Deleted Added
full compact
regex.h (30578) regex.h (126212)
1/* Definitions for data structures and routines for the regular
2 expression library, version 0.12.
1/* $FreeBSD: head/gnu/lib/libregex/regex.h 126212 2004-02-25 02:39:10Z ache $ */
2#ifndef _REGEX_H
3#include <posix/regex.h>
3
4
4 Copyright (C) 1985, 1989, 1990, 1991, 1992, 1993 Free Software Foundation, Inc.
5/* Document internal interfaces. */
6extern reg_syntax_t __re_set_syntax _RE_ARGS ((reg_syntax_t syntax));
5
7
6 This program is free software; you can redistribute it and/or modify
7 it under the terms of the GNU General Public License as published by
8 the Free Software Foundation; either version 2, or (at your option)
9 any later version.
10
11 This program 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
14 GNU General Public License for more details.
15
16 You should have received a copy of the GNU General Public License
17 along with this program; if not, write to the Free Software
18 Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. */
19
20#ifndef __REGEXP_LIBRARY_H__
21#define __REGEXP_LIBRARY_H__
22
23/* Allow the use in C++ code. */
24#ifdef __cplusplus
25extern "C" {
26#endif
27
28/* POSIX says that <sys/types.h> must be included (by the caller) before
29 <regex.h>. */
30
31#ifdef VMS
32/* VMS doesn't have `size_t' in <sys/types.h>, even though POSIX says it
33 should be there. */
34#include <stddef.h>
35#endif
36
37
38/* The following two types have to be signed and unsigned integer type
39 wide enough to hold a value of a pointer. For most ANSI compilers
40 ptrdiff_t and size_t should be likely OK. Still size of these two
41 types is 2 for Microsoft C. Ugh... */
42typedef long s_reg_t;
43typedef unsigned long active_reg_t;
44
45/* The following bits are used to determine the regexp syntax we
46 recognize. The set/not-set meanings are chosen so that Emacs syntax
47 remains the value 0. The bits are given in alphabetical order, and
48 the definitions shifted by one from the previous bit; thus, when we
49 add or remove a bit, only one other definition need change. */
50typedef unsigned long reg_syntax_t;
51
52/* If this bit is not set, then \ inside a bracket expression is literal.
53 If set, then such a \ quotes the following character. */
54#define RE_BACKSLASH_ESCAPE_IN_LISTS (1L)
55
56/* If this bit is not set, then + and ? are operators, and \+ and \? are
57 literals.
58 If set, then \+ and \? are operators and + and ? are literals. */
59#define RE_BK_PLUS_QM (RE_BACKSLASH_ESCAPE_IN_LISTS << 1)
60
61/* If this bit is set, then character classes are supported. They are:
62 [:alpha:], [:upper:], [:lower:], [:digit:], [:alnum:], [:xdigit:],
63 [:space:], [:print:], [:punct:], [:graph:], and [:cntrl:].
64 If not set, then character classes are not supported. */
65#define RE_CHAR_CLASSES (RE_BK_PLUS_QM << 1)
66
67/* If this bit is set, then ^ and $ are always anchors (outside bracket
68 expressions, of course).
69 If this bit is not set, then it depends:
70 ^ is an anchor if it is at the beginning of a regular
71 expression or after an open-group or an alternation operator;
72 $ is an anchor if it is at the end of a regular expression, or
73 before a close-group or an alternation operator.
74
75 This bit could be (re)combined with RE_CONTEXT_INDEP_OPS, because
76 POSIX draft 11.2 says that * etc. in leading positions is undefined.
77 We already implemented a previous draft which made those constructs
78 invalid, though, so we haven't changed the code back. */
79#define RE_CONTEXT_INDEP_ANCHORS (RE_CHAR_CLASSES << 1)
80
81/* If this bit is set, then special characters are always special
82 regardless of where they are in the pattern.
83 If this bit is not set, then special characters are special only in
84 some contexts; otherwise they are ordinary. Specifically,
85 * + ? and intervals are only special when not after the beginning,
86 open-group, or alternation operator. */
87#define RE_CONTEXT_INDEP_OPS (RE_CONTEXT_INDEP_ANCHORS << 1)
88
89/* If this bit is set, then *, +, ?, and { cannot be first in an re or
90 immediately after an alternation or begin-group operator. */
91#define RE_CONTEXT_INVALID_OPS (RE_CONTEXT_INDEP_OPS << 1)
92
93/* If this bit is set, then . matches newline.
