1 /* Definitions for data structures and routines for the regular
2 expression library, version 0.12.
3 Copyright (C) 1985,89,90,91,92,93,95,96 Free Software Foundation, Inc.
5 This file is part of the GNU C Library. Its master source is NOT part of
6 the C library, however. The master source lives in /gd/gnu/lib.
8 The GNU C Library is free software; you can redistribute it and/or
9 modify it under the terms of the GNU Library General Public License as
10 published by the Free Software Foundation; either version 2 of the
11 License, or (at your option) any later version.
13 The GNU C Library is distributed in the hope that it will be useful,
14 but WITHOUT ANY WARRANTY; without even the implied warranty of
15 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
16 Library General Public License for more details.
18 You should have received a copy of the GNU Library General Public
19 License along with the GNU C Library; see the file COPYING.LIB. If not,
20 write to the Free Software Foundation, Inc., 59 Temple Place - Suite 330,
21 Boston, MA 02111-1307, USA. */
23 #ifndef __REGEXP_LIBRARY_H__
24 #define __REGEXP_LIBRARY_H__
26 /* POSIX says that <sys/types.h> must be included (by the caller) before
29 #if !defined (_POSIX_C_SOURCE) && !defined (_POSIX_SOURCE) && defined (VMS)
30 /* VMS doesn't have `size_t' in <sys/types.h>, even though POSIX says it
36 /* The following bits are used to determine the regexp syntax we
37 recognize. The set/not-set meanings are chosen so that Emacs syntax
38 remains the value 0. The bits are given in alphabetical order, and
39 the definitions shifted by one from the previous bit; thus, when we
40 add or remove a bit, only one other definition need change. */
41 typedef unsigned reg_syntax_t;
43 /* If this bit is not set, then \ inside a bracket expression is literal.
44 If set, then such a \ quotes the following character. */
45 #define RE_BACKSLASH_ESCAPE_IN_LISTS (1)
47 /* If this bit is not set, then + and ? are operators, and \+ and \? are
49 If set, then \+ and \? are operators and + and ? are literals. */
50 #define RE_BK_PLUS_QM (RE_BACKSLASH_ESCAPE_IN_LISTS << 1)
52 /* If this bit is set, then character classes are supported. They are:
53 [:alpha:], [:upper:], [:lower:], [:digit:], [:alnum:], [:xdigit:],
54 [:space:], [:print:], [:punct:], [:graph:], and [:cntrl:].
55 If not set, then character classes are not supported. */
56 #define RE_CHAR_CLASSES (RE_BK_PLUS_QM << 1)
58 /* If this bit is set, then ^ and $ are always anchors (outside bracket
59 expressions, of course).
60 If this bit is not set, then it depends:
61 ^ is an anchor if it is at the beginning of a regular
62 expression or after an open-group or an alternation operator;
63 $ is an anchor if it is at the end of a regular expression, or
64 before a close-group or an alternation operator.
66 This bit could be (re)combined with RE_CONTEXT_INDEP_OPS, because
67 POSIX draft 11.2 says that * etc. in leading positions is undefined.
68 We already implemented a previous draft which made those constructs
69 invalid, though, so we haven't changed the code back. */
70 #define RE_CONTEXT_INDEP_ANCHORS (RE_CHAR_CLASSES << 1)
72 /* If this bit is set, then special characters are always special
73 regardless of where they are in the pattern.
74 If this bit is not set, then special characters are special only in
75 some contexts; otherwise they are ordinary. Specifically,
76 * + ? and intervals are only special when not after the beginning,
77 open-group, or alternation operator. */
78 #define RE_CONTEXT_INDEP_OPS (RE_CONTEXT_INDEP_ANCHORS << 1)
80 /* If this bit is set, then *, +, ?, and { cannot be first in an re or
81 immediately after an alternation or begin-group operator. */
82 #define RE_CONTEXT_INVALID_OPS (RE_CONTEXT_INDEP_OPS << 1)
84 /* If this bit is set, then . matches newline.
85 If not set, then it doesn't. */
86 #define RE_DOT_NEWLINE (RE_CONTEXT_INVALID_OPS << 1)
88 /* If this bit is set, then . doesn't match NUL.
