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; if not, write to the Free Software
18 Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307,
21 #ifndef __REGEXP_LIBRARY_H__
22 #define __REGEXP_LIBRARY_H__
24 /* POSIX says that <sys/types.h> must be included (by the caller) before
27 #if !defined (_POSIX_C_SOURCE) && !defined (_POSIX_SOURCE) && defined (VMS)
28 /* VMS doesn't have `size_t' in <sys/types.h>, even though POSIX says it
34 /* The following bits are used to determine the regexp syntax we
35 recognize. The set/not-set meanings are chosen so that Emacs syntax
36 remains the value 0. The bits are given in alphabetical order, and
37 the definitions shifted by one from the previous bit; thus, when we
38 add or remove a bit, only one other definition need change. */
39 typedef unsigned reg_syntax_t;
41 /* If this bit is set, then even *?, +? and ?? match greedily. */
42 #define RE_ALL_GREEDY (1)
44 /* If this bit is not set, then \ inside a bracket expression is literal.
45 If set, then such a \ quotes the following character. */
46 #define RE_BACKSLASH_ESCAPE_IN_LISTS (RE_ALL_GREEDY << 1)
48 /* If this bit is not set, then + and ? are operators, and \+ and \? are
50 If set, then \+ and \? are operators and + and ? are literals. */
51 #define RE_BK_PLUS_QM (RE_BACKSLASH_ESCAPE_IN_LISTS << 1)
53 /* If this bit is set, then character classes are supported. They are:
54 [:alpha:], [:upper:], [:lower:], [:digit:], [:alnum:], [:xdigit:],
55 [:space:], [:print:], [:punct:], [:graph:], and [:cntrl:].
56 If not set, then character classes are not supported. */
57 #define RE_CHAR_CLASSES (RE_BK_PLUS_QM << 1)
59 /* If this bit is set, then ^ and $ are always anchors (outside bracket
60 expressions, of course).
61 If this bit is not set, then it depends:
62 ^ is an anchor if it is at the beginning of a regular
63 expression or after an open-group or an alternation operator;
64 $ is an anchor if it is at the end of a regular expression, or
65 before a close-group or an alternation operator.
67 This bit could be (re)combined with RE_CONTEXT_INDEP_OPS, because
68 POSIX draft 11.2 says that * etc. in leading positions is undefined.
69 We already implemented a previous draft which made those constructs
70 invalid, though, so we haven't changed the code back. */
71 #define RE_CONTEXT_INDEP_ANCHORS (RE_CHAR_CLASSES << 1)
73 /* If this bit is set, then special characters are always special
74 regardless of where they are in the pattern.
75 If this bit is not set, then special characters are special only in
76 some contexts; otherwise they are ordinary. Specifically,
77 * + ? and intervals are only special when not after the beginning,
78 open-group, or alternation operator. */
79 #define RE_CONTEXT_INDEP_OPS (RE_CONTEXT_INDEP_ANCHORS << 1)
81 /* If this bit is set, then *, +, ?, and { cannot be first in an re or
82 immediately after an alternation or begin-group operator. */
83 #define RE_CONTEXT_INVALID_OPS (RE_CONTEXT_INDEP_OPS << 1)
85 /* If this bit is set, then . matches newline.
86 If not set, then it doesn't. */
87 #define RE_DOT_NEWLINE (RE_CONTEXT_INVALID_OPS << 1)
89 /* If this bit is set, then . doesn't match NUL.
90 If not set, then it does. */
91 #define RE_DOT_NOT_NULL (RE_DOT_NEWLINE << 1)
93 /* If this bit is set, nonmatching lists [^...] do not match newline.
94 If not set, they do. */
95 #define RE_HAT_LISTS_NOT_NEWLINE (RE_DOT_NOT_NULL << 1)
97 /* If this bit is set, either \{...\} or {...} defines an
98 interval, depending on RE_NO_BK_BRACES.
99 If not set, \{, \}, {, and } are literals. */
100 #define RE_INTERVALS (RE_HAT_LISTS_NOT_NEWLINE << 1)
102 /* If this bit is set, +, ? and | aren't recognized as operators.
103 If not set, they are. */
104 #define RE_LIMITED_OPS (RE_INTERVALS << 1)
106 /* If this bit is set, newline is an alternation operator.
