1 /* Definitions for data structures and routines for the regular
2 expression library, version 0.12.
3 Copyright (C) 1985,89,90,91,92,93,95,96,97,98 Free Software Foundation, Inc.
5 NOTE: The canonical source of this file is maintained with the
6 GNU C Library. Bugs can be reported to bug-glibc@prep.ai.mit.edu.
8 This program is free software; you can redistribute it and/or modify it
9 under the terms of the GNU General Public License as published by the
10 Free Software Foundation; either version 2, or (at your option) any
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.
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 Foundation,
20 Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
25 /* Allow the use in C++ code. */
30 /* POSIX says that <sys/types.h> must be included (by the caller) before
33 #if !defined _POSIX_C_SOURCE && !defined _POSIX_SOURCE && defined VMS
34 /* VMS doesn't have `size_t' in <sys/types.h>, even though POSIX says it
39 /* The following two types have to be signed and unsigned integer type
40 wide enough to hold a value of a pointer. For most ANSI compilers
41 ptrdiff_t and size_t should be likely OK. Still size of these two
42 types is 2 for Microsoft C. Ugh... */
43 typedef long int s_reg_t;
44 typedef unsigned long int active_reg_t;
46 /* The following bits are used to determine the regexp syntax we
47 recognize. The set/not-set meanings are chosen so that Emacs syntax
48 remains the value 0. The bits are given in alphabetical order, and
49 the definitions shifted by one from the previous bit; thus, when we
50 add or remove a bit, only one other definition need change. */
51 typedef unsigned long int reg_syntax_t;
53 /* If this bit is not set, then \ inside a bracket expression is literal.
54 If set, then such a \ quotes the following character. */
55 #define RE_BACKSLASH_ESCAPE_IN_LISTS ((unsigned long int) 1)
57 /* If this bit is not set, then + and ? are operators, and \+ and \? are
59 If set, then \+ and \? are operators and + and ? are literals. */
60 #define RE_BK_PLUS_QM (RE_BACKSLASH_ESCAPE_IN_LISTS << 1)
62 /* If this bit is set, then character classes are supported. They are:
63 [:alpha:], [:upper:], [:lower:], [:digit:], [:alnum:], [:xdigit:],
64 [:space:], [:print:], [:punct:], [:graph:], and [:cntrl:].
65 If not set, then character classes are not supported. */
66 #define RE_CHAR_CLASSES (RE_BK_PLUS_QM << 1)
68 /* If this bit is set, then ^ and $ are always anchors (outside bracket
69 expressions, of course).
70 If this bit is not set, then it depends:
71 ^ is an anchor if it is at the beginning of a regular
72 expression or after an open-group or an alternation operator;
73 $ is an anchor if it is at the end of a regular expression, or
74 before a close-group or an alternation operator.
76 This bit could be (re)combined with RE_CONTEXT_INDEP_OPS, because
77 POSIX draft 11.2 says that * etc. in leading positions is undefined.
78 We already implemented a previous draft which made those constructs
79 invalid, though, so we haven't changed the code back. */
80 #define RE_CONTEXT_INDEP_ANCHORS (RE_CHAR_CLASSES << 1)
82 /* If this bit is set, then special characters are always special
83 regardless of where they are in the pattern.
84 If this bit is not set, then special characters are special only in
85 some contexts; otherwise they are ordinary. Specifically,
86 * + ? and intervals are only special when not after the beginning,
87 open-group, or alternation operator. */
88 #define RE_CONTEXT_INDEP_OPS (RE_CONTEXT_INDEP_ANCHORS << 1)
90 /* If this bit is set, then *, +, ?, and { cannot be first in an re or
91 immediately after an alternation or begin-group operator. */
92 #define RE_CONTEXT_INVALID_OPS (RE_CONTEXT_INDEP_OPS << 1)
94 /* If this bit is set, then . matches newline.
95 If not set, then it doesn't. */
96 #define RE_DOT_NEWLINE (RE_CONTEXT_INVALID_OPS << 1)
98 /* If this bit is set, then . doesn't match NUL.
