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

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

Parent Directory Parent Directory | Revision Log Revision Log


Revision 155 - (show annotations)
Tue Apr 24 13:36:11 2007 UTC (12 years, 5 months ago) by ph10
File MIME type: text/plain
File size: 267172 byte(s)
Error occurred while calculating annotation data.
Tidies and final updates for 7.1.
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. (Certain features that appeared in Python and
22 PCRE before they appeared in Perl are also available using the Python
23 syntax.)
24
25 The current implementation of PCRE (release 7.x) corresponds approxi-
26 mately with Perl 5.10, including support for UTF-8 encoded strings and
27 Unicode general category properties. However, UTF-8 and Unicode support
28 has to be explicitly enabled; it is not the default. The Unicode tables
29 correspond to Unicode release 5.0.0.
30
31 In addition to the Perl-compatible matching function, PCRE contains an
32 alternative matching function that matches the same compiled patterns
33 in a different way. In certain circumstances, the alternative function
34 has some advantages. For a discussion of the two matching algorithms,
35 see the pcrematching page.
36
37 PCRE is written in C and released as a C library. A number of people
38 have written wrappers and interfaces of various kinds. In particular,
39 Google Inc. have provided a comprehensive C++ wrapper. This is now
40 included as part of the PCRE distribution. The pcrecpp page has details
41 of this interface. Other people's contributions can be found in the
42 Contrib directory at the primary FTP site, which is:
43
44 ftp://ftp.csx.cam.ac.uk/pub/software/programming/pcre
45
46 Details of exactly which Perl regular expression features are and are
47 not supported by PCRE are given in separate documents. See the pcrepat-
48 tern and pcrecompat pages.
49
50 Some features of PCRE can be included, excluded, or changed when the
51 library is built. The pcre_config() function makes it possible for a
52 client to discover which features are available. The features them-
53 selves are described in the pcrebuild page. Documentation about build-
54 ing PCRE for various operating systems can be found in the README file
55 in the source distribution.
56
57 The library contains a number of undocumented internal functions and
58 data tables that are used by more than one of the exported external
59 functions, but which are not intended for use by external callers.
60 Their names all begin with "_pcre_", which hopefully will not provoke
61 any name clashes. In some environments, it is possible to control which
62 external symbols are exported when a shared library is built, and in
63 these cases the undocumented symbols are not exported.
64
65
66 USER DOCUMENTATION
67
68 The user documentation for PCRE comprises a number of different sec-
69 tions. In the "man" format, each of these is a separate "man page". In
70 the HTML format, each is a separate page, linked from the index page.
71 In the plain text format, all the sections are concatenated, for ease
72 of searching. The sections are as follows:
73
74 pcre this document
75 pcre-config show PCRE installation configuration information
76 pcreapi details of PCRE's native C API
77 pcrebuild options for building PCRE
78 pcrecallout details of the callout feature
79 pcrecompat discussion of Perl compatibility
80 pcrecpp details of the C++ wrapper
81 pcregrep description of the pcregrep command
82 pcrematching discussion of the two matching algorithms
83 pcrepartial details of the partial matching facility
84 pcrepattern syntax and semantics of supported
85 regular expressions
86 pcreperform discussion of performance issues
87 pcreposix the POSIX-compatible C API
88 pcreprecompile details of saving and re-using precompiled patterns
89 pcresample discussion of the sample program
90 pcrestack discussion of stack usage
91 pcretest description of the pcretest testing command
92
93 In addition, in the "man" and HTML formats, there is a short page for
94 each C library function, listing its arguments and results.
95
96
97 LIMITATIONS
98
99 There are some size limitations in PCRE but it is hoped that they will
100 never in practice be relevant.
101
102 The maximum length of a compiled pattern is 65539 (sic) bytes if PCRE
103 is compiled with the default internal linkage size of 2. If you want to
104 process regular expressions that are truly enormous, you can compile
105 PCRE with an internal linkage size of 3 or 4 (see the README file in
106 the source distribution and the pcrebuild documentation for details).
107 In these cases the limit is substantially larger. However, the speed
108 of execution is slower.
109
110 All values in repeating quantifiers must be less than 65536. The maxi-
111 mum compiled length of subpattern with an explicit repeat count is
112 30000 bytes. The maximum number of capturing subpatterns is 65535.
113
114 There is no limit to the number of parenthesized subpatterns, but there
115 can be no more than 65535 capturing subpatterns.
116
117 The maximum length of name for a named subpattern is 32 characters, and
118 the maximum number of named subpatterns is 10000.
119
120 The maximum length of a subject string is the largest positive number
121 that an integer variable can hold. However, when using the traditional
122 matching function, PCRE uses recursion to handle subpatterns and indef-
123 inite repetition. This means that the available stack space may limit
124 the size of a subject string that can be processed by certain patterns.
125 For a discussion of stack issues, see the pcrestack documentation.
126
127
128 UTF-8 AND UNICODE PROPERTY SUPPORT
129
130 From release 3.3, PCRE has had some support for character strings
131 encoded in the UTF-8 format. For release 4.0 this was greatly extended
132 to cover most common requirements, and in release 5.0 additional sup-
133 port for Unicode general category properties was added.
134
135 In order process UTF-8 strings, you must build PCRE to include UTF-8
136 support in the code, and, in addition, you must call pcre_compile()
137 with the PCRE_UTF8 option flag. When you do this, both the pattern and
138 any subject strings that are matched against it are treated as UTF-8
139 strings instead of just strings of bytes.
140
141 If you compile PCRE with UTF-8 support, but do not use it at run time,
142 the library will be a bit bigger, but the additional run time overhead
143 is limited to testing the PCRE_UTF8 flag occasionally, so should not be
144 very big.
145
146 If PCRE is built with Unicode character property support (which implies
147 UTF-8 support), the escape sequences \p{..}, \P{..}, and \X are sup-
148 ported. The available properties that can be tested are limited to the
149 general category properties such as Lu for an upper case letter or Nd
150 for a decimal number, the Unicode script names such as Arabic or Han,
151 and the derived properties Any and L&. A full list is given in the
152 pcrepattern documentation. Only the short names for properties are sup-
153 ported. For example, \p{L} matches a letter. Its Perl synonym, \p{Let-
154 ter}, is not supported. Furthermore, in Perl, many properties may
155 optionally be prefixed by "Is", for compatibility with Perl 5.6. PCRE
156 does not support this.
157
158 The following comments apply when PCRE is running in UTF-8 mode:
159
160 1. When you set the PCRE_UTF8 flag, the strings passed as patterns and
161 subjects are checked for validity on entry to the relevant functions.
162 If an invalid UTF-8 string is passed, an error return is given. In some
163 situations, you may already know that your strings are valid, and
164 therefore want to skip these checks in order to improve performance. If
165 you set the PCRE_NO_UTF8_CHECK flag at compile time or at run time,
166 PCRE assumes that the pattern or subject it is given (respectively)
167 contains only valid UTF-8 codes. In this case, it does not diagnose an
168 invalid UTF-8 string. If you pass an invalid UTF-8 string to PCRE when
169 PCRE_NO_UTF8_CHECK is set, the results are undefined. Your program may
170 crash.
171
172 2. An unbraced hexadecimal escape sequence (such as \xb3) matches a
173 two-byte UTF-8 character if the value is greater than 127.
174
175 3. Octal numbers up to \777 are recognized, and match two-byte UTF-8
176 characters for values greater than \177.
177
178 4. Repeat quantifiers apply to complete UTF-8 characters, not to indi-
179 vidual bytes, for example: \x{100}{3}.
180
181 5. The dot metacharacter matches one UTF-8 character instead of a sin-
182 gle byte.
183
184 6. The escape sequence \C can be used to match a single byte in UTF-8
185 mode, but its use can lead to some strange effects. This facility is
186 not available in the alternative matching function, pcre_dfa_exec().
187
188 7. The character escapes \b, \B, \d, \D, \s, \S, \w, and \W correctly
189 test characters of any code value, but the characters that PCRE recog-
190 nizes as digits, spaces, or word characters remain the same set as
191 before, all with values less than 256. This remains true even when PCRE
192 includes Unicode property support, because to do otherwise would slow
193 down PCRE in many common cases. If you really want to test for a wider
194 sense of, say, "digit", you must use Unicode property tests such as
195 \p{Nd}.
196
197 8. Similarly, characters that match the POSIX named character classes
198 are all low-valued characters.
199
200 9. Case-insensitive matching applies only to characters whose values
201 are less than 128, unless PCRE is built with Unicode property support.
202 Even when Unicode property support is available, PCRE still uses its
203 own character tables when checking the case of low-valued characters,
204 so as not to degrade performance. The Unicode property information is
205 used only for characters with higher values. Even when Unicode property
206 support is available, PCRE supports case-insensitive matching only when
207 there is a one-to-one mapping between a letter's cases. There are a
208 small number of many-to-one mappings in Unicode; these are not sup-
209 ported by PCRE.
210
211
212 AUTHOR
213
214 Philip Hazel
215 University Computing Service
216 Cambridge CB2 3QH, England.
217
218 Putting an actual email address here seems to have been a spam magnet,
219 so I've taken it away. If you want to email me, use my two initials,
220 followed by the two digits 10, at the domain cam.ac.uk.
221
222
223 REVISION
224
225 Last updated: 18 April 2007
226 Copyright (c) 1997-2007 University of Cambridge.
227 ------------------------------------------------------------------------------
228
229
230 PCREBUILD(3) PCREBUILD(3)
231
232
233 NAME
234 PCRE - Perl-compatible regular expressions
235
236
237 PCRE BUILD-TIME OPTIONS
238
239 This document describes the optional features of PCRE that can be
240 selected when the library is compiled. They are all selected, or dese-
241 lected, by providing options to the configure script that is run before
242 the make command. The complete list of options for configure (which
243 includes the standard ones such as the selection of the installation
244 directory) can be obtained by running
245
246 ./configure --help
247
248 The following sections include descriptions of options whose names
249 begin with --enable or --disable. These settings specify changes to the
250 defaults for the configure command. Because of the way that configure
251 works, --enable and --disable always come in pairs, so the complemen-
252 tary option always exists as well, but as it specifies the default, it
253 is not described.
254
255
256 C++ SUPPORT
257
258 By default, the configure script will search for a C++ compiler and C++
259 header files. If it finds them, it automatically builds the C++ wrapper
260 library for PCRE. You can disable this by adding
261
262 --disable-cpp
263
264 to the configure command.
265
266
267 UTF-8 SUPPORT
268
269 To build PCRE with support for UTF-8 character strings, add
270
271 --enable-utf8
272
273 to the configure command. Of itself, this does not make PCRE treat
274 strings as UTF-8. As well as compiling PCRE with this option, you also
275 have have to set the PCRE_UTF8 option when you call the pcre_compile()
276 function.
277
278
279 UNICODE CHARACTER PROPERTY SUPPORT
280
281 UTF-8 support allows PCRE to process character values greater than 255
282 in the strings that it handles. On its own, however, it does not pro-
283 vide any facilities for accessing the properties of such characters. If
284 you want to be able to use the pattern escapes \P, \p, and \X, which
285 refer to Unicode character properties, you must add
286
287 --enable-unicode-properties
288
289 to the configure command. This implies UTF-8 support, even if you have
290 not explicitly requested it.
291
292 Including Unicode property support adds around 30K of tables to the
293 PCRE library. Only the general category properties such as Lu and Nd
294 are supported. Details are given in the pcrepattern documentation.
295
296
297 CODE VALUE OF NEWLINE
298
299 By default, PCRE interprets character 10 (linefeed, LF) as indicating
300 the end of a line. This is the normal newline character on Unix-like
301 systems. You can compile PCRE to use character 13 (carriage return, CR)
302 instead, by adding
303
304 --enable-newline-is-cr
305
306 to the configure command. There is also a --enable-newline-is-lf
307 option, which explicitly specifies linefeed as the newline character.
308
309 Alternatively, you can specify that line endings are to be indicated by
310 the two character sequence CRLF. If you want this, add
311
312 --enable-newline-is-crlf
313
314 to the configure command. There is a fourth option, specified by
315
316 --enable-newline-is-anycrlf
317
318 which causes PCRE to recognize any of the three sequences CR, LF, or
319 CRLF as indicating a line ending. Finally, a fifth option, specified by
320
321 --enable-newline-is-any
322
323 causes PCRE to recognize any Unicode newline sequence.
324
325 Whatever line ending convention is selected when PCRE is built can be
326 overridden when the library functions are called. At build time it is
327 conventional to use the standard for your operating system.
328
329
330 BUILDING SHARED AND STATIC LIBRARIES
331
332 The PCRE building process uses libtool to build both shared and static
333 Unix libraries by default. You can suppress one of these by adding one
334 of
335
336 --disable-shared
337 --disable-static
338
339 to the configure command, as required.
340
341
342 POSIX MALLOC USAGE
343
344 When PCRE is called through the POSIX interface (see the pcreposix doc-
345 umentation), additional working storage is required for holding the
346 pointers to capturing substrings, because PCRE requires three integers
347 per substring, whereas the POSIX interface provides only two. If the
348 number of expected substrings is small, the wrapper function uses space
349 on the stack, because this is faster than using malloc() for each call.
350 The default threshold above which the stack is no longer used is 10; it
351 can be changed by adding a setting such as
352
353 --with-posix-malloc-threshold=20
354
355 to the configure command.
356
357
358 HANDLING VERY LARGE PATTERNS
359
360 Within a compiled pattern, offset values are used to point from one
361 part to another (for example, from an opening parenthesis to an alter-
362 nation metacharacter). By default, two-byte values are used for these
363 offsets, leading to a maximum size for a compiled pattern of around
364 64K. This is sufficient to handle all but the most gigantic patterns.
365 Nevertheless, some people do want to process enormous patterns, so it
366 is possible to compile PCRE to use three-byte or four-byte offsets by
367 adding a setting such as
368
369 --with-link-size=3
370
371 to the configure command. The value given must be 2, 3, or 4. Using
372 longer offsets slows down the operation of PCRE because it has to load
373 additional bytes when handling them.
374
375
376 AVOIDING EXCESSIVE STACK USAGE
377
378 When matching with the pcre_exec() function, PCRE implements backtrack-
379 ing by making recursive calls to an internal function called match().
380 In environments where the size of the stack is limited, this can se-
381 verely limit PCRE's operation. (The Unix environment does not usually
382 suffer from this problem, but it may sometimes be necessary to increase
383 the maximum stack size. There is a discussion in the pcrestack docu-
384 mentation.) An alternative approach to recursion that uses memory from
385 the heap to remember data, instead of using recursive function calls,
386 has been implemented to work round the problem of limited stack size.
387 If you want to build a version of PCRE that works this way, add
388
389 --disable-stack-for-recursion
390
391 to the configure command. With this configuration, PCRE will use the
392 pcre_stack_malloc and pcre_stack_free variables to call memory manage-
393 ment functions. Separate functions are provided because the usage is
394 very predictable: the block sizes requested are always the same, and
395 the blocks are always freed in reverse order. A calling program might
396 be able to implement optimized functions that perform better than the
397 standard malloc() and free() functions. PCRE runs noticeably more
398 slowly when built in this way. This option affects only the pcre_exec()
399 function; it is not relevant for the the pcre_dfa_exec() function.
400
401
402 LIMITING PCRE RESOURCE USAGE
403
404 Internally, PCRE has a function called match(), which it calls repeat-
405 edly (sometimes recursively) when matching a pattern with the
406 pcre_exec() function. By controlling the maximum number of times this
407 function may be called during a single matching operation, a limit can
408 be placed on the resources used by a single call to pcre_exec(). The
409 limit can be changed at run time, as described in the pcreapi documen-
410 tation. The default is 10 million, but this can be changed by adding a
411 setting such as
412
413 --with-match-limit=500000
414
415 to the configure command. This setting has no effect on the
416 pcre_dfa_exec() matching function.
417
418 In some environments it is desirable to limit the depth of recursive
419 calls of match() more strictly than the total number of calls, in order
420 to restrict the maximum amount of stack (or heap, if --disable-stack-
421 for-recursion is specified) that is used. A second limit controls this;
422 it defaults to the value that is set for --with-match-limit, which
423 imposes no additional constraints. However, you can set a lower limit
424 by adding, for example,
425
426 --with-match-limit-recursion=10000
427
428 to the configure command. This value can also be overridden at run
429 time.
430
431
432 CREATING CHARACTER TABLES AT BUILD TIME
433
434 PCRE uses fixed tables for processing characters whose code values are
435 less than 256. By default, PCRE is built with a set of tables that are
436 distributed in the file pcre_chartables.c.dist. These tables are for
437 ASCII codes only. If you add
438
439 --enable-rebuild-chartables
440
441 to the configure command, the distributed tables are no longer used.
442 Instead, a program called dftables is compiled and run. This outputs
443 the source for new set of tables, created in the default locale of your
444 C runtime system. (This method of replacing the tables does not work if
445 you are cross compiling, because dftables is run on the local host. If
446 you need to create alternative tables when cross compiling, you will
447 have to do so "by hand".)
448
449
450 USING EBCDIC CODE
451
452 PCRE assumes by default that it will run in an environment where the
453 character code is ASCII (or Unicode, which is a superset of ASCII).
454 PCRE can, however, be compiled to run in an EBCDIC environment by
455 adding
456
457 --enable-ebcdic
458
459 to the configure command. This setting implies --enable-rebuild-charta-
460 bles.
461
462
463 SEE ALSO
464
465 pcreapi(3), pcre_config(3).
466
467
468 AUTHOR
469
470 Philip Hazel
471 University Computing Service
472 Cambridge CB2 3QH, England.
473
474
475 REVISION
476
477 Last updated: 16 April 2007
478 Copyright (c) 1997-2007 University of Cambridge.
479 ------------------------------------------------------------------------------
480
481
482 PCREMATCHING(3) PCREMATCHING(3)
483
484
485 NAME
486 PCRE - Perl-compatible regular expressions
487
488
489 PCRE MATCHING ALGORITHMS
490
491 This document describes the two different algorithms that are available
492 in PCRE for matching a compiled regular expression against a given sub-
493 ject string. The "standard" algorithm is the one provided by the
494 pcre_exec() function. This works in the same was as Perl's matching
495 function, and provides a Perl-compatible matching operation.
496
497 An alternative algorithm is provided by the pcre_dfa_exec() function;
498 this operates in a different way, and is not Perl-compatible. It has
499 advantages and disadvantages compared with the standard algorithm, and
500 these are described below.
501
502 When there is only one possible way in which a given subject string can
503 match a pattern, the two algorithms give the same answer. A difference
504 arises, however, when there are multiple possibilities. For example, if
505 the pattern
506
507 ^<.*>
508
509 is matched against the string
510
511 <something> <something else> <something further>
512
513 there are three possible answers. The standard algorithm finds only one
514 of them, whereas the alternative algorithm finds all three.
515
516
517 REGULAR EXPRESSIONS AS TREES
518
519 The set of strings that are matched by a regular expression can be rep-
520 resented as a tree structure. An unlimited repetition in the pattern
521 makes the tree of infinite size, but it is still a tree. Matching the
522 pattern to a given subject string (from a given starting point) can be
523 thought of as a search of the tree. There are two ways to search a
524 tree: depth-first and breadth-first, and these correspond to the two
525 matching algorithms provided by PCRE.
526
527
528 THE STANDARD MATCHING ALGORITHM
529
530 In the terminology of Jeffrey Friedl's book "Mastering Regular Expres-
531 sions", the standard algorithm is an "NFA algorithm". It conducts a
532 depth-first search of the pattern tree. That is, it proceeds along a
533 single path through the tree, checking that the subject matches what is
534 required. When there is a mismatch, the algorithm tries any alterna-
535 tives at the current point, and if they all fail, it backs up to the
536 previous branch point in the tree, and tries the next alternative
537 branch at that level. This often involves backing up (moving to the
538 left) in the subject string as well. The order in which repetition
539 branches are tried is controlled by the greedy or ungreedy nature of
540 the quantifier.
541
542 If a leaf node is reached, a matching string has been found, and at
543 that point the algorithm stops. Thus, if there is more than one possi-
544 ble match, this algorithm returns the first one that it finds. Whether
545 this is the shortest, the longest, or some intermediate length depends
546 on the way the greedy and ungreedy repetition quantifiers are specified
547 in the pattern.
548
549 Because it ends up with a single path through the tree, it is rela-
550 tively straightforward for this algorithm to keep track of the sub-
551 strings that are matched by portions of the pattern in parentheses.
552 This provides support for capturing parentheses and back references.
553
554
555 THE ALTERNATIVE MATCHING ALGORITHM
556
557 This algorithm conducts a breadth-first search of the tree. Starting
558 from the first matching point in the subject, it scans the subject
559 string from left to right, once, character by character, and as it does
560 this, it remembers all the paths through the tree that represent valid
561 matches. In Friedl's terminology, this is a kind of "DFA algorithm",
562 though it is not implemented as a traditional finite state machine (it
563 keeps multiple states active simultaneously).
564
565 The scan continues until either the end of the subject is reached, or
566 there are no more unterminated paths. At this point, terminated paths
567 represent the different matching possibilities (if there are none, the
568 match has failed). Thus, if there is more than one possible match,
569 this algorithm finds all of them, and in particular, it finds the long-
570 est. In PCRE, there is an option to stop the algorithm after the first
571 match (which is necessarily the shortest) has been found.
572
573 Note that all the matches that are found start at the same point in the
574 subject. If the pattern
575
576 cat(er(pillar)?)
577
578 is matched against the string "the caterpillar catchment", the result
579 will be the three strings "cat", "cater", and "caterpillar" that start
580 at the fourth character of the subject. The algorithm does not automat-
581 ically move on to find matches that start at later positions.
582
583 There are a number of features of PCRE regular expressions that are not
584 supported by the alternative matching algorithm. They are as follows:
585
586 1. Because the algorithm finds all possible matches, the greedy or
587 ungreedy nature of repetition quantifiers is not relevant. Greedy and
588 ungreedy quantifiers are treated in exactly the same way. However, pos-
589 sessive quantifiers can make a difference when what follows could also
590 match what is quantified, for example in a pattern like this:
591
592 ^a++\w!
593
594 This pattern matches "aaab!" but not "aaa!", which would be matched by
595 a non-possessive quantifier. Similarly, if an atomic group is present,
596 it is matched as if it were a standalone pattern at the current point,
597 and the longest match is then "locked in" for the rest of the overall
598 pattern.
599
600 2. When dealing with multiple paths through the tree simultaneously, it
601 is not straightforward to keep track of captured substrings for the
602 different matching possibilities, and PCRE's implementation of this
603 algorithm does not attempt to do this. This means that no captured sub-
604 strings are available.
605
606 3. Because no substrings are captured, back references within the pat-
607 tern are not supported, and cause errors if encountered.
608
609 4. For the same reason, conditional expressions that use a backrefer-
610 ence as the condition or test for a specific group recursion are not
611 supported.
612
613 5. Callouts are supported, but the value of the capture_top field is
614 always 1, and the value of the capture_last field is always -1.
615
616 6. The \C escape sequence, which (in the standard algorithm) matches a
617 single byte, even in UTF-8 mode, is not supported because the alterna-
618 tive algorithm moves through the subject string one character at a
619 time, for all active paths through the tree.
620
621
622 ADVANTAGES OF THE ALTERNATIVE ALGORITHM
623
624 Using the alternative matching algorithm provides the following advan-
625 tages:
626
627 1. All possible matches (at a single point in the subject) are automat-
628 ically found, and in particular, the longest match is found. To find
629 more than one match using the standard algorithm, you have to do kludgy
630 things with callouts.
631
632 2. There is much better support for partial matching. The restrictions
633 on the content of the pattern that apply when using the standard algo-
634 rithm for partial matching do not apply to the alternative algorithm.
635 For non-anchored patterns, the starting position of a partial match is
636 available.
637
638 3. Because the alternative algorithm scans the subject string just
639 once, and never needs to backtrack, it is possible to pass very long
640 subject strings to the matching function in several pieces, checking
641 for partial matching each time.
642
643
644 DISADVANTAGES OF THE ALTERNATIVE ALGORITHM
645
646 The alternative algorithm suffers from a number of disadvantages:
647
648 1. It is substantially slower than the standard algorithm. This is
649 partly because it has to search for all possible matches, but is also
650 because it is less susceptible to optimization.
651
652 2. Capturing parentheses and back references are not supported.
653
654 3. Although atomic groups are supported, their use does not provide the
655 performance advantage that it does for the standard algorithm.
