Annotation of embedaddon/php/ext/mbstring/oniguruma/oniguruma.h, revision 1.1.1.1
1.1 misho 1: #ifndef ONIGURUMA_H
2: #define ONIGURUMA_H
3: /**********************************************************************
4: oniguruma.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: #include "../php_onig_compat.h"
33:
34: #ifdef __cplusplus
35: extern "C" {
36: #endif
37:
38: #define ONIGURUMA
39: #define ONIGURUMA_VERSION_MAJOR 4
40: #define ONIGURUMA_VERSION_MINOR 7
41: #define ONIGURUMA_VERSION_TEENY 1
42:
43: #ifdef __cplusplus
44: # ifndef HAVE_PROTOTYPES
45: # define HAVE_PROTOTYPES 1
46: # endif
47: # ifndef HAVE_STDARG_PROTOTYPES
48: # define HAVE_STDARG_PROTOTYPES 1
49: # endif
50: #endif
51:
52: /* escape Mac OS X/Xcode 2.4/gcc 4.0.1 problem */
53: #if defined(__APPLE__) && defined(__GNUC__) && __GNUC__ >= 4
54: # ifndef HAVE_STDARG_PROTOTYPES
55: # define HAVE_STDARG_PROTOTYPES 1
56: # endif
57: #endif
58:
59: #ifndef P_
60: #if defined(__STDC__) || defined(_WIN32)
61: # define P_(args) args
62: #else
63: # define P_(args) ()
64: #endif
65: #endif
66:
67: #ifndef PV_
68: #ifdef HAVE_STDARG_PROTOTYPES
69: # define PV_(args) args
70: #else
71: # define PV_(args) ()
72: #endif
73: #endif
74:
75: #ifndef ONIG_EXTERN
76: #if defined(_WIN32) && !defined(__GNUC__)
77: #if defined(EXPORT) || defined(RUBY_EXPORT)
78: #define ONIG_EXTERN extern __declspec(dllexport)
79: #else
80: #define ONIG_EXTERN extern __declspec(dllimport)
81: #endif
82: #endif
83: #endif
84:
85: #ifndef ONIG_EXTERN
86: #define ONIG_EXTERN extern
87: #endif
88:
89: /* PART: character encoding */
90:
91: #ifndef ONIG_ESCAPE_UCHAR_COLLISION
92: #define UChar OnigUChar
93: #endif
94:
95: typedef unsigned char OnigUChar;
96: typedef unsigned long OnigCodePoint;
97: typedef unsigned int OnigDistance;
98:
99: #define ONIG_INFINITE_DISTANCE ~((OnigDistance )0)
100:
101: /* ambiguous match flag */
102: typedef unsigned int OnigAmbigType;
103:
104: ONIG_EXTERN OnigAmbigType OnigDefaultAmbigFlag;
105:
106: #define ONIGENC_AMBIGUOUS_MATCH_NONE 0
107: #define ONIGENC_AMBIGUOUS_MATCH_ASCII_CASE (1<<0)
108: #define ONIGENC_AMBIGUOUS_MATCH_NONASCII_CASE (1<<1)
109:
110: #define ONIGENC_AMBIGUOUS_MATCH_LIMIT (1<<1)
111:
112: #define ONIGENC_AMBIGUOUS_MATCH_FULL \
113: ( ONIGENC_AMBIGUOUS_MATCH_ASCII_CASE | ONIGENC_AMBIGUOUS_MATCH_NONASCII_CASE )
114: #define ONIGENC_AMBIGUOUS_MATCH_DEFAULT OnigDefaultAmbigFlag
115:
116:
117: #define ONIGENC_MAX_COMP_AMBIG_CODE_LEN 3
118: #define ONIGENC_MAX_COMP_AMBIG_CODE_ITEM_NUM 4
119:
120: /* code range */
121: #define ONIGENC_CODE_RANGE_NUM(range) ((int )range[0])
122: #define ONIGENC_CODE_RANGE_FROM(range,i) range[((i)*2) + 1]
123: #define ONIGENC_CODE_RANGE_TO(range,i) range[((i)*2) + 2]
124:
125: typedef struct {
126: int len;
127: OnigCodePoint code[ONIGENC_MAX_COMP_AMBIG_CODE_LEN];
128: } OnigCompAmbigCodeItem;
129:
130: typedef struct {
131: int n;
132: OnigCodePoint code;
133: OnigCompAmbigCodeItem items[ONIGENC_MAX_COMP_AMBIG_CODE_ITEM_NUM];
134: } OnigCompAmbigCodes;
135:
136: typedef struct {
137: OnigCodePoint from;
138: OnigCodePoint to;
139: } OnigPairAmbigCodes;
140:
141: typedef struct {
142: OnigCodePoint esc;
143: OnigCodePoint anychar;
144: OnigCodePoint anytime;
145: OnigCodePoint zero_or_one_time;
146: OnigCodePoint one_or_more_time;
147: OnigCodePoint anychar_anytime;
148: } OnigMetaCharTableType;
149:
150:
151: #if defined(RUBY_PLATFORM) && defined(M17N_H)
152:
153: #define ONIG_RUBY_M17N
154: typedef m17n_encoding* OnigEncoding;
155:
156: #else
157:
158: typedef struct {
159: int (*mbc_enc_len)(const OnigUChar* p);
160: const char* name;
161: int max_enc_len;
162: int min_enc_len;
163: OnigAmbigType support_ambig_flag;
164: OnigMetaCharTableType meta_char_table;
165: int (*is_mbc_newline)(const OnigUChar* p, const OnigUChar* end);
166: OnigCodePoint (*mbc_to_code)(const OnigUChar* p, const OnigUChar* end);
167: int (*code_to_mbclen)(OnigCodePoint code);
168: int (*code_to_mbc)(OnigCodePoint code, OnigUChar *buf);
169: int (*mbc_to_normalize)(OnigAmbigType flag, const OnigUChar** pp, const OnigUChar* end, OnigUChar* to);
170: int (*is_mbc_ambiguous)(OnigAmbigType flag, const OnigUChar** pp, const OnigUChar* end);
171: int (*get_all_pair_ambig_codes)(OnigAmbigType flag, const OnigPairAmbigCodes** acs);
172: int (*get_all_comp_ambig_codes)(OnigAmbigType flag, const OnigCompAmbigCodes** acs);
173: int (*is_code_ctype)(OnigCodePoint code, unsigned int ctype);
174: int (*get_ctype_code_range)(int ctype, const OnigCodePoint* sb_range[], const OnigCodePoint* mb_range[]);
175: OnigUChar* (*left_adjust_char_head)(const OnigUChar* start, const OnigUChar* p);
176: int (*is_allowed_reverse_match)(const OnigUChar* p, const OnigUChar* end);
177: } OnigEncodingType;
178:
179: typedef OnigEncodingType* OnigEncoding;
180:
181: ONIG_EXTERN OnigEncodingType OnigEncodingASCII;
182: ONIG_EXTERN OnigEncodingType OnigEncodingISO_8859_1;
183: ONIG_EXTERN OnigEncodingType OnigEncodingISO_8859_2;
184: ONIG_EXTERN OnigEncodingType OnigEncodingISO_8859_3;
185: ONIG_EXTERN OnigEncodingType OnigEncodingISO_8859_4;
186: ONIG_EXTERN OnigEncodingType OnigEncodingISO_8859_5;
187: ONIG_EXTERN OnigEncodingType OnigEncodingISO_8859_6;
188: ONIG_EXTERN OnigEncodingType OnigEncodingISO_8859_7;
189: ONIG_EXTERN OnigEncodingType OnigEncodingISO_8859_8;
190: ONIG_EXTERN OnigEncodingType OnigEncodingISO_8859_9;
191: ONIG_EXTERN OnigEncodingType OnigEncodingISO_8859_10;
192: ONIG_EXTERN OnigEncodingType OnigEncodingISO_8859_11;
193: ONIG_EXTERN OnigEncodingType OnigEncodingISO_8859_13;
194: ONIG_EXTERN OnigEncodingType OnigEncodingISO_8859_14;
195: ONIG_EXTERN OnigEncodingType OnigEncodingISO_8859_15;
196: ONIG_EXTERN OnigEncodingType OnigEncodingISO_8859_16;
197: ONIG_EXTERN OnigEncodingType OnigEncodingUTF8;
198: ONIG_EXTERN OnigEncodingType OnigEncodingUTF16_BE;
199: ONIG_EXTERN OnigEncodingType OnigEncodingUTF16_LE;
200: ONIG_EXTERN OnigEncodingType OnigEncodingUTF32_BE;
201: ONIG_EXTERN OnigEncodingType OnigEncodingUTF32_LE;
202: ONIG_EXTERN OnigEncodingType OnigEncodingEUC_JP;
203: ONIG_EXTERN OnigEncodingType OnigEncodingEUC_TW;
204: ONIG_EXTERN OnigEncodingType OnigEncodingEUC_KR;
205: ONIG_EXTERN OnigEncodingType OnigEncodingEUC_CN;
206: ONIG_EXTERN OnigEncodingType OnigEncodingSJIS;
207: ONIG_EXTERN OnigEncodingType OnigEncodingKOI8;
208: ONIG_EXTERN OnigEncodingType OnigEncodingKOI8_R;
209: ONIG_EXTERN OnigEncodingType OnigEncodingBIG5;
210: ONIG_EXTERN OnigEncodingType OnigEncodingGB18030;
211:
212: #define ONIG_ENCODING_ASCII (&OnigEncodingASCII)
213: #define ONIG_ENCODING_ISO_8859_1 (&OnigEncodingISO_8859_1)
214: #define ONIG_ENCODING_ISO_8859_2 (&OnigEncodingISO_8859_2)
215: #define ONIG_ENCODING_ISO_8859_3 (&OnigEncodingISO_8859_3)
216: #define ONIG_ENCODING_ISO_8859_4 (&OnigEncodingISO_8859_4)
217: #define ONIG_ENCODING_ISO_8859_5 (&OnigEncodingISO_8859_5)
218: #define ONIG_ENCODING_ISO_8859_6 (&OnigEncodingISO_8859_6)
219: #define ONIG_ENCODING_ISO_8859_7 (&OnigEncodingISO_8859_7)
220: #define ONIG_ENCODING_ISO_8859_8 (&OnigEncodingISO_8859_8)
221: #define ONIG_ENCODING_ISO_8859_9 (&OnigEncodingISO_8859_9)
222: #define ONIG_ENCODING_ISO_8859_10 (&OnigEncodingISO_8859_10)
223: #define ONIG_ENCODING_ISO_8859_11 (&OnigEncodingISO_8859_11)
224: #define ONIG_ENCODING_ISO_8859_13 (&OnigEncodingISO_8859_13)
225: #define ONIG_ENCODING_ISO_8859_14 (&OnigEncodingISO_8859_14)
226: #define ONIG_ENCODING_ISO_8859_15 (&OnigEncodingISO_8859_15)
227: #define ONIG_ENCODING_ISO_8859_16 (&OnigEncodingISO_8859_16)
228: #define ONIG_ENCODING_UTF8 (&OnigEncodingUTF8)
229: #define ONIG_ENCODING_UTF16_BE (&OnigEncodingUTF16_BE)
230: #define ONIG_ENCODING_UTF16_LE (&OnigEncodingUTF16_LE)
231: #define ONIG_ENCODING_UTF32_BE (&OnigEncodingUTF32_BE)
232: #define ONIG_ENCODING_UTF32_LE (&OnigEncodingUTF32_LE)
233: #define ONIG_ENCODING_EUC_JP (&OnigEncodingEUC_JP)
234: #define ONIG_ENCODING_EUC_TW (&OnigEncodingEUC_TW)
235: #define ONIG_ENCODING_EUC_KR (&OnigEncodingEUC_KR)
236: #define ONIG_ENCODING_EUC_CN (&OnigEncodingEUC_CN)
237: #define ONIG_ENCODING_SJIS (&OnigEncodingSJIS)
238: #define ONIG_ENCODING_KOI8 (&OnigEncodingKOI8)
239: #define ONIG_ENCODING_KOI8_R (&OnigEncodingKOI8_R)
240: #define ONIG_ENCODING_BIG5 (&OnigEncodingBIG5)
241: #define ONIG_ENCODING_GB18030 (&OnigEncodingGB18030)
242:
243: #endif /* else RUBY && M17N */
244:
245: #define ONIG_ENCODING_UNDEF ((OnigEncoding )0)
246:
247:
248: /* work size */
249: #define ONIGENC_CODE_TO_MBC_MAXLEN 7
250: #define ONIGENC_MBC_NORMALIZE_MAXLEN ONIGENC_CODE_TO_MBC_MAXLEN
251:
252: /* character types */
253: #define ONIGENC_CTYPE_NEWLINE (1<< 0)
254: #define ONIGENC_CTYPE_ALPHA (1<< 1)
255: #define ONIGENC_CTYPE_BLANK (1<< 2)
256: #define ONIGENC_CTYPE_CNTRL (1<< 3)
257: #define ONIGENC_CTYPE_DIGIT (1<< 4)
258: #define ONIGENC_CTYPE_GRAPH (1<< 5)
259: #define ONIGENC_CTYPE_LOWER (1<< 6)
260: #define ONIGENC_CTYPE_PRINT (1<< 7)
261: #define ONIGENC_CTYPE_PUNCT (1<< 8)
262: #define ONIGENC_CTYPE_SPACE (1<< 9)
263: #define ONIGENC_CTYPE_UPPER (1<<10)
264: #define ONIGENC_CTYPE_XDIGIT (1<<11)
265: #define ONIGENC_CTYPE_WORD (1<<12)
266: #define ONIGENC_CTYPE_ASCII (1<<13)
267: #define ONIGENC_CTYPE_ALNUM (ONIGENC_CTYPE_ALPHA | ONIGENC_CTYPE_DIGIT)
268:
269: #define enc_len(enc,p) ONIGENC_MBC_ENC_LEN(enc, p)
270:
271: #define ONIGENC_IS_UNDEF(enc) ((enc) == ONIG_ENCODING_UNDEF)
272: #define ONIGENC_IS_SINGLEBYTE(enc) (ONIGENC_MBC_MAXLEN(enc) == 1)
273: #define ONIGENC_IS_MBC_HEAD(enc,p) (ONIGENC_MBC_ENC_LEN(enc,p) != 1)
274: #define ONIGENC_IS_MBC_ASCII(p) (*(p) < 128)
275: #define ONIGENC_IS_CODE_ASCII(code) ((code) < 128)
276: #define ONIGENC_IS_CODE_SB_WORD(enc,code) \
