Regular expressions implemented under Windows
[public/netxms.git] / include / netxms-regex.h
1 /* Definitions for data structures and routines for the regular
2 expression library, version 0.12.
3
4 Copyright (C) 1985, 89, 90, 91, 92, 93, 95 Free Software Foundation, Inc.
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
18 Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. */
19
20 #ifndef __REGEXP_LIBRARY_H__
21 #define __REGEXP_LIBRARY_H__
22
23 #ifdef __cplusplus
24 #define EXTERN_DECL extern "C"
25 #else
26 #define EXTERN_DECL extern
27 #endif
28
29 #include <nms_util.h>
30
31
32 /* POSIX says that <sys/types.h> must be included (by the caller) before
33 <regex.h>. */
34
35 #if !defined (_POSIX_C_SOURCE) && !defined (_POSIX_SOURCE) && defined (VMS)
36 /* VMS doesn't have `size_t' in <sys/types.h>, even though POSIX says it
37 should be there. */
38 #include <stddef.h>
39 #endif
40
41
42 /* The following bits are used to determine the regexp syntax we
43 recognize. The set/not-set meanings are chosen so that Emacs syntax
44 remains the value 0. The bits are given in alphabetical order, and
45 the definitions shifted by one from the previous bit; thus, when we
46 add or remove a bit, only one other definition need change. */
47 typedef unsigned reg_syntax_t;
48
49 /* If this bit is not set, then \ inside a bracket expression is literal.
50 If set, then such a \ quotes the following character. */
51 #define RE_BACKSLASH_ESCAPE_IN_LISTS (1)
52
53 /* If this bit is not set, then + and ? are operators, and \+ and \? are
54 literals.
55 If set, then \+ and \? are operators and + and ? are literals. */
56 #define RE_BK_PLUS_QM (RE_BACKSLASH_ESCAPE_IN_LISTS << 1)
57
58 /* If this bit is set, then character classes are supported. They are:
59 [:alpha:], [:upper:], [:lower:], [:digit:], [:alnum:], [:xdigit:],
60 [:space:], [:print:], [:punct:], [:graph:], and [:cntrl:].
61 If not set, then character classes are not supported. */
62 #define RE_CHAR_CLASSES (RE_BK_PLUS_QM << 1)
63
64 /* If this bit is set, then ^ and $ are always anchors (outside bracket
65 expressions, of course).
66 If this bit is not set, then it depends:
67 ^ is an anchor if it is at the beginning of a regular
68 expression or after an open-group or an alternation operator;
69 $ is an anchor if it is at the end of a regular expression, or
70 before a close-group or an alternation operator.
71
72 This bit could be (re)combined with RE_CONTEXT_INDEP_OPS, because
73 POSIX draft 11.2 says that * etc. in leading positions is undefined.
74 We already implemented a previous draft which made those constructs
75 invalid, though, so we haven't changed the code back. */
76 #define RE_CONTEXT_INDEP_ANCHORS (RE_CHAR_CLASSES << 1)
77
78 /* If this bit is set, then special characters are always special
79 regardless of where they are in the pattern.
80 If this bit is not set, then special characters are special only in
81 some contexts; otherwise they are ordinary. Specifically,
82 * + ? and intervals are only special when not after the beginning,
83 open-group, or alternation operator. */
84 #define RE_CONTEXT_INDEP_OPS (RE_CONTEXT_INDEP_ANCHORS << 1)
85
86 /* If this bit is set, then *, +, ?, and { cannot be first in an re or
87 immediately after an alternation or begin-group operator. */
88 #define RE_CONTEXT_INVALID_OPS (RE_CONTEXT_INDEP_OPS << 1)
89
90 /* If this bit is set, then . matches newline.
91 If not set, then it doesn't. */
92 #define RE_DOT_NEWLINE (RE_CONTEXT_INVALID_OPS << 1)
93
94 /* If this bit is set, then . doesn't match NUL.
