]>
Commit | Line | Data |
---|---|---|
01d90158 CF |
1 | /* Definitions for data structures and routines for the regular |
2 | expression library. | |
3 | Copyright (C) 1985,1989-93,1995-98,2000,2001,2002,2003,2005,2006 | |
4 | Free Software Foundation, Inc. | |
5 | This file is part of the GNU C Library. | |
6 | ||
7 | The GNU C Library is free software; you can redistribute it and/or | |
8 | modify it under the terms of the GNU Lesser General Public | |
9 | License as published by the Free Software Foundation; either | |
10 | version 2.1 of the License, or (at your option) any later version. | |
11 | ||
12 | The GNU C Library is distributed in the hope that it will be useful, | |
13 | but WITHOUT ANY WARRANTY; without even the implied warranty of | |
14 | MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU | |
15 | Lesser General Public License for more details. | |
16 | ||
17 | You should have received a copy of the GNU Lesser General Public | |
18 | License along with the GNU C Library; if not, write to the Free | |
19 | Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA | |
20 | 02111-1307 USA. */ | |
21 | ||
22 | #ifndef _REGEX_H | |
23 | #define _REGEX_H 1 | |
24 | ||
25 | #include <sys/types.h> | |
26 | ||
27 | #define __DLL_IMPORT__ | |
28 | #define __DLL_EXPORT__ | |
29 | ||
30 | #if (defined __WIN32__) || (defined _WIN32) | |
31 | # ifdef BUILD_REGEX_DLL | |
32 | # define REGEX_DLL_IMPEXP __DLL_EXPORT__ | |
33 | # elif defined(REGEX_STATIC) | |
34 | # define REGEX_DLL_IMPEXP | |
35 | # elif defined (USE_REGEX_DLL) | |
36 | # define REGEX_DLL_IMPEXP __DLL_IMPORT__ | |
37 | # elif defined (USE_REGEX_STATIC) | |
38 | # define REGEX_DLL_IMPEXP | |
39 | # else /* assume USE_REGEX_DLL */ | |
40 | # define REGEX_DLL_IMPEXP __DLL_IMPORT__ | |
41 | # endif | |
42 | #else /* __WIN32__ */ | |
43 | # define REGEX_DLL_IMPEXP | |
44 | #endif | |
45 | ||
46 | /* Allow the use in C++ code. */ | |
47 | #ifdef __cplusplus | |
48 | extern "C" { | |
49 | #endif | |
50 | ||
51 | /* The following two types have to be signed and unsigned integer type | |
52 | wide enough to hold a value of a pointer. For most ANSI compilers | |
53 | ptrdiff_t and size_t should be likely OK. Still size of these two | |
54 | types is 2 for Microsoft C. Ugh... */ | |
55 | typedef long int s_reg_t; | |
56 | typedef unsigned long int active_reg_t; | |
57 | ||
58 | /* The following bits are used to determine the regexp syntax we | |
59 | recognize. The set/not-set meanings are chosen so that Emacs syntax | |
60 | remains the value 0. The bits are given in alphabetical order, and | |
61 | the definitions shifted by one from the previous bit; thus, when we | |
62 | add or remove a bit, only one other definition need change. */ | |
63 | typedef unsigned long int reg_syntax_t; | |
64 | ||
65 | /* If this bit is not set, then \ inside a bracket expression is literal. | |
66 | If set, then such a \ quotes the following character. */ | |
67 | #define RE_BACKSLASH_ESCAPE_IN_LISTS ((unsigned long int) 1) | |
68 | ||
69 | /* If this bit is not set, then + and ? are operators, and \+ and \? are | |
70 | literals. | |
71 | If set, then \+ and \? are operators and + and ? are literals. */ | |
72 | #define RE_BK_PLUS_QM (RE_BACKSLASH_ESCAPE_IN_LISTS << 1) | |