99 If not set, then it does. */
100 #define RE_DOT_NOT_NULL (RE_DOT_NEWLINE << 1)
102 /* If this bit is set, nonmatching lists [^...] do not match newline.
103 If not set, they do. */
104 #define RE_HAT_LISTS_NOT_NEWLINE (RE_DOT_NOT_NULL << 1)
106 /* If this bit is set, either \{...\} or {...} defines an
107 interval, depending on RE_NO_BK_BRACES.
108 If not set, \{, \}, {, and } are literals. */
109 #define RE_INTERVALS (RE_HAT_LISTS_NOT_NEWLINE << 1)
111 /* If this bit is set, +, ? and | aren't recognized as operators.
112 If not set, they are. */
113 #define RE_LIMITED_OPS (RE_INTERVALS << 1)
115 /* If this bit is set, newline is an alternation operator.
116 If not set, newline is literal. */
117 #define RE_NEWLINE_ALT (RE_LIMITED_OPS << 1)
119 /* If this bit is set, then `{...}' defines an interval, and \{ and \}
121 If not set, then `\{...\}' defines an interval. */
122 #define RE_NO_BK_BRACES (RE_NEWLINE_ALT << 1)
124 /* If this bit is set, (...) defines a group, and \( and \) are literals.
125 If not set, \(...\) defines a group, and ( and ) are literals. */
126 #define RE_NO_BK_PARENS (RE_NO_BK_BRACES << 1)
128 /* If this bit is set, then \<digit> matches <digit>.
129 If not set, then \<digit> is a back-reference. */
130 #define RE_NO_BK_REFS (RE_NO_BK_PARENS << 1)
132 /* If this bit is set, then | is an alternation operator, and \| is literal.
133 If not set, then \| is an alternation operator, and | is literal. */
134 #define RE_NO_BK_VBAR (RE_NO_BK_REFS << 1)
136 /* If this bit is set, then an ending range point collating higher
137 than the starting range point, as in [z-a], is invalid.
138 If not set, then when ending range point collates higher than the
139 starting range point, the range is ignored. */
140 #define RE_NO_EMPTY_RANGES (RE_NO_BK_VBAR << 1)
142 /* If this bit is set, then an unmatched ) is ordinary.
143 If not set, then an unmatched ) is invalid. */
144 #define RE_UNMATCHED_RIGHT_PAREN_ORD (RE_NO_EMPTY_RANGES << 1)
146 /* If this bit is set, succeed as soon as we match the whole pattern,
147 without further backtracking. */
148 #define RE_NO_POSIX_BACKTRACKING (RE_UNMATCHED_RIGHT_PAREN_ORD << 1)
150 /* If this bit is set, do not process the GNU regex operators.
151 If not set, then the GNU regex operators are recognized. */
152 #define RE_NO_GNU_OPS (RE_NO_POSIX_BACKTRACKING << 1)
154 /* If this bit is set, turn on internal regex debugging.
155 If not set, and debugging was on, turn it off.
156 This only works if regex.c is compiled -DDEBUG.
157 We define this bit always, so that all that's needed to turn on
158 debugging is to recompile regex.c; the calling code can always have
159 this bit set, and it won't affect anything in the normal case. */
160 #define RE_DEBUG (RE_NO_GNU_OPS << 1)
162 /* This global variable defines the particular regexp syntax to use (for
163 some interfaces). When a regexp is compiled, the syntax used is
164 stored in the pattern buffer, so changing this does not affect
165 already-compiled regexps. */
166 extern reg_syntax_t re_syntax_options;
168 /* Define combinations of the above bits for the standard possibilities.
