sender 1.5.1 → 1.5.3

Sign up to get free protection for your applications and to get access to all the features.
data/ext/sender/regint.h DELETED
@@ -1,842 +0,0 @@
1
- #ifndef ONIGURUMA_REGINT_H
2
- #define ONIGURUMA_REGINT_H
3
- /**********************************************************************
4
- regint.h - Oniguruma (regular expression library)
5
- **********************************************************************/
6
- /*-
7
- * Copyright (c) 2002-2007 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__) || defined(_M_AMD86) || \
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
- #ifndef RUBY_DEFINES_H
87
- #include "ruby/ruby.h"
88
- #undef xmalloc
89
- #undef xrealloc
90
- #undef xcalloc
91
- #undef xfree
92
- #endif
93
- #ifdef ONIG_ESCAPE_UCHAR_COLLISION
94
- #undef ONIG_ESCAPE_UCHAR_COLLISION
95
- #endif
96
- #undef USE_MATCH_RANGE_IS_COMPLETE_RANGE
97
- #undef USE_CAPTURE_HISTORY
98
- #define USE_VARIABLE_META_CHARS
99
- #define USE_WORD_BEGIN_END /* "\<": word-begin, "\>": word-end */
100
- #define USE_POSIX_REGION_OPTION /* needed for POSIX API support */
101
- #define USE_FIND_LONGEST_SEARCH_ALL_OF_RANGE
102
- /* #define USE_COMBINATION_EXPLOSION_CHECK */ /* (X*)* */
103
- /* #define USE_MULTI_THREAD_SYSTEM */
104
- #define THREAD_SYSTEM_INIT /* depend on thread system */
105
- #define THREAD_SYSTEM_END /* depend on thread system */
106
- #define THREAD_ATOMIC_START /* depend on thread system */
107
- #define THREAD_ATOMIC_END /* depend on thread system */
108
- #define THREAD_PASS /* depend on thread system */
109
- #ifndef xmalloc
110
- #define xmalloc malloc
111
- #define xrealloc realloc
112
- #define xcalloc calloc
113
- #define xfree free
114
- #endif
115
-
116
- #ifdef RUBY
117
-
118
- #define CHECK_INTERRUPT_IN_MATCH_AT rb_thread_check_ints()
119
- #define onig_st_init_table st_init_table
120
- #define onig_st_init_table_with_size st_init_table_with_size
121
- #define onig_st_init_numtable st_init_numtable
122
- #define onig_st_init_numtable_with_size st_init_numtable_with_size
123
- #define onig_st_init_strtable st_init_strtable
124
- #define onig_st_init_strtable_with_size st_init_strtable_with_size
125
- #define onig_st_delete st_delete
126
- #define onig_st_delete_safe st_delete_safe
127
- #define onig_st_insert st_insert
128
- #define onig_st_lookup st_lookup
129
- #define onig_st_foreach st_foreach
130
- #define onig_st_add_direct st_add_direct
131
- #define onig_st_free_table st_free_table
132
- #define onig_st_cleanup_safe st_cleanup_safe
133
- #define onig_st_copy st_copy
134
- #define onig_st_nothing_key_clone st_nothing_key_clone
135
- #define onig_st_nothing_key_free st_nothing_key_free
136
- #define onig_st_is_member st_is_member
137
-
138
- #define USE_UPPER_CASE_TABLE
139
- #else
140
-
141
- #define st_init_table onig_st_init_table
142
- #define st_init_table_with_size onig_st_init_table_with_size
143
- #define st_init_numtable onig_st_init_numtable
144
- #define st_init_numtable_with_size onig_st_init_numtable_with_size
145
- #define st_init_strtable onig_st_init_strtable
146
- #define st_init_strtable_with_size onig_st_init_strtable_with_size
147
- #define st_delete onig_st_delete
148
- #define st_delete_safe onig_st_delete_safe
149
- #define st_insert onig_st_insert
150
- #define st_lookup onig_st_lookup
151
- #define st_foreach onig_st_foreach
152
- #define st_add_direct onig_st_add_direct
153
- #define st_free_table onig_st_free_table
154
- #define st_cleanup_safe onig_st_cleanup_safe
155
- #define st_copy onig_st_copy
156
- #define st_nothing_key_clone onig_st_nothing_key_clone
157
- #define st_nothing_key_free onig_st_nothing_key_free
158
- /* */
159