73 | ||
74 | /* If this bit is set, then character classes are supported. They are: | |
75 | [:alpha:], [:upper:], [:lower:], [:digit:], [:alnum:], [:xdigit:], | |
76 | [:space:], [:print:], [:punct:], [:graph:], and [:cntrl:]. | |
77 | If not set, then character classes are not supported. */ | |
78 | #define RE_CHAR_CLASSES (RE_BK_PLUS_QM << 1) | |
79 | ||
80 | /* If this bit is set, then ^ and $ are always anchors (outside bracket | |
81 | expressions, of course). | |
82 | If this bit is not set, then it depends: | |
83 | ^ is an anchor if it is at the beginning of a regular | |
84 | expression or after an open-group or an alternation operator; | |
85 | $ is an anchor if it is at the end of a regular expression, or | |
86 | before a close-group or an alternation operator. | |
87 | ||
88 | This bit could be (re)combined with RE_CONTEXT_INDEP_OPS, because | |
89 | POSIX draft 11.2 says that * etc. in leading positions is undefined. | |
90 | We already implemented a previous draft which made those constructs | |
91 | invalid, though, so we haven't changed the code back. */ | |
92 | #define RE_CONTEXT_INDEP_ANCHORS (RE_CHAR_CLASSES << 1) | |
93 | ||
94 | /* If this bit is set, then special characters are always special | |
95 | regardless of where they are in the pattern. | |
96 | If this bit is not set, then special characters are special only in | |
97 | some contexts; otherwise they are ordinary. Specifically, | |
98 | * + ? and intervals are only special when not after the beginning, | |
99 | open-group, or alternation operator. */ | |
100 | #define RE_CONTEXT_INDEP_OPS (RE_CONTEXT_INDEP_ANCHORS << 1) | |
101 | ||
102 | /* If this bit is set, then *, +, ?, and { cannot be first in an re or | |
103 | immediately after an alternation or begin-group operator. */ | |
104 | #define RE_CONTEXT_INVALID_OPS (RE_CONTEXT_INDEP_OPS << 1) | |
105 | ||
106 | /* If this bit is set, then . matches newline. | |
107 | If not set, then it doesn't. */ | |
108 | #define RE_DOT_NEWLINE (RE_CONTEXT_INVALID_OPS << 1) | |
109 | ||
110 | /* If this bit is set, then . doesn't match NUL. | |
111 | If not set, then it does. */ | |
112 | #define RE_DOT_NOT_NULL (RE_DOT_NEWLINE << 1) | |
113 | ||
114 | /* If this bit is set, nonmatching lists [^...] do not match newline. | |
115 | If not set, they do. */ | |
116 | #define RE_HAT_LISTS_NOT_NEWLINE (RE_DOT_NOT_NULL << 1) | |
117 | ||
118 | /* If this bit is set, either \{...\} or {...} defines an | |
119 | interval, depending on RE_NO_BK_BRACES. | |
120 | If not set, \{, \}, {, and } are literals. */ | |
121 | #define RE_INTERVALS (RE_HAT_LISTS_NOT_NEWLINE << 1) | |
122 | ||
123 | /* If this bit is set, +, ? and | aren't recognized as operators. | |
124 | If not set, they are. */ | |
125 | #define RE_LIMITED_OPS (RE_INTERVALS << 1) | |
126 | ||
127 | /* If this bit is set, newline is an alternation operator. | |
128 | If not set, newline is literal. */ | |
129 | #define RE_NEWLINE_ALT (RE_LIMITED_OPS << 1) | |
130 | ||
131 | /* If this bit is set, then `{...}' defines an interval, and \{ and \} | |
132 | are literals. | |
133 | If not set, then `\{...\}' defines an interval. */ | |
134 | #define RE_NO_BK_BRACES (RE_NEWLINE_ALT << 1) | |
135 | ||