169 (The [[[ comments delimit what gets put into the Texinfo file, so
170 don't delete them!) */
171 /* [[[begin syntaxes]]] */
172 #define RE_SYNTAX_EMACS 0
174 #define RE_SYNTAX_AWK \
175 (RE_BACKSLASH_ESCAPE_IN_LISTS | RE_DOT_NOT_NULL \
176 | RE_NO_BK_PARENS | RE_NO_BK_REFS \
177 | RE_NO_BK_VBAR | RE_NO_EMPTY_RANGES \
178 | RE_DOT_NEWLINE | RE_CONTEXT_INDEP_ANCHORS \
179 | RE_UNMATCHED_RIGHT_PAREN_ORD | RE_NO_GNU_OPS)
181 #define RE_SYNTAX_GNU_AWK \
182 ((RE_SYNTAX_POSIX_EXTENDED | RE_BACKSLASH_ESCAPE_IN_LISTS | RE_DEBUG) \
183 & ~(RE_DOT_NOT_NULL | RE_INTERVALS | RE_CONTEXT_INDEP_OPS))
185 #define RE_SYNTAX_POSIX_AWK \
186 (RE_SYNTAX_POSIX_EXTENDED | RE_BACKSLASH_ESCAPE_IN_LISTS \
187 | RE_INTERVALS | RE_NO_GNU_OPS)
189 #define RE_SYNTAX_GREP \
190 (RE_BK_PLUS_QM | RE_CHAR_CLASSES \
191 | RE_HAT_LISTS_NOT_NEWLINE | RE_INTERVALS \
194 #define RE_SYNTAX_EGREP \
195 (RE_CHAR_CLASSES | RE_CONTEXT_INDEP_ANCHORS \
196 | RE_CONTEXT_INDEP_OPS | RE_HAT_LISTS_NOT_NEWLINE \
197 | RE_NEWLINE_ALT | RE_NO_BK_PARENS \
200 #define RE_SYNTAX_POSIX_EGREP \
201 (RE_SYNTAX_EGREP | RE_INTERVALS | RE_NO_BK_BRACES)
203 /* P1003.2/D11.2, section 4.20.7.1, lines 5078ff. */
204 #define RE_SYNTAX_ED RE_SYNTAX_POSIX_BASIC
206 #define RE_SYNTAX_SED RE_SYNTAX_POSIX_BASIC
208 /* Syntax bits common to both basic and extended POSIX regex syntax. */
209 #define _RE_SYNTAX_POSIX_COMMON \
210 (RE_CHAR_CLASSES | RE_DOT_NEWLINE | RE_DOT_NOT_NULL \
211 | RE_INTERVALS | RE_NO_EMPTY_RANGES)
213 #define RE_SYNTAX_POSIX_BASIC \
214 (_RE_SYNTAX_POSIX_COMMON | RE_BK_PLUS_QM)
216 /* Differs from ..._POSIX_BASIC only in that RE_BK_PLUS_QM becomes
217 RE_LIMITED_OPS, i.e., \? \+ \| are not recognized. Actually, this
218 isn't minimal, since other operators, such as \`, aren't disabled. */
219 #define RE_SYNTAX_POSIX_MINIMAL_BASIC \
220 (_RE_SYNTAX_POSIX_COMMON | RE_LIMITED_OPS)
222 #define RE_SYNTAX_POSIX_EXTENDED \
223 (_RE_SYNTAX_POSIX_COMMON | RE_CONTEXT_INDEP_ANCHORS \
224 | RE_CONTEXT_INDEP_OPS | RE_NO_BK_BRACES \
225 | RE_NO_BK_PARENS | RE_NO_BK_VBAR \
226 | RE_UNMATCHED_RIGHT_PAREN_ORD)
228 /* Differs from ..._POSIX_EXTENDED in that RE_CONTEXT_INVALID_OPS
229 replaces RE_CONTEXT_INDEP_OPS and RE_NO_BK_REFS is added. */
230 #define RE_SYNTAX_POSIX_MINIMAL_EXTENDED \
231 (_RE_SYNTAX_POSIX_COMMON | RE_CONTEXT_INDEP_ANCHORS \
232 | RE_CONTEXT_INVALID_OPS | RE_NO_BK_BRACES \
233 | RE_NO_BK_PARENS | RE_NO_BK_REFS \
234 | RE_NO_BK_VBAR | RE_UNMATCHED_RIGHT_PAREN_ORD)
235 /* [[[end syntaxes]]] */
237 /* Maximum number of duplicates an interval can allow. Some systems
238 (erroneously) define this in other header files, but we want our
239 value, so remove any previous define. */
243 /* If sizeof(int) == 2, then ((1 << 15) - 1) overflows. */
244 #define RE_DUP_MAX (0x7fff)
247 /* POSIX `cflags' bits (i.e., information for `regcomp'). */
249 /* If this bit is set, then use extended regular expression syntax.