- #define onig_st_is_member st_is_member
160
-
161
- #define CHECK_INTERRUPT_IN_MATCH_AT
162
-
163
- #endif
164
-
165
- #define STATE_CHECK_STRING_THRESHOLD_LEN 7
166
- #define STATE_CHECK_BUFF_MAX_SIZE 0x4000
167
-
168
- #define THREAD_PASS_LIMIT_COUNT 8
169
- #define xmemset memset
170
- #define xmemcpy memcpy
171
- #define xmemmove memmove
172
-
173
- #if defined(_WIN32) && !defined(__GNUC__)
174
- #define xalloca _alloca
175
- #define xvsnprintf _vsnprintf
176
- #else
177
- #define xalloca alloca
178
- #define xvsnprintf vsnprintf
179
- #endif
180
-
181
-
182
- #if defined(USE_RECOMPILE_API) && defined(USE_MULTI_THREAD_SYSTEM)
183
- #define ONIG_STATE_INC(reg) (reg)->state++
184
- #define ONIG_STATE_DEC(reg) (reg)->state--
185
-
186
- #define ONIG_STATE_INC_THREAD(reg) do {\
187
- THREAD_ATOMIC_START;\
188
- (reg)->state++;\
189
- THREAD_ATOMIC_END;\
190
- } while(0)
191
- #define ONIG_STATE_DEC_THREAD(reg) do {\
192
- THREAD_ATOMIC_START;\
193
- (reg)->state--;\
194
- THREAD_ATOMIC_END;\
195
- } while(0)
196
- #else
197
- #define ONIG_STATE_INC(reg) /* Nothing */
198
- #define ONIG_STATE_DEC(reg) /* Nothing */
199
- #define ONIG_STATE_INC_THREAD(reg) /* Nothing */
200
- #define ONIG_STATE_DEC_THREAD(reg) /* Nothing */
201
- #endif /* USE_RECOMPILE_API && USE_MULTI_THREAD_SYSTEM */
202
-
203
- #ifdef HAVE_STDLIB_H
204
- #include <stdlib.h>
205
- #endif
206
-
207
- #if defined(HAVE_ALLOCA_H) && (defined(_AIX) || !defined(__GNUC__))
208
- #include <alloca.h>
209
- #endif
210
-
211
- #ifdef HAVE_STRING_H
212
- # include <string.h>
213
- #else
214
- # include <strings.h>
215
- #endif
216
-
217
- #include <ctype.h>
218
- #ifdef HAVE_SYS_TYPES_H
219
- #include <sys/types.h>
220
- #endif
221
-
222
- #ifdef ONIG_DEBUG
223
- # include <stdio.h>
224
- #endif
225
-
226
- #include "regenc.h"
227
-
228
- #ifdef MIN
229
- #undef MIN
230
- #endif
231
- #ifdef MAX
232
- #undef MAX
233
- #endif
234
- #define MIN(a,b) (((a)>(b))?(b):(a))
235
- #define MAX(a,b) (((a)<(b))?(b):(a))
236
-
237
- #define IS_NULL(p) (((void*)(p)) == (void*)0)
238
- #define IS_NOT_NULL(p) (((void*)(p)) != (void*)0)
239
- #define CHECK_NULL_RETURN(p) if (IS_NULL(p)) return NULL
240
- #define CHECK_NULL_RETURN_MEMERR(p) if (IS_NULL(p)) return ONIGERR_MEMORY
241
- #define NULL_UCHARP ((UChar* )0)
242
-
243
- #ifdef PLATFORM_UNALIGNED_WORD_ACCESS
244
-
245
- #define PLATFORM_GET_INC(val,p,type) do{\
246
- val = *(type* )p;\
247
- (p) += sizeof(type);\
248
- } while(0)
249
-
250
- #else
251
-
252
- #define PLATFORM_GET_INC(val,p,type) do{\
253
- xmemcpy(&val, (p), sizeof(type));\
254
- (p) += sizeof(type);\
255
- } while(0)
256
-
257
- /* sizeof(OnigCodePoint) */
258
- #define WORD_ALIGNMENT_SIZE SIZEOF_LONG
259
-
260
- #define GET_ALIGNMENT_PAD_SIZE(addr,pad_size) do {\
261
- (pad_size) = WORD_ALIGNMENT_SIZE \
262
- - ((uintptr_t )(addr) % WORD_ALIGNMENT_SIZE);\
263
- if ((pad_size) == WORD_ALIGNMENT_SIZE) (pad_size) = 0;\
264
- } while (0)
265
-
266
- #define ALIGNMENT_RIGHT(addr) do {\
267
- (addr) += (WORD_ALIGNMENT_SIZE - 1);\
268
- (addr) -= ((uintptr_t )(addr) % WORD_ALIGNMENT_SIZE);\
269
- } while (0)
270
-
271
- #endif /* PLATFORM_UNALIGNED_WORD_ACCESS */
272
-
273
- /* stack pop level */
274
- #define STACK_POP_LEVEL_FREE 0
275
- #define STACK_POP_LEVEL_MEM_START 1
276
- #define STACK_POP_LEVEL_ALL 2
277
-
278
- /* optimize flags */
279
- #define ONIG_OPTIMIZE_NONE 0
280
- #define ONIG_OPTIMIZE_EXACT 1 /* Slow Search */
281
- #define ONIG_OPTIMIZE_EXACT_BM 2 /* Boyer Moore Search */
282
- #define ONIG_OPTIMIZE_EXACT_BM_NOT_REV 3 /* BM (but not simple match) */
283
- #define ONIG_OPTIMIZE_EXACT_IC 4 /* Slow Search (ignore case) */
284
- #define ONIG_OPTIMIZE_MAP 5 /* char map */
