1 | /* Definitions for data structures and routines for the regular |
2 | expression library. |
3 | Copyright (C) 1985, 1989-2014 Free Software Foundation, Inc. |
4 | This file is part of the GNU C Library. |
5 | |
6 | The GNU C Library is free software; you can redistribute it and/or |
7 | modify it under the terms of the GNU Lesser General Public |
8 | License as published by the Free Software Foundation; either |
9 | version 2.1 of the License, or (at your option) any later version. |
10 | |
11 | The GNU C Library is distributed in the hope that it will be useful, |
12 | but WITHOUT ANY WARRANTY; without even the implied warranty of |
13 | MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU |
14 | Lesser General Public License for more details. |
15 | |
16 | You should have received a copy of the GNU Lesser General Public |
17 | License along with the GNU C Library; if not, see |
18 | <http://www.gnu.org/licenses/>. */ |
19 | |
20 | #ifndef _REGEX_H |
21 | #define _REGEX_H 1 |
22 | |
23 | #include <sys/types.h> |
24 | #include <gnu/option-groups.h> |
25 | |
26 | /* Allow the use in C++ code. */ |
27 | #ifdef __cplusplus |
28 | extern "C" { |
29 | #endif |
30 | |
31 | /* The following two types have to be signed and unsigned integer type |
32 | wide enough to hold a value of a pointer. For most ANSI compilers |
33 | ptrdiff_t and size_t should be likely OK. Still size of these two |
34 | types is 2 for Microsoft C. Ugh... */ |
35 | typedef long int s_reg_t; |
36 | typedef unsigned long int active_reg_t; |
37 | |
38 | /* The following bits are used to determine the regexp syntax we |
39 | recognize. The set/not-set meanings are chosen so that Emacs syntax |
40 | remains the value 0. The bits are given in alphabetical order, and |
41 | the definitions shifted by one from the previous bit; thus, when we |
42 | add or remove a bit, only one other definition need change. */ |
43 | typedef unsigned long int reg_syntax_t; |
44 | |
45 | #ifdef __USE_GNU |
46 | /* If this bit is not set, then \ inside a bracket expression is literal. |
47 | If set, then such a \ quotes the following character. */ |
48 | # define RE_BACKSLASH_ESCAPE_IN_LISTS ((unsigned long int) 1) |
49 | |
50 | /* If this bit is not set, then + and ? are operators, and \+ and \? are |
51 | literals. |
52 | If set, then \+ and \? are operators and + and ? are literals. */ |
53 | # define RE_BK_PLUS_QM (RE_BACKSLASH_ESCAPE_IN_LISTS << 1) |
54 | |
55 | /* If this bit is set, then character classes are supported. They are: |
56 | [:alpha:], [:upper:], [:lower:], [:digit:], [:alnum:], [:xdigit:], |
57 | [:space:], [:print:], [:punct:], [:graph:], and [:cntrl:]. |
58 | If not set, then character classes are not supported. */ |
59 | # define RE_CHAR_CLASSES (RE_BK_PLUS_QM << 1) |
60 | |
61 | /* If this bit is set, then ^ and $ are always anchors (outside bracket |
62 | expressions, of course). |
63 | If this bit is not set, then it depends: |
64 | ^ is an anchor if it is at the beginning of a regular |
65 | expression or after an open-group or an alternation operator; |
66 | $ is an anchor if it is at the end of a regular expression, or |
67 | before a close-group or an alternation operator. |
