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 */
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) || 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
60 #define USE_BACKREF_WITH_LEVEL /* \k<name+n>, \k<name-n> */
61 #define USE_INSISTENT_CHECK_CAPTURES_STATUS_IN_ENDLESS_REPEAT /* /(?:()|())*\2/ */
62 #define USE_NEWLINE_AT_END_OF_STRING_HAS_EMPTY_LINE /* /\n$/ =~ "\n" */
63 #define USE_WARNING_REDUNDANT_NESTED_REPEAT_OPERATOR
64 /* !!! moved to regenc.h. */ /* #define USE_CRNL_AS_LINE_TERMINATOR */
67 #define USE_OP_PUSH_OR_JUMP_EXACT
68 #define USE_QUANT_PEEK_NEXT
69 #define USE_ST_LIBRARY
71 #define INIT_MATCH_STACK_SIZE 160
72 #define DEFAULT_MATCH_STACK_LIMIT_SIZE 0 /* unlimited */
73 #define DEFAULT_PARSE_DEPTH_LIMIT 4096
76 # define ARG_UNUSED __attribute__ ((unused))
82 /* escape other system UChar definition */
84 #ifdef ONIG_ESCAPE_UCHAR_COLLISION
85 #undef ONIG_ESCAPE_UCHAR_COLLISION
88 #define USE_WORD_BEGIN_END /* "\<", "\>" */
89 #define USE_CAPTURE_HISTORY
90 #define USE_VARIABLE_META_CHARS
91 #define USE_POSIX_API_REGION_OPTION
92 #define USE_FIND_LONGEST_SEARCH_ALL_OF_RANGE
93 /* #define USE_COMBINATION_EXPLOSION_CHECK */ /* (X*)* */
95 #define xmalloc malloc
96 #define xrealloc realloc
97 #define xcalloc calloc
100 #define CHECK_INTERRUPT_IN_MATCH_AT
102 #define st_init_table onig_st_init_table
103 #define st_init_table_with_size onig_st_init_table_with_size
104 #define st_init_numtable onig_st_init_numtable
105 #define st_init_numtable_with_size onig_st_init_numtable_with_size
106 #define st_init_strtable onig_st_init_strtable
107 #define st_init_strtable_with_size onig_st_init_strtable_with_size
108 #define st_delete onig_st_delete
109 #define st_delete_safe onig_st_delete_safe
110 #define st_insert onig_st_insert
111 #define st_lookup onig_st_lookup
112 #define st_foreach onig_st_foreach
113 #define st_add_direct onig_st_add_direct
114 #define st_free_table onig_st_free_table
115 #define st_cleanup_safe onig_st_cleanup_safe
116 #define st_copy onig_st_copy
117 #define st_nothing_key_clone onig_st_nothing_key_clone
118 #define st_nothing_key_free onig_st_nothing_key_free
120 #define onig_st_is_member st_is_member
122 #define STATE_CHECK_STRING_THRESHOLD_LEN 7
123 #define STATE_CHECK_BUFF_MAX_SIZE 0x4000
125 #define xmemset memset
126 #define xmemcpy memcpy
127 #define xmemmove memmove
129 #if defined(_WIN32) && !defined(__GNUC__)
130 #define xalloca _alloca
131 #define xvsnprintf(buf,size,fmt,args) _vsnprintf_s(buf,size,_TRUNCATE,fmt,args)
132 #define xsnprintf sprintf_s
133 #define xstrcat(dest,src,size) strcat_s(dest,size,src)
135 #define xalloca alloca
136 #define xvsnprintf vsnprintf
137 #define xsnprintf snprintf
138 #define xstrcat(dest,src,size) strcat(dest,src)
146 #if defined(HAVE_ALLOCA_H) && !defined(__GNUC__)
153 # include <strings.h>
157 #ifdef HAVE_SYS_TYPES_H
159 #include <sys/types.h>
163 #ifdef HAVE_INTTYPES_H
164 #include <inttypes.h>
176 #if defined(_MSC_VER) && (_MSC_VER < 1300)
177 typedef int intptr_t;
189 #define MIN(a,b) (((a)>(b))?(b):(a))
190 #define MAX(a,b) (((a)<(b))?(b):(a))
192 #define IS_NULL(p) (((void*)(p)) == (void*)0)
193 #define IS_NOT_NULL(p) (((void*)(p)) != (void*)0)
194 #define CHECK_NULL_RETURN(p) if (IS_NULL(p)) return NULL
195 #define CHECK_NULL_RETURN_MEMERR(p) if (IS_NULL(p)) return ONIGERR_MEMORY
