Commit | Line | Data |
---|---|---|
b1fe2cfc | 1 | /* Definitions for data structures and routines for the regular |
8b20806d | 2 | expression library, version 0.12. |
b1fe2cfc | 3 | |
429ab54e | 4 | Copyright (C) 1985, 1989, 1990, 1991, 1992, 1993, 1995, 2000, 2001, |
fff23de2 | 5 | 2002, 2003, 2004, 2005, 2006, 2007, 2008 |
429ab54e | 6 | Free Software Foundation, Inc. |
b1fe2cfc KB |
7 | |
8 | This program is free software; you can redistribute it and/or modify | |
9 | it under the terms of the GNU General Public License as published by | |
e468b87f | 10 | the Free Software Foundation; either version 3, or (at your option) |
b1fe2cfc KB |
11 | any later version. |
12 | ||
13 | This program 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. | |
17 | ||
18 | You should have received a copy of the GNU General Public License | |
19 | along with this program; if not, write to the Free Software | |
4fc5845f | 20 | Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, |
ba4a8e51 | 21 | USA. */ |
b1fe2cfc | 22 | |
0b32bf0e SM |
23 | #ifndef _REGEX_H |
24 | #define _REGEX_H 1 | |
25 | ||
26 | /* Allow the use in C++ code. */ | |
27 | #ifdef __cplusplus | |
28 | extern "C" { | |
29 | #endif | |
b1fe2cfc | 30 | |
9114e279 KB |
31 | /* POSIX says that <sys/types.h> must be included (by the caller) before |
32 | <regex.h>. */ | |
33 | ||
0b32bf0e | 34 | #if !defined _POSIX_C_SOURCE && !defined _POSIX_SOURCE && defined VMS |
9114e279 KB |
35 | /* VMS doesn't have `size_t' in <sys/types.h>, even though POSIX says it |
36 | should be there. */ | |
0b32bf0e | 37 | # include <stddef.h> |
d074788d JB |
38 | #endif |
39 | ||
b1fe2cfc | 40 | /* The following bits are used to determine the regexp syntax we |
fb4a568d SM |
41 | recognize. The set/not-set meanings where historically chosen so |
42 | that Emacs syntax had the value 0. | |
43 | The bits are given in alphabetical order, and | |
b1fe2cfc KB |
44 | the definitions shifted by one from the previous bit; thus, when we |
45 | add or remove a bit, only one other definition need change. */ | |
0b32bf0e | 46 | typedef unsigned long int reg_syntax_t; |
f95d599c | 47 | |
b1fe2cfc KB |
48 | /* If this bit is not set, then \ inside a bracket expression is literal. |
49 | If set, then such a \ quotes the following character. */ | |
0b32bf0e | 50 | #define RE_BACKSLASH_ESCAPE_IN_LISTS ((unsigned long int) 1) |
b1fe2cfc KB |
51 | |
52 | /* If this bit is not set, then + and ? are operators, and \+ and \? are | |
0b32bf0e | 53 | literals. |
b1fe2cfc KB |
54 | If set, then \+ and \? are operators and + and ? are literals. */ |
55 | #define RE_BK_PLUS_QM (RE_BACKSLASH_ESCAPE_IN_LISTS << 1) | |
56 | ||
57 | /* If this bit is set, then character classes are supported. They are: | |
58 | [:alpha:], [:upper:], [:lower:], [:digit:], [:alnum:], [:xdigit:], | |
59 | [:space:], [:print:], [:punct:], [:graph:], and [:cntrl:]. | |
60 | If not set, then character classes are not supported. */ | |
61 | #define RE_CHAR_CLASSES (RE_BK_PLUS_QM << 1) | |
62 | ||
63 | /* If this bit is set, then ^ and $ are always anchors (outside bracket | |
64 | expressions, of course). | |
65 | If this bit is not set, then it depends: | |
66 | ^ is an anchor if it is at the beginning of a regular | |
67 | expression or after an open-group or an alternation operator; | |
68 | $ is an anchor if it is at the end of a regular expression, or | |
0b32bf0e | 69 | before a close-group or an alternation operator. |
b1fe2cfc KB |
70 | |
71 | This bit could be (re)combined with RE_CONTEXT_INDEP_OPS, because | |
72 | POSIX draft 11.2 says that * etc. in leading positions is undefined. | |
