3
/**********************************************************************
4
oniguruma.h - Oniguruma (regular expression library)
5
**********************************************************************/
7
* Copyright (c) 2002-2007 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
35
} /* satisfy cc-mode */
40
#define ONIGURUMA_VERSION_MAJOR 5
41
#define ONIGURUMA_VERSION_MINOR 9
42
#define ONIGURUMA_VERSION_TEENY 0
45
# ifndef HAVE_PROTOTYPES
46
# define HAVE_PROTOTYPES 1
48
# ifndef HAVE_STDARG_PROTOTYPES
49
# define HAVE_STDARG_PROTOTYPES 1
53
/* escape Mac OS X/Xcode 2.4/gcc 4.0.1 problem */
54
#if defined(__APPLE__) && defined(__GNUC__) && __GNUC__ >= 4
55
# ifndef HAVE_STDARG_PROTOTYPES
56
# define HAVE_STDARG_PROTOTYPES 1
61
#if defined(__STDC__) || defined(_WIN32)
62
# define P_(args) args
69
#ifdef HAVE_STDARG_PROTOTYPES
70
# define PV_(args) args
78
#define ONIG_EXTERN RUBY_EXTERN
80
#if defined(_WIN32) && !defined(__GNUC__)
81
#if defined(EXPORT) || defined(RUBY_EXPORT)
82
#define ONIG_EXTERN extern __declspec(dllexport)
84
#define ONIG_EXTERN extern __declspec(dllimport)
91
#define ONIG_EXTERN extern
94
/* PART: character encoding */
96
#ifndef ONIG_ESCAPE_UCHAR_COLLISION
97
#define UChar OnigUChar
100
typedef unsigned char OnigUChar;
101
typedef unsigned long OnigCodePoint;
102
typedef unsigned int OnigDistance;
104
#define ONIG_INFINITE_DISTANCE ~((OnigDistance )0)
107
typedef unsigned int OnigCaseFoldType;
109
ONIG_EXTERN OnigCaseFoldType OnigDefaultCaseFoldFlag;
111
/* #define ONIGENC_CASE_FOLD_HIRAGANA_KATAKANA (1<<1) */
112
/* #define ONIGENC_CASE_FOLD_KATAKANA_WIDTH (1<<2) */
113
#define ONIGENC_CASE_FOLD_TURKISH_AZERI (1<<20)
114
#define INTERNAL_ONIGENC_CASE_FOLD_MULTI_CHAR (1<<30)
116
#define ONIGENC_CASE_FOLD_MIN INTERNAL_ONIGENC_CASE_FOLD_MULTI_CHAR
117
#define ONIGENC_CASE_FOLD_DEFAULT OnigDefaultCaseFoldFlag
120
#define ONIGENC_MAX_COMP_CASE_FOLD_CODE_LEN 3
121
#define ONIGENC_GET_CASE_FOLD_CODES_MAX_NUM 13
122
/* 13 => Unicode:0x1ffc */
125
#define ONIGENC_CODE_RANGE_NUM(range) ((int )range[0])
126
#define ONIGENC_CODE_RANGE_FROM(range,i) range[((i)*2) + 1]
127
#define ONIGENC_CODE_RANGE_TO(range,i) range[((i)*2) + 2]
130
int byte_len; /* argument(original) character(s) byte length */
131
int code_len; /* number of code */
132
OnigCodePoint code[ONIGENC_MAX_COMP_CASE_FOLD_CODE_LEN];
133
} OnigCaseFoldCodeItem;
137
OnigCodePoint anychar;
138
OnigCodePoint anytime;
139
OnigCodePoint zero_or_one_time;
140
OnigCodePoint one_or_more_time;
141
OnigCodePoint anychar_anytime;
142
} OnigMetaCharTableType;
144
typedef int (*OnigApplyAllCaseFoldFunc)(OnigCodePoint from, OnigCodePoint* to, int to_len, void* arg);
146
typedef struct OnigEncodingTypeST {
147
int (*mbc_enc_len)(const OnigUChar* p);
151
int (*is_mbc_newline)(const OnigUChar* p, const OnigUChar* end);
152
OnigCodePoint (*mbc_to_code)(const OnigUChar* p, const OnigUChar* end);
153
int (*code_to_mbclen)(OnigCodePoint code);
154
int (*code_to_mbc)(OnigCodePoint code, OnigUChar *buf);
155
int (*mbc_case_fold)(OnigCaseFoldType flag, const OnigUChar** pp, const OnigUChar* end, OnigUChar* to);
156
int (*apply_all_case_fold)(OnigCaseFoldType flag, OnigApplyAllCaseFoldFunc f, void* arg);
157
int (*get_case_fold_codes_by_str)(OnigCaseFoldType flag, const OnigUChar* p, const OnigUChar* end, OnigCaseFoldCodeItem acs[]);
158
int (*property_name_to_ctype)(struct OnigEncodingTypeST* enc, OnigUChar* p, OnigUChar* end);
159
int (*is_code_ctype)(OnigCodePoint code, unsigned int ctype);
160
int (*get_ctype_code_range)(int ctype, OnigCodePoint* sb_out, const OnigCodePoint* ranges[]);
161
OnigUChar* (*left_adjust_char_head)(const OnigUChar* start, const OnigUChar* p);
162
int (*is_allowed_reverse_match)(const OnigUChar* p, const OnigUChar* end);
165
typedef OnigEncodingType* OnigEncoding;
167
ONIG_EXTERN OnigEncodingType OnigEncodingASCII;
168
