/[pcre]/code/trunk/doc/pcre.txt
ViewVC logotype

Contents of /code/trunk/doc/pcre.txt

Parent Directory Parent Directory | Revision Log Revision Log


Revision 87 - (show annotations)
Sat Feb 24 21:41:21 2007 UTC (8 years, 4 months ago) by nigel
File MIME type: text/plain
File size: 231262 byte(s)
Error occurred while calculating annotation data.
Load pcre-6.5 into code/trunk.
1 -----------------------------------------------------------------------------
2 This file contains a concatenation of the PCRE man pages, converted to plain
3 text format for ease of searching with a text editor, or for use on systems
4 that do not have a man page processor. The small individual files that give
5 synopses of each function in the library have not been included. There are
6 separate text files for the pcregrep and pcretest commands.
7 -----------------------------------------------------------------------------
8
9
10 PCRE(3) PCRE(3)
11
12
13 NAME
14 PCRE - Perl-compatible regular expressions
15
16
17 INTRODUCTION
18
19 The PCRE library is a set of functions that implement regular expres-
20 sion pattern matching using the same syntax and semantics as Perl, with
21 just a few differences. The current implementation of PCRE (release
22 6.x) corresponds approximately with Perl 5.8, including support for
23 UTF-8 encoded strings and Unicode general category properties. However,
24 this support has to be explicitly enabled; it is not the default.
25
26 In addition to the Perl-compatible matching function, PCRE also con-
27 tains an alternative matching function that matches the same compiled
28 patterns in a different way. In certain circumstances, the alternative
29 function has some advantages. For a discussion of the two matching
30 algorithms, see the pcrematching page.
31
32 PCRE is written in C and released as a C library. A number of people
33 have written wrappers and interfaces of various kinds. In particular,
34 Google Inc. have provided a comprehensive C++ wrapper. This is now
35 included as part of the PCRE distribution. The pcrecpp page has details
36 of this interface. Other people's contributions can be found in the
37 Contrib directory at the primary FTP site, which is:
38
39 ftp://ftp.csx.cam.ac.uk/pub/software/programming/pcre
40
41 Details of exactly which Perl regular expression features are and are
42 not supported by PCRE are given in separate documents. See the pcrepat-
43 tern and pcrecompat pages.
44
45 Some features of PCRE can be included, excluded, or changed when the
46 library is built. The pcre_config() function makes it possible for a
47 client to discover which features are available. The features them-
48 selves are described in the pcrebuild page. Documentation about build-
49 ing PCRE for various operating systems can be found in the README file
50 in the source distribution.
51
52 The library contains a number of undocumented internal functions and
53 data tables that are used by more than one of the exported external
54 functions, but which are not intended for use by external callers.
55 Their names all begin with "_pcre_", which hopefully will not provoke
56 any name clashes. In some environments, it is possible to control which
57 external symbols are exported when a shared library is built, and in
58 these cases the undocumented symbols are not exported.
59
60
61 USER DOCUMENTATION
62
63 The user documentation for PCRE comprises a number of different sec-
64 tions. In the "man" format, each of these is a separate "man page". In
65 the HTML format, each is a separate page, linked from the index page.
66 In the plain text format, all the sections are concatenated, for ease
67 of searching. The sections are as follows:
68
69 pcre this document
70 pcreapi details of PCRE's native C API
71 pcrebuild options for building PCRE
72 pcrecallout details of the callout feature
73 pcrecompat discussion of Perl compatibility
74 pcrecpp details of the C++ wrapper
75 pcregrep description of the pcregrep command
76 pcrematching discussion of the two matching algorithms
77 pcrepartial details of the partial matching facility
78 pcrepattern syntax and semantics of supported
79 regular expressions
80 pcreperform discussion of performance issues
81 pcreposix the POSIX-compatible C API
82 pcreprecompile details of saving and re-using precompiled patterns
83 pcresample discussion of the sample program
84 pcretest description of the pcretest testing command
85
86 In addition, in the "man" and HTML formats, there is a short page for
87 each C library function, listing its arguments and results.
88
89
90 LIMITATIONS
91
92 There are some size limitations in PCRE but it is hoped that they will
93 never in practice be relevant.
94
95 The maximum length of a compiled pattern is 65539 (sic) bytes if PCRE
96 is compiled with the default internal linkage size of 2. If you want to
97 process regular expressions that are truly enormous, you can compile
98 PCRE with an internal linkage size of 3 or 4 (see the README file in
99 the source distribution and the pcrebuild documentation for details).
100 In these cases the limit is substantially larger. However, the speed
101 of execution will be slower.
102
103 All values in repeating quantifiers must be less than 65536. The maxi-
104 mum number of capturing subpatterns is 65535.
105
106 There is no limit to the number of non-capturing subpatterns, but the
107 maximum depth of nesting of all kinds of parenthesized subpattern,
108 including capturing subpatterns, assertions, and other types of subpat-
109 tern, is 200.
110
111 The maximum length of a subject string is the largest positive number
112 that an integer variable can hold. However, when using the traditional
113 matching function, PCRE uses recursion to handle subpatterns and indef-
114 inite repetition. This means that the available stack space may limit
115 the size of a subject string that can be processed by certain patterns.
116
117
118 UTF-8 AND UNICODE PROPERTY SUPPORT
119
120 From release 3.3, PCRE has had some support for character strings
121 encoded in the UTF-8 format. For release 4.0 this was greatly extended
122 to cover most common requirements, and in release 5.0 additional sup-
123 port for Unicode general category properties was added.
124
125 In order process UTF-8 strings, you must build PCRE to include UTF-8
126 support in the code, and, in addition, you must call pcre_compile()
127 with the PCRE_UTF8 option flag. When you do this, both the pattern and
128 any subject strings that are matched against it are treated as UTF-8
129 strings instead of just strings of bytes.
130
131 If you compile PCRE with UTF-8 support, but do not use it at run time,
132 the library will be a bit bigger, but the additional run time overhead
133 is limited to testing the PCRE_UTF8 flag in several places, so should
134 not be very large.
135
136 If PCRE is built with Unicode character property support (which implies
137 UTF-8 support), the escape sequences \p{..}, \P{..}, and \X are sup-
138 ported. The available properties that can be tested are limited to the
139 general category properties such as Lu for an upper case letter or Nd
140 for a decimal number, the Unicode script names such as Arabic or Han,
141 and the derived properties Any and L&. A full list is given in the
142 pcrepattern documentation. Only the short names for properties are sup-
143 ported. For example, \p{L} matches a letter. Its Perl synonym, \p{Let-
144 ter}, is not supported. Furthermore, in Perl, many properties may
145 optionally be prefixed by "Is", for compatibility with Perl 5.6. PCRE
146 does not support this.
147
148 The following comments apply when PCRE is running in UTF-8 mode:
149
150 1. When you set the PCRE_UTF8 flag, the strings passed as patterns and
151 subjects are checked for validity on entry to the relevant functions.
152 If an invalid UTF-8 string is passed, an error return is given. In some
153 situations, you may already know that your strings are valid, and
154 therefore want to skip these checks in order to improve performance. If
155 you set the PCRE_NO_UTF8_CHECK flag at compile time or at run time,
156 PCRE assumes that the pattern or subject it is given (respectively)
157 contains only valid UTF-8 codes. In this case, it does not diagnose an
158 invalid UTF-8 string. If you pass an invalid UTF-8 string to PCRE when
159 PCRE_NO_UTF8_CHECK is set, the results are undefined. Your program may
160 crash.
161
162 2. An unbraced hexadecimal escape sequence (such as \xb3) matches a
163 two-byte UTF-8 character if the value is greater than 127.
164
165 3. Repeat quantifiers apply to complete UTF-8 characters, not to indi-
166 vidual bytes, for example: \x{100}{3}.
167
168 4. The dot metacharacter matches one UTF-8 character instead of a sin-
169 gle byte.
170
171 5. The escape sequence \C can be used to match a single byte in UTF-8
172 mode, but its use can lead to some strange effects. This facility is
173 not available in the alternative matching function, pcre_dfa_exec().
174
175 6. The character escapes \b, \B, \d, \D, \s, \S, \w, and \W correctly
176 test characters of any code value, but the characters that PCRE recog-
177 nizes as digits, spaces, or word characters remain the same set as
178 before, all with values less than 256. This remains true even when PCRE
179 includes Unicode property support, because to do otherwise would slow
180 down PCRE in many common cases. If you really want to test for a wider
181 sense of, say, "digit", you must use Unicode property tests such as
182 \p{Nd}.
183
184 7. Similarly, characters that match the POSIX named character classes
185 are all low-valued characters.
186
187 8. Case-insensitive matching applies only to characters whose values
188 are less than 128, unless PCRE is built with Unicode property support.
189 Even when Unicode property support is available, PCRE still uses its
190 own character tables when checking the case of low-valued characters,
191 so as not to degrade performance. The Unicode property information is
192 used only for characters with higher values. Even when Unicode property
193 support is available, PCRE supports case-insensitive matching only when
194 there is a one-to-one mapping between a letter's cases. There are a
195 small number of many-to-one mappings in Unicode; these are not sup-
196 ported by PCRE.
197
198
199 AUTHOR
200
201 Philip Hazel
202 University Computing Service,
203 Cambridge CB2 3QG, England.
204
205 Putting an actual email address here seems to have been a spam magnet,
206 so I've taken it away. If you want to email me, use my initial and sur-
207 name, separated by a dot, at the domain ucs.cam.ac.uk.
208
209 Last updated: 24 January 2006
210 Copyright (c) 1997-2006 University of Cambridge.
211 ------------------------------------------------------------------------------
212
213
214 PCREBUILD(3) PCREBUILD(3)
215
216
217 NAME
218 PCRE - Perl-compatible regular expressions
219
220
221 PCRE BUILD-TIME OPTIONS
222
223 This document describes the optional features of PCRE that can be
224 selected when the library is compiled. They are all selected, or dese-
225 lected, by providing options to the configure script that is run before
226 the make command. The complete list of options for configure (which
227 includes the standard ones such as the selection of the installation
228 directory) can be obtained by running
229
230 ./configure --help
231
232 The following sections describe certain options whose names begin with
233 --enable or --disable. These settings specify changes to the defaults
234 for the configure command. Because of the way that configure works,
235 --enable and --disable always come in pairs, so the complementary
236 option always exists as well, but as it specifies the default, it is
237 not described.
238
239
240 C++ SUPPORT
241
242 By default, the configure script will search for a C++ compiler and C++
243 header files. If it finds them, it automatically builds the C++ wrapper
244 library for PCRE. You can disable this by adding
245
246 --disable-cpp
247
248 to the configure command.
249
250
251 UTF-8 SUPPORT
252
253 To build PCRE with support for UTF-8 character strings, add
254
255 --enable-utf8
256
257 to the configure command. Of itself, this does not make PCRE treat
258 strings as UTF-8. As well as compiling PCRE with this option, you also
259 have have to set the PCRE_UTF8 option when you call the pcre_compile()
260 function.
261
262
263 UNICODE CHARACTER PROPERTY SUPPORT
264
265 UTF-8 support allows PCRE to process character values greater than 255
266 in the strings that it handles. On its own, however, it does not pro-
267 vide any facilities for accessing the properties of such characters. If
268 you want to be able to use the pattern escapes \P, \p, and \X, which
269 refer to Unicode character properties, you must add
270
271 --enable-unicode-properties
272
273 to the configure command. This implies UTF-8 support, even if you have
274 not explicitly requested it.
275
276 Including Unicode property support adds around 90K of tables to the
277 PCRE library, approximately doubling its size. Only the general cate-
278 gory properties such as Lu and Nd are supported. Details are given in
279 the pcrepattern documentation.
280
281
282 CODE VALUE OF NEWLINE
283
284 By default, PCRE treats character 10 (linefeed) as the newline charac-
285 ter. This is the normal newline character on Unix-like systems. You can
286 compile PCRE to use character 13 (carriage return) instead by adding
287
288 --enable-newline-is-cr
289
290 to the configure command. For completeness there is also a --enable-
291 newline-is-lf option, which explicitly specifies linefeed as the new-
292 line character.
293
294
295 BUILDING SHARED AND STATIC LIBRARIES
296
297 The PCRE building process uses libtool to build both shared and static
298 Unix libraries by default. You can suppress one of these by adding one
299 of
300
301 --disable-shared
302 --disable-static
303
304 to the configure command, as required.
305
306
307 POSIX MALLOC USAGE
308
309 When PCRE is called through the POSIX interface (see the pcreposix doc-
310 umentation), additional working storage is required for holding the
311 pointers to capturing substrings, because PCRE requires three integers
312 per substring, whereas the POSIX interface provides only two. If the
313 number of expected substrings is small, the wrapper function uses space
314 on the stack, because this is faster than using malloc() for each call.
315 The default threshold above which the stack is no longer used is 10; it
316 can be changed by adding a setting such as
317
318 --with-posix-malloc-threshold=20
319
320 to the configure command.
321
322
323 LIMITING PCRE RESOURCE USAGE
324
325 Internally, PCRE has a function called match(), which it calls repeat-
326 edly (possibly recursively) when matching a pattern with the
327 pcre_exec() function. By controlling the maximum number of times this
328 function may be called during a single matching operation, a limit can
329 be placed on the resources used by a single call to pcre_exec(). The
330 limit can be changed at run time, as described in the pcreapi documen-
331 tation. The default is 10 million, but this can be changed by adding a
332 setting such as
333
334 --with-match-limit=500000
335
336 to the configure command. This setting has no effect on the
337 pcre_dfa_exec() matching function.
338
339
340 HANDLING VERY LARGE PATTERNS
341
342 Within a compiled pattern, offset values are used to point from one
343 part to another (for example, from an opening parenthesis to an alter-
344 nation metacharacter). By default, two-byte values are used for these
345 offsets, leading to a maximum size for a compiled pattern of around
346 64K. This is sufficient to handle all but the most gigantic patterns.
347 Nevertheless, some people do want to process enormous patterns, so it
348 is possible to compile PCRE to use three-byte or four-byte offsets by
349 adding a setting such as
350
351 --with-link-size=3
352
353 to the configure command. The value given must be 2, 3, or 4. Using
354 longer offsets slows down the operation of PCRE because it has to load
355 additional bytes when handling them.
356
357 If you build PCRE with an increased link size, test 2 (and test 5 if
358 you are using UTF-8) will fail. Part of the output of these tests is a
359 representation of the compiled pattern, and this changes with the link
360 size.
361
362
363 AVOIDING EXCESSIVE STACK USAGE
364
365 When matching with the pcre_exec() function, PCRE implements backtrack-
366 ing by making recursive calls to an internal function called match().
367 In environments where the size of the stack is limited, this can se-
368 verely limit PCRE's operation. (The Unix environment does not usually
369 suffer from this problem.) An alternative approach that uses memory
370 from the heap to remember data, instead of using recursive function
371 calls, has been implemented to work round this problem. If you want to
372 build a version of PCRE that works this way, add
373
374 --disable-stack-for-recursion
375
376 to the configure command. With this configuration, PCRE will use the
377 pcre_stack_malloc and pcre_stack_free variables to call memory manage-
378 ment functions. Separate functions are provided because the usage is
379 very predictable: the block sizes requested are always the same, and
380 the blocks are always freed in reverse order. A calling program might
381 be able to implement optimized functions that perform better than the
382 standard malloc() and free() functions. PCRE runs noticeably more
383 slowly when built in this way. This option affects only the pcre_exec()
384 function; it is not relevant for the the pcre_dfa_exec() function.
385
386
387 USING EBCDIC CODE
388
389 PCRE assumes by default that it will run in an environment where the
390 character code is ASCII (or Unicode, which is a superset of ASCII).
391 PCRE can, however, be compiled to run in an EBCDIC environment by
392 adding
393
394 --enable-ebcdic
395
396 to the configure command.
397
398 Last updated: 15 August 2005
399 Copyright (c) 1997-2005 University of Cambridge.
400 ------------------------------------------------------------------------------
401
402
403 PCREMATCHING(3) PCREMATCHING(3)
404
405
406 NAME
407 PCRE - Perl-compatible regular expressions
408
409
410 PCRE MATCHING ALGORITHMS
411
412 This document describes the two different algorithms that are available
413 in PCRE for matching a compiled regular expression against a given sub-
414 ject string. The "standard" algorithm is the one provided by the
415 pcre_exec() function. This works in the same was as Perl's matching
416 function, and provides a Perl-compatible matching operation.
417
418 An alternative algorithm is provided by the pcre_dfa_exec() function;
419 this operates in a different way, and is not Perl-compatible. It has
420 advantages and disadvantages compared with the standard algorithm, and
421 these are described below.
422
423 When there is only one possible way in which a given subject string can
424 match a pattern, the two algorithms give the same answer. A difference
425 arises, however, when there are multiple possibilities. For example, if
426 the pattern
427
428 ^<.*>
429
430 is matched against the string
431
432 <something> <something else> <something further>
433
434 there are three possible answers. The standard algorithm finds only one
435 of them, whereas the DFA algorithm finds all three.
436
437
438 REGULAR EXPRESSIONS AS TREES
439
440 The set of strings that are matched by a regular expression can be rep-
441 resented as a tree structure. An unlimited repetition in the pattern
442 makes the tree of infinite size, but it is still a tree. Matching the
443 pattern to a given subject string (from a given starting point) can be
444 thought of as a search of the tree. There are two standard ways to
445 search a tree: depth-first and breadth-first, and these correspond to
446 the two matching algorithms provided by PCRE.
447
448
449 THE STANDARD MATCHING ALGORITHM
450
451 In the terminology of Jeffrey Friedl's book Mastering Regular Expres-
452 sions, the standard algorithm is an "NFA algorithm". It conducts a
453 depth-first search of the pattern tree. That is, it proceeds along a
454 single path through the tree, checking that the subject matches what is
455 required. When there is a mismatch, the algorithm tries any alterna-
456 tives at the current point, and if they all fail, it backs up to the
457 previous branch point in the tree, and tries the next alternative
458 branch at that level. This often involves backing up (moving to the
459 left) in the subject string as well. The order in which repetition
460 branches are tried is controlled by the greedy or ungreedy nature of
461 the quantifier.
462
463 If a leaf node is reached, a matching string has been found, and at
464 that point the algorithm stops. Thus, if there is more than one possi-
465 ble match, this algorithm returns the first one that it finds. Whether
466 this is the shortest, the longest, or some intermediate length depends
467 on the way the greedy and ungreedy repetition quantifiers are specified
468 in the pattern.
469
470 Because it ends up with a single path through the tree, it is rela-
471 tively straightforward for this algorithm to keep track of the sub-
472 strings that are matched by portions of the pattern in parentheses.
473 This provides support for capturing parentheses and back references.
474
475
476 THE DFA MATCHING ALGORITHM
477
478 DFA stands for "deterministic finite automaton", but you do not need to
479 understand the origins of that name. This algorithm conducts a breadth-
480 first search of the tree. Starting from the first matching point in the
481 subject, it scans the subject string from left to right, once, charac-
482 ter by character, and as it does this, it remembers all the paths
483 through the tree that represent valid matches.
484
485 The scan continues until either the end of the subject is reached, or
486 there are no more unterminated paths. At this point, terminated paths
487 represent the different matching possibilities (if there are none, the
488 match has failed). Thus, if there is more than one possible match,
489 this algorithm finds all of them, and in particular, it finds the long-
490 est. In PCRE, there is an option to stop the algorithm after the first
491 match (which is necessarily the shortest) has been found.
492
493 Note that all the matches that are found start at the same point in the
494 subject. If the pattern
495
496 cat(er(pillar)?)
497
498 is matched against the string "the caterpillar catchment", the result
499 will be the three strings "cat", "cater", and "caterpillar" that start
500 at the fourth character of the subject. The algorithm does not automat-
501 ically move on to find matches that start at later positions.
502
503 There are a number of features of PCRE regular expressions that are not
504 supported by the DFA matching algorithm. They are as follows:
505
506 1. Because the algorithm finds all possible matches, the greedy or
507 ungreedy nature of repetition quantifiers is not relevant. Greedy and
508 ungreedy quantifiers are treated in exactly the same way.
509
510 2. When dealing with multiple paths through the tree simultaneously, it
511 is not straightforward to keep track of captured substrings for the
512 different matching possibilities, and PCRE's implementation of this
513 algorithm does not attempt to do this. This means that no captured sub-
514 strings are available.
515
516 3. Because no substrings are captured, back references within the pat-
517 tern are not supported, and cause errors if encountered.
518
519 4. For the same reason, conditional expressions that use a backrefer-
520 ence as the condition are not supported.
521
522 5. Callouts are supported, but the value of the capture_top field is
523 always 1, and the value of the capture_last field is always -1.
524
525 6. The \C escape sequence, which (in the standard algorithm) matches a
526 single byte, even in UTF-8 mode, is not supported because the DFA algo-
527 rithm moves through the subject string one character at a time, for all
528 active paths through the tree.
529
530
531 ADVANTAGES OF THE DFA ALGORITHM
532
533 Using the DFA matching algorithm provides the following advantages:
534
535 1. All possible matches (at a single point in the subject) are automat-
536 ically found, and in particular, the longest match is found. To find
537 more than one match using the standard algorithm, you have to do kludgy
538 things with callouts.
539
540 2. There is much better support for partial matching. The restrictions
541 on the content of the pattern that apply when using the standard algo-
542 rithm for partial matching do not apply to the DFA algorithm. For non-
543 anchored patterns, the starting position of a partial match is avail-
544 able.
545
546 3. Because the DFA algorithm scans the subject string just once, and
547 never needs to backtrack, it is possible to pass very long subject
548 strings to the matching function in several pieces, checking for par-
549 tial matching each time.
550
551
552 DISADVANTAGES OF THE DFA ALGORITHM
553
554 The DFA algorithm suffers from a number of disadvantages:
555
556 1. It is substantially slower than the standard algorithm. This is
557 partly because it has to search for all possible matches, but is also
558 because it is less susceptible to optimization.
559
560 2. Capturing parentheses and back references are not supported.
561
562 3. The "atomic group" feature of PCRE regular expressions is supported,
563 but does not provide the advantage that it does for the standard algo-
564 rithm.
565
566 Last updated: 28 February 2005
567 Copyright (c) 1997-2005 University of Cambridge.
568 ------------------------------------------------------------------------------
569
570
571 PCREAPI(3) PCREAPI(3)
572
573
574 NAME
575 PCRE - Perl-compatible regular expressions
576
577
578 PCRE NATIVE API
579
580 #include <pcre.h>
581
582 pcre *pcre_compile(const char *pattern, int options,
583 const char **errptr, int *erroffset,
584 const unsigned char *tableptr);
585
586 pcre *pcre_compile2(const char *pattern, int options,
587 int *errorcodeptr,
588 const char **errptr, int *erroffset,
589 const unsigned char *tableptr);
590
591 pcre_extra *pcre_study(const pcre *code, int options,
592 const char **errptr);
593
594 int pcre_exec(const pcre *code, const pcre_extra *extra,
595 const char *subject, int length, int startoffset,
596 int options, int *ovector, int ovecsize);
597
598 int pcre_dfa_exec(const pcre *code, const pcre_extra *extra,
599 const char *subject, int length, int startoffset,
600 int options, int *ovector, int ovecsize,
601 int *workspace, int wscount);
602
603 int pcre_copy_named_substring(const pcre *code,
604 const char *subject, int *ovector,
605 int stringcount, const char *stringname,
606 char *buffer, int buffersize);
607
608 int pcre_copy_substring(const char *subject, int *ovector,
609 int stringcount, int stringnumber, char *buffer,
610 int buffersize);
611
612 int pcre_get_named_substring(const pcre *code,
613 const char *subject, int *ovector,
614 int stringcount, const char *stringname,
615 const char **stringptr);
616
617 int pcre_get_stringnumber(const pcre *code,
618 const char *name);
619
620 int pcre_get_substring(const char *subject, int *ovector,
621 int stringcount, int stringnumber,
622 const char **stringptr);
623
624 int pcre_get_substring_list(const char *subject,
625 int *ovector, int stringcount, const char ***listptr);
626
627 void pcre_free_substring(const char *stringptr);
628
629 void pcre_free_substring_list(const char **stringptr);
630
631 const unsigned char *pcre_maketables(void);
632
633 int pcre_fullinfo(const pcre *code, const pcre_extra *extra,
634 int what, void *where);
635
636 int pcre_info(const pcre *code, int *optptr, int *firstcharptr);
637
638 int pcre_refcount(pcre *code, int adjust);
639
640 int pcre_config(int what, void *where);
641
642 char *pcre_version(void);
643
644 void *(*pcre_malloc)(size_t);
645
646 void (*pcre_free)(void *);
647
648 void *(*pcre_stack_malloc)(size_t);
649
650 void (*pcre_stack_free)(void *);
651
652 int (*pcre_callout)(pcre_callout_block *);
653
654
655 PCRE API OVERVIEW
656
657 PCRE has its own native API, which is described in this document. There
658 is also a set of wrapper functions that correspond to the POSIX regular
659 expression API. These are described in the pcreposix documentation.
660 Both of these APIs define a set of C function calls. A C++ wrapper is
661 distributed with PCRE. It is documented in the pcrecpp page.
