]>
Commit | Line | Data |
---|---|---|
a997bf42 ÆAB |
1 | #include <stdio.h> |
2 | #include <stddef.h> | |
3 | ||
f05951fe | 4 | /* Definitions for data structures and routines for the regular |
d18f76dc ÆAB |
5 | expression library. |
6 | Copyright (C) 1985,1989-93,1995-98,2000,2001,2002,2003,2005,2006,2008 | |
7 | Free Software Foundation, Inc. | |
8 | This file is part of the GNU C Library. | |
f05951fe | 9 | |
d18f76dc ÆAB |
10 | The GNU C Library is free software; you can redistribute it and/or |
11 | modify it under the terms of the GNU Lesser General Public | |
12 | License as published by the Free Software Foundation; either | |
13 | version 2.1 of the License, or (at your option) any later version. | |
f05951fe | 14 | |
d18f76dc | 15 | The GNU C Library is distributed in the hope that it will be useful, |
f05951fe | 16 | but WITHOUT ANY WARRANTY; without even the implied warranty of |
d18f76dc ÆAB |
17 | MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU |
18 | Lesser General Public License for more details. | |
f05951fe | 19 | |
d18f76dc | 20 | You should have received a copy of the GNU Lesser General Public |
48425792 TZ |
21 | License along with the GNU C Library; if not, see |
22 | <http://www.gnu.org/licenses/>. */ | |
f05951fe | 23 | |
d18f76dc ÆAB |
24 | #ifndef _REGEX_H |
25 | #define _REGEX_H 1 | |
f05951fe | 26 | |
d18f76dc | 27 | #ifdef HAVE_STDDEF_H |
f05951fe JS |
28 | #include <stddef.h> |
29 | #endif | |
30 | ||
d18f76dc ÆAB |
31 | #ifdef HAVE_SYS_TYPES_H |
32 | #include <sys/types.h> | |
33 | #endif | |
34 | ||
35 | #ifndef _LIBC | |
36 | #define __USE_GNU 1 | |
37 | #endif | |
38 | ||
39 | /* Allow the use in C++ code. */ | |
40 | #ifdef __cplusplus | |
41 | extern "C" { | |
42 | #endif | |
43 | ||
44 | /* The following two types have to be signed and unsigned integer type | |
45 | wide enough to hold a value of a pointer. For most ANSI compilers | |
46 | ptrdiff_t and size_t should be likely OK. Still size of these two | |
47 | types is 2 for Microsoft C. Ugh... */ | |
48 | typedef long int s_reg_t; | |
49 | typedef unsigned long int active_reg_t; | |
f05951fe JS |
50 | |
51 | /* The following bits are used to determine the regexp syntax we | |
52 | recognize. The set/not-set meanings are chosen so that Emacs syntax | |
53 | remains the value 0. The bits are given in alphabetical order, and | |
54 | the definitions shifted by one from the previous bit; thus, when we | |
55 | add or remove a bit, only one other definition need change. */ | |
d18f76dc | 56 | typedef unsigned long int reg_syntax_t; |
f05951fe | 57 | |
d18f76dc | 58 | #ifdef __USE_GNU |
f05951fe JS |
59 | /* If this bit is not set, then \ inside a bracket expression is literal. |
60 | If set, then such a \ quotes the following character. */ | |
d18f76dc | 61 | # define RE_BACKSLASH_ESCAPE_IN_LISTS ((unsigned long int) 1) |
f05951fe JS |
62 | |
63 | /* If this bit is not set, then + and ? are operators, and \+ and \? are | |
64 | literals. | |
65 | If set, then \+ and \? are operators and + and ? are literals. */ | |
d18f76dc | 66 | # define RE_BK_PLUS_QM (RE_BACKSLASH_ESCAPE_IN_LISTS << 1) |
f05951fe JS |
67 | |
68 | /* If this bit is set, then character classes are supported. They are: | |
69 | [:alpha:], [:upper:], [:lower:], [:digit:], [:alnum:], [:xdigit:], | |
70 | [:space:], [:print:], [:punct:], [:graph:], and [:cntrl:]. | |
71 | If not set, then character classes are not supported. */ | |