ONIG_EXTERN OnigEncodingType OnigEncodingISO_8859_1;
169
ONIG_EXTERN OnigEncodingType OnigEncodingISO_8859_2;
170
ONIG_EXTERN OnigEncodingType OnigEncodingISO_8859_3;
171
ONIG_EXTERN OnigEncodingType OnigEncodingISO_8859_4;
172
ONIG_EXTERN OnigEncodingType OnigEncodingISO_8859_5;
173
ONIG_EXTERN OnigEncodingType OnigEncodingISO_8859_6;
174
ONIG_EXTERN OnigEncodingType OnigEncodingISO_8859_7;
175
ONIG_EXTERN OnigEncodingType OnigEncodingISO_8859_8;
176
ONIG_EXTERN OnigEncodingType OnigEncodingISO_8859_9;
177
ONIG_EXTERN OnigEncodingType OnigEncodingISO_8859_10;
178
ONIG_EXTERN OnigEncodingType OnigEncodingISO_8859_11;
179
ONIG_EXTERN OnigEncodingType OnigEncodingISO_8859_13;
180
ONIG_EXTERN OnigEncodingType OnigEncodingISO_8859_14;
181
ONIG_EXTERN OnigEncodingType OnigEncodingISO_8859_15;
182
ONIG_EXTERN OnigEncodingType OnigEncodingISO_8859_16;
183
ONIG_EXTERN OnigEncodingType OnigEncodingUTF8;
184
ONIG_EXTERN OnigEncodingType OnigEncodingUTF16_BE;
185
ONIG_EXTERN OnigEncodingType OnigEncodingUTF16_LE;
186
ONIG_EXTERN OnigEncodingType OnigEncodingUTF32_BE;
187
ONIG_EXTERN OnigEncodingType OnigEncodingUTF32_LE;
188
ONIG_EXTERN OnigEncodingType OnigEncodingEUC_JP;
189
ONIG_EXTERN OnigEncodingType OnigEncodingEUC_TW;
190
ONIG_EXTERN OnigEncodingType OnigEncodingEUC_KR;
191
ONIG_EXTERN OnigEncodingType OnigEncodingEUC_CN;
192
ONIG_EXTERN OnigEncodingType OnigEncodingSJIS;
193
ONIG_EXTERN OnigEncodingType OnigEncodingKOI8;
194
ONIG_EXTERN OnigEncodingType OnigEncodingKOI8_R;
195
ONIG_EXTERN OnigEncodingType OnigEncodingCP1251;
196
ONIG_EXTERN OnigEncodingType OnigEncodingBIG5;
197
ONIG_EXTERN OnigEncodingType OnigEncodingGB18030;
199
#define ONIG_ENCODING_ASCII (&OnigEncodingASCII)
200
#define ONIG_ENCODING_ISO_8859_1 (&OnigEncodingISO_8859_1)
201
#define ONIG_ENCODING_ISO_8859_2 (&OnigEncodingISO_8859_2)
202
#define ONIG_ENCODING_ISO_8859_3 (&OnigEncodingISO_8859_3)
203
#define ONIG_ENCODING_ISO_8859_4 (&OnigEncodingISO_8859_4)
204
#define ONIG_ENCODING_ISO_8859_5 (&OnigEncodingISO_8859_5)
205
#define ONIG_ENCODING_ISO_8859_6 (&OnigEncodingISO_8859_6)
206
#define ONIG_ENCODING_ISO_8859_7 (&OnigEncodingISO_8859_7)
207
#define ONIG_ENCODING_ISO_8859_8 (&OnigEncodingISO_8859_8)
208
#define ONIG_ENCODING_ISO_8859_9 (&OnigEncodingISO_8859_9)
209
#define ONIG_ENCODING_ISO_8859_10 (&OnigEncodingISO_8859_10)
210
#define ONIG_ENCODING_ISO_8859_11 (&OnigEncodingISO_8859_11)
211
#define ONIG_ENCODING_ISO_8859_13 (&OnigEncodingISO_8859_13)
212
#define ONIG_ENCODING_ISO_8859_14 (&OnigEncodingISO_8859_14)
213
#define ONIG_ENCODING_ISO_8859_15 (&OnigEncodingISO_8859_15)
214
#define ONIG_ENCODING_ISO_8859_16 (&OnigEncodingISO_8859_16)
215
#define ONIG_ENCODING_UTF8 (&OnigEncodingUTF8)
216
#define ONIG_ENCODING_UTF16_BE (&OnigEncodingUTF16_BE)
217
#define ONIG_ENCODING_UTF16_LE (&OnigEncodingUTF16_LE)
218
#define ONIG_ENCODING_UTF32_BE (&OnigEncodingUTF32_BE)
219
#define ONIG_ENCODING_UTF32_LE (&OnigEncodingUTF32_LE)
220
#define ONIG_ENCODING_EUC_JP (&OnigEncodingEUC_JP)
221
#define ONIG_ENCODING_EUC_TW (&OnigEncodingEUC_TW)
222
#define ONIG_ENCODING_EUC_KR (&OnigEncodingEUC_KR)
223
#define ONIG_ENCODING_EUC_CN (&OnigEncodingEUC_CN)
224
#define ONIG_ENCODING_SJIS (&OnigEncodingSJIS)
225
#define ONIG_ENCODING_KOI8 (&OnigEncodingKOI8)
226
#define ONIG_ENCODING_KOI8_R (&OnigEncodingKOI8_R)
227
#define ONIG_ENCODING_CP1251 (&OnigEncodingCP1251)
228
#define ONIG_ENCODING_BIG5 (&OnigEncodingBIG5)
229
#define ONIG_ENCODING_GB18030 (&OnigEncodingGB18030)
231
#define ONIG_ENCODING_UNDEF ((OnigEncoding )0)
235
#define ONIGENC_CODE_TO_MBC_MAXLEN 7
236
#define ONIGENC_MBC_CASE_FOLD_MAXLEN 18
237
/* 18: 6(max-byte) * 3(case-fold chars) */
239
/* character types */
240
#define ONIGENC_CTYPE_NEWLINE 0
241
#define ONIGENC_CTYPE_ALPHA 1
242
#define ONIGENC_CTYPE_BLANK 2
243
#define ONIGENC_CTYPE_CNTRL 3
244
#define ONIGENC_CTYPE_DIGIT 4
245