662
663 The native API C function prototypes are defined in the header file
664 pcre.h, and on Unix systems the library itself is called libpcre. It
665 can normally be accessed by adding -lpcre to the command for linking an
666 application that uses PCRE. The header file defines the macros
667 PCRE_MAJOR and PCRE_MINOR to contain the major and minor release num-
668 bers for the library. Applications can use these to include support
669 for different releases of PCRE.
670
671 The functions pcre_compile(), pcre_compile2(), pcre_study(), and
672 pcre_exec() are used for compiling and matching regular expressions in
673 a Perl-compatible manner. A sample program that demonstrates the sim-
674 plest way of using them is provided in the file called pcredemo.c in
675 the source distribution. The pcresample documentation describes how to
676 run it.
677
678 A second matching function, pcre_dfa_exec(), which is not Perl-compati-
679 ble, is also provided. This uses a different algorithm for the match-
680 ing. This allows it to find all possible matches (at a given point in
681 the subject), not just one. However, this algorithm does not return
682 captured substrings. A description of the two matching algorithms and
683 their advantages and disadvantages is given in the pcrematching docu-
684 mentation.
685
686 In addition to the main compiling and matching functions, there are
687 convenience functions for extracting captured substrings from a subject
688 string that is matched by pcre_exec(). They are:
689
690 pcre_copy_substring()
691 pcre_copy_named_substring()
692 pcre_get_substring()
693 pcre_get_named_substring()
694 pcre_get_substring_list()
695 pcre_get_stringnumber()
696
697 pcre_free_substring() and pcre_free_substring_list() are also provided,
698 to free the memory used for extracted strings.
699
700 The function pcre_maketables() is used to build a set of character
701 tables in the current locale for passing to pcre_compile(),
702 pcre_exec(), or pcre_dfa_exec(). This is an optional facility that is
703 provided for specialist use. Most commonly, no special tables are
704 passed, in which case internal tables that are generated when PCRE is
705 built are used.
706
707 The function pcre_fullinfo() is used to find out information about a
708 compiled pattern; pcre_info() is an obsolete version that returns only
709 some of the available information, but is retained for backwards com-
710 patibility. The function pcre_version() returns a pointer to a string
711 containing the version of PCRE and its date of release.
712
713 The function pcre_refcount() maintains a reference count in a data
714 block containing a compiled pattern. This is provided for the benefit
715 of object-oriented applications.
716
717 The global variables pcre_malloc and pcre_free initially contain the
718 entry points of the standard malloc() and free() functions, respec-
719 tively. PCRE calls the memory management functions via these variables,
720 so a calling program can replace them if it wishes to intercept the
721 calls. This should be done before calling any PCRE functions.
722
723 The global variables pcre_stack_malloc and pcre_stack_free are also
724 indirections to memory management functions. These special functions
725 are used only when PCRE is compiled to use the heap for remembering
726 data, instead of recursive function calls, when running the pcre_exec()
727 function. This is a non-standard way of building PCRE, for use in envi-
728 ronments that have limited stacks. Because of the greater use of memory
729 management, it runs more slowly. Separate functions are provided so
730 that special-purpose external code can be used for this case. When
731 used, these functions are always called in a stack-like manner (last
732 obtained, first freed), and always for memory blocks of the same size.
733
734 The global variable pcre_callout initially contains NULL. It can be set
735 by the caller to a "callout" function, which PCRE will then call at
736 specified points during a matching operation. Details are given in the
737 pcrecallout documentation.
738
739
740 MULTITHREADING
741
742 The PCRE functions can be used in multi-threading applications, with
743 the proviso that the memory management functions pointed to by
744 pcre_malloc, pcre_free, pcre_stack_malloc, and pcre_stack_free, and the
745 callout function pointed to by pcre_callout, are shared by all threads.
746
747 The compiled form of a regular expression is not altered during match-
748 ing, so the same compiled pattern can safely be used by several threads
749 at once.
750
751
752 SAVING PRECOMPILED PATTERNS FOR LATER USE
753
754 The compiled form of a regular expression can be saved and re-used at a
755 later time, possibly by a different program, and even on a host other
756 than the one on which it was compiled. Details are given in the
757 pcreprecompile documentation.
758
759
760 CHECKING BUILD-TIME OPTIONS
761
762 int pcre_config(int what, void *where);
763
764 The function pcre_config() makes it possible for a PCRE client to dis-
765 cover which optional features have been compiled into the PCRE library.
766 The pcrebuild documentation has more details about these optional fea-
767 tures.
768
769 The first argument for pcre_config() is an integer, specifying which
770 information is required; the second argument is a pointer to a variable
771 into which the information is placed. The following information is
772 available:
773
774 PCRE_CONFIG_UTF8
775
776 The output is an integer that is set to one if UTF-8 support is avail-
777 able; otherwise it is set to zero.
778
779 PCRE_CONFIG_UNICODE_PROPERTIES
780
781 The output is an integer that is set to one if support for Unicode
782 character properties is available; otherwise it is set to zero.
783
784 PCRE_CONFIG_NEWLINE
785
786 The output is an integer that is set to the value of the code that is
787 used for the newline character. It is either linefeed (10) or carriage
788 return (13), and should normally be the standard character for your
789 operating system.
790
791 PCRE_CONFIG_LINK_SIZE
792
793 The output is an integer that contains the number of bytes used for
794 internal linkage in compiled regular expressions. The value is 2, 3, or
795 4. Larger values allow larger regular expressions to be compiled, at
796 the expense of slower matching. The default value of 2 is sufficient
797 for all but the most massive patterns, since it allows the compiled
798 pattern to be up to 64K in size.
799
800 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
801
802 The output is an integer that contains the threshold above which the
803 POSIX interface uses malloc() for output vectors. Further details are
804 given in the pcreposix documentation.
805
806 PCRE_CONFIG_MATCH_LIMIT
807
808 The output is an integer that gives the default limit for the number of
809 internal matching function calls in a pcre_exec() execution. Further
810 details are given with pcre_exec() below.
811
812 PCRE_CONFIG_MATCH_LIMIT_RECURSION
813
814 The output is an integer that gives the default limit for the depth of
815 recursion when calling the internal matching function in a pcre_exec()
816 execution. Further details are given with pcre_exec() below.
817
818 PCRE_CONFIG_STACKRECURSE
819
820 The output is an integer that is set to one if internal recursion when
821 running pcre_exec() is implemented by recursive function calls that use
822 the stack to remember their state. This is the usual way that PCRE is
823 compiled. The output is zero if PCRE was compiled to use blocks of data
824 on the heap instead of recursive function calls. In this case,
825 pcre_stack_malloc and pcre_stack_free are called to manage memory
826 blocks on the heap, thus avoiding the use of the stack.
827
828
829 COMPILING A PATTERN
830
831 pcre *pcre_compile(const char *pattern, int options,
832 const char **errptr, int *erroffset,
833 const unsigned char *tableptr);
834
835 pcre *pcre_compile2(const char *pattern, int options,
836 int *errorcodeptr,
837 const char **errptr, int *erroffset,
838 const unsigned char *tableptr);
839
840 Either of the functions pcre_compile() or pcre_compile2() can be called
841 to compile a pattern into an internal form. The only difference between
842 the two interfaces is that pcre_compile2() has an additional argument,
843 errorcodeptr, via which a numerical error code can be returned.
844
845 The pattern is a C string terminated by a binary zero, and is passed in
846 the pattern argument. A pointer to a single block of memory that is
847 obtained via pcre_malloc is returned. This contains the compiled code
848 and related data. The pcre type is defined for the returned block; this
849 is a typedef for a structure whose contents are not externally defined.
850 It is up to the caller to free the memory when it is no longer
851 required.
852
853 Although the compiled code of a PCRE regex is relocatable, that is, it
854 does not depend on memory location, the complete pcre data block is not
855 fully relocatable, because it may contain a copy of the tableptr argu-
856 ment, which is an address (see below).
857
858 The options argument contains independent bits that affect the compila-
859 tion. It should be zero if no options are required. The available
860 options are described below. Some of them, in particular, those that
861 are compatible with Perl, can also be set and unset from within the
862 pattern (see the detailed description in the pcrepattern documenta-
863 tion). For these options, the contents of the options argument speci-
864 fies their initial settings at the start of compilation and execution.
865 The PCRE_ANCHORED option can be set at the time of matching as well as
866 at compile time.
867
868 If errptr is NULL, pcre_compile() returns NULL immediately. Otherwise,
869 if compilation of a pattern fails, pcre_compile() returns NULL, and
870 sets the variable pointed to by errptr to point to a textual error mes-
871 sage. This is a static string that is part of the library. You must not
872 try to free it. The offset from the start of the pattern to the charac-
873 ter where the error was discovered is placed in the variable pointed to
874 by erroffset, which must not be NULL. If it is, an immediate error is
875 given.
876
877 If pcre_compile2() is used instead of pcre_compile(), and the error-
878 codeptr argument is not NULL, a non-zero error code number is returned
879 via this argument in the event of an error. This is in addition to the
880 textual error message. Error codes and messages are listed below.
881
882 If the final argument, tableptr, is NULL, PCRE uses a default set of
883 character tables that are built when PCRE is compiled, using the
884 default C locale. Otherwise, tableptr must be an address that is the
885 result of a call to pcre_maketables(). This value is stored with the
886 compiled pattern, and used again by pcre_exec(), unless another table
887 pointer is passed to it. For more discussion, see the section on locale
888 support below.
889
890 This code fragment shows a typical straightforward call to pcre_com-
891 pile():
892
893 pcre *re;
894 const char *error;
895 int erroffset;
896 re = pcre_compile(
897 "^A.*Z", /* the pattern */
898 0, /* default options */
899 &error, /* for error message */
900 &erroffset, /* for error offset */
901 NULL); /* use default character tables */
902
903 The following names for option bits are defined in the pcre.h header
904 file:
905
906 PCRE_ANCHORED
907
908 If this bit is set, the pattern is forced to be "anchored", that is, it
909 is constrained to match only at the first matching point in the string
910 that is being searched (the "subject string"). This effect can also be
911 achieved by appropriate constructs in the pattern itself, which is the
912 only way to do it in Perl.
913
914 PCRE_AUTO_CALLOUT
915
916 If this bit is set, pcre_compile() automatically inserts callout items,
917 all with number 255, before each pattern item. For discussion of the
918 callout facility, see the pcrecallout documentation.
919
920 PCRE_CASELESS
921
922 If this bit is set, letters in the pattern match both upper and lower
923 case letters. It is equivalent to Perl's /i option, and it can be
924 changed within a pattern by a (?i) option setting. In UTF-8 mode, PCRE
925 always understands the concept of case for characters whose values are
926 less than 128, so caseless matching is always possible. For characters
927 with higher values, the concept of case is supported if PCRE is com-
928 piled with Unicode property support, but not otherwise. If you want to
929 use caseless matching for characters 128 and above, you must ensure
930 that PCRE is compiled with Unicode property support as well as with
931 UTF-8 support.
932
933 PCRE_DOLLAR_ENDONLY
934
935 If this bit is set, a dollar metacharacter in the pattern matches only
936 at the end of the subject string. Without this option, a dollar also
937 matches immediately before the final character if it is a newline (but
938 not before any other newlines). The PCRE_DOLLAR_ENDONLY option is
939 ignored if PCRE_MULTILINE is set. There is no equivalent to this option
940 in Perl, and no way to set it within a pattern.
941
942 PCRE_DOTALL
943
944 If this bit is set, a dot metacharater in the pattern matches all char-
945 acters, including newlines. Without it, newlines are excluded. This
946 option is equivalent to Perl's /s option, and it can be changed within
947 a pattern by a (?s) option setting. A negative class such as [^a]
948 always matches a newline character, independent of the setting of this
949 option.
950
951 PCRE_EXTENDED
952
953 If this bit is set, whitespace data characters in the pattern are
954 totally ignored except when escaped or inside a character class. White-
955 space does not include the VT character (code 11). In addition, charac-
956 ters between an unescaped # outside a character class and the next new-
957 line character, inclusive, are also ignored. This is equivalent to
958 Perl's /x option, and it can be changed within a pattern by a (?x)
959 option setting.
960
961 This option makes it possible to include comments inside complicated
962 patterns. Note, however, that this applies only to data characters.
963 Whitespace characters may never appear within special character
964 sequences in a pattern, for example within the sequence (?( which
965 introduces a conditional subpattern.
966
967 PCRE_EXTRA
968
969 This option was invented in order to turn on additional functionality
970 of PCRE that is incompatible with Perl, but it is currently of very
971 little use. When set, any backslash in a pattern that is followed by a
972 letter that has no special meaning causes an error, thus reserving
973 these combinations for future expansion. By default, as in Perl, a
974 backslash followed by a letter with no special meaning is treated as a
975 literal. There are at present no other features controlled by this
976 option. It can also be set by a (?X) option setting within a pattern.
977
978 PCRE_FIRSTLINE
979
980 If this option is set, an unanchored pattern is required to match
981 before or at the first newline character in the subject string, though
982 the matched text may continue over the newline.
983
984 PCRE_MULTILINE
985
986 By default, PCRE treats the subject string as consisting of a single
987 line of characters (even if it actually contains newlines). The "start
988 of line" metacharacter (^) matches only at the start of the string,
989 while the "end of line" metacharacter ($) matches only at the end of
990 the string, or before a terminating newline (unless PCRE_DOLLAR_ENDONLY
991 is set). This is the same as Perl.
992
993 When PCRE_MULTILINE it is set, the "start of line" and "end of line"
994 constructs match immediately following or immediately before any new-
995 line in the subject string, respectively, as well as at the very start
996 and end. This is equivalent to Perl's /m option, and it can be changed
997 within a pattern by a (?m) option setting. If there are no "\n" charac-
998 ters in a subject string, or no occurrences of ^ or $ in a pattern,
999 setting PCRE_MULTILINE has no effect.
1000
1001 PCRE_NO_AUTO_CAPTURE
1002
1003 If this option is set, it disables the use of numbered capturing paren-
1004 theses in the pattern. Any opening parenthesis that is not followed by
1005 ? behaves as if it were followed by ?: but named parentheses can still
1006 be used for capturing (and they acquire numbers in the usual way).
1007 There is no equivalent of this option in Perl.
1008
1009 PCRE_UNGREEDY
1010
1011 This option inverts the "greediness" of the quantifiers so that they
1012 are not greedy by default, but become greedy if followed by "?". It is
1013 not compatible with Perl. It can also be set by a (?U) option setting
1014 within the pattern.
1015
1016 PCRE_UTF8
1017
1018 This option causes PCRE to regard both the pattern and the subject as
1019 strings of UTF-8 characters instead of single-byte character strings.
1020 However, it is available only when PCRE is built to include UTF-8 sup-
1021 port. If not, the use of this option provokes an error. Details of how
1022 this option changes the behaviour of PCRE are given in the section on
1023 UTF-8 support in the main pcre page.
1024
1025 PCRE_NO_UTF8_CHECK
1026
1027 When PCRE_UTF8 is set, the validity of the pattern as a UTF-8 string is
1028 automatically checked. If an invalid UTF-8 sequence of bytes is found,
1029 pcre_compile() returns an error. If you already know that your pattern
1030 is valid, and you want to skip this check for performance reasons, you
1031 can set the PCRE_NO_UTF8_CHECK option. When it is set, the effect of
1032 passing an invalid UTF-8 string as a pattern is undefined. It may cause
1033 your program to crash. Note that this option can also be passed to
1034 pcre_exec() and pcre_dfa_exec(), to suppress the UTF-8 validity check-
1035 ing of subject strings.
1036
1037
1038 COMPILATION ERROR CODES
1039
1040 The following table lists the error codes than may be returned by
1041 pcre_compile2(), along with the error messages that may be returned by
1042 both compiling functions.
1043
1044 0 no error
1045 1 \ at end of pattern
1046 2 \c at end of pattern
1047 3 unrecognized character follows \
1048 4 numbers out of order in {} quantifier
1049 5 number too big in {} quantifier
1050 6 missing terminating ] for character class
1051 7 invalid escape sequence in character class
1052 8 range out of order in character class
1053 9 nothing to repeat
1054 10 operand of unlimited repeat could match the empty string
1055 11 internal error: unexpected repeat
1056 12 unrecognized character after (?
1057 13 POSIX named classes are supported only within a class
1058 14 missing )
1059 15 reference to non-existent subpattern
1060 16 erroffset passed as NULL
1061 17 unknown option bit(s) set
1062 18 missing ) after comment
1063 19 parentheses nested too deeply
1064 20 regular expression too large
1065 21 failed to get memory
1066 22 unmatched parentheses
1067 23 internal error: code overflow
1068 24 unrecognized character after (?<
1069 25 lookbehind assertion is not fixed length
1070 26 malformed number after (?(
1071 27 conditional group contains more than two branches
1072 28 assertion expected after (?(
1073 29 (?R or (?digits must be followed by )
1074 30 unknown POSIX class name
1075 31 POSIX collating elements are not supported
1076 32 this version of PCRE is not compiled with PCRE_UTF8 support
1077 33 spare error
1078 34 character value in \x{...} sequence is too large
1079 35 invalid condition (?(0)
1080 36 \C not allowed in lookbehind assertion
1081 37 PCRE does not support \L, \l, \N, \U, or \u
1082 38 number after (?C is > 255
1083 39 closing ) for (?C expected
1084 40 recursive call could loop indefinitely
1085 41 unrecognized character after (?P
1086 42 syntax error after (?P
1087 43 two named groups have the same name
1088 44 invalid UTF-8 string
1089 45 support for \P, \p, and \X has not been compiled
1090 46 malformed \P or \p sequence
1091 47 unknown property name after \P or \p
1092
1093
1094 STUDYING A PATTERN
1095
1096 pcre_extra *pcre_study(const pcre *code, int options
1097 const char **errptr);
1098
1099 If a compiled pattern is going to be used several times, it is worth
1100 spending more time analyzing it in order to speed up the time taken for
1101 matching. The function pcre_study() takes a pointer to a compiled pat-
1102 tern as its first argument. If studying the pattern produces additional
1103 information that will help speed up matching, pcre_study() returns a
1104 pointer to a pcre_extra block, in which the study_data field points to
1105 the results of the study.
1106
1107 The returned value from pcre_study() can be passed directly to
1108 pcre_exec(). However, a pcre_extra block also contains other fields
1109 that can be set by the caller before the block is passed; these are
1110 described below in the section on matching a pattern.
1111
1112 If studying the pattern does not produce any additional information
1113 pcre_study() returns NULL. In that circumstance, if the calling program
1114 wants to pass any of the other fields to pcre_exec(), it must set up
1115 its own pcre_extra block.
1116
1117 The second argument of pcre_study() contains option bits. At present,
1118 no options are defined, and this argument should always be zero.
1119
1120 The third argument for pcre_study() is a pointer for an error message.
1121 If studying succeeds (even if no data is returned), the variable it
1122 points to is set to NULL. Otherwise it is set to point to a textual
1123 error message. This is a static string that is part of the library. You
1124 must not try to free it. You should test the error pointer for NULL
1125 after calling pcre_study(), to be sure that it has run successfully.
1126
1127 This is a typical call to pcre_study():
1128
1129 pcre_extra *pe;
1130 pe = pcre_study(
1131 re, /* result of pcre_compile() */
1132 0, /* no options exist */
1133 &error); /* set to NULL or points to a message */
1134
1135 At present, studying a pattern is useful only for non-anchored patterns
1136 that do not have a single fixed starting character. A bitmap of possi-
1137 ble starting bytes is created.
1138
1139
1140 LOCALE SUPPORT
1141
1142 PCRE handles caseless matching, and determines whether characters are
1143 letters digits, or whatever, by reference to a set of tables, indexed
1144 by character value. When running in UTF-8 mode, this applies only to
1145 characters with codes less than 128. Higher-valued codes never match
1146 escapes such as \w or \d, but can be tested with \p if PCRE is built
1147 with Unicode character property support. The use of locales with Uni-
1148 code is discouraged.
1149
1150 An internal set of tables is created in the default C locale when PCRE
1151 is built. This is used when the final argument of pcre_compile() is
1152 NULL, and is sufficient for many applications. An alternative set of
1153 tables can, however, be supplied. These may be created in a different
1154 locale from the default. As more and more applications change to using
1155 Unicode, the need for this locale support is expected to die away.
1156
1157 External tables are built by calling the pcre_maketables() function,
1158 which has no arguments, in the relevant locale. The result can then be
1159 passed to pcre_compile() or pcre_exec() as often as necessary. For
1160 example, to build and use tables that are appropriate for the French
1161 locale (where accented characters with values greater than 128 are
1162 treated as letters), the following code could be used:
1163
1164 setlocale(LC_CTYPE, "fr_FR");
1165 tables = pcre_maketables();
1166 re = pcre_compile(..., tables);
1167
1168 When pcre_maketables() runs, the tables are built in memory that is
1169 obtained via pcre_malloc. It is the caller's responsibility to ensure
1170 that the memory containing the tables remains available for as long as
1171 it is needed.
1172
1173 The pointer that is passed to pcre_compile() is saved with the compiled
1174 pattern, and the same tables are used via this pointer by pcre_study()
1175 and normally also by pcre_exec(). Thus, by default, for any single pat-
1176 tern, compilation, studying and matching all happen in the same locale,
1177 but different patterns can be compiled in different locales.
1178
1179 It is possible to pass a table pointer or NULL (indicating the use of
1180 the internal tables) to pcre_exec(). Although not intended for this
1181 purpose, this facility could be used to match a pattern in a different
1182 locale from the one in which it was compiled. Passing table pointers at
1183 run time is discussed below in the section on matching a pattern.
1184
1185
1186 INFORMATION ABOUT A PATTERN
1187
1188 int pcre_fullinfo(const pcre *code, const pcre_extra *extra,
1189 int what, void *where);
1190
1191 The pcre_fullinfo() function returns information about a compiled pat-
1192 tern. It replaces the obsolete pcre_info() function, which is neverthe-
1193 less retained for backwards compability (and is documented below).
1194
1195 The first argument for pcre_fullinfo() is a pointer to the compiled
1196 pattern. The second argument is the result of pcre_study(), or NULL if
1197 the pattern was not studied. The third argument specifies which piece
1198 of information is required, and the fourth argument is a pointer to a
1199 variable to receive the data. The yield of the function is zero for
1200 success, or one of the following negative numbers:
1201
1202 PCRE_ERROR_NULL the argument code was NULL
1203 the argument where was NULL
1204 PCRE_ERROR_BADMAGIC the "magic number" was not found
1205 PCRE_ERROR_BADOPTION the value of what was invalid
1206
1207 The "magic number" is placed at the start of each compiled pattern as
1208 an simple check against passing an arbitrary memory pointer. Here is a
1209 typical call of pcre_fullinfo(), to obtain the length of the compiled
1210 pattern:
1211
1212 int rc;
1213 unsigned long int length;
1214 rc = pcre_fullinfo(
1215 re, /* result of pcre_compile() */
1216 pe, /* result of pcre_study(), or NULL */
1217 PCRE_INFO_SIZE, /* what is required */
1218 &length); /* where to put the data */
1219
1220 The possible values for the third argument are defined in pcre.h, and
1221 are as follows:
1222
1223 PCRE_INFO_BACKREFMAX
1224
1225 Return the number of the highest back reference in the pattern. The
1226 fourth argument should point to an int variable. Zero is returned if
1227 there are no back references.
1228
1229 PCRE_INFO_CAPTURECOUNT
1230
1231 Return the number of capturing subpatterns in the pattern. The fourth
1232 argument should point to an int variable.
1233
1234 PCRE_INFO_DEFAULT_TABLES
1235
1236 Return a pointer to the internal default character tables within PCRE.
1237 The fourth argument should point to an unsigned char * variable. This
1238 information call is provided for internal use by the pcre_study() func-
1239 tion. External callers can cause PCRE to use its internal tables by
1240 passing a NULL table pointer.
1241
1242 PCRE_INFO_FIRSTBYTE
1243
1244 Return information about the first byte of any matched string, for a
1245 non-anchored pattern. (This option used to be called
1246 PCRE_INFO_FIRSTCHAR; the old name is still recognized for backwards
1247 compatibility.)
1248
1249 If there is a fixed first byte, for example, from a pattern such as
1250 (cat|cow|coyote), it is returned in the integer pointed to by where.
1251 Otherwise, if either
1252
1253 (a) the pattern was compiled with the PCRE_MULTILINE option, and every
1254 branch starts with "^", or
1255
1256 (b) every branch of the pattern starts with ".*" and PCRE_DOTALL is not
1257 set (if it were set, the pattern would be anchored),
1258
1259 -1 is returned, indicating that the pattern matches only at the start
1260 of a subject string or after any newline within the string. Otherwise
1261 -2 is returned. For anchored patterns, -2 is returned.
1262
1263 PCRE_INFO_FIRSTTABLE
1264
1265 If the pattern was studied, and this resulted in the construction of a
1266 256-bit table indicating a fixed set of bytes for the first byte in any
1267 matching string, a pointer to the table is returned. Otherwise NULL is
1268 returned. The fourth argument should point to an unsigned char * vari-
1269 able.