250 If not set, then use basic regular expression syntax. */
251 #define REG_EXTENDED 1
253 /* If this bit is set, then ignore case when matching.
254 If not set, then case is significant. */
255 #define REG_ICASE (REG_EXTENDED << 1)
257 /* If this bit is set, then anchors do not match at newline
258 characters in the string.
259 If not set, then anchors do match at newlines. */
260 #define REG_NEWLINE (REG_ICASE << 1)
262 /* If this bit is set, then report only success or fail in regexec.
263 If not set, then returns differ between not matching and errors. */
264 #define REG_NOSUB (REG_NEWLINE << 1)
267 /* POSIX `eflags' bits (i.e., information for regexec). */
269 /* If this bit is set, then the beginning-of-line operator doesn't match
270 the beginning of the string (presumably because it's not the
271 beginning of a line).
272 If not set, then the beginning-of-line operator does match the
273 beginning of the string. */
276 /* Like REG_NOTBOL, except for the end-of-line. */
277 #define REG_NOTEOL (1 << 1)
280 /* If any error codes are removed, changed, or added, update the
281 `re_error_msg' table in regex.c. */
284 #if (_XOPEN_SOURCE - 0) == 500
285 REG_NOSYS = -1, /* This will never happen for this implementation. */
288 REG_NOERROR = 0, /* Success. */
289 REG_NOMATCH, /* Didn't find a match (for regexec). */
291 /* POSIX regcomp return error codes. (In the order listed in the
293 REG_BADPAT, /* Invalid pattern. */
294 REG_ECOLLATE, /* Not implemented. */
295 REG_ECTYPE, /* Invalid character class name. */
296 REG_EESCAPE, /* Trailing backslash. */
297 REG_ESUBREG, /* Invalid back reference. */
298 REG_EBRACK, /* Unmatched left bracket. */
299 REG_EPAREN, /* Parenthesis imbalance. */
300 REG_EBRACE, /* Unmatched \{. */
301 REG_BADBR, /* Invalid contents of \{\}. */
302 REG_ERANGE, /* Invalid range end. */
303 REG_ESPACE, /* Ran out of memory. */
304 REG_BADRPT, /* No preceding re for repetition op. */
306 /* Error codes we've added. */
307 REG_EEND, /* Premature end. */
308 REG_ESIZE, /* Compiled pattern bigger than 2^16 bytes. */
309 REG_ERPAREN /* Unmatched ) or \); not returned from regcomp. */
312 /* This data structure represents a compiled pattern. Before calling
313 the pattern compiler, the fields `buffer', `allocated', `fastmap',
314 `translate', and `no_sub' can be set. After the pattern has been
315 compiled, the `re_nsub' field is available. All other fields are
316 private to the regex routines. */
318 #ifndef RE_TRANSLATE_TYPE
319 # define RE_TRANSLATE_TYPE char *
322 struct re_pattern_buffer
324 /* [[[begin pattern_buffer]]] */
325 /* Space that holds the compiled pattern. It is declared as
326 `unsigned char *' because its elements are
327 sometimes used as array indexes. */
328 unsigned char *buffer;
330 /* Number of bytes to which `buffer' points. */
331 unsigned long int allocated;
333 /* Number of bytes actually used in `buffer'. */
334 unsigned long int used;
336 /* Syntax setting with which the pattern was compiled. */
339 /* Pointer to a fastmap, if any, otherwise zero. re_search uses
340 the fastmap, if there is one, to skip over impossible
341 starting points for matches. */
344 /* Either a translate table to apply to all characters before
345 comparing them, or zero for no translation. The translation
346 is applied to a pattern when it is compiled and to a string
347 when it is matched. */
348 RE_TRANSLATE_TYPE translate;
350 /* Number of subexpressions found by the compiler. */
353 /* Zero if this pattern cannot match the empty string, one else.
