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 Free Software Foundation, Inc.
6 NOTE: The canonical source of this file is maintained with the GNU C Library.
7 Bugs can be reported to bug-glibc@prep.ai.mit.edu.
9 This program is free software; you can redistribute it and/or modify it
10 under the terms of the GNU General Public License as published by the
11 Free Software Foundation; either version 2, or (at your option) any
14 This program is distributed in the hope that it will be useful,
15 but WITHOUT ANY WARRANTY; without even the implied warranty of
16 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
17 GNU General Public License for more details.
19 You should have received a copy of the GNU General Public License
20 along with this program; if not, write to the Free Software
21 Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307,
27 /* Allow the use in C++ code. */
32 /* POSIX says that <sys/types.h> must be included (by the caller) before
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
41 /* The following two types have to be signed and unsigned integer type
42 wide enough to hold a value of a pointer. For most ANSI compilers
43 ptrdiff_t and size_t should be likely OK. Still size of these two
44 types is 2 for Microsoft C. Ugh... */
45 typedef long int s_reg_t
;
46 typedef unsigned long int active_reg_t
;
48 /* The following bits are used to determine the regexp syntax we
49 recognize. The set/not-set meanings are chosen so that Emacs syntax
50 remains the value 0. The bits are given in alphabetical order, and
51 the definitions shifted by one from the previous bit; thus, when we
52 add or remove a bit, only one other definition need change. */
53 typedef unsigned long int reg_syntax_t
;
55 /* If this bit is not set, then \ inside a bracket expression is literal.
56 If set, then such a \ quotes the following character. */
57 #define RE_BACKSLASH_ESCAPE_IN_LISTS ((unsigned long int) 1)
59 /* If this bit is not set, then + and ? are operators, and \+ and \? are
61 If set, then \+ and \? are operators and + and ? are literals. */
62 #define RE_BK_PLUS_QM (RE_BACKSLASH_ESCAPE_IN_LISTS << 1)
64 /* If this bit is set, then character classes are supported. They are:
65 [:alpha:], [:upper:], [:lower:], [:digit:], [:alnum:], [:xdigit:],
66 [:space:], [:print:], [:punct:], [:graph:], and [:cntrl:].
67 If not set, then character classes are not supported. */
68 #define RE_CHAR_CLASSES (RE_BK_PLUS_QM << 1)
70 /* If this bit is set, then ^ and $ are always anchors (outside bracket
71 expressions, of course).
72 If this bit is not set, then it depends:
73 ^ is an anchor if it is at the beginning of a regular
74 expression or after an open-group or an alternation operator;
75 $ is an anchor if it is at the end of a regular expression, or
76 before a close-group or an alternation operator.
78 This bit could be (re)combined with RE_CONTEXT_INDEP_OPS, because
79 POSIX draft 11.2 says that * etc. in leading positions is undefined.
80 We already implemented a previous draft which made those constructs
81 invalid, though, so we haven't changed the code back. */
82 #define RE_CONTEXT_INDEP_ANCHORS (RE_CHAR_CLASSES << 1)
84 /* If this bit is set, then special characters are always special
85 regardless of where they are in the pattern.
86 If this bit is not set, then special characters are special only in
87 some contexts; otherwise they are ordinary. Specifically,
88 * + ? and intervals are only special when not after the beginning,
89 open-group, or alternation operator. */
90 #define RE_CONTEXT_INDEP_OPS (RE_CONTEXT_INDEP_ANCHORS << 1)
92 /* If this bit is set, then *, +, ?, and { cannot be first in an re or
93 immediately after an alternation or begin-group operator. */
94 #define RE_CONTEXT_INVALID_OPS (RE_CONTEXT_INDEP_OPS << 1)
96 /* If this bit is set, then . matches newline.
97 If not set, then it doesn't. */
98 #define RE_DOT_NEWLINE (RE_CONTEXT_INVALID_OPS << 1)
100 /* If this bit is set, then . doesn't match NUL.
101 If not set, then it does. */
102 #define RE_DOT_NOT_NULL (RE_DOT_NEWLINE << 1)
104 /* If this bit is set, nonmatching lists [^...] do not match newline.
