Debian Codegraph @

lib/regex.h

raw
1 /* Definitions for data structures and routines for the regular
2 expression library.
3 Copyright (C) 1985, 1989-2026 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 <https://www.gnu.org/licenses/>. */
19
20 #ifndef _REGEX_H
21 #define _REGEX_H 1
22
23 #include <sys/types.h>
24
25 /* Allow the use in C++ code. */
26 #ifdef __cplusplus
27 extern "C" {
28 #endif
29
30 /* Define __USE_GNU to declare GNU extensions that violate the
31 POSIX name space rules. */
32 #ifdef _GNU_SOURCE
33 # define __USE_GNU 1
34 #endif
35
36 #ifdef _REGEX_LARGE_OFFSETS
37
38 /* Use types and values that are wide enough to represent signed and
39 unsigned byte offsets in memory. This currently works only when
40 the regex code is used outside of the GNU C library; it is not yet
41 supported within glibc itself, and glibc users should not define
42 _REGEX_LARGE_OFFSETS. */
43
44 /* The type of object sizes. */
45 typedef size_t __re_size_t;
46
47 /* The type of object sizes, in places where the traditional code
48 uses unsigned long int. */
49 typedef size_t __re_long_size_t;
50
51 #else
52
53 /* The traditional GNU regex implementation mishandles strings longer
54 than INT_MAX. */
55 typedef unsigned int __re_size_t;
56 typedef unsigned long int __re_long_size_t;
57
58 #endif
59
60 /* The following two types have to be signed and unsigned integer type
61 wide enough to hold a value of a pointer. For most ANSI compilers
62 ptrdiff_t and size_t should be likely OK. Still size of these two
63 types is 2 for Microsoft C. Ugh... */
64 typedef long int s_reg_t;
65 typedef unsigned long int active_reg_t;
66
67 /* The following bits are used to determine the regexp syntax we
68 recognize. The set/not-set meanings are chosen so that Emacs syntax
69 is the value 0 for Emacs 20 (2000) and earlier, and the value
70 RE_SYNTAX_EMACS for Emacs 21 (2001) and later. */
71 typedef unsigned long int reg_syntax_t;
72
73 #ifdef __USE_GNU
74 /* If this bit is not set, then \ inside a bracket expression is literal.
75 If set, then such a \ quotes the following character. */
76 # define RE_BACKSLASH_ESCAPE_IN_LISTS 1ul
77
78 /* If this bit is not set, then + and ? are operators, and \+ and \? are
79 literals.
80 If set, then \+ and \? are operators and + and ? are literals. */
81 # define RE_BK_PLUS_QM (RE_BACKSLASH_ESCAPE_IN_LISTS << 1)
82
83 /* If this bit is set, then character classes are supported. They are:
84 [:alpha:], [:upper:], [:lower:], [:digit:], [:alnum:], [:xdigit:],
85 [:space:], [:print:], [:punct:], [:graph:], and [:cntrl:].
86 If not set, then character classes are not supported. */
87 # define RE_CHAR_CLASSES (RE_BK_PLUS_QM << 1)
88
89 /* If this bit is set, then ^ and $ are always anchors (outside bracket
90 expressions, of course).
91 If this bit is not set, then it depends:
92 ^ is an anchor if it is at the beginning of a regular
93 expression or after an open-group or an alternation operator;
94 $ is an anchor if it is at the end of a regular expression, or
95 before a close-group or an alternation operator.
96
97 This bit could be (re)combined with RE_CONTEXT_INDEP_OPS, because
98 POSIX draft 11.2 says that * etc. in leading positions is undefined.
99 We already implemented a previous draft which made those constructs
100 invalid, though, so we haven't changed the code back. */
101 # define RE_CONTEXT_INDEP_ANCHORS (RE_CHAR_CLASSES << 1)
102
103 /* If this bit is set, then special characters are always special
104 regardless of where they are in the pattern.