#define ONIGENC_CTYPE_GRAPH 5
246
#define ONIGENC_CTYPE_LOWER 6
247
#define ONIGENC_CTYPE_PRINT 7
248
#define ONIGENC_CTYPE_PUNCT 8
249
#define ONIGENC_CTYPE_SPACE 9
250
#define ONIGENC_CTYPE_UPPER 10
251
#define ONIGENC_CTYPE_XDIGIT 11
252
#define ONIGENC_CTYPE_WORD 12
253
#define ONIGENC_CTYPE_ALNUM 13 /* alpha || digit */
254
#define ONIGENC_CTYPE_ASCII 14
255
#define ONIGENC_MAX_STD_CTYPE ONIGENC_CTYPE_ASCII
258
#define enc_len(enc,p) ONIGENC_MBC_ENC_LEN(enc, p)
260
#define ONIGENC_IS_UNDEF(enc) ((enc) == ONIG_ENCODING_UNDEF)
261
#define ONIGENC_IS_SINGLEBYTE(enc) (ONIGENC_MBC_MAXLEN(enc) == 1)
262
#define ONIGENC_IS_MBC_HEAD(enc,p) (ONIGENC_MBC_ENC_LEN(enc,p) != 1)
263
#define ONIGENC_IS_MBC_ASCII(p) (*(p) < 128)
264
#define ONIGENC_IS_CODE_ASCII(code) ((code) < 128)
265
#define ONIGENC_IS_MBC_WORD(enc,s,end) \
266
ONIGENC_IS_CODE_WORD(enc,ONIGENC_MBC_TO_CODE(enc,s,end))
269
#define ONIGENC_NAME(enc) ((enc)->name)
271
#define ONIGENC_MBC_CASE_FOLD(enc,flag,pp,end,buf) \
272
(enc)->mbc_case_fold(flag,(const OnigUChar** )pp,end,buf)
273
#define ONIGENC_IS_ALLOWED_REVERSE_MATCH(enc,s,end) \
274
(enc)->is_allowed_reverse_match(s,end)
275
#define ONIGENC_LEFT_ADJUST_CHAR_HEAD(enc,start,s) \
276
(enc)->left_adjust_char_head(start, s)
277
#define ONIGENC_APPLY_ALL_CASE_FOLD(enc,case_fold_flag,f,arg) \
278
(enc)->apply_all_case_fold(case_fold_flag,f,arg)
279
#define ONIGENC_GET_CASE_FOLD_CODES_BY_STR(enc,case_fold_flag,p,end,acs) \
280
(enc)->get_case_fold_codes_by_str(case_fold_flag,p,end,acs)
281
#define ONIGENC_STEP_BACK(enc,start,s,n) \
282
onigenc_step_back((enc),(start),(s),(n))
284
#define ONIGENC_MBC_ENC_LEN(enc,p) (enc)->mbc_enc_len(p)
285
#define ONIGENC_MBC_MAXLEN(enc) ((enc)->max_enc_len)
286
#define ONIGENC_MBC_MAXLEN_DIST(enc) ONIGENC_MBC_MAXLEN(enc)
287
#define ONIGENC_MBC_MINLEN(enc) ((enc)->min_enc_len)
288
#define ONIGENC_IS_MBC_NEWLINE(enc,p,end) (enc)->is_mbc_newline((p),(end))
289
#define ONIGENC_MBC_TO_CODE(enc,p,end) (enc)->mbc_to_code((p),(end))
290
#define ONIGENC_CODE_TO_MBCLEN(enc,code) (enc)->code_to_mbclen(code)
291
#define ONIGENC_CODE_TO_MBC(enc,code,buf) (enc)->code_to_mbc(code,buf)
292
#define ONIGENC_PROPERTY_NAME_TO_CTYPE(enc,p,end) \
293
(enc)->property_name_to_ctype(enc,p,end)
295
#define ONIGENC_IS_CODE_CTYPE(enc,code,ctype) (enc)->is_code_ctype(code,ctype)
297
#define ONIGENC_IS_CODE_NEWLINE(enc,code) \
298
ONIGENC_IS_CODE_CTYPE(enc,code,ONIGENC_CTYPE_NEWLINE)
299
#define ONIGENC_IS_CODE_GRAPH(enc,code) \
300
ONIGENC_IS_CODE_CTYPE(enc,code,ONIGENC_CTYPE_GRAPH)
301
#define ONIGENC_IS_CODE_PRINT(enc,code) \
302
ONIGENC_IS_CODE_CTYPE(enc,code,ONIGENC_CTYPE_PRINT)
303
#define ONIGENC_IS_CODE_ALNUM(enc,code) \
304
ONIGENC_IS_CODE_CTYPE(enc,code,ONIGENC_CTYPE_ALNUM)
305
#define ONIGENC_IS_CODE_ALPHA(enc,code) \
306
ONIGENC_IS_CODE_CTYPE(enc,code,ONIGENC_CTYPE_ALPHA)
307
#define ONIGENC_IS_CODE_LOWER(enc,code) \
308
ONIGENC_IS_CODE_CTYPE(enc,code,ONIGENC_CTYPE_LOWER)
309
#define ONIGENC_IS_CODE_UPPER(enc,code) \
310
ONIGENC_IS_CODE_CTYPE(enc,code,ONIGENC_CTYPE_UPPER)
311
#define ONIGENC_IS_CODE_CNTRL(enc,code) \
312
ONIGENC_IS_CODE_CTYPE(enc,code,ONIGENC_CTYPE_CNTRL)
313
#define ONIGENC_IS_CODE_PUNCT(enc,code) \
314
ONIGENC_IS_CODE_CTYPE(enc,code,ONIGENC_CTYPE_PUNCT)
315
#define ONIGENC_IS_CODE_SPACE(enc,code) \
316
ONIGENC_IS_CODE_CTYPE(enc,code,ONIGENC_CTYPE_SPACE)
317
#define ONIGENC_IS_CODE_BLANK(enc,code) \
318
ONIGENC_IS_CODE_CTYPE(enc,code,ONIGENC_CTYPE_BLANK)
319
#define ONIGENC_IS_CODE_DIGIT(enc,code) \
320
ONIGENC_IS_CODE_CTYPE(enc,code,ONIGENC_CTYPE_DIGIT)
321
#define ONIGENC_IS_CODE_XDIGIT(enc,code) \
322
ONIGENC_IS_CODE_CTYPE(enc,code,ONIGENC_CTYPE_XDIGIT)
323
#define ONIGENC_IS_CODE_WORD(enc,code) \
324
ONIGENC_IS_CODE_CTYPE(enc,code,ONIGENC_CTYPE_WORD)
326
#define ONIGENC_GET_CTYPE_CODE_RANGE(enc,ctype,sbout,ranges) \
327
(enc)->get_ctype_code_range(ctype,sbout,ranges)
330
