1 #ifndef REGINT_H 2 #define REGINT_H 3 /********************************************************************** 4 regint.h - Oniguruma (regular expression library) 5 **********************************************************************/ 6 /*- 7 * Copyright (c) 2002-2013 K.Kosako <sndgk393 AT ybb DOT ne DOT jp> 8 * All rights reserved. 9 * 10 * Redistribution and use in source and binary forms, with or without 11 * modification, are permitted provided that the following conditions 12 * are met: 13 * 1. Redistributions of source code must retain the above copyright 14 * notice, this list of conditions and the following disclaimer. 15 * 2. Redistributions in binary form must reproduce the above copyright 16 * notice, this list of conditions and the following disclaimer in the 17 * documentation and/or other materials provided with the distribution. 18 * 19 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND 20 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 21 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 22 * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE 23 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 24 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 25 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 26 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 27 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 28 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 29 * SUCH DAMAGE. 30 */ 31 32 /* for debug */ 33 /* #define ONIG_DEBUG_PARSE_TREE */ 34 /* #define ONIG_DEBUG_COMPILE */ 35 /* #define ONIG_DEBUG_SEARCH */ 36 /* #define ONIG_DEBUG_MATCH */ 37 /* #define ONIG_DONT_OPTIMIZE */ 38 39 /* for byte-code statistical data. */ 40 /* #define ONIG_DEBUG_STATISTICS */ 41 42 #if defined(ONIG_DEBUG_PARSE_TREE) || defined(ONIG_DEBUG_MATCH) || \ 43 defined(ONIG_DEBUG_SEARCH) || defined(ONIG_DEBUG_COMPILE) || \ 44 defined(ONIG_DEBUG_STATISTICS) 45 #ifndef ONIG_DEBUG 46 #define ONIG_DEBUG 47 #endif 48 #endif 49 50 #if defined(__i386) || defined(__i386__) || defined(_M_IX86) || \ 51 (defined(__ppc__) && defined(__APPLE__)) || \ 52 defined(__x86_64) || defined(__x86_64__) || \ 53 defined(__mc68020__) 54 #define PLATFORM_UNALIGNED_WORD_ACCESS 55 #endif 56 57 /* config */ 58 /* spec. config */ 59 #define USE_NAMED_GROUP 60 #define USE_SUBEXP_CALL 61 #define USE_BACKREF_WITH_LEVEL /* \k<name+n>, \k<name-n> */ 62 #define USE_MONOMANIAC_CHECK_CAPTURES_IN_ENDLESS_REPEAT /* /(?:()|())*\2/ */ 63 #define USE_NEWLINE_AT_END_OF_STRING_HAS_EMPTY_LINE /* /\n$/ =~ "\n" */ 64 #define USE_WARNING_REDUNDANT_NESTED_REPEAT_OPERATOR 65 /* #define USE_RECOMPILE_API */ 66 /* !!! moved to regenc.h. */ /* #define USE_CRNL_AS_LINE_TERMINATOR */ 67 68 /* internal config */ 69 #define USE_PARSE_TREE_NODE_RECYCLE 70 #define USE_OP_PUSH_OR_JUMP_EXACT 71 #define USE_QTFR_PEEK_NEXT 72 #define USE_ST_LIBRARY 73 #define USE_SHARED_CCLASS_TABLE 74 75 #define INIT_MATCH_STACK_SIZE 160 76 #define DEFAULT_MATCH_STACK_LIMIT_SIZE 0 /* unlimited */ 77 78 #if defined(__GNUC__) 79 # define ARG_UNUSED __attribute__ ((unused)) 80 #else 81 # define ARG_UNUSED 82 #endif 83 84 /* */ 85 /* escape other system UChar definition */ 86 //#include "config.h" 87 #ifdef ONIG_ESCAPE_UCHAR_COLLISION 88 #undef ONIG_ESCAPE_UCHAR_COLLISION 89 #endif 90 91 #define USE_WORD_BEGIN_END /* "\<", "\>" */ 92 #define USE_CAPTURE_HISTORY 93 #define USE_VARIABLE_META_CHARS 94 #define USE_POSIX_API_REGION_OPTION 95 #define USE_FIND_LONGEST_SEARCH_ALL_OF_RANGE 96 /* #define USE_COMBINATION_EXPLOSION_CHECK */ /* (X*)* */ 97 98 /* #define USE_MULTI_THREAD_SYSTEM */ 99 #define THREAD_SYSTEM_INIT /* depend on thread system */ 100 #define THREAD_SYSTEM_END /* depend on thread system */ 101 #define THREAD_ATOMIC_START /* depend on thread system */ 102 #define THREAD_ATOMIC_END /* depend on thread system */ 103 #define THREAD_PASS /* depend on thread system */ 104 #define xmalloc malloc 105 #define xrealloc realloc 106 #define xcalloc calloc 107 #define xfree free 108 109 #define CHECK_INTERRUPT_IN_MATCH_AT 110 111 #define st_init_table onig_st_init_table 112 #define st_init_table_with_size onig_st_init_table_with_size 113 #define st_init_numtable onig_st_init_numtable 114 #define st_init_numtable_with_size onig_st_init_numtable_with_size 115 #define st_init_strtable onig_st_init_strtable 116 #define st_init_strtable_with_size onig_st_init_strtable_with_size 117 #define st_delete onig_st_delete 118 #define st_delete_safe onig_st_delete_safe 119 #define st_insert onig_st_insert 120 #define st_lookup onig_st_lookup 121 #define st_foreach onig_st_foreach 122 #define st_add_direct onig_st_add_direct 123 #define st_free_table onig_st_free_table 124 #define st_cleanup_safe onig_st_cleanup_safe 125 #define st_copy onig_st_copy 126 #define st_nothing_key_clone onig_st_nothing_key_clone 127 #define st_nothing_key_free onig_st_nothing_key_free 128 /* */ 129 #define onig_st_is_member st_is_member 130 131 #define STATE_CHECK_STRING_THRESHOLD_LEN 7 132 #define STATE_CHECK_BUFF_MAX_SIZE 0x4000 133 134 #define THREAD_PASS_LIMIT_COUNT 8 135 //#define xmemset memset 136 //#define xmemcpy memcpy 137 //#define xmemmove memmove 138 139 #if defined(_WIN32) && !defined(__GNUC__) 140 #define xalloca _alloca 141 #define xvsnprintf _vsnprintf 142 #else 143 #define xalloca alloca 144 #define xvsnprintf vsnprintf 145 #endif 146 147 148 #if defined(USE_RECOMPILE_API) && defined(USE_MULTI_THREAD_SYSTEM) 149 #define ONIG_STATE_INC(reg) (reg)->state++ 150 #define ONIG_STATE_DEC(reg) (reg)->state-- 151 152 #define ONIG_STATE_INC_THREAD(reg) do {\ 153 THREAD_ATOMIC_START;\ 154 (reg)->state++;\ 155 THREAD_ATOMIC_END;\ 156 } while(0) 157 #define ONIG_STATE_DEC_THREAD(reg) do {\ 158 THREAD_ATOMIC_START;\ 159 (reg)->state--;\ 160 THREAD_ATOMIC_END;\ 161 } while(0) 162 #else 163 #define ONIG_STATE_INC(reg) /* Nothing */ 164 #define ONIG_STATE_DEC(reg) /* Nothing */ 165 #define ONIG_STATE_INC_THREAD(reg) /* Nothing */ 166 #define ONIG_STATE_DEC_THREAD(reg) /* Nothing */ 167 #endif /* USE_RECOMPILE_API && USE_MULTI_THREAD_SYSTEM */ 168 169 #if 0 170 #ifdef HAVE_STDLIB_H 171 #include <stdlib.h> 172 #endif 173 174 #if defined(HAVE_ALLOCA_H) && !defined(__GNUC__) 175 #include <alloca.h> 176 #endif 177 178 #ifdef HAVE_STRING_H 179 # include <string.h> 180 #else 181 # include <strings.h> 182 #endif 183 184 #include <ctype.h> 185 #ifdef HAVE_SYS_TYPES_H 186 #ifndef __BORLANDC__ 187 #include <sys/types.h> 188 #endif 189 #endif 190 191 #ifdef __BORLANDC__ 192 #include <malloc.h> 193 #endif 194 195 #ifdef ONIG_DEBUG 196 # include <stdio.h> 197 #endif 198 #endif 199 200 #include "regenc.h" 201 202 #ifdef MIN 203 #undef MIN 204 #endif 205 #ifdef MAX 206 #undef MAX 207 #endif 208 #define MIN(a,b) (((a)>(b))?(b):(a)) 209 #define MAX(a,b) (((a)<(b))?(b):(a)) 210 211 #define IS_NULL(p) (((void*)(p)) == (void*)0) 212 #define IS_NOT_NULL(p) (((void*)(p)) != (void*)0) 213 #define CHECK_NULL_RETURN(p) if (IS_NULL(p)) return NULL 214 #define CHECK_NULL_RETURN_MEMERR(p) if (IS_NULL(p)) return ONIGERR_MEMORY 215 #define NULL_UCHARP ((UChar* )0) 216 217 #ifdef PLATFORM_UNALIGNED_WORD_ACCESS 218 219 #define PLATFORM_GET_INC(val,p,type) do{\ 220 val = *(type* )p;\ 221 (p) += sizeof(type);\ 222 } while(0) 223 224 #else 225 226 #define PLATFORM_GET_INC(val,p,type) do{\ 227 xmemcpy(&val, (p), sizeof(type));\ 228 (p) += sizeof(type);\ 