105 If not set, they do. */
106 #define RE_HAT_LISTS_NOT_NEWLINE (RE_DOT_NOT_NULL << 1)
108 /* If this bit is set, either \{...\} or {...} defines an
109 interval, depending on RE_NO_BK_BRACES.
110 If not set, \{, \}, {, and } are literals. */
111 #define RE_INTERVALS (RE_HAT_LISTS_NOT_NEWLINE << 1)
113 /* If this bit is set, +, ? and | aren't recognized as operators.
114 If not set, they are. */
115 #define RE_LIMITED_OPS (RE_INTERVALS << 1)
117 /* If this bit is set, newline is an alternation operator.
118 If not set, newline is literal. */
119 #define RE_NEWLINE_ALT (RE_LIMITED_OPS << 1)
121 /* If this bit is set, then `{...}' defines an interval, and \{ and \}
123 If not set, then `\{...\}' defines an interval. */
124 #define RE_NO_BK_BRACES (RE_NEWLINE_ALT << 1)
126 /* If this bit is set, (...) defines a group, and \( and \) are literals.
127 If not set, \(...\) defines a group, and ( and ) are literals. */
128 #define RE_NO_BK_PARENS (RE_NO_BK_BRACES << 1)
130 /* If this bit is set, then \<digit> matches <digit>.
131 If not set, then \<digit> is a back-reference. */
132 #define RE_NO_BK_REFS (RE_NO_BK_PARENS << 1)
134 /* If this bit is set, then | is an alternation operator, and \| is literal.
135 If not set, then \| is an alternation operator, and | is literal. */
136 #define RE_NO_BK_VBAR (RE_NO_BK_REFS << 1)
138 /* If this bit is set, then an ending range point collating higher
139 than the starting range point, as in [z-a], is invalid.
140 If not set, then when ending range point collates higher than the
141 starting range point, the range is ignored. */
142 #define RE_NO_EMPTY_RANGES (RE_NO_BK_VBAR << 1)
144 /* If this bit is set, then an unmatched ) is ordinary.
145 If not set, then an unmatched ) is invalid. */
146 #define RE_UNMATCHED_RIGHT_PAREN_ORD (RE_NO_EMPTY_RANGES << 1)
148 /* If this bit is set, succeed as soon as we match the whole pattern,
149 without further backtracking. */
150 #define RE_NO_POSIX_BACKTRACKING (RE_UNMATCHED_RIGHT_PAREN_ORD << 1)
152 /* If this bit is set, do not process the GNU regex operators.
153 If not set, then the GNU regex operators are recognized. */
154 #define RE_NO_GNU_OPS (RE_NO_POSIX_BACKTRACKING << 1)
156 /* If this bit is set, turn on internal regex debugging.
157 If not set, and debugging was on, turn it off.
158 This only works if regex.c is compiled -DDEBUG.
159 We define this bit always, so that all that's needed to turn on
160 debugging is to recompile regex.c; the calling code can always have
161 this bit set, and it won't affect anything in the normal case. */
162 #define RE_DEBUG (RE_NO_GNU_OPS << 1)
164 /* This global variable defines the particular regexp syntax to use (for
165 some interfaces). When a regexp is compiled, the syntax used is
166 stored in the pattern buffer, so changing this does not affect
167 already-compiled regexps. */
168 extern reg_syntax_t re_syntax_options
;
170 /* Define combinations of the above bits for the standard possibilities.