95 If not set, then it does. */
96 #define RE_DOT_NOT_NULL (RE_DOT_NEWLINE << 1)
97
98 /* If this bit is set, nonmatching lists [^...] do not match newline.
99 If not set, they do. */
100 #define RE_HAT_LISTS_NOT_NEWLINE (RE_DOT_NOT_NULL << 1)
101
102 /* If this bit is set, either \{...\} or {...} defines an
103 interval, depending on RE_NO_BK_BRACES.
104 If not set, \{, \}, {, and } are literals. */
105 #define RE_INTERVALS (RE_HAT_LISTS_NOT_NEWLINE << 1)
106
107 /* If this bit is set, +, ? and | aren't recognized as operators.
108 If not set, they are. */
109 #define RE_LIMITED_OPS (RE_INTERVALS << 1)
110
111 /* If this bit is set, newline is an alternation operator.
112 If not set, newline is literal. */
113 #define RE_NEWLINE_ALT (RE_LIMITED_OPS << 1)
114
115 /* If this bit is set, then `{...}' defines an interval, and \{ and \}
116 are literals.
117 If not set, then `\{...\}' defines an interval. */
118 #define RE_NO_BK_BRACES (RE_NEWLINE_ALT << 1)
119
120 /* If this bit is set, (...) defines a group, and \( and \) are literals.
121 If not set, \(...\) defines a group, and ( and ) are literals. */
122 #define RE_NO_BK_PARENS (RE_NO_BK_BRACES << 1)
123
124 /* If this bit is set, then \<digit> matches <digit>.
125 If not set, then \<digit> is a back-reference. */
126 #define RE_NO_BK_REFS (RE_NO_BK_PARENS << 1)
127
128 /* If this bit is set, then | is an alternation operator, and \| is literal.
129 If not set, then \| is an alternation operator, and | is literal. */
130 #define RE_NO_BK_VBAR (RE_NO_BK_REFS << 1)
131
132 /* If this bit is set, then an ending range point collating higher
133 than the starting range point, as in [z-a], is invalid.
134 If not set, then when ending range point collates higher than the
135 starting range point, the range is ignored. */
136 #define RE_NO_EMPTY_RANGES (RE_NO_BK_VBAR << 1)
137
138 /* If this bit is set, then an unmatched ) is ordinary.
139 If not set, then an unmatched ) is invalid. */
140 #define RE_UNMATCHED_RIGHT_PAREN_ORD (RE_NO_EMPTY_RANGES << 1)
141
142 /* If this bit is set, succeed as soon as we match the whole pattern,
143 without further backtracking. */
144 #define RE_NO_POSIX_BACKTRACKING (RE_UNMATCHED_RIGHT_PAREN_ORD << 1)
145
146 /* This global variable defines the particular regexp syntax to use (for
147 some interfaces). When a regexp is compiled, the syntax used is
148 stored in the pattern buffer, so changing this does not affect
149 already-compiled regexps. */
150 EXTERN_DECL reg_syntax_t re_syntax_options;
151
152 /* Define combinations of the above bits for the standard possibilities.