229 } while(0) 230 231 /* sizeof(OnigCodePoint) */ 232 #define WORD_ALIGNMENT_SIZE SIZEOF_LONG 233 234 #define GET_ALIGNMENT_PAD_SIZE(addr,pad_size) do {\ 235 (pad_size) = WORD_ALIGNMENT_SIZE \ 236 - ((unsigned int )(UINTN)(addr) % WORD_ALIGNMENT_SIZE);\ 237 if ((pad_size) == WORD_ALIGNMENT_SIZE) (pad_size) = 0;\ 238 } while (0) 239 240 #define ALIGNMENT_RIGHT(addr) do {\ 241 (addr) += (WORD_ALIGNMENT_SIZE - 1);\ 242 (addr) -= ((unsigned int )(UINTN)(addr) % WORD_ALIGNMENT_SIZE);\ 243 } while (0) 244 245 #endif /* PLATFORM_UNALIGNED_WORD_ACCESS */ 246 247 /* stack pop level */ 248 #define STACK_POP_LEVEL_FREE 0 249 #define STACK_POP_LEVEL_MEM_START 1 250 #define STACK_POP_LEVEL_ALL 2 251 252 /* optimize flags */ 253 #define ONIG_OPTIMIZE_NONE 0 254 #define ONIG_OPTIMIZE_EXACT 1 /* Slow Search */ 255 #define ONIG_OPTIMIZE_EXACT_BM 2 /* Boyer Moore Search */ 256 #define ONIG_OPTIMIZE_EXACT_BM_NOT_REV 3 /* BM (but not simple match) */ 257 #define ONIG_OPTIMIZE_EXACT_IC 4 /* Slow Search (ignore case) */ 258 #define ONIG_OPTIMIZE_MAP 5 /* char map */ 259 260 /* bit status */ 261 typedef unsigned int BitStatusType; 262 263 #define BIT_STATUS_BITS_NUM (sizeof(BitStatusType) * 8) 264 #define BIT_STATUS_CLEAR(stats) (stats) = 0 265 #define BIT_STATUS_ON_ALL(stats) (stats) = ~((BitStatusType )0) 266 #define BIT_STATUS_AT(stats,n) \ 267 ((n) < (int )BIT_STATUS_BITS_NUM ? ((stats) & (1 << n)) : ((stats) & 1)) 268 269 #define BIT_STATUS_ON_AT(stats,n) do {\ 270 if ((n) < (int )BIT_STATUS_BITS_NUM) \ 271 (stats) |= (1 << (n));\ 272 else\ 273 (stats) |= 1;\ 274 } while (0) 275 276 #define BIT_STATUS_ON_AT_SIMPLE(stats,n) do {\ 277 if ((n) < (int )BIT_STATUS_BITS_NUM)\ 278 (stats) |= (1 << (n));\ 279 } while (0) 280 281 282 #define INT_MAX_LIMIT ((1UL << (SIZEOF_INT * 8 - 1)) - 1) 283 284 #define DIGITVAL(code) ((code) - '0') 285 #define ODIGITVAL(code) DIGITVAL(code) 286 #define XDIGITVAL(enc,code) \ 287 (ONIGENC_IS_CODE_DIGIT(enc,code) ? DIGITVAL(code) \ 288 : (ONIGENC_IS_CODE_UPPER(enc,code) ? (code) - 'A' + 10 : (code) - 'a' + 10)) 289 290 #define IS_SINGLELINE(option) ((option) & ONIG_OPTION_SINGLELINE) 291 #define IS_MULTILINE(option) ((option) & ONIG_OPTION_MULTILINE) 292 #define IS_IGNORECASE(option) ((option) & ONIG_OPTION_IGNORECASE) 293 #define IS_EXTEND(option) ((option) & ONIG_OPTION_EXTEND) 294 #define IS_FIND_LONGEST(option) ((option) & ONIG_OPTION_FIND_LONGEST) 295 #define IS_FIND_NOT_EMPTY(option) ((option) & ONIG_OPTION_FIND_NOT_EMPTY) 296 #define IS_FIND_CONDITION(option) ((option) & \ 297 (ONIG_OPTION_FIND_LONGEST | ONIG_OPTION_FIND_NOT_EMPTY)) 298 #define IS_NOTBOL(option) ((option) & ONIG_OPTION_NOTBOL) 299 #define IS_NOTEOL(option) ((option) & ONIG_OPTION_NOTEOL) 300 #define IS_POSIX_REGION(option) ((option) & ONIG_OPTION_POSIX_REGION) 301 302 /* OP_SET_OPTION is required for these options. 303 #define IS_DYNAMIC_OPTION(option) \ 304 (((option) & (ONIG_OPTION_MULTILINE | ONIG_OPTION_IGNORECASE)) != 0) 305 */ 306 /* ignore-case and multibyte status are included in compiled code. */ 307 #define IS_DYNAMIC_OPTION(option) 0 308 309 #define DISABLE_CASE_FOLD_MULTI_CHAR(case_fold_flag) \ 310 ((case_fold_flag) & ~INTERNAL_ONIGENC_CASE_FOLD_MULTI_CHAR) 311 312 #define REPEAT_INFINITE -1 313 #define IS_REPEAT_INFINITE(n) ((n) == REPEAT_INFINITE) 314 315 /* bitset */ 316 #define BITS_PER_BYTE 8 317 #define SINGLE_BYTE_SIZE (1 << BITS_PER_BYTE) 318 #define BITS_IN_ROOM (sizeof(Bits) * BITS_PER_BYTE) 319 #define BITSET_SIZE (SINGLE_BYTE_SIZE / BITS_IN_ROOM) 320 321 #ifdef PLATFORM_UNALIGNED_WORD_ACCESS 322 typedef unsigned int Bits; 323 #else 324 typedef unsigned char Bits; 325 #endif 326 typedef Bits BitSet[BITSET_SIZE]; 