171 (The [[[ comments delimit what gets put into the Texinfo file, so
172 don't delete them!) */
173 /* [[[begin syntaxes]]] */
174 #define RE_SYNTAX_EMACS 0
176 #define RE_SYNTAX_AWK \
177 (RE_BACKSLASH_ESCAPE_IN_LISTS | RE_DOT_NOT_NULL \
178 | RE_NO_BK_PARENS | RE_NO_BK_REFS \
179 | RE_NO_BK_VBAR | RE_NO_EMPTY_RANGES \
180 | RE_DOT_NEWLINE | RE_CONTEXT_INDEP_ANCHORS \
181 | RE_UNMATCHED_RIGHT_PAREN_ORD | RE_NO_GNU_OPS)
183 #define RE_SYNTAX_GNU_AWK \
184 ((RE_SYNTAX_POSIX_EXTENDED | RE_BACKSLASH_ESCAPE_IN_LISTS | RE_DEBUG) \
185 & ~(RE_DOT_NOT_NULL | RE_INTERVALS | RE_CONTEXT_INDEP_OPS))
187 #define RE_SYNTAX_POSIX_AWK \
188 (RE_SYNTAX_POSIX_EXTENDED | RE_BACKSLASH_ESCAPE_IN_LISTS \
189 | RE_INTERVALS | RE_NO_GNU_OPS)
191 #define RE_SYNTAX_GREP \
192 (RE_BK_PLUS_QM | RE_CHAR_CLASSES \
193 | RE_HAT_LISTS_NOT_NEWLINE | RE_INTERVALS \
196 #define RE_SYNTAX_EGREP \
197 (RE_CHAR_CLASSES | RE_CONTEXT_INDEP_ANCHORS \
198 | RE_CONTEXT_INDEP_OPS | RE_HAT_LISTS_NOT_NEWLINE \
199 | RE_NEWLINE_ALT | RE_NO_BK_PARENS \
202 #define RE_SYNTAX_POSIX_EGREP \
203 (RE_SYNTAX_EGREP | RE_INTERVALS | RE_NO_BK_BRACES)
205 /* P1003.2/D11.2, section 4.20.7.1, lines 5078ff. */
206 #define RE_SYNTAX_ED RE_SYNTAX_POSIX_BASIC
208 #define RE_SYNTAX_SED RE_SYNTAX_POSIX_BASIC
210 /* Syntax bits common to both basic and extended POSIX regex syntax. */
211 #define _RE_SYNTAX_POSIX_COMMON \
212 (RE_CHAR_CLASSES | RE_DOT_NEWLINE | RE_DOT_NOT_NULL \
213 | RE_INTERVALS | RE_NO_EMPTY_RANGES)
215 #define RE_SYNTAX_POSIX_BASIC \
216 (_RE_SYNTAX_POSIX_COMMON | RE_BK_PLUS_QM)
218 /* Differs from ..._POSIX_BASIC only in that RE_BK_PLUS_QM becomes
219 RE_LIMITED_OPS, i.e., \? \+ \| are not recognized. Actually, this
220 isn't minimal, since other operators, such as \`, aren't disabled. */
221 #define RE_SYNTAX_POSIX_MINIMAL_BASIC \
222 (_RE_SYNTAX_POSIX_COMMON | RE_LIMITED_OPS)
224 #define RE_SYNTAX_POSIX_EXTENDED \
225 (_RE_SYNTAX_POSIX_COMMON | RE_CONTEXT_INDEP_ANCHORS \
226 | RE_CONTEXT_INDEP_OPS | RE_NO_BK_BRACES \
227 | RE_NO_BK_PARENS | RE_NO_BK_VBAR \
228 | RE_UNMATCHED_RIGHT_PAREN_ORD)
230 /* Differs from ..._POSIX_EXTENDED in that RE_CONTEXT_INVALID_OPS
231 replaces RE_CONTEXT_INDEP_OPS and RE_NO_BK_REFS is added. */
232 #define RE_SYNTAX_POSIX_MINIMAL_EXTENDED \
233 (_RE_SYNTAX_POSIX_COMMON | RE_CONTEXT_INDEP_ANCHORS \
234 | RE_CONTEXT_INVALID_OPS | RE_NO_BK_BRACES \
235 | RE_NO_BK_PARENS | RE_NO_BK_REFS \
236 | RE_NO_BK_VBAR | RE_UNMATCHED_RIGHT_PAREN_ORD)
237 /* [[[end syntaxes]]] */
239 /* Maximum number of duplicates an interval can allow. Some systems
240 (erroneously) define this in other header files, but we want our
241 value, so remove any previous define. */
245 /* If sizeof(int) == 2, then ((1 << 15) - 1) overflows. */
246 #define RE_DUP_MAX (0x7fff)
249 /* POSIX `cflags' bits (i.e., information for `regcomp'). */
251 /* If this bit is set, then use extended regular expression syntax.