1270
1271 PCRE_INFO_LASTLITERAL
1272
1273 Return the value of the rightmost literal byte that must exist in any
1274 matched string, other than at its start, if such a byte has been
1275 recorded. The fourth argument should point to an int variable. If there
1276 is no such byte, -1 is returned. For anchored patterns, a last literal
1277 byte is recorded only if it follows something of variable length. For
1278 example, for the pattern /^a\d+z\d+/ the returned value is "z", but for
1279 /^a\dz\d/ the returned value is -1.
1280
1281 PCRE_INFO_NAMECOUNT
1282 PCRE_INFO_NAMEENTRYSIZE
1283 PCRE_INFO_NAMETABLE
1284
1285 PCRE supports the use of named as well as numbered capturing parenthe-
1286 ses. The names are just an additional way of identifying the parenthe-
1287 ses, which still acquire numbers. A convenience function called
1288 pcre_get_named_substring() is provided for extracting an individual
1289 captured substring by name. It is also possible to extract the data
1290 directly, by first converting the name to a number in order to access
1291 the correct pointers in the output vector (described with pcre_exec()
1292 below). To do the conversion, you need to use the name-to-number map,
1293 which is described by these three values.
1294
1295 The map consists of a number of fixed-size entries. PCRE_INFO_NAMECOUNT
1296 gives the number of entries, and PCRE_INFO_NAMEENTRYSIZE gives the size
1297 of each entry; both of these return an int value. The entry size
1298 depends on the length of the longest name. PCRE_INFO_NAMETABLE returns
1299 a pointer to the first entry of the table (a pointer to char). The
1300 first two bytes of each entry are the number of the capturing parenthe-
1301 sis, most significant byte first. The rest of the entry is the corre-
1302 sponding name, zero terminated. The names are in alphabetical order.
1303 For example, consider the following pattern (assume PCRE_EXTENDED is
1304 set, so white space - including newlines - is ignored):
1305
1306 (?P<date> (?P<year>(\d\d)?\d\d) -
1307 (?P<month>\d\d) - (?P<day>\d\d) )
1308
1309 There are four named subpatterns, so the table has four entries, and
1310 each entry in the table is eight bytes long. The table is as follows,
1311 with non-printing bytes shows in hexadecimal, and undefined bytes shown
1312 as ??:
1313
1314 00 01 d a t e 00 ??
1315 00 05 d a y 00 ?? ??
1316 00 04 m o n t h 00
1317 00 02 y e a r 00 ??
1318
1319 When writing code to extract data from named subpatterns using the
1320 name-to-number map, remember that the length of each entry is likely to
1321 be different for each compiled pattern.
1322
1323 PCRE_INFO_OPTIONS
1324
1325 Return a copy of the options with which the pattern was compiled. The
1326 fourth argument should point to an unsigned long int variable. These
1327 option bits are those specified in the call to pcre_compile(), modified
1328 by any top-level option settings within the pattern itself.
1329
1330 A pattern is automatically anchored by PCRE if all of its top-level
1331 alternatives begin with one of the following:
1332
1333 ^ unless PCRE_MULTILINE is set
1334 \A always
1335 \G always
1336 .* if PCRE_DOTALL is set and there are no back
1337 references to the subpattern in which .* appears
1338
1339 For such patterns, the PCRE_ANCHORED bit is set in the options returned
1340 by pcre_fullinfo().
1341
1342 PCRE_INFO_SIZE
1343
1344 Return the size of the compiled pattern, that is, the value that was
1345 passed as the argument to pcre_malloc() when PCRE was getting memory in
1346 which to place the compiled data. The fourth argument should point to a
1347 size_t variable.
1348
1349 PCRE_INFO_STUDYSIZE
1350
1351 Return the size of the data block pointed to by the study_data field in
1352 a pcre_extra block. That is, it is the value that was passed to
1353 pcre_malloc() when PCRE was getting memory into which to place the data
1354 created by pcre_study(). The fourth argument should point to a size_t
1355 variable.
1356
1357
1358 OBSOLETE INFO FUNCTION
1359
1360 int pcre_info(const pcre *code, int *optptr, int *firstcharptr);
1361
1362 The pcre_info() function is now obsolete because its interface is too
1363 restrictive to return all the available data about a compiled pattern.
1364 New programs should use pcre_fullinfo() instead. The yield of
1365 pcre_info() is the number of capturing subpatterns, or one of the fol-
1366 lowing negative numbers:
1367
1368 PCRE_ERROR_NULL the argument code was NULL
1369 PCRE_ERROR_BADMAGIC the "magic number" was not found
1370
1371 If the optptr argument is not NULL, a copy of the options with which
1372 the pattern was compiled is placed in the integer it points to (see
1373 PCRE_INFO_OPTIONS above).
1374
1375 If the pattern is not anchored and the firstcharptr argument is not
1376 NULL, it is used to pass back information about the first character of
1377 any matched string (see PCRE_INFO_FIRSTBYTE above).
1378
1379
1380 REFERENCE COUNTS
1381
1382 int pcre_refcount(pcre *code, int adjust);
1383
1384 The pcre_refcount() function is used to maintain a reference count in
1385 the data block that contains a compiled pattern. It is provided for the
1386 benefit of applications that operate in an object-oriented manner,
1387 where different parts of the application may be using the same compiled
1388 pattern, but you want to free the block when they are all done.
1389
1390 When a pattern is compiled, the reference count field is initialized to
1391 zero. It is changed only by calling this function, whose action is to
1392 add the adjust value (which may be positive or negative) to it. The
1393 yield of the function is the new value. However, the value of the count
1394 is constrained to lie between 0 and 65535, inclusive. If the new value
1395 is outside these limits, it is forced to the appropriate limit value.
1396
1397 Except when it is zero, the reference count is not correctly preserved
1398 if a pattern is compiled on one host and then transferred to a host
1399 whose byte-order is different. (This seems a highly unlikely scenario.)
1400
1401
1402 MATCHING A PATTERN: THE TRADITIONAL FUNCTION
1403
1404 int pcre_exec(const pcre *code, const pcre_extra *extra,
1405 const char *subject, int length, int startoffset,
1406 int options, int *ovector, int ovecsize);
1407
1408 The function pcre_exec() is called to match a subject string against a
1409 compiled pattern, which is passed in the code argument. If the pattern
1410 has been studied, the result of the study should be passed in the extra
1411 argument. This function is the main matching facility of the library,
1412 and it operates in a Perl-like manner. For specialist use there is also
1413 an alternative matching function, which is described below in the sec-
1414 tion about the pcre_dfa_exec() function.
1415
1416 In most applications, the pattern will have been compiled (and option-
1417 ally studied) in the same process that calls pcre_exec(). However, it
1418 is possible to save compiled patterns and study data, and then use them
1419 later in different processes, possibly even on different hosts. For a
1420 discussion about this, see the pcreprecompile documentation.
1421
1422 Here is an example of a simple call to pcre_exec():
1423
1424 int rc;
1425 int ovector[30];
1426 rc = pcre_exec(
1427 re, /* result of pcre_compile() */
1428 NULL, /* we didn't study the pattern */
1429 "some string", /* the subject string */
1430 11, /* the length of the subject string */
1431 0, /* start at offset 0 in the subject */
1432 0, /* default options */
1433 ovector, /* vector of integers for substring information */
1434 30); /* number of elements (NOT size in bytes) */
1435
1436 Extra data for pcre_exec()
1437
1438 If the extra argument is not NULL, it must point to a pcre_extra data
1439 block. The pcre_study() function returns such a block (when it doesn't
1440 return NULL), but you can also create one for yourself, and pass addi-
1441 tional information in it. The pcre_extra block contains the following
1442 fields (not necessarily in this order):
1443
1444 unsigned long int flags;
1445 void *study_data;
1446 unsigned long int match_limit;
1447 unsigned long int match_limit_recursion;
1448 void *callout_data;
1449 const unsigned char *tables;
1450
1451 The flags field is a bitmap that specifies which of the other fields
1452 are set. The flag bits are:
1453
1454 PCRE_EXTRA_STUDY_DATA
1455 PCRE_EXTRA_MATCH_LIMIT
1456 PCRE_EXTRA_MATCH_LIMIT_RECURSION
1457 PCRE_EXTRA_CALLOUT_DATA
1458 PCRE_EXTRA_TABLES
1459
1460 Other flag bits should be set to zero. The study_data field is set in
1461 the pcre_extra block that is returned by pcre_study(), together with
1462 the appropriate flag bit. You should not set this yourself, but you may
1463 add to the block by setting the other fields and their corresponding
1464 flag bits.
1465
1466 The match_limit field provides a means of preventing PCRE from using up
1467 a vast amount of resources when running patterns that are not going to
1468 match, but which have a very large number of possibilities in their
1469 search trees. The classic example is the use of nested unlimited
1470 repeats.
1471
1472 Internally, PCRE uses a function called match() which it calls repeat-
1473 edly (sometimes recursively). The limit set by match_limit is imposed
1474 on the number of times this function is called during a match, which
1475 has the effect of limiting the amount of backtracking that can take
1476 place. For patterns that are not anchored, the count restarts from zero
1477 for each position in the subject string.
1478
1479 The default value for the limit can be set when PCRE is built; the
1480 default default is 10 million, which handles all but the most extreme
1481 cases. You can override the default by suppling pcre_exec() with a
1482 pcre_extra block in which match_limit is set, and
1483 PCRE_EXTRA_MATCH_LIMIT is set in the flags field. If the limit is
1484 exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
1485
1486 The match_limit_recursion field is similar to match_limit, but instead
1487 of limiting the total number of times that match() is called, it limits
1488 the depth of recursion. The recursion depth is a smaller number than
1489 the total number of calls, because not all calls to match() are recur-
1490 sive. This limit is of use only if it is set smaller than match_limit.
1491
1492 Limiting the recursion depth limits the amount of stack that can be
1493 used, or, when PCRE has been compiled to use memory on the heap instead
1494 of the stack, the amount of heap memory that can be used.
1495
1496 The default value for match_limit_recursion can be set when PCRE is
1497 built; the default default is the same value as the default for
1498 match_limit. You can override the default by suppling pcre_exec() with
1499 a pcre_extra block in which match_limit_recursion is set, and
1500 PCRE_EXTRA_MATCH_LIMIT_RECURSION is set in the flags field. If the
1501 limit is exceeded, pcre_exec() returns PCRE_ERROR_RECURSIONLIMIT.
1502
1503 The pcre_callout field is used in conjunction with the "callout" fea-
1504 ture, which is described in the pcrecallout documentation.
1505
1506 The tables field is used to pass a character tables pointer to
1507 pcre_exec(); this overrides the value that is stored with the compiled
1508 pattern. A non-NULL value is stored with the compiled pattern only if
1509 custom tables were supplied to pcre_compile() via its tableptr argu-
1510 ment. If NULL is passed to pcre_exec() using this mechanism, it forces
1511 PCRE's internal tables to be used. This facility is helpful when re-
1512 using patterns that have been saved after compiling with an external
1513 set of tables, because the external tables might be at a different
1514 address when pcre_exec() is called. See the pcreprecompile documenta-
1515 tion for a discussion of saving compiled patterns for later use.
1516
1517 Option bits for pcre_exec()
1518
1519 The unused bits of the options argument for pcre_exec() must be zero.
1520 The only bits that may be set are PCRE_ANCHORED, PCRE_NOTBOL,
1521 PCRE_NOTEOL, PCRE_NOTEMPTY, PCRE_NO_UTF8_CHECK and PCRE_PARTIAL.
1522
1523 PCRE_ANCHORED
1524
1525 The PCRE_ANCHORED option limits pcre_exec() to matching at the first
1526 matching position. If a pattern was compiled with PCRE_ANCHORED, or
1527 turned out to be anchored by virtue of its contents, it cannot be made
1528 unachored at matching time.
1529
1530 PCRE_NOTBOL
1531
1532 This option specifies that first character of the subject string is not
1533 the beginning of a line, so the circumflex metacharacter should not
1534 match before it. Setting this without PCRE_MULTILINE (at compile time)
1535 causes circumflex never to match. This option affects only the behav-
1536 iour of the circumflex metacharacter. It does not affect \A.
1537
1538 PCRE_NOTEOL
1539
1540 This option specifies that the end of the subject string is not the end
1541 of a line, so the dollar metacharacter should not match it nor (except
1542 in multiline mode) a newline immediately before it. Setting this with-
1543 out PCRE_MULTILINE (at compile time) causes dollar never to match. This
1544 option affects only the behaviour of the dollar metacharacter. It does
1545 not affect \Z or \z.
1546
1547 PCRE_NOTEMPTY
1548
1549 An empty string is not considered to be a valid match if this option is
1550 set. If there are alternatives in the pattern, they are tried. If all
1551 the alternatives match the empty string, the entire match fails. For
1552 example, if the pattern
1553
1554 a?b?
1555
1556 is applied to a string not beginning with "a" or "b", it matches the
1557 empty string at the start of the subject. With PCRE_NOTEMPTY set, this
1558 match is not valid, so PCRE searches further into the string for occur-
1559 rences of "a" or "b".
1560
1561 Perl has no direct equivalent of PCRE_NOTEMPTY, but it does make a spe-
1562 cial case of a pattern match of the empty string within its split()
1563 function, and when using the /g modifier. It is possible to emulate
1564 Perl's behaviour after matching a null string by first trying the match
1565 again at the same offset with PCRE_NOTEMPTY and PCRE_ANCHORED, and then
1566 if that fails by advancing the starting offset (see below) and trying
1567 an ordinary match again. There is some code that demonstrates how to do
1568 this in the pcredemo.c sample program.
1569
1570 PCRE_NO_UTF8_CHECK
1571
1572 When PCRE_UTF8 is set at compile time, the validity of the subject as a
1573 UTF-8 string is automatically checked when pcre_exec() is subsequently
1574 called. The value of startoffset is also checked to ensure that it
1575 points to the start of a UTF-8 character. If an invalid UTF-8 sequence
1576 of bytes is found, pcre_exec() returns the error PCRE_ERROR_BADUTF8. If
1577 startoffset contains an invalid value, PCRE_ERROR_BADUTF8_OFFSET is
1578 returned.
1579
1580 If you already know that your subject is valid, and you want to skip
1581 these checks for performance reasons, you can set the
1582 PCRE_NO_UTF8_CHECK option when calling pcre_exec(). You might want to
1583 do this for the second and subsequent calls to pcre_exec() if you are
1584 making repeated calls to find all the matches in a single subject
1585 string. However, you should be sure that the value of startoffset
1586 points to the start of a UTF-8 character. When PCRE_NO_UTF8_CHECK is
1587 set, the effect of passing an invalid UTF-8 string as a subject, or a
1588 value of startoffset that does not point to the start of a UTF-8 char-
1589 acter, is undefined. Your program may crash.
1590
1591 PCRE_PARTIAL
1592
1593 This option turns on the partial matching feature. If the subject
1594 string fails to match the pattern, but at some point during the match-
1595 ing process the end of the subject was reached (that is, the subject
1596 partially matches the pattern and the failure to match occurred only
1597 because there were not enough subject characters), pcre_exec() returns
1598 PCRE_ERROR_PARTIAL instead of PCRE_ERROR_NOMATCH. When PCRE_PARTIAL is
1599 used, there are restrictions on what may appear in the pattern. These
1600 are discussed in the pcrepartial documentation.
1601
1602 The string to be matched by pcre_exec()
1603
1604 The subject string is passed to pcre_exec() as a pointer in subject, a
1605 length in length, and a starting byte offset in startoffset. In UTF-8
1606 mode, the byte offset must point to the start of a UTF-8 character.
1607 Unlike the pattern string, the subject may contain binary zero bytes.
1608 When the starting offset is zero, the search for a match starts at the
1609 beginning of the subject, and this is by far the most common case.
1610
1611 A non-zero starting offset is useful when searching for another match
1612 in the same subject by calling pcre_exec() again after a previous suc-
1613 cess. Setting startoffset differs from just passing over a shortened
1614 string and setting PCRE_NOTBOL in the case of a pattern that begins
1615 with any kind of lookbehind. For example, consider the pattern
1616
1617 \Biss\B
1618
1619 which finds occurrences of "iss" in the middle of words. (\B matches
1620 only if the current position in the subject is not a word boundary.)
1621 When applied to the string "Mississipi" the first call to pcre_exec()
1622 finds the first occurrence. If pcre_exec() is called again with just
1623 the remainder of the subject, namely "issipi", it does not match,
1624 because \B is always false at the start of the subject, which is deemed
1625 to be a word boundary. However, if pcre_exec() is passed the entire
1626 string again, but with startoffset set to 4, it finds the second occur-
1627 rence of "iss" because it is able to look behind the starting point to
1628 discover that it is preceded by a letter.
1629
1630 If a non-zero starting offset is passed when the pattern is anchored,
1631 one attempt to match at the given offset is made. This can only succeed
1632 if the pattern does not require the match to be at the start of the
1633 subject.
1634
1635 How pcre_exec() returns captured substrings
1636
1637 In general, a pattern matches a certain portion of the subject, and in
1638 addition, further substrings from the subject may be picked out by
1639 parts of the pattern. Following the usage in Jeffrey Friedl's book,
1640 this is called "capturing" in what follows, and the phrase "capturing
1641 subpattern" is used for a fragment of a pattern that picks out a sub-
1642 string. PCRE supports several other kinds of parenthesized subpattern
1643 that do not cause substrings to be captured.
1644
1645 Captured substrings are returned to the caller via a vector of integer
1646 offsets whose address is passed in ovector. The number of elements in
1647 the vector is passed in ovecsize, which must be a non-negative number.
1648 Note: this argument is NOT the size of ovector in bytes.
1649
1650 The first two-thirds of the vector is used to pass back captured sub-
1651 strings, each substring using a pair of integers. The remaining third
1652 of the vector is used as workspace by pcre_exec() while matching cap-
1653 turing subpatterns, and is not available for passing back information.
1654 The length passed in ovecsize should always be a multiple of three. If
1655 it is not, it is rounded down.
1656
1657 When a match is successful, information about captured substrings is
1658 returned in pairs of integers, starting at the beginning of ovector,
1659 and continuing up to two-thirds of its length at the most. The first
1660 element of a pair is set to the offset of the first character in a sub-
1661 string, and the second is set to the offset of the first character
1662 after the end of a substring. The first pair, ovector[0] and ovec-
1663 tor[1], identify the portion of the subject string matched by the
1664 entire pattern. The next pair is used for the first capturing subpat-
1665 tern, and so on. The value returned by pcre_exec() is the number of
1666 pairs that have been set. If there are no capturing subpatterns, the
1667 return value from a successful match is 1, indicating that just the
1668 first pair of offsets has been set.
1669
1670 Some convenience functions are provided for extracting the captured
1671 substrings as separate strings. These are described in the following
1672 section.
1673
1674 It is possible for an capturing subpattern number n+1 to match some
1675 part of the subject when subpattern n has not been used at all. For
1676 example, if the string "abc" is matched against the pattern (a|(z))(bc)
1677 subpatterns 1 and 3 are matched, but 2 is not. When this happens, both
1678 offset values corresponding to the unused subpattern are set to -1.
1679
1680 If a capturing subpattern is matched repeatedly, it is the last portion
1681 of the string that it matched that is returned.
1682
1683 If the vector is too small to hold all the captured substring offsets,
1684 it is used as far as possible (up to two-thirds of its length), and the
1685 function returns a value of zero. In particular, if the substring off-
1686 sets are not of interest, pcre_exec() may be called with ovector passed
1687 as NULL and ovecsize as zero. However, if the pattern contains back
1688 references and the ovector is not big enough to remember the related
1689 substrings, PCRE has to get additional memory for use during matching.
1690 Thus it is usually advisable to supply an ovector.
1691
1692 Note that pcre_info() can be used to find out how many capturing sub-
1693 patterns there are in a compiled pattern. The smallest size for ovector
1694 that will allow for n captured substrings, in addition to the offsets
1695 of the substring matched by the whole pattern, is (n+1)*3.
1696
1697 Return values from pcre_exec()
1698
1699 If pcre_exec() fails, it returns a negative number. The following are
1700 defined in the header file:
1701
1702 PCRE_ERROR_NOMATCH (-1)
1703
1704 The subject string did not match the pattern.
1705
1706 PCRE_ERROR_NULL (-2)
1707
1708 Either code or subject was passed as NULL, or ovector was NULL and
1709 ovecsize was not zero.
1710
1711 PCRE_ERROR_BADOPTION (-3)
1712
1713 An unrecognized bit was set in the options argument.
1714
1715 PCRE_ERROR_BADMAGIC (-4)
1716
1717 PCRE stores a 4-byte "magic number" at the start of the compiled code,
1718 to catch the case when it is passed a junk pointer and to detect when a
1719 pattern that was compiled in an environment of one endianness is run in
1720 an environment with the other endianness. This is the error that PCRE
1721 gives when the magic number is not present.
1722
1723 PCRE_ERROR_UNKNOWN_NODE (-5)
1724
1725 While running the pattern match, an unknown item was encountered in the
1726 compiled pattern. This error could be caused by a bug in PCRE or by
1727 overwriting of the compiled pattern.
1728
1729 PCRE_ERROR_NOMEMORY (-6)
1730
1731 If a pattern contains back references, but the ovector that is passed
1732 to pcre_exec() is not big enough to remember the referenced substrings,
1733 PCRE gets a block of memory at the start of matching to use for this
1734 purpose. If the call via pcre_malloc() fails, this error is given. The
1735 memory is automatically freed at the end of matching.
1736
1737 PCRE_ERROR_NOSUBSTRING (-7)
1738
1739 This error is used by the pcre_copy_substring(), pcre_get_substring(),
1740 and pcre_get_substring_list() functions (see below). It is never
1741 returned by pcre_exec().
1742
1743 PCRE_ERROR_MATCHLIMIT (-8)
1744
1745 The backtracking limit, as specified by the match_limit field in a
1746 pcre_extra structure (or defaulted) was reached. See the description
1747 above.
1748
1749 PCRE_ERROR_RECURSIONLIMIT (-21)
1750
1751 The internal recursion limit, as specified by the match_limit_recursion
1752 field in a pcre_extra structure (or defaulted) was reached. See the
1753 description above.
1754
1755 PCRE_ERROR_CALLOUT (-9)
1756
1757 This error is never generated by pcre_exec() itself. It is provided for
1758 use by callout functions that want to yield a distinctive error code.
1759 See the pcrecallout documentation for details.
1760
1761 PCRE_ERROR_BADUTF8 (-10)
1762
1763 A string that contains an invalid UTF-8 byte sequence was passed as a
1764 subject.
1765
1766 PCRE_ERROR_BADUTF8_OFFSET (-11)
1767
1768 The UTF-8 byte sequence that was passed as a subject was valid, but the
1769 value of startoffset did not point to the beginning of a UTF-8 charac-
1770 ter.
1771
1772 PCRE_ERROR_PARTIAL (-12)
1773
1774 The subject string did not match, but it did match partially. See the
1775 pcrepartial documentation for details of partial matching.
1776
1777 PCRE_ERROR_BADPARTIAL (-13)
1778
1779 The PCRE_PARTIAL option was used with a compiled pattern containing
1780 items that are not supported for partial matching. See the pcrepartial
1781 documentation for details of partial matching.
1782
1783 PCRE_ERROR_INTERNAL (-14)
1784
1785 An unexpected internal error has occurred. This error could be caused
1786 by a bug in PCRE or by overwriting of the compiled pattern.
1787
1788 PCRE_ERROR_BADCOUNT (-15)
1789
1790 This error is given if the value of the ovecsize argument is negative.
1791
1792
1793 EXTRACTING CAPTURED SUBSTRINGS BY NUMBER
1794
1795 int pcre_copy_substring(const char *subject, int *ovector,
1796 int stringcount, int stringnumber, char *buffer,
1797 int buffersize);
1798
1799 int pcre_get_substring(const char *subject, int *ovector,
1800 int stringcount, int stringnumber,
1801 const char **stringptr);
1802
1803 int pcre_get_substring_list(const char *subject,
1804 int *ovector, int stringcount, const char ***listptr);
1805
1806 Captured substrings can be accessed directly by using the offsets
1807 returned by pcre_exec() in ovector. For convenience, the functions
1808 pcre_copy_substring(), pcre_get_substring(), and pcre_get_sub-
1809 string_list() are provided for extracting captured substrings as new,
1810 separate, zero-terminated strings. These functions identify substrings
1811 by number. The next section describes functions for extracting named
1812 substrings. A substring that contains a binary zero is correctly
1813 extracted and has a further zero added on the end, but the result is
1814 not, of course, a C string.
1815
1816 The first three arguments are the same for all three of these func-
1817 tions: subject is the subject string that has just been successfully
1818 matched, ovector is a pointer to the vector of integer offsets that was
1819 passed to pcre_exec(), and stringcount is the number of substrings that
1820 were captured by the match, including the substring that matched the
1821 entire regular expression. This is the value returned by pcre_exec() if
1822 it is greater than zero. If pcre_exec() returned zero, indicating that
1823 it ran out of space in ovector, the value passed as stringcount should
1824 be the number of elements in the vector divided by three.