656
657
658 AUTHOR
659
660 Philip Hazel
661 University Computing Service
662 Cambridge CB2 3QH, England.
663
664
665 REVISION
666
667 Last updated: 06 March 2007
668 Copyright (c) 1997-2007 University of Cambridge.
669 ------------------------------------------------------------------------------
670
671
672 PCREAPI(3) PCREAPI(3)
673
674
675 NAME
676 PCRE - Perl-compatible regular expressions
677
678
679 PCRE NATIVE API
680
681 #include <pcre.h>
682
683 pcre *pcre_compile(const char *pattern, int options,
684 const char **errptr, int *erroffset,
685 const unsigned char *tableptr);
686
687 pcre *pcre_compile2(const char *pattern, int options,
688 int *errorcodeptr,
689 const char **errptr, int *erroffset,
690 const unsigned char *tableptr);
691
692 pcre_extra *pcre_study(const pcre *code, int options,
693 const char **errptr);
694
695 int pcre_exec(const pcre *code, const pcre_extra *extra,
696 const char *subject, int length, int startoffset,
697 int options, int *ovector, int ovecsize);
698
699 int pcre_dfa_exec(const pcre *code, const pcre_extra *extra,
700 const char *subject, int length, int startoffset,
701 int options, int *ovector, int ovecsize,
702 int *workspace, int wscount);
703
704 int pcre_copy_named_substring(const pcre *code,
705 const char *subject, int *ovector,
706 int stringcount, const char *stringname,
707 char *buffer, int buffersize);
708
709 int pcre_copy_substring(const char *subject, int *ovector,
710 int stringcount, int stringnumber, char *buffer,
711 int buffersize);
712
713 int pcre_get_named_substring(const pcre *code,
714 const char *subject, int *ovector,
715 int stringcount, const char *stringname,
716 const char **stringptr);
717
718 int pcre_get_stringnumber(const pcre *code,
719 const char *name);
720
721 int pcre_get_stringtable_entries(const pcre *code,
722 const char *name, char **first, char **last);
723
724 int pcre_get_substring(const char *subject, int *ovector,
725 int stringcount, int stringnumber,
726 const char **stringptr);
727
728 int pcre_get_substring_list(const char *subject,
729 int *ovector, int stringcount, const char ***listptr);
730
731 void pcre_free_substring(const char *stringptr);
732
733 void pcre_free_substring_list(const char **stringptr);
734
735 const unsigned char *pcre_maketables(void);
736
737 int pcre_fullinfo(const pcre *code, const pcre_extra *extra,
738 int what, void *where);
739
740 int pcre_info(const pcre *code, int *optptr, int *firstcharptr);
741
742 int pcre_refcount(pcre *code, int adjust);
743
744 int pcre_config(int what, void *where);
745
746 char *pcre_version(void);
747
748 void *(*pcre_malloc)(size_t);
749
750 void (*pcre_free)(void *);
751
752 void *(*pcre_stack_malloc)(size_t);
753
754 void (*pcre_stack_free)(void *);
755
756 int (*pcre_callout)(pcre_callout_block *);
757
758
759 PCRE API OVERVIEW
760
761 PCRE has its own native API, which is described in this document. There
762 are also some wrapper functions that correspond to the POSIX regular
763 expression API. These are described in the pcreposix documentation.
764 Both of these APIs define a set of C function calls. A C++ wrapper is
765 distributed with PCRE. It is documented in the pcrecpp page.
766
767 The native API C function prototypes are defined in the header file
768 pcre.h, and on Unix systems the library itself is called libpcre. It
769 can normally be accessed by adding -lpcre to the command for linking an
770 application that uses PCRE. The header file defines the macros
771 PCRE_MAJOR and PCRE_MINOR to contain the major and minor release num-
772 bers for the library. Applications can use these to include support
773 for different releases of PCRE.
774
775 The functions pcre_compile(), pcre_compile2(), pcre_study(), and
776 pcre_exec() are used for compiling and matching regular expressions in
777 a Perl-compatible manner. A sample program that demonstrates the sim-
778 plest way of using them is provided in the file called pcredemo.c in
779 the source distribution. The pcresample documentation describes how to
780 run it.
781
782 A second matching function, pcre_dfa_exec(), which is not Perl-compati-
783 ble, is also provided. This uses a different algorithm for the match-
784 ing. The alternative algorithm finds all possible matches (at a given
785 point in the subject), and scans the subject just once. However, this
786 algorithm does not return captured substrings. A description of the two
787 matching algorithms and their advantages and disadvantages is given in
788 the pcrematching documentation.
789
790 In addition to the main compiling and matching functions, there are
791 convenience functions for extracting captured substrings from a subject
792 string that is matched by pcre_exec(). They are:
793
794 pcre_copy_substring()
795 pcre_copy_named_substring()
796 pcre_get_substring()
797 pcre_get_named_substring()
798 pcre_get_substring_list()
799 pcre_get_stringnumber()
800 pcre_get_stringtable_entries()
801
802 pcre_free_substring() and pcre_free_substring_list() are also provided,
803 to free the memory used for extracted strings.
804
805 The function pcre_maketables() is used to build a set of character
806 tables in the current locale for passing to pcre_compile(),
807 pcre_exec(), or pcre_dfa_exec(). This is an optional facility that is
808 provided for specialist use. Most commonly, no special tables are
809 passed, in which case internal tables that are generated when PCRE is
810 built are used.
811
812 The function pcre_fullinfo() is used to find out information about a
813 compiled pattern; pcre_info() is an obsolete version that returns only
814 some of the available information, but is retained for backwards com-
815 patibility. The function pcre_version() returns a pointer to a string
816 containing the version of PCRE and its date of release.
817
818 The function pcre_refcount() maintains a reference count in a data
819 block containing a compiled pattern. This is provided for the benefit
820 of object-oriented applications.
821
822 The global variables pcre_malloc and pcre_free initially contain the
823 entry points of the standard malloc() and free() functions, respec-
824 tively. PCRE calls the memory management functions via these variables,
825 so a calling program can replace them if it wishes to intercept the
826 calls. This should be done before calling any PCRE functions.
827
828 The global variables pcre_stack_malloc and pcre_stack_free are also
829 indirections to memory management functions. These special functions
830 are used only when PCRE is compiled to use the heap for remembering
831 data, instead of recursive function calls, when running the pcre_exec()
832 function. See the pcrebuild documentation for details of how to do
833 this. It is a non-standard way of building PCRE, for use in environ-
834 ments that have limited stacks. Because of the greater use of memory
835 management, it runs more slowly. Separate functions are provided so
836 that special-purpose external code can be used for this case. When
837 used, these functions are always called in a stack-like manner (last
838 obtained, first freed), and always for memory blocks of the same size.
839 There is a discussion about PCRE's stack usage in the pcrestack docu-
840 mentation.
841
842 The global variable pcre_callout initially contains NULL. It can be set
843 by the caller to a "callout" function, which PCRE will then call at
844 specified points during a matching operation. Details are given in the
845 pcrecallout documentation.
846
847
848 NEWLINES
849
850 PCRE supports five different conventions for indicating line breaks in
851 strings: a single CR (carriage return) character, a single LF (line-
852 feed) character, the two-character sequence CRLF, any of the three pre-
853 ceding, or any Unicode newline sequence. The Unicode newline sequences
854 are the three just mentioned, plus the single characters VT (vertical
855 tab, U+000B), FF (formfeed, U+000C), NEL (next line, U+0085), LS (line
856 separator, U+2028), and PS (paragraph separator, U+2029).
857
858 Each of the first three conventions is used by at least one operating
859 system as its standard newline sequence. When PCRE is built, a default
860 can be specified. The default default is LF, which is the Unix stan-
861 dard. When PCRE is run, the default can be overridden, either when a
862 pattern is compiled, or when it is matched.
863
864 In the PCRE documentation the word "newline" is used to mean "the char-
865 acter or pair of characters that indicate a line break". The choice of
866 newline convention affects the handling of the dot, circumflex, and
867 dollar metacharacters, the handling of #-comments in /x mode, and, when
868 CRLF is a recognized line ending sequence, the match position advance-
869 ment for a non-anchored pattern. The choice of newline convention does
870 not affect the interpretation of the \n or \r escape sequences.
871
872
873 MULTITHREADING
874
875 The PCRE functions can be used in multi-threading applications, with
876 the proviso that the memory management functions pointed to by
877 pcre_malloc, pcre_free, pcre_stack_malloc, and pcre_stack_free, and the
878 callout function pointed to by pcre_callout, are shared by all threads.
879
880 The compiled form of a regular expression is not altered during match-
881 ing, so the same compiled pattern can safely be used by several threads
882 at once.
883
884
885 SAVING PRECOMPILED PATTERNS FOR LATER USE
886
887 The compiled form of a regular expression can be saved and re-used at a
888 later time, possibly by a different program, and even on a host other
889 than the one on which it was compiled. Details are given in the
890 pcreprecompile documentation. However, compiling a regular expression
891 with one version of PCRE for use with a different version is not guar-
892 anteed to work and may cause crashes.
893
894
895 CHECKING BUILD-TIME OPTIONS
896
897 int pcre_config(int what, void *where);
898
899 The function pcre_config() makes it possible for a PCRE client to dis-
900 cover which optional features have been compiled into the PCRE library.
901 The pcrebuild documentation has more details about these optional fea-
902 tures.
903
904 The first argument for pcre_config() is an integer, specifying which
905 information is required; the second argument is a pointer to a variable
906 into which the information is placed. The following information is
907 available:
908
909 PCRE_CONFIG_UTF8
910
911 The output is an integer that is set to one if UTF-8 support is avail-
912 able; otherwise it is set to zero.
913
914 PCRE_CONFIG_UNICODE_PROPERTIES
915
916 The output is an integer that is set to one if support for Unicode
917 character properties is available; otherwise it is set to zero.
918
919 PCRE_CONFIG_NEWLINE
920
921 The output is an integer whose value specifies the default character
922 sequence that is recognized as meaning "newline". The four values that
923 are supported are: 10 for LF, 13 for CR, 3338 for CRLF, -2 for ANYCRLF,
924 and -1 for ANY. The default should normally be the standard sequence
925 for your operating system.
926
927 PCRE_CONFIG_LINK_SIZE
928
929 The output is an integer that contains the number of bytes used for
930 internal linkage in compiled regular expressions. The value is 2, 3, or
931 4. Larger values allow larger regular expressions to be compiled, at
932 the expense of slower matching. The default value of 2 is sufficient
933 for all but the most massive patterns, since it allows the compiled
934 pattern to be up to 64K in size.
935
936 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
937
938 The output is an integer that contains the threshold above which the
939 POSIX interface uses malloc() for output vectors. Further details are
940 given in the pcreposix documentation.
941
942 PCRE_CONFIG_MATCH_LIMIT
943
944 The output is an integer that gives the default limit for the number of
945 internal matching function calls in a pcre_exec() execution. Further
946 details are given with pcre_exec() below.
947
948 PCRE_CONFIG_MATCH_LIMIT_RECURSION
949
950 The output is an integer that gives the default limit for the depth of
951 recursion when calling the internal matching function in a pcre_exec()
952 execution. Further details are given with pcre_exec() below.
953
954 PCRE_CONFIG_STACKRECURSE
955
956 The output is an integer that is set to one if internal recursion when
957 running pcre_exec() is implemented by recursive function calls that use
958 the stack to remember their state. This is the usual way that PCRE is
959 compiled. The output is zero if PCRE was compiled to use blocks of data
960 on the heap instead of recursive function calls. In this case,
961 pcre_stack_malloc and pcre_stack_free are called to manage memory
962 blocks on the heap, thus avoiding the use of the stack.
963
964
965 COMPILING A PATTERN
966
967 pcre *pcre_compile(const char *pattern, int options,
968 const char **errptr, int *erroffset,
969 const unsigned char *tableptr);
970
971 pcre *pcre_compile2(const char *pattern, int options,
972 int *errorcodeptr,
973 const char **errptr, int *erroffset,
974 const unsigned char *tableptr);
975
976 Either of the functions pcre_compile() or pcre_compile2() can be called
977 to compile a pattern into an internal form. The only difference between
978 the two interfaces is that pcre_compile2() has an additional argument,
979 errorcodeptr, via which a numerical error code can be returned.
980
981 The pattern is a C string terminated by a binary zero, and is passed in
982 the pattern argument. A pointer to a single block of memory that is
983 obtained via pcre_malloc is returned. This contains the compiled code
984 and related data. The pcre type is defined for the returned block; this
985 is a typedef for a structure whose contents are not externally defined.
986 It is up to the caller to free the memory (via pcre_free) when it is no
987 longer required.
988
989 Although the compiled code of a PCRE regex is relocatable, that is, it
990 does not depend on memory location, the complete pcre data block is not
991 fully relocatable, because it may contain a copy of the tableptr argu-
992 ment, which is an address (see below).
993
994 The options argument contains various bit settings that affect the com-
995 pilation. It should be zero if no options are required. The available
996 options are described below. Some of them, in particular, those that
997 are compatible with Perl, can also be set and unset from within the
998 pattern (see the detailed description in the pcrepattern documenta-
999 tion). For these options, the contents of the options argument speci-
1000 fies their initial settings at the start of compilation and execution.
1001 The PCRE_ANCHORED and PCRE_NEWLINE_xxx options can be set at the time
1002 of matching as well as at compile time.
1003
1004 If errptr is NULL, pcre_compile() returns NULL immediately. Otherwise,
1005 if compilation of a pattern fails, pcre_compile() returns NULL, and
1006 sets the variable pointed to by errptr to point to a textual error mes-
1007 sage. This is a static string that is part of the library. You must not
1008 try to free it. The offset from the start of the pattern to the charac-
1009 ter where the error was discovered is placed in the variable pointed to
1010 by erroffset, which must not be NULL. If it is, an immediate error is
1011 given.
1012
1013 If pcre_compile2() is used instead of pcre_compile(), and the error-
1014 codeptr argument is not NULL, a non-zero error code number is returned
1015 via this argument in the event of an error. This is in addition to the
1016 textual error message. Error codes and messages are listed below.
1017
1018 If the final argument, tableptr, is NULL, PCRE uses a default set of
1019 character tables that are built when PCRE is compiled, using the
1020 default C locale. Otherwise, tableptr must be an address that is the
1021 result of a call to pcre_maketables(). This value is stored with the
1022 compiled pattern, and used again by pcre_exec(), unless another table
1023 pointer is passed to it. For more discussion, see the section on locale
1024 support below.
1025
1026 This code fragment shows a typical straightforward call to pcre_com-
1027 pile():
1028
1029 pcre *re;
1030 const char *error;
1031 int erroffset;
1032 re = pcre_compile(
1033 "^A.*Z", /* the pattern */
1034 0, /* default options */
1035 &error, /* for error message */
1036 &erroffset, /* for error offset */
1037 NULL); /* use default character tables */
1038
1039 The following names for option bits are defined in the pcre.h header
1040 file:
1041
1042 PCRE_ANCHORED
1043
1044 If this bit is set, the pattern is forced to be "anchored", that is, it
1045 is constrained to match only at the first matching point in the string
1046 that is being searched (the "subject string"). This effect can also be
1047 achieved by appropriate constructs in the pattern itself, which is the
1048 only way to do it in Perl.
1049
1050 PCRE_AUTO_CALLOUT
1051
1052 If this bit is set, pcre_compile() automatically inserts callout items,
1053 all with number 255, before each pattern item. For discussion of the
1054 callout facility, see the pcrecallout documentation.
1055
1056 PCRE_CASELESS
1057
1058 If this bit is set, letters in the pattern match both upper and lower
1059 case letters. It is equivalent to Perl's /i option, and it can be
1060 changed within a pattern by a (?i) option setting. In UTF-8 mode, PCRE
1061 always understands the concept of case for characters whose values are
1062 less than 128, so caseless matching is always possible. For characters
1063 with higher values, the concept of case is supported if PCRE is com-
1064 piled with Unicode property support, but not otherwise. If you want to
1065 use caseless matching for characters 128 and above, you must ensure
1066 that PCRE is compiled with Unicode property support as well as with
1067 UTF-8 support.
1068
1069 PCRE_DOLLAR_ENDONLY
1070
1071 If this bit is set, a dollar metacharacter in the pattern matches only
1072 at the end of the subject string. Without this option, a dollar also
1073 matches immediately before a newline at the end of the string (but not
1074 before any other newlines). The PCRE_DOLLAR_ENDONLY option is ignored
1075 if PCRE_MULTILINE is set. There is no equivalent to this option in
1076 Perl, and no way to set it within a pattern.
1077
1078 PCRE_DOTALL
1079
1080 If this bit is set, a dot metacharater in the pattern matches all char-
1081 acters, including those that indicate newline. Without it, a dot does
1082 not match when the current position is at a newline. This option is
1083 equivalent to Perl's /s option, and it can be changed within a pattern
1084 by a (?s) option setting. A negative class such as [^a] always matches
1085 newline characters, independent of the setting of this option.
1086
1087 PCRE_DUPNAMES
1088
1089 If this bit is set, names used to identify capturing subpatterns need
1090 not be unique. This can be helpful for certain types of pattern when it
1091 is known that only one instance of the named subpattern can ever be
1092 matched. There are more details of named subpatterns below; see also
1093 the pcrepattern documentation.
1094
1095 PCRE_EXTENDED
1096
1097 If this bit is set, whitespace data characters in the pattern are
1098 totally ignored except when escaped or inside a character class. White-
1099 space does not include the VT character (code 11). In addition, charac-
1100 ters between an unescaped # outside a character class and the next new-
1101 line, inclusive, are also ignored. This is equivalent to Perl's /x
1102 option, and it can be changed within a pattern by a (?x) option set-
1103 ting.
1104
1105 This option makes it possible to include comments inside complicated
1106 patterns. Note, however, that this applies only to data characters.
1107 Whitespace characters may never appear within special character
1108 sequences in a pattern, for example within the sequence (?( which
1109 introduces a conditional subpattern.
1110
1111 PCRE_EXTRA
1112
1113 This option was invented in order to turn on additional functionality
1114 of PCRE that is incompatible with Perl, but it is currently of very
1115 little use. When set, any backslash in a pattern that is followed by a
1116 letter that has no special meaning causes an error, thus reserving
1117 these combinations for future expansion. By default, as in Perl, a
1118 backslash followed by a letter with no special meaning is treated as a
1119 literal. (Perl can, however, be persuaded to give a warning for this.)
1120 There are at present no other features controlled by this option. It
1121 can also be set by a (?X) option setting within a pattern.
1122
1123 PCRE_FIRSTLINE
1124
1125 If this option is set, an unanchored pattern is required to match
1126 before or at the first newline in the subject string, though the
1127 matched text may continue over the newline.
1128
1129 PCRE_MULTILINE
1130
1131 By default, PCRE treats the subject string as consisting of a single
1132 line of characters (even if it actually contains newlines). The "start
1133 of line" metacharacter (^) matches only at the start of the string,
1134 while the "end of line" metacharacter ($) matches only at the end of
1135 the string, or before a terminating newline (unless PCRE_DOLLAR_ENDONLY
1136 is set). This is the same as Perl.
1137
1138 When PCRE_MULTILINE it is set, the "start of line" and "end of line"
1139 constructs match immediately following or immediately before internal
1140 newlines in the subject string, respectively, as well as at the very
1141 start and end. This is equivalent to Perl's /m option, and it can be
1142 changed within a pattern by a (?m) option setting. If there are no new-
1143 lines in a subject string, or no occurrences of ^ or $ in a pattern,
1144 setting PCRE_MULTILINE has no effect.
1145
1146 PCRE_NEWLINE_CR
1147 PCRE_NEWLINE_LF
1148 PCRE_NEWLINE_CRLF
1149 PCRE_NEWLINE_ANYCRLF
1150 PCRE_NEWLINE_ANY
1151
1152 These options override the default newline definition that was chosen
1153 when PCRE was built. Setting the first or the second specifies that a
1154 newline is indicated by a single character (CR or LF, respectively).
1155 Setting PCRE_NEWLINE_CRLF specifies that a newline is indicated by the
1156 two-character CRLF sequence. Setting PCRE_NEWLINE_ANYCRLF specifies
1157 that any of the three preceding sequences should be recognized. Setting
1158 PCRE_NEWLINE_ANY specifies that any Unicode newline sequence should be
1159 recognized. The Unicode newline sequences are the three just mentioned,
1160 plus the single characters VT (vertical tab, U+000B), FF (formfeed,
1161 U+000C), NEL (next line, U+0085), LS (line separator, U+2028), and PS
1162 (paragraph separator, U+2029). The last two are recognized only in
1163 UTF-8 mode.
1164
1165 The newline setting in the options word uses three bits that are
1166 treated as a number, giving eight possibilities. Currently only six are
1167 used (default plus the five values above). This means that if you set
1168 more than one newline option, the combination may or may not be sensi-
1169 ble. For example, PCRE_NEWLINE_CR with PCRE_NEWLINE_LF is equivalent to
1170 PCRE_NEWLINE_CRLF, but other combinations may yield unused numbers and
1171 cause an error.
1172
1173 The only time that a line break is specially recognized when compiling
1174 a pattern is if PCRE_EXTENDED is set, and an unescaped # outside a
1175 character class is encountered. This indicates a comment that lasts
1176 until after the next line break sequence. In other circumstances, line
1177 break sequences are treated as literal data, except that in
1178 PCRE_EXTENDED mode, both CR and LF are treated as whitespace characters
1179 and are therefore ignored.
1180
1181 The newline option that is set at compile time becomes the default that
1182 is used for pcre_exec() and pcre_dfa_exec(), but it can be overridden.
1183
1184 PCRE_NO_AUTO_CAPTURE
1185
1186 If this option is set, it disables the use of numbered capturing paren-
1187 theses in the pattern. Any opening parenthesis that is not followed by
1188 ? behaves as if it were followed by ?: but named parentheses can still
1189 be used for capturing (and they acquire numbers in the usual way).
1190 There is no equivalent of this option in Perl.
1191
1192 PCRE_UNGREEDY
1193
1194 This option inverts the "greediness" of the quantifiers so that they
1195 are not greedy by default, but become greedy if followed by "?". It is
1196 not compatible with Perl. It can also be set by a (?U) option setting
1197 within the pattern.
1198
1199 PCRE_UTF8
1200
1201 This option causes PCRE to regard both the pattern and the subject as
1202 strings of UTF-8 characters instead of single-byte character strings.
1203 However, it is available only when PCRE is built to include UTF-8 sup-
1204 port. If not, the use of this option provokes an error. Details of how
1205 this option changes the behaviour of PCRE are given in the section on
1206 UTF-8 support in the main pcre page.
1207
1208 PCRE_NO_UTF8_CHECK
1209
1210 When PCRE_UTF8 is set, the validity of the pattern as a UTF-8 string is
1211 automatically checked. If an invalid UTF-8 sequence of bytes is found,
1212 pcre_compile() returns an error. If you already know that your pattern
1213 is valid, and you want to skip this check for performance reasons, you
1214 can set the PCRE_NO_UTF8_CHECK option. When it is set, the effect of
1215 passing an invalid UTF-8 string as a pattern is undefined. It may cause
1216 your program to crash. Note that this option can also be passed to
1217 pcre_exec() and pcre_dfa_exec(), to suppress the UTF-8 validity check-
1218 ing of subject strings.
1219
1220
1221 COMPILATION ERROR CODES
1222
1223 The following table lists the error codes than may be returned by
1224 pcre_compile2(), along with the error messages that may be returned by
1225 both compiling functions. As PCRE has developed, some error codes have
1226 fallen out of use. To avoid confusion, they have not been re-used.
1227
1228 0 no error
1229 1 \ at end of pattern
1230 2 \c at end of pattern
1231 3 unrecognized character follows \
1232 4 numbers out of order in {} quantifier
1233 5 number too big in {} quantifier
1234 6 missing terminating ] for character class
1235 7 invalid escape sequence in character class
1236 8 range out of order in character class
1237 9 nothing to repeat
1238 10 [this code is not in use]
1239 11 internal error: unexpected repeat
1240 12 unrecognized character after (?
1241 13 POSIX named classes are supported only within a class
1242 14 missing )
1243 15 reference to non-existent subpattern
1244 16 erroffset passed as NULL
1245 17 unknown option bit(s) set
1246 18 missing ) after comment
1247 19 [this code is not in use]
1248 20 regular expression too large
1249 21 failed to get memory
1250 22 unmatched parentheses
1251 23 internal error: code overflow
1252 24 unrecognized character after (?<
1253 25 lookbehind assertion is not fixed length
1254 26 malformed number or name after (?(
1255 27 conditional group contains more than two branches
1256 28 assertion expected after (?(
1257 29 (?R or (?digits must be followed by )
1258 30 unknown POSIX class name
1259 31 POSIX collating elements are not supported
1260 32 this version of PCRE is not compiled with PCRE_UTF8 support
1261 33 [this code is not in use]
1262 34 character value in \x{...} sequence is too large
1263 35 invalid condition (?(0)
1264 36 \C not allowed in lookbehind assertion
1265 37 PCRE does not support \L, \l, \N, \U, or \u
1266 38 number after (?C is > 255
1267 39 closing ) for (?C expected
1268 40 recursive call could loop indefinitely
1269 41 unrecognized character after (?P
1270 42 syntax error in subpattern name (missing terminator)
1271 43 two named subpatterns have the same name
1272 44 invalid UTF-8 string
1273 45 support for \P, \p, and \X has not been compiled
1274 46 malformed \P or \p sequence
1275 47 unknown property name after \P or \p
1276 48 subpattern name is too long (maximum 32 characters)
1277 49 too many named subpatterns (maximum 10,000)
1278 50 repeated subpattern is too long
1279 51 octal value is greater than \377 (not in UTF-8 mode)
1280 52 internal error: overran compiling workspace
1281 53 internal error: previously-checked referenced subpattern not
1282 found
1283 54 DEFINE group contains more than one branch
1284 55 repeating a DEFINE group is not allowed
1285 56 inconsistent NEWLINE options"
1286
1287
1288 STUDYING A PATTERN
1289
1290 pcre_extra *pcre_study(const pcre *code, int options
1291 const char **errptr);
1292
1293 If a compiled pattern is going to be used several times, it is worth
1294 spending more time analyzing it in order to speed up the time taken for
1295 matching. The function pcre_study() takes a pointer to a compiled pat-
1296 tern as its first argument. If studying the pattern produces additional
1297 information that will help speed up matching, pcre_study() returns a
1298 pointer to a pcre_extra block, in which the study_data field points to
1299 the results of the study.