OnigUChar* onigenc_step_back P_((OnigEncoding enc, const OnigUChar* start, const OnigUChar* s, int n));
335
int onigenc_init P_((void));
337
int onigenc_set_default_encoding P_((OnigEncoding enc));
339
OnigEncoding onigenc_get_default_encoding P_((void));
341
void onigenc_set_default_caseconv_table P_((const OnigUChar* table));
343
OnigUChar* onigenc_get_right_adjust_char_head_with_prev P_((OnigEncoding enc, const OnigUChar* start, const OnigUChar* s, const OnigUChar** prev));
345
OnigUChar* onigenc_get_prev_char_head P_((OnigEncoding enc, const OnigUChar* start, const OnigUChar* s));
347
OnigUChar* onigenc_get_left_adjust_char_head P_((OnigEncoding enc, const OnigUChar* start, const OnigUChar* s));
349
OnigUChar* onigenc_get_right_adjust_char_head P_((OnigEncoding enc, const OnigUChar* start, const OnigUChar* s));
351
int onigenc_strlen P_((OnigEncoding enc, const OnigUChar* p, const OnigUChar* end));
353
int onigenc_strlen_null P_((OnigEncoding enc, const OnigUChar* p));
355
int onigenc_str_bytelen_null P_((OnigEncoding enc, const OnigUChar* p));
359
/* PART: regular expression */
361
/* config parameters */
362
#define ONIG_NREGION 10
363
#define ONIG_MAX_BACKREF_NUM 1000
364
#define ONIG_MAX_REPEAT_NUM 100000
365
#define ONIG_MAX_MULTI_BYTE_RANGES_NUM 10000
367
#define ONIG_MAX_ERROR_MESSAGE_LEN 90
369
typedef unsigned int OnigOptionType;
371
#define ONIG_OPTION_DEFAULT ONIG_OPTION_NONE
374
#define ONIG_OPTION_NONE 0U
375
#define ONIG_OPTION_IGNORECASE 1U
376
#define ONIG_OPTION_EXTEND (ONIG_OPTION_IGNORECASE << 1)
377
#define ONIG_OPTION_MULTILINE (ONIG_OPTION_EXTEND << 1)
378
#define ONIG_OPTION_SINGLELINE (ONIG_OPTION_MULTILINE << 1)
379
#define ONIG_OPTION_FIND_LONGEST (ONIG_OPTION_SINGLELINE << 1)
380
#define ONIG_OPTION_FIND_NOT_EMPTY (ONIG_OPTION_FIND_LONGEST << 1)
381
#define ONIG_OPTION_NEGATE_SINGLELINE (ONIG_OPTION_FIND_NOT_EMPTY << 1)
382
#define ONIG_OPTION_DONT_CAPTURE_GROUP (ONIG_OPTION_NEGATE_SINGLELINE << 1)
383
#define ONIG_OPTION_CAPTURE_GROUP (ONIG_OPTION_DONT_CAPTURE_GROUP << 1)
384
/* options (search time) */
385
#define ONIG_OPTION_NOTBOL (ONIG_OPTION_CAPTURE_GROUP << 1)
386
#define ONIG_OPTION_NOTEOL (ONIG_OPTION_NOTBOL << 1)
387
#define ONIG_OPTION_POSIX_REGION (ONIG_OPTION_NOTEOL << 1)
388
#define ONIG_OPTION_MAXBIT ONIG_OPTION_POSIX_REGION /* limit */
390
#define ONIG_OPTION_ON(options,regopt) ((options) |= (regopt))
391
#define ONIG_OPTION_OFF(options,regopt) ((options) &= ~(regopt))
392
#define ONIG_IS_OPTION_ON(options,option) ((options) & (option))
398
unsigned int behavior;
399
OnigOptionType options; /* default option */
400
OnigMetaCharTableType meta_char_table;
403
ONIG_EXTERN OnigSyntaxType OnigSyntaxASIS;
404
ONIG_EXTERN OnigSyntaxType OnigSyntaxPosixBasic;
405
ONIG_EXTERN OnigSyntaxType OnigSyntaxPosixExtended;
406
ONIG_EXTERN OnigSyntaxType OnigSyntaxEmacs;
407
ONIG_EXTERN OnigSyntaxType OnigSyntaxGrep;
408
ONIG_EXTERN OnigSyntaxType OnigSyntaxGnuRegex;
409
ONIG_EXTERN OnigSyntaxType OnigSyntaxJava;
410
ONIG_EXTERN OnigSyntaxType OnigSyntaxPerl;
411
ONIG_EXTERN OnigSyntaxType OnigSyntaxPerl_NG;
412
ONIG_EXTERN OnigSyntaxType OnigSyntaxRuby;
414
/* predefined syntaxes (see regsyntax.c) */
415
#define ONIG_SYNTAX_ASIS (&OnigSyntaxASIS)
416
#define ONIG_SYNTAX_POSIX_BASIC (&OnigSyntaxPosixBasic)
417
#define ONIG_SYNTAX_POSIX_EXTENDED (&OnigSyntaxPosixExtended)
418
#define ONIG_SYNTAX_EMACS (&OnigSyntaxEmacs)
419
#define ONIG_SYNTAX_GREP (&OnigSyntaxGrep)
420
#define ONIG_SYNTAX_GNU_REGEX (&OnigSyntaxGnuRegex)
421
#define ONIG_SYNTAX_JAVA (&OnigSyntaxJava)
422
#define ONIG_SYNTAX_PERL (&OnigSyntaxPerl)
423
#define ONIG_SYNTAX_PERL_NG (&OnigSyntaxPerl_NG)
424
#define ONIG_SYNTAX_RUBY (&OnigSyntaxRuby)
427
ONIG_EXTERN OnigSyntaxType* OnigDefaultSyntax;
428
#define ONIG_SYNTAX_DEFAULT OnigDefaultSyntax
430
/* syntax (operators) */
431
#define ONIG_SYN_OP_VARIABLE_META_CHARACTERS (1U<<0)
