Commit | Line | Data |
---|---|---|
b1fe2cfc | 1 | /* Definitions for data structures and routines for the regular |
8b20806d | 2 | expression library, version 0.12. |
b1fe2cfc | 3 | |
dd52d970 | 4 | Copyright (C) 1985,89,90,91,92,93,95,2000 Free Software Foundation, Inc. |
b1fe2cfc KB |
5 | |
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 | |
ba4a8e51 KH |
18 | Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, |
19 | USA. */ | |
b1fe2cfc | 20 | |
0b32bf0e SM |
21 | #ifndef _REGEX_H |
22 | #define _REGEX_H 1 | |
23 | ||
24 | /* Allow the use in C++ code. */ | |
25 | #ifdef __cplusplus | |
26 | extern "C" { | |
27 | #endif | |
b1fe2cfc | 28 | |
9114e279 KB |
29 | /* POSIX says that <sys/types.h> must be included (by the caller) before |
30 | <regex.h>. */ | |
31 | ||
0b32bf0e | 32 | #if !defined _POSIX_C_SOURCE && !defined _POSIX_SOURCE && defined VMS |
9114e279 KB |
33 | /* VMS doesn't have `size_t' in <sys/types.h>, even though POSIX says it |
34 | should be there. */ | |
0b32bf0e | 35 | # include <stddef.h> |
d074788d JB |
36 | #endif |
37 | ||
b1fe2cfc | 38 | /* The following bits are used to determine the regexp syntax we |
fb4a568d SM |
39 | recognize. The set/not-set meanings where historically chosen so |
40 | that Emacs syntax had the value 0. | |
41 | The bits are given in alphabetical order, and | |
b1fe2cfc KB |
42 | the definitions shifted by one from the previous bit; thus, when we |
43 | add or remove a bit, only one other definition need change. */ | |
0b32bf0e | 44 | typedef unsigned long int reg_syntax_t; |
f95d599c | 45 | |
b1fe2cfc KB |
46 | /* If this bit is not set, then \ inside a bracket expression is literal. |
47 | If set, then such a \ quotes the following character. */ | |
0b32bf0e | 48 | #define RE_BACKSLASH_ESCAPE_IN_LISTS ((unsigned long int) 1) |
b1fe2cfc KB |
49 | |
50 | /* If this bit is not set, then + and ? are operators, and \+ and \? are | |
0b32bf0e | 51 | literals. |
b1fe2cfc KB |
52 | If set, then \+ and \? are operators and + and ? are literals. */ |
53 | #define RE_BK_PLUS_QM (RE_BACKSLASH_ESCAPE_IN_LISTS << 1) | |
54 | ||
55 | /* If this bit is set, then character classes are supported. They are: | |
56 | [:alpha:], [:upper:], [:lower:], [:digit:], [:alnum:], [:xdigit:], | |
57 | [:space:], [:print:], [:punct:], [:graph:], and [:cntrl:]. | |
58 | If not set, then character classes are not supported. */ | |
59 | #define RE_CHAR_CLASSES (RE_BK_PLUS_QM << 1) | |
60 | ||
61 | /* If this bit is set, then ^ and $ are always anchors (outside bracket | |
62 | expressions, of course). | |
63 | If this bit is not set, then it depends: | |
64 | ^ is an anchor if it is at the beginning of a regular | |
65 | expression or after an open-group or an alternation operator; | |
66 | $ is an anchor if it is at the end of a regular expression, or | |
0b32bf0e | 67 | before a close-group or an alternation operator. |
b1fe2cfc KB |
68 | |
69 | This bit could be (re)combined with RE_CONTEXT_INDEP_OPS, because | |
70 | POSIX draft 11.2 says that * etc. in leading positions is undefined. | |
71 | We already implemented a previous draft which made those constructs | |
72 | invalid, though, so we haven't changed the code back. */ | |
