NULL) { |
| 545 | for (i = 0;i < ret->nbStates;i++) |
| 546 | xmlRegFreeState(ret->states[i]); |
| 547 | xmlFree(ret->states); |
| 548 | } |
| 549 | ret->states = NULL; |
| 550 | ret->nbStates = 0; |
| 551 | if (ret->atoms != NULL) { |
| 552 | for (i = 0;i < ret->nbAtoms;i++) |
| 553 | xmlRegFreeAtom(ret->atoms[i]); |
| 554 | xmlFree(ret->atoms); |
| 555 | } |
| 556 | ret->atoms = NULL; |
| 557 | ret->nbAtoms = 0; |
| 558 | |
| 559 | ret->compact = transitions; |
Daniel Veillard | 118aed7 | 2002-09-24 14:13:13 +0000 | [diff] [blame] | 560 | ret->transdata = transdata; |
Daniel Veillard | 23e7357 | 2002-09-19 19:56:43 +0000 | [diff] [blame] | 561 | ret->stringMap = stringMap; |
| 562 | ret->nbstrings = nbatoms; |
| 563 | ret->nbstates = nbstates; |
| 564 | xmlFree(stateRemap); |
| 565 | xmlFree(stringRemap); |
| 566 | } |
Daniel Veillard | a76fe5c | 2003-04-24 16:06:47 +0000 | [diff] [blame] | 567 | not_determ: |
| 568 | ctxt->string = NULL; |
| 569 | ctxt->nbStates = 0; |
| 570 | ctxt->states = NULL; |
| 571 | ctxt->nbAtoms = 0; |
| 572 | ctxt->atoms = NULL; |
| 573 | ctxt->nbCounters = 0; |
| 574 | ctxt->counters = NULL; |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 575 | return(ret); |
| 576 | } |
| 577 | |
| 578 | /** |
| 579 | * xmlRegNewParserCtxt: |
| 580 | * @string: the string to parse |
| 581 | * |
| 582 | * Allocate a new regexp parser context |
| 583 | * |
| 584 | * Returns the new context or NULL in case of error |
| 585 | */ |
| 586 | static xmlRegParserCtxtPtr |
| 587 | xmlRegNewParserCtxt(const xmlChar *string) { |
| 588 | xmlRegParserCtxtPtr ret; |
| 589 | |
| 590 | ret = (xmlRegParserCtxtPtr) xmlMalloc(sizeof(xmlRegParserCtxt)); |
| 591 | if (ret == NULL) |
| 592 | return(NULL); |
| 593 | memset(ret, 0, sizeof(xmlRegParserCtxt)); |
| 594 | if (string != NULL) |
| 595 | ret->string = xmlStrdup(string); |
| 596 | ret->cur = ret->string; |
| 597 | ret->neg = 0; |
| 598 | ret->error = 0; |
Daniel Veillard | e19fc23 | 2002-04-22 16:01:24 +0000 | [diff] [blame] | 599 | ret->determinist = -1; |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 600 | return(ret); |
| 601 | } |
| 602 | |
| 603 | /** |
| 604 | * xmlRegNewRange: |
| 605 | * @ctxt: the regexp parser context |
| 606 | * @neg: is that negative |
| 607 | * @type: the type of range |
| 608 | * @start: the start codepoint |
| 609 | * @end: the end codepoint |
| 610 | * |
| 611 | * Allocate a new regexp range |
| 612 | * |
| 613 | * Returns the new range or NULL in case of error |
| 614 | */ |
| 615 | static xmlRegRangePtr |
| 616 | xmlRegNewRange(xmlRegParserCtxtPtr ctxt, |
| 617 | int neg, xmlRegAtomType type, int start, int end) { |
| 618 | xmlRegRangePtr ret; |
| 619 | |
| 620 | ret = (xmlRegRangePtr) xmlMalloc(sizeof(xmlRegRange)); |
| 621 | if (ret == NULL) { |
| 622 | ERROR("failed to allocate regexp range"); |
| 623 | return(NULL); |
| 624 | } |
| 625 | ret->neg = neg; |
| 626 | ret->type = type; |
| 627 | ret->start = start; |
| 628 | ret->end = end; |
| 629 | return(ret); |
| 630 | } |
| 631 | |
| 632 | /** |
| 633 | * xmlRegFreeRange: |
| 634 | * @range: the regexp range |
| 635 | * |
| 636 | * Free a regexp range |
| 637 | */ |
| 638 | static void |
| 639 | xmlRegFreeRange(xmlRegRangePtr range) { |
| 640 | if (range == NULL) |
| 641 | return; |
| 642 | |
| 643 | if (range->blockName != NULL) |
| 644 | xmlFree(range->blockName); |
| 645 | xmlFree(range); |
| 646 | } |
| 647 | |
| 648 | /** |
| 649 | * xmlRegNewAtom: |
| 650 | * @ctxt: the regexp parser context |
| 651 | * @type: the type of atom |
| 652 | * |
| 653 | * Allocate a new regexp range |
| 654 | * |
| 655 | * Returns the new atom or NULL in case of error |
| 656 | */ |
| 657 | static xmlRegAtomPtr |
| 658 | xmlRegNewAtom(xmlRegParserCtxtPtr ctxt, xmlRegAtomType type) { |
| 659 | xmlRegAtomPtr ret; |
| 660 | |
| 661 | ret = (xmlRegAtomPtr) xmlMalloc(sizeof(xmlRegAtom)); |
| 662 | if (ret == NULL) { |
| 663 | ERROR("failed to allocate regexp atom"); |
| 664 | return(NULL); |
| 665 | } |
| 666 | memset(ret, 0, sizeof(xmlRegAtom)); |
| 667 | ret->type = type; |
| 668 | ret->quant = XML_REGEXP_QUANT_ONCE; |
| 669 | ret->min = 0; |
| 670 | ret->max = 0; |
| 671 | return(ret); |
| 672 | } |
| 673 | |
| 674 | /** |
| 675 | * xmlRegFreeAtom: |
| 676 | * @atom: the regexp atom |
| 677 | * |
| 678 | * Free a regexp atom |
| 679 | */ |
| 680 | static void |
| 681 | xmlRegFreeAtom(xmlRegAtomPtr atom) { |
| 682 | int i; |
| 683 | |
| 684 | if (atom == NULL) |
| 685 | return; |
| 686 | |
| 687 | for (i = 0;i < atom->nbRanges;i++) |
| 688 | xmlRegFreeRange(atom->ranges[i]); |
| 689 | if (atom->ranges != NULL) |
| 690 | xmlFree(atom->ranges); |
| 691 | if (atom->type == XML_REGEXP_STRING) |
| 692 | xmlFree(atom->valuep); |
| 693 | xmlFree(atom); |
| 694 | } |
| 695 | |
| 696 | static xmlRegStatePtr |
| 697 | xmlRegNewState(xmlRegParserCtxtPtr ctxt) { |
| 698 | xmlRegStatePtr ret; |
| 699 | |
| 700 | ret = (xmlRegStatePtr) xmlMalloc(sizeof(xmlRegState)); |
| 701 | if (ret == NULL) { |
| 702 | ERROR("failed to allocate regexp state"); |
| 703 | return(NULL); |
| 704 | } |
| 705 | memset(ret, 0, sizeof(xmlRegState)); |
| 706 | ret->type = XML_REGEXP_TRANS_STATE; |
| 707 | ret->mark = XML_REGEXP_MARK_NORMAL; |
| 708 | return(ret); |
| 709 | } |
| 710 | |
| 711 | /** |
| 712 | * xmlRegFreeState: |
| 713 | * @state: the regexp state |
| 714 | * |
| 715 | * Free a regexp state |
| 716 | */ |
| 717 | static void |
| 718 | xmlRegFreeState(xmlRegStatePtr state) { |
| 719 | if (state == NULL) |
| 720 | return; |
| 721 | |
| 722 | if (state->trans != NULL) |
| 723 | xmlFree(state->trans); |
| 724 | xmlFree(state); |
| 725 | } |
| 726 | |
| 727 | /** |
| 728 | * xmlRegFreeParserCtxt: |
| 729 | * @ctxt: the regexp parser context |
| 730 | * |
| 731 | * Free a regexp parser context |
| 732 | */ |
| 733 | static void |
| 734 | xmlRegFreeParserCtxt(xmlRegParserCtxtPtr ctxt) { |
| 735 | int i; |
| 736 | if (ctxt == NULL) |
| 737 | return; |
| 738 | |
| 739 | if (ctxt->string != NULL) |
| 740 | xmlFree(ctxt->string); |
| 741 | if (ctxt->states != NULL) { |
| 742 | for (i = 0;i < ctxt->nbStates;i++) |
| 743 | xmlRegFreeState(ctxt->states[i]); |
| 744 | xmlFree(ctxt->states); |
| 745 | } |
| 746 | if (ctxt->atoms != NULL) { |
| 747 | for (i = 0;i < ctxt->nbAtoms;i++) |
| 748 | xmlRegFreeAtom(ctxt->atoms[i]); |
| 749 | xmlFree(ctxt->atoms); |
| 750 | } |
| 751 | if (ctxt->counters != NULL) |
| 752 | xmlFree(ctxt->counters); |
| 753 | xmlFree(ctxt); |
| 754 | } |
| 755 | |
| 756 | /************************************************************************ |
| 757 | * * |
| 758 | * Display of Data structures * |
| 759 | * * |
| 760 | ************************************************************************/ |
| 761 | |
| 762 | static void |
| 763 | xmlRegPrintAtomType(FILE *output, xmlRegAtomType type) { |
| 764 | switch (type) { |
| 765 | case XML_REGEXP_EPSILON: |
| 766 | fprintf(output, "epsilon "); break; |
| 767 | case XML_REGEXP_CHARVAL: |
| 768 | fprintf(output, "charval "); break; |
| 769 | case XML_REGEXP_RANGES: |
| 770 | fprintf(output, "ranges "); break; |
| 771 | case XML_REGEXP_SUBREG: |
| 772 | fprintf(output, "subexpr "); break; |
| 773 | case XML_REGEXP_STRING: |
| 774 | fprintf(output, "string "); break; |
| 775 | case XML_REGEXP_ANYCHAR: |
| 776 | fprintf(output, "anychar "); break; |
| 777 | case XML_REGEXP_ANYSPACE: |
| 778 | fprintf(output, "anyspace "); break; |
| 779 | case XML_REGEXP_NOTSPACE: |
| 780 | fprintf(output, "notspace "); break; |
| 781 | case XML_REGEXP_INITNAME: |
| 782 | fprintf(output, "initname "); break; |
| 783 | case XML_REGEXP_NOTINITNAME: |
| 784 | fprintf(output, "notinitname "); break; |
| 785 | case XML_REGEXP_NAMECHAR: |
| 786 | fprintf(output, "namechar "); break; |
| 787 | case XML_REGEXP_NOTNAMECHAR: |
| 788 | fprintf(output, "notnamechar "); break; |
| 789 | case XML_REGEXP_DECIMAL: |
| 790 | fprintf(output, "decimal "); break; |
| 791 | case XML_REGEXP_NOTDECIMAL: |
| 792 | fprintf(output, "notdecimal "); break; |
| 793 | case XML_REGEXP_REALCHAR: |
| 794 | fprintf(output, "realchar "); break; |
| 795 | case XML_REGEXP_NOTREALCHAR: |
| 796 | fprintf(output, "notrealchar "); break; |
| 797 | case XML_REGEXP_LETTER: |
| 798 | fprintf(output, "LETTER "); break; |
| 799 | case XML_REGEXP_LETTER_UPPERCASE: |
| 800 | fprintf(output, "LETTER_UPPERCASE "); break; |
| 801 | case XML_REGEXP_LETTER_LOWERCASE: |
| 802 | fprintf(output, "LETTER_LOWERCASE "); break; |
| 803 | case XML_REGEXP_LETTER_TITLECASE: |
| 804 | fprintf(output, "LETTER_TITLECASE "); break; |
| 805 | case XML_REGEXP_LETTER_MODIFIER: |
| 806 | fprintf(output, "LETTER_MODIFIER "); break; |
| 807 | case XML_REGEXP_LETTER_OTHERS: |
| 808 | fprintf(output, "LETTER_OTHERS "); break; |
| 809 | case XML_REGEXP_MARK: |
| 810 | fprintf(output, "MARK "); break; |
| 811 | case XML_REGEXP_MARK_NONSPACING: |
| 812 | fprintf(output, "MARK_NONSPACING "); break; |
| 813 | case XML_REGEXP_MARK_SPACECOMBINING: |
| 814 | fprintf(output, "MARK_SPACECOMBINING "); break; |
| 815 | case XML_REGEXP_MARK_ENCLOSING: |
| 816 | fprintf(output, "MARK_ENCLOSING "); break; |
| 817 | case XML_REGEXP_NUMBER: |
| 818 | fprintf(output, "NUMBER "); break; |
| 819 | case XML_REGEXP_NUMBER_DECIMAL: |
| 820 | fprintf(output, "NUMBER_DECIMAL "); break; |
| 821 | case XML_REGEXP_NUMBER_LETTER: |
| 822 | fprintf(output, "NUMBER_LETTER "); break; |
| 823 | case XML_REGEXP_NUMBER_OTHERS: |
| 824 | fprintf(output, "NUMBER_OTHERS "); break; |
| 825 | case XML_REGEXP_PUNCT: |
| 826 | fprintf(output, "PUNCT "); break; |
| 827 | case XML_REGEXP_PUNCT_CONNECTOR: |
| 828 | fprintf(output, "PUNCT_CONNECTOR "); break; |
| 829 | case XML_REGEXP_PUNCT_DASH: |
| 830 | fprintf(output, "PUNCT_DASH "); break; |
| 831 | case XML_REGEXP_PUNCT_OPEN: |
| 832 | fprintf(output, "PUNCT_OPEN "); break; |
| 833 | case XML_REGEXP_PUNCT_CLOSE: |
| 834 | fprintf(output, "PUNCT_CLOSE "); break; |
| 835 | case XML_REGEXP_PUNCT_INITQUOTE: |
| 836 | fprintf(output, "PUNCT_INITQUOTE "); break; |
| 837 | case XML_REGEXP_PUNCT_FINQUOTE: |
| 838 | fprintf(output, "PUNCT_FINQUOTE "); break; |
| 839 | case XML_REGEXP_PUNCT_OTHERS: |
| 840 | fprintf(output, "PUNCT_OTHERS "); break; |
| 841 | case XML_REGEXP_SEPAR: |
| 842 | fprintf(output, "SEPAR "); break; |
| 843 | case XML_REGEXP_SEPAR_SPACE: |
| 844 | fprintf(output, "SEPAR_SPACE "); break; |
| 845 | case XML_REGEXP_SEPAR_LINE: |
| 846 | fprintf(output, "SEPAR_LINE "); break; |
| 847 | case XML_REGEXP_SEPAR_PARA: |
| 848 | fprintf(output, "SEPAR_PARA "); break; |
| 849 | case XML_REGEXP_SYMBOL: |
| 850 | fprintf(output, "SYMBOL "); break; |
| 851 | case XML_REGEXP_SYMBOL_MATH: |
| 852 | fprintf(output, "SYMBOL_MATH "); break; |
| 853 | case XML_REGEXP_SYMBOL_CURRENCY: |
| 854 | fprintf(output, "SYMBOL_CURRENCY "); break; |
| 855 | case XML_REGEXP_SYMBOL_MODIFIER: |
| 856 | fprintf(output, "SYMBOL_MODIFIER "); break; |
| 857 | case XML_REGEXP_SYMBOL_OTHERS: |
| 858 | fprintf(output, "SYMBOL_OTHERS "); break; |
| 859 | case XML_REGEXP_OTHER: |
| 860 | fprintf(output, "OTHER "); break; |
| 861 | case XML_REGEXP_OTHER_CONTROL: |
| 862 | fprintf(output, "OTHER_CONTROL "); break; |
| 863 | case XML_REGEXP_OTHER_FORMAT: |
| 864 | fprintf(output, "OTHER_FORMAT "); break; |
| 865 | case XML_REGEXP_OTHER_PRIVATE: |
| 866 | fprintf(output, "OTHER_PRIVATE "); break; |
| 867 | case XML_REGEXP_OTHER_NA: |
| 868 | fprintf(output, "OTHER_NA "); break; |
| 869 | case XML_REGEXP_BLOCK_NAME: |
| 870 | fprintf(output, "BLOCK "); break; |
| 871 | } |
| 872 | } |
| 873 | |
| 874 | static void |
| 875 | xmlRegPrintQuantType(FILE *output, xmlRegQuantType type) { |
| 876 | switch (type) { |
| 877 | case XML_REGEXP_QUANT_EPSILON: |
| 878 | fprintf(output, "epsilon "); break; |
| 879 | case XML_REGEXP_QUANT_ONCE: |
| 880 | fprintf(output, "once "); break; |
| 881 | case XML_REGEXP_QUANT_OPT: |
| 882 | fprintf(output, "? "); break; |
| 883 | case XML_REGEXP_QUANT_MULT: |
| 884 | fprintf(output, "* "); break; |
| 885 | case XML_REGEXP_QUANT_PLUS: |
| 886 | fprintf(output, "+ "); break; |
| 887 | case XML_REGEXP_QUANT_RANGE: |
| 888 | fprintf(output, "range "); break; |
Daniel Veillard | 7646b18 | 2002-04-20 06:41:40 +0000 | [diff] [blame] | 889 | case XML_REGEXP_QUANT_ONCEONLY: |
| 890 | fprintf(output, "onceonly "); break; |
| 891 | case XML_REGEXP_QUANT_ALL: |
| 892 | fprintf(output, "all "); break; |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 893 | } |
| 894 | } |
| 895 | static void |
| 896 | xmlRegPrintRange(FILE *output, xmlRegRangePtr range) { |
| 897 | fprintf(output, " range: "); |
| 898 | if (range->neg) |
| 899 | fprintf(output, "negative "); |
| 900 | xmlRegPrintAtomType(output, range->type); |
| 901 | fprintf(output, "%c - %c\n", range->start, range->end); |
| 902 | } |
| 903 | |
| 904 | static void |
| 905 | xmlRegPrintAtom(FILE *output, xmlRegAtomPtr atom) { |
| 906 | fprintf(output, " atom: "); |
| 907 | if (atom == NULL) { |
| 908 | fprintf(output, "NULL\n"); |
| 909 | return; |
| 910 | } |
| 911 | xmlRegPrintAtomType(output, atom->type); |
| 912 | xmlRegPrintQuantType(output, atom->quant); |
| 913 | if (atom->quant == XML_REGEXP_QUANT_RANGE) |
| 914 | fprintf(output, "%d-%d ", atom->min, atom->max); |
| 915 | if (atom->type == XML_REGEXP_STRING) |
| 916 | fprintf(output, "'%s' ", (char *) atom->valuep); |
| 917 | if (atom->type == XML_REGEXP_CHARVAL) |
| 918 | fprintf(output, "char %c\n", atom->codepoint); |
| 919 | else if (atom->type == XML_REGEXP_RANGES) { |
| 920 | int i; |
| 921 | fprintf(output, "%d entries\n", atom->nbRanges); |
| 922 | for (i = 0; i < atom->nbRanges;i++) |
| 923 | xmlRegPrintRange(output, atom->ranges[i]); |
| 924 | } else if (atom->type == XML_REGEXP_SUBREG) { |
| 925 | fprintf(output, "start %d end %d\n", atom->start->no, atom->stop->no); |
| 926 | } else { |
| 927 | fprintf(output, "\n"); |
| 928 | } |
| 929 | } |
| 930 | |
| 931 | static void |
| 932 | xmlRegPrintTrans(FILE *output, xmlRegTransPtr trans) { |
| 933 | fprintf(output, " trans: "); |
| 934 | if (trans == NULL) { |
| 935 | fprintf(output, "NULL\n"); |
| 936 | return; |
| 937 | } |
| 938 | if (trans->to < 0) { |
| 939 | fprintf(output, "removed\n"); |
| 940 | return; |
| 941 | } |
| 942 | if (trans->counter >= 0) { |
| 943 | fprintf(output, "counted %d, ", trans->counter); |
| 944 | } |
Daniel Veillard | 8a001f6 | 2002-04-20 07:24:11 +0000 | [diff] [blame] | 945 | if (trans->count == REGEXP_ALL_COUNTER) { |
| 946 | fprintf(output, "all transition, "); |
| 947 | } else if (trans->count >= 0) { |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 948 | fprintf(output, "count based %d, ", trans->count); |
| 949 | } |
| 950 | if (trans->atom == NULL) { |
| 951 | fprintf(output, "epsilon to %d\n", trans->to); |
| 952 | return; |
| 953 | } |
| 954 | if (trans->atom->type == XML_REGEXP_CHARVAL) |
| 955 | fprintf(output, "char %c ", trans->atom->codepoint); |
| 956 | fprintf(output, "atom %d, to %d\n", trans->atom->no, trans->to); |
| 957 | } |
| 958 | |
| 959 | static void |
| 960 | xmlRegPrintState(FILE *output, xmlRegStatePtr state) { |
| 961 | int i; |
| 962 | |
| 963 | fprintf(output, " state: "); |
| 964 | if (state == NULL) { |
| 965 | fprintf(output, "NULL\n"); |
| 966 | return; |
| 967 | } |
| 968 | if (state->type == XML_REGEXP_START_STATE) |
| 969 | fprintf(output, "START "); |
| 970 | if (state->type == XML_REGEXP_FINAL_STATE) |
| 971 | fprintf(output, "FINAL "); |
| 972 | |
| 973 | fprintf(output, "%d, %d transitions:\n", state->no, state->nbTrans); |
| 974 | for (i = 0;i < state->nbTrans; i++) { |
| 975 | xmlRegPrintTrans(output, &(state->trans[i])); |
| 976 | } |
| 977 | } |
| 978 | |
Daniel Veillard | 23e7357 | 2002-09-19 19:56:43 +0000 | [diff] [blame] | 979 | #ifdef DEBUG_REGEXP_GRAPH |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 980 | static void |
| 981 | xmlRegPrintCtxt(FILE *output, xmlRegParserCtxtPtr ctxt) { |
| 982 | int i; |
| 983 | |
| 984 | fprintf(output, " ctxt: "); |
| 985 | if (ctxt == NULL) { |
| 986 | fprintf(output, "NULL\n"); |
| 987 | return; |
| 988 | } |
| 989 | fprintf(output, "'%s' ", ctxt->string); |
| 990 | if (ctxt->error) |
| 991 | fprintf(output, "error "); |
| 992 | if (ctxt->neg) |
| 993 | fprintf(output, "neg "); |
| 994 | fprintf(output, "\n"); |
| 995 | fprintf(output, "%d atoms:\n", ctxt->nbAtoms); |
| 996 | for (i = 0;i < ctxt->nbAtoms; i++) { |
| 997 | fprintf(output, " %02d ", i); |
| 998 | xmlRegPrintAtom(output, ctxt->atoms[i]); |
| 999 | } |
| 1000 | if (ctxt->atom != NULL) { |
| 1001 | fprintf(output, "current atom:\n"); |
| 1002 | xmlRegPrintAtom(output, ctxt->atom); |
| 1003 | } |
| 1004 | fprintf(output, "%d states:", ctxt->nbStates); |
| 1005 | if (ctxt->start != NULL) |
| 1006 | fprintf(output, " start: %d", ctxt->start->no); |
| 1007 | if (ctxt->end != NULL) |
| 1008 | fprintf(output, " end: %d", ctxt->end->no); |
| 1009 | fprintf(output, "\n"); |
| 1010 | for (i = 0;i < ctxt->nbStates; i++) { |
| 1011 | xmlRegPrintState(output, ctxt->states[i]); |
| 1012 | } |
| 1013 | fprintf(output, "%d counters:\n", ctxt->nbCounters); |
| 1014 | for (i = 0;i < ctxt->nbCounters; i++) { |
| 1015 | fprintf(output, " %d: min %d max %d\n", i, ctxt->counters[i].min, |
| 1016 | ctxt->counters[i].max); |
| 1017 | } |
| 1018 | } |
Daniel Veillard | 23e7357 | 2002-09-19 19:56:43 +0000 | [diff] [blame] | 1019 | #endif |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 1020 | |
| 1021 | /************************************************************************ |
| 1022 | * * |
| 1023 | * Finite Automata structures manipulations * |
| 1024 | * * |
| 1025 | ************************************************************************/ |
| 1026 | |
| 1027 | static void |
| 1028 | xmlRegAtomAddRange(xmlRegParserCtxtPtr ctxt, xmlRegAtomPtr atom, |
| 1029 | int neg, xmlRegAtomType type, int start, int end, |
| 1030 | xmlChar *blockName) { |
| 1031 | xmlRegRangePtr range; |
| 1032 | |
| 1033 | if (atom == NULL) { |
| 1034 | ERROR("add range: atom is NULL"); |
| 1035 | return; |
| 1036 | } |
| 1037 | if (atom->type != XML_REGEXP_RANGES) { |
| 1038 | ERROR("add range: atom is not ranges"); |
| 1039 | return; |
| 1040 | } |
| 1041 | if (atom->maxRanges == 0) { |
| 1042 | atom->maxRanges = 4; |
| 1043 | atom->ranges = (xmlRegRangePtr *) xmlMalloc(atom->maxRanges * |
| 1044 | sizeof(xmlRegRangePtr)); |
| 1045 | if (atom->ranges == NULL) { |
| 1046 | ERROR("add range: allocation failed"); |
| 1047 | atom->maxRanges = 0; |