153 (The [[[ comments delimit what gets put into the Texinfo file, so
154 don't delete them!) */
155 /* [[[begin syntaxes]]] */
156 #define RE_SYNTAX_EMACS 0
157
158 #define RE_SYNTAX_AWK \
159 (RE_BACKSLASH_ESCAPE_IN_LISTS | RE_DOT_NOT_NULL \
160 | RE_NO_BK_PARENS | RE_NO_BK_REFS \
161 | RE_NO_BK_VBAR | RE_NO_EMPTY_RANGES \
162 | RE_UNMATCHED_RIGHT_PAREN_ORD)
163
164 #define RE_SYNTAX_POSIX_AWK \
165 (RE_SYNTAX_POSIX_EXTENDED | RE_BACKSLASH_ESCAPE_IN_LISTS)
166
167 #define RE_SYNTAX_GREP \
168 (RE_BK_PLUS_QM | RE_CHAR_CLASSES \
169 | RE_HAT_LISTS_NOT_NEWLINE | RE_INTERVALS \
170 | RE_NEWLINE_ALT)
171
172 #define RE_SYNTAX_EGREP \
173 (RE_CHAR_CLASSES | RE_CONTEXT_INDEP_ANCHORS \
174 | RE_CONTEXT_INDEP_OPS | RE_HAT_LISTS_NOT_NEWLINE \
175 | RE_NEWLINE_ALT | RE_NO_BK_PARENS \
176 | RE_NO_BK_VBAR)
177
178 #define RE_SYNTAX_POSIX_EGREP \
179 (RE_SYNTAX_EGREP | RE_INTERVALS | RE_NO_BK_BRACES)
180
181 /* P1003.2/D11.2, section 4.20.7.1, lines 5078ff. */
182 #define RE_SYNTAX_ED RE_SYNTAX_POSIX_BASIC
183
184 #define RE_SYNTAX_SED RE_SYNTAX_POSIX_BASIC
185
186 /* Syntax bits common to both basic and extended POSIX regex syntax. */
187 #define _RE_SYNTAX_POSIX_COMMON \
188 (RE_CHAR_CLASSES | RE_DOT_NEWLINE | RE_DOT_NOT_NULL \
189 | RE_INTERVALS | RE_NO_EMPTY_RANGES)
190
191 #define RE_SYNTAX_POSIX_BASIC \
192 (_RE_SYNTAX_POSIX_COMMON | RE_BK_PLUS_QM)
193
194 /* Differs from ..._POSIX_BASIC only in that RE_BK_PLUS_QM becomes
195 RE_LIMITED_OPS, i.e., \? \+ \| are not recognized. Actually, this
196 isn't minimal, since other operators, such as \`, aren't disabled. */
197 #define RE_SYNTAX_POSIX_MINIMAL_BASIC \
198 (_RE_SYNTAX_POSIX_COMMON | RE_LIMITED_OPS)
199
200 #define RE_SYNTAX_POSIX_EXTENDED \
201 (_RE_SYNTAX_POSIX_COMMON | RE_CONTEXT_INDEP_ANCHORS \
202 | RE_CONTEXT_INDEP_OPS | RE_NO_BK_BRACES \
203 | RE_NO_BK_PARENS | RE_NO_BK_VBAR \
204 | RE_UNMATCHED_RIGHT_PAREN_ORD)
205
206 /* Differs from ..._POSIX_EXTENDED in that RE_CONTEXT_INVALID_OPS
207 replaces RE_CONTEXT_INDEP_OPS and RE_NO_BK_REFS is added. */
208 #define RE_SYNTAX_POSIX_MINIMAL_EXTENDED \
209 (_RE_SYNTAX_POSIX_COMMON | RE_CONTEXT_INDEP_ANCHORS \
210 | RE_CONTEXT_INVALID_OPS | RE_NO_BK_BRACES \
211 | RE_NO_BK_PARENS | RE_NO_BK_REFS \
212 | RE_NO_BK_VBAR | RE_UNMATCHED_RIGHT_PAREN_ORD)
213 /* [[[end syntaxes]]] */
214
215 /* Maximum number of duplicates an interval can allow. Some systems
216 (erroneously) define this in other header files, but we want our
217 value, so remove any previous define. */
218 #ifdef RE_DUP_MAX
219 #undef RE_DUP_MAX
220 #endif
221 #define RE_DUP_MAX ((1 << 15) - 1)
222
223
224 /* POSIX `cflags' bits (i.e., information for `regcomp'). */
225
226 /* If this bit is set, then use extended regular expression syntax.
227 If not set, then use basic regular expression syntax. */
228 #define REG_EXTENDED 1
229
230 /* If this bit is set, then ignore case when matching.