196 #define NULL_UCHARP ((UChar* )0)
198 #define INFINITE_LEN ONIG_INFINITE_DISTANCE
200 #ifdef PLATFORM_UNALIGNED_WORD_ACCESS
202 #define PLATFORM_GET_INC(val,p,type) do{\
204 (p) += sizeof(type);\
209 #define PLATFORM_GET_INC(val,p,type) do{\
210 xmemcpy(&val, (p), sizeof(type));\
211 (p) += sizeof(type);\
214 /* sizeof(OnigCodePoint) */
216 # define WORD_ALIGNMENT_SIZE SIZEOF_SIZE_T
218 # define WORD_ALIGNMENT_SIZE SIZEOF_LONG
221 #define GET_ALIGNMENT_PAD_SIZE(addr,pad_size) do {\
222 (pad_size) = WORD_ALIGNMENT_SIZE \
223 - ((unsigned int )(addr) % WORD_ALIGNMENT_SIZE);\
224 if ((pad_size) == WORD_ALIGNMENT_SIZE) (pad_size) = 0;\
227 #define ALIGNMENT_RIGHT(addr) do {\
228 (addr) += (WORD_ALIGNMENT_SIZE - 1);\
229 (addr) -= ((unsigned int )(addr) % WORD_ALIGNMENT_SIZE);\
232 #endif /* PLATFORM_UNALIGNED_WORD_ACCESS */
239 #define REG_EXTP(reg) (RegExt* )((reg)->chain)
240 #define REG_EXTPL(reg) ((reg)->chain)
242 /* stack pop level */
244 STACK_POP_LEVEL_FREE = 0,
245 STACK_POP_LEVEL_MEM_START = 1,
246 STACK_POP_LEVEL_ALL =2
250 #define ONIG_OPTIMIZE_NONE 0
251 #define ONIG_OPTIMIZE_EXACT 1 /* Slow Search */
252 #define ONIG_OPTIMIZE_EXACT_BM 2 /* Boyer Moore Search */
253 #define ONIG_OPTIMIZE_EXACT_BM_NOT_REV 3 /* BM (but not simple match) */
254 #define ONIG_OPTIMIZE_EXACT_IC 4 /* Slow Search (ignore case) */
255 #define ONIG_OPTIMIZE_MAP 5 /* char map */
258 typedef unsigned int MemStatusType;
260 #define MEM_STATUS_BITS_NUM (sizeof(MemStatusType) * 8)
261 #define MEM_STATUS_CLEAR(stats) (stats) = 0
262 #define MEM_STATUS_ON_ALL(stats) (stats) = ~((MemStatusType )0)
263 #define MEM_STATUS_AT(stats,n) \
264 ((n) < (int )MEM_STATUS_BITS_NUM ? ((stats) & ((MemStatusType )1 << n)) : ((stats) & 1))
265 #define MEM_STATUS_AT0(stats,n) \
266 ((n) > 0 && (n) < (int )MEM_STATUS_BITS_NUM ? ((stats) & ((MemStatusType )1 << n)) : ((stats) & 1))
268 #define MEM_STATUS_ON(stats,n) do {\
269 if ((n) < (int )MEM_STATUS_BITS_NUM) {\
271 (stats) |= ((MemStatusType )1 << (n));\
277 #define MEM_STATUS_ON_SIMPLE(stats,n) do {\
278 if ((n) < (int )MEM_STATUS_BITS_NUM)\
279 (stats) |= ((MemStatusType )1 << (n));\
283 #define INT_MAX_LIMIT ((1UL << (SIZEOF_INT * 8 - 1)) - 1)
285 #define ONIGENC_IS_CODE_DIGIT_ASCII(enc, code) \
286 ((code) < 128 && ONIGENC_IS_CODE_CTYPE(enc,code,ONIGENC_CTYPE_DIGIT))
288 #define DIGITVAL(code) ((code) - '0')
289 #define ODIGITVAL(code) DIGITVAL(code)
290 #define XDIGITVAL(enc,code) \
291 (ONIGENC_IS_CODE_DIGIT_ASCII(enc,code) ? DIGITVAL(code) \
292 : (ONIGENC_IS_CODE_UPPER(enc,code) ? (code) - 'A' + 10 : (code) - 'a' + 10))
294 #define IS_SINGLELINE(option) ((option) & ONIG_OPTION_SINGLELINE)
295 #define IS_MULTILINE(option) ((option) & ONIG_OPTION_MULTILINE)
296 #define IS_IGNORECASE(option) ((option) & ONIG_OPTION_IGNORECASE)
297 #define IS_EXTEND(option) ((option) & ONIG_OPTION_EXTEND)
298 #define IS_FIND_LONGEST(option) ((option) & ONIG_OPTION_FIND_LONGEST)
299 #define IS_FIND_NOT_EMPTY(option) ((option) & ONIG_OPTION_FIND_NOT_EMPTY)
300 #define IS_FIND_CONDITION(option) ((option) & \
301 (ONIG_OPTION_FIND_LONGEST | ONIG_OPTION_FIND_NOT_EMPTY))
302 #define IS_NOTBOL(option) ((option) & ONIG_OPTION_NOTBOL)
303 #define IS_NOTEOL(option) ((option) & ONIG_OPTION_NOTEOL)
304 #define IS_POSIX_REGION(option) ((option) & ONIG_OPTION_POSIX_REGION)
306 /* OP_SET_OPTION is required for these options.