73 | We already implemented a previous draft which made those constructs | |
74 | invalid, though, so we haven't changed the code back. */ | |
75 | #define RE_CONTEXT_INDEP_ANCHORS (RE_CHAR_CLASSES << 1) | |
76 | ||
77 | /* If this bit is set, then special characters are always special | |
78 | regardless of where they are in the pattern. | |
79 | If this bit is not set, then special characters are special only in | |
0b32bf0e | 80 | some contexts; otherwise they are ordinary. Specifically, |
b1fe2cfc KB |
81 | * + ? and intervals are only special when not after the beginning, |
82 | open-group, or alternation operator. */ | |
83 | #define RE_CONTEXT_INDEP_OPS (RE_CONTEXT_INDEP_ANCHORS << 1) | |
84 | ||
85 | /* If this bit is set, then *, +, ?, and { cannot be first in an re or | |
86 | immediately after an alternation or begin-group operator. */ | |
87 | #define RE_CONTEXT_INVALID_OPS (RE_CONTEXT_INDEP_OPS << 1) | |
88 | ||
89 | /* If this bit is set, then . matches newline. | |
90 | If not set, then it doesn't. */ | |
91 | #define RE_DOT_NEWLINE (RE_CONTEXT_INVALID_OPS << 1) | |
92 | ||
93 | /* If this bit is set, then . doesn't match NUL. | |
94 | If not set, then it does. */ | |
95 | #define RE_DOT_NOT_NULL (RE_DOT_NEWLINE << 1) | |
96 | ||
97 | /* If this bit is set, nonmatching lists [^...] do not match newline. | |
98 | If not set, they do. */ | |
99 | #define RE_HAT_LISTS_NOT_NEWLINE (RE_DOT_NOT_NULL << 1) | |
100 | ||
101 | /* If this bit is set, either \{...\} or {...} defines an | |
0b32bf0e | 102 | interval, depending on RE_NO_BK_BRACES. |
b1fe2cfc KB |
103 | If not set, \{, \}, {, and } are literals. */ |
104 | #define RE_INTERVALS (RE_HAT_LISTS_NOT_NEWLINE << 1) | |
105 | ||
106 | /* If this bit is set, +, ? and | aren't recognized as operators. | |
107 | If not set, they are. */ | |
108 | #define RE_LIMITED_OPS (RE_INTERVALS << 1) | |
109 | ||
110 | /* If this bit is set, newline is an alternation operator. | |
111 | If not set, newline is literal. */ | |
112 | #define RE_NEWLINE_ALT (RE_LIMITED_OPS << 1) | |
113 | ||
114 | /* If this bit is set, then `{...}' defines an interval, and \{ and \} | |
115 | are literals. | |
116 | If not set, then `\{...\}' defines an interval. */ | |
117 | #define RE_NO_BK_BRACES (RE_NEWLINE_ALT << 1) | |
118 | ||
119 | /* If this bit is set, (...) defines a group, and \( and \) are literals. | |
120 | If not set, \(...\) defines a group, and ( and ) are literals. */ | |
121 | #define RE_NO_BK_PARENS (RE_NO_BK_BRACES << 1) | |
122 | ||
123 | /* If this bit is set, then \<digit> matches <digit>. | |
124 | If not set, then \<digit> is a back-reference. */ | |
125 | #define RE_NO_BK_REFS (RE_NO_BK_PARENS << 1) | |
126 | ||
0b32bf0e | 127 | /* If this bit is set, then | is an alternation operator, and \| is literal. |
b1fe2cfc KB |
128 | If not set, then \| is an alternation operator, and | is literal. */ |
129 | #define RE_NO_BK_VBAR (RE_NO_BK_REFS << 1) | |
130 | ||
131 | /* If this bit is set, then an ending range point collating higher | |
132 | than the starting range point, as in [z-a], is invalid. | |
133 | If not set, then when ending range point collates higher than the | |
134 | starting range point, the range is ignored. */ | |
135 | #define RE_NO_EMPTY_RANGES (RE_NO_BK_VBAR << 1) | |
136 | ||
0b32bf0e SM |
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_EMPTY_RANGES << 1) | |
140 | ||
5cc05562 RS |
141 | /* If this bit is set, succeed as soon as we match the whole pattern, |
142 | without further backtracking. */ | |
0b32bf0e | 143 | #define RE_NO_POSIX_BACKTRACKING (RE_UNMATCHED_RIGHT_PAREN_ORD << 1) |
f95d599c | 144 | |
0b32bf0e SM |