252 If not set, then use basic regular expression syntax. */
253 #define REG_EXTENDED 1
255 /* If this bit is set, then ignore case when matching.
256 If not set, then case is significant. */
257 #define REG_ICASE (REG_EXTENDED << 1)
259 /* If this bit is set, then anchors do not match at newline
260 characters in the string.
261 If not set, then anchors do match at newlines. */
262 #define REG_NEWLINE (REG_ICASE << 1)
264 /* If this bit is set, then report only success or fail in regexec.
265 If not set, then returns differ between not matching and errors. */
266 #define REG_NOSUB (REG_NEWLINE << 1)
269 /* POSIX `eflags' bits (i.e., information for regexec). */
271 /* If this bit is set, then the beginning-of-line operator doesn't match
272 the beginning of the string (presumably because it's not the
273 beginning of a line).
274 If not set, then the beginning-of-line operator does match the
275 beginning of the string. */
278 /* Like REG_NOTBOL, except for the end-of-line. */
279 #define REG_NOTEOL (1 << 1)
282 /* If any error codes are removed, changed, or added, update the
283 `re_error_msg' table in regex.c. */
286 #if (_XOPEN_SOURCE - 0) == 500
287 REG_NOSYS
= -1, /* This will never happen for this implementation. */
290 REG_NOERROR
= 0, /* Success. */
291 REG_NOMATCH
, /* Didn't find a match (for regexec). */
293 /* POSIX regcomp return error codes. (In the order listed in the
295 REG_BADPAT
, /* Invalid pattern. */
296 REG_ECOLLATE
, /* Not implemented. */
297 REG_ECTYPE
, /* Invalid character class name. */
298 REG_EESCAPE
, /* Trailing backslash. */
299 REG_ESUBREG
, /* Invalid back reference. */
300 REG_EBRACK
, /* Unmatched left bracket. */
301 REG_EPAREN
, /* Parenthesis imbalance. */
302 REG_EBRACE
, /* Unmatched \{. */
303 REG_BADBR
, /* Invalid contents of \{\}. */
304 REG_ERANGE
, /* Invalid range end. */
305 REG_ESPACE
, /* Ran out of memory. */
306 REG_BADRPT
, /* No preceding re for repetition op. */
308 /* Error codes we've added. */
309 REG_EEND
, /* Premature end. */
310 REG_ESIZE
, /* Compiled pattern bigger than 2^16 bytes. */
311 REG_ERPAREN
/* Unmatched ) or \); not returned from regcomp. */
314 /* This data structure represents a compiled pattern. Before calling
315 the pattern compiler, the fields `buffer', `allocated', `fastmap',
316 `translate', and `no_sub' can be set. After the pattern has been
317 compiled, the `re_nsub' field is available. All other fields are
318 private to the regex routines. */
320 #ifndef RE_TRANSLATE_TYPE
321 # define RE_TRANSLATE_TYPE char *
324 struct re_pattern_buffer
326 /* [[[begin pattern_buffer]]] */
327 /* Space that holds the compiled pattern. It is declared as
328 `unsigned char *' because its elements are
329 sometimes used as array indexes. */
330 unsigned char *buffer
;
332 /* Number of bytes to which `buffer' points. */
333 unsigned long int allocated
;
335 /* Number of bytes actually used in `buffer'. */
336 unsigned long int used
;
338 /* Syntax setting with which the pattern was compiled. */
341 /* Pointer to a fastmap, if any, otherwise zero. re_search uses
342 the fastmap, if there is one, to skip over impossible
343 starting points for matches. */
346 /* Either a translate table to apply to all characters before
347 comparing them, or zero for no translation. The translation
348 is applied to a pattern when it is compiled and to a string
349 when it is matched. */
350 RE_TRANSLATE_TYPE translate
;
352 /* Number of subexpressions found by the compiler. */
355 /* Zero if this pattern cannot match the empty string, one else.