307 #define IS_DYNAMIC_OPTION(option) \
308 (((option) & (ONIG_OPTION_MULTILINE | ONIG_OPTION_IGNORECASE)) != 0)
310 /* ignore-case and multibyte status are included in compiled code. */
311 #define IS_DYNAMIC_OPTION(option) 0
313 #define DISABLE_CASE_FOLD_MULTI_CHAR(case_fold_flag) \
314 ((case_fold_flag) & ~INTERNAL_ONIGENC_CASE_FOLD_MULTI_CHAR)
316 #define REPEAT_INFINITE -1
317 #define IS_REPEAT_INFINITE(n) ((n) == REPEAT_INFINITE)
320 #define BITS_PER_BYTE 8
321 #define SINGLE_BYTE_SIZE (1 << BITS_PER_BYTE)
322 #define BITS_IN_ROOM (sizeof(Bits) * BITS_PER_BYTE)
323 #define BITSET_SIZE (SINGLE_BYTE_SIZE / BITS_IN_ROOM)
325 #ifdef PLATFORM_UNALIGNED_WORD_ACCESS
326 typedef unsigned int Bits;
328 typedef unsigned char Bits;
330 typedef Bits BitSet[BITSET_SIZE];
331 typedef Bits* BitSetRef;
333 #define SIZE_BITSET sizeof(BitSet)
335 #define BITSET_CLEAR(bs) do {\
337 for (i = 0; i < (int )BITSET_SIZE; i++) { (bs)[i] = 0; } \
340 #define BS_ROOM(bs,pos) (bs)[pos / BITS_IN_ROOM]
341 #define BS_BIT(pos) (1 << (pos % BITS_IN_ROOM))
343 #define BITSET_AT(bs, pos) (BS_ROOM(bs,pos) & BS_BIT(pos))
344 #define BITSET_SET_BIT(bs, pos) BS_ROOM(bs,pos) |= BS_BIT(pos)
345 #define BITSET_CLEAR_BIT(bs, pos) BS_ROOM(bs,pos) &= ~(BS_BIT(pos))
346 #define BITSET_INVERT_BIT(bs, pos) BS_ROOM(bs,pos) ^= BS_BIT(pos)
349 typedef struct _BBuf {
355 #define BB_INIT(buf,size) onig_bbuf_init((BBuf* )(buf), (size))
357 #define BB_SIZE_INC(buf,inc) do{\
358 (buf)->alloc += (inc);\
359 (buf)->p = (UChar* )xrealloc((buf)->p, (buf)->alloc);\
360 if (IS_NULL((buf)->p)) return(ONIGERR_MEMORY);\
363 #define BB_EXPAND(buf,low) do{\
364 do { (buf)->alloc *= 2; } while ((buf)->alloc < (unsigned int )low);\
365 (buf)->p = (UChar* )xrealloc((buf)->p, (buf)->alloc);\
366 if (IS_NULL((buf)->p)) return(ONIGERR_MEMORY);\
369 #define BB_ENSURE_SIZE(buf,size) do{\
370 unsigned int new_alloc = (buf)->alloc;\
371 while (new_alloc < (unsigned int )(size)) { new_alloc *= 2; }\
372 if ((buf)->alloc != new_alloc) {\
373 (buf)->p = (UChar* )xrealloc((buf)->p, new_alloc);\
374 if (IS_NULL((buf)->p)) return(ONIGERR_MEMORY);\
375 (buf)->alloc = new_alloc;\
379 #define BB_WRITE(buf,pos,bytes,n) do{\
380 int used = (pos) + (n);\
381 if ((buf)->alloc < (unsigned int )used) BB_EXPAND((buf),used);\
382 xmemcpy((buf)->p + (pos), (bytes), (n));\
383 if ((buf)->used < (unsigned int )used) (buf)->used = used;\
386 #define BB_WRITE1(buf,pos,byte) do{\
387 int used = (pos) + 1;\
388 if ((buf)->alloc < (unsigned int )used) BB_EXPAND((buf),used);\
389 (buf)->p[(pos)] = (byte);\
390 if ((buf)->used < (unsigned int )used) (buf)->used = used;\
393 #define BB_ADD(buf,bytes,n) BB_WRITE((buf),(buf)->used,(bytes),(n))
