1 /* Definitions for data structures and routines for the regular
2 expression library, version 0.12.
4 Copyright (C) 1985, 89, 90, 91, 92, 93, 95 Free Software Foundation, Inc.
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)
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.
16 You should have received a copy of the GNU General Public License
17 along with this program; see the file COPYING. If not, write to
18 the Free Software Foundation, Inc., 59 Temple Place - Suite 330,
19 Boston, MA 02111-1307, USA. */
21 /* Synched up with: FSF 19.29. */
23 #ifndef __REGEXP_LIBRARY_H__
24 #define __REGEXP_LIBRARY_H__
26 /* POSIX says that <sys/types.h> must be included (by the caller) before
30 /* The following bits are used to determine the regexp syntax we
31 recognize. The not-set meaning typically corresponds to the syntax
32 used by Emacs (the exception is RE_INTERVAL, made for historical
33 reasons). The bits are given in alphabetical order, and the
34 definitions shifted by one from the previous bit; thus, when we add or
35 remove a bit, only one other definition need change. */
36 typedef unsigned reg_syntax_t;
38 /* If this bit is not set, then \ inside a bracket expression is literal.
39 If set, then such a \ quotes the following character. */
40 #define RE_BACKSLASH_ESCAPE_IN_LISTS (1)
42 /* If this bit is not set, then + and ? are operators, and \+ and \? are
44 If set, then \+ and \? are operators and + and ? are literals. */
45 #define RE_BK_PLUS_QM (RE_BACKSLASH_ESCAPE_IN_LISTS << 1)
47 /* If this bit is set, then character classes are supported. They are:
48 [:alpha:], [:upper:], [:lower:], [:digit:], [:alnum:], [:xdigit:],
49 [:space:], [:print:], [:punct:], [:graph:], and [:cntrl:].
50 If not set, then character classes are not supported. */
51 #define RE_CHAR_CLASSES (RE_BK_PLUS_QM << 1)
53 /* If this bit is set, then ^ and $ are always anchors (outside bracket
54 expressions, of course).
55 If this bit is not set, then it depends:
56 ^ is an anchor if it is at the beginning of a regular
57 expression or after an open-group or an alternation operator;
58 $ is an anchor if it is at the end of a regular expression, or
59 before a close-group or an alternation operator.
61 This bit could be (re)combined with RE_CONTEXT_INDEP_OPS, because
62 POSIX draft 11.2 says that * etc. in leading positions is undefined.
63 We already implemented a previous draft which made those constructs
64 invalid, though, so we haven't changed the code back. */
65 #define RE_CONTEXT_INDEP_ANCHORS (RE_CHAR_CLASSES << 1)
67 /* If this bit is set, then special characters are always special
68 regardless of where they are in the pattern.
69 If this bit is not set, then special characters are special only in
70 some contexts; otherwise they are ordinary. Specifically,
71 * + ? and intervals are only special when not after the beginning,
72 open-group, or alternation operator. */
73 #define RE_CONTEXT_INDEP_OPS (RE_CONTEXT_INDEP_ANCHORS << 1)
75 /* If this bit is set, then *, +, ?, and { cannot be first in an re or
76 immediately after an alternation or begin-group operator. */
77 #define RE_CONTEXT_INVALID_OPS (RE_CONTEXT_INDEP_OPS << 1)
79 /* If this bit is set, then . matches newline.
80 If not set, then it doesn't. */
81 #define RE_DOT_NEWLINE (RE_CONTEXT_INVALID_OPS << 1)
83 /* If this bit is set, then . doesn't match NUL.
84 If not set, then it does. */
85 #define RE_DOT_NOT_NULL (RE_DOT_NEWLINE << 1)
87 /* If this bit is set, nonmatching lists [^...] do not match newline.
88 If not set, they do. */
89 #define RE_HAT_LISTS_NOT_NEWLINE (RE_DOT_NOT_NULL << 1)
91 /* If this bit is set, either \{...\} or {...} defines an
92 interval, depending on RE_NO_BK_BRACES.