354 Well, in truth it's used only in `re_search_2', to see
355 whether or not we should use the fastmap, so we don't set
356 this absolutely perfectly; see `re_compile_fastmap' (the
357 `duplicate' case). */
358 unsigned can_be_null : 1;
360 /* If REGS_UNALLOCATED, allocate space in the `regs' structure
361 for `max (RE_NREGS, re_nsub + 1)' groups.
362 If REGS_REALLOCATE, reallocate space if necessary.
363 If REGS_FIXED, use what's there. */
364 #define REGS_UNALLOCATED 0
365 #define REGS_REALLOCATE 1
367 unsigned regs_allocated : 2;
369 /* Set to zero when `regex_compile' compiles a pattern; set to one
370 by `re_compile_fastmap' if it updates the fastmap. */
371 unsigned fastmap_accurate : 1;
373 /* If set, `re_match_2' does not return information about
377 /* If set, a beginning-of-line anchor doesn't match at the
378 beginning of the string. */
379 unsigned not_bol : 1;
381 /* Similarly for an end-of-line anchor. */
382 unsigned not_eol : 1;
384 /* If true, an anchor at a newline matches. */
385 unsigned newline_anchor : 1;
387 /* [[[end pattern_buffer]]] */
390 typedef struct re_pattern_buffer regex_t;
392 /* Type for byte offsets within the string. POSIX mandates this. */
393 typedef int regoff_t;
396 /* This is the structure we store register match data in. See
397 regex.texinfo for a full description of what registers match. */
406 /* If `regs_allocated' is REGS_UNALLOCATED in the pattern buffer,
407 `re_match_2' returns information about at least this many registers
408 the first time a `regs' structure is passed. */
414 /* POSIX specification for registers. Aside from the different names than
415 `re_registers', POSIX uses an array of structures, instead of a
416 structure of arrays. */
419 regoff_t rm_so; /* Byte offset from string's start to substring's start. */
420 regoff_t rm_eo; /* Byte offset from string's start to substring's end. */
423 /* Declarations for routines. */
425 /* To avoid duplicating every routine declaration -- once with a
426 prototype (if we are ANSI), and once without (if we aren't) -- we
427 use the following macro to declare argument types. This
428 unfortunately clutters up the declarations a bit, but I think it's
433 # define _RE_ARGS(args) args
435 #else /* not __STDC__ */
437 # define _RE_ARGS(args) ()
439 #endif /* not __STDC__ */
441 /* Sets the current default syntax to SYNTAX, and return the old syntax.
442 You can also simply assign to the `re_syntax_options' variable. */
443 extern reg_syntax_t __re_set_syntax _RE_ARGS ((reg_syntax_t syntax));
444 extern reg_syntax_t re_set_syntax _RE_ARGS ((reg_syntax_t syntax));
446 /* Compile the regular expression PATTERN, with length LENGTH
447 and syntax given by the global `re_syntax_options', into the buffer
448 BUFFER. Return NULL if successful, and an error string if not. */
449 extern const char *__re_compile_pattern
450 _RE_ARGS ((const char *pattern, size_t length,
451 struct re_pattern_buffer *buffer));
452 extern const char *re_compile_pattern
453 _RE_ARGS ((const char *pattern, size_t length,
454 struct re_pattern_buffer *buffer));
457 /* Compile a fastmap for the compiled pattern in BUFFER; used to
458 accelerate searches. Return 0 if successful and -2 if was an
460 extern int __re_compile_fastmap _RE_ARGS ((struct re_pattern_buffer *buffer));
461 extern int re_compile_fastmap _RE_ARGS ((struct re_pattern_buffer *buffer));
464 /* Search in the string STRING (with length LENGTH) for the pattern
465 compiled into BUFFER. Start searching at position START, for RANGE
466 characters. Return the starting position of the match, -1 for no
467 match, or -2 for an internal error. Also return register
