1 .\" Copyright (c) 1992, 1993, 1994 Henry Spencer.
2 .\" Copyright (c) 1992, 1993, 1994
3 .\" The Regents of the University of California. All rights reserved.
5 .\" This code is derived from software contributed to Berkeley by
8 .\" Redistribution and use in source and binary forms, with or without
9 .\" modification, are permitted provided that the following conditions
11 .\" 1. Redistributions of source code must retain the above copyright
12 .\" notice, this list of conditions and the following disclaimer.
13 .\" 2. Redistributions in binary form must reproduce the above copyright
14 .\" notice, this list of conditions and the following disclaimer in the
15 .\" documentation and/or other materials provided with the distribution.
16 .\" 4. Neither the name of the University nor the names of its contributors
17 .\" may be used to endorse or promote products derived from this software
18 .\" without specific prior written permission.
20 .\" THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
21 .\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
22 .\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
23 .\" ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
24 .\" FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
25 .\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
26 .\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
27 .\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
28 .\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
29 .\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
32 .\" @(#)regex.3 8.4 (Berkeley) 3/20/94
43 .Nd regular-expression library
50 .Fa "regex_t * restrict preg" "const char * restrict pattern" "int cflags"
54 .Fa "const regex_t * restrict preg" "const char * restrict string"
55 .Fa "size_t nmatch" "regmatch_t pmatch[restrict]" "int eflags"
59 .Fa "int errcode" "const regex_t * restrict preg"
60 .Fa "char * restrict errbuf" "size_t errbuf_size"
63 .Fn regfree "regex_t *preg"
65 These routines implement
74 compiles an RE written as a string into an internal form,
76 matches that internal form against a string and reports results,
78 transforms error codes from either into human-readable messages,
81 frees any dynamically-allocated storage used by the internal form
86 declares two structure types,
90 the former for compiled internal forms and the latter for match reporting.
91 It also declares the four functions,
94 and a number of constants with names starting with
100 compiles the regular expression contained in the
103 subject to the flags in
105 and places the results in the
107 structure pointed to by
112 is the bitwise OR of zero or more of the following flags:
113 .Bl -tag -width REG_EXTENDED
118 rather than the obsolete
123 This is a synonym for 0,
124 provided as a counterpart to
126 to improve readability.
128 Compile with recognition of all special characters turned off.
129 All characters are thus considered ordinary,
133 This is an extension,
134 compatible with but not specified by
136 and should be used with
137 caution in software intended to be portable to other systems.
145 Compile for matching that ignores upper/lower case distinctions.
149 Compile for matching that need only report success or failure,
150 not what was matched.
152 Compile for newline-sensitive matching.
153 By default, newline is a completely ordinary character with no special
154 meaning in either REs or strings.
157 bracket expressions and
162 anchor matches the null string after any newline in the string
163 in addition to its normal function,
166 anchor matches the null string before any newline in the
167 string in addition to its normal function.
169 The regular expression ends,
170 not at the first NUL,
171 but just before the character pointed to by the
173 member of the structure pointed to by
179 This flag permits inclusion of NULs in the RE;
180 they are considered ordinary characters.
181 This is an extension,
182 compatible with but not specified by
184 and should be used with
185 caution in software intended to be portable to other systems.
190 returns 0 and fills in the structure pointed to by
192 One member of that structure
199 contains the number of parenthesized subexpressions within the RE
200 (except that the value of this member is undefined if the
205 fails, it returns a non-zero error code;
212 matches the compiled RE pointed to by
216 subject to the flags in
218 and reports results using
221 and the returned value.
222 The RE must have been compiled by a previous invocation of
224 The compiled form is not altered during execution of
226 so a single compiled RE can be used simultaneously by multiple threads.
229 the NUL-terminated string pointed to by
231 is considered to be the text of an entire line, minus any terminating
235 argument is the bitwise OR of zero or more of the following flags:
236 .Bl -tag -width REG_STARTEND
238 The first character of
240 is not the beginning of a line, so the
242 anchor should not match before it.
243 This does not affect the behavior of newlines under
248 does not end a line, so the
250 anchor should not match before it.
251 This does not affect the behavior of newlines under
254 The string is considered to start at
257 .Fa pmatch Ns [0]. Ns Va rm_so
258 and to have a terminating NUL located at
261 .Fa pmatch Ns [0]. Ns Va rm_eo
262 (there need not actually be a NUL at that location),
263 regardless of the value of
265 See below for the definition of
269 This is an extension,
270 compatible with but not specified by
272 and should be used with
273 caution in software intended to be portable to other systems.
279 affects only the location of the string,
280 not how it is matched.
285 for a discussion of what is matched in situations where an RE or a
286 portion thereof could match any of several substrings of
291 returns 0 for success and the non-zero code
294 Other non-zero error codes may be returned in exceptional situations;
300 was specified in the compilation of the RE,
307 argument (but see below for the case where
312 points to an array of
316 Such a structure has at least the members
322 (a signed arithmetic type at least as large as an
326 containing respectively the offset of the first character of a substring
327 and the offset of the first character after the end of the substring.
328 Offsets are measured from the beginning of the
332 An empty substring is denoted by equal offsets,
333 both indicating the character following the empty substring.
335 The 0th member of the
337 array is filled in to indicate what substring of
339 was matched by the entire RE.