1300
1301 The returned value from pcre_study() can be passed directly to
1302 pcre_exec(). However, a pcre_extra block also contains other fields
1303 that can be set by the caller before the block is passed; these are
1304 described below in the section on matching a pattern.
1305
1306 If studying the pattern does not produce any additional information
1307 pcre_study() returns NULL. In that circumstance, if the calling program
1308 wants to pass any of the other fields to pcre_exec(), it must set up
1309 its own pcre_extra block.
1310
1311 The second argument of pcre_study() contains option bits. At present,
1312 no options are defined, and this argument should always be zero.
1313
1314 The third argument for pcre_study() is a pointer for an error message.
1315 If studying succeeds (even if no data is returned), the variable it
1316 points to is set to NULL. Otherwise it is set to point to a textual
1317 error message. This is a static string that is part of the library. You
1318 must not try to free it. You should test the error pointer for NULL
1319 after calling pcre_study(), to be sure that it has run successfully.
1320
1321 This is a typical call to pcre_study():
1322
1323 pcre_extra *pe;
1324 pe = pcre_study(
1325 re, /* result of pcre_compile() */
1326 0, /* no options exist */
1327 &error); /* set to NULL or points to a message */
1328
1329 At present, studying a pattern is useful only for non-anchored patterns
1330 that do not have a single fixed starting character. A bitmap of possi-
1331 ble starting bytes is created.
1332
1333
1334 LOCALE SUPPORT
1335
1336 PCRE handles caseless matching, and determines whether characters are
1337 letters, digits, or whatever, by reference to a set of tables, indexed
1338 by character value. When running in UTF-8 mode, this applies only to
1339 characters with codes less than 128. Higher-valued codes never match
1340 escapes such as \w or \d, but can be tested with \p if PCRE is built
1341 with Unicode character property support. The use of locales with Uni-
1342 code is discouraged. If you are handling characters with codes greater
1343 than 128, you should either use UTF-8 and Unicode, or use locales, but
1344 not try to mix the two.
1345
1346 PCRE contains an internal set of tables that are used when the final
1347 argument of pcre_compile() is NULL. These are sufficient for many
1348 applications. Normally, the internal tables recognize only ASCII char-
1349 acters. However, when PCRE is built, it is possible to cause the inter-
1350 nal tables to be rebuilt in the default "C" locale of the local system,
1351 which may cause them to be different.
1352
1353 The internal tables can always be overridden by tables supplied by the
1354 application that calls PCRE. These may be created in a different locale
1355 from the default. As more and more applications change to using Uni-
1356 code, the need for this locale support is expected to die away.
1357
1358 External tables are built by calling the pcre_maketables() function,
1359 which has no arguments, in the relevant locale. The result can then be
1360 passed to pcre_compile() or pcre_exec() as often as necessary. For
1361 example, to build and use tables that are appropriate for the French
1362 locale (where accented characters with values greater than 128 are
1363 treated as letters), the following code could be used:
1364
1365 setlocale(LC_CTYPE, "fr_FR");
1366 tables = pcre_maketables();
1367 re = pcre_compile(..., tables);
1368
1369 The locale name "fr_FR" is used on Linux and other Unix-like systems;
1370 if you are using Windows, the name for the French locale is "french".
1371
1372 When pcre_maketables() runs, the tables are built in memory that is
1373 obtained via pcre_malloc. It is the caller's responsibility to ensure
1374 that the memory containing the tables remains available for as long as
1375 it is needed.
1376
1377 The pointer that is passed to pcre_compile() is saved with the compiled
1378 pattern, and the same tables are used via this pointer by pcre_study()
1379 and normally also by pcre_exec(). Thus, by default, for any single pat-
1380 tern, compilation, studying and matching all happen in the same locale,
1381 but different patterns can be compiled in different locales.
1382
1383 It is possible to pass a table pointer or NULL (indicating the use of
1384 the internal tables) to pcre_exec(). Although not intended for this
1385 purpose, this facility could be used to match a pattern in a different
1386 locale from the one in which it was compiled. Passing table pointers at
1387 run time is discussed below in the section on matching a pattern.
1388
1389
1390 INFORMATION ABOUT A PATTERN
1391
1392 int pcre_fullinfo(const pcre *code, const pcre_extra *extra,
1393 int what, void *where);
1394
1395 The pcre_fullinfo() function returns information about a compiled pat-
1396 tern. It replaces the obsolete pcre_info() function, which is neverthe-
1397 less retained for backwards compability (and is documented below).
1398
1399 The first argument for pcre_fullinfo() is a pointer to the compiled
1400 pattern. The second argument is the result of pcre_study(), or NULL if
1401 the pattern was not studied. The third argument specifies which piece
1402 of information is required, and the fourth argument is a pointer to a
1403 variable to receive the data. The yield of the function is zero for
1404 success, or one of the following negative numbers:
1405
1406 PCRE_ERROR_NULL the argument code was NULL
1407 the argument where was NULL
1408 PCRE_ERROR_BADMAGIC the "magic number" was not found
1409 PCRE_ERROR_BADOPTION the value of what was invalid
1410
1411 The "magic number" is placed at the start of each compiled pattern as
1412 an simple check against passing an arbitrary memory pointer. Here is a
1413 typical call of pcre_fullinfo(), to obtain the length of the compiled
1414 pattern:
1415
1416 int rc;
1417 size_t length;
1418 rc = pcre_fullinfo(
1419 re, /* result of pcre_compile() */
1420 pe, /* result of pcre_study(), or NULL */
1421 PCRE_INFO_SIZE, /* what is required */
1422 &length); /* where to put the data */
1423
1424 The possible values for the third argument are defined in pcre.h, and
1425 are as follows:
1426
1427 PCRE_INFO_BACKREFMAX
1428
1429 Return the number of the highest back reference in the pattern. The
1430 fourth argument should point to an int variable. Zero is returned if
1431 there are no back references.
1432
1433 PCRE_INFO_CAPTURECOUNT
1434
1435 Return the number of capturing subpatterns in the pattern. The fourth
1436 argument should point to an int variable.
1437
1438 PCRE_INFO_DEFAULT_TABLES
1439
1440 Return a pointer to the internal default character tables within PCRE.
1441 The fourth argument should point to an unsigned char * variable. This
1442 information call is provided for internal use by the pcre_study() func-
1443 tion. External callers can cause PCRE to use its internal tables by
1444 passing a NULL table pointer.
1445
1446 PCRE_INFO_FIRSTBYTE
1447
1448 Return information about the first byte of any matched string, for a
1449 non-anchored pattern. The fourth argument should point to an int vari-
1450 able. (This option used to be called PCRE_INFO_FIRSTCHAR; the old name
1451 is still recognized for backwards compatibility.)
1452
1453 If there is a fixed first byte, for example, from a pattern such as
1454 (cat|cow|coyote), its value is returned. Otherwise, if either
1455
1456 (a) the pattern was compiled with the PCRE_MULTILINE option, and every
1457 branch starts with "^", or
1458
1459 (b) every branch of the pattern starts with ".*" and PCRE_DOTALL is not
1460 set (if it were set, the pattern would be anchored),
1461
1462 -1 is returned, indicating that the pattern matches only at the start
1463 of a subject string or after any newline within the string. Otherwise
1464 -2 is returned. For anchored patterns, -2 is returned.
1465
1466 PCRE_INFO_FIRSTTABLE
1467
1468 If the pattern was studied, and this resulted in the construction of a
1469 256-bit table indicating a fixed set of bytes for the first byte in any
1470 matching string, a pointer to the table is returned. Otherwise NULL is
1471 returned. The fourth argument should point to an unsigned char * vari-
1472 able.
1473
1474 PCRE_INFO_LASTLITERAL
1475
1476 Return the value of the rightmost literal byte that must exist in any
1477 matched string, other than at its start, if such a byte has been
1478 recorded. The fourth argument should point to an int variable. If there
1479 is no such byte, -1 is returned. For anchored patterns, a last literal
1480 byte is recorded only if it follows something of variable length. For
1481 example, for the pattern /^a\d+z\d+/ the returned value is "z", but for
1482 /^a\dz\d/ the returned value is -1.
1483
1484 PCRE_INFO_NAMECOUNT
1485 PCRE_INFO_NAMEENTRYSIZE
1486 PCRE_INFO_NAMETABLE
1487
1488 PCRE supports the use of named as well as numbered capturing parenthe-
1489 ses. The names are just an additional way of identifying the parenthe-
1490 ses, which still acquire numbers. Several convenience functions such as
1491 pcre_get_named_substring() are provided for extracting captured sub-
1492 strings by name. It is also possible to extract the data directly, by
1493 first converting the name to a number in order to access the correct
1494 pointers in the output vector (described with pcre_exec() below). To do
1495 the conversion, you need to use the name-to-number map, which is
1496 described by these three values.
1497
1498 The map consists of a number of fixed-size entries. PCRE_INFO_NAMECOUNT
1499 gives the number of entries, and PCRE_INFO_NAMEENTRYSIZE gives the size
1500 of each entry; both of these return an int value. The entry size
1501 depends on the length of the longest name. PCRE_INFO_NAMETABLE returns
1502 a pointer to the first entry of the table (a pointer to char). The
1503 first two bytes of each entry are the number of the capturing parenthe-
1504 sis, most significant byte first. The rest of the entry is the corre-
1505 sponding name, zero terminated. The names are in alphabetical order.
1506 When PCRE_DUPNAMES is set, duplicate names are in order of their paren-
1507 theses numbers. For example, consider the following pattern (assume
1508 PCRE_EXTENDED is set, so white space - including newlines - is
1509 ignored):
1510
1511 (?<date> (?<year>(\d\d)?\d\d) -
1512 (?<month>\d\d) - (?<day>\d\d) )
1513
1514 There are four named subpatterns, so the table has four entries, and
1515 each entry in the table is eight bytes long. The table is as follows,
1516 with non-printing bytes shows in hexadecimal, and undefined bytes shown
1517 as ??:
1518
1519 00 01 d a t e 00 ??
1520 00 05 d a y 00 ?? ??
1521 00 04 m o n t h 00
1522 00 02 y e a r 00 ??
1523
1524 When writing code to extract data from named subpatterns using the
1525 name-to-number map, remember that the length of the entries is likely
1526 to be different for each compiled pattern.
1527
1528 PCRE_INFO_OPTIONS
1529
1530 Return a copy of the options with which the pattern was compiled. The
1531 fourth argument should point to an unsigned long int variable. These
1532 option bits are those specified in the call to pcre_compile(), modified
1533 by any top-level option settings within the pattern itself.
1534
1535 A pattern is automatically anchored by PCRE if all of its top-level
1536 alternatives begin with one of the following:
1537
1538 ^ unless PCRE_MULTILINE is set
1539 \A always
1540 \G always
1541 .* if PCRE_DOTALL is set and there are no back
1542 references to the subpattern in which .* appears
1543
1544 For such patterns, the PCRE_ANCHORED bit is set in the options returned
1545 by pcre_fullinfo().
1546
1547 PCRE_INFO_SIZE
1548
1549 Return the size of the compiled pattern, that is, the value that was
1550 passed as the argument to pcre_malloc() when PCRE was getting memory in
1551 which to place the compiled data. The fourth argument should point to a
1552 size_t variable.
1553
1554 PCRE_INFO_STUDYSIZE
1555
1556 Return the size of the data block pointed to by the study_data field in
1557 a pcre_extra block. That is, it is the value that was passed to
1558 pcre_malloc() when PCRE was getting memory into which to place the data
1559 created by pcre_study(). The fourth argument should point to a size_t
1560 variable.
1561
1562
1563 OBSOLETE INFO FUNCTION
1564
1565 int pcre_info(const pcre *code, int *optptr, int *firstcharptr);
1566
1567 The pcre_info() function is now obsolete because its interface is too
1568 restrictive to return all the available data about a compiled pattern.
1569 New programs should use pcre_fullinfo() instead. The yield of
1570 pcre_info() is the number of capturing subpatterns, or one of the fol-
1571 lowing negative numbers:
1572
1573 PCRE_ERROR_NULL the argument code was NULL
1574 PCRE_ERROR_BADMAGIC the "magic number" was not found
1575
1576 If the optptr argument is not NULL, a copy of the options with which
1577 the pattern was compiled is placed in the integer it points to (see
1578 PCRE_INFO_OPTIONS above).
1579
1580 If the pattern is not anchored and the firstcharptr argument is not
1581 NULL, it is used to pass back information about the first character of
1582 any matched string (see PCRE_INFO_FIRSTBYTE above).
1583
1584
1585 REFERENCE COUNTS
1586
1587 int pcre_refcount(pcre *code, int adjust);
1588
1589 The pcre_refcount() function is used to maintain a reference count in
1590 the data block that contains a compiled pattern. It is provided for the
1591 benefit of applications that operate in an object-oriented manner,
1592 where different parts of the application may be using the same compiled
1593 pattern, but you want to free the block when they are all done.
1594
1595 When a pattern is compiled, the reference count field is initialized to
1596 zero. It is changed only by calling this function, whose action is to
1597 add the adjust value (which may be positive or negative) to it. The
1598 yield of the function is the new value. However, the value of the count
1599 is constrained to lie between 0 and 65535, inclusive. If the new value
1600 is outside these limits, it is forced to the appropriate limit value.
1601
1602 Except when it is zero, the reference count is not correctly preserved
1603 if a pattern is compiled on one host and then transferred to a host
1604 whose byte-order is different. (This seems a highly unlikely scenario.)
1605
1606
1607 MATCHING A PATTERN: THE TRADITIONAL FUNCTION
1608
1609 int pcre_exec(const pcre *code, const pcre_extra *extra,
1610 const char *subject, int length, int startoffset,
1611 int options, int *ovector, int ovecsize);
1612
1613 The function pcre_exec() is called to match a subject string against a
1614 compiled pattern, which is passed in the code argument. If the pattern
1615 has been studied, the result of the study should be passed in the extra
1616 argument. This function is the main matching facility of the library,
1617 and it operates in a Perl-like manner. For specialist use there is also
1618 an alternative matching function, which is described below in the sec-
1619 tion about the pcre_dfa_exec() function.
1620
1621 In most applications, the pattern will have been compiled (and option-
1622 ally studied) in the same process that calls pcre_exec(). However, it
1623 is possible to save compiled patterns and study data, and then use them
1624 later in different processes, possibly even on different hosts. For a
1625 discussion about this, see the pcreprecompile documentation.
1626
1627 Here is an example of a simple call to pcre_exec():
1628
1629 int rc;
1630 int ovector[30];
1631 rc = pcre_exec(
1632 re, /* result of pcre_compile() */
1633 NULL, /* we didn't study the pattern */
1634 "some string", /* the subject string */
1635 11, /* the length of the subject string */
1636 0, /* start at offset 0 in the subject */
1637 0, /* default options */
1638 ovector, /* vector of integers for substring information */
1639 30); /* number of elements (NOT size in bytes) */
1640
1641 Extra data for pcre_exec()
1642
1643 If the extra argument is not NULL, it must point to a pcre_extra data
1644 block. The pcre_study() function returns such a block (when it doesn't
1645 return NULL), but you can also create one for yourself, and pass addi-
1646 tional information in it. The pcre_extra block contains the following
1647 fields (not necessarily in this order):
1648
1649 unsigned long int flags;
1650 void *study_data;
1651 unsigned long int match_limit;
1652 unsigned long int match_limit_recursion;
1653 void *callout_data;
1654 const unsigned char *tables;
1655
1656 The flags field is a bitmap that specifies which of the other fields
1657 are set. The flag bits are:
1658
1659 PCRE_EXTRA_STUDY_DATA
1660 PCRE_EXTRA_MATCH_LIMIT
1661 PCRE_EXTRA_MATCH_LIMIT_RECURSION
1662 PCRE_EXTRA_CALLOUT_DATA
1663 PCRE_EXTRA_TABLES
1664
1665 Other flag bits should be set to zero. The study_data field is set in
1666 the pcre_extra block that is returned by pcre_study(), together with
1667 the appropriate flag bit. You should not set this yourself, but you may
1668 add to the block by setting the other fields and their corresponding
1669 flag bits.
1670
1671 The match_limit field provides a means of preventing PCRE from using up
1672 a vast amount of resources when running patterns that are not going to
1673 match, but which have a very large number of possibilities in their
1674 search trees. The classic example is the use of nested unlimited
1675 repeats.
1676
1677 Internally, PCRE uses a function called match() which it calls repeat-
1678 edly (sometimes recursively). The limit set by match_limit is imposed
1679 on the number of times this function is called during a match, which
1680 has the effect of limiting the amount of backtracking that can take
1681 place. For patterns that are not anchored, the count restarts from zero
1682 for each position in the subject string.
1683
1684 The default value for the limit can be set when PCRE is built; the
1685 default default is 10 million, which handles all but the most extreme
1686 cases. You can override the default by suppling pcre_exec() with a
1687 pcre_extra block in which match_limit is set, and
1688 PCRE_EXTRA_MATCH_LIMIT is set in the flags field. If the limit is
1689 exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
1690
1691 The match_limit_recursion field is similar to match_limit, but instead
1692 of limiting the total number of times that match() is called, it limits
1693 the depth of recursion. The recursion depth is a smaller number than
1694 the total number of calls, because not all calls to match() are recur-
1695 sive. This limit is of use only if it is set smaller than match_limit.
1696
1697 Limiting the recursion depth limits the amount of stack that can be
1698 used, or, when PCRE has been compiled to use memory on the heap instead
1699 of the stack, the amount of heap memory that can be used.
1700
1701 The default value for match_limit_recursion can be set when PCRE is
1702 built; the default default is the same value as the default for
1703 match_limit. You can override the default by suppling pcre_exec() with
1704 a pcre_extra block in which match_limit_recursion is set, and
1705 PCRE_EXTRA_MATCH_LIMIT_RECURSION is set in the flags field. If the
1706 limit is exceeded, pcre_exec() returns PCRE_ERROR_RECURSIONLIMIT.
1707
1708 The pcre_callout field is used in conjunction with the "callout" fea-
1709 ture, which is described in the pcrecallout documentation.
1710
1711 The tables field is used to pass a character tables pointer to
1712 pcre_exec(); this overrides the value that is stored with the compiled
1713 pattern. A non-NULL value is stored with the compiled pattern only if
1714 custom tables were supplied to pcre_compile() via its tableptr argu-
1715 ment. If NULL is passed to pcre_exec() using this mechanism, it forces
1716 PCRE's internal tables to be used. This facility is helpful when re-
1717 using patterns that have been saved after compiling with an external
1718 set of tables, because the external tables might be at a different
1719 address when pcre_exec() is called. See the pcreprecompile documenta-
1720 tion for a discussion of saving compiled patterns for later use.
1721
1722 Option bits for pcre_exec()
1723
1724 The unused bits of the options argument for pcre_exec() must be zero.
1725 The only bits that may be set are PCRE_ANCHORED, PCRE_NEWLINE_xxx,
1726 PCRE_NOTBOL, PCRE_NOTEOL, PCRE_NOTEMPTY, PCRE_NO_UTF8_CHECK and
1727 PCRE_PARTIAL.
1728
1729 PCRE_ANCHORED
1730
1731 The PCRE_ANCHORED option limits pcre_exec() to matching at the first
1732 matching position. If a pattern was compiled with PCRE_ANCHORED, or
1733 turned out to be anchored by virtue of its contents, it cannot be made
1734 unachored at matching time.
1735
1736 PCRE_NEWLINE_CR
1737 PCRE_NEWLINE_LF
1738 PCRE_NEWLINE_CRLF
1739 PCRE_NEWLINE_ANYCRLF
1740 PCRE_NEWLINE_ANY
1741
1742 These options override the newline definition that was chosen or
1743 defaulted when the pattern was compiled. For details, see the descrip-
1744 tion of pcre_compile() above. During matching, the newline choice
1745 affects the behaviour of the dot, circumflex, and dollar metacharac-
1746 ters. It may also alter the way the match position is advanced after a
1747 match failure for an unanchored pattern. When PCRE_NEWLINE_CRLF,
1748 PCRE_NEWLINE_ANYCRLF, or PCRE_NEWLINE_ANY is set, and a match attempt
1749 fails when the current position is at a CRLF sequence, the match posi-
1750 tion is advanced by two characters instead of one, in other words, to
1751 after the CRLF.
1752
1753 PCRE_NOTBOL
1754
1755 This option specifies that first character of the subject string is not
1756 the beginning of a line, so the circumflex metacharacter should not
1757 match before it. Setting this without PCRE_MULTILINE (at compile time)
1758 causes circumflex never to match. This option affects only the behav-
1759 iour of the circumflex metacharacter. It does not affect \A.
1760
1761 PCRE_NOTEOL
1762
1763 This option specifies that the end of the subject string is not the end
1764 of a line, so the dollar metacharacter should not match it nor (except
1765 in multiline mode) a newline immediately before it. Setting this with-
1766 out PCRE_MULTILINE (at compile time) causes dollar never to match. This
1767 option affects only the behaviour of the dollar metacharacter. It does
1768 not affect \Z or \z.
1769
1770 PCRE_NOTEMPTY
1771
1772 An empty string is not considered to be a valid match if this option is
1773 set. If there are alternatives in the pattern, they are tried. If all
1774 the alternatives match the empty string, the entire match fails. For
1775 example, if the pattern
1776
1777 a?b?
1778
1779 is applied to a string not beginning with "a" or "b", it matches the
1780 empty string at the start of the subject. With PCRE_NOTEMPTY set, this
1781 match is not valid, so PCRE searches further into the string for occur-
1782 rences of "a" or "b".
1783
1784 Perl has no direct equivalent of PCRE_NOTEMPTY, but it does make a spe-
1785 cial case of a pattern match of the empty string within its split()
1786 function, and when using the /g modifier. It is possible to emulate
1787 Perl's behaviour after matching a null string by first trying the match
1788 again at the same offset with PCRE_NOTEMPTY and PCRE_ANCHORED, and then
1789 if that fails by advancing the starting offset (see below) and trying
1790 an ordinary match again. There is some code that demonstrates how to do
1791 this in the pcredemo.c sample program.
1792
1793 PCRE_NO_UTF8_CHECK
1794
1795 When PCRE_UTF8 is set at compile time, the validity of the subject as a
1796 UTF-8 string is automatically checked when pcre_exec() is subsequently
1797 called. The value of startoffset is also checked to ensure that it
1798 points to the start of a UTF-8 character. If an invalid UTF-8 sequence
1799 of bytes is found, pcre_exec() returns the error PCRE_ERROR_BADUTF8. If
1800 startoffset contains an invalid value, PCRE_ERROR_BADUTF8_OFFSET is
1801 returned.
1802
1803 If you already know that your subject is valid, and you want to skip
1804 these checks for performance reasons, you can set the
1805 PCRE_NO_UTF8_CHECK option when calling pcre_exec(). You might want to
1806 do this for the second and subsequent calls to pcre_exec() if you are
1807 making repeated calls to find all the matches in a single subject
1808 string. However, you should be sure that the value of startoffset
1809 points to the start of a UTF-8 character. When PCRE_NO_UTF8_CHECK is
1810 set, the effect of passing an invalid UTF-8 string as a subject, or a
1811 value of startoffset that does not point to the start of a UTF-8 char-
1812 acter, is undefined. Your program may crash.
1813
1814 PCRE_PARTIAL
1815
1816 This option turns on the partial matching feature. If the subject
1817 string fails to match the pattern, but at some point during the match-
1818 ing process the end of the subject was reached (that is, the subject
1819 partially matches the pattern and the failure to match occurred only
1820 because there were not enough subject characters), pcre_exec() returns
1821 PCRE_ERROR_PARTIAL instead of PCRE_ERROR_NOMATCH. When PCRE_PARTIAL is
1822 used, there are restrictions on what may appear in the pattern. These
1823 are discussed in the pcrepartial documentation.
1824
1825 The string to be matched by pcre_exec()
1826
1827 The subject string is passed to pcre_exec() as a pointer in subject, a
1828 length in length, and a starting byte offset in startoffset. In UTF-8
1829 mode, the byte offset must point to the start of a UTF-8 character.
1830 Unlike the pattern string, the subject may contain binary zero bytes.
1831 When the starting offset is zero, the search for a match starts at the
1832 beginning of the subject, and this is by far the most common case.
1833
1834 A non-zero starting offset is useful when searching for another match
1835 in the same subject by calling pcre_exec() again after a previous suc-
1836 cess. Setting startoffset differs from just passing over a shortened
1837 string and setting PCRE_NOTBOL in the case of a pattern that begins
1838 with any kind of lookbehind. For example, consider the pattern
1839
1840 \Biss\B
1841
1842 which finds occurrences of "iss" in the middle of words. (\B matches
1843 only if the current position in the subject is not a word boundary.)
1844 When applied to the string "Mississipi" the first call to pcre_exec()
1845 finds the first occurrence. If pcre_exec() is called again with just
1846 the remainder of the subject, namely "issipi", it does not match,
1847 because \B is always false at the start of the subject, which is deemed
1848 to be a word boundary. However, if pcre_exec() is passed the entire
1849 string again, but with startoffset set to 4, it finds the second occur-
1850 rence of "iss" because it is able to look behind the starting point to
1851 discover that it is preceded by a letter.
1852
1853 If a non-zero starting offset is passed when the pattern is anchored,
1854 one attempt to match at the given offset is made. This can only succeed
1855 if the pattern does not require the match to be at the start of the
1856 subject.