89 If not set, then it does. */
90 #define RE_DOT_NOT_NULL (RE_DOT_NEWLINE << 1)
92 /* If this bit is set, nonmatching lists [^...] do not match newline.
93 If not set, they do. */
94 #define RE_HAT_LISTS_NOT_NEWLINE (RE_DOT_NOT_NULL << 1)
96 /* If this bit is set, either \{...\} or {...} defines an
97 interval, depending on RE_NO_BK_BRACES.
98 If not set, \{, \}, {, and } are literals. */
99 #define RE_INTERVALS (RE_HAT_LISTS_NOT_NEWLINE << 1)
101 /* If this bit is set, +, ? and | aren't recognized as operators.
102 If not set, they are. */
103 #define RE_LIMITED_OPS (RE_INTERVALS << 1)
105 /* If this bit is set, newline is an alternation operator.
106 If not set, newline is literal. */
107 #define RE_NEWLINE_ALT (RE_LIMITED_OPS << 1)
109 /* If this bit is set, then `{...}' defines an interval, and \{ and \}
111 If not set, then `\{...\}' defines an interval. */
112 #define RE_NO_BK_BRACES (RE_NEWLINE_ALT << 1)
114 /* If this bit is set, (...) defines a group, and \( and \) are literals.
115 If not set, \(...\) defines a group, and ( and ) are literals. */
116 #define RE_NO_BK_PARENS (RE_NO_BK_BRACES << 1)
118 /* If this bit is set, then \<digit> matches <digit>.
119 If not set, then \<digit> is a back-reference. */
120 #define RE_NO_BK_REFS (RE_NO_BK_PARENS << 1)
122 /* If this bit is set, then | is an alternation operator, and \| is literal.
123 If not set, then \| is an alternation operator, and | is literal. */
124 #define RE_NO_BK_VBAR (RE_NO_BK_REFS << 1)
126 /* If this bit is set, then an ending range point collating higher
127 than the starting range point, as in [z-a], is invalid.
128 If not set, then when ending range point collates higher than the
129 starting range point, the range is ignored. */
130 #define RE_NO_EMPTY_RANGES (RE_NO_BK_VBAR << 1)
132 /* If this bit is set, then an unmatched ) is ordinary.
133 If not set, then an unmatched ) is invalid. */
134 #define RE_UNMATCHED_RIGHT_PAREN_ORD (RE_NO_EMPTY_RANGES << 1)
136 /* If this bit is set, succeed as soon as we match the whole pattern,
137 without further backtracking. */
138 #define RE_NO_POSIX_BACKTRACKING (RE_UNMATCHED_RIGHT_PAREN_ORD << 1)
140 /* This global variable defines the particular regexp syntax to use (for
141 some interfaces). When a regexp is compiled, the syntax used is
142 stored in the pattern buffer, so changing this does not affect
143 already-compiled regexps. */
144 extern reg_syntax_t re_syntax_options;
146 /* Define combinations of the above bits for the standard possibilities.
147 (The [[[ comments delimit what gets put into the Texinfo file, so
148 don't delete them!) */
149 /* [[[begin syntaxes]]] */
150 #define RE_SYNTAX_EMACS 0
152 #define RE_SYNTAX_AWK \
153 (RE_BACKSLASH_ESCAPE_IN_LISTS | RE_DOT_NOT_NULL \
154 | RE_NO_BK_PARENS | RE_NO_BK_REFS \
155 | RE_NO_BK_VBAR | RE_NO_EMPTY_RANGES \
156 | RE_UNMATCHED_RIGHT_PAREN_ORD)
158 #define RE_SYNTAX_POSIX_AWK \
159 (RE_SYNTAX_POSIX_EXTENDED | RE_BACKSLASH_ESCAPE_IN_LISTS)