285
-
286
- /* bit status */
287
- typedef unsigned int BitStatusType;
288
-
289
- #define BIT_STATUS_BITS_NUM (sizeof(BitStatusType) * 8)
290
- #define BIT_STATUS_CLEAR(stats) (stats) = 0
291
- #define BIT_STATUS_ON_ALL(stats) (stats) = ~((BitStatusType )0)
292
- #define BIT_STATUS_AT(stats,n) \
293
- ((n) < (int )BIT_STATUS_BITS_NUM ? ((stats) & (1 << n)) : ((stats) & 1))
294
-
295
- #define BIT_STATUS_ON_AT(stats,n) do {\
296
- if ((n) < (int )BIT_STATUS_BITS_NUM) \
297
- (stats) |= (1 << (n));\
298
- else\
299
- (stats) |= 1;\
300
- } while (0)
301
-
302
- #define BIT_STATUS_ON_AT_SIMPLE(stats,n) do {\
303
- if ((n) < (int )BIT_STATUS_BITS_NUM)\
304
- (stats) |= (1 << (n));\
305
- } while (0)
306
-
307
-
308
- #define INT_MAX_LIMIT ((1UL << (SIZEOF_INT * 8 - 1)) - 1)
309
-
310
- #define DIGITVAL(code) ((code) - '0')
311
- #define ODIGITVAL(code) DIGITVAL(code)
312
- #define XDIGITVAL(enc,code) \
313
- (ONIGENC_IS_CODE_DIGIT(enc,code) ? DIGITVAL(code) \
314
- : (ONIGENC_IS_CODE_UPPER(enc,code) ? (code) - 'A' + 10 : (code) - 'a' + 10))
315
-
316
- #define IS_SINGLELINE(option) ((option) & ONIG_OPTION_SINGLELINE)
317
- #define IS_MULTILINE(option) ((option) & ONIG_OPTION_MULTILINE)
318
- #define IS_IGNORECASE(option) ((option) & ONIG_OPTION_IGNORECASE)
319
- #define IS_EXTEND(option) ((option) & ONIG_OPTION_EXTEND)
320
- #define IS_FIND_LONGEST(option) ((option) & ONIG_OPTION_FIND_LONGEST)
321
- #define IS_FIND_NOT_EMPTY(option) ((option) & ONIG_OPTION_FIND_NOT_EMPTY)
322
- #define IS_FIND_CONDITION(option) ((option) & \
323
- (ONIG_OPTION_FIND_LONGEST | ONIG_OPTION_FIND_NOT_EMPTY))
324
- #define IS_NOTBOL(option) ((option) & ONIG_OPTION_NOTBOL)
325
- #define IS_NOTEOL(option) ((option) & ONIG_OPTION_NOTEOL)
326
- #define IS_POSIX_REGION(option) ((option) & ONIG_OPTION_POSIX_REGION)
327
-
328
- /* OP_SET_OPTION is required for these options.
329
- #define IS_DYNAMIC_OPTION(option) \
330
- (((option) & (ONIG_OPTION_MULTILINE | ONIG_OPTION_IGNORECASE)) != 0)
331
- */
332
- /* ignore-case and multibyte status are included in compiled code. */
333
- #define IS_DYNAMIC_OPTION(option) 0
334
-
335
- #define DISABLE_CASE_FOLD_MULTI_CHAR(case_fold_flag) \
336
- ((case_fold_flag) & ~INTERNAL_ONIGENC_CASE_FOLD_MULTI_CHAR)
337
-
338
- #define REPEAT_INFINITE -1
339
- #define IS_REPEAT_INFINITE(n) ((n) == REPEAT_INFINITE)
340
-
341
- /* bitset */
342
- #define BITS_PER_BYTE 8
343
- #define SINGLE_BYTE_SIZE (1 << BITS_PER_BYTE)
344
- #define BITS_IN_ROOM (sizeof(Bits) * BITS_PER_BYTE)
345
- #define BITSET_SIZE (SINGLE_BYTE_SIZE / BITS_IN_ROOM)
346
-
347
- #ifdef PLATFORM_UNALIGNED_WORD_ACCESS
348
- typedef unsigned int Bits;
349
- #else
350
- typedef unsigned char Bits;
351
- #endif
352
- typedef Bits BitSet[BITSET_SIZE];
353
- typedef Bits* BitSetRef;
354
-
355
- #define SIZE_BITSET sizeof(BitSet)
356
-
357
- #define BITSET_CLEAR(bs) do {\
358
- int i;\
359
- for (i = 0; i < (int )BITSET_SIZE; i++) { (bs)[i] = 0; } \
360
- } while (0)
361
-
362
- #define BS_ROOM(bs,pos) (bs)[pos / BITS_IN_ROOM]
363
- #define BS_BIT(pos) (1 << (pos % BITS_IN_ROOM))
364
-
365
- #define BITSET_AT(bs, pos) (BS_ROOM(bs,pos) & BS_BIT(pos))
366
- #define BITSET_SET_BIT(bs, pos) BS_ROOM(bs,pos) |= BS_BIT(pos)
367
- #define BITSET_CLEAR_BIT(bs, pos) BS_ROOM(bs,pos) &= ~(BS_BIT(pos))
368
- #define BITSET_INVERT_BIT(bs, pos) BS_ROOM(bs,pos) ^= BS_BIT(pos)
369
-
370
- /* bytes buffer */
371
- typedef struct _BBuf {
372
- UChar* p;
373
- unsigned int used;
374
- unsigned int alloc;
375
- } BBuf;
376
-
377
- #define BBUF_INIT(buf,size) onig_bbuf_init((BBuf* )(buf), (size))
378
-
379
- #define BBUF_SIZE_INC(buf,inc) do{\
380
- (buf)->alloc += (inc);\
381