94 If not set, then it doesn't. */
95#define RE_DOT_NEWLINE (RE_CONTEXT_INVALID_OPS << 1)
96
97/* If this bit is set, then . doesn't match NUL.
98 If not set, then it does. */
99#define RE_DOT_NOT_NULL (RE_DOT_NEWLINE << 1)
100
101/* If this bit is set, nonmatching lists [^...] do not match newline.
102 If not set, they do. */
103#define RE_HAT_LISTS_NOT_NEWLINE (RE_DOT_NOT_NULL << 1)
104
105/* If this bit is set, either \{...\} or {...} defines an
106 interval, depending on RE_NO_BK_BRACES.
107 If not set, \{, \}, {, and } are literals. */
108#define RE_INTERVALS (RE_HAT_LISTS_NOT_NEWLINE << 1)
109
110/* If this bit is set, +, ? and | aren't recognized as operators.
111 If not set, they are. */
112#define RE_LIMITED_OPS (RE_INTERVALS << 1)
113
114/* If this bit is set, newline is an alternation operator.
115 If not set, newline is literal. */
116#define RE_NEWLINE_ALT (RE_LIMITED_OPS << 1)
117
118/* If this bit is set, then `{...}' defines an interval, and \{ and \}
119 are literals.
120 If not set, then `\{...\}' defines an interval. */
121#define RE_NO_BK_BRACES (RE_NEWLINE_ALT << 1)
122
123/* If this bit is set, (...) defines a group, and \( and \) are literals.
124 If not set, \(...\) defines a group, and ( and ) are literals. */
125#define RE_NO_BK_PARENS (RE_NO_BK_BRACES << 1)
126
127/* If this bit is set, then \<digit> matches <digit>.
128 If not set, then \<digit> is a back-reference. */
129#define RE_NO_BK_REFS (RE_NO_BK_PARENS << 1)
130
131/* If this bit is set, then | is an alternation operator, and \| is literal.
132 If not set, then \| is an alternation operator, and | is literal. */
133#define RE_NO_BK_VBAR (RE_NO_BK_REFS << 1)
134
135/* If this bit is set, then an ending range point collating higher
136 than the starting range point, as in [z-a], is invalid.
137 If not set, then when ending range point collates higher than the
138 starting range point, the range is ignored. */
139#define RE_NO_EMPTY_RANGES (RE_NO_BK_VBAR << 1)
140
141/* If this bit is set, then an unmatched ) is ordinary.
142 If not set, then an unmatched ) is invalid. */
143#define RE_UNMATCHED_RIGHT_PAREN_ORD (RE_NO_EMPTY_RANGES << 1)
144
145/* If this bit is set, do not process the GNU regex operators.
146 IF not set, then the GNU regex operators are recognized. */
147#define RE_NO_GNU_OPS (RE_UNMATCHED_RIGHT_PAREN_ORD << 1)
148
149/* This global variable defines the particular regexp syntax to use (for
150 some interfaces). When a regexp is compiled, the syntax used is
151 stored in the pattern buffer, so changing this does not affect
152 already-compiled regexps. */
153extern reg_syntax_t re_syntax_options;
154
155/* Define combinations of the above bits for the standard possibilities.