145 | /* If this bit is set, do not process the GNU regex operators. |
146 | If not set, then the GNU regex operators are recognized. */ | |
147 | #define RE_NO_GNU_OPS (RE_NO_POSIX_BACKTRACKING << 1) | |
dd52d970 | 148 | |
0b32bf0e | 149 | /* If this bit is set, then *?, +? and ?? match non greedily. */ |
98222d27 | 150 | #define RE_FRUGAL (RE_NO_GNU_OPS << 1) |
0b32bf0e SM |
151 | |
152 | /* If this bit is set, then (?:...) is treated as a shy group. */ | |
153 | #define RE_SHY_GROUPS (RE_FRUGAL << 1) | |
5cc05562 | 154 | |
c0f9ea08 SM |
155 | /* If this bit is set, ^ and $ only match at beg/end of buffer. */ |
156 | #define RE_NO_NEWLINE_ANCHOR (RE_SHY_GROUPS << 1) | |
157 | ||
158 | /* If this bit is set, turn on internal regex debugging. | |
159 | If not set, and debugging was on, turn it off. | |
160 | This only works if regex.c is compiled -DDEBUG. | |
161 | We define this bit always, so that all that's needed to turn on | |
162 | debugging is to recompile regex.c; the calling code can always have | |
163 | this bit set, and it won't affect anything in the normal case. */ | |
164 | #define RE_DEBUG (RE_NO_NEWLINE_ANCHOR << 1) | |
165 | ||
b1fe2cfc KB |
166 | /* This global variable defines the particular regexp syntax to use (for |
167 | some interfaces). When a regexp is compiled, the syntax used is | |
168 | stored in the pattern buffer, so changing this does not affect | |
169 | already-compiled regexps. */ | |
170 | extern reg_syntax_t re_syntax_options; | |
766ff627 RS |
171 | |
172 | #ifdef emacs | |
173 | /* In Emacs, this is the string or buffer in which we | |
174 | are matching. It is used for looking up syntax properties. */ | |
175 | extern Lisp_Object re_match_object; | |
176 | #endif | |
177 | ||
b1fe2cfc KB |
178 | \f |
179 | /* Define combinations of the above bits for the standard possibilities. | |
180 | (The [[[ comments delimit what gets put into the Texinfo file, so | |
fb4a568d | 181 | don't delete them!) */ |
b1fe2cfc | 182 | /* [[[begin syntaxes]]] */ |
0b32bf0e SM |
183 | #define RE_SYNTAX_EMACS \ |
184 | (RE_CHAR_CLASSES | RE_INTERVALS | RE_SHY_GROUPS | RE_FRUGAL) | |
b1fe2cfc KB |
185 | |
186 | #define RE_SYNTAX_AWK \ | |
0b32bf0e SM |
187 | (RE_BACKSLASH_ESCAPE_IN_LISTS | RE_DOT_NOT_NULL \ |
188 | | RE_NO_BK_PARENS | RE_NO_BK_REFS \ | |
189 | | RE_NO_BK_VBAR | RE_NO_EMPTY_RANGES \ | |
190 | | RE_DOT_NEWLINE | RE_CONTEXT_INDEP_ANCHORS \ | |
191 | | RE_UNMATCHED_RIGHT_PAREN_ORD | RE_NO_GNU_OPS) | |
192 | ||
193 | #define RE_SYNTAX_GNU_AWK \ | |
194 | ((RE_SYNTAX_POSIX_EXTENDED | RE_BACKSLASH_ESCAPE_IN_LISTS | RE_DEBUG) \ | |
195 | & ~(RE_DOT_NOT_NULL | RE_INTERVALS | RE_CONTEXT_INDEP_OPS)) | |
b1fe2cfc KB |
196 | |
197 | #define RE_SYNTAX_POSIX_AWK \ | |
0b32bf0e SM |
198 | (RE_SYNTAX_POSIX_EXTENDED | RE_BACKSLASH_ESCAPE_IN_LISTS \ |
199 | | RE_INTERVALS | RE_NO_GNU_OPS) | |
b1fe2cfc KB |
200 | |
201 | #define RE_SYNTAX_GREP \ | |
202 | (RE_BK_PLUS_QM | RE_CHAR_CLASSES \ | |
203 | | RE_HAT_LISTS_NOT_NEWLINE | RE_INTERVALS \ | |
0b32bf0e | 204 | | RE_NEWLINE_ALT) |
b1fe2cfc KB |
205 | |
206 | #define RE_SYNTAX_EGREP \ | |
207 | (RE_CHAR_CLASSES | RE_CONTEXT_INDEP_ANCHORS \ | |
208 | | RE_CONTEXT_INDEP_OPS | RE_HAT_LISTS_NOT_NEWLINE \ | |
209 | | RE_NEWLINE_ALT | RE_NO_BK_PARENS \ | |
0b32bf0e | 210 | | RE_NO_BK_VBAR) |
b1fe2cfc KB |
211 | |
212 | #define RE_SYNTAX_POSIX_EGREP \ | |
213 | (RE_SYNTAX_EGREP | RE_INTERVALS | RE_NO_BK_BRACES) | |
214 | ||
9114e279 KB |
215 | /* P1003.2/D11.2, section 4.20.7.1, lines 5078ff. */ |
216 | #define RE_SYNTAX_ED RE_SYNTAX_POSIX_BASIC | |
217 | ||
b1fe2cfc KB |
218 | #define RE_SYNTAX_SED RE_SYNTAX_POSIX_BASIC |