- (buf)->p = (UChar* )xrealloc((buf)->p, (buf)->alloc);\
382
- if (IS_NULL((buf)->p)) return(ONIGERR_MEMORY);\
383
- } while (0)
384
-
385
- #define BBUF_EXPAND(buf,low) do{\
386
- do { (buf)->alloc *= 2; } while ((buf)->alloc < (unsigned int )low);\
387
- (buf)->p = (UChar* )xrealloc((buf)->p, (buf)->alloc);\
388
- if (IS_NULL((buf)->p)) return(ONIGERR_MEMORY);\
389
- } while (0)
390
-
391
- #define BBUF_ENSURE_SIZE(buf,size) do{\
392
- unsigned int new_alloc = (buf)->alloc;\
393
- while (new_alloc < (unsigned int )(size)) { new_alloc *= 2; }\
394
- if ((buf)->alloc != new_alloc) {\
395
- (buf)->p = (UChar* )xrealloc((buf)->p, new_alloc);\
396
- if (IS_NULL((buf)->p)) return(ONIGERR_MEMORY);\
397
- (buf)->alloc = new_alloc;\
398
- }\
399
- } while (0)
400
-
401
- #define BBUF_WRITE(buf,pos,bytes,n) do{\
402
- int used = (pos) + (n);\
403
- if ((buf)->alloc < (unsigned int )used) BBUF_EXPAND((buf),used);\
404
- xmemcpy((buf)->p + (pos), (bytes), (n));\
405
- if ((buf)->used < (unsigned int )used) (buf)->used = used;\
406
- } while (0)
407
-
408
- #define BBUF_WRITE1(buf,pos,byte) do{\
409
- int used = (pos) + 1;\
410
- if ((buf)->alloc < (unsigned int )used) BBUF_EXPAND((buf),used);\
411
- (buf)->p[(pos)] = (byte);\
412
- if ((buf)->used < (unsigned int )used) (buf)->used = used;\
413
- } while (0)
414
-
415
- #define BBUF_ADD(buf,bytes,n) BBUF_WRITE((buf),(buf)->used,(bytes),(n))
416
- #define BBUF_ADD1(buf,byte) BBUF_WRITE1((buf),(buf)->used,(byte))
417
- #define BBUF_GET_ADD_ADDRESS(buf) ((buf)->p + (buf)->used)
418
- #define BBUF_GET_OFFSET_POS(buf) ((buf)->used)
419
-
420
- /* from < to */
421
- #define BBUF_MOVE_RIGHT(buf,from,to,n) do {\
422
- if ((unsigned int )((to)+(n)) > (buf)->alloc) BBUF_EXPAND((buf),(to) + (n));\
423
- xmemmove((buf)->p + (to), (buf)->p + (from), (n));\
424
- if ((unsigned int )((to)+(n)) > (buf)->used) (buf)->used = (to) + (n);\
425
- } while (0)
426
-
427
- /* from > to */
428
- #define BBUF_MOVE_LEFT(buf,from,to,n) do {\
429
- xmemmove((buf)->p + (to), (buf)->p + (from), (n));\
430
- } while (0)
431
-
432
- /* from > to */
433
- #define BBUF_MOVE_LEFT_REDUCE(buf,from,to) do {\
434
- xmemmove((buf)->p + (to), (buf)->p + (from), (buf)->used - (from));\
435
- (buf)->used -= (from - to);\
436
- } while (0)
437
-
438
- #define BBUF_INSERT(buf,pos,bytes,n) do {\
439
- if (pos >= (buf)->used) {\
440
- BBUF_WRITE(buf,pos,bytes,n);\
441
- }\
442
- else {\
443
- BBUF_MOVE_RIGHT((buf),(pos),(pos) + (n),((buf)->used - (pos)));\
444
- xmemcpy((buf)->p + (pos), (bytes), (n));\
445
- }\
446
- } while (0)
447
-
448
- #define BBUF_GET_BYTE(buf, pos) (buf)->p[(pos)]
449
-
450
-
451
- #define ANCHOR_BEGIN_BUF (1<<0)
452
- #define ANCHOR_BEGIN_LINE (1<<1)
453
- #define ANCHOR_BEGIN_POSITION (1<<2)
454
- #define ANCHOR_END_BUF (1<<3)
455
- #define ANCHOR_SEMI_END_BUF (1<<4)
456
- #define ANCHOR_END_LINE (1<<5)
457
-
458
- #define ANCHOR_WORD_BOUND (1<<6)
459
- #define ANCHOR_NOT_WORD_BOUND (1<<7)
460
- #define ANCHOR_WORD_BEGIN (1<<8)
461
- #define ANCHOR_WORD_END (1<<9)
462
- #define ANCHOR_PREC_READ (1<<10)
463
- #define ANCHOR_PREC_READ_NOT (1<<11)
464
- #define ANCHOR_LOOK_BEHIND (1<<12)
465
- #define ANCHOR_LOOK_BEHIND_NOT (1<<13)
466
-
467
- #define ANCHOR_ANYCHAR_STAR (1<<14) /* ".*" optimize info */
468
- #define ANCHOR_ANYCHAR_STAR_ML (1<<15) /* ".*" optimize info (multi-line) */
469
-
470
- /* operation code */
471
- enum OpCode {
472
- OP_FINISH = 0, /* matching process terminator (no more alternative) */
473
- OP_END = 1, /* pattern code terminator (success end) */
474
-
475
- OP_EXACT1 = 2, /* single byte, N = 1 */
476
- OP_EXACT2, /* single byte, N = 2 */
477
- OP_EXACT3, /* single byte, N = 3 */