93 If not set, \{, \}, {, and } are literals. */
94 #define RE_INTERVALS (RE_HAT_LISTS_NOT_NEWLINE << 1)
96 /* If this bit is set, +, ? and | aren't recognized as operators.
97 If not set, they are. */
98 #define RE_LIMITED_OPS (RE_INTERVALS << 1)
100 /* If this bit is set, newline is an alternation operator.
101 If not set, newline is literal. */
102 #define RE_NEWLINE_ALT (RE_LIMITED_OPS << 1)
104 /* If this bit is set, then `{...}' defines an interval, and \{ and \}
106 If not set, then `\{...\}' defines an interval. */
107 #define RE_NO_BK_BRACES (RE_NEWLINE_ALT << 1)
109 /* If this bit is set, (...) defines a group, and \( and \) are literals.
110 If not set, \(...\) defines a group, and ( and ) are literals. */
111 #define RE_NO_BK_PARENS (RE_NO_BK_BRACES << 1)
113 /* If this bit is set, then \<digit> matches <digit>.
114 If not set, then \<digit> is a back-reference. */
115 #define RE_NO_BK_REFS (RE_NO_BK_PARENS << 1)
117 /* If this bit is set, then | is an alternation operator, and \| is literal.
118 If not set, then \| is an alternation operator, and | is literal. */
119 #define RE_NO_BK_VBAR (RE_NO_BK_REFS << 1)
121 /* If this bit is set, then an ending range point collating higher
122 than the starting range point, as in [z-a], is invalid.
123 If not set, then when ending range point collates higher than the
124 starting range point, the range is ignored. */
125 #define RE_NO_EMPTY_RANGES (RE_NO_BK_VBAR << 1)
127 /* If this bit is not set, allow minimal matching:
128 - a*? and a+? and a?? perform shortest-possible matching (compare with a*
129 and a+ and a?, respectively, which perform longest-possible matching)
130 - other juxtaposing of * + and ? is rejected.
131 If this bit is set, consecutive * + and ?'s are collapsed in a logical
133 - a*? and a+? are the same as a*
134 - a?? is the same as a?
136 #define RE_NO_MINIMAL_MATCHING (RE_NO_EMPTY_RANGES << 1)
138 /* If this bit is set, succeed as soon as we match the whole pattern,
139 without further backtracking. */
140 #define RE_NO_POSIX_BACKTRACKING (RE_NO_MINIMAL_MATCHING << 1)
142 /* If this bit is not set, (?:re) behaves like (re) (or \(?:re\) behaves like
143 \(re\)) except that the matched string is not registered. */
144 #define RE_NO_SHY_GROUPS (RE_NO_POSIX_BACKTRACKING << 1)
146 /* If this bit is set, then an unmatched ) is ordinary.
147 If not set, then an unmatched ) is invalid. */
148 #define RE_UNMATCHED_RIGHT_PAREN_ORD (RE_NO_SHY_GROUPS << 1)
150 /* This global variable defines the particular regexp syntax to use (for
151 some interfaces). When a regexp is compiled, the syntax used is
152 stored in the pattern buffer, so changing this does not affect
153 already-compiled regexps. */
154 extern reg_syntax_t re_syntax_options;
156 /* Define combinations of the above bits for the standard possibilities.