394 #define BB_ADD1(buf,byte) BB_WRITE1((buf),(buf)->used,(byte))
395 #define BB_GET_ADD_ADDRESS(buf) ((buf)->p + (buf)->used)
396 #define BB_GET_OFFSET_POS(buf) ((buf)->used)
399 #define BB_MOVE_RIGHT(buf,from,to,n) do {\
400 if ((unsigned int )((to)+(n)) > (buf)->alloc) BB_EXPAND((buf),(to) + (n));\
401 xmemmove((buf)->p + (to), (buf)->p + (from), (n));\
402 if ((unsigned int )((to)+(n)) > (buf)->used) (buf)->used = (to) + (n);\
406 #define BB_MOVE_LEFT(buf,from,to,n) do {\
407 xmemmove((buf)->p + (to), (buf)->p + (from), (n));\
411 #define BB_MOVE_LEFT_REDUCE(buf,from,to) do {\
412 xmemmove((buf)->p + (to), (buf)->p + (from), (buf)->used - (from));\
413 (buf)->used -= (from - to);\
416 #define BB_INSERT(buf,pos,bytes,n) do {\
417 if (pos >= (buf)->used) {\
418 BB_WRITE(buf,pos,bytes,n);\
421 BB_MOVE_RIGHT((buf),(pos),(pos) + (n),((buf)->used - (pos)));\
422 xmemcpy((buf)->p + (pos), (bytes), (n));\
426 #define BB_GET_BYTE(buf, pos) (buf)->p[(pos)]
430 #define ANCHOR_PREC_READ (1<<0)
431 #define ANCHOR_PREC_READ_NOT (1<<1)
432 #define ANCHOR_LOOK_BEHIND (1<<2)
433 #define ANCHOR_LOOK_BEHIND_NOT (1<<3)
435 #define ANCHOR_BEGIN_BUF (1<<4)
436 #define ANCHOR_BEGIN_LINE (1<<5)
437 #define ANCHOR_BEGIN_POSITION (1<<6)
438 #define ANCHOR_END_BUF (1<<7)
439 #define ANCHOR_SEMI_END_BUF (1<<8)
440 #define ANCHOR_END_LINE (1<<9)
441 #define ANCHOR_WORD_BOUND (1<<10)
442 #define ANCHOR_NOT_WORD_BOUND (1<<11)
443 #define ANCHOR_WORD_BEGIN (1<<12)
444 #define ANCHOR_WORD_END (1<<13)
445 #define ANCHOR_ANYCHAR_STAR (1<<14) /* ".*" optimize info */
446 #define ANCHOR_ANYCHAR_STAR_ML (1<<15) /* ".*" optimize info (multi-line) */
448 #define ANCHOR_HAS_BODY(a) ((a)->type < ANCHOR_BEGIN_BUF)
453 OP_FINISH = 0, /* matching process terminator (no more alternative) */
454 OP_END = 1, /* pattern code terminator (success end) */
456 OP_EXACT1 = 2, /* single byte, N = 1 */
457 OP_EXACT2, /* single byte, N = 2 */
458 OP_EXACT3, /* single byte, N = 3 */
459 OP_EXACT4, /* single byte, N = 4 */
460 OP_EXACT5, /* single byte, N = 5 */
461 OP_EXACTN, /* single byte */
462 OP_EXACTMB2N1, /* mb-length = 2 N = 1 */
463 OP_EXACTMB2N2, /* mb-length = 2 N = 2 */
464 OP_EXACTMB2N3, /* mb-length = 2 N = 3 */
465 OP_EXACTMB2N, /* mb-length = 2 */
466 OP_EXACTMB3N, /* mb-length = 3 */
467 OP_EXACTMBN, /* other length */
469 OP_EXACT1_IC, /* single byte, N = 1, ignore case */
470 OP_EXACTN_IC, /* single byte, ignore case */
478 OP_CCLASS_NODE, /* pointer to CClassNode node */
480 OP_ANYCHAR, /* "." */
481 OP_ANYCHAR_ML, /* "." multi-line */
482 OP_ANYCHAR_STAR, /* ".*" */
483 OP_ANYCHAR_ML_STAR, /* ".*" multi-line */