219 | ||
220 | /* Syntax bits common to both basic and extended POSIX regex syntax. */ | |
221 | #define _RE_SYNTAX_POSIX_COMMON \ | |
222 | (RE_CHAR_CLASSES | RE_DOT_NEWLINE | RE_DOT_NOT_NULL \ | |
0b32bf0e | 223 | | RE_INTERVALS | RE_NO_EMPTY_RANGES) |
b1fe2cfc KB |
224 | |
225 | #define RE_SYNTAX_POSIX_BASIC \ | |
226 | (_RE_SYNTAX_POSIX_COMMON | RE_BK_PLUS_QM) | |
227 | ||
228 | /* Differs from ..._POSIX_BASIC only in that RE_BK_PLUS_QM becomes | |
229 | RE_LIMITED_OPS, i.e., \? \+ \| are not recognized. Actually, this | |
230 | isn't minimal, since other operators, such as \`, aren't disabled. */ | |
231 | #define RE_SYNTAX_POSIX_MINIMAL_BASIC \ | |
232 | (_RE_SYNTAX_POSIX_COMMON | RE_LIMITED_OPS) | |
233 | ||
234 | #define RE_SYNTAX_POSIX_EXTENDED \ | |
0b32bf0e SM |
235 | (_RE_SYNTAX_POSIX_COMMON | RE_CONTEXT_INDEP_ANCHORS \ |
236 | | RE_CONTEXT_INDEP_OPS | RE_NO_BK_BRACES \ | |
237 | | RE_NO_BK_PARENS | RE_NO_BK_VBAR \ | |
238 | | RE_CONTEXT_INVALID_OPS | RE_UNMATCHED_RIGHT_PAREN_ORD) | |
b1fe2cfc | 239 | |
0b32bf0e SM |
240 | /* Differs from ..._POSIX_EXTENDED in that RE_CONTEXT_INDEP_OPS is |
241 | removed and RE_NO_BK_REFS is added. */ | |
b1fe2cfc KB |
242 | #define RE_SYNTAX_POSIX_MINIMAL_EXTENDED \ |
243 | (_RE_SYNTAX_POSIX_COMMON | RE_CONTEXT_INDEP_ANCHORS \ | |
244 | | RE_CONTEXT_INVALID_OPS | RE_NO_BK_BRACES \ | |
245 | | RE_NO_BK_PARENS | RE_NO_BK_REFS \ | |
246 | | RE_NO_BK_VBAR | RE_UNMATCHED_RIGHT_PAREN_ORD) | |
247 | /* [[[end syntaxes]]] */ | |
248 | \f | |
249 | /* Maximum number of duplicates an interval can allow. Some systems | |
250 | (erroneously) define this in other header files, but we want our | |
251 | value, so remove any previous define. */ | |
252 | #ifdef RE_DUP_MAX | |
0b32bf0e | 253 | # undef RE_DUP_MAX |
b1fe2cfc | 254 | #endif |
0b32bf0e SM |
255 | /* If sizeof(int) == 2, then ((1 << 15) - 1) overflows. */ |
256 | #define RE_DUP_MAX (0x7fff) | |
b1fe2cfc KB |
257 | |
258 | ||
259 | /* POSIX `cflags' bits (i.e., information for `regcomp'). */ | |
260 | ||
261 | /* If this bit is set, then use extended regular expression syntax. | |
262 | If not set, then use basic regular expression syntax. */ | |
263 | #define REG_EXTENDED 1 | |
264 | ||
265 | /* If this bit is set, then ignore case when matching. | |
266 | If not set, then case is significant. */ | |
267 | #define REG_ICASE (REG_EXTENDED << 1) | |
0b32bf0e | 268 | |
b1fe2cfc KB |
269 | /* If this bit is set, then anchors do not match at newline |
270 | characters in the string. | |
271 | If not set, then anchors do match at newlines. */ | |
272 | #define REG_NEWLINE (REG_ICASE << 1) | |
273 | ||
274 | /* If this bit is set, then report only success or fail in regexec. | |
275 | If not set, then returns differ between not matching and errors. */ | |
276 | #define REG_NOSUB (REG_NEWLINE << 1) | |
277 | ||
278 | ||
279 | /* POSIX `eflags' bits (i.e., information for regexec). */ | |
280 | ||
281 | /* If this bit is set, then the beginning-of-line operator doesn't match | |
282 | the beginning of the string (presumably because it's not the | |
283 | beginning of a line). | |
284 | If not set, then the beginning-of-line operator does match the | |
285 | beginning of the string. */ | |
286 | #define REG_NOTBOL 1 | |
287 | ||
288 | /* Like REG_NOTBOL, except for the end-of-line. */ | |
289 | #define REG_NOTEOL (1 << 1) | |
290 | ||
291 | ||
292 | /* If any error codes are removed, changed, or added, update the | |
293 | `re_error_msg' table in regex.c. */ | |
294 | typedef enum | |
295 | { | |
0b32bf0e SM |
296 | #ifdef _XOPEN_SOURCE |
297 | REG_ENOSYS = -1, /* This will never happen for this implementation. */ | |