68 | |
69 | This bit could be (re)combined with RE_CONTEXT_INDEP_OPS, because |
70 | POSIX draft 11.2 says that * etc. in leading positions is undefined. |
71 | We already implemented a previous draft which made those constructs |
72 | invalid, though, so we haven't changed the code back. */ |
73 | # define RE_CONTEXT_INDEP_ANCHORS (RE_CHAR_CLASSES << 1) |
74 | |
75 | /* If this bit is set, then special characters are always special |
76 | regardless of where they are in the pattern. |
77 | If this bit is not set, then special characters are special only in |
78 | some contexts; otherwise they are ordinary. Specifically, |
79 | * + ? and intervals are only special when not after the beginning, |
80 | open-group, or alternation operator. */ |
81 | # define RE_CONTEXT_INDEP_OPS (RE_CONTEXT_INDEP_ANCHORS << 1) |
82 | |
83 | /* If this bit is set, then *, +, ?, and { cannot be first in an re or |
84 | immediately after an alternation or begin-group operator. */ |
85 | # define RE_CONTEXT_INVALID_OPS (RE_CONTEXT_INDEP_OPS << 1) |
86 | |
87 | /* If this bit is set, then . matches newline. |
88 | If not set, then it doesn't. */ |
89 | # define RE_DOT_NEWLINE (RE_CONTEXT_INVALID_OPS << 1) |
90 | |
91 | /* If this bit is set, then . doesn't match NUL. |
92 | If not set, then it does. */ |
93 | # define RE_DOT_NOT_NULL (RE_DOT_NEWLINE << 1) |
94 | |
95 | /* If this bit is set, nonmatching lists [^...] do not match newline. |
96 | If not set, they do. */ |
97 | # define RE_HAT_LISTS_NOT_NEWLINE (RE_DOT_NOT_NULL << 1) |
98 | |
99 | /* If this bit is set, either \{...\} or {...} defines an |
100 | interval, depending on RE_NO_BK_BRACES. |
101 | If not set, \{, \}, {, and } are literals. */ |
102 | # define RE_INTERVALS (RE_HAT_LISTS_NOT_NEWLINE << 1) |
103 | |
104 | /* If this bit is set, +, ? and | aren't recognized as operators. |
105 | If not set, they are. */ |
106 | # define RE_LIMITED_OPS (RE_INTERVALS << 1) |
107 | |
108 | /* If this bit is set, newline is an alternation operator. |
109 | If not set, newline is literal. */ |
110 | # define RE_NEWLINE_ALT (RE_LIMITED_OPS << 1) |
111 | |
112 | /* If this bit is set, then `{...}' defines an interval, and \{ and \} |
113 | are literals. |
114 | If not set, then `\{...\}' defines an interval. */ |
115 | # define RE_NO_BK_BRACES (RE_NEWLINE_ALT << 1) |
116 | |
117 | /* If this bit is set, (...) defines a group, and \( and \) are literals. |
118 | If not set, \(...\) defines a group, and ( and ) are literals. */ |
119 | # define RE_NO_BK_PARENS (RE_NO_BK_BRACES << 1) |
120 | |
121 | /* If this bit is set, then \<digit> matches <digit>. |
122 | If not set, then \<digit> is a back-reference. */ |
123 | # define RE_NO_BK_REFS (RE_NO_BK_PARENS << 1) |
124 | |
125 | /* If this bit is set, then | is an alternation operator, and \| is literal. |
126 | If not set, then \| is an alternation operator, and | is literal. */ |
127 | # define RE_NO_BK_VBAR (RE_NO_BK_REFS << 1) |
128 | |
129 | /* If this bit is set, then an ending range point collating higher |
130 | than the starting range point, as in [z-a], is invalid. |
131 | If not set, then when ending range point collates higher than the |
132 | starting range point, the range is ignored. */ |