136 | /* If this bit is set, (...) defines a group, and \( and \) are literals. | |
137 | If not set, \(...\) defines a group, and ( and ) are literals. */ | |
138 | #define RE_NO_BK_PARENS (RE_NO_BK_BRACES << 1) | |
139 | ||
140 | /* If this bit is set, then \<digit> matches <digit>. | |
141 | If not set, then \<digit> is a back-reference. */ | |
142 | #define RE_NO_BK_REFS (RE_NO_BK_PARENS << 1) | |
143 | ||
144 | /* If this bit is set, then | is an alternation operator, and \| is literal. | |
145 | If not set, then \| is an alternation operator, and | is literal. */ | |
146 | #define RE_NO_BK_VBAR (RE_NO_BK_REFS << 1) | |
147 | ||
148 | /* If this bit is set, then an ending range point collating higher | |
149 | than the starting range point, as in [z-a], is invalid. | |
150 | If not set, then when ending range point collates higher than the | |
151 | starting range point, the range is ignored. */ | |
152 | #define RE_NO_EMPTY_RANGES (RE_NO_BK_VBAR << 1) | |
153 | ||
154 | /* If this bit is set, then an unmatched ) is ordinary. | |
155 | If not set, then an unmatched ) is invalid. */ | |
156 | #define RE_UNMATCHED_RIGHT_PAREN_ORD (RE_NO_EMPTY_RANGES << 1) | |
157 | ||
158 | /* If this bit is set, succeed as soon as we match the whole pattern, | |
159 | without further backtracking. */ | |
160 | #define RE_NO_POSIX_BACKTRACKING (RE_UNMATCHED_RIGHT_PAREN_ORD << 1) | |
161 | ||
162 | /* If this bit is set, do not process the GNU regex operators. | |
163 | If not set, then the GNU regex operators are recognized. */ | |
164 | #define RE_NO_GNU_OPS (RE_NO_POSIX_BACKTRACKING << 1) | |
165 | ||
166 | /* If this bit is set, turn on internal regex debugging. | |
167 | If not set, and debugging was on, turn it off. | |
168 | This only works if regex.c is compiled -DDEBUG. | |
169 | We define this bit always, so that all that's needed to turn on | |
170 | debugging is to recompile regex.c; the calling code can always have | |
171 | this bit set, and it won't affect anything in the normal case. */ | |
172 | #define RE_DEBUG (RE_NO_GNU_OPS << 1) | |
173 | ||
174 | /* If this bit is set, a syntactically invalid interval is treated as | |
175 | a string of ordinary characters. For example, the ERE 'a{1' is | |
176 | treated as 'a\{1'. */ | |
177 | #define RE_INVALID_INTERVAL_ORD (RE_DEBUG << 1) | |
178 | ||
179 | /* If this bit is set, then ignore case when matching. | |
180 | If not set, then case is significant. */ | |
181 | #define RE_ICASE (RE_INVALID_INTERVAL_ORD << 1) | |
182 | ||
183 | /* This bit is used internally like RE_CONTEXT_INDEP_ANCHORS but only | |
184 | for ^, because it is difficult to scan the regex backwards to find | |
185 | whether ^ should be special. */ | |
186 | #define RE_CARET_ANCHORS_HERE (RE_ICASE << 1) | |
187 | ||
188 | /* If this bit is set, then \{ cannot be first in an bre or | |
189 | immediately after an alternation or begin-group operator. */ | |
190 | #define RE_CONTEXT_INVALID_DUP (RE_CARET_ANCHORS_HERE << 1) | |
191 | ||
192 | /* If this bit is set, then no_sub will be set to 1 during | |
193 | re_compile_pattern. */ | |
194 | #define RE_NO_SUB (RE_CONTEXT_INVALID_DUP << 1) | |
195 | ||
196 | /* This global variable defines the particular regexp syntax to use (for | |
197 | some interfaces). When a regexp is compiled, the syntax used is | |