| 1048 | return; |
| 1049 | } |
| 1050 | } else if (atom->nbRanges >= atom->maxRanges) { |
| 1051 | xmlRegRangePtr *tmp; |
| 1052 | atom->maxRanges *= 2; |
| 1053 | tmp = (xmlRegRangePtr *) xmlRealloc(atom->ranges, atom->maxRanges * |
| 1054 | sizeof(xmlRegRangePtr)); |
| 1055 | if (tmp == NULL) { |
| 1056 | ERROR("add range: allocation failed"); |
| 1057 | atom->maxRanges /= 2; |
| 1058 | return; |
| 1059 | } |
| 1060 | atom->ranges = tmp; |
| 1061 | } |
| 1062 | range = xmlRegNewRange(ctxt, neg, type, start, end); |
| 1063 | if (range == NULL) |
| 1064 | return; |
| 1065 | range->blockName = blockName; |
| 1066 | atom->ranges[atom->nbRanges++] = range; |
| 1067 | |
| 1068 | } |
| 1069 | |
| 1070 | static int |
| 1071 | xmlRegGetCounter(xmlRegParserCtxtPtr ctxt) { |
| 1072 | if (ctxt->maxCounters == 0) { |
| 1073 | ctxt->maxCounters = 4; |
| 1074 | ctxt->counters = (xmlRegCounter *) xmlMalloc(ctxt->maxCounters * |
| 1075 | sizeof(xmlRegCounter)); |
| 1076 | if (ctxt->counters == NULL) { |
| 1077 | ERROR("reg counter: allocation failed"); |
| 1078 | ctxt->maxCounters = 0; |
| 1079 | return(-1); |
| 1080 | } |
| 1081 | } else if (ctxt->nbCounters >= ctxt->maxCounters) { |
| 1082 | xmlRegCounter *tmp; |
| 1083 | ctxt->maxCounters *= 2; |
| 1084 | tmp = (xmlRegCounter *) xmlRealloc(ctxt->counters, ctxt->maxCounters * |
| 1085 | sizeof(xmlRegCounter)); |
| 1086 | if (tmp == NULL) { |
| 1087 | ERROR("reg counter: allocation failed"); |
| 1088 | ctxt->maxCounters /= 2; |
| 1089 | return(-1); |
| 1090 | } |
| 1091 | ctxt->counters = tmp; |
| 1092 | } |
| 1093 | ctxt->counters[ctxt->nbCounters].min = -1; |
| 1094 | ctxt->counters[ctxt->nbCounters].max = -1; |
| 1095 | return(ctxt->nbCounters++); |
| 1096 | } |
| 1097 | |
Daniel Veillard | a76fe5c | 2003-04-24 16:06:47 +0000 | [diff] [blame] | 1098 | static int |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 1099 | xmlRegAtomPush(xmlRegParserCtxtPtr ctxt, xmlRegAtomPtr atom) { |
| 1100 | if (atom == NULL) { |
| 1101 | ERROR("atom push: atom is NULL"); |
Daniel Veillard | a76fe5c | 2003-04-24 16:06:47 +0000 | [diff] [blame] | 1102 | return(-1); |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 1103 | } |
| 1104 | if (ctxt->maxAtoms == 0) { |
| 1105 | ctxt->maxAtoms = 4; |
| 1106 | ctxt->atoms = (xmlRegAtomPtr *) xmlMalloc(ctxt->maxAtoms * |
| 1107 | sizeof(xmlRegAtomPtr)); |
| 1108 | if (ctxt->atoms == NULL) { |
| 1109 | ERROR("atom push: allocation failed"); |
| 1110 | ctxt->maxAtoms = 0; |
Daniel Veillard | a76fe5c | 2003-04-24 16:06:47 +0000 | [diff] [blame] | 1111 | return(-1); |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 1112 | } |
| 1113 | } else if (ctxt->nbAtoms >= ctxt->maxAtoms) { |
| 1114 | xmlRegAtomPtr *tmp; |
| 1115 | ctxt->maxAtoms *= 2; |
| 1116 | tmp = (xmlRegAtomPtr *) xmlRealloc(ctxt->atoms, ctxt->maxAtoms * |
| 1117 | sizeof(xmlRegAtomPtr)); |
| 1118 | if (tmp == NULL) { |
| 1119 | ERROR("atom push: allocation failed"); |
| 1120 | ctxt->maxAtoms /= 2; |
Daniel Veillard | a76fe5c | 2003-04-24 16:06:47 +0000 | [diff] [blame] | 1121 | return(-1); |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 1122 | } |
| 1123 | ctxt->atoms = tmp; |
| 1124 | } |
| 1125 | atom->no = ctxt->nbAtoms; |
| 1126 | ctxt->atoms[ctxt->nbAtoms++] = atom; |
Daniel Veillard | a76fe5c | 2003-04-24 16:06:47 +0000 | [diff] [blame] | 1127 | return(0); |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 1128 | } |
| 1129 | |
| 1130 | static void |
| 1131 | xmlRegStateAddTrans(xmlRegParserCtxtPtr ctxt, xmlRegStatePtr state, |
| 1132 | xmlRegAtomPtr atom, xmlRegStatePtr target, |
| 1133 | int counter, int count) { |
| 1134 | if (state == NULL) { |
| 1135 | ERROR("add state: state is NULL"); |
| 1136 | return; |
| 1137 | } |
| 1138 | if (target == NULL) { |
| 1139 | ERROR("add state: target is NULL"); |
| 1140 | return; |
| 1141 | } |
| 1142 | if (state->maxTrans == 0) { |
| 1143 | state->maxTrans = 4; |
| 1144 | state->trans = (xmlRegTrans *) xmlMalloc(state->maxTrans * |
| 1145 | sizeof(xmlRegTrans)); |
| 1146 | if (state->trans == NULL) { |
| 1147 | ERROR("add range: allocation failed"); |
| 1148 | state->maxTrans = 0; |
| 1149 | return; |
| 1150 | } |
| 1151 | } else if (state->nbTrans >= state->maxTrans) { |
| 1152 | xmlRegTrans *tmp; |
| 1153 | state->maxTrans *= 2; |
| 1154 | tmp = (xmlRegTrans *) xmlRealloc(state->trans, state->maxTrans * |
| 1155 | sizeof(xmlRegTrans)); |
| 1156 | if (tmp == NULL) { |
| 1157 | ERROR("add range: allocation failed"); |
| 1158 | state->maxTrans /= 2; |
| 1159 | return; |
| 1160 | } |
| 1161 | state->trans = tmp; |
| 1162 | } |
| 1163 | #ifdef DEBUG_REGEXP_GRAPH |
| 1164 | printf("Add trans from %d to %d ", state->no, target->no); |
Daniel Veillard | 8a001f6 | 2002-04-20 07:24:11 +0000 | [diff] [blame] | 1165 | if (count == REGEXP_ALL_COUNTER) |
| 1166 | printf("all transition"); |
Daniel Veillard | 4402ab4 | 2002-09-12 16:02:56 +0000 | [diff] [blame] | 1167 | else if (count >= 0) |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 1168 | printf("count based %d", count); |
| 1169 | else if (counter >= 0) |
| 1170 | printf("counted %d", counter); |
| 1171 | else if (atom == NULL) |
| 1172 | printf("epsilon transition"); |
| 1173 | printf("\n"); |
| 1174 | #endif |
| 1175 | |
| 1176 | state->trans[state->nbTrans].atom = atom; |
| 1177 | state->trans[state->nbTrans].to = target->no; |
| 1178 | state->trans[state->nbTrans].counter = counter; |
| 1179 | state->trans[state->nbTrans].count = count; |
| 1180 | state->nbTrans++; |
| 1181 | } |
| 1182 | |
Daniel Veillard | a76fe5c | 2003-04-24 16:06:47 +0000 | [diff] [blame] | 1183 | static int |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 1184 | xmlRegStatePush(xmlRegParserCtxtPtr ctxt, xmlRegStatePtr state) { |
Daniel Veillard | a76fe5c | 2003-04-24 16:06:47 +0000 | [diff] [blame] | 1185 | if (state == NULL) return(-1); |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 1186 | if (ctxt->maxStates == 0) { |
| 1187 | ctxt->maxStates = 4; |
| 1188 | ctxt->states = (xmlRegStatePtr *) xmlMalloc(ctxt->maxStates * |
| 1189 | sizeof(xmlRegStatePtr)); |
| 1190 | if (ctxt->states == NULL) { |
| 1191 | ERROR("add range: allocation failed"); |
| 1192 | ctxt->maxStates = 0; |
Daniel Veillard | a76fe5c | 2003-04-24 16:06:47 +0000 | [diff] [blame] | 1193 | return(-1); |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 1194 | } |
| 1195 | } else if (ctxt->nbStates >= ctxt->maxStates) { |
| 1196 | xmlRegStatePtr *tmp; |
| 1197 | ctxt->maxStates *= 2; |
| 1198 | tmp = (xmlRegStatePtr *) xmlRealloc(ctxt->states, ctxt->maxStates * |
| 1199 | sizeof(xmlRegStatePtr)); |
| 1200 | if (tmp == NULL) { |
| 1201 | ERROR("add range: allocation failed"); |
| 1202 | ctxt->maxStates /= 2; |
Daniel Veillard | a76fe5c | 2003-04-24 16:06:47 +0000 | [diff] [blame] | 1203 | return(-1); |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 1204 | } |
| 1205 | ctxt->states = tmp; |
| 1206 | } |
| 1207 | state->no = ctxt->nbStates; |
| 1208 | ctxt->states[ctxt->nbStates++] = state; |
Daniel Veillard | a76fe5c | 2003-04-24 16:06:47 +0000 | [diff] [blame] | 1209 | return(0); |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 1210 | } |
| 1211 | |
| 1212 | /** |
Daniel Veillard | 7646b18 | 2002-04-20 06:41:40 +0000 | [diff] [blame] | 1213 | * xmlFAGenerateAllTransition: |
Daniel Veillard | 441bc32 | 2002-04-20 17:38:48 +0000 | [diff] [blame] | 1214 | * @ctxt: a regexp parser context |
| 1215 | * @from: the from state |
| 1216 | * @to: the target state or NULL for building a new one |
| 1217 | * @lax: |
Daniel Veillard | 7646b18 | 2002-04-20 06:41:40 +0000 | [diff] [blame] | 1218 | * |
| 1219 | */ |
| 1220 | static void |
| 1221 | xmlFAGenerateAllTransition(xmlRegParserCtxtPtr ctxt, |
Daniel Veillard | 441bc32 | 2002-04-20 17:38:48 +0000 | [diff] [blame] | 1222 | xmlRegStatePtr from, xmlRegStatePtr to, |
| 1223 | int lax) { |
Daniel Veillard | 7646b18 | 2002-04-20 06:41:40 +0000 | [diff] [blame] | 1224 | if (to == NULL) { |
| 1225 | to = xmlRegNewState(ctxt); |
| 1226 | xmlRegStatePush(ctxt, to); |
| 1227 | ctxt->state = to; |
| 1228 | } |
Daniel Veillard | 441bc32 | 2002-04-20 17:38:48 +0000 | [diff] [blame] | 1229 | if (lax) |
| 1230 | xmlRegStateAddTrans(ctxt, from, NULL, to, -1, REGEXP_ALL_LAX_COUNTER); |
| 1231 | else |
| 1232 | xmlRegStateAddTrans(ctxt, from, NULL, to, -1, REGEXP_ALL_COUNTER); |
Daniel Veillard | 7646b18 | 2002-04-20 06:41:40 +0000 | [diff] [blame] | 1233 | } |
| 1234 | |
| 1235 | /** |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 1236 | * xmlFAGenerateEpsilonTransition: |
Daniel Veillard | 441bc32 | 2002-04-20 17:38:48 +0000 | [diff] [blame] | 1237 | * @ctxt: a regexp parser context |
| 1238 | * @from: the from state |
| 1239 | * @to: the target state or NULL for building a new one |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 1240 | * |
| 1241 | */ |
| 1242 | static void |
| 1243 | xmlFAGenerateEpsilonTransition(xmlRegParserCtxtPtr ctxt, |
| 1244 | xmlRegStatePtr from, xmlRegStatePtr to) { |
| 1245 | if (to == NULL) { |
| 1246 | to = xmlRegNewState(ctxt); |
| 1247 | xmlRegStatePush(ctxt, to); |
| 1248 | ctxt->state = to; |
| 1249 | } |
| 1250 | xmlRegStateAddTrans(ctxt, from, NULL, to, -1, -1); |
| 1251 | } |
| 1252 | |
| 1253 | /** |
| 1254 | * xmlFAGenerateCountedEpsilonTransition: |
Daniel Veillard | 441bc32 | 2002-04-20 17:38:48 +0000 | [diff] [blame] | 1255 | * @ctxt: a regexp parser context |
| 1256 | * @from: the from state |
| 1257 | * @to: the target state or NULL for building a new one |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 1258 | * counter: the counter for that transition |
| 1259 | * |
| 1260 | */ |
| 1261 | static void |
| 1262 | xmlFAGenerateCountedEpsilonTransition(xmlRegParserCtxtPtr ctxt, |
| 1263 | xmlRegStatePtr from, xmlRegStatePtr to, int counter) { |
| 1264 | if (to == NULL) { |
| 1265 | to = xmlRegNewState(ctxt); |
| 1266 | xmlRegStatePush(ctxt, to); |
| 1267 | ctxt->state = to; |
| 1268 | } |
| 1269 | xmlRegStateAddTrans(ctxt, from, NULL, to, counter, -1); |
| 1270 | } |
| 1271 | |
| 1272 | /** |
| 1273 | * xmlFAGenerateCountedTransition: |
Daniel Veillard | 441bc32 | 2002-04-20 17:38:48 +0000 | [diff] [blame] | 1274 | * @ctxt: a regexp parser context |
| 1275 | * @from: the from state |
| 1276 | * @to: the target state or NULL for building a new one |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 1277 | * counter: the counter for that transition |
| 1278 | * |
| 1279 | */ |
| 1280 | static void |
| 1281 | xmlFAGenerateCountedTransition(xmlRegParserCtxtPtr ctxt, |
| 1282 | xmlRegStatePtr from, xmlRegStatePtr to, int counter) { |
| 1283 | if (to == NULL) { |
| 1284 | to = xmlRegNewState(ctxt); |
| 1285 | xmlRegStatePush(ctxt, to); |
| 1286 | ctxt->state = to; |
| 1287 | } |
| 1288 | xmlRegStateAddTrans(ctxt, from, NULL, to, -1, counter); |
| 1289 | } |
| 1290 | |
| 1291 | /** |
| 1292 | * xmlFAGenerateTransitions: |
Daniel Veillard | 441bc32 | 2002-04-20 17:38:48 +0000 | [diff] [blame] | 1293 | * @ctxt: a regexp parser context |
| 1294 | * @from: the from state |
| 1295 | * @to: the target state or NULL for building a new one |
| 1296 | * @atom: the atom generating the transition |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 1297 | * |
Daniel Veillard | a76fe5c | 2003-04-24 16:06:47 +0000 | [diff] [blame] | 1298 | * Returns 0 if succes and -1 in case of error. |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 1299 | */ |
Daniel Veillard | a76fe5c | 2003-04-24 16:06:47 +0000 | [diff] [blame] | 1300 | static int |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 1301 | xmlFAGenerateTransitions(xmlRegParserCtxtPtr ctxt, xmlRegStatePtr from, |
| 1302 | xmlRegStatePtr to, xmlRegAtomPtr atom) { |
| 1303 | if (atom == NULL) { |
| 1304 | ERROR("genrate transition: atom == NULL"); |
Daniel Veillard | a76fe5c | 2003-04-24 16:06:47 +0000 | [diff] [blame] | 1305 | return(-1); |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 1306 | } |
| 1307 | if (atom->type == XML_REGEXP_SUBREG) { |
| 1308 | /* |
| 1309 | * this is a subexpression handling one should not need to |
| 1310 | * create a new node excep for XML_REGEXP_QUANT_RANGE. |
| 1311 | */ |
Daniel Veillard | a76fe5c | 2003-04-24 16:06:47 +0000 | [diff] [blame] | 1312 | if (xmlRegAtomPush(ctxt, atom) < 0) { |
| 1313 | return(-1); |
| 1314 | } |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 1315 | if ((to != NULL) && (atom->stop != to) && |
| 1316 | (atom->quant != XML_REGEXP_QUANT_RANGE)) { |
| 1317 | /* |
| 1318 | * Generate an epsilon transition to link to the target |
| 1319 | */ |
| 1320 | xmlFAGenerateEpsilonTransition(ctxt, atom->stop, to); |
| 1321 | } |
| 1322 | switch (atom->quant) { |
| 1323 | case XML_REGEXP_QUANT_OPT: |
| 1324 | atom->quant = XML_REGEXP_QUANT_ONCE; |
| 1325 | xmlFAGenerateEpsilonTransition(ctxt, atom->start, atom->stop); |
| 1326 | break; |
| 1327 | case XML_REGEXP_QUANT_MULT: |
| 1328 | atom->quant = XML_REGEXP_QUANT_ONCE; |
| 1329 | xmlFAGenerateEpsilonTransition(ctxt, atom->start, atom->stop); |
| 1330 | xmlFAGenerateEpsilonTransition(ctxt, atom->stop, atom->start); |
| 1331 | break; |
| 1332 | case XML_REGEXP_QUANT_PLUS: |
| 1333 | atom->quant = XML_REGEXP_QUANT_ONCE; |
| 1334 | xmlFAGenerateEpsilonTransition(ctxt, atom->stop, atom->start); |
| 1335 | break; |
| 1336 | case XML_REGEXP_QUANT_RANGE: { |
| 1337 | int counter; |
| 1338 | xmlRegStatePtr newstate; |
| 1339 | |
| 1340 | /* |
| 1341 | * This one is nasty: |
| 1342 | * 1/ register a new counter |
| 1343 | * 2/ register an epsilon transition associated to |
| 1344 | * this counter going from atom->stop to atom->start |
| 1345 | * 3/ create a new state |
| 1346 | * 4/ generate a counted transition from atom->stop to |
| 1347 | * that state |
| 1348 | */ |
| 1349 | counter = xmlRegGetCounter(ctxt); |
| 1350 | ctxt->counters[counter].min = atom->min - 1; |
| 1351 | ctxt->counters[counter].max = atom->max - 1; |
| 1352 | atom->min = 0; |
| 1353 | atom->max = 0; |
| 1354 | atom->quant = XML_REGEXP_QUANT_ONCE; |
| 1355 | xmlFAGenerateCountedEpsilonTransition(ctxt, atom->stop, |
| 1356 | atom->start, counter); |
| 1357 | if (to != NULL) { |
| 1358 | newstate = to; |
| 1359 | } else { |
| 1360 | newstate = xmlRegNewState(ctxt); |
| 1361 | xmlRegStatePush(ctxt, newstate); |
| 1362 | ctxt->state = newstate; |
| 1363 | } |
| 1364 | xmlFAGenerateCountedTransition(ctxt, atom->stop, |
| 1365 | newstate, counter); |
| 1366 | } |
| 1367 | default: |
| 1368 | break; |
| 1369 | } |
Daniel Veillard | a76fe5c | 2003-04-24 16:06:47 +0000 | [diff] [blame] | 1370 | return(0); |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 1371 | } else { |
| 1372 | if (to == NULL) { |
| 1373 | to = xmlRegNewState(ctxt); |
Daniel Veillard | a76fe5c | 2003-04-24 16:06:47 +0000 | [diff] [blame] | 1374 | if (to != NULL) |
| 1375 | xmlRegStatePush(ctxt, to); |
| 1376 | else { |
| 1377 | return(-1); |
| 1378 | } |
| 1379 | } |
| 1380 | if (xmlRegAtomPush(ctxt, atom) < 0) { |
| 1381 | return(-1); |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 1382 | } |
| 1383 | xmlRegStateAddTrans(ctxt, from, atom, to, -1, -1); |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 1384 | ctxt->state = to; |
| 1385 | } |
| 1386 | switch (atom->quant) { |
| 1387 | case XML_REGEXP_QUANT_OPT: |
| 1388 | atom->quant = XML_REGEXP_QUANT_ONCE; |
| 1389 | xmlFAGenerateEpsilonTransition(ctxt, from, to); |
| 1390 | break; |
| 1391 | case XML_REGEXP_QUANT_MULT: |
| 1392 | atom->quant = XML_REGEXP_QUANT_ONCE; |
| 1393 | xmlFAGenerateEpsilonTransition(ctxt, from, to); |
| 1394 | xmlRegStateAddTrans(ctxt, to, atom, to, -1, -1); |
| 1395 | break; |
| 1396 | case XML_REGEXP_QUANT_PLUS: |
| 1397 | atom->quant = XML_REGEXP_QUANT_ONCE; |
| 1398 | xmlRegStateAddTrans(ctxt, to, atom, to, -1, -1); |
| 1399 | break; |
| 1400 | default: |
| 1401 | break; |
| 1402 | } |
Daniel Veillard | a76fe5c | 2003-04-24 16:06:47 +0000 | [diff] [blame] | 1403 | return(0); |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 1404 | } |
| 1405 | |
| 1406 | /** |
| 1407 | * xmlFAReduceEpsilonTransitions: |
Daniel Veillard | 441bc32 | 2002-04-20 17:38:48 +0000 | [diff] [blame] | 1408 | * @ctxt: a regexp parser context |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 1409 | * @fromnr: the from state |
| 1410 | * @tonr: the to state |
| 1411 | * @cpunter: should that transition be associted to a counted |
| 1412 | * |
| 1413 | */ |
| 1414 | static void |
| 1415 | xmlFAReduceEpsilonTransitions(xmlRegParserCtxtPtr ctxt, int fromnr, |
| 1416 | int tonr, int counter) { |
| 1417 | int transnr; |
| 1418 | xmlRegStatePtr from; |
| 1419 | xmlRegStatePtr to; |
| 1420 | |
| 1421 | #ifdef DEBUG_REGEXP_GRAPH |
| 1422 | printf("xmlFAReduceEpsilonTransitions(%d, %d)\n", fromnr, tonr); |
| 1423 | #endif |
| 1424 | from = ctxt->states[fromnr]; |
| 1425 | if (from == NULL) |
| 1426 | return; |
| 1427 | to = ctxt->states[tonr]; |
| 1428 | if (to == NULL) |
| 1429 | return; |
| 1430 | if ((to->mark == XML_REGEXP_MARK_START) || |
| 1431 | (to->mark == XML_REGEXP_MARK_VISITED)) |
| 1432 | return; |
| 1433 | |
| 1434 | to->mark = XML_REGEXP_MARK_VISITED; |
| 1435 | if (to->type == XML_REGEXP_FINAL_STATE) { |
| 1436 | #ifdef DEBUG_REGEXP_GRAPH |
| 1437 | printf("State %d is final, so %d becomes final\n", tonr, fromnr); |
| 1438 | #endif |
| 1439 | from->type = XML_REGEXP_FINAL_STATE; |
| 1440 | } |
| 1441 | for (transnr = 0;transnr < to->nbTrans;transnr++) { |
| 1442 | if (to->trans[transnr].atom == NULL) { |
| 1443 | /* |
| 1444 | * Don't remove counted transitions |
| 1445 | * Don't loop either |
| 1446 | */ |
Daniel Veillard | b509f15 | 2002-04-17 16:28:10 +0000 | [diff] [blame] | 1447 | if (to->trans[transnr].to != fromnr) { |
| 1448 | if (to->trans[transnr].count >= 0) { |
| 1449 | int newto = to->trans[transnr].to; |
| 1450 | |
| 1451 | xmlRegStateAddTrans(ctxt, from, NULL, |
| 1452 | ctxt->states[newto], |
| 1453 | -1, to->trans[transnr].count); |
| 1454 | } else { |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 1455 | #ifdef DEBUG_REGEXP_GRAPH |
Daniel Veillard | b509f15 | 2002-04-17 16:28:10 +0000 | [diff] [blame] | 1456 | printf("Found epsilon trans %d from %d to %d\n", |
| 1457 | transnr, tonr, to->trans[transnr].to); |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 1458 | #endif |
Daniel Veillard | b509f15 | 2002-04-17 16:28:10 +0000 | [diff] [blame] | 1459 | if (to->trans[transnr].counter >= 0) { |
| 1460 | xmlFAReduceEpsilonTransitions(ctxt, fromnr, |