356 Well, in truth it's used only in `re_search_2', to see
357 whether or not we should use the fastmap, so we don't set
358 this absolutely perfectly; see `re_compile_fastmap' (the
359 `duplicate' case). */
360 unsigned can_be_null
: 1;
362 /* If REGS_UNALLOCATED, allocate space in the `regs' structure
363 for `max (RE_NREGS, re_nsub + 1)' groups.
364 If REGS_REALLOCATE, reallocate space if necessary.
365 If REGS_FIXED, use what's there. */
366 #define REGS_UNALLOCATED 0
367 #define REGS_REALLOCATE 1
369 unsigned regs_allocated
: 2;
371 /* Set to zero when `regex_compile' compiles a pattern; set to one
372 by `re_compile_fastmap' if it updates the fastmap. */
373 unsigned fastmap_accurate
: 1;
375 /* If set, `re_match_2' does not return information about
379 /* If set, a beginning-of-line anchor doesn't match at the
380 beginning of the string. */
381 unsigned not_bol
: 1;
383 /* Similarly for an end-of-line anchor. */
384 unsigned not_eol
: 1;
386 /* If true, an anchor at a newline matches. */
387 unsigned newline_anchor
: 1;
389 /* [[[end pattern_buffer]]] */
392 typedef struct re_pattern_buffer regex_t
;
394 /* Type for byte offsets within the string. POSIX mandates this. */
395 typedef int regoff_t
;
398 /* This is the structure we store register match data in. See
399 regex.texinfo for a full description of what registers match. */
408 /* If `regs_allocated' is REGS_UNALLOCATED in the pattern buffer,
409 `re_match_2' returns information about at least this many registers
410 the first time a `regs' structure is passed. */
416 /* POSIX specification for registers. Aside from the different names than
417 `re_registers', POSIX uses an array of structures, instead of a
418 structure of arrays. */
421 regoff_t rm_so
; /* Byte offset from string's start to substring's start. */
422 regoff_t rm_eo
; /* Byte offset from string's start to substring's end. */
425 /* Declarations for routines. */
427 /* To avoid duplicating every routine declaration -- once with a
428 prototype (if we are ANSI), and once without (if we aren't) -- we
429 use the following macro to declare argument types. This
430 unfortunately clutters up the declarations a bit, but I think it's
435 # define _RE_ARGS(args) args
437 #else /* not __STDC__ */
439 # define _RE_ARGS(args) ()
441 #endif /* not __STDC__ */
443 /* Sets the current default syntax to SYNTAX, and return the old syntax.
444 You can also simply assign to the `re_syntax_options' variable. */
445 extern reg_syntax_t __re_set_syntax
_RE_ARGS ((reg_syntax_t syntax
));
446 extern reg_syntax_t re_set_syntax
_RE_ARGS ((reg_syntax_t syntax
));
448 /* Compile the regular expression PATTERN, with length LENGTH
449 and syntax given by the global `re_syntax_options', into the buffer
450 BUFFER. Return NULL if successful, and an error string if not. */
451 extern const char *__re_compile_pattern
452 _RE_ARGS ((const char *pattern
, size_t length
,
453 struct re_pattern_buffer
*buffer
));
454 extern const char *re_compile_pattern
455 _RE_ARGS ((const char *pattern
, size_t length
,
456 struct re_pattern_buffer
*buffer
));
459 /* Compile a fastmap for the compiled pattern in BUFFER; used to
460 accelerate searches. Return 0 if successful and -2 if was an
462 extern int __re_compile_fastmap
_RE_ARGS ((struct re_pattern_buffer
*buffer
));
463 extern int re_compile_fastmap
_RE_ARGS ((struct re_pattern_buffer
*buffer
));
466 /* Search in the string STRING (with length LENGTH) for the pattern
467 compiled into BUFFER. Start searching at position START, for RANGE
468 characters. Return the starting position of the match, -1 for no
469 match, or -2 for an internal error. Also return register
470 information in REGS (if REGS and BUFFER->no_sub are nonzero). */