107 If not set, newline is literal. */
108 #define RE_NEWLINE_ALT (RE_LIMITED_OPS << 1)
110 /* If this bit is set, then `{...}' defines an interval, and \{ and \}
112 If not set, then `\{...\}' defines an interval. */
113 #define RE_NO_BK_BRACES (RE_NEWLINE_ALT << 1)
115 /* If this bit is set, (...) defines a group, and \( and \) are literals.
116 If not set, \(...\) defines a group, and ( and ) are literals. */
117 #define RE_NO_BK_PARENS (RE_NO_BK_BRACES << 1)
119 /* If this bit is set, then \<digit> matches <digit>.
120 If not set, then \<digit> is a back-reference. */
121 #define RE_NO_BK_REFS (RE_NO_BK_PARENS << 1)
123 /* If this bit is set, then | is an alternation operator, and \| is literal.
124 If not set, then \| is an alternation operator, and | is literal. */
125 #define RE_NO_BK_VBAR (RE_NO_BK_REFS << 1)
127 /* If this bit is set, then an ending range point collating higher
128 than the starting range point, as in [z-a], is invalid.
129 If not set, then when ending range point collates higher than the
130 starting range point, the range is ignored. */
131 #define RE_NO_EMPTY_RANGES (RE_NO_BK_VBAR << 1)
133 /* If this bit is set, succeed as soon as we match the whole pattern,
134 without further backtracking. */
135 #define RE_NO_POSIX_BACKTRACKING (RE_NO_EMPTY_RANGES << 1)
137 /* If this bit is set, then an unmatched ) is ordinary.
138 If not set, then an unmatched ) is invalid. */
139 #define RE_UNMATCHED_RIGHT_PAREN_ORD (RE_NO_POSIX_BACKTRACKING << 1)
141 /* This global variable defines the particular regexp syntax to use (for
142 some interfaces). When a regexp is compiled, the syntax used is
143 stored in the pattern buffer, so changing this does not affect
144 already-compiled regexps. */
145 extern reg_syntax_t re_syntax_options;
148 /* In Emacs, this is the string or buffer in which we
149 are matching. It is used for looking up syntax properties. */
150 extern Lisp_Object re_match_object;
154 /* Define combinations of the above bits for the standard possibilities.
155 (The [[[ comments delimit what gets put into the Texinfo file, so
156 don't delete them!) */
157 /* [[[begin syntaxes]]] */
158 #define RE_SYNTAX_EMACS 0
160 #define RE_SYNTAX_AWK \
161 (RE_BACKSLASH_ESCAPE_IN_LISTS | RE_DOT_NOT_NULL \
162 | RE_NO_BK_PARENS | RE_NO_BK_REFS \
163 | RE_NO_BK_VBAR | RE_NO_EMPTY_RANGES \
164 | RE_UNMATCHED_RIGHT_PAREN_ORD | RE_ALL_GREEDY)
166 #define RE_SYNTAX_POSIX_AWK \
167 (RE_SYNTAX_POSIX_EXTENDED | RE_BACKSLASH_ESCAPE_IN_LISTS)
169 #define RE_SYNTAX_GREP \
170 (RE_BK_PLUS_QM | RE_CHAR_CLASSES \
171 | RE_HAT_LISTS_NOT_NEWLINE | RE_INTERVALS \
172 | RE_NEWLINE_ALT | RE_ALL_GREEDY)
174 #define RE_SYNTAX_EGREP \
175 (RE_CHAR_CLASSES | RE_CONTEXT_INDEP_ANCHORS \
176 | RE_CONTEXT_INDEP_OPS | RE_HAT_LISTS_NOT_NEWLINE \