277: (ONIGENC_IS_CODE_ASCII(code) && ONIGENC_IS_CODE_WORD(enc,code))
278: #define ONIGENC_IS_MBC_WORD(enc,s,end) \
279: ONIGENC_IS_CODE_WORD(enc,ONIGENC_MBC_TO_CODE(enc,s,end))
280:
281:
282: #ifdef ONIG_RUBY_M17N
283:
284: #include <ctype.h> /* for isblank(), isgraph() */
285:
286: #define ONIGENC_MBC_TO_NORMALIZE(enc,flag,pp,end,buf) \
287: onigenc_mbc_to_normalize(enc,flag,pp,end,buf)
288: #define ONIGENC_IS_MBC_AMBIGUOUS(enc,flag,pp,end) \
289: onigenc_is_mbc_ambiguous(enc,flag,pp,end)
290:
291: #define ONIGENC_SUPPORT_AMBIG_FLAG(enc) ONIGENC_AMBIGUOUS_MATCH_ASCII_CASE
292: #define ONIGENC_IS_ALLOWED_REVERSE_MATCH(enc,s,end) \
293: onigenc_is_allowed_reverse_match(enc, s, end)
294: #define ONIGENC_LEFT_ADJUST_CHAR_HEAD(enc,start,s) \
295: onigenc_get_left_adjust_char_head(enc, start, s)
296: #define ONIGENC_GET_ALL_PAIR_AMBIG_CODES(enc, ambig_flag, acs) 0
297: #define ONIGENC_GET_ALL_COMP_AMBIG_CODES(enc, ambig_flag, acs) 0
298: #define ONIGENC_GET_CTYPE_CODE_RANGE(enc,ctype,sbr,mbr) \
299: ONIG_NO_SUPPORT_CONFIG
300: #define ONIGENC_MBC_ENC_LEN(enc,p) m17n_mbclen(enc,(int )(*p))
301: #define ONIGENC_MBC_MAXLEN(enc) m17n_mbmaxlen(enc)
302: #define ONIGENC_MBC_MAXLEN_DIST(enc) \
303: (ONIGENC_MBC_MAXLEN(enc) > 0 ? ONIGENC_MBC_MAXLEN(enc) \
304: : ONIG_INFINITE_DISTANCE)
305: #define ONIGENC_MBC_MINLEN(enc) 1
306: #define ONIGENC_MBC_TO_CODE(enc,p,e) m17n_codepoint((enc),(p),(e))
307: #define ONIGENC_CODE_TO_MBCLEN(enc,code) m17n_codelen((enc),(code))
308: #define ONIGENC_CODE_TO_MBC(enc,code,buf) onigenc_code_to_mbc(enc, code, buf)
309:
310: #if 0 /* !! not supported !! */
311: #define ONIGENC_IS_MBC_NEWLINE(enc,p,end)
312: #define ONIGENC_STEP_BACK(enc,start,s,n)
313: #endif
314:
315: #define ONIGENC_IS_CODE_CTYPE(enc,code,ctype) \
316: onigenc_is_code_ctype(enc,code,ctype)
317:
318: #ifdef isblank
319: # define ONIGENC_IS_CODE_BLANK(enc,code) isblank((int )code)
320: #else
321: # define ONIGENC_IS_CODE_BLANK(enc,code) ((code) == ' ' || (code) == '\t')
322: #endif
323: #ifdef isgraph
324: # define ONIGENC_IS_CODE_GRAPH(enc,code) isgraph((int )code)
325: #else
326: # define ONIGENC_IS_CODE_GRAPH(enc,code) \
327: (isprint((int )code) && !isspace((int )code))
328: #endif
329:
330: #define ONIGENC_IS_CODE_PRINT(enc,code) m17n_isprint(enc,code)
331: #define ONIGENC_IS_CODE_ALNUM(enc,code) m17n_isalnum(enc,code)
332: #define ONIGENC_IS_CODE_ALPHA(enc,code) m17n_isalpha(enc,code)
333: #define ONIGENC_IS_CODE_LOWER(enc,code) m17n_islower(enc,code)
334: #define ONIGENC_IS_CODE_UPPER(enc,code) m17n_isupper(enc,code)
335: #define ONIGENC_IS_CODE_CNTRL(enc,code) m17n_iscntrl(enc,code)
336: #define ONIGENC_IS_CODE_PUNCT(enc,code) m17n_ispunct(enc,code)
337: #define ONIGENC_IS_CODE_SPACE(enc,code) m17n_isspace(enc,code)
338: #define ONIGENC_IS_CODE_DIGIT(enc,code) m17n_isdigit(enc,code)
339: #define ONIGENC_IS_CODE_XDIGIT(enc,code) m17n_isxdigit(enc,code)
340: #define ONIGENC_IS_CODE_WORD(enc,code) m17n_iswchar(enc,code)
341:
342: ONIG_EXTERN
343: int onigenc_is_code_ctype P_((OnigEncoding enc, OnigCodePoint code, int ctype));
344: ONIG_EXTERN
345: int onigenc_code_to_mbc P_((OnigEncoding enc, OnigCodePoint code, OnigUChar *buf));
346: ONIG_EXTERN
347: int onigenc_mbc_to_normalize P_((OnigEncoding enc, OnigAmbigType flag, const OnigUChar** pp, const OnigUChar* end, OnigUChar* buf));
348: ONIG_EXTERN
349: int onigenc_is_mbc_ambiguous P_((OnigEncoding enc, OnigAmbigType flag, const OnigUChar** pp, const OnigUChar* end));
350: ONIG_EXTERN
351: int onigenc_is_allowed_reverse_match P_((OnigEncoding enc, const OnigUChar* s, const OnigUChar* end));
352:
353: #else /* ONIG_RUBY_M17N */
354:
355: #define ONIGENC_NAME(enc) ((enc)->name)
356:
357: #define ONIGENC_MBC_TO_NORMALIZE(enc,flag,pp,end,buf) \
358: (enc)->mbc_to_normalize(flag,(const OnigUChar** )pp,end,buf)
359: #define ONIGENC_IS_MBC_AMBIGUOUS(enc,flag,pp,end) \
360: (enc)->is_mbc_ambiguous(flag,(const OnigUChar** )pp,end)
361: #define ONIGENC_SUPPORT_AMBIG_FLAG(enc) ((enc)->support_ambig_flag)
362: #define ONIGENC_IS_ALLOWED_REVERSE_MATCH(enc,s,end) \
363: (enc)->is_allowed_reverse_match(s,end)
364: #define ONIGENC_LEFT_ADJUST_CHAR_HEAD(enc,start,s) \
365: (enc)->left_adjust_char_head(start, s)
366: #define ONIGENC_GET_ALL_PAIR_AMBIG_CODES(enc,ambig_flag,acs) \
367: (enc)->get_all_pair_ambig_codes(ambig_flag,acs)
368: #define ONIGENC_GET_ALL_COMP_AMBIG_CODES(enc,ambig_flag,acs) \
369: (enc)->get_all_comp_ambig_codes(ambig_flag,acs)
370: #define ONIGENC_STEP_BACK(enc,start,s,n) \
371: onigenc_step_back((enc),(start),(s),(n))
372:
373: #define ONIGENC_MBC_ENC_LEN(enc,p) (enc)->mbc_enc_len(p)
374: #define ONIGENC_MBC_MAXLEN(enc) ((enc)->max_enc_len)
375: #define ONIGENC_MBC_MAXLEN_DIST(enc) ONIGENC_MBC_MAXLEN(enc)
376: #define ONIGENC_MBC_MINLEN(enc) ((enc)->min_enc_len)