| 1461 | to->trans[transnr].to, |
| 1462 | to->trans[transnr].counter); |
| 1463 | } else { |
| 1464 | xmlFAReduceEpsilonTransitions(ctxt, fromnr, |
| 1465 | to->trans[transnr].to, |
| 1466 | counter); |
| 1467 | } |
| 1468 | } |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 1469 | } |
| 1470 | } else { |
| 1471 | int newto = to->trans[transnr].to; |
| 1472 | |
Daniel Veillard | b509f15 | 2002-04-17 16:28:10 +0000 | [diff] [blame] | 1473 | if (to->trans[transnr].counter >= 0) { |
| 1474 | xmlRegStateAddTrans(ctxt, from, to->trans[transnr].atom, |
| 1475 | ctxt->states[newto], |
| 1476 | to->trans[transnr].counter, -1); |
| 1477 | } else { |
| 1478 | xmlRegStateAddTrans(ctxt, from, to->trans[transnr].atom, |
| 1479 | ctxt->states[newto], counter, -1); |
| 1480 | } |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 1481 | } |
| 1482 | } |
| 1483 | to->mark = XML_REGEXP_MARK_NORMAL; |
| 1484 | } |
| 1485 | |
| 1486 | /** |
| 1487 | * xmlFAEliminateEpsilonTransitions: |
Daniel Veillard | 441bc32 | 2002-04-20 17:38:48 +0000 | [diff] [blame] | 1488 | * @ctxt: a regexp parser context |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 1489 | * |
| 1490 | */ |
| 1491 | static void |
| 1492 | xmlFAEliminateEpsilonTransitions(xmlRegParserCtxtPtr ctxt) { |
| 1493 | int statenr, transnr; |
| 1494 | xmlRegStatePtr state; |
| 1495 | |
Daniel Veillard | a76fe5c | 2003-04-24 16:06:47 +0000 | [diff] [blame] | 1496 | if (ctxt->states == NULL) return; |
| 1497 | |
| 1498 | |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 1499 | /* |
| 1500 | * build the completed transitions bypassing the epsilons |
| 1501 | * Use a marking algorithm to avoid loops |
| 1502 | */ |
| 1503 | for (statenr = 0;statenr < ctxt->nbStates;statenr++) { |
| 1504 | state = ctxt->states[statenr]; |
| 1505 | if (state == NULL) |
| 1506 | continue; |
| 1507 | for (transnr = 0;transnr < state->nbTrans;transnr++) { |
| 1508 | if ((state->trans[transnr].atom == NULL) && |
| 1509 | (state->trans[transnr].to >= 0)) { |
| 1510 | if (state->trans[transnr].to == statenr) { |
| 1511 | state->trans[transnr].to = -1; |
| 1512 | #ifdef DEBUG_REGEXP_GRAPH |
| 1513 | printf("Removed loopback epsilon trans %d on %d\n", |
| 1514 | transnr, statenr); |
| 1515 | #endif |
| 1516 | } else if (state->trans[transnr].count < 0) { |
| 1517 | int newto = state->trans[transnr].to; |
| 1518 | |
| 1519 | #ifdef DEBUG_REGEXP_GRAPH |
| 1520 | printf("Found epsilon trans %d from %d to %d\n", |
| 1521 | transnr, statenr, newto); |
| 1522 | #endif |
| 1523 | state->mark = XML_REGEXP_MARK_START; |
| 1524 | xmlFAReduceEpsilonTransitions(ctxt, statenr, |
| 1525 | newto, state->trans[transnr].counter); |
| 1526 | state->mark = XML_REGEXP_MARK_NORMAL; |
| 1527 | #ifdef DEBUG_REGEXP_GRAPH |
| 1528 | } else { |
| 1529 | printf("Found counted transition %d on %d\n", |
| 1530 | transnr, statenr); |
| 1531 | #endif |
| 1532 | } |
| 1533 | } |
| 1534 | } |
| 1535 | } |
| 1536 | /* |
| 1537 | * Eliminate the epsilon transitions |
| 1538 | */ |
| 1539 | for (statenr = 0;statenr < ctxt->nbStates;statenr++) { |
| 1540 | state = ctxt->states[statenr]; |
| 1541 | if (state == NULL) |
| 1542 | continue; |
| 1543 | for (transnr = 0;transnr < state->nbTrans;transnr++) { |
| 1544 | if ((state->trans[transnr].atom == NULL) && |
| 1545 | (state->trans[transnr].count < 0) && |
| 1546 | (state->trans[transnr].to >= 0)) { |
| 1547 | state->trans[transnr].to = -1; |
| 1548 | } |
| 1549 | } |
| 1550 | } |
Daniel Veillard | 23e7357 | 2002-09-19 19:56:43 +0000 | [diff] [blame] | 1551 | |
| 1552 | /* |
| 1553 | * Use this pass to detect unreachable states too |
| 1554 | */ |
| 1555 | for (statenr = 0;statenr < ctxt->nbStates;statenr++) { |
| 1556 | state = ctxt->states[statenr]; |
| 1557 | if (state != NULL) |
| 1558 | state->reached = 0; |
| 1559 | } |
| 1560 | state = ctxt->states[0]; |
| 1561 | if (state != NULL) |
| 1562 | state->reached = 1; |
| 1563 | while (state != NULL) { |
| 1564 | xmlRegStatePtr target = NULL; |
| 1565 | state->reached = 2; |
| 1566 | /* |
| 1567 | * Mark all state reachable from the current reachable state |
| 1568 | */ |
| 1569 | for (transnr = 0;transnr < state->nbTrans;transnr++) { |
| 1570 | if ((state->trans[transnr].to >= 0) && |
| 1571 | ((state->trans[transnr].atom != NULL) || |
| 1572 | (state->trans[transnr].count >= 0))) { |
| 1573 | int newto = state->trans[transnr].to; |
| 1574 | |
| 1575 | if (ctxt->states[newto] == NULL) |
| 1576 | continue; |
| 1577 | if (ctxt->states[newto]->reached == 0) { |
| 1578 | ctxt->states[newto]->reached = 1; |
| 1579 | target = ctxt->states[newto]; |
| 1580 | } |
| 1581 | } |
| 1582 | } |
| 1583 | /* |
| 1584 | * find the next accessible state not explored |
| 1585 | */ |
| 1586 | if (target == NULL) { |
| 1587 | for (statenr = 1;statenr < ctxt->nbStates;statenr++) { |
| 1588 | state = ctxt->states[statenr]; |
| 1589 | if ((state != NULL) && (state->reached == 1)) { |
| 1590 | target = state; |
| 1591 | break; |
| 1592 | } |
| 1593 | } |
| 1594 | } |
| 1595 | state = target; |
| 1596 | } |
| 1597 | for (statenr = 0;statenr < ctxt->nbStates;statenr++) { |
| 1598 | state = ctxt->states[statenr]; |
| 1599 | if ((state != NULL) && (state->reached == 0)) { |
| 1600 | #ifdef DEBUG_REGEXP_GRAPH |
| 1601 | printf("Removed unreachable state %d\n", statenr); |
| 1602 | #endif |
| 1603 | xmlRegFreeState(state); |
| 1604 | ctxt->states[statenr] = NULL; |
| 1605 | } |
| 1606 | } |
| 1607 | |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 1608 | } |
| 1609 | |
Daniel Veillard | e19fc23 | 2002-04-22 16:01:24 +0000 | [diff] [blame] | 1610 | /** |
| 1611 | * xmlFACompareAtoms: |
| 1612 | * @atom1: an atom |
| 1613 | * @atom2: an atom |
| 1614 | * |
| 1615 | * Compares two atoms to check whether they are equivatents |
| 1616 | * |
| 1617 | * Returns 1 if yes and 0 otherwise |
| 1618 | */ |
| 1619 | static int |
| 1620 | xmlFACompareAtoms(xmlRegAtomPtr atom1, xmlRegAtomPtr atom2) { |
| 1621 | if (atom1 == atom2) |
| 1622 | return(1); |
| 1623 | if ((atom1 == NULL) || (atom2 == NULL)) |
| 1624 | return(0); |
| 1625 | |
| 1626 | if (atom1->type != atom2->type) |
| 1627 | return(0); |
| 1628 | switch (atom1->type) { |
| 1629 | case XML_REGEXP_STRING: |
| 1630 | return(xmlStrEqual((xmlChar *)atom1->valuep, |
| 1631 | (xmlChar *)atom2->valuep)); |
| 1632 | case XML_REGEXP_EPSILON: |
| 1633 | return(1); |
| 1634 | case XML_REGEXP_CHARVAL: |
| 1635 | return(atom1->codepoint == atom2->codepoint); |
| 1636 | case XML_REGEXP_RANGES: |
| 1637 | TODO; |
| 1638 | return(0); |
| 1639 | default: |
| 1640 | break; |
| 1641 | } |
| 1642 | return(1); |
| 1643 | } |
| 1644 | |
| 1645 | /** |
| 1646 | * xmlFARecurseDeterminism: |
| 1647 | * @ctxt: a regexp parser context |
| 1648 | * |
| 1649 | * Check whether the associated regexp is determinist, |
| 1650 | * should be called after xmlFAEliminateEpsilonTransitions() |
| 1651 | * |
| 1652 | */ |
| 1653 | static int |
| 1654 | xmlFARecurseDeterminism(xmlRegParserCtxtPtr ctxt, xmlRegStatePtr state, |
| 1655 | int to, xmlRegAtomPtr atom) { |
| 1656 | int ret = 1; |
| 1657 | int transnr; |
| 1658 | xmlRegTransPtr t1; |
| 1659 | |
| 1660 | if (state == NULL) |
| 1661 | return(ret); |
| 1662 | for (transnr = 0;transnr < state->nbTrans;transnr++) { |
| 1663 | t1 = &(state->trans[transnr]); |
| 1664 | /* |
| 1665 | * check transitions conflicting with the one looked at |
| 1666 | */ |
| 1667 | if (t1->atom == NULL) { |
| 1668 | if (t1->to == -1) |
| 1669 | continue; |
| 1670 | ret = xmlFARecurseDeterminism(ctxt, ctxt->states[t1->to], |
| 1671 | to, atom); |
| 1672 | if (ret == 0) |
| 1673 | return(0); |
| 1674 | continue; |
| 1675 | } |
| 1676 | if (t1->to != to) |
| 1677 | continue; |
| 1678 | if (xmlFACompareAtoms(t1->atom, atom)) |
| 1679 | return(0); |
| 1680 | } |
| 1681 | return(ret); |
| 1682 | } |
| 1683 | |
| 1684 | /** |
| 1685 | * xmlFAComputesDeterminism: |
| 1686 | * @ctxt: a regexp parser context |
| 1687 | * |
| 1688 | * Check whether the associated regexp is determinist, |
| 1689 | * should be called after xmlFAEliminateEpsilonTransitions() |
| 1690 | * |
| 1691 | */ |
| 1692 | static int |
| 1693 | xmlFAComputesDeterminism(xmlRegParserCtxtPtr ctxt) { |
| 1694 | int statenr, transnr; |
| 1695 | xmlRegStatePtr state; |
| 1696 | xmlRegTransPtr t1, t2; |
| 1697 | int i; |
| 1698 | int ret = 1; |
| 1699 | |
Daniel Veillard | 4402ab4 | 2002-09-12 16:02:56 +0000 | [diff] [blame] | 1700 | #ifdef DEBUG_REGEXP_GRAPH |
| 1701 | printf("xmlFAComputesDeterminism\n"); |
| 1702 | xmlRegPrintCtxt(stdout, ctxt); |
| 1703 | #endif |
Daniel Veillard | e19fc23 | 2002-04-22 16:01:24 +0000 | [diff] [blame] | 1704 | if (ctxt->determinist != -1) |
| 1705 | return(ctxt->determinist); |
| 1706 | |
| 1707 | /* |
| 1708 | * Check for all states that there isn't 2 transitions |
| 1709 | * with the same atom and a different target. |
| 1710 | */ |
| 1711 | for (statenr = 0;statenr < ctxt->nbStates;statenr++) { |
| 1712 | state = ctxt->states[statenr]; |
| 1713 | if (state == NULL) |
| 1714 | continue; |
| 1715 | for (transnr = 0;transnr < state->nbTrans;transnr++) { |
| 1716 | t1 = &(state->trans[transnr]); |
| 1717 | /* |
| 1718 | * Determinism checks in case of counted or all transitions |
| 1719 | * will have to be handled separately |
| 1720 | */ |
| 1721 | if (t1->atom == NULL) |
| 1722 | continue; |
| 1723 | if (t1->to == -1) /* eliminated */ |
| 1724 | continue; |
| 1725 | for (i = 0;i < transnr;i++) { |
| 1726 | t2 = &(state->trans[i]); |
| 1727 | if (t2->to == -1) /* eliminated */ |
| 1728 | continue; |
| 1729 | if (t2->atom != NULL) { |
| 1730 | if (t1->to == t2->to) { |
| 1731 | if (xmlFACompareAtoms(t1->atom, t2->atom)) |
| 1732 | t2->to = -1; /* eliminate */ |
| 1733 | } else { |
| 1734 | /* not determinist ! */ |
| 1735 | if (xmlFACompareAtoms(t1->atom, t2->atom)) |
| 1736 | ret = 0; |
| 1737 | } |
| 1738 | } else if (t1->to != -1) { |
| 1739 | /* |
| 1740 | * do the closure in case of remaining specific |
| 1741 | * epsilon transitions like choices or all |
| 1742 | */ |
| 1743 | ret = xmlFARecurseDeterminism(ctxt, ctxt->states[t1->to], |
| 1744 | t2->to, t2->atom); |
| 1745 | if (ret == 0) |
| 1746 | return(0); |
| 1747 | } |
| 1748 | } |
| 1749 | if (ret == 0) |
| 1750 | break; |
| 1751 | } |
| 1752 | if (ret == 0) |
| 1753 | break; |
| 1754 | } |
| 1755 | ctxt->determinist = ret; |
| 1756 | return(ret); |
| 1757 | } |
| 1758 | |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 1759 | /************************************************************************ |
| 1760 | * * |
| 1761 | * Routines to check input against transition atoms * |
| 1762 | * * |
| 1763 | ************************************************************************/ |
| 1764 | |
| 1765 | static int |
| 1766 | xmlRegCheckCharacterRange(xmlRegAtomType type, int codepoint, int neg, |
| 1767 | int start, int end, const xmlChar *blockName) { |
| 1768 | int ret = 0; |
| 1769 | |
| 1770 | switch (type) { |
| 1771 | case XML_REGEXP_STRING: |
| 1772 | case XML_REGEXP_SUBREG: |
| 1773 | case XML_REGEXP_RANGES: |
| 1774 | case XML_REGEXP_EPSILON: |
| 1775 | return(-1); |
| 1776 | case XML_REGEXP_ANYCHAR: |
| 1777 | ret = ((codepoint != '\n') && (codepoint != '\r')); |
| 1778 | break; |
| 1779 | case XML_REGEXP_CHARVAL: |
| 1780 | ret = ((codepoint >= start) && (codepoint <= end)); |
| 1781 | break; |
| 1782 | case XML_REGEXP_NOTSPACE: |
| 1783 | neg = !neg; |
| 1784 | case XML_REGEXP_ANYSPACE: |
| 1785 | ret = ((codepoint == '\n') || (codepoint == '\r') || |
| 1786 | (codepoint == '\t') || (codepoint == ' ')); |
| 1787 | break; |
| 1788 | case XML_REGEXP_NOTINITNAME: |
| 1789 | neg = !neg; |
| 1790 | case XML_REGEXP_INITNAME: |
| 1791 | ret = (xmlIsLetter(codepoint) || |
| 1792 | (codepoint == '_') || (codepoint == ':')); |
| 1793 | break; |
| 1794 | case XML_REGEXP_NOTNAMECHAR: |
| 1795 | neg = !neg; |
| 1796 | case XML_REGEXP_NAMECHAR: |
| 1797 | ret = (xmlIsLetter(codepoint) || xmlIsDigit(codepoint) || |
| 1798 | (codepoint == '.') || (codepoint == '-') || |
| 1799 | (codepoint == '_') || (codepoint == ':') || |
| 1800 | xmlIsCombining(codepoint) || xmlIsExtender(codepoint)); |
| 1801 | break; |
| 1802 | case XML_REGEXP_NOTDECIMAL: |
| 1803 | neg = !neg; |
| 1804 | case XML_REGEXP_DECIMAL: |
| 1805 | ret = xmlUCSIsCatNd(codepoint); |
| 1806 | break; |
| 1807 | case XML_REGEXP_REALCHAR: |
| 1808 | neg = !neg; |
| 1809 | case XML_REGEXP_NOTREALCHAR: |
| 1810 | ret = xmlUCSIsCatP(codepoint); |
| 1811 | if (ret == 0) |
| 1812 | ret = xmlUCSIsCatZ(codepoint); |
| 1813 | if (ret == 0) |
| 1814 | ret = xmlUCSIsCatC(codepoint); |
| 1815 | break; |
| 1816 | case XML_REGEXP_LETTER: |
| 1817 | ret = xmlUCSIsCatL(codepoint); |
| 1818 | break; |
| 1819 | case XML_REGEXP_LETTER_UPPERCASE: |
| 1820 | ret = xmlUCSIsCatLu(codepoint); |
| 1821 | break; |
| 1822 | case XML_REGEXP_LETTER_LOWERCASE: |
| 1823 | ret = xmlUCSIsCatLl(codepoint); |
| 1824 | break; |
| 1825 | case XML_REGEXP_LETTER_TITLECASE: |
| 1826 | ret = xmlUCSIsCatLt(codepoint); |
| 1827 | break; |
| 1828 | case XML_REGEXP_LETTER_MODIFIER: |
| 1829 | ret = xmlUCSIsCatLm(codepoint); |
| 1830 | break; |
| 1831 | case XML_REGEXP_LETTER_OTHERS: |
| 1832 | ret = xmlUCSIsCatLo(codepoint); |
| 1833 | break; |
| 1834 | case XML_REGEXP_MARK: |
| 1835 | ret = xmlUCSIsCatM(codepoint); |
| 1836 | break; |
| 1837 | case XML_REGEXP_MARK_NONSPACING: |
| 1838 | ret = xmlUCSIsCatMn(codepoint); |
| 1839 | break; |
| 1840 | case XML_REGEXP_MARK_SPACECOMBINING: |
| 1841 | ret = xmlUCSIsCatMc(codepoint); |
| 1842 | break; |
| 1843 | case XML_REGEXP_MARK_ENCLOSING: |
| 1844 | ret = xmlUCSIsCatMe(codepoint); |
| 1845 | break; |
| 1846 | case XML_REGEXP_NUMBER: |
| 1847 | ret = xmlUCSIsCatN(codepoint); |
| 1848 | break; |
| 1849 | case XML_REGEXP_NUMBER_DECIMAL: |
| 1850 | ret = xmlUCSIsCatNd(codepoint); |
| 1851 | break; |
| 1852 | case XML_REGEXP_NUMBER_LETTER: |
| 1853 | ret = xmlUCSIsCatNl(codepoint); |
| 1854 | break; |
| 1855 | case XML_REGEXP_NUMBER_OTHERS: |
| 1856 | ret = xmlUCSIsCatNo(codepoint); |
| 1857 | break; |
| 1858 | case XML_REGEXP_PUNCT: |
| 1859 | ret = xmlUCSIsCatP(codepoint); |
| 1860 | break; |
| 1861 | case XML_REGEXP_PUNCT_CONNECTOR: |
| 1862 | ret = xmlUCSIsCatPc(codepoint); |
| 1863 | break; |
| 1864 | case XML_REGEXP_PUNCT_DASH: |
| 1865 | ret = xmlUCSIsCatPd(codepoint); |
| 1866 | break; |
| 1867 | case XML_REGEXP_PUNCT_OPEN: |
| 1868 | ret = xmlUCSIsCatPs(codepoint); |
| 1869 | break; |
| 1870 | case XML_REGEXP_PUNCT_CLOSE: |
| 1871 | ret = xmlUCSIsCatPe(codepoint); |
| 1872 | break; |
| 1873 | case XML_REGEXP_PUNCT_INITQUOTE: |
| 1874 | ret = xmlUCSIsCatPi(codepoint); |
| 1875 | break; |
| 1876 | case XML_REGEXP_PUNCT_FINQUOTE: |
| 1877 | ret = xmlUCSIsCatPf(codepoint); |
| 1878 | break; |
| 1879 | case XML_REGEXP_PUNCT_OTHERS: |
| 1880 | ret = xmlUCSIsCatPo(codepoint); |
| 1881 | break; |
| 1882 | case XML_REGEXP_SEPAR: |
| 1883 | ret = xmlUCSIsCatZ(codepoint); |
| 1884 | break; |
| 1885 | case XML_REGEXP_SEPAR_SPACE: |
| 1886 | ret = xmlUCSIsCatZs(codepoint); |
| 1887 | break; |
| 1888 | case XML_REGEXP_SEPAR_LINE: |
| 1889 | ret = xmlUCSIsCatZl(codepoint); |
| 1890 | break; |
| 1891 | case XML_REGEXP_SEPAR_PARA: |
| 1892 | ret = xmlUCSIsCatZp(codepoint); |
| 1893 | break; |
| 1894 | case XML_REGEXP_SYMBOL: |
| 1895 | ret = xmlUCSIsCatS(codepoint); |
| 1896 | break; |
| 1897 | case XML_REGEXP_SYMBOL_MATH: |
| 1898 | ret = xmlUCSIsCatSm(codepoint); |
| 1899 | break; |
| 1900 | case XML_REGEXP_SYMBOL_CURRENCY: |
| 1901 | ret = xmlUCSIsCatSc(codepoint); |
| 1902 | break; |
| 1903 | case XML_REGEXP_SYMBOL_MODIFIER: |
| 1904 | ret = xmlUCSIsCatSk(codepoint); |
| 1905 | break; |
| 1906 | case XML_REGEXP_SYMBOL_OTHERS: |
| 1907 | ret = xmlUCSIsCatSo(codepoint); |
| 1908 | break; |
| 1909 | case XML_REGEXP_OTHER: |
| 1910 | ret = xmlUCSIsCatC(codepoint); |
| 1911 | break; |
| 1912 | case XML_REGEXP_OTHER_CONTROL: |
| 1913 | ret = xmlUCSIsCatCc(codepoint); |
| 1914 | break; |
| 1915 | case XML_REGEXP_OTHER_FORMAT: |
| 1916 | ret = xmlUCSIsCatCf(codepoint); |
| 1917 | break; |
| 1918 | case XML_REGEXP_OTHER_PRIVATE: |
| 1919 | ret = xmlUCSIsCatCo(codepoint); |
| 1920 | break; |
| 1921 | case XML_REGEXP_OTHER_NA: |
| 1922 | /* ret = xmlUCSIsCatCn(codepoint); */ |
| 1923 | /* Seems it doesn't exist anymore in recent Unicode releases */ |
| 1924 | ret = 0; |
| 1925 | break; |
| 1926 | case XML_REGEXP_BLOCK_NAME: |
| 1927 | ret = xmlUCSIsBlock(codepoint, (const char *) blockName); |
| 1928 | break; |
| 1929 | } |
| 1930 | if (neg) |
| 1931 | return(!ret); |
| 1932 | return(ret); |
| 1933 | } |
| 1934 | |
| 1935 | static int |
| 1936 | xmlRegCheckCharacter(xmlRegAtomPtr atom, int codepoint) { |
| 1937 | int i, ret = 0; |
| 1938 | xmlRegRangePtr range; |
| 1939 | |
| 1940 | if ((atom == NULL) || (!xmlIsChar(codepoint))) |
| 1941 | return(-1); |
| 1942 | |
| 1943 | switch (atom->type) { |
| 1944 | case XML_REGEXP_SUBREG: |
| 1945 | case XML_REGEXP_EPSILON: |
| 1946 | return(-1); |
| 1947 | case XML_REGEXP_CHARVAL: |
| 1948 | return(codepoint == atom->codepoint); |
| 1949 | case XML_REGEXP_RANGES: { |
| 1950 | int accept = 0; |
| 1951 | for (i = 0;i < atom->nbRanges;i++) { |
| 1952 | range = atom->ranges[i]; |
| 1953 | if (range->neg) { |
| 1954 | ret = xmlRegCheckCharacterRange(range->type, codepoint, |
| 1955 | 0, range->start, range->end, |
| 1956 | range->blockName); |
| 1957 | if (ret != 0) |
| 1958 | return(0); /* excluded char */ |
| 1959 | } else { |
| 1960 | ret = xmlRegCheckCharacterRange(range->type, codepoint, |
| 1961 | 0, range->start, range->end, |
| 1962 | range->blockName); |
| 1963 | if (ret != 0) |
| 1964 | accept = 1; /* might still be excluded */ |
| 1965 | } |
| 1966 | } |
| 1967 | return(accept); |
| 1968 | } |
| 1969 | case XML_REGEXP_STRING: |
| 1970 | printf("TODO: XML_REGEXP_STRING\n"); |
| 1971 | return(-1); |
| 1972 | case XML_REGEXP_ANYCHAR: |
| 1973 | case XML_REGEXP_ANYSPACE: |
| 1974 | case XML_REGEXP_NOTSPACE: |
| 1975 | case XML_REGEXP_INITNAME: |
| 1976 | case XML_REGEXP_NOTINITNAME: |
| 1977 | case XML_REGEXP_NAMECHAR: |
| 1978 | case XML_REGEXP_NOTNAMECHAR: |
| 1979 | case XML_REGEXP_DECIMAL: |
| 1980 | case XML_REGEXP_NOTDECIMAL: |
| 1981 | case XML_REGEXP_REALCHAR: |
| 1982 | case XML_REGEXP_NOTREALCHAR: |
| 1983 | case XML_REGEXP_LETTER: |
| 1984 | case XML_REGEXP_LETTER_UPPERCASE: |
| 1985 | case XML_REGEXP_LETTER_LOWERCASE: |
| 1986 | case XML_REGEXP_LETTER_TITLECASE: |