156 (The [[[ comments delimit what gets put into the Texinfo file, so
157 don't delete them!) */
158/* [[[begin syntaxes]]] */
159#define RE_SYNTAX_EMACS 0
160
161#define RE_SYNTAX_AWK \
162 (RE_BACKSLASH_ESCAPE_IN_LISTS | RE_DOT_NOT_NULL \
163 | RE_NO_BK_PARENS | RE_NO_BK_REFS \
164 | RE_NO_BK_VBAR | RE_NO_EMPTY_RANGES \
165 | RE_DOT_NEWLINE | RE_CONTEXT_INDEP_ANCHORS \
166 | RE_UNMATCHED_RIGHT_PAREN_ORD | RE_NO_GNU_OPS)
167
168#define RE_SYNTAX_GNU_AWK \
169 ((RE_SYNTAX_POSIX_EXTENDED | RE_BACKSLASH_ESCAPE_IN_LISTS) \
170 & ~(RE_DOT_NOT_NULL | RE_INTERVALS | RE_CONTEXT_INDEP_OPS))
171
172#define RE_SYNTAX_POSIX_AWK \
173 (RE_SYNTAX_POSIX_EXTENDED | RE_BACKSLASH_ESCAPE_IN_LISTS \
174 | RE_INTERVALS | RE_NO_GNU_OPS)
175
176#define RE_SYNTAX_GREP \
177 (RE_BK_PLUS_QM | RE_CHAR_CLASSES \
178 | RE_HAT_LISTS_NOT_NEWLINE | RE_INTERVALS \
179 | RE_NEWLINE_ALT)
180
181#define RE_SYNTAX_EGREP \
182 (RE_CHAR_CLASSES | RE_CONTEXT_INDEP_ANCHORS \
183 | RE_CONTEXT_INDEP_OPS | RE_HAT_LISTS_NOT_NEWLINE \
184 | RE_NEWLINE_ALT | RE_NO_BK_PARENS \
185 | RE_NO_BK_VBAR)
186
187#define RE_SYNTAX_POSIX_EGREP \
188 (RE_SYNTAX_EGREP | RE_INTERVALS | RE_NO_BK_BRACES)
189
190/* P1003.2/D11.2, section 4.20.7.1, lines 5078ff. */
191#define RE_SYNTAX_ED RE_SYNTAX_POSIX_BASIC
192
193#define RE_SYNTAX_SED RE_SYNTAX_POSIX_BASIC
194
195/* Syntax bits common to both basic and extended POSIX regex syntax. */
196#define _RE_SYNTAX_POSIX_COMMON \
197 (RE_CHAR_CLASSES | RE_DOT_NEWLINE | RE_DOT_NOT_NULL \
198 | RE_INTERVALS | RE_NO_EMPTY_RANGES)
199
200#define RE_SYNTAX_POSIX_BASIC \
201 (_RE_SYNTAX_POSIX_COMMON | RE_BK_PLUS_QM)
202
203/* Differs from ..._POSIX_BASIC only in that RE_BK_PLUS_QM becomes
204 RE_LIMITED_OPS, i.e., \? \+ \| are not recognized. Actually, this
205 isn't minimal, since other operators, such as \`, aren't disabled. */
206#define RE_SYNTAX_POSIX_MINIMAL_BASIC \
207 (_RE_SYNTAX_POSIX_COMMON | RE_LIMITED_OPS)
208
209#define RE_SYNTAX_POSIX_EXTENDED \
210 (_RE_SYNTAX_POSIX_COMMON | RE_CONTEXT_INDEP_ANCHORS \
211 | RE_CONTEXT_INDEP_OPS | RE_NO_BK_BRACES \
212 | RE_NO_BK_PARENS | RE_NO_BK_VBAR \
213 | RE_UNMATCHED_RIGHT_PAREN_ORD)
214
215/* Differs from ..._POSIX_EXTENDED in that RE_CONTEXT_INVALID_OPS
216 replaces RE_CONTEXT_INDEP_OPS and RE_NO_BK_REFS is added. */
217#define RE_SYNTAX_POSIX_MINIMAL_EXTENDED \
218 (_RE_SYNTAX_POSIX_COMMON | RE_CONTEXT_INDEP_ANCHORS \
219 | RE_CONTEXT_INVALID_OPS | RE_NO_BK_BRACES \
220 | RE_NO_BK_PARENS | RE_NO_BK_REFS \
221 | RE_NO_BK_VBAR | RE_UNMATCHED_RIGHT_PAREN_ORD)
222/* [[[end syntaxes]]] */
223
224/* Maximum number of duplicates an interval can allow. Some systems
225 (erroneously) define this in other header files, but we want our
226 value, so remove any previous define. */
227#ifdef RE_DUP_MAX
228#undef RE_DUP_MAX
229#endif
230/* if sizeof(int) == 2, then ((1 << 15) - 1) overflows */
231#define RE_DUP_MAX (0x7fff)
232
233
234/* POSIX `cflags' bits (i.e., information for `regcomp'). */
235
236/* If this bit is set, then use extended regular expression syntax.