432
#define ONIG_SYN_OP_DOT_ANYCHAR (1U<<1) /* . */
433
#define ONIG_SYN_OP_ASTERISK_ZERO_INF (1U<<2) /* * */
434
#define ONIG_SYN_OP_ESC_ASTERISK_ZERO_INF (1U<<3)
435
#define ONIG_SYN_OP_PLUS_ONE_INF (1U<<4) /* + */
436
#define ONIG_SYN_OP_ESC_PLUS_ONE_INF (1U<<5)
437
#define ONIG_SYN_OP_QMARK_ZERO_ONE (1U<<6) /* ? */
438
#define ONIG_SYN_OP_ESC_QMARK_ZERO_ONE (1U<<7)
439
#define ONIG_SYN_OP_BRACE_INTERVAL (1U<<8) /* {lower,upper} */
440
#define ONIG_SYN_OP_ESC_BRACE_INTERVAL (1U<<9) /* \{lower,upper\} */
441
#define ONIG_SYN_OP_VBAR_ALT (1U<<10) /* | */
442
#define ONIG_SYN_OP_ESC_VBAR_ALT (1U<<11) /* \| */
443
#define ONIG_SYN_OP_LPAREN_SUBEXP (1U<<12) /* (...) */
444
#define ONIG_SYN_OP_ESC_LPAREN_SUBEXP (1U<<13) /* \(...\) */
445
#define ONIG_SYN_OP_ESC_AZ_BUF_ANCHOR (1U<<14) /* \A, \Z, \z */
446
#define ONIG_SYN_OP_ESC_CAPITAL_G_BEGIN_ANCHOR (1U<<15) /* \G */
447
#define ONIG_SYN_OP_DECIMAL_BACKREF (1U<<16) /* \num */
448
#define ONIG_SYN_OP_BRACKET_CC (1U<<17) /* [...] */
449
#define ONIG_SYN_OP_ESC_W_WORD (1U<<18) /* \w, \W */
450
#define ONIG_SYN_OP_ESC_LTGT_WORD_BEGIN_END (1U<<19) /* \<. \> */
451
#define ONIG_SYN_OP_ESC_B_WORD_BOUND (1U<<20) /* \b, \B */
452
#define ONIG_SYN_OP_ESC_S_WHITE_SPACE (1U<<21) /* \s, \S */
453
#define ONIG_SYN_OP_ESC_D_DIGIT (1U<<22) /* \d, \D */
454
#define ONIG_SYN_OP_LINE_ANCHOR (1U<<23) /* ^, $ */
455
#define ONIG_SYN_OP_POSIX_BRACKET (1U<<24) /* [:xxxx:] */
456
#define ONIG_SYN_OP_QMARK_NON_GREEDY (1U<<25) /* ??,*?,+?,{n,m}? */
457
#define ONIG_SYN_OP_ESC_CONTROL_CHARS (1U<<26) /* \n,\r,\t,\a ... */
458
#define ONIG_SYN_OP_ESC_C_CONTROL (1U<<27) /* \cx */
459
#define ONIG_SYN_OP_ESC_OCTAL3 (1U<<28) /* \OOO */
460
#define ONIG_SYN_OP_ESC_X_HEX2 (1U<<29) /* \xHH */
461
#define ONIG_SYN_OP_ESC_X_BRACE_HEX8 (1U<<30) /* \x{7HHHHHHH} */
463
#define ONIG_SYN_OP2_ESC_CAPITAL_Q_QUOTE (1U<<0) /* \Q...\E */
464
#define ONIG_SYN_OP2_QMARK_GROUP_EFFECT (1U<<1) /* (?...) */
465
#define ONIG_SYN_OP2_OPTION_PERL (1U<<2) /* (?imsx),(?-imsx) */
466
#define ONIG_SYN_OP2_OPTION_RUBY (1U<<3) /* (?imx), (?-imx) */
467
#define ONIG_SYN_OP2_PLUS_POSSESSIVE_REPEAT (1U<<4) /* ?+,*+,++ */
468
#define ONIG_SYN_OP2_PLUS_POSSESSIVE_INTERVAL (1U<<5) /* {n,m}+ */
469
#define ONIG_SYN_OP2_CCLASS_SET_OP (1U<<6) /* [...&&..[..]..] */
470
#define ONIG_SYN_OP2_QMARK_LT_NAMED_GROUP (1U<<7) /* (?<name>...) */
471
#define ONIG_SYN_OP2_ESC_K_NAMED_BACKREF (1U<<8) /* \k<name> */
472
#define ONIG_SYN_OP2_ESC_G_SUBEXP_CALL (1U<<9) /* \g<name>, \g<n> */
473
#define ONIG_SYN_OP2_ATMARK_CAPTURE_HISTORY (1U<<10) /* (?@..),(?@<x>..) */
474
#define ONIG_SYN_OP2_ESC_CAPITAL_C_BAR_CONTROL (1U<<11) /* \C-x */
475
#define ONIG_SYN_OP2_ESC_CAPITAL_M_BAR_META (1U<<12) /* \M-x */
476
#define ONIG_SYN_OP2_ESC_V_VTAB (1U<<13) /* \v as VTAB */
477
#define ONIG_SYN_OP2_ESC_U_HEX4 (1U<<14) /* \uHHHH */
478
#define ONIG_SYN_OP2_ESC_GNU_BUF_ANCHOR (1U<<15) /* \`, \' */
479
#define ONIG_SYN_OP2_ESC_P_BRACE_CHAR_PROPERTY (1U<<16) /* \p{...}, \P{...} */
480
#define ONIG_SYN_OP2_ESC_P_BRACE_CIRCUMFLEX_NOT (1U<<17) /* \p{^..}, \P{^..} */
481
/* #define ONIG_SYN_OP2_CHAR_PROPERTY_PREFIX_IS (1U<<18) */
482
#define ONIG_SYN_OP2_ESC_H_XDIGIT (1U<<19) /* \h, \H */
483
#define ONIG_SYN_OP2_INEFFECTIVE_ESCAPE (1U<<20) /* \ */
485
/* syntax (behavior) */
486
#define ONIG_SYN_CONTEXT_INDEP_ANCHORS (1U<<31) /* not implemented */
487
#define ONIG_SYN_CONTEXT_INDEP_REPEAT_OPS (1U<<0) /* ?, *, +, {n,m} */
488
#define ONIG_SYN_CONTEXT_INVALID_REPEAT_OPS (1U<<1) /* error or ignore */
489
#define ONIG_SYN_ALLOW_UNMATCHED_CLOSE_SUBEXP (1U<<2) /* ...)... */
490
#define ONIG_SYN_ALLOW_INVALID_INTERVAL (1U<<3) /* {??? */
491
#define ONIG_SYN_ALLOW_INTERVAL_LOW_ABBREV (1U<<4) /* {,n} => {0,n} */
492
#define ONIG_SYN_STRICT_CHECK_BACKREF (1U<<5) /* /(\1)/,/\1()/ ..*/
493
#define ONIG_SYN_DIFFERENT_LEN_ALT_LOOK_BEHIND (1U<<6) /* (?<=a|bc) */
494