177 | RE_NEWLINE_ALT | RE_NO_BK_PARENS \
178 | RE_NO_BK_VBAR | RE_ALL_GREEDY)
180 #define RE_SYNTAX_POSIX_EGREP \
181 (RE_SYNTAX_EGREP | RE_INTERVALS | RE_NO_BK_BRACES)
183 /* P1003.2/D11.2, section 4.20.7.1, lines 5078ff. */
184 #define RE_SYNTAX_ED RE_SYNTAX_POSIX_BASIC
186 #define RE_SYNTAX_SED RE_SYNTAX_POSIX_BASIC
188 /* Syntax bits common to both basic and extended POSIX regex syntax. */
189 #define _RE_SYNTAX_POSIX_COMMON \
190 (RE_CHAR_CLASSES | RE_DOT_NEWLINE | RE_DOT_NOT_NULL \
191 | RE_INTERVALS | RE_NO_EMPTY_RANGES | RE_ALL_GREEDY)
193 #define RE_SYNTAX_POSIX_BASIC \
194 (_RE_SYNTAX_POSIX_COMMON | RE_BK_PLUS_QM)
196 /* Differs from ..._POSIX_BASIC only in that RE_BK_PLUS_QM becomes
197 RE_LIMITED_OPS, i.e., \? \+ \| are not recognized. Actually, this
198 isn't minimal, since other operators, such as \`, aren't disabled. */
199 #define RE_SYNTAX_POSIX_MINIMAL_BASIC \
200 (_RE_SYNTAX_POSIX_COMMON | RE_LIMITED_OPS)
202 #define RE_SYNTAX_POSIX_EXTENDED \
203 (_RE_SYNTAX_POSIX_COMMON | RE_CONTEXT_INDEP_ANCHORS \
204 | RE_CONTEXT_INDEP_OPS | RE_NO_BK_BRACES \
205 | RE_NO_BK_PARENS | RE_NO_BK_VBAR \
206 | RE_UNMATCHED_RIGHT_PAREN_ORD)
208 /* Differs from ..._POSIX_EXTENDED in that RE_CONTEXT_INVALID_OPS
209 replaces RE_CONTEXT_INDEP_OPS and RE_NO_BK_REFS is added. */
210 #define RE_SYNTAX_POSIX_MINIMAL_EXTENDED \
211 (_RE_SYNTAX_POSIX_COMMON | RE_CONTEXT_INDEP_ANCHORS \
212 | RE_CONTEXT_INVALID_OPS | RE_NO_BK_BRACES \
213 | RE_NO_BK_PARENS | RE_NO_BK_REFS \
214 | RE_NO_BK_VBAR | RE_UNMATCHED_RIGHT_PAREN_ORD)
215 /* [[[end syntaxes]]] */
217 /* Maximum number of duplicates an interval can allow. Some systems
218 (erroneously) define this in other header files, but we want our
219 value, so remove any previous define. */
223 #define RE_DUP_MAX ((1 << 15) - 1)
226 /* POSIX `cflags' bits (i.e., information for `regcomp'). */
228 /* If this bit is set, then use extended regular expression syntax.
229 If not set, then use basic regular expression syntax. */
230 #define REG_EXTENDED 1
232 /* If this bit is set, then ignore case when matching.
233 If not set, then case is significant. */
234 #define REG_ICASE (REG_EXTENDED << 1)
236 /* If this bit is set, then anchors do not match at newline
237 characters in the string.
238 If not set, then anchors do match at newlines. */
239 #define REG_NEWLINE (REG_ICASE << 1)
241 /* If this bit is set, then report only success or fail in regexec.
242 If not set, then returns differ between not matching and errors. */
243 #define REG_NOSUB (REG_NEWLINE << 1)
246 /* POSIX `eflags' bits (i.e., information for regexec). */
248 /* If this bit is set, then the beginning-of-line operator doesn't match
249 the beginning of the string (presumably because it's not the
250 beginning of a line).