198 | stored in the pattern buffer, so changing this does not affect | |
199 | already-compiled regexps. */ | |
200 | REGEX_DLL_IMPEXP reg_syntax_t re_syntax_options; | |
201 | \f | |
202 | /* Define combinations of the above bits for the standard possibilities. | |
203 | (The [[[ comments delimit what gets put into the Texinfo file, so | |
204 | don't delete them!) */ | |
205 | /* [[[begin syntaxes]]] */ | |
206 | #define RE_SYNTAX_EMACS 0 | |
207 | ||
208 | #define RE_SYNTAX_AWK \ | |
209 | (RE_BACKSLASH_ESCAPE_IN_LISTS | RE_DOT_NOT_NULL \ | |
210 | | RE_NO_BK_PARENS | RE_NO_BK_REFS \ | |
211 | | RE_NO_BK_VBAR | RE_NO_EMPTY_RANGES \ | |
212 | | RE_DOT_NEWLINE | RE_CONTEXT_INDEP_ANCHORS \ | |
213 | | RE_UNMATCHED_RIGHT_PAREN_ORD | RE_NO_GNU_OPS) | |
214 | ||
215 | #define RE_SYNTAX_GNU_AWK \ | |
216 | ((RE_SYNTAX_POSIX_EXTENDED | RE_BACKSLASH_ESCAPE_IN_LISTS | RE_DEBUG) \ | |
217 | & ~(RE_DOT_NOT_NULL | RE_INTERVALS | RE_CONTEXT_INDEP_OPS \ | |
218 | | RE_CONTEXT_INVALID_OPS )) | |
219 | ||
220 | #define RE_SYNTAX_POSIX_AWK \ | |
221 | (RE_SYNTAX_POSIX_EXTENDED | RE_BACKSLASH_ESCAPE_IN_LISTS \ | |
222 | | RE_INTERVALS | RE_NO_GNU_OPS) | |
223 | ||
224 | #define RE_SYNTAX_GREP \ | |
225 | (RE_BK_PLUS_QM | RE_CHAR_CLASSES \ | |
226 | | RE_HAT_LISTS_NOT_NEWLINE | RE_INTERVALS \ | |
227 | | RE_NEWLINE_ALT) | |
228 | ||
229 | #define RE_SYNTAX_EGREP \ | |
230 | (RE_CHAR_CLASSES | RE_CONTEXT_INDEP_ANCHORS \ | |
231 | | RE_CONTEXT_INDEP_OPS | RE_HAT_LISTS_NOT_NEWLINE \ | |
232 | | RE_NEWLINE_ALT | RE_NO_BK_PARENS \ | |
233 | | RE_NO_BK_VBAR) | |
234 | ||
235 | #define RE_SYNTAX_POSIX_EGREP \ | |
236 | (RE_SYNTAX_EGREP | RE_INTERVALS | RE_NO_BK_BRACES \ | |
237 | | RE_INVALID_INTERVAL_ORD) | |
238 | ||
239 | /* P1003.2/D11.2, section 4.20.7.1, lines 5078ff. */ | |
240 | #define RE_SYNTAX_ED RE_SYNTAX_POSIX_BASIC | |
241 | ||
242 | #define RE_SYNTAX_SED RE_SYNTAX_POSIX_BASIC | |
243 | ||
244 | /* Syntax bits common to both basic and extended POSIX regex syntax. */ | |
245 | #define _RE_SYNTAX_POSIX_COMMON \ | |
246 | (RE_CHAR_CLASSES | RE_DOT_NEWLINE | RE_DOT_NOT_NULL \ | |
247 | | RE_INTERVALS | RE_NO_EMPTY_RANGES) | |
248 | ||
249 | #define RE_SYNTAX_POSIX_BASIC \ | |
250 | (_RE_SYNTAX_POSIX_COMMON | RE_BK_PLUS_QM | RE_CONTEXT_INVALID_DUP) | |
251 | ||
252 | /* Differs from ..._POSIX_BASIC only in that RE_BK_PLUS_QM becomes | |
253 | RE_LIMITED_OPS, i.e., \? \+ \| are not recognized. Actually, this | |
254 | isn't minimal, since other operators, such as \`, aren't disabled. */ | |
255 | #define RE_SYNTAX_POSIX_MINIMAL_BASIC \ | |
256 | (_RE_SYNTAX_POSIX_COMMON | RE_LIMITED_OPS) | |
257 | ||
258 | #define RE_SYNTAX_POSIX_EXTENDED \ | |
259 | (_RE_SYNTAX_POSIX_COMMON | RE_CONTEXT_INDEP_ANCHORS \ | |
260 | | RE_CONTEXT_INDEP_OPS | RE_NO_BK_BRACES \ | |
261 | | RE_NO_BK_PARENS | RE_NO_BK_VBAR \ | |
262 | | RE_CONTEXT_INVALID_OPS | RE_UNMATCHED_RIGHT_PAREN_ORD) | |
263 | ||
264 | /* Differs from ..._POSIX_EXTENDED in that RE_CONTEXT_INDEP_OPS is | |
265 | removed and RE_NO_BK_REFS is added. */ | |
266 | #define RE_SYNTAX_POSIX_MINIMAL_EXTENDED \ | |
267 | (_RE_SYNTAX_POSIX_COMMON | RE_CONTEXT_INDEP_ANCHORS \ | |
268 | | RE_CONTEXT_INVALID_OPS | RE_NO_BK_BRACES \ | |
269 | | RE_NO_BK_PARENS | RE_NO_BK_REFS \ | |