157 (The [[[ comments delimit what gets put into the Texinfo file, so
158 don't delete them!) */
159 /* [[[begin syntaxes]]] */
160 #define RE_SYNTAX_EMACS RE_INTERVALS
162 #define RE_SYNTAX_AWK \
163 (RE_BACKSLASH_ESCAPE_IN_LISTS | RE_DOT_NOT_NULL \
164 | RE_NO_BK_PARENS | RE_NO_BK_REFS \
165 | RE_NO_BK_VBAR | RE_NO_EMPTY_RANGES \
166 | RE_UNMATCHED_RIGHT_PAREN_ORD | RE_NO_SHY_GROUPS \
167 | RE_NO_MINIMAL_MATCHING)
169 #define RE_SYNTAX_POSIX_AWK \
170 (RE_SYNTAX_POSIX_EXTENDED | RE_BACKSLASH_ESCAPE_IN_LISTS)
172 #define RE_SYNTAX_GREP \
173 (RE_BK_PLUS_QM | RE_CHAR_CLASSES \
174 | RE_HAT_LISTS_NOT_NEWLINE | RE_INTERVALS \
175 | RE_NEWLINE_ALT | RE_NO_SHY_GROUPS \
176 | RE_NO_MINIMAL_MATCHING)
178 #define RE_SYNTAX_EGREP \
179 (RE_CHAR_CLASSES | RE_CONTEXT_INDEP_ANCHORS \
180 | RE_CONTEXT_INDEP_OPS | RE_HAT_LISTS_NOT_NEWLINE \
181 | RE_NEWLINE_ALT | RE_NO_BK_PARENS \
182 | RE_NO_BK_VBAR | RE_NO_SHY_GROUPS \
183 | RE_NO_MINIMAL_MATCHING)
185 #define RE_SYNTAX_POSIX_EGREP \
186 (RE_SYNTAX_EGREP | RE_INTERVALS | RE_NO_BK_BRACES)
188 /* P1003.2/D11.2, section 4.20.7.1, lines 5078ff. */
189 #define RE_SYNTAX_ED RE_SYNTAX_POSIX_BASIC
191 #define RE_SYNTAX_SED RE_SYNTAX_POSIX_BASIC
193 /* Syntax bits common to both basic and extended POSIX regex syntax. */
194 #define _RE_SYNTAX_POSIX_COMMON \
195 (RE_CHAR_CLASSES | RE_DOT_NEWLINE | RE_DOT_NOT_NULL \
196 | RE_INTERVALS | RE_NO_EMPTY_RANGES | RE_NO_SHY_GROUPS \
197 | RE_NO_MINIMAL_MATCHING)
199 #define RE_SYNTAX_POSIX_BASIC \
200 (_RE_SYNTAX_POSIX_COMMON | RE_BK_PLUS_QM)
202 /* Differs from ..._POSIX_BASIC only in that RE_BK_PLUS_QM becomes
203 RE_LIMITED_OPS, i.e., \? \+ \| are not recognized. Actually, this
204 isn't minimal, since other operators, such as \`, aren't disabled. */
205 #define RE_SYNTAX_POSIX_MINIMAL_BASIC \
206 (_RE_SYNTAX_POSIX_COMMON | RE_LIMITED_OPS)
208 #define RE_SYNTAX_POSIX_EXTENDED \
209 (_RE_SYNTAX_POSIX_COMMON | RE_CONTEXT_INDEP_ANCHORS \
210 | RE_CONTEXT_INDEP_OPS | RE_NO_BK_BRACES \
211 | RE_NO_BK_PARENS | RE_NO_BK_VBAR \
212 | RE_UNMATCHED_RIGHT_PAREN_ORD)
214 /* Differs from ..._POSIX_EXTENDED in that RE_CONTEXT_INVALID_OPS
215 replaces RE_CONTEXT_INDEP_OPS and RE_NO_BK_REFS is added. */
216 #define RE_SYNTAX_POSIX_MINIMAL_EXTENDED \
217 (_RE_SYNTAX_POSIX_COMMON | RE_CONTEXT_INDEP_ANCHORS \
218 | RE_CONTEXT_INVALID_OPS | RE_NO_BK_BRACES \
219 | RE_NO_BK_PARENS | RE_NO_BK_REFS \
220 | RE_NO_BK_VBAR | RE_UNMATCHED_RIGHT_PAREN_ORD)
221 /* [[[end syntaxes]]] */
223 /* Maximum number of duplicates an interval can allow. Some systems
224 (erroneously) define this in other header files, but we want our
225 value, so remove any previous define. */
229 #define RE_DUP_MAX ((1 << 15) - 1)
232 /* POSIX `cflags' bits (i.e., information for `regcomp'). */
234 /* If this bit is set, then use extended regular expression syntax.