231 If not set, then case is significant. */
232 #define REG_ICASE (REG_EXTENDED << 1)
233
234 /* If this bit is set, then anchors do not match at newline
235 characters in the string.
236 If not set, then anchors do match at newlines. */
237 #define REG_NEWLINE (REG_ICASE << 1)
238
239 /* If this bit is set, then report only success or fail in regexec.
240 If not set, then returns differ between not matching and errors. */
241 #define REG_NOSUB (REG_NEWLINE << 1)
242
243
244 /* POSIX `eflags' bits (i.e., information for regexec). */
245
246 /* If this bit is set, then the beginning-of-line operator doesn't match
247 the beginning of the string (presumably because it's not the
248 beginning of a line).
249 If not set, then the beginning-of-line operator does match the
250 beginning of the string. */
251 #define REG_NOTBOL 1
252
253 /* Like REG_NOTBOL, except for the end-of-line. */
254 #define REG_NOTEOL (1 << 1)
255
256
257 /* If any error codes are removed, changed, or added, update the
258 `re_error_msg' table in regex.c. */
259 typedef enum
260 {
261 REG_NOERROR = 0, /* Success. */
262 REG_NOMATCH, /* Didn't find a match (for regexec). */
263
264 /* POSIX regcomp return error codes. (In the order listed in the
265 standard.) */
266 REG_BADPAT, /* Invalid pattern. */
267 REG_ECOLLATE, /* Not implemented. */
268 REG_ECTYPE, /* Invalid character class name. */
269 REG_EESCAPE, /* Trailing backslash. */
270 REG_ESUBREG, /* Invalid back reference. */
271 REG_EBRACK, /* Unmatched left bracket. */
272 REG_EPAREN, /* Parenthesis imbalance. */
273 REG_EBRACE, /* Unmatched \{. */
274 REG_BADBR, /* Invalid contents of \{\}. */
275 REG_ERANGE, /* Invalid range end. */
276 REG_ESPACE, /* Ran out of memory. */
277 REG_BADRPT, /* No preceding re for repetition op. */
278
279 /* Error codes we've added. */
280 REG_EEND, /* Premature end. */
281 REG_ESIZE, /* Compiled pattern bigger than 2^16 bytes. */
282 REG_ERPAREN /* Unmatched ) or \); not returned from regcomp. */
283 } reg_errcode_t;
284
285 /* This data structure represents a compiled pattern. Before calling
286 the pattern compiler, the fields `buffer', `allocated', `fastmap',
287 `translate', and `no_sub' can be set. After the pattern has been
288 compiled, the `re_nsub' field is available. All other fields are
289 private to the regex routines. */
290
291 struct re_pattern_buffer
292 {
293 /* [[[begin pattern_buffer]]] */
294 /* Space that holds the compiled pattern. It is declared as
295 `unsigned char *' because its elements are
296 sometimes used as array indexes. */
297 unsigned char *buffer;
298
299 /* Number of bytes to which `buffer' points. */
300 unsigned long allocated;
301
302 /* Number of bytes actually used in `buffer'. */
303 unsigned long used;
304
305 /* Syntax setting with which the pattern was compiled. */
306 reg_syntax_t syntax;
307
308 /* Pointer to a fastmap, if any, otherwise zero. re_search uses
309 the fastmap, if there is one, to skip over impossible
310 starting points for matches. */
311 char *fastmap;
312
313 /* Either a translate table to apply to all characters before
314 comparing them, or zero for no translation. The translation
315 is applied to a pattern when it is compiled and to a string
316 when it is matched. */
317 char *translate;
318
319 /* Number of subexpressions found by the compiler. */
320 size_t re_nsub;
321
322 /* Zero if this pattern cannot match the empty string, one else.