1825
1826 The functions pcre_copy_substring() and pcre_get_substring() extract a
1827 single substring, whose number is given as stringnumber. A value of
1828 zero extracts the substring that matched the entire pattern, whereas
1829 higher values extract the captured substrings. For pcre_copy_sub-
1830 string(), the string is placed in buffer, whose length is given by
1831 buffersize, while for pcre_get_substring() a new block of memory is
1832 obtained via pcre_malloc, and its address is returned via stringptr.
1833 The yield of the function is the length of the string, not including
1834 the terminating zero, or one of
1835
1836 PCRE_ERROR_NOMEMORY (-6)
1837
1838 The buffer was too small for pcre_copy_substring(), or the attempt to
1839 get memory failed for pcre_get_substring().
1840
1841 PCRE_ERROR_NOSUBSTRING (-7)
1842
1843 There is no substring whose number is stringnumber.
1844
1845 The pcre_get_substring_list() function extracts all available sub-
1846 strings and builds a list of pointers to them. All this is done in a
1847 single block of memory that is obtained via pcre_malloc. The address of
1848 the memory block is returned via listptr, which is also the start of
1849 the list of string pointers. The end of the list is marked by a NULL
1850 pointer. The yield of the function is zero if all went well, or
1851
1852 PCRE_ERROR_NOMEMORY (-6)
1853
1854 if the attempt to get the memory block failed.
1855
1856 When any of these functions encounter a substring that is unset, which
1857 can happen when capturing subpattern number n+1 matches some part of
1858 the subject, but subpattern n has not been used at all, they return an
1859 empty string. This can be distinguished from a genuine zero-length sub-
1860 string by inspecting the appropriate offset in ovector, which is nega-
1861 tive for unset substrings.
1862
1863 The two convenience functions pcre_free_substring() and pcre_free_sub-
1864 string_list() can be used to free the memory returned by a previous
1865 call of pcre_get_substring() or pcre_get_substring_list(), respec-
1866 tively. They do nothing more than call the function pointed to by
1867 pcre_free, which of course could be called directly from a C program.
1868 However, PCRE is used in some situations where it is linked via a spe-
1869 cial interface to another programming language which cannot use
1870 pcre_free directly; it is for these cases that the functions are pro-
1871 vided.
1872
1873
1874 EXTRACTING CAPTURED SUBSTRINGS BY NAME
1875
1876 int pcre_get_stringnumber(const pcre *code,
1877 const char *name);
1878
1879 int pcre_copy_named_substring(const pcre *code,
1880 const char *subject, int *ovector,
1881 int stringcount, const char *stringname,
1882 char *buffer, int buffersize);
1883
1884 int pcre_get_named_substring(const pcre *code,
1885 const char *subject, int *ovector,
1886 int stringcount, const char *stringname,
1887 const char **stringptr);
1888
1889 To extract a substring by name, you first have to find associated num-
1890 ber. For example, for this pattern
1891
1892 (a+)b(?P<xxx>\d+)...
1893
1894 the number of the subpattern called "xxx" is 2. You can find the number
1895 from the name by calling pcre_get_stringnumber(). The first argument is
1896 the compiled pattern, and the second is the name. The yield of the
1897 function is the subpattern number, or PCRE_ERROR_NOSUBSTRING (-7) if
1898 there is no subpattern of that name.
1899
1900 Given the number, you can extract the substring directly, or use one of
1901 the functions described in the previous section. For convenience, there
1902 are also two functions that do the whole job.
1903
1904 Most of the arguments of pcre_copy_named_substring() and
1905 pcre_get_named_substring() are the same as those for the similarly
1906 named functions that extract by number. As these are described in the
1907 previous section, they are not re-described here. There are just two
1908 differences:
1909
1910 First, instead of a substring number, a substring name is given. Sec-
1911 ond, there is an extra argument, given at the start, which is a pointer
1912 to the compiled pattern. This is needed in order to gain access to the
1913 name-to-number translation table.
1914
1915 These functions call pcre_get_stringnumber(), and if it succeeds, they
1916 then call pcre_copy_substring() or pcre_get_substring(), as appropri-
1917 ate.
1918
1919
1920 FINDING ALL POSSIBLE MATCHES
1921
1922 The traditional matching function uses a similar algorithm to Perl,
1923 which stops when it finds the first match, starting at a given point in
1924 the subject. If you want to find all possible matches, or the longest
1925 possible match, consider using the alternative matching function (see
1926 below) instead. If you cannot use the alternative function, but still
1927 need to find all possible matches, you can kludge it up by making use
1928 of the callout facility, which is described in the pcrecallout documen-
1929 tation.
1930
1931 What you have to do is to insert a callout right at the end of the pat-
1932 tern. When your callout function is called, extract and save the cur-
1933 rent matched substring. Then return 1, which forces pcre_exec() to
1934 backtrack and try other alternatives. Ultimately, when it runs out of
1935 matches, pcre_exec() will yield PCRE_ERROR_NOMATCH.
1936
1937
1938 MATCHING A PATTERN: THE ALTERNATIVE FUNCTION
1939
1940 int pcre_dfa_exec(const pcre *code, const pcre_extra *extra,
1941 const char *subject, int length, int startoffset,
1942 int options, int *ovector, int ovecsize,
1943 int *workspace, int wscount);
1944
1945 The function pcre_dfa_exec() is called to match a subject string
1946 against a compiled pattern, using a "DFA" matching algorithm. This has
1947 different characteristics to the normal algorithm, and is not compati-
1948 ble with Perl. Some of the features of PCRE patterns are not supported.
1949 Nevertheless, there are times when this kind of matching can be useful.
1950 For a discussion of the two matching algorithms, see the pcrematching
1951 documentation.
1952
1953 The arguments for the pcre_dfa_exec() function are the same as for
1954 pcre_exec(), plus two extras. The ovector argument is used in a differ-
1955 ent way, and this is described below. The other common arguments are
1956 used in the same way as for pcre_exec(), so their description is not
1957 repeated here.
1958
1959 The two additional arguments provide workspace for the function. The
1960 workspace vector should contain at least 20 elements. It is used for
1961 keeping track of multiple paths through the pattern tree. More
1962 workspace will be needed for patterns and subjects where there are a
1963 lot of possible matches.
1964
1965 Here is an example of a simple call to pcre_dfa_exec():
1966
1967 int rc;
1968 int ovector[10];
1969 int wspace[20];
1970 rc = pcre_dfa_exec(
1971 re, /* result of pcre_compile() */
1972 NULL, /* we didn't study the pattern */
1973 "some string", /* the subject string */
1974 11, /* the length of the subject string */
1975 0, /* start at offset 0 in the subject */
1976 0, /* default options */
1977 ovector, /* vector of integers for substring information */
1978 10, /* number of elements (NOT size in bytes) */
1979 wspace, /* working space vector */
1980 20); /* number of elements (NOT size in bytes) */
1981
1982 Option bits for pcre_dfa_exec()
1983
1984 The unused bits of the options argument for pcre_dfa_exec() must be
1985 zero. The only bits that may be set are PCRE_ANCHORED, PCRE_NOTBOL,
1986 PCRE_NOTEOL, PCRE_NOTEMPTY, PCRE_NO_UTF8_CHECK, PCRE_PARTIAL,
1987 PCRE_DFA_SHORTEST, and PCRE_DFA_RESTART. All but the last three of
1988 these are the same as for pcre_exec(), so their description is not
1989 repeated here.
1990
1991 PCRE_PARTIAL
1992
1993 This has the same general effect as it does for pcre_exec(), but the
1994 details are slightly different. When PCRE_PARTIAL is set for
1995 pcre_dfa_exec(), the return code PCRE_ERROR_NOMATCH is converted into
1996 PCRE_ERROR_PARTIAL if the end of the subject is reached, there have
1997 been no complete matches, but there is still at least one matching pos-
1998 sibility. The portion of the string that provided the partial match is
1999 set as the first matching string.
2000
2001 PCRE_DFA_SHORTEST
2002
2003 Setting the PCRE_DFA_SHORTEST option causes the matching algorithm to
2004 stop as soon as it has found one match. Because of the way the DFA
2005 algorithm works, this is necessarily the shortest possible match at the
2006 first possible matching point in the subject string.
2007
2008 PCRE_DFA_RESTART
2009
2010 When pcre_dfa_exec() is called with the PCRE_PARTIAL option, and
2011 returns a partial match, it is possible to call it again, with addi-
2012 tional subject characters, and have it continue with the same match.
2013 The PCRE_DFA_RESTART option requests this action; when it is set, the
2014 workspace and wscount options must reference the same vector as before
2015 because data about the match so far is left in them after a partial
2016 match. There is more discussion of this facility in the pcrepartial
2017 documentation.
2018
2019 Successful returns from pcre_dfa_exec()
2020
2021 When pcre_dfa_exec() succeeds, it may have matched more than one sub-
2022 string in the subject. Note, however, that all the matches from one run
2023 of the function start at the same point in the subject. The shorter
2024 matches are all initial substrings of the longer matches. For example,
2025 if the pattern
2026
2027 <.*>
2028
2029 is matched against the string
2030
2031 This is <something> <something else> <something further> no more
2032
2033 the three matched strings are
2034
2035 <something>
2036 <something> <something else>
2037 <something> <something else> <something further>
2038
2039 On success, the yield of the function is a number greater than zero,
2040 which is the number of matched substrings. The substrings themselves
2041 are returned in ovector. Each string uses two elements; the first is
2042 the offset to the start, and the second is the offset to the end. All
2043 the strings have the same start offset. (Space could have been saved by
2044 giving this only once, but it was decided to retain some compatibility
2045 with the way pcre_exec() returns data, even though the meaning of the
2046 strings is different.)
2047
2048 The strings are returned in reverse order of length; that is, the long-
2049 est matching string is given first. If there were too many matches to
2050 fit into ovector, the yield of the function is zero, and the vector is
2051 filled with the longest matches.
2052
2053 Error returns from pcre_dfa_exec()
2054
2055 The pcre_dfa_exec() function returns a negative number when it fails.
2056 Many of the errors are the same as for pcre_exec(), and these are
2057 described above. There are in addition the following errors that are
2058 specific to pcre_dfa_exec():
2059
2060 PCRE_ERROR_DFA_UITEM (-16)
2061
2062 This return is given if pcre_dfa_exec() encounters an item in the pat-
2063 tern that it does not support, for instance, the use of \C or a back
2064 reference.
2065
2066 PCRE_ERROR_DFA_UCOND (-17)
2067
2068 This return is given if pcre_dfa_exec() encounters a condition item in
2069 a pattern that uses a back reference for the condition. This is not
2070 supported.
2071
2072 PCRE_ERROR_DFA_UMLIMIT (-18)
2073
2074 This return is given if pcre_dfa_exec() is called with an extra block
2075 that contains a setting of the match_limit field. This is not supported
2076 (it is meaningless).
2077
2078 PCRE_ERROR_DFA_WSSIZE (-19)
2079
2080 This return is given if pcre_dfa_exec() runs out of space in the
2081 workspace vector.
2082
2083 PCRE_ERROR_DFA_RECURSE (-20)
2084
2085 When a recursive subpattern is processed, the matching function calls
2086 itself recursively, using private vectors for ovector and workspace.
2087 This error is given if the output vector is not large enough. This
2088 should be extremely rare, as a vector of size 1000 is used.
2089
2090 Last updated: 18 January 2006
2091 Copyright (c) 1997-2006 University of Cambridge.
2092 ------------------------------------------------------------------------------
2093
2094
2095 PCRECALLOUT(3) PCRECALLOUT(3)
2096
2097
2098 NAME
2099 PCRE - Perl-compatible regular expressions
2100
2101
2102 PCRE CALLOUTS
2103
2104 int (*pcre_callout)(pcre_callout_block *);
2105
2106 PCRE provides a feature called "callout", which is a means of temporar-
2107 ily passing control to the caller of PCRE in the middle of pattern
2108 matching. The caller of PCRE provides an external function by putting
2109 its entry point in the global variable pcre_callout. By default, this
2110 variable contains NULL, which disables all calling out.
2111
2112 Within a regular expression, (?C) indicates the points at which the
2113 external function is to be called. Different callout points can be
2114 identified by putting a number less than 256 after the letter C. The
2115 default value is zero. For example, this pattern has two callout
2116 points:
2117
2118 (?C1)eabc(?C2)def
2119
2120 If the PCRE_AUTO_CALLOUT option bit is set when pcre_compile() is
2121 called, PCRE automatically inserts callouts, all with number 255,
2122 before each item in the pattern. For example, if PCRE_AUTO_CALLOUT is
2123 used with the pattern
2124
2125 A(\d{2}|--)
2126
2127 it is processed as if it were
2128
2129 (?C255)A(?C255)((?C255)\d{2}(?C255)|(?C255)-(?C255)-(?C255))(?C255)
2130
2131 Notice that there is a callout before and after each parenthesis and
2132 alternation bar. Automatic callouts can be used for tracking the
2133 progress of pattern matching. The pcretest command has an option that
2134 sets automatic callouts; when it is used, the output indicates how the
2135 pattern is matched. This is useful information when you are trying to
2136 optimize the performance of a particular pattern.
2137
2138
2139 MISSING CALLOUTS
2140
2141 You should be aware that, because of optimizations in the way PCRE
2142 matches patterns, callouts sometimes do not happen. For example, if the
2143 pattern is
2144
2145 ab(?C4)cd
2146
2147 PCRE knows that any matching string must contain the letter "d". If the
2148 subject string is "abyz", the lack of "d" means that matching doesn't
2149 ever start, and the callout is never reached. However, with "abyd",
2150 though the result is still no match, the callout is obeyed.
2151
2152
2153 THE CALLOUT INTERFACE
2154
2155 During matching, when PCRE reaches a callout point, the external func-
2156 tion defined by pcre_callout is called (if it is set). This applies to
2157 both the pcre_exec() and the pcre_dfa_exec() matching functions. The
2158 only argument to the callout function is a pointer to a pcre_callout
2159 block. This structure contains the following fields:
2160
2161 int version;
2162 int callout_number;
2163 int *offset_vector;
2164 const char *subject;
2165 int subject_length;
2166 int start_match;
2167 int current_position;
2168 int capture_top;
2169 int capture_last;
2170 void *callout_data;
2171 int pattern_position;
2172 int next_item_length;
2173
2174 The version field is an integer containing the version number of the
2175 block format. The initial version was 0; the current version is 1. The
2176 version number will change again in future if additional fields are
2177 added, but the intention is never to remove any of the existing fields.
2178
2179 The callout_number field contains the number of the callout, as com-
2180 piled into the pattern (that is, the number after ?C for manual call-
2181 outs, and 255 for automatically generated callouts).
2182
2183 The offset_vector field is a pointer to the vector of offsets that was
2184 passed by the caller to pcre_exec() or pcre_dfa_exec(). When
2185 pcre_exec() is used, the contents can be inspected in order to extract
2186 substrings that have been matched so far, in the same way as for
2187 extracting substrings after a match has completed. For pcre_dfa_exec()
2188 this field is not useful.
2189
2190 The subject and subject_length fields contain copies of the values that
2191 were passed to pcre_exec().
2192
2193 The start_match field contains the offset within the subject at which
2194 the current match attempt started. If the pattern is not anchored, the
2195 callout function may be called several times from the same point in the
2196 pattern for different starting points in the subject.
2197
2198 The current_position field contains the offset within the subject of
2199 the current match pointer.
2200
2201 When the pcre_exec() function is used, the capture_top field contains
2202 one more than the number of the highest numbered captured substring so
2203 far. If no substrings have been captured, the value of capture_top is
2204 one. This is always the case when pcre_dfa_exec() is used, because it
2205 does not support captured substrings.
2206
2207 The capture_last field contains the number of the most recently cap-
2208 tured substring. If no substrings have been captured, its value is -1.
2209 This is always the case when pcre_dfa_exec() is used.
2210
2211 The callout_data field contains a value that is passed to pcre_exec()
2212 or pcre_dfa_exec() specifically so that it can be passed back in call-
2213 outs. It is passed in the pcre_callout field of the pcre_extra data
2214 structure. If no such data was passed, the value of callout_data in a
2215 pcre_callout block is NULL. There is a description of the pcre_extra
2216 structure in the pcreapi documentation.
2217
2218 The pattern_position field is present from version 1 of the pcre_call-
2219 out structure. It contains the offset to the next item to be matched in
2220 the pattern string.
2221
2222 The next_item_length field is present from version 1 of the pcre_call-
2223 out structure. It contains the length of the next item to be matched in
2224 the pattern string. When the callout immediately precedes an alterna-
2225 tion bar, a closing parenthesis, or the end of the pattern, the length
2226 is zero. When the callout precedes an opening parenthesis, the length
2227 is that of the entire subpattern.
2228
2229 The pattern_position and next_item_length fields are intended to help
2230 in distinguishing between different automatic callouts, which all have
2231 the same callout number. However, they are set for all callouts.
2232
2233
2234 RETURN VALUES
2235
2236 The external callout function returns an integer to PCRE. If the value
2237 is zero, matching proceeds as normal. If the value is greater than
2238 zero, matching fails at the current point, but the testing of other
2239 matching possibilities goes ahead, just as if a lookahead assertion had
2240 failed. If the value is less than zero, the match is abandoned, and
2241 pcre_exec() (or pcre_dfa_exec()) returns the negative value.
2242
2243 Negative values should normally be chosen from the set of
2244 PCRE_ERROR_xxx values. In particular, PCRE_ERROR_NOMATCH forces a stan-
2245 dard "no match" failure. The error number PCRE_ERROR_CALLOUT is
2246 reserved for use by callout functions; it will never be used by PCRE
2247 itself.
2248
2249 Last updated: 28 February 2005
2250 Copyright (c) 1997-2005 University of Cambridge.
2251 ------------------------------------------------------------------------------
2252
2253
2254 PCRECOMPAT(3) PCRECOMPAT(3)
2255
2256
2257 NAME
2258 PCRE - Perl-compatible regular expressions
2259
2260
2261 DIFFERENCES BETWEEN PCRE AND PERL
2262
2263 This document describes the differences in the ways that PCRE and Perl
2264 handle regular expressions. The differences described here are with
2265 respect to Perl 5.8.
2266
2267 1. PCRE has only a subset of Perl's UTF-8 and Unicode support. Details
2268 of what it does have are given in the section on UTF-8 support in the
2269 main pcre page.
2270
2271 2. PCRE does not allow repeat quantifiers on lookahead assertions. Perl
2272 permits them, but they do not mean what you might think. For example,
2273 (?!a){3} does not assert that the next three characters are not "a". It
2274 just asserts that the next character is not "a" three times.
2275
2276 3. Capturing subpatterns that occur inside negative lookahead asser-
2277 tions are counted, but their entries in the offsets vector are never
2278 set. Perl sets its numerical variables from any such patterns that are
2279 matched before the assertion fails to match something (thereby succeed-
2280 ing), but only if the negative lookahead assertion contains just one
2281 branch.
2282
2283 4. Though binary zero characters are supported in the subject string,
2284 they are not allowed in a pattern string because it is passed as a nor-
2285 mal C string, terminated by zero. The escape sequence \0 can be used in
2286 the pattern to represent a binary zero.
2287
2288 5. The following Perl escape sequences are not supported: \l, \u, \L,
2289 \U, and \N. In fact these are implemented by Perl's general string-han-
2290 dling and are not part of its pattern matching engine. If any of these
2291 are encountered by PCRE, an error is generated.
2292
2293 6. The Perl escape sequences \p, \P, and \X are supported only if PCRE
2294 is built with Unicode character property support. The properties that
2295 can be tested with \p and \P are limited to the general category prop-
2296 erties such as Lu and Nd, script names such as Greek or Han, and the
2297 derived properties Any and L&.
2298
2299 7. PCRE does support the \Q...\E escape for quoting substrings. Charac-
2300 ters in between are treated as literals. This is slightly different
2301 from Perl in that $ and @ are also handled as literals inside the
2302 quotes. In Perl, they cause variable interpolation (but of course PCRE
2303 does not have variables). Note the following examples:
2304
2305 Pattern PCRE matches Perl matches
2306
2307 \Qabc$xyz\E abc$xyz abc followed by the
2308 contents of $xyz
2309 \Qabc\$xyz\E abc\$xyz abc\$xyz
2310 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
2311
2312 The \Q...\E sequence is recognized both inside and outside character
2313 classes.
2314
2315 8. Fairly obviously, PCRE does not support the (?{code}) and (?p{code})
2316 constructions. However, there is support for recursive patterns using
2317 the non-Perl items (?R), (?number), and (?P>name). Also, the PCRE
2318 "callout" feature allows an external function to be called during pat-
2319 tern matching. See the pcrecallout documentation for details.
2320
2321 9. There are some differences that are concerned with the settings of
2322 captured strings when part of a pattern is repeated. For example,
2323 matching "aba" against the pattern /^(a(b)?)+$/ in Perl leaves $2
2324 unset, but in PCRE it is set to "b".
2325
2326 10. PCRE provides some extensions to the Perl regular expression facil-
2327 ities:
2328
2329 (a) Although lookbehind assertions must match fixed length strings,
2330 each alternative branch of a lookbehind assertion can match a different
2331 length of string. Perl requires them all to have the same length.
2332
2333 (b) If PCRE_DOLLAR_ENDONLY is set and PCRE_MULTILINE is not set, the $
2334 meta-character matches only at the very end of the string.
2335
2336 (c) If PCRE_EXTRA is set, a backslash followed by a letter with no spe-
2337 cial meaning is faulted.
2338
2339 (d) If PCRE_UNGREEDY is set, the greediness of the repetition quanti-
2340 fiers is inverted, that is, by default they are not greedy, but if fol-
2341 lowed by a question mark they are.
2342
2343 (e) PCRE_ANCHORED can be used at matching time to force a pattern to be
2344 tried only at the first matching position in the subject string.
2345
2346 (f) The PCRE_NOTBOL, PCRE_NOTEOL, PCRE_NOTEMPTY, and PCRE_NO_AUTO_CAP-
2347 TURE options for pcre_exec() have no Perl equivalents.
2348
2349 (g) The (?R), (?number), and (?P>name) constructs allows for recursive
2350 pattern matching (Perl can do this using the (?p{code}) construct,
2351 which PCRE cannot support.)
2352
2353 (h) PCRE supports named capturing substrings, using the Python syntax.
2354
2355 (i) PCRE supports the possessive quantifier "++" syntax, taken from
2356 Sun's Java package.
2357
2358 (j) The (R) condition, for testing recursion, is a PCRE extension.
2359
2360 (k) The callout facility is PCRE-specific.
2361
2362 (l) The partial matching facility is PCRE-specific.
2363
2364 (m) Patterns compiled by PCRE can be saved and re-used at a later time,
2365 even on different hosts that have the other endianness.
2366
2367 (n) The alternative matching function (pcre_dfa_exec()) matches in a
2368 different way and is not Perl-compatible.
2369
2370 Last updated: 24 January 2006
2371 Copyright (c) 1997-2006 University of Cambridge.
2372 ------------------------------------------------------------------------------
2373
2374
2375 PCREPATTERN(3) PCREPATTERN(3)
2376
2377
2378 NAME
2379 PCRE - Perl-compatible regular expressions
2380
2381
2382 PCRE REGULAR EXPRESSION DETAILS
2383
2384 The syntax and semantics of the regular expressions supported by PCRE
2385 are described below. Regular expressions are also described in the Perl
2386 documentation and in a number of books, some of which have copious
2387 examples. Jeffrey Friedl's "Mastering Regular Expressions", published
2388 by O'Reilly, covers regular expressions in great detail. This descrip-
2389 tion of PCRE's regular expressions is intended as reference material.
2390
2391 The original operation of PCRE was on strings of one-byte characters.
2392 However, there is now also support for UTF-8 character strings. To use
2393 this, you must build PCRE to include UTF-8 support, and then call
2394 pcre_compile() with the PCRE_UTF8 option. How this affects pattern
2395 matching is mentioned in several places below. There is also a summary
2396 of UTF-8 features in the section on UTF-8 support in the main pcre
2397 page.
2398
2399 The remainder of this document discusses the patterns that are sup-
2400 ported by PCRE when its main matching function, pcre_exec(), is used.
2401 From release 6.0, PCRE offers a second matching function,
2402 pcre_dfa_exec(), which matches using a different algorithm that is not
2403 Perl-compatible. The advantages and disadvantages of the alternative
2404 function, and how it differs from the normal function, are discussed in
2405 the pcrematching page.
2406
2407 A regular expression is a pattern that is matched against a subject
2408 string from left to right. Most characters stand for themselves in a
2409 pattern, and match the corresponding characters in the subject. As a
2410 trivial example, the pattern
2411
2412 The quick brown fox
2413
2414 matches a portion of a subject string that is identical to itself. When
2415 caseless matching is specified (the PCRE_CASELESS option), letters are
2416 matched independently of case. In UTF-8 mode, PCRE always understands
2417 the concept of case for characters whose values are less than 128, so
2418 caseless matching is always possible. For characters with higher val-
2419 ues, the concept of case is supported if PCRE is compiled with Unicode
2420 property support, but not otherwise. If you want to use caseless
2421 matching for characters 128 and above, you must ensure that PCRE is
2422 compiled with Unicode property support as well as with UTF-8 support.