161 #define RE_SYNTAX_GREP \
162 (RE_BK_PLUS_QM | RE_CHAR_CLASSES \
163 | RE_HAT_LISTS_NOT_NEWLINE | RE_INTERVALS \
166 #define RE_SYNTAX_EGREP \
167 (RE_CHAR_CLASSES | RE_CONTEXT_INDEP_ANCHORS \
168 | RE_CONTEXT_INDEP_OPS | RE_HAT_LISTS_NOT_NEWLINE \
169 | RE_NEWLINE_ALT | RE_NO_BK_PARENS \
172 #define RE_SYNTAX_POSIX_EGREP \
173 (RE_SYNTAX_EGREP | RE_INTERVALS | RE_NO_BK_BRACES)
175 /* P1003.2/D11.2, section 4.20.7.1, lines 5078ff. */
176 #define RE_SYNTAX_ED RE_SYNTAX_POSIX_BASIC
178 #define RE_SYNTAX_SED RE_SYNTAX_POSIX_BASIC
180 /* Syntax bits common to both basic and extended POSIX regex syntax. */
181 #define _RE_SYNTAX_POSIX_COMMON \
182 (RE_CHAR_CLASSES | RE_DOT_NEWLINE | RE_DOT_NOT_NULL \
183 | RE_INTERVALS | RE_NO_EMPTY_RANGES)
185 #define RE_SYNTAX_POSIX_BASIC \
186 (_RE_SYNTAX_POSIX_COMMON | RE_BK_PLUS_QM)
188 /* Differs from ..._POSIX_BASIC only in that RE_BK_PLUS_QM becomes
189 RE_LIMITED_OPS, i.e., \? \+ \| are not recognized. Actually, this
190 isn't minimal, since other operators, such as \`, aren't disabled. */
191 #define RE_SYNTAX_POSIX_MINIMAL_BASIC \
192 (_RE_SYNTAX_POSIX_COMMON | RE_LIMITED_OPS)
194 #define RE_SYNTAX_POSIX_EXTENDED \
195 (_RE_SYNTAX_POSIX_COMMON | RE_CONTEXT_INDEP_ANCHORS \
196 | RE_CONTEXT_INDEP_OPS | RE_NO_BK_BRACES \
197 | RE_NO_BK_PARENS | RE_NO_BK_VBAR \
198 | RE_UNMATCHED_RIGHT_PAREN_ORD)
200 /* Differs from ..._POSIX_EXTENDED in that RE_CONTEXT_INVALID_OPS
201 replaces RE_CONTEXT_INDEP_OPS and RE_NO_BK_REFS is added. */
202 #define RE_SYNTAX_POSIX_MINIMAL_EXTENDED \
203 (_RE_SYNTAX_POSIX_COMMON | RE_CONTEXT_INDEP_ANCHORS \
204 | RE_CONTEXT_INVALID_OPS | RE_NO_BK_BRACES \
205 | RE_NO_BK_PARENS | RE_NO_BK_REFS \
206 | RE_NO_BK_VBAR | RE_UNMATCHED_RIGHT_PAREN_ORD)
207 /* [[[end syntaxes]]] */
209 /* Maximum number of duplicates an interval can allow. Some systems
210 (erroneously) define this in other header files, but we want our
211 value, so remove any previous define. */
215 #define RE_DUP_MAX ((1 << 15) - 1)
218 /* POSIX `cflags' bits (i.e., information for `regcomp'). */
220 /* If this bit is set, then use extended regular expression syntax.
221 If not set, then use basic regular expression syntax. */
222 #define REG_EXTENDED 1
224 /* If this bit is set, then ignore case when matching.
225 If not set, then case is significant. */
226 #define REG_ICASE (REG_EXTENDED << 1)
228 /* If this bit is set, then anchors do not match at newline
229 characters in the string.
230 If not set, then anchors do match at newlines. */
231 #define REG_NEWLINE (REG_ICASE << 1)
233 /* If this bit is set, then report only success or fail in regexec.
234 If not set, then returns differ between not matching and errors. */
235 #define REG_NOSUB (REG_NEWLINE << 1)
238 /* POSIX `eflags' bits (i.e., information for regexec). */
240 /* If this bit is set, then the beginning-of-line operator doesn't match
241 the beginning of the string (presumably because it's not the
242 beginning of a line).