d18f76dc | 72 | # define RE_CHAR_CLASSES (RE_BK_PLUS_QM << 1) |
f05951fe JS |
73 | |
74 | /* If this bit is set, then ^ and $ are always anchors (outside bracket | |
75 | expressions, of course). | |
76 | If this bit is not set, then it depends: | |
d18f76dc ÆAB |
77 | ^ is an anchor if it is at the beginning of a regular |
78 | expression or after an open-group or an alternation operator; | |
79 | $ is an anchor if it is at the end of a regular expression, or | |
80 | before a close-group or an alternation operator. | |
f05951fe JS |
81 | |
82 | This bit could be (re)combined with RE_CONTEXT_INDEP_OPS, because | |
83 | POSIX draft 11.2 says that * etc. in leading positions is undefined. | |
84 | We already implemented a previous draft which made those constructs | |
85 | invalid, though, so we haven't changed the code back. */ | |
d18f76dc | 86 | # define RE_CONTEXT_INDEP_ANCHORS (RE_CHAR_CLASSES << 1) |
f05951fe JS |
87 | |
88 | /* If this bit is set, then special characters are always special | |
89 | regardless of where they are in the pattern. | |
90 | If this bit is not set, then special characters are special only in | |
91 | some contexts; otherwise they are ordinary. Specifically, | |
92 | * + ? and intervals are only special when not after the beginning, | |
93 | open-group, or alternation operator. */ | |
d18f76dc | 94 | # define RE_CONTEXT_INDEP_OPS (RE_CONTEXT_INDEP_ANCHORS << 1) |
f05951fe JS |
95 | |
96 | /* If this bit is set, then *, +, ?, and { cannot be first in an re or | |
97 | immediately after an alternation or begin-group operator. */ | |
d18f76dc | 98 | # define RE_CONTEXT_INVALID_OPS (RE_CONTEXT_INDEP_OPS << 1) |
f05951fe JS |
99 | |
100 | /* If this bit is set, then . matches newline. | |
101 | If not set, then it doesn't. */ | |
d18f76dc | 102 | # define RE_DOT_NEWLINE (RE_CONTEXT_INVALID_OPS << 1) |
f05951fe JS |
103 | |
104 | /* If this bit is set, then . doesn't match NUL. | |
105 | If not set, then it does. */ | |
d18f76dc | 106 | # define RE_DOT_NOT_NULL (RE_DOT_NEWLINE << 1) |
f05951fe JS |
107 | |
108 | /* If this bit is set, nonmatching lists [^...] do not match newline. | |
109 | If not set, they do. */ | |
d18f76dc | 110 | # define RE_HAT_LISTS_NOT_NEWLINE (RE_DOT_NOT_NULL << 1) |
f05951fe JS |
111 | |
112 | /* If this bit is set, either \{...\} or {...} defines an | |
113 | interval, depending on RE_NO_BK_BRACES. | |
114 | If not set, \{, \}, {, and } are literals. */ | |
d18f76dc | 115 | # define RE_INTERVALS (RE_HAT_LISTS_NOT_NEWLINE << 1) |
f05951fe JS |
116 | |
117 | /* If this bit is set, +, ? and | aren't recognized as operators. | |
118 | If not set, they are. */ | |
d18f76dc | 119 | # define RE_LIMITED_OPS (RE_INTERVALS << 1) |
f05951fe JS |
120 | |
121 | /* If this bit is set, newline is an alternation operator. | |
122 | If not set, newline is literal. */ | |
d18f76dc | 123 | # define RE_NEWLINE_ALT (RE_LIMITED_OPS << 1) |
f05951fe JS |
124 | |
125 | /* If this bit is set, then `{...}' defines an interval, and \{ and \} | |
126 | are literals. | |
127 | If not set, then `\{...\}' defines an interval. */ | |
d18f76dc | 128 | # define RE_NO_BK_BRACES (RE_NEWLINE_ALT << 1) |
f05951fe JS |
129 | |
130 | /* If this bit is set, (...) defines a group, and \( and \) are literals. | |
131 | If not set, \(...\) defines a group, and ( and ) are literals. */ | |
d18f76dc | 132 | # define RE_NO_BK_PARENS (RE_NO_BK_BRACES << 1) |