73 | #define RE_CONTEXT_INDEP_ANCHORS (RE_CHAR_CLASSES << 1) | |
74 | ||
75 | /* If this bit is set, then special characters are always special | |
76 | regardless of where they are in the pattern. | |
77 | If this bit is not set, then special characters are special only in | |
0b32bf0e | 78 | some contexts; otherwise they are ordinary. Specifically, |
b1fe2cfc KB |
79 | * + ? and intervals are only special when not after the beginning, |
80 | open-group, or alternation operator. */ | |
81 | #define RE_CONTEXT_INDEP_OPS (RE_CONTEXT_INDEP_ANCHORS << 1) | |
82 | ||
83 | /* If this bit is set, then *, +, ?, and { cannot be first in an re or | |
84 | immediately after an alternation or begin-group operator. */ | |
85 | #define RE_CONTEXT_INVALID_OPS (RE_CONTEXT_INDEP_OPS << 1) | |
86 | ||
87 | /* If this bit is set, then . matches newline. | |
88 | If not set, then it doesn't. */ | |
89 | #define RE_DOT_NEWLINE (RE_CONTEXT_INVALID_OPS << 1) | |
90 | ||
91 | /* If this bit is set, then . doesn't match NUL. | |
92 | If not set, then it does. */ | |
93 | #define RE_DOT_NOT_NULL (RE_DOT_NEWLINE << 1) | |
94 | ||
95 | /* If this bit is set, nonmatching lists [^...] do not match newline. | |
96 | If not set, they do. */ | |
97 | #define RE_HAT_LISTS_NOT_NEWLINE (RE_DOT_NOT_NULL << 1) | |
98 | ||
99 | /* If this bit is set, either \{...\} or {...} defines an | |
0b32bf0e | 100 | interval, depending on RE_NO_BK_BRACES. |
b1fe2cfc KB |
101 | If not set, \{, \}, {, and } are literals. */ |
102 | #define RE_INTERVALS (RE_HAT_LISTS_NOT_NEWLINE << 1) | |
103 | ||
104 | /* If this bit is set, +, ? and | aren't recognized as operators. | |
105 | If not set, they are. */ | |
106 | #define RE_LIMITED_OPS (RE_INTERVALS << 1) | |
107 | ||
108 | /* If this bit is set, newline is an alternation operator. | |
109 | If not set, newline is literal. */ | |
110 | #define RE_NEWLINE_ALT (RE_LIMITED_OPS << 1) | |
111 | ||
112 | /* If this bit is set, then `{...}' defines an interval, and \{ and \} | |
113 | are literals. | |
114 | If not set, then `\{...\}' defines an interval. */ | |
115 | #define RE_NO_BK_BRACES (RE_NEWLINE_ALT << 1) | |
116 | ||
117 | /* If this bit is set, (...) defines a group, and \( and \) are literals. | |
118 | If not set, \(...\) defines a group, and ( and ) are literals. */ | |
119 | #define RE_NO_BK_PARENS (RE_NO_BK_BRACES << 1) | |
120 | ||
121 | /* If this bit is set, then \<digit> matches <digit>. | |
122 | If not set, then \<digit> is a back-reference. */ | |
123 | #define RE_NO_BK_REFS (RE_NO_BK_PARENS << 1) | |
124 | ||
0b32bf0e | 125 | /* If this bit is set, then | is an alternation operator, and \| is literal. |
b1fe2cfc KB |
126 | If not set, then \| is an alternation operator, and | is literal. */ |
127 | #define RE_NO_BK_VBAR (RE_NO_BK_REFS << 1) | |
128 | ||
129 | /* If this bit is set, then an ending range point collating higher | |
130 | than the starting range point, as in [z-a], is invalid. | |
131 | If not set, then when ending range point collates higher than the | |
132 | starting range point, the range is ignored. */ | |
133 | #define RE_NO_EMPTY_RANGES (RE_NO_BK_VBAR << 1) | |
134 | ||
0b32bf0e SM |