105 If this bit is not set, then special characters are special only in
106 some contexts; otherwise they are ordinary. Specifically,
107 * + ? and intervals are only special when not after the beginning,
108 open-group, or alternation operator. */
109 # define RE_CONTEXT_INDEP_OPS (RE_CONTEXT_INDEP_ANCHORS << 1)
110
111 /* If this bit is set, then *, +, ?, and { cannot be first in an re or
112 immediately after an alternation or begin-group operator. */
113 # define RE_CONTEXT_INVALID_OPS (RE_CONTEXT_INDEP_OPS << 1)
114
115 /* If this bit is set, then . matches newline.
116 If not set, then it doesn't. */
117 # define RE_DOT_NEWLINE (RE_CONTEXT_INVALID_OPS << 1)
118
119 /* If this bit is set, then . doesn't match NUL.
120 If not set, then it does. */
121 # define RE_DOT_NOT_NULL (RE_DOT_NEWLINE << 1)
122
123 /* If this bit is set, nonmatching lists [^...] do not match newline.
124 If not set, they do. */
125 # define RE_HAT_LISTS_NOT_NEWLINE (RE_DOT_NOT_NULL << 1)
126
127 /* If this bit is set, either \{...\} or {...} defines an
128 interval, depending on RE_NO_BK_BRACES.
129 If not set, \{, \}, {, and } are literals. */
130 # define RE_INTERVALS (RE_HAT_LISTS_NOT_NEWLINE << 1)
131
132 /* If this bit is set, +, ? and | aren't recognized as operators.
133 If not set, they are. */
134 # define RE_LIMITED_OPS (RE_INTERVALS << 1)
135
136 /* If this bit is set, newline is an alternation operator.
137 If not set, newline is literal. */
138 # define RE_NEWLINE_ALT (RE_LIMITED_OPS << 1)
139
140 /* If this bit is set, then '{...}' defines an interval, and \{ and \}
141 are literals.
142 If not set, then '\{...\}' defines an interval. */
143 # define RE_NO_BK_BRACES (RE_NEWLINE_ALT << 1)
144
145 /* If this bit is set, (...) defines a group, and \( and \) are literals.
146 If not set, \(...\) defines a group, and ( and ) are literals. */
147 # define RE_NO_BK_PARENS (RE_NO_BK_BRACES << 1)
148
149 /* If this bit is set, then \<digit> matches <digit>.
150 If not set, then \<digit> is a back-reference. */
151 # define RE_NO_BK_REFS (RE_NO_BK_PARENS << 1)
152
153 /* If this bit is set, then | is an alternation operator, and \| is literal.
154 If not set, then \| is an alternation operator, and | is literal. */
155 # define RE_NO_BK_VBAR (RE_NO_BK_REFS << 1)
156
157 /* If this bit is set, then an ending range point collating higher
158 than the starting range point, as in [z-a], is invalid.
159 If not set, then when ending range point collates higher than the
160 starting range point, the range is ignored. */
161 # define RE_NO_EMPTY_RANGES (RE_NO_BK_VBAR << 1)
162
163 /* If this bit is set, then an unmatched ) is ordinary.
164 If not set, then an unmatched ) is invalid. */
165 # define RE_UNMATCHED_RIGHT_PAREN_ORD (RE_NO_EMPTY_RANGES << 1)
166
167 /* If this bit is set, succeed as soon as we match the whole pattern,
168 without further backtracking. */
169 # define RE_NO_POSIX_BACKTRACKING (RE_UNMATCHED_RIGHT_PAREN_ORD << 1)
170
171 /* If this bit is set, do not process the GNU regex operators.
172 If not set, then the GNU regex operators are recognized. */
173 # define RE_NO_GNU_OPS (RE_NO_POSIX_BACKTRACKING << 1)
174
175 /* If this bit is set, turn on internal regex debugging.