237 If not set, then use basic regular expression syntax. */
238#define REG_EXTENDED 1
239
240/* If this bit is set, then ignore case when matching.
241 If not set, then case is significant. */
242#define REG_ICASE (REG_EXTENDED << 1)
243
244/* If this bit is set, then anchors do not match at newline
245 characters in the string.
246 If not set, then anchors do match at newlines. */
247#define REG_NEWLINE (REG_ICASE << 1)
248
249/* If this bit is set, then report only success or fail in regexec.
250 If not set, then returns differ between not matching and errors. */
251#define REG_NOSUB (REG_NEWLINE << 1)
252
253
254/* POSIX `eflags' bits (i.e., information for regexec). */
255
256/* If this bit is set, then the beginning-of-line operator doesn't match
257 the beginning of the string (presumably because it's not the
258 beginning of a line).
259 If not set, then the beginning-of-line operator does match the
260 beginning of the string. */
261#define REG_NOTBOL 1
262
263/* Like REG_NOTBOL, except for the end-of-line. */
264#define REG_NOTEOL (1 << 1)
265
266
267/* If any error codes are removed, changed, or added, update the
268 `re_error_msg' table in regex.c. */
269typedef enum
270{
271 REG_NOERROR = 0, /* Success. */
272 REG_NOMATCH, /* Didn't find a match (for regexec). */
273
274 /* POSIX regcomp return error codes. (In the order listed in the
275 standard.) */
276 REG_BADPAT, /* Invalid pattern. */
277 REG_ECOLLATE, /* Not implemented. */
278 REG_ECTYPE, /* Invalid character class name. */
279 REG_EESCAPE, /* Trailing backslash. */
280 REG_ESUBREG, /* Invalid back reference. */
281 REG_EBRACK, /* Unmatched left bracket. */
282 REG_EPAREN, /* Parenthesis imbalance. */
283 REG_EBRACE, /* Unmatched \{. */
284 REG_BADBR, /* Invalid contents of \{\}. */
285 REG_ERANGE, /* Invalid range end. */
286 REG_ESPACE, /* Ran out of memory. */
287 REG_BADRPT, /* No preceding re for repetition op. */
288
289 /* Error codes we've added. */
290 REG_EEND, /* Premature end. */
291 REG_ESIZE, /* Compiled pattern bigger than 2^16 bytes. */
292 REG_ERPAREN /* Unmatched ) or \); not returned from regcomp. */
293} reg_errcode_t;
294
295/* This data structure represents a compiled pattern. Before calling
296 the pattern compiler, the fields `buffer', `allocated', `fastmap',
297 `translate', and `no_sub' can be set. After the pattern has been
298 compiled, the `re_nsub' field is available. All other fields are
299 private to the regex routines. */
300
301struct re_pattern_buffer
302{
303/* [[[begin pattern_buffer]]] */
304 /* Space that holds the compiled pattern. It is declared as
305 `unsigned char *' because its elements are
306 sometimes used as array indexes. */
307 unsigned char *buffer;
308
309 /* Number of bytes to which `buffer' points. */
310 unsigned long allocated;
311
312 /* Number of bytes actually used in `buffer'. */
313 unsigned long used;
314
315 /* Syntax setting with which the pattern was compiled. */
316 reg_syntax_t syntax;
317
318 /* Pointer to a fastmap, if any, otherwise zero. re_search uses
319 the fastmap, if there is one, to skip over impossible
320 starting points for matches. */
321 char *fastmap;
322
323 /* Either a translate table to apply to all characters before
324 comparing them, or zero for no translation. The translation
325 is applied to a pattern when it is compiled and to a string
326 when it is matched. */
327 char *translate;
328
329 /* Number of subexpressions found by the compiler. */
330 size_t re_nsub;
331
332 /* Zero if this pattern cannot match the empty string, one else.
333 Well, in truth it's used only in `re_search_2', to see
334 whether or not we should use the fastmap, so we don't set
335 this absolutely perfectly; see `re_compile_fastmap' (the
336 `duplicate' case). */
337 unsigned can_be_null : 1;
338
339 /* If REGS_UNALLOCATED, allocate space in the `regs' structure
340 for `max (RE_NREGS, re_nsub + 1)' groups.