1857
1858 How pcre_exec() returns captured substrings
1859
1860 In general, a pattern matches a certain portion of the subject, and in
1861 addition, further substrings from the subject may be picked out by
1862 parts of the pattern. Following the usage in Jeffrey Friedl's book,
1863 this is called "capturing" in what follows, and the phrase "capturing
1864 subpattern" is used for a fragment of a pattern that picks out a sub-
1865 string. PCRE supports several other kinds of parenthesized subpattern
1866 that do not cause substrings to be captured.
1867
1868 Captured substrings are returned to the caller via a vector of integer
1869 offsets whose address is passed in ovector. The number of elements in
1870 the vector is passed in ovecsize, which must be a non-negative number.
1871 Note: this argument is NOT the size of ovector in bytes.
1872
1873 The first two-thirds of the vector is used to pass back captured sub-
1874 strings, each substring using a pair of integers. The remaining third
1875 of the vector is used as workspace by pcre_exec() while matching cap-
1876 turing subpatterns, and is not available for passing back information.
1877 The length passed in ovecsize should always be a multiple of three. If
1878 it is not, it is rounded down.
1879
1880 When a match is successful, information about captured substrings is
1881 returned in pairs of integers, starting at the beginning of ovector,
1882 and continuing up to two-thirds of its length at the most. The first
1883 element of a pair is set to the offset of the first character in a sub-
1884 string, and the second is set to the offset of the first character
1885 after the end of a substring. The first pair, ovector[0] and ovec-
1886 tor[1], identify the portion of the subject string matched by the
1887 entire pattern. The next pair is used for the first capturing subpat-
1888 tern, and so on. The value returned by pcre_exec() is one more than the
1889 highest numbered pair that has been set. For example, if two substrings
1890 have been captured, the returned value is 3. If there are no capturing
1891 subpatterns, the return value from a successful match is 1, indicating
1892 that just the first pair of offsets has been set.
1893
1894 If a capturing subpattern is matched repeatedly, it is the last portion
1895 of the string that it matched that is returned.
1896
1897 If the vector is too small to hold all the captured substring offsets,
1898 it is used as far as possible (up to two-thirds of its length), and the
1899 function returns a value of zero. In particular, if the substring off-
1900 sets are not of interest, pcre_exec() may be called with ovector passed
1901 as NULL and ovecsize as zero. However, if the pattern contains back
1902 references and the ovector is not big enough to remember the related
1903 substrings, PCRE has to get additional memory for use during matching.
1904 Thus it is usually advisable to supply an ovector.
1905
1906 The pcre_info() function can be used to find out how many capturing
1907 subpatterns there are in a compiled pattern. The smallest size for
1908 ovector that will allow for n captured substrings, in addition to the
1909 offsets of the substring matched by the whole pattern, is (n+1)*3.
1910
1911 It is possible for capturing subpattern number n+1 to match some part
1912 of the subject when subpattern n has not been used at all. For example,
1913 if the string "abc" is matched against the pattern (a|(z))(bc) the
1914 return from the function is 4, and subpatterns 1 and 3 are matched, but
1915 2 is not. When this happens, both values in the offset pairs corre-
1916 sponding to unused subpatterns are set to -1.
1917
1918 Offset values that correspond to unused subpatterns at the end of the
1919 expression are also set to -1. For example, if the string "abc" is
1920 matched against the pattern (abc)(x(yz)?)? subpatterns 2 and 3 are not
1921 matched. The return from the function is 2, because the highest used
1922 capturing subpattern number is 1. However, you can refer to the offsets
1923 for the second and third capturing subpatterns if you wish (assuming
1924 the vector is large enough, of course).
1925
1926 Some convenience functions are provided for extracting the captured
1927 substrings as separate strings. These are described below.
1928
1929 Error return values from pcre_exec()
1930
1931 If pcre_exec() fails, it returns a negative number. The following are
1932 defined in the header file:
1933
1934 PCRE_ERROR_NOMATCH (-1)
1935
1936 The subject string did not match the pattern.
1937
1938 PCRE_ERROR_NULL (-2)
1939
1940 Either code or subject was passed as NULL, or ovector was NULL and
1941 ovecsize was not zero.
1942
1943 PCRE_ERROR_BADOPTION (-3)
1944
1945 An unrecognized bit was set in the options argument.
1946
1947 PCRE_ERROR_BADMAGIC (-4)
1948
1949 PCRE stores a 4-byte "magic number" at the start of the compiled code,
1950 to catch the case when it is passed a junk pointer and to detect when a
1951 pattern that was compiled in an environment of one endianness is run in
1952 an environment with the other endianness. This is the error that PCRE
1953 gives when the magic number is not present.
1954
1955 PCRE_ERROR_UNKNOWN_OPCODE (-5)
1956
1957 While running the pattern match, an unknown item was encountered in the
1958 compiled pattern. This error could be caused by a bug in PCRE or by
1959 overwriting of the compiled pattern.
1960
1961 PCRE_ERROR_NOMEMORY (-6)
1962
1963 If a pattern contains back references, but the ovector that is passed
1964 to pcre_exec() is not big enough to remember the referenced substrings,
1965 PCRE gets a block of memory at the start of matching to use for this
1966 purpose. If the call via pcre_malloc() fails, this error is given. The
1967 memory is automatically freed at the end of matching.
1968
1969 PCRE_ERROR_NOSUBSTRING (-7)
1970
1971 This error is used by the pcre_copy_substring(), pcre_get_substring(),
1972 and pcre_get_substring_list() functions (see below). It is never
1973 returned by pcre_exec().
1974
1975 PCRE_ERROR_MATCHLIMIT (-8)
1976
1977 The backtracking limit, as specified by the match_limit field in a
1978 pcre_extra structure (or defaulted) was reached. See the description
1979 above.
1980
1981 PCRE_ERROR_CALLOUT (-9)
1982
1983 This error is never generated by pcre_exec() itself. It is provided for
1984 use by callout functions that want to yield a distinctive error code.
1985 See the pcrecallout documentation for details.
1986
1987 PCRE_ERROR_BADUTF8 (-10)
1988
1989 A string that contains an invalid UTF-8 byte sequence was passed as a
1990 subject.
1991
1992 PCRE_ERROR_BADUTF8_OFFSET (-11)
1993
1994 The UTF-8 byte sequence that was passed as a subject was valid, but the
1995 value of startoffset did not point to the beginning of a UTF-8 charac-
1996 ter.
1997
1998 PCRE_ERROR_PARTIAL (-12)
1999
2000 The subject string did not match, but it did match partially. See the
2001 pcrepartial documentation for details of partial matching.
2002
2003 PCRE_ERROR_BADPARTIAL (-13)
2004
2005 The PCRE_PARTIAL option was used with a compiled pattern containing
2006 items that are not supported for partial matching. See the pcrepartial
2007 documentation for details of partial matching.
2008
2009 PCRE_ERROR_INTERNAL (-14)
2010
2011 An unexpected internal error has occurred. This error could be caused
2012 by a bug in PCRE or by overwriting of the compiled pattern.
2013
2014 PCRE_ERROR_BADCOUNT (-15)
2015
2016 This error is given if the value of the ovecsize argument is negative.
2017
2018 PCRE_ERROR_RECURSIONLIMIT (-21)
2019
2020 The internal recursion limit, as specified by the match_limit_recursion
2021 field in a pcre_extra structure (or defaulted) was reached. See the
2022 description above.
2023
2024 PCRE_ERROR_NULLWSLIMIT (-22)
2025
2026 When a group that can match an empty substring is repeated with an
2027 unbounded upper limit, the subject position at the start of the group
2028 must be remembered, so that a test for an empty string can be made when
2029 the end of the group is reached. Some workspace is required for this;
2030 if it runs out, this error is given.
2031
2032 PCRE_ERROR_BADNEWLINE (-23)
2033
2034 An invalid combination of PCRE_NEWLINE_xxx options was given.
2035
2036 Error numbers -16 to -20 are not used by pcre_exec().
2037
2038
2039 EXTRACTING CAPTURED SUBSTRINGS BY NUMBER
2040
2041 int pcre_copy_substring(const char *subject, int *ovector,
2042 int stringcount, int stringnumber, char *buffer,
2043 int buffersize);
2044
2045 int pcre_get_substring(const char *subject, int *ovector,
2046 int stringcount, int stringnumber,
2047 const char **stringptr);
2048
2049 int pcre_get_substring_list(const char *subject,
2050 int *ovector, int stringcount, const char ***listptr);
2051
2052 Captured substrings can be accessed directly by using the offsets
2053 returned by pcre_exec() in ovector. For convenience, the functions
2054 pcre_copy_substring(), pcre_get_substring(), and pcre_get_sub-
2055 string_list() are provided for extracting captured substrings as new,
2056 separate, zero-terminated strings. These functions identify substrings
2057 by number. The next section describes functions for extracting named
2058 substrings.
2059
2060 A substring that contains a binary zero is correctly extracted and has
2061 a further zero added on the end, but the result is not, of course, a C
2062 string. However, you can process such a string by referring to the
2063 length that is returned by pcre_copy_substring() and pcre_get_sub-
2064 string(). Unfortunately, the interface to pcre_get_substring_list() is
2065 not adequate for handling strings containing binary zeros, because the
2066 end of the final string is not independently indicated.
2067
2068 The first three arguments are the same for all three of these func-
2069 tions: subject is the subject string that has just been successfully
2070 matched, ovector is a pointer to the vector of integer offsets that was
2071 passed to pcre_exec(), and stringcount is the number of substrings that
2072 were captured by the match, including the substring that matched the
2073 entire regular expression. This is the value returned by pcre_exec() if
2074 it is greater than zero. If pcre_exec() returned zero, indicating that
2075 it ran out of space in ovector, the value passed as stringcount should
2076 be the number of elements in the vector divided by three.
2077
2078 The functions pcre_copy_substring() and pcre_get_substring() extract a
2079 single substring, whose number is given as stringnumber. A value of
2080 zero extracts the substring that matched the entire pattern, whereas
2081 higher values extract the captured substrings. For pcre_copy_sub-
2082 string(), the string is placed in buffer, whose length is given by
2083 buffersize, while for pcre_get_substring() a new block of memory is
2084 obtained via pcre_malloc, and its address is returned via stringptr.
2085 The yield of the function is the length of the string, not including
2086 the terminating zero, or one of these error codes:
2087
2088 PCRE_ERROR_NOMEMORY (-6)
2089
2090 The buffer was too small for pcre_copy_substring(), or the attempt to
2091 get memory failed for pcre_get_substring().
2092
2093 PCRE_ERROR_NOSUBSTRING (-7)
2094
2095 There is no substring whose number is stringnumber.
2096
2097 The pcre_get_substring_list() function extracts all available sub-
2098 strings and builds a list of pointers to them. All this is done in a
2099 single block of memory that is obtained via pcre_malloc. The address of
2100 the memory block is returned via listptr, which is also the start of
2101 the list of string pointers. The end of the list is marked by a NULL
2102 pointer. The yield of the function is zero if all went well, or the
2103 error code
2104
2105 PCRE_ERROR_NOMEMORY (-6)
2106
2107 if the attempt to get the memory block failed.
2108
2109 When any of these functions encounter a substring that is unset, which
2110 can happen when capturing subpattern number n+1 matches some part of
2111 the subject, but subpattern n has not been used at all, they return an
2112 empty string. This can be distinguished from a genuine zero-length sub-
2113 string by inspecting the appropriate offset in ovector, which is nega-
2114 tive for unset substrings.
2115
2116 The two convenience functions pcre_free_substring() and pcre_free_sub-
2117 string_list() can be used to free the memory returned by a previous
2118 call of pcre_get_substring() or pcre_get_substring_list(), respec-
2119 tively. They do nothing more than call the function pointed to by
2120 pcre_free, which of course could be called directly from a C program.
2121 However, PCRE is used in some situations where it is linked via a spe-
2122 cial interface to another programming language that cannot use
2123 pcre_free directly; it is for these cases that the functions are pro-
2124 vided.
2125
2126
2127 EXTRACTING CAPTURED SUBSTRINGS BY NAME
2128
2129 int pcre_get_stringnumber(const pcre *code,
2130 const char *name);
2131
2132 int pcre_copy_named_substring(const pcre *code,
2133 const char *subject, int *ovector,
2134 int stringcount, const char *stringname,
2135 char *buffer, int buffersize);
2136
2137 int pcre_get_named_substring(const pcre *code,
2138 const char *subject, int *ovector,
2139 int stringcount, const char *stringname,
2140 const char **stringptr);
2141
2142 To extract a substring by name, you first have to find associated num-
2143 ber. For example, for this pattern
2144
2145 (a+)b(?<xxx>\d+)...
2146
2147 the number of the subpattern called "xxx" is 2. If the name is known to
2148 be unique (PCRE_DUPNAMES was not set), you can find the number from the
2149 name by calling pcre_get_stringnumber(). The first argument is the com-
2150 piled pattern, and the second is the name. The yield of the function is
2151 the subpattern number, or PCRE_ERROR_NOSUBSTRING (-7) if there is no
2152 subpattern of that name.
2153
2154 Given the number, you can extract the substring directly, or use one of
2155 the functions described in the previous section. For convenience, there
2156 are also two functions that do the whole job.
2157
2158 Most of the arguments of pcre_copy_named_substring() and
2159 pcre_get_named_substring() are the same as those for the similarly
2160 named functions that extract by number. As these are described in the
2161 previous section, they are not re-described here. There are just two
2162 differences:
2163
2164 First, instead of a substring number, a substring name is given. Sec-
2165 ond, there is an extra argument, given at the start, which is a pointer
2166 to the compiled pattern. This is needed in order to gain access to the
2167 name-to-number translation table.
2168
2169 These functions call pcre_get_stringnumber(), and if it succeeds, they
2170 then call pcre_copy_substring() or pcre_get_substring(), as appropri-
2171 ate. NOTE: If PCRE_DUPNAMES is set and there are duplicate names, the
2172 behaviour may not be what you want (see the next section).
2173
2174
2175 DUPLICATE SUBPATTERN NAMES
2176
2177 int pcre_get_stringtable_entries(const pcre *code,
2178 const char *name, char **first, char **last);
2179
2180 When a pattern is compiled with the PCRE_DUPNAMES option, names for
2181 subpatterns are not required to be unique. Normally, patterns with
2182 duplicate names are such that in any one match, only one of the named
2183 subpatterns participates. An example is shown in the pcrepattern docu-
2184 mentation. When duplicates are present, pcre_copy_named_substring() and
2185 pcre_get_named_substring() return the first substring corresponding to
2186 the given name that is set. If none are set, an empty string is
2187 returned. The pcre_get_stringnumber() function returns one of the num-
2188 bers that are associated with the name, but it is not defined which it
2189 is.
2190
2191 If you want to get full details of all captured substrings for a given
2192 name, you must use the pcre_get_stringtable_entries() function. The
2193 first argument is the compiled pattern, and the second is the name. The
2194 third and fourth are pointers to variables which are updated by the
2195 function. After it has run, they point to the first and last entries in
2196 the name-to-number table for the given name. The function itself
2197 returns the length of each entry, or PCRE_ERROR_NOSUBSTRING (-7) if
2198 there are none. The format of the table is described above in the sec-
2199 tion entitled Information about a pattern. Given all the relevant
2200 entries for the name, you can extract each of their numbers, and hence
2201 the captured data, if any.
2202
2203
2204 FINDING ALL POSSIBLE MATCHES
2205
2206 The traditional matching function uses a similar algorithm to Perl,
2207 which stops when it finds the first match, starting at a given point in
2208 the subject. If you want to find all possible matches, or the longest
2209 possible match, consider using the alternative matching function (see
2210 below) instead. If you cannot use the alternative function, but still
2211 need to find all possible matches, you can kludge it up by making use
2212 of the callout facility, which is described in the pcrecallout documen-
2213 tation.
2214
2215 What you have to do is to insert a callout right at the end of the pat-
2216 tern. When your callout function is called, extract and save the cur-
2217 rent matched substring. Then return 1, which forces pcre_exec() to
2218 backtrack and try other alternatives. Ultimately, when it runs out of
2219 matches, pcre_exec() will yield PCRE_ERROR_NOMATCH.
2220
2221
2222 MATCHING A PATTERN: THE ALTERNATIVE FUNCTION
2223
2224 int pcre_dfa_exec(const pcre *code, const pcre_extra *extra,
2225 const char *subject, int length, int startoffset,
2226 int options, int *ovector, int ovecsize,
2227 int *workspace, int wscount);
2228
2229 The function pcre_dfa_exec() is called to match a subject string
2230 against a compiled pattern, using a matching algorithm that scans the
2231 subject string just once, and does not backtrack. This has different
2232 characteristics to the normal algorithm, and is not compatible with
2233 Perl. Some of the features of PCRE patterns are not supported. Never-
2234 theless, there are times when this kind of matching can be useful. For
2235 a discussion of the two matching algorithms, see the pcrematching docu-
2236 mentation.
2237
2238 The arguments for the pcre_dfa_exec() function are the same as for
2239 pcre_exec(), plus two extras. The ovector argument is used in a differ-
2240 ent way, and this is described below. The other common arguments are
2241 used in the same way as for pcre_exec(), so their description is not
2242 repeated here.
2243
2244 The two additional arguments provide workspace for the function. The
2245 workspace vector should contain at least 20 elements. It is used for
2246 keeping track of multiple paths through the pattern tree. More
2247 workspace will be needed for patterns and subjects where there are a
2248 lot of potential matches.
2249
2250 Here is an example of a simple call to pcre_dfa_exec():
2251
2252 int rc;
2253 int ovector[10];
2254 int wspace[20];
2255 rc = pcre_dfa_exec(
2256 re, /* result of pcre_compile() */
2257 NULL, /* we didn't study the pattern */
2258 "some string", /* the subject string */
2259 11, /* the length of the subject string */
2260 0, /* start at offset 0 in the subject */
2261 0, /* default options */
2262 ovector, /* vector of integers for substring information */
2263 10, /* number of elements (NOT size in bytes) */
2264 wspace, /* working space vector */
2265 20); /* number of elements (NOT size in bytes) */
2266
2267 Option bits for pcre_dfa_exec()
2268
2269 The unused bits of the options argument for pcre_dfa_exec() must be
2270 zero. The only bits that may be set are PCRE_ANCHORED, PCRE_NEW-
2271 LINE_xxx, PCRE_NOTBOL, PCRE_NOTEOL, PCRE_NOTEMPTY, PCRE_NO_UTF8_CHECK,
2272 PCRE_PARTIAL, PCRE_DFA_SHORTEST, and PCRE_DFA_RESTART. All but the last
2273 three of these are the same as for pcre_exec(), so their description is
2274 not repeated here.
2275
2276 PCRE_PARTIAL
2277
2278 This has the same general effect as it does for pcre_exec(), but the
2279 details are slightly different. When PCRE_PARTIAL is set for
2280 pcre_dfa_exec(), the return code PCRE_ERROR_NOMATCH is converted into
2281 PCRE_ERROR_PARTIAL if the end of the subject is reached, there have
2282 been no complete matches, but there is still at least one matching pos-
2283 sibility. The portion of the string that provided the partial match is
2284 set as the first matching string.
2285
2286 PCRE_DFA_SHORTEST
2287
2288 Setting the PCRE_DFA_SHORTEST option causes the matching algorithm to
2289 stop as soon as it has found one match. Because of the way the alterna-
2290 tive algorithm works, this is necessarily the shortest possible match
2291 at the first possible matching point in the subject string.
2292
2293 PCRE_DFA_RESTART
2294
2295 When pcre_dfa_exec() is called with the PCRE_PARTIAL option, and
2296 returns a partial match, it is possible to call it again, with addi-
2297 tional subject characters, and have it continue with the same match.
2298 The PCRE_DFA_RESTART option requests this action; when it is set, the
2299 workspace and wscount options must reference the same vector as before
2300 because data about the match so far is left in them after a partial
2301 match. There is more discussion of this facility in the pcrepartial
2302 documentation.
2303
2304 Successful returns from pcre_dfa_exec()
2305
2306 When pcre_dfa_exec() succeeds, it may have matched more than one sub-
2307 string in the subject. Note, however, that all the matches from one run
2308 of the function start at the same point in the subject. The shorter
2309 matches are all initial substrings of the longer matches. For example,
2310 if the pattern
2311
2312 <.*>
2313
2314 is matched against the string
2315
2316 This is <something> <something else> <something further> no more
2317
2318 the three matched strings are
2319
2320 <something>
2321 <something> <something else>
2322 <something> <something else> <something further>
2323
2324 On success, the yield of the function is a number greater than zero,
2325 which is the number of matched substrings. The substrings themselves
2326 are returned in ovector. Each string uses two elements; the first is
2327 the offset to the start, and the second is the offset to the end. In
2328 fact, all the strings have the same start offset. (Space could have
2329 been saved by giving this only once, but it was decided to retain some
2330 compatibility with the way pcre_exec() returns data, even though the
2331 meaning of the strings is different.)
2332
2333 The strings are returned in reverse order of length; that is, the long-
2334 est matching string is given first. If there were too many matches to
2335 fit into ovector, the yield of the function is zero, and the vector is
2336 filled with the longest matches.
2337
2338 Error returns from pcre_dfa_exec()
2339
2340 The pcre_dfa_exec() function returns a negative number when it fails.
2341 Many of the errors are the same as for pcre_exec(), and these are
2342 described above. There are in addition the following errors that are
2343 specific to pcre_dfa_exec():
2344
2345 PCRE_ERROR_DFA_UITEM (-16)
2346
2347 This return is given if pcre_dfa_exec() encounters an item in the pat-
2348 tern that it does not support, for instance, the use of \C or a back
2349 reference.
2350
2351 PCRE_ERROR_DFA_UCOND (-17)
2352
2353 This return is given if pcre_dfa_exec() encounters a condition item
2354 that uses a back reference for the condition, or a test for recursion
2355 in a specific group. These are not supported.
2356
2357 PCRE_ERROR_DFA_UMLIMIT (-18)
2358
2359 This return is given if pcre_dfa_exec() is called with an extra block
2360 that contains a setting of the match_limit field. This is not supported
2361 (it is meaningless).
2362
2363 PCRE_ERROR_DFA_WSSIZE (-19)
2364
2365 This return is given if pcre_dfa_exec() runs out of space in the
2366 workspace vector.
2367
2368 PCRE_ERROR_DFA_RECURSE (-20)
2369
2370 When a recursive subpattern is processed, the matching function calls
2371 itself recursively, using private vectors for ovector and workspace.
2372 This error is given if the output vector is not large enough. This
2373 should be extremely rare, as a vector of size 1000 is used.
2374
2375
2376 SEE ALSO
2377
2378 pcrebuild(3), pcrecallout(3), pcrecpp(3)(3), pcrematching(3), pcrepar-
2379 tial(3), pcreposix(3), pcreprecompile(3), pcresample(3), pcrestack(3).
2380
2381
2382 AUTHOR
2383
2384 Philip Hazel
2385 University Computing Service
2386 Cambridge CB2 3QH, England.
2387
2388
2389 REVISION
2390
2391 Last updated: 24 April 2007
2392 Copyright (c) 1997-2007 University of Cambridge.
2393 ------------------------------------------------------------------------------
2394
2395
2396 PCRECALLOUT(3) PCRECALLOUT(3)
2397
2398
2399 NAME
2400 PCRE - Perl-compatible regular expressions
2401
2402
2403 PCRE CALLOUTS
2404
2405 int (*pcre_callout)(pcre_callout_block *);
2406
2407 PCRE provides a feature called "callout", which is a means of temporar-
2408 ily passing control to the caller of PCRE in the middle of pattern
2409 matching. The caller of PCRE provides an external function by putting
2410 its entry point in the global variable pcre_callout. By default, this
2411 variable contains NULL, which disables all calling out.
2412
2413 Within a regular expression, (?C) indicates the points at which the
2414 external function is to be called. Different callout points can be
2415 identified by putting a number less than 256 after the letter C. The
2416 default value is zero. For example, this pattern has two callout
2417 points:
2418
2419 (?C1)abc(?C2)def
2420
2421 If the PCRE_AUTO_CALLOUT option bit is set when pcre_compile() is
2422 called, PCRE automatically inserts callouts, all with number 255,
2423 before each item in the pattern. For example, if PCRE_AUTO_CALLOUT is
2424 used with the pattern
2425
2426 A(\d{2}|--)
2427
2428 it is processed as if it were
2429
2430 (?C255)A(?C255)((?C255)\d{2}(?C255)|(?C255)-(?C255)-(?C255))(?C255)
2431
2432 Notice that there is a callout before and after each parenthesis and
2433 alternation bar. Automatic callouts can be used for tracking the
2434 progress of pattern matching. The pcretest command has an option that
2435 sets automatic callouts; when it is used, the output indicates how the
2436 pattern is matched. This is useful information when you are trying to
2437 optimize the performance of a particular pattern.
2438
2439
2440 MISSING CALLOUTS
2441
2442 You should be aware that, because of optimizations in the way PCRE
2443 matches patterns, callouts sometimes do not happen. For example, if the
2444 pattern is
2445
2446 ab(?C4)cd
2447
2448 PCRE knows that any matching string must contain the letter "d". If the
2449 subject string is "abyz", the lack of "d" means that matching doesn't
2450 ever start, and the callout is never reached. However, with "abyd",
2451 though the result is still no match, the callout is obeyed.