298 | #endif | |
299 | ||
b1fe2cfc KB |
300 | REG_NOERROR = 0, /* Success. */ |
301 | REG_NOMATCH, /* Didn't find a match (for regexec). */ | |
302 | ||
303 | /* POSIX regcomp return error codes. (In the order listed in the | |
304 | standard.) */ | |
305 | REG_BADPAT, /* Invalid pattern. */ | |
306 | REG_ECOLLATE, /* Not implemented. */ | |
307 | REG_ECTYPE, /* Invalid character class name. */ | |
308 | REG_EESCAPE, /* Trailing backslash. */ | |
309 | REG_ESUBREG, /* Invalid back reference. */ | |
310 | REG_EBRACK, /* Unmatched left bracket. */ | |
0b32bf0e | 311 | REG_EPAREN, /* Parenthesis imbalance. */ |
b1fe2cfc KB |
312 | REG_EBRACE, /* Unmatched \{. */ |
313 | REG_BADBR, /* Invalid contents of \{\}. */ | |
314 | REG_ERANGE, /* Invalid range end. */ | |
315 | REG_ESPACE, /* Ran out of memory. */ | |
316 | REG_BADRPT, /* No preceding re for repetition op. */ | |
317 | ||
318 | /* Error codes we've added. */ | |
319 | REG_EEND, /* Premature end. */ | |
320 | REG_ESIZE, /* Compiled pattern bigger than 2^16 bytes. */ | |
be13f5e5 KH |
321 | REG_ERPAREN, /* Unmatched ) or \); not returned from regcomp. */ |
322 | REG_ERANGEX /* Range striding over charsets. */ | |
b1fe2cfc KB |
323 | } reg_errcode_t; |
324 | \f | |
325 | /* This data structure represents a compiled pattern. Before calling | |
326 | the pattern compiler, the fields `buffer', `allocated', `fastmap', | |
327 | `translate', and `no_sub' can be set. After the pattern has been | |
328 | compiled, the `re_nsub' field is available. All other fields are | |
329 | private to the regex routines. */ | |
330 | ||
fc4e96c8 | 331 | #ifndef RE_TRANSLATE_TYPE |
0b32bf0e | 332 | # define RE_TRANSLATE_TYPE char * |
6fffb26e RS |
333 | #endif |
334 | ||
b1fe2cfc KB |
335 | struct re_pattern_buffer |
336 | { | |
337 | /* [[[begin pattern_buffer]]] */ | |
338 | /* Space that holds the compiled pattern. It is declared as | |
339 | `unsigned char *' because its elements are | |
340 | sometimes used as array indexes. */ | |
341 | unsigned char *buffer; | |
342 | ||
343 | /* Number of bytes to which `buffer' points. */ | |
98222d27 | 344 | size_t allocated; |
b1fe2cfc KB |
345 | |
346 | /* Number of bytes actually used in `buffer'. */ | |
98222d27 | 347 | size_t used; |
b1fe2cfc KB |
348 | |
349 | /* Syntax setting with which the pattern was compiled. */ | |
350 | reg_syntax_t syntax; | |
351 | ||
352 | /* Pointer to a fastmap, if any, otherwise zero. re_search uses | |
353 | the fastmap, if there is one, to skip over impossible | |
354 | starting points for matches. */ | |
355 | char *fastmap; | |
356 | ||
357 | /* Either a translate table to apply to all characters before | |
358 | comparing them, or zero for no translation. The translation | |
359 | is applied to a pattern when it is compiled and to a string | |
360 | when it is matched. */ | |
6fffb26e | 361 | RE_TRANSLATE_TYPE translate; |
b1fe2cfc KB |
362 | |
363 | /* Number of subexpressions found by the compiler. */ | |
364 | size_t re_nsub; | |
365 | ||
366 | /* Zero if this pattern cannot match the empty string, one else. | |
367 | Well, in truth it's used only in `re_search_2', to see | |
368 | whether or not we should use the fastmap, so we don't set | |
dd52d970 | 369 | this absolutely perfectly; see `re_compile_fastmap'. */ |
b1fe2cfc KB |
370 | unsigned can_be_null : 1; |
371 | ||
372 | /* If REGS_UNALLOCATED, allocate space in the `regs' structure | |
373 | for `max (RE_NREGS, re_nsub + 1)' groups. | |
374 | If REGS_REALLOCATE, reallocate space if necessary. | |
375 | If REGS_FIXED, use what's there. */ | |
376 | #define REGS_UNALLOCATED 0 | |