135 | /* If this bit is set, then an unmatched ) is ordinary. |
136 | If not set, then an unmatched ) is invalid. */ | |
137 | #define RE_UNMATCHED_RIGHT_PAREN_ORD (RE_NO_EMPTY_RANGES << 1) | |
138 | ||
5cc05562 RS |
139 | /* If this bit is set, succeed as soon as we match the whole pattern, |
140 | without further backtracking. */ | |
0b32bf0e | 141 | #define RE_NO_POSIX_BACKTRACKING (RE_UNMATCHED_RIGHT_PAREN_ORD << 1) |
f95d599c | 142 | |
0b32bf0e SM |
143 | /* If this bit is set, do not process the GNU regex operators. |
144 | If not set, then the GNU regex operators are recognized. */ | |
145 | #define RE_NO_GNU_OPS (RE_NO_POSIX_BACKTRACKING << 1) | |
dd52d970 | 146 | |
0b32bf0e | 147 | /* If this bit is set, then *?, +? and ?? match non greedily. */ |
98222d27 | 148 | #define RE_FRUGAL (RE_NO_GNU_OPS << 1) |
0b32bf0e SM |
149 | |
150 | /* If this bit is set, then (?:...) is treated as a shy group. */ | |
151 | #define RE_SHY_GROUPS (RE_FRUGAL << 1) | |
5cc05562 | 152 | |
b1fe2cfc KB |
153 | /* This global variable defines the particular regexp syntax to use (for |
154 | some interfaces). When a regexp is compiled, the syntax used is | |
155 | stored in the pattern buffer, so changing this does not affect | |
156 | already-compiled regexps. */ | |
157 | extern reg_syntax_t re_syntax_options; | |
766ff627 RS |
158 | |
159 | #ifdef emacs | |
160 | /* In Emacs, this is the string or buffer in which we | |
161 | are matching. It is used for looking up syntax properties. */ | |
162 | extern Lisp_Object re_match_object; | |
163 | #endif | |
164 | ||
b1fe2cfc KB |
165 | \f |
166 | /* Define combinations of the above bits for the standard possibilities. | |
167 | (The [[[ comments delimit what gets put into the Texinfo file, so | |
fb4a568d | 168 | don't delete them!) */ |
b1fe2cfc | 169 | /* [[[begin syntaxes]]] */ |
0b32bf0e SM |
170 | #define RE_SYNTAX_EMACS \ |
171 | (RE_CHAR_CLASSES | RE_INTERVALS | RE_SHY_GROUPS | RE_FRUGAL) | |
b1fe2cfc KB |
172 | |
173 | #define RE_SYNTAX_AWK \ | |
0b32bf0e SM |
174 | (RE_BACKSLASH_ESCAPE_IN_LISTS | RE_DOT_NOT_NULL \ |
175 | | RE_NO_BK_PARENS | RE_NO_BK_REFS \ | |
176 | | RE_NO_BK_VBAR | RE_NO_EMPTY_RANGES \ | |
177 | | RE_DOT_NEWLINE | RE_CONTEXT_INDEP_ANCHORS \ | |
178 | | RE_UNMATCHED_RIGHT_PAREN_ORD | RE_NO_GNU_OPS) | |
179 | ||
180 | #define RE_SYNTAX_GNU_AWK \ | |
181 | ((RE_SYNTAX_POSIX_EXTENDED | RE_BACKSLASH_ESCAPE_IN_LISTS | RE_DEBUG) \ | |
182 | & ~(RE_DOT_NOT_NULL | RE_INTERVALS | RE_CONTEXT_INDEP_OPS)) | |
b1fe2cfc KB |
183 | |
184 | #define RE_SYNTAX_POSIX_AWK \ | |
0b32bf0e SM |
185 | (RE_SYNTAX_POSIX_EXTENDED | RE_BACKSLASH_ESCAPE_IN_LISTS \ |
186 | | RE_INTERVALS | RE_NO_GNU_OPS) | |
b1fe2cfc KB |
187 | |
188 | #define RE_SYNTAX_GREP \ | |
189 | (RE_BK_PLUS_QM | RE_CHAR_CLASSES \ | |
190 | | RE_HAT_LISTS_NOT_NEWLINE | RE_INTERVALS \ | |
0b32bf0e | 191 | | RE_NEWLINE_ALT) |
b1fe2cfc KB |
192 | |
193 | #define RE_SYNTAX_EGREP \ | |
194 | (RE_CHAR_CLASSES | RE_CONTEXT_INDEP_ANCHORS \ | |
195 | | RE_CONTEXT_INDEP_OPS | RE_HAT_LISTS_NOT_NEWLINE \ | |
196 | | RE_NEWLINE_ALT | RE_NO_BK_PARENS \ | |
0b32bf0e | 197 | | RE_NO_BK_VBAR) |
b1fe2cfc KB |
198 | |
199 | #define RE_SYNTAX_POSIX_EGREP \ | |