f05951fe JS |
133 | |
134 | /* If this bit is set, then \<digit> matches <digit>. | |
135 | If not set, then \<digit> is a back-reference. */ | |
d18f76dc | 136 | # define RE_NO_BK_REFS (RE_NO_BK_PARENS << 1) |
f05951fe JS |
137 | |
138 | /* If this bit is set, then | is an alternation operator, and \| is literal. | |
139 | If not set, then \| is an alternation operator, and | is literal. */ | |
d18f76dc | 140 | # define RE_NO_BK_VBAR (RE_NO_BK_REFS << 1) |
f05951fe JS |
141 | |
142 | /* If this bit is set, then an ending range point collating higher | |
143 | than the starting range point, as in [z-a], is invalid. | |
144 | If not set, then when ending range point collates higher than the | |
145 | starting range point, the range is ignored. */ | |
d18f76dc | 146 | # define RE_NO_EMPTY_RANGES (RE_NO_BK_VBAR << 1) |
f05951fe JS |
147 | |
148 | /* If this bit is set, then an unmatched ) is ordinary. | |
149 | If not set, then an unmatched ) is invalid. */ | |
d18f76dc ÆAB |
150 | # define RE_UNMATCHED_RIGHT_PAREN_ORD (RE_NO_EMPTY_RANGES << 1) |
151 | ||
152 | /* If this bit is set, succeed as soon as we match the whole pattern, | |
153 | without further backtracking. */ | |
154 | # define RE_NO_POSIX_BACKTRACKING (RE_UNMATCHED_RIGHT_PAREN_ORD << 1) | |
155 | ||
156 | /* If this bit is set, do not process the GNU regex operators. | |
157 | If not set, then the GNU regex operators are recognized. */ | |
158 | # define RE_NO_GNU_OPS (RE_NO_POSIX_BACKTRACKING << 1) | |
159 | ||
160 | /* If this bit is set, a syntactically invalid interval is treated as | |
161 | a string of ordinary characters. For example, the ERE 'a{1' is | |
162 | treated as 'a\{1'. */ | |
163 | # define RE_INVALID_INTERVAL_ORD (RE_NO_GNU_OPS << 1) | |
164 | ||
165 | /* If this bit is set, then ignore case when matching. | |
166 | If not set, then case is significant. */ | |
167 | # define RE_ICASE (RE_INVALID_INTERVAL_ORD << 1) | |
168 | ||
169 | /* This bit is used internally like RE_CONTEXT_INDEP_ANCHORS but only | |
170 | for ^, because it is difficult to scan the regex backwards to find | |
171 | whether ^ should be special. */ | |
172 | # define RE_CARET_ANCHORS_HERE (RE_ICASE << 1) | |
173 | ||
174 | /* If this bit is set, then \{ cannot be first in an bre or | |
175 | immediately after an alternation or begin-group operator. */ | |
176 | # define RE_CONTEXT_INVALID_DUP (RE_CARET_ANCHORS_HERE << 1) | |
177 | ||
178 | /* If this bit is set, then no_sub will be set to 1 during | |
179 | re_compile_pattern. */ | |
180 | #define RE_NO_SUB (RE_CONTEXT_INVALID_DUP << 1) | |
181 | #endif | |
f05951fe JS |
182 | |
183 | /* This global variable defines the particular regexp syntax to use (for | |
184 | some interfaces). When a regexp is compiled, the syntax used is | |
185 | stored in the pattern buffer, so changing this does not affect | |
186 | already-compiled regexps. */ | |
187 | extern reg_syntax_t re_syntax_options; | |
188 | \f | |
d18f76dc | 189 | #ifdef __USE_GNU |
f05951fe JS |
190 | /* Define combinations of the above bits for the standard possibilities. |
191 | (The [[[ comments delimit what gets put into the Texinfo file, so | |
192 | don't delete them!) */ | |
193 | /* [[[begin syntaxes]]] */ | |
194 | #define RE_SYNTAX_EMACS 0 | |
195 | ||
196 | #define RE_SYNTAX_AWK \ | |
d18f76dc ÆAB |
197 | (RE_BACKSLASH_ESCAPE_IN_LISTS | RE_DOT_NOT_NULL \ |