235 If not set, then use basic regular expression syntax. */
236 #define REG_EXTENDED 1
238 /* If this bit is set, then ignore case when matching.
239 If not set, then case is significant. */
240 #define REG_ICASE (REG_EXTENDED << 1)
242 /* If this bit is set, then anchors do not match at newline
243 characters in the string.
244 If not set, then anchors do match at newlines. */
245 #define REG_NEWLINE (REG_ICASE << 1)
247 /* If this bit is set, then report only success or fail in regexec.
248 If not set, then returns differ between not matching and errors. */
249 #define REG_NOSUB (REG_NEWLINE << 1)
252 /* POSIX `eflags' bits (i.e., information for regexec). */
254 /* If this bit is set, then the beginning-of-line operator doesn't match
255 the beginning of the string (presumably because it's not the
256 beginning of a line).
257 If not set, then the beginning-of-line operator does match the
258 beginning of the string. */
261 /* Like REG_NOTBOL, except for the end-of-line. */
262 #define REG_NOTEOL (1 << 1)
265 /* If any error codes are removed, changed, or added, update the
266 `re_error_msg' table in regex.c. */
269 REG_NOERROR = 0, /* Success. */
270 REG_NOMATCH, /* Didn't find a match (for regexec). */
272 /* POSIX regcomp return error codes. (In the order listed in the
274 REG_BADPAT, /* Invalid pattern. */
275 REG_ECOLLATE, /* Not implemented. */
276 REG_ECTYPE, /* Invalid character class name. */
277 REG_EESCAPE, /* Trailing backslash. */
278 REG_ESUBREG, /* Invalid back reference. */
279 REG_EBRACK, /* Unmatched left bracket. */
280 REG_EPAREN, /* Parenthesis imbalance. */
281 REG_EBRACE, /* Unmatched \{. */
282 REG_BADBR, /* Invalid contents of \{\}. */
283 REG_ERANGE, /* Invalid range end. */
284 REG_ESPACE, /* Ran out of memory. */
285 REG_BADRPT, /* No preceding re for repetition op. */
287 /* Error codes we've added. */
288 REG_EEND, /* Premature end. */
289 REG_ESIZE, /* Compiled pattern bigger than 2^16 bytes. */
290 REG_ERPAREN /* Unmatched ) or \); not returned from regcomp. */
292 ,REG_ESYNTAX /* Invalid syntax designator. */
295 ,REG_ERANGESPAN /* Ranges may not span charsets. */
296 ,REG_ECATEGORY /* Invalid category designator */
300 /* This data structure represents a compiled pattern. Before calling
301 the pattern compiler, the fields `buffer', `allocated', `fastmap',
302 `translate', and `no_sub' can be set. After the pattern has been
303 compiled, the `re_nsub' field is available. All other fields are
304 private to the regex routines. */
306 struct re_pattern_buffer
308 /* [[[begin pattern_buffer]]] */
309 /* Space that holds the compiled pattern. It is declared as
310 `unsigned char *' because its elements are
311 sometimes used as array indexes. */
312 unsigned char *buffer;
314 /* Number of bytes to which `buffer' points. */
315 unsigned long allocated;
317 /* Number of bytes actually used in `buffer'. */
320 /* Syntax setting with which the pattern was compiled. */
323 /* Pointer to a fastmap, if any, otherwise zero. re_search uses
324 the fastmap, if there is one, to skip over impossible
325 starting points for matches. */
328 /* Either a translate table to apply to all characters before
329 comparing them, or zero for no translation. The translation
330 is applied to a pattern when it is compiled and to a string
331 when it is matched. */
334 /* Number of subexpressions found by the compiler. */
337 /* Zero if this pattern cannot match the empty string, one else.
338 Well, in truth it's used only in `re_search_2', to see
339 whether or not we should use the fastmap, so we don't set
340 this absolutely perfectly; see `re_compile_fastmap' (the
341 `duplicate' case). */
342 unsigned can_be_null : 1;
344 /* If REGS_UNALLOCATED, allocate space in the `regs' structure
345 for `max (RE_NREGS, re_nsub + 1)' groups.