270 | | RE_NO_BK_VBAR | RE_UNMATCHED_RIGHT_PAREN_ORD) | |
271 | /* [[[end syntaxes]]] */ | |
272 | \f | |
273 | /* Maximum number of duplicates an interval can allow. Some systems | |
274 | (erroneously) define this in other header files, but we want our | |
275 | value, so remove any previous define. */ | |
276 | #ifdef RE_DUP_MAX | |
277 | # undef RE_DUP_MAX | |
278 | #endif | |
279 | /* If sizeof(int) == 2, then ((1 << 15) - 1) overflows. */ | |
280 | #define RE_DUP_MAX (0x7fff) | |
281 | ||
282 | ||
283 | /* POSIX `cflags' bits (i.e., information for `regcomp'). */ | |
284 | ||
285 | /* If this bit is set, then use extended regular expression syntax. | |
286 | If not set, then use basic regular expression syntax. */ | |
287 | #define REG_EXTENDED 1 | |
288 | ||
289 | /* If this bit is set, then ignore case when matching. | |
290 | If not set, then case is significant. */ | |
291 | #define REG_ICASE (REG_EXTENDED << 1) | |
292 | ||
293 | /* If this bit is set, then anchors do not match at newline | |
294 | characters in the string. | |
295 | If not set, then anchors do match at newlines. */ | |
296 | #define REG_NEWLINE (REG_ICASE << 1) | |
297 | ||
298 | /* If this bit is set, then report only success or fail in regexec. | |
299 | If not set, then returns differ between not matching and errors. */ | |
300 | #define REG_NOSUB (REG_NEWLINE << 1) | |
301 | ||
302 | ||
303 | /* POSIX `eflags' bits (i.e., information for regexec). */ | |
304 | ||
305 | /* If this bit is set, then the beginning-of-line operator doesn't match | |
306 | the beginning of the string (presumably because it's not the | |
307 | beginning of a line). | |
308 | If not set, then the beginning-of-line operator does match the | |
309 | beginning of the string. */ | |
310 | #define REG_NOTBOL 1 | |
311 | ||
312 | /* Like REG_NOTBOL, except for the end-of-line. */ | |
313 | #define REG_NOTEOL (1 << 1) | |
314 | ||
315 | /* Use PMATCH[0] to delimit the start and end of the search in the | |
316 | buffer. */ | |
317 | #define REG_STARTEND (1 << 2) | |
318 | ||
319 | ||
320 | /* If any error codes are removed, changed, or added, update the | |
321 | `re_error_msg' table in regex.c. */ | |
322 | typedef enum | |
323 | { | |
324 | #ifdef _XOPEN_SOURCE | |
325 | REG_ENOSYS = -1, /* This will never happen for this implementation. */ | |
326 | #endif | |
327 | ||
328 | REG_NOERROR = 0, /* Success. */ | |
329 | REG_NOMATCH, /* Didn't find a match (for regexec). */ | |
330 | ||
331 | /* POSIX regcomp return error codes. (In the order listed in the | |
332 | standard.) */ | |
333 | REG_BADPAT, /* Invalid pattern. */ | |
334 | REG_ECOLLATE, /* Inalid collating element. */ | |
335 | REG_ECTYPE, /* Invalid character class name. */ | |
336 | REG_EESCAPE, /* Trailing backslash. */ | |
337 | REG_ESUBREG, /* Invalid back reference. */ | |
338 | REG_EBRACK, /* Unmatched left bracket. */ | |
339 | REG_EPAREN, /* Parenthesis imbalance. */ | |
340 | REG_EBRACE, /* Unmatched \{. */ | |
341 | REG_BADBR, /* Invalid contents of \{\}. */ | |
342 | REG_ERANGE, /* Invalid range end. */ | |
343 | REG_ESPACE, /* Ran out of memory. */ | |
344 | REG_BADRPT, /* No preceding re for repetition op. */ | |
345 | ||
346 | /* Error codes we've added. */ | |
