3 /**********************************************************************
4 regint.h - Oniguruma (regular expression library)
5 **********************************************************************/
7 * Copyright (c) 2002-2017 K.Kosako <sndgk393 AT ybb DOT ne DOT jp>
10 * Redistribution and use in source and binary forms, with or without
11 * modification, are permitted provided that the following conditions
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.
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
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 */
39 /* for byte-code statistical data. */
40 /* #define ONIG_DEBUG_STATISTICS */
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)
50 #if defined(__i386) || defined(__i386__) || defined(_M_IX86) || \
51 (defined(__ppc__) && defined(__APPLE__)) || \
52 defined(__x86_64) || defined(__x86_64__) || \
54 #define PLATFORM_UNALIGNED_WORD_ACCESS
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 /* !!! moved to regenc.h. */ /* #define USE_CRNL_AS_LINE_TERMINATOR */
68 #define USE_OP_PUSH_OR_JUMP_EXACT
69 #define USE_QTFR_PEEK_NEXT
70 #define USE_ST_LIBRARY
72 #define INIT_MATCH_STACK_SIZE 160
73 #define DEFAULT_MATCH_STACK_LIMIT_SIZE 0 /* unlimited */
74 #define DEFAULT_PARSE_DEPTH_LIMIT 4096
77 # define ARG_UNUSED __attribute__ ((unused))
83 /* escape other system UChar definition */
85 #ifdef ONIG_ESCAPE_UCHAR_COLLISION
86 #undef ONIG_ESCAPE_UCHAR_COLLISION
89 #define USE_WORD_BEGIN_END /* "\<", "\>" */
90 #define USE_CAPTURE_HISTORY
91 #define USE_VARIABLE_META_CHARS
92 #define USE_POSIX_API_REGION_OPTION
93 #define USE_FIND_LONGEST_SEARCH_ALL_OF_RANGE
94 /* #define USE_COMBINATION_EXPLOSION_CHECK */ /* (X*)* */
96 #define xmalloc malloc
97 #define xrealloc realloc
98 #define xcalloc calloc
101 #define CHECK_INTERRUPT_IN_MATCH_AT
103 #define st_init_table onig_st_init_table
104 #define st_init_table_with_size onig_st_init_table_with_size
105 #define st_init_numtable onig_st_init_numtable
106 #define st_init_numtable_with_size onig_st_init_numtable_with_size
107 #define st_init_strtable onig_st_init_strtable
108 #define st_init_strtable_with_size onig_st_init_strtable_with_size
109 #define st_delete onig_st_delete
110 #define st_delete_safe onig_st_delete_safe
111 #define st_insert onig_st_insert
112 #define st_lookup onig_st_lookup
113 #define st_foreach onig_st_foreach
114 #define st_add_direct onig_st_add_direct
115 #define st_free_table onig_st_free_table
116 #define st_cleanup_safe onig_st_cleanup_safe
117 #define st_copy onig_st_copy
118 #define st_nothing_key_clone onig_st_nothing_key_clone
119 #define st_nothing_key_free onig_st_nothing_key_free
121 #define onig_st_is_member st_is_member
123 #define STATE_CHECK_STRING_THRESHOLD_LEN 7
124 #define STATE_CHECK_BUFF_MAX_SIZE 0x4000
126 #define xmemset memset
127 #define xmemcpy memcpy
128 #define xmemmove memmove
130 #if defined(_WIN32) && !defined(__GNUC__)
131 #define xalloca _alloca
132 #define xvsnprintf(buf,size,fmt,args) _vsnprintf_s(buf,size,_TRUNCATE,fmt,args)
133 #define xsnprintf sprintf_s
134 #define xstrcat(dest,src,size) strcat_s(dest,size,src)
136 #define xalloca alloca
137 #define xvsnprintf vsnprintf
138 #define xsnprintf snprintf
139 #define xstrcat(dest,src,size) strcat(dest,src)
147 #if defined(HAVE_ALLOCA_H) && !defined(__GNUC__)