#define ONIG_SYN_CAPTURE_ONLY_NAMED_GROUP (1U<<7) /* see doc/RE */
495
#define ONIG_SYN_ALLOW_MULTIPLEX_DEFINITION_NAME (1U<<8) /* (?<x>)(?<x>) */
496
#define ONIG_SYN_FIXED_INTERVAL_IS_GREEDY_ONLY (1U<<9) /* a{n}?=(?:a{n})? */
498
/* syntax (behavior) in char class [...] */
499
#define ONIG_SYN_NOT_NEWLINE_IN_NEGATIVE_CC (1U<<20) /* [^...] */
500
#define ONIG_SYN_BACKSLASH_ESCAPE_IN_CC (1U<<21) /* [..\w..] etc.. */
501
#define ONIG_SYN_ALLOW_EMPTY_RANGE_IN_CC (1U<<22)
502
#define ONIG_SYN_ALLOW_DOUBLE_RANGE_OP_IN_CC (1U<<23) /* [0-9-a]=[0-9\-a] */
503
/* syntax (behavior) warning */
504
#define ONIG_SYN_WARN_CC_OP_NOT_ESCAPED (1U<<24) /* [,-,] */
505
#define ONIG_SYN_WARN_REDUNDANT_NESTED_REPEAT (1U<<25) /* (?:a*)+ */
507
/* meta character specifiers (onig_set_meta_char()) */
508
#define ONIG_META_CHAR_ESCAPE 0
509
#define ONIG_META_CHAR_ANYCHAR 1
510
#define ONIG_META_CHAR_ANYTIME 2
511
#define ONIG_META_CHAR_ZERO_OR_ONE_TIME 3
512
#define ONIG_META_CHAR_ONE_OR_MORE_TIME 4
513
#define ONIG_META_CHAR_ANYCHAR_ANYTIME 5
515
#define ONIG_INEFFECTIVE_META_CHAR 0
518
#define ONIG_IS_PATTERN_ERROR(ecode) ((ecode) <= -100 && (ecode) > -1000)
520
#define ONIG_NORMAL 0
521
#define ONIG_MISMATCH -1
522
#define ONIG_NO_SUPPORT_CONFIG -2
525
#define ONIGERR_MEMORY -5
526
#define ONIGERR_TYPE_BUG -6
527
#define ONIGERR_PARSER_BUG -11
528
#define ONIGERR_STACK_BUG -12
529
#define ONIGERR_UNDEFINED_BYTECODE -13
530
#define ONIGERR_UNEXPECTED_BYTECODE -14
531
#define ONIGERR_MATCH_STACK_LIMIT_OVER -15
532
#define ONIGERR_DEFAULT_ENCODING_IS_NOT_SETTED -21
533
#define ONIGERR_SPECIFIED_ENCODING_CANT_CONVERT_TO_WIDE_CHAR -22
535
#define ONIGERR_INVALID_ARGUMENT -30
537
#define ONIGERR_END_PATTERN_AT_LEFT_BRACE -100
538
#define ONIGERR_END_PATTERN_AT_LEFT_BRACKET -101
539
#define ONIGERR_EMPTY_CHAR_CLASS -102
540
#define ONIGERR_PREMATURE_END_OF_CHAR_CLASS -103
541
#define ONIGERR_END_PATTERN_AT_ESCAPE -104
542
#define ONIGERR_END_PATTERN_AT_META -105
543
#define ONIGERR_END_PATTERN_AT_CONTROL -106
544
#define ONIGERR_META_CODE_SYNTAX -108
545
#define ONIGERR_CONTROL_CODE_SYNTAX -109
546
#define ONIGERR_CHAR_CLASS_VALUE_AT_END_OF_RANGE -110
547
#define ONIGERR_CHAR_CLASS_VALUE_AT_START_OF_RANGE -111
548
#define ONIGERR_UNMATCHED_RANGE_SPECIFIER_IN_CHAR_CLASS -112
549
#define ONIGERR_TARGET_OF_REPEAT_OPERATOR_NOT_SPECIFIED -113
550
#define ONIGERR_TARGET_OF_REPEAT_OPERATOR_INVALID -114
551
#define ONIGERR_NESTED_REPEAT_OPERATOR -115
552
#define ONIGERR_UNMATCHED_CLOSE_PARENTHESIS -116
553
#define ONIGERR_END_PATTERN_WITH_UNMATCHED_PARENTHESIS -117
554
#define ONIGERR_END_PATTERN_IN_GROUP -118
555
#define ONIGERR_UNDEFINED_GROUP_OPTION -119
556
#define ONIGERR_INVALID_POSIX_BRACKET_TYPE -121
557
#define ONIGERR_INVALID_LOOK_BEHIND_PATTERN -122
558
#define ONIGERR_INVALID_REPEAT_RANGE_PATTERN -123
559
/* values error (syntax error) */
560
#define ONIGERR_TOO_BIG_NUMBER -200
561
#define ONIGERR_TOO_BIG_NUMBER_FOR_REPEAT_RANGE -201
562
#define ONIGERR_UPPER_SMALLER_THAN_LOWER_IN_REPEAT_RANGE -202
563
#define ONIGERR_EMPTY_RANGE_IN_CHAR_CLASS -203
564
#define ONIGERR_MISMATCH_CODE_LENGTH_IN_CLASS_RANGE -204
565
#define ONIGERR_TOO_MANY_MULTI_BYTE_RANGES -205
566
#define ONIGERR_TOO_SHORT_MULTI_BYTE_STRING -206
567
#define ONIGERR_TOO_BIG_BACKREF_NUMBER -207
568
#define ONIGERR_INVALID_BACKREF -208
569
#define ONIGERR_NUMBERED_BACKREF_OR_CALL_NOT_ALLOWED -209
570
#define ONIGERR_TOO_LONG_WIDE_CHAR_VALUE -212
571
#define ONIGERR_EMPTY_GROUP_NAME -214
572
#define ONIGERR_INVALID_GROUP_NAME -215
573
#define ONIGERR_INVALID_CHAR_IN_GROUP_NAME -216
574
#define ONIGERR_UNDEFINED_NAME_REFERENCE -217
575
#define ONIGERR_UNDEFINED_GROUP_REFERENCE -218
576
#define ONIGERR_MULTIPLEX_DEFINED_NAME -219
577
#define ONIGERR_MULTIPLEX_DEFINITION_NAME_CALL -220
578
#define ONIGERR_NEVER_ENDING_RECURSION -221
579
#define ONIGERR_GROUP_NUMBER_OVER_FOR_CAPTURE_HISTORY -222