200 | (RE_SYNTAX_EGREP | RE_INTERVALS | RE_NO_BK_BRACES) | |
201 | ||
9114e279 KB |
202 | /* P1003.2/D11.2, section 4.20.7.1, lines 5078ff. */ |
203 | #define RE_SYNTAX_ED RE_SYNTAX_POSIX_BASIC | |
204 | ||
b1fe2cfc KB |
205 | #define RE_SYNTAX_SED RE_SYNTAX_POSIX_BASIC |
206 | ||
207 | /* Syntax bits common to both basic and extended POSIX regex syntax. */ | |
208 | #define _RE_SYNTAX_POSIX_COMMON \ | |
209 | (RE_CHAR_CLASSES | RE_DOT_NEWLINE | RE_DOT_NOT_NULL \ | |
0b32bf0e | 210 | | RE_INTERVALS | RE_NO_EMPTY_RANGES) |
b1fe2cfc KB |
211 | |
212 | #define RE_SYNTAX_POSIX_BASIC \ | |
213 | (_RE_SYNTAX_POSIX_COMMON | RE_BK_PLUS_QM) | |
214 | ||
215 | /* Differs from ..._POSIX_BASIC only in that RE_BK_PLUS_QM becomes | |
216 | RE_LIMITED_OPS, i.e., \? \+ \| are not recognized. Actually, this | |
217 | isn't minimal, since other operators, such as \`, aren't disabled. */ | |
218 | #define RE_SYNTAX_POSIX_MINIMAL_BASIC \ | |
219 | (_RE_SYNTAX_POSIX_COMMON | RE_LIMITED_OPS) | |
220 | ||
221 | #define RE_SYNTAX_POSIX_EXTENDED \ | |
0b32bf0e SM |
222 | (_RE_SYNTAX_POSIX_COMMON | RE_CONTEXT_INDEP_ANCHORS \ |
223 | | RE_CONTEXT_INDEP_OPS | RE_NO_BK_BRACES \ | |
224 | | RE_NO_BK_PARENS | RE_NO_BK_VBAR \ | |
225 | | RE_CONTEXT_INVALID_OPS | RE_UNMATCHED_RIGHT_PAREN_ORD) | |
b1fe2cfc | 226 | |
0b32bf0e SM |
227 | /* Differs from ..._POSIX_EXTENDED in that RE_CONTEXT_INDEP_OPS is |
228 | removed and RE_NO_BK_REFS is added. */ | |
b1fe2cfc KB |
229 | #define RE_SYNTAX_POSIX_MINIMAL_EXTENDED \ |
230 | (_RE_SYNTAX_POSIX_COMMON | RE_CONTEXT_INDEP_ANCHORS \ | |
231 | | RE_CONTEXT_INVALID_OPS | RE_NO_BK_BRACES \ | |
232 | | RE_NO_BK_PARENS | RE_NO_BK_REFS \ | |
233 | | RE_NO_BK_VBAR | RE_UNMATCHED_RIGHT_PAREN_ORD) | |
234 | /* [[[end syntaxes]]] */ | |
235 | \f | |
236 | /* Maximum number of duplicates an interval can allow. Some systems | |
237 | (erroneously) define this in other header files, but we want our | |
238 | value, so remove any previous define. */ | |
239 | #ifdef RE_DUP_MAX | |
0b32bf0e | 240 | # undef RE_DUP_MAX |
b1fe2cfc | 241 | #endif |
0b32bf0e SM |
242 | /* If sizeof(int) == 2, then ((1 << 15) - 1) overflows. */ |
243 | #define RE_DUP_MAX (0x7fff) | |
b1fe2cfc KB |
244 | |
245 | ||
246 | /* POSIX `cflags' bits (i.e., information for `regcomp'). */ | |
247 | ||
248 | /* If this bit is set, then use extended regular expression syntax. | |
249 | If not set, then use basic regular expression syntax. */ | |
250 | #define REG_EXTENDED 1 | |
251 | ||
252 | /* If this bit is set, then ignore case when matching. | |
253 | If not set, then case is significant. */ | |
254 | #define REG_ICASE (REG_EXTENDED << 1) | |
0b32bf0e | 255 | |
b1fe2cfc KB |
256 | /* If this bit is set, then anchors do not match at newline |
257 | characters in the string. | |
258 | If not set, then anchors do match at newlines. */ | |
259 | #define REG_NEWLINE (REG_ICASE << 1) | |
260 | ||
261 | /* If this bit is set, then report only success or fail in regexec. | |
262 | If not set, then returns differ between not matching and errors. */ | |
263 | #define REG_NOSUB (REG_NEWLINE << 1) | |
264 | ||
265 | ||
266 | /* POSIX `eflags' bits (i.e., information for regexec). */ | |
267 | ||