347 | REG_EEND, /* Premature end. */ | |
348 | REG_ESIZE, /* Compiled pattern bigger than 2^16 bytes. */ | |
349 | REG_ERPAREN /* Unmatched ) or \); not returned from regcomp. */ | |
350 | } reg_errcode_t; | |
351 | \f | |
352 | /* This data structure represents a compiled pattern. Before calling | |
353 | the pattern compiler, the fields `buffer', `allocated', `fastmap', | |
354 | `translate', and `no_sub' can be set. After the pattern has been | |
355 | compiled, the `re_nsub' field is available. All other fields are | |
356 | private to the regex routines. */ | |
357 | ||
358 | #ifndef RE_TRANSLATE_TYPE | |
359 | # define RE_TRANSLATE_TYPE unsigned char * | |
360 | #endif | |
361 | ||
362 | struct re_pattern_buffer | |
363 | { | |
364 | /* Space that holds the compiled pattern. It is declared as | |
365 | `unsigned char *' because its elements are sometimes used as | |
366 | array indexes. */ | |
367 | unsigned char *buffer; | |
368 | ||
369 | /* Number of bytes to which `buffer' points. */ | |
370 | unsigned long int allocated; | |
371 | ||
372 | /* Number of bytes actually used in `buffer'. */ | |
373 | unsigned long int used; | |
374 | ||
375 | /* Syntax setting with which the pattern was compiled. */ | |
376 | reg_syntax_t syntax; | |
377 | ||
378 | /* Pointer to a fastmap, if any, otherwise zero. re_search uses the | |
379 | fastmap, if there is one, to skip over impossible starting points | |
380 | for matches. */ | |
381 | char *fastmap; | |
382 | ||
383 | /* Either a translate table to apply to all characters before | |
384 | comparing them, or zero for no translation. The translation is | |
385 | applied to a pattern when it is compiled and to a string when it | |
386 | is matched. */ | |
387 | RE_TRANSLATE_TYPE translate; | |
388 | ||
389 | /* Number of subexpressions found by the compiler. */ | |
390 | size_t re_nsub; | |
391 | ||
392 | /* Zero if this pattern cannot match the empty string, one else. | |
393 | Well, in truth it's used only in `re_search_2', to see whether or | |
394 | not we should use the fastmap, so we don't set this absolutely | |
395 | perfectly; see `re_compile_fastmap' (the `duplicate' case). */ | |
396 | unsigned can_be_null : 1; | |
397 | ||
398 | /* If REGS_UNALLOCATED, allocate space in the `regs' structure | |
399 | for `max (RE_NREGS, re_nsub + 1)' groups. | |
400 | If REGS_REALLOCATE, reallocate space if necessary. | |
401 | If REGS_FIXED, use what's there. */ | |
402 | #define REGS_UNALLOCATED 0 | |
403 | #define REGS_REALLOCATE 1 | |
404 | #define REGS_FIXED 2 | |
405 | unsigned regs_allocated : 2; | |
406 | ||
407 | /* Set to zero when `regex_compile' compiles a pattern; set to one | |
408 | by `re_compile_fastmap' if it updates the fastmap. */ | |
409 | unsigned fastmap_accurate : 1; | |
410 | ||
411 | /* If set, `re_match_2' does not return information about | |
412 | subexpressions. */ | |
413 | unsigned no_sub : 1; | |
414 | ||
415 | /* If set, a beginning-of-line anchor doesn't match at the beginning | |
416 | of the string. */ | |
417 | unsigned not_bol : 1; | |
418 | ||
419 | /* Similarly for an end-of-line anchor. */ | |
420 | unsigned not_eol : 1; | |
421 | ||
422 | /* If true, an anchor at a newline matches. */ | |
423 | unsigned newline_anchor : 1; | |
424 | }; | |