484 OP_ANYCHAR_STAR_PEEK_NEXT,
485 OP_ANYCHAR_ML_STAR_PEEK_NEXT,
507 OP_BACKREF_WITH_LEVEL, /* \k<xxx+n>, \k<xxx-n> */
508 OP_BACKREF_CHECK, /* (?(n)), (?('name')) */
509 OP_BACKREF_CHECK_WITH_LEVEL, /* (?(n)), (?('name')) */
512 OP_MEMORY_START_PUSH, /* push back-tracker to stack */
513 OP_MEMORY_END_PUSH, /* push back-tracker to stack */
514 OP_MEMORY_END_PUSH_REC, /* push back-tracker to stack */
516 OP_MEMORY_END_REC, /* push marker to stack */
518 OP_FAIL, /* pop stack and move */
523 OP_PUSH_OR_JUMP_EXACT1, /* if match exact then push, else jump. */
524 OP_PUSH_IF_PEEK_NEXT, /* if match exact then push, else none. */
525 OP_REPEAT, /* {n,m} */
526 OP_REPEAT_NG, /* {n,m}? (non greedy) */
528 OP_REPEAT_INC_NG, /* non greedy */
529 OP_REPEAT_INC_SG, /* search and get in stack */
530 OP_REPEAT_INC_NG_SG, /* search and get in stack (non greedy) */
531 OP_EMPTY_CHECK_START, /* null loop checker start */
532 OP_EMPTY_CHECK_END, /* null loop checker end */
533 OP_EMPTY_CHECK_END_MEMST, /* null loop checker end (with capture status) */
534 OP_EMPTY_CHECK_END_MEMST_PUSH, /* with capture status and push check-end */
536 OP_PREC_READ_START, /* (?=...) start */
537 OP_PREC_READ_END, /* (?=...) end */
538 OP_PUSH_PREC_READ_NOT, /* (?!...) start */
539 OP_FAIL_PREC_READ_NOT, /* (?!...) end */
540 OP_PUSH_STOP_BT, /* (?>...) start */
541 OP_POP_STOP_BT, /* (?>...) end */
542 OP_LOOK_BEHIND, /* (?<=...) start (no needs end opcode) */
543 OP_PUSH_LOOK_BEHIND_NOT, /* (?<!...) start */
544 OP_FAIL_LOOK_BEHIND_NOT, /* (?<!...) end */
546 OP_CALL, /* \g<name> */
551 OP_STATE_CHECK_PUSH, /* combination explosion check and push */
552 OP_STATE_CHECK_PUSH_OR_JUMP, /* check ok -> push, else jump */
553 OP_STATE_CHECK, /* check only */
554 OP_STATE_CHECK_ANYCHAR_STAR,
555 OP_STATE_CHECK_ANYCHAR_ML_STAR,
557 /* no need: IS_DYNAMIC_OPTION() == 0 */
558 OP_SET_OPTION_PUSH, /* set option and push recover option */
559 OP_SET_OPTION /* set option */
563 SAVE_KEEP = 0, /* SAVE S */
565 SAVE_RIGHT_RANGE = 2,
569 UPDATE_VAR_KEEP_FROM_STACK_LAST = 0,
570 UPDATE_VAR_S_FROM_STACK = 1,
571 UPDATE_VAR_RIGHT_RANGE_FROM_STACK = 2,
572 UPDATE_VAR_RIGHT_RANGE_FROM_S_STACK = 3,
573 UPDATE_VAR_RIGHT_RANGE_INIT = 4,
576 typedef int RelAddrType;
577 typedef int AbsAddrType;
578 typedef int LengthType;
579 typedef int RepeatNumType;
580 typedef int MemNumType;
581 typedef short int StateCheckNumType;
582 typedef void* PointerType;
583 typedef int SaveType;
584 typedef int UpdateVarType;
586 #define SIZE_OPCODE 1
587 #define SIZE_RELADDR sizeof(RelAddrType)
588 #define SIZE_ABSADDR sizeof(AbsAddrType)
589 #define SIZE_LENGTH sizeof(LengthType)