154 # include <strings.h>
158 #ifdef HAVE_SYS_TYPES_H
160 #include <sys/types.h>
180 #define MIN(a,b) (((a)>(b))?(b):(a))
181 #define MAX(a,b) (((a)<(b))?(b):(a))
183 #define IS_NULL(p) (((void*)(p)) == (void*)0)
184 #define IS_NOT_NULL(p) (((void*)(p)) != (void*)0)
185 #define CHECK_NULL_RETURN(p) if (IS_NULL(p)) return NULL
186 #define CHECK_NULL_RETURN_MEMERR(p) if (IS_NULL(p)) return ONIGERR_MEMORY
187 #define NULL_UCHARP ((UChar* )0)
189 #ifdef PLATFORM_UNALIGNED_WORD_ACCESS
191 #define PLATFORM_GET_INC(val,p,type) do{\
193 (p) += sizeof(type);\
198 #define PLATFORM_GET_INC(val,p,type) do{\
199 xmemcpy(&val, (p), sizeof(type));\
200 (p) += sizeof(type);\
203 /* sizeof(OnigCodePoint) */
204 #define WORD_ALIGNMENT_SIZE SIZEOF_LONG
206 #define GET_ALIGNMENT_PAD_SIZE(addr,pad_size) do {\
207 (pad_size) = WORD_ALIGNMENT_SIZE \
208 - ((unsigned int )(addr) % WORD_ALIGNMENT_SIZE);\
209 if ((pad_size) == WORD_ALIGNMENT_SIZE) (pad_size) = 0;\
212 #define ALIGNMENT_RIGHT(addr) do {\
213 (addr) += (WORD_ALIGNMENT_SIZE - 1);\
214 (addr) -= ((unsigned int )(addr) % WORD_ALIGNMENT_SIZE);\
217 #endif /* PLATFORM_UNALIGNED_WORD_ACCESS */
219 /* stack pop level */
220 #define STACK_POP_LEVEL_FREE 0
221 #define STACK_POP_LEVEL_MEM_START 1
222 #define STACK_POP_LEVEL_ALL 2
225 #define ONIG_OPTIMIZE_NONE 0
226 #define ONIG_OPTIMIZE_EXACT 1 /* Slow Search */
227 #define ONIG_OPTIMIZE_EXACT_BM 2 /* Boyer Moore Search */
228 #define ONIG_OPTIMIZE_EXACT_BM_NOT_REV 3 /* BM (but not simple match) */
229 #define ONIG_OPTIMIZE_EXACT_IC 4 /* Slow Search (ignore case) */
230 #define ONIG_OPTIMIZE_MAP 5 /* char map */
233 typedef unsigned int BitStatusType;
235 #define BIT_STATUS_BITS_NUM (sizeof(BitStatusType) * 8)
236 #define BIT_STATUS_CLEAR(stats) (stats) = 0
237 #define BIT_STATUS_ON_ALL(stats) (stats) = ~((BitStatusType )0)
238 #define BIT_STATUS_AT(stats,n) \
239 ((n) < (int )BIT_STATUS_BITS_NUM ? ((stats) & (1 << n)) : ((stats) & 1))
241 #define BIT_STATUS_ON_AT(stats,n) do {\
242 if ((n) < (int )BIT_STATUS_BITS_NUM) \
243 (stats) |= (1 << (n));\
248 #define BIT_STATUS_ON_AT_SIMPLE(stats,n) do {\
249 if ((n) < (int )BIT_STATUS_BITS_NUM)\
250 (stats) |= (1 << (n));\
254 #define INT_MAX_LIMIT ((1UL << (SIZEOF_INT * 8 - 1)) - 1)
256 #define DIGITVAL(code) ((code) - '0')
257 #define ODIGITVAL(code) DIGITVAL(code)
258 #define XDIGITVAL(enc,code) \
259 (ONIGENC_IS_CODE_DIGIT(enc,code) ? DIGITVAL(code) \
260 : (ONIGENC_IS_CODE_UPPER(enc,code) ? (code) - 'A' + 10 : (code) - 'a' + 10))
262 #define IS_SINGLELINE(option) ((option) & ONIG_OPTION_SINGLELINE)
263 #define IS_MULTILINE(option) ((option) & ONIG_OPTION_MULTILINE)
264 #define IS_IGNORECASE(option) ((option) & ONIG_OPTION_IGNORECASE)
265 #define IS_EXTEND(option) ((option) & ONIG_OPTION_EXTEND)
266 #define IS_FIND_LONGEST(option) ((option) & ONIG_OPTION_FIND_LONGEST)
267 #define IS_FIND_NOT_EMPTY(option) ((option) & ONIG_OPTION_FIND_NOT_EMPTY)
268 #define IS_FIND_CONDITION(option) ((option) & \
269 (ONIG_OPTION_FIND_LONGEST | ONIG_OPTION_FIND_NOT_EMPTY))
270 #define IS_NOTBOL(option) ((option) & ONIG_OPTION_NOTBOL)
271 #define IS_NOTEOL(option) ((option) & ONIG_OPTION_NOTEOL)
272 #define IS_POSIX_REGION(option) ((option) & ONIG_OPTION_POSIX_REGION)
274 /* OP_SET_OPTION is required for these options.