478
- OP_EXACT4, /* single byte, N = 4 */
479
- OP_EXACT5, /* single byte, N = 5 */
480
- OP_EXACTN, /* single byte */
481
- OP_EXACTMB2N1, /* mb-length = 2 N = 1 */
482
- OP_EXACTMB2N2, /* mb-length = 2 N = 2 */
483
- OP_EXACTMB2N3, /* mb-length = 2 N = 3 */
484
- OP_EXACTMB2N, /* mb-length = 2 */
485
- OP_EXACTMB3N, /* mb-length = 3 */
486
- OP_EXACTMBN, /* other length */
487
-
488
- OP_EXACT1_IC, /* single byte, N = 1, ignore case */
489
- OP_EXACTN_IC, /* single byte, ignore case */
490
-
491
- OP_CCLASS,
492
- OP_CCLASS_MB,
493
- OP_CCLASS_MIX,
494
- OP_CCLASS_NOT,
495
- OP_CCLASS_MB_NOT,
496
- OP_CCLASS_MIX_NOT,
497
- OP_CCLASS_NODE, /* pointer to CClassNode node */
498
-
499
- OP_ANYCHAR, /* "." */
500
- OP_ANYCHAR_ML, /* "." multi-line */
501
- OP_ANYCHAR_STAR, /* ".*" */
502
- OP_ANYCHAR_ML_STAR, /* ".*" multi-line */
503
- OP_ANYCHAR_STAR_PEEK_NEXT,
504
- OP_ANYCHAR_ML_STAR_PEEK_NEXT,
505
-
506
- OP_WORD,
507
- OP_NOT_WORD,
508
- OP_WORD_BOUND,
509
- OP_NOT_WORD_BOUND,
510
- OP_WORD_BEGIN,
511
- OP_WORD_END,
512
-
513
- OP_BEGIN_BUF,
514
- OP_END_BUF,
515
- OP_BEGIN_LINE,
516
- OP_END_LINE,
517
- OP_SEMI_END_BUF,
518
- OP_BEGIN_POSITION,
519
-
520
- OP_BACKREF1,
521
- OP_BACKREF2,
522
- OP_BACKREFN,
523
- OP_BACKREFN_IC,
524
- OP_BACKREF_MULTI,
525
- OP_BACKREF_MULTI_IC,
526
- OP_BACKREF_WITH_LEVEL, /* \k<xxx+n>, \k<xxx-n> */
527
-
528
- OP_MEMORY_START,
529
- OP_MEMORY_START_PUSH, /* push back-tracker to stack */
530
- OP_MEMORY_END_PUSH, /* push back-tracker to stack */
531
- OP_MEMORY_END_PUSH_REC, /* push back-tracker to stack */
532
- OP_MEMORY_END,
533
- OP_MEMORY_END_REC, /* push marker to stack */
534
-
535
- OP_FAIL, /* pop stack and move */
536
- OP_JUMP,
537
- OP_PUSH,
538
- OP_POP,
539
- OP_PUSH_OR_JUMP_EXACT1, /* if match exact then push, else jump. */
540
- OP_PUSH_IF_PEEK_NEXT, /* if match exact then push, else none. */
541
- OP_REPEAT, /* {n,m} */
542
- OP_REPEAT_NG, /* {n,m}? (non greedy) */
543
- OP_REPEAT_INC,
544
- OP_REPEAT_INC_NG, /* non greedy */
545
- OP_REPEAT_INC_SG, /* search and get in stack */
546
- OP_REPEAT_INC_NG_SG, /* search and get in stack (non greedy) */
547
- OP_NULL_CHECK_START, /* null loop checker start */
548
- OP_NULL_CHECK_END, /* null loop checker end */
549
- OP_NULL_CHECK_END_MEMST, /* null loop checker end (with capture status) */
550
- OP_NULL_CHECK_END_MEMST_PUSH, /* with capture status and push check-end */
551
-
552
- OP_PUSH_POS, /* (?=...) start */
553
- OP_POP_POS, /* (?=...) end */
554
- OP_PUSH_POS_NOT, /* (?!...) start */
555
- OP_FAIL_POS, /* (?!...) end */
556
- OP_PUSH_STOP_BT, /* (?>...) start */
557
- OP_POP_STOP_BT, /* (?>...) end */
558
- OP_LOOK_BEHIND, /* (?<=...) start (no needs end opcode) */
559
- OP_PUSH_LOOK_BEHIND_NOT, /* (?<!...) start */
560
- OP_FAIL_LOOK_BEHIND_NOT, /* (?<!...) end */
561
-
562
- OP_CALL, /* \g<name> */
563
- OP_RETURN,
564
-
565
- OP_STATE_CHECK_PUSH, /* combination explosion check and push */
566
- OP_STATE_CHECK_PUSH_OR_JUMP, /* check ok -> push, else jump */
567
- OP_STATE_CHECK, /* check only */
568
- OP_STATE_CHECK_ANYCHAR_STAR,
569
- OP_STATE_CHECK_ANYCHAR_ML_STAR,
570
-
571
- /* no need: IS_DYNAMIC_OPTION() == 0 */
572
- OP_SET_OPTION_PUSH, /* set option and push recover option */
573
- OP_SET_OPTION /* set option */
574
- };
575
-
576
- typedef int RelAddrType;
577
- typedef int AbsAddrType;
578
- typedef int LengthType;
579
- typedef int RepeatNumType;
580
- typedef short int MemNumType;
581
- typedef short int StateCheckNumType;
582
- typedef void* PointerType;
583
-
584
- #define SIZE_OPCODE 1
585
- #define SIZE_RELADDR sizeof(RelAddrType)
586
- #define SIZE_ABSADDR sizeof(AbsAddrType)
587