590 #define SIZE_MEMNUM sizeof(MemNumType)
591 #define SIZE_STATE_CHECK_NUM sizeof(StateCheckNumType)
592 #define SIZE_REPEATNUM sizeof(RepeatNumType)
593 #define SIZE_OPTION sizeof(OnigOptionType)
594 #define SIZE_CODE_POINT sizeof(OnigCodePoint)
595 #define SIZE_POINTER sizeof(PointerType)
596 #define SIZE_SAVE_TYPE sizeof(SaveType)
597 #define SIZE_UPDATE_VAR_TYPE sizeof(UpdateVarType)
599 #define GET_RELADDR_INC(addr,p) PLATFORM_GET_INC(addr, p, RelAddrType)
600 #define GET_ABSADDR_INC(addr,p) PLATFORM_GET_INC(addr, p, AbsAddrType)
601 #define GET_LENGTH_INC(len,p) PLATFORM_GET_INC(len, p, LengthType)
602 #define GET_MEMNUM_INC(num,p) PLATFORM_GET_INC(num, p, MemNumType)
603 #define GET_REPEATNUM_INC(num,p) PLATFORM_GET_INC(num, p, RepeatNumType)
604 #define GET_OPTION_INC(option,p) PLATFORM_GET_INC(option, p, OnigOptionType)
605 #define GET_POINTER_INC(ptr,p) PLATFORM_GET_INC(ptr, p, PointerType)
606 #define GET_STATE_CHECK_NUM_INC(num,p) PLATFORM_GET_INC(num, p, StateCheckNumType)
607 #define GET_SAVE_TYPE_INC(type,p) PLATFORM_GET_INC(type, p, SaveType)
608 #define GET_UPDATE_VAR_TYPE_INC(type,p) PLATFORM_GET_INC(type, p, UpdateVarType)
610 /* code point's address must be aligned address. */
611 #define GET_CODE_POINT(code,p) code = *((OnigCodePoint* )(p))
612 #define GET_BYTE_INC(byte,p) do{\
618 /* op-code + arg size */
619 #define SIZE_OP_ANYCHAR_STAR SIZE_OPCODE
620 #define SIZE_OP_ANYCHAR_STAR_PEEK_NEXT (SIZE_OPCODE + 1)
621 #define SIZE_OP_JUMP (SIZE_OPCODE + SIZE_RELADDR)
622 #define SIZE_OP_PUSH (SIZE_OPCODE + SIZE_RELADDR)
623 #define SIZE_OP_PUSH_SUPER (SIZE_OPCODE + SIZE_RELADDR)
624 #define SIZE_OP_POP SIZE_OPCODE
625 #define SIZE_OP_PUSH_OR_JUMP_EXACT1 (SIZE_OPCODE + SIZE_RELADDR + 1)
626 #define SIZE_OP_PUSH_IF_PEEK_NEXT (SIZE_OPCODE + SIZE_RELADDR + 1)
627 #define SIZE_OP_REPEAT_INC (SIZE_OPCODE + SIZE_MEMNUM)
628 #define SIZE_OP_REPEAT_INC_NG (SIZE_OPCODE + SIZE_MEMNUM)
629 #define SIZE_OP_PREC_READ_START SIZE_OPCODE
630 #define SIZE_OP_PUSH_PREC_READ_NOT (SIZE_OPCODE + SIZE_RELADDR)
631 #define SIZE_OP_PREC_READ_END SIZE_OPCODE
632 #define SIZE_OP_FAIL_PREC_READ_NOT SIZE_OPCODE
633 #define SIZE_OP_SET_OPTION (SIZE_OPCODE + SIZE_OPTION)
634 #define SIZE_OP_SET_OPTION_PUSH (SIZE_OPCODE + SIZE_OPTION)
635 #define SIZE_OP_FAIL SIZE_OPCODE
636 #define SIZE_OP_MEMORY_START (SIZE_OPCODE + SIZE_MEMNUM)
637 #define SIZE_OP_MEMORY_START_PUSH (SIZE_OPCODE + SIZE_MEMNUM)
638 #define SIZE_OP_MEMORY_END_PUSH (SIZE_OPCODE + SIZE_MEMNUM)
639 #define SIZE_OP_MEMORY_END_PUSH_REC (SIZE_OPCODE + SIZE_MEMNUM)
640 #define SIZE_OP_MEMORY_END (SIZE_OPCODE + SIZE_MEMNUM)
641 #define SIZE_OP_MEMORY_END_REC (SIZE_OPCODE + SIZE_MEMNUM)