275 #define IS_DYNAMIC_OPTION(option) \
276 (((option) & (ONIG_OPTION_MULTILINE | ONIG_OPTION_IGNORECASE)) != 0)
278 /* ignore-case and multibyte status are included in compiled code. */
279 #define IS_DYNAMIC_OPTION(option) 0
281 #define DISABLE_CASE_FOLD_MULTI_CHAR(case_fold_flag) \
282 ((case_fold_flag) & ~INTERNAL_ONIGENC_CASE_FOLD_MULTI_CHAR)
284 #define REPEAT_INFINITE -1
285 #define IS_REPEAT_INFINITE(n) ((n) == REPEAT_INFINITE)
288 #define BITS_PER_BYTE 8
289 #define SINGLE_BYTE_SIZE (1 << BITS_PER_BYTE)
290 #define BITS_IN_ROOM (sizeof(Bits) * BITS_PER_BYTE)
291 #define BITSET_SIZE (SINGLE_BYTE_SIZE / BITS_IN_ROOM)
293 #ifdef PLATFORM_UNALIGNED_WORD_ACCESS
294 typedef unsigned int Bits;
296 typedef unsigned char Bits;
298 typedef Bits BitSet[BITSET_SIZE];
299 typedef Bits* BitSetRef;
301 #define SIZE_BITSET sizeof(BitSet)
303 #define BITSET_CLEAR(bs) do {\
305 for (i = 0; i < (int )BITSET_SIZE; i++) { (bs)[i] = 0; } \
308 #define BS_ROOM(bs,pos) (bs)[pos / BITS_IN_ROOM]
309 #define BS_BIT(pos) (1 << (pos % BITS_IN_ROOM))
311 #define BITSET_AT(bs, pos) (BS_ROOM(bs,pos) & BS_BIT(pos))
312 #define BITSET_SET_BIT(bs, pos) BS_ROOM(bs,pos) |= BS_BIT(pos)
313 #define BITSET_CLEAR_BIT(bs, pos) BS_ROOM(bs,pos) &= ~(BS_BIT(pos))
314 #define BITSET_INVERT_BIT(bs, pos) BS_ROOM(bs,pos) ^= BS_BIT(pos)
317 typedef struct _BBuf {
323 #define BBUF_INIT(buf,size) onig_bbuf_init((BBuf* )(buf), (size))
325 #define BBUF_SIZE_INC(buf,inc) do{\
326 (buf)->alloc += (inc);\
327 (buf)->p = (UChar* )xrealloc((buf)->p, (buf)->alloc);\
328 if (IS_NULL((buf)->p)) return(ONIGERR_MEMORY);\
331 #define BBUF_EXPAND(buf,low) do{\
332 do { (buf)->alloc *= 2; } while ((buf)->alloc < (unsigned int )low);\
333 (buf)->p = (UChar* )xrealloc((buf)->p, (buf)->alloc);\
334 if (IS_NULL((buf)->p)) return(ONIGERR_MEMORY);\
337 #define BBUF_ENSURE_SIZE(buf,size) do{\
338 unsigned int new_alloc = (buf)->alloc;\
339 while (new_alloc < (unsigned int )(size)) { new_alloc *= 2; }\
340 if ((buf)->alloc != new_alloc) {\
341 (buf)->p = (UChar* )xrealloc((buf)->p, new_alloc);\
342 if (IS_NULL((buf)->p)) return(ONIGERR_MEMORY);\
343 (buf)->alloc = new_alloc;\
347 #define BBUF_WRITE(buf,pos,bytes,n) do{\
348 int used = (pos) + (n);\
349 if ((buf)->alloc < (unsigned int )used) BBUF_EXPAND((buf),used);\
350 xmemcpy((buf)->p + (pos), (bytes), (n));\
351 if ((buf)->used < (unsigned int )used) (buf)->used = used;\
354 #define BBUF_WRITE1(buf,pos,byte) do{\
355 int used = (pos) + 1;\
356 if ((buf)->alloc < (unsigned int )used) BBUF_EXPAND((buf),used);\
357 (buf)->p[(pos)] = (byte);\
358 if ((buf)->used < (unsigned int )used) (buf)->used = used;\
361 #define BBUF_ADD(buf,bytes,n) BBUF_WRITE((buf),(buf)->used,(bytes),(n))