323 Well, in truth it's used only in `re_search_2', to see
324 whether or not we should use the fastmap, so we don't set
325 this absolutely perfectly; see `re_compile_fastmap' (the
326 `duplicate' case). */
327 unsigned can_be_null : 1;
328
329 /* If REGS_UNALLOCATED, allocate space in the `regs' structure
330 for `max (RE_NREGS, re_nsub + 1)' groups.
331 If REGS_REALLOCATE, reallocate space if necessary.
332 If REGS_FIXED, use what's there. */
333 #define REGS_UNALLOCATED 0
334 #define REGS_REALLOCATE 1
335 #define REGS_FIXED 2
336 unsigned regs_allocated : 2;
337
338 /* Set to zero when `regex_compile' compiles a pattern; set to one
339 by `re_compile_fastmap' if it updates the fastmap. */
340 unsigned fastmap_accurate : 1;
341
342 /* If set, `re_match_2' does not return information about
343 subexpressions. */
344 unsigned no_sub : 1;
345
346 /* If set, a beginning-of-line anchor doesn't match at the
347 beginning of the string. */
348 unsigned not_bol : 1;
349
350 /* Similarly for an end-of-line anchor. */
351 unsigned not_eol : 1;
352
353 /* If true, an anchor at a newline matches. */
354 unsigned newline_anchor : 1;
355
356 /* [[[end pattern_buffer]]] */
357 };
358
359 typedef struct re_pattern_buffer regex_t;
360
361 /* Type for byte offsets within the string. POSIX mandates this. */
362 typedef int regoff_t;
363
364
365 /* This is the structure we store register match data in. See
366 regex.texinfo for a full description of what registers match. */
367 struct re_registers
368 {
369 unsigned num_regs;
370 regoff_t *start;
371 regoff_t *end;
372 };
373
374
375 /* If `regs_allocated' is REGS_UNALLOCATED in the pattern buffer,
376 `re_match_2' returns information about at least this many registers
377 the first time a `regs' structure is passed. */
378 #ifndef RE_NREGS
379 #define RE_NREGS 30
380 #endif
381
382
383 /* POSIX specification for registers. Aside from the different names than
384 `re_registers', POSIX uses an array of structures, instead of a
385 structure of arrays. */
386 typedef struct
387 {
388 regoff_t rm_so; /* Byte offset from string's start to substring's start. */
389 regoff_t rm_eo; /* Byte offset from string's start to substring's end. */
390 } regmatch_t;
391
392 /* Declarations for routines. */
393
394 /* To avoid duplicating every routine declaration -- once with a
395 prototype (if we are ANSI), and once without (if we aren't) -- we
396 use the following macro to declare argument types. This
397 unfortunately clutters up the declarations a bit, but I think it's
398 worth it. */
399
400 #if __STDC__ || defined(_WIN32)
401
402 #define _RE_ARGS(args) args
403
404 #else /* not __STDC__ */
405
406 #define _RE_ARGS(args) ()
407
408 #endif /* not __STDC__ */
409
410 /* Sets the current default syntax to SYNTAX, and return the old syntax.