| 1987 | case XML_REGEXP_LETTER_MODIFIER: |
| 1988 | case XML_REGEXP_LETTER_OTHERS: |
| 1989 | case XML_REGEXP_MARK: |
| 1990 | case XML_REGEXP_MARK_NONSPACING: |
| 1991 | case XML_REGEXP_MARK_SPACECOMBINING: |
| 1992 | case XML_REGEXP_MARK_ENCLOSING: |
| 1993 | case XML_REGEXP_NUMBER: |
| 1994 | case XML_REGEXP_NUMBER_DECIMAL: |
| 1995 | case XML_REGEXP_NUMBER_LETTER: |
| 1996 | case XML_REGEXP_NUMBER_OTHERS: |
| 1997 | case XML_REGEXP_PUNCT: |
| 1998 | case XML_REGEXP_PUNCT_CONNECTOR: |
| 1999 | case XML_REGEXP_PUNCT_DASH: |
| 2000 | case XML_REGEXP_PUNCT_OPEN: |
| 2001 | case XML_REGEXP_PUNCT_CLOSE: |
| 2002 | case XML_REGEXP_PUNCT_INITQUOTE: |
| 2003 | case XML_REGEXP_PUNCT_FINQUOTE: |
| 2004 | case XML_REGEXP_PUNCT_OTHERS: |
| 2005 | case XML_REGEXP_SEPAR: |
| 2006 | case XML_REGEXP_SEPAR_SPACE: |
| 2007 | case XML_REGEXP_SEPAR_LINE: |
| 2008 | case XML_REGEXP_SEPAR_PARA: |
| 2009 | case XML_REGEXP_SYMBOL: |
| 2010 | case XML_REGEXP_SYMBOL_MATH: |
| 2011 | case XML_REGEXP_SYMBOL_CURRENCY: |
| 2012 | case XML_REGEXP_SYMBOL_MODIFIER: |
| 2013 | case XML_REGEXP_SYMBOL_OTHERS: |
| 2014 | case XML_REGEXP_OTHER: |
| 2015 | case XML_REGEXP_OTHER_CONTROL: |
| 2016 | case XML_REGEXP_OTHER_FORMAT: |
| 2017 | case XML_REGEXP_OTHER_PRIVATE: |
| 2018 | case XML_REGEXP_OTHER_NA: |
| 2019 | case XML_REGEXP_BLOCK_NAME: |
| 2020 | ret = xmlRegCheckCharacterRange(atom->type, codepoint, 0, 0, 0, |
| 2021 | (const xmlChar *)atom->valuep); |
| 2022 | if (atom->neg) |
| 2023 | ret = !ret; |
| 2024 | break; |
| 2025 | } |
| 2026 | return(ret); |
| 2027 | } |
| 2028 | |
| 2029 | /************************************************************************ |
| 2030 | * * |
| 2031 | * Saving an restoring state of an execution context * |
| 2032 | * * |
| 2033 | ************************************************************************/ |
| 2034 | |
| 2035 | #ifdef DEBUG_REGEXP_EXEC |
| 2036 | static void |
| 2037 | xmlFARegDebugExec(xmlRegExecCtxtPtr exec) { |
| 2038 | printf("state: %d:%d:idx %d", exec->state->no, exec->transno, exec->index); |
| 2039 | if (exec->inputStack != NULL) { |
| 2040 | int i; |
| 2041 | printf(": "); |
| 2042 | for (i = 0;(i < 3) && (i < exec->inputStackNr);i++) |
| 2043 | printf("%s ", exec->inputStack[exec->inputStackNr - (i + 1)]); |
| 2044 | } else { |
| 2045 | printf(": %s", &(exec->inputString[exec->index])); |
| 2046 | } |
| 2047 | printf("\n"); |
| 2048 | } |
| 2049 | #endif |
| 2050 | |
| 2051 | static void |
| 2052 | xmlFARegExecSave(xmlRegExecCtxtPtr exec) { |
| 2053 | #ifdef DEBUG_REGEXP_EXEC |
| 2054 | printf("saving "); |
| 2055 | exec->transno++; |
| 2056 | xmlFARegDebugExec(exec); |
| 2057 | exec->transno--; |
| 2058 | #endif |
| 2059 | |
| 2060 | if (exec->maxRollbacks == 0) { |
| 2061 | exec->maxRollbacks = 4; |
| 2062 | exec->rollbacks = (xmlRegExecRollback *) xmlMalloc(exec->maxRollbacks * |
| 2063 | sizeof(xmlRegExecRollback)); |
| 2064 | if (exec->rollbacks == NULL) { |
| 2065 | fprintf(stderr, "exec save: allocation failed"); |
| 2066 | exec->maxRollbacks = 0; |
| 2067 | return; |
| 2068 | } |
| 2069 | memset(exec->rollbacks, 0, |
| 2070 | exec->maxRollbacks * sizeof(xmlRegExecRollback)); |
| 2071 | } else if (exec->nbRollbacks >= exec->maxRollbacks) { |
| 2072 | xmlRegExecRollback *tmp; |
| 2073 | int len = exec->maxRollbacks; |
| 2074 | |
| 2075 | exec->maxRollbacks *= 2; |
| 2076 | tmp = (xmlRegExecRollback *) xmlRealloc(exec->rollbacks, |
| 2077 | exec->maxRollbacks * sizeof(xmlRegExecRollback)); |
| 2078 | if (tmp == NULL) { |
| 2079 | fprintf(stderr, "exec save: allocation failed"); |
| 2080 | exec->maxRollbacks /= 2; |
| 2081 | return; |
| 2082 | } |
| 2083 | exec->rollbacks = tmp; |
| 2084 | tmp = &exec->rollbacks[len]; |
| 2085 | memset(tmp, 0, (exec->maxRollbacks - len) * sizeof(xmlRegExecRollback)); |
| 2086 | } |
| 2087 | exec->rollbacks[exec->nbRollbacks].state = exec->state; |
| 2088 | exec->rollbacks[exec->nbRollbacks].index = exec->index; |
| 2089 | exec->rollbacks[exec->nbRollbacks].nextbranch = exec->transno + 1; |
| 2090 | if (exec->comp->nbCounters > 0) { |
| 2091 | if (exec->rollbacks[exec->nbRollbacks].counts == NULL) { |
| 2092 | exec->rollbacks[exec->nbRollbacks].counts = (int *) |
| 2093 | xmlMalloc(exec->comp->nbCounters * sizeof(int)); |
| 2094 | if (exec->rollbacks[exec->nbRollbacks].counts == NULL) { |
| 2095 | fprintf(stderr, "exec save: allocation failed"); |
| 2096 | exec->status = -5; |
| 2097 | return; |
| 2098 | } |
| 2099 | } |
| 2100 | memcpy(exec->rollbacks[exec->nbRollbacks].counts, exec->counts, |
| 2101 | exec->comp->nbCounters * sizeof(int)); |
| 2102 | } |
| 2103 | exec->nbRollbacks++; |
| 2104 | } |
| 2105 | |
| 2106 | static void |
| 2107 | xmlFARegExecRollBack(xmlRegExecCtxtPtr exec) { |
| 2108 | if (exec->nbRollbacks <= 0) { |
| 2109 | exec->status = -1; |
| 2110 | #ifdef DEBUG_REGEXP_EXEC |
| 2111 | printf("rollback failed on empty stack\n"); |
| 2112 | #endif |
| 2113 | return; |
| 2114 | } |
| 2115 | exec->nbRollbacks--; |
| 2116 | exec->state = exec->rollbacks[exec->nbRollbacks].state; |
| 2117 | exec->index = exec->rollbacks[exec->nbRollbacks].index; |
| 2118 | exec->transno = exec->rollbacks[exec->nbRollbacks].nextbranch; |
| 2119 | if (exec->comp->nbCounters > 0) { |
| 2120 | if (exec->rollbacks[exec->nbRollbacks].counts == NULL) { |
| 2121 | fprintf(stderr, "exec save: allocation failed"); |
| 2122 | exec->status = -6; |
| 2123 | return; |
| 2124 | } |
| 2125 | memcpy(exec->counts, exec->rollbacks[exec->nbRollbacks].counts, |
| 2126 | exec->comp->nbCounters * sizeof(int)); |
| 2127 | } |
| 2128 | |
| 2129 | #ifdef DEBUG_REGEXP_EXEC |
| 2130 | printf("restored "); |
| 2131 | xmlFARegDebugExec(exec); |
| 2132 | #endif |
| 2133 | } |
| 2134 | |
| 2135 | /************************************************************************ |
| 2136 | * * |
| 2137 | * Verifyer, running an input against a compiled regexp * |
| 2138 | * * |
| 2139 | ************************************************************************/ |
| 2140 | |
| 2141 | static int |
| 2142 | xmlFARegExec(xmlRegexpPtr comp, const xmlChar *content) { |
| 2143 | xmlRegExecCtxt execval; |
| 2144 | xmlRegExecCtxtPtr exec = &execval; |
| 2145 | int ret, codepoint, len; |
| 2146 | |
| 2147 | exec->inputString = content; |
| 2148 | exec->index = 0; |
| 2149 | exec->determinist = 1; |
| 2150 | exec->maxRollbacks = 0; |
| 2151 | exec->nbRollbacks = 0; |
| 2152 | exec->rollbacks = NULL; |
| 2153 | exec->status = 0; |
| 2154 | exec->comp = comp; |
| 2155 | exec->state = comp->states[0]; |
| 2156 | exec->transno = 0; |
| 2157 | exec->transcount = 0; |
| 2158 | if (comp->nbCounters > 0) { |
| 2159 | exec->counts = (int *) xmlMalloc(comp->nbCounters * sizeof(int)); |
| 2160 | if (exec->counts == NULL) |
| 2161 | return(-1); |
| 2162 | memset(exec->counts, 0, comp->nbCounters * sizeof(int)); |
| 2163 | } else |
| 2164 | exec->counts = NULL; |
| 2165 | while ((exec->status == 0) && |
| 2166 | ((exec->inputString[exec->index] != 0) || |
| 2167 | (exec->state->type != XML_REGEXP_FINAL_STATE))) { |
| 2168 | xmlRegTransPtr trans; |
| 2169 | xmlRegAtomPtr atom; |
| 2170 | |
| 2171 | /* |
| 2172 | * End of input on non-terminal state, rollback, however we may |
| 2173 | * still have epsilon like transition for counted transitions |
| 2174 | * on counters, in that case don't break too early. |
| 2175 | */ |
| 2176 | if ((exec->inputString[exec->index] == 0) && (exec->counts == NULL)) |
| 2177 | goto rollback; |
| 2178 | |
| 2179 | exec->transcount = 0; |
| 2180 | for (;exec->transno < exec->state->nbTrans;exec->transno++) { |
| 2181 | trans = &exec->state->trans[exec->transno]; |
| 2182 | if (trans->to < 0) |
| 2183 | continue; |
| 2184 | atom = trans->atom; |
| 2185 | ret = 0; |
| 2186 | if (trans->count >= 0) { |
| 2187 | int count; |
| 2188 | xmlRegCounterPtr counter; |
| 2189 | |
| 2190 | /* |
| 2191 | * A counted transition. |
| 2192 | */ |
| 2193 | |
| 2194 | count = exec->counts[trans->count]; |
| 2195 | counter = &exec->comp->counters[trans->count]; |
| 2196 | #ifdef DEBUG_REGEXP_EXEC |
| 2197 | printf("testing count %d: val %d, min %d, max %d\n", |
| 2198 | trans->count, count, counter->min, counter->max); |
| 2199 | #endif |
| 2200 | ret = ((count >= counter->min) && (count <= counter->max)); |
| 2201 | } else if (atom == NULL) { |
| 2202 | fprintf(stderr, "epsilon transition left at runtime\n"); |
| 2203 | exec->status = -2; |
| 2204 | break; |
| 2205 | } else if (exec->inputString[exec->index] != 0) { |
| 2206 | codepoint = CUR_SCHAR(&(exec->inputString[exec->index]), len); |
| 2207 | ret = xmlRegCheckCharacter(atom, codepoint); |
| 2208 | if ((ret == 1) && (atom->min > 0) && (atom->max > 0)) { |
| 2209 | xmlRegStatePtr to = comp->states[trans->to]; |
| 2210 | |
| 2211 | /* |
| 2212 | * this is a multiple input sequence |
| 2213 | */ |
| 2214 | if (exec->state->nbTrans > exec->transno + 1) { |
| 2215 | xmlFARegExecSave(exec); |
| 2216 | } |
| 2217 | exec->transcount = 1; |
| 2218 | do { |
| 2219 | /* |
| 2220 | * Try to progress as much as possible on the input |
| 2221 | */ |
| 2222 | if (exec->transcount == atom->max) { |
| 2223 | break; |
| 2224 | } |
| 2225 | exec->index += len; |
| 2226 | /* |
| 2227 | * End of input: stop here |
| 2228 | */ |
| 2229 | if (exec->inputString[exec->index] == 0) { |
| 2230 | exec->index -= len; |
| 2231 | break; |
| 2232 | } |
| 2233 | if (exec->transcount >= atom->min) { |
| 2234 | int transno = exec->transno; |
| 2235 | xmlRegStatePtr state = exec->state; |
| 2236 | |
| 2237 | /* |
| 2238 | * The transition is acceptable save it |
| 2239 | */ |
| 2240 | exec->transno = -1; /* trick */ |
| 2241 | exec->state = to; |
| 2242 | xmlFARegExecSave(exec); |
| 2243 | exec->transno = transno; |
| 2244 | exec->state = state; |
| 2245 | } |
| 2246 | codepoint = CUR_SCHAR(&(exec->inputString[exec->index]), |
| 2247 | len); |
| 2248 | ret = xmlRegCheckCharacter(atom, codepoint); |
| 2249 | exec->transcount++; |
| 2250 | } while (ret == 1); |
| 2251 | if (exec->transcount < atom->min) |
| 2252 | ret = 0; |
| 2253 | |
| 2254 | /* |
| 2255 | * If the last check failed but one transition was found |
| 2256 | * possible, rollback |
| 2257 | */ |
| 2258 | if (ret < 0) |
| 2259 | ret = 0; |
| 2260 | if (ret == 0) { |
| 2261 | goto rollback; |
| 2262 | } |
| 2263 | } |
| 2264 | } |
| 2265 | if (ret == 1) { |
| 2266 | if (exec->state->nbTrans > exec->transno + 1) { |
| 2267 | xmlFARegExecSave(exec); |
| 2268 | } |
| 2269 | if (trans->counter >= 0) { |
| 2270 | #ifdef DEBUG_REGEXP_EXEC |
| 2271 | printf("Increasing count %d\n", trans->counter); |
| 2272 | #endif |
| 2273 | exec->counts[trans->counter]++; |
| 2274 | } |
| 2275 | #ifdef DEBUG_REGEXP_EXEC |
| 2276 | printf("entering state %d\n", trans->to); |
| 2277 | #endif |
| 2278 | exec->state = comp->states[trans->to]; |
| 2279 | exec->transno = 0; |
| 2280 | if (trans->atom != NULL) { |
| 2281 | exec->index += len; |
| 2282 | } |
| 2283 | goto progress; |
| 2284 | } else if (ret < 0) { |
| 2285 | exec->status = -4; |
| 2286 | break; |
| 2287 | } |
| 2288 | } |
| 2289 | if ((exec->transno != 0) || (exec->state->nbTrans == 0)) { |
| 2290 | rollback: |
| 2291 | /* |
| 2292 | * Failed to find a way out |
| 2293 | */ |
| 2294 | exec->determinist = 0; |
| 2295 | xmlFARegExecRollBack(exec); |
| 2296 | } |
| 2297 | progress: |
| 2298 | continue; |
| 2299 | } |
| 2300 | if (exec->rollbacks != NULL) { |
| 2301 | if (exec->counts != NULL) { |
| 2302 | int i; |
| 2303 | |
| 2304 | for (i = 0;i < exec->maxRollbacks;i++) |
| 2305 | if (exec->rollbacks[i].counts != NULL) |
| 2306 | xmlFree(exec->rollbacks[i].counts); |
| 2307 | } |
| 2308 | xmlFree(exec->rollbacks); |
| 2309 | } |
| 2310 | if (exec->counts != NULL) |
| 2311 | xmlFree(exec->counts); |
| 2312 | if (exec->status == 0) |
| 2313 | return(1); |
| 2314 | if (exec->status == -1) |
| 2315 | return(0); |
| 2316 | return(exec->status); |
| 2317 | } |
| 2318 | |
| 2319 | /************************************************************************ |
| 2320 | * * |
| 2321 | * Progressive interface to the verifyer one atom at a time * |
| 2322 | * * |
| 2323 | ************************************************************************/ |
| 2324 | |
| 2325 | /** |
Daniel Veillard | 01c13b5 | 2002-12-10 15:19:08 +0000 | [diff] [blame] | 2326 | * xmlRegNewExecCtxt: |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 2327 | * @comp: a precompiled regular expression |
| 2328 | * @callback: a callback function used for handling progresses in the |
| 2329 | * automata matching phase |
| 2330 | * @data: the context data associated to the callback in this context |
| 2331 | * |
| 2332 | * Build a context used for progressive evaluation of a regexp. |
Daniel Veillard | 01c13b5 | 2002-12-10 15:19:08 +0000 | [diff] [blame] | 2333 | * |
| 2334 | * Returns the new context |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 2335 | */ |
| 2336 | xmlRegExecCtxtPtr |
| 2337 | xmlRegNewExecCtxt(xmlRegexpPtr comp, xmlRegExecCallbacks callback, void *data) { |
| 2338 | xmlRegExecCtxtPtr exec; |
| 2339 | |
| 2340 | if (comp == NULL) |
| 2341 | return(NULL); |
Daniel Veillard | a76fe5c | 2003-04-24 16:06:47 +0000 | [diff] [blame] | 2342 | if ((comp->compact == NULL) && (comp->states == NULL)) |
| 2343 | return(NULL); |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 2344 | exec = (xmlRegExecCtxtPtr) xmlMalloc(sizeof(xmlRegExecCtxt)); |
| 2345 | if (exec == NULL) { |
| 2346 | return(NULL); |
| 2347 | } |
| 2348 | memset(exec, 0, sizeof(xmlRegExecCtxt)); |
| 2349 | exec->inputString = NULL; |
| 2350 | exec->index = 0; |
| 2351 | exec->determinist = 1; |
| 2352 | exec->maxRollbacks = 0; |
| 2353 | exec->nbRollbacks = 0; |
| 2354 | exec->rollbacks = NULL; |
| 2355 | exec->status = 0; |
| 2356 | exec->comp = comp; |
Daniel Veillard | 23e7357 | 2002-09-19 19:56:43 +0000 | [diff] [blame] | 2357 | if (comp->compact == NULL) |
| 2358 | exec->state = comp->states[0]; |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 2359 | exec->transno = 0; |
| 2360 | exec->transcount = 0; |
| 2361 | exec->callback = callback; |
| 2362 | exec->data = data; |
| 2363 | if (comp->nbCounters > 0) { |
| 2364 | exec->counts = (int *) xmlMalloc(comp->nbCounters * sizeof(int)); |
| 2365 | if (exec->counts == NULL) { |
| 2366 | xmlFree(exec); |
| 2367 | return(NULL); |
| 2368 | } |
| 2369 | memset(exec->counts, 0, comp->nbCounters * sizeof(int)); |
| 2370 | } else |
| 2371 | exec->counts = NULL; |
| 2372 | exec->inputStackMax = 0; |
| 2373 | exec->inputStackNr = 0; |
| 2374 | exec->inputStack = NULL; |
| 2375 | return(exec); |
| 2376 | } |
| 2377 | |
| 2378 | /** |
| 2379 | * xmlRegFreeExecCtxt: |
| 2380 | * @exec: a regular expression evaulation context |
| 2381 | * |
| 2382 | * Free the structures associated to a regular expression evaulation context. |
| 2383 | */ |
| 2384 | void |
| 2385 | xmlRegFreeExecCtxt(xmlRegExecCtxtPtr exec) { |
| 2386 | if (exec == NULL) |
| 2387 | return; |
| 2388 | |
| 2389 | if (exec->rollbacks != NULL) { |
| 2390 | if (exec->counts != NULL) { |
| 2391 | int i; |
| 2392 | |
| 2393 | for (i = 0;i < exec->maxRollbacks;i++) |
| 2394 | if (exec->rollbacks[i].counts != NULL) |
| 2395 | xmlFree(exec->rollbacks[i].counts); |
| 2396 | } |
| 2397 | xmlFree(exec->rollbacks); |
| 2398 | } |
| 2399 | if (exec->counts != NULL) |
| 2400 | xmlFree(exec->counts); |
| 2401 | if (exec->inputStack != NULL) { |
| 2402 | int i; |
| 2403 | |
Daniel Veillard | 3237023 | 2002-10-16 14:08:14 +0000 | [diff] [blame] | 2404 | for (i = 0;i < exec->inputStackNr;i++) { |
| 2405 | if (exec->inputStack[i].value != NULL) |
| 2406 | xmlFree(exec->inputStack[i].value); |
| 2407 | } |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 2408 | xmlFree(exec->inputStack); |
| 2409 | } |
| 2410 | xmlFree(exec); |
| 2411 | } |
| 2412 | |
| 2413 | static void |
| 2414 | xmlFARegExecSaveInputString(xmlRegExecCtxtPtr exec, const xmlChar *value, |
| 2415 | void *data) { |
| 2416 | #ifdef DEBUG_PUSH |
| 2417 | printf("saving value: %d:%s\n", exec->inputStackNr, value); |
| 2418 | #endif |
| 2419 | if (exec->inputStackMax == 0) { |
| 2420 | exec->inputStackMax = 4; |
| 2421 | exec->inputStack = (xmlRegInputTokenPtr) |
| 2422 | xmlMalloc(exec->inputStackMax * sizeof(xmlRegInputToken)); |
| 2423 | if (exec->inputStack == NULL) { |
| 2424 | fprintf(stderr, "push input: allocation failed"); |
| 2425 | exec->inputStackMax = 0; |
| 2426 | return; |
| 2427 | } |
| 2428 | } else if (exec->inputStackNr + 1 >= exec->inputStackMax) { |
| 2429 | xmlRegInputTokenPtr tmp; |
| 2430 | |
| 2431 | exec->inputStackMax *= 2; |
| 2432 | tmp = (xmlRegInputTokenPtr) xmlRealloc(exec->inputStack, |
| 2433 | exec->inputStackMax * sizeof(xmlRegInputToken)); |
| 2434 | if (tmp == NULL) { |
| 2435 | fprintf(stderr, "push input: allocation failed"); |
| 2436 | exec->inputStackMax /= 2; |
| 2437 | return; |
| 2438 | } |
| 2439 | exec->inputStack = tmp; |
| 2440 | } |
| 2441 | exec->inputStack[exec->inputStackNr].value = xmlStrdup(value); |
| 2442 | exec->inputStack[exec->inputStackNr].data = data; |
| 2443 | exec->inputStackNr++; |
| 2444 | exec->inputStack[exec->inputStackNr].value = NULL; |
| 2445 | exec->inputStack[exec->inputStackNr].data = NULL; |
| 2446 | } |
| 2447 | |
| 2448 | |
| 2449 | /** |
Daniel Veillard | 23e7357 | 2002-09-19 19:56:43 +0000 | [diff] [blame] | 2450 | * xmlRegCompactPushString: |
| 2451 | * @exec: a regexp execution context |
| 2452 | * @comp: the precompiled exec with a compact table |
| 2453 | * @value: a string token input |
| 2454 | * @data: data associated to the token to reuse in callbacks |
| 2455 | * |
| 2456 | * Push one input token in the execution context |
| 2457 | * |
| 2458 | * Returns: 1 if the regexp reached a final state, 0 if non-final, and |
| 2459 | * a negative value in case of error. |
| 2460 | */ |
| 2461 | static int |
| 2462 | xmlRegCompactPushString(xmlRegExecCtxtPtr exec, |
| 2463 | xmlRegexpPtr comp, |
| 2464 | const xmlChar *value, |
| 2465 | void *data) { |
| 2466 | int state = exec->index; |
| 2467 | int i, target; |
| 2468 | |
| 2469 | if ((comp == NULL) || (comp->compact == NULL) || (comp->stringMap == NULL)) |
| 2470 | return(-1); |
| 2471 | |
| 2472 | if (value == NULL) { |
| 2473 | /* |
| 2474 | * are we at a final state ? |
| 2475 | */ |
| 2476 | if (comp->compact[state * (comp->nbstrings + 1)] == |
| 2477 | XML_REGEXP_FINAL_STATE) |
| 2478 | return(1); |
| 2479 | return(0); |
| 2480 | } |
| 2481 | |
| 2482 | #ifdef DEBUG_PUSH |