362 #define BBUF_ADD1(buf,byte) BBUF_WRITE1((buf),(buf)->used,(byte))
363 #define BBUF_GET_ADD_ADDRESS(buf) ((buf)->p + (buf)->used)
364 #define BBUF_GET_OFFSET_POS(buf) ((buf)->used)
367 #define BBUF_MOVE_RIGHT(buf,from,to,n) do {\
368 if ((unsigned int )((to)+(n)) > (buf)->alloc) BBUF_EXPAND((buf),(to) + (n));\
369 xmemmove((buf)->p + (to), (buf)->p + (from), (n));\
370 if ((unsigned int )((to)+(n)) > (buf)->used) (buf)->used = (to) + (n);\
374 #define BBUF_MOVE_LEFT(buf,from,to,n) do {\
375 xmemmove((buf)->p + (to), (buf)->p + (from), (n));\
379 #define BBUF_MOVE_LEFT_REDUCE(buf,from,to) do {\
380 xmemmove((buf)->p + (to), (buf)->p + (from), (buf)->used - (from));\
381 (buf)->used -= (from - to);\
384 #define BBUF_INSERT(buf,pos,bytes,n) do {\
385 if (pos >= (buf)->used) {\
386 BBUF_WRITE(buf,pos,bytes,n);\
389 BBUF_MOVE_RIGHT((buf),(pos),(pos) + (n),((buf)->used - (pos)));\
390 xmemcpy((buf)->p + (pos), (bytes), (n));\
394 #define BBUF_GET_BYTE(buf, pos) (buf)->p[(pos)]
398 #define ANCHOR_PREC_READ (1<<0)
399 #define ANCHOR_PREC_READ_NOT (1<<1)
400 #define ANCHOR_LOOK_BEHIND (1<<2)
401 #define ANCHOR_LOOK_BEHIND_NOT (1<<3)
403 #define ANCHOR_BEGIN_BUF (1<<4)
404 #define ANCHOR_BEGIN_LINE (1<<5)
405 #define ANCHOR_BEGIN_POSITION (1<<6)
406 #define ANCHOR_END_BUF (1<<7)
407 #define ANCHOR_SEMI_END_BUF (1<<8)
408 #define ANCHOR_END_LINE (1<<9)
409 #define ANCHOR_WORD_BOUND (1<<10)
410 #define ANCHOR_NOT_WORD_BOUND (1<<11)
411 #define ANCHOR_WORD_BEGIN (1<<12)
412 #define ANCHOR_WORD_END (1<<13)
413 #define ANCHOR_ANYCHAR_STAR (1<<14) /* ".*" optimize info */
414 #define ANCHOR_ANYCHAR_STAR_ML (1<<15) /* ".*" optimize info (multi-line) */
416 #define ANCHOR_HAS_BODY(a) ((a)->type < ANCHOR_BEGIN_BUF)
421 OP_FINISH = 0, /* matching process terminator (no more alternative) */
422 OP_END = 1, /* pattern code terminator (success end) */
424 OP_EXACT1 = 2, /* single byte, N = 1 */
425 OP_EXACT2, /* single byte, N = 2 */
426 OP_EXACT3, /* single byte, N = 3 */
427 OP_EXACT4, /* single byte, N = 4 */
428 OP_EXACT5, /* single byte, N = 5 */
429 OP_EXACTN, /* single byte */
430 OP_EXACTMB2N1, /* mb-length = 2 N = 1 */
431 OP_EXACTMB2N2, /* mb-length = 2 N = 2 */
432 OP_EXACTMB2N3, /* mb-length = 2 N = 3 */
433 OP_EXACTMB2N, /* mb-length = 2 */
434 OP_EXACTMB3N, /* mb-length = 3 */
435 OP_EXACTMBN, /* other length */
437 OP_EXACT1_IC, /* single byte, N = 1, ignore case */
438 OP_EXACTN_IC, /* single byte, ignore case */
446 OP_CCLASS_NODE, /* pointer to CClassNode node */
448 OP_ANYCHAR, /* "." */
449 OP_ANYCHAR_ML, /* "." multi-line */
450 OP_ANYCHAR_STAR, /* ".*" */
451 OP_ANYCHAR_ML_STAR, /* ".*" multi-line */
452 OP_ANYCHAR_STAR_PEEK_NEXT,
453 OP_ANYCHAR_ML_STAR_PEEK_NEXT,