346 If REGS_REALLOCATE, reallocate space if necessary.
347 If REGS_FIXED, use what's there. */
348 #define REGS_UNALLOCATED 0
349 #define REGS_REALLOCATE 1
351 unsigned regs_allocated : 2;
353 /* Set to zero when `regex_compile' compiles a pattern; set to one
354 by `re_compile_fastmap' if it updates the fastmap. */
355 unsigned fastmap_accurate : 1;
357 /* If set, `re_match_2' does not return information about
361 /* If set, a beginning-of-line anchor doesn't match at the
362 beginning of the string. */
363 unsigned not_bol : 1;
365 /* Similarly for an end-of-line anchor. */
366 unsigned not_eol : 1;
368 /* If true, an anchor at a newline matches. */
369 unsigned newline_anchor : 1;
371 /* [[[end pattern_buffer]]] */
374 typedef struct re_pattern_buffer regex_t;
376 /* Type for byte offsets within the string. POSIX mandates this. */
377 typedef int regoff_t;
380 /* This is the structure we store register match data in. See
381 regex.texinfo for a full description of what registers match. */
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. */
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. */
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. */
407 /* Declarations for routines. */
409 /* Sets the current default syntax to SYNTAX, and return the old syntax.
410 You can also simply assign to the `re_syntax_options' variable. */
411 reg_syntax_t re_set_syntax (reg_syntax_t syntax);
413 /* Compile the regular expression PATTERN, with length LENGTH
414 and syntax given by the global `re_syntax_options', into the buffer
415 BUFFER. Return NULL if successful, and an error string if not. */
416 CONST char *re_compile_pattern (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 int re_compile_fastmap (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). */
431 int re_search (struct re_pattern_buffer *buffer, CONST char *string,
432 int length, int start, int range,
433 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 int re_search_2 (struct re_pattern_buffer *buffer, CONST char *string1,
439 int length1, CONST char *string2, int length2, int start,
440 int range, struct re_registers *regs, int stop);
443 /* Like `re_search', but return how many characters in STRING the regexp
444 in BUFFER matched, starting at position START. */
445 int re_match (struct re_pattern_buffer *buffer, CONST char *string,
446 int length, int start, struct re_registers *regs);
449 /* Relates to `re_match' as `re_search_2' relates to `re_search'. */
450 int re_match_2 (struct re_pattern_buffer *buffer, CONST char *string1,
451 int length1, CONST char *string2, int length2,
452 int start, struct re_registers *regs, int stop);
455 /* Set REGS to hold NUM_REGS registers, storing them in STARTS and
456 ENDS. Subsequent matches using BUFFER and REGS will use this memory
457 for recording register information. STARTS and ENDS must be
458 allocated with malloc, and must each be at least `NUM_REGS * sizeof
459 (regoff_t)' bytes long.
461 If NUM_REGS == 0, then subsequent matches should allocate their own
464 Unless this function is called, the first search or match using
465 PATTERN_BUFFER will allocate its own register data, without
466 freeing the old data. */
467 void re_set_registers (struct re_pattern_buffer *buffer,
468 struct re_registers *regs, unsigned num_regs,
469 regoff_t *starts, regoff_t *ends);
471 #ifdef _REGEX_RE_COMP
472 /* 4.2 bsd compatibility. */
473 char *re_comp (CONST char *);
474 int re_exec (CONST char *);
477 /* POSIX compatibility. */
478 int regcomp (regex_t *preg, CONST char *pattern, int cflags);
479 int regexec (CONST regex_t *preg, CONST char *string, size_t nmatch,
480 regmatch_t pmatch[], int eflags);
481 size_t regerror (int errcode, CONST regex_t *preg, char *errbuf,
483 void regfree (regex_t *preg);
485 #endif /* not __REGEXP_LIBRARY_H__ */