133 | # define RE_NO_EMPTY_RANGES (RE_NO_BK_VBAR << 1) |
134 | |
135 | /* If this bit is set, then an unmatched ) is ordinary. |
136 | If not set, then an unmatched ) is invalid. */ |
137 | # define RE_UNMATCHED_RIGHT_PAREN_ORD (RE_NO_EMPTY_RANGES << 1) |
138 | |
139 | /* If this bit is set, succeed as soon as we match the whole pattern, |
140 | without further backtracking. */ |
141 | # define RE_NO_POSIX_BACKTRACKING (RE_UNMATCHED_RIGHT_PAREN_ORD << 1) |
142 | |
143 | /* If this bit is set, do not process the GNU regex operators. |
144 | If not set, then the GNU regex operators are recognized. */ |
145 | # define RE_NO_GNU_OPS (RE_NO_POSIX_BACKTRACKING << 1) |
146 | |
147 | /* If this bit is set, turn on internal regex debugging. |
148 | If not set, and debugging was on, turn it off. |
149 | This only works if regex.c is compiled -DDEBUG. |
150 | We define this bit always, so that all that's needed to turn on |
151 | debugging is to recompile regex.c; the calling code can always have |
152 | this bit set, and it won't affect anything in the normal case. */ |
153 | # define RE_DEBUG (RE_NO_GNU_OPS << 1) |
154 | |
155 | /* If this bit is set, a syntactically invalid interval is treated as |
156 | a string of ordinary characters. For example, the ERE 'a{1' is |
157 | treated as 'a\{1'. */ |
158 | # define RE_INVALID_INTERVAL_ORD (RE_DEBUG << 1) |
159 | |
160 | /* EGLIBC: Old regex implementation does not support these. */ |
161 | # ifdef __OPTION_POSIX_REGEXP_GLIBC |
162 | /* If this bit is set, then ignore case when matching. |
163 | If not set, then case is significant. */ |
164 | # define RE_ICASE (RE_INVALID_INTERVAL_ORD << 1) |
165 | |
166 | /* This bit is used internally like RE_CONTEXT_INDEP_ANCHORS but only |
167 | for ^, because it is difficult to scan the regex backwards to find |
168 | whether ^ should be special. */ |
169 | # define RE_CARET_ANCHORS_HERE (RE_ICASE << 1) |
170 | |
171 | /* If this bit is set, then \{ cannot be first in an bre or |
172 | immediately after an alternation or begin-group operator. */ |
173 | # define RE_CONTEXT_INVALID_DUP (RE_CARET_ANCHORS_HERE << 1) |
174 | |
175 | /* If this bit is set, then no_sub will be set to 1 during |
176 | re_compile_pattern. */ |
177 | # define RE_NO_SUB (RE_CONTEXT_INVALID_DUP << 1) |
178 | # endif /* __OPTION_POSIX_REGEXP_GLIBC */ |
179 | #endif |
180 | |
181 | /* This global variable defines the particular regexp syntax to use (for |
182 | some interfaces). When a regexp is compiled, the syntax used is |
183 | stored in the pattern buffer, so changing this does not affect |
184 | already-compiled regexps. */ |
185 | extern reg_syntax_t re_syntax_options; |
186 | |
187 | #ifdef __USE_GNU |
188 | /* Define combinations of the above bits for the standard possibilities. |
189 | (The [[[ comments delimit what gets put into the Texinfo file, so |
190 | don't delete them!) */ |
191 | /* [[[begin syntaxes]]] */ |
192 | #define RE_SYNTAX_EMACS 0 |
193 | |
194 | #define RE_SYNTAX_AWK \ |
195 | (RE_BACKSLASH_ESCAPE_IN_LISTS | RE_DOT_NOT_NULL \ |
196 | | RE_NO_BK_PARENS | RE_NO_BK_REFS \ |
197 | | RE_NO_BK_VBAR | RE_NO_EMPTY_RANGES \ |
198 | | RE_DOT_NEWLINE | RE_CONTEXT_INDEP_ANCHORS \ |