425 | ||
426 | typedef struct re_pattern_buffer regex_t; | |
427 | \f | |
428 | /* Type for byte offsets within the string. POSIX mandates this. */ | |
429 | typedef int regoff_t; | |
430 | ||
431 | ||
432 | /* This is the structure we store register match data in. See | |
433 | regex.texinfo for a full description of what registers match. */ | |
434 | struct re_registers | |
435 | { | |
436 | unsigned num_regs; | |
437 | regoff_t *start; | |
438 | regoff_t *end; | |
439 | }; | |
440 | ||
441 | ||
442 | /* If `regs_allocated' is REGS_UNALLOCATED in the pattern buffer, | |
443 | `re_match_2' returns information about at least this many registers | |
444 | the first time a `regs' structure is passed. */ | |
445 | #ifndef RE_NREGS | |
446 | # define RE_NREGS 30 | |
447 | #endif | |
448 | ||
449 | ||
450 | /* POSIX specification for registers. Aside from the different names than | |
451 | `re_registers', POSIX uses an array of structures, instead of a | |
452 | structure of arrays. */ | |
453 | typedef struct | |
454 | { | |
455 | regoff_t rm_so; /* Byte offset from string's start to substring's start. */ | |
456 | regoff_t rm_eo; /* Byte offset from string's start to substring's end. */ | |
457 | } regmatch_t; | |
458 | \f | |
459 | /* Declarations for routines. */ | |
460 | ||
461 | /* Sets the current default syntax to SYNTAX, and return the old syntax. | |
462 | You can also simply assign to the `re_syntax_options' variable. */ | |
463 | REGEX_DLL_IMPEXP reg_syntax_t re_set_syntax (reg_syntax_t __syntax); | |
464 | ||
465 | /* Compile the regular expression PATTERN, with length LENGTH | |
466 | and syntax given by the global `re_syntax_options', into the buffer | |
467 | BUFFER. Return NULL if successful, and an error string if not. */ | |
468 | REGEX_DLL_IMPEXP const char *re_compile_pattern (const char *__pattern, size_t __length, | |
469 | struct re_pattern_buffer *__buffer); | |
470 | ||
471 | ||
472 | /* Compile a fastmap for the compiled pattern in BUFFER; used to | |
473 | accelerate searches. Return 0 if successful and -2 if was an | |
474 | internal error. */ | |
475 | REGEX_DLL_IMPEXP int re_compile_fastmap (struct re_pattern_buffer *__buffer); | |
476 | ||
477 | ||
478 | /* Search in the string STRING (with length LENGTH) for the pattern | |
479 | compiled into BUFFER. Start searching at position START, for RANGE | |
480 | characters. Return the starting position of the match, -1 for no | |
481 | match, or -2 for an internal error. Also return register | |
482 | information in REGS (if REGS and BUFFER->no_sub are nonzero). */ | |
483 | REGEX_DLL_IMPEXP int re_search (struct re_pattern_buffer *__buffer, const char *__string, | |
484 | int __length, int __start, int __range, | |
485 | struct re_registers *__regs); | |
486 | ||
487 | ||
488 | /* Like `re_search', but search in the concatenation of STRING1 and | |
489 | STRING2. Also, stop searching at index START + STOP. */ | |
490 | REGEX_DLL_IMPEXP int re_search_2 (struct re_pattern_buffer *__buffer, | |
491 | const char *__string1, int __length1, | |
492 | const char *__string2, int __length2, int __start, | |
493 | int __range, struct re_registers *__regs, int __stop); | |
494 | ||
495 | ||
496 | /* Like `re_search', but return how many characters in STRING the regexp | |