377 | #define REGS_REALLOCATE 1 | |
378 | #define REGS_FIXED 2 | |
379 | unsigned regs_allocated : 2; | |
380 | ||
9114e279 KB |
381 | /* Set to zero when `regex_compile' compiles a pattern; set to one |
382 | by `re_compile_fastmap' if it updates the fastmap. */ | |
b1fe2cfc KB |
383 | unsigned fastmap_accurate : 1; |
384 | ||
9114e279 KB |
385 | /* If set, `re_match_2' does not return information about |
386 | subexpressions. */ | |
b1fe2cfc KB |
387 | unsigned no_sub : 1; |
388 | ||
389 | /* If set, a beginning-of-line anchor doesn't match at the | |
0b32bf0e | 390 | beginning of the string. */ |
b1fe2cfc KB |
391 | unsigned not_bol : 1; |
392 | ||
393 | /* Similarly for an end-of-line anchor. */ | |
394 | unsigned not_eol : 1; | |
395 | ||
1351dd3b SM |
396 | /* If true, the compilation of the pattern had to look up the syntax table, |
397 | so the compiled pattern is only valid for the current syntax table. */ | |
398 | unsigned used_syntax : 1; | |
399 | ||
2d1675e4 | 400 | #ifdef emacs |
66f089b2 | 401 | /* If true, multi-byte form in the regexp pattern should be |
0f4b87bf | 402 | recognized as a multibyte character. */ |
766ff627 | 403 | unsigned multibyte : 1; |
66f089b2 KH |
404 | |
405 | /* If true, multi-byte form in the target of match should be | |
406 | recognized as a multibyte character. */ | |
407 | unsigned target_multibyte : 1; | |
0f4b87bf KH |
408 | |
409 | /* Charset of unibyte characters at compiling time. */ | |
410 | int charset_unibyte; | |
2d1675e4 | 411 | #endif |
766ff627 | 412 | |
b1fe2cfc KB |
413 | /* [[[end pattern_buffer]]] */ |
414 | }; | |
415 | ||
416 | typedef struct re_pattern_buffer regex_t; | |
b1fe2cfc KB |
417 | \f |
418 | /* Type for byte offsets within the string. POSIX mandates this. */ | |
419 | typedef int regoff_t; | |
420 | ||
421 | ||
422 | /* This is the structure we store register match data in. See | |
423 | regex.texinfo for a full description of what registers match. */ | |
424 | struct re_registers | |
425 | { | |
426 | unsigned num_regs; | |
427 | regoff_t *start; | |
428 | regoff_t *end; | |
429 | }; | |
430 | ||
431 | ||
432 | /* If `regs_allocated' is REGS_UNALLOCATED in the pattern buffer, | |
433 | `re_match_2' returns information about at least this many registers | |
434 | the first time a `regs' structure is passed. */ | |
435 | #ifndef RE_NREGS | |
0b32bf0e | 436 | # define RE_NREGS 30 |
b1fe2cfc KB |
437 | #endif |
438 | ||
439 | ||
440 | /* POSIX specification for registers. Aside from the different names than | |
441 | `re_registers', POSIX uses an array of structures, instead of a | |
442 | structure of arrays. */ | |
443 | typedef struct | |
444 | { | |
445 | regoff_t rm_so; /* Byte offset from string's start to substring's start. */ | |
446 | regoff_t rm_eo; /* Byte offset from string's start to substring's end. */ | |
447 | } regmatch_t; | |
448 | \f | |
449 | /* Declarations for routines. */ | |
450 | ||
451 | /* To avoid duplicating every routine declaration -- once with a | |
452 | prototype (if we are ANSI), and once without (if we aren't) -- we | |
453 | use the following macro to declare argument types. This | |
454 | unfortunately clutters up the declarations a bit, but I think it's | |
65565c68 | 455 | worth it. */ |
b1fe2cfc | 456 | |
2a90cd71 | 457 | #if defined __STDC__ || defined PROTOTYPES |
0783dc9a | 458 | |
0b32bf0e | 459 | # define _RE_ARGS(args) args |
0783dc9a | 460 | |
2a90cd71 | 461 | #else /* not __STDC__ || PROTOTYPES */ |
0783dc9a | 462 | |
0b32bf0e | 463 | # define _RE_ARGS(args) () |
0783dc9a | 464 | |
2a90cd71 | 465 | #endif /* not __STDC__ || PROTOTYPES */ |
b1fe2cfc KB |
466 | |