268 | /* If this bit is set, then the beginning-of-line operator doesn't match | |
269 | the beginning of the string (presumably because it's not the | |
270 | beginning of a line). | |
271 | If not set, then the beginning-of-line operator does match the | |
272 | beginning of the string. */ | |
273 | #define REG_NOTBOL 1 | |
274 | ||
275 | /* Like REG_NOTBOL, except for the end-of-line. */ | |
276 | #define REG_NOTEOL (1 << 1) | |
277 | ||
278 | ||
279 | /* If any error codes are removed, changed, or added, update the | |
280 | `re_error_msg' table in regex.c. */ | |
281 | typedef enum | |
282 | { | |
0b32bf0e SM |
283 | #ifdef _XOPEN_SOURCE |
284 | REG_ENOSYS = -1, /* This will never happen for this implementation. */ | |
285 | #endif | |
286 | ||
b1fe2cfc KB |
287 | REG_NOERROR = 0, /* Success. */ |
288 | REG_NOMATCH, /* Didn't find a match (for regexec). */ | |
289 | ||
290 | /* POSIX regcomp return error codes. (In the order listed in the | |
291 | standard.) */ | |
292 | REG_BADPAT, /* Invalid pattern. */ | |
293 | REG_ECOLLATE, /* Not implemented. */ | |
294 | REG_ECTYPE, /* Invalid character class name. */ | |
295 | REG_EESCAPE, /* Trailing backslash. */ | |
296 | REG_ESUBREG, /* Invalid back reference. */ | |
297 | REG_EBRACK, /* Unmatched left bracket. */ | |
0b32bf0e | 298 | REG_EPAREN, /* Parenthesis imbalance. */ |
b1fe2cfc KB |
299 | REG_EBRACE, /* Unmatched \{. */ |
300 | REG_BADBR, /* Invalid contents of \{\}. */ | |
301 | REG_ERANGE, /* Invalid range end. */ | |
302 | REG_ESPACE, /* Ran out of memory. */ | |
303 | REG_BADRPT, /* No preceding re for repetition op. */ | |
304 | ||
305 | /* Error codes we've added. */ | |
306 | REG_EEND, /* Premature end. */ | |
307 | REG_ESIZE, /* Compiled pattern bigger than 2^16 bytes. */ | |
308 | REG_ERPAREN /* Unmatched ) or \); not returned from regcomp. */ | |
309 | } reg_errcode_t; | |
310 | \f | |
311 | /* This data structure represents a compiled pattern. Before calling | |
312 | the pattern compiler, the fields `buffer', `allocated', `fastmap', | |
313 | `translate', and `no_sub' can be set. After the pattern has been | |
314 | compiled, the `re_nsub' field is available. All other fields are | |
315 | private to the regex routines. */ | |
316 | ||
fc4e96c8 | 317 | #ifndef RE_TRANSLATE_TYPE |
0b32bf0e | 318 | # define RE_TRANSLATE_TYPE char * |
6fffb26e RS |
319 | #endif |
320 | ||
b1fe2cfc KB |
321 | struct re_pattern_buffer |
322 | { | |
323 | /* [[[begin pattern_buffer]]] */ | |
324 | /* Space that holds the compiled pattern. It is declared as | |
325 | `unsigned char *' because its elements are | |
326 | sometimes used as array indexes. */ | |
327 | unsigned char *buffer; | |
328 | ||
329 | /* Number of bytes to which `buffer' points. */ | |
98222d27 | 330 | size_t allocated; |
b1fe2cfc KB |
331 | |
332 | /* Number of bytes actually used in `buffer'. */ | |
98222d27 | 333 | size_t used; |
b1fe2cfc KB |
334 | |
335 | /* Syntax setting with which the pattern was compiled. */ | |
336 | reg_syntax_t syntax; | |
337 | ||
338 | /* Pointer to a fastmap, if any, otherwise zero. re_search uses | |
339 | the fastmap, if there is one, to skip over impossible | |
340 | starting points for matches. */ | |
341 | char *fastmap; | |