- #define SIZE_LENGTH sizeof(LengthType)
588
- #define SIZE_MEMNUM sizeof(MemNumType)
589
- #define SIZE_STATE_CHECK_NUM sizeof(StateCheckNumType)
590
- #define SIZE_REPEATNUM sizeof(RepeatNumType)
591
- #define SIZE_OPTION sizeof(OnigOptionType)
592
- #define SIZE_CODE_POINT sizeof(OnigCodePoint)
593
- #define SIZE_POINTER sizeof(PointerType)
594
-
595
-
596
- #define GET_RELADDR_INC(addr,p) PLATFORM_GET_INC(addr, p, RelAddrType)
597
- #define GET_ABSADDR_INC(addr,p) PLATFORM_GET_INC(addr, p, AbsAddrType)
598
- #define GET_LENGTH_INC(len,p) PLATFORM_GET_INC(len, p, LengthType)
599
- #define GET_MEMNUM_INC(num,p) PLATFORM_GET_INC(num, p, MemNumType)
600
- #define GET_REPEATNUM_INC(num,p) PLATFORM_GET_INC(num, p, RepeatNumType)
601
- #define GET_OPTION_INC(option,p) PLATFORM_GET_INC(option, p, OnigOptionType)
602
- #define GET_POINTER_INC(ptr,p) PLATFORM_GET_INC(ptr, p, PointerType)
603
- #define GET_STATE_CHECK_NUM_INC(num,p) PLATFORM_GET_INC(num, p, StateCheckNumType)
604
-
605
- /* code point's address must be aligned address. */
606
- #define GET_CODE_POINT(code,p) code = *((OnigCodePoint* )(p))
607
- #define GET_BYTE_INC(byte,p) do{\
608
- byte = *(p);\
609
- (p)++;\
610
- } while(0)
611
-
612
-
613
- /* op-code + arg size */
614
- #define SIZE_OP_ANYCHAR_STAR SIZE_OPCODE
615
- #define SIZE_OP_ANYCHAR_STAR_PEEK_NEXT (SIZE_OPCODE + 1)
616
- #define SIZE_OP_JUMP (SIZE_OPCODE + SIZE_RELADDR)
617
- #define SIZE_OP_PUSH (SIZE_OPCODE + SIZE_RELADDR)
618
- #define SIZE_OP_POP SIZE_OPCODE
619
- #define SIZE_OP_PUSH_OR_JUMP_EXACT1 (SIZE_OPCODE + SIZE_RELADDR + 1)
620
- #define SIZE_OP_PUSH_IF_PEEK_NEXT (SIZE_OPCODE + SIZE_RELADDR + 1)
621
- #define SIZE_OP_REPEAT_INC (SIZE_OPCODE + SIZE_MEMNUM)
622
- #define SIZE_OP_REPEAT_INC_NG (SIZE_OPCODE + SIZE_MEMNUM)
623
- #define SIZE_OP_PUSH_POS SIZE_OPCODE
624
- #define SIZE_OP_PUSH_POS_NOT (SIZE_OPCODE + SIZE_RELADDR)
625
- #define SIZE_OP_POP_POS SIZE_OPCODE
626
- #define SIZE_OP_FAIL_POS SIZE_OPCODE
627
- #define SIZE_OP_SET_OPTION (SIZE_OPCODE + SIZE_OPTION)
628
- #define SIZE_OP_SET_OPTION_PUSH (SIZE_OPCODE + SIZE_OPTION)
629
- #define SIZE_OP_FAIL SIZE_OPCODE
630
- #define SIZE_OP_MEMORY_START (SIZE_OPCODE + SIZE_MEMNUM)
631
- #define SIZE_OP_MEMORY_START_PUSH (SIZE_OPCODE + SIZE_MEMNUM)
632
- #define SIZE_OP_MEMORY_END_PUSH (SIZE_OPCODE + SIZE_MEMNUM)
633
- #define SIZE_OP_MEMORY_END_PUSH_REC (SIZE_OPCODE + SIZE_MEMNUM)
634
- #define SIZE_OP_MEMORY_END (SIZE_OPCODE + SIZE_MEMNUM)
635
- #define SIZE_OP_MEMORY_END_REC (SIZE_OPCODE + SIZE_MEMNUM)
636
- #define SIZE_OP_PUSH_STOP_BT SIZE_OPCODE
637
- #define SIZE_OP_POP_STOP_BT SIZE_OPCODE
638
- #define SIZE_OP_NULL_CHECK_START (SIZE_OPCODE + SIZE_MEMNUM)
639
- #define SIZE_OP_NULL_CHECK_END (SIZE_OPCODE + SIZE_MEMNUM)
640
- #define SIZE_OP_LOOK_BEHIND (SIZE_OPCODE + SIZE_LENGTH)
641
- #define SIZE_OP_PUSH_LOOK_BEHIND_NOT (SIZE_OPCODE + SIZE_RELADDR + SIZE_LENGTH)
642
- #define SIZE_OP_FAIL_LOOK_BEHIND_NOT SIZE_OPCODE
643
- #define SIZE_OP_CALL (SIZE_OPCODE + SIZE_ABSADDR)
644
- #define SIZE_OP_RETURN SIZE_OPCODE
645
-
646
- #ifdef USE_COMBINATION_EXPLOSION_CHECK
647
- #define SIZE_OP_STATE_CHECK (SIZE_OPCODE + SIZE_STATE_CHECK_NUM)
648
- #define SIZE_OP_STATE_CHECK_PUSH (SIZE_OPCODE + SIZE_STATE_CHECK_NUM + SIZE_RELADDR)
649
- #define SIZE_OP_STATE_CHECK_PUSH_OR_JUMP (SIZE_OPCODE + SIZE_STATE_CHECK_NUM + SIZE_RELADDR)
650
- #define SIZE_OP_STATE_CHECK_ANYCHAR_STAR (SIZE_OPCODE + SIZE_STATE_CHECK_NUM)
651
- #endif
652
-
653
- #define MC_ESC(syn) (syn)->meta_char_table.esc
654
- #define MC_ANYCHAR(syn) (syn)->meta_char_table.anychar
655
- #define MC_ANYTIME(syn) (syn)->meta_char_table.anytime
656
- #define MC_ZERO_OR_ONE_TIME(syn) (syn)->meta_char_table.zero_or_one_time