176 If not set, and debugging was on, turn it off.
177 This only works if regex.c is compiled -DDEBUG.
178 We define this bit always, so that all that's needed to turn on
179 debugging is to recompile regex.c; the calling code can always have
180 this bit set, and it won't affect anything in the normal case. */
181 # define RE_DEBUG (RE_NO_GNU_OPS << 1)
182
183 /* If this bit is set, a syntactically invalid interval is treated as
184 a string of ordinary characters. For example, the ERE 'a{1' is
185 treated as 'a\{1'. */
186 # define RE_INVALID_INTERVAL_ORD (RE_DEBUG << 1)
187
188 /* If this bit is set, then ignore case when matching.
189 If not set, then case is significant. */
190 # define RE_ICASE (RE_INVALID_INTERVAL_ORD << 1)
191
192 /* This bit is used internally like RE_CONTEXT_INDEP_ANCHORS but only
193 for ^, because it is difficult to scan the regex backwards to find
194 whether ^ should be special. */
195 # define RE_CARET_ANCHORS_HERE (RE_ICASE << 1)
196
197 /* If this bit is set, then \{ cannot be first in a regex or
198 immediately after an alternation, open-group or \} operator. */
199 # define RE_CONTEXT_INVALID_DUP (RE_CARET_ANCHORS_HERE << 1)
200
201 /* If this bit is set, then no_sub will be set to 1 during
202 re_compile_pattern. */
203 # define RE_NO_SUB (RE_CONTEXT_INVALID_DUP << 1)
204 #endif
205
206 /* This global variable defines the particular regexp syntax to use (for
207 some interfaces). When a regexp is compiled, the syntax used is
208 stored in the pattern buffer, so changing this does not affect
209 already-compiled regexps. */
210 extern reg_syntax_t re_syntax_options;
211
212 #ifdef __USE_GNU
213 /* Define combinations of the above bits for the standard possibilities.
214 (The [[[ comments delimit what gets put into the Texinfo file, so
215 don't delete them!) */
216 /* [[[begin syntaxes]]] */
217 # define RE_SYNTAX_EMACS \
218 (RE_CHAR_CLASSES | RE_INTERVALS)
219
220 # define RE_SYNTAX_AWK \
221 (RE_BACKSLASH_ESCAPE_IN_LISTS | RE_DOT_NOT_NULL \
222 | RE_NO_BK_PARENS | RE_NO_BK_REFS \
223 | RE_NO_BK_VBAR | RE_NO_EMPTY_RANGES \
224 | RE_DOT_NEWLINE | RE_CONTEXT_INDEP_ANCHORS \
225 | RE_CHAR_CLASSES \
226 | RE_UNMATCHED_RIGHT_PAREN_ORD | RE_NO_GNU_OPS)
227
228 # define RE_SYNTAX_GNU_AWK \
229 ((RE_SYNTAX_POSIX_EXTENDED | RE_BACKSLASH_ESCAPE_IN_LISTS \
230 | RE_INVALID_INTERVAL_ORD) \
231 & ~(RE_DOT_NOT_NULL | RE_CONTEXT_INDEP_OPS \
232 | RE_CONTEXT_INVALID_OPS ))
233
234 # define RE_SYNTAX_POSIX_AWK \
235 (RE_SYNTAX_POSIX_EXTENDED | RE_BACKSLASH_ESCAPE_IN_LISTS \
236 | RE_INTERVALS | RE_NO_GNU_OPS \
237 | RE_INVALID_INTERVAL_ORD)
238
239 # define RE_SYNTAX_GREP \
240 ((RE_SYNTAX_POSIX_BASIC | RE_NEWLINE_ALT) \
241 & ~(RE_CONTEXT_INVALID_DUP | RE_DOT_NOT_NULL))
242
243 # define RE_SYNTAX_EGREP \
244 ((RE_SYNTAX_POSIX_EXTENDED | RE_INVALID_INTERVAL_ORD | RE_NEWLINE_ALT) \