411 You can also simply assign to the `re_syntax_options' variable. */
412 EXTERN_DECL reg_syntax_t LIBNETXMS_EXPORTABLE re_set_syntax _RE_ARGS ((reg_syntax_t syntax));
413
414 /* Compile the regular expression PATTERN, with length LENGTH
415 and syntax given by the global `re_syntax_options', into the buffer
416 BUFFER. Return NULL if successful, and an error string if not. */
417 EXTERN_DECL const char LIBNETXMS_EXPORTABLE *re_compile_pattern
418 _RE_ARGS ((const char *pattern, int length,
419 struct re_pattern_buffer *buffer));
420
421
422 /* Compile a fastmap for the compiled pattern in BUFFER; used to
423 accelerate searches. Return 0 if successful and -2 if was an
424 internal error. */
425 EXTERN_DECL int LIBNETXMS_EXPORTABLE re_compile_fastmap _RE_ARGS ((struct re_pattern_buffer *buffer));
426
427
428 /* Search in the string STRING (with length LENGTH) for the pattern
429 compiled into BUFFER. Start searching at position START, for RANGE
430 characters. Return the starting position of the match, -1 for no
431 match, or -2 for an internal error. Also return register
432 information in REGS (if REGS and BUFFER->no_sub are nonzero). */
433 EXTERN_DECL int LIBNETXMS_EXPORTABLE re_search
434 _RE_ARGS ((struct re_pattern_buffer *buffer, const char *string,
435 int length, int start, int range, struct re_registers *regs));
436
437
438 /* Like `re_search', but search in the concatenation of STRING1 and
439 STRING2. Also, stop searching at index START + STOP. */
440 EXTERN_DECL int LIBNETXMS_EXPORTABLE re_search_2
441 _RE_ARGS ((struct re_pattern_buffer *buffer, const char *string1,
442 int length1, const char *string2, int length2,
443 int start, int range, struct re_registers *regs, int stop));
444
445
446 /* Like `re_search', but return how many characters in STRING the regexp
447 in BUFFER matched, starting at position START. */
448 EXTERN_DECL int LIBNETXMS_EXPORTABLE re_match
449 _RE_ARGS ((struct re_pattern_buffer *buffer, const char *string,
450 int length, int start, struct re_registers *regs));
451
452
453 /* Relates to `re_match' as `re_search_2' relates to `re_search'. */
454 EXTERN_DECL int LIBNETXMS_EXPORTABLE re_match_2
455 _RE_ARGS ((struct re_pattern_buffer *buffer, const char *string1,
456 int length1, const char *string2, int length2,
457 int start, struct re_registers *regs, int stop));
458
459
460 /* Set REGS to hold NUM_REGS registers, storing them in STARTS and
461 ENDS. Subsequent matches using BUFFER and REGS will use this memory
462 for recording register information. STARTS and ENDS must be
463 allocated with malloc, and must each be at least `NUM_REGS * sizeof
464 (regoff_t)' bytes long.
465
466 If NUM_REGS == 0, then subsequent matches should allocate their own
467 register data.
468
469 Unless this function is called, the first search or match using
470 PATTERN_BUFFER will allocate its own register data, without
471 freeing the old data. */
472 EXTERN_DECL void LIBNETXMS_EXPORTABLE re_set_registers
473 _RE_ARGS ((struct re_pattern_buffer *buffer, struct re_registers *regs,
474 unsigned num_regs, regoff_t *starts, regoff_t *ends));
475
476 #ifdef _REGEX_RE_COMP
477 /* 4.2 bsd compatibility. */
478 EXTERN_DECL char LIBNETXMS_EXPORTABLE *re_comp _RE_ARGS ((const char *));
479 EXTERN_DECL int LIBNETXMS_EXPORTABLE re_exec _RE_ARGS ((const char *));
480 #endif
481
482 /* POSIX compatibility. */
483 EXTERN_DECL int LIBNETXMS_EXPORTABLE regcomp _RE_ARGS ((regex_t *preg, const char *pattern, int cflags));
484 EXTERN_DECL int LIBNETXMS_EXPORTABLE regexec
485 _RE_ARGS ((const regex_t *preg, const char *string, size_t nmatch,
486 regmatch_t pmatch[], int eflags));
487 EXTERN_DECL size_t LIBNETXMS_EXPORTABLE regerror
488 _RE_ARGS ((int errcode, const regex_t *preg, char *errbuf,
489 size_t errbuf_size));
490 EXTERN_DECL void LIBNETXMS_EXPORTABLE regfree _RE_ARGS ((regex_t *preg));
491
492 #endif /* not __REGEXP_LIBRARY_H__ */
493
494 /*
495 Local variables:
496 make-backup-files: t
497 version-control: t
498 trim-versions-without-asking: nil
499 End:
500 */