199 | | RE_CHAR_CLASSES \ |
200 | | RE_UNMATCHED_RIGHT_PAREN_ORD | RE_NO_GNU_OPS) |
201 | |
202 | #define RE_SYNTAX_GNU_AWK \ |
203 | ((RE_SYNTAX_POSIX_EXTENDED | RE_BACKSLASH_ESCAPE_IN_LISTS \ |
204 | | RE_INVALID_INTERVAL_ORD) \ |
205 | & ~(RE_DOT_NOT_NULL | RE_CONTEXT_INDEP_OPS \ |
206 | | RE_CONTEXT_INVALID_OPS )) |
207 | |
208 | #define RE_SYNTAX_POSIX_AWK \ |
209 | (RE_SYNTAX_POSIX_EXTENDED | RE_BACKSLASH_ESCAPE_IN_LISTS \ |
210 | | RE_INTERVALS | RE_NO_GNU_OPS \ |
211 | | RE_INVALID_INTERVAL_ORD) |
212 | |
213 | #define RE_SYNTAX_GREP \ |
214 | (RE_BK_PLUS_QM | RE_CHAR_CLASSES \ |
215 | | RE_HAT_LISTS_NOT_NEWLINE | RE_INTERVALS \ |
216 | | RE_NEWLINE_ALT) |
217 | |
218 | #define RE_SYNTAX_EGREP \ |
219 | (RE_CHAR_CLASSES | RE_CONTEXT_INDEP_ANCHORS \ |
220 | | RE_CONTEXT_INDEP_OPS | RE_HAT_LISTS_NOT_NEWLINE \ |
221 | | RE_NEWLINE_ALT | RE_NO_BK_PARENS \ |
222 | | RE_NO_BK_VBAR) |
223 | |
224 | #define RE_SYNTAX_POSIX_EGREP \ |
225 | (RE_SYNTAX_EGREP | RE_INTERVALS | RE_NO_BK_BRACES \ |
226 | | RE_INVALID_INTERVAL_ORD) |
227 | |
228 | /* P1003.2/D11.2, section 4.20.7.1, lines 5078ff. */ |
229 | #define RE_SYNTAX_ED RE_SYNTAX_POSIX_BASIC |
230 | |
231 | #define RE_SYNTAX_SED RE_SYNTAX_POSIX_BASIC |
232 | |
233 | /* Syntax bits common to both basic and extended POSIX regex syntax. */ |
234 | #define _RE_SYNTAX_POSIX_COMMON \ |
235 | (RE_CHAR_CLASSES | RE_DOT_NEWLINE | RE_DOT_NOT_NULL \ |
236 | | RE_INTERVALS | RE_NO_EMPTY_RANGES) |
237 | |
238 | #ifdef __OPTION_POSIX_REGEXP_GLIBC |
239 | #define RE_SYNTAX_POSIX_BASIC \ |
240 | (_RE_SYNTAX_POSIX_COMMON | RE_BK_PLUS_QM | RE_CONTEXT_INVALID_DUP) |
241 | #else |
242 | #define RE_SYNTAX_POSIX_BASIC \ |
243 | (_RE_SYNTAX_POSIX_COMMON | RE_BK_PLUS_QM) |
244 | #endif |
245 | |
246 | /* Differs from ..._POSIX_BASIC only in that RE_BK_PLUS_QM becomes |
247 | RE_LIMITED_OPS, i.e., \? \+ \| are not recognized. Actually, this |
248 | isn't minimal, since other operators, such as \`, aren't disabled. */ |
249 | #define RE_SYNTAX_POSIX_MINIMAL_BASIC \ |
250 | (_RE_SYNTAX_POSIX_COMMON | RE_LIMITED_OPS) |
251 | |
252 | #define RE_SYNTAX_POSIX_EXTENDED \ |
253 | (_RE_SYNTAX_POSIX_COMMON | RE_CONTEXT_INDEP_ANCHORS \ |
254 | | RE_CONTEXT_INDEP_OPS | RE_NO_BK_BRACES \ |
255 | | RE_NO_BK_PARENS | RE_NO_BK_VBAR \ |
256 | | RE_CONTEXT_INVALID_OPS | RE_UNMATCHED_RIGHT_PAREN_ORD) |
257 | |
258 | /* Differs from ..._POSIX_EXTENDED in that RE_CONTEXT_INDEP_OPS is |
259 | removed and RE_NO_BK_REFS is added. */ |
260 | #define RE_SYNTAX_POSIX_MINIMAL_EXTENDED \ |
261 | (_RE_SYNTAX_POSIX_COMMON | RE_CONTEXT_INDEP_ANCHORS \ |
262 | | RE_CONTEXT_INVALID_OPS | RE_NO_BK_BRACES \ |
263 | | RE_NO_BK_PARENS | RE_NO_BK_REFS \ |
264 | | RE_NO_BK_VBAR | RE_UNMATCHED_RIGHT_PAREN_ORD) |
265 | /* [[[end syntaxes]]] */ |
266 | |
267 | /* Maximum number of duplicates an interval can allow. Some systems |
268 | (erroneously) define this in other header files, but we want our |
269 | value, so remove any previous define. */ |
270 | # ifdef RE_DUP_MAX |
271 | # undef RE_DUP_MAX |
272 | # endif |
273 | /* If sizeof(int) == 2, then ((1 << 15) - 1) overflows. */ |
274 | # define RE_DUP_MAX (0x7fff) |
275 | #endif |
276 | |
277 | |