497 | in BUFFER matched, starting at position START. */ | |
498 | REGEX_DLL_IMPEXP int re_match (struct re_pattern_buffer *__buffer, const char *__string, | |
499 | int __length, int __start, struct re_registers *__regs); | |
500 | ||
501 | ||
502 | /* Relates to `re_match' as `re_search_2' relates to `re_search'. */ | |
503 | REGEX_DLL_IMPEXP int re_match_2 (struct re_pattern_buffer *__buffer, | |
504 | const char *__string1, int __length1, | |
505 | const char *__string2, int __length2, int __start, | |
506 | struct re_registers *__regs, int __stop); | |
507 | ||
508 | ||
509 | /* Set REGS to hold NUM_REGS registers, storing them in STARTS and | |
510 | ENDS. Subsequent matches using BUFFER and REGS will use this memory | |
511 | for recording register information. STARTS and ENDS must be | |
512 | allocated with malloc, and must each be at least `NUM_REGS * sizeof | |
513 | (regoff_t)' bytes long. | |
514 | ||
515 | If NUM_REGS == 0, then subsequent matches should allocate their own | |
516 | register data. | |
517 | ||
518 | Unless this function is called, the first search or match using | |
519 | PATTERN_BUFFER will allocate its own register data, without | |
520 | freeing the old data. */ | |
521 | REGEX_DLL_IMPEXP void re_set_registers (struct re_pattern_buffer *__buffer, | |
522 | struct re_registers *__regs, | |
523 | unsigned int __num_regs, | |
524 | regoff_t *__starts, regoff_t *__ends); | |
525 | ||
526 | #if defined _REGEX_RE_COMP || defined _LIBC | |
527 | # ifndef _CRAY | |
528 | /* 4.2 bsd compatibility. */ | |
529 | REGEX_DLL_IMPEXP char *re_comp (const char *); | |
530 | REGEX_DLL_IMPEXP int re_exec (const char *); | |
531 | # endif | |
532 | #endif | |
533 | ||
534 | /* GCC 2.95 and later have "__restrict"; C99 compilers have | |
535 | "restrict", and "configure" may have defined "restrict". */ | |
536 | #ifndef __restrict | |
537 | # if ! (2 < __GNUC__ || (2 == __GNUC__ && 95 <= __GNUC_MINOR__)) | |
538 | # if defined restrict || 199901L <= __STDC_VERSION__ | |
539 | # define __restrict restrict | |
540 | # else | |
541 | # define __restrict | |
542 | # endif | |
543 | # endif | |
544 | #endif | |
545 | /* gcc 3.1 and up support the [restrict] syntax. */ | |
546 | #ifndef __restrict_arr | |
547 | # if (__GNUC__ > 3 || (__GNUC__ == 3 && __GNUC_MINOR__ >= 1)) \ | |
548 | && !defined __GNUG__ | |
549 | # define __restrict_arr __restrict | |
550 | # else | |
551 | # define __restrict_arr | |
552 | # endif | |
553 | #endif | |
554 | ||
555 | /* POSIX compatibility. */ | |
556 | REGEX_DLL_IMPEXP int regcomp (regex_t *__restrict __preg, | |
557 | const char *__restrict __pattern, | |
558 | int __cflags); | |
559 | ||
560 | REGEX_DLL_IMPEXP int regexec (const regex_t *__restrict __preg, | |
561 | const char *__restrict __string, size_t __nmatch, | |
562 | regmatch_t __pmatch[__restrict_arr], | |
563 | int __eflags); | |
564 | ||
565 | REGEX_DLL_IMPEXP size_t regerror (int __errcode, const regex_t *__restrict __preg, | |
566 | char *__restrict __errbuf, size_t __errbuf_size); | |
567 | ||
568 | REGEX_DLL_IMPEXP void regfree (regex_t *__preg); | |
569 | ||
570 | ||
571 | #ifdef __cplusplus | |
572 | } | |
573 | #endif /* C++ */ | |
574 | ||
575 | #endif /* regex.h */ |