475 OP_BACKREF_WITH_LEVEL, /* \k<xxx+n>, \k<xxx-n> */
478 OP_MEMORY_START_PUSH, /* push back-tracker to stack */
479 OP_MEMORY_END_PUSH, /* push back-tracker to stack */
480 OP_MEMORY_END_PUSH_REC, /* push back-tracker to stack */
482 OP_MEMORY_END_REC, /* push marker to stack */
484 OP_FAIL, /* pop stack and move */
488 OP_PUSH_OR_JUMP_EXACT1, /* if match exact then push, else jump. */
489 OP_PUSH_IF_PEEK_NEXT, /* if match exact then push, else none. */
490 OP_REPEAT, /* {n,m} */
491 OP_REPEAT_NG, /* {n,m}? (non greedy) */
493 OP_REPEAT_INC_NG, /* non greedy */
494 OP_REPEAT_INC_SG, /* search and get in stack */
495 OP_REPEAT_INC_NG_SG, /* search and get in stack (non greedy) */
496 OP_NULL_CHECK_START, /* null loop checker start */
497 OP_NULL_CHECK_END, /* null loop checker end */
498 OP_NULL_CHECK_END_MEMST, /* null loop checker end (with capture status) */
499 OP_NULL_CHECK_END_MEMST_PUSH, /* with capture status and push check-end */
501 OP_PUSH_POS, /* (?=...) start */
502 OP_POP_POS, /* (?=...) end */
503 OP_PUSH_POS_NOT, /* (?!...) start */
504 OP_FAIL_POS, /* (?!...) end */
505 OP_PUSH_STOP_BT, /* (?>...) start */
506 OP_POP_STOP_BT, /* (?>...) end */
507 OP_LOOK_BEHIND, /* (?<=...) start (no needs end opcode) */
508 OP_PUSH_LOOK_BEHIND_NOT, /* (?<!...) start */
509 OP_FAIL_LOOK_BEHIND_NOT, /* (?<!...) end */
511 OP_CALL, /* \g<name> */
514 OP_STATE_CHECK_PUSH, /* combination explosion check and push */
515 OP_STATE_CHECK_PUSH_OR_JUMP, /* check ok -> push, else jump */
516 OP_STATE_CHECK, /* check only */
517 OP_STATE_CHECK_ANYCHAR_STAR,
518 OP_STATE_CHECK_ANYCHAR_ML_STAR,
520 /* no need: IS_DYNAMIC_OPTION() == 0 */
521 OP_SET_OPTION_PUSH, /* set option and push recover option */
522 OP_SET_OPTION /* set option */
525 typedef int RelAddrType;
526 typedef int AbsAddrType;
527 typedef int LengthType;
528 typedef int RepeatNumType;
529 typedef int MemNumType;
530 typedef short int StateCheckNumType;
531 typedef void* PointerType;
533 #define SIZE_OPCODE 1
534 #define SIZE_RELADDR sizeof(RelAddrType)
535 #define SIZE_ABSADDR sizeof(AbsAddrType)
536 #define SIZE_LENGTH sizeof(LengthType)
537 #define SIZE_MEMNUM sizeof(MemNumType)
538 #define SIZE_STATE_CHECK_NUM sizeof(StateCheckNumType)
539 #define SIZE_REPEATNUM sizeof(RepeatNumType)
540 #define SIZE_OPTION sizeof(OnigOptionType)
541 #define SIZE_CODE_POINT sizeof(OnigCodePoint)
542 #define SIZE_POINTER sizeof(PointerType)
545 #define GET_RELADDR_INC(addr,p) PLATFORM_GET_INC(addr, p, RelAddrType)
546 #define GET_ABSADDR_INC(addr,p) PLATFORM_GET_INC(addr, p, AbsAddrType)
547 #define GET_LENGTH_INC(len,p) PLATFORM_GET_INC(len, p, LengthType)
548 #define GET_MEMNUM_INC(num,p) PLATFORM_GET_INC(num, p, MemNumType)