327 typedef Bits* BitSetRef; 328 329 #define SIZE_BITSET sizeof(BitSet) 330 331 #define BITSET_CLEAR(bs) do {\ 332 int id;\ 333 for (id = 0; id < (int )BITSET_SIZE; id++) { (bs)[id] = 0; } \ 334 } while (0) 335 336 #define BS_ROOM(bs,pos) (bs)[pos / BITS_IN_ROOM] 337 #define BS_BIT(pos) (1 << (pos % BITS_IN_ROOM)) 338 339 #define BITSET_AT(bs, pos) (BS_ROOM(bs,pos) & BS_BIT(pos)) 340 #define BITSET_SET_BIT(bs, pos) BS_ROOM(bs,pos) |= BS_BIT(pos) 341 #define BITSET_CLEAR_BIT(bs, pos) BS_ROOM(bs,pos) &= ~(BS_BIT(pos)) 342 #define BITSET_INVERT_BIT(bs, pos) BS_ROOM(bs,pos) ^= BS_BIT(pos) 343 344 /* bytes buffer */ 345 typedef struct _BBuf { 346 UChar* p; 347 unsigned int used; 348 unsigned int alloc; 349 } BBuf; 350 351 #define BBUF_INIT(buf,size) onig_bbuf_init((BBuf* )(buf), (size)) 352 353 #define BBUF_SIZE_INC(buf,inc) do{\ 354 (buf)->alloc += (inc);\ 355 (buf)->p = (UChar* )xrealloc((buf)->p, (buf)->alloc);\ 356 if (IS_NULL((buf)->p)) return(ONIGERR_MEMORY);\ 357 } while (0) 358 359 #define BBUF_EXPAND(buf,low) do{\ 360 unsigned int OldSize_ = (buf)->alloc * sizeof((buf)->p[0]);\ 361 do { (buf)->alloc *= 2; } while ((buf)->alloc < (unsigned int )low);\ 362 (buf)->p = (UChar* )xrealloc((buf)->p, (buf)->alloc, OldSize_);\ 363 if (IS_NULL((buf)->p)) return(ONIGERR_MEMORY);\ 364 } while (0) 365 366 #define BBUF_ENSURE_SIZE(buf,size) do{\ 367 unsigned int new_alloc = (buf)->alloc;\ 368 while (new_alloc < (unsigned int )(size)) { new_alloc *= 2; }\ 369 if ((buf)->alloc != new_alloc) {\ 370 (buf)->p = (UChar* )xrealloc((buf)->p, new_alloc, (buf)->alloc);\ 371 if (IS_NULL((buf)->p)) return(ONIGERR_MEMORY);\ 372 (buf)->alloc = new_alloc;\ 373 }\ 374 } while (0) 375 376 #define BBUF_WRITE(buf,pos,bytes,n) do{\ 377 int used = (pos) + (n);\ 378 if ((buf)->alloc < (unsigned int )used) BBUF_EXPAND((buf),used);\ 379 xmemcpy((buf)->p + (pos), (bytes), (n));\ 380 if ((buf)->used < (unsigned int )used) (buf)->used = used;\ 381 } while (0) 382 383 #define BBUF_WRITE1(buf,pos,byte) do{\ 384 int used = (pos) + 1;\ 385 if ((buf)->alloc < (unsigned int )used) BBUF_EXPAND((buf),used);\ 386 (buf)->p[(pos)] = (byte);\ 387 if ((buf)->used < (unsigned int )used) (buf)->used = used;\ 388 } while (0) 389 390 #define BBUF_ADD(buf,bytes,n) BBUF_WRITE((buf),(buf)->used,(bytes),(n)) 391 #define BBUF_ADD1(buf,byte) BBUF_WRITE1((buf),(buf)->used,(byte)) 392 #define BBUF_GET_ADD_ADDRESS(buf) ((buf)->p + (buf)->used) 393 #define BBUF_GET_OFFSET_POS(buf) ((buf)->used) 394 395 /* from < to */ 396 #define BBUF_MOVE_RIGHT(buf,from,to,n) do {\ 397 if ((unsigned int )((to)+(n)) > (buf)->alloc) BBUF_EXPAND((buf),(to) + (n));\ 398 xmemmove((buf)->p + (to), (buf)->p + (from), (n));\ 399 if ((unsigned int )((to)+(n)) > (buf)->used) (buf)->used = (to) + (n);\ 400 } while (0) 401 402 /* from > to */ 403 #define BBUF_MOVE_LEFT(buf,from,to,n) do {\ 404 xmemmove((buf)->p + (to), (buf)->p + (from), (n));\ 405 } while (0) 406 407 /* from > to */ 408 #define BBUF_MOVE_LEFT_REDUCE(buf,from,to) do {\ 409 xmemmove((buf)->p + (to), (buf)->p + (from), (buf)->used - (from));\ 410 (buf)->used -= (from - to);\ 411 } while (0) 412 413 #define BBUF_INSERT(buf,pos,bytes,n) do {\ 414 if (pos >= (buf)->used) {\ 415 BBUF_WRITE(buf,pos,bytes,n);\ 416 }\ 417 else {\ 418 BBUF_MOVE_RIGHT((buf),(pos),(pos) + (n),((buf)->used - (pos)));\ 419 xmemcpy((buf)->p + (pos), (bytes), (n));\ 420 }\ 421 } while (0) 422 423 #define BBUF_GET_BYTE(buf, pos) (buf)->p[(pos)] 424 425 426 #define ANCHOR_BEGIN_BUF (1<<0) 427 #define ANCHOR_BEGIN_LINE (1<<1) 428 #define ANCHOR_BEGIN_POSITION (1<<2) 429 #define ANCHOR_END_BUF (1<<3) 430 #define ANCHOR_SEMI_END_BUF (1<<4) 431 #define ANCHOR_END_LINE (1<<5) 432 433 #define ANCHOR_WORD_BOUND (1<<6) 434 #define ANCHOR_NOT_WORD_BOUND (1<<7) 435 #define ANCHOR_WORD_BEGIN (1<<8) 436 #define ANCHOR_WORD_END (1<<9) 437 #define ANCHOR_PREC_READ (1<<10) 438 #define ANCHOR_PREC_READ_NOT (1<<11) 439 #define ANCHOR_LOOK_BEHIND (1<<12) 440 #define ANCHOR_LOOK_BEHIND_NOT (1<<13) 441 442 #define ANCHOR_ANYCHAR_STAR (1<<14) /* ".