341 If REGS_REALLOCATE, reallocate space if necessary.
342 If REGS_FIXED, use what's there. */
343#define REGS_UNALLOCATED 0
344#define REGS_REALLOCATE 1
345#define REGS_FIXED 2
346 unsigned regs_allocated : 2;
347
348 /* Set to zero when `regex_compile' compiles a pattern; set to one
349 by `re_compile_fastmap' if it updates the fastmap. */
350 unsigned fastmap_accurate : 1;
351
352 /* If set, `re_match_2' does not return information about
353 subexpressions. */
354 unsigned no_sub : 1;
355
356 /* If set, a beginning-of-line anchor doesn't match at the
357 beginning of the string. */
358 unsigned not_bol : 1;
359
360 /* Similarly for an end-of-line anchor. */
361 unsigned not_eol : 1;
362
363 /* If true, an anchor at a newline matches. */
364 unsigned newline_anchor : 1;
365
366/* [[[end pattern_buffer]]] */
367};
368
369typedef struct re_pattern_buffer regex_t;
370
371
372/* search.c (search_buffer) in Emacs needs this one opcode value. It is
373 defined both in `regex.c' and here. */
374#define RE_EXACTN_VALUE 1
375
376/* Type for byte offsets within the string. POSIX mandates this. */
377typedef int regoff_t;
378
379
380/* This is the structure we store register match data in. See
381 regex.texinfo for a full description of what registers match. */
382struct re_registers
383{
384 unsigned num_regs;
385 regoff_t *start;
386 regoff_t *end;
387};
388
389
390/* If `regs_allocated' is REGS_UNALLOCATED in the pattern buffer,
391 `re_match_2' returns information about at least this many registers
392 the first time a `regs' structure is passed. */
393#ifndef RE_NREGS
394#define RE_NREGS 30
395#endif
396
397
398/* POSIX specification for registers. Aside from the different names than
399 `re_registers', POSIX uses an array of structures, instead of a
400 structure of arrays. */
401typedef struct
402{
403 regoff_t rm_so; /* Byte offset from string's start to substring's start. */
404 regoff_t rm_eo; /* Byte offset from string's start to substring's end. */
405} regmatch_t;
406
407/* Declarations for routines. */
408
409/* To avoid duplicating every routine declaration -- once with a
410 prototype (if we are ANSI), and once without (if we aren't) -- we
411 use the following macro to declare argument types. This
412 unfortunately clutters up the declarations a bit, but I think it's
413 worth it. */
414
415#ifdef __STDC__
416
417#define _RE_ARGS(args) args
418
419#else /* not __STDC__ */
420
421#define _RE_ARGS(args) ()
422
423#endif /* not __STDC__ */
424
425/* Sets the current default syntax to SYNTAX, and return the old syntax.
426 You can also simply assign to the `re_syntax_options' variable. */
427extern reg_syntax_t re_set_syntax _RE_ARGS ((reg_syntax_t syntax));
428
429/* Compile the regular expression PATTERN, with length LENGTH
430 and syntax given by the global `re_syntax_options', into the buffer
431 BUFFER. Return NULL if successful, and an error string if not. */
432extern const char *re_compile_pattern
8extern const char *__re_compile_pattern
433 _RE_ARGS ((const char *pattern, size_t length,
434 struct re_pattern_buffer *buffer));
435
9 _RE_ARGS ((const char *pattern, size_t length,
10 struct re_pattern_buffer *buffer));
11
12extern int __re_compile_fastmap _RE_ARGS ((struct re_pattern_buffer *buffer));
436
13
437/* Compile a fastmap for the compiled pattern in BUFFER; used to
438 accelerate searches. Return 0 if successful and -2 if was an
439 internal error. */
440extern int re_compile_fastmap _RE_ARGS ((struct re_pattern_buffer *buffer));
441
442
443/* Search in the string STRING (with length LENGTH) for the pattern
444 compiled into BUFFER. Start searching at position START, for RANGE
445 characters. Return the starting position of the match, -1 for no
446 match, or -2 for an internal error. Also return register
447 information in REGS (if REGS and BUFFER->no_sub are nonzero). */
448extern int re_search