342 | ||
343 | /* Either a translate table to apply to all characters before | |
344 | comparing them, or zero for no translation. The translation | |
345 | is applied to a pattern when it is compiled and to a string | |
346 | when it is matched. */ | |
6fffb26e | 347 | RE_TRANSLATE_TYPE translate; |
b1fe2cfc KB |
348 | |
349 | /* Number of subexpressions found by the compiler. */ | |
350 | size_t re_nsub; | |
351 | ||
352 | /* Zero if this pattern cannot match the empty string, one else. | |
353 | Well, in truth it's used only in `re_search_2', to see | |
354 | whether or not we should use the fastmap, so we don't set | |
dd52d970 | 355 | this absolutely perfectly; see `re_compile_fastmap'. */ |
b1fe2cfc KB |
356 | unsigned can_be_null : 1; |
357 | ||
358 | /* If REGS_UNALLOCATED, allocate space in the `regs' structure | |
359 | for `max (RE_NREGS, re_nsub + 1)' groups. | |
360 | If REGS_REALLOCATE, reallocate space if necessary. | |
361 | If REGS_FIXED, use what's there. */ | |
362 | #define REGS_UNALLOCATED 0 | |
363 | #define REGS_REALLOCATE 1 | |
364 | #define REGS_FIXED 2 | |
365 | unsigned regs_allocated : 2; | |
366 | ||
9114e279 KB |
367 | /* Set to zero when `regex_compile' compiles a pattern; set to one |
368 | by `re_compile_fastmap' if it updates the fastmap. */ | |
b1fe2cfc KB |
369 | unsigned fastmap_accurate : 1; |
370 | ||
9114e279 KB |
371 | /* If set, `re_match_2' does not return information about |
372 | subexpressions. */ | |
b1fe2cfc KB |
373 | unsigned no_sub : 1; |
374 | ||
375 | /* If set, a beginning-of-line anchor doesn't match at the | |
0b32bf0e | 376 | beginning of the string. */ |
b1fe2cfc KB |
377 | unsigned not_bol : 1; |
378 | ||
379 | /* Similarly for an end-of-line anchor. */ | |
380 | unsigned not_eol : 1; | |
381 | ||
382 | /* If true, an anchor at a newline matches. */ | |
383 | unsigned newline_anchor : 1; | |
384 | ||
2d1675e4 | 385 | #ifdef emacs |
766ff627 RS |
386 | /* If true, multi-byte form in the `buffer' should be recognized as a |
387 | multibyte character. */ | |
388 | unsigned multibyte : 1; | |
2d1675e4 | 389 | #endif |
766ff627 | 390 | |
b1fe2cfc KB |
391 | /* [[[end pattern_buffer]]] */ |
392 | }; | |
393 | ||
394 | typedef struct re_pattern_buffer regex_t; | |
b1fe2cfc KB |
395 | \f |
396 | /* Type for byte offsets within the string. POSIX mandates this. */ | |
397 | typedef int regoff_t; | |
398 | ||
399 | ||
400 | /* This is the structure we store register match data in. See | |
401 | regex.texinfo for a full description of what registers match. */ | |
402 | struct re_registers | |
403 | { | |
404 | unsigned num_regs; | |
405 | regoff_t *start; | |
406 | regoff_t *end; | |
407 | }; | |
408 | ||
409 | ||
410 | /* If `regs_allocated' is REGS_UNALLOCATED in the pattern buffer, | |
411 | `re_match_2' returns information about at least this many registers | |
412 | the first time a `regs' structure is passed. */ | |
413 | #ifndef RE_NREGS | |
0b32bf0e | 414 | # define RE_NREGS 30 |
b1fe2cfc KB |
415 | #endif |
416 | ||
417 | ||
418 | /* POSIX specification for registers. Aside from the different names than | |
419 | `re_registers', POSIX uses an array of structures, instead of a | |