377: #define ONIGENC_IS_MBC_NEWLINE(enc,p,end) (enc)->is_mbc_newline((p),(end))
378: #define ONIGENC_MBC_TO_CODE(enc,p,end) (enc)->mbc_to_code((p),(end))
379: #define ONIGENC_CODE_TO_MBCLEN(enc,code) (enc)->code_to_mbclen(code)
380: #define ONIGENC_CODE_TO_MBC(enc,code,buf) (enc)->code_to_mbc(code,buf)
381:
382: #define ONIGENC_IS_CODE_CTYPE(enc,code,ctype) (enc)->is_code_ctype(code,ctype)
383:
384: #define ONIGENC_IS_CODE_NEWLINE(enc,code) \
385: ONIGENC_IS_CODE_CTYPE(enc,code,ONIGENC_CTYPE_NEWLINE)
386: #define ONIGENC_IS_CODE_GRAPH(enc,code) \
387: ONIGENC_IS_CODE_CTYPE(enc,code,ONIGENC_CTYPE_GRAPH)
388: #define ONIGENC_IS_CODE_PRINT(enc,code) \
389: ONIGENC_IS_CODE_CTYPE(enc,code,ONIGENC_CTYPE_PRINT)
390: #define ONIGENC_IS_CODE_ALNUM(enc,code) \
391: ONIGENC_IS_CODE_CTYPE(enc,code,ONIGENC_CTYPE_ALNUM)
392: #define ONIGENC_IS_CODE_ALPHA(enc,code) \
393: ONIGENC_IS_CODE_CTYPE(enc,code,ONIGENC_CTYPE_ALPHA)
394: #define ONIGENC_IS_CODE_LOWER(enc,code) \
395: ONIGENC_IS_CODE_CTYPE(enc,code,ONIGENC_CTYPE_LOWER)
396: #define ONIGENC_IS_CODE_UPPER(enc,code) \
397: ONIGENC_IS_CODE_CTYPE(enc,code,ONIGENC_CTYPE_UPPER)
398: #define ONIGENC_IS_CODE_CNTRL(enc,code) \
399: ONIGENC_IS_CODE_CTYPE(enc,code,ONIGENC_CTYPE_CNTRL)
400: #define ONIGENC_IS_CODE_PUNCT(enc,code) \
401: ONIGENC_IS_CODE_CTYPE(enc,code,ONIGENC_CTYPE_PUNCT)
402: #define ONIGENC_IS_CODE_SPACE(enc,code) \
403: ONIGENC_IS_CODE_CTYPE(enc,code,ONIGENC_CTYPE_SPACE)
404: #define ONIGENC_IS_CODE_BLANK(enc,code) \
405: ONIGENC_IS_CODE_CTYPE(enc,code,ONIGENC_CTYPE_BLANK)
406: #define ONIGENC_IS_CODE_DIGIT(enc,code) \
407: ONIGENC_IS_CODE_CTYPE(enc,code,ONIGENC_CTYPE_DIGIT)
408: #define ONIGENC_IS_CODE_XDIGIT(enc,code) \
409: ONIGENC_IS_CODE_CTYPE(enc,code,ONIGENC_CTYPE_XDIGIT)
410: #define ONIGENC_IS_CODE_WORD(enc,code) \
411: ONIGENC_IS_CODE_CTYPE(enc,code,ONIGENC_CTYPE_WORD)
412:
413: #define ONIGENC_GET_CTYPE_CODE_RANGE(enc,ctype,sbr,mbr) \
414: (enc)->get_ctype_code_range(ctype,sbr,mbr)
415:
416: ONIG_EXTERN
417: OnigUChar* onigenc_step_back P_((OnigEncoding enc, const OnigUChar* start, const OnigUChar* s, int n));
418:
419: #endif /* is not ONIG_RUBY_M17N */
420:
421:
422: /* encoding API */
423: ONIG_EXTERN
424: int onigenc_init P_((void));
425: ONIG_EXTERN
426: int onigenc_set_default_encoding P_((OnigEncoding enc));
427: ONIG_EXTERN
428: OnigEncoding onigenc_get_default_encoding P_((void));
429: ONIG_EXTERN
430: void onigenc_set_default_caseconv_table P_((const OnigUChar* table));
431: ONIG_EXTERN
432: OnigUChar* onigenc_get_right_adjust_char_head_with_prev P_((OnigEncoding enc, const OnigUChar* start, const OnigUChar* s, const OnigUChar** prev));
433: ONIG_EXTERN
434: OnigUChar* onigenc_get_prev_char_head P_((OnigEncoding enc, const OnigUChar* start, const OnigUChar* s));
435: ONIG_EXTERN
436: OnigUChar* onigenc_get_left_adjust_char_head P_((OnigEncoding enc, const OnigUChar* start, const OnigUChar* s));
437: ONIG_EXTERN
438: OnigUChar* onigenc_get_right_adjust_char_head P_((OnigEncoding enc, const OnigUChar* start, const OnigUChar* s));
439: ONIG_EXTERN
440: int onigenc_strlen P_((OnigEncoding enc, const OnigUChar* p, const OnigUChar* end));
441: ONIG_EXTERN
442: int onigenc_strlen_null P_((OnigEncoding enc, const OnigUChar* p));
443: ONIG_EXTERN
444: int onigenc_str_bytelen_null P_((OnigEncoding enc, const OnigUChar* p));
445:
446:
447:
448: /* PART: regular expression */
449:
450: /* config parameters */
451: #define ONIG_NREGION 10
452: #define ONIG_MAX_BACKREF_NUM 1000
453: #define ONIG_MAX_REPEAT_NUM 100000
454: #define ONIG_MAX_MULTI_BYTE_RANGES_NUM 10000
455: /* constants */
456: #define ONIG_MAX_ERROR_MESSAGE_LEN 90
457:
458: typedef unsigned int OnigOptionType;
459:
460: #define ONIG_OPTION_DEFAULT ONIG_OPTION_NONE
461:
462: /* options */
463: #define ONIG_OPTION_NONE 0U
464: #define ONIG_OPTION_IGNORECASE 1U
465: #define ONIG_OPTION_EXTEND (ONIG_OPTION_IGNORECASE << 1)
466: #define ONIG_OPTION_MULTILINE (ONIG_OPTION_EXTEND << 1)
467: #define ONIG_OPTION_SINGLELINE (ONIG_OPTION_MULTILINE << 1)
468: #define ONIG_OPTION_FIND_LONGEST (ONIG_OPTION_SINGLELINE << 1)
469: #define ONIG_OPTION_FIND_NOT_EMPTY (ONIG_OPTION_FIND_LONGEST << 1)
470: #define ONIG_OPTION_NEGATE_SINGLELINE (ONIG_OPTION_FIND_NOT_EMPTY << 1)
471: #define ONIG_OPTION_DONT_CAPTURE_GROUP (ONIG_OPTION_NEGATE_SINGLELINE << 1)
472: #define ONIG_OPTION_CAPTURE_GROUP (ONIG_OPTION_DONT_CAPTURE_GROUP << 1)
473: /* options (search time) */
474: #define ONIG_OPTION_NOTBOL (ONIG_OPTION_CAPTURE_GROUP << 1)
475: #define ONIG_OPTION_NOTEOL (ONIG_OPTION_NOTBOL << 1)
476: #define ONIG_OPTION_POSIX_REGION (ONIG_OPTION_NOTEOL << 1)
477: #define ONIG_OPTION_MAXBIT ONIG_OPTION_POSIX_REGION /* limit */
478:
479: #define ONIG_OPTION_ON(options,regopt) ((options) |= (regopt))
480: #define ONIG_OPTION_OFF(options,regopt) ((options) &= ~(regopt))