245 & ~(RE_CONTEXT_INVALID_OPS | RE_DOT_NOT_NULL))
246
247 /* POSIX grep -E behavior is no longer incompatible with GNU. */
248 # define RE_SYNTAX_POSIX_EGREP \
249 RE_SYNTAX_EGREP
250
251 /* P1003.2/D11.2, section 4.20.7.1, lines 5078ff. */
252 # define RE_SYNTAX_ED RE_SYNTAX_POSIX_BASIC
253
254 # define RE_SYNTAX_SED RE_SYNTAX_POSIX_BASIC
255
256 /* Syntax bits common to both basic and extended POSIX regex syntax. */
257 # define _RE_SYNTAX_POSIX_COMMON \
258 (RE_CHAR_CLASSES | RE_DOT_NEWLINE | RE_DOT_NOT_NULL \
259 | RE_INTERVALS | RE_NO_EMPTY_RANGES)
260
261 # define RE_SYNTAX_POSIX_BASIC \
262 (_RE_SYNTAX_POSIX_COMMON | RE_BK_PLUS_QM | RE_CONTEXT_INVALID_DUP)
263
264 /* Differs from ..._POSIX_BASIC only in that RE_BK_PLUS_QM becomes
265 RE_LIMITED_OPS, i.e., \? \+ \| are not recognized. Actually, this
266 isn't minimal, since other operators, such as \`, aren't disabled. */
267 # define RE_SYNTAX_POSIX_MINIMAL_BASIC \
268 (_RE_SYNTAX_POSIX_COMMON | RE_LIMITED_OPS)
269
270 # define RE_SYNTAX_POSIX_EXTENDED \
271 (_RE_SYNTAX_POSIX_COMMON | RE_CONTEXT_INDEP_ANCHORS \
272 | RE_CONTEXT_INDEP_OPS | RE_NO_BK_BRACES \
273 | RE_NO_BK_PARENS | RE_NO_BK_VBAR \
274 | RE_CONTEXT_INVALID_OPS | RE_UNMATCHED_RIGHT_PAREN_ORD)
275
276 /* Differs from ..._POSIX_EXTENDED in that RE_CONTEXT_INDEP_OPS is
277 removed and RE_NO_BK_REFS is added. */
278 # define RE_SYNTAX_POSIX_MINIMAL_EXTENDED \
279 (_RE_SYNTAX_POSIX_COMMON | RE_CONTEXT_INDEP_ANCHORS \
280 | RE_CONTEXT_INVALID_OPS | RE_NO_BK_BRACES \
281 | RE_NO_BK_PARENS | RE_NO_BK_REFS \
282 | RE_NO_BK_VBAR | RE_UNMATCHED_RIGHT_PAREN_ORD)
283 /* [[[end syntaxes]]] */
284
285 /* Maximum number of duplicates an interval can allow. POSIX-conforming
286 systems might define this in <limits.h>, but we want our
287 value, so remove any previous define. */
288 # ifdef _REGEX_INCLUDE_LIMITS_H
289 # include <limits.h>
290 # endif
291 # ifdef RE_DUP_MAX
292 # undef RE_DUP_MAX
293 # endif
294
295 /* RE_DUP_MAX is 2**15 - 1 because an earlier implementation stored
296 the counter as a 2-byte signed integer. This is no longer true, so
297 RE_DUP_MAX could be increased to (INT_MAX / 10 - 1), or to
298 ((SIZE_MAX - 9) / 10) if _REGEX_LARGE_OFFSETS is defined.
299 However, there would be a huge performance problem if someone
300 actually used a pattern like a\{214748363\}, so RE_DUP_MAX retains
301 its historical value. */
302 # define RE_DUP_MAX (0x7fff)
303 #endif
304
305
306 /* POSIX 'cflags' bits (i.e., information for 'regcomp'). */
307
308 /* If this bit is set, then use extended regular expression syntax.
309 If not set, then use basic regular expression syntax. */
310 #define REG_EXTENDED 1
311
312 /* If this bit is set, then ignore case when matching.
313 If not set, then case is significant. */
314 #define REG_ICASE (1 << 1)
315
316 /* If this bit is set, then anchors do not match at newline
317 characters in the string.
