VirtualBox

source: kBuild/vendor/grep/2.12/lib/regex.h@ 2595

Last change on this file since 2595 was 2595, checked in by bird, 12 years ago

gnu grep version 2.12 (grep-2.12.tar.xz, md5sum=8d2f0346d08b13c18afb81f0e8aa1e2f)

  • Property svn:eol-style set to native
File size: 24.3 KB
Line 
1/* Definitions for data structures and routines for the regular
2 expression library.
3 Copyright (C) 1985, 1989-1993, 1995-1998, 2000-2003, 2005-2012
4 Free Software Foundation, Inc.
5 This file is part of the GNU C Library.
6
7 This program is free software; you can redistribute it and/or modify
8 it under the terms of the GNU General Public License as published by
9 the Free Software Foundation; either version 3, or (at your option)
10 any later version.
11
12 This program is distributed in the hope that it will be useful,
13 but WITHOUT ANY WARRANTY; without even the implied warranty of
14 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
15 GNU General Public License for more details.
16
17 You should have received a copy of the GNU General Public License along
18 with this program; if not, see <http://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
27extern "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 nonnegative object indexes. Traditionally, GNU regex
45 uses 'int' for these. Code that uses __re_idx_t should work
46 regardless of whether the type is signed. */
47typedef size_t __re_idx_t;
48
49/* The type of object sizes. */
50typedef size_t __re_size_t;
51
52/* The type of object sizes, in places where the traditional code
53 uses unsigned long int. */
54typedef size_t __re_long_size_t;
55
56#else
57
58/* The traditional GNU regex implementation mishandles strings longer
59 than INT_MAX. */
60typedef int __re_idx_t;
61typedef unsigned int __re_size_t;
62typedef unsigned long int __re_long_size_t;
63
64#endif
65
66/* The following two types have to be signed and unsigned integer type
67 wide enough to hold a value of a pointer. For most ANSI compilers
68 ptrdiff_t and size_t should be likely OK. Still size of these two
69 types is 2 for Microsoft C. Ugh... */
70typedef long int s_reg_t;
71typedef unsigned long int active_reg_t;
72
73/* The following bits are used to determine the regexp syntax we
74 recognize. The set/not-set meanings are chosen so that Emacs syntax
75 remains the value 0. The bits are given in alphabetical order, and
76 the definitions shifted by one from the previous bit; thus, when we
77 add or remove a bit, only one other definition need change. */
78typedef unsigned long int reg_syntax_t;
79
80#ifdef __USE_GNU
81/* If this bit is not set, then \ inside a bracket expression is literal.
82 If set, then such a \ quotes the following character. */
83# define RE_BACKSLASH_ESCAPE_IN_LISTS ((unsigned long int) 1)
84
85/* If this bit is not set, then + and ? are operators, and \+ and \? are
86 literals.
87 If set, then \+ and \? are operators and + and ? are literals. */
88# define RE_BK_PLUS_QM (RE_BACKSLASH_ESCAPE_IN_LISTS << 1)
89
90/* If this bit is set, then character classes are supported. They are:
91 [:alpha:], [:upper:], [:lower:], [:digit:], [:alnum:], [:xdigit:],
92 [:space:], [:print:], [:punct:], [:graph:], and [:cntrl:].
93 If not set, then character classes are not supported. */
94# define RE_CHAR_CLASSES (RE_BK_PLUS_QM << 1)
95
96/* If this bit is set, then ^ and $ are always anchors (outside bracket
97 expressions, of course).
98 If this bit is not set, then it depends:
99 ^ is an anchor if it is at the beginning of a regular
100 expression or after an open-group or an alternation operator;
101 $ is an anchor if it is at the end of a regular expression, or
102 before a close-group or an alternation operator.
103
104 This bit could be (re)combined with RE_CONTEXT_INDEP_OPS, because
105 POSIX draft 11.2 says that * etc. in leading positions is undefined.
106 We already implemented a previous draft which made those constructs
107 invalid, though, so we haven't changed the code back. */
108# define RE_CONTEXT_INDEP_ANCHORS (RE_CHAR_CLASSES << 1)
109
110/* If this bit is set, then special characters are always special
111 regardless of where they are in the pattern.