642 #define SIZE_OP_PUSH_STOP_BT SIZE_OPCODE
643 #define SIZE_OP_POP_STOP_BT SIZE_OPCODE
644 #define SIZE_OP_EMPTY_CHECK_START (SIZE_OPCODE + SIZE_MEMNUM)
645 #define SIZE_OP_EMPTY_CHECK_END (SIZE_OPCODE + SIZE_MEMNUM)
646 #define SIZE_OP_LOOK_BEHIND (SIZE_OPCODE + SIZE_LENGTH)
647 #define SIZE_OP_PUSH_LOOK_BEHIND_NOT (SIZE_OPCODE + SIZE_RELADDR + SIZE_LENGTH)
648 #define SIZE_OP_FAIL_LOOK_BEHIND_NOT SIZE_OPCODE
649 #define SIZE_OP_CALL (SIZE_OPCODE + SIZE_ABSADDR)
650 #define SIZE_OP_RETURN SIZE_OPCODE
651 #define SIZE_OP_PUSH_SAVE_VAL (SIZE_OPCODE + SIZE_SAVE_TYPE + SIZE_MEMNUM)
652 #define SIZE_OP_UPDATE_VAR (SIZE_OPCODE + SIZE_UPDATE_VAR_TYPE + SIZE_MEMNUM)
654 #ifdef USE_COMBINATION_EXPLOSION_CHECK
655 #define SIZE_OP_STATE_CHECK (SIZE_OPCODE + SIZE_STATE_CHECK_NUM)
656 #define SIZE_OP_STATE_CHECK_PUSH (SIZE_OPCODE + SIZE_STATE_CHECK_NUM + SIZE_RELADDR)
657 #define SIZE_OP_STATE_CHECK_PUSH_OR_JUMP (SIZE_OPCODE + SIZE_STATE_CHECK_NUM + SIZE_RELADDR)
658 #define SIZE_OP_STATE_CHECK_ANYCHAR_STAR (SIZE_OPCODE + SIZE_STATE_CHECK_NUM)
661 #define MC_ESC(syn) (syn)->meta_char_table.esc
662 #define MC_ANYCHAR(syn) (syn)->meta_char_table.anychar
663 #define MC_ANYTIME(syn) (syn)->meta_char_table.anytime
664 #define MC_ZERO_OR_ONE_TIME(syn) (syn)->meta_char_table.zero_or_one_time
665 #define MC_ONE_OR_MORE_TIME(syn) (syn)->meta_char_table.one_or_more_time
666 #define MC_ANYCHAR_ANYTIME(syn) (syn)->meta_char_table.anychar_anytime
668 #define IS_MC_ESC_CODE(code, syn) \
669 ((code) == MC_ESC(syn) && \
670 !IS_SYNTAX_OP2((syn), ONIG_SYN_OP2_INEFFECTIVE_ESCAPE))
673 #define SYN_POSIX_COMMON_OP \
674 ( ONIG_SYN_OP_DOT_ANYCHAR | ONIG_SYN_OP_POSIX_BRACKET | \
675 ONIG_SYN_OP_DECIMAL_BACKREF | \
676 ONIG_SYN_OP_BRACKET_CC | ONIG_SYN_OP_ASTERISK_ZERO_INF | \
677 ONIG_SYN_OP_LINE_ANCHOR | \
678 ONIG_SYN_OP_ESC_CONTROL_CHARS )
680 #define SYN_GNU_REGEX_OP \
681 ( ONIG_SYN_OP_DOT_ANYCHAR | ONIG_SYN_OP_BRACKET_CC | \
682 ONIG_SYN_OP_POSIX_BRACKET | ONIG_SYN_OP_DECIMAL_BACKREF | \
683 ONIG_SYN_OP_BRACE_INTERVAL | ONIG_SYN_OP_LPAREN_SUBEXP | \
684 ONIG_SYN_OP_VBAR_ALT | \
685 ONIG_SYN_OP_ASTERISK_ZERO_INF | ONIG_SYN_OP_PLUS_ONE_INF | \
686 ONIG_SYN_OP_QMARK_ZERO_ONE | \
687 ONIG_SYN_OP_ESC_AZ_BUF_ANCHOR | ONIG_SYN_OP_ESC_CAPITAL_G_BEGIN_ANCHOR | \
688 ONIG_SYN_OP_ESC_W_WORD | \
689 ONIG_SYN_OP_ESC_B_WORD_BOUND | ONIG_SYN_OP_ESC_LTGT_WORD_BEGIN_END | \
690 ONIG_SYN_OP_ESC_S_WHITE_SPACE | ONIG_SYN_OP_ESC_D_DIGIT | \
691 ONIG_SYN_OP_LINE_ANCHOR )
693 #define SYN_GNU_REGEX_BV \
694 ( ONIG_SYN_CONTEXT_INDEP_ANCHORS | ONIG_SYN_CONTEXT_INDEP_REPEAT_OPS | \
695 ONIG_SYN_CONTEXT_INVALID_REPEAT_OPS | ONIG_SYN_ALLOW_INVALID_INTERVAL | \