243 If not set, then the beginning-of-line operator does match the
244 beginning of the string. */
247 /* Like REG_NOTBOL, except for the end-of-line. */
248 #define REG_NOTEOL (1 << 1)
251 /* If any error codes are removed, changed, or added, update the
252 `re_error_msg' table in regex.c. */
255 REG_NOERROR = 0, /* Success. */
256 REG_NOMATCH, /* Didn't find a match (for regexec). */
258 /* POSIX regcomp return error codes. (In the order listed in the
260 REG_BADPAT, /* Invalid pattern. */
261 REG_ECOLLATE, /* Not implemented. */
262 REG_ECTYPE, /* Invalid character class name. */
263 REG_EESCAPE, /* Trailing backslash. */
264 REG_ESUBREG, /* Invalid back reference. */
265 REG_EBRACK, /* Unmatched left bracket. */
266 REG_EPAREN, /* Parenthesis imbalance. */
267 REG_EBRACE, /* Unmatched \{. */
268 REG_BADBR, /* Invalid contents of \{\}. */
269 REG_ERANGE, /* Invalid range end. */
270 REG_ESPACE, /* Ran out of memory. */
271 REG_BADRPT, /* No preceding re for repetition op. */
273 /* Error codes we've added. */
274 REG_EEND, /* Premature end. */
275 REG_ESIZE, /* Compiled pattern bigger than 2^16 bytes. */
276 REG_ERPAREN /* Unmatched ) or \); not returned from regcomp. */
279 /* This data structure represents a compiled pattern. Before calling
280 the pattern compiler, the fields `buffer', `allocated', `fastmap',
281 `translate', and `no_sub' can be set. After the pattern has been
282 compiled, the `re_nsub' field is available. All other fields are
283 private to the regex routines. */
285 #ifndef RE_TRANSLATE_TYPE
286 #define RE_TRANSLATE_TYPE char *
289 struct re_pattern_buffer
291 /* [[[begin pattern_buffer]]] */
292 /* Space that holds the compiled pattern. It is declared as
293 `unsigned char *' because its elements are
294 sometimes used as array indexes. */
295 unsigned char *buffer;
297 /* Number of bytes to which `buffer' points. */
298 unsigned long allocated;
300 /* Number of bytes actually used in `buffer'. */
303 /* Syntax setting with which the pattern was compiled. */
306 /* Pointer to a fastmap, if any, otherwise zero. re_search uses
307 the fastmap, if there is one, to skip over impossible
308 starting points for matches. */
311 /* Either a translate table to apply to all characters before
312 comparing them, or zero for no translation. The translation
313 is applied to a pattern when it is compiled and to a string
314 when it is matched. */
315 RE_TRANSLATE_TYPE translate;
317 /* Number of subexpressions found by the compiler. */
320 /* Zero if this pattern cannot match the empty string, one else.
321 Well, in truth it's used only in `re_search_2', to see
322 whether or not we should use the fastmap, so we don't set
323 this absolutely perfectly; see `re_compile_fastmap' (the
324 `duplicate' case). */
325 unsigned can_be_null : 1;
327 /* If REGS_UNALLOCATED, allocate space in the `regs' structure
328 for `max (RE_NREGS, re_nsub + 1)' groups.
329 If REGS_REALLOCATE, reallocate space if necessary.
330 If REGS_FIXED, use what's there. */
331 #define REGS_UNALLOCATED 0
332 #define REGS_REALLOCATE 1
334 unsigned regs_allocated : 2;
336 /* Set to zero when `regex_compile' compiles a pattern; set to one
337 by `re_compile_fastmap' if it updates the fastmap. */
338 unsigned fastmap_accurate : 1;
340 /* If set, `re_match_2' does not return information about
344 /* If set, a beginning-of-line anchor doesn't match at the
345 beginning of the string. */
346 unsigned not_bol : 1;
348 /* Similarly for an end-of-line anchor. */
349 unsigned not_eol : 1;
351 /* If true, an anchor at a newline matches. */
352 unsigned newline_anchor : 1;
354 /* [[[end pattern_buffer]]] */
357 typedef struct re_pattern_buffer regex_t;
359 /* Type for byte offsets within the string. POSIX mandates this. */
360 typedef int regoff_t;
363 /* This is the structure we store register match data in. See
364 regex.texinfo for a full description of what registers match. */
373 /* If `regs_allocated' is REGS_UNALLOCATED in the pattern buffer,
374 `re_match_2' returns information about at least this many registers
375 the first time a `regs' structure is passed. */
381 /* POSIX specification for registers. Aside from the different names than
382 `re_registers', POSIX uses an array of structures, instead of a
383 structure of arrays. */
386 regoff_t rm_so; /* Byte offset from string's start to substring's start. */
387 regoff_t rm_eo; /* Byte offset from string's start to substring's end. */
390 /* Declarations for routines. */
392 /* To avoid duplicating every routine declaration -- once with a
393 prototype (if we are ANSI), and once without (if we aren't) -- we
394 use the following macro to declare argument types. This
395 unfortunately clutters up the declarations a bit, but I think it's
400 #define _RE_ARGS(args) args
402 #else /* not __STDC__ */
404 #define _RE_ARGS(args) ()
406 #endif /* not __STDC__ */
408 /* Sets the current default syntax to SYNTAX, and return the old syntax.