*" optimize info */ 443 #define ANCHOR_ANYCHAR_STAR_ML (1<<15) /* ".*" optimize info (multi-line) */ 444 445 /* operation code */ 446 enum OpCode { 447 OP_FINISH = 0, /* matching process terminator (no more alternative) */ 448 OP_END = 1, /* pattern code terminator (success end) */ 449 450 OP_EXACT1 = 2, /* single byte, N = 1 */ 451 OP_EXACT2, /* single byte, N = 2 */ 452 OP_EXACT3, /* single byte, N = 3 */ 453 OP_EXACT4, /* single byte, N = 4 */ 454 OP_EXACT5, /* single byte, N = 5 */ 455 OP_EXACTN, /* single byte */ 456 OP_EXACTMB2N1, /* mb-length = 2 N = 1 */ 457 OP_EXACTMB2N2, /* mb-length = 2 N = 2 */ 458 OP_EXACTMB2N3, /* mb-length = 2 N = 3 */ 459 OP_EXACTMB2N, /* mb-length = 2 */ 460 OP_EXACTMB3N, /* mb-length = 3 */ 461 OP_EXACTMBN, /* other length */ 462 463 OP_EXACT1_IC, /* single byte, N = 1, ignore case */ 464 OP_EXACTN_IC, /* single byte, ignore case */ 465 466 OP_CCLASS, 467 OP_CCLASS_MB, 468 OP_CCLASS_MIX, 469 OP_CCLASS_NOT, 470 OP_CCLASS_MB_NOT, 471 OP_CCLASS_MIX_NOT, 472 OP_CCLASS_NODE, /* pointer to CClassNode node */ 473 474 OP_ANYCHAR, /* "." */ 475 OP_ANYCHAR_ML, /* "." multi-line */ 476 OP_ANYCHAR_STAR, /* ".*" */ 477 OP_ANYCHAR_ML_STAR, /* ".*" multi-line */ 478 OP_ANYCHAR_STAR_PEEK_NEXT, 479 OP_ANYCHAR_ML_STAR_PEEK_NEXT, 480 481 OP_WORD, 482 OP_NOT_WORD, 483 OP_WORD_BOUND, 484 OP_NOT_WORD_BOUND, 485 OP_WORD_BEGIN, 486 OP_WORD_END, 487 488 OP_BEGIN_BUF, 489 OP_END_BUF, 490 OP_BEGIN_LINE, 491 OP_END_LINE, 492 OP_SEMI_END_BUF, 493 OP_BEGIN_POSITION, 494 495 OP_BACKREF1, 496 OP_BACKREF2, 497 OP_BACKREFN, 498 OP_BACKREFN_IC, 499 OP_BACKREF_MULTI, 500 OP_BACKREF_MULTI_IC, 501 OP_BACKREF_WITH_LEVEL, /* \k<xxx+n>, \k<xxx-n> */ 502 503 OP_MEMORY_START, 504 OP_MEMORY_START_PUSH, /* push back-tracker to stack */ 505 OP_MEMORY_END_PUSH, /* push back-tracker to stack */ 506 OP_MEMORY_END_PUSH_REC, /* push back-tracker to stack */ 507 OP_MEMORY_END, 508 OP_MEMORY_END_REC, /* push marker to stack */ 509 510 OP_FAIL, /* pop stack and move */ 511 OP_JUMP, 512 OP_PUSH, 513 OP_POP, 514 OP_PUSH_OR_JUMP_EXACT1, /* if match exact then push, else jump. */ 515 OP_PUSH_IF_PEEK_NEXT, /* if match exact then push, else none. */ 516 OP_REPEAT, /* {n,m} */ 517 OP_REPEAT_NG, /* {n,m}? (non greedy) */ 518 OP_REPEAT_INC, 519 OP_REPEAT_INC_NG, /* non greedy */ 520 OP_REPEAT_INC_SG, /* search and get in stack */ 521 OP_REPEAT_INC_NG_SG, /* search and get in stack (non greedy) */ 522 OP_NULL_CHECK_START, /* null loop checker start */ 523 OP_NULL_CHECK_END, /* null loop checker end */ 524 OP_NULL_CHECK_END_MEMST, /* null loop checker end (with capture status) */ 525 OP_NULL_CHECK_END_MEMST_PUSH, /* with capture status and push check-end */ 526 527 OP_PUSH_POS, /* (?=...) start */ 528 OP_POP_POS, /* (?=...) end */ 529 OP_PUSH_POS_NOT, /* (?!...) start */ 530 OP_FAIL_POS, /* (?!...) end */ 531 OP_PUSH_STOP_BT, /* (?>...) start */ 532 OP_POP_STOP_BT, /* (?>...) end */ 533 OP_LOOK_BEHIND, /* (?<=...) start (no needs end opcode) */ 534 OP_PUSH_LOOK_BEHIND_NOT, /* (?<!...) start */ 535 OP_FAIL_LOOK_BEHIND_NOT, /* (?