468 information in REGS (if REGS and BUFFER->no_sub are nonzero). */
469 extern int __re_search
470 _RE_ARGS ((struct re_pattern_buffer *buffer, const char *string,
471 int length, int start, int range, struct re_registers *regs));
473 _RE_ARGS ((struct re_pattern_buffer *buffer, const char *string,
474 int length, int start, int range, struct re_registers *regs));
477 /* Like `re_search', but search in the concatenation of STRING1 and
478 STRING2. Also, stop searching at index START + STOP. */
479 extern int __re_search_2
480 _RE_ARGS ((struct re_pattern_buffer *buffer, const char *string1,
481 int length1, const char *string2, int length2,
482 int start, int range, struct re_registers *regs, int stop));
483 extern int re_search_2
484 _RE_ARGS ((struct re_pattern_buffer *buffer, const char *string1,
485 int length1, const char *string2, int length2,
486 int start, int range, struct re_registers *regs, int stop));
489 /* Like `re_search', but return how many characters in STRING the regexp
490 in BUFFER matched, starting at position START. */
491 extern int __re_match
492 _RE_ARGS ((struct re_pattern_buffer *buffer, const char *string,
493 int length, int start, struct re_registers *regs));
495 _RE_ARGS ((struct re_pattern_buffer *buffer, const char *string,
496 int length, int start, struct re_registers *regs));
499 /* Relates to `re_match' as `re_search_2' relates to `re_search'. */
500 extern int __re_match_2
501 _RE_ARGS ((struct re_pattern_buffer *buffer, const char *string1,
502 int length1, const char *string2, int length2,
503 int start, struct re_registers *regs, int stop));
504 extern int re_match_2
505 _RE_ARGS ((struct re_pattern_buffer *buffer, const char *string1,
506 int length1, const char *string2, int length2,
507 int start, struct re_registers *regs, int stop));
510 /* Set REGS to hold NUM_REGS registers, storing them in STARTS and
511 ENDS. Subsequent matches using BUFFER and REGS will use this memory
512 for recording register information. STARTS and ENDS must be
513 allocated with malloc, and must each be at least `NUM_REGS * sizeof
514 (regoff_t)' bytes long.
516 If NUM_REGS == 0, then subsequent matches should allocate their own
519 Unless this function is called, the first search or match using
520 PATTERN_BUFFER will allocate its own register data, without
521 freeing the old data. */
522 extern void __re_set_registers
523 _RE_ARGS ((struct re_pattern_buffer *buffer, struct re_registers *regs,
524 unsigned num_regs, regoff_t *starts, regoff_t *ends));
525 extern void re_set_registers
526 _RE_ARGS ((struct re_pattern_buffer *buffer, struct re_registers *regs,
527 unsigned num_regs, regoff_t *starts, regoff_t *ends));
529 #ifdef _REGEX_RE_COMP
531 /* 4.2 bsd compatibility. */
532 extern char *re_comp _RE_ARGS ((const char *));
533 extern int re_exec _RE_ARGS ((const char *));
537 /* POSIX compatibility. */
538 extern int __regcomp _RE_ARGS ((regex_t *__preg, const char *__pattern,
540 extern int regcomp _RE_ARGS ((regex_t *__preg, const char *__pattern,
543 extern int __regexec _RE_ARGS ((const regex_t *__preg,
544 const char *__string, size_t __nmatch,
545 regmatch_t __pmatch[], int __eflags));
546 extern int regexec _RE_ARGS ((const regex_t *__preg,
547 const char *__string, size_t __nmatch,
548 regmatch_t __pmatch[], int __eflags));
550 extern size_t __regerror _RE_ARGS ((int __errcode, const regex_t *__preg,
551 char *__errbuf, size_t __errbuf_size));
552 extern size_t regerror _RE_ARGS ((int __errcode, const regex_t *__preg,
553 char *__errbuf, size_t __errbuf_size));
555 extern void __regfree _RE_ARGS ((regex_t *__preg));
556 extern void regfree _RE_ARGS ((regex_t *__preg));
569 trim-versions-without-asking: nil