251 If not set, then the beginning-of-line operator does match the
252 beginning of the string. */
255 /* Like REG_NOTBOL, except for the end-of-line. */
256 #define REG_NOTEOL (1 << 1)
259 /* If any error codes are removed, changed, or added, update the
260 `re_error_msg' table in regex.c. */
263 REG_NOERROR = 0, /* Success. */
264 REG_NOMATCH, /* Didn't find a match (for regexec). */
266 /* POSIX regcomp return error codes. (In the order listed in the
268 REG_BADPAT, /* Invalid pattern. */
269 REG_ECOLLATE, /* Not implemented. */
270 REG_ECTYPE, /* Invalid character class name. */
271 REG_EESCAPE, /* Trailing backslash. */
272 REG_ESUBREG, /* Invalid back reference. */
273 REG_EBRACK, /* Unmatched left bracket. */
274 REG_EPAREN, /* Parenthesis imbalance. */
275 REG_EBRACE, /* Unmatched \{. */
276 REG_BADBR, /* Invalid contents of \{\}. */
277 REG_ERANGE, /* Invalid range end. */
278 REG_ESPACE, /* Ran out of memory. */
279 REG_BADRPT, /* No preceding re for repetition op. */
281 /* Error codes we've added. */
282 REG_EEND, /* Premature end. */
283 REG_ESIZE, /* Compiled pattern bigger than 2^16 bytes. */
284 REG_ERPAREN /* Unmatched ) or \); not returned from regcomp. */
287 /* This data structure represents a compiled pattern. Before calling
288 the pattern compiler, the fields `buffer', `allocated', `fastmap',
289 `translate', and `no_sub' can be set. After the pattern has been
290 compiled, the `re_nsub' field is available. All other fields are
291 private to the regex routines. */
293 #ifndef RE_TRANSLATE_TYPE
294 #define RE_TRANSLATE_TYPE char *
295 #define RE_TRANSLATE(TBL, C) ((TBL)[C])
296 #define RE_TRANSLATE_P(TBL) (TBL)
299 struct re_pattern_buffer
301 /* [[[begin pattern_buffer]]] */
302 /* Space that holds the compiled pattern. It is declared as
303 `unsigned char *' because its elements are
304 sometimes used as array indexes. */
305 unsigned char *buffer;
307 /* Number of bytes to which `buffer' points. */
308 unsigned long allocated;
310 /* Number of bytes actually used in `buffer'. */
313 /* Syntax setting with which the pattern was compiled. */
316 /* Pointer to a fastmap, if any, otherwise zero. re_search uses
317 the fastmap, if there is one, to skip over impossible
318 starting points for matches. */
321 /* Either a translate table to apply to all characters before
322 comparing them, or zero for no translation. The translation
323 is applied to a pattern when it is compiled and to a string
324 when it is matched. */
325 RE_TRANSLATE_TYPE translate;
327 /* Number of subexpressions found by the compiler. */
330 /* Zero if this pattern cannot match the empty string, one else.
331 Well, in truth it's used only in `re_search_2', to see
332 whether or not we should use the fastmap, so we don't set
333 this absolutely perfectly; see `re_compile_fastmap' (the
334 `duplicate' case). */
335 unsigned can_be_null : 1;
337 /* If REGS_UNALLOCATED, allocate space in the `regs' structure
338 for `max (RE_NREGS, re_nsub + 1)' groups.
339 If REGS_REALLOCATE, reallocate space if necessary.
340 If REGS_FIXED, use what's there. */
341 #define REGS_UNALLOCATED 0
342 #define REGS_REALLOCATE 1
344 unsigned regs_allocated : 2;
346 /* Set to zero when `regex_compile' compiles a pattern; set to one
347 by `re_compile_fastmap' if it updates the fastmap. */
348 unsigned fastmap_accurate : 1;
350 /* If set, `re_match_2' does not return information about
354 /* If set, a beginning-of-line anchor doesn't match at the
355 beginning of the string. */
356 unsigned not_bol : 1;
358 /* Similarly for an end-of-line anchor. */
359 unsigned not_eol : 1;
361 /* If true, an anchor at a newline matches. */
362 unsigned newline_anchor : 1;
364 /* If true, multi-byte form in the `buffer' should be recognized as a
365 multibyte character. */
366 unsigned multibyte : 1;
368 /* [[[end pattern_buffer]]] */
371 typedef struct re_pattern_buffer regex_t;
373 /* Type for byte offsets within the string. POSIX mandates this. */
374 typedef int regoff_t;
377 /* This is the structure we store register match data in. See
378 regex.texinfo for a full description of what registers match. */
387 /* If `regs_allocated' is REGS_UNALLOCATED in the pattern buffer,
388 `re_match_2' returns information about at least this many registers
389 the first time a `regs' structure is passed. */
395 /* POSIX specification for registers. Aside from the different names than
396 `re_registers', POSIX uses an array of structures, instead of a
397 structure of arrays. */
400 regoff_t rm_so; /* Byte offset from string's start to substring's start. */
401 regoff_t rm_eo; /* Byte offset from string's start to substring's end. */
404 /* Declarations for routines. */
406 /* To avoid duplicating every routine declaration -- once with a
407 prototype (if we are ANSI), and once without (if we aren't) -- we
408 use the following macro to declare argument types. This
409 unfortunately clutters up the declarations a bit, but I think it's
414 #define _RE_ARGS(args) args
416 #else /* not __STDC__ */
418 #define _RE_ARGS(args) ()
420 #endif /* not __STDC__ */
422 /* Sets the current default syntax to SYNTAX, and return the old syntax.