2423
2424 The power of regular expressions comes from the ability to include
2425 alternatives and repetitions in the pattern. These are encoded in the
2426 pattern by the use of metacharacters, which do not stand for themselves
2427 but instead are interpreted in some special way.
2428
2429 There are two different sets of metacharacters: those that are recog-
2430 nized anywhere in the pattern except within square brackets, and those
2431 that are recognized in square brackets. Outside square brackets, the
2432 metacharacters are as follows:
2433
2434 \ general escape character with several uses
2435 ^ assert start of string (or line, in multiline mode)
2436 $ assert end of string (or line, in multiline mode)
2437 . match any character except newline (by default)
2438 [ start character class definition
2439 | start of alternative branch
2440 ( start subpattern
2441 ) end subpattern
2442 ? extends the meaning of (
2443 also 0 or 1 quantifier
2444 also quantifier minimizer
2445 * 0 or more quantifier
2446 + 1 or more quantifier
2447 also "possessive quantifier"
2448 { start min/max quantifier
2449
2450 Part of a pattern that is in square brackets is called a "character
2451 class". In a character class the only metacharacters are:
2452
2453 \ general escape character
2454 ^ negate the class, but only if the first character
2455 - indicates character range
2456 [ POSIX character class (only if followed by POSIX
2457 syntax)
2458 ] terminates the character class
2459
2460 The following sections describe the use of each of the metacharacters.
2461
2462
2463 BACKSLASH
2464
2465 The backslash character has several uses. Firstly, if it is followed by
2466 a non-alphanumeric character, it takes away any special meaning that
2467 character may have. This use of backslash as an escape character
2468 applies both inside and outside character classes.
2469
2470 For example, if you want to match a * character, you write \* in the
2471 pattern. This escaping action applies whether or not the following
2472 character would otherwise be interpreted as a metacharacter, so it is
2473 always safe to precede a non-alphanumeric with backslash to specify
2474 that it stands for itself. In particular, if you want to match a back-
2475 slash, you write \\.
2476
2477 If a pattern is compiled with the PCRE_EXTENDED option, whitespace in
2478 the pattern (other than in a character class) and characters between a
2479 # outside a character class and the next newline character are ignored.
2480 An escaping backslash can be used to include a whitespace or # charac-
2481 ter as part of the pattern.
2482
2483 If you want to remove the special meaning from a sequence of charac-
2484 ters, you can do so by putting them between \Q and \E. This is differ-
2485 ent from Perl in that $ and @ are handled as literals in \Q...\E
2486 sequences in PCRE, whereas in Perl, $ and @ cause variable interpola-
2487 tion. Note the following examples:
2488
2489 Pattern PCRE matches Perl matches
2490
2491 \Qabc$xyz\E abc$xyz abc followed by the
2492 contents of $xyz
2493 \Qabc\$xyz\E abc\$xyz abc\$xyz
2494 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
2495
2496 The \Q...\E sequence is recognized both inside and outside character
2497 classes.
2498
2499 Non-printing characters
2500
2501 A second use of backslash provides a way of encoding non-printing char-
2502 acters in patterns in a visible manner. There is no restriction on the
2503 appearance of non-printing characters, apart from the binary zero that
2504 terminates a pattern, but when a pattern is being prepared by text
2505 editing, it is usually easier to use one of the following escape
2506 sequences than the binary character it represents:
2507
2508 \a alarm, that is, the BEL character (hex 07)
2509 \cx "control-x", where x is any character
2510 \e escape (hex 1B)
2511 \f formfeed (hex 0C)
2512 \n newline (hex 0A)
2513 \r carriage return (hex 0D)
2514 \t tab (hex 09)
2515 \ddd character with octal code ddd, or backreference
2516 \xhh character with hex code hh
2517 \x{hhh..} character with hex code hhh..
2518
2519 The precise effect of \cx is as follows: if x is a lower case letter,
2520 it is converted to upper case. Then bit 6 of the character (hex 40) is
2521 inverted. Thus \cz becomes hex 1A, but \c{ becomes hex 3B, while \c;
2522 becomes hex 7B.
2523
2524 After \x, from zero to two hexadecimal digits are read (letters can be
2525 in upper or lower case). Any number of hexadecimal digits may appear
2526 between \x{ and }, but the value of the character code must be less
2527 than 256 in non-UTF-8 mode, and less than 2**31 in UTF-8 mode (that is,
2528 the maximum hexadecimal value is 7FFFFFFF). If characters other than
2529 hexadecimal digits appear between \x{ and }, or if there is no termi-
2530 nating }, this form of escape is not recognized. Instead, the initial
2531 \x will be interpreted as a basic hexadecimal escape, with no following
2532 digits, giving a character whose value is zero.
2533
2534 Characters whose value is less than 256 can be defined by either of the
2535 two syntaxes for \x. There is no difference in the way they are han-
2536 dled. For example, \xdc is exactly the same as \x{dc}.
2537
2538 After \0 up to two further octal digits are read. In both cases, if
2539 there are fewer than two digits, just those that are present are used.
2540 Thus the sequence \0\x\07 specifies two binary zeros followed by a BEL
2541 character (code value 7). Make sure you supply two digits after the
2542 initial zero if the pattern character that follows is itself an octal
2543 digit.
2544
2545 The handling of a backslash followed by a digit other than 0 is compli-
2546 cated. Outside a character class, PCRE reads it and any following dig-
2547 its as a decimal number. If the number is less than 10, or if there
2548 have been at least that many previous capturing left parentheses in the
2549 expression, the entire sequence is taken as a back reference. A
2550 description of how this works is given later, following the discussion
2551 of parenthesized subpatterns.
2552
2553 Inside a character class, or if the decimal number is greater than 9
2554 and there have not been that many capturing subpatterns, PCRE re-reads
2555 up to three octal digits following the backslash, and generates a sin-
2556 gle byte from the least significant 8 bits of the value. Any subsequent
2557 digits stand for themselves. For example:
2558
2559 \040 is another way of writing a space
2560 \40 is the same, provided there are fewer than 40
2561 previous capturing subpatterns
2562 \7 is always a back reference
2563 \11 might be a back reference, or another way of
2564 writing a tab
2565 \011 is always a tab
2566 \0113 is a tab followed by the character "3"
2567 \113 might be a back reference, otherwise the
2568 character with octal code 113
2569 \377 might be a back reference, otherwise
2570 the byte consisting entirely of 1 bits
2571 \81 is either a back reference, or a binary zero
2572 followed by the two characters "8" and "1"
2573
2574 Note that octal values of 100 or greater must not be introduced by a
2575 leading zero, because no more than three octal digits are ever read.
2576
2577 All the sequences that define a single byte value or a single UTF-8
2578 character (in UTF-8 mode) can be used both inside and outside character
2579 classes. In addition, inside a character class, the sequence \b is
2580 interpreted as the backspace character (hex 08), and the sequence \X is
2581 interpreted as the character "X". Outside a character class, these
2582 sequences have different meanings (see below).
2583
2584 Generic character types
2585
2586 The third use of backslash is for specifying generic character types.
2587 The following are always recognized:
2588
2589 \d any decimal digit
2590 \D any character that is not a decimal digit
2591 \s any whitespace character
2592 \S any character that is not a whitespace character
2593 \w any "word" character
2594 \W any "non-word" character
2595
2596 Each pair of escape sequences partitions the complete set of characters
2597 into two disjoint sets. Any given character matches one, and only one,
2598 of each pair.
2599
2600 These character type sequences can appear both inside and outside char-
2601 acter classes. They each match one character of the appropriate type.
2602 If the current matching point is at the end of the subject string, all
2603 of them fail, since there is no character to match.
2604
2605 For compatibility with Perl, \s does not match the VT character (code
2606 11). This makes it different from the the POSIX "space" class. The \s
2607 characters are HT (9), LF (10), FF (12), CR (13), and space (32).
2608
2609 A "word" character is an underscore or any character less than 256 that
2610 is a letter or digit. The definition of letters and digits is con-
2611 trolled by PCRE's low-valued character tables, and may vary if locale-
2612 specific matching is taking place (see "Locale support" in the pcreapi
2613 page). For example, in the "fr_FR" (French) locale, some character
2614 codes greater than 128 are used for accented letters, and these are
2615 matched by \w.
2616
2617 In UTF-8 mode, characters with values greater than 128 never match \d,
2618 \s, or \w, and always match \D, \S, and \W. This is true even when Uni-
2619 code character property support is available. The use of locales with
2620 Unicode is discouraged.
2621
2622 Unicode character properties
2623
2624 When PCRE is built with Unicode character property support, three addi-
2625 tional escape sequences to match character properties are available
2626 when UTF-8 mode is selected. They are:
2627
2628 \p{xx} a character with the xx property
2629 \P{xx} a character without the xx property
2630 \X an extended Unicode sequence
2631
2632 The property names represented by xx above are limited to the Unicode
2633 script names, the general category properties, and "Any", which matches
2634 any character (including newline). Other properties such as "InMusical-
2635 Symbols" are not currently supported by PCRE. Note that \P{Any} does
2636 not match any characters, so always causes a match failure.
2637
2638 Sets of Unicode characters are defined as belonging to certain scripts.
2639 A character from one of these sets can be matched using a script name.
2640 For example:
2641
2642 \p{Greek}
2643 \P{Han}
2644
2645 Those that are not part of an identified script are lumped together as
2646 "Common". The current list of scripts is:
2647
2648 Arabic, Armenian, Bengali, Bopomofo, Braille, Buginese, Buhid, Cana-
2649 dian_Aboriginal, Cherokee, Common, Coptic, Cypriot, Cyrillic, Deseret,
2650 Devanagari, Ethiopic, Georgian, Glagolitic, Gothic, Greek, Gujarati,
2651 Gurmukhi, Han, Hangul, Hanunoo, Hebrew, Hiragana, Inherited, Kannada,
2652 Katakana, Kharoshthi, Khmer, Lao, Latin, Limbu, Linear_B, Malayalam,
2653 Mongolian, Myanmar, New_Tai_Lue, Ogham, Old_Italic, Old_Persian, Oriya,
2654 Osmanya, Runic, Shavian, Sinhala, Syloti_Nagri, Syriac, Tagalog, Tag-
2655 banwa, Tai_Le, Tamil, Telugu, Thaana, Thai, Tibetan, Tifinagh,
2656 Ugaritic, Yi.
2657
2658 Each character has exactly one general category property, specified by
2659 a two-letter abbreviation. For compatibility with Perl, negation can be
2660 specified by including a circumflex between the opening brace and the
2661 property name. For example, \p{^Lu} is the same as \P{Lu}.
2662
2663 If only one letter is specified with \p or \P, it includes all the gen-
2664 eral category properties that start with that letter. In this case, in
2665 the absence of negation, the curly brackets in the escape sequence are
2666 optional; these two examples have the same effect:
2667
2668 \p{L}
2669 \pL
2670
2671 The following general category property codes are supported:
2672
2673 C Other
2674 Cc Control
2675 Cf Format
2676 Cn Unassigned
2677 Co Private use
2678 Cs Surrogate
2679
2680 L Letter
2681 Ll Lower case letter
2682 Lm Modifier letter
2683 Lo Other letter
2684 Lt Title case letter
2685 Lu Upper case letter
2686
2687 M Mark
2688 Mc Spacing mark
2689 Me Enclosing mark
2690 Mn Non-spacing mark
2691
2692 N Number
2693 Nd Decimal number
2694 Nl Letter number
2695 No Other number
2696
2697 P Punctuation
2698 Pc Connector punctuation
2699 Pd Dash punctuation
2700 Pe Close punctuation
2701 Pf Final punctuation
2702 Pi Initial punctuation
2703 Po Other punctuation
2704 Ps Open punctuation
2705
2706 S Symbol
2707 Sc Currency symbol
2708 Sk Modifier symbol
2709 Sm Mathematical symbol
2710 So Other symbol
2711
2712 Z Separator
2713 Zl Line separator
2714 Zp Paragraph separator
2715 Zs Space separator
2716
2717 The special property L& is also supported: it matches a character that
2718 has the Lu, Ll, or Lt property, in other words, a letter that is not
2719 classified as a modifier or "other".
2720
2721 The long synonyms for these properties that Perl supports (such as
2722 \p{Letter}) are not supported by PCRE. Nor is is permitted to prefix
2723 any of these properties with "Is".
2724
2725 No character that is in the Unicode table has the Cn (unassigned) prop-
2726 erty. Instead, this property is assumed for any code point that is not
2727 in the Unicode table.
2728
2729 Specifying caseless matching does not affect these escape sequences.
2730 For example, \p{Lu} always matches only upper case letters.
2731
2732 The \X escape matches any number of Unicode characters that form an
2733 extended Unicode sequence. \X is equivalent to
2734
2735 (?>\PM\pM*)
2736
2737 That is, it matches a character without the "mark" property, followed
2738 by zero or more characters with the "mark" property, and treats the
2739 sequence as an atomic group (see below). Characters with the "mark"
2740 property are typically accents that affect the preceding character.
2741
2742 Matching characters by Unicode property is not fast, because PCRE has
2743 to search a structure that contains data for over fifteen thousand
2744 characters. That is why the traditional escape sequences such as \d and
2745 \w do not use Unicode properties in PCRE.
2746
2747 Simple assertions
2748
2749 The fourth use of backslash is for certain simple assertions. An asser-
2750 tion specifies a condition that has to be met at a particular point in
2751 a match, without consuming any characters from the subject string. The
2752 use of subpatterns for more complicated assertions is described below.
2753 The backslashed assertions are:
2754
2755 \b matches at a word boundary
2756 \B matches when not at a word boundary
2757 \A matches at start of subject
2758 \Z matches at end of subject or before newline at end
2759 \z matches at end of subject
2760 \G matches at first matching position in subject
2761
2762 These assertions may not appear in character classes (but note that \b
2763 has a different meaning, namely the backspace character, inside a char-
2764 acter class).
2765
2766 A word boundary is a position in the subject string where the current
2767 character and the previous character do not both match \w or \W (i.e.
2768 one matches \w and the other matches \W), or the start or end of the
2769 string if the first or last character matches \w, respectively.
2770
2771 The \A, \Z, and \z assertions differ from the traditional circumflex
2772 and dollar (described in the next section) in that they only ever match
2773 at the very start and end of the subject string, whatever options are
2774 set. Thus, they are independent of multiline mode. These three asser-
2775 tions are not affected by the PCRE_NOTBOL or PCRE_NOTEOL options, which
2776 affect only the behaviour of the circumflex and dollar metacharacters.
2777 However, if the startoffset argument of pcre_exec() is non-zero, indi-
2778 cating that matching is to start at a point other than the beginning of
2779 the subject, \A can never match. The difference between \Z and \z is
2780 that \Z matches before a newline that is the last character of the
2781 string as well as at the end of the string, whereas \z matches only at
2782 the end.
2783
2784 The \G assertion is true only when the current matching position is at
2785 the start point of the match, as specified by the startoffset argument
2786 of pcre_exec(). It differs from \A when the value of startoffset is
2787 non-zero. By calling pcre_exec() multiple times with appropriate argu-
2788 ments, you can mimic Perl's /g option, and it is in this kind of imple-
2789 mentation where \G can be useful.
2790
2791 Note, however, that PCRE's interpretation of \G, as the start of the
2792 current match, is subtly different from Perl's, which defines it as the
2793 end of the previous match. In Perl, these can be different when the
2794 previously matched string was empty. Because PCRE does just one match
2795 at a time, it cannot reproduce this behaviour.
2796
2797 If all the alternatives of a pattern begin with \G, the expression is
2798 anchored to the starting match position, and the "anchored" flag is set
2799 in the compiled regular expression.
2800
2801
2802 CIRCUMFLEX AND DOLLAR
2803
2804 Outside a character class, in the default matching mode, the circumflex
2805 character is an assertion that is true only if the current matching
2806 point is at the start of the subject string. If the startoffset argu-
2807 ment of pcre_exec() is non-zero, circumflex can never match if the
2808 PCRE_MULTILINE option is unset. Inside a character class, circumflex
2809 has an entirely different meaning (see below).
2810
2811 Circumflex need not be the first character of the pattern if a number
2812 of alternatives are involved, but it should be the first thing in each
2813 alternative in which it appears if the pattern is ever to match that
2814 branch. If all possible alternatives start with a circumflex, that is,
2815 if the pattern is constrained to match only at the start of the sub-
2816 ject, it is said to be an "anchored" pattern. (There are also other
2817 constructs that can cause a pattern to be anchored.)
2818
2819 A dollar character is an assertion that is true only if the current
2820 matching point is at the end of the subject string, or immediately
2821 before a newline character that is the last character in the string (by
2822 default). Dollar need not be the last character of the pattern if a
2823 number of alternatives are involved, but it should be the last item in
2824 any branch in which it appears. Dollar has no special meaning in a
2825 character class.
2826
2827 The meaning of dollar can be changed so that it matches only at the
2828 very end of the string, by setting the PCRE_DOLLAR_ENDONLY option at
2829 compile time. This does not affect the \Z assertion.
2830
2831 The meanings of the circumflex and dollar characters are changed if the
2832 PCRE_MULTILINE option is set. When this is the case, they match immedi-
2833 ately after and immediately before an internal newline character,
2834 respectively, in addition to matching at the start and end of the sub-
2835 ject string. For example, the pattern /^abc$/ matches the subject
2836 string "def\nabc" (where \n represents a newline character) in multi-
2837 line mode, but not otherwise. Consequently, patterns that are anchored
2838 in single line mode because all branches start with ^ are not anchored
2839 in multiline mode, and a match for circumflex is possible when the
2840 startoffset argument of pcre_exec() is non-zero. The PCRE_DOL-
2841 LAR_ENDONLY option is ignored if PCRE_MULTILINE is set.
2842
2843 Note that the sequences \A, \Z, and \z can be used to match the start
2844 and end of the subject in both modes, and if all branches of a pattern
2845 start with \A it is always anchored, whether PCRE_MULTILINE is set or
2846 not.
2847
2848
2849 FULL STOP (PERIOD, DOT)
2850
2851 Outside a character class, a dot in the pattern matches any one charac-
2852 ter in the subject, including a non-printing character, but not (by
2853 default) newline. In UTF-8 mode, a dot matches any UTF-8 character,
2854 which might be more than one byte long, except (by default) newline. If
2855 the PCRE_DOTALL option is set, dots match newlines as well. The han-
2856 dling of dot is entirely independent of the handling of circumflex and
2857 dollar, the only relationship being that they both involve newline
2858 characters. Dot has no special meaning in a character class.
2859
2860
2861 MATCHING A SINGLE BYTE
2862
2863 Outside a character class, the escape sequence \C matches any one byte,
2864 both in and out of UTF-8 mode. Unlike a dot, it can match a newline.
2865 The feature is provided in Perl in order to match individual bytes in
2866 UTF-8 mode. Because it breaks up UTF-8 characters into individual
2867 bytes, what remains in the string may be a malformed UTF-8 string. For
2868 this reason, the \C escape sequence is best avoided.
2869
2870 PCRE does not allow \C to appear in lookbehind assertions (described
2871 below), because in UTF-8 mode this would make it impossible to calcu-
2872 late the length of the lookbehind.
2873
2874
2875 SQUARE BRACKETS AND CHARACTER CLASSES
2876
2877 An opening square bracket introduces a character class, terminated by a
2878 closing square bracket. A closing square bracket on its own is not spe-
2879 cial. If a closing square bracket is required as a member of the class,
2880 it should be the first data character in the class (after an initial
2881 circumflex, if present) or escaped with a backslash.
2882
2883 A character class matches a single character in the subject. In UTF-8
2884 mode, the character may occupy more than one byte. A matched character
2885 must be in the set of characters defined by the class, unless the first
2886 character in the class definition is a circumflex, in which case the
2887 subject character must not be in the set defined by the class. If a
2888 circumflex is actually required as a member of the class, ensure it is
2889 not the first character, or escape it with a backslash.
2890
2891 For example, the character class [aeiou] matches any lower case vowel,
2892 while [^aeiou] matches any character that is not a lower case vowel.
2893 Note that a circumflex is just a convenient notation for specifying the
2894 characters that are in the class by enumerating those that are not. A
2895 class that starts with a circumflex is not an assertion: it still con-
2896 sumes a character from the subject string, and therefore it fails if
2897 the current pointer is at the end of the string.
2898
2899 In UTF-8 mode, characters with values greater than 255 can be included
2900 in a class as a literal string of bytes, or by using the \x{ escaping
2901 mechanism.
2902
2903 When caseless matching is set, any letters in a class represent both
2904 their upper case and lower case versions, so for example, a caseless
2905 [aeiou] matches "A" as well as "a", and a caseless [^aeiou] does not
2906 match "A", whereas a caseful version would. In UTF-8 mode, PCRE always
2907 understands the concept of case for characters whose values are less
2908 than 128, so caseless matching is always possible. For characters with
2909 higher values, the concept of case is supported if PCRE is compiled
2910 with Unicode property support, but not otherwise. If you want to use
2911 caseless matching for characters 128 and above, you must ensure that
2912 PCRE is compiled with Unicode property support as well as with UTF-8
2913 support.
2914
2915 The newline character is never treated in any special way in character
2916 classes, whatever the setting of the PCRE_DOTALL or PCRE_MULTILINE
2917 options is. A class such as [^a] will always match a newline.
2918
2919 The minus (hyphen) character can be used to specify a range of charac-
2920 ters in a character class. For example, [d-m] matches any letter
2921 between d and m, inclusive. If a minus character is required in a
2922 class, it must be escaped with a backslash or appear in a position
2923 where it cannot be interpreted as indicating a range, typically as the
2924 first or last character in the class.
2925
2926 It is not possible to have the literal character "]" as the end charac-
2927 ter of a range. A pattern such as [W-]46] is interpreted as a class of
2928 two characters ("W" and "-") followed by a literal string "46]", so it
2929 would match "W46]" or "-46]". However, if the "]" is escaped with a
2930 backslash it is interpreted as the end of range, so [W-\]46] is inter-
2931 preted as a class containing a range followed by two other characters.
2932 The octal or hexadecimal representation of "]" can also be used to end
2933 a range.
2934
2935 Ranges operate in the collating sequence of character values. They can
2936 also be used for characters specified numerically, for example
2937 [\000-\037]. In UTF-8 mode, ranges can include characters whose values
2938 are greater than 255, for example [\x{100}-\x{2ff}].
2939
2940 If a range that includes letters is used when caseless matching is set,
2941 it matches the letters in either case. For example, [W-c] is equivalent
2942 to [][\\^_`wxyzabc], matched caselessly, and in non-UTF-8 mode, if
2943 character tables for the "fr_FR" locale are in use, [\xc8-\xcb] matches
2944 accented E characters in both cases. In UTF-8 mode, PCRE supports the
2945 concept of case for characters with values greater than 128 only when
2946 it is compiled with Unicode property support.
2947
2948 The character types \d, \D, \p, \P, \s, \S, \w, and \W may also appear
2949 in a character class, and add the characters that they match to the
2950 class. For example, [\dABCDEF] matches any hexadecimal digit. A circum-
2951 flex can conveniently be used with the upper case character types to
2952 specify a more restricted set of characters than the matching lower
2953 case type. For example, the class [^\W_] matches any letter or digit,
2954 but not underscore.
2955
2956 The only metacharacters that are recognized in character classes are
2957 backslash, hyphen (only where it can be interpreted as specifying a
2958 range), circumflex (only at the start), opening square bracket (only
2959 when it can be interpreted as introducing a POSIX class name - see the
2960 next section), and the terminating closing square bracket. However,
2961 escaping other non-alphanumeric characters does no harm.
2962
2963
2964 POSIX CHARACTER CLASSES
2965
2966 Perl supports the POSIX notation for character classes. This uses names
2967 enclosed by [: and :] within the enclosing square brackets. PCRE also
2968 supports this notation. For example,
2969
2970 [01[:alpha:]%]
2971
2972 matches "0", "1", any alphabetic character, or "%". The supported class
2973 names are
2974
2975 alnum letters and digits
2976 alpha letters
2977 ascii character codes 0 - 127
2978 blank space or tab only
2979 cntrl control characters
2980 digit decimal digits (same as \d)
2981 graph printing characters, excluding space
2982 lower lower case letters
2983 print printing characters, including space
2984 punct printing characters, excluding letters and digits
2985 space white space (not quite the same as \s)
2986 upper upper case letters
2987 word "word" characters (same as \w)
2988 xdigit hexadecimal digits
2989
2990 The "space" characters are HT (9), LF (10), VT (11), FF (12), CR (13),
2991 and space (32). Notice that this list includes the VT character (code
2992 11). This makes "space" different to \s, which does not include VT (for
2993 Perl compatibility).
2994
2995 The name "word" is a Perl extension, and "blank" is a GNU extension
2996 from Perl 5.8. Another Perl extension is negation, which is indicated
2997 by a ^ character after the colon. For example,
2998
2999 [12[:^digit:]]
3000
3001 matches "1", "2", or any non-digit. PCRE (and Perl) also recognize the
3002 POSIX syntax [.ch.] and [=ch=] where "ch" is a "collating element", but
3003 these are not supported, and an error is given if they are encountered.