278 | /* POSIX `cflags' bits (i.e., information for `regcomp'). */ |
279 | |
280 | /* If this bit is set, then use extended regular expression syntax. |
281 | If not set, then use basic regular expression syntax. */ |
282 | #define REG_EXTENDED 1 |
283 | |
284 | /* If this bit is set, then ignore case when matching. |
285 | If not set, then case is significant. */ |
286 | #define REG_ICASE (REG_EXTENDED << 1) |
287 | |
288 | /* If this bit is set, then anchors do not match at newline |
289 | characters in the string. |
290 | If not set, then anchors do match at newlines. */ |
291 | #define REG_NEWLINE (REG_ICASE << 1) |
292 | |
293 | /* If this bit is set, then report only success or fail in regexec. |
294 | If not set, then returns differ between not matching and errors. */ |
295 | #define REG_NOSUB (REG_NEWLINE << 1) |
296 | |
297 | |
298 | /* POSIX `eflags' bits (i.e., information for regexec). */ |
299 | |
300 | /* If this bit is set, then the beginning-of-line operator doesn't match |
301 | the beginning of the string (presumably because it's not the |
302 | beginning of a line). |
303 | If not set, then the beginning-of-line operator does match the |
304 | beginning of the string. */ |
305 | #define REG_NOTBOL 1 |
306 | |
307 | /* Like REG_NOTBOL, except for the end-of-line. */ |
308 | #define REG_NOTEOL (1 << 1) |
309 | |
310 | #ifdef __OPTION_POSIX_REGEXP_GLIBC |
311 | /* Use PMATCH[0] to delimit the start and end of the search in the |
312 | buffer. */ |
313 | #define REG_STARTEND (1 << 2) |
314 | #endif |
315 | |
316 | |
317 | /* If any error codes are removed, changed, or added, update the |
318 | `re_error_msg' table in regex.c. */ |
319 | typedef enum |
320 | { |
321 | #if defined _XOPEN_SOURCE || defined __USE_XOPEN2K |
322 | REG_ENOSYS = -1, /* This will never happen for this implementation. */ |
323 | #endif |
324 | |
325 | REG_NOERROR = 0, /* Success. */ |
326 | REG_NOMATCH, /* Didn't find a match (for regexec). */ |
327 | |
328 | /* POSIX regcomp return error codes. (In the order listed in the |
329 | standard.) */ |
330 | REG_BADPAT, /* Invalid pattern. */ |
331 | REG_ECOLLATE, /* Inalid collating element. */ |
332 | REG_ECTYPE, /* Invalid character class name. */ |
333 | REG_EESCAPE, /* Trailing backslash. */ |
334 | REG_ESUBREG, /* Invalid back reference. */ |
335 | REG_EBRACK, /* Unmatched left bracket. */ |
336 | REG_EPAREN, /* Parenthesis imbalance. */ |
337 | REG_EBRACE, /* Unmatched \{. */ |
338 | REG_BADBR, /* Invalid contents of \{\}. */ |
339 | REG_ERANGE, /* Invalid range end. */ |
340 | REG_ESPACE, /* Ran out of memory. */ |
341 | REG_BADRPT, /* No preceding re for repetition op. */ |
342 | |
343 | /* Error codes we've added. */ |
344 | REG_EEND, /* Premature end. */ |
345 | REG_ESIZE, /* Compiled pattern bigger than 2^16 bytes. */ |
346 | REG_ERPAREN /* Unmatched ) or \); not returned from regcomp. */ |
347 | } reg_errcode_t; |
348 | |
349 | /* This data structure represents a compiled pattern. Before calling |
350 | the pattern compiler, the fields `buffer', `allocated', `fastmap', |
351 | and `translate' can be set. After the pattern has been compiled, |
352 | the fields `re_nsub', `not_bol' and `not_eol' are available. All |