14extern int __re_search
449 _RE_ARGS ((struct re_pattern_buffer *buffer, const char *string,
450 int length, int start, int range, struct re_registers *regs));
451
15 _RE_ARGS ((struct re_pattern_buffer *buffer, const char *string,
16 int length, int start, int range, struct re_registers *regs));
17
452
453/* Like `re_search', but search in the concatenation of STRING1 and
454 STRING2. Also, stop searching at index START + STOP. */
455extern int re_search_2
18extern int __re_search_2
456 _RE_ARGS ((struct re_pattern_buffer *buffer, const char *string1,
457 int length1, const char *string2, int length2,
458 int start, int range, struct re_registers *regs, int stop));
459
19 _RE_ARGS ((struct re_pattern_buffer *buffer, const char *string1,
20 int length1, const char *string2, int length2,
21 int start, int range, struct re_registers *regs, int stop));
22
460
461/* Like `re_search', but return how many characters in STRING the regexp
462 in BUFFER matched, starting at position START. */
463extern int re_match
23extern int __re_match
464 _RE_ARGS ((struct re_pattern_buffer *buffer, const char *string,
465 int length, int start, struct re_registers *regs));
466
24 _RE_ARGS ((struct re_pattern_buffer *buffer, const char *string,
25 int length, int start, struct re_registers *regs));
26
467
468/* Relates to `re_match' as `re_search_2' relates to `re_search'. */
469extern int re_match_2
27extern int __re_match_2
470 _RE_ARGS ((struct re_pattern_buffer *buffer, const char *string1,
471 int length1, const char *string2, int length2,
472 int start, struct re_registers *regs, int stop));
473
28 _RE_ARGS ((struct re_pattern_buffer *buffer, const char *string1,
29 int length1, const char *string2, int length2,
30 int start, struct re_registers *regs, int stop));
31
474
475/* Set REGS to hold NUM_REGS registers, storing them in STARTS and
476 ENDS. Subsequent matches using BUFFER and REGS will use this memory
477 for recording register information. STARTS and ENDS must be
478 allocated with malloc, and must each be at least `NUM_REGS * sizeof
479 (regoff_t)' bytes long.
480
481 If NUM_REGS == 0, then subsequent matches should allocate their own
482 register data.
483
484 Unless this function is called, the first search or match using
485 PATTERN_BUFFER will allocate its own register data, without
486 freeing the old data. */
487extern void re_set_registers
32extern void __re_set_registers
488 _RE_ARGS ((struct re_pattern_buffer *buffer, struct re_registers *regs,
489 unsigned num_regs, regoff_t *starts, regoff_t *ends));
490
33 _RE_ARGS ((struct re_pattern_buffer *buffer, struct re_registers *regs,
34 unsigned num_regs, regoff_t *starts, regoff_t *ends));
35
491/* 4.2 bsd compatibility. */
492extern char *re_comp _RE_ARGS ((const char *));
493extern int re_exec _RE_ARGS ((const char *));
36extern int __regcomp _RE_ARGS ((regex_t *__preg, const char *__pattern,
37 int __cflags));
494
38
495/* POSIX compatibility. */
496extern int regcomp _RE_ARGS ((regex_t *preg, const char *pattern, int cflags));
497extern int regexec
498 _RE_ARGS ((const regex_t *preg, const char *string, size_t nmatch,
499 regmatch_t pmatch[], int eflags));
500extern size_t regerror
501 _RE_ARGS ((int errcode, const regex_t *preg, char *errbuf,
502 size_t errbuf_size));
503extern void regfree _RE_ARGS ((regex_t *preg));
39extern int __regexec _RE_ARGS ((const regex_t *__preg,
40 const char *__string, size_t __nmatch,
41 regmatch_t __pmatch[], int __eflags));
504
42
505#ifdef __cplusplus
506}
507#endif /* C++ */
43extern size_t __regerror _RE_ARGS ((int __errcode, const regex_t *__preg,
44 char *__errbuf, size_t __errbuf_size));
508
45
509#endif /* not __REGEXP_LIBRARY_H__ */
510
511/*
512Local variables:
513make-backup-files: t
514version-control: t
515trim-versions-without-asking: nil
516End:
517*/
46extern void __regfree _RE_ARGS ((regex_t *__preg));
47#endif