112 If this bit is not set, then special characters are special only in
113 some contexts; otherwise they are ordinary. Specifically,
114 * + ? and intervals are only special when not after the beginning,
115 open-group, or alternation operator. */
116# define RE_CONTEXT_INDEP_OPS (RE_CONTEXT_INDEP_ANCHORS << 1)
117
118/* If this bit is set, then *, +, ?, and { cannot be first in an re or
119 immediately after an alternation or begin-group operator. */
120# define RE_CONTEXT_INVALID_OPS (RE_CONTEXT_INDEP_OPS << 1)
121
122/* If this bit is set, then . matches newline.
123 If not set, then it doesn't. */
124# define RE_DOT_NEWLINE (RE_CONTEXT_INVALID_OPS << 1)
125
126/* If this bit is set, then . doesn't match NUL.
127 If not set, then it does. */
128# define RE_DOT_NOT_NULL (RE_DOT_NEWLINE << 1)
129
130/* If this bit is set, nonmatching lists [^...] do not match newline.
131 If not set, they do. */
132# define RE_HAT_LISTS_NOT_NEWLINE (RE_DOT_NOT_NULL << 1)
133
134/* If this bit is set, either \{...\} or {...} defines an
135 interval, depending on RE_NO_BK_BRACES.
136 If not set, \{, \}, {, and } are literals. */
137# define RE_INTERVALS (RE_HAT_LISTS_NOT_NEWLINE << 1)
138
139/* If this bit is set, +, ? and | aren't recognized as operators.
140 If not set, they are. */
141# define RE_LIMITED_OPS (RE_INTERVALS << 1)
142
143/* If this bit is set, newline is an alternation operator.
144 If not set, newline is literal. */
145# define RE_NEWLINE_ALT (RE_LIMITED_OPS << 1)
146
147/* If this bit is set, then '{...}' defines an interval, and \{ and \}
148 are literals.
149 If not set, then '\{...\}' defines an interval. */
150# define RE_NO_BK_BRACES (RE_NEWLINE_ALT << 1)
151
152/* If this bit is set, (...) defines a group, and \( and \) are literals.
153 If not set, \(...\) defines a group, and ( and ) are literals. */
154# define RE_NO_BK_PARENS (RE_NO_BK_BRACES << 1)
155
156/* If this bit is set, then \<digit> matches <digit>.
157 If not set, then \<digit> is a back-reference. */
158# define RE_NO_BK_REFS (RE_NO_BK_PARENS << 1)
159
160/* If this bit is set, then | is an alternation operator, and \| is literal.
161 If not set, then \| is an alternation operator, and | is literal. */
162# define RE_NO_BK_VBAR (RE_NO_BK_REFS << 1)
163
164/* If this bit is set, then an ending range point collating higher
165 than the starting range point, as in [z-a], is invalid.
166 If not set, then when ending range point collates higher than the
167 starting range point, the range is ignored. */
168# define RE_NO_EMPTY_RANGES (RE_NO_BK_VBAR << 1)
169
170/* If this bit is set, then an unmatched ) is ordinary.
171 If not set, then an unmatched ) is invalid. */
172# define RE_UNMATCHED_RIGHT_PAREN_ORD (RE_NO_EMPTY_RANGES << 1)
173
174/* If this bit is set, succeed as soon as we match the whole pattern,
175 without further backtracking. */
176# define RE_NO_POSIX_BACKTRACKING (RE_UNMATCHED_RIGHT_PAREN_ORD << 1)
177
178/* If this bit is set, do not process the GNU regex operators.
179 If not set, then the GNU regex operators are recognized. */
180# define RE_NO_GNU_OPS (RE_NO_POSIX_BACKTRACKING << 1)
181
182/* If this bit is set, turn on internal regex debugging.
183 If not set, and debugging was on, turn it off.
184 This only works if regex.c is compiled -DDEBUG.
185 We define this bit always, so that all that's needed to turn on
186 debugging is to recompile regex.c; the calling code can always have
187 this bit set, and it won't affect anything in the normal case. */
188# define RE_DEBUG (RE_NO_GNU_OPS << 1)
189
190/* If this bit is set, a syntactically invalid interval is treated as
191 a string of ordinary characters. For example, the ERE 'a{1' is
192 treated as 'a\{1'. */
193# define RE_INVALID_INTERVAL_ORD (RE_DEBUG << 1)
194
195/* If this bit is set, then ignore case when matching.