353 | other fields are private to the regex routines. */ |
354 | |
355 | #ifndef RE_TRANSLATE_TYPE |
356 | # define __RE_TRANSLATE_TYPE unsigned char * |
357 | # ifdef __USE_GNU |
358 | # define RE_TRANSLATE_TYPE __RE_TRANSLATE_TYPE |
359 | # endif |
360 | #endif |
361 | |
362 | #ifdef __USE_GNU |
363 | # define __REPB_PREFIX(name) name |
364 | #else |
365 | # define __REPB_PREFIX(name) __##name |
366 | #endif |
367 | |
368 | struct re_pattern_buffer |
369 | { |
370 | /* Space that holds the compiled pattern. It is declared as |
371 | `unsigned char *' because its elements are sometimes used as |
372 | array indexes. */ |
373 | unsigned char *__REPB_PREFIX(buffer); |
374 | |
375 | /* Number of bytes to which `buffer' points. */ |
376 | unsigned long int __REPB_PREFIX(allocated); |
377 | |
378 | /* Number of bytes actually used in `buffer'. */ |
379 | unsigned long int __REPB_PREFIX(used); |
380 | |
381 | /* Syntax setting with which the pattern was compiled. */ |
382 | reg_syntax_t __REPB_PREFIX(syntax); |
383 | |
384 | /* Pointer to a fastmap, if any, otherwise zero. re_search uses the |
385 | fastmap, if there is one, to skip over impossible starting points |
386 | for matches. */ |
387 | char *__REPB_PREFIX(fastmap); |
388 | |
389 | /* Either a translate table to apply to all characters before |
390 | comparing them, or zero for no translation. The translation is |
391 | applied to a pattern when it is compiled and to a string when it |
392 | is matched. */ |
393 | __RE_TRANSLATE_TYPE __REPB_PREFIX(translate); |
394 | |
395 | /* Number of subexpressions found by the compiler. */ |
396 | size_t re_nsub; |
397 | |
398 | /* Zero if this pattern cannot match the empty string, one else. |
399 | Well, in truth it's used only in `re_search_2', to see whether or |
400 | not we should use the fastmap, so we don't set this absolutely |
401 | perfectly; see `re_compile_fastmap' (the `duplicate' case). */ |
402 | unsigned __REPB_PREFIX(can_be_null) : 1; |
403 | |
404 | /* If REGS_UNALLOCATED, allocate space in the `regs' structure |
405 | for `max (RE_NREGS, re_nsub + 1)' groups. |
406 | If REGS_REALLOCATE, reallocate space if necessary. |
407 | If REGS_FIXED, use what's there. */ |
408 | #ifdef __USE_GNU |
409 | # define REGS_UNALLOCATED 0 |
410 | # define REGS_REALLOCATE 1 |
411 | # define REGS_FIXED 2 |
412 | #endif |
413 | unsigned __REPB_PREFIX(regs_allocated) : 2; |
414 | |
415 | /* Set to zero when `regex_compile' compiles a pattern; set to one |
416 | by `re_compile_fastmap' if it updates the fastmap. */ |
417 | unsigned __REPB_PREFIX(fastmap_accurate) : 1; |
418 | |
419 | /* If set, `re_match_2' does not return information about |
420 | subexpressions. */ |
421 | unsigned __REPB_PREFIX(no_sub) : 1; |
422 | |
423 | /* If set, a beginning-of-line anchor doesn't match at the beginning |
424 | of the string. */ |
425 | unsigned __REPB_PREFIX(not_bol) : 1; |
426 | |
427 | /* Similarly for an end-of-line anchor. */ |
428 | unsigned __REPB_PREFIX(not_eol) : 1; |
429 | |
430 | /* If true, an anchor at a newline matches. */ |
431 | unsigned __REPB_PREFIX(newline_anchor) : 1; |
432 | }; |
433 | |
434 | typedef struct re_pattern_buffer regex_t; |
435 | |