| 2483 | printf("value pushed: %s\n", value); |
| 2484 | #endif |
| 2485 | |
| 2486 | /* |
| 2487 | * Examine all outside transition from current state |
| 2488 | */ |
| 2489 | for (i = 0;i < comp->nbstrings;i++) { |
| 2490 | target = comp->compact[state * (comp->nbstrings + 1) + i + 1]; |
| 2491 | if ((target > 0) && (target <= comp->nbstates)) { |
| 2492 | target--; /* to avoid 0 */ |
| 2493 | if (xmlStrEqual(comp->stringMap[i], value)) { |
| 2494 | exec->index = target; |
Daniel Veillard | 118aed7 | 2002-09-24 14:13:13 +0000 | [diff] [blame] | 2495 | if ((exec->callback != NULL) && (comp->transdata != NULL)) { |
| 2496 | exec->callback(exec->data, value, |
| 2497 | comp->transdata[state * comp->nbstrings + i], data); |
| 2498 | } |
Daniel Veillard | 23e7357 | 2002-09-19 19:56:43 +0000 | [diff] [blame] | 2499 | #ifdef DEBUG_PUSH |
| 2500 | printf("entering state %d\n", target); |
| 2501 | #endif |
| 2502 | if (comp->compact[target * (comp->nbstrings + 1)] == |
| 2503 | XML_REGEXP_FINAL_STATE) |
| 2504 | return(1); |
| 2505 | return(0); |
| 2506 | } |
| 2507 | } |
| 2508 | } |
| 2509 | /* |
| 2510 | * Failed to find an exit transition out from current state for the |
| 2511 | * current token |
| 2512 | */ |
| 2513 | #ifdef DEBUG_PUSH |
| 2514 | printf("failed to find a transition for %s on state %d\n", value, state); |
| 2515 | #endif |
| 2516 | exec->status = -1; |
| 2517 | return(-1); |
| 2518 | } |
| 2519 | |
| 2520 | /** |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 2521 | * xmlRegExecPushString: |
Daniel Veillard | ea7751d | 2002-12-20 00:16:24 +0000 | [diff] [blame] | 2522 | * @exec: a regexp execution context or NULL to indicate the end |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 2523 | * @value: a string token input |
| 2524 | * @data: data associated to the token to reuse in callbacks |
| 2525 | * |
| 2526 | * Push one input token in the execution context |
| 2527 | * |
| 2528 | * Returns: 1 if the regexp reached a final state, 0 if non-final, and |
| 2529 | * a negative value in case of error. |
| 2530 | */ |
| 2531 | int |
| 2532 | xmlRegExecPushString(xmlRegExecCtxtPtr exec, const xmlChar *value, |
| 2533 | void *data) { |
| 2534 | xmlRegTransPtr trans; |
| 2535 | xmlRegAtomPtr atom; |
| 2536 | int ret; |
| 2537 | int final = 0; |
| 2538 | |
| 2539 | if (exec == NULL) |
| 2540 | return(-1); |
Daniel Veillard | 23e7357 | 2002-09-19 19:56:43 +0000 | [diff] [blame] | 2541 | if (exec->comp == NULL) |
| 2542 | return(-1); |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 2543 | if (exec->status != 0) |
| 2544 | return(exec->status); |
| 2545 | |
Daniel Veillard | 23e7357 | 2002-09-19 19:56:43 +0000 | [diff] [blame] | 2546 | if (exec->comp->compact != NULL) |
| 2547 | return(xmlRegCompactPushString(exec, exec->comp, value, data)); |
| 2548 | |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 2549 | if (value == NULL) { |
| 2550 | if (exec->state->type == XML_REGEXP_FINAL_STATE) |
| 2551 | return(1); |
| 2552 | final = 1; |
| 2553 | } |
| 2554 | |
| 2555 | #ifdef DEBUG_PUSH |
| 2556 | printf("value pushed: %s\n", value); |
| 2557 | #endif |
| 2558 | /* |
| 2559 | * If we have an active rollback stack push the new value there |
| 2560 | * and get back to where we were left |
| 2561 | */ |
| 2562 | if ((value != NULL) && (exec->inputStackNr > 0)) { |
| 2563 | xmlFARegExecSaveInputString(exec, value, data); |
| 2564 | value = exec->inputStack[exec->index].value; |
| 2565 | data = exec->inputStack[exec->index].data; |
| 2566 | #ifdef DEBUG_PUSH |
| 2567 | printf("value loaded: %s\n", value); |
| 2568 | #endif |
| 2569 | } |
| 2570 | |
| 2571 | while ((exec->status == 0) && |
| 2572 | ((value != NULL) || |
| 2573 | ((final == 1) && |
| 2574 | (exec->state->type != XML_REGEXP_FINAL_STATE)))) { |
| 2575 | |
| 2576 | /* |
| 2577 | * End of input on non-terminal state, rollback, however we may |
| 2578 | * still have epsilon like transition for counted transitions |
| 2579 | * on counters, in that case don't break too early. |
| 2580 | */ |
Daniel Veillard | b509f15 | 2002-04-17 16:28:10 +0000 | [diff] [blame] | 2581 | if ((value == NULL) && (exec->counts == NULL)) |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 2582 | goto rollback; |
| 2583 | |
| 2584 | exec->transcount = 0; |
| 2585 | for (;exec->transno < exec->state->nbTrans;exec->transno++) { |
| 2586 | trans = &exec->state->trans[exec->transno]; |
| 2587 | if (trans->to < 0) |
| 2588 | continue; |
| 2589 | atom = trans->atom; |
| 2590 | ret = 0; |
Daniel Veillard | 441bc32 | 2002-04-20 17:38:48 +0000 | [diff] [blame] | 2591 | if (trans->count == REGEXP_ALL_LAX_COUNTER) { |
| 2592 | int i; |
| 2593 | int count; |
| 2594 | xmlRegTransPtr t; |
| 2595 | xmlRegCounterPtr counter; |
| 2596 | |
| 2597 | ret = 0; |
| 2598 | |
| 2599 | #ifdef DEBUG_PUSH |
| 2600 | printf("testing all lax %d\n", trans->count); |
| 2601 | #endif |
| 2602 | /* |
| 2603 | * Check all counted transitions from the current state |
| 2604 | */ |
| 2605 | if ((value == NULL) && (final)) { |
| 2606 | ret = 1; |
| 2607 | } else if (value != NULL) { |
| 2608 | for (i = 0;i < exec->state->nbTrans;i++) { |
| 2609 | t = &exec->state->trans[i]; |
| 2610 | if ((t->counter < 0) || (t == trans)) |
| 2611 | continue; |
| 2612 | counter = &exec->comp->counters[t->counter]; |
| 2613 | count = exec->counts[t->counter]; |
| 2614 | if ((count < counter->max) && |
| 2615 | (t->atom != NULL) && |
| 2616 | (xmlStrEqual(value, t->atom->valuep))) { |
| 2617 | ret = 0; |
| 2618 | break; |
| 2619 | } |
| 2620 | if ((count >= counter->min) && |
| 2621 | (count < counter->max) && |
| 2622 | (xmlStrEqual(value, t->atom->valuep))) { |
| 2623 | ret = 1; |
| 2624 | break; |
| 2625 | } |
| 2626 | } |
| 2627 | } |
| 2628 | } else if (trans->count == REGEXP_ALL_COUNTER) { |
Daniel Veillard | 8a001f6 | 2002-04-20 07:24:11 +0000 | [diff] [blame] | 2629 | int i; |
| 2630 | int count; |
| 2631 | xmlRegTransPtr t; |
| 2632 | xmlRegCounterPtr counter; |
| 2633 | |
| 2634 | ret = 1; |
| 2635 | |
| 2636 | #ifdef DEBUG_PUSH |
| 2637 | printf("testing all %d\n", trans->count); |
| 2638 | #endif |
| 2639 | /* |
| 2640 | * Check all counted transitions from the current state |
| 2641 | */ |
| 2642 | for (i = 0;i < exec->state->nbTrans;i++) { |
| 2643 | t = &exec->state->trans[i]; |
| 2644 | if ((t->counter < 0) || (t == trans)) |
| 2645 | continue; |
| 2646 | counter = &exec->comp->counters[t->counter]; |
| 2647 | count = exec->counts[t->counter]; |
| 2648 | if ((count < counter->min) || (count > counter->max)) { |
| 2649 | ret = 0; |
| 2650 | break; |
| 2651 | } |
| 2652 | } |
| 2653 | } else if (trans->count >= 0) { |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 2654 | int count; |
| 2655 | xmlRegCounterPtr counter; |
| 2656 | |
| 2657 | /* |
| 2658 | * A counted transition. |
| 2659 | */ |
| 2660 | |
| 2661 | count = exec->counts[trans->count]; |
| 2662 | counter = &exec->comp->counters[trans->count]; |
| 2663 | #ifdef DEBUG_PUSH |
| 2664 | printf("testing count %d: val %d, min %d, max %d\n", |
| 2665 | trans->count, count, counter->min, counter->max); |
| 2666 | #endif |
| 2667 | ret = ((count >= counter->min) && (count <= counter->max)); |
| 2668 | } else if (atom == NULL) { |
| 2669 | fprintf(stderr, "epsilon transition left at runtime\n"); |
| 2670 | exec->status = -2; |
| 2671 | break; |
| 2672 | } else if (value != NULL) { |
| 2673 | ret = xmlStrEqual(value, atom->valuep); |
Daniel Veillard | 441bc32 | 2002-04-20 17:38:48 +0000 | [diff] [blame] | 2674 | if ((ret == 1) && (trans->counter >= 0)) { |
| 2675 | xmlRegCounterPtr counter; |
| 2676 | int count; |
| 2677 | |
| 2678 | count = exec->counts[trans->counter]; |
| 2679 | counter = &exec->comp->counters[trans->counter]; |
| 2680 | if (count >= counter->max) |
| 2681 | ret = 0; |
| 2682 | } |
| 2683 | |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 2684 | if ((ret == 1) && (atom->min > 0) && (atom->max > 0)) { |
| 2685 | xmlRegStatePtr to = exec->comp->states[trans->to]; |
| 2686 | |
| 2687 | /* |
| 2688 | * this is a multiple input sequence |
| 2689 | */ |
| 2690 | if (exec->state->nbTrans > exec->transno + 1) { |
| 2691 | if (exec->inputStackNr <= 0) { |
| 2692 | xmlFARegExecSaveInputString(exec, value, data); |
| 2693 | } |
| 2694 | xmlFARegExecSave(exec); |
| 2695 | } |
| 2696 | exec->transcount = 1; |
| 2697 | do { |
| 2698 | /* |
| 2699 | * Try to progress as much as possible on the input |
| 2700 | */ |
| 2701 | if (exec->transcount == atom->max) { |
| 2702 | break; |
| 2703 | } |
| 2704 | exec->index++; |
| 2705 | value = exec->inputStack[exec->index].value; |
| 2706 | data = exec->inputStack[exec->index].data; |
| 2707 | #ifdef DEBUG_PUSH |
| 2708 | printf("value loaded: %s\n", value); |
| 2709 | #endif |
| 2710 | |
| 2711 | /* |
| 2712 | * End of input: stop here |
| 2713 | */ |
| 2714 | if (value == NULL) { |
| 2715 | exec->index --; |
| 2716 | break; |
| 2717 | } |
| 2718 | if (exec->transcount >= atom->min) { |
| 2719 | int transno = exec->transno; |
| 2720 | xmlRegStatePtr state = exec->state; |
| 2721 | |
| 2722 | /* |
| 2723 | * The transition is acceptable save it |
| 2724 | */ |
| 2725 | exec->transno = -1; /* trick */ |
| 2726 | exec->state = to; |
| 2727 | if (exec->inputStackNr <= 0) { |
| 2728 | xmlFARegExecSaveInputString(exec, value, data); |
| 2729 | } |
| 2730 | xmlFARegExecSave(exec); |
| 2731 | exec->transno = transno; |
| 2732 | exec->state = state; |
| 2733 | } |
| 2734 | ret = xmlStrEqual(value, atom->valuep); |
| 2735 | exec->transcount++; |
| 2736 | } while (ret == 1); |
| 2737 | if (exec->transcount < atom->min) |
| 2738 | ret = 0; |
| 2739 | |
| 2740 | /* |
| 2741 | * If the last check failed but one transition was found |
| 2742 | * possible, rollback |
| 2743 | */ |
| 2744 | if (ret < 0) |
| 2745 | ret = 0; |
| 2746 | if (ret == 0) { |
| 2747 | goto rollback; |
| 2748 | } |
| 2749 | } |
| 2750 | } |
| 2751 | if (ret == 1) { |
| 2752 | if ((exec->callback != NULL) && (atom != NULL)) { |
| 2753 | exec->callback(exec->data, atom->valuep, |
| 2754 | atom->data, data); |
| 2755 | } |
| 2756 | if (exec->state->nbTrans > exec->transno + 1) { |
| 2757 | if (exec->inputStackNr <= 0) { |
| 2758 | xmlFARegExecSaveInputString(exec, value, data); |
| 2759 | } |
| 2760 | xmlFARegExecSave(exec); |
| 2761 | } |
| 2762 | if (trans->counter >= 0) { |
| 2763 | #ifdef DEBUG_PUSH |
| 2764 | printf("Increasing count %d\n", trans->counter); |
| 2765 | #endif |
| 2766 | exec->counts[trans->counter]++; |
| 2767 | } |
| 2768 | #ifdef DEBUG_PUSH |
| 2769 | printf("entering state %d\n", trans->to); |
| 2770 | #endif |
| 2771 | exec->state = exec->comp->states[trans->to]; |
| 2772 | exec->transno = 0; |
| 2773 | if (trans->atom != NULL) { |
| 2774 | if (exec->inputStack != NULL) { |
| 2775 | exec->index++; |
| 2776 | if (exec->index < exec->inputStackNr) { |
| 2777 | value = exec->inputStack[exec->index].value; |
| 2778 | data = exec->inputStack[exec->index].data; |
| 2779 | #ifdef DEBUG_PUSH |
| 2780 | printf("value loaded: %s\n", value); |
| 2781 | #endif |
| 2782 | } else { |
| 2783 | value = NULL; |
| 2784 | data = NULL; |
| 2785 | #ifdef DEBUG_PUSH |
| 2786 | printf("end of input\n"); |
| 2787 | #endif |
| 2788 | } |
| 2789 | } else { |
| 2790 | value = NULL; |
| 2791 | data = NULL; |
| 2792 | #ifdef DEBUG_PUSH |
| 2793 | printf("end of input\n"); |
| 2794 | #endif |
| 2795 | } |
| 2796 | } |
| 2797 | goto progress; |
| 2798 | } else if (ret < 0) { |
| 2799 | exec->status = -4; |
| 2800 | break; |
| 2801 | } |
| 2802 | } |
| 2803 | if ((exec->transno != 0) || (exec->state->nbTrans == 0)) { |
| 2804 | rollback: |
| 2805 | /* |
| 2806 | * Failed to find a way out |
| 2807 | */ |
| 2808 | exec->determinist = 0; |
| 2809 | xmlFARegExecRollBack(exec); |
| 2810 | if (exec->status == 0) { |
| 2811 | value = exec->inputStack[exec->index].value; |
| 2812 | data = exec->inputStack[exec->index].data; |
| 2813 | #ifdef DEBUG_PUSH |
| 2814 | printf("value loaded: %s\n", value); |
| 2815 | #endif |
| 2816 | } |
| 2817 | } |
| 2818 | progress: |
| 2819 | continue; |
| 2820 | } |
| 2821 | if (exec->status == 0) { |
| 2822 | return(exec->state->type == XML_REGEXP_FINAL_STATE); |
| 2823 | } |
| 2824 | return(exec->status); |
| 2825 | } |
| 2826 | |
Daniel Veillard | 52b48c7 | 2003-04-13 19:53:42 +0000 | [diff] [blame] | 2827 | /** |
| 2828 | * xmlRegExecPushString2: |
| 2829 | * @exec: a regexp execution context or NULL to indicate the end |
| 2830 | * @value: the first string token input |
| 2831 | * @value2: the second string token input |
| 2832 | * @data: data associated to the token to reuse in callbacks |
| 2833 | * |
| 2834 | * Push one input token in the execution context |
| 2835 | * |
| 2836 | * Returns: 1 if the regexp reached a final state, 0 if non-final, and |
| 2837 | * a negative value in case of error. |
| 2838 | */ |
| 2839 | int |
| 2840 | xmlRegExecPushString2(xmlRegExecCtxtPtr exec, const xmlChar *value, |
| 2841 | const xmlChar *value2, void *data) { |
| 2842 | xmlChar buf[150]; |
| 2843 | int lenn, lenp, ret; |
| 2844 | xmlChar *str; |
| 2845 | |
| 2846 | if (exec == NULL) |
| 2847 | return(-1); |
| 2848 | if (exec->comp == NULL) |
| 2849 | return(-1); |
| 2850 | if (exec->status != 0) |
| 2851 | return(exec->status); |
| 2852 | |
| 2853 | if (value2 == NULL) |
| 2854 | return(xmlRegExecPushString(exec, value, data)); |
| 2855 | |
| 2856 | lenn = strlen((char *) value2); |
| 2857 | lenp = strlen((char *) value); |
| 2858 | |
| 2859 | if (150 < lenn + lenp + 2) { |
Daniel Veillard | 3c908dc | 2003-04-19 00:07:51 +0000 | [diff] [blame] | 2860 | str = (xmlChar *) xmlMallocAtomic(lenn + lenp + 2); |
Daniel Veillard | 52b48c7 | 2003-04-13 19:53:42 +0000 | [diff] [blame] | 2861 | if (str == NULL) { |
| 2862 | exec->status = -1; |
| 2863 | return(-1); |
| 2864 | } |
| 2865 | } else { |
| 2866 | str = buf; |
| 2867 | } |
| 2868 | memcpy(&str[0], value, lenp); |
| 2869 | str[lenp] = '|'; |
| 2870 | memcpy(&str[lenp + 1], value2, lenn); |
| 2871 | str[lenn + lenp + 1] = 0; |
| 2872 | |
| 2873 | if (exec->comp->compact != NULL) |
| 2874 | ret = xmlRegCompactPushString(exec, exec->comp, str, data); |
| 2875 | else |
| 2876 | ret = xmlRegExecPushString(exec, str, data); |
| 2877 | |
| 2878 | if (str != buf) |
| 2879 | xmlFree(buf); |
| 2880 | return(ret); |
| 2881 | } |
| 2882 | |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 2883 | #if 0 |
| 2884 | static int |
| 2885 | xmlRegExecPushChar(xmlRegExecCtxtPtr exec, int UCS) { |
| 2886 | xmlRegTransPtr trans; |
| 2887 | xmlRegAtomPtr atom; |
| 2888 | int ret; |
| 2889 | int codepoint, len; |
| 2890 | |
| 2891 | if (exec == NULL) |
| 2892 | return(-1); |
| 2893 | if (exec->status != 0) |
| 2894 | return(exec->status); |
| 2895 | |
| 2896 | while ((exec->status == 0) && |
| 2897 | ((exec->inputString[exec->index] != 0) || |
| 2898 | (exec->state->type != XML_REGEXP_FINAL_STATE))) { |
| 2899 | |
| 2900 | /* |
| 2901 | * End of input on non-terminal state, rollback, however we may |
| 2902 | * still have epsilon like transition for counted transitions |
| 2903 | * on counters, in that case don't break too early. |
| 2904 | */ |
| 2905 | if ((exec->inputString[exec->index] == 0) && (exec->counts == NULL)) |
| 2906 | goto rollback; |
| 2907 | |
| 2908 | exec->transcount = 0; |
| 2909 | for (;exec->transno < exec->state->nbTrans;exec->transno++) { |
| 2910 | trans = &exec->state->trans[exec->transno]; |
| 2911 | if (trans->to < 0) |
| 2912 | continue; |
| 2913 | atom = trans->atom; |
| 2914 | ret = 0; |
| 2915 | if (trans->count >= 0) { |
| 2916 | int count; |
| 2917 | xmlRegCounterPtr counter; |
| 2918 | |
| 2919 | /* |
| 2920 | * A counted transition. |
| 2921 | */ |
| 2922 | |
| 2923 | count = exec->counts[trans->count]; |
| 2924 | counter = &exec->comp->counters[trans->count]; |
| 2925 | #ifdef DEBUG_REGEXP_EXEC |
| 2926 | printf("testing count %d: val %d, min %d, max %d\n", |
| 2927 | trans->count, count, counter->min, counter->max); |
| 2928 | #endif |
| 2929 | ret = ((count >= counter->min) && (count <= counter->max)); |
| 2930 | } else if (atom == NULL) { |
| 2931 | fprintf(stderr, "epsilon transition left at runtime\n"); |
| 2932 | exec->status = -2; |
| 2933 | break; |
| 2934 | } else if (exec->inputString[exec->index] != 0) { |
| 2935 | codepoint = CUR_SCHAR(&(exec->inputString[exec->index]), len); |
| 2936 | ret = xmlRegCheckCharacter(atom, codepoint); |
| 2937 | if ((ret == 1) && (atom->min > 0) && (atom->max > 0)) { |
| 2938 | xmlRegStatePtr to = exec->comp->states[trans->to]; |
| 2939 | |
| 2940 | /* |
| 2941 | * this is a multiple input sequence |
| 2942 | */ |
| 2943 | if (exec->state->nbTrans > exec->transno + 1) { |
| 2944 | xmlFARegExecSave(exec); |
| 2945 | } |
| 2946 | exec->transcount = 1; |
| 2947 | do { |
| 2948 | /* |
| 2949 | * Try to progress as much as possible on the input |
| 2950 | */ |
| 2951 | if (exec->transcount == atom->max) { |
| 2952 | break; |
| 2953 | } |
| 2954 | exec->index += len; |
| 2955 | /* |
| 2956 | * End of input: stop here |
| 2957 | */ |
| 2958 | if (exec->inputString[exec->index] == 0) { |
| 2959 | exec->index -= len; |
| 2960 | break; |
| 2961 | } |
| 2962 | if (exec->transcount >= atom->min) { |
| 2963 | int transno = exec->transno; |
| 2964 | xmlRegStatePtr state = exec->state; |
| 2965 | |
| 2966 | /* |
| 2967 | * The transition is acceptable save it |
| 2968 | */ |
| 2969 | exec->transno = -1; /* trick */ |
| 2970 | exec->state = to; |
| 2971 | xmlFARegExecSave(exec); |
| 2972 | exec->transno = transno; |
| 2973 | exec->state = state; |
| 2974 | } |
| 2975 | codepoint = CUR_SCHAR(&(exec->inputString[exec->index]), |
| 2976 | len); |
| 2977 | ret = xmlRegCheckCharacter(atom, codepoint); |
| 2978 | exec->transcount++; |
| 2979 | } while (ret == 1); |
| 2980 | if (exec->transcount < atom->min) |
| 2981 | ret = 0; |
| 2982 | |
| 2983 | /* |
| 2984 | * If the last check failed but one transition was found |
| 2985 | * possible, rollback |
| 2986 | */ |
| 2987 | if (ret < 0) |
| 2988 | ret = 0; |
| 2989 | if (ret == 0) { |
| 2990 | goto rollback; |
| 2991 | } |
| 2992 | } |
| 2993 | } |
| 2994 | if (ret == 1) { |
| 2995 | if (exec->state->nbTrans > exec->transno + 1) { |
| 2996 | xmlFARegExecSave(exec); |
| 2997 | } |
| 2998 | if (trans->counter >= 0) { |
| 2999 | #ifdef DEBUG_REGEXP_EXEC |
| 3000 | printf("Increasing count %d\n", trans->counter); |
| 3001 | #endif |
| 3002 | exec->counts[trans->counter]++; |
| 3003 | } |