<!...) end */ 536 537 OP_CALL, /* \g<name> */ 538 OP_RETURN, 539 540 OP_STATE_CHECK_PUSH, /* combination explosion check and push */ 541 OP_STATE_CHECK_PUSH_OR_JUMP, /* check ok -> push, else jump */ 542 OP_STATE_CHECK, /* check only */ 543 OP_STATE_CHECK_ANYCHAR_STAR, 544 OP_STATE_CHECK_ANYCHAR_ML_STAR, 545 546 /* no need: IS_DYNAMIC_OPTION() == 0 */ 547 OP_SET_OPTION_PUSH, /* set option and push recover option */ 548 OP_SET_OPTION /* set option */ 549 }; 550 551 typedef int RelAddrType; 552 typedef int AbsAddrType; 553 typedef int LengthType; 554 typedef int RepeatNumType; 555 typedef short int MemNumType; 556 typedef short int StateCheckNumType; 557 typedef void* PointerType; 558 559 #define SIZE_OPCODE 1 560 #define SIZE_RELADDR sizeof(RelAddrType) 561 #define SIZE_ABSADDR sizeof(AbsAddrType) 562 #define SIZE_LENGTH sizeof(LengthType) 563 #define SIZE_MEMNUM sizeof(MemNumType) 564 #define SIZE_STATE_CHECK_NUM sizeof(StateCheckNumType) 565 #define SIZE_REPEATNUM sizeof(RepeatNumType) 566 #define SIZE_OPTION sizeof(OnigOptionType) 567 #define SIZE_CODE_POINT sizeof(OnigCodePoint) 568 #define SIZE_POINTER sizeof(PointerType) 569 570 571 #define GET_RELADDR_INC(addr,p) PLATFORM_GET_INC(addr, p, RelAddrType) 572 #define GET_ABSADDR_INC(addr,p) PLATFORM_GET_INC(addr, p, AbsAddrType) 573 #define GET_LENGTH_INC(len,p) PLATFORM_GET_INC(len, p, LengthType) 574 #define GET_MEMNUM_INC(num,p) PLATFORM_GET_INC(num, p, MemNumType) 575 #define GET_REPEATNUM_INC(num,p) PLATFORM_GET_INC(num, p, RepeatNumType) 576 #define GET_OPTION_INC(option,p) PLATFORM_GET_INC(option, p, OnigOptionType) 577 #define GET_POINTER_INC(ptr,p) PLATFORM_GET_INC(ptr, p, PointerType) 578 #define GET_STATE_CHECK_NUM_INC(num,p) PLATFORM_GET_INC(num, p, StateCheckNumType) 579 580 /* code point's address must be aligned address. */ 581 #define GET_CODE_POINT(code,p) code = *((OnigCodePoint* )(p)) 582 #define GET_BYTE_INC(byte,p) do{\ 583 byte = *(p);\ 584 (p)++;\ 585 } while(0) 586 587 588 /* op-code + arg size */ 589 #define SIZE_OP_ANYCHAR_STAR SIZE_OPCODE 590 #define SIZE_OP_ANYCHAR_STAR_PEEK_NEXT (SIZE_OPCODE + 1) 591 #define SIZE_OP_JUMP (SIZE_OPCODE + SIZE_RELADDR) 592 #define SIZE_OP_PUSH (SIZE_OPCODE + SIZE_RELADDR) 593 #define SIZE_OP_POP SIZE_OPCODE 594 #define SIZE_OP_PUSH_OR_JUMP_EXACT1 (SIZE_OPCODE + SIZE_RELADDR + 1) 595 #define SIZE_OP_PUSH_IF_PEEK_NEXT (SIZE_OPCODE + SIZE_RELADDR + 1) 596 #define SIZE_OP_REPEAT_INC (SIZE_OPCODE + SIZE_MEMNUM) 597 #define SIZE_OP_REPEAT_INC_NG (SIZE_OPCODE + SIZE_MEMNUM) 598 #define SIZE_OP_PUSH_POS SIZE_OPCODE 599 #define SIZE_OP_PUSH_POS_NOT (SIZE_OPCODE + SIZE_RELADDR) 600 #define SIZE_OP_POP_POS SIZE_OPCODE 601 #define SIZE_OP_FAIL_POS SIZE_OPCODE 602 #define SIZE_OP_SET_OPTION (SIZE_OPCODE + SIZE_OPTION) 603 #define SIZE_OP_SET_OPTION_PUSH (SIZE_OPCODE + SIZE_OPTION) 604 #define SIZE_OP_FAIL SIZE_OPCODE 605 #define SIZE_OP_MEMORY_START (SIZE_OPCODE + SIZE_MEMNUM) 606 #define SIZE_OP_MEMORY_START_PUSH (SIZE_OPCODE + SIZE_MEMNUM) 607 #define SIZE_OP_MEMORY_END_PUSH (SIZE_OPCODE + SIZE_MEMNUM) 608 #define SIZE_OP_MEMORY_END_PUSH_REC (SIZE_OPCODE + SIZE_MEMNUM) 609 #define SIZE_OP_MEMORY_END (SIZE_OPCODE + SIZE_MEMNUM) 610 #define SIZE_OP_MEMORY_END_REC (SIZE_OPCODE + SIZE_MEMNUM) 611 #define SIZE_OP_PUSH_STOP_BT SIZE_OPCODE 612 #define SIZE_OP_POP_STOP_BT SIZE_OPCODE 613 #define SIZE_OP_NULL_CHECK_START (SIZE_OPCODE + SIZE_MEMNUM) 614 #define SIZE_OP_NULL_CHECK_END (SIZE_OPCODE + SIZE_MEMNUM) 615 #define SIZE_OP_LOOK_BEHIND (SIZE_OPCODE + SIZE_LENGTH) 616 #define SIZE_OP_PUSH_LOOK_BEHIND_NOT (SIZE_OPCODE + SIZE_RELADDR + SIZE_LENGTH) 617 #define SIZE_OP_FAIL_LOOK_BEHIND_NOT SIZE_OPCODE 618 #define SIZE_OP_CALL (SIZE_OPCODE + SIZE_ABSADDR) 619 #define SIZE_OP_RETURN SIZE_OPCODE 620 621 #ifdef USE_COMBINATION_EXPLOSION_CHECK 622 #define SIZE_OP_STATE_CHECK (SIZE_OPCODE + SIZE_STATE_CHECK_NUM) 623 #define SIZE_OP_STATE_CHECK_PUSH (SIZE_OPCODE + SIZE_STATE_CHECK_NUM + SIZE_RELADDR) 624 #define SIZE_OP_STATE_CHECK_PUSH_OR_JUMP (SIZE_OPCODE + SIZE_STATE_CHECK_NUM + SIZE_RELADDR) 625 #define SIZE_OP_STATE_CHECK_ANYCHAR_STAR (SIZE_OPCODE + SIZE_STATE_CHECK_NUM) 626 #endif 627 628 #define MC_ESC(syn) (syn)->meta_char_table.esc 629 #define MC_ANYCHAR(syn) (syn)->meta_char_table.anychar 630 #define MC_ANYTIME(syn) (syn)->meta_char_table.anytime 631 #define MC_ZERO_OR_ONE_TIME(syn) (syn)->meta_char_table.zero_or_one_time 632 #define MC_ONE_OR_MORE_TIME(syn) (syn)->meta_char_table.one_or_more_time 633 #define MC_ANYCHAR_ANYTIME(syn) (syn)->meta_char_table.anychar_anytime 634 635 #define IS_MC_ESC_CODE(code, syn) \ 636 ((code) == MC_ESC(syn) && \ 637 !IS_SYNTAX_OP2((syn), ONIG_SYN_OP2_INEFFECTIVE_ESCAPE)) 638 639 640 #define SYN_POSIX_COMMON_OP \ 641 ( ONIG_SYN_OP_DOT_ANYCHAR | ONIG_SYN_OP_POSIX_BRACKET | \ 642 ONIG_SYN_OP_DECIMAL_BACKREF | \ 643 ONIG_SYN_OP_BRACKET_CC | ONIG_SYN_OP_ASTERISK_ZERO_INF | \ 644 ONIG_SYN_OP_LINE_ANCHOR | \ 645 ONIG_SYN_OP_ESC_CONTROL_CHARS ) 646 647 #define SYN_GNU_REGEX_OP \ 648 ( ONIG_SYN_OP_DOT_ANYCHAR | ONIG_SYN_OP_BRACKET_CC | \ 649 ONIG_SYN_OP_POSIX_BRACKET | ONIG_SYN_OP_DECIMAL_BACKREF | \ 650 ONIG_SYN_OP_BRACE_INTERVAL | ONIG_SYN_OP_LPAREN_SUBEXP | \ 651 ONIG_SYN_OP_VBAR_ALT | \ 652 ONIG_SYN_OP_ASTERISK_ZERO_INF | ONIG_SYN_OP_PLUS_ONE_INF | \ 653 ONIG_SYN_OP_QMARK_ZERO_ONE | \ 654 ONIG_SYN_OP_ESC_AZ_BUF_ANCHOR | ONIG_SYN_OP_ESC_CAPITAL_G_BEGIN_ANCHOR | \ 655 ONIG_SYN_OP_ESC_W_WORD | \ 656 ONIG_SYN_OP_ESC_B_WORD_BOUND | ONIG_SYN_OP_ESC_LTGT_WORD_BEGIN_END | \ 657 ONIG_SYN_OP_ESC_S_WHITE_SPACE | ONIG_SYN_OP_ESC_D_DIGIT | \ 658 ONIG_SYN_OP_LINE_ANCHOR ) 659 660 #define SYN_GNU_REGEX_BV \ 661 ( ONIG_SYN_CONTEXT_INDEP_ANCHORS | ONIG_SYN_CONTEXT_INDEP_REPEAT_OPS | \ 662 ONIG_SYN_CONTEXT_INVALID_REPEAT_OPS | ONIG_SYN_ALLOW_INVALID_INTERVAL | \ 663 ONIG_SYN_BACKSLASH_ESCAPE_IN_CC | ONIG_SYN_ALLOW_DOUBLE_RANGE_OP_IN_CC ) 664 665 666 #define NCCLASS_FLAGS(cc) ((cc)->flags) 667 #define NCCLASS_FLAG_SET(cc,flag) (NCCLASS_FLAGS(cc) |= (flag)) 668 #define NCCLASS_FLAG_CLEAR(cc,flag) (NCCLASS_FLAGS(cc) &= ~(flag)) 669 #define IS_NCCLASS_FLAG_ON(cc,flag) ((NCCLASS_FLAGS(cc) & (flag)) != 0) 670 671 /* cclass node */ 672 #define FLAG_NCCLASS_NOT (1<<0) 673 #define FLAG_NCCLASS_SHARE (1<<1) 674 675 #define NCCLASS_SET_NOT(nd) NCCLASS_FLAG_SET(nd, FLAG_NCCLASS_NOT) 676 #define NCCLASS_SET_SHARE(nd) NCCLASS_FLAG_SET(nd, FLAG_NCCLASS_SHARE) 677 #define NCCLASS_CLEAR_NOT(nd) NCCLASS_FLAG_CLEAR(nd, FLAG_NCCLASS_NOT) 678 #define IS_NCCLASS_NOT(nd) IS_NCCLASS_FLAG_ON(nd, FLAG_NCCLASS_NOT) 679 #define IS_NCCLASS_SHARE(nd) IS_NCCLASS_FLAG_ON(nd, FLAG_NCCLASS_SHARE) 680 681 typedef struct { 682 int type; 683 /* struct _Node* next; */ 684 /* unsigned int flags; */ 685 } NodeBase; 686 687 typedef struct { 688 NodeBase base; 689 unsigned int flags; 690 BitSet bs; 691 BBuf* mbuf; /* multi-byte info or NULL */ 692 } CClassNode; 693 694 typedef long OnigStackIndex; 695 696 typedef struct _OnigStackType { 697 unsigned int type; 698 union { 699 struct { 700 UChar *pcode; /* byte code position */ 701 UChar *pstr; /* string position */ 702 UChar *pstr_prev; /* previous char position of pstr */ 703 #ifdef USE_COMBINATION_EXPLOSION_CHECK 704 unsigned int state_check; 705 #endif 706 } state; 707 struct { 708 int count; /* for OP_REPEAT_INC, OP_REPEAT_INC_NG */ 709 UChar *pcode; /* byte code position (head of repeated target) */ 710 int num; /* repeat id */ 711 } repeat; 712 struct { 713 OnigStackIndex si; /* index of stack */ 714 } repeat_inc; 715 struct { 716 int num; /* memory num */ 717 UChar *pstr; /* start/end position */ 718 /* Following information is setted, if this stack type is MEM-START */ 719 OnigStackIndex start; /* prev. info (for backtrack "(...)