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 file is part of SXEmacs
8 SXEmacs is free software: you can redistribute it and/or modify
9 it under the terms of the GNU General Public License as published by
10 the Free Software Foundation, either version 3 of the License, or
11 (at your option) any later version.
13 SXEmacs 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
16 GNU General Public License for more details.
18 You should have received a copy of the GNU General Public License
19 along with this program. If not, see <http://www.gnu.org/licenses/>. */
22 /* Synched up with: FSF 19.29. */
24 #ifndef INCLUDED_regex_h_
25 #define INCLUDED_regex_h_
28 #define RE_TRANSLATE_TYPE Lisp_Object
30 #define RE_TRANSLATE_TYPE char *
32 /* type definitions copied from lisp.h */
33 #ifndef SIZEOF_EMACS_INT
34 # define SIZEOF_EMACS_INT SIZEOF_VOID_P
38 # if SIZEOF_EMACS_INT == SIZEOF_LONG
39 # define EMACS_INT long
40 # elif SIZEOF_EMACS_INT == SIZEOF_INT
41 # define EMACS_INT int
42 # elif SIZEOF_EMACS_INT == SIZEOF_LONG_LONG_INT
43 # define EMACS_INT long long
45 # error Unable to determine suitable type for EMACS_INT
49 /* Counts of bytes or array elements */
50 typedef EMACS_INT Memory_count;
51 typedef EMACS_INT Element_count;
55 /* POSIX says that <sys/types.h> must be included (by the caller) before
58 /* The following bits are used to determine the regexp syntax we
59 recognize. The not-set meaning typically corresponds to the syntax
60 used by Emacs (the exception is RE_INTERVAL, made for historical
61 reasons). The bits are given in alphabetical order, and the
62 definitions shifted by one from the previous bit; thus, when we add or
63 remove a bit, only one other definition need change. */
64 typedef unsigned reg_syntax_t;
66 /* If this bit is not set, then \ inside a bracket expression is literal.
67 If set, then such a \ quotes the following character. */
68 #define RE_BACKSLASH_ESCAPE_IN_LISTS (1)
70 /* If this bit is not set, then + and ? are operators, and \+ and \? are
72 If set, then \+ and \? are operators and + and ? are literals. */
73 #define RE_BK_PLUS_QM (RE_BACKSLASH_ESCAPE_IN_LISTS << 1)
75 /* If this bit is set, then character classes are supported. They are:
76 [:alpha:], [:upper:], [:lower:], [:digit:], [:alnum:], [:xdigit:],
77 [:space:], [:print:], [:punct:], [:graph:], and [:cntrl:].
78 If not set, then character classes are not supported. */
79 #define RE_CHAR_CLASSES (RE_BK_PLUS_QM << 1)
81 /* If this bit is set, then ^ and $ are always anchors (outside bracket
82 expressions, of course).
83 If this bit is not set, then it depends:
84 ^ is an anchor if it is at the beginning of a regular
85 expression or after an open-group or an alternation operator;
86 $ is an anchor if it is at the end of a regular expression, or
87 before a close-group or an alternation operator.
89 This bit could be (re)combined with RE_CONTEXT_INDEP_OPS, because
90 POSIX draft 11.2 says that * etc. in leading positions is undefined.
91 We already implemented a previous draft which made those constructs
92 invalid, though, so we haven't changed the code back. */
93 #define RE_CONTEXT_INDEP_ANCHORS (RE_CHAR_CLASSES << 1)
95 /* If this bit is set, then special characters are always special
96 regardless of where they are in the pattern.
97 If this bit is not set, then special characters are special only in
98 some contexts; otherwise they are ordinary. Specifically,
99 * + ? and intervals are only special when not after the beginning,
100 open-group, or alternation operator. */
101 #define RE_CONTEXT_INDEP_OPS (RE_CONTEXT_INDEP_ANCHORS << 1)
103 /* If this bit is set, then *, +, ?, and { cannot be first in an re or
104 immediately after an alternation or begin-group operator. */
105 #define RE_CONTEXT_INVALID_OPS (RE_CONTEXT_INDEP_OPS << 1)
107 /* If this bit is set, then . matches newline.
108 If not set, then it doesn't. */
109 #define RE_DOT_NEWLINE (RE_CONTEXT_INVALID_OPS << 1)
111 /* If this bit is set, then . doesn't match NUL.
112 If not set, then it does. */
113 #define RE_DOT_NOT_NULL (RE_DOT_NEWLINE << 1)
115 /* If this bit is set, nonmatching lists [^...] do not match newline.