436 | /* Type for byte offsets within the string. POSIX mandates this. */ |
437 | typedef int regoff_t; |
438 | |
439 | |
440 | #ifdef __USE_GNU |
441 | /* This is the structure we store register match data in. See |
442 | regex.texinfo for a full description of what registers match. */ |
443 | struct re_registers |
444 | { |
445 | unsigned num_regs; |
446 | regoff_t *start; |
447 | regoff_t *end; |
448 | }; |
449 | |
450 | |
451 | /* If `regs_allocated' is REGS_UNALLOCATED in the pattern buffer, |
452 | `re_match_2' returns information about at least this many registers |
453 | the first time a `regs' structure is passed. */ |
454 | # ifndef RE_NREGS |
455 | # define RE_NREGS 30 |
456 | # endif |
457 | #endif |
458 | |
459 | |
460 | /* POSIX specification for registers. Aside from the different names than |
461 | `re_registers', POSIX uses an array of structures, instead of a |
462 | structure of arrays. */ |
463 | typedef struct |
464 | { |
465 | regoff_t rm_so; /* Byte offset from string's start to substring's start. */ |
466 | regoff_t rm_eo; /* Byte offset from string's start to substring's end. */ |
467 | } regmatch_t; |
468 | |
469 | /* Declarations for routines. */ |
470 | |
471 | #ifdef __USE_GNU |
472 | /* Sets the current default syntax to SYNTAX, and return the old syntax. |
473 | You can also simply assign to the `re_syntax_options' variable. */ |
474 | extern reg_syntax_t re_set_syntax (reg_syntax_t __syntax); |
475 | |
476 | /* Compile the regular expression PATTERN, with length LENGTH |
477 | and syntax given by the global `re_syntax_options', into the buffer |
478 | BUFFER. Return NULL if successful, and an error string if not. |
479 | |
480 | To free the allocated storage, you must call `regfree' on BUFFER. |
481 | Note that the translate table must either have been initialised by |
482 | `regcomp', with a malloc'ed value, or set to NULL before calling |
483 | `regfree'. */ |
484 | extern const char *re_compile_pattern (const char *__pattern, size_t __length, |
485 | struct re_pattern_buffer *__buffer); |
486 | |
487 | |
488 | /* Compile a fastmap for the compiled pattern in BUFFER; used to |
489 | accelerate searches. Return 0 if successful and -2 if was an |
490 | internal error. */ |
491 | extern int re_compile_fastmap (struct re_pattern_buffer *__buffer); |
492 | |
493 | |
494 | /* Search in the string STRING (with length LENGTH) for the pattern |
495 | compiled into BUFFER. Start searching at position START, for RANGE |
496 | characters. Return the starting position of the match, -1 for no |
497 | match, or -2 for an internal error. Also return register |
498 | information in REGS (if REGS and BUFFER->no_sub are nonzero). */ |
499 | extern int re_search (struct re_pattern_buffer *__buffer, const char *__string, |
500 | int __length, int __start, int __range, |
501 | struct re_registers *__regs); |
502 | |
503 | |
504 | /* Like `re_search', but search in the concatenation of STRING1 and |
505 | STRING2. Also, stop searching at index START + STOP. */ |
506 | extern int re_search_2 (struct re_pattern_buffer *__buffer, |
507 | const char *__string1, int __length1, |
508 | const char *__string2, int __length2, int __start, |