3004
3005 In UTF-8 mode, characters with values greater than 128 do not match any
3006 of the POSIX character classes.
3007
3008
3009 VERTICAL BAR
3010
3011 Vertical bar characters are used to separate alternative patterns. For
3012 example, the pattern
3013
3014 gilbert|sullivan
3015
3016 matches either "gilbert" or "sullivan". Any number of alternatives may
3017 appear, and an empty alternative is permitted (matching the empty
3018 string). The matching process tries each alternative in turn, from
3019 left to right, and the first one that succeeds is used. If the alterna-
3020 tives are within a subpattern (defined below), "succeeds" means match-
3021 ing the rest of the main pattern as well as the alternative in the sub-
3022 pattern.
3023
3024
3025 INTERNAL OPTION SETTING
3026
3027 The settings of the PCRE_CASELESS, PCRE_MULTILINE, PCRE_DOTALL, and
3028 PCRE_EXTENDED options can be changed from within the pattern by a
3029 sequence of Perl option letters enclosed between "(?" and ")". The
3030 option letters are
3031
3032 i for PCRE_CASELESS
3033 m for PCRE_MULTILINE
3034 s for PCRE_DOTALL
3035 x for PCRE_EXTENDED
3036
3037 For example, (?im) sets caseless, multiline matching. It is also possi-
3038 ble to unset these options by preceding the letter with a hyphen, and a
3039 combined setting and unsetting such as (?im-sx), which sets PCRE_CASE-
3040 LESS and PCRE_MULTILINE while unsetting PCRE_DOTALL and PCRE_EXTENDED,
3041 is also permitted. If a letter appears both before and after the
3042 hyphen, the option is unset.
3043
3044 When an option change occurs at top level (that is, not inside subpat-
3045 tern parentheses), the change applies to the remainder of the pattern
3046 that follows. If the change is placed right at the start of a pattern,
3047 PCRE extracts it into the global options (and it will therefore show up
3048 in data extracted by the pcre_fullinfo() function).
3049
3050 An option change within a subpattern affects only that part of the cur-
3051 rent pattern that follows it, so
3052
3053 (a(?i)b)c
3054
3055 matches abc and aBc and no other strings (assuming PCRE_CASELESS is not
3056 used). By this means, options can be made to have different settings
3057 in different parts of the pattern. Any changes made in one alternative
3058 do carry on into subsequent branches within the same subpattern. For
3059 example,
3060
3061 (a(?i)b|c)
3062
3063 matches "ab", "aB", "c", and "C", even though when matching "C" the
3064 first branch is abandoned before the option setting. This is because
3065 the effects of option settings happen at compile time. There would be
3066 some very weird behaviour otherwise.
3067
3068 The PCRE-specific options PCRE_UNGREEDY and PCRE_EXTRA can be changed
3069 in the same way as the Perl-compatible options by using the characters
3070 U and X respectively. The (?X) flag setting is special in that it must
3071 always occur earlier in the pattern than any of the additional features
3072 it turns on, even when it is at top level. It is best to put it at the
3073 start.
3074
3075
3076 SUBPATTERNS
3077
3078 Subpatterns are delimited by parentheses (round brackets), which can be
3079 nested. Turning part of a pattern into a subpattern does two things:
3080
3081 1. It localizes a set of alternatives. For example, the pattern
3082
3083 cat(aract|erpillar|)
3084
3085 matches one of the words "cat", "cataract", or "caterpillar". Without
3086 the parentheses, it would match "cataract", "erpillar" or the empty
3087 string.
3088
3089 2. It sets up the subpattern as a capturing subpattern. This means
3090 that, when the whole pattern matches, that portion of the subject
3091 string that matched the subpattern is passed back to the caller via the
3092 ovector argument of pcre_exec(). Opening parentheses are counted from
3093 left to right (starting from 1) to obtain numbers for the capturing
3094 subpatterns.
3095
3096 For example, if the string "the red king" is matched against the pat-
3097 tern
3098
3099 the ((red|white) (king|queen))
3100
3101 the captured substrings are "red king", "red", and "king", and are num-
3102 bered 1, 2, and 3, respectively.
3103
3104 The fact that plain parentheses fulfil two functions is not always
3105 helpful. There are often times when a grouping subpattern is required
3106 without a capturing requirement. If an opening parenthesis is followed
3107 by a question mark and a colon, the subpattern does not do any captur-
3108 ing, and is not counted when computing the number of any subsequent
3109 capturing subpatterns. For example, if the string "the white queen" is
3110 matched against the pattern
3111
3112 the ((?:red|white) (king|queen))
3113
3114 the captured substrings are "white queen" and "queen", and are numbered
3115 1 and 2. The maximum number of capturing subpatterns is 65535, and the
3116 maximum depth of nesting of all subpatterns, both capturing and non-
3117 capturing, is 200.
3118
3119 As a convenient shorthand, if any option settings are required at the
3120 start of a non-capturing subpattern, the option letters may appear
3121 between the "?" and the ":". Thus the two patterns
3122
3123 (?i:saturday|sunday)
3124 (?:(?i)saturday|sunday)
3125
3126 match exactly the same set of strings. Because alternative branches are
3127 tried from left to right, and options are not reset until the end of
3128 the subpattern is reached, an option setting in one branch does affect
3129 subsequent branches, so the above patterns match "SUNDAY" as well as
3130 "Saturday".
3131
3132
3133 NAMED SUBPATTERNS
3134
3135 Identifying capturing parentheses by number is simple, but it can be
3136 very hard to keep track of the numbers in complicated regular expres-
3137 sions. Furthermore, if an expression is modified, the numbers may
3138 change. To help with this difficulty, PCRE supports the naming of sub-
3139 patterns, something that Perl does not provide. The Python syntax
3140 (?P<name>...) is used. Names consist of alphanumeric characters and
3141 underscores, and must be unique within a pattern.
3142
3143 Named capturing parentheses are still allocated numbers as well as
3144 names. The PCRE API provides function calls for extracting the name-to-
3145 number translation table from a compiled pattern. There is also a con-
3146 venience function for extracting a captured substring by name. For fur-
3147 ther details see the pcreapi documentation.
3148
3149
3150 REPETITION
3151
3152 Repetition is specified by quantifiers, which can follow any of the
3153 following items:
3154
3155 a literal data character
3156 the . metacharacter
3157 the \C escape sequence
3158 the \X escape sequence (in UTF-8 mode with Unicode properties)
3159 an escape such as \d that matches a single character
3160 a character class
3161 a back reference (see next section)
3162 a parenthesized subpattern (unless it is an assertion)
3163
3164 The general repetition quantifier specifies a minimum and maximum num-
3165 ber of permitted matches, by giving the two numbers in curly brackets
3166 (braces), separated by a comma. The numbers must be less than 65536,
3167 and the first must be less than or equal to the second. For example:
3168
3169 z{2,4}
3170
3171 matches "zz", "zzz", or "zzzz". A closing brace on its own is not a
3172 special character. If the second number is omitted, but the comma is
3173 present, there is no upper limit; if the second number and the comma
3174 are both omitted, the quantifier specifies an exact number of required
3175 matches. Thus
3176
3177 [aeiou]{3,}
3178
3179 matches at least 3 successive vowels, but may match many more, while
3180
3181 \d{8}
3182
3183 matches exactly 8 digits. An opening curly bracket that appears in a
3184 position where a quantifier is not allowed, or one that does not match
3185 the syntax of a quantifier, is taken as a literal character. For exam-
3186 ple, {,6} is not a quantifier, but a literal string of four characters.
3187
3188 In UTF-8 mode, quantifiers apply to UTF-8 characters rather than to
3189 individual bytes. Thus, for example, \x{100}{2} matches two UTF-8 char-
3190 acters, each of which is represented by a two-byte sequence. Similarly,
3191 when Unicode property support is available, \X{3} matches three Unicode
3192 extended sequences, each of which may be several bytes long (and they
3193 may be of different lengths).
3194
3195 The quantifier {0} is permitted, causing the expression to behave as if
3196 the previous item and the quantifier were not present.
3197
3198 For convenience (and historical compatibility) the three most common
3199 quantifiers have single-character abbreviations:
3200
3201 * is equivalent to {0,}
3202 + is equivalent to {1,}
3203 ? is equivalent to {0,1}
3204
3205 It is possible to construct infinite loops by following a subpattern
3206 that can match no characters with a quantifier that has no upper limit,
3207 for example:
3208
3209 (a?)*
3210
3211 Earlier versions of Perl and PCRE used to give an error at compile time
3212 for such patterns. However, because there are cases where this can be
3213 useful, such patterns are now accepted, but if any repetition of the
3214 subpattern does in fact match no characters, the loop is forcibly bro-
3215 ken.
3216
3217 By default, the quantifiers are "greedy", that is, they match as much
3218 as possible (up to the maximum number of permitted times), without
3219 causing the rest of the pattern to fail. The classic example of where
3220 this gives problems is in trying to match comments in C programs. These
3221 appear between /* and */ and within the comment, individual * and /
3222 characters may appear. An attempt to match C comments by applying the
3223 pattern
3224
3225 /\*.*\*/
3226
3227 to the string
3228
3229 /* first comment */ not comment /* second comment */
3230
3231 fails, because it matches the entire string owing to the greediness of
3232 the .* item.
3233
3234 However, if a quantifier is followed by a question mark, it ceases to
3235 be greedy, and instead matches the minimum number of times possible, so
3236 the pattern
3237
3238 /\*.*?\*/
3239
3240 does the right thing with the C comments. The meaning of the various
3241 quantifiers is not otherwise changed, just the preferred number of
3242 matches. Do not confuse this use of question mark with its use as a
3243 quantifier in its own right. Because it has two uses, it can sometimes
3244 appear doubled, as in
3245
3246 \d??\d
3247
3248 which matches one digit by preference, but can match two if that is the
3249 only way the rest of the pattern matches.
3250
3251 If the PCRE_UNGREEDY option is set (an option which is not available in
3252 Perl), the quantifiers are not greedy by default, but individual ones
3253 can be made greedy by following them with a question mark. In other
3254 words, it inverts the default behaviour.
3255
3256 When a parenthesized subpattern is quantified with a minimum repeat
3257 count that is greater than 1 or with a limited maximum, more memory is
3258 required for the compiled pattern, in proportion to the size of the
3259 minimum or maximum.
3260
3261 If a pattern starts with .* or .{0,} and the PCRE_DOTALL option (equiv-
3262 alent to Perl's /s) is set, thus allowing the . to match newlines, the
3263 pattern is implicitly anchored, because whatever follows will be tried
3264 against every character position in the subject string, so there is no
3265 point in retrying the overall match at any position after the first.
3266 PCRE normally treats such a pattern as though it were preceded by \A.
3267
3268 In cases where it is known that the subject string contains no new-
3269 lines, it is worth setting PCRE_DOTALL in order to obtain this opti-
3270 mization, or alternatively using ^ to indicate anchoring explicitly.
3271
3272 However, there is one situation where the optimization cannot be used.
3273 When .* is inside capturing parentheses that are the subject of a
3274 backreference elsewhere in the pattern, a match at the start may fail,
3275 and a later one succeed. Consider, for example:
3276
3277 (.*)abc\1
3278
3279 If the subject is "xyz123abc123" the match point is the fourth charac-
3280 ter. For this reason, such a pattern is not implicitly anchored.
3281
3282 When a capturing subpattern is repeated, the value captured is the sub-
3283 string that matched the final iteration. For example, after
3284
3285 (tweedle[dume]{3}\s*)+
3286
3287 has matched "tweedledum tweedledee" the value of the captured substring
3288 is "tweedledee". However, if there are nested capturing subpatterns,
3289 the corresponding captured values may have been set in previous itera-
3290 tions. For example, after
3291
3292 /(a|(b))+/
3293
3294 matches "aba" the value of the second captured substring is "b".
3295
3296
3297 ATOMIC GROUPING AND POSSESSIVE QUANTIFIERS
3298
3299 With both maximizing and minimizing repetition, failure of what follows
3300 normally causes the repeated item to be re-evaluated to see if a dif-
3301 ferent number of repeats allows the rest of the pattern to match. Some-
3302 times it is useful to prevent this, either to change the nature of the
3303 match, or to cause it fail earlier than it otherwise might, when the
3304 author of the pattern knows there is no point in carrying on.
3305
3306 Consider, for example, the pattern \d+foo when applied to the subject
3307 line
3308
3309 123456bar
3310
3311 After matching all 6 digits and then failing to match "foo", the normal
3312 action of the matcher is to try again with only 5 digits matching the
3313 \d+ item, and then with 4, and so on, before ultimately failing.
3314 "Atomic grouping" (a term taken from Jeffrey Friedl's book) provides
3315 the means for specifying that once a subpattern has matched, it is not
3316 to be re-evaluated in this way.
3317
3318 If we use atomic grouping for the previous example, the matcher would
3319 give up immediately on failing to match "foo" the first time. The nota-
3320 tion is a kind of special parenthesis, starting with (?> as in this
3321 example:
3322
3323 (?>\d+)foo
3324
3325 This kind of parenthesis "locks up" the part of the pattern it con-
3326 tains once it has matched, and a failure further into the pattern is
3327 prevented from backtracking into it. Backtracking past it to previous
3328 items, however, works as normal.
3329
3330 An alternative description is that a subpattern of this type matches
3331 the string of characters that an identical standalone pattern would
3332 match, if anchored at the current point in the subject string.
3333
3334 Atomic grouping subpatterns are not capturing subpatterns. Simple cases
3335 such as the above example can be thought of as a maximizing repeat that
3336 must swallow everything it can. So, while both \d+ and \d+? are pre-
3337 pared to adjust the number of digits they match in order to make the
3338 rest of the pattern match, (?>\d+) can only match an entire sequence of
3339 digits.
3340
3341 Atomic groups in general can of course contain arbitrarily complicated
3342 subpatterns, and can be nested. However, when the subpattern for an
3343 atomic group is just a single repeated item, as in the example above, a
3344 simpler notation, called a "possessive quantifier" can be used. This
3345 consists of an additional + character following a quantifier. Using
3346 this notation, the previous example can be rewritten as
3347
3348 \d++foo
3349
3350 Possessive quantifiers are always greedy; the setting of the
3351 PCRE_UNGREEDY option is ignored. They are a convenient notation for the
3352 simpler forms of atomic group. However, there is no difference in the
3353 meaning or processing of a possessive quantifier and the equivalent
3354 atomic group.
3355
3356 The possessive quantifier syntax is an extension to the Perl syntax. It
3357 originates in Sun's Java package.
3358
3359 When a pattern contains an unlimited repeat inside a subpattern that
3360 can itself be repeated an unlimited number of times, the use of an
3361 atomic group is the only way to avoid some failing matches taking a
3362 very long time indeed. The pattern
3363
3364 (\D+|<\d+>)*[!?]
3365
3366 matches an unlimited number of substrings that either consist of non-
3367 digits, or digits enclosed in <>, followed by either ! or ?. When it
3368 matches, it runs quickly. However, if it is applied to
3369
3370 aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
3371
3372 it takes a long time before reporting failure. This is because the
3373 string can be divided between the internal \D+ repeat and the external
3374 * repeat in a large number of ways, and all have to be tried. (The
3375 example uses [!?] rather than a single character at the end, because
3376 both PCRE and Perl have an optimization that allows for fast failure
3377 when a single character is used. They remember the last single charac-
3378 ter that is required for a match, and fail early if it is not present
3379 in the string.) If the pattern is changed so that it uses an atomic
3380 group, like this:
3381
3382 ((?>\D+)|<\d+>)*[!?]
3383
3384 sequences of non-digits cannot be broken, and failure happens quickly.
3385
3386
3387 BACK REFERENCES
3388
3389 Outside a character class, a backslash followed by a digit greater than
3390 0 (and possibly further digits) is a back reference to a capturing sub-
3391 pattern earlier (that is, to its left) in the pattern, provided there
3392 have been that many previous capturing left parentheses.
3393
3394 However, if the decimal number following the backslash is less than 10,
3395 it is always taken as a back reference, and causes an error only if
3396 there are not that many capturing left parentheses in the entire pat-
3397 tern. In other words, the parentheses that are referenced need not be
3398 to the left of the reference for numbers less than 10. See the subsec-
3399 tion entitled "Non-printing characters" above for further details of
3400 the handling of digits following a backslash.
3401
3402 A back reference matches whatever actually matched the capturing sub-
3403 pattern in the current subject string, rather than anything matching
3404 the subpattern itself (see "Subpatterns as subroutines" below for a way
3405 of doing that). So the pattern
3406
3407 (sens|respons)e and \1ibility
3408
3409 matches "sense and sensibility" and "response and responsibility", but
3410 not "sense and responsibility". If caseful matching is in force at the
3411 time of the back reference, the case of letters is relevant. For exam-
3412 ple,
3413
3414 ((?i)rah)\s+\1
3415
3416 matches "rah rah" and "RAH RAH", but not "RAH rah", even though the
3417 original capturing subpattern is matched caselessly.
3418
3419 Back references to named subpatterns use the Python syntax (?P=name).
3420 We could rewrite the above example as follows:
3421
3422 (?<p1>(?i)rah)\s+(?P=p1)
3423
3424 There may be more than one back reference to the same subpattern. If a
3425 subpattern has not actually been used in a particular match, any back
3426 references to it always fail. For example, the pattern
3427
3428 (a|(bc))\2
3429
3430 always fails if it starts to match "a" rather than "bc". Because there
3431 may be many capturing parentheses in a pattern, all digits following
3432 the backslash are taken as part of a potential back reference number.
3433 If the pattern continues with a digit character, some delimiter must be
3434 used to terminate the back reference. If the PCRE_EXTENDED option is
3435 set, this can be whitespace. Otherwise an empty comment (see "Com-
3436 ments" below) can be used.
3437
3438 A back reference that occurs inside the parentheses to which it refers
3439 fails when the subpattern is first used, so, for example, (a\1) never
3440 matches. However, such references can be useful inside repeated sub-
3441 patterns. For example, the pattern
3442
3443 (a|b\1)+
3444
3445 matches any number of "a"s and also "aba", "ababbaa" etc. At each iter-
3446 ation of the subpattern, the back reference matches the character
3447 string corresponding to the previous iteration. In order for this to
3448 work, the pattern must be such that the first iteration does not need
3449 to match the back reference. This can be done using alternation, as in
3450 the example above, or by a quantifier with a minimum of zero.
3451
3452
3453 ASSERTIONS
3454
3455 An assertion is a test on the characters following or preceding the
3456 current matching point that does not actually consume any characters.
3457 The simple assertions coded as \b, \B, \A, \G, \Z, \z, ^ and $ are
3458 described above.
3459
3460 More complicated assertions are coded as subpatterns. There are two
3461 kinds: those that look ahead of the current position in the subject
3462 string, and those that look behind it. An assertion subpattern is
3463 matched in the normal way, except that it does not cause the current
3464 matching position to be changed.
3465
3466 Assertion subpatterns are not capturing subpatterns, and may not be
3467 repeated, because it makes no sense to assert the same thing several
3468 times. If any kind of assertion contains capturing subpatterns within
3469 it, these are counted for the purposes of numbering the capturing sub-
3470 patterns in the whole pattern. However, substring capturing is carried
3471 out only for positive assertions, because it does not make sense for
3472 negative assertions.
3473
3474 Lookahead assertions
3475
3476 Lookahead assertions start with (?= for positive assertions and (?! for
3477 negative assertions. For example,
3478
3479 \w+(?=;)
3480
3481 matches a word followed by a semicolon, but does not include the semi-
3482 colon in the match, and
3483
3484 foo(?!bar)
3485
3486 matches any occurrence of "foo" that is not followed by "bar". Note
3487 that the apparently similar pattern
3488
3489 (?!foo)bar
3490
3491 does not find an occurrence of "bar" that is preceded by something
3492 other than "foo"; it finds any occurrence of "bar" whatsoever, because
3493 the assertion (?!foo) is always true when the next three characters are
3494 "bar". A lookbehind assertion is needed to achieve the other effect.
3495
3496 If you want to force a matching failure at some point in a pattern, the
3497 most convenient way to do it is with (?!) because an empty string
3498 always matches, so an assertion that requires there not to be an empty
3499 string must always fail.
3500
3501 Lookbehind assertions
3502
3503 Lookbehind assertions start with (?<= for positive assertions and (?<!
3504 for negative assertions. For example,
3505
3506 (?<!foo)bar
3507
3508 does find an occurrence of "bar" that is not preceded by "foo". The
3509 contents of a lookbehind assertion are restricted such that all the
3510 strings it matches must have a fixed length. However, if there are sev-
3511 eral alternatives, they do not all have to have the same fixed length.
3512 Thus
3513
3514 (?<=bullock|donkey)
3515
3516 is permitted, but
3517
3518 (?<!dogs?|cats?)
3519
3520 causes an error at compile time. Branches that match different length
3521 strings are permitted only at the top level of a lookbehind assertion.
3522 This is an extension compared with Perl (at least for 5.8), which
3523 requires all branches to match the same length of string. An assertion
3524 such as
3525
3526 (?<=ab(c|de))
3527
3528 is not permitted, because its single top-level branch can match two
3529 different lengths, but it is acceptable if rewritten to use two top-
3530 level branches:
3531
3532 (?<=abc|abde)
3533
3534 The implementation of lookbehind assertions is, for each alternative,
3535 to temporarily move the current position back by the fixed width and
3536 then try to match. If there are insufficient characters before the cur-
3537 rent position, the match is deemed to fail.
3538
3539 PCRE does not allow the \C escape (which matches a single byte in UTF-8
3540 mode) to appear in lookbehind assertions, because it makes it impossi-
3541 ble to calculate the length of the lookbehind. The \X escape, which can
3542 match different numbers of bytes, is also not permitted.
3543
3544 Atomic groups can be used in conjunction with lookbehind assertions to
3545 specify efficient matching at the end of the subject string. Consider a
3546 simple pattern such as
3547
3548 abcd$
3549
3550 when applied to a long string that does not match. Because matching
3551 proceeds from left to right, PCRE will look for each "a" in the subject
3552 and then see if what follows matches the rest of the pattern. If the
3553 pattern is specified as
3554
3555 ^.*abcd$
3556
3557 the initial .* matches the entire string at first, but when this fails
3558 (because there is no following "a"), it backtracks to match all but the
3559 last character, then all but the last two characters, and so on. Once
3560 again the search for "a" covers the entire string, from right to left,
3561 so we are no better off. However, if the pattern is written as
3562
3563 ^(?>.*)(?<=abcd)
3564
3565 or, equivalently, using the possessive quantifier syntax,
3566
3567 ^.*+(?<=abcd)
3568
3569 there can be no backtracking for the .* item; it can match only the
3570 entire string. The subsequent lookbehind assertion does a single test
3571 on the last four characters. If it fails, the match fails immediately.
3572 For long strings, this approach makes a significant difference to the
3573 processing time.
3574
3575 Using multiple assertions
3576
3577 Several assertions (of any sort) may occur in succession. For example,
3578
3579 (?<=\d{3})(?<!999)foo
3580
3581 matches "foo" preceded by three digits that are not "999". Notice that
3582 each of the assertions is applied independently at the same point in
3583 the subject string. First there is a check that the previous three
3584 characters are all digits, and then there is a check that the same
3585 three characters are not "999". This pattern does not match "foo" pre-
3586 ceded by six characters, the first of which are digits and the last
3587 three of which are not "999". For example, it doesn't match "123abc-
3588 foo". A pattern to do that is
3589
3590 (?<=\d{3}...)(?<!999)foo
3591
3592 This time the first assertion looks at the preceding six characters,
3593 checking that the first three are digits, and then the second assertion
3594 checks that the preceding three characters are not "999".
3595
3596 Assertions can be nested in any combination. For example,
3597
3598 (?<=(?<!foo)bar)baz
3599
3600 matches an occurrence of "baz" that is preceded by "bar" which in turn
3601 is not preceded by "foo", while
3602
3603 (?<=\d{3}(?!999)...)foo
3604
3605 is another pattern that matches "foo" preceded by three digits and any
3606 three characters that are not "999".
3607
3608
3609 CONDITIONAL SUBPATTERNS
3610
3611 It is possible to cause the matching process to obey a subpattern con-
3612 ditionally or to choose between two alternative subpatterns, depending
3613 on the result of an assertion, or whether a previous capturing subpat-
3614 tern matched or not. The two possible forms of conditional subpattern
3615 are
3616
3617 (?(condition)yes-pattern)
3618 (?(condition)yes-pattern|no-pattern)
3619
3620 If the condition is satisfied, the yes-pattern is used; otherwise the
3621 no-pattern (if present) is used. If there are more than two alterna-
3622 tives in the subpattern, a compile-time error occurs.