116 If not set, they do. */
117 #define RE_HAT_LISTS_NOT_NEWLINE (RE_DOT_NOT_NULL << 1)
119 /* If this bit is set, either \{...\} or {...} defines an
120 interval, depending on RE_NO_BK_BRACES.
121 If not set, \{, \}, {, and } are literals. */
122 #define RE_INTERVALS (RE_HAT_LISTS_NOT_NEWLINE << 1)
124 /* If this bit is set, +, ? and | aren't recognized as operators.
125 If not set, they are. */
126 #define RE_LIMITED_OPS (RE_INTERVALS << 1)
128 /* If this bit is set, newline is an alternation operator.
129 If not set, newline is literal. */
130 #define RE_NEWLINE_ALT (RE_LIMITED_OPS << 1)
132 /* If this bit is set, then `{...}' defines an interval, and \{ and \}
134 If not set, then `\{...\}' defines an interval. */
135 #define RE_NO_BK_BRACES (RE_NEWLINE_ALT << 1)
137 /* If this bit is set, (...) defines a group, and \( and \) are literals.
138 If not set, \(...\) defines a group, and ( and ) are literals. */
139 #define RE_NO_BK_PARENS (RE_NO_BK_BRACES << 1)
141 /* If this bit is set, then \<digit> matches <digit>.
142 If not set, then \<digit> is a back-reference. */
143 #define RE_NO_BK_REFS (RE_NO_BK_PARENS << 1)
145 /* If this bit is set, then | is an alternation operator, and \| is literal.
146 If not set, then \| is an alternation operator, and | is literal. */
147 #define RE_NO_BK_VBAR (RE_NO_BK_REFS << 1)
149 /* If this bit is set, then an ending range point collating higher
150 than the starting range point, as in [z-a], is invalid.
151 If not set, then when ending range point collates higher than the
152 starting range point, the range is ignored. */
153 #define RE_NO_EMPTY_RANGES (RE_NO_BK_VBAR << 1)
155 /* If this bit is not set, allow minimal matching:
156 - a*? and a+? and a?? perform shortest-possible matching (compare with a*
157 and a+ and a?, respectively, which perform longest-possible matching)
158 - other juxtaposing of * + and ? is rejected.
159 If this bit is set, consecutive * + and ?'s are collapsed in a logical
161 - a*? and a+? are the same as a*
162 - a?? is the same as a?
164 #define RE_NO_MINIMAL_MATCHING (RE_NO_EMPTY_RANGES << 1)
166 /* If this bit is set, succeed as soon as we match the whole pattern,
167 without further backtracking. */
168 #define RE_NO_POSIX_BACKTRACKING (RE_NO_MINIMAL_MATCHING << 1)
170 /* If this bit is not set, (?:re) behaves like (re) (or \(?:re\) behaves like
171 \(re\)) except that the matched string is not registered. */
172 #define RE_NO_SHY_GROUPS (RE_NO_POSIX_BACKTRACKING << 1)
174 /* If this bit is set, then an unmatched ) is ordinary.
175 If not set, then an unmatched ) is invalid. */
176 #define RE_UNMATCHED_RIGHT_PAREN_ORD (RE_NO_SHY_GROUPS << 1)
178 /* This global variable defines the particular regexp syntax to use (for
179 some interfaces). When a regexp is compiled, the syntax used is
180 stored in the pattern buffer, so changing this does not affect
181 already-compiled regexps. */
182 extern reg_syntax_t re_syntax_options;
184 /* Define combinations of the above bits for the standard possibilities.