481: #define ONIG_IS_OPTION_ON(options,option) ((options) & (option))
482:
483: /* syntax */
484: typedef struct {
485: unsigned int op;
486: unsigned int op2;
487: unsigned int behavior;
488: OnigOptionType options; /* default option */
489: } OnigSyntaxType;
490:
491: ONIG_EXTERN OnigSyntaxType OnigSyntaxASIS;
492: ONIG_EXTERN OnigSyntaxType OnigSyntaxPosixBasic;
493: ONIG_EXTERN OnigSyntaxType OnigSyntaxPosixExtended;
494: ONIG_EXTERN OnigSyntaxType OnigSyntaxEmacs;
495: ONIG_EXTERN OnigSyntaxType OnigSyntaxGrep;
496: ONIG_EXTERN OnigSyntaxType OnigSyntaxGnuRegex;
497: ONIG_EXTERN OnigSyntaxType OnigSyntaxJava;
498: ONIG_EXTERN OnigSyntaxType OnigSyntaxPerl;
499: ONIG_EXTERN OnigSyntaxType OnigSyntaxPerl_NG;
500: ONIG_EXTERN OnigSyntaxType OnigSyntaxRuby;
501:
502: /* predefined syntaxes (see regsyntax.c) */
503: #define ONIG_SYNTAX_ASIS (&OnigSyntaxASIS)
504: #define ONIG_SYNTAX_POSIX_BASIC (&OnigSyntaxPosixBasic)
505: #define ONIG_SYNTAX_POSIX_EXTENDED (&OnigSyntaxPosixExtended)
506: #define ONIG_SYNTAX_EMACS (&OnigSyntaxEmacs)
507: #define ONIG_SYNTAX_GREP (&OnigSyntaxGrep)
508: #define ONIG_SYNTAX_GNU_REGEX (&OnigSyntaxGnuRegex)
509: #define ONIG_SYNTAX_JAVA (&OnigSyntaxJava)
510: #define ONIG_SYNTAX_PERL (&OnigSyntaxPerl)
511: #define ONIG_SYNTAX_PERL_NG (&OnigSyntaxPerl_NG)
512: #define ONIG_SYNTAX_RUBY (&OnigSyntaxRuby)
513:
514: /* default syntax */
515: ONIG_EXTERN OnigSyntaxType* OnigDefaultSyntax;
516: #define ONIG_SYNTAX_DEFAULT OnigDefaultSyntax
517:
518: /* syntax (operators) */
519: #define ONIG_SYN_OP_VARIABLE_META_CHARACTERS (1U<<0)
520: #define ONIG_SYN_OP_DOT_ANYCHAR (1U<<1) /* . */
521: #define ONIG_SYN_OP_ASTERISK_ZERO_INF (1U<<2) /* * */
522: #define ONIG_SYN_OP_ESC_ASTERISK_ZERO_INF (1U<<3)
523: #define ONIG_SYN_OP_PLUS_ONE_INF (1U<<4) /* + */
524: #define ONIG_SYN_OP_ESC_PLUS_ONE_INF (1U<<5)
525: #define ONIG_SYN_OP_QMARK_ZERO_ONE (1U<<6) /* ? */
526: #define ONIG_SYN_OP_ESC_QMARK_ZERO_ONE (1U<<7)
527: #define ONIG_SYN_OP_BRACE_INTERVAL (1U<<8) /* {lower,upper} */
528: #define ONIG_SYN_OP_ESC_BRACE_INTERVAL (1U<<9) /* \{lower,upper\} */
529: #define ONIG_SYN_OP_VBAR_ALT (1U<<10) /* | */
530: #define ONIG_SYN_OP_ESC_VBAR_ALT (1U<<11) /* \| */
531: #define ONIG_SYN_OP_LPAREN_SUBEXP (1U<<12) /* (...) */
532: #define ONIG_SYN_OP_ESC_LPAREN_SUBEXP (1U<<13) /* \(...\) */
533: #define ONIG_SYN_OP_ESC_AZ_BUF_ANCHOR (1U<<14) /* \A, \Z, \z */
534: #define ONIG_SYN_OP_ESC_CAPITAL_G_BEGIN_ANCHOR (1U<<15) /* \G */
535: #define ONIG_SYN_OP_DECIMAL_BACKREF (1U<<16) /* \num */
536: #define ONIG_SYN_OP_BRACKET_CC (1U<<17) /* [...] */
537: #define ONIG_SYN_OP_ESC_W_WORD (1U<<18) /* \w, \W */
538: #define ONIG_SYN_OP_ESC_LTGT_WORD_BEGIN_END (1U<<19) /* \<. \> */
539: #define ONIG_SYN_OP_ESC_B_WORD_BOUND (1U<<20) /* \b, \B */
540: #define ONIG_SYN_OP_ESC_S_WHITE_SPACE (1U<<21) /* \s, \S */
541: #define ONIG_SYN_OP_ESC_D_DIGIT (1U<<22) /* \d, \D */
542: #define ONIG_SYN_OP_LINE_ANCHOR (1U<<23) /* ^, $ */
543: #define ONIG_SYN_OP_POSIX_BRACKET (1U<<24) /* [:xxxx:] */
544: #define ONIG_SYN_OP_QMARK_NON_GREEDY (1U<<25) /* ??,*?,+?,{n,m}? */
545: #define ONIG_SYN_OP_ESC_CONTROL_CHARS (1U<<26) /* \n,\r,\t,\a ... */
546: #define ONIG_SYN_OP_ESC_C_CONTROL (1U<<27) /* \cx */
547: #define ONIG_SYN_OP_ESC_OCTAL3 (1U<<28) /* \OOO */
548: #define ONIG_SYN_OP_ESC_X_HEX2 (1U<<29) /* \xHH */
549: #define ONIG_SYN_OP_ESC_X_BRACE_HEX8 (1U<<30) /* \x{7HHHHHHH} */
550:
551: #define ONIG_SYN_OP2_ESC_CAPITAL_Q_QUOTE (1U<<0) /* \Q...\E */
552: #define ONIG_SYN_OP2_QMARK_GROUP_EFFECT (1U<<1) /* (?...) */
553: #define ONIG_SYN_OP2_OPTION_PERL (1U<<2) /* (?imsx),(?-imsx) */
554: #define ONIG_SYN_OP2_OPTION_RUBY (1U<<3) /* (?imx), (?-imx) */
555: #define ONIG_SYN_OP2_PLUS_POSSESSIVE_REPEAT (1U<<4) /* ?+,*+,++ */
556: #define ONIG_SYN_OP2_PLUS_POSSESSIVE_INTERVAL (1U<<5) /* {n,m}+ */
557: #define ONIG_SYN_OP2_CCLASS_SET_OP (1U<<6) /* [...&&..[..]..] */
558: #define ONIG_SYN_OP2_QMARK_LT_NAMED_GROUP (1U<<7) /* (?<name>...) */
559: #define ONIG_SYN_OP2_ESC_K_NAMED_BACKREF (1U<<8) /* \k<name> */
560: #define ONIG_SYN_OP2_ESC_G_SUBEXP_CALL (1U<<9) /* \g<name>, \g<n> */
561: #define ONIG_SYN_OP2_ATMARK_CAPTURE_HISTORY (1U<<10) /* (?@..),(?@<x>..) */
562: #define ONIG_SYN_OP2_ESC_CAPITAL_C_BAR_CONTROL (1U<<11) /* \C-x */
563: #define ONIG_SYN_OP2_ESC_CAPITAL_M_BAR_META (1U<<12) /* \M-x */
564: #define ONIG_SYN_OP2_ESC_V_VTAB (1U<<13) /* \v as VTAB */
565: #define ONIG_SYN_OP2_ESC_U_HEX4 (1U<<14) /* \uHHHH */
566: #define ONIG_SYN_OP2_ESC_GNU_BUF_ANCHOR (1U<<15) /* \`, \' */
567: #define ONIG_SYN_OP2_ESC_P_BRACE_CHAR_PROPERTY (1U<<16) /* \p{...}, \P{...} */
568: #define ONIG_SYN_OP2_ESC_P_BRACE_CIRCUMFLEX_NOT (1U<<17) /* \p{^..}, \P{^..} */
569: #define ONIG_SYN_OP2_CHAR_PROPERTY_PREFIX_IS (1U<<18) /* \p{IsXDigit} */
570: #define ONIG_SYN_OP2_ESC_H_XDIGIT (1U<<19) /* \h, \H */