3623
3624 There are three kinds of condition. If the text between the parentheses
3625 consists of a sequence of digits, the condition is satisfied if the
3626 capturing subpattern of that number has previously matched. The number
3627 must be greater than zero. Consider the following pattern, which con-
3628 tains non-significant white space to make it more readable (assume the
3629 PCRE_EXTENDED option) and to divide it into three parts for ease of
3630 discussion:
3631
3632 ( \( )? [^()]+ (?(1) \) )
3633
3634 The first part matches an optional opening parenthesis, and if that
3635 character is present, sets it as the first captured substring. The sec-
3636 ond part matches one or more characters that are not parentheses. The
3637 third part is a conditional subpattern that tests whether the first set
3638 of parentheses matched or not. If they did, that is, if subject started
3639 with an opening parenthesis, the condition is true, and so the yes-pat-
3640 tern is executed and a closing parenthesis is required. Otherwise,
3641 since no-pattern is not present, the subpattern matches nothing. In
3642 other words, this pattern matches a sequence of non-parentheses,
3643 optionally enclosed in parentheses.
3644
3645 If the condition is the string (R), it is satisfied if a recursive call
3646 to the pattern or subpattern has been made. At "top level", the condi-
3647 tion is false. This is a PCRE extension. Recursive patterns are
3648 described in the next section.
3649
3650 If the condition is not a sequence of digits or (R), it must be an
3651 assertion. This may be a positive or negative lookahead or lookbehind
3652 assertion. Consider this pattern, again containing non-significant
3653 white space, and with the two alternatives on the second line:
3654
3655 (?(?=[^a-z]*[a-z])
3656 \d{2}-[a-z]{3}-\d{2} | \d{2}-\d{2}-\d{2} )
3657
3658 The condition is a positive lookahead assertion that matches an
3659 optional sequence of non-letters followed by a letter. In other words,
3660 it tests for the presence of at least one letter in the subject. If a
3661 letter is found, the subject is matched against the first alternative;
3662 otherwise it is matched against the second. This pattern matches
3663 strings in one of the two forms dd-aaa-dd or dd-dd-dd, where aaa are
3664 letters and dd are digits.
3665
3666
3667 COMMENTS
3668
3669 The sequence (?# marks the start of a comment that continues up to the
3670 next closing parenthesis. Nested parentheses are not permitted. The
3671 characters that make up a comment play no part in the pattern matching
3672 at all.
3673
3674 If the PCRE_EXTENDED option is set, an unescaped # character outside a
3675 character class introduces a comment that continues up to the next new-
3676 line character in the pattern.
3677
3678
3679 RECURSIVE PATTERNS
3680
3681 Consider the problem of matching a string in parentheses, allowing for
3682 unlimited nested parentheses. Without the use of recursion, the best
3683 that can be done is to use a pattern that matches up to some fixed
3684 depth of nesting. It is not possible to handle an arbitrary nesting
3685 depth. Perl provides a facility that allows regular expressions to
3686 recurse (amongst other things). It does this by interpolating Perl code
3687 in the expression at run time, and the code can refer to the expression
3688 itself. A Perl pattern to solve the parentheses problem can be created
3689 like this:
3690
3691 $re = qr{\( (?: (?>[^()]+) | (?p{$re}) )* \)}x;
3692
3693 The (?p{...}) item interpolates Perl code at run time, and in this case
3694 refers recursively to the pattern in which it appears. Obviously, PCRE
3695 cannot support the interpolation of Perl code. Instead, it supports
3696 some special syntax for recursion of the entire pattern, and also for
3697 individual subpattern recursion.
3698
3699 The special item that consists of (? followed by a number greater than
3700 zero and a closing parenthesis is a recursive call of the subpattern of
3701 the given number, provided that it occurs inside that subpattern. (If
3702 not, it is a "subroutine" call, which is described in the next sec-
3703 tion.) The special item (?R) is a recursive call of the entire regular
3704 expression.
3705
3706 A recursive subpattern call is always treated as an atomic group. That
3707 is, once it has matched some of the subject string, it is never re-
3708 entered, even if it contains untried alternatives and there is a subse-
3709 quent matching failure.
3710
3711 This PCRE pattern solves the nested parentheses problem (assume the
3712 PCRE_EXTENDED option is set so that white space is ignored):
3713
3714 \( ( (?>[^()]+) | (?R) )* \)
3715
3716 First it matches an opening parenthesis. Then it matches any number of
3717 substrings which can either be a sequence of non-parentheses, or a
3718 recursive match of the pattern itself (that is, a correctly parenthe-
3719 sized substring). Finally there is a closing parenthesis.
3720
3721 If this were part of a larger pattern, you would not want to recurse
3722 the entire pattern, so instead you could use this:
3723
3724 ( \( ( (?>[^()]+) | (?1) )* \) )
3725
3726 We have put the pattern into parentheses, and caused the recursion to
3727 refer to them instead of the whole pattern. In a larger pattern, keep-
3728 ing track of parenthesis numbers can be tricky. It may be more conve-
3729 nient to use named parentheses instead. For this, PCRE uses (?P>name),
3730 which is an extension to the Python syntax that PCRE uses for named
3731 parentheses (Perl does not provide named parentheses). We could rewrite
3732 the above example as follows:
3733
3734 (?P<pn> \( ( (?>[^()]+) | (?P>pn) )* \) )
3735
3736 This particular example pattern contains nested unlimited repeats, and
3737 so the use of atomic grouping for matching strings of non-parentheses
3738 is important when applying the pattern to strings that do not match.
3739 For example, when this pattern is applied to
3740
3741 (aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa()
3742
3743 it yields "no match" quickly. However, if atomic grouping is not used,
3744 the match runs for a very long time indeed because there are so many
3745 different ways the + and * repeats can carve up the subject, and all
3746 have to be tested before failure can be reported.
3747
3748 At the end of a match, the values set for any capturing subpatterns are
3749 those from the outermost level of the recursion at which the subpattern
3750 value is set. If you want to obtain intermediate values, a callout
3751 function can be used (see the next section and the pcrecallout documen-
3752 tation). If the pattern above is matched against
3753
3754 (ab(cd)ef)
3755
3756 the value for the capturing parentheses is "ef", which is the last
3757 value taken on at the top level. If additional parentheses are added,
3758 giving
3759
3760 \( ( ( (?>[^()]+) | (?R) )* ) \)
3761 ^ ^
3762 ^ ^
3763
3764 the string they capture is "ab(cd)ef", the contents of the top level
3765 parentheses. If there are more than 15 capturing parentheses in a pat-
3766 tern, PCRE has to obtain extra memory to store data during a recursion,
3767 which it does by using pcre_malloc, freeing it via pcre_free after-
3768 wards. If no memory can be obtained, the match fails with the
3769 PCRE_ERROR_NOMEMORY error.
3770
3771 Do not confuse the (?R) item with the condition (R), which tests for
3772 recursion. Consider this pattern, which matches text in angle brack-
3773 ets, allowing for arbitrary nesting. Only digits are allowed in nested
3774 brackets (that is, when recursing), whereas any characters are permit-
3775 ted at the outer level.
3776
3777 < (?: (?(R) \d++ | [^<>]*+) | (?R)) * >
3778
3779 In this pattern, (?(R) is the start of a conditional subpattern, with
3780 two different alternatives for the recursive and non-recursive cases.
3781 The (?R) item is the actual recursive call.
3782
3783
3784 SUBPATTERNS AS SUBROUTINES
3785
3786 If the syntax for a recursive subpattern reference (either by number or
3787 by name) is used outside the parentheses to which it refers, it oper-
3788 ates like a subroutine in a programming language. An earlier example
3789 pointed out that the pattern
3790
3791 (sens|respons)e and \1ibility
3792
3793 matches "sense and sensibility" and "response and responsibility", but
3794 not "sense and responsibility". If instead the pattern
3795
3796 (sens|respons)e and (?1)ibility
3797
3798 is used, it does match "sense and responsibility" as well as the other
3799 two strings. Such references must, however, follow the subpattern to
3800 which they refer.
3801
3802 Like recursive subpatterns, a "subroutine" call is always treated as an
3803 atomic group. That is, once it has matched some of the subject string,
3804 it is never re-entered, even if it contains untried alternatives and
3805 there is a subsequent matching failure.
3806
3807
3808 CALLOUTS
3809
3810 Perl has a feature whereby using the sequence (?{...}) causes arbitrary
3811 Perl code to be obeyed in the middle of matching a regular expression.
3812 This makes it possible, amongst other things, to extract different sub-
3813 strings that match the same pair of parentheses when there is a repeti-
3814 tion.
3815
3816 PCRE provides a similar feature, but of course it cannot obey arbitrary
3817 Perl code. The feature is called "callout". The caller of PCRE provides
3818 an external function by putting its entry point in the global variable
3819 pcre_callout. By default, this variable contains NULL, which disables
3820 all calling out.
3821
3822 Within a regular expression, (?C) indicates the points at which the
3823 external function is to be called. If you want to identify different
3824 callout points, you can put a number less than 256 after the letter C.
3825 The default value is zero. For example, this pattern has two callout
3826 points:
3827
3828 (?C1)abc(?C2)def
3829
3830 If the PCRE_AUTO_CALLOUT flag is passed to pcre_compile(), callouts are
3831 automatically installed before each item in the pattern. They are all
3832 numbered 255.
3833
3834 During matching, when PCRE reaches a callout point (and pcre_callout is
3835 set), the external function is called. It is provided with the number
3836 of the callout, the position in the pattern, and, optionally, one item
3837 of data originally supplied by the caller of pcre_exec(). The callout
3838 function may cause matching to proceed, to backtrack, or to fail alto-
3839 gether. A complete description of the interface to the callout function
3840 is given in the pcrecallout documentation.
3841
3842 Last updated: 24 January 2006
3843 Copyright (c) 1997-2006 University of Cambridge.
3844 ------------------------------------------------------------------------------
3845
3846
3847 PCREPARTIAL(3) PCREPARTIAL(3)
3848
3849
3850 NAME
3851 PCRE - Perl-compatible regular expressions
3852
3853
3854 PARTIAL MATCHING IN PCRE
3855
3856 In normal use of PCRE, if the subject string that is passed to
3857 pcre_exec() or pcre_dfa_exec() matches as far as it goes, but is too
3858 short to match the entire pattern, PCRE_ERROR_NOMATCH is returned.
3859 There are circumstances where it might be helpful to distinguish this
3860 case from other cases in which there is no match.
3861
3862 Consider, for example, an application where a human is required to type
3863 in data for a field with specific formatting requirements. An example
3864 might be a date in the form ddmmmyy, defined by this pattern:
3865
3866 ^\d?\d(jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)\d\d$
3867
3868 If the application sees the user's keystrokes one by one, and can check
3869 that what has been typed so far is potentially valid, it is able to
3870 raise an error as soon as a mistake is made, possibly beeping and not
3871 reflecting the character that has been typed. This immediate feedback
3872 is likely to be a better user interface than a check that is delayed
3873 until the entire string has been entered.
3874
3875 PCRE supports the concept of partial matching by means of the PCRE_PAR-
3876 TIAL option, which can be set when calling pcre_exec() or
3877 pcre_dfa_exec(). When this flag is set for pcre_exec(), the return code
3878 PCRE_ERROR_NOMATCH is converted into PCRE_ERROR_PARTIAL if at any time
3879 during the matching process the last part of the subject string matched
3880 part of the pattern. Unfortunately, for non-anchored matching, it is
3881 not possible to obtain the position of the start of the partial match.
3882 No captured data is set when PCRE_ERROR_PARTIAL is returned.
3883
3884 When PCRE_PARTIAL is set for pcre_dfa_exec(), the return code
3885 PCRE_ERROR_NOMATCH is converted into PCRE_ERROR_PARTIAL if the end of
3886 the subject is reached, there have been no complete matches, but there
3887 is still at least one matching possibility. The portion of the string
3888 that provided the partial match is set as the first matching string.
3889
3890 Using PCRE_PARTIAL disables one of PCRE's optimizations. PCRE remembers
3891 the last literal byte in a pattern, and abandons matching immediately
3892 if such a byte is not present in the subject string. This optimization
3893 cannot be used for a subject string that might match only partially.
3894
3895
3896 RESTRICTED PATTERNS FOR PCRE_PARTIAL
3897
3898 Because of the way certain internal optimizations are implemented in
3899 the pcre_exec() function, the PCRE_PARTIAL option cannot be used with
3900 all patterns. These restrictions do not apply when pcre_dfa_exec() is
3901 used. For pcre_exec(), repeated single characters such as
3902
3903 a{2,4}
3904
3905 and repeated single metasequences such as
3906
3907 \d+
3908
3909 are not permitted if the maximum number of occurrences is greater than
3910 one. Optional items such as \d? (where the maximum is one) are permit-
3911 ted. Quantifiers with any values are permitted after parentheses, so
3912 the invalid examples above can be coded thus:
3913
3914 (a){2,4}
3915 (\d)+
3916
3917 These constructions run more slowly, but for the kinds of application
3918 that are envisaged for this facility, this is not felt to be a major
3919 restriction.
3920
3921 If PCRE_PARTIAL is set for a pattern that does not conform to the
3922 restrictions, pcre_exec() returns the error code PCRE_ERROR_BADPARTIAL
3923 (-13).
3924
3925
3926 EXAMPLE OF PARTIAL MATCHING USING PCRETEST
3927
3928 If the escape sequence \P is present in a pcretest data line, the
3929 PCRE_PARTIAL flag is used for the match. Here is a run of pcretest that
3930 uses the date example quoted above:
3931
3932 re> /^\d?\d(jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)\d\d$/
3933 data> 25jun04\P
3934 0: 25jun04
3935 1: jun
3936 data> 25dec3\P
3937 Partial match
3938 data> 3ju\P
3939 Partial match
3940 data> 3juj\P
3941 No match
3942 data> j\P
3943 No match
3944
3945 The first data string is matched completely, so pcretest shows the
3946 matched substrings. The remaining four strings do not match the com-
3947 plete pattern, but the first two are partial matches. The same test,
3948 using DFA matching (by means of the \D escape sequence), produces the
3949 following output:
3950
3951 re> /^?(jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)$/
3952 data> 25jun04\P\D
3953 0: 25jun04
3954 data> 23dec3\P\D
3955 Partial match: 23dec3
3956 data> 3ju\P\D
3957 Partial match: 3ju
3958 data> 3juj\P\D
3959 No match
3960 data> j\P\D
3961 No match
3962
3963 Notice that in this case the portion of the string that was matched is
3964 made available.
3965
3966
3967 MULTI-SEGMENT MATCHING WITH pcre_dfa_exec()
3968
3969 When a partial match has been found using pcre_dfa_exec(), it is possi-
3970 ble to continue the match by providing additional subject data and
3971 calling pcre_dfa_exec() again with the PCRE_DFA_RESTART option and the
3972 same working space (where details of the previous partial match are
3973 stored). Here is an example using pcretest, where the \R escape
3974 sequence sets the PCRE_DFA_RESTART option and the \D escape sequence
3975 requests the use of pcre_dfa_exec():
3976
3977 re> /^?(jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)$/
3978 data> 23ja\P\D
3979 Partial match: 23ja
3980 data> n05\R\D
3981 0: n05
3982
3983 The first call has "23ja" as the subject, and requests partial match-
3984 ing; the second call has "n05" as the subject for the continued
3985 (restarted) match. Notice that when the match is complete, only the
3986 last part is shown; PCRE does not retain the previously partially-
3987 matched string. It is up to the calling program to do that if it needs
3988 to.
3989
3990 This facility can be used to pass very long subject strings to
3991 pcre_dfa_exec(). However, some care is needed for certain types of pat-
3992 tern.
3993
3994 1. If the pattern contains tests for the beginning or end of a line,
3995 you need to pass the PCRE_NOTBOL or PCRE_NOTEOL options, as appropri-
3996 ate, when the subject string for any call does not contain the begin-
3997 ning or end of a line.
3998
3999 2. If the pattern contains backward assertions (including \b or \B),
4000 you need to arrange for some overlap in the subject strings to allow
4001 for this. For example, you could pass the subject in chunks that were
4002 500 bytes long, but in a buffer of 700 bytes, with the starting offset
4003 set to 200 and the previous 200 bytes at the start of the buffer.
4004
4005 3. Matching a subject string that is split into multiple segments does
4006 not always produce exactly the same result as matching over one single
4007 long string. The difference arises when there are multiple matching
4008 possibilities, because a partial match result is given only when there
4009 are no completed matches in a call to fBpcre_dfa_exec(). This means
4010 that as soon as the shortest match has been found, continuation to a
4011 new subject segment is no longer possible. Consider this pcretest
4012 example:
4013
4014 re> /dog(sbody)?/
4015 data> do\P\D
4016 Partial match: do
4017 data> gsb\R\P\D
4018 0: g
4019 data> dogsbody\D
4020 0: dogsbody
4021 1: dog
4022
4023 The pattern matches the words "dog" or "dogsbody". When the subject is
4024 presented in several parts ("do" and "gsb" being the first two) the
4025 match stops when "dog" has been found, and it is not possible to con-
4026 tinue. On the other hand, if "dogsbody" is presented as a single
4027 string, both matches are found.
4028
4029 Because of this phenomenon, it does not usually make sense to end a
4030 pattern that is going to be matched in this way with a variable repeat.
4031
4032 4. Patterns that contain alternatives at the top level which do not all
4033 start with the same pattern item may not work as expected. For example,
4034 consider this pattern:
4035
4036 1234|3789
4037
4038 If the first part of the subject is "ABC123", a partial match of the
4039 first alternative is found at offset 3. There is no partial match for
4040 the second alternative, because such a match does not start at the same
4041 point in the subject string. Attempting to continue with the string
4042 "789" does not yield a match because only those alternatives that match
4043 at one point in the subject are remembered. The problem arises because
4044 the start of the second alternative matches within the first alterna-
4045 tive. There is no problem with anchored patterns or patterns such as:
4046
4047 1234|ABCD
4048
4049 where no string can be a partial match for both alternatives.
4050
4051 Last updated: 16 January 2006
4052 Copyright (c) 1997-2006 University of Cambridge.
4053 ------------------------------------------------------------------------------
4054
4055
4056 PCREPRECOMPILE(3) PCREPRECOMPILE(3)
4057
4058
4059 NAME
4060 PCRE - Perl-compatible regular expressions
4061
4062
4063 SAVING AND RE-USING PRECOMPILED PCRE PATTERNS
4064
4065 If you are running an application that uses a large number of regular
4066 expression patterns, it may be useful to store them in a precompiled
4067 form instead of having to compile them every time the application is
4068 run. If you are not using any private character tables (see the
4069 pcre_maketables() documentation), this is relatively straightforward.
4070 If you are using private tables, it is a little bit more complicated.
4071
4072 If you save compiled patterns to a file, you can copy them to a differ-
4073 ent host and run them there. This works even if the new host has the
4074 opposite endianness to the one on which the patterns were compiled.
4075 There may be a small performance penalty, but it should be insignifi-
4076 cant.
4077
4078
4079 SAVING A COMPILED PATTERN
4080 The value returned by pcre_compile() points to a single block of memory
4081 that holds the compiled pattern and associated data. You can find the
4082 length of this block in bytes by calling pcre_fullinfo() with an argu-
4083 ment of PCRE_INFO_SIZE. You can then save the data in any appropriate
4084 manner. Here is sample code that compiles a pattern and writes it to a
4085 file. It assumes that the variable fd refers to a file that is open for
4086 output:
4087
4088 int erroroffset, rc, size;
4089 char *error;
4090 pcre *re;
4091
4092 re = pcre_compile("my pattern", 0, &error, &erroroffset, NULL);
4093 if (re == NULL) { ... handle errors ... }
4094 rc = pcre_fullinfo(re, NULL, PCRE_INFO_SIZE, &size);
4095 if (rc < 0) { ... handle errors ... }
4096 rc = fwrite(re, 1, size, fd);
4097 if (rc != size) { ... handle errors ... }
4098
4099 In this example, the bytes that comprise the compiled pattern are
4100 copied exactly. Note that this is binary data that may contain any of
4101 the 256 possible byte values. On systems that make a distinction
4102 between binary and non-binary data, be sure that the file is opened for
4103 binary output.
4104
4105 If you want to write more than one pattern to a file, you will have to
4106 devise a way of separating them. For binary data, preceding each pat-
4107 tern with its length is probably the most straightforward approach.
4108 Another possibility is to write out the data in hexadecimal instead of
4109 binary, one pattern to a line.
4110
4111 Saving compiled patterns in a file is only one possible way of storing
4112 them for later use. They could equally well be saved in a database, or
4113 in the memory of some daemon process that passes them via sockets to
4114 the processes that want them.
4115
4116 If the pattern has been studied, it is also possible to save the study
4117 data in a similar way to the compiled pattern itself. When studying
4118 generates additional information, pcre_study() returns a pointer to a
4119 pcre_extra data block. Its format is defined in the section on matching
4120 a pattern in the pcreapi documentation. The study_data field points to
4121 the binary study data, and this is what you must save (not the
4122 pcre_extra block itself). The length of the study data can be obtained
4123 by calling pcre_fullinfo() with an argument of PCRE_INFO_STUDYSIZE.
4124 Remember to check that pcre_study() did return a non-NULL value before
4125 trying to save the study data.
4126
4127
4128 RE-USING A PRECOMPILED PATTERN
4129
4130 Re-using a precompiled pattern is straightforward. Having reloaded it
4131 into main memory, you pass its pointer to pcre_exec() or
4132 pcre_dfa_exec() in the usual way. This should work even on another
4133 host, and even if that host has the opposite endianness to the one
4134 where the pattern was compiled.
4135
4136 However, if you passed a pointer to custom character tables when the
4137 pattern was compiled (the tableptr argument of pcre_compile()), you
4138 must now pass a similar pointer to pcre_exec() or pcre_dfa_exec(),
4139 because the value saved with the compiled pattern will obviously be
4140 nonsense. A field in a pcre_extra() block is used to pass this data, as
4141 described in the section on matching a pattern in the pcreapi documen-
4142 tation.
4143
4144 If you did not provide custom character tables when the pattern was
4145 compiled, the pointer in the compiled pattern is NULL, which causes
4146 pcre_exec() to use PCRE's internal tables. Thus, you do not need to
4147 take any special action at run time in this case.
4148
4149 If you saved study data with the compiled pattern, you need to create
4150 your own pcre_extra data block and set the study_data field to point to
4151 the reloaded study data. You must also set the PCRE_EXTRA_STUDY_DATA
4152 bit in the flags field to indicate that study data is present. Then
4153 pass the pcre_extra block to pcre_exec() or pcre_dfa_exec() in the
4154 usual way.
4155
4156
4157 COMPATIBILITY WITH DIFFERENT PCRE RELEASES
4158
4159 The layout of the control block that is at the start of the data that
4160 makes up a compiled pattern was changed for release 5.0. If you have
4161 any saved patterns that were compiled with previous releases (not a
4162 facility that was previously advertised), you will have to recompile
4163 them for release 5.0. However, from now on, it should be possible to
4164 make changes in a compatible manner.
4165
4166 Notwithstanding the above, if you have any saved patterns in UTF-8 mode
4167 that use \p or \P that were compiled with any release up to and includ-
4168 ing 6.4, you will have to recompile them for release 6.5 and above.
4169
4170 Last updated: 01 February 2006
4171 Copyright (c) 1997-2006 University of Cambridge.
4172 ------------------------------------------------------------------------------
4173
4174
4175 PCREPERFORM(3) PCREPERFORM(3)
4176
4177
4178 NAME
4179 PCRE - Perl-compatible regular expressions
4180
4181
4182 PCRE PERFORMANCE
4183
4184 Certain items that may appear in regular expression patterns are more
4185 efficient than others. It is more efficient to use a character class
4186 like [aeiou] than a set of alternatives such as (a|e|i|o|u). In gen-
4187 eral, the simplest construction that provides the required behaviour is
4188 usually the most efficient. Jeffrey Friedl's book contains a lot of
4189 useful general discussion about optimizing regular expressions for
4190 efficient performance. This document contains a few observations about
4191 PCRE.
4192
4193 Using Unicode character properties (the \p, \P, and \X escapes) is
4194 slow, because PCRE has to scan a structure that contains data for over
4195 fifteen thousand characters whenever it needs a character's property.
4196 If you can find an alternative pattern that does not use character
4197 properties, it will probably be faster.
4198
4199 When a pattern begins with .* not in parentheses, or in parentheses
4200 that are not the subject of a backreference, and the PCRE_DOTALL option
4201 is set, the pattern is implicitly anchored by PCRE, since it can match
4202 only at the start of a subject string. However, if PCRE_DOTALL is not
4203 set, PCRE cannot make this optimization, because the . metacharacter
4204 does not then match a newline, and if the subject string contains new-
4205 lines, the pattern may match from the character immediately following
4206 one of them instead of from the very start. For example, the pattern
4207
4208 .*second
4209
4210 matches the subject "first\nand second" (where \n stands for a newline
4211 character), with the match starting at the seventh character. In order
4212 to do this, PCRE has to retry the match starting after every newline in
4213 the subject.
4214
4215 If you are using such a pattern with subject strings that do not con-
4216 tain newlines, the best performance is obtained by setting PCRE_DOTALL,
4217 or starting the pattern with ^.* or ^.*? to indicate explicit anchor-
4218 ing. That saves PCRE from having to scan along the subject looking for
4219 a newline to restart at.