657
- #define MC_ONE_OR_MORE_TIME(syn) (syn)->meta_char_table.one_or_more_time
658
- #define MC_ANYCHAR_ANYTIME(syn) (syn)->meta_char_table.anychar_anytime
659
-
660
- #define IS_MC_ESC_CODE(code, syn) \
661
- ((code) == MC_ESC(syn) && \
662
- !IS_SYNTAX_OP2((syn), ONIG_SYN_OP2_INEFFECTIVE_ESCAPE))
663
-
664
-
665
- #define SYN_POSIX_COMMON_OP \
666
- ( ONIG_SYN_OP_DOT_ANYCHAR | ONIG_SYN_OP_POSIX_BRACKET | \
667
- ONIG_SYN_OP_DECIMAL_BACKREF | \
668
- ONIG_SYN_OP_BRACKET_CC | ONIG_SYN_OP_ASTERISK_ZERO_INF | \
669
- ONIG_SYN_OP_LINE_ANCHOR | \
670
- ONIG_SYN_OP_ESC_CONTROL_CHARS )
671
-
672
- #define SYN_GNU_REGEX_OP \
673
- ( ONIG_SYN_OP_DOT_ANYCHAR | ONIG_SYN_OP_BRACKET_CC | \
674
- ONIG_SYN_OP_POSIX_BRACKET | ONIG_SYN_OP_DECIMAL_BACKREF | \
675
- ONIG_SYN_OP_BRACE_INTERVAL | ONIG_SYN_OP_LPAREN_SUBEXP | \
676
- ONIG_SYN_OP_VBAR_ALT | \
677
- ONIG_SYN_OP_ASTERISK_ZERO_INF | ONIG_SYN_OP_PLUS_ONE_INF | \
678
- ONIG_SYN_OP_QMARK_ZERO_ONE | \
679
- ONIG_SYN_OP_ESC_AZ_BUF_ANCHOR | ONIG_SYN_OP_ESC_CAPITAL_G_BEGIN_ANCHOR | \
680
- ONIG_SYN_OP_ESC_W_WORD | \
681
- ONIG_SYN_OP_ESC_B_WORD_BOUND | ONIG_SYN_OP_ESC_LTGT_WORD_BEGIN_END | \
682
- ONIG_SYN_OP_ESC_S_WHITE_SPACE | ONIG_SYN_OP_ESC_D_DIGIT | \
683
- ONIG_SYN_OP_LINE_ANCHOR )
684
-
685
- #define SYN_GNU_REGEX_BV \
686
- ( ONIG_SYN_CONTEXT_INDEP_ANCHORS | ONIG_SYN_CONTEXT_INDEP_REPEAT_OPS | \
687
- ONIG_SYN_CONTEXT_INVALID_REPEAT_OPS | ONIG_SYN_ALLOW_INVALID_INTERVAL | \
688
- ONIG_SYN_BACKSLASH_ESCAPE_IN_CC | ONIG_SYN_ALLOW_DOUBLE_RANGE_OP_IN_CC )
689
-
690
-
691
- #define NCCLASS_FLAGS(cc) ((cc)->flags)
692
- #define NCCLASS_FLAG_SET(cc,flag) (NCCLASS_FLAGS(cc) |= (flag))
693
- #define NCCLASS_FLAG_CLEAR(cc,flag) (NCCLASS_FLAGS(cc) &= ~(flag))
694
- #define IS_NCCLASS_FLAG_ON(cc,flag) ((NCCLASS_FLAGS(cc) & (flag)) != 0)
695
-
696
- /* cclass node */
697
- #define FLAG_NCCLASS_NOT (1<<0)
698
- #define FLAG_NCCLASS_SHARE (1<<1)
699
-
700
- #define NCCLASS_SET_NOT(nd) NCCLASS_FLAG_SET(nd, FLAG_NCCLASS_NOT)
701
- #define NCCLASS_SET_SHARE(nd) NCCLASS_FLAG_SET(nd, FLAG_NCCLASS_SHARE)
702
- #define NCCLASS_CLEAR_NOT(nd) NCCLASS_FLAG_CLEAR(nd, FLAG_NCCLASS_NOT)
703
- #define IS_NCCLASS_NOT(nd) IS_NCCLASS_FLAG_ON(nd, FLAG_NCCLASS_NOT)
704
- #define IS_NCCLASS_SHARE(nd) IS_NCCLASS_FLAG_ON(nd, FLAG_NCCLASS_SHARE)
705
-
706
- typedef struct {
707
- int type;
708
- /* struct _Node* next; */
709
- /* unsigned int flags; */
710
- } NodeBase;
711
-
712
- typedef struct {
713
- NodeBase base;
714
- unsigned int flags;
715
- BitSet bs;
716
- BBuf* mbuf; /* multi-byte info or NULL */
717
- } CClassNode;
718
-
719
- typedef long OnigStackIndex;
720
-
721
- typedef struct _OnigStackType {
722
- unsigned int type;
723
- union {
724
- struct {
725
- UChar *pcode; /* byte code position */
726
- UChar *pstr; /* string position */
727
- UChar *pstr_prev; /* previous char position of pstr */
728
- #ifdef USE_COMBINATION_EXPLOSION_CHECK
729
- unsigned int state_check;
730
- #endif
731
- } state;
732
- struct {
733
- int count; /* for OP_REPEAT_INC, OP_REPEAT_INC_NG */
734
- UChar *pcode; /* byte code position (head of repeated target) */
735
- int num; /* repeat id */
736
- } repeat;
737
- struct {
738
- OnigStackIndex si; /* index of stack */
739
- } repeat_inc;
740
- struct {
741
- int num; /* memory num */
742
- UChar *pstr; /* start/end position */
743
- /* Following information is setted, if this stack type is MEM-START */
744
- OnigStackIndex start; /* prev. info (for backtrack "(...)*" ) */
745
- OnigStackIndex end; /* prev. info (for backtrack "(...)*" ) */
746
- } mem;
747
- struct {
748
- int num; /* null check id */
749
- UChar *pstr; /* start position */
750
- } null_check;