196 If not set, then case is significant. */
197# define RE_ICASE (RE_INVALID_INTERVAL_ORD << 1)
198
199/* This bit is used internally like RE_CONTEXT_INDEP_ANCHORS but only
200 for ^, because it is difficult to scan the regex backwards to find
201 whether ^ should be special. */
202# define RE_CARET_ANCHORS_HERE (RE_ICASE << 1)
203
204/* If this bit is set, then \{ cannot be first in a regex or
205 immediately after an alternation, open-group or \} operator. */
206# define RE_CONTEXT_INVALID_DUP (RE_CARET_ANCHORS_HERE << 1)
207
208/* If this bit is set, then no_sub will be set to 1 during
209 re_compile_pattern. */
210# define RE_NO_SUB (RE_CONTEXT_INVALID_DUP << 1)
211#endif
212
213/* This global variable defines the particular regexp syntax to use (for
214 some interfaces). When a regexp is compiled, the syntax used is
215 stored in the pattern buffer, so changing this does not affect
216 already-compiled regexps. */
217extern reg_syntax_t re_syntax_options;
218
219
220#ifdef __USE_GNU
221/* Define combinations of the above bits for the standard possibilities.
222 (The [[[ comments delimit what gets put into the Texinfo file, so
223 don't delete them!) */
224/* [[[begin syntaxes]]] */
225# define RE_SYNTAX_EMACS 0
226
227# define RE_SYNTAX_AWK \
228 (RE_BACKSLASH_ESCAPE_IN_LISTS | RE_DOT_NOT_NULL \
229 | RE_NO_BK_PARENS | RE_NO_BK_REFS \
230 | RE_NO_BK_VBAR | RE_NO_EMPTY_RANGES \
231 | RE_DOT_NEWLINE | RE_CONTEXT_INDEP_ANCHORS \
232 | RE_CHAR_CLASSES \
233 | RE_UNMATCHED_RIGHT_PAREN_ORD | RE_NO_GNU_OPS)
234
235# define RE_SYNTAX_GNU_AWK \
236 ((RE_SYNTAX_POSIX_EXTENDED | RE_BACKSLASH_ESCAPE_IN_LISTS \
237 | RE_INVALID_INTERVAL_ORD) \
238 & ~(RE_DOT_NOT_NULL | RE_CONTEXT_INDEP_OPS \
239 | RE_CONTEXT_INVALID_OPS ))
240
241# define RE_SYNTAX_POSIX_AWK \
242 (RE_SYNTAX_POSIX_EXTENDED | RE_BACKSLASH_ESCAPE_IN_LISTS \
243 | RE_INTERVALS | RE_NO_GNU_OPS \
244 | RE_INVALID_INTERVAL_ORD)
245
246# define RE_SYNTAX_GREP \
247 (RE_BK_PLUS_QM | RE_CHAR_CLASSES \
248 | RE_HAT_LISTS_NOT_NEWLINE | RE_INTERVALS \
249 | RE_NEWLINE_ALT)
250
251# define RE_SYNTAX_EGREP \
252 (RE_CHAR_CLASSES | RE_CONTEXT_INDEP_ANCHORS \
253 | RE_CONTEXT_INDEP_OPS | RE_HAT_LISTS_NOT_NEWLINE \
254 | RE_NEWLINE_ALT | RE_NO_BK_PARENS \
255 | RE_NO_BK_VBAR)
256
257# define RE_SYNTAX_POSIX_EGREP \
258 (RE_SYNTAX_EGREP | RE_INTERVALS | RE_NO_BK_BRACES \
259 | RE_INVALID_INTERVAL_ORD)
260
261/* P1003.2/D11.2, section 4.20.7.1, lines 5078ff. */
262# define RE_SYNTAX_ED RE_SYNTAX_POSIX_BASIC
263
264# define RE_SYNTAX_SED RE_SYNTAX_POSIX_BASIC
265
266/* Syntax bits common to both basic and extended POSIX regex syntax. */
267# define _RE_SYNTAX_POSIX_COMMON \
268 (RE_CHAR_CLASSES | RE_DOT_NEWLINE | RE_DOT_NOT_NULL \
269 | RE_INTERVALS | RE_NO_EMPTY_RANGES)
270
271# define RE_SYNTAX_POSIX_BASIC \