4220
4221 Beware of patterns that contain nested indefinite repeats. These can
4222 take a long time to run when applied to a string that does not match.
4223 Consider the pattern fragment
4224
4225 (a+)*
4226
4227 This can match "aaaa" in 33 different ways, and this number increases
4228 very rapidly as the string gets longer. (The * repeat can match 0, 1,
4229 2, 3, or 4 times, and for each of those cases other than 0, the +
4230 repeats can match different numbers of times.) When the remainder of
4231 the pattern is such that the entire match is going to fail, PCRE has in
4232 principle to try every possible variation, and this can take an
4233 extremely long time.
4234
4235 An optimization catches some of the more simple cases such as
4236
4237 (a+)*b
4238
4239 where a literal character follows. Before embarking on the standard
4240 matching procedure, PCRE checks that there is a "b" later in the sub-
4241 ject string, and if there is not, it fails the match immediately. How-
4242 ever, when there is no following literal this optimization cannot be
4243 used. You can see the difference by comparing the behaviour of
4244
4245 (a+)*\d
4246
4247 with the pattern above. The former gives a failure almost instantly
4248 when applied to a whole line of "a" characters, whereas the latter
4249 takes an appreciable time with strings longer than about 20 characters.
4250
4251 In many cases, the solution to this kind of performance issue is to use
4252 an atomic group or a possessive quantifier.
4253
4254 Last updated: 28 February 2005
4255 Copyright (c) 1997-2005 University of Cambridge.
4256 ------------------------------------------------------------------------------
4257
4258
4259 PCREPOSIX(3) PCREPOSIX(3)
4260
4261
4262 NAME
4263 PCRE - Perl-compatible regular expressions.
4264
4265
4266 SYNOPSIS OF POSIX API
4267
4268 #include <pcreposix.h>
4269
4270 int regcomp(regex_t *preg, const char *pattern,
4271 int cflags);
4272
4273 int regexec(regex_t *preg, const char *string,
4274 size_t nmatch, regmatch_t pmatch[], int eflags);
4275
4276 size_t regerror(int errcode, const regex_t *preg,
4277 char *errbuf, size_t errbuf_size);
4278
4279 void regfree(regex_t *preg);
4280
4281
4282 DESCRIPTION
4283
4284 This set of functions provides a POSIX-style API to the PCRE regular
4285 expression package. See the pcreapi documentation for a description of
4286 PCRE's native API, which contains much additional functionality.
4287
4288 The functions described here are just wrapper functions that ultimately
4289 call the PCRE native API. Their prototypes are defined in the
4290 pcreposix.h header file, and on Unix systems the library itself is
4291 called pcreposix.a, so can be accessed by adding -lpcreposix to the
4292 command for linking an application that uses them. Because the POSIX
4293 functions call the native ones, it is also necessary to add -lpcre.
4294
4295 I have implemented only those option bits that can be reasonably mapped
4296 to PCRE native options. In addition, the option REG_EXTENDED is defined
4297 with the value zero. This has no effect, but since programs that are
4298 written to the POSIX interface often use it, this makes it easier to
4299 slot in PCRE as a replacement library. Other POSIX options are not even
4300 defined.
4301
4302 When PCRE is called via these functions, it is only the API that is
4303 POSIX-like in style. The syntax and semantics of the regular expres-
4304 sions themselves are still those of Perl, subject to the setting of
4305 various PCRE options, as described below. "POSIX-like in style" means
4306 that the API approximates to the POSIX definition; it is not fully
4307 POSIX-compatible, and in multi-byte encoding domains it is probably
4308 even less compatible.
4309
4310 The header for these functions is supplied as pcreposix.h to avoid any
4311 potential clash with other POSIX libraries. It can, of course, be
4312 renamed or aliased as regex.h, which is the "correct" name. It provides
4313 two structure types, regex_t for compiled internal forms, and reg-
4314 match_t for returning captured substrings. It also defines some con-
4315 stants whose names start with "REG_"; these are used for setting
4316 options and identifying error codes.
4317
4318
4319 COMPILING A PATTERN
4320
4321 The function regcomp() is called to compile a pattern into an internal
4322 form. The pattern is a C string terminated by a binary zero, and is
4323 passed in the argument pattern. The preg argument is a pointer to a
4324 regex_t structure that is used as a base for storing information about
4325 the compiled regular expression.
4326
4327 The argument cflags is either zero, or contains one or more of the bits
4328 defined by the following macros:
4329
4330 REG_DOTALL
4331
4332 The PCRE_DOTALL option is set when the regular expression is passed for
4333 compilation to the native function. Note that REG_DOTALL is not part of
4334 the POSIX standard.
4335
4336 REG_ICASE
4337
4338 The PCRE_CASELESS option is set when the regular expression is passed
4339 for compilation to the native function.
4340
4341 REG_NEWLINE
4342
4343 The PCRE_MULTILINE option is set when the regular expression is passed
4344 for compilation to the native function. Note that this does not mimic
4345 the defined POSIX behaviour for REG_NEWLINE (see the following sec-
4346 tion).
4347
4348 REG_NOSUB
4349
4350 The PCRE_NO_AUTO_CAPTURE option is set when the regular expression is
4351 passed for compilation to the native function. In addition, when a pat-
4352 tern that is compiled with this flag is passed to regexec() for match-
4353 ing, the nmatch and pmatch arguments are ignored, and no captured
4354 strings are returned.
4355
4356 REG_UTF8
4357
4358 The PCRE_UTF8 option is set when the regular expression is passed for
4359 compilation to the native function. This causes the pattern itself and
4360 all data strings used for matching it to be treated as UTF-8 strings.
4361 Note that REG_UTF8 is not part of the POSIX standard.
4362
4363 In the absence of these flags, no options are passed to the native
4364 function. This means the the regex is compiled with PCRE default
4365 semantics. In particular, the way it handles newline characters in the
4366 subject string is the Perl way, not the POSIX way. Note that setting
4367 PCRE_MULTILINE has only some of the effects specified for REG_NEWLINE.
4368 It does not affect the way newlines are matched by . (they aren't) or
4369 by a negative class such as [^a] (they are).
4370
4371 The yield of regcomp() is zero on success, and non-zero otherwise. The
4372 preg structure is filled in on success, and one member of the structure
4373 is public: re_nsub contains the number of capturing subpatterns in the
4374 regular expression. Various error codes are defined in the header file.
4375
4376
4377 MATCHING NEWLINE CHARACTERS
4378
4379 This area is not simple, because POSIX and Perl take different views of
4380 things. It is not possible to get PCRE to obey POSIX semantics, but
4381 then PCRE was never intended to be a POSIX engine. The following table
4382 lists the different possibilities for matching newline characters in
4383 PCRE:
4384
4385 Default Change with
4386
4387 . matches newline no PCRE_DOTALL
4388 newline matches [^a] yes not changeable
4389 $ matches \n at end yes PCRE_DOLLARENDONLY
4390 $ matches \n in middle no PCRE_MULTILINE
4391 ^ matches \n in middle no PCRE_MULTILINE
4392
4393 This is the equivalent table for POSIX:
4394
4395 Default Change with
4396
4397 . matches newline yes REG_NEWLINE
4398 newline matches [^a] yes REG_NEWLINE
4399 $ matches \n at end no REG_NEWLINE
4400 $ matches \n in middle no REG_NEWLINE
4401 ^ matches \n in middle no REG_NEWLINE
4402
4403 PCRE's behaviour is the same as Perl's, except that there is no equiva-
4404 lent for PCRE_DOLLAR_ENDONLY in Perl. In both PCRE and Perl, there is
4405 no way to stop newline from matching [^a].
4406
4407 The default POSIX newline handling can be obtained by setting
4408 PCRE_DOTALL and PCRE_DOLLAR_ENDONLY, but there is no way to make PCRE
4409 behave exactly as for the REG_NEWLINE action.
4410
4411
4412 MATCHING A PATTERN
4413
4414 The function regexec() is called to match a compiled pattern preg
4415 against a given string, which is terminated by a zero byte, subject to
4416 the options in eflags. These can be:
4417
4418 REG_NOTBOL
4419
4420 The PCRE_NOTBOL option is set when calling the underlying PCRE matching
4421 function.
4422
4423 REG_NOTEOL
4424
4425 The PCRE_NOTEOL option is set when calling the underlying PCRE matching
4426 function.
4427
4428 If the pattern was compiled with the REG_NOSUB flag, no data about any
4429 matched strings is returned. The nmatch and pmatch arguments of
4430 regexec() are ignored.
4431
4432 Otherwise,the portion of the string that was matched, and also any cap-
4433 tured substrings, are returned via the pmatch argument, which points to
4434 an array of nmatch structures of type regmatch_t, containing the mem-
4435 bers rm_so and rm_eo. These contain the offset to the first character
4436 of each substring and the offset to the first character after the end
4437 of each substring, respectively. The 0th element of the vector relates
4438 to the entire portion of string that was matched; subsequent elements
4439 relate to the capturing subpatterns of the regular expression. Unused
4440 entries in the array have both structure members set to -1.
4441
4442 A successful match yields a zero return; various error codes are
4443 defined in the header file, of which REG_NOMATCH is the "expected"
4444 failure code.
4445
4446
4447 ERROR MESSAGES
4448
4449 The regerror() function maps a non-zero errorcode from either regcomp()
4450 or regexec() to a printable message. If preg is not NULL, the error
4451 should have arisen from the use of that structure. A message terminated
4452 by a binary zero is placed in errbuf. The length of the message,
4453 including the zero, is limited to errbuf_size. The yield of the func-
4454 tion is the size of buffer needed to hold the whole message.
4455
4456
4457 MEMORY USAGE
4458
4459 Compiling a regular expression causes memory to be allocated and asso-
4460 ciated with the preg structure. The function regfree() frees all such
4461 memory, after which preg may no longer be used as a compiled expres-
4462 sion.
4463
4464
4465 AUTHOR
4466
4467 Philip Hazel
4468 University Computing Service,
4469 Cambridge CB2 3QG, England.
4470
4471 Last updated: 16 January 2006
4472 Copyright (c) 1997-2006 University of Cambridge.
4473 ------------------------------------------------------------------------------
4474
4475
4476 PCRECPP(3) PCRECPP(3)
4477
4478
4479 NAME
4480 PCRE - Perl-compatible regular expressions.
4481
4482
4483 SYNOPSIS OF C++ WRAPPER
4484
4485 #include <pcrecpp.h>
4486
4487
4488 DESCRIPTION
4489
4490 The C++ wrapper for PCRE was provided by Google Inc. Some additional
4491 functionality was added by Giuseppe Maxia. This brief man page was con-
4492 structed from the notes in the pcrecpp.h file, which should be con-
4493 sulted for further details.
4494
4495
4496 MATCHING INTERFACE
4497
4498 The "FullMatch" operation checks that supplied text matches a supplied
4499 pattern exactly. If pointer arguments are supplied, it copies matched
4500 sub-strings that match sub-patterns into them.
4501
4502 Example: successful match
4503 pcrecpp::RE re("h.*o");
4504 re.FullMatch("hello");
4505
4506 Example: unsuccessful match (requires full match):
4507 pcrecpp::RE re("e");
4508 !re.FullMatch("hello");
4509
4510 Example: creating a temporary RE object:
4511 pcrecpp::RE("h.*o").FullMatch("hello");
4512
4513 You can pass in a "const char*" or a "string" for "text". The examples
4514 below tend to use a const char*. You can, as in the different examples
4515 above, store the RE object explicitly in a variable or use a temporary
4516 RE object. The examples below use one mode or the other arbitrarily.
4517 Either could correctly be used for any of these examples.
4518
4519 You must supply extra pointer arguments to extract matched subpieces.
4520
4521 Example: extracts "ruby" into "s" and 1234 into "i"
4522 int i;
4523 string s;
4524 pcrecpp::RE re("(\\w+):(\\d+)");
4525 re.FullMatch("ruby:1234", &s, &i);
4526
4527 Example: does not try to extract any extra sub-patterns
4528 re.FullMatch("ruby:1234", &s);
4529
4530 Example: does not try to extract into NULL
4531 re.FullMatch("ruby:1234", NULL, &i);
4532
4533 Example: integer overflow causes failure
4534 !re.FullMatch("ruby:1234567891234", NULL, &i);
4535
4536 Example: fails because there aren't enough sub-patterns:
4537 !pcrecpp::RE("\\w+:\\d+").FullMatch("ruby:1234", &s);
4538
4539 Example: fails because string cannot be stored in integer
4540 !pcrecpp::RE("(.*)").FullMatch("ruby", &i);
4541
4542 The provided pointer arguments can be pointers to any scalar numeric
4543 type, or one of:
4544
4545 string (matched piece is copied to string)
4546 StringPiece (StringPiece is mutated to point to matched piece)
4547 T (where "bool T::ParseFrom(const char*, int)" exists)
4548 NULL (the corresponding matched sub-pattern is not copied)
4549
4550 The function returns true iff all of the following conditions are sat-
4551 isfied:
4552
4553 a. "text" matches "pattern" exactly;
4554
4555 b. The number of matched sub-patterns is >= number of supplied
4556 pointers;
4557
4558 c. The "i"th argument has a suitable type for holding the
4559 string captured as the "i"th sub-pattern. If you pass in
4560 NULL for the "i"th argument, or pass fewer arguments than
4561 number of sub-patterns, "i"th captured sub-pattern is
4562 ignored.
4563
4564 The matching interface supports at most 16 arguments per call. If you
4565 need more, consider using the more general interface
4566 pcrecpp::RE::DoMatch. See pcrecpp.h for the signature for DoMatch.
4567
4568
4569 PARTIAL MATCHES
4570
4571 You can use the "PartialMatch" operation when you want the pattern to
4572 match any substring of the text.
4573
4574 Example: simple search for a string:
4575 pcrecpp::RE("ell").PartialMatch("hello");
4576
4577 Example: find first number in a string:
4578 int number;
4579 pcrecpp::RE re("(\\d+)");
4580 re.PartialMatch("x*100 + 20", &number);
4581 assert(number == 100);
4582
4583
4584 UTF-8 AND THE MATCHING INTERFACE
4585
4586 By default, pattern and text are plain text, one byte per character.
4587 The UTF8 flag, passed to the constructor, causes both pattern and
4588 string to be treated as UTF-8 text, still a byte stream but potentially
4589 multiple bytes per character. In practice, the text is likelier to be
4590 UTF-8 than the pattern, but the match returned may depend on the UTF8
4591 flag, so always use it when matching UTF8 text. For example, "." will
4592 match one byte normally but with UTF8 set may match up to three bytes
4593 of a multi-byte character.
4594
4595 Example:
4596 pcrecpp::RE_Options options;
4597 options.set_utf8();
4598 pcrecpp::RE re(utf8_pattern, options);
4599 re.FullMatch(utf8_string);
4600
4601 Example: using the convenience function UTF8():
4602 pcrecpp::RE re(utf8_pattern, pcrecpp::UTF8());
4603 re.FullMatch(utf8_string);
4604
4605 NOTE: The UTF8 flag is ignored if pcre was not configured with the
4606 --enable-utf8 flag.
4607
4608
4609 PASSING MODIFIERS TO THE REGULAR EXPRESSION ENGINE
4610
4611 PCRE defines some modifiers to change the behavior of the regular
4612 expression engine. The C++ wrapper defines an auxiliary class,
4613 RE_Options, as a vehicle to pass such modifiers to a RE class. Cur-
4614 rently, the following modifiers are supported:
4615
4616 modifier description Perl corresponding
4617
4618 PCRE_CASELESS case insensitive match /i
4619 PCRE_MULTILINE multiple lines match /m
4620 PCRE_DOTALL dot matches newlines /s
4621 PCRE_DOLLAR_ENDONLY $ matches only at end N/A
4622 PCRE_EXTRA strict escape parsing N/A
4623 PCRE_EXTENDED ignore whitespaces /x
4624 PCRE_UTF8 handles UTF8 chars built-in
4625 PCRE_UNGREEDY reverses * and *? N/A
4626 PCRE_NO_AUTO_CAPTURE disables capturing parens N/A (*)
4627
4628 (*) Both Perl and PCRE allow non capturing parentheses by means of the
4629 "?:" modifier within the pattern itself. e.g. (?:ab|cd) does not cap-
4630 ture, while (ab|cd) does.
4631
4632 For a full account on how each modifier works, please check the PCRE
4633 API reference page.
4634
4635 For each modifier, there are two member functions whose name is made
4636 out of the modifier in lowercase, without the "PCRE_" prefix. For
4637 instance, PCRE_CASELESS is handled by
4638
4639 bool caseless()
4640
4641 which returns true if the modifier is set, and
4642
4643 RE_Options & set_caseless(bool)
4644
4645 which sets or unsets the modifier. Moreover, PCRE_EXTRA_MATCH_LIMIT can
4646 be accessed through the set_match_limit() and match_limit() member
4647 functions. Setting match_limit to a non-zero value will limit the exe-
4648 cution of pcre to keep it from doing bad things like blowing the stack
4649 or taking an eternity to return a result. A value of 5000 is good
4650 enough to stop stack blowup in a 2MB thread stack. Setting match_limit
4651 to zero disables match limiting. Alternatively, you can call
4652 match_limit_recursion() which uses PCRE_EXTRA_MATCH_LIMIT_RECURSION to
4653 limit how much PCRE recurses. match_limit() limits the number of
4654 matches PCRE does; match_limit_recursion() limits the depth of internal
4655 recursion, and therefore the amount of stack that is used.
4656
4657 Normally, to pass one or more modifiers to a RE class, you declare a
4658 RE_Options object, set the appropriate options, and pass this object to
4659 a RE constructor. Example:
4660
4661 RE_options opt;
4662 opt.set_caseless(true);
4663 if (RE("HELLO", opt).PartialMatch("hello world")) ...
4664
4665 RE_options has two constructors. The default constructor takes no argu-
4666 ments and creates a set of flags that are off by default. The optional
4667 parameter option_flags is to facilitate transfer of legacy code from C
4668 programs. This lets you do
4669
4670 RE(pattern,
4671 RE_Options(PCRE_CASELESS|PCRE_MULTILINE)).PartialMatch(str);
4672
4673 However, new code is better off doing
4674
4675 RE(pattern,
4676 RE_Options().set_caseless(true).set_multiline(true))
4677 .PartialMatch(str);
4678
4679 If you are going to pass one of the most used modifiers, there are some
4680 convenience functions that return a RE_Options class with the appropri-
4681 ate modifier already set: CASELESS(), UTF8(), MULTILINE(), DOTALL(),
4682 and EXTENDED().
4683
4684 If you need to set several options at once, and you don't want to go
4685 through the pains of declaring a RE_Options object and setting several
4686 options, there is a parallel method that give you such ability on the
4687 fly. You can concatenate several set_xxxxx() member functions, since
4688 each of them returns a reference to its class object. For example, to
4689 pass PCRE_CASELESS, PCRE_EXTENDED, and PCRE_MULTILINE to a RE with one
4690 statement, you may write:
4691
4692 RE(" ^ xyz \\s+ .* blah$",
4693 RE_Options()
4694 .set_caseless(true)
4695 .set_extended(true)
4696 .set_multiline(true)).PartialMatch(sometext);
4697
4698
4699 SCANNING TEXT INCREMENTALLY
4700
4701 The "Consume" operation may be useful if you want to repeatedly match
4702 regular expressions at the front of a string and skip over them as they
4703 match. This requires use of the "StringPiece" type, which represents a
4704 sub-range of a real string. Like RE, StringPiece is defined in the
4705 pcrecpp namespace.
4706
4707 Example: read lines of the form "var = value" from a string.
4708 string contents = ...; // Fill string somehow
4709 pcrecpp::StringPiece input(contents); // Wrap in a StringPiece
4710
4711 string var;
4712 int value;
4713 pcrecpp::RE re("(\\w+) = (\\d+)\n");
4714 while (re.Consume(&input, &var, &value)) {
4715 ...;
4716 }
4717
4718 Each successful call to "Consume" will set "var/value", and also
4719 advance "input" so it points past the matched text.
4720
4721 The "FindAndConsume" operation is similar to "Consume" but does not
4722 anchor your match at the beginning of the string. For example, you
4723 could extract all words from a string by repeatedly calling
4724
4725 pcrecpp::RE("(\\w+)").FindAndConsume(&input, &word)
4726
4727
4728 PARSING HEX/OCTAL/C-RADIX NUMBERS
4729
4730 By default, if you pass a pointer to a numeric value, the corresponding
4731 text is interpreted as a base-10 number. You can instead wrap the
4732 pointer with a call to one of the operators Hex(), Octal(), or CRadix()
4733 to interpret the text in another base. The CRadix operator interprets
4734 C-style "0" (base-8) and "0x" (base-16) prefixes, but defaults to
4735 base-10.
4736
4737 Example:
4738 int a, b, c, d;
4739 pcrecpp::RE re("(.*) (.*) (.*) (.*)");
4740 re.FullMatch("100 40 0100 0x40",
4741 pcrecpp::Octal(&a), pcrecpp::Hex(&b),
4742 pcrecpp::CRadix(&c), pcrecpp::CRadix(&d));
4743
4744 will leave 64 in a, b, c, and d.
4745
4746
4747 REPLACING PARTS OF STRINGS
4748
4749 You can replace the first match of "pattern" in "str" with "rewrite".
4750 Within "rewrite", backslash-escaped digits (\1 to \9) can be used to
4751 insert text matching corresponding parenthesized group from the pat-
4752 tern. \0 in "rewrite" refers to the entire matching text. For example:
4753
4754 string s = "yabba dabba doo";
4755 pcrecpp::RE("b+").Replace("d", &s);
4756
4757 will leave "s" containing "yada dabba doo". The result is true if the
4758 pattern matches and a replacement occurs, false otherwise.
4759
4760 GlobalReplace is like Replace except that it replaces all occurrences
4761 of the pattern in the string with the rewrite. Replacements are not
4762 subject to re-matching. For example:
4763
4764 string s = "yabba dabba doo";
4765 pcrecpp::RE("b+").GlobalReplace("d", &s);
4766
4767 will leave "s" containing "yada dada doo". It returns the number of
4768 replacements made.
4769
4770 Extract is like Replace, except that if the pattern matches, "rewrite"
4771 is copied into "out" (an additional argument) with substitutions. The
4772 non-matching portions of "text" are ignored. Returns true iff a match
4773 occurred and the extraction happened successfully; if no match occurs,
4774 the string is left unaffected.
4775
4776
4777 AUTHOR
4778
4779 The C++ wrapper was contributed by Google Inc.
4780 Copyright (c) 2005 Google Inc.
4781 ------------------------------------------------------------------------------
4782
4783
4784 PCRESAMPLE(3) PCRESAMPLE(3)
4785
4786
4787 NAME
4788 PCRE - Perl-compatible regular expressions
4789
4790
4791 PCRE SAMPLE PROGRAM
4792
4793 A simple, complete demonstration program, to get you started with using
4794 PCRE, is supplied in the file pcredemo.c in the PCRE distribution.
4795
4796 The program compiles the regular expression that is its first argument,
4797 and matches it against the subject string in its second argument. No
4798 PCRE options are set, and default character tables are used. If match-
4799 ing succeeds, the program outputs the portion of the subject that
4800 matched, together with the contents of any captured substrings.
4801
4802 If the -g option is given on the command line, the program then goes on
4803 to check for further matches of the same regular expression in the same
4804 subject string. The logic is a little bit tricky because of the possi-
4805 bility of matching an empty string. Comments in the code explain what
4806 is going on.
4807
4808 If PCRE is installed in the standard include and library directories
4809 for your system, you should be able to compile the demonstration pro-
4810 gram using this command:
4811
4812 gcc -o pcredemo pcredemo.c -lpcre
4813
4814 If PCRE is installed elsewhere, you may need to add additional options
4815 to the command line. For example, on a Unix-like system that has PCRE
4816 installed in /usr/local, you can compile the demonstration program
4817 using a command like this:
4818
4819 gcc -o pcredemo -I/usr/local/include pcredemo.c \
4820 -L/usr/local/lib -lpcre
4821
4822 Once you have compiled the demonstration program, you can run simple
4823 tests like this:
4824
4825 ./pcredemo 'cat|dog' 'the cat sat on the mat'
4826 ./pcredemo -g 'cat|dog' 'the dog sat on the cat'
4827
4828 Note that there is a much more comprehensive test program, called
4829 pcretest, which supports many more facilities for testing regular
4830 expressions and the PCRE library. The pcredemo program is provided as a
4831 simple coding example.
4832
4833 On some operating systems (e.g. Solaris), when PCRE is not installed in
4834 the standard library directory, you may get an error like this when you
4835 try to run pcredemo:
4836
4837 ld.so.1: a.out: fatal: libpcre.so.0: open failed: No such file or
4838 directory
4839
4840 This is caused by the way shared library support works on those sys-
4841 tems. You need to add
4842
4843 -R/usr/local/lib
4844
4845 (for example) to the compile command to get round this problem.
4846
4847 Last updated: 09 September 2004
4848 Copyright (c) 1997-2004 University of Cambridge.
4849 ------------------------------------------------------------------------------

  ViewVC Help
Powered by ViewVC 1.1.5