198 | | RE_NO_BK_PARENS | RE_NO_BK_REFS \ | |
199 | | RE_NO_BK_VBAR | RE_NO_EMPTY_RANGES \ | |
200 | | RE_DOT_NEWLINE | RE_CONTEXT_INDEP_ANCHORS \ | |
201 | | RE_UNMATCHED_RIGHT_PAREN_ORD | RE_NO_GNU_OPS) | |
202 | ||
203 | #define RE_SYNTAX_GNU_AWK \ | |
204 | ((RE_SYNTAX_POSIX_EXTENDED | RE_BACKSLASH_ESCAPE_IN_LISTS \ | |
205 | | RE_INVALID_INTERVAL_ORD) \ | |
206 | & ~(RE_DOT_NOT_NULL | RE_CONTEXT_INDEP_OPS \ | |
207 | | RE_CONTEXT_INVALID_OPS )) | |
208 | ||
209 | #define RE_SYNTAX_POSIX_AWK \ | |
210 | (RE_SYNTAX_POSIX_EXTENDED | RE_BACKSLASH_ESCAPE_IN_LISTS \ | |
211 | | RE_INTERVALS | RE_NO_GNU_OPS \ | |
212 | | RE_INVALID_INTERVAL_ORD) | |
f05951fe JS |
213 | |
214 | #define RE_SYNTAX_GREP \ | |
215 | (RE_BK_PLUS_QM | RE_CHAR_CLASSES \ | |
216 | | RE_HAT_LISTS_NOT_NEWLINE | RE_INTERVALS \ | |
217 | | RE_NEWLINE_ALT) | |
218 | ||
219 | #define RE_SYNTAX_EGREP \ | |
220 | (RE_CHAR_CLASSES | RE_CONTEXT_INDEP_ANCHORS \ | |
221 | | RE_CONTEXT_INDEP_OPS | RE_HAT_LISTS_NOT_NEWLINE \ | |
222 | | RE_NEWLINE_ALT | RE_NO_BK_PARENS \ | |
223 | | RE_NO_BK_VBAR) | |
224 | ||
225 | #define RE_SYNTAX_POSIX_EGREP \ | |
d18f76dc ÆAB |
226 | (RE_SYNTAX_EGREP | RE_INTERVALS | RE_NO_BK_BRACES \ |
227 | | RE_INVALID_INTERVAL_ORD) | |
f05951fe JS |
228 | |
229 | /* P1003.2/D11.2, section 4.20.7.1, lines 5078ff. */ | |
230 | #define RE_SYNTAX_ED RE_SYNTAX_POSIX_BASIC | |
231 | ||
232 | #define RE_SYNTAX_SED RE_SYNTAX_POSIX_BASIC | |
233 | ||
234 | /* Syntax bits common to both basic and extended POSIX regex syntax. */ | |
235 | #define _RE_SYNTAX_POSIX_COMMON \ | |
236 | (RE_CHAR_CLASSES | RE_DOT_NEWLINE | RE_DOT_NOT_NULL \ | |
237 | | RE_INTERVALS | RE_NO_EMPTY_RANGES) | |
238 | ||
239 | #define RE_SYNTAX_POSIX_BASIC \ | |
d18f76dc | 240 | (_RE_SYNTAX_POSIX_COMMON | RE_BK_PLUS_QM | RE_CONTEXT_INVALID_DUP) |
f05951fe JS |
241 | |
242 | /* Differs from ..._POSIX_BASIC only in that RE_BK_PLUS_QM becomes | |
243 | RE_LIMITED_OPS, i.e., \? \+ \| are not recognized. Actually, this | |
244 | isn't minimal, since other operators, such as \`, aren't disabled. */ | |
245 | #define RE_SYNTAX_POSIX_MINIMAL_BASIC \ | |
246 | (_RE_SYNTAX_POSIX_COMMON | RE_LIMITED_OPS) | |
247 | ||
248 | #define RE_SYNTAX_POSIX_EXTENDED \ | |
d18f76dc ÆAB |
249 | (_RE_SYNTAX_POSIX_COMMON | RE_CONTEXT_INDEP_ANCHORS \ |
250 | | RE_CONTEXT_INDEP_OPS | RE_NO_BK_BRACES \ | |
251 | | RE_NO_BK_PARENS | RE_NO_BK_VBAR \ | |
252 | | RE_CONTEXT_INVALID_OPS | RE_UNMATCHED_RIGHT_PAREN_ORD) | |
f05951fe | 253 | |
d18f76dc ÆAB |
254 | /* Differs from ..._POSIX_EXTENDED in that RE_CONTEXT_INDEP_OPS is |
255 | removed and RE_NO_BK_REFS is added. */ | |
f05951fe JS |
256 | #define RE_SYNTAX_POSIX_MINIMAL_EXTENDED \ |
257 | (_RE_SYNTAX_POSIX_COMMON | RE_CONTEXT_INDEP_ANCHORS \ | |
258 | | RE_CONTEXT_INVALID_OPS | RE_NO_BK_BRACES \ | |
259 | | RE_NO_BK_PARENS | RE_NO_BK_REFS \ | |
260 | | RE_NO_BK_VBAR | RE_UNMATCHED_RIGHT_PAREN_ORD) | |
261 | /* [[[end syntaxes]]] */ | |
262 | \f | |
263 | /* Maximum number of duplicates an interval can allow. Some systems | |
264 | (erroneously) define this in other header files, but we want our | |
265 | value, so remove any previous define. */ | |
d18f76dc ÆAB |
266 | # ifdef RE_DUP_MAX |
267 | # undef RE_DUP_MAX | |
268 | # endif | |
269 | /* If sizeof(int) == 2, then ((1 << 15) - 1) overflows. */ | |
270 | # define RE_DUP_MAX (0x7fff) | |
f05951fe | 271 | #endif |
f05951fe JS |
272 | |
273 | ||