2452
2453
2454 THE CALLOUT INTERFACE
2455
2456 During matching, when PCRE reaches a callout point, the external func-
2457 tion defined by pcre_callout is called (if it is set). This applies to
2458 both the pcre_exec() and the pcre_dfa_exec() matching functions. The
2459 only argument to the callout function is a pointer to a pcre_callout
2460 block. This structure contains the following fields:
2461
2462 int version;
2463 int callout_number;
2464 int *offset_vector;
2465 const char *subject;
2466 int subject_length;
2467 int start_match;
2468 int current_position;
2469 int capture_top;
2470 int capture_last;
2471 void *callout_data;
2472 int pattern_position;
2473 int next_item_length;
2474
2475 The version field is an integer containing the version number of the
2476 block format. The initial version was 0; the current version is 1. The
2477 version number will change again in future if additional fields are
2478 added, but the intention is never to remove any of the existing fields.
2479
2480 The callout_number field contains the number of the callout, as com-
2481 piled into the pattern (that is, the number after ?C for manual call-
2482 outs, and 255 for automatically generated callouts).
2483
2484 The offset_vector field is a pointer to the vector of offsets that was
2485 passed by the caller to pcre_exec() or pcre_dfa_exec(). When
2486 pcre_exec() is used, the contents can be inspected in order to extract
2487 substrings that have been matched so far, in the same way as for
2488 extracting substrings after a match has completed. For pcre_dfa_exec()
2489 this field is not useful.
2490
2491 The subject and subject_length fields contain copies of the values that
2492 were passed to pcre_exec().
2493
2494 The start_match field contains the offset within the subject at which
2495 the current match attempt started. If the pattern is not anchored, the
2496 callout function may be called several times from the same point in the
2497 pattern for different starting points in the subject.
2498
2499 The current_position field contains the offset within the subject of
2500 the current match pointer.
2501
2502 When the pcre_exec() function is used, the capture_top field contains
2503 one more than the number of the highest numbered captured substring so
2504 far. If no substrings have been captured, the value of capture_top is
2505 one. This is always the case when pcre_dfa_exec() is used, because it
2506 does not support captured substrings.
2507
2508 The capture_last field contains the number of the most recently cap-
2509 tured substring. If no substrings have been captured, its value is -1.
2510 This is always the case when pcre_dfa_exec() is used.
2511
2512 The callout_data field contains a value that is passed to pcre_exec()
2513 or pcre_dfa_exec() specifically so that it can be passed back in call-
2514 outs. It is passed in the pcre_callout field of the pcre_extra data
2515 structure. If no such data was passed, the value of callout_data in a
2516 pcre_callout block is NULL. There is a description of the pcre_extra
2517 structure in the pcreapi documentation.
2518
2519 The pattern_position field is present from version 1 of the pcre_call-
2520 out structure. It contains the offset to the next item to be matched in
2521 the pattern string.
2522
2523 The next_item_length field is present from version 1 of the pcre_call-
2524 out structure. It contains the length of the next item to be matched in
2525 the pattern string. When the callout immediately precedes an alterna-
2526 tion bar, a closing parenthesis, or the end of the pattern, the length
2527 is zero. When the callout precedes an opening parenthesis, the length
2528 is that of the entire subpattern.
2529
2530 The pattern_position and next_item_length fields are intended to help
2531 in distinguishing between different automatic callouts, which all have
2532 the same callout number. However, they are set for all callouts.
2533
2534
2535 RETURN VALUES
2536
2537 The external callout function returns an integer to PCRE. If the value
2538 is zero, matching proceeds as normal. If the value is greater than
2539 zero, matching fails at the current point, but the testing of other
2540 matching possibilities goes ahead, just as if a lookahead assertion had
2541 failed. If the value is less than zero, the match is abandoned, and
2542 pcre_exec() (or pcre_dfa_exec()) returns the negative value.
2543
2544 Negative values should normally be chosen from the set of
2545 PCRE_ERROR_xxx values. In particular, PCRE_ERROR_NOMATCH forces a stan-
2546 dard "no match" failure. The error number PCRE_ERROR_CALLOUT is
2547 reserved for use by callout functions; it will never be used by PCRE
2548 itself.
2549
2550
2551 AUTHOR
2552
2553 Philip Hazel
2554 University Computing Service
2555 Cambridge CB2 3QH, England.
2556
2557
2558 REVISION
2559
2560 Last updated: 06 March 2007
2561 Copyright (c) 1997-2007 University of Cambridge.
2562 ------------------------------------------------------------------------------
2563
2564
2565 PCRECOMPAT(3) PCRECOMPAT(3)
2566
2567
2568 NAME
2569 PCRE - Perl-compatible regular expressions
2570
2571
2572 DIFFERENCES BETWEEN PCRE AND PERL
2573
2574 This document describes the differences in the ways that PCRE and Perl
2575 handle regular expressions. The differences described here are mainly
2576 with respect to Perl 5.8, though PCRE version 7.0 contains some fea-
2577 tures that are expected to be in the forthcoming Perl 5.10.
2578
2579 1. PCRE has only a subset of Perl's UTF-8 and Unicode support. Details
2580 of what it does have are given in the section on UTF-8 support in the
2581 main pcre page.
2582
2583 2. PCRE does not allow repeat quantifiers on lookahead assertions. Perl
2584 permits them, but they do not mean what you might think. For example,
2585 (?!a){3} does not assert that the next three characters are not "a". It
2586 just asserts that the next character is not "a" three times.
2587
2588 3. Capturing subpatterns that occur inside negative lookahead asser-
2589 tions are counted, but their entries in the offsets vector are never
2590 set. Perl sets its numerical variables from any such patterns that are
2591 matched before the assertion fails to match something (thereby succeed-
2592 ing), but only if the negative lookahead assertion contains just one
2593 branch.
2594
2595 4. Though binary zero characters are supported in the subject string,
2596 they are not allowed in a pattern string because it is passed as a nor-
2597 mal C string, terminated by zero. The escape sequence \0 can be used in
2598 the pattern to represent a binary zero.
2599
2600 5. The following Perl escape sequences are not supported: \l, \u, \L,
2601 \U, and \N. In fact these are implemented by Perl's general string-han-
2602 dling and are not part of its pattern matching engine. If any of these
2603 are encountered by PCRE, an error is generated.
2604
2605 6. The Perl escape sequences \p, \P, and \X are supported only if PCRE
2606 is built with Unicode character property support. The properties that
2607 can be tested with \p and \P are limited to the general category prop-
2608 erties such as Lu and Nd, script names such as Greek or Han, and the
2609 derived properties Any and L&.
2610
2611 7. PCRE does support the \Q...\E escape for quoting substrings. Charac-
2612 ters in between are treated as literals. This is slightly different
2613 from Perl in that $ and @ are also handled as literals inside the
2614 quotes. In Perl, they cause variable interpolation (but of course PCRE
2615 does not have variables). Note the following examples:
2616
2617 Pattern PCRE matches Perl matches
2618
2619 \Qabc$xyz\E abc$xyz abc followed by the
2620 contents of $xyz
2621 \Qabc\$xyz\E abc\$xyz abc\$xyz
2622 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
2623
2624 The \Q...\E sequence is recognized both inside and outside character
2625 classes.
2626
2627 8. Fairly obviously, PCRE does not support the (?{code}) and (??{code})
2628 constructions. However, there is support for recursive patterns. This
2629 is not available in Perl 5.8, but will be in Perl 5.10. Also, the PCRE
2630 "callout" feature allows an external function to be called during pat-
2631 tern matching. See the pcrecallout documentation for details.
2632
2633 9. Subpatterns that are called recursively or as "subroutines" are
2634 always treated as atomic groups in PCRE. This is like Python, but
2635 unlike Perl.
2636
2637 10. There are some differences that are concerned with the settings of
2638 captured strings when part of a pattern is repeated. For example,
2639 matching "aba" against the pattern /^(a(b)?)+$/ in Perl leaves $2
2640 unset, but in PCRE it is set to "b".
2641
2642 11. PCRE provides some extensions to the Perl regular expression facil-
2643 ities. Perl 5.10 will include new features that are not in earlier
2644 versions, some of which (such as named parentheses) have been in PCRE
2645 for some time. This list is with respect to Perl 5.10:
2646
2647 (a) Although lookbehind assertions must match fixed length strings,
2648 each alternative branch of a lookbehind assertion can match a different
2649 length of string. Perl requires them all to have the same length.
2650
2651 (b) If PCRE_DOLLAR_ENDONLY is set and PCRE_MULTILINE is not set, the $
2652 meta-character matches only at the very end of the string.
2653
2654 (c) If PCRE_EXTRA is set, a backslash followed by a letter with no spe-
2655 cial meaning is faulted. Otherwise, like Perl, the backslash is
2656 ignored. (Perl can be made to issue a warning.)
2657
2658 (d) If PCRE_UNGREEDY is set, the greediness of the repetition quanti-
2659 fiers is inverted, that is, by default they are not greedy, but if fol-
2660 lowed by a question mark they are.
2661
2662 (e) PCRE_ANCHORED can be used at matching time to force a pattern to be
2663 tried only at the first matching position in the subject string.
2664
2665 (f) The PCRE_NOTBOL, PCRE_NOTEOL, PCRE_NOTEMPTY, and PCRE_NO_AUTO_CAP-
2666 TURE options for pcre_exec() have no Perl equivalents.
2667
2668 (g) The callout facility is PCRE-specific.
2669
2670 (h) The partial matching facility is PCRE-specific.
2671
2672 (i) Patterns compiled by PCRE can be saved and re-used at a later time,
2673 even on different hosts that have the other endianness.
2674
2675 (j) The alternative matching function (pcre_dfa_exec()) matches in a
2676 different way and is not Perl-compatible.
2677
2678
2679 AUTHOR
2680
2681 Philip Hazel
2682 University Computing Service
2683 Cambridge CB2 3QH, England.
2684
2685
2686 REVISION
2687
2688 Last updated: 06 March 2007
2689 Copyright (c) 1997-2007 University of Cambridge.
2690 ------------------------------------------------------------------------------
2691
2692
2693 PCREPATTERN(3) PCREPATTERN(3)
2694
2695
2696 NAME
2697 PCRE - Perl-compatible regular expressions
2698
2699
2700 PCRE REGULAR EXPRESSION DETAILS
2701
2702 The syntax and semantics of the regular expressions supported by PCRE
2703 are described below. Regular expressions are also described in the Perl
2704 documentation and in a number of books, some of which have copious
2705 examples. Jeffrey Friedl's "Mastering Regular Expressions", published
2706 by O'Reilly, covers regular expressions in great detail. This descrip-
2707 tion of PCRE's regular expressions is intended as reference material.
2708
2709 The original operation of PCRE was on strings of one-byte characters.
2710 However, there is now also support for UTF-8 character strings. To use
2711 this, you must build PCRE to include UTF-8 support, and then call
2712 pcre_compile() with the PCRE_UTF8 option. How this affects pattern
2713 matching is mentioned in several places below. There is also a summary
2714 of UTF-8 features in the section on UTF-8 support in the main pcre
2715 page.
2716
2717 The remainder of this document discusses the patterns that are sup-
2718 ported by PCRE when its main matching function, pcre_exec(), is used.
2719 From release 6.0, PCRE offers a second matching function,
2720 pcre_dfa_exec(), which matches using a different algorithm that is not
2721 Perl-compatible. The advantages and disadvantages of the alternative
2722 function, and how it differs from the normal function, are discussed in
2723 the pcrematching page.
2724
2725
2726 CHARACTERS AND METACHARACTERS
2727
2728 A regular expression is a pattern that is matched against a subject
2729 string from left to right. Most characters stand for themselves in a
2730 pattern, and match the corresponding characters in the subject. As a
2731 trivial example, the pattern
2732
2733 The quick brown fox
2734
2735 matches a portion of a subject string that is identical to itself. When
2736 caseless matching is specified (the PCRE_CASELESS option), letters are
2737 matched independently of case. In UTF-8 mode, PCRE always understands
2738 the concept of case for characters whose values are less than 128, so
2739 caseless matching is always possible. For characters with higher val-
2740 ues, the concept of case is supported if PCRE is compiled with Unicode
2741 property support, but not otherwise. If you want to use caseless
2742 matching for characters 128 and above, you must ensure that PCRE is
2743 compiled with Unicode property support as well as with UTF-8 support.
2744
2745 The power of regular expressions comes from the ability to include
2746 alternatives and repetitions in the pattern. These are encoded in the
2747 pattern by the use of metacharacters, which do not stand for themselves
2748 but instead are interpreted in some special way.
2749
2750 There are two different sets of metacharacters: those that are recog-
2751 nized anywhere in the pattern except within square brackets, and those
2752 that are recognized within square brackets. Outside square brackets,
2753 the metacharacters are as follows:
2754
2755 \ general escape character with several uses
2756 ^ assert start of string (or line, in multiline mode)
2757 $ assert end of string (or line, in multiline mode)
2758 . match any character except newline (by default)
2759 [ start character class definition
2760 | start of alternative branch
2761 ( start subpattern
2762 ) end subpattern
2763 ? extends the meaning of (
2764 also 0 or 1 quantifier
2765 also quantifier minimizer
2766 * 0 or more quantifier
2767 + 1 or more quantifier
2768 also "possessive quantifier"
2769 { start min/max quantifier
2770
2771 Part of a pattern that is in square brackets is called a "character
2772 class". In a character class the only metacharacters are:
2773
2774 \ general escape character
2775 ^ negate the class, but only if the first character
2776 - indicates character range
2777 [ POSIX character class (only if followed by POSIX
2778 syntax)
2779 ] terminates the character class
2780
2781 The following sections describe the use of each of the metacharacters.
2782
2783
2784 BACKSLASH
2785
2786 The backslash character has several uses. Firstly, if it is followed by
2787 a non-alphanumeric character, it takes away any special meaning that
2788 character may have. This use of backslash as an escape character
2789 applies both inside and outside character classes.
2790
2791 For example, if you want to match a * character, you write \* in the
2792 pattern. This escaping action applies whether or not the following
2793 character would otherwise be interpreted as a metacharacter, so it is
2794 always safe to precede a non-alphanumeric with backslash to specify
2795 that it stands for itself. In particular, if you want to match a back-
2796 slash, you write \\.
2797
2798 If a pattern is compiled with the PCRE_EXTENDED option, whitespace in
2799 the pattern (other than in a character class) and characters between a
2800 # outside a character class and the next newline are ignored. An escap-
2801 ing backslash can be used to include a whitespace or # character as
2802 part of the pattern.
2803
2804 If you want to remove the special meaning from a sequence of charac-
2805 ters, you can do so by putting them between \Q and \E. This is differ-
2806 ent from Perl in that $ and @ are handled as literals in \Q...\E
2807 sequences in PCRE, whereas in Perl, $ and @ cause variable interpola-
2808 tion. Note the following examples:
2809
2810 Pattern PCRE matches Perl matches
2811
2812 \Qabc$xyz\E abc$xyz abc followed by the
2813 contents of $xyz
2814 \Qabc\$xyz\E abc\$xyz abc\$xyz
2815 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
2816
2817 The \Q...\E sequence is recognized both inside and outside character
2818 classes.
2819
2820 Non-printing characters
2821
2822 A second use of backslash provides a way of encoding non-printing char-
2823 acters in patterns in a visible manner. There is no restriction on the
2824 appearance of non-printing characters, apart from the binary zero that
2825 terminates a pattern, but when a pattern is being prepared by text
2826 editing, it is usually easier to use one of the following escape
2827 sequences than the binary character it represents:
2828
2829 \a alarm, that is, the BEL character (hex 07)
2830 \cx "control-x", where x is any character
2831 \e escape (hex 1B)
2832 \f formfeed (hex 0C)
2833 \n newline (hex 0A)
2834 \r carriage return (hex 0D)
2835 \t tab (hex 09)
2836 \ddd character with octal code ddd, or backreference
2837 \xhh character with hex code hh
2838 \x{hhh..} character with hex code hhh..
2839
2840 The precise effect of \cx is as follows: if x is a lower case letter,
2841 it is converted to upper case. Then bit 6 of the character (hex 40) is
2842 inverted. Thus \cz becomes hex 1A, but \c{ becomes hex 3B, while \c;
2843 becomes hex 7B.
2844
2845 After \x, from zero to two hexadecimal digits are read (letters can be
2846 in upper or lower case). Any number of hexadecimal digits may appear
2847 between \x{ and }, but the value of the character code must be less
2848 than 256 in non-UTF-8 mode, and less than 2**31 in UTF-8 mode (that is,
2849 the maximum hexadecimal value is 7FFFFFFF). If characters other than
2850 hexadecimal digits appear between \x{ and }, or if there is no termi-
2851 nating }, this form of escape is not recognized. Instead, the initial
2852 \x will be interpreted as a basic hexadecimal escape, with no following
2853 digits, giving a character whose value is zero.
2854
2855 Characters whose value is less than 256 can be defined by either of the
2856 two syntaxes for \x. There is no difference in the way they are han-
2857 dled. For example, \xdc is exactly the same as \x{dc}.
2858
2859 After \0 up to two further octal digits are read. If there are fewer
2860 than two digits, just those that are present are used. Thus the
2861 sequence \0\x\07 specifies two binary zeros followed by a BEL character
2862 (code value 7). Make sure you supply two digits after the initial zero
2863 if the pattern character that follows is itself an octal digit.
2864
2865 The handling of a backslash followed by a digit other than 0 is compli-
2866 cated. Outside a character class, PCRE reads it and any following dig-
2867 its as a decimal number. If the number is less than 10, or if there
2868 have been at least that many previous capturing left parentheses in the
2869 expression, the entire sequence is taken as a back reference. A
2870 description of how this works is given later, following the discussion
2871 of parenthesized subpatterns.
2872
2873 Inside a character class, or if the decimal number is greater than 9
2874 and there have not been that many capturing subpatterns, PCRE re-reads
2875 up to three octal digits following the backslash, and uses them to gen-
2876 erate a data character. Any subsequent digits stand for themselves. In
2877 non-UTF-8 mode, the value of a character specified in octal must be
2878 less than \400. In UTF-8 mode, values up to \777 are permitted. For
2879 example:
2880
2881 \040 is another way of writing a space
2882 \40 is the same, provided there are fewer than 40
2883 previous capturing subpatterns
2884 \7 is always a back reference
2885 \11 might be a back reference, or another way of
2886 writing a tab
2887 \011 is always a tab
2888 \0113 is a tab followed by the character "3"
2889 \113 might be a back reference, otherwise the
2890 character with octal code 113
2891 \377 might be a back reference, otherwise
2892 the byte consisting entirely of 1 bits
2893 \81 is either a back reference, or a binary zero
2894 followed by the two characters "8" and "1"
2895
2896 Note that octal values of 100 or greater must not be introduced by a
2897 leading zero, because no more than three octal digits are ever read.
2898
2899 All the sequences that define a single character value can be used both
2900 inside and outside character classes. In addition, inside a character
2901 class, the sequence \b is interpreted as the backspace character (hex
2902 08), and the sequences \R and \X are interpreted as the characters "R"
2903 and "X", respectively. Outside a character class, these sequences have
2904 different meanings (see below).
2905
2906 Absolute and relative back references
2907
2908 The sequence \g followed by a positive or negative number, optionally
2909 enclosed in braces, is an absolute or relative back reference. Back
2910 references are discussed later, following the discussion of parenthe-
2911 sized subpatterns.
2912
2913 Generic character types
2914
2915 Another use of backslash is for specifying generic character types. The
2916 following are always recognized:
2917
2918 \d any decimal digit
2919 \D any character that is not a decimal digit
2920 \s any whitespace character
2921 \S any character that is not a whitespace character
2922 \w any "word" character
2923 \W any "non-word" character
2924
2925 Each pair of escape sequences partitions the complete set of characters
2926 into two disjoint sets. Any given character matches one, and only one,
2927 of each pair.
2928
2929 These character type sequences can appear both inside and outside char-
2930 acter classes. They each match one character of the appropriate type.
2931 If the current matching point is at the end of the subject string, all
2932 of them fail, since there is no character to match.
2933
2934 For compatibility with Perl, \s does not match the VT character (code
2935 11). This makes it different from the the POSIX "space" class. The \s
2936 characters are HT (9), LF (10), FF (12), CR (13), and space (32). (If
2937 "use locale;" is included in a Perl script, \s may match the VT charac-
2938 ter. In PCRE, it never does.)
2939
2940 A "word" character is an underscore or any character less than 256 that
2941 is a letter or digit. The definition of letters and digits is con-
2942 trolled by PCRE's low-valued character tables, and may vary if locale-
2943 specific matching is taking place (see "Locale support" in the pcreapi
2944 page). For example, in a French locale such as "fr_FR" in Unix-like
2945 systems, or "french" in Windows, some character codes greater than 128
2946 are used for accented letters, and these are matched by \w.
2947
2948 In UTF-8 mode, characters with values greater than 128 never match \d,
2949 \s, or \w, and always match \D, \S, and \W. This is true even when Uni-
2950 code character property support is available. The use of locales with
2951 Unicode is discouraged.
2952
2953 Newline sequences
2954
2955 Outside a character class, the escape sequence \R matches any Unicode
2956 newline sequence. This is an extension to Perl. In non-UTF-8 mode \R is
2957 equivalent to the following:
2958
2959 (?>\r\n|\n|\x0b|\f|\r|\x85)
2960
2961 This is an example of an "atomic group", details of which are given
2962 below. This particular group matches either the two-character sequence
2963 CR followed by LF, or one of the single characters LF (linefeed,
2964 U+000A), VT (vertical tab, U+000B), FF (formfeed, U+000C), CR (carriage
2965 return, U+000D), or NEL (next line, U+0085). The two-character sequence
2966 is treated as a single unit that cannot be split.
2967
2968 In UTF-8 mode, two additional characters whose codepoints are greater
2969 than 255 are added: LS (line separator, U+2028) and PS (paragraph sepa-
2970 rator, U+2029). Unicode character property support is not needed for
2971 these characters to be recognized.
2972
2973 Inside a character class, \R matches the letter "R".
2974
2975 Unicode character properties
2976
2977 When PCRE is built with Unicode character property support, three addi-
2978 tional escape sequences to match character properties are available
2979 when UTF-8 mode is selected. They are:
2980
2981 \p{xx} a character with the xx property
2982 \P{xx} a character without the xx property
2983 \X an extended Unicode sequence
2984
2985 The property names represented by xx above are limited to the Unicode
2986 script names, the general category properties, and "Any", which matches
2987 any character (including newline). Other properties such as "InMusical-
2988 Symbols" are not currently supported by PCRE. Note that \P{Any} does
2989 not match any characters, so always causes a match failure.
2990
2991 Sets of Unicode characters are defined as belonging to certain scripts.
2992 A character from one of these sets can be matched using a script name.
2993 For example:
2994
2995 \p{Greek}
2996 \P{Han}
2997
2998 Those that are not part of an identified script are lumped together as
2999 "Common". The current list of scripts is:
3000
3001 Arabic, Armenian, Balinese, Bengali, Bopomofo, Braille, Buginese,
3002 Buhid, Canadian_Aboriginal, Cherokee, Common, Coptic, Cuneiform,
3003 Cypriot, Cyrillic, Deseret, Devanagari, Ethiopic, Georgian, Glagolitic,
3004 Gothic, Greek, Gujarati, Gurmukhi, Han, Hangul, Hanunoo, Hebrew, Hira-
3005 gana, Inherited, Kannada, Katakana, Kharoshthi, Khmer, Lao, Latin,
3006 Limbu, Linear_B, Malayalam, Mongolian, Myanmar, New_Tai_Lue, Nko,
3007 Ogham, Old_Italic, Old_Persian, Oriya, Osmanya, Phags_Pa, Phoenician,
3008 Runic, Shavian, Sinhala, Syloti_Nagri, Syriac, Tagalog, Tagbanwa,
3009 Tai_Le, Tamil, Telugu, Thaana, Thai, Tibetan, Tifinagh, Ugaritic, Yi.
3010
3011 Each character has exactly one general category property, specified by
3012 a two-letter abbreviation. For compatibility with Perl, negation can be
3013 specified by including a circumflex between the opening brace and the
3014 property name. For example, \p{^Lu} is the same as \P{Lu}.
3015
3016 If only one letter is specified with \p or \P, it includes all the gen-
3017 eral category properties that start with that letter. In this case, in
3018 the absence of negation, the curly brackets in the escape sequence are
3019 optional; these two examples have the same effect:
3020
3021 \p{L}
3022 \pL
3023
3024 The following general category property codes are supported:
3025
3026 C Other
3027 Cc Control
3028 Cf Format
3029 Cn Unassigned
3030 Co Private use
3031 Cs Surrogate
3032
3033 L Letter
3034 Ll Lower case letter
3035 Lm Modifier letter
3036 Lo Other letter
3037 Lt Title case letter
3038 Lu Upper case letter
3039
3040 M Mark
3041 Mc Spacing mark
3042 Me Enclosing mark
3043 Mn Non-spacing mark
3044
3045 N Number
3046 Nd Decimal number
3047 Nl Letter number
3048 No Other number
3049
3050 P Punctuation
3051 Pc Connector punctuation
3052 Pd Dash punctuation
3053 Pe Close punctuation
3054 Pf Final punctuation
3055 Pi Initial punctuation
3056 Po Other punctuation
3057 Ps Open punctuation
3058
3059 S Symbol
3060 Sc Currency symbol
3061 Sk Modifier symbol
3062 Sm Mathematical symbol
3063 So Other symbol
3064
3065 Z Separator
3066 Zl Line separator
3067 Zp Paragraph separator
3068 Zs Space separator
3069
3070 The special property L& is also supported: it matches a character that
3071 has the Lu, Ll, or Lt property, in other words, a letter that is not
3072 classified as a modifier or "other".
3073
3074 The long synonyms for these properties that Perl supports (such as
3075 \p{Letter}) are not supported by PCRE, nor is it permitted to prefix
3076 any of these properties with "Is".