580
#define ONIGERR_INVALID_CHAR_PROPERTY_NAME -223
581
#define ONIGERR_INVALID_WIDE_CHAR_VALUE -400
582
#define ONIGERR_TOO_BIG_WIDE_CHAR_VALUE -401
583
#define ONIGERR_NOT_SUPPORTED_ENCODING_COMBINATION -402
584
#define ONIGERR_INVALID_COMBINATION_OF_OPTIONS -403
586
/* errors related to thread */
587
#define ONIGERR_OVER_THREAD_PASS_LIMIT_COUNT -1001
590
/* must be smaller than BIT_STATUS_BITS_NUM (unsigned int * 8) */
591
#define ONIG_MAX_CAPTURE_HISTORY_GROUP 31
592
#define ONIG_IS_CAPTURE_HISTORY_GROUP(r, i) \
593
((i) <= ONIG_MAX_CAPTURE_HISTORY_GROUP && (r)->list && (r)->list[i])
595
typedef struct OnigCaptureTreeNodeStruct {
596
int group; /* group number */
601
struct OnigCaptureTreeNodeStruct** childs;
602
} OnigCaptureTreeNode;
604
/* match result region type */
605
struct re_registers {
611
OnigCaptureTreeNode* history_root; /* capture history tree root */
614
/* capture tree traverse */
615
#define ONIG_TRAVERSE_CALLBACK_AT_FIRST 1
616
#define ONIG_TRAVERSE_CALLBACK_AT_LAST 2
617
#define ONIG_TRAVERSE_CALLBACK_AT_BOTH \
618
( ONIG_TRAVERSE_CALLBACK_AT_FIRST | ONIG_TRAVERSE_CALLBACK_AT_LAST )
621
#define ONIG_REGION_NOTPOS -1
623
typedef struct re_registers OnigRegion;
636
typedef void (*OnigWarnFunc) P_((const char* s));
637
extern void onig_null_warn P_((const char* s));
638
#define ONIG_NULL_WARN onig_null_warn
640
#define ONIG_CHAR_TABLE_SIZE 256
643
#define ONIG_STATE_NORMAL 0
644
#define ONIG_STATE_SEARCHING 1
645
#define ONIG_STATE_COMPILING -1
646
#define ONIG_STATE_MODIFY -2
648
#define ONIG_STATE(reg) \
649
((reg)->state > 0 ? ONIG_STATE_SEARCHING : (reg)->state)
651
typedef struct re_pattern_buffer {
652
/* common members of BBuf(bytes-buffer) */
653
unsigned char* p; /* compiled pattern */
654
unsigned int used; /* used space for p */
655
unsigned int alloc; /* allocated space for p */
657
int state; /* normal, searching, compiling */
658
int num_mem; /* used memory(...) num counted from 1 */
659
int num_repeat; /* OP_REPEAT/OP_REPEAT_NG id-counter */
660
int num_null_check; /* OP_NULL_CHECK_START/END id counter */
661
int num_comb_exp_check; /* combination explosion check */
662
int num_call; /* number of subexp call */
663
unsigned int capture_history; /* (?@...) flag (1-31) */
664
unsigned int bt_mem_start; /* need backtrack flag */
665
unsigned int bt_mem_end; /* need backtrack flag */
667
int repeat_range_alloc;
668
OnigRepeatRange* repeat_range;
671
OnigOptionType options;
672
OnigSyntaxType* syntax;
673
OnigCaseFoldType case_fold_flag;
676
/* optimization info (string search, char-map and anchors) */
677
int optimize; /* optimize flag */
678
int threshold_len; /* search str-length for apply optimize */
679
int anchor; /* BEGIN_BUF, BEGIN_POS, (SEMI_)END_BUF */
680
OnigDistance anchor_dmin; /* (SEMI_)END_BUF anchor distance */
681
OnigDistance anchor_dmax; /* (SEMI_)END_BUF anchor distance */
682
int sub_anchor; /* start-anchor for exact or map */
683
unsigned char *exact;
684
unsigned char *exact_end;
685
unsigned char map[ONIG_CHAR_TABLE_SIZE]; /* used as BM skip or char-map */
686
int *int_map; /* BM skip for exact_len > 255 */
687
int *int_map_backward; /* BM skip for backward search */
688
OnigDistance dmin; /* min-distance of exact or map */
689
OnigDistance dmax; /* max-distance of exact or map */
691
/* regex_t link chain */
692
struct re_pattern_buffer* chain; /* escape compile-conflict */
695
typedef OnigRegexType* OnigRegex;
697
#ifndef ONIG_ESCAPE_REGEX_T_COLLISION
698
typedef OnigRegexType regex_t;
704
OnigEncoding pattern_enc;
705
OnigEncoding target_enc;
706
OnigSyntaxType* syntax;
707
OnigOptionType option;
708
OnigCaseFoldType case_fold_flag;
711
/* Oniguruma Native API */
713
int onig_init P_((void));
715
int onig_error_code_to_str PV_((OnigUChar* s, int err_code, ...));
717
void onig_set_warn_func P_((OnigWarnFunc f));
719