571: #define ONIG_SYN_OP2_INEFFECTIVE_ESCAPE (1U<<20) /* \ */
572:
573: /* syntax (behavior) */
574: #define ONIG_SYN_CONTEXT_INDEP_ANCHORS (1U<<31) /* not implemented */
575: #define ONIG_SYN_CONTEXT_INDEP_REPEAT_OPS (1U<<0) /* ?, *, +, {n,m} */
576: #define ONIG_SYN_CONTEXT_INVALID_REPEAT_OPS (1U<<1) /* error or ignore */
577: #define ONIG_SYN_ALLOW_UNMATCHED_CLOSE_SUBEXP (1U<<2) /* ...)... */
578: #define ONIG_SYN_ALLOW_INVALID_INTERVAL (1U<<3) /* {??? */
579: #define ONIG_SYN_ALLOW_INTERVAL_LOW_ABBREV (1U<<4) /* {,n} => {0,n} */
580: #define ONIG_SYN_STRICT_CHECK_BACKREF (1U<<5) /* /(\1)/,/\1()/ ..*/
581: #define ONIG_SYN_DIFFERENT_LEN_ALT_LOOK_BEHIND (1U<<6) /* (?<=a|bc) */
582: #define ONIG_SYN_CAPTURE_ONLY_NAMED_GROUP (1U<<7) /* see doc/RE */
583: #define ONIG_SYN_ALLOW_MULTIPLEX_DEFINITION_NAME (1U<<8) /* (?<x>)(?<x>) */
584: #define ONIG_SYN_FIXED_INTERVAL_IS_GREEDY_ONLY (1U<<9) /* a{n}?=(?:a{n})? */
585:
586: /* syntax (behavior) in char class [...] */
587: #define ONIG_SYN_NOT_NEWLINE_IN_NEGATIVE_CC (1U<<20) /* [^...] */
588: #define ONIG_SYN_BACKSLASH_ESCAPE_IN_CC (1U<<21) /* [..\w..] etc.. */
589: #define ONIG_SYN_ALLOW_EMPTY_RANGE_IN_CC (1U<<22)
590: #define ONIG_SYN_ALLOW_DOUBLE_RANGE_OP_IN_CC (1U<<23) /* [0-9-a]=[0-9\-a] */
591: /* syntax (behavior) warning */
592: #define ONIG_SYN_WARN_CC_OP_NOT_ESCAPED (1U<<24) /* [,-,] */
593: #define ONIG_SYN_WARN_REDUNDANT_NESTED_REPEAT (1U<<25) /* (?:a*)+ */
594:
595: /* meta character specifiers (onig_set_meta_char()) */
596: #define ONIG_META_CHAR_ESCAPE 0
597: #define ONIG_META_CHAR_ANYCHAR 1
598: #define ONIG_META_CHAR_ANYTIME 2
599: #define ONIG_META_CHAR_ZERO_OR_ONE_TIME 3
600: #define ONIG_META_CHAR_ONE_OR_MORE_TIME 4
601: #define ONIG_META_CHAR_ANYCHAR_ANYTIME 5
602:
603: #define ONIG_INEFFECTIVE_META_CHAR 0
604:
605: /* error codes */
606: #define ONIG_IS_PATTERN_ERROR(ecode) ((ecode) <= -100 && (ecode) > -1000)
607: /* normal return */
608: #define ONIG_NORMAL 0
609: #define ONIG_MISMATCH -1
610: #define ONIG_NO_SUPPORT_CONFIG -2
611:
612: /* internal error */
613: #define ONIGERR_MEMORY -5
614: #define ONIGERR_TYPE_BUG -6
615: #define ONIGERR_PARSER_BUG -11
616: #define ONIGERR_STACK_BUG -12
617: #define ONIGERR_UNDEFINED_BYTECODE -13
618: #define ONIGERR_UNEXPECTED_BYTECODE -14
619: #define ONIGERR_MATCH_STACK_LIMIT_OVER -15
620: #define ONIGERR_DEFAULT_ENCODING_IS_NOT_SETTED -21
621: #define ONIGERR_SPECIFIED_ENCODING_CANT_CONVERT_TO_WIDE_CHAR -22
622: /* general error */
623: #define ONIGERR_INVALID_ARGUMENT -30
624: /* syntax error */
625: #define ONIGERR_END_PATTERN_AT_LEFT_BRACE -100
626: #define ONIGERR_END_PATTERN_AT_LEFT_BRACKET -101
627: #define ONIGERR_EMPTY_CHAR_CLASS -102
628: #define ONIGERR_PREMATURE_END_OF_CHAR_CLASS -103
629: #define ONIGERR_END_PATTERN_AT_ESCAPE -104
630: #define ONIGERR_END_PATTERN_AT_META -105
631: #define ONIGERR_END_PATTERN_AT_CONTROL -106
632: #define ONIGERR_META_CODE_SYNTAX -108
633: #define ONIGERR_CONTROL_CODE_SYNTAX -109
634: #define ONIGERR_CHAR_CLASS_VALUE_AT_END_OF_RANGE -110
635: #define ONIGERR_CHAR_CLASS_VALUE_AT_START_OF_RANGE -111
636: #define ONIGERR_UNMATCHED_RANGE_SPECIFIER_IN_CHAR_CLASS -112
637: #define ONIGERR_TARGET_OF_REPEAT_OPERATOR_NOT_SPECIFIED -113
638: #define ONIGERR_TARGET_OF_REPEAT_OPERATOR_INVALID -114
639: #define ONIGERR_NESTED_REPEAT_OPERATOR -115
640: #define ONIGERR_UNMATCHED_CLOSE_PARENTHESIS -116
641: #define ONIGERR_END_PATTERN_WITH_UNMATCHED_PARENTHESIS -117
642: #define ONIGERR_END_PATTERN_IN_GROUP -118
643: #define ONIGERR_UNDEFINED_GROUP_OPTION -119
644: #define ONIGERR_INVALID_POSIX_BRACKET_TYPE -121
645: #define ONIGERR_INVALID_LOOK_BEHIND_PATTERN -122
646: #define ONIGERR_INVALID_REPEAT_RANGE_PATTERN -123
647: /* values error (syntax error) */
648: #define ONIGERR_TOO_BIG_NUMBER -200
649: #define ONIGERR_TOO_BIG_NUMBER_FOR_REPEAT_RANGE -201
650: #define ONIGERR_UPPER_SMALLER_THAN_LOWER_IN_REPEAT_RANGE -202
651: #define ONIGERR_EMPTY_RANGE_IN_CHAR_CLASS -203
652: #define ONIGERR_MISMATCH_CODE_LENGTH_IN_CLASS_RANGE -204
653: #define ONIGERR_TOO_MANY_MULTI_BYTE_RANGES -205
654: #define ONIGERR_TOO_SHORT_MULTI_BYTE_STRING -206
655: #define ONIGERR_TOO_BIG_BACKREF_NUMBER -207
656: #define ONIGERR_INVALID_BACKREF -208
657: #define ONIGERR_NUMBERED_BACKREF_OR_CALL_NOT_ALLOWED -209
658: #define ONIGERR_TOO_LONG_WIDE_CHAR_VALUE -212
659: #define ONIGERR_EMPTY_GROUP_NAME -214
660: #define ONIGERR_INVALID_GROUP_NAME -215
661: #define ONIGERR_INVALID_CHAR_IN_GROUP_NAME -216
662: #define ONIGERR_UNDEFINED_NAME_REFERENCE -217
663: #define ONIGERR_UNDEFINED_GROUP_REFERENCE -218
664: #define ONIGERR_MULTIPLEX_DEFINED_NAME -219
665: #define ONIGERR_MULTIPLEX_DEFINITION_NAME_CALL -220
666: #define ONIGERR_NEVER_ENDING_RECURSION -221
667: #define ONIGERR_GROUP_NUMBER_OVER_FOR_CAPTURE_HISTORY -222