318 If not set, then anchors do match at newlines. */
319 #define REG_NEWLINE (1 << 2)
320
321 /* If this bit is set, then report only success or fail in regexec.
322 If not set, then returns differ between not matching and errors. */
323 #define REG_NOSUB (1 << 3)
324
325
326 /* POSIX 'eflags' bits (i.e., information for regexec). */
327
328 /* If this bit is set, then the beginning-of-line operator doesn't match
329 the beginning of the string (presumably because it's not the
330 beginning of a line).
331 If not set, then the beginning-of-line operator does match the
332 beginning of the string. */
333 #define REG_NOTBOL 1
334
335 /* Like REG_NOTBOL, except for the end-of-line. */
336 #define REG_NOTEOL (1 << 1)
337
338 /* Use PMATCH[0] to delimit the start and end of the search in the
339 buffer. */
340 #define REG_STARTEND (1 << 2)
341
342
343 /* If any error codes are removed, changed, or added, update the
344 '__re_error_msgid' table in regcomp.c. */
345
346 typedef enum
347 {
348 _REG_ENOSYS = -1, /* This will never happen for this implementation. */
349 _REG_NOERROR = 0, /* Success. */
350 _REG_NOMATCH, /* Didn't find a match (for regexec). */
351
352 /* POSIX regcomp return error codes. (In the order listed in the
353 standard.) */
354 _REG_BADPAT, /* Invalid pattern. */
355 _REG_ECOLLATE, /* Invalid collating element. */
356 _REG_ECTYPE, /* Invalid character class name. */
357 _REG_EESCAPE, /* Trailing backslash. */
358 _REG_ESUBREG, /* Invalid back reference. */
359 _REG_EBRACK, /* Unmatched left bracket. */
360 _REG_EPAREN, /* Parenthesis imbalance. */
361 _REG_EBRACE, /* Unmatched \{. */
362 _REG_BADBR, /* Invalid contents of \{\}. */
363 _REG_ERANGE, /* Invalid range end. */
364 _REG_ESPACE, /* Ran out of memory. */
365 _REG_BADRPT, /* No preceding re for repetition op. */
366
367 /* Error codes we've added. */
368 _REG_EEND, /* Premature end. */
369 _REG_ESIZE, /* Too large (e.g., repeat count too large). */
370 _REG_ERPAREN /* Unmatched ) or \); not returned from regcomp. */
371 } reg_errcode_t;
372
373 #if defined _XOPEN_SOURCE || defined __USE_XOPEN2K
374 # define REG_ENOSYS _REG_ENOSYS
375 #endif
376 #define REG_NOERROR _REG_NOERROR
377 #define REG_NOMATCH _REG_NOMATCH
378 #define REG_BADPAT _REG_BADPAT
379 #define REG_ECOLLATE _REG_ECOLLATE
380 #define REG_ECTYPE _REG_ECTYPE
381 #define REG_EESCAPE _REG_EESCAPE
382 #define REG_ESUBREG _REG_ESUBREG
383 #define REG_EBRACK _REG_EBRACK
384 #define REG_EPAREN _REG_EPAREN
385 #define REG_EBRACE _REG_EBRACE
386 #define REG_BADBR _REG_BADBR
387 #define REG_ERANGE _REG_ERANGE
388 #define REG_ESPACE _REG_ESPACE
389 #define REG_BADRPT _REG_BADRPT
390 #define REG_EEND _REG_EEND
391 #define REG_ESIZE _REG_ESIZE
392 #define REG_ERPAREN _REG_ERPAREN
393
394 /* This data structure represents a compiled pattern. Before calling
395 the pattern compiler, the fields 'buffer', 'allocated', 'fastmap',
396 and 'translate' can be set. After the pattern has been compiled,
397 the fields 're_nsub', 'not_bol' and 'not_eol' are available. All
398 other fields are private to the regex routines. */
399
400 #ifndef RE_TRANSLATE_TYPE
401 # define __RE_TRANSLATE_TYPE unsigned char *
402 # ifdef __USE_GNU
403 # define RE_TRANSLATE_TYPE __RE_TRANSLATE_TYPE
404 # endif
405 #endif
406
407 #ifdef __USE_GNU
408 # define __REPB_PREFIX(name) name
409 #else
410 # define __REPB_PREFIX(name) __##name
411 #endif
412
413 struct re_pattern_buffer
414 {
415 /* Space that holds the compiled pattern. The type
416 'struct re_dfa_t' is private and is not declared here. */
417 struct re_dfa_t *__REPB_PREFIX(buffer);
418