420 | structure of arrays. */ | |
421 | typedef struct | |
422 | { | |
423 | regoff_t rm_so; /* Byte offset from string's start to substring's start. */ | |
424 | regoff_t rm_eo; /* Byte offset from string's start to substring's end. */ | |
425 | } regmatch_t; | |
426 | \f | |
427 | /* Declarations for routines. */ | |
428 | ||
429 | /* To avoid duplicating every routine declaration -- once with a | |
430 | prototype (if we are ANSI), and once without (if we aren't) -- we | |
431 | use the following macro to declare argument types. This | |
432 | unfortunately clutters up the declarations a bit, but I think it's | |
65565c68 | 433 | worth it. */ |
b1fe2cfc | 434 | |
2a90cd71 | 435 | #if defined __STDC__ || defined PROTOTYPES |
0783dc9a | 436 | |
0b32bf0e | 437 | # define _RE_ARGS(args) args |
0783dc9a | 438 | |
2a90cd71 | 439 | #else /* not __STDC__ || PROTOTYPES */ |
0783dc9a | 440 | |
0b32bf0e | 441 | # define _RE_ARGS(args) () |
0783dc9a | 442 | |
2a90cd71 | 443 | #endif /* not __STDC__ || PROTOTYPES */ |
b1fe2cfc KB |
444 | |
445 | /* Sets the current default syntax to SYNTAX, and return the old syntax. | |
446 | You can also simply assign to the `re_syntax_options' variable. */ | |
447 | extern reg_syntax_t re_set_syntax _RE_ARGS ((reg_syntax_t syntax)); | |
448 | ||
449 | /* Compile the regular expression PATTERN, with length LENGTH | |
450 | and syntax given by the global `re_syntax_options', into the buffer | |
451 | BUFFER. Return NULL if successful, and an error string if not. */ | |
452 | extern const char *re_compile_pattern | |
0b32bf0e | 453 | _RE_ARGS ((const char *pattern, size_t length, |
b1fe2cfc KB |
454 | struct re_pattern_buffer *buffer)); |
455 | ||
456 | ||
457 | /* Compile a fastmap for the compiled pattern in BUFFER; used to | |
458 | accelerate searches. Return 0 if successful and -2 if was an | |
459 | internal error. */ | |
460 | extern int re_compile_fastmap _RE_ARGS ((struct re_pattern_buffer *buffer)); | |
461 | ||
462 | ||
463 | /* Search in the string STRING (with length LENGTH) for the pattern | |
464 | compiled into BUFFER. Start searching at position START, for RANGE | |
465 | characters. Return the starting position of the match, -1 for no | |
466 | match, or -2 for an internal error. Also return register | |
467 | information in REGS (if REGS and BUFFER->no_sub are nonzero). */ | |
468 | extern int re_search | |
469 | _RE_ARGS ((struct re_pattern_buffer *buffer, const char *string, | |
470 | int length, int start, int range, struct re_registers *regs)); | |
471 | ||
472 | ||
473 | /* Like `re_search', but search in the concatenation of STRING1 and | |
474 | STRING2. Also, stop searching at index START + STOP. */ | |
475 | extern int re_search_2 | |
476 | _RE_ARGS ((struct re_pattern_buffer *buffer, const char *string1, | |
477 | int length1, const char *string2, int length2, | |
478 | int start, int range, struct re_registers *regs, int stop)); | |
479 | ||
480 | ||
481 | /* Like `re_search', but return how many characters in STRING the regexp | |
482 | in BUFFER matched, starting at position START. */ | |
483 | extern int re_match | |
484 | _RE_ARGS ((struct re_pattern_buffer *buffer, const char *string, | |