471 extern int __re_search
472 _RE_ARGS ((struct re_pattern_buffer
*buffer
, const char *string
,
473 int length
, int start
, int range
, struct re_registers
*regs
));
475 _RE_ARGS ((struct re_pattern_buffer
*buffer
, const char *string
,
476 int length
, int start
, int range
, struct re_registers
*regs
));
479 /* Like `re_search', but search in the concatenation of STRING1 and
480 STRING2. Also, stop searching at index START + STOP. */
481 extern int __re_search_2
482 _RE_ARGS ((struct re_pattern_buffer
*buffer
, const char *string1
,
483 int length1
, const char *string2
, int length2
,
484 int start
, int range
, struct re_registers
*regs
, int stop
));
485 extern int re_search_2
486 _RE_ARGS ((struct re_pattern_buffer
*buffer
, const char *string1
,
487 int length1
, const char *string2
, int length2
,
488 int start
, int range
, struct re_registers
*regs
, int stop
));
491 /* Like `re_search', but return how many characters in STRING the regexp
492 in BUFFER matched, starting at position START. */
493 extern int __re_match
494 _RE_ARGS ((struct re_pattern_buffer
*buffer
, const char *string
,
495 int length
, int start
, struct re_registers
*regs
));
497 _RE_ARGS ((struct re_pattern_buffer
*buffer
, const char *string
,
498 int length
, int start
, struct re_registers
*regs
));
501 /* Relates to `re_match' as `re_search_2' relates to `re_search'. */
502 extern int __re_match_2
503 _RE_ARGS ((struct re_pattern_buffer
*buffer
, const char *string1
,
504 int length1
, const char *string2
, int length2
,
505 int start
, struct re_registers
*regs
, int stop
));
506 extern int re_match_2
507 _RE_ARGS ((struct re_pattern_buffer
*buffer
, const char *string1
,
508 int length1
, const char *string2
, int length2
,
509 int start
, struct re_registers
*regs
, int stop
));
512 /* Set REGS to hold NUM_REGS registers, storing them in STARTS and
513 ENDS. Subsequent matches using BUFFER and REGS will use this memory
514 for recording register information. STARTS and ENDS must be
515 allocated with malloc, and must each be at least `NUM_REGS * sizeof
516 (regoff_t)' bytes long.
518 If NUM_REGS == 0, then subsequent matches should allocate their own
521 Unless this function is called, the first search or match using
522 PATTERN_BUFFER will allocate its own register data, without
523 freeing the old data. */
524 extern void __re_set_registers
525 _RE_ARGS ((struct re_pattern_buffer
*buffer
, struct re_registers
*regs
,
526 unsigned num_regs
, regoff_t
*starts
, regoff_t
*ends
));
527 extern void re_set_registers
528 _RE_ARGS ((struct re_pattern_buffer
*buffer
, struct re_registers
*regs
,
529 unsigned num_regs
, regoff_t
*starts
, regoff_t
*ends
));
531 #ifdef _REGEX_RE_COMP
533 /* 4.2 bsd compatibility. */
534 extern char *re_comp
_RE_ARGS ((const char *));
535 extern int re_exec
_RE_ARGS ((const char *));
539 /* POSIX compatibility. */
540 extern int __regcomp
_RE_ARGS ((regex_t
*__preg
, const char *__pattern
,
542 extern int regcomp
_RE_ARGS ((regex_t
*__preg
, const char *__pattern
,
545 extern int __regexec
_RE_ARGS ((const regex_t
*__preg
,
546 const char *__string
, size_t __nmatch
,
547 regmatch_t __pmatch
[], int __eflags
));
548 extern int regexec
_RE_ARGS ((const regex_t
*__preg
,
549 const char *__string
, size_t __nmatch
,
550 regmatch_t __pmatch
[], int __eflags
));
552 extern size_t __regerror
_RE_ARGS ((int __errcode
, const regex_t
*__preg
,
553 char *__errbuf
, size_t __errbuf_size
));
554 extern size_t regerror
_RE_ARGS ((int __errcode
, const regex_t
*__preg
,
555 char *__errbuf
, size_t __errbuf_size
));
557 extern void __regfree
_RE_ARGS ((regex_t
*__preg
));
558 extern void regfree
_RE_ARGS ((regex_t
*__preg
));
571 trim-versions-without-asking: nil