| 3004 | #ifdef DEBUG_REGEXP_EXEC |
| 3005 | printf("entering state %d\n", trans->to); |
| 3006 | #endif |
| 3007 | exec->state = exec->comp->states[trans->to]; |
| 3008 | exec->transno = 0; |
| 3009 | if (trans->atom != NULL) { |
| 3010 | exec->index += len; |
| 3011 | } |
| 3012 | goto progress; |
| 3013 | } else if (ret < 0) { |
| 3014 | exec->status = -4; |
| 3015 | break; |
| 3016 | } |
| 3017 | } |
| 3018 | if ((exec->transno != 0) || (exec->state->nbTrans == 0)) { |
| 3019 | rollback: |
| 3020 | /* |
| 3021 | * Failed to find a way out |
| 3022 | */ |
| 3023 | exec->determinist = 0; |
| 3024 | xmlFARegExecRollBack(exec); |
| 3025 | } |
| 3026 | progress: |
| 3027 | continue; |
| 3028 | } |
| 3029 | } |
| 3030 | #endif |
| 3031 | /************************************************************************ |
| 3032 | * * |
| 3033 | * Parser for the Shemas Datatype Regular Expressions * |
| 3034 | * http://www.w3.org/TR/2001/REC-xmlschema-2-20010502/#regexs * |
| 3035 | * * |
| 3036 | ************************************************************************/ |
| 3037 | |
| 3038 | /** |
| 3039 | * xmlFAIsChar: |
Daniel Veillard | 441bc32 | 2002-04-20 17:38:48 +0000 | [diff] [blame] | 3040 | * @ctxt: a regexp parser context |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 3041 | * |
| 3042 | * [10] Char ::= [^.\?*+()|#x5B#x5D] |
| 3043 | */ |
| 3044 | static int |
| 3045 | xmlFAIsChar(xmlRegParserCtxtPtr ctxt) { |
| 3046 | int cur; |
| 3047 | int len; |
| 3048 | |
| 3049 | cur = CUR_SCHAR(ctxt->cur, len); |
| 3050 | if ((cur == '.') || (cur == '\\') || (cur == '?') || |
| 3051 | (cur == '*') || (cur == '+') || (cur == '(') || |
| 3052 | (cur == ')') || (cur == '|') || (cur == 0x5B) || |
| 3053 | (cur == 0x5D) || (cur == 0)) |
| 3054 | return(-1); |
| 3055 | return(cur); |
| 3056 | } |
| 3057 | |
| 3058 | /** |
| 3059 | * xmlFAParseCharProp: |
Daniel Veillard | 441bc32 | 2002-04-20 17:38:48 +0000 | [diff] [blame] | 3060 | * @ctxt: a regexp parser context |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 3061 | * |
| 3062 | * [27] charProp ::= IsCategory | IsBlock |
| 3063 | * [28] IsCategory ::= Letters | Marks | Numbers | Punctuation | |
| 3064 | * Separators | Symbols | Others |
| 3065 | * [29] Letters ::= 'L' [ultmo]? |
| 3066 | * [30] Marks ::= 'M' [nce]? |
| 3067 | * [31] Numbers ::= 'N' [dlo]? |
| 3068 | * [32] Punctuation ::= 'P' [cdseifo]? |
| 3069 | * [33] Separators ::= 'Z' [slp]? |
| 3070 | * [34] Symbols ::= 'S' [mcko]? |
| 3071 | * [35] Others ::= 'C' [cfon]? |
| 3072 | * [36] IsBlock ::= 'Is' [a-zA-Z0-9#x2D]+ |
| 3073 | */ |
| 3074 | static void |
| 3075 | xmlFAParseCharProp(xmlRegParserCtxtPtr ctxt) { |
| 3076 | int cur; |
| 3077 | xmlRegAtomType type = 0; |
| 3078 | xmlChar *blockName = NULL; |
| 3079 | |
| 3080 | cur = CUR; |
| 3081 | if (cur == 'L') { |
| 3082 | NEXT; |
| 3083 | cur = CUR; |
| 3084 | if (cur == 'u') { |
| 3085 | NEXT; |
| 3086 | type = XML_REGEXP_LETTER_UPPERCASE; |
| 3087 | } else if (cur == 'l') { |
| 3088 | NEXT; |
| 3089 | type = XML_REGEXP_LETTER_LOWERCASE; |
| 3090 | } else if (cur == 't') { |
| 3091 | NEXT; |
| 3092 | type = XML_REGEXP_LETTER_TITLECASE; |
| 3093 | } else if (cur == 'm') { |
| 3094 | NEXT; |
| 3095 | type = XML_REGEXP_LETTER_MODIFIER; |
| 3096 | } else if (cur == 'o') { |
| 3097 | NEXT; |
| 3098 | type = XML_REGEXP_LETTER_OTHERS; |
| 3099 | } else { |
| 3100 | type = XML_REGEXP_LETTER; |
| 3101 | } |
| 3102 | } else if (cur == 'M') { |
| 3103 | NEXT; |
| 3104 | cur = CUR; |
| 3105 | if (cur == 'n') { |
| 3106 | NEXT; |
| 3107 | /* nonspacing */ |
| 3108 | type = XML_REGEXP_MARK_NONSPACING; |
| 3109 | } else if (cur == 'c') { |
| 3110 | NEXT; |
| 3111 | /* spacing combining */ |
| 3112 | type = XML_REGEXP_MARK_SPACECOMBINING; |
| 3113 | } else if (cur == 'e') { |
| 3114 | NEXT; |
| 3115 | /* enclosing */ |
| 3116 | type = XML_REGEXP_MARK_ENCLOSING; |
| 3117 | } else { |
| 3118 | /* all marks */ |
| 3119 | type = XML_REGEXP_MARK; |
| 3120 | } |
| 3121 | } else if (cur == 'N') { |
| 3122 | NEXT; |
| 3123 | cur = CUR; |
| 3124 | if (cur == 'd') { |
| 3125 | NEXT; |
| 3126 | /* digital */ |
| 3127 | type = XML_REGEXP_NUMBER_DECIMAL; |
| 3128 | } else if (cur == 'l') { |
| 3129 | NEXT; |
| 3130 | /* letter */ |
| 3131 | type = XML_REGEXP_NUMBER_LETTER; |
| 3132 | } else if (cur == 'o') { |
| 3133 | NEXT; |
| 3134 | /* other */ |
| 3135 | type = XML_REGEXP_NUMBER_OTHERS; |
| 3136 | } else { |
| 3137 | /* all numbers */ |
| 3138 | type = XML_REGEXP_NUMBER; |
| 3139 | } |
| 3140 | } else if (cur == 'P') { |
| 3141 | NEXT; |
| 3142 | cur = CUR; |
| 3143 | if (cur == 'c') { |
| 3144 | NEXT; |
| 3145 | /* connector */ |
| 3146 | type = XML_REGEXP_PUNCT_CONNECTOR; |
| 3147 | } else if (cur == 'd') { |
| 3148 | NEXT; |
| 3149 | /* dash */ |
| 3150 | type = XML_REGEXP_PUNCT_DASH; |
| 3151 | } else if (cur == 's') { |
| 3152 | NEXT; |
| 3153 | /* open */ |
| 3154 | type = XML_REGEXP_PUNCT_OPEN; |
| 3155 | } else if (cur == 'e') { |
| 3156 | NEXT; |
| 3157 | /* close */ |
| 3158 | type = XML_REGEXP_PUNCT_CLOSE; |
| 3159 | } else if (cur == 'i') { |
| 3160 | NEXT; |
| 3161 | /* initial quote */ |
| 3162 | type = XML_REGEXP_PUNCT_INITQUOTE; |
| 3163 | } else if (cur == 'f') { |
| 3164 | NEXT; |
| 3165 | /* final quote */ |
| 3166 | type = XML_REGEXP_PUNCT_FINQUOTE; |
| 3167 | } else if (cur == 'o') { |
| 3168 | NEXT; |
| 3169 | /* other */ |
| 3170 | type = XML_REGEXP_PUNCT_OTHERS; |
| 3171 | } else { |
| 3172 | /* all punctuation */ |
| 3173 | type = XML_REGEXP_PUNCT; |
| 3174 | } |
| 3175 | } else if (cur == 'Z') { |
| 3176 | NEXT; |
| 3177 | cur = CUR; |
| 3178 | if (cur == 's') { |
| 3179 | NEXT; |
| 3180 | /* space */ |
| 3181 | type = XML_REGEXP_SEPAR_SPACE; |
| 3182 | } else if (cur == 'l') { |
| 3183 | NEXT; |
| 3184 | /* line */ |
| 3185 | type = XML_REGEXP_SEPAR_LINE; |
| 3186 | } else if (cur == 'p') { |
| 3187 | NEXT; |
| 3188 | /* paragraph */ |
| 3189 | type = XML_REGEXP_SEPAR_PARA; |
| 3190 | } else { |
| 3191 | /* all separators */ |
| 3192 | type = XML_REGEXP_SEPAR; |
| 3193 | } |
| 3194 | } else if (cur == 'S') { |
| 3195 | NEXT; |
| 3196 | cur = CUR; |
| 3197 | if (cur == 'm') { |
| 3198 | NEXT; |
| 3199 | type = XML_REGEXP_SYMBOL_MATH; |
| 3200 | /* math */ |
| 3201 | } else if (cur == 'c') { |
| 3202 | NEXT; |
| 3203 | type = XML_REGEXP_SYMBOL_CURRENCY; |
| 3204 | /* currency */ |
| 3205 | } else if (cur == 'k') { |
| 3206 | NEXT; |
| 3207 | type = XML_REGEXP_SYMBOL_MODIFIER; |
| 3208 | /* modifiers */ |
| 3209 | } else if (cur == 'o') { |
| 3210 | NEXT; |
| 3211 | type = XML_REGEXP_SYMBOL_OTHERS; |
| 3212 | /* other */ |
| 3213 | } else { |
| 3214 | /* all symbols */ |
| 3215 | type = XML_REGEXP_SYMBOL; |
| 3216 | } |
| 3217 | } else if (cur == 'C') { |
| 3218 | NEXT; |
| 3219 | cur = CUR; |
| 3220 | if (cur == 'c') { |
| 3221 | NEXT; |
| 3222 | /* control */ |
| 3223 | type = XML_REGEXP_OTHER_CONTROL; |
| 3224 | } else if (cur == 'f') { |
| 3225 | NEXT; |
| 3226 | /* format */ |
| 3227 | type = XML_REGEXP_OTHER_FORMAT; |
| 3228 | } else if (cur == 'o') { |
| 3229 | NEXT; |
| 3230 | /* private use */ |
| 3231 | type = XML_REGEXP_OTHER_PRIVATE; |
| 3232 | } else if (cur == 'n') { |
| 3233 | NEXT; |
| 3234 | /* not assigned */ |
| 3235 | type = XML_REGEXP_OTHER_NA; |
| 3236 | } else { |
| 3237 | /* all others */ |
| 3238 | type = XML_REGEXP_OTHER; |
| 3239 | } |
| 3240 | } else if (cur == 'I') { |
| 3241 | const xmlChar *start; |
| 3242 | NEXT; |
| 3243 | cur = CUR; |
| 3244 | if (cur != 's') { |
| 3245 | ERROR("IsXXXX expected"); |
| 3246 | return; |
| 3247 | } |
| 3248 | NEXT; |
| 3249 | start = ctxt->cur; |
| 3250 | cur = CUR; |
| 3251 | if (((cur >= 'a') && (cur <= 'z')) || |
| 3252 | ((cur >= 'A') && (cur <= 'Z')) || |
| 3253 | ((cur >= '0') && (cur <= '9')) || |
| 3254 | (cur == 0x2D)) { |
| 3255 | NEXT; |
| 3256 | cur = CUR; |
| 3257 | while (((cur >= 'a') && (cur <= 'z')) || |
| 3258 | ((cur >= 'A') && (cur <= 'Z')) || |
| 3259 | ((cur >= '0') && (cur <= '9')) || |
| 3260 | (cur == 0x2D)) { |
| 3261 | NEXT; |
| 3262 | cur = CUR; |
| 3263 | } |
| 3264 | } |
| 3265 | type = XML_REGEXP_BLOCK_NAME; |
| 3266 | blockName = xmlStrndup(start, ctxt->cur - start); |
| 3267 | } else { |
| 3268 | ERROR("Unknown char property"); |
| 3269 | return; |
| 3270 | } |
| 3271 | if (ctxt->atom == NULL) { |
| 3272 | ctxt->atom = xmlRegNewAtom(ctxt, type); |
| 3273 | if (ctxt->atom != NULL) |
| 3274 | ctxt->atom->valuep = blockName; |
| 3275 | } else if (ctxt->atom->type == XML_REGEXP_RANGES) { |
| 3276 | xmlRegAtomAddRange(ctxt, ctxt->atom, ctxt->neg, |
| 3277 | type, 0, 0, blockName); |
| 3278 | } |
| 3279 | } |
| 3280 | |
| 3281 | /** |
| 3282 | * xmlFAParseCharClassEsc: |
Daniel Veillard | 441bc32 | 2002-04-20 17:38:48 +0000 | [diff] [blame] | 3283 | * @ctxt: a regexp parser context |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 3284 | * |
| 3285 | * [23] charClassEsc ::= ( SingleCharEsc | MultiCharEsc | catEsc | complEsc ) |
| 3286 | * [24] SingleCharEsc ::= '\' [nrt\|.?*+(){}#x2D#x5B#x5D#x5E] |
| 3287 | * [25] catEsc ::= '\p{' charProp '}' |
| 3288 | * [26] complEsc ::= '\P{' charProp '}' |
| 3289 | * [37] MultiCharEsc ::= '.' | ('\' [sSiIcCdDwW]) |
| 3290 | */ |
| 3291 | static void |
| 3292 | xmlFAParseCharClassEsc(xmlRegParserCtxtPtr ctxt) { |
| 3293 | int cur; |
| 3294 | |
| 3295 | if (CUR == '.') { |
| 3296 | if (ctxt->atom == NULL) { |
| 3297 | ctxt->atom = xmlRegNewAtom(ctxt, XML_REGEXP_ANYCHAR); |
| 3298 | } else if (ctxt->atom->type == XML_REGEXP_RANGES) { |
| 3299 | xmlRegAtomAddRange(ctxt, ctxt->atom, ctxt->neg, |
| 3300 | XML_REGEXP_ANYCHAR, 0, 0, NULL); |
| 3301 | } |
| 3302 | NEXT; |
| 3303 | return; |
| 3304 | } |
| 3305 | if (CUR != '\\') { |
| 3306 | ERROR("Escaped sequence: expecting \\"); |
| 3307 | return; |
| 3308 | } |
| 3309 | NEXT; |
| 3310 | cur = CUR; |
| 3311 | if (cur == 'p') { |
| 3312 | NEXT; |
| 3313 | if (CUR != '{') { |
| 3314 | ERROR("Expecting '{'"); |
| 3315 | return; |
| 3316 | } |
| 3317 | NEXT; |
| 3318 | xmlFAParseCharProp(ctxt); |
| 3319 | if (CUR != '}') { |
| 3320 | ERROR("Expecting '}'"); |
| 3321 | return; |
| 3322 | } |
| 3323 | NEXT; |
| 3324 | } else if (cur == 'P') { |
| 3325 | NEXT; |
| 3326 | if (CUR != '{') { |
| 3327 | ERROR("Expecting '{'"); |
| 3328 | return; |
| 3329 | } |
| 3330 | NEXT; |
| 3331 | xmlFAParseCharProp(ctxt); |
| 3332 | ctxt->atom->neg = 1; |
| 3333 | if (CUR != '}') { |
| 3334 | ERROR("Expecting '}'"); |
| 3335 | return; |
| 3336 | } |
| 3337 | NEXT; |
| 3338 | } else if ((cur == 'n') || (cur == 'r') || (cur == 't') || (cur == '\\') || |
| 3339 | (cur == '|') || (cur == '.') || (cur == '?') || (cur == '*') || |
| 3340 | (cur == '+') || (cur == '(') || (cur == ')') || (cur == '{') || |
| 3341 | (cur == '}') || (cur == 0x2D) || (cur == 0x5B) || (cur == 0x5D) || |
| 3342 | (cur == 0x5E)) { |
| 3343 | if (ctxt->atom == NULL) { |
| 3344 | ctxt->atom = xmlRegNewAtom(ctxt, XML_REGEXP_CHARVAL); |
| 3345 | if (ctxt->atom != NULL) |
| 3346 | ctxt->atom->codepoint = cur; |
| 3347 | } else if (ctxt->atom->type == XML_REGEXP_RANGES) { |
| 3348 | xmlRegAtomAddRange(ctxt, ctxt->atom, ctxt->neg, |
| 3349 | XML_REGEXP_CHARVAL, cur, cur, NULL); |
| 3350 | } |
| 3351 | NEXT; |
| 3352 | } else if ((cur == 's') || (cur == 'S') || (cur == 'i') || (cur == 'I') || |
| 3353 | (cur == 'c') || (cur == 'C') || (cur == 'd') || (cur == 'D') || |
| 3354 | (cur == 'w') || (cur == 'W')) { |
Daniel Veillard | b509f15 | 2002-04-17 16:28:10 +0000 | [diff] [blame] | 3355 | xmlRegAtomType type = XML_REGEXP_ANYSPACE; |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 3356 | |
| 3357 | switch (cur) { |
| 3358 | case 's': |
| 3359 | type = XML_REGEXP_ANYSPACE; |
| 3360 | break; |
| 3361 | case 'S': |
| 3362 | type = XML_REGEXP_NOTSPACE; |
| 3363 | break; |
| 3364 | case 'i': |
| 3365 | type = XML_REGEXP_INITNAME; |
| 3366 | break; |
| 3367 | case 'I': |
| 3368 | type = XML_REGEXP_NOTINITNAME; |
| 3369 | break; |
| 3370 | case 'c': |
| 3371 | type = XML_REGEXP_NAMECHAR; |
| 3372 | break; |
| 3373 | case 'C': |
| 3374 | type = XML_REGEXP_NOTNAMECHAR; |
| 3375 | break; |
| 3376 | case 'd': |
| 3377 | type = XML_REGEXP_DECIMAL; |
| 3378 | break; |
| 3379 | case 'D': |
| 3380 | type = XML_REGEXP_NOTDECIMAL; |
| 3381 | break; |
| 3382 | case 'w': |
| 3383 | type = XML_REGEXP_REALCHAR; |
| 3384 | break; |
| 3385 | case 'W': |
| 3386 | type = XML_REGEXP_NOTREALCHAR; |
| 3387 | break; |
| 3388 | } |
| 3389 | NEXT; |
| 3390 | if (ctxt->atom == NULL) { |
| 3391 | ctxt->atom = xmlRegNewAtom(ctxt, type); |
| 3392 | } else if (ctxt->atom->type == XML_REGEXP_RANGES) { |
| 3393 | xmlRegAtomAddRange(ctxt, ctxt->atom, ctxt->neg, |
| 3394 | type, 0, 0, NULL); |
| 3395 | } |
| 3396 | } |
| 3397 | } |
| 3398 | |
| 3399 | /** |
| 3400 | * xmlFAParseCharRef: |
Daniel Veillard | 441bc32 | 2002-04-20 17:38:48 +0000 | [diff] [blame] | 3401 | * @ctxt: a regexp parser context |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 3402 | * |
| 3403 | * [19] XmlCharRef ::= ( '&#' [0-9]+ ';' ) | (' &#x' [0-9a-fA-F]+ ';' ) |
| 3404 | */ |
| 3405 | static int |
| 3406 | xmlFAParseCharRef(xmlRegParserCtxtPtr ctxt) { |
| 3407 | int ret = 0, cur; |
| 3408 | |
| 3409 | if ((CUR != '&') || (NXT(1) != '#')) |
| 3410 | return(-1); |
| 3411 | NEXT; |
| 3412 | NEXT; |
| 3413 | cur = CUR; |
| 3414 | if (cur == 'x') { |
| 3415 | NEXT; |
| 3416 | cur = CUR; |
| 3417 | if (((cur >= '0') && (cur <= '9')) || |
| 3418 | ((cur >= 'a') && (cur <= 'f')) || |
| 3419 | ((cur >= 'A') && (cur <= 'F'))) { |
| 3420 | while (((cur >= '0') && (cur <= '9')) || |
| 3421 | ((cur >= 'A') && (cur <= 'F'))) { |
| 3422 | if ((cur >= '0') && (cur <= '9')) |
| 3423 | ret = ret * 16 + cur - '0'; |
| 3424 | else if ((cur >= 'a') && (cur <= 'f')) |
| 3425 | ret = ret * 16 + 10 + (cur - 'a'); |
| 3426 | else |
| 3427 | ret = ret * 16 + 10 + (cur - 'A'); |
| 3428 | NEXT; |
| 3429 | cur = CUR; |
| 3430 | } |
| 3431 | } else { |
| 3432 | ERROR("Char ref: expecting [0-9A-F]"); |
| 3433 | return(-1); |
| 3434 | } |
| 3435 | } else { |
| 3436 | if ((cur >= '0') && (cur <= '9')) { |
| 3437 | while ((cur >= '0') && (cur <= '9')) { |
| 3438 | ret = ret * 10 + cur - '0'; |
| 3439 | NEXT; |
| 3440 | cur = CUR; |
| 3441 | } |
| 3442 | } else { |
| 3443 | ERROR("Char ref: expecting [0-9]"); |
| 3444 | return(-1); |
| 3445 | } |
| 3446 | } |
| 3447 | if (cur != ';') { |
| 3448 | ERROR("Char ref: expecting ';'"); |
| 3449 | return(-1); |
| 3450 | } else { |
| 3451 | NEXT; |
| 3452 | } |
| 3453 | return(ret); |
| 3454 | } |
| 3455 | |
| 3456 | /** |
| 3457 | * xmlFAParseCharRange: |
Daniel Veillard | 441bc32 | 2002-04-20 17:38:48 +0000 | [diff] [blame] | 3458 | * @ctxt: a regexp parser context |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 3459 | * |
| 3460 | * [17] charRange ::= seRange | XmlCharRef | XmlCharIncDash |
| 3461 | * [18] seRange ::= charOrEsc '-' charOrEsc |
| 3462 | * [20] charOrEsc ::= XmlChar | SingleCharEsc |
| 3463 | * [21] XmlChar ::= [^\#x2D#x5B#x5D] |
| 3464 | * [22] XmlCharIncDash ::= [^\#x5B#x5D] |
| 3465 | */ |
| 3466 | static void |
| 3467 | xmlFAParseCharRange(xmlRegParserCtxtPtr ctxt) { |
| 3468 | int cur; |
| 3469 | int start = -1; |
| 3470 | int end = -1; |
| 3471 | |
| 3472 | if ((CUR == '&') && (NXT(1) == '#')) { |
| 3473 | end = start = xmlFAParseCharRef(ctxt); |
| 3474 | xmlRegAtomAddRange(ctxt, ctxt->atom, ctxt->neg, |
| 3475 | XML_REGEXP_CHARVAL, start, end, NULL); |
| 3476 | return; |
| 3477 | } |
| 3478 | cur = CUR; |
| 3479 | if (cur == '\\') { |
| 3480 | NEXT; |
| 3481 | cur = CUR; |
| 3482 | switch (cur) { |
| 3483 | case 'n': start = 0xA; break; |
| 3484 | case 'r': start = 0xD; break; |
| 3485 | case 't': start = 0x9; break; |
| 3486 | case '\\': case '|': case '.': case '-': case '^': case '?': |
| 3487 | case '*': case '+': case '{': case '}': case '(': case ')': |
| 3488 | case '[': case ']': |
| 3489 | start = cur; break; |
| 3490 | default: |
| 3491 | ERROR("Invalid escape value"); |
| 3492 | return; |
| 3493 | } |
| 3494 | end = start; |
| 3495 | } else if ((cur != 0x5B) && (cur != 0x5D)) { |
| 3496 | end = start = cur; |
| 3497 | } else { |
| 3498 | ERROR("Expecting a char range"); |
| 3499 | return; |
| 3500 | } |
| 3501 | NEXT; |
| 3502 | if (start == '-') { |
| 3503 | return; |
| 3504 | } |
| 3505 | cur = CUR; |
| 3506 | if (cur != '-') { |
| 3507 | xmlRegAtomAddRange(ctxt, ctxt->atom, ctxt->neg, |
| 3508 | XML_REGEXP_CHARVAL, start, end, NULL); |
| 3509 | return; |
| 3510 | } |
| 3511 | NEXT; |
| 3512 | cur = CUR; |
| 3513 | if (cur == '\\') { |
| 3514 | NEXT; |
| 3515 | cur = CUR; |
| 3516 | switch (cur) { |
| 3517 | case 'n': end = 0xA; break; |
| 3518 | case 'r': end = 0xD; break; |
| 3519 | case 't': end = 0x9; break; |
| 3520 | case '\\': case '|': case '.': case '-': case '^': case '?': |
| 3521 | case '*': case '+': case '{': case '}': case '(': case ')': |
| 3522 | case '[': case ']': |
| 3523 | end = cur; break; |
| 3524 | default: |
| 3525 | ERROR("Invalid escape value"); |
| 3526 | return; |
| 3527 | } |
| 3528 | } else if ((cur != 0x5B) && (cur != 0x5D)) { |
| 3529 | end = cur; |
| 3530 | } else { |
| 3531 | ERROR("Expecting the end of a char range"); |
| 3532 | return; |
| 3533 | } |
| 3534 | NEXT; |
| 3535 | /* TODO check that the values are acceptable character ranges for XML */ |
| 3536 | if (end < start) { |
| 3537 | ERROR("End of range is before start of range"); |
| 3538 | } else { |
| 3539 | xmlRegAtomAddRange(ctxt, ctxt->atom, ctxt->neg, |
| 3540 | XML_REGEXP_CHARVAL, start, end, NULL); |
| 3541 | } |
| 3542 | return; |
| 3543 | } |
| 3544 | |
| 3545 | /** |
| 3546 | * xmlFAParsePosCharGroup: |
Daniel Veillard | 441bc32 | 2002-04-20 17:38:48 +0000 | [diff] [blame] | 3547 | * @ctxt: a regexp parser context |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 3548 | * |
| 3549 | * [14] posCharGroup ::= ( charRange | charClassEsc )+ |
| 3550 | */ |
| 3551 | static void |
| 3552 | xmlFAParsePosCharGroup(xmlRegParserCtxtPtr ctxt) { |
| 3553 | do { |
| 3554 | if ((CUR == '\\') || (CUR == '.')) { |
| 3555 | xmlFAParseCharClassEsc(ctxt); |
| 3556 | } else { |
| 3557 | xmlFAParseCharRange(ctxt); |
| 3558 | } |
| 3559 | } while ((CUR != ']') && (CUR != '^') && (CUR != '-') && |
| 3560 | (ctxt->error == 0)); |
| 3561 | } |
| 3562 | |
| 3563 | /** |
| 3564 | * xmlFAParseCharGroup: |
Daniel Veillard | 441bc32 | 2002-04-20 17:38:48 +0000 | [diff] [blame] | 3565 | * @ctxt: a regexp parser context |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 3566 | * |