549 #define GET_REPEATNUM_INC(num,p) PLATFORM_GET_INC(num, p, RepeatNumType)
550 #define GET_OPTION_INC(option,p) PLATFORM_GET_INC(option, p, OnigOptionType)
551 #define GET_POINTER_INC(ptr,p) PLATFORM_GET_INC(ptr, p, PointerType)
552 #define GET_STATE_CHECK_NUM_INC(num,p) PLATFORM_GET_INC(num, p, StateCheckNumType)
554 /* code point's address must be aligned address. */
555 #define GET_CODE_POINT(code,p) code = *((OnigCodePoint* )(p))
556 #define GET_BYTE_INC(byte,p) do{\
562 /* op-code + arg size */
563 #define SIZE_OP_ANYCHAR_STAR SIZE_OPCODE
564 #define SIZE_OP_ANYCHAR_STAR_PEEK_NEXT (SIZE_OPCODE + 1)
565 #define SIZE_OP_JUMP (SIZE_OPCODE + SIZE_RELADDR)
566 #define SIZE_OP_PUSH (SIZE_OPCODE + SIZE_RELADDR)
567 #define SIZE_OP_POP SIZE_OPCODE
568 #define SIZE_OP_PUSH_OR_JUMP_EXACT1 (SIZE_OPCODE + SIZE_RELADDR + 1)
569 #define SIZE_OP_PUSH_IF_PEEK_NEXT (SIZE_OPCODE + SIZE_RELADDR + 1)
570 #define SIZE_OP_REPEAT_INC (SIZE_OPCODE + SIZE_MEMNUM)
571 #define SIZE_OP_REPEAT_INC_NG (SIZE_OPCODE + SIZE_MEMNUM)
572 #define SIZE_OP_PUSH_POS SIZE_OPCODE
573 #define SIZE_OP_PUSH_POS_NOT (SIZE_OPCODE + SIZE_RELADDR)
574 #define SIZE_OP_POP_POS SIZE_OPCODE
575 #define SIZE_OP_FAIL_POS SIZE_OPCODE
576 #define SIZE_OP_SET_OPTION (SIZE_OPCODE + SIZE_OPTION)
577 #define SIZE_OP_SET_OPTION_PUSH (SIZE_OPCODE + SIZE_OPTION)
578 #define SIZE_OP_FAIL SIZE_OPCODE
579 #define SIZE_OP_MEMORY_START (SIZE_OPCODE + SIZE_MEMNUM)
580 #define SIZE_OP_MEMORY_START_PUSH (SIZE_OPCODE + SIZE_MEMNUM)
581 #define SIZE_OP_MEMORY_END_PUSH (SIZE_OPCODE + SIZE_MEMNUM)
582 #define SIZE_OP_MEMORY_END_PUSH_REC (SIZE_OPCODE + SIZE_MEMNUM)
583 #define SIZE_OP_MEMORY_END (SIZE_OPCODE + SIZE_MEMNUM)
584 #define SIZE_OP_MEMORY_END_REC (SIZE_OPCODE + SIZE_MEMNUM)
585 #define SIZE_OP_PUSH_STOP_BT SIZE_OPCODE
586 #define SIZE_OP_POP_STOP_BT SIZE_OPCODE
587 #define SIZE_OP_NULL_CHECK_START (SIZE_OPCODE + SIZE_MEMNUM)
588 #define SIZE_OP_NULL_CHECK_END (SIZE_OPCODE + SIZE_MEMNUM)
589 #define SIZE_OP_LOOK_BEHIND (SIZE_OPCODE + SIZE_LENGTH)
590 #define SIZE_OP_PUSH_LOOK_BEHIND_NOT (SIZE_OPCODE + SIZE_RELADDR + SIZE_LENGTH)
591 #define SIZE_OP_FAIL_LOOK_BEHIND_NOT SIZE_OPCODE
592 #define SIZE_OP_CALL (SIZE_OPCODE + SIZE_ABSADDR)
593 #define SIZE_OP_RETURN SIZE_OPCODE
595 #ifdef USE_COMBINATION_EXPLOSION_CHECK
596 #define SIZE_OP_STATE_CHECK (SIZE_OPCODE + SIZE_STATE_CHECK_NUM)
597 #define SIZE_OP_STATE_CHECK_PUSH (SIZE_OPCODE + SIZE_STATE_CHECK_NUM + SIZE_RELADDR)
598 #define SIZE_OP_STATE_CHECK_PUSH_OR_JUMP (SIZE_OPCODE + SIZE_STATE_CHECK_NUM + SIZE_RELADDR)
599 #define SIZE_OP_STATE_CHECK_ANYCHAR_STAR (SIZE_OPCODE + SIZE_STATE_CHECK_NUM)
602 #define MC_ESC(syn) (syn)->meta_char_table.esc
603 #define MC_ANYCHAR(syn) (syn)->meta_char_table.anychar