751
- #ifdef USE_SUBEXP_CALL
752
- struct {
753
- UChar *ret_addr; /* byte code position */
754
- int num; /* null check id */
755
- UChar *pstr; /* string position */
756
- } call_frame;
757
- #endif
758
- } u;
759
- } OnigStackType;
760
-
761
- typedef struct {
762
- void* stack_p;
763
- int stack_n;
764
- OnigOptionType options;
765
- OnigRegion* region;
766
- const UChar* start; /* search start position (for \G: BEGIN_POSITION) */
767
- #ifdef USE_FIND_LONGEST_SEARCH_ALL_OF_RANGE
768
- int best_len; /* for ONIG_OPTION_FIND_LONGEST */
769
- UChar* best_s;
770
- #endif
771
- #ifdef USE_COMBINATION_EXPLOSION_CHECK
772
- void* state_check_buff;
773
- int state_check_buff_size;
774
- #endif
775
- } OnigMatchArg;
776
-
777
-
778
- #define IS_CODE_SB_WORD(enc,code) \
779
- (ONIGENC_IS_CODE_ASCII(code) && ONIGENC_IS_CODE_WORD(enc,code))
780
-
781
- #ifdef ONIG_DEBUG
782
-
783
- typedef struct {
784
- short int opcode;
785
- char* name;
786
- short int arg_type;
787
- } OnigOpInfoType;
788
-
789
- extern OnigOpInfoType OnigOpInfo[];
790
-
791
- extern void onig_print_compiled_byte_code P_((FILE* f, UChar* bp, UChar** nextp, OnigEncoding enc));
792
-
793
- #ifdef ONIG_DEBUG_STATISTICS
794
- extern void onig_statistics_init P_((void));
795
- extern void onig_print_statistics P_((FILE* f));
796
- #endif
797
- #endif
798
-
799
- extern UChar* onig_error_code_to_format P_((int code));
800
- extern void onig_snprintf_with_pattern PV_((UChar buf[], int bufsize, OnigEncoding enc, UChar* pat, UChar* pat_end, const UChar *fmt, ...));
801
- extern int onig_bbuf_init P_((BBuf* buf, int size));
802
- extern int onig_alloc_init P_((regex_t** reg, OnigOptionType option, OnigCaseFoldType case_fold_flag, OnigEncoding enc, const OnigSyntaxType* syntax));
803
- extern int onig_compile P_((regex_t* reg, const UChar* pattern, const UChar* pattern_end, OnigErrorInfo* einfo));
804
- extern void onig_chain_reduce P_((regex_t* reg));
805
- extern void onig_chain_link_add P_((regex_t* to, regex_t* add));
806
- extern void onig_transfer P_((regex_t* to, regex_t* from));
807
- extern int onig_is_code_in_cc P_((OnigEncoding enc, OnigCodePoint code, CClassNode* cc));
808
- extern int onig_is_code_in_cc_len P_((int enclen, OnigCodePoint code, CClassNode* cc));
809
-
810
- /* strend hash */
811
- typedef void hash_table_type;
812
- #ifdef RUBY
813
- #include "ruby/st.h"
814
- typedef st_data_t hash_data_type;
815
- #else
816
- typedef unsigned long hash_data_type;
817
- #endif
818
-
819
- extern hash_table_type* onig_st_init_strend_table_with_size P_((int size));
820
- extern int onig_st_lookup_strend P_((hash_table_type* table, const UChar* str_key, const UChar* end_key, hash_data_type *value));
821
- extern int onig_st_insert_strend P_((hash_table_type* table, const UChar* str_key, const UChar* end_key, hash_data_type value));
822
-
823
- /* encoding property management */
824
- #define PROPERTY_LIST_ADD_PROP(Name, CR) \
825
- r = onigenc_property_list_add_property((UChar* )Name, CR,\
826
- &PropertyNameTable, &PropertyList, &PropertyListNum,\
827
- &PropertyListSize);\
828
- if (r != 0) goto end
829
-
830
- #define PROPERTY_LIST_INIT_CHECK \
831
- if (PropertyInited == 0) {\
832
- int r = onigenc_property_list_init(init_property_list);\
833
- if (r != 0) return r;\
834
- }
835
-
836
- extern int onigenc_property_list_add_property P_((UChar* name, const OnigCodePoint* prop, hash_table_type **table, const OnigCodePoint*** plist, int *pnum, int *psize));
837
-
838
- typedef int (*ONIGENC_INIT_PROPERTY_LIST_FUNC_TYPE)(void);
839
-
840
- extern int onigenc_property_list_init P_((ONIGENC_INIT_PROPERTY_LIST_FUNC_TYPE));
841
-
842
- #endif /* ONIGURUMA_REGINT_H */