3077
3078 No character that is in the Unicode table has the Cn (unassigned) prop-
3079 erty. Instead, this property is assumed for any code point that is not
3080 in the Unicode table.
3081
3082 Specifying caseless matching does not affect these escape sequences.
3083 For example, \p{Lu} always matches only upper case letters.
3084
3085 The \X escape matches any number of Unicode characters that form an
3086 extended Unicode sequence. \X is equivalent to
3087
3088 (?>\PM\pM*)
3089
3090 That is, it matches a character without the "mark" property, followed
3091 by zero or more characters with the "mark" property, and treats the
3092 sequence as an atomic group (see below). Characters with the "mark"
3093 property are typically accents that affect the preceding character.
3094
3095 Matching characters by Unicode property is not fast, because PCRE has
3096 to search a structure that contains data for over fifteen thousand
3097 characters. That is why the traditional escape sequences such as \d and
3098 \w do not use Unicode properties in PCRE.
3099
3100 Simple assertions
3101
3102 The final use of backslash is for certain simple assertions. An asser-
3103 tion specifies a condition that has to be met at a particular point in
3104 a match, without consuming any characters from the subject string. The
3105 use of subpatterns for more complicated assertions is described below.
3106 The backslashed assertions are:
3107
3108 \b matches at a word boundary
3109 \B matches when not at a word boundary
3110 \A matches at the start of the subject
3111 \Z matches at the end of the subject
3112 also matches before a newline at the end of the subject
3113 \z matches only at the end of the subject
3114 \G matches at the first matching position in the subject
3115
3116 These assertions may not appear in character classes (but note that \b
3117 has a different meaning, namely the backspace character, inside a char-
3118 acter class).
3119
3120 A word boundary is a position in the subject string where the current
3121 character and the previous character do not both match \w or \W (i.e.
3122 one matches \w and the other matches \W), or the start or end of the
3123 string if the first or last character matches \w, respectively.
3124
3125 The \A, \Z, and \z assertions differ from the traditional circumflex
3126 and dollar (described in the next section) in that they only ever match
3127 at the very start and end of the subject string, whatever options are
3128 set. Thus, they are independent of multiline mode. These three asser-
3129 tions are not affected by the PCRE_NOTBOL or PCRE_NOTEOL options, which
3130 affect only the behaviour of the circumflex and dollar metacharacters.
3131 However, if the startoffset argument of pcre_exec() is non-zero, indi-
3132 cating that matching is to start at a point other than the beginning of
3133 the subject, \A can never match. The difference between \Z and \z is
3134 that \Z matches before a newline at the end of the string as well as at
3135 the very end, whereas \z matches only at the end.
3136
3137 The \G assertion is true only when the current matching position is at
3138 the start point of the match, as specified by the startoffset argument
3139 of pcre_exec(). It differs from \A when the value of startoffset is
3140 non-zero. By calling pcre_exec() multiple times with appropriate argu-
3141 ments, you can mimic Perl's /g option, and it is in this kind of imple-
3142 mentation where \G can be useful.
3143
3144 Note, however, that PCRE's interpretation of \G, as the start of the
3145 current match, is subtly different from Perl's, which defines it as the
3146 end of the previous match. In Perl, these can be different when the
3147 previously matched string was empty. Because PCRE does just one match
3148 at a time, it cannot reproduce this behaviour.
3149
3150 If all the alternatives of a pattern begin with \G, the expression is
3151 anchored to the starting match position, and the "anchored" flag is set
3152 in the compiled regular expression.
3153
3154
3155 CIRCUMFLEX AND DOLLAR
3156
3157 Outside a character class, in the default matching mode, the circumflex
3158 character is an assertion that is true only if the current matching
3159 point is at the start of the subject string. If the startoffset argu-
3160 ment of pcre_exec() is non-zero, circumflex can never match if the
3161 PCRE_MULTILINE option is unset. Inside a character class, circumflex
3162 has an entirely different meaning (see below).
3163
3164 Circumflex need not be the first character of the pattern if a number
3165 of alternatives are involved, but it should be the first thing in each
3166 alternative in which it appears if the pattern is ever to match that
3167 branch. If all possible alternatives start with a circumflex, that is,
3168 if the pattern is constrained to match only at the start of the sub-
3169 ject, it is said to be an "anchored" pattern. (There are also other
3170 constructs that can cause a pattern to be anchored.)
3171
3172 A dollar character is an assertion that is true only if the current
3173 matching point is at the end of the subject string, or immediately
3174 before a newline at the end of the string (by default). Dollar need not
3175 be the last character of the pattern if a number of alternatives are
3176 involved, but it should be the last item in any branch in which it
3177 appears. Dollar has no special meaning in a character class.
3178
3179 The meaning of dollar can be changed so that it matches only at the
3180 very end of the string, by setting the PCRE_DOLLAR_ENDONLY option at
3181 compile time. This does not affect the \Z assertion.
3182
3183 The meanings of the circumflex and dollar characters are changed if the
3184 PCRE_MULTILINE option is set. When this is the case, a circumflex
3185 matches immediately after internal newlines as well as at the start of
3186 the subject string. It does not match after a newline that ends the
3187 string. A dollar matches before any newlines in the string, as well as
3188 at the very end, when PCRE_MULTILINE is set. When newline is specified
3189 as the two-character sequence CRLF, isolated CR and LF characters do
3190 not indicate newlines.
3191
3192 For example, the pattern /^abc$/ matches the subject string "def\nabc"
3193 (where \n represents a newline) in multiline mode, but not otherwise.
3194 Consequently, patterns that are anchored in single line mode because
3195 all branches start with ^ are not anchored in multiline mode, and a
3196 match for circumflex is possible when the startoffset argument of
3197 pcre_exec() is non-zero. The PCRE_DOLLAR_ENDONLY option is ignored if
3198 PCRE_MULTILINE is set.
3199
3200 Note that the sequences \A, \Z, and \z can be used to match the start
3201 and end of the subject in both modes, and if all branches of a pattern
3202 start with \A it is always anchored, whether or not PCRE_MULTILINE is
3203 set.
3204
3205
3206 FULL STOP (PERIOD, DOT)
3207
3208 Outside a character class, a dot in the pattern matches any one charac-
3209 ter in the subject string except (by default) a character that signi-
3210 fies the end of a line. In UTF-8 mode, the matched character may be
3211 more than one byte long.
3212
3213 When a line ending is defined as a single character, dot never matches
3214 that character; when the two-character sequence CRLF is used, dot does
3215 not match CR if it is immediately followed by LF, but otherwise it
3216 matches all characters (including isolated CRs and LFs). When any Uni-
3217 code line endings are being recognized, dot does not match CR or LF or
3218 any of the other line ending characters.
3219
3220 The behaviour of dot with regard to newlines can be changed. If the
3221 PCRE_DOTALL option is set, a dot matches any one character, without
3222 exception. If the two-character sequence CRLF is present in the subject
3223 string, it takes two dots to match it.
3224
3225 The handling of dot is entirely independent of the handling of circum-
3226 flex and dollar, the only relationship being that they both involve
3227 newlines. Dot has no special meaning in a character class.
3228
3229
3230 MATCHING A SINGLE BYTE
3231
3232 Outside a character class, the escape sequence \C matches any one byte,
3233 both in and out of UTF-8 mode. Unlike a dot, it always matches any
3234 line-ending characters. The feature is provided in Perl in order to
3235 match individual bytes in UTF-8 mode. Because it breaks up UTF-8 char-
3236 acters into individual bytes, what remains in the string may be a mal-
3237 formed UTF-8 string. For this reason, the \C escape sequence is best
3238 avoided.
3239
3240 PCRE does not allow \C to appear in lookbehind assertions (described
3241 below), because in UTF-8 mode this would make it impossible to calcu-
3242 late the length of the lookbehind.
3243
3244
3245 SQUARE BRACKETS AND CHARACTER CLASSES
3246
3247 An opening square bracket introduces a character class, terminated by a
3248 closing square bracket. A closing square bracket on its own is not spe-
3249 cial. If a closing square bracket is required as a member of the class,
3250 it should be the first data character in the class (after an initial
3251 circumflex, if present) or escaped with a backslash.
3252
3253 A character class matches a single character in the subject. In UTF-8
3254 mode, the character may occupy more than one byte. A matched character
3255 must be in the set of characters defined by the class, unless the first
3256 character in the class definition is a circumflex, in which case the
3257 subject character must not be in the set defined by the class. If a
3258 circumflex is actually required as a member of the class, ensure it is
3259 not the first character, or escape it with a backslash.
3260
3261 For example, the character class [aeiou] matches any lower case vowel,
3262 while [^aeiou] matches any character that is not a lower case vowel.
3263 Note that a circumflex is just a convenient notation for specifying the
3264 characters that are in the class by enumerating those that are not. A
3265 class that starts with a circumflex is not an assertion: it still con-
3266 sumes a character from the subject string, and therefore it fails if
3267 the current pointer is at the end of the string.
3268
3269 In UTF-8 mode, characters with values greater than 255 can be included
3270 in a class as a literal string of bytes, or by using the \x{ escaping
3271 mechanism.
3272
3273 When caseless matching is set, any letters in a class represent both
3274 their upper case and lower case versions, so for example, a caseless
3275 [aeiou] matches "A" as well as "a", and a caseless [^aeiou] does not
3276 match "A", whereas a caseful version would. In UTF-8 mode, PCRE always
3277 understands the concept of case for characters whose values are less
3278 than 128, so caseless matching is always possible. For characters with
3279 higher values, the concept of case is supported if PCRE is compiled
3280 with Unicode property support, but not otherwise. If you want to use
3281 caseless matching for characters 128 and above, you must ensure that
3282 PCRE is compiled with Unicode property support as well as with UTF-8
3283 support.
3284
3285 Characters that might indicate line breaks are never treated in any
3286 special way when matching character classes, whatever line-ending
3287 sequence is in use, and whatever setting of the PCRE_DOTALL and
3288 PCRE_MULTILINE options is used. A class such as [^a] always matches one
3289 of these characters.
3290
3291 The minus (hyphen) character can be used to specify a range of charac-
3292 ters in a character class. For example, [d-m] matches any letter
3293 between d and m, inclusive. If a minus character is required in a
3294 class, it must be escaped with a backslash or appear in a position
3295 where it cannot be interpreted as indicating a range, typically as the
3296 first or last character in the class.
3297
3298 It is not possible to have the literal character "]" as the end charac-
3299 ter of a range. A pattern such as [W-]46] is interpreted as a class of
3300 two characters ("W" and "-") followed by a literal string "46]", so it
3301 would match "W46]" or "-46]". However, if the "]" is escaped with a
3302 backslash it is interpreted as the end of range, so [W-\]46] is inter-
3303 preted as a class containing a range followed by two other characters.
3304 The octal or hexadecimal representation of "]" can also be used to end
3305 a range.
3306
3307 Ranges operate in the collating sequence of character values. They can
3308 also be used for characters specified numerically, for example
3309 [\000-\037]. In UTF-8 mode, ranges can include characters whose values
3310 are greater than 255, for example [\x{100}-\x{2ff}].
3311
3312 If a range that includes letters is used when caseless matching is set,
3313 it matches the letters in either case. For example, [W-c] is equivalent
3314 to [][\\^_`wxyzabc], matched caselessly, and in non-UTF-8 mode, if
3315 character tables for a French locale are in use, [\xc8-\xcb] matches
3316 accented E characters in both cases. In UTF-8 mode, PCRE supports the
3317 concept of case for characters with values greater than 128 only when
3318 it is compiled with Unicode property support.
3319
3320 The character types \d, \D, \p, \P, \s, \S, \w, and \W may also appear
3321 in a character class, and add the characters that they match to the
3322 class. For example, [\dABCDEF] matches any hexadecimal digit. A circum-
3323 flex can conveniently be used with the upper case character types to
3324 specify a more restricted set of characters than the matching lower
3325 case type. For example, the class [^\W_] matches any letter or digit,
3326 but not underscore.
3327
3328 The only metacharacters that are recognized in character classes are
3329 backslash, hyphen (only where it can be interpreted as specifying a
3330 range), circumflex (only at the start), opening square bracket (only
3331 when it can be interpreted as introducing a POSIX class name - see the
3332 next section), and the terminating closing square bracket. However,
3333 escaping other non-alphanumeric characters does no harm.
3334
3335
3336 POSIX CHARACTER CLASSES
3337
3338 Perl supports the POSIX notation for character classes. This uses names
3339 enclosed by [: and :] within the enclosing square brackets. PCRE also
3340 supports this notation. For example,
3341
3342 [01[:alpha:]%]
3343
3344 matches "0", "1", any alphabetic character, or "%". The supported class
3345 names are
3346
3347 alnum letters and digits
3348 alpha letters
3349 ascii character codes 0 - 127
3350 blank space or tab only
3351 cntrl control characters
3352 digit decimal digits (same as \d)
3353 graph printing characters, excluding space
3354 lower lower case letters
3355 print printing characters, including space
3356 punct printing characters, excluding letters and digits
3357 space white space (not quite the same as \s)
3358 upper upper case letters
3359 word "word" characters (same as \w)
3360 xdigit hexadecimal digits
3361
3362 The "space" characters are HT (9), LF (10), VT (11), FF (12), CR (13),
3363 and space (32). Notice that this list includes the VT character (code
3364 11). This makes "space" different to \s, which does not include VT (for
3365 Perl compatibility).
3366
3367 The name "word" is a Perl extension, and "blank" is a GNU extension
3368 from Perl 5.8. Another Perl extension is negation, which is indicated
3369 by a ^ character after the colon. For example,
3370
3371 [12[:^digit:]]
3372
3373 matches "1", "2", or any non-digit. PCRE (and Perl) also recognize the
3374 POSIX syntax [.ch.] and [=ch=] where "ch" is a "collating element", but
3375 these are not supported, and an error is given if they are encountered.
3376
3377 In UTF-8 mode, characters with values greater than 128 do not match any
3378 of the POSIX character classes.
3379
3380
3381 VERTICAL BAR
3382
3383 Vertical bar characters are used to separate alternative patterns. For
3384 example, the pattern
3385
3386 gilbert|sullivan
3387
3388 matches either "gilbert" or "sullivan". Any number of alternatives may
3389 appear, and an empty alternative is permitted (matching the empty
3390 string). The matching process tries each alternative in turn, from left
3391 to right, and the first one that succeeds is used. If the alternatives
3392 are within a subpattern (defined below), "succeeds" means matching the
3393 rest of the main pattern as well as the alternative in the subpattern.
3394
3395
3396 INTERNAL OPTION SETTING
3397
3398 The settings of the PCRE_CASELESS, PCRE_MULTILINE, PCRE_DOTALL, and
3399 PCRE_EXTENDED options can be changed from within the pattern by a
3400 sequence of Perl option letters enclosed between "(?" and ")". The
3401 option letters are
3402
3403 i for PCRE_CASELESS
3404 m for PCRE_MULTILINE
3405 s for PCRE_DOTALL
3406 x for PCRE_EXTENDED
3407
3408 For example, (?im) sets caseless, multiline matching. It is also possi-
3409 ble to unset these options by preceding the letter with a hyphen, and a
3410 combined setting and unsetting such as (?im-sx), which sets PCRE_CASE-
3411 LESS and PCRE_MULTILINE while unsetting PCRE_DOTALL and PCRE_EXTENDED,
3412 is also permitted. If a letter appears both before and after the
3413 hyphen, the option is unset.
3414
3415 When an option change occurs at top level (that is, not inside subpat-
3416 tern parentheses), the change applies to the remainder of the pattern
3417 that follows. If the change is placed right at the start of a pattern,
3418 PCRE extracts it into the global options (and it will therefore show up
3419 in data extracted by the pcre_fullinfo() function).
3420
3421 An option change within a subpattern (see below for a description of
3422 subpatterns) affects only that part of the current pattern that follows
3423 it, so
3424
3425 (a(?i)b)c
3426
3427 matches abc and aBc and no other strings (assuming PCRE_CASELESS is not
3428 used). By this means, options can be made to have different settings
3429 in different parts of the pattern. Any changes made in one alternative
3430 do carry on into subsequent branches within the same subpattern. For
3431 example,
3432
3433 (a(?i)b|c)
3434
3435 matches "ab", "aB", "c", and "C", even though when matching "C" the
3436 first branch is abandoned before the option setting. This is because
3437 the effects of option settings happen at compile time. There would be
3438 some very weird behaviour otherwise.
3439
3440 The PCRE-specific options PCRE_DUPNAMES, PCRE_UNGREEDY, and PCRE_EXTRA
3441 can be changed in the same way as the Perl-compatible options by using
3442 the characters J, U and X respectively.
3443
3444
3445 SUBPATTERNS
3446
3447 Subpatterns are delimited by parentheses (round brackets), which can be
3448 nested. Turning part of a pattern into a subpattern does two things:
3449
3450 1. It localizes a set of alternatives. For example, the pattern
3451
3452 cat(aract|erpillar|)
3453
3454 matches one of the words "cat", "cataract", or "caterpillar". Without
3455 the parentheses, it would match "cataract", "erpillar" or an empty
3456 string.
3457
3458 2. It sets up the subpattern as a capturing subpattern. This means
3459 that, when the whole pattern matches, that portion of the subject
3460 string that matched the subpattern is passed back to the caller via the
3461 ovector argument of pcre_exec(). Opening parentheses are counted from
3462 left to right (starting from 1) to obtain numbers for the capturing
3463 subpatterns.
3464
3465 For example, if the string "the red king" is matched against the pat-
3466 tern
3467
3468 the ((red|white) (king|queen))
3469
3470 the captured substrings are "red king", "red", and "king", and are num-
3471 bered 1, 2, and 3, respectively.
3472
3473 The fact that plain parentheses fulfil two functions is not always
3474 helpful. There are often times when a grouping subpattern is required
3475 without a capturing requirement. If an opening parenthesis is followed
3476 by a question mark and a colon, the subpattern does not do any captur-
3477 ing, and is not counted when computing the number of any subsequent
3478 capturing subpatterns. For example, if the string "the white queen" is
3479 matched against the pattern
3480
3481 the ((?:red|white) (king|queen))
3482
3483 the captured substrings are "white queen" and "queen", and are numbered
3484 1 and 2. The maximum number of capturing subpatterns is 65535.
3485
3486 As a convenient shorthand, if any option settings are required at the
3487 start of a non-capturing subpattern, the option letters may appear
3488 between the "?" and the ":". Thus the two patterns
3489
3490 (?i:saturday|sunday)
3491 (?:(?i)saturday|sunday)
3492
3493 match exactly the same set of strings. Because alternative branches are
3494 tried from left to right, and options are not reset until the end of
3495 the subpattern is reached, an option setting in one branch does affect
3496 subsequent branches, so the above patterns match "SUNDAY" as well as
3497 "Saturday".
3498
3499
3500 NAMED SUBPATTERNS
3501
3502 Identifying capturing parentheses by number is simple, but it can be
3503 very hard to keep track of the numbers in complicated regular expres-
3504 sions. Furthermore, if an expression is modified, the numbers may
3505 change. To help with this difficulty, PCRE supports the naming of sub-
3506 patterns. This feature was not added to Perl until release 5.10. Python
3507 had the feature earlier, and PCRE introduced it at release 4.0, using
3508 the Python syntax. PCRE now supports both the Perl and the Python syn-
3509 tax.
3510
3511 In PCRE, a subpattern can be named in one of three ways: (?<name>...)
3512 or (?'name'...) as in Perl, or (?P<name>...) as in Python. References
3513 to capturing parentheses from other parts of the pattern, such as back-
3514 references, recursion, and conditions, can be made by name as well as
3515 by number.
3516
3517 Names consist of up to 32 alphanumeric characters and underscores.
3518 Named capturing parentheses are still allocated numbers as well as
3519 names, exactly as if the names were not present. The PCRE API provides
3520 function calls for extracting the name-to-number translation table from
3521 a compiled pattern. There is also a convenience function for extracting
3522 a captured substring by name.
3523
3524 By default, a name must be unique within a pattern, but it is possible
3525 to relax this constraint by setting the PCRE_DUPNAMES option at compile
3526 time. This can be useful for patterns where only one instance of the
3527 named parentheses can match. Suppose you want to match the name of a
3528 weekday, either as a 3-letter abbreviation or as the full name, and in
3529 both cases you want to extract the abbreviation. This pattern (ignoring
3530 the line breaks) does the job:
3531
3532 (?<DN>Mon|Fri|Sun)(?:day)?|
3533 (?<DN>Tue)(?:sday)?|
3534 (?<DN>Wed)(?:nesday)?|
3535 (?<DN>Thu)(?:rsday)?|
3536 (?<DN>Sat)(?:urday)?
3537
3538 There are five capturing substrings, but only one is ever set after a
3539 match. The convenience function for extracting the data by name
3540 returns the substring for the first (and in this example, the only)
3541 subpattern of that name that matched. This saves searching to find
3542 which numbered subpattern it was. If you make a reference to a non-
3543 unique named subpattern from elsewhere in the pattern, the one that
3544 corresponds to the lowest number is used. For further details of the
3545 interfaces for handling named subpatterns, see the pcreapi documenta-
3546 tion.
3547
3548
3549 REPETITION
3550
3551 Repetition is specified by quantifiers, which can follow any of the
3552 following items:
3553
3554 a literal data character
3555 the dot metacharacter
3556 the \C escape sequence
3557 the \X escape sequence (in UTF-8 mode with Unicode properties)
3558 the \R escape sequence
3559 an escape such as \d that matches a single character
3560 a character class
3561 a back reference (see next section)
3562 a parenthesized subpattern (unless it is an assertion)
3563
3564 The general repetition quantifier specifies a minimum and maximum num-
3565 ber of permitted matches, by giving the two numbers in curly brackets
3566 (braces), separated by a comma. The numbers must be less than 65536,
3567 and the first must be less than or equal to the second. For example:
3568
3569 z{2,4}
3570
3571 matches "zz", "zzz", or "zzzz". A closing brace on its own is not a
3572 special character. If the second number is omitted, but the comma is
3573 present, there is no upper limit; if the second number and the comma
3574 are both omitted, the quantifier specifies an exact number of required
3575 matches. Thus
3576
3577 [aeiou]{3,}
3578
3579 matches at least 3 successive vowels, but may match many more, while
3580
3581 \d{8}
3582
3583 matches exactly 8 digits. An opening curly bracket that appears in a
3584 position where a quantifier is not allowed, or one that does not match
3585 the syntax of a quantifier, is taken as a literal character. For exam-
3586 ple, {,6} is not a quantifier, but a literal string of four characters.
3587
3588 In UTF-8 mode, quantifiers apply to UTF-8 characters rather than to
3589 individual bytes. Thus, for example, \x{100}{2} matches two UTF-8 char-
3590 acters, each of which is represented by a two-byte sequence. Similarly,
3591 when Unicode property support is available, \X{3} matches three Unicode
3592 extended sequences, each of which may be several bytes long (and they
3593 may be of different lengths).
3594
3595 The quantifier {0} is permitted, causing the expression to behave as if
3596 the previous item and the quantifier were not present.
3597
3598 For convenience, the three most common quantifiers have single-charac-
3599 ter abbreviations:
3600
3601 * is equivalent to {0,}
3602 + is equivalent to {1,}
3603 ? is equivalent to {0,1}
3604
3605 It is possible to construct infinite loops by following a subpattern
3606 that can match no characters with a quantifier that has no upper limit,
3607 for example:
3608
3609 (a?)*
3610
3611 Earlier versions of Perl and PCRE used to give an error at compile time
3612 for such patterns. However, because there are cases where this can be
3613 useful, such patterns are now accepted, but if any repetition of the
3614 subpattern does in fact match no characters, the loop is forcibly bro-
3615 ken.
3616
3617 By default, the quantifiers are "greedy", that is, they match as much
3618 as possible (up to the maximum number of permitted times), without
3619 causing the rest of the pattern to fail. The classic example of where
3620 this gives problems is in trying to match comments in C programs. These
3621 appear between /* and */ and within the comment, individual * and /
3622 characters may appear. An attempt to match C comments by applying the
3623 pattern
3624
3625 /\*.*\*/
3626
3627 to the string
3628
3629 /* first comment */ not comment /* second comment */
3630
3631 fails, because it matches the entire string owing to the greediness of
3632 the .* item.
3633
3634 However, if a quantifier is followed by a question mark, it ceases to
3635 be greedy, and instead matches the minimum number of times possible, so
3636 the pattern
3637
3638 /\*.*?\*/
3639
3640 does the right thing with the C comments. The meaning of the various
3641 quantifiers is not otherwise changed, just the preferred number of
3642 matches. Do not confuse this use of question mark with its use as a
3643 quantifier in its own right. Because it has two uses, it can sometimes
3644 appear doubled, as in
3645
3646 \d??\d
3647
3648 which matches one digit by preference, but can match two if that is the
3649 only way the rest of the pattern matches.
3650
3651 If the PCRE_UNGREEDY option is set (an option that is not available in
3652 Perl), the quantifiers are not greedy by default, but individual ones
3653 can be made greedy by following them with a question mark. In other
3654 words, it inverts the default behaviour.
3655
3656 When a parenthesized subpattern is quantified with a minimum repeat
3657 count that is greater than 1 or with a limited maximum, more memory is
3658 required for the compiled pattern, in proportion to the size of the
3659 minimum or maximum.
3660
3661 If a pattern starts with .* or .{0,} and the PCRE_DOTALL option (equiv-
3662 alent to Perl's /s) is set, thus allowing the dot to match newlines,
3663 the pattern is implicitly anchored, because whatever follows will be
3664 tried against every character position in the subject string, so there
3665 is no point in retrying the overall match at any position after the
3666 first. PCRE normally treats such a pattern as though it were preceded
3667 by \A.