485 | int length, int start, struct re_registers *regs)); | |
486 | ||
487 | ||
488 | /* Relates to `re_match' as `re_search_2' relates to `re_search'. */ | |
0b32bf0e | 489 | extern int re_match_2 |
b1fe2cfc KB |
490 | _RE_ARGS ((struct re_pattern_buffer *buffer, const char *string1, |
491 | int length1, const char *string2, int length2, | |
492 | int start, struct re_registers *regs, int stop)); | |
493 | ||
494 | ||
495 | /* Set REGS to hold NUM_REGS registers, storing them in STARTS and | |
496 | ENDS. Subsequent matches using BUFFER and REGS will use this memory | |
497 | for recording register information. STARTS and ENDS must be | |
498 | allocated with malloc, and must each be at least `NUM_REGS * sizeof | |
499 | (regoff_t)' bytes long. | |
500 | ||
501 | If NUM_REGS == 0, then subsequent matches should allocate their own | |
502 | register data. | |
503 | ||
504 | Unless this function is called, the first search or match using | |
505 | PATTERN_BUFFER will allocate its own register data, without | |
506 | freeing the old data. */ | |
507 | extern void re_set_registers | |
508 | _RE_ARGS ((struct re_pattern_buffer *buffer, struct re_registers *regs, | |
509 | unsigned num_regs, regoff_t *starts, regoff_t *ends)); | |
510 | ||
0b32bf0e SM |
511 | #if defined _REGEX_RE_COMP || defined _LIBC |
512 | # ifndef _CRAY | |
b1fe2cfc KB |
513 | /* 4.2 bsd compatibility. */ |
514 | extern char *re_comp _RE_ARGS ((const char *)); | |
515 | extern int re_exec _RE_ARGS ((const char *)); | |
0b32bf0e SM |
516 | # endif |
517 | #endif | |
518 | ||
519 | /* GCC 2.95 and later have "__restrict"; C99 compilers have | |
520 | "restrict", and "configure" may have defined "restrict". */ | |
521 | #ifndef __restrict | |
522 | # if ! (2 < __GNUC__ || (2 == __GNUC__ && 95 <= __GNUC_MINOR__)) | |
523 | # if defined restrict || 199901L <= __STDC_VERSION__ | |
524 | # define __restrict restrict | |
525 | # else | |
526 | # define __restrict | |
527 | # endif | |
528 | # endif | |
273c0c50 | 529 | #endif |
0b32bf0e SM |
530 | /* For now unconditionally define __restrict_arr to expand to nothing. |
531 | Ideally we would have a test for the compiler which allows defining | |
532 | it to restrict. */ | |
533 | #define __restrict_arr | |
b1fe2cfc KB |
534 | |
535 | /* POSIX compatibility. */ | |
0b32bf0e SM |
536 | extern int regcomp _RE_ARGS ((regex_t *__restrict __preg, |
537 | const char *__restrict __pattern, | |
538 | int __cflags)); | |
539 | ||
540 | extern int regexec _RE_ARGS ((const regex_t *__restrict __preg, | |
541 | const char *__restrict __string, size_t __nmatch, | |
542 | regmatch_t __pmatch[__restrict_arr], | |
543 | int __eflags)); | |
544 | ||
545 | extern size_t regerror _RE_ARGS ((int __errcode, const regex_t *__preg, | |
546 | char *__errbuf, size_t __errbuf_size)); | |
547 | ||
548 | extern void regfree _RE_ARGS ((regex_t *__preg)); | |
549 | ||
550 | ||
551 | #ifdef __cplusplus | |
552 | } | |
553 | #endif /* C++ */ | |
554 | ||
555 | #endif /* regex.h */ | |
b1fe2cfc KB |
556 | \f |
557 | /* | |
558 | Local variables: | |
559 | make-backup-files: t | |
560 | version-control: t | |
561 | trim-versions-without-asking: nil | |
562 | End: | |
563 | */ |