*" ) */ 720 OnigStackIndex end; /* prev. info (for backtrack "(...)*" ) */ 721 } mem; 722 struct { 723 int num; /* null check id */ 724 UChar *pstr; /* start position */ 725 } null_check; 726 #ifdef USE_SUBEXP_CALL 727 struct { 728 UChar *ret_addr; /* byte code position */ 729 int num; /* null check id */ 730 UChar *pstr; /* string position */ 731 } call_frame; 732 #endif 733 } u; 734 } OnigStackType; 735 736 typedef struct { 737 void* stack_p; 738 int stack_n; 739 OnigOptionType options; 740 OnigRegion* region; 741 const UChar* start; /* search start position (for \G: BEGIN_POSITION) */ 742 #ifdef USE_FIND_LONGEST_SEARCH_ALL_OF_RANGE 743 int best_len; /* for ONIG_OPTION_FIND_LONGEST */ 744 UChar* best_s; 745 #endif 746 #ifdef USE_COMBINATION_EXPLOSION_CHECK 747 void* state_check_buff; 748 int state_check_buff_size; 749 #endif 750 } OnigMatchArg; 751 752 753 #define IS_CODE_SB_WORD(enc,code) \ 754 (ONIGENC_IS_CODE_ASCII(code) && ONIGENC_IS_CODE_WORD(enc,code)) 755 756 typedef struct OnigEndCallListItem { 757 struct OnigEndCallListItem* next; 758 void (*func)(void); 759 } OnigEndCallListItemType; 760 761 extern void onig_add_end_call(void (*func)(void)); 762 763 764 #ifdef ONIG_DEBUG 765 766 typedef struct { 767 short int opcode; 768 char* name; 769 short int arg_type; 770 } OnigOpInfoType; 771 772 extern OnigOpInfoType OnigOpInfo[]; 773 774 775 extern void onig_print_compiled_byte_code P_((FILE* f, UChar* bp, UChar** nextp, OnigEncoding enc)); 776 777 #ifdef ONIG_DEBUG_STATISTICS 778 extern void onig_statistics_init P_((void)); 779 extern void onig_print_statistics P_((FILE* f)); 780 #endif 781 #endif 782 783 extern UChar* onig_error_code_to_format P_((int code)); 784 extern void onig_snprintf_with_pattern PV_((UChar buf[], int bufsize, OnigEncoding enc, UChar* pat, UChar* pat_end, const UChar *fmt, ...)); 785 extern int onig_bbuf_init P_((BBuf* buf, int size)); 786 extern int onig_compile P_((regex_t* reg, const UChar* pattern, const UChar* pattern_end, OnigErrorInfo* einfo)); 787 extern void onig_chain_reduce P_((regex_t* reg)); 788 extern void onig_chain_link_add P_((regex_t* to, regex_t* add)); 789 extern void onig_transfer P_((regex_t* to, regex_t* from)); 790 extern int onig_is_code_in_cc P_((OnigEncoding enc, OnigCodePoint code, CClassNode* cc)); 791 extern int onig_is_code_in_cc_len P_((int enclen, OnigCodePoint code, CClassNode* cc)); 792 793 /* strend hash */ 794 typedef void hash_table_type; 795 typedef unsigned long hash_data_type; 796 797 extern hash_table_type* onig_st_init_strend_table_with_size P_((int size)); 798 extern int onig_st_lookup_strend P_((hash_table_type* table, const UChar* str_key, const UChar* end_key, hash_data_type *value)); 799 extern int onig_st_insert_strend P_((hash_table_type* table, const UChar* str_key, const UChar* end_key, hash_data_type value)); 800 801 /* encoding property management */ 802 #define PROPERTY_LIST_ADD_PROP(Name, CR) \ 803 r = onigenc_property_list_add_property((UChar* )Name, CR,\ 804 &PropertyNameTable, &PropertyList, &PropertyListNum,\ 805 &PropertyListSize);\ 806 if (r != 0) goto end 807 808 #define PROPERTY_LIST_INIT_CHECK \ 809 if (PropertyInited == 0) {\ 810 int r = onigenc_property_list_init(init_property_list);\ 811 if (r != 0) return r;\ 812 } 813 814 extern int onigenc_property_list_add_property P_((UChar* name, const OnigCodePoint* prop, hash_table_type **table, const OnigCodePoint*** plist, int *pnum, int *psize)); 815 816 typedef int (*ONIGENC_INIT_PROPERTY_LIST_FUNC_TYPE)(void); 817 818 extern int onigenc_property_list_init P_((ONIGENC_INIT_PROPERTY_LIST_FUNC_TYPE)); 819 820 #endif /* REGINT_H */ 821