668: #define ONIGERR_INVALID_CHAR_PROPERTY_NAME -223
669: #define ONIGERR_INVALID_WIDE_CHAR_VALUE -400
670: #define ONIGERR_TOO_BIG_WIDE_CHAR_VALUE -401
671: #define ONIGERR_NOT_SUPPORTED_ENCODING_COMBINATION -402
672: #define ONIGERR_INVALID_COMBINATION_OF_OPTIONS -403
673:
674: /* errors related to thread */
675: #define ONIGERR_OVER_THREAD_PASS_LIMIT_COUNT -1001
676:
677:
678: /* must be smaller than BIT_STATUS_BITS_NUM (unsigned int * 8) */
679: #define ONIG_MAX_CAPTURE_HISTORY_GROUP 31
680: #define ONIG_IS_CAPTURE_HISTORY_GROUP(r, i) \
681: ((i) <= ONIG_MAX_CAPTURE_HISTORY_GROUP && (r)->list && (r)->list[i])
682:
683: typedef struct OnigCaptureTreeNodeStruct {
684: int group; /* group number */
685: int beg;
686: int end;
687: int allocated;
688: int num_childs;
689: struct OnigCaptureTreeNodeStruct** childs;
690: } OnigCaptureTreeNode;
691:
692: /* match result region type */
693: struct re_registers {
694: int allocated;
695: int num_regs;
696: int* beg;
697: int* end;
698: /* extended */
699: OnigCaptureTreeNode* history_root; /* capture history tree root */
700: };
701:
702: /* capture tree traverse */
703: #define ONIG_TRAVERSE_CALLBACK_AT_FIRST 1
704: #define ONIG_TRAVERSE_CALLBACK_AT_LAST 2
705: #define ONIG_TRAVERSE_CALLBACK_AT_BOTH \
706: ( ONIG_TRAVERSE_CALLBACK_AT_FIRST | ONIG_TRAVERSE_CALLBACK_AT_LAST )
707:
708:
709: #define ONIG_REGION_NOTPOS -1
710:
711: typedef struct re_registers OnigRegion;
712:
713: typedef struct {
714: OnigEncoding enc;
715: OnigUChar* par;
716: OnigUChar* par_end;
717: } OnigErrorInfo;
718:
719: typedef struct {
720: int lower;
721: int upper;
722: } OnigRepeatRange;
723:
724: typedef void (*OnigWarnFunc) P_((const char* s));
725: extern void onig_null_warn P_((const char* s));
726: #define ONIG_NULL_WARN onig_null_warn
727:
728: #define ONIG_CHAR_TABLE_SIZE 256
729:
730: /* regex_t state */
731: #define ONIG_STATE_NORMAL 0
732: #define ONIG_STATE_SEARCHING 1
733: #define ONIG_STATE_COMPILING -1
734: #define ONIG_STATE_MODIFY -2
735:
736: #define ONIG_STATE(reg) \
737: ((reg)->state > 0 ? ONIG_STATE_SEARCHING : (reg)->state)
738:
739: typedef struct re_pattern_buffer {
740: /* common members of BBuf(bytes-buffer) */
741: unsigned char* p; /* compiled pattern */
742: unsigned int used; /* used space for p */
743: unsigned int alloc; /* allocated space for p */
744:
745: int state; /* normal, searching, compiling */
746: int num_mem; /* used memory(...) num counted from 1 */
747: int num_repeat; /* OP_REPEAT/OP_REPEAT_NG id-counter */
748: int num_null_check; /* OP_NULL_CHECK_START/END id counter */
749: int num_comb_exp_check; /* combination explosion check */
750: int num_call; /* number of subexp call */
751: unsigned int capture_history; /* (?@...) flag (1-31) */
752: unsigned int bt_mem_start; /* need backtrack flag */
753: unsigned int bt_mem_end; /* need backtrack flag */
754: int stack_pop_level;
755: int repeat_range_alloc;
756: OnigRepeatRange* repeat_range;
757:
758: OnigEncoding enc;
759: OnigOptionType options;
760: OnigSyntaxType* syntax;
761: OnigAmbigType ambig_flag;
762: void* name_table;
763:
764: /* optimization info (string search, char-map and anchors) */
765: int optimize; /* optimize flag */
766: int threshold_len; /* search str-length for apply optimize */
767: int anchor; /* BEGIN_BUF, BEGIN_POS, (SEMI_)END_BUF */
768: OnigDistance anchor_dmin; /* (SEMI_)END_BUF anchor distance */
769: OnigDistance anchor_dmax; /* (SEMI_)END_BUF anchor distance */
770: int sub_anchor; /* start-anchor for exact or map */
771: unsigned char *exact;
772: unsigned char *exact_end;
773: unsigned char map[ONIG_CHAR_TABLE_SIZE]; /* used as BM skip or char-map */
774: int *int_map; /* BM skip for exact_len > 255 */
775: int *int_map_backward; /* BM skip for backward search */
776: OnigDistance dmin; /* min-distance of exact or map */
777: OnigDistance dmax; /* max-distance of exact or map */
778:
779: /* regex_t link chain */
780: struct re_pattern_buffer* chain; /* escape compile-conflict */
781: } OnigRegexType;
782:
783: typedef OnigRegexType* OnigRegex;
784:
785: #ifndef ONIG_ESCAPE_REGEX_T_COLLISION
786: typedef OnigRegexType regex_t;
787: #endif
788:
789:
790: typedef struct {
791: int num_of_elements;
792: OnigEncoding pattern_enc;
793: OnigEncoding target_enc;
794: OnigSyntaxType* syntax;
795: OnigOptionType option;
796: OnigAmbigType ambig_flag;
797: } OnigCompileInfo;
798:
799: /* Oniguruma Native API */
800: ONIG_EXTERN
801: int onig_init P_((void));
802: ONIG_EXTERN
803: int onig_error_code_to_str PV_((OnigUChar* s, int err_code, ...));
804: ONIG_EXTERN
805: void onig_set_warn_func P_((OnigWarnFunc f));
806: ONIG_EXTERN
807: void onig_set_verb_warn_func P_((OnigWarnFunc f));
808: ONIG_EXTERN
809: int onig_new P_((OnigRegex*, const OnigUChar* pattern, const OnigUChar* pattern_end, OnigOptionType option, OnigEncoding enc, OnigSyntaxType* syntax, OnigErrorInfo* einfo));