340 Remaining members report what substring was matched by parenthesized
341 subexpressions within the RE;
344 reports subexpression
346 with subexpressions counted (starting at 1) by the order of their opening
347 parentheses in the RE, left to right.
348 Unused entries in the array (corresponding either to subexpressions that
349 did not participate in the match at all, or to subexpressions that do not
350 exist in the RE (that is,
353 .Fa preg Ns -> Ns Va re_nsub ) )
359 If a subexpression participated in the match several times,
360 the reported substring is the last one it matched.
361 (Note, as an example in particular, that when the RE
365 the parenthesized subexpression matches each of the three
368 an infinite number of empty strings following the last
370 so the reported substring is one of the empties.)
376 must point to at least one
383 to hold the input offsets for
385 Use for output is still entirely controlled by
394 will not be changed by a successful
406 to a human-readable, printable message.
410 .No non\- Ns Dv NULL ,
411 the error code should have arisen from use of
416 and if the error code came from
418 it should have been the result from the most recent
425 may be able to supply a more detailed message using information
432 places the NUL-terminated message into the buffer pointed to by
434 limiting the length (including the NUL) to at most
437 If the whole message will not fit,
438 as much of it as will fit before the terminating NUL is supplied.
440 the returned value is the size of buffer needed to hold the whole
441 message (including terminating NUL).
446 is ignored but the return value is still correct.
456 that results is the printable name of the error code,
459 rather than an explanation thereof.
470 member of the structure it points to
471 must point to the printable name of an error code;
472 in this case, the result in
474 is the decimal digits of
475 the numeric value of the error code
476 (0 if the name is not recognized).
480 are intended primarily as debugging facilities;
482 compatible with but not specified by
484 and should be used with
485 caution in software intended to be portable to other systems.
486 Be warned also that they are considered experimental and changes are possible.
491 frees any dynamically-allocated storage associated with the compiled RE
496 is no longer a valid compiled RE
497 and the effect of supplying it to
503 None of these functions references global variables except for tables
505 all are safe for use from multiple threads if the arguments are safe.
506 .Sh IMPLEMENTATION CHOICES
507 There are a number of decisions that
509 leaves up to the implementor,
510 either by explicitly saying
512 or by virtue of them being
513 forbidden by the RE grammar.
514 This implementation treats them as follows.
518 for a discussion of the definition of case-independent matching.
520 There is no particular limit on the length of REs,
521 except insofar as memory is limited.
522 Memory usage is approximately linear in RE size, and largely insensitive
523 to RE complexity, except for bounded repetitions.
526 for one short RE using them
527 that will run almost any system out of memory.
529 A backslashed character other than one specifically given a magic meaning
532 (such magic meanings occur only in obsolete
535 is taken as an ordinary character.
543 Equivalence classes cannot begin or end bracket-expression ranges.
544 The endpoint of one range cannot begin another.
547 the limit on repetition counts in bounded repetitions, is 255.
549 A repetition operator
554 cannot follow another
556 A repetition operator cannot begin an expression or subexpression
563 cannot appear first or last in a (sub)expression or after another
567 cannot be an empty subexpression.
568 An empty parenthesized subexpression,
570 is legal and matches an
572 An empty string is not a legal RE.
576 followed by a digit is considered the beginning of bounds for a
577 bounded repetition, which must then follow the syntax for bounds.
581 followed by a digit is considered an ordinary character.
586 beginning and ending subexpressions in obsolete
588 REs are anchors, not ordinary characters.
590 Non-zero error codes from
594 include the following:
596 .Bl -tag -width REG_ECOLLATE -compact
603 invalid regular expression
605 invalid collating element
607 invalid character class
610 applied to unescapable character
612 invalid backreference number
626 invalid repetition count(s) in
629 invalid character range in
640 empty (sub)expression
642 cannot happen - you found a bug
644 invalid argument, e.g.\& negative-length string
646 illegal byte sequence (bad multibyte character)
653 sections 2.8 (Regular Expression Notation)
655 B.5 (C Binding for Regular Expression Matching).
657 Originally written by
659 Altered for inclusion in the
663 This is an alpha release with known defects.
664 Please report problems.
666 The back-reference code is subtle and doubts linger about its correctness
673 This will improve with later releases.
677 exceeding 0 is expensive;
679 exceeding 1 is worse.
683 is largely insensitive to RE complexity
686 references are massively expensive.
687 RE length does matter; in particular, there is a strong speed bonus
688 for keeping RE length under about 30 characters,
689 with most special characters counting roughly double.
694 implements bounded repetitions by macro expansion,
695 which is costly in time and space if counts are large
696 or bounded repetitions are nested.
698 .Ql "((((a{1,100}){1,100}){1,100}){1,100}){1,100}"
699 will (eventually) run almost any existing machine out of swap space.
701 There are suspected problems with response to obscure error conditions.
703 certain kinds of internal overflow,
704 produced only by truly enormous REs or by multiply nested bounded repetitions,
705 are probably not handled well.
711 are legal REs because
714 a special character only in the presence of a previous unmatched
716 This cannot be fixed until the spec is fixed.
718 The standard's definition of back references is vague.
720 .Ql "a\e(\e(b\e)*\e2\e)*d"
723 Until the standard is clarified,
724 behavior in such cases should not be relied on.
726 The implementation of word-boundary matching is a bit of a kludge,
727 and bugs may lurk in combinations of word-boundary matching and anchoring.
729 Word-boundary matching does not work properly in multibyte locales.