3668
3669 In cases where it is known that the subject string contains no new-
3670 lines, it is worth setting PCRE_DOTALL in order to obtain this opti-
3671 mization, or alternatively using ^ to indicate anchoring explicitly.
3672
3673 However, there is one situation where the optimization cannot be used.
3674 When .* is inside capturing parentheses that are the subject of a
3675 backreference elsewhere in the pattern, a match at the start may fail
3676 where a later one succeeds. Consider, for example:
3677
3678 (.*)abc\1
3679
3680 If the subject is "xyz123abc123" the match point is the fourth charac-
3681 ter. For this reason, such a pattern is not implicitly anchored.
3682
3683 When a capturing subpattern is repeated, the value captured is the sub-
3684 string that matched the final iteration. For example, after
3685
3686 (tweedle[dume]{3}\s*)+
3687
3688 has matched "tweedledum tweedledee" the value of the captured substring
3689 is "tweedledee". However, if there are nested capturing subpatterns,
3690 the corresponding captured values may have been set in previous itera-
3691 tions. For example, after
3692
3693 /(a|(b))+/
3694
3695 matches "aba" the value of the second captured substring is "b".
3696
3697
3698 ATOMIC GROUPING AND POSSESSIVE QUANTIFIERS
3699
3700 With both maximizing ("greedy") and minimizing ("ungreedy" or "lazy")
3701 repetition, failure of what follows normally causes the repeated item
3702 to be re-evaluated to see if a different number of repeats allows the
3703 rest of the pattern to match. Sometimes it is useful to prevent this,
3704 either to change the nature of the match, or to cause it fail earlier
3705 than it otherwise might, when the author of the pattern knows there is
3706 no point in carrying on.
3707
3708 Consider, for example, the pattern \d+foo when applied to the subject
3709 line
3710
3711 123456bar
3712
3713 After matching all 6 digits and then failing to match "foo", the normal
3714 action of the matcher is to try again with only 5 digits matching the
3715 \d+ item, and then with 4, and so on, before ultimately failing.
3716 "Atomic grouping" (a term taken from Jeffrey Friedl's book) provides
3717 the means for specifying that once a subpattern has matched, it is not
3718 to be re-evaluated in this way.
3719
3720 If we use atomic grouping for the previous example, the matcher gives
3721 up immediately on failing to match "foo" the first time. The notation
3722 is a kind of special parenthesis, starting with (?> as in this example:
3723
3724 (?>\d+)foo
3725
3726 This kind of parenthesis "locks up" the part of the pattern it con-
3727 tains once it has matched, and a failure further into the pattern is
3728 prevented from backtracking into it. Backtracking past it to previous
3729 items, however, works as normal.
3730
3731 An alternative description is that a subpattern of this type matches
3732 the string of characters that an identical standalone pattern would
3733 match, if anchored at the current point in the subject string.
3734
3735 Atomic grouping subpatterns are not capturing subpatterns. Simple cases
3736 such as the above example can be thought of as a maximizing repeat that
3737 must swallow everything it can. So, while both \d+ and \d+? are pre-
3738 pared to adjust the number of digits they match in order to make the
3739 rest of the pattern match, (?>\d+) can only match an entire sequence of
3740 digits.
3741
3742 Atomic groups in general can of course contain arbitrarily complicated
3743 subpatterns, and can be nested. However, when the subpattern for an
3744 atomic group is just a single repeated item, as in the example above, a
3745 simpler notation, called a "possessive quantifier" can be used. This
3746 consists of an additional + character following a quantifier. Using
3747 this notation, the previous example can be rewritten as
3748
3749 \d++foo
3750
3751 Possessive quantifiers are always greedy; the setting of the
3752 PCRE_UNGREEDY option is ignored. They are a convenient notation for the
3753 simpler forms of atomic group. However, there is no difference in the
3754 meaning of a possessive quantifier and the equivalent atomic group,
3755 though there may be a performance difference; possessive quantifiers
3756 should be slightly faster.
3757
3758 The possessive quantifier syntax is an extension to the Perl 5.8 syn-
3759 tax. Jeffrey Friedl originated the idea (and the name) in the first
3760 edition of his book. Mike McCloskey liked it, so implemented it when he
3761 built Sun's Java package, and PCRE copied it from there. It ultimately
3762 found its way into Perl at release 5.10.
3763
3764 PCRE has an optimization that automatically "possessifies" certain sim-
3765 ple pattern constructs. For example, the sequence A+B is treated as
3766 A++B because there is no point in backtracking into a sequence of A's
3767 when B must follow.
3768
3769 When a pattern contains an unlimited repeat inside a subpattern that
3770 can itself be repeated an unlimited number of times, the use of an
3771 atomic group is the only way to avoid some failing matches taking a
3772 very long time indeed. The pattern
3773
3774 (\D+|<\d+>)*[!?]
3775
3776 matches an unlimited number of substrings that either consist of non-
3777 digits, or digits enclosed in <>, followed by either ! or ?. When it
3778 matches, it runs quickly. However, if it is applied to
3779
3780 aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
3781
3782 it takes a long time before reporting failure. This is because the
3783 string can be divided between the internal \D+ repeat and the external
3784 * repeat in a large number of ways, and all have to be tried. (The
3785 example uses [!?] rather than a single character at the end, because
3786 both PCRE and Perl have an optimization that allows for fast failure
3787 when a single character is used. They remember the last single charac-
3788 ter that is required for a match, and fail early if it is not present
3789 in the string.) If the pattern is changed so that it uses an atomic
3790 group, like this:
3791
3792 ((?>\D+)|<\d+>)*[!?]
3793
3794 sequences of non-digits cannot be broken, and failure happens quickly.
3795
3796
3797 BACK REFERENCES
3798
3799 Outside a character class, a backslash followed by a digit greater than
3800 0 (and possibly further digits) is a back reference to a capturing sub-
3801 pattern earlier (that is, to its left) in the pattern, provided there
3802 have been that many previous capturing left parentheses.
3803
3804 However, if the decimal number following the backslash is less than 10,
3805 it is always taken as a back reference, and causes an error only if
3806 there are not that many capturing left parentheses in the entire pat-
3807 tern. In other words, the parentheses that are referenced need not be
3808 to the left of the reference for numbers less than 10. A "forward back
3809 reference" of this type can make sense when a repetition is involved
3810 and the subpattern to the right has participated in an earlier itera-
3811 tion.
3812
3813 It is not possible to have a numerical "forward back reference" to a
3814 subpattern whose number is 10 or more using this syntax because a
3815 sequence such as \50 is interpreted as a character defined in octal.
3816 See the subsection entitled "Non-printing characters" above for further
3817 details of the handling of digits following a backslash. There is no
3818 such problem when named parentheses are used. A back reference to any
3819 subpattern is possible using named parentheses (see below).
3820
3821 Another way of avoiding the ambiguity inherent in the use of digits
3822 following a backslash is to use the \g escape sequence, which is a fea-
3823 ture introduced in Perl 5.10. This escape must be followed by a posi-
3824 tive or a negative number, optionally enclosed in braces. These exam-
3825 ples are all identical:
3826
3827 (ring), \1
3828 (ring), \g1
3829 (ring), \g{1}
3830
3831 A positive number specifies an absolute reference without the ambiguity
3832 that is present in the older syntax. It is also useful when literal
3833 digits follow the reference. A negative number is a relative reference.
3834 Consider this example:
3835
3836 (abc(def)ghi)\g{-1}
3837
3838 The sequence \g{-1} is a reference to the most recently started captur-
3839 ing subpattern before \g, that is, is it equivalent to \2. Similarly,
3840 \g{-2} would be equivalent to \1. The use of relative references can be
3841 helpful in long patterns, and also in patterns that are created by
3842 joining together fragments that contain references within themselves.
3843
3844 A back reference matches whatever actually matched the capturing sub-
3845 pattern in the current subject string, rather than anything matching
3846 the subpattern itself (see "Subpatterns as subroutines" below for a way
3847 of doing that). So the pattern
3848
3849 (sens|respons)e and \1ibility
3850
3851 matches "sense and sensibility" and "response and responsibility", but
3852 not "sense and responsibility". If caseful matching is in force at the
3853 time of the back reference, the case of letters is relevant. For exam-
3854 ple,
3855
3856 ((?i)rah)\s+\1
3857
3858 matches "rah rah" and "RAH RAH", but not "RAH rah", even though the
3859 original capturing subpattern is matched caselessly.
3860
3861 Back references to named subpatterns use the Perl syntax \k<name> or
3862 \k'name' or the Python syntax (?P=name). We could rewrite the above
3863 example in either of the following ways:
3864
3865 (?<p1>(?i)rah)\s+\k<p1>
3866 (?P<p1>(?i)rah)\s+(?P=p1)
3867
3868 A subpattern that is referenced by name may appear in the pattern
3869 before or after the reference.
3870
3871 There may be more than one back reference to the same subpattern. If a
3872 subpattern has not actually been used in a particular match, any back
3873 references to it always fail. For example, the pattern
3874
3875 (a|(bc))\2
3876
3877 always fails if it starts to match "a" rather than "bc". Because there
3878 may be many capturing parentheses in a pattern, all digits following
3879 the backslash are taken as part of a potential back reference number.
3880 If the pattern continues with a digit character, some delimiter must be
3881 used to terminate the back reference. If the PCRE_EXTENDED option is
3882 set, this can be whitespace. Otherwise an empty comment (see "Com-
3883 ments" below) can be used.
3884
3885 A back reference that occurs inside the parentheses to which it refers
3886 fails when the subpattern is first used, so, for example, (a\1) never
3887 matches. However, such references can be useful inside repeated sub-
3888 patterns. For example, the pattern
3889
3890 (a|b\1)+
3891
3892 matches any number of "a"s and also "aba", "ababbaa" etc. At each iter-
3893 ation of the subpattern, the back reference matches the character
3894 string corresponding to the previous iteration. In order for this to
3895 work, the pattern must be such that the first iteration does not need
3896 to match the back reference. This can be done using alternation, as in
3897 the example above, or by a quantifier with a minimum of zero.
3898
3899
3900 ASSERTIONS
3901
3902 An assertion is a test on the characters following or preceding the
3903 current matching point that does not actually consume any characters.
3904 The simple assertions coded as \b, \B, \A, \G, \Z, \z, ^ and $ are
3905 described above.
3906
3907 More complicated assertions are coded as subpatterns. There are two
3908 kinds: those that look ahead of the current position in the subject
3909 string, and those that look behind it. An assertion subpattern is
3910 matched in the normal way, except that it does not cause the current
3911 matching position to be changed.
3912
3913 Assertion subpatterns are not capturing subpatterns, and may not be
3914 repeated, because it makes no sense to assert the same thing several
3915 times. If any kind of assertion contains capturing subpatterns within
3916 it, these are counted for the purposes of numbering the capturing sub-
3917 patterns in the whole pattern. However, substring capturing is carried
3918 out only for positive assertions, because it does not make sense for
3919 negative assertions.
3920
3921 Lookahead assertions
3922
3923 Lookahead assertions start with (?= for positive assertions and (?! for
3924 negative assertions. For example,
3925
3926 \w+(?=;)
3927
3928 matches a word followed by a semicolon, but does not include the semi-
3929 colon in the match, and
3930
3931 foo(?!bar)
3932
3933 matches any occurrence of "foo" that is not followed by "bar". Note
3934 that the apparently similar pattern
3935
3936 (?!foo)bar
3937
3938 does not find an occurrence of "bar" that is preceded by something
3939 other than "foo"; it finds any occurrence of "bar" whatsoever, because
3940 the assertion (?!foo) is always true when the next three characters are
3941 "bar". A lookbehind assertion is needed to achieve the other effect.
3942
3943 If you want to force a matching failure at some point in a pattern, the
3944 most convenient way to do it is with (?!) because an empty string
3945 always matches, so an assertion that requires there not to be an empty
3946 string must always fail.
3947
3948 Lookbehind assertions
3949
3950 Lookbehind assertions start with (?<= for positive assertions and (?<!
3951 for negative assertions. For example,
3952
3953 (?<!foo)bar
3954
3955 does find an occurrence of "bar" that is not preceded by "foo". The
3956 contents of a lookbehind assertion are restricted such that all the
3957 strings it matches must have a fixed length. However, if there are sev-
3958 eral top-level alternatives, they do not all have to have the same
3959 fixed length. Thus
3960
3961 (?<=bullock|donkey)
3962
3963 is permitted, but
3964
3965 (?<!dogs?|cats?)
3966
3967 causes an error at compile time. Branches that match different length
3968 strings are permitted only at the top level of a lookbehind assertion.
3969 This is an extension compared with Perl (at least for 5.8), which
3970 requires all branches to match the same length of string. An assertion
3971 such as
3972
3973 (?<=ab(c|de))
3974
3975 is not permitted, because its single top-level branch can match two
3976 different lengths, but it is acceptable if rewritten to use two top-
3977 level branches:
3978
3979 (?<=abc|abde)
3980
3981 The implementation of lookbehind assertions is, for each alternative,
3982 to temporarily move the current position back by the fixed length and
3983 then try to match. If there are insufficient characters before the cur-
3984 rent position, the assertion fails.
3985
3986 PCRE does not allow the \C escape (which matches a single byte in UTF-8
3987 mode) to appear in lookbehind assertions, because it makes it impossi-
3988 ble to calculate the length of the lookbehind. The \X and \R escapes,
3989 which can match different numbers of bytes, are also not permitted.
3990
3991 Possessive quantifiers can be used in conjunction with lookbehind
3992 assertions to specify efficient matching at the end of the subject
3993 string. Consider a simple pattern such as
3994
3995 abcd$
3996
3997 when applied to a long string that does not match. Because matching
3998 proceeds from left to right, PCRE will look for each "a" in the subject
3999 and then see if what follows matches the rest of the pattern. If the
4000 pattern is specified as
4001
4002 ^.*abcd$
4003
4004 the initial .* matches the entire string at first, but when this fails
4005 (because there is no following "a"), it backtracks to match all but the
4006 last character, then all but the last two characters, and so on. Once
4007 again the search for "a" covers the entire string, from right to left,
4008 so we are no better off. However, if the pattern is written as
4009
4010 ^.*+(?<=abcd)
4011
4012 there can be no backtracking for the .*+ item; it can match only the
4013 entire string. The subsequent lookbehind assertion does a single test
4014 on the last four characters. If it fails, the match fails immediately.
4015 For long strings, this approach makes a significant difference to the
4016 processing time.
4017
4018 Using multiple assertions
4019
4020 Several assertions (of any sort) may occur in succession. For example,
4021
4022 (?<=\d{3})(?<!999)foo
4023
4024 matches "foo" preceded by three digits that are not "999". Notice that
4025 each of the assertions is applied independently at the same point in
4026 the subject string. First there is a check that the previous three
4027 characters are all digits, and then there is a check that the same
4028 three characters are not "999". This pattern does not match "foo" pre-
4029 ceded by six characters, the first of which are digits and the last
4030 three of which are not "999". For example, it doesn't match "123abc-
4031 foo". A pattern to do that is
4032
4033 (?<=\d{3}...)(?<!999)foo
4034
4035 This time the first assertion looks at the preceding six characters,
4036 checking that the first three are digits, and then the second assertion
4037 checks that the preceding three characters are not "999".
4038
4039 Assertions can be nested in any combination. For example,
4040
4041 (?<=(?<!foo)bar)baz
4042
4043 matches an occurrence of "baz" that is preceded by "bar" which in turn
4044 is not preceded by "foo", while
4045
4046 (?<=\d{3}(?!999)...)foo
4047
4048 is another pattern that matches "foo" preceded by three digits and any
4049 three characters that are not "999".
4050
4051
4052 CONDITIONAL SUBPATTERNS
4053
4054 It is possible to cause the matching process to obey a subpattern con-
4055 ditionally or to choose between two alternative subpatterns, depending
4056 on the result of an assertion, or whether a previous capturing subpat-
4057 tern matched or not. The two possible forms of conditional subpattern
4058 are
4059
4060 (?(condition)yes-pattern)
4061 (?(condition)yes-pattern|no-pattern)
4062
4063 If the condition is satisfied, the yes-pattern is used; otherwise the
4064 no-pattern (if present) is used. If there are more than two alterna-
4065 tives in the subpattern, a compile-time error occurs.
4066
4067 There are four kinds of condition: references to subpatterns, refer-
4068 ences to recursion, a pseudo-condition called DEFINE, and assertions.
4069
4070 Checking for a used subpattern by number
4071
4072 If the text between the parentheses consists of a sequence of digits,
4073 the condition is true if the capturing subpattern of that number has
4074 previously matched.
4075
4076 Consider the following pattern, which contains non-significant white
4077 space to make it more readable (assume the PCRE_EXTENDED option) and to
4078 divide it into three parts for ease of discussion:
4079
4080 ( \( )? [^()]+ (?(1) \) )
4081
4082 The first part matches an optional opening parenthesis, and if that
4083 character is present, sets it as the first captured substring. The sec-
4084 ond part matches one or more characters that are not parentheses. The
4085 third part is a conditional subpattern that tests whether the first set
4086 of parentheses matched or not. If they did, that is, if subject started
4087 with an opening parenthesis, the condition is true, and so the yes-pat-
4088 tern is executed and a closing parenthesis is required. Otherwise,
4089 since no-pattern is not present, the subpattern matches nothing. In
4090 other words, this pattern matches a sequence of non-parentheses,
4091 optionally enclosed in parentheses.
4092
4093 Checking for a used subpattern by name
4094
4095 Perl uses the syntax (?(<name>)...) or (?('name')...) to test for a
4096 used subpattern by name. For compatibility with earlier versions of
4097 PCRE, which had this facility before Perl, the syntax (?(name)...) is
4098 also recognized. However, there is a possible ambiguity with this syn-
4099 tax, because subpattern names may consist entirely of digits. PCRE
4100 looks first for a named subpattern; if it cannot find one and the name
4101 consists entirely of digits, PCRE looks for a subpattern of that num-
4102 ber, which must be greater than zero. Using subpattern names that con-
4103 sist entirely of digits is not recommended.
4104
4105 Rewriting the above example to use a named subpattern gives this:
4106
4107 (?<OPEN> \( )? [^()]+ (?(<OPEN>) \) )
4108
4109
4110 Checking for pattern recursion
4111
4112 If the condition is the string (R), and there is no subpattern with the
4113 name R, the condition is true if a recursive call to the whole pattern
4114 or any subpattern has been made. If digits or a name preceded by amper-
4115 sand follow the letter R, for example:
4116
4117 (?(R3)...) or (?(R&name)...)
4118
4119 the condition is true if the most recent recursion is into the subpat-
4120 tern whose number or name is given. This condition does not check the
4121 entire recursion stack.
4122
4123 At "top level", all these recursion test conditions are false. Recur-
4124 sive patterns are described below.
4125
4126 Defining subpatterns for use by reference only
4127
4128 If the condition is the string (DEFINE), and there is no subpattern
4129 with the name DEFINE, the condition is always false. In this case,
4130 there may be only one alternative in the subpattern. It is always
4131 skipped if control reaches this point in the pattern; the idea of
4132 DEFINE is that it can be used to define "subroutines" that can be ref-
4133 erenced from elsewhere. (The use of "subroutines" is described below.)
4134 For example, a pattern to match an IPv4 address could be written like
4135 this (ignore whitespace and line breaks):
4136
4137 (?(DEFINE) (?<byte> 2[0-4]\d | 25[0-5] | 1\d\d | [1-9]?\d) )
4138 \b (?&byte) (\.(?&byte)){3} \b
4139
4140 The first part of the pattern is a DEFINE group inside which a another
4141 group named "byte" is defined. This matches an individual component of
4142 an IPv4 address (a number less than 256). When matching takes place,
4143 this part of the pattern is skipped because DEFINE acts like a false
4144 condition.
4145
4146 The rest of the pattern uses references to the named group to match the
4147 four dot-separated components of an IPv4 address, insisting on a word
4148 boundary at each end.
4149
4150 Assertion conditions
4151
4152 If the condition is not in any of the above formats, it must be an
4153 assertion. This may be a positive or negative lookahead or lookbehind
4154 assertion. Consider this pattern, again containing non-significant
4155 white space, and with the two alternatives on the second line:
4156
4157 (?(?=[^a-z]*[a-z])
4158 \d{2}-[a-z]{3}-\d{2} | \d{2}-\d{2}-\d{2} )
4159
4160 The condition is a positive lookahead assertion that matches an
4161 optional sequence of non-letters followed by a letter. In other words,
4162 it tests for the presence of at least one letter in the subject. If a
4163 letter is found, the subject is matched against the first alternative;
4164 otherwise it is matched against the second. This pattern matches
4165 strings in one of the two forms dd-aaa-dd or dd-dd-dd, where aaa are
4166 letters and dd are digits.
4167
4168
4169 COMMENTS
4170
4171 The sequence (?# marks the start of a comment that continues up to the
4172 next closing parenthesis. Nested parentheses are not permitted. The
4173 characters that make up a comment play no part in the pattern matching
4174 at all.
4175
4176 If the PCRE_EXTENDED option is set, an unescaped # character outside a
4177 character class introduces a comment that continues to immediately
4178 after the next newline in the pattern.
4179
4180
4181 RECURSIVE PATTERNS
4182
4183 Consider the problem of matching a string in parentheses, allowing for
4184 unlimited nested parentheses. Without the use of recursion, the best
4185 that can be done is to use a pattern that matches up to some fixed
4186 depth of nesting. It is not possible to handle an arbitrary nesting
4187 depth.
4188
4189 For some time, Perl has provided a facility that allows regular expres-
4190 sions to recurse (amongst other things). It does this by interpolating
4191 Perl code in the expression at run time, and the code can refer to the
4192 expression itself. A Perl pattern using code interpolation to solve the
4193 parentheses problem can be created like this:
4194
4195 $re = qr{\( (?: (?>[^()]+) | (?p{$re}) )* \)}x;
4196
4197 The (?p{...}) item interpolates Perl code at run time, and in this case
4198 refers recursively to the pattern in which it appears.
4199
4200 Obviously, PCRE cannot support the interpolation of Perl code. Instead,
4201 it supports special syntax for recursion of the entire pattern, and
4202 also for individual subpattern recursion. After its introduction in
4203 PCRE and Python, this kind of recursion was introduced into Perl at
4204 release 5.10.
4205
4206 A special item that consists of (? followed by a number greater than
4207 zero and a closing parenthesis is a recursive call of the subpattern of
4208 the given number, provided that it occurs inside that subpattern. (If
4209 not, it is a "subroutine" call, which is described in the next sec-
4210 tion.) The special item (?R) or (?0) is a recursive call of the entire
4211 regular expression.
4212
4213 In PCRE (like Python, but unlike Perl), a recursive subpattern call is
4214 always treated as an atomic group. That is, once it has matched some of
4215 the subject string, it is never re-entered, even if it contains untried
4216 alternatives and there is a subsequent matching failure.
4217
4218 This PCRE pattern solves the nested parentheses problem (assume the
4219 PCRE_EXTENDED option is set so that white space is ignored):
4220
4221 \( ( (?>[^()]+) | (?R) )* \)
4222
4223 First it matches an opening parenthesis. Then it matches any number of
4224 substrings which can either be a sequence of non-parentheses, or a
4225 recursive match of the pattern itself (that is, a correctly parenthe-
4226 sized substring). Finally there is a closing parenthesis.
4227
4228 If this were part of a larger pattern, you would not want to recurse
4229 the entire pattern, so instead you could use this:
4230
4231 ( \( ( (?>[^()]+) | (?1) )* \) )
4232
4233 We have put the pattern into parentheses, and caused the recursion to
4234 refer to them instead of the whole pattern. In a larger pattern, keep-
4235 ing track of parenthesis numbers can be tricky. It may be more conve-
4236 nient to use named parentheses instead. The Perl syntax for this is
4237 (?&name); PCRE's earlier syntax (?P>name) is also supported. We could
4238 rewrite the above example as follows:
4239
4240 (?<pn> \( ( (?>[^()]+) | (?&pn) )* \) )
4241
4242 If there is more than one subpattern with the same name, the earliest
4243 one is used. This particular example pattern contains nested unlimited
4244 repeats, and so the use of atomic grouping for matching strings of non-
4245 parentheses is important when applying the pattern to strings that do
4246 not match. For example, when this pattern is applied to
4247
4248 (aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa()
4249
4250 it yields "no match" quickly. However, if atomic grouping is not used,
4251 the match runs for a very long time indeed because there are so many
4252 different ways the + and * repeats can carve up the subject, and all
4253 have to be tested before failure can be reported.
4254
4255 At the end of a match, the values set for any capturing subpatterns are
4256 those from the outermost level of the recursion at which the subpattern
4257 value is set. If you want to obtain intermediate values, a callout
4258 function can be used (see below and the pcrecallout documentation). If
4259 the pattern above is matched against
4260
4261 (ab(cd)ef)
4262
4263 the value for the capturing parentheses is "ef", which is the last
4264 value taken on at the top level. If additional parentheses are added,
4265 giving
4266
4267 \( ( ( (?>[^()]+) | (?R) )* ) \)
4268 ^ ^
4269 ^ ^
4270
4271 the string they capture is "ab(cd)ef", the contents of the top level
4272 parentheses. If there are more than 15 capturing parentheses in a pat-
4273 tern, PCRE has to obtain extra memory to store data during a recursion,
4274 which it does by using pcre_malloc, freeing it via pcre_free after-
4275 wards. If no memory can be obtained, the match fails with the
4276 PCRE_ERROR_NOMEMORY error.
4277
4278 Do not confuse the (?R) item with the condition (R), which tests for
4279 recursion. Consider this pattern, which matches text in angle brack-
4280 ets, allowing for arbitrary nesting. Only digits are allowed in nested
4281 brackets (that is, when recursing), whereas any characters are permit-
4282 ted at the outer level.