| 3567 | * [13] charGroup ::= posCharGroup | negCharGroup | charClassSub |
| 3568 | * [15] negCharGroup ::= '^' posCharGroup |
| 3569 | * [16] charClassSub ::= ( posCharGroup | negCharGroup ) '-' charClassExpr |
| 3570 | * [12] charClassExpr ::= '[' charGroup ']' |
| 3571 | */ |
| 3572 | static void |
| 3573 | xmlFAParseCharGroup(xmlRegParserCtxtPtr ctxt) { |
| 3574 | int n = ctxt->neg; |
| 3575 | while ((CUR != ']') && (ctxt->error == 0)) { |
| 3576 | if (CUR == '^') { |
| 3577 | int neg = ctxt->neg; |
| 3578 | |
| 3579 | NEXT; |
| 3580 | ctxt->neg = !ctxt->neg; |
| 3581 | xmlFAParsePosCharGroup(ctxt); |
| 3582 | ctxt->neg = neg; |
| 3583 | } else if (CUR == '-') { |
| 3584 | NEXT; |
| 3585 | ctxt->neg = !ctxt->neg; |
| 3586 | if (CUR != '[') { |
| 3587 | ERROR("charClassExpr: '[' expected"); |
| 3588 | break; |
| 3589 | } |
| 3590 | NEXT; |
| 3591 | xmlFAParseCharGroup(ctxt); |
| 3592 | if (CUR == ']') { |
| 3593 | NEXT; |
| 3594 | } else { |
| 3595 | ERROR("charClassExpr: ']' expected"); |
| 3596 | break; |
| 3597 | } |
| 3598 | break; |
| 3599 | } else if (CUR != ']') { |
| 3600 | xmlFAParsePosCharGroup(ctxt); |
| 3601 | } |
| 3602 | } |
| 3603 | ctxt->neg = n; |
| 3604 | } |
| 3605 | |
| 3606 | /** |
| 3607 | * xmlFAParseCharClass: |
Daniel Veillard | 441bc32 | 2002-04-20 17:38:48 +0000 | [diff] [blame] | 3608 | * @ctxt: a regexp parser context |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 3609 | * |
| 3610 | * [11] charClass ::= charClassEsc | charClassExpr |
| 3611 | * [12] charClassExpr ::= '[' charGroup ']' |
| 3612 | */ |
| 3613 | static void |
| 3614 | xmlFAParseCharClass(xmlRegParserCtxtPtr ctxt) { |
| 3615 | if (CUR == '[') { |
| 3616 | NEXT; |
| 3617 | ctxt->atom = xmlRegNewAtom(ctxt, XML_REGEXP_RANGES); |
| 3618 | if (ctxt->atom == NULL) |
| 3619 | return; |
| 3620 | xmlFAParseCharGroup(ctxt); |
| 3621 | if (CUR == ']') { |
| 3622 | NEXT; |
| 3623 | } else { |
| 3624 | ERROR("xmlFAParseCharClass: ']' expected"); |
| 3625 | } |
| 3626 | } else { |
| 3627 | xmlFAParseCharClassEsc(ctxt); |
| 3628 | } |
| 3629 | } |
| 3630 | |
| 3631 | /** |
| 3632 | * xmlFAParseQuantExact: |
Daniel Veillard | 441bc32 | 2002-04-20 17:38:48 +0000 | [diff] [blame] | 3633 | * @ctxt: a regexp parser context |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 3634 | * |
| 3635 | * [8] QuantExact ::= [0-9]+ |
Daniel Veillard | a76fe5c | 2003-04-24 16:06:47 +0000 | [diff] [blame] | 3636 | * |
| 3637 | * Returns 0 if success or -1 in case of error |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 3638 | */ |
| 3639 | static int |
| 3640 | xmlFAParseQuantExact(xmlRegParserCtxtPtr ctxt) { |
| 3641 | int ret = 0; |
| 3642 | int ok = 0; |
| 3643 | |
| 3644 | while ((CUR >= '0') && (CUR <= '9')) { |
| 3645 | ret = ret * 10 + (CUR - '0'); |
| 3646 | ok = 1; |
| 3647 | NEXT; |
| 3648 | } |
| 3649 | if (ok != 1) { |
| 3650 | return(-1); |
| 3651 | } |
| 3652 | return(ret); |
| 3653 | } |
| 3654 | |
| 3655 | /** |
| 3656 | * xmlFAParseQuantifier: |
Daniel Veillard | 441bc32 | 2002-04-20 17:38:48 +0000 | [diff] [blame] | 3657 | * @ctxt: a regexp parser context |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 3658 | * |
| 3659 | * [4] quantifier ::= [?*+] | ( '{' quantity '}' ) |
| 3660 | * [5] quantity ::= quantRange | quantMin | QuantExact |
| 3661 | * [6] quantRange ::= QuantExact ',' QuantExact |
| 3662 | * [7] quantMin ::= QuantExact ',' |
| 3663 | * [8] QuantExact ::= [0-9]+ |
| 3664 | */ |
| 3665 | static int |
| 3666 | xmlFAParseQuantifier(xmlRegParserCtxtPtr ctxt) { |
| 3667 | int cur; |
| 3668 | |
| 3669 | cur = CUR; |
| 3670 | if ((cur == '?') || (cur == '*') || (cur == '+')) { |
| 3671 | if (ctxt->atom != NULL) { |
| 3672 | if (cur == '?') |
| 3673 | ctxt->atom->quant = XML_REGEXP_QUANT_OPT; |
| 3674 | else if (cur == '*') |
| 3675 | ctxt->atom->quant = XML_REGEXP_QUANT_MULT; |
| 3676 | else if (cur == '+') |
| 3677 | ctxt->atom->quant = XML_REGEXP_QUANT_PLUS; |
| 3678 | } |
| 3679 | NEXT; |
| 3680 | return(1); |
| 3681 | } |
| 3682 | if (cur == '{') { |
| 3683 | int min = 0, max = 0; |
| 3684 | |
| 3685 | NEXT; |
| 3686 | cur = xmlFAParseQuantExact(ctxt); |
| 3687 | if (cur >= 0) |
| 3688 | min = cur; |
| 3689 | if (CUR == ',') { |
| 3690 | NEXT; |
| 3691 | cur = xmlFAParseQuantExact(ctxt); |
| 3692 | if (cur >= 0) |
| 3693 | max = cur; |
| 3694 | } |
| 3695 | if (CUR == '}') { |
| 3696 | NEXT; |
| 3697 | } else { |
| 3698 | ERROR("Unterminated quantifier"); |
| 3699 | } |
| 3700 | if (max == 0) |
| 3701 | max = min; |
| 3702 | if (ctxt->atom != NULL) { |
| 3703 | ctxt->atom->quant = XML_REGEXP_QUANT_RANGE; |
| 3704 | ctxt->atom->min = min; |
| 3705 | ctxt->atom->max = max; |
| 3706 | } |
| 3707 | return(1); |
| 3708 | } |
| 3709 | return(0); |
| 3710 | } |
| 3711 | |
| 3712 | /** |
| 3713 | * xmlFAParseAtom: |
Daniel Veillard | 441bc32 | 2002-04-20 17:38:48 +0000 | [diff] [blame] | 3714 | * @ctxt: a regexp parser context |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 3715 | * |
| 3716 | * [9] atom ::= Char | charClass | ( '(' regExp ')' ) |
| 3717 | */ |
| 3718 | static int |
| 3719 | xmlFAParseAtom(xmlRegParserCtxtPtr ctxt) { |
| 3720 | int codepoint, len; |
| 3721 | |
| 3722 | codepoint = xmlFAIsChar(ctxt); |
| 3723 | if (codepoint > 0) { |
| 3724 | ctxt->atom = xmlRegNewAtom(ctxt, XML_REGEXP_CHARVAL); |
| 3725 | if (ctxt->atom == NULL) |
| 3726 | return(-1); |
| 3727 | codepoint = CUR_SCHAR(ctxt->cur, len); |
| 3728 | ctxt->atom->codepoint = codepoint; |
| 3729 | NEXTL(len); |
| 3730 | return(1); |
| 3731 | } else if (CUR == '|') { |
| 3732 | return(0); |
| 3733 | } else if (CUR == 0) { |
| 3734 | return(0); |
| 3735 | } else if (CUR == ')') { |
| 3736 | return(0); |
| 3737 | } else if (CUR == '(') { |
| 3738 | xmlRegStatePtr start, oldend; |
| 3739 | |
| 3740 | NEXT; |
| 3741 | xmlFAGenerateEpsilonTransition(ctxt, ctxt->state, NULL); |
| 3742 | start = ctxt->state; |
| 3743 | oldend = ctxt->end; |
| 3744 | ctxt->end = NULL; |
| 3745 | ctxt->atom = NULL; |
| 3746 | xmlFAParseRegExp(ctxt, 0); |
| 3747 | if (CUR == ')') { |
| 3748 | NEXT; |
| 3749 | } else { |
| 3750 | ERROR("xmlFAParseAtom: expecting ')'"); |
| 3751 | } |
| 3752 | ctxt->atom = xmlRegNewAtom(ctxt, XML_REGEXP_SUBREG); |
| 3753 | if (ctxt->atom == NULL) |
| 3754 | return(-1); |
| 3755 | ctxt->atom->start = start; |
| 3756 | ctxt->atom->stop = ctxt->state; |
| 3757 | ctxt->end = oldend; |
| 3758 | return(1); |
| 3759 | } else if ((CUR == '[') || (CUR == '\\') || (CUR == '.')) { |
| 3760 | xmlFAParseCharClass(ctxt); |
| 3761 | return(1); |
| 3762 | } |
| 3763 | return(0); |
| 3764 | } |
| 3765 | |
| 3766 | /** |
| 3767 | * xmlFAParsePiece: |
Daniel Veillard | 441bc32 | 2002-04-20 17:38:48 +0000 | [diff] [blame] | 3768 | * @ctxt: a regexp parser context |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 3769 | * |
| 3770 | * [3] piece ::= atom quantifier? |
| 3771 | */ |
| 3772 | static int |
| 3773 | xmlFAParsePiece(xmlRegParserCtxtPtr ctxt) { |
| 3774 | int ret; |
| 3775 | |
| 3776 | ctxt->atom = NULL; |
| 3777 | ret = xmlFAParseAtom(ctxt); |
| 3778 | if (ret == 0) |
| 3779 | return(0); |
| 3780 | if (ctxt->atom == NULL) { |
| 3781 | ERROR("internal: no atom generated"); |
| 3782 | } |
| 3783 | xmlFAParseQuantifier(ctxt); |
| 3784 | return(1); |
| 3785 | } |
| 3786 | |
| 3787 | /** |
| 3788 | * xmlFAParseBranch: |
Daniel Veillard | 441bc32 | 2002-04-20 17:38:48 +0000 | [diff] [blame] | 3789 | * @ctxt: a regexp parser context |
| 3790 | * @first: is taht the first |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 3791 | * |
| 3792 | * [2] branch ::= piece* |
Daniel Veillard | a76fe5c | 2003-04-24 16:06:47 +0000 | [diff] [blame] | 3793 | 8 |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 3794 | */ |
Daniel Veillard | a76fe5c | 2003-04-24 16:06:47 +0000 | [diff] [blame] | 3795 | static int |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 3796 | xmlFAParseBranch(xmlRegParserCtxtPtr ctxt, int first) { |
| 3797 | xmlRegStatePtr previous; |
| 3798 | xmlRegAtomPtr prevatom = NULL; |
| 3799 | int ret; |
| 3800 | |
| 3801 | previous = ctxt->state; |
| 3802 | ret = xmlFAParsePiece(ctxt); |
| 3803 | if (ret != 0) { |
| 3804 | if (first) { |
Daniel Veillard | a76fe5c | 2003-04-24 16:06:47 +0000 | [diff] [blame] | 3805 | if (xmlFAGenerateTransitions(ctxt, previous, NULL, ctxt->atom) < 0) |
| 3806 | return(-1); |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 3807 | previous = ctxt->state; |
| 3808 | } else { |
| 3809 | prevatom = ctxt->atom; |
| 3810 | } |
| 3811 | ctxt->atom = NULL; |
| 3812 | } |
| 3813 | while ((ret != 0) && (ctxt->error == 0)) { |
| 3814 | ret = xmlFAParsePiece(ctxt); |
| 3815 | if (ret != 0) { |
| 3816 | if (first) { |
Daniel Veillard | a76fe5c | 2003-04-24 16:06:47 +0000 | [diff] [blame] | 3817 | if (xmlFAGenerateTransitions(ctxt, previous, NULL, |
| 3818 | ctxt->atom) < 0) |
| 3819 | return(-1); |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 3820 | } else { |
Daniel Veillard | a76fe5c | 2003-04-24 16:06:47 +0000 | [diff] [blame] | 3821 | if (xmlFAGenerateTransitions(ctxt, previous, NULL, |
| 3822 | prevatom) < 0) |
| 3823 | return(-1); |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 3824 | prevatom = ctxt->atom; |
| 3825 | } |
| 3826 | previous = ctxt->state; |
| 3827 | ctxt->atom = NULL; |
| 3828 | } |
| 3829 | } |
| 3830 | if (!first) { |
Daniel Veillard | a76fe5c | 2003-04-24 16:06:47 +0000 | [diff] [blame] | 3831 | if (xmlFAGenerateTransitions(ctxt, previous, ctxt->end, prevatom) < 0) |
| 3832 | return(-1); |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 3833 | } |
Daniel Veillard | a76fe5c | 2003-04-24 16:06:47 +0000 | [diff] [blame] | 3834 | return(0); |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 3835 | } |
| 3836 | |
| 3837 | /** |
| 3838 | * xmlFAParseRegExp: |
Daniel Veillard | 441bc32 | 2002-04-20 17:38:48 +0000 | [diff] [blame] | 3839 | * @ctxt: a regexp parser context |
| 3840 | * @top: is that the top-level expressions ? |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 3841 | * |
| 3842 | * [1] regExp ::= branch ( '|' branch )* |
| 3843 | */ |
| 3844 | static void |
| 3845 | xmlFAParseRegExp(xmlRegParserCtxtPtr ctxt, int top) { |
| 3846 | xmlRegStatePtr start, end, oldend; |
| 3847 | |
| 3848 | oldend = ctxt->end; |
| 3849 | |
| 3850 | start = ctxt->state; |
| 3851 | xmlFAParseBranch(ctxt, (ctxt->end == NULL)); |
| 3852 | if (CUR != '|') { |
| 3853 | ctxt->end = ctxt->state; |
| 3854 | return; |
| 3855 | } |
| 3856 | end = ctxt->state; |
| 3857 | while ((CUR == '|') && (ctxt->error == 0)) { |
| 3858 | NEXT; |
| 3859 | ctxt->state = start; |
| 3860 | ctxt->end = end; |
| 3861 | xmlFAParseBranch(ctxt, 0); |
| 3862 | } |
| 3863 | if (!top) |
| 3864 | ctxt->end = oldend; |
| 3865 | } |
| 3866 | |
| 3867 | /************************************************************************ |
| 3868 | * * |
| 3869 | * The basic API * |
| 3870 | * * |
| 3871 | ************************************************************************/ |
| 3872 | |
| 3873 | /** |
| 3874 | * xmlRegexpPrint: |
| 3875 | * @output: the file for the output debug |
| 3876 | * @regexp: the compiled regexp |
| 3877 | * |
| 3878 | * Print the content of the compiled regular expression |
| 3879 | */ |
| 3880 | void |
| 3881 | xmlRegexpPrint(FILE *output, xmlRegexpPtr regexp) { |
| 3882 | int i; |
| 3883 | |
| 3884 | fprintf(output, " regexp: "); |
| 3885 | if (regexp == NULL) { |
| 3886 | fprintf(output, "NULL\n"); |
| 3887 | return; |
| 3888 | } |
| 3889 | fprintf(output, "'%s' ", regexp->string); |
| 3890 | fprintf(output, "\n"); |
| 3891 | fprintf(output, "%d atoms:\n", regexp->nbAtoms); |
| 3892 | for (i = 0;i < regexp->nbAtoms; i++) { |
| 3893 | fprintf(output, " %02d ", i); |
| 3894 | xmlRegPrintAtom(output, regexp->atoms[i]); |
| 3895 | } |
| 3896 | fprintf(output, "%d states:", regexp->nbStates); |
| 3897 | fprintf(output, "\n"); |
| 3898 | for (i = 0;i < regexp->nbStates; i++) { |
| 3899 | xmlRegPrintState(output, regexp->states[i]); |
| 3900 | } |
| 3901 | fprintf(output, "%d counters:\n", regexp->nbCounters); |
| 3902 | for (i = 0;i < regexp->nbCounters; i++) { |
| 3903 | fprintf(output, " %d: min %d max %d\n", i, regexp->counters[i].min, |
| 3904 | regexp->counters[i].max); |
| 3905 | } |
| 3906 | } |
| 3907 | |
| 3908 | /** |
| 3909 | * xmlRegexpCompile: |
| 3910 | * @regexp: a regular expression string |
| 3911 | * |
| 3912 | * Parses a regular expression conforming to XML Schemas Part 2 Datatype |
| 3913 | * Appendix F and build an automata suitable for testing strings against |
| 3914 | * that regular expression |
| 3915 | * |
| 3916 | * Returns the compiled expression or NULL in case of error |
| 3917 | */ |
| 3918 | xmlRegexpPtr |
| 3919 | xmlRegexpCompile(const xmlChar *regexp) { |
| 3920 | xmlRegexpPtr ret; |
| 3921 | xmlRegParserCtxtPtr ctxt; |
| 3922 | |
| 3923 | ctxt = xmlRegNewParserCtxt(regexp); |
| 3924 | if (ctxt == NULL) |
| 3925 | return(NULL); |
| 3926 | |
| 3927 | /* initialize the parser */ |
| 3928 | ctxt->end = NULL; |
| 3929 | ctxt->start = ctxt->state = xmlRegNewState(ctxt); |
| 3930 | xmlRegStatePush(ctxt, ctxt->start); |
| 3931 | |
| 3932 | /* parse the expression building an automata */ |
| 3933 | xmlFAParseRegExp(ctxt, 1); |
| 3934 | if (CUR != 0) { |
| 3935 | ERROR("xmlFAParseRegExp: extra characters"); |
| 3936 | } |
| 3937 | ctxt->end = ctxt->state; |
| 3938 | ctxt->start->type = XML_REGEXP_START_STATE; |
| 3939 | ctxt->end->type = XML_REGEXP_FINAL_STATE; |
| 3940 | |
| 3941 | /* remove the Epsilon except for counted transitions */ |
| 3942 | xmlFAEliminateEpsilonTransitions(ctxt); |
| 3943 | |
| 3944 | |
| 3945 | if (ctxt->error != 0) { |
| 3946 | xmlRegFreeParserCtxt(ctxt); |
| 3947 | return(NULL); |
| 3948 | } |
| 3949 | ret = xmlRegEpxFromParse(ctxt); |
| 3950 | xmlRegFreeParserCtxt(ctxt); |
| 3951 | return(ret); |
| 3952 | } |
| 3953 | |
| 3954 | /** |
| 3955 | * xmlRegexpExec: |
| 3956 | * @comp: the compiled regular expression |
| 3957 | * @content: the value to check against the regular expression |
| 3958 | * |
| 3959 | * Check if the regular expression generate the value |
| 3960 | * |
| 3961 | * Returns 1 if it matches, 0 if not and a negativa value in case of error |
| 3962 | */ |
| 3963 | int |
| 3964 | xmlRegexpExec(xmlRegexpPtr comp, const xmlChar *content) { |
| 3965 | if ((comp == NULL) || (content == NULL)) |
| 3966 | return(-1); |
| 3967 | return(xmlFARegExec(comp, content)); |
| 3968 | } |
| 3969 | |
| 3970 | /** |
Daniel Veillard | 23e7357 | 2002-09-19 19:56:43 +0000 | [diff] [blame] | 3971 | * xmlRegexpIsDeterminist: |
| 3972 | * @comp: the compiled regular expression |
| 3973 | * |
| 3974 | * Check if the regular expression is determinist |
| 3975 | * |
| 3976 | * Returns 1 if it yes, 0 if not and a negativa value in case of error |
| 3977 | */ |
| 3978 | int |
| 3979 | xmlRegexpIsDeterminist(xmlRegexpPtr comp) { |
| 3980 | xmlAutomataPtr am; |
| 3981 | int ret; |
| 3982 | |
| 3983 | if (comp == NULL) |
| 3984 | return(-1); |
| 3985 | if (comp->determinist != -1) |
| 3986 | return(comp->determinist); |
| 3987 | |
| 3988 | am = xmlNewAutomata(); |
Daniel Veillard | bd9afb5 | 2002-09-25 22:25:35 +0000 | [diff] [blame] | 3989 | if (am->states != NULL) { |
| 3990 | int i; |
| 3991 | |
| 3992 | for (i = 0;i < am->nbStates;i++) |
| 3993 | xmlRegFreeState(am->states[i]); |
| 3994 | xmlFree(am->states); |
| 3995 | } |
Daniel Veillard | 23e7357 | 2002-09-19 19:56:43 +0000 | [diff] [blame] | 3996 | am->nbAtoms = comp->nbAtoms; |
| 3997 | am->atoms = comp->atoms; |
| 3998 | am->nbStates = comp->nbStates; |
| 3999 | am->states = comp->states; |
| 4000 | am->determinist = -1; |
| 4001 | ret = xmlFAComputesDeterminism(am); |
| 4002 | am->atoms = NULL; |
| 4003 | am->states = NULL; |
| 4004 | xmlFreeAutomata(am); |
| 4005 | return(ret); |
| 4006 | } |
| 4007 | |
| 4008 | /** |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 4009 | * xmlRegFreeRegexp: |
| 4010 | * @regexp: the regexp |
| 4011 | * |
| 4012 | * Free a regexp |
| 4013 | */ |
| 4014 | void |
| 4015 | xmlRegFreeRegexp(xmlRegexpPtr regexp) { |
| 4016 | int i; |
| 4017 | if (regexp == NULL) |
| 4018 | return; |
| 4019 | |
| 4020 | if (regexp->string != NULL) |
| 4021 | xmlFree(regexp->string); |
| 4022 | if (regexp->states != NULL) { |
| 4023 | for (i = 0;i < regexp->nbStates;i++) |
| 4024 | xmlRegFreeState(regexp->states[i]); |
| 4025 | xmlFree(regexp->states); |
| 4026 | } |
| 4027 | if (regexp->atoms != NULL) { |
| 4028 | for (i = 0;i < regexp->nbAtoms;i++) |
| 4029 | xmlRegFreeAtom(regexp->atoms[i]); |
| 4030 | xmlFree(regexp->atoms); |
| 4031 | } |
| 4032 | if (regexp->counters != NULL) |
| 4033 | xmlFree(regexp->counters); |
Daniel Veillard | 23e7357 | 2002-09-19 19:56:43 +0000 | [diff] [blame] | 4034 | if (regexp->compact != NULL) |
| 4035 | xmlFree(regexp->compact); |
Daniel Veillard | 118aed7 | 2002-09-24 14:13:13 +0000 | [diff] [blame] | 4036 | if (regexp->transdata != NULL) |
| 4037 | xmlFree(regexp->transdata); |
Daniel Veillard | 23e7357 | 2002-09-19 19:56:43 +0000 | [diff] [blame] | 4038 | if (regexp->stringMap != NULL) { |
| 4039 | for (i = 0; i < regexp->nbstrings;i++) |
| 4040 | xmlFree(regexp->stringMap[i]); |
| 4041 | xmlFree(regexp->stringMap); |
| 4042 | } |
| 4043 | |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 4044 | xmlFree(regexp); |
| 4045 | } |
| 4046 | |
| 4047 | #ifdef LIBXML_AUTOMATA_ENABLED |
| 4048 | /************************************************************************ |
| 4049 | * * |
| 4050 | * The Automata interface * |
| 4051 | * * |
| 4052 | ************************************************************************/ |
| 4053 | |
| 4054 | /** |
| 4055 | * xmlNewAutomata: |
| 4056 | * |
| 4057 | * Create a new automata |
| 4058 | * |
| 4059 | * Returns the new object or NULL in case of failure |
| 4060 | */ |
| 4061 | xmlAutomataPtr |
| 4062 | xmlNewAutomata(void) { |
| 4063 | xmlAutomataPtr ctxt; |
| 4064 | |
| 4065 | ctxt = xmlRegNewParserCtxt(NULL); |
| 4066 | if (ctxt == NULL) |
| 4067 | return(NULL); |
| 4068 | |
| 4069 | /* initialize the parser */ |
| 4070 | ctxt->end = NULL; |
| 4071 | ctxt->start = ctxt->state = xmlRegNewState(ctxt); |
Daniel Veillard | a76fe5c | 2003-04-24 16:06:47 +0000 | [diff] [blame] | 4072 | if (ctxt->start == NULL) { |
| 4073 | xmlFreeAutomata(ctxt); |
| 4074 | return(NULL); |
| 4075 | } |
| 4076 | if (xmlRegStatePush(ctxt, ctxt->start) < 0) { |
| 4077 | xmlRegFreeState(ctxt->start); |