419 /* Number of bytes to which 'buffer' points. */
420 __re_long_size_t __REPB_PREFIX(allocated);
421
422 /* Number of bytes actually used in 'buffer'. */
423 __re_long_size_t __REPB_PREFIX(used);
424
425 /* Syntax setting with which the pattern was compiled. */
426 reg_syntax_t __REPB_PREFIX(syntax);
427
428 /* Pointer to a fastmap, if any, otherwise zero. re_search uses the
429 fastmap, if there is one, to skip over impossible starting points
430 for matches. */
431 char *__REPB_PREFIX(fastmap);
432
433 /* Either a translate table to apply to all characters before
434 comparing them, or zero for no translation. The translation is
435 applied to a pattern when it is compiled and to a string when it
436 is matched. */
437 __RE_TRANSLATE_TYPE __REPB_PREFIX(translate);
438
439 /* Number of subexpressions found by the compiler. */
440 size_t re_nsub;
441
442 /* Zero if this pattern cannot match the empty string, one else.
443 Well, in truth it's used only in 're_search_2', to see whether or
444 not we should use the fastmap, so we don't set this absolutely
445 perfectly; see 're_compile_fastmap' (the "duplicate" case). */
446 unsigned __REPB_PREFIX(can_be_null) : 1;
447
448 /* If REGS_UNALLOCATED, allocate space in the 'regs' structure
449 for 'max (RE_NREGS, re_nsub + 1)' groups.
450 If REGS_REALLOCATE, reallocate space if necessary.
451 If REGS_FIXED, use what's there. */
452 #ifdef __USE_GNU
453 # define REGS_UNALLOCATED 0
454 # define REGS_REALLOCATE 1
455 # define REGS_FIXED 2
456 #endif
457 unsigned __REPB_PREFIX(regs_allocated) : 2;
458
459 /* Set to zero when 're_compile_pattern' compiles a pattern; set to
460 one by 're_compile_fastmap' if it updates the fastmap. */
461 unsigned __REPB_PREFIX(fastmap_accurate) : 1;
462
463 /* If set, 're_match_2' does not return information about
464 subexpressions. */
465 unsigned __REPB_PREFIX(no_sub) : 1;
466
467 /* If set, a beginning-of-line anchor doesn't match at the beginning
468 of the string. */
469 unsigned __REPB_PREFIX(not_bol) : 1;
470
471 /* Similarly for an end-of-line anchor. */
472 unsigned __REPB_PREFIX(not_eol) : 1;
473
474 /* If true, an anchor at a newline matches. */
475 unsigned __REPB_PREFIX(newline_anchor) : 1;
476 };
477
478 typedef struct re_pattern_buffer regex_t;
479
480 /* Type for byte offsets within the string. POSIX mandates this. */
481 #ifdef _REGEX_LARGE_OFFSETS
482 /* POSIX 1003.1-2008 requires that regoff_t be at least as wide as
483 ptrdiff_t and ssize_t. We don't know of any hosts where ptrdiff_t
484 is wider than ssize_t, so ssize_t is safe. ptrdiff_t is not
485 visible here, so use ssize_t. */
486 typedef ssize_t regoff_t;
487 #else
488 /* The traditional GNU regex implementation mishandles strings longer
489 than INT_MAX. */
490 typedef int regoff_t;
491 #endif
492
493
494 #ifdef __USE_GNU
495 /* This is the structure we store register match data in. See
496 regex.texinfo for a full description of what registers match. */
497 struct re_registers
498 {
499 __re_size_t num_regs;
500 regoff_t *start;
501 regoff_t *end;
502 };
503
504
505 /* If 'regs_allocated' is REGS_UNALLOCATED in the pattern buffer,
506 're_match_2' returns information about at least this many registers
507 the first time a 'regs' structure is passed. */
508 # ifndef RE_NREGS
509 # define RE_NREGS 30
510 # endif
511 #endif
512
513
514 /* POSIX specification for registers. Aside from the different names than
515 're_registers', POSIX uses an array of structures, instead of a
516 structure of arrays. */
517 typedef struct
518 {
519 regoff_t rm_so; /* Byte offset from string's start to substring's start. */