4283
4284 < (?: (?(R) \d++ | [^<>]*+) | (?R)) * >
4285
4286 In this pattern, (?(R) is the start of a conditional subpattern, with
4287 two different alternatives for the recursive and non-recursive cases.
4288 The (?R) item is the actual recursive call.
4289
4290
4291 SUBPATTERNS AS SUBROUTINES
4292
4293 If the syntax for a recursive subpattern reference (either by number or
4294 by name) is used outside the parentheses to which it refers, it oper-
4295 ates like a subroutine in a programming language. The "called" subpat-
4296 tern may be defined before or after the reference. An earlier example
4297 pointed out that the pattern
4298
4299 (sens|respons)e and \1ibility
4300
4301 matches "sense and sensibility" and "response and responsibility", but
4302 not "sense and responsibility". If instead the pattern
4303
4304 (sens|respons)e and (?1)ibility
4305
4306 is used, it does match "sense and responsibility" as well as the other
4307 two strings. Another example is given in the discussion of DEFINE
4308 above.
4309
4310 Like recursive subpatterns, a "subroutine" call is always treated as an
4311 atomic group. That is, once it has matched some of the subject string,
4312 it is never re-entered, even if it contains untried alternatives and
4313 there is a subsequent matching failure.
4314
4315 When a subpattern is used as a subroutine, processing options such as
4316 case-independence are fixed when the subpattern is defined. They cannot
4317 be changed for different calls. For example, consider this pattern:
4318
4319 (abc)(?i:(?1))
4320
4321 It matches "abcabc". It does not match "abcABC" because the change of
4322 processing option does not affect the called subpattern.
4323
4324
4325 CALLOUTS
4326
4327 Perl has a feature whereby using the sequence (?{...}) causes arbitrary
4328 Perl code to be obeyed in the middle of matching a regular expression.
4329 This makes it possible, amongst other things, to extract different sub-
4330 strings that match the same pair of parentheses when there is a repeti-
4331 tion.
4332
4333 PCRE provides a similar feature, but of course it cannot obey arbitrary
4334 Perl code. The feature is called "callout". The caller of PCRE provides
4335 an external function by putting its entry point in the global variable
4336 pcre_callout. By default, this variable contains NULL, which disables
4337 all calling out.
4338
4339 Within a regular expression, (?C) indicates the points at which the
4340 external function is to be called. If you want to identify different
4341 callout points, you can put a number less than 256 after the letter C.
4342 The default value is zero. For example, this pattern has two callout
4343 points:
4344
4345 (?C1)abc(?C2)def
4346
4347 If the PCRE_AUTO_CALLOUT flag is passed to pcre_compile(), callouts are
4348 automatically installed before each item in the pattern. They are all
4349 numbered 255.
4350
4351 During matching, when PCRE reaches a callout point (and pcre_callout is
4352 set), the external function is called. It is provided with the number
4353 of the callout, the position in the pattern, and, optionally, one item
4354 of data originally supplied by the caller of pcre_exec(). The callout
4355 function may cause matching to proceed, to backtrack, or to fail alto-
4356 gether. A complete description of the interface to the callout function
4357 is given in the pcrecallout documentation.
4358
4359
4360 SEE ALSO
4361
4362 pcreapi(3), pcrecallout(3), pcrematching(3), pcre(3).
4363
4364
4365 AUTHOR
4366
4367 Philip Hazel
4368 University Computing Service
4369 Cambridge CB2 3QH, England.
4370
4371
4372 REVISION
4373
4374 Last updated: 06 March 2007
4375 Copyright (c) 1997-2007 University of Cambridge.
4376 ------------------------------------------------------------------------------
4377
4378
4379 PCREPARTIAL(3) PCREPARTIAL(3)
4380
4381
4382 NAME
4383 PCRE - Perl-compatible regular expressions
4384
4385
4386 PARTIAL MATCHING IN PCRE
4387
4388 In normal use of PCRE, if the subject string that is passed to
4389 pcre_exec() or pcre_dfa_exec() matches as far as it goes, but is too
4390 short to match the entire pattern, PCRE_ERROR_NOMATCH is returned.
4391 There are circumstances where it might be helpful to distinguish this
4392 case from other cases in which there is no match.
4393
4394 Consider, for example, an application where a human is required to type
4395 in data for a field with specific formatting requirements. An example
4396 might be a date in the form ddmmmyy, defined by this pattern:
4397
4398 ^\d?\d(jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)\d\d$
4399
4400 If the application sees the user's keystrokes one by one, and can check
4401 that what has been typed so far is potentially valid, it is able to
4402 raise an error as soon as a mistake is made, possibly beeping and not
4403 reflecting the character that has been typed. This immediate feedback
4404 is likely to be a better user interface than a check that is delayed
4405 until the entire string has been entered.
4406
4407 PCRE supports the concept of partial matching by means of the PCRE_PAR-
4408 TIAL option, which can be set when calling pcre_exec() or
4409 pcre_dfa_exec(). When this flag is set for pcre_exec(), the return code
4410 PCRE_ERROR_NOMATCH is converted into PCRE_ERROR_PARTIAL if at any time
4411 during the matching process the last part of the subject string matched
4412 part of the pattern. Unfortunately, for non-anchored matching, it is
4413 not possible to obtain the position of the start of the partial match.
4414 No captured data is set when PCRE_ERROR_PARTIAL is returned.
4415
4416 When PCRE_PARTIAL is set for pcre_dfa_exec(), the return code
4417 PCRE_ERROR_NOMATCH is converted into PCRE_ERROR_PARTIAL if the end of
4418 the subject is reached, there have been no complete matches, but there
4419 is still at least one matching possibility. The portion of the string
4420 that provided the partial match is set as the first matching string.
4421
4422 Using PCRE_PARTIAL disables one of PCRE's optimizations. PCRE remembers
4423 the last literal byte in a pattern, and abandons matching immediately
4424 if such a byte is not present in the subject string. This optimization
4425 cannot be used for a subject string that might match only partially.
4426
4427
4428 RESTRICTED PATTERNS FOR PCRE_PARTIAL
4429
4430 Because of the way certain internal optimizations are implemented in
4431 the pcre_exec() function, the PCRE_PARTIAL option cannot be used with
4432 all patterns. These restrictions do not apply when pcre_dfa_exec() is
4433 used. For pcre_exec(), repeated single characters such as
4434
4435 a{2,4}
4436
4437 and repeated single metasequences such as
4438
4439 \d+
4440
4441 are not permitted if the maximum number of occurrences is greater than
4442 one. Optional items such as \d? (where the maximum is one) are permit-
4443 ted. Quantifiers with any values are permitted after parentheses, so
4444 the invalid examples above can be coded thus:
4445
4446 (a){2,4}
4447 (\d)+
4448
4449 These constructions run more slowly, but for the kinds of application
4450 that are envisaged for this facility, this is not felt to be a major
4451 restriction.
4452
4453 If PCRE_PARTIAL is set for a pattern that does not conform to the
4454 restrictions, pcre_exec() returns the error code PCRE_ERROR_BADPARTIAL
4455 (-13).
4456
4457
4458 EXAMPLE OF PARTIAL MATCHING USING PCRETEST
4459
4460 If the escape sequence \P is present in a pcretest data line, the
4461 PCRE_PARTIAL flag is used for the match. Here is a run of pcretest that
4462 uses the date example quoted above:
4463
4464 re> /^\d?\d(jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)\d\d$/
4465 data> 25jun04\P
4466 0: 25jun04
4467 1: jun
4468 data> 25dec3\P
4469 Partial match
4470 data> 3ju\P
4471 Partial match
4472 data> 3juj\P
4473 No match
4474 data> j\P
4475 No match
4476
4477 The first data string is matched completely, so pcretest shows the
4478 matched substrings. The remaining four strings do not match the com-
4479 plete pattern, but the first two are partial matches. The same test,
4480 using pcre_dfa_exec() matching (by means of the \D escape sequence),
4481 produces the following output:
4482
4483 re> /^\d?\d(jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)\d\d$/
4484 data> 25jun04\P\D
4485 0: 25jun04
4486 data> 23dec3\P\D
4487 Partial match: 23dec3
4488 data> 3ju\P\D
4489 Partial match: 3ju
4490 data> 3juj\P\D
4491 No match
4492 data> j\P\D
4493 No match
4494
4495 Notice that in this case the portion of the string that was matched is
4496 made available.
4497
4498
4499 MULTI-SEGMENT MATCHING WITH pcre_dfa_exec()
4500
4501 When a partial match has been found using pcre_dfa_exec(), it is possi-
4502 ble to continue the match by providing additional subject data and
4503 calling pcre_dfa_exec() again with the same compiled regular expres-
4504 sion, this time setting the PCRE_DFA_RESTART option. You must also pass
4505 the same working space as before, because this is where details of the
4506 previous partial match are stored. Here is an example using pcretest,
4507 using the \R escape sequence to set the PCRE_DFA_RESTART option (\P and
4508 \D are as above):
4509
4510 re> /^\d?\d(jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)\d\d$/
4511 data> 23ja\P\D
4512 Partial match: 23ja
4513 data> n05\R\D
4514 0: n05
4515
4516 The first call has "23ja" as the subject, and requests partial match-
4517 ing; the second call has "n05" as the subject for the continued
4518 (restarted) match. Notice that when the match is complete, only the
4519 last part is shown; PCRE does not retain the previously partially-
4520 matched string. It is up to the calling program to do that if it needs
4521 to.
4522
4523 You can set PCRE_PARTIAL with PCRE_DFA_RESTART to continue partial
4524 matching over multiple segments. This facility can be used to pass very
4525 long subject strings to pcre_dfa_exec(). However, some care is needed
4526 for certain types of pattern.
4527
4528 1. If the pattern contains tests for the beginning or end of a line,
4529 you need to pass the PCRE_NOTBOL or PCRE_NOTEOL options, as appropri-
4530 ate, when the subject string for any call does not contain the begin-
4531 ning or end of a line.
4532
4533 2. If the pattern contains backward assertions (including \b or \B),
4534 you need to arrange for some overlap in the subject strings to allow
4535 for this. For example, you could pass the subject in chunks that are
4536 500 bytes long, but in a buffer of 700 bytes, with the starting offset
4537 set to 200 and the previous 200 bytes at the start of the buffer.
4538
4539 3. Matching a subject string that is split into multiple segments does
4540 not always produce exactly the same result as matching over one single
4541 long string. The difference arises when there are multiple matching
4542 possibilities, because a partial match result is given only when there
4543 are no completed matches in a call to pcre_dfa_exec(). This means that
4544 as soon as the shortest match has been found, continuation to a new
4545 subject segment is no longer possible. Consider this pcretest example:
4546
4547 re> /dog(sbody)?/
4548 data> do\P\D
4549 Partial match: do
4550 data> gsb\R\P\D
4551 0: g
4552 data> dogsbody\D
4553 0: dogsbody
4554 1: dog
4555
4556 The pattern matches the words "dog" or "dogsbody". When the subject is
4557 presented in several parts ("do" and "gsb" being the first two) the
4558 match stops when "dog" has been found, and it is not possible to con-
4559 tinue. On the other hand, if "dogsbody" is presented as a single
4560 string, both matches are found.
4561
4562 Because of this phenomenon, it does not usually make sense to end a
4563 pattern that is going to be matched in this way with a variable repeat.
4564
4565 4. Patterns that contain alternatives at the top level which do not all
4566 start with the same pattern item may not work as expected. For example,
4567 consider this pattern:
4568
4569 1234|3789
4570
4571 If the first part of the subject is "ABC123", a partial match of the
4572 first alternative is found at offset 3. There is no partial match for
4573 the second alternative, because such a match does not start at the same
4574 point in the subject string. Attempting to continue with the string
4575 "789" does not yield a match because only those alternatives that match
4576 at one point in the subject are remembered. The problem arises because
4577 the start of the second alternative matches within the first alterna-
4578 tive. There is no problem with anchored patterns or patterns such as:
4579
4580 1234|ABCD
4581
4582 where no string can be a partial match for both alternatives.
4583
4584
4585 AUTHOR
4586
4587 Philip Hazel
4588 University Computing Service
4589 Cambridge CB2 3QH, England.
4590
4591
4592 REVISION
4593
4594 Last updated: 06 March 2007
4595 Copyright (c) 1997-2007 University of Cambridge.
4596 ------------------------------------------------------------------------------
4597
4598
4599 PCREPRECOMPILE(3) PCREPRECOMPILE(3)
4600
4601
4602 NAME
4603 PCRE - Perl-compatible regular expressions
4604
4605
4606 SAVING AND RE-USING PRECOMPILED PCRE PATTERNS
4607
4608 If you are running an application that uses a large number of regular
4609 expression patterns, it may be useful to store them in a precompiled
4610 form instead of having to compile them every time the application is
4611 run. If you are not using any private character tables (see the
4612 pcre_maketables() documentation), this is relatively straightforward.
4613 If you are using private tables, it is a little bit more complicated.
4614
4615 If you save compiled patterns to a file, you can copy them to a differ-
4616 ent host and run them there. This works even if the new host has the
4617 opposite endianness to the one on which the patterns were compiled.
4618 There may be a small performance penalty, but it should be insignifi-
4619 cant. However, compiling regular expressions with one version of PCRE
4620 for use with a different version is not guaranteed to work and may
4621 cause crashes.
4622
4623
4624 SAVING A COMPILED PATTERN
4625 The value returned by pcre_compile() points to a single block of memory
4626 that holds the compiled pattern and associated data. You can find the
4627 length of this block in bytes by calling pcre_fullinfo() with an argu-
4628 ment of PCRE_INFO_SIZE. You can then save the data in any appropriate
4629 manner. Here is sample code that compiles a pattern and writes it to a
4630 file. It assumes that the variable fd refers to a file that is open for
4631 output:
4632
4633 int erroroffset, rc, size;
4634 char *error;
4635 pcre *re;
4636
4637 re = pcre_compile("my pattern", 0, &error, &erroroffset, NULL);
4638 if (re == NULL) { ... handle errors ... }
4639 rc = pcre_fullinfo(re, NULL, PCRE_INFO_SIZE, &size);
4640 if (rc < 0) { ... handle errors ... }
4641 rc = fwrite(re, 1, size, fd);
4642 if (rc != size) { ... handle errors ... }
4643
4644 In this example, the bytes that comprise the compiled pattern are
4645 copied exactly. Note that this is binary data that may contain any of
4646 the 256 possible byte values. On systems that make a distinction
4647 between binary and non-binary data, be sure that the file is opened for
4648 binary output.
4649
4650 If you want to write more than one pattern to a file, you will have to
4651 devise a way of separating them. For binary data, preceding each pat-
4652 tern with its length is probably the most straightforward approach.
4653 Another possibility is to write out the data in hexadecimal instead of
4654 binary, one pattern to a line.
4655
4656 Saving compiled patterns in a file is only one possible way of storing
4657 them for later use. They could equally well be saved in a database, or
4658 in the memory of some daemon process that passes them via sockets to
4659 the processes that want them.
4660
4661 If the pattern has been studied, it is also possible to save the study
4662 data in a similar way to the compiled pattern itself. When studying
4663 generates additional information, pcre_study() returns a pointer to a
4664 pcre_extra data block. Its format is defined in the section on matching
4665 a pattern in the pcreapi documentation. The study_data field points to
4666 the binary study data, and this is what you must save (not the
4667 pcre_extra block itself). The length of the study data can be obtained
4668 by calling pcre_fullinfo() with an argument of PCRE_INFO_STUDYSIZE.
4669 Remember to check that pcre_study() did return a non-NULL value before
4670 trying to save the study data.
4671
4672
4673 RE-USING A PRECOMPILED PATTERN
4674
4675 Re-using a precompiled pattern is straightforward. Having reloaded it
4676 into main memory, you pass its pointer to pcre_exec() or
4677 pcre_dfa_exec() in the usual way. This should work even on another
4678 host, and even if that host has the opposite endianness to the one
4679 where the pattern was compiled.
4680
4681 However, if you passed a pointer to custom character tables when the
4682 pattern was compiled (the tableptr argument of pcre_compile()), you
4683 must now pass a similar pointer to pcre_exec() or pcre_dfa_exec(),
4684 because the value saved with the compiled pattern will obviously be
4685 nonsense. A field in a pcre_extra() block is used to pass this data, as
4686 described in the section on matching a pattern in the pcreapi documen-
4687 tation.
4688
4689 If you did not provide custom character tables when the pattern was
4690 compiled, the pointer in the compiled pattern is NULL, which causes
4691 pcre_exec() to use PCRE's internal tables. Thus, you do not need to
4692 take any special action at run time in this case.
4693
4694 If you saved study data with the compiled pattern, you need to create
4695 your own pcre_extra data block and set the study_data field to point to
4696 the reloaded study data. You must also set the PCRE_EXTRA_STUDY_DATA
4697 bit in the flags field to indicate that study data is present. Then
4698 pass the pcre_extra block to pcre_exec() or pcre_dfa_exec() in the
4699 usual way.
4700
4701
4702 COMPATIBILITY WITH DIFFERENT PCRE RELEASES
4703
4704 The layout of the control block that is at the start of the data that
4705 makes up a compiled pattern was changed for release 5.0. If you have
4706 any saved patterns that were compiled with previous releases (not a
4707 facility that was previously advertised), you will have to recompile
4708 them for release 5.0 and above.
4709
4710 If you have any saved patterns in UTF-8 mode that use \p or \P that
4711 were compiled with any release up to and including 6.4, you will have
4712 to recompile them for release 6.5 and above.
4713
4714 All saved patterns from earlier releases must be recompiled for release
4715 7.0 or higher, because there was an internal reorganization at that
4716 release.
4717
4718
4719 AUTHOR
4720
4721 Philip Hazel
4722 University Computing Service
4723 Cambridge CB2 3QH, England.
4724
4725
4726 REVISION
4727
4728 Last updated: 24 April 2007
4729 Copyright (c) 1997-2007 University of Cambridge.
4730 ------------------------------------------------------------------------------
4731
4732
4733 PCREPERFORM(3) PCREPERFORM(3)
4734
4735
4736 NAME
4737 PCRE - Perl-compatible regular expressions
4738
4739
4740 PCRE PERFORMANCE
4741
4742 Two aspects of performance are discussed below: memory usage and pro-
4743 cessing time. The way you express your pattern as a regular expression
4744 can affect both of them.
4745
4746
4747 MEMORY USAGE
4748
4749 Patterns are compiled by PCRE into a reasonably efficient byte code, so
4750 that most simple patterns do not use much memory. However, there is one
4751 case where memory usage can be unexpectedly large. When a parenthesized
4752 subpattern has a quantifier with a minimum greater than 1 and/or a lim-
4753 ited maximum, the whole subpattern is repeated in the compiled code.
4754 For example, the pattern
4755
4756 (abc|def){2,4}
4757
4758 is compiled as if it were
4759
4760 (abc|def)(abc|def)((abc|def)(abc|def)?)?
4761
4762 (Technical aside: It is done this way so that backtrack points within
4763 each of the repetitions can be independently maintained.)
4764
4765 For regular expressions whose quantifiers use only small numbers, this
4766 is not usually a problem. However, if the numbers are large, and par-
4767 ticularly if such repetitions are nested, the memory usage can become
4768 an embarrassment. For example, the very simple pattern
4769
4770 ((ab){1,1000}c){1,3}
4771
4772 uses 51K bytes when compiled. When PCRE is compiled with its default
4773 internal pointer size of two bytes, the size limit on a compiled pat-
4774 tern is 64K, and this is reached with the above pattern if the outer
4775 repetition is increased from 3 to 4. PCRE can be compiled to use larger
4776 internal pointers and thus handle larger compiled patterns, but it is
4777 better to try to rewrite your pattern to use less memory if you can.
4778
4779 One way of reducing the memory usage for such patterns is to make use
4780 of PCRE's "subroutine" facility. Re-writing the above pattern as
4781
4782 ((ab)(?2){0,999}c)(?1){0,2}
4783
4784 reduces the memory requirements to 18K, and indeed it remains under 20K
4785 even with the outer repetition increased to 100. However, this pattern
4786 is not exactly equivalent, because the "subroutine" calls are treated
4787 as atomic groups into which there can be no backtracking if there is a
4788 subsequent matching failure. Therefore, PCRE cannot do this kind of
4789 rewriting automatically. Furthermore, there is a noticeable loss of
4790 speed when executing the modified pattern. Nevertheless, if the atomic
4791 grouping is not a problem and the loss of speed is acceptable, this
4792 kind of rewriting will allow you to process patterns that PCRE cannot
4793 otherwise handle.
4794
4795
4796 PROCESSING TIME
4797
4798 Certain items in regular expression patterns are processed more effi-
4799 ciently than others. It is more efficient to use a character class like
4800 [aeiou] than a set of single-character alternatives such as
4801 (a|e|i|o|u). In general, the simplest construction that provides the
4802 required behaviour is usually the most efficient. Jeffrey Friedl's book
4803 contains a lot of useful general discussion about optimizing regular
4804 expressions for efficient performance. This document contains a few
4805 observations about PCRE.
4806
4807 Using Unicode character properties (the \p, \P, and \X escapes) is
4808 slow, because PCRE has to scan a structure that contains data for over
4809 fifteen thousand characters whenever it needs a character's property.
4810 If you can find an alternative pattern that does not use character
4811 properties, it will probably be faster.
4812
4813 When a pattern begins with .* not in parentheses, or in parentheses
4814 that are not the subject of a backreference, and the PCRE_DOTALL option
4815 is set, the pattern is implicitly anchored by PCRE, since it can match
4816 only at the start of a subject string. However, if PCRE_DOTALL is not
4817 set, PCRE cannot make this optimization, because the . metacharacter
4818 does not then match a newline, and if the subject string contains new-
4819 lines, the pattern may match from the character immediately following
4820 one of them instead of from the very start. For example, the pattern
4821
4822 .*second
4823
4824 matches the subject "first\nand second" (where \n stands for a newline
4825 character), with the match starting at the seventh character. In order
4826 to do this, PCRE has to retry the match starting after every newline in
4827 the subject.
4828
4829 If you are using such a pattern with subject strings that do not con-
4830 tain newlines, the best performance is obtained by setting PCRE_DOTALL,
4831 or starting the pattern with ^.* or ^.*? to indicate explicit anchor-
4832 ing. That saves PCRE from having to scan along the subject looking for
4833 a newline to restart at.
4834
4835 Beware of patterns that contain nested indefinite repeats. These can
4836 take a long time to run when applied to a string that does not match.
4837 Consider the pattern fragment
4838
4839 ^(a+)*
4840
4841 This can match "aaaa" in 16 different ways, and this number increases
4842 very rapidly as the string gets longer. (The * repeat can match 0, 1,
4843 2, 3, or 4 times, and for each of those cases other than 0 or 4, the +
4844 repeats can match different numbers of times.) When the remainder of
4845 the pattern is such that the entire match is going to fail, PCRE has in
4846 principle to try every possible variation, and this can take an
4847 extremely long time, even for relatively short strings.
4848
4849 An optimization catches some of the more simple cases such as
4850
4851 (a+)*b
4852
4853 where a literal character follows. Before embarking on the standard
4854 matching procedure, PCRE checks that there is a "b" later in the sub-
4855 ject string, and if there is not, it fails the match immediately. How-
4856 ever, when there is no following literal this optimization cannot be
4857 used. You can see the difference by comparing the behaviour of
4858
4859 (a+)*\d
4860
4861 with the pattern above. The former gives a failure almost instantly
4862 when applied to a whole line of "a" characters, whereas the latter
4863 takes an appreciable time with strings longer than about 20 characters.
4864
4865 In many cases, the solution to this kind of performance issue is to use
4866 an atomic group or a possessive quantifier.
4867
4868
4869 AUTHOR
4870
4871 Philip Hazel
4872 University Computing Service
4873 Cambridge CB2 3QH, England.
4874
4875
4876 REVISION
4877
4878 Last updated: 06 March 2007
4879 Copyright (c) 1997-2007 University of Cambridge.
4880 ------------------------------------------------------------------------------
4881
4882
4883 PCREPOSIX(3) PCREPOSIX(3)
4884
4885
4886 NAME
4887 PCRE - Perl-compatible regular expressions.
4888
4889
4890 SYNOPSIS OF POSIX API
4891
4892 #include <pcreposix.h>
4893
4894 int regcomp(regex_t *preg, const char *pattern,
4895 int cflags);
4896
4897 int regexec(regex_t *preg, const char *string,
4898 size_t nmatch, regmatch_t pmatch[], int eflags);
4899
4900 size_t regerror(int errcode, const regex_t *preg,
4901 char *errbuf, size_t errbuf_size);
4902
4903 void regfree(regex_t *preg);
4904
4905
4906 DESCRIPTION
4907
4908 This set of functions provides a POSIX-style API to the PCRE regular
4909 expression package. See the pcreapi documentation for a description of
4910 PCRE's native API, which contains much additional functionality.
4911
4912 The functions described here are just wrapper functions that ultimately
4913 call the PCRE native API. Their prototypes are defined in the
4914 pcreposix.h header file, and on Unix systems the library itself is
4915 called pcreposix.a, so can be accessed by adding