| 4078 | xmlFreeAutomata(ctxt); |
| 4079 | return(NULL); |
| 4080 | } |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 4081 | |
| 4082 | return(ctxt); |
| 4083 | } |
| 4084 | |
| 4085 | /** |
| 4086 | * xmlFreeAutomata: |
| 4087 | * @am: an automata |
| 4088 | * |
| 4089 | * Free an automata |
| 4090 | */ |
| 4091 | void |
| 4092 | xmlFreeAutomata(xmlAutomataPtr am) { |
| 4093 | if (am == NULL) |
| 4094 | return; |
| 4095 | xmlRegFreeParserCtxt(am); |
| 4096 | } |
| 4097 | |
| 4098 | /** |
| 4099 | * xmlAutomataGetInitState: |
| 4100 | * @am: an automata |
| 4101 | * |
Daniel Veillard | a9b66d0 | 2002-12-11 14:23:49 +0000 | [diff] [blame] | 4102 | * Initial state lookup |
| 4103 | * |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 4104 | * Returns the initial state of the automata |
| 4105 | */ |
| 4106 | xmlAutomataStatePtr |
| 4107 | xmlAutomataGetInitState(xmlAutomataPtr am) { |
| 4108 | if (am == NULL) |
| 4109 | return(NULL); |
| 4110 | return(am->start); |
| 4111 | } |
| 4112 | |
| 4113 | /** |
| 4114 | * xmlAutomataSetFinalState: |
| 4115 | * @am: an automata |
| 4116 | * @state: a state in this automata |
| 4117 | * |
| 4118 | * Makes that state a final state |
| 4119 | * |
| 4120 | * Returns 0 or -1 in case of error |
| 4121 | */ |
| 4122 | int |
| 4123 | xmlAutomataSetFinalState(xmlAutomataPtr am, xmlAutomataStatePtr state) { |
| 4124 | if ((am == NULL) || (state == NULL)) |
| 4125 | return(-1); |
| 4126 | state->type = XML_REGEXP_FINAL_STATE; |
| 4127 | return(0); |
| 4128 | } |
| 4129 | |
| 4130 | /** |
| 4131 | * xmlAutomataNewTransition: |
| 4132 | * @am: an automata |
| 4133 | * @from: the starting point of the transition |
| 4134 | * @to: the target point of the transition or NULL |
| 4135 | * @token: the input string associated to that transition |
| 4136 | * @data: data passed to the callback function if the transition is activated |
| 4137 | * |
| 4138 | * If @to is NULL, this create first a new target state in the automata |
| 4139 | * and then adds a transition from the @from state to the target state |
| 4140 | * activated by the value of @token |
| 4141 | * |
| 4142 | * Returns the target state or NULL in case of error |
| 4143 | */ |
| 4144 | xmlAutomataStatePtr |
| 4145 | xmlAutomataNewTransition(xmlAutomataPtr am, xmlAutomataStatePtr from, |
| 4146 | xmlAutomataStatePtr to, const xmlChar *token, |
| 4147 | void *data) { |
| 4148 | xmlRegAtomPtr atom; |
| 4149 | |
| 4150 | if ((am == NULL) || (from == NULL) || (token == NULL)) |
| 4151 | return(NULL); |
| 4152 | atom = xmlRegNewAtom(am, XML_REGEXP_STRING); |
Daniel Veillard | a76fe5c | 2003-04-24 16:06:47 +0000 | [diff] [blame] | 4153 | if (atom == NULL) |
| 4154 | return(NULL); |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 4155 | atom->data = data; |
| 4156 | if (atom == NULL) |
| 4157 | return(NULL); |
| 4158 | atom->valuep = xmlStrdup(token); |
| 4159 | |
Daniel Veillard | a76fe5c | 2003-04-24 16:06:47 +0000 | [diff] [blame] | 4160 | if (xmlFAGenerateTransitions(am, from, to, atom) < 0) { |
| 4161 | xmlRegFreeAtom(atom); |
| 4162 | return(NULL); |
| 4163 | } |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 4164 | if (to == NULL) |
| 4165 | return(am->state); |
| 4166 | return(to); |
| 4167 | } |
| 4168 | |
| 4169 | /** |
Daniel Veillard | 52b48c7 | 2003-04-13 19:53:42 +0000 | [diff] [blame] | 4170 | * xmlAutomataNewTransition2: |
| 4171 | * @am: an automata |
| 4172 | * @from: the starting point of the transition |
| 4173 | * @to: the target point of the transition or NULL |
| 4174 | * @token: the first input string associated to that transition |
| 4175 | * @token2: the second input string associated to that transition |
| 4176 | * @data: data passed to the callback function if the transition is activated |
| 4177 | * |
| 4178 | * If @to is NULL, this create first a new target state in the automata |
| 4179 | * and then adds a transition from the @from state to the target state |
| 4180 | * activated by the value of @token |
| 4181 | * |
| 4182 | * Returns the target state or NULL in case of error |
| 4183 | */ |
| 4184 | xmlAutomataStatePtr |
| 4185 | xmlAutomataNewTransition2(xmlAutomataPtr am, xmlAutomataStatePtr from, |
| 4186 | xmlAutomataStatePtr to, const xmlChar *token, |
| 4187 | const xmlChar *token2, void *data) { |
| 4188 | xmlRegAtomPtr atom; |
| 4189 | |
| 4190 | if ((am == NULL) || (from == NULL) || (token == NULL)) |
| 4191 | return(NULL); |
| 4192 | atom = xmlRegNewAtom(am, XML_REGEXP_STRING); |
| 4193 | atom->data = data; |
| 4194 | if (atom == NULL) |
| 4195 | return(NULL); |
| 4196 | if ((token2 == NULL) || (*token2 == 0)) { |
| 4197 | atom->valuep = xmlStrdup(token); |
| 4198 | } else { |
| 4199 | int lenn, lenp; |
| 4200 | xmlChar *str; |
| 4201 | |
| 4202 | lenn = strlen((char *) token2); |
| 4203 | lenp = strlen((char *) token); |
| 4204 | |
Daniel Veillard | 3c908dc | 2003-04-19 00:07:51 +0000 | [diff] [blame] | 4205 | str = (xmlChar *) xmlMallocAtomic(lenn + lenp + 2); |
Daniel Veillard | 52b48c7 | 2003-04-13 19:53:42 +0000 | [diff] [blame] | 4206 | if (str == NULL) { |
| 4207 | xmlRegFreeAtom(atom); |
| 4208 | return(NULL); |
| 4209 | } |
| 4210 | memcpy(&str[0], token, lenp); |
| 4211 | str[lenp] = '|'; |
| 4212 | memcpy(&str[lenp + 1], token2, lenn); |
| 4213 | str[lenn + lenp + 1] = 0; |
| 4214 | |
| 4215 | atom->valuep = str; |
| 4216 | } |
| 4217 | |
Daniel Veillard | a76fe5c | 2003-04-24 16:06:47 +0000 | [diff] [blame] | 4218 | if (xmlFAGenerateTransitions(am, from, to, atom) < 0) { |
| 4219 | xmlRegFreeAtom(atom); |
| 4220 | return(NULL); |
| 4221 | } |
Daniel Veillard | 52b48c7 | 2003-04-13 19:53:42 +0000 | [diff] [blame] | 4222 | if (to == NULL) |
| 4223 | return(am->state); |
| 4224 | return(to); |
| 4225 | } |
| 4226 | |
| 4227 | /** |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 4228 | * xmlAutomataNewCountTrans: |
| 4229 | * @am: an automata |
| 4230 | * @from: the starting point of the transition |
| 4231 | * @to: the target point of the transition or NULL |
| 4232 | * @token: the input string associated to that transition |
| 4233 | * @min: the minimum successive occurences of token |
Daniel Veillard | a9b66d0 | 2002-12-11 14:23:49 +0000 | [diff] [blame] | 4234 | * @max: the maximum successive occurences of token |
| 4235 | * @data: data associated to the transition |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 4236 | * |
| 4237 | * If @to is NULL, this create first a new target state in the automata |
| 4238 | * and then adds a transition from the @from state to the target state |
| 4239 | * activated by a succession of input of value @token and whose number |
| 4240 | * is between @min and @max |
| 4241 | * |
| 4242 | * Returns the target state or NULL in case of error |
| 4243 | */ |
| 4244 | xmlAutomataStatePtr |
| 4245 | xmlAutomataNewCountTrans(xmlAutomataPtr am, xmlAutomataStatePtr from, |
| 4246 | xmlAutomataStatePtr to, const xmlChar *token, |
| 4247 | int min, int max, void *data) { |
| 4248 | xmlRegAtomPtr atom; |
| 4249 | |
| 4250 | if ((am == NULL) || (from == NULL) || (token == NULL)) |
| 4251 | return(NULL); |
| 4252 | if (min < 0) |
| 4253 | return(NULL); |
| 4254 | if ((max < min) || (max < 1)) |
| 4255 | return(NULL); |
| 4256 | atom = xmlRegNewAtom(am, XML_REGEXP_STRING); |
| 4257 | if (atom == NULL) |
| 4258 | return(NULL); |
| 4259 | atom->valuep = xmlStrdup(token); |
| 4260 | atom->data = data; |
| 4261 | if (min == 0) |
| 4262 | atom->min = 1; |
| 4263 | else |
| 4264 | atom->min = min; |
| 4265 | atom->max = max; |
| 4266 | |
Daniel Veillard | a76fe5c | 2003-04-24 16:06:47 +0000 | [diff] [blame] | 4267 | if (xmlFAGenerateTransitions(am, from, to, atom) < 0) { |
| 4268 | xmlRegFreeAtom(atom); |
| 4269 | return(NULL); |
| 4270 | } |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 4271 | if (to == NULL) |
| 4272 | to = am->state; |
| 4273 | if (to == NULL) |
| 4274 | return(NULL); |
| 4275 | if (min == 0) |
| 4276 | xmlFAGenerateEpsilonTransition(am, from, to); |
| 4277 | return(to); |
| 4278 | } |
| 4279 | |
| 4280 | /** |
Daniel Veillard | 7646b18 | 2002-04-20 06:41:40 +0000 | [diff] [blame] | 4281 | * xmlAutomataNewOnceTrans: |
| 4282 | * @am: an automata |
| 4283 | * @from: the starting point of the transition |
| 4284 | * @to: the target point of the transition or NULL |
| 4285 | * @token: the input string associated to that transition |
| 4286 | * @min: the minimum successive occurences of token |
Daniel Veillard | a9b66d0 | 2002-12-11 14:23:49 +0000 | [diff] [blame] | 4287 | * @max: the maximum successive occurences of token |
| 4288 | * @data: data associated to the transition |
Daniel Veillard | 7646b18 | 2002-04-20 06:41:40 +0000 | [diff] [blame] | 4289 | * |
| 4290 | * If @to is NULL, this create first a new target state in the automata |
| 4291 | * and then adds a transition from the @from state to the target state |
| 4292 | * activated by a succession of input of value @token and whose number |
| 4293 | * is between @min and @max, moreover that transistion can only be crossed |
| 4294 | * once. |
| 4295 | * |
| 4296 | * Returns the target state or NULL in case of error |
| 4297 | */ |
| 4298 | xmlAutomataStatePtr |
| 4299 | xmlAutomataNewOnceTrans(xmlAutomataPtr am, xmlAutomataStatePtr from, |
| 4300 | xmlAutomataStatePtr to, const xmlChar *token, |
| 4301 | int min, int max, void *data) { |
| 4302 | xmlRegAtomPtr atom; |
| 4303 | int counter; |
| 4304 | |
| 4305 | if ((am == NULL) || (from == NULL) || (token == NULL)) |
| 4306 | return(NULL); |
| 4307 | if (min < 1) |
| 4308 | return(NULL); |
| 4309 | if ((max < min) || (max < 1)) |
| 4310 | return(NULL); |
| 4311 | atom = xmlRegNewAtom(am, XML_REGEXP_STRING); |
| 4312 | if (atom == NULL) |
| 4313 | return(NULL); |
| 4314 | atom->valuep = xmlStrdup(token); |
| 4315 | atom->data = data; |
| 4316 | atom->quant = XML_REGEXP_QUANT_ONCEONLY; |
| 4317 | if (min == 0) |
| 4318 | atom->min = 1; |
| 4319 | else |
| 4320 | atom->min = min; |
| 4321 | atom->max = max; |
| 4322 | /* |
| 4323 | * associate a counter to the transition. |
| 4324 | */ |
| 4325 | counter = xmlRegGetCounter(am); |
| 4326 | am->counters[counter].min = 1; |
| 4327 | am->counters[counter].max = 1; |
| 4328 | |
| 4329 | /* xmlFAGenerateTransitions(am, from, to, atom); */ |
| 4330 | if (to == NULL) { |
| 4331 | to = xmlRegNewState(am); |
| 4332 | xmlRegStatePush(am, to); |
| 4333 | } |
| 4334 | xmlRegStateAddTrans(am, from, atom, to, counter, -1); |
| 4335 | xmlRegAtomPush(am, atom); |
| 4336 | am->state = to; |
| 4337 | if (to == NULL) |
| 4338 | to = am->state; |
| 4339 | if (to == NULL) |
| 4340 | return(NULL); |
| 4341 | return(to); |
| 4342 | } |
| 4343 | |
| 4344 | /** |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 4345 | * xmlAutomataNewState: |
| 4346 | * @am: an automata |
| 4347 | * |
| 4348 | * Create a new disconnected state in the automata |
| 4349 | * |
| 4350 | * Returns the new state or NULL in case of error |
| 4351 | */ |
| 4352 | xmlAutomataStatePtr |
| 4353 | xmlAutomataNewState(xmlAutomataPtr am) { |
| 4354 | xmlAutomataStatePtr to; |
| 4355 | |
| 4356 | if (am == NULL) |
| 4357 | return(NULL); |
| 4358 | to = xmlRegNewState(am); |
| 4359 | xmlRegStatePush(am, to); |
| 4360 | return(to); |
| 4361 | } |
| 4362 | |
| 4363 | /** |
Daniel Veillard | a9b66d0 | 2002-12-11 14:23:49 +0000 | [diff] [blame] | 4364 | * xmlAutomataNewEpsilon: |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 4365 | * @am: an automata |
| 4366 | * @from: the starting point of the transition |
| 4367 | * @to: the target point of the transition or NULL |
| 4368 | * |
| 4369 | * If @to is NULL, this create first a new target state in the automata |
| 4370 | * and then adds a an epsilon transition from the @from state to the |
| 4371 | * target state |
| 4372 | * |
| 4373 | * Returns the target state or NULL in case of error |
| 4374 | */ |
| 4375 | xmlAutomataStatePtr |
| 4376 | xmlAutomataNewEpsilon(xmlAutomataPtr am, xmlAutomataStatePtr from, |
| 4377 | xmlAutomataStatePtr to) { |
| 4378 | if ((am == NULL) || (from == NULL)) |
| 4379 | return(NULL); |
| 4380 | xmlFAGenerateEpsilonTransition(am, from, to); |
| 4381 | if (to == NULL) |
| 4382 | return(am->state); |
| 4383 | return(to); |
| 4384 | } |
| 4385 | |
Daniel Veillard | b509f15 | 2002-04-17 16:28:10 +0000 | [diff] [blame] | 4386 | /** |
Daniel Veillard | 7646b18 | 2002-04-20 06:41:40 +0000 | [diff] [blame] | 4387 | * xmlAutomataNewAllTrans: |
| 4388 | * @am: an automata |
| 4389 | * @from: the starting point of the transition |
| 4390 | * @to: the target point of the transition or NULL |
Daniel Veillard | a9b66d0 | 2002-12-11 14:23:49 +0000 | [diff] [blame] | 4391 | * @lax: allow to transition if not all all transitions have been activated |
Daniel Veillard | 7646b18 | 2002-04-20 06:41:40 +0000 | [diff] [blame] | 4392 | * |
| 4393 | * If @to is NULL, this create first a new target state in the automata |
| 4394 | * and then adds a an ALL transition from the @from state to the |
| 4395 | * target state. That transition is an epsilon transition allowed only when |
| 4396 | * all transitions from the @from node have been activated. |
| 4397 | * |
| 4398 | * Returns the target state or NULL in case of error |
| 4399 | */ |
| 4400 | xmlAutomataStatePtr |
| 4401 | xmlAutomataNewAllTrans(xmlAutomataPtr am, xmlAutomataStatePtr from, |
Daniel Veillard | 441bc32 | 2002-04-20 17:38:48 +0000 | [diff] [blame] | 4402 | xmlAutomataStatePtr to, int lax) { |
Daniel Veillard | 7646b18 | 2002-04-20 06:41:40 +0000 | [diff] [blame] | 4403 | if ((am == NULL) || (from == NULL)) |
| 4404 | return(NULL); |
Daniel Veillard | 441bc32 | 2002-04-20 17:38:48 +0000 | [diff] [blame] | 4405 | xmlFAGenerateAllTransition(am, from, to, lax); |
Daniel Veillard | 7646b18 | 2002-04-20 06:41:40 +0000 | [diff] [blame] | 4406 | if (to == NULL) |
| 4407 | return(am->state); |
| 4408 | return(to); |
| 4409 | } |
| 4410 | |
| 4411 | /** |
Daniel Veillard | b509f15 | 2002-04-17 16:28:10 +0000 | [diff] [blame] | 4412 | * xmlAutomataNewCounter: |
| 4413 | * @am: an automata |
| 4414 | * @min: the minimal value on the counter |
| 4415 | * @max: the maximal value on the counter |
| 4416 | * |
| 4417 | * Create a new counter |
| 4418 | * |
| 4419 | * Returns the counter number or -1 in case of error |
| 4420 | */ |
| 4421 | int |
| 4422 | xmlAutomataNewCounter(xmlAutomataPtr am, int min, int max) { |
| 4423 | int ret; |
| 4424 | |
| 4425 | if (am == NULL) |
| 4426 | return(-1); |
| 4427 | |
| 4428 | ret = xmlRegGetCounter(am); |
| 4429 | if (ret < 0) |
| 4430 | return(-1); |
| 4431 | am->counters[ret].min = min; |
| 4432 | am->counters[ret].max = max; |
| 4433 | return(ret); |
| 4434 | } |
| 4435 | |
| 4436 | /** |
| 4437 | * xmlAutomataNewCountedTrans: |
| 4438 | * @am: an automata |
| 4439 | * @from: the starting point of the transition |
| 4440 | * @to: the target point of the transition or NULL |
| 4441 | * @counter: the counter associated to that transition |
| 4442 | * |
| 4443 | * If @to is NULL, this create first a new target state in the automata |
| 4444 | * and then adds an epsilon transition from the @from state to the target state |
| 4445 | * which will increment the counter provided |
| 4446 | * |
| 4447 | * Returns the target state or NULL in case of error |
| 4448 | */ |
| 4449 | xmlAutomataStatePtr |
| 4450 | xmlAutomataNewCountedTrans(xmlAutomataPtr am, xmlAutomataStatePtr from, |
| 4451 | xmlAutomataStatePtr to, int counter) { |
| 4452 | if ((am == NULL) || (from == NULL) || (counter < 0)) |
| 4453 | return(NULL); |
| 4454 | xmlFAGenerateCountedEpsilonTransition(am, from, to, counter); |
| 4455 | if (to == NULL) |
| 4456 | return(am->state); |
| 4457 | return(to); |
| 4458 | } |
| 4459 | |
| 4460 | /** |
| 4461 | * xmlAutomataNewCounterTrans: |
| 4462 | * @am: an automata |
| 4463 | * @from: the starting point of the transition |
| 4464 | * @to: the target point of the transition or NULL |
| 4465 | * @counter: the counter associated to that transition |
| 4466 | * |
| 4467 | * If @to is NULL, this create first a new target state in the automata |
| 4468 | * and then adds an epsilon transition from the @from state to the target state |
| 4469 | * which will be allowed only if the counter is within the right range. |
| 4470 | * |
| 4471 | * Returns the target state or NULL in case of error |
| 4472 | */ |
| 4473 | xmlAutomataStatePtr |
| 4474 | xmlAutomataNewCounterTrans(xmlAutomataPtr am, xmlAutomataStatePtr from, |
| 4475 | xmlAutomataStatePtr to, int counter) { |
| 4476 | if ((am == NULL) || (from == NULL) || (counter < 0)) |
| 4477 | return(NULL); |
| 4478 | xmlFAGenerateCountedTransition(am, from, to, counter); |
| 4479 | if (to == NULL) |
| 4480 | return(am->state); |
| 4481 | return(to); |
| 4482 | } |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 4483 | |
| 4484 | /** |
| 4485 | * xmlAutomataCompile: |
| 4486 | * @am: an automata |
| 4487 | * |
| 4488 | * Compile the automata into a Reg Exp ready for being executed. |
| 4489 | * The automata should be free after this point. |
| 4490 | * |
| 4491 | * Returns the compiled regexp or NULL in case of error |
| 4492 | */ |
| 4493 | xmlRegexpPtr |
| 4494 | xmlAutomataCompile(xmlAutomataPtr am) { |
| 4495 | xmlRegexpPtr ret; |
| 4496 | |
Daniel Veillard | a76fe5c | 2003-04-24 16:06:47 +0000 | [diff] [blame] | 4497 | if ((am == NULL) || (am->error != 0)) return(NULL); |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 4498 | xmlFAEliminateEpsilonTransitions(am); |
Daniel Veillard | 23e7357 | 2002-09-19 19:56:43 +0000 | [diff] [blame] | 4499 | /* xmlFAComputesDeterminism(am); */ |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 4500 | ret = xmlRegEpxFromParse(am); |
| 4501 | |
| 4502 | return(ret); |
| 4503 | } |
Daniel Veillard | e19fc23 | 2002-04-22 16:01:24 +0000 | [diff] [blame] | 4504 | |
| 4505 | /** |
| 4506 | * xmlAutomataIsDeterminist: |
| 4507 | * @am: an automata |
| 4508 | * |
| 4509 | * Checks if an automata is determinist. |
| 4510 | * |
| 4511 | * Returns 1 if true, 0 if not, and -1 in case of error |
| 4512 | */ |
| 4513 | int |
| 4514 | xmlAutomataIsDeterminist(xmlAutomataPtr am) { |
| 4515 | int ret; |
| 4516 | |
| 4517 | if (am == NULL) |
| 4518 | return(-1); |
| 4519 | |
| 4520 | ret = xmlFAComputesDeterminism(am); |
| 4521 | return(ret); |
| 4522 | } |
Daniel Veillard | 4255d50 | 2002-04-16 15:50:10 +0000 | [diff] [blame] | 4523 | #endif /* LIBXML_AUTOMATA_ENABLED */ |
| 4524 | #endif /* LIBXML_REGEXP_ENABLED */ |