185 (The [[[ comments delimit what gets put into the Texinfo file, so
186 don't delete them!) */
187 /* [[[begin syntaxes]]] */
188 #define RE_SYNTAX_EMACS \
189 (RE_CHAR_CLASSES | RE_INTERVALS)
191 #define RE_SYNTAX_AWK \
192 (RE_BACKSLASH_ESCAPE_IN_LISTS | RE_DOT_NOT_NULL \
193 | RE_NO_BK_PARENS | RE_NO_BK_REFS \
194 | RE_NO_BK_VBAR | RE_NO_EMPTY_RANGES \
195 | RE_DOT_NEWLINE | RE_CONTEXT_INDEP_ANCHORS \
196 | RE_UNMATCHED_RIGHT_PAREN_ORD | RE_NO_GNU_OPS)
198 #define RE_SYNTAX_GNU_AWK \
199 ((RE_SYNTAX_POSIX_EXTENDED | RE_BACKSLASH_ESCAPE_IN_LISTS | RE_DEBUG) \
200 & ~(RE_DOT_NOT_NULL | RE_INTERVALS | RE_CONTEXT_INDEP_OPS))
202 #define RE_SYNTAX_POSIX_AWK \
203 (RE_SYNTAX_POSIX_EXTENDED | RE_BACKSLASH_ESCAPE_IN_LISTS \
204 | RE_INTERVALS | RE_NO_GNU_OPS)
206 #define RE_SYNTAX_GREP \
207 (RE_BK_PLUS_QM | RE_CHAR_CLASSES \
208 | RE_HAT_LISTS_NOT_NEWLINE | RE_INTERVALS \
211 #define RE_SYNTAX_EGREP \
212 (RE_CHAR_CLASSES | RE_CONTEXT_INDEP_ANCHORS \
213 | RE_CONTEXT_INDEP_OPS | RE_HAT_LISTS_NOT_NEWLINE \
214 | RE_NEWLINE_ALT | RE_NO_BK_PARENS \
217 #define RE_SYNTAX_POSIX_EGREP \
218 (RE_SYNTAX_EGREP | RE_INTERVALS | RE_NO_BK_BRACES)
220 /* P1003.2/D11.2, section 4.20.7.1, lines 5078ff. */
221 #define RE_SYNTAX_ED RE_SYNTAX_POSIX_BASIC
223 #define RE_SYNTAX_SED RE_SYNTAX_POSIX_BASIC
225 /* Syntax bits common to both basic and extended POSIX regex syntax. */
226 #define _RE_SYNTAX_POSIX_COMMON \
227 (RE_CHAR_CLASSES | RE_DOT_NEWLINE | RE_DOT_NOT_NULL \
228 | RE_INTERVALS | RE_NO_EMPTY_RANGES)
230 #define RE_SYNTAX_POSIX_BASIC \
231 (_RE_SYNTAX_POSIX_COMMON | RE_BK_PLUS_QM)
233 /* Differs from ..._POSIX_BASIC only in that RE_BK_PLUS_QM becomes
234 RE_LIMITED_OPS, i.e., \? \+ \| are not recognized. Actually, this
235 isn't minimal, since other operators, such as \`, aren't disabled. */
236 #define RE_SYNTAX_POSIX_MINIMAL_BASIC \
237 (_RE_SYNTAX_POSIX_COMMON | RE_LIMITED_OPS)
239 #define RE_SYNTAX_POSIX_EXTENDED \
240 (_RE_SYNTAX_POSIX_COMMON | RE_CONTEXT_INDEP_ANCHORS \
241 | RE_CONTEXT_INDEP_OPS | RE_NO_BK_BRACES \
242 | RE_NO_BK_PARENS | RE_NO_BK_VBAR \
243 | RE_CONTEXT_INVALID_OPS | RE_UNMATCHED_RIGHT_PAREN_ORD)
245 /* Differs from ..._POSIX_EXTENDED in that RE_CONTEXT_INDEP_OPS is
246 removed and RE_NO_BK_REFS is added. */
247 #define RE_SYNTAX_POSIX_MINIMAL_EXTENDED \
248 (_RE_SYNTAX_POSIX_COMMON | RE_CONTEXT_INDEP_ANCHORS \
249 | RE_CONTEXT_INVALID_OPS | RE_NO_BK_BRACES \
250 | RE_NO_BK_PARENS | RE_NO_BK_REFS \
251 | RE_NO_BK_VBAR | RE_UNMATCHED_RIGHT_PAREN_ORD)
252 /* [[[end syntaxes]]] */
254 /* Maximum number of duplicates an interval can allow. Some systems
255 (erroneously) define this in other header files, but we want our
256 value, so remove any previous define. */
260 #define RE_DUP_MAX ((1 << 15) - 1)
262 /* POSIX `cflags' bits (i.e., information for `regcomp'). */
264 /* If this bit is set, then use extended regular expression syntax.
265 If not set, then use basic regular expression syntax. */
266 #define REG_EXTENDED 1
268 /* If this bit is set, then ignore case when matching.
269 If not set, then case is significant. */
270 #define REG_ICASE (REG_EXTENDED << 1)
272 /* If this bit is set, then anchors do not match at newline
273 characters in the string.
274 If not set, then anchors do match at newlines. */
275 #define REG_NEWLINE (REG_ICASE << 1)
277 /* If this bit is set, then report only success or fail in regexec.