467 | /* Sets the current default syntax to SYNTAX, and return the old syntax. | |
468 | You can also simply assign to the `re_syntax_options' variable. */ | |
469 | extern reg_syntax_t re_set_syntax _RE_ARGS ((reg_syntax_t syntax)); | |
470 | ||
471 | /* Compile the regular expression PATTERN, with length LENGTH | |
472 | and syntax given by the global `re_syntax_options', into the buffer | |
473 | BUFFER. Return NULL if successful, and an error string if not. */ | |
474 | extern const char *re_compile_pattern | |
0b32bf0e | 475 | _RE_ARGS ((const char *pattern, size_t length, |
b1fe2cfc KB |
476 | struct re_pattern_buffer *buffer)); |
477 | ||
478 | ||
479 | /* Compile a fastmap for the compiled pattern in BUFFER; used to | |
480 | accelerate searches. Return 0 if successful and -2 if was an | |
481 | internal error. */ | |
482 | extern int re_compile_fastmap _RE_ARGS ((struct re_pattern_buffer *buffer)); | |
483 | ||
484 | ||
485 | /* Search in the string STRING (with length LENGTH) for the pattern | |
486 | compiled into BUFFER. Start searching at position START, for RANGE | |
487 | characters. Return the starting position of the match, -1 for no | |
488 | match, or -2 for an internal error. Also return register | |
489 | information in REGS (if REGS and BUFFER->no_sub are nonzero). */ | |
490 | extern int re_search | |
491 | _RE_ARGS ((struct re_pattern_buffer *buffer, const char *string, | |
492 | int length, int start, int range, struct re_registers *regs)); | |
493 | ||
494 | ||
495 | /* Like `re_search', but search in the concatenation of STRING1 and | |
496 | STRING2. Also, stop searching at index START + STOP. */ | |
497 | extern int re_search_2 | |
498 | _RE_ARGS ((struct re_pattern_buffer *buffer, const char *string1, | |
499 | int length1, const char *string2, int length2, | |
500 | int start, int range, struct re_registers *regs, int stop)); | |
501 | ||
502 | ||
503 | /* Like `re_search', but return how many characters in STRING the regexp | |
504 | in BUFFER matched, starting at position START. */ | |
505 | extern int re_match | |
506 | _RE_ARGS ((struct re_pattern_buffer *buffer, const char *string, | |
507 | int length, int start, struct re_registers *regs)); | |
508 | ||
509 | ||
510 | /* Relates to `re_match' as `re_search_2' relates to `re_search'. */ | |
0b32bf0e | 511 | extern int re_match_2 |
b1fe2cfc KB |
512 | _RE_ARGS ((struct re_pattern_buffer *buffer, const char *string1, |
513 | int length1, const char *string2, int length2, | |
514 | int start, struct re_registers *regs, int stop)); | |
515 | ||
516 | ||
517 | /* Set REGS to hold NUM_REGS registers, storing them in STARTS and | |
518 | ENDS. Subsequent matches using BUFFER and REGS will use this memory | |
519 | for recording register information. STARTS and ENDS must be | |
520 | allocated with malloc, and must each be at least `NUM_REGS * sizeof | |
521 | (regoff_t)' bytes long. | |
522 | ||
523 | If NUM_REGS == 0, then subsequent matches should allocate their own | |
524 | register data. | |
525 | ||
526 | Unless this function is called, the first search or match using | |
527 | PATTERN_BUFFER will allocate its own register data, without | |
528 | freeing the old data. */ | |
529 | extern void re_set_registers | |
530 | _RE_ARGS ((struct re_pattern_buffer *buffer, struct re_registers *regs, | |
531 | unsigned num_regs, regoff_t *starts, regoff_t *ends)); | |
532 | ||
0b32bf0e SM |
533 | #if defined _REGEX_RE_COMP || defined _LIBC |
534 | # ifndef _CRAY | |
b1fe2cfc KB |
535 | /* 4.2 bsd compatibility. */ |
536 | extern char *re_comp _RE_ARGS ((const char *)); | |
537 | extern int re_exec _RE_ARGS ((const char *)); | |