274 | /* POSIX `cflags' bits (i.e., information for `regcomp'). */ | |
275 | ||
276 | /* If this bit is set, then use extended regular expression syntax. | |
277 | If not set, then use basic regular expression syntax. */ | |
278 | #define REG_EXTENDED 1 | |
279 | ||
280 | /* If this bit is set, then ignore case when matching. | |
281 | If not set, then case is significant. */ | |
282 | #define REG_ICASE (REG_EXTENDED << 1) | |
283 | ||
284 | /* If this bit is set, then anchors do not match at newline | |
285 | characters in the string. | |
286 | If not set, then anchors do match at newlines. */ | |
287 | #define REG_NEWLINE (REG_ICASE << 1) | |
288 | ||
289 | /* If this bit is set, then report only success or fail in regexec. | |
290 | If not set, then returns differ between not matching and errors. */ | |
291 | #define REG_NOSUB (REG_NEWLINE << 1) | |
292 | ||
293 | ||
294 | /* POSIX `eflags' bits (i.e., information for regexec). */ | |
295 | ||
296 | /* If this bit is set, then the beginning-of-line operator doesn't match | |
297 | the beginning of the string (presumably because it's not the | |
298 | beginning of a line). | |
299 | If not set, then the beginning-of-line operator does match the | |
300 | beginning of the string. */ | |
301 | #define REG_NOTBOL 1 | |
302 | ||
303 | /* Like REG_NOTBOL, except for the end-of-line. */ | |
304 | #define REG_NOTEOL (1 << 1) | |
305 | ||
d18f76dc ÆAB |
306 | /* Use PMATCH[0] to delimit the start and end of the search in the |
307 | buffer. */ | |
308 | #define REG_STARTEND (1 << 2) | |
309 | ||
f05951fe JS |
310 | |
311 | /* If any error codes are removed, changed, or added, update the | |
312 | `re_error_msg' table in regex.c. */ | |
313 | typedef enum | |
314 | { | |
d18f76dc ÆAB |
315 | #if defined _XOPEN_SOURCE || defined __USE_XOPEN2K |
316 | REG_ENOSYS = -1, /* This will never happen for this implementation. */ | |
317 | #endif | |
318 | ||
f05951fe JS |
319 | REG_NOERROR = 0, /* Success. */ |
320 | REG_NOMATCH, /* Didn't find a match (for regexec). */ | |
321 | ||
322 | /* POSIX regcomp return error codes. (In the order listed in the | |
323 | standard.) */ | |
324 | REG_BADPAT, /* Invalid pattern. */ | |
d18f76dc | 325 | REG_ECOLLATE, /* Inalid collating element. */ |
f05951fe JS |
326 | REG_ECTYPE, /* Invalid character class name. */ |
327 | REG_EESCAPE, /* Trailing backslash. */ | |
328 | REG_ESUBREG, /* Invalid back reference. */ | |
329 | REG_EBRACK, /* Unmatched left bracket. */ | |
330 | REG_EPAREN, /* Parenthesis imbalance. */ | |
331 | REG_EBRACE, /* Unmatched \{. */ | |
332 | REG_BADBR, /* Invalid contents of \{\}. */ | |
333 | REG_ERANGE, /* Invalid range end. */ | |
334 | REG_ESPACE, /* Ran out of memory. */ | |
335 | REG_BADRPT, /* No preceding re for repetition op. */ | |
336 | ||
337 | /* Error codes we've added. */ | |
338 | REG_EEND, /* Premature end. */ | |
339 | REG_ESIZE, /* Compiled pattern bigger than 2^16 bytes. */ | |
340 | REG_ERPAREN /* Unmatched ) or \); not returned from regcomp. */ | |
341 | } reg_errcode_t; | |
342 | \f | |
343 | /* This data structure represents a compiled pattern. Before calling | |
344 | the pattern compiler, the fields `buffer', `allocated', `fastmap', | |
345 | `translate', and `no_sub' can be set. After the pattern has been | |
346 | compiled, the `re_nsub' field is available. All other fields are | |
347 | private to the regex routines. */ | |
348 | ||
d18f76dc ÆAB |