423 You can also simply assign to the `re_syntax_options' variable. */
424 extern reg_syntax_t re_set_syntax _RE_ARGS ((reg_syntax_t syntax));
426 /* Compile the regular expression PATTERN, with length LENGTH
427 and syntax given by the global `re_syntax_options', into the buffer
428 BUFFER. Return NULL if successful, and an error string if not. */
429 extern const char *re_compile_pattern
430 _RE_ARGS ((const char *pattern, int length,
431 struct re_pattern_buffer *buffer));
434 /* Compile a fastmap for the compiled pattern in BUFFER; used to
435 accelerate searches. Return 0 if successful and -2 if was an
437 extern int re_compile_fastmap _RE_ARGS ((struct re_pattern_buffer *buffer));
440 /* Search in the string STRING (with length LENGTH) for the pattern
441 compiled into BUFFER. Start searching at position START, for RANGE
442 characters. Return the starting position of the match, -1 for no
443 match, or -2 for an internal error. Also return register
444 information in REGS (if REGS and BUFFER->no_sub are nonzero). */
446 _RE_ARGS ((struct re_pattern_buffer *buffer, const char *string,
447 int length, int start, int range, struct re_registers *regs));
450 /* Like `re_search', but search in the concatenation of STRING1 and
451 STRING2. Also, stop searching at index START + STOP. */
452 extern int re_search_2
453 _RE_ARGS ((struct re_pattern_buffer *buffer, const char *string1,
454 int length1, const char *string2, int length2,
455 int start, int range, struct re_registers *regs, int stop));
458 /* Like `re_search', but return how many characters in STRING the regexp
459 in BUFFER matched, starting at position START. */
461 _RE_ARGS ((struct re_pattern_buffer *buffer, const char *string,
462 int length, int start, struct re_registers *regs));
465 /* Relates to `re_match' as `re_search_2' relates to `re_search'. */
466 extern int re_match_2
467 _RE_ARGS ((struct re_pattern_buffer *buffer, const char *string1,
468 int length1, const char *string2, int length2,
469 int start, struct re_registers *regs, int stop));
472 /* Set REGS to hold NUM_REGS registers, storing them in STARTS and
473 ENDS. Subsequent matches using BUFFER and REGS will use this memory
474 for recording register information. STARTS and ENDS must be
475 allocated with malloc, and must each be at least `NUM_REGS * sizeof
476 (regoff_t)' bytes long.
478 If NUM_REGS == 0, then subsequent matches should allocate their own
481 Unless this function is called, the first search or match using
482 PATTERN_BUFFER will allocate its own register data, without
483 freeing the old data. */
484 extern void re_set_registers
485 _RE_ARGS ((struct re_pattern_buffer *buffer, struct re_registers *regs,
486 unsigned num_regs, regoff_t *starts, regoff_t *ends));
488 #ifdef _REGEX_RE_COMP
489 /* 4.2 bsd compatibility. */
490 extern char *re_comp _RE_ARGS ((const char *));
491 extern int re_exec _RE_ARGS ((const char *));
494 /* POSIX compatibility. */
495 extern int regcomp _RE_ARGS ((regex_t *preg, const char *pattern, int cflags));
497 _RE_ARGS ((const regex_t *preg, const char *string, size_t nmatch,
498 regmatch_t pmatch[], int eflags));
499 extern size_t regerror
500 _RE_ARGS ((int errcode, const regex_t *preg, char *errbuf,
501 size_t errbuf_size));
502 extern void regfree _RE_ARGS ((regex_t *preg));
504 #endif /* not __REGEXP_LIBRARY_H__ */
510 trim-versions-without-asking: nil