272 (_RE_SYNTAX_POSIX_COMMON | RE_BK_PLUS_QM | RE_CONTEXT_INVALID_DUP)
273
274/* Differs from ..._POSIX_BASIC only in that RE_BK_PLUS_QM becomes
275 RE_LIMITED_OPS, i.e., \? \+ \| are not recognized. Actually, this
276 isn't minimal, since other operators, such as \`, aren't disabled. */
277# define RE_SYNTAX_POSIX_MINIMAL_BASIC \
278 (_RE_SYNTAX_POSIX_COMMON | RE_LIMITED_OPS)
279
280# define RE_SYNTAX_POSIX_EXTENDED \
281 (_RE_SYNTAX_POSIX_COMMON | RE_CONTEXT_INDEP_ANCHORS \
282 | RE_CONTEXT_INDEP_OPS | RE_NO_BK_BRACES \
283 | RE_NO_BK_PARENS | RE_NO_BK_VBAR \
284 | RE_CONTEXT_INVALID_OPS | RE_UNMATCHED_RIGHT_PAREN_ORD)
285
286/* Differs from ..._POSIX_EXTENDED in that RE_CONTEXT_INDEP_OPS is
287 removed and RE_NO_BK_REFS is added. */
288# define RE_SYNTAX_POSIX_MINIMAL_EXTENDED \
289 (_RE_SYNTAX_POSIX_COMMON | RE_CONTEXT_INDEP_ANCHORS \
290 | RE_CONTEXT_INVALID_OPS | RE_NO_BK_BRACES \
291 | RE_NO_BK_PARENS | RE_NO_BK_REFS \
292 | RE_NO_BK_VBAR | RE_UNMATCHED_RIGHT_PAREN_ORD)
293/* [[[end syntaxes]]] */
294
295/* Maximum number of duplicates an interval can allow. POSIX-conforming
296 systems might define this in <limits.h>, but we want our
297 value, so remove any previous define. */
298# ifdef _REGEX_INCLUDE_LIMITS_H
299# include <limits.h>
300# endif
301# ifdef RE_DUP_MAX
302# undef RE_DUP_MAX
303# endif
304
305/* RE_DUP_MAX is 2**15 - 1 because an earlier implementation stored
306 the counter as a 2-byte signed integer. This is no longer true, so
307 RE_DUP_MAX could be increased to (INT_MAX / 10 - 1), or to
308 ((SIZE_MAX - 9) / 10) if _REGEX_LARGE_OFFSETS is defined.
309 However, there would be a huge performance problem if someone
310 actually used a pattern like a\{214748363\}, so RE_DUP_MAX retains
311 its historical value. */
312# define RE_DUP_MAX (0x7fff)
313#endif
314
315
316/* POSIX 'cflags' bits (i.e., information for 'regcomp'). */
317
318/* If this bit is set, then use extended regular expression syntax.
319 If not set, then use basic regular expression syntax. */
320#define REG_EXTENDED 1
321
322/* If this bit is set, then ignore case when matching.
323 If not set, then case is significant. */
324#define REG_ICASE (1 << 1)
325
326/* If this bit is set, then anchors do not match at newline
327 characters in the string.
328 If not set, then anchors do match at newlines. */
329#define REG_NEWLINE (1 << 2)
330
331/* If this bit is set, then report only success or fail in regexec.
332 If not set, then returns differ between not matching and errors. */
333#define REG_NOSUB (1 << 3)
334
335
336/* POSIX 'eflags' bits (i.e., information for regexec). */
337
338/* If this bit is set, then the beginning-of-line operator doesn't match
339 the beginning of the string (presumably because it's not the
340 beginning of a line).