349 | #ifndef RE_TRANSLATE_TYPE |
350 | # define __RE_TRANSLATE_TYPE unsigned char * | |
351 | # ifdef __USE_GNU | |
352 | # define RE_TRANSLATE_TYPE __RE_TRANSLATE_TYPE | |
353 | # endif | |
354 | #endif | |
355 | ||
356 | #ifdef __USE_GNU | |
357 | # define __REPB_PREFIX(name) name | |
358 | #else | |
359 | # define __REPB_PREFIX(name) __##name | |
360 | #endif | |
361 | ||
f05951fe JS |
362 | struct re_pattern_buffer |
363 | { | |
d18f76dc ÆAB |
364 | /* Space that holds the compiled pattern. It is declared as |
365 | `unsigned char *' because its elements are sometimes used as | |
366 | array indexes. */ | |
367 | unsigned char *__REPB_PREFIX(buffer); | |
f05951fe | 368 | |
d18f76dc ÆAB |
369 | /* Number of bytes to which `buffer' points. */ |
370 | unsigned long int __REPB_PREFIX(allocated); | |
f05951fe | 371 | |
d18f76dc ÆAB |
372 | /* Number of bytes actually used in `buffer'. */ |
373 | unsigned long int __REPB_PREFIX(used); | |
f05951fe | 374 | |
d18f76dc ÆAB |
375 | /* Syntax setting with which the pattern was compiled. */ |
376 | reg_syntax_t __REPB_PREFIX(syntax); | |
f05951fe | 377 | |
d18f76dc ÆAB |
378 | /* Pointer to a fastmap, if any, otherwise zero. re_search uses the |
379 | fastmap, if there is one, to skip over impossible starting points | |
380 | for matches. */ | |
381 | char *__REPB_PREFIX(fastmap); | |
f05951fe | 382 | |
d18f76dc ÆAB |
383 | /* Either a translate table to apply to all characters before |
384 | comparing them, or zero for no translation. The translation is | |
385 | applied to a pattern when it is compiled and to a string when it | |
386 | is matched. */ | |
387 | __RE_TRANSLATE_TYPE __REPB_PREFIX(translate); | |
f05951fe | 388 | |
d18f76dc | 389 | /* Number of subexpressions found by the compiler. */ |
f05951fe JS |
390 | size_t re_nsub; |
391 | ||
d18f76dc ÆAB |
392 | /* Zero if this pattern cannot match the empty string, one else. |
393 | Well, in truth it's used only in `re_search_2', to see whether or | |
394 | not we should use the fastmap, so we don't set this absolutely | |
395 | perfectly; see `re_compile_fastmap' (the `duplicate' case). */ | |
396 | unsigned __REPB_PREFIX(can_be_null) : 1; | |
397 | ||
398 | /* If REGS_UNALLOCATED, allocate space in the `regs' structure | |
399 | for `max (RE_NREGS, re_nsub + 1)' groups. | |
400 | If REGS_REALLOCATE, reallocate space if necessary. | |
401 | If REGS_FIXED, use what's there. */ | |
402 | #ifdef __USE_GNU | |
403 | # define REGS_UNALLOCATED 0 | |
404 | # define REGS_REALLOCATE 1 | |
405 | # define REGS_FIXED 2 | |
406 | #endif | |
407 | unsigned __REPB_PREFIX(regs_allocated) : 2; | |
f05951fe | 408 | |
d18f76dc ÆAB |
409 | /* Set to zero when `regex_compile' compiles a pattern; set to one |
410 | by `re_compile_fastmap' if it updates the fastmap. */ | |
411 | unsigned __REPB_PREFIX(fastmap_accurate) : 1; | |
412 | ||
413 | /* If set, `re_match_2' does not return information about | |
414 | subexpressions. */ | |
415 | unsigned __REPB_PREFIX(no_sub) : 1; | |
f05951fe | 416 | |
d18f76dc ÆAB |
417 | /* If set, a beginning-of-line anchor doesn't match at the beginning |
418 | of the string. */ | |
419 | unsigned __REPB_PREFIX(not_bol) : 1; | |
420 | ||
421 | /* Similarly for an end-of-line anchor. */ | |
422 | unsigned __REPB_PREFIX(not_eol) : 1; | |
423 | ||
424 | /* If true, an anchor at a newline matches. */ | |
425 | unsigned __REPB_PREFIX(newline_anchor) : 1; | |
426 | }; | |