520 regoff_t rm_eo; /* Byte offset from string's start to substring's end. */
521 } regmatch_t;
522
523 /* Declarations for routines. */
524
525 #ifndef _Attr_access_
526 # ifdef __attr_access
527 # define _Attr_access_(arg) __attr_access (arg)
528 # elif defined __GNUC__ && 10 <= __GNUC__
529 # define _Attr_access_(x) __attribute__ ((__access__ x))
530 # else
531 # define _Attr_access_(x)
532 # endif
533 #endif
534
535 #ifdef __USE_GNU
536 /* Sets the current default syntax to SYNTAX, and return the old syntax.
537 You can also simply assign to the 're_syntax_options' variable. */
538 extern reg_syntax_t re_set_syntax (reg_syntax_t __syntax);
539
540 /* Compile the regular expression PATTERN, with length LENGTH
541 and syntax given by the global 're_syntax_options', into the buffer
542 BUFFER. Return NULL if successful, and an error string if not.
543
544 To free the allocated storage, you must call 'regfree' on BUFFER.
545 Note that the translate table must either have been initialized by
546 'regcomp', with a malloc'ed value, or set to NULL before calling
547 'regfree'. */
548 extern const char *re_compile_pattern (const char *__pattern, size_t __length,
549 struct re_pattern_buffer *__buffer)
550 _Attr_access_ ((__read_only__, 1, 2));
551
552
553 /* Compile a fastmap for the compiled pattern in BUFFER; used to
554 accelerate searches. Return 0 if successful and -2 if was an
555 internal error. */
556 extern int re_compile_fastmap (struct re_pattern_buffer *__buffer);
557
558
559 /* Search in the string STRING (with length LENGTH) for the pattern
560 compiled into BUFFER. Start searching at position START, for RANGE
561 characters. Return the starting position of the match, -1 for no
562 match, or -2 for an internal error. Also return register
563 information in REGS (if REGS and BUFFER->no_sub are nonzero). */
564 extern regoff_t re_search (struct re_pattern_buffer *__buffer,
565 const char *__String, regoff_t __length,
566 regoff_t __start, regoff_t __range,
567 struct re_registers *__regs)
568 _Attr_access_ ((__read_only__, 2, 3));
569
570
571 /* Like 're_search', but search in the concatenation of STRING1 and
572 STRING2. Also, stop searching at index START + STOP. */
573 extern regoff_t re_search_2 (struct re_pattern_buffer *__buffer,
574 const char *__string1, regoff_t __length1,
575 const char *__string2, regoff_t __length2,
576 regoff_t __start, regoff_t __range,
577 struct re_registers *__regs,
578 regoff_t __stop)
579 _Attr_access_ ((__read_only__, 2, 3))
580 _Attr_access_ ((__read_only__, 4, 5));
581
582
583 /* Like 're_search', but return how many characters in STRING the regexp
584 in BUFFER matched, starting at position START. */
585 extern regoff_t re_match (struct re_pattern_buffer *__buffer,
586 const char *__String, regoff_t __length,
587 regoff_t __start, struct re_registers *__regs)
588 _Attr_access_ ((__read_only__, 2, 3));
589
590
591 /* Relates to 're_match' as 're_search_2' relates to 're_search'. */
592 extern regoff_t re_match_2 (struct re_pattern_buffer *__buffer,
593 const char *__string1, regoff_t __length1,
594 const char *__string2, regoff_t __length2,
595 regoff_t __start, struct re_registers *__regs,
596 regoff_t __stop)
597 _Attr_access_ ((__read_only__, 2, 3))
598 _Attr_access_ ((__read_only__, 4, 5));
599
600
601 /* Set REGS to hold NUM_REGS registers, storing them in STARTS and
602 ENDS. Subsequent matches using BUFFER and REGS will use this memory
603 for recording register information. STARTS and ENDS must be
604 allocated with malloc, and must each be at least 'NUM_REGS * sizeof
605 (regoff_t)' bytes long.