278 If not set, then returns differ between not matching and errors. */
279 #define REG_NOSUB (REG_NEWLINE << 1)
281 /* POSIX `eflags' bits (i.e., information for regexec). */
283 /* If this bit is set, then the beginning-of-line operator doesn't match
284 the beginning of the string (presumably because it's not the
285 beginning of a line).
286 If not set, then the beginning-of-line operator does match the
287 beginning of the string. */
290 /* Like REG_NOTBOL, except for the end-of-line. */
291 #define REG_NOTEOL (1 << 1)
293 /* If any error codes are removed, changed, or added, update the
294 `re_error_msg' table in regex.c. */
296 REG_NOERROR = 0, /* Success. */
297 REG_NOMATCH, /* Didn't find a match (for regexec). */
299 /* POSIX regcomp return error codes. (In the order listed in the
301 REG_BADPAT, /* Invalid pattern. */
302 REG_ECOLLATE, /* Not implemented. */
303 REG_ECTYPE, /* Invalid character class name. */
304 REG_EESCAPE, /* Trailing backslash. */
305 REG_ESUBREG, /* Invalid back reference. */
306 REG_EBRACK, /* Unmatched left bracket. */
307 REG_EPAREN, /* Parenthesis imbalance. */
308 REG_EBRACE, /* Unmatched \{. */
309 REG_BADBR, /* Invalid contents of \{\}. */
310 REG_ERANGE, /* Invalid range end. */
311 REG_ESPACE, /* Ran out of memory. */
312 REG_BADRPT, /* No preceding re for repetition op. */
314 /* Error codes we've added. */
315 REG_EEND, /* Premature end. */
316 REG_ESIZE, /* Compiled pattern bigger than 2^16 bytes. */
317 REG_ERPAREN /* Unmatched ) or \); not returned from regcomp. */
319 , REG_ESYNTAX /* Invalid syntax designator. */
322 , REG_ERANGESPAN /* Ranges may not span charsets. */
323 , REG_ECATEGORY /* Invalid category designator */
327 /* This data structure represents a compiled pattern. Before calling
328 the pattern compiler, the fields `buffer', `allocated', `fastmap',
329 `translate', and `no_sub' can be set. After the pattern has been
330 compiled, the `re_nsub' field is available. All other fields are
331 private to the regex routines. */
333 struct re_pattern_buffer {
334 /* [[[begin pattern_buffer]]] */
335 /* Space that holds the compiled pattern. It is declared as
336 `unsigned char *' because its elements are
337 sometimes used as array indexes. */
338 unsigned char *buffer;
340 /* Number of bytes to which `buffer' points. */
341 unsigned long allocated;
343 /* Number of bytes actually used in `buffer'. */
346 /* Syntax setting with which the pattern was compiled. */
349 /* Pointer to a fastmap, if any, otherwise zero. re_search uses
350 the fastmap, if there is one, to skip over impossible
351 starting points for matches. */
354 /* Either a translate table to apply to all characters before
355 comparing them, or zero for no translation. The translation
356 is applied to a pattern when it is compiled and to a string
357 when it is matched. */
358 RE_TRANSLATE_TYPE translate;
360 /* Number of returnable groups found by the compiler. (This does
361 not count shy groups.) */
362 Element_count re_nsub;
364 /* Total number of groups found by the compiler. (Including
366 Element_count re_ngroups;
368 /* Zero if this pattern cannot match the empty string, one else.
369 Well, in truth it's used only in `re_search_2', to see
370 whether or not we should use the fastmap, so we don't set
371 this absolutely perfectly; see `re_compile_fastmap' (the
372 `duplicate' case). */
373 unsigned can_be_null:1;
375 /* If REGS_UNALLOCATED, allocate space in the `regs' structure
376 for `max (RE_NREGS, re_nsub + 1)' groups.