f05951fe | 427 | |
d18f76dc | 428 | typedef struct re_pattern_buffer regex_t; |
f05951fe JS |
429 | \f |
430 | /* Type for byte offsets within the string. POSIX mandates this. */ | |
431 | typedef int regoff_t; | |
432 | ||
433 | ||
d18f76dc | 434 | #ifdef __USE_GNU |
f05951fe JS |
435 | /* This is the structure we store register match data in. See |
436 | regex.texinfo for a full description of what registers match. */ | |
437 | struct re_registers | |
438 | { | |
439 | unsigned num_regs; | |
440 | regoff_t *start; | |
441 | regoff_t *end; | |
442 | }; | |
443 | ||
444 | ||
445 | /* If `regs_allocated' is REGS_UNALLOCATED in the pattern buffer, | |
446 | `re_match_2' returns information about at least this many registers | |
447 | the first time a `regs' structure is passed. */ | |
d18f76dc ÆAB |
448 | # ifndef RE_NREGS |
449 | # define RE_NREGS 30 | |
450 | # endif | |
f05951fe JS |
451 | #endif |
452 | ||
453 | ||
454 | /* POSIX specification for registers. Aside from the different names than | |
455 | `re_registers', POSIX uses an array of structures, instead of a | |
456 | structure of arrays. */ | |
457 | typedef struct | |
458 | { | |
459 | regoff_t rm_so; /* Byte offset from string's start to substring's start. */ | |
460 | regoff_t rm_eo; /* Byte offset from string's start to substring's end. */ | |
461 | } regmatch_t; | |
462 | \f | |
463 | /* Declarations for routines. */ | |
464 | ||
d18f76dc | 465 | #ifdef __USE_GNU |
f05951fe JS |
466 | /* Sets the current default syntax to SYNTAX, and return the old syntax. |
467 | You can also simply assign to the `re_syntax_options' variable. */ | |
d18f76dc | 468 | extern reg_syntax_t re_set_syntax (reg_syntax_t __syntax); |
f05951fe JS |
469 | |
470 | /* Compile the regular expression PATTERN, with length LENGTH | |
471 | and syntax given by the global `re_syntax_options', into the buffer | |
472 | BUFFER. Return NULL if successful, and an error string if not. */ | |
d18f76dc ÆAB |
473 | extern const char *re_compile_pattern (const char *__pattern, size_t __length, |
474 | struct re_pattern_buffer *__buffer); | |
f05951fe JS |
475 | |
476 | ||
477 | /* Compile a fastmap for the compiled pattern in BUFFER; used to | |
478 | accelerate searches. Return 0 if successful and -2 if was an | |
479 | internal error. */ | |
d18f76dc | 480 | extern int re_compile_fastmap (struct re_pattern_buffer *__buffer); |
f05951fe JS |
481 | |
482 | ||
483 | /* Search in the string STRING (with length LENGTH) for the pattern | |
484 | compiled into BUFFER. Start searching at position START, for RANGE | |
485 | characters. Return the starting position of the match, -1 for no | |
486 | match, or -2 for an internal error. Also return register | |
487 | information in REGS (if REGS and BUFFER->no_sub are nonzero). */ | |
d18f76dc ÆAB |
488 | extern int re_search (struct re_pattern_buffer *__buffer, const char *__cstring, |
489 | int __length, int __start, int __range, | |
490 | struct re_registers *__regs); | |
f05951fe JS |
491 | |
492 | ||
493 | /* Like `re_search', but search in the concatenation of STRING1 and | |
494 | STRING2. Also, stop searching at index START + STOP. */ | |
d18f76dc ÆAB |
495 | extern int re_search_2 (struct re_pattern_buffer *__buffer, |
496 | const char *__string1, int __length1, | |
497 | const char *__string2, int __length2, int __start, | |
498 | int __range, struct re_registers *__regs, int __stop); | |