409 You can also simply assign to the `re_syntax_options' variable. */
410 extern reg_syntax_t re_set_syntax _RE_ARGS ((reg_syntax_t syntax));
412 /* Compile the regular expression PATTERN, with length LENGTH
413 and syntax given by the global `re_syntax_options', into the buffer
414 BUFFER. Return NULL if successful, and an error string if not. */
415 extern const char *re_compile_pattern
416 _RE_ARGS ((const char *pattern, int length,
417 struct re_pattern_buffer *buffer));
420 /* Compile a fastmap for the compiled pattern in BUFFER; used to
421 accelerate searches. Return 0 if successful and -2 if was an
423 extern int re_compile_fastmap _RE_ARGS ((struct re_pattern_buffer *buffer));
426 /* Search in the string STRING (with length LENGTH) for the pattern
427 compiled into BUFFER. Start searching at position START, for RANGE
428 characters. Return the starting position of the match, -1 for no
429 match, or -2 for an internal error. Also return register
430 information in REGS (if REGS and BUFFER->no_sub are nonzero). */
432 _RE_ARGS ((struct re_pattern_buffer *buffer, const char *string,
433 int length, int start, int range, struct re_registers *regs));
436 /* Like `re_search', but search in the concatenation of STRING1 and
437 STRING2. Also, stop searching at index START + STOP. */
438 extern int re_search_2
439 _RE_ARGS ((struct re_pattern_buffer *buffer, const char *string1,
440 int length1, const char *string2, int length2,
441 int start, int range, struct re_registers *regs, int stop));
444 /* Like `re_search', but return how many characters in STRING the regexp
445 in BUFFER matched, starting at position START. */
447 _RE_ARGS ((struct re_pattern_buffer *buffer, const char *string,
448 int length, int start, struct re_registers *regs));
451 /* Relates to `re_match' as `re_search_2' relates to `re_search'. */
452 extern int re_match_2
453 _RE_ARGS ((struct re_pattern_buffer *buffer, const char *string1,
454 int length1, const char *string2, int length2,
455 int start, struct re_registers *regs, int stop));
458 /* Set REGS to hold NUM_REGS registers, storing them in STARTS and
459 ENDS. Subsequent matches using BUFFER and REGS will use this memory
460 for recording register information. STARTS and ENDS must be
461 allocated with malloc, and must each be at least `NUM_REGS * sizeof
462 (regoff_t)' bytes long.
464 If NUM_REGS == 0, then subsequent matches should allocate their own
467 Unless this function is called, the first search or match using
468 PATTERN_BUFFER will allocate its own register data, without
469 freeing the old data. */
470 extern void re_set_registers
471 _RE_ARGS ((struct re_pattern_buffer *buffer, struct re_registers *regs,
472 unsigned num_regs, regoff_t *starts, regoff_t *ends));
474 #ifdef _REGEX_RE_COMP
475 /* 4.2 bsd compatibility. */
476 extern char *re_comp _RE_ARGS ((const char *));
477 extern int re_exec _RE_ARGS ((const char *));
480 /* POSIX compatibility. */
481 extern int regcomp _RE_ARGS ((regex_t *preg, const char *pattern, int cflags));
483 _RE_ARGS ((const regex_t *preg, const char *string, size_t nmatch,
484 regmatch_t pmatch[], int eflags));
485 extern size_t regerror
486 _RE_ARGS ((int errcode, const regex_t *preg, char *errbuf,
487 size_t errbuf_size));
488 extern void regfree _RE_ARGS ((regex_t *preg));
490 #endif /* not __REGEXP_LIBRARY_H__ */
496 trim-versions-without-asking: nil