0b32bf0e SM |
538 | # endif |
539 | #endif | |
540 | ||
541 | /* GCC 2.95 and later have "__restrict"; C99 compilers have | |
542 | "restrict", and "configure" may have defined "restrict". */ | |
543 | #ifndef __restrict | |
544 | # if ! (2 < __GNUC__ || (2 == __GNUC__ && 95 <= __GNUC_MINOR__)) | |
545 | # if defined restrict || 199901L <= __STDC_VERSION__ | |
546 | # define __restrict restrict | |
547 | # else | |
548 | # define __restrict | |
549 | # endif | |
550 | # endif | |
273c0c50 | 551 | #endif |
b90e26e1 | 552 | /* For now conditionally define __restrict_arr to expand to nothing. |
0b32bf0e SM |
553 | Ideally we would have a test for the compiler which allows defining |
554 | it to restrict. */ | |
27847466 | 555 | #ifndef __restrict_arr |
b90e26e1 | 556 | # define __restrict_arr |
27847466 | 557 | #endif |
b1fe2cfc KB |
558 | |
559 | /* POSIX compatibility. */ | |
0b32bf0e SM |
560 | extern int regcomp _RE_ARGS ((regex_t *__restrict __preg, |
561 | const char *__restrict __pattern, | |
562 | int __cflags)); | |
563 | ||
564 | extern int regexec _RE_ARGS ((const regex_t *__restrict __preg, | |
565 | const char *__restrict __string, size_t __nmatch, | |
566 | regmatch_t __pmatch[__restrict_arr], | |
567 | int __eflags)); | |
568 | ||
569 | extern size_t regerror _RE_ARGS ((int __errcode, const regex_t *__preg, | |
570 | char *__errbuf, size_t __errbuf_size)); | |
571 | ||
572 | extern void regfree _RE_ARGS ((regex_t *__preg)); | |
573 | ||
574 | ||
575 | #ifdef __cplusplus | |
576 | } | |
577 | #endif /* C++ */ | |
578 | ||
c13ecd8f RS |
579 | /* For platform which support the ISO C amendement 1 functionality we |
580 | support user defined character classes. */ | |
581 | #if WIDE_CHAR_SUPPORT | |
582 | /* Solaris 2.5 has a bug: <wchar.h> must be included before <wctype.h>. */ | |
583 | # include <wchar.h> | |
584 | # include <wctype.h> | |
585 | #endif | |
586 | ||
587 | #if WIDE_CHAR_SUPPORT | |
588 | /* The GNU C library provides support for user-defined character classes | |
589 | and the functions from ISO C amendement 1. */ | |
590 | # ifdef CHARCLASS_NAME_MAX | |
591 | # define CHAR_CLASS_MAX_LENGTH CHARCLASS_NAME_MAX | |
592 | # else | |
593 | /* This shouldn't happen but some implementation might still have this | |
594 | problem. Use a reasonable default value. */ | |
595 | # define CHAR_CLASS_MAX_LENGTH 256 | |
596 | # endif | |
597 | typedef wctype_t re_wctype_t; | |
598 | typedef wchar_t re_wchar_t; | |
599 | # define re_wctype wctype | |
600 | # define re_iswctype iswctype | |
601 | # define re_wctype_to_bit(cc) 0 | |
602 | #else | |
603 | # define CHAR_CLASS_MAX_LENGTH 9 /* Namely, `multibyte'. */ | |
604 | # define btowc(c) c | |
605 | ||
606 | /* Character classes. */ | |
607 | typedef enum { RECC_ERROR = 0, | |
608 | RECC_ALNUM, RECC_ALPHA, RECC_WORD, | |
609 | RECC_GRAPH, RECC_PRINT, | |
610 | RECC_LOWER, RECC_UPPER, | |
611 | RECC_PUNCT, RECC_CNTRL, | |
612 | RECC_DIGIT, RECC_XDIGIT, | |
613 | RECC_BLANK, RECC_SPACE, | |
614 | RECC_MULTIBYTE, RECC_NONASCII, | |
615 | RECC_ASCII, RECC_UNIBYTE | |
616 | } re_wctype_t; | |
617 | ||
fa8459a3 DN |
618 | extern char re_iswctype (int ch, re_wctype_t cc); |
619 | extern re_wctype_t re_wctype (const unsigned char* str); | |
620 | ||
c13ecd8f RS |
621 | typedef int re_wchar_t; |
622 | ||
4e2f601f | 623 | extern void re_set_whitespace_regexp (const char *regexp); |
fa8459a3 | 624 | |
c13ecd8f RS |
625 | #endif /* not WIDE_CHAR_SUPPORT */ |
626 | ||
0b32bf0e | 627 | #endif /* regex.h */ |
b1fe2cfc | 628 | \f |
839966f3 KH |
629 | /* arch-tag: bda6e3ec-3c02-4237-a55a-01ad2e120083 |
630 | (do not change this comment) */ |