f05951fe JS |
499 | |
500 | ||
501 | /* Like `re_search', but return how many characters in STRING the regexp | |
502 | in BUFFER matched, starting at position START. */ | |
d18f76dc ÆAB |
503 | extern int re_match (struct re_pattern_buffer *__buffer, const char *__cstring, |
504 | int __length, int __start, struct re_registers *__regs); | |
f05951fe JS |
505 | |
506 | ||
507 | /* Relates to `re_match' as `re_search_2' relates to `re_search'. */ | |
d18f76dc ÆAB |
508 | extern int re_match_2 (struct re_pattern_buffer *__buffer, |
509 | const char *__string1, int __length1, | |
510 | const char *__string2, int __length2, int __start, | |
511 | struct re_registers *__regs, int __stop); | |
f05951fe JS |
512 | |
513 | ||
514 | /* Set REGS to hold NUM_REGS registers, storing them in STARTS and | |
515 | ENDS. Subsequent matches using BUFFER and REGS will use this memory | |
516 | for recording register information. STARTS and ENDS must be | |
517 | allocated with malloc, and must each be at least `NUM_REGS * sizeof | |
518 | (regoff_t)' bytes long. | |
519 | ||
520 | If NUM_REGS == 0, then subsequent matches should allocate their own | |
521 | register data. | |
522 | ||
523 | Unless this function is called, the first search or match using | |
524 | PATTERN_BUFFER will allocate its own register data, without | |
525 | freeing the old data. */ | |
d18f76dc ÆAB |
526 | extern void re_set_registers (struct re_pattern_buffer *__buffer, |
527 | struct re_registers *__regs, | |
528 | unsigned int __num_regs, | |
529 | regoff_t *__starts, regoff_t *__ends); | |
530 | #endif /* Use GNU */ | |
531 | ||
532 | #if defined _REGEX_RE_COMP || (defined _LIBC && defined __USE_BSD) | |
533 | # ifndef _CRAY | |
f05951fe | 534 | /* 4.2 bsd compatibility. */ |
d18f76dc ÆAB |
535 | extern char *re_comp (const char *); |
536 | extern int re_exec (const char *); | |
537 | # endif | |
538 | #endif | |
539 | ||
540 | /* GCC 2.95 and later have "__restrict"; C99 compilers have | |
541 | "restrict", and "configure" may have defined "restrict". */ | |
542 | #ifndef __restrict | |
543 | # if ! (2 < __GNUC__ || (2 == __GNUC__ && 95 <= __GNUC_MINOR__)) | |
544 | # if defined restrict || 199901L <= __STDC_VERSION__ | |
545 | # define __restrict restrict | |
546 | # else | |
547 | # define __restrict | |
548 | # endif | |
549 | # endif | |
550 | #endif | |
551 | /* gcc 3.1 and up support the [restrict] syntax. */ | |
552 | #ifndef __restrict_arr | |
553 | # if (__GNUC__ > 3 || (__GNUC__ == 3 && __GNUC_MINOR__ >= 1)) \ | |
554 | && !defined __GNUG__ | |
555 | # define __restrict_arr __restrict | |
556 | # else | |
557 | # define __restrict_arr | |
558 | # endif | |
559 | #endif | |
f05951fe JS |
560 | |
561 | /* POSIX compatibility. */ | |
d18f76dc ÆAB |
562 | extern int regcomp (regex_t *__restrict __preg, |
563 | const char *__restrict __pattern, | |
564 | int __cflags); | |
565 | ||
566 | extern int regexec (const regex_t *__restrict __preg, | |
567 | const char *__restrict __cstring, size_t __nmatch, | |
568 | regmatch_t __pmatch[__restrict_arr], | |
569 | int __eflags); | |
570 | ||
571 | extern size_t regerror (int __errcode, const regex_t *__restrict __preg, | |
572 | char *__restrict __errbuf, size_t __errbuf_size); | |
573 | ||
574 | extern void regfree (regex_t *__preg); | |
575 | ||
576 | ||
577 | #ifdef __cplusplus | |
578 | } | |
579 | #endif /* C++ */ | |
580 | ||
581 | #endif /* regex.h */ |