696 ONIG_SYN_BACKSLASH_ESCAPE_IN_CC | ONIG_SYN_ALLOW_DOUBLE_RANGE_OP_IN_CC )
699 #define NCCLASS_FLAGS(cc) ((cc)->flags)
700 #define NCCLASS_FLAG_SET(cc,flag) (NCCLASS_FLAGS(cc) |= (flag))
701 #define NCCLASS_FLAG_CLEAR(cc,flag) (NCCLASS_FLAGS(cc) &= ~(flag))
702 #define IS_NCCLASS_FLAG_ON(cc,flag) ((NCCLASS_FLAGS(cc) & (flag)) != 0)
705 #define FLAG_NCCLASS_NOT (1<<0)
706 #define FLAG_NCCLASS_SHARE (1<<1)
708 #define NCCLASS_SET_NOT(nd) NCCLASS_FLAG_SET(nd, FLAG_NCCLASS_NOT)
709 #define NCCLASS_CLEAR_NOT(nd) NCCLASS_FLAG_CLEAR(nd, FLAG_NCCLASS_NOT)
710 #define IS_NCCLASS_NOT(nd) IS_NCCLASS_FLAG_ON(nd, FLAG_NCCLASS_NOT)
715 OnigOptionType options;
718 const UChar* start; /* search start position (for \G: BEGIN_POSITION) */
719 #ifdef USE_FIND_LONGEST_SEARCH_ALL_OF_RANGE
720 int best_len; /* for ONIG_OPTION_FIND_LONGEST */
723 #ifdef USE_COMBINATION_EXPLOSION_CHECK
724 void* state_check_buff;
725 int state_check_buff_size;
730 #define IS_CODE_SB_WORD(enc,code) \
731 (ONIGENC_IS_CODE_ASCII(code) && ONIGENC_IS_CODE_WORD(enc,code))
733 typedef struct OnigEndCallListItem {
734 struct OnigEndCallListItem* next;
736 } OnigEndCallListItemType;
738 extern void onig_add_end_call(void (*func)(void));
749 extern OnigOpInfoType OnigOpInfo[];
752 extern void onig_print_compiled_byte_code P_((FILE* f, UChar* bp, UChar** nextp, UChar* start, OnigEncoding enc));
754 #ifdef ONIG_DEBUG_STATISTICS
755 extern void onig_statistics_init P_((void));
756 extern int onig_print_statistics P_((FILE* f));
760 extern void onig_warning(const char* s);
761 extern UChar* onig_error_code_to_format P_((int code));
762 extern void onig_snprintf_with_pattern PV_((UChar buf[], int bufsize, OnigEncoding enc, UChar* pat, UChar* pat_end, const UChar *fmt, ...));
763 extern int onig_bbuf_init P_((BBuf* buf, int size));
764 extern int onig_compile P_((regex_t* reg, const UChar* pattern, const UChar* pattern_end, OnigErrorInfo* einfo));
765 extern void onig_transfer P_((regex_t* to, regex_t* from));
766 extern int onig_is_code_in_cc_len P_((int enclen, OnigCodePoint code, void* /* CClassNode* */ cc));
769 typedef void hash_table_type;
771 # include <windows.h>
772 typedef ULONG_PTR hash_data_type;
774 typedef unsigned long hash_data_type;
777 extern hash_table_type* onig_st_init_strend_table_with_size P_((int size));
778 extern int onig_st_lookup_strend P_((hash_table_type* table, const UChar* str_key, const UChar* end_key, hash_data_type *value));
779 extern int onig_st_insert_strend P_((hash_table_type* table, const UChar* str_key, const UChar* end_key, hash_data_type value));
781 typedef int (*ONIGENC_INIT_PROPERTY_LIST_FUNC_TYPE)(void);
783 #endif /* REGINT_H */