void onig_set_verb_warn_func P_((OnigWarnFunc f));
721
int onig_new P_((OnigRegex*, const OnigUChar* pattern, const OnigUChar* pattern_end, OnigOptionType option, OnigEncoding enc, OnigSyntaxType* syntax, OnigErrorInfo* einfo));
723
int onig_new_deluxe P_((OnigRegex* reg, const OnigUChar* pattern, const OnigUChar* pattern_end, OnigCompileInfo* ci, OnigErrorInfo* einfo));
725
void onig_free P_((OnigRegex));
727
int onig_recompile P_((OnigRegex, const OnigUChar* pattern, const OnigUChar* pattern_end, OnigOptionType option, OnigEncoding enc, OnigSyntaxType* syntax, OnigErrorInfo* einfo));
729
int onig_recompile_deluxe P_((OnigRegex reg, const OnigUChar* pattern, const OnigUChar* pattern_end, OnigCompileInfo* ci, OnigErrorInfo* einfo));
731
int onig_search P_((OnigRegex, const OnigUChar* str, const OnigUChar* end, const OnigUChar* start, const OnigUChar* range, OnigRegion* region, OnigOptionType option));
733
int onig_match P_((OnigRegex, const OnigUChar* str, const OnigUChar* end, const OnigUChar* at, OnigRegion* region, OnigOptionType option));
735
OnigRegion* onig_region_new P_((void));
737
void onig_region_init P_((OnigRegion* region));
739
void onig_region_free P_((OnigRegion* region, int free_self));
741
void onig_region_copy P_((OnigRegion* to, OnigRegion* from));
743
void onig_region_clear P_((OnigRegion* region));
745
int onig_region_resize P_((OnigRegion* region, int n));
747
int onig_region_set P_((OnigRegion* region, int at, int beg, int end));
749
int onig_name_to_group_numbers P_((OnigRegex reg, const OnigUChar* name, const OnigUChar* name_end, int** nums));
751
int onig_name_to_backref_number P_((OnigRegex reg, const OnigUChar* name, const OnigUChar* name_end, OnigRegion *region));
753
int onig_foreach_name P_((OnigRegex reg, int (*func)(const OnigUChar*, const OnigUChar*,int,int*,OnigRegex,void*), void* arg));
755
int onig_number_of_names P_((OnigRegex reg));
757
int onig_number_of_captures P_((OnigRegex reg));
759
int onig_number_of_capture_histories P_((OnigRegex reg));
761
OnigCaptureTreeNode* onig_get_capture_tree P_((OnigRegion* region));
763
int onig_capture_tree_traverse P_((OnigRegion* region, int at, int(*callback_func)(int,int,int,int,int,void*), void* arg));
765
int onig_noname_group_capture_is_active P_((OnigRegex reg));
767
OnigEncoding onig_get_encoding P_((OnigRegex reg));
769
OnigOptionType onig_get_options P_((OnigRegex reg));
771
OnigCaseFoldType onig_get_case_fold_flag P_((OnigRegex reg));
773
OnigSyntaxType* onig_get_syntax P_((OnigRegex reg));
775
int onig_set_default_syntax P_((OnigSyntaxType* syntax));
777
void onig_copy_syntax P_((OnigSyntaxType* to, OnigSyntaxType* from));
779
unsigned int onig_get_syntax_op P_((OnigSyntaxType* syntax));
781
unsigned int onig_get_syntax_op2 P_((OnigSyntaxType* syntax));
783
unsigned int onig_get_syntax_behavior P_((OnigSyntaxType* syntax));
785
OnigOptionType onig_get_syntax_options P_((OnigSyntaxType* syntax));
787
void onig_set_syntax_op P_((OnigSyntaxType* syntax, unsigned int op));
789
void onig_set_syntax_op2 P_((OnigSyntaxType* syntax, unsigned int op2));
791
void onig_set_syntax_behavior P_((OnigSyntaxType* syntax, unsigned int behavior));
793
void onig_set_syntax_options P_((OnigSyntaxType* syntax, OnigOptionType options));
795
int onig_set_meta_char P_((OnigSyntaxType* syntax, unsigned int what, OnigCodePoint code));
797
void onig_copy_encoding P_((OnigEncoding to, OnigEncoding from));
799
OnigCaseFoldType onig_get_default_case_fold_flag P_((void));
801
int onig_set_default_case_fold_flag P_((OnigCaseFoldType case_fold_flag));
803
unsigned int onig_get_match_stack_limit_size P_((void));
805
int onig_set_match_stack_limit_size P_((unsigned int size));
807
int onig_end P_((void));
809
const char* onig_version P_((void));
811
const char* onig_copyright P_((void));
815
{ /* satisfy cc-mode */
820
#endif /* ONIGURUMA_H */
added
removed
include/ruby/oniguruma.h