341 If not set, then the beginning-of-line operator does match the
342 beginning of the string. */
343#define REG_NOTBOL 1
344
345/* Like REG_NOTBOL, except for the end-of-line. */
346#define REG_NOTEOL (1 << 1)
347
348/* Use PMATCH[0] to delimit the start and end of the search in the
349 buffer. */
350#define REG_STARTEND (1 << 2)
351
352
353/* If any error codes are removed, changed, or added, update the
354 '__re_error_msgid' table in regcomp.c. */
355
356typedef enum
357{
358 _REG_ENOSYS = -1, /* This will never happen for this implementation. */
359 _REG_NOERROR = 0, /* Success. */
360 _REG_NOMATCH, /* Didn't find a match (for regexec). */
361
362 /* POSIX regcomp return error codes. (In the order listed in the
363 standard.) */
364 _REG_BADPAT, /* Invalid pattern. */
365 _REG_ECOLLATE, /* Invalid collating element. */
366 _REG_ECTYPE, /* Invalid character class name. */
367 _REG_EESCAPE, /* Trailing backslash. */
368 _REG_ESUBREG, /* Invalid back reference. */
369 _REG_EBRACK, /* Unmatched left bracket. */
370 _REG_EPAREN, /* Parenthesis imbalance. */
371 _REG_EBRACE, /* Unmatched \{. */
372 _REG_BADBR, /* Invalid contents of \{\}. */
373 _REG_ERANGE, /* Invalid range end. */
374 _REG_ESPACE, /* Ran out of memory. */
375 _REG_BADRPT, /* No preceding re for repetition op. */
376
377 /* Error codes we've added. */
378 _REG_EEND, /* Premature end. */
379 _REG_ESIZE, /* Too large (e.g., repeat count too large). */
380 _REG_ERPAREN /* Unmatched ) or \); not returned from regcomp. */
381} reg_errcode_t;
382
383#if defined _XOPEN_SOURCE || defined __USE_XOPEN2K
384# define REG_ENOSYS _REG_ENOSYS
385#endif
386#define REG_NOERROR _REG_NOERROR
387#define REG_NOMATCH _REG_NOMATCH
388#define REG_BADPAT _REG_BADPAT
389#define REG_ECOLLATE _REG_ECOLLATE
390#define REG_ECTYPE _REG_ECTYPE
391#define REG_EESCAPE _REG_EESCAPE
392#define REG_ESUBREG _REG_ESUBREG
393#define REG_EBRACK _REG_EBRACK
394#define REG_EPAREN _REG_EPAREN
395#define REG_EBRACE _REG_EBRACE
396#define REG_BADBR _REG_BADBR
397#define REG_ERANGE _REG_ERANGE
398#define REG_ESPACE _REG_ESPACE
399#define REG_BADRPT _REG_BADRPT
400#define REG_EEND _REG_EEND
401#define REG_ESIZE _REG_ESIZE
402#define REG_ERPAREN _REG_ERPAREN
403
404
405/* This data structure represents a compiled pattern. Before calling
406 the pattern compiler, the fields 'buffer', 'allocated', 'fastmap',
407 and 'translate' can be set. After the pattern has been compiled,
408 the fields 're_nsub', 'not_bol' and 'not_eol' are available. All
409 other fields are private to the regex routines. */
410
411#ifndef RE_TRANSLATE_TYPE
412# define __RE_TRANSLATE_TYPE unsigned char *
413# ifdef __USE_GNU
414# define RE_TRANSLATE_TYPE __RE_TRANSLATE_TYPE
415# endif
416#endif
417
418#ifdef __USE_GNU
419# define __REPB_PREFIX(name) name
420#else
421# define __REPB_PREFIX(name) __##name
422#endif
423
424struct re_pattern_buffer
425{
426 /* Space that holds the compiled pattern. The type
427 'struct re_dfa_t' is private and is not declared here. */
428 struct re_dfa_t *__REPB_PREFIX(buffer);
429
430 /* Number of bytes to which 'buffer' points. */
431 __re_long_size_t __REPB_PREFIX(allocated);
432
433 /* Number of bytes actually used in 'buffer'. */
434 __re_long_size_t __REPB_PREFIX(used);
435
436 /* Syntax setting with which the pattern was compiled. */
437 reg_syntax_t __REPB_PREFIX(syntax);
438
439 /* Pointer to a fastmap, if any, otherwise zero. re_search uses the
440 fastmap, if there is one, to skip over impossible starting points
441 for matches. */
442 char *__REPB_PREFIX(fastmap);
443
444 /* Either a translate table to apply to all characters before
445 comparing them, or zero for no translation. The translation is
446 applied to a pattern when it is compiled and to a string when it
447 is matched. */
448 __RE_TRANSLATE_TYPE __REPB_PREFIX(translate);
449
450 /* Number of subexpressions found by the compiler. */
451 size_t re_nsub;
452
453 /* Zero if this pattern cannot match the empty string, one else.
454 Well, in truth it's used only in 're_search_2', to see whether or
455 not we should use the fastmap, so we don't set this absolutely
456 perfectly; see 're_compile_fastmap' (the "duplicate" case). */
457 unsigned __REPB_PREFIX(can_be_null) : 1;
458
459 /* If REGS_UNALLOCATED, allocate space in the 'regs' structure
460 for 'max (RE_NREGS, re_nsub + 1)' groups.