604 #define MC_ANYTIME(syn) (syn)->meta_char_table.anytime
605 #define MC_ZERO_OR_ONE_TIME(syn) (syn)->meta_char_table.zero_or_one_time
606 #define MC_ONE_OR_MORE_TIME(syn) (syn)->meta_char_table.one_or_more_time
607 #define MC_ANYCHAR_ANYTIME(syn) (syn)->meta_char_table.anychar_anytime
609 #define IS_MC_ESC_CODE(code, syn) \
610 ((code) == MC_ESC(syn) && \
611 !IS_SYNTAX_OP2((syn), ONIG_SYN_OP2_INEFFECTIVE_ESCAPE))
614 #define SYN_POSIX_COMMON_OP \
615 ( ONIG_SYN_OP_DOT_ANYCHAR | ONIG_SYN_OP_POSIX_BRACKET | \
616 ONIG_SYN_OP_DECIMAL_BACKREF | \
617 ONIG_SYN_OP_BRACKET_CC | ONIG_SYN_OP_ASTERISK_ZERO_INF | \
618 ONIG_SYN_OP_LINE_ANCHOR | \
619 ONIG_SYN_OP_ESC_CONTROL_CHARS )
621 #define SYN_GNU_REGEX_OP \
622 ( ONIG_SYN_OP_DOT_ANYCHAR | ONIG_SYN_OP_BRACKET_CC | \
623 ONIG_SYN_OP_POSIX_BRACKET | ONIG_SYN_OP_DECIMAL_BACKREF | \
624 ONIG_SYN_OP_BRACE_INTERVAL | ONIG_SYN_OP_LPAREN_SUBEXP | \
625 ONIG_SYN_OP_VBAR_ALT | \
626 ONIG_SYN_OP_ASTERISK_ZERO_INF | ONIG_SYN_OP_PLUS_ONE_INF | \
627 ONIG_SYN_OP_QMARK_ZERO_ONE | \
628 ONIG_SYN_OP_ESC_AZ_BUF_ANCHOR | ONIG_SYN_OP_ESC_CAPITAL_G_BEGIN_ANCHOR | \
629 ONIG_SYN_OP_ESC_W_WORD | \
630 ONIG_SYN_OP_ESC_B_WORD_BOUND | ONIG_SYN_OP_ESC_LTGT_WORD_BEGIN_END | \
631 ONIG_SYN_OP_ESC_S_WHITE_SPACE | ONIG_SYN_OP_ESC_D_DIGIT | \
632 ONIG_SYN_OP_LINE_ANCHOR )
634 #define SYN_GNU_REGEX_BV \
635 ( ONIG_SYN_CONTEXT_INDEP_ANCHORS | ONIG_SYN_CONTEXT_INDEP_REPEAT_OPS | \
636 ONIG_SYN_CONTEXT_INVALID_REPEAT_OPS | ONIG_SYN_ALLOW_INVALID_INTERVAL | \
637 ONIG_SYN_BACKSLASH_ESCAPE_IN_CC | ONIG_SYN_ALLOW_DOUBLE_RANGE_OP_IN_CC )
640 #define NCCLASS_FLAGS(cc) ((cc)->flags)
641 #define NCCLASS_FLAG_SET(cc,flag) (NCCLASS_FLAGS(cc) |= (flag))
642 #define NCCLASS_FLAG_CLEAR(cc,flag) (NCCLASS_FLAGS(cc) &= ~(flag))
643 #define IS_NCCLASS_FLAG_ON(cc,flag) ((NCCLASS_FLAGS(cc) & (flag)) != 0)
646 #define FLAG_NCCLASS_NOT (1<<0)
647 #define FLAG_NCCLASS_SHARE (1<<1)
649 #define NCCLASS_SET_NOT(nd) NCCLASS_FLAG_SET(nd, FLAG_NCCLASS_NOT)
650 #define NCCLASS_SET_SHARE(nd) NCCLASS_FLAG_SET(nd, FLAG_NCCLASS_SHARE)
651 #define NCCLASS_CLEAR_NOT(nd) NCCLASS_FLAG_CLEAR(nd, FLAG_NCCLASS_NOT)
652 #define IS_NCCLASS_NOT(nd) IS_NCCLASS_FLAG_ON(nd, FLAG_NCCLASS_NOT)
653 #define IS_NCCLASS_SHARE(nd) IS_NCCLASS_FLAG_ON(nd, FLAG_NCCLASS_SHARE)
669 BBuf* mbuf; /* multi-byte info or NULL */
672 typedef long OnigStackIndex;
674 typedef struct _OnigStackType {
678 UChar *pcode; /* byte code position */
679 UChar *pstr; /* string position */
680 UChar *pstr_prev; /* previous char position of pstr */
681 #ifdef USE_COMBINATION_EXPLOSION_CHECK
682 unsigned int state_check;
686 int count; /* for OP_REPEAT_INC, OP_REPEAT_INC_NG */
687 UChar *pcode; /* byte code position (head of repeated target) */