509 | int __range, struct re_registers *__regs, int __stop); |
510 | |
511 | |
512 | /* Like `re_search', but return how many characters in STRING the regexp |
513 | in BUFFER matched, starting at position START. */ |
514 | extern int re_match (struct re_pattern_buffer *__buffer, const char *__string, |
515 | int __length, int __start, struct re_registers *__regs); |
516 | |
517 | |
518 | /* Relates to `re_match' as `re_search_2' relates to `re_search'. */ |
519 | extern int re_match_2 (struct re_pattern_buffer *__buffer, |
520 | const char *__string1, int __length1, |
521 | const char *__string2, int __length2, int __start, |
522 | struct re_registers *__regs, int __stop); |
523 | |
524 | |
525 | /* Set REGS to hold NUM_REGS registers, storing them in STARTS and |
526 | ENDS. Subsequent matches using BUFFER and REGS will use this memory |
527 | for recording register information. STARTS and ENDS must be |
528 | allocated with malloc, and must each be at least `NUM_REGS * sizeof |
529 | (regoff_t)' bytes long. |
530 | |
531 | If NUM_REGS == 0, then subsequent matches should allocate their own |
532 | register data. |
533 | |
534 | Unless this function is called, the first search or match using |
535 | PATTERN_BUFFER will allocate its own register data, without |
536 | freeing the old data. */ |
537 | extern void re_set_registers (struct re_pattern_buffer *__buffer, |
538 | struct re_registers *__regs, |
539 | unsigned int __num_regs, |
540 | regoff_t *__starts, regoff_t *__ends); |
541 | #endif /* Use GNU */ |
542 | |
543 | #if defined _REGEX_RE_COMP || (defined _LIBC && defined __USE_BSD) |
544 | # ifndef _CRAY |
545 | /* 4.2 bsd compatibility. */ |
546 | extern char *re_comp (const char *); |
547 | extern int re_exec (const char *); |
548 | # endif |
549 | #endif |
550 | |
551 | /* GCC 2.95 and later have "__restrict"; C99 compilers have |
552 | "restrict", and "configure" may have defined "restrict". */ |
553 | #ifndef __restrict |
554 | # if ! (2 < __GNUC__ || (2 == __GNUC__ && 95 <= __GNUC_MINOR__)) |
555 | # if defined restrict || 199901L <= __STDC_VERSION__ |
556 | # define __restrict restrict |
557 | # else |
558 | # define __restrict |
559 | # endif |
560 | # endif |
561 | #endif |
562 | /* gcc 3.1 and up support the [restrict] syntax. */ |
563 | #ifndef __restrict_arr |
564 | # if (__GNUC__ > 3 || (__GNUC__ == 3 && __GNUC_MINOR__ >= 1)) \ |
565 | && !defined __GNUG__ |
566 | # define __restrict_arr __restrict |
567 | # else |
568 | # define __restrict_arr |
569 | # endif |
570 | #endif |
571 | |
572 | /* POSIX compatibility. */ |
573 | extern int regcomp (regex_t *__restrict __preg, |
574 | const char *__restrict __pattern, |
575 | int __cflags); |
576 | |
577 | extern int regexec (const regex_t *__restrict __preg, |
578 | const char *__restrict __string, size_t __nmatch, |
579 | regmatch_t __pmatch[__restrict_arr], |
580 | int __eflags); |
581 | |
582 | extern size_t regerror (int __errcode, const regex_t *__restrict __preg, |
583 | char *__restrict __errbuf, size_t __errbuf_size); |
584 | |
585 | extern void regfree (regex_t *__preg); |
586 | |
587 | |
588 | #ifdef __cplusplus |
589 | } |
590 | #endif /* C++ */ |
591 | |
592 | #endif /* regex.h */ |
593 | |