461 If REGS_REALLOCATE, reallocate space if necessary.
462 If REGS_FIXED, use what's there. */
463#ifdef __USE_GNU
464# define REGS_UNALLOCATED 0
465# define REGS_REALLOCATE 1
466# define REGS_FIXED 2
467#endif
468 unsigned __REPB_PREFIX(regs_allocated) : 2;
469
470 /* Set to zero when 're_compile_pattern' compiles a pattern; set to
471 one by 're_compile_fastmap' if it updates the fastmap. */
472 unsigned __REPB_PREFIX(fastmap_accurate) : 1;
473
474 /* If set, 're_match_2' does not return information about
475 subexpressions. */
476 unsigned __REPB_PREFIX(no_sub) : 1;
477
478 /* If set, a beginning-of-line anchor doesn't match at the beginning
479 of the string. */
480 unsigned __REPB_PREFIX(not_bol) : 1;
481
482 /* Similarly for an end-of-line anchor. */
483 unsigned __REPB_PREFIX(not_eol) : 1;
484
485 /* If true, an anchor at a newline matches. */
486 unsigned __REPB_PREFIX(newline_anchor) : 1;
487};
488
489typedef struct re_pattern_buffer regex_t;
490
491
492/* Type for byte offsets within the string. POSIX mandates this. */
493#ifdef _REGEX_LARGE_OFFSETS
494/* POSIX 1003.1-2008 requires that regoff_t be at least as wide as
495 ptrdiff_t and ssize_t. We don't know of any hosts where ptrdiff_t
496 is wider than ssize_t, so ssize_t is safe. */
497typedef ssize_t regoff_t;
498#else
499/* The traditional GNU regex implementation mishandles strings longer
500 than INT_MAX. */
501typedef int regoff_t;
502#endif
503
504
505#ifdef __USE_GNU
506/* This is the structure we store register match data in. See
507 regex.texinfo for a full description of what registers match. */
508struct re_registers
509{
510 __re_size_t num_regs;
511 regoff_t *start;
512 regoff_t *end;
513};
514
515
516/* If 'regs_allocated' is REGS_UNALLOCATED in the pattern buffer,
517 're_match_2' returns information about at least this many registers
518 the first time a 'regs' structure is passed. */
519# ifndef RE_NREGS
520# define RE_NREGS 30
521# endif
522#endif
523
524
525/* POSIX specification for registers. Aside from the different names than
526 're_registers', POSIX uses an array of structures, instead of a
527 structure of arrays. */
528typedef struct
529{
530 regoff_t rm_so; /* Byte offset from string's start to substring's start. */
531 regoff_t rm_eo; /* Byte offset from string's start to substring's end. */
532} regmatch_t;
533
534
535/* Declarations for routines. */
536
537#ifdef __USE_GNU
538/* Sets the current default syntax to SYNTAX, and return the old syntax.
539 You can also simply assign to the 're_syntax_options' variable. */
540extern reg_syntax_t re_set_syntax (reg_syntax_t __syntax);