688 int num; /* repeat id */
691 OnigStackIndex si; /* index of stack */
694 int num; /* memory num */
695 UChar *pstr; /* start/end position */
696 /* Following information is set, if this stack type is MEM-START */
697 OnigStackIndex start; /* prev. info (for backtrack "(...)*" ) */
698 OnigStackIndex end; /* prev. info (for backtrack "(...)*" ) */
701 int num; /* null check id */
702 UChar *pstr; /* start position */
704 #ifdef USE_SUBEXP_CALL
706 UChar *ret_addr; /* byte code position */
707 int num; /* null check id */
708 UChar *pstr; /* string position */
717 OnigOptionType options;
720 const UChar* start; /* search start position (for \G: BEGIN_POSITION) */
721 #ifdef USE_FIND_LONGEST_SEARCH_ALL_OF_RANGE
722 int best_len; /* for ONIG_OPTION_FIND_LONGEST */
725 #ifdef USE_COMBINATION_EXPLOSION_CHECK
726 void* state_check_buff;
727 int state_check_buff_size;
732 #define IS_CODE_SB_WORD(enc,code) \
733 (ONIGENC_IS_CODE_ASCII(code) && ONIGENC_IS_CODE_WORD(enc,code))
735 typedef struct OnigEndCallListItem {
736 struct OnigEndCallListItem* next;
738 } OnigEndCallListItemType;
740 extern void onig_add_end_call(void (*func)(void));
751 extern OnigOpInfoType OnigOpInfo[];
754 extern void onig_print_compiled_byte_code P_((FILE* f, UChar* bp, UChar** nextp, UChar* start, OnigEncoding enc));
756 #ifdef ONIG_DEBUG_STATISTICS
757 extern void onig_statistics_init P_((void));
758 extern int onig_print_statistics P_((FILE* f));
762 extern void onig_warning(const char* s);
763 extern UChar* onig_error_code_to_format P_((int code));
764 extern void onig_snprintf_with_pattern PV_((UChar buf[], int bufsize, OnigEncoding enc, UChar* pat, UChar* pat_end, const UChar *fmt, ...));
765 extern int onig_bbuf_init P_((BBuf* buf, int size));
766 extern int onig_compile P_((regex_t* reg, const UChar* pattern, const UChar* pattern_end, OnigErrorInfo* einfo));
767 extern void onig_transfer P_((regex_t* to, regex_t* from));
768 extern int onig_is_code_in_cc P_((OnigEncoding enc, OnigCodePoint code, CClassNode* cc));
769 extern int onig_is_code_in_cc_len P_((int enclen, OnigCodePoint code, CClassNode* cc));
772 typedef void hash_table_type;
774 # include <windows.h>
775 typedef ULONG_PTR hash_data_type;
777 typedef unsigned long hash_data_type;
780 extern hash_table_type* onig_st_init_strend_table_with_size P_((int size));
781 extern int onig_st_lookup_strend P_((hash_table_type* table, const UChar* str_key, const UChar* end_key, hash_data_type *value));
782 extern int onig_st_insert_strend P_((hash_table_type* table, const UChar* str_key, const UChar* end_key, hash_data_type value));
784 typedef int (*ONIGENC_INIT_PROPERTY_LIST_FUNC_TYPE)(void);
786 #endif /* REGINT_H */