377 If REGS_REALLOCATE, reallocate space if necessary.
378 If REGS_FIXED, use what's there. */
379 #define REGS_UNALLOCATED 0
380 #define REGS_REALLOCATE 1
382 unsigned regs_allocated:2;
384 /* Set to zero when `regex_compile' compiles a pattern; set to one
385 by `re_compile_fastmap' if it updates the fastmap. */
386 unsigned fastmap_accurate:1;
388 /* If set, `re_match_2' does not return information about
392 /* If set, a beginning-of-line anchor doesn't match at the
393 beginning of the string. */
396 /* Similarly for an end-of-line anchor. */
399 /* If true, an anchor at a newline matches. */
400 unsigned newline_anchor:1;
402 /* Mapping between back references and groups (may not be
403 equivalent with shy groups). */
404 int *external_to_internal_register;
406 int external_to_internal_register_size;
408 /* [[[end pattern_buffer]]] */
411 typedef struct re_pattern_buffer regex_t;
413 /* Type for byte offsets within the string. POSIX mandates this. */
414 typedef int regoff_t;
416 /* This is the structure we store register match data in. See
417 regex.texinfo for a full description of what registers match. */
418 struct re_registers {
424 /* If `regs_allocated' is REGS_UNALLOCATED in the pattern buffer,
425 `re_match_2' returns information about at least this many registers
426 the first time a `regs' structure is passed. */
431 /* POSIX specification for registers. Aside from the different names than
432 `re_registers', POSIX uses an array of structures, instead of a
433 structure of arrays. */
435 regoff_t rm_so; /* Byte offset from string's start to substring's start. */
436 regoff_t rm_eo; /* Byte offset from string's start to substring's end. */
439 /* Declarations for routines. */
441 /* Sets the current default syntax to SYNTAX, and return the old syntax.
442 You can also simply assign to the `re_syntax_options' variable. */
443 reg_syntax_t re_set_syntax(reg_syntax_t syntax);
445 /* Compile the regular expression PATTERN, with length LENGTH
446 and syntax given by the global `re_syntax_options', into the buffer
447 BUFFER. Return NULL if successful, and an error string if not. */
448 const char *re_compile_pattern(const char *pattern, int length,
449 struct re_pattern_buffer *buffer);
451 /* Compile a fastmap for the compiled pattern in BUFFER; used to
452 accelerate searches. Return 0 if successful and -2 if was an
454 int re_compile_fastmap(struct re_pattern_buffer *buffer);
456 /* Search in the string STRING (with length LENGTH) for the pattern
457 compiled into BUFFER. Start searching at position START, for RANGE
458 characters. Return the starting position of the match, -1 for no
459 match, or -2 for an internal error. Also return register
460 information in REGS (if REGS and BUFFER->no_sub are nonzero). */
461 int re_search(struct re_pattern_buffer *buffer, const char *string,
462 int length, int start, int range, struct re_registers *regs);
464 /* Like `re_search', but search in the concatenation of STRING1 and
465 STRING2. Also, stop searching at index START + STOP. */
466 int re_search_2(struct re_pattern_buffer *buffer, const char *string1,
467 int length1, const char *string2, int length2, int start,
468 int range, struct re_registers *regs, int stop);
470 /* Like `re_search', but return how many characters in STRING the regexp
471 in BUFFER matched, starting at position START. */
472 int re_match(struct re_pattern_buffer *buffer, const char *string,
473 int length, int start, struct re_registers *regs);
475 /* Relates to `re_match' as `re_search_2' relates to `re_search'. */
476 int re_match_2(struct re_pattern_buffer *buffer, const char *string1,
477 int length1, const char *string2, int length2,
478 int start, struct re_registers *regs, int stop);
480 /* Set REGS to hold NUM_REGS registers, storing them in STARTS and
481 ENDS. Subsequent matches using BUFFER and REGS will use this memory
482 for recording register information. STARTS and ENDS must be
483 allocated with malloc, and must each be at least `NUM_REGS * sizeof
484 (regoff_t)' bytes long.
486 If NUM_REGS == 0, then subsequent matches should allocate their own
489 Unless this function is called, the first search or match using
490 PATTERN_BUFFER will allocate its own register data, without
491 freeing the old data. */
492 void re_set_registers(struct re_pattern_buffer *buffer,
493 struct re_registers *regs, unsigned num_regs,
494 regoff_t * starts, regoff_t * ends);
496 #ifdef _REGEX_RE_COMP
497 /* 4.2 bsd compatibility. */
498 char *re_comp(const char *);
499 int re_exec(const char *);
502 /* POSIX compatibility. */
503 /* #### Arrgh, not any more. But I don't have time to deal with this
504 properly, and I don't know if we should care. - sjt */
505 int regcomp(regex_t * preg, const char *pattern, int cflags);
506 int regexec(const regex_t * preg, const char *string, Element_count nmatch,
507 regmatch_t pmatch[], int eflags);
508 Memory_count regerror(int errcode, const regex_t * preg, char *errbuf,
509 Memory_count errbuf_size);
510 void regfree(regex_t * preg);
512 #endif /* INCLUDED_regex_h_ */