541
542/* Compile the regular expression PATTERN, with length LENGTH
543 and syntax given by the global 're_syntax_options', into the buffer
544 BUFFER. Return NULL if successful, and an error string if not.
545
546 To free the allocated storage, you must call 'regfree' on BUFFER.
547 Note that the translate table must either have been initialised by
548 'regcomp', with a malloc'ed value, or set to NULL before calling
549 'regfree'. */
550extern const char *re_compile_pattern (const char *__pattern, size_t __length,
551 struct re_pattern_buffer *__buffer);
552
553
554/* Compile a fastmap for the compiled pattern in BUFFER; used to
555 accelerate searches. Return 0 if successful and -2 if was an
556 internal error. */
557extern int re_compile_fastmap (struct re_pattern_buffer *__buffer);
558
559
560/* Search in the string STRING (with length LENGTH) for the pattern
561 compiled into BUFFER. Start searching at position START, for RANGE
562 characters. Return the starting position of the match, -1 for no
563 match, or -2 for an internal error. Also return register
564 information in REGS (if REGS and BUFFER->no_sub are nonzero). */
565extern regoff_t re_search (struct re_pattern_buffer *__buffer,
566 const char *__string, __re_idx_t __length,
567 __re_idx_t __start, regoff_t __range,
568 struct re_registers *__regs);
569
570
571/* Like 're_search', but search in the concatenation of STRING1 and
572 STRING2. Also, stop searching at index START + STOP. */
573extern regoff_t re_search_2 (struct re_pattern_buffer *__buffer,
574 const char *__string1, __re_idx_t __length1,
575 const char *__string2, __re_idx_t __length2,
576 __re_idx_t __start, regoff_t __range,
577 struct re_registers *__regs,
578 __re_idx_t __stop);
579
580
581/* Like 're_search', but return how many characters in STRING the regexp
582 in BUFFER matched, starting at position START. */
583extern regoff_t re_match (struct re_pattern_buffer *__buffer,
584 const char *__string, __re_idx_t __length,
585 __re_idx_t __start, struct re_registers *__regs);
586
587
588/* Relates to 're_match' as 're_search_2' relates to 're_search'. */
589extern regoff_t re_match_2 (struct re_pattern_buffer *__buffer,
590 const char *__string1, __re_idx_t __length1,
591 const char *__string2, __re_idx_t __length2,
592 __re_idx_t __start, struct re_registers *__regs,
593 __re_idx_t __stop);
594
595
596/* Set REGS to hold NUM_REGS registers, storing them in STARTS and
597 ENDS. Subsequent matches using BUFFER and REGS will use this memory
598 for recording register information. STARTS and ENDS must be
599 allocated with malloc, and must each be at least 'NUM_REGS * sizeof
600 (regoff_t)' bytes long.
601
602 If NUM_REGS == 0, then subsequent matches should allocate their own
603 register data.
604
605 Unless this function is called, the first search or match using
606 BUFFER will allocate its own register data, without
607 freeing the old data. */
608extern void re_set_registers (struct re_pattern_buffer *__buffer,
609 struct re_registers *__regs,
610 __re_size_t __num_regs,
611 regoff_t *__starts, regoff_t *__ends);
612#endif /* Use GNU */
613
614#if defined _REGEX_RE_COMP || (defined _LIBC && defined __USE_BSD)
615# ifndef _CRAY
616/* 4.2 bsd compatibility. */
617extern char *re_comp (const char *);
618extern int re_exec (const char *);
619# endif
620#endif
621
622/* GCC 2.95 and later have "__restrict"; C99 compilers have
623 "restrict", and "configure" may have defined "restrict".
624 Other compilers use __restrict, __restrict__, and _Restrict, and
625 'configure' might #define 'restrict' to those words, so pick a
626 different name. */
627#ifndef _Restrict_
628# if 199901L <= __STDC_VERSION__
629# define _Restrict_ restrict
630# elif 2 < __GNUC__ || (2 == __GNUC__ && 95 <= __GNUC_MINOR__)
631# define _Restrict_ __restrict
632# else
633# define _Restrict_
634# endif
635#endif
636/* gcc 3.1 and up support the [restrict] syntax. Don't trust
637 sys/cdefs.h's definition of __restrict_arr, though, as it
638 mishandles gcc -ansi -pedantic. */
639#ifndef _Restrict_arr_
640# if ((199901L <= __STDC_VERSION__ \
641 || ((3 < __GNUC__ || (3 == __GNUC__ && 1 <= __GNUC_MINOR__)) \
642 && !defined __STRICT_ANSI__)) \
643 && !defined __GNUG__)
644# define _Restrict_arr_ _Restrict_
645# else
646# define _Restrict_arr_
647# endif
648#endif
649
650/* POSIX compatibility. */
651extern int regcomp (regex_t *_Restrict_ __preg,
652 const char *_Restrict_ __pattern,
653 int __cflags);
654
655extern int regexec (const regex_t *_Restrict_ __preg,
656 const char *_Restrict_ __string, size_t __nmatch,
657 regmatch_t __pmatch[_Restrict_arr_],
658 int __eflags);
659
660extern size_t regerror (int __errcode, const regex_t *_Restrict_ __preg,
661 char *_Restrict_ __errbuf, size_t __errbuf_size);
662
663extern void regfree (regex_t *__preg);
664
665
666#ifdef __cplusplus
667}
668#endif /* C++ */
669
670#endif /* regex.h */
Note: See TracBrowser for help on using the repository browser.

© 2024 Oracle Support Privacy / Do Not Sell My Info Terms of Use Trademark Policy Automated Access Etiquette