810: ONIG_EXTERN
811: int onig_new_deluxe P_((OnigRegex* reg, const OnigUChar* pattern, const OnigUChar* pattern_end, OnigCompileInfo* ci, OnigErrorInfo* einfo));
812: ONIG_EXTERN
813: void onig_free P_((OnigRegex));
814: ONIG_EXTERN
815: int onig_recompile P_((OnigRegex, const OnigUChar* pattern, const OnigUChar* pattern_end, OnigOptionType option, OnigEncoding enc, OnigSyntaxType* syntax, OnigErrorInfo* einfo));
816: ONIG_EXTERN
817: int onig_recompile_deluxe P_((OnigRegex reg, const OnigUChar* pattern, const OnigUChar* pattern_end, OnigCompileInfo* ci, OnigErrorInfo* einfo));
818: ONIG_EXTERN
819: int onig_search P_((OnigRegex, const OnigUChar* str, const OnigUChar* end, const OnigUChar* start, const OnigUChar* range, OnigRegion* region, OnigOptionType option));
820: ONIG_EXTERN
821: int onig_match P_((OnigRegex, const OnigUChar* str, const OnigUChar* end, const OnigUChar* at, OnigRegion* region, OnigOptionType option));
822: ONIG_EXTERN
823: OnigRegion* onig_region_new P_((void));
824: ONIG_EXTERN
825: void onig_region_init P_((OnigRegion* region));
826: ONIG_EXTERN
827: void onig_region_free P_((OnigRegion* region, int free_self));
828: ONIG_EXTERN
829: void onig_region_copy P_((OnigRegion* to, OnigRegion* from));
830: ONIG_EXTERN
831: void onig_region_clear P_((OnigRegion* region));
832: ONIG_EXTERN
833: int onig_region_resize P_((OnigRegion* region, int n));
834: ONIG_EXTERN
835: int onig_region_set P_((OnigRegion* region, int at, int beg, int end));
836: ONIG_EXTERN
837: int onig_name_to_group_numbers P_((OnigRegex reg, const OnigUChar* name, const OnigUChar* name_end, int** nums));
838: ONIG_EXTERN
839: int onig_name_to_backref_number P_((OnigRegex reg, const OnigUChar* name, const OnigUChar* name_end, OnigRegion *region));
840: ONIG_EXTERN
841: int onig_foreach_name P_((OnigRegex reg, int (*func)(const OnigUChar*, const OnigUChar*,int,int*,OnigRegex,void*), void* arg));
842: ONIG_EXTERN
843: int onig_number_of_names P_((OnigRegex reg));
844: ONIG_EXTERN
845: int onig_number_of_captures P_((OnigRegex reg));
846: ONIG_EXTERN
847: int onig_number_of_capture_histories P_((OnigRegex reg));
848: ONIG_EXTERN
849: OnigCaptureTreeNode* onig_get_capture_tree P_((OnigRegion* region));
850: ONIG_EXTERN
851: int onig_capture_tree_traverse P_((OnigRegion* region, int at, int(*callback_func)(int,int,int,int,int,void*), void* arg));
852: ONIG_EXTERN
853: int onig_noname_group_capture_is_active P_((OnigRegex reg));
854: ONIG_EXTERN
855: OnigEncoding onig_get_encoding P_((OnigRegex reg));
856: ONIG_EXTERN
857: OnigOptionType onig_get_options P_((OnigRegex reg));
858: ONIG_EXTERN
859: OnigAmbigType onig_get_ambig_flag P_((OnigRegex reg));
860: ONIG_EXTERN
861: OnigSyntaxType* onig_get_syntax P_((OnigRegex reg));
862: ONIG_EXTERN
863: int onig_set_default_syntax P_((OnigSyntaxType* syntax));
864: ONIG_EXTERN
865: void onig_copy_syntax P_((OnigSyntaxType* to, OnigSyntaxType* from));
866: ONIG_EXTERN
867: unsigned int onig_get_syntax_op P_((OnigSyntaxType* syntax));
868: ONIG_EXTERN
869: unsigned int onig_get_syntax_op2 P_((OnigSyntaxType* syntax));
870: ONIG_EXTERN
871: unsigned int onig_get_syntax_behavior P_((OnigSyntaxType* syntax));
872: ONIG_EXTERN
873: OnigOptionType onig_get_syntax_options P_((OnigSyntaxType* syntax));
874: ONIG_EXTERN
875: void onig_set_syntax_op P_((OnigSyntaxType* syntax, unsigned int op));
876: ONIG_EXTERN
877: void onig_set_syntax_op2 P_((OnigSyntaxType* syntax, unsigned int op2));
878: ONIG_EXTERN
879: void onig_set_syntax_behavior P_((OnigSyntaxType* syntax, unsigned int behavior));
880: ONIG_EXTERN
881: void onig_set_syntax_options P_((OnigSyntaxType* syntax, OnigOptionType options));
882: ONIG_EXTERN
883: int onig_set_meta_char P_((OnigEncoding enc, unsigned int what, OnigCodePoint code));
884: ONIG_EXTERN
885: void onig_copy_encoding P_((OnigEncoding to, OnigEncoding from));
886: ONIG_EXTERN
887: OnigAmbigType onig_get_default_ambig_flag P_((void));
888: ONIG_EXTERN
889: int onig_set_default_ambig_flag P_((OnigAmbigType ambig_flag));
890: ONIG_EXTERN
891: unsigned int onig_get_match_stack_limit_size P_((void));
892: ONIG_EXTERN
893: int onig_set_match_stack_limit_size P_((unsigned int size));
894: ONIG_EXTERN
895: int onig_end P_((void));
896: ONIG_EXTERN
897: const char* onig_version P_((void));
898: ONIG_EXTERN
899: const char* onig_copyright P_((void));
900:
901: #ifdef __cplusplus
902: }
903: #endif
904:
905: #endif /* ONIGURUMA_H */
FreeBSD-CVSweb <freebsd-cvsweb@FreeBSD.org>
![[BACK]](/icons/cvsweb/back.gif){kind=icon}
Return to oniguruma.h CVS log ![[TXT]](/icons/cvsweb/text.gif){kind=icon}
![[DIR]](/icons/cvsweb/dir.gif){kind=icon}
Up to [ELWIX - Embedded LightWeight unIX -] / embedaddon / php / ext / mbstring / oniguruma