606
607 If NUM_REGS == 0, then subsequent matches should allocate their own
608 register data.
609
610 Unless this function is called, the first search or match using
611 BUFFER will allocate its own register data, without
612 freeing the old data. */
613 extern void re_set_registers (struct re_pattern_buffer *__buffer,
614 struct re_registers *__regs,
615 __re_size_t __num_regs,
616 regoff_t *__starts, regoff_t *__ends);
617 #endif /* Use GNU */
618
619 #if defined _REGEX_RE_COMP || (defined _LIBC && defined __USE_MISC)
620 /* 4.2 bsd compatibility. */
621 extern char *re_comp (const char *);
622 extern int re_exec (const char *);
623 #endif
624
625 /* For plain 'restrict', use glibc's __restrict if defined.
626 Otherwise, GCC 2.95 and later have "__restrict"; C99 compilers have
627 "restrict", and "configure" may have defined "restrict".
628 Other compilers use __restrict, __restrict__, and _Restrict, and
629 'configure' might #define 'restrict' to those words, so pick a
630 different name. */
631 #ifndef _Restrict_
632 # if defined __restrict \
633 || 2 < __GNUC__ + (95 <= __GNUC_MINOR__) \
634 || __clang_major__ >= 3
635 # define _Restrict_ __restrict
636 # else
637 # if 199901L <= __STDC_VERSION__ || defined restrict
638 # define _Restrict_ restrict
639 # else
640 # define _Restrict_
641 # endif
642 # endif
643 #endif
644 /* For the ISO C99 syntax
645 array_name[restrict]
646 use glibc's __restrict_arr if available.
647 Otherwise, GCC 3.1 and clang support this syntax (but not in C++ mode).
648 Other ISO C99 compilers support it as well. */
649 #ifndef _Restrict_arr_
650 # ifdef __restrict_arr
651 # define _Restrict_arr_ __restrict_arr
652 # else
653 # if ((199901L <= __STDC_VERSION__ \
654 || 3 < __GNUC__ + (1 <= __GNUC_MINOR__) \
655 || __clang_major__ >= 3) \
656 && !defined __cplusplus)
657 # define _Restrict_arr_ _Restrict_
658 # else
659 # define _Restrict_arr_
660 # endif
661 # endif
662 #endif
663
664 /* POSIX compatibility. */
665 extern int regcomp (regex_t *_Restrict_ __preg,
666 const char *_Restrict_ __pattern,
667 int __cflags);
668
669 extern int regexec (const regex_t *_Restrict_ __preg,
670 const char *_Restrict_ __String, size_t __nmatch,
671 regmatch_t __pmatch[_Restrict_arr_],
672 int __eflags);
673
674 extern size_t regerror (int __errcode, const regex_t *_Restrict_ __preg,
675 char *_Restrict_ __errbuf, size_t __errbuf_size)
676 _Attr_access_ ((__write_only__, 3, 4));
677
678 extern void regfree (regex_t *__preg);
679
680 #ifdef __cplusplus
681 }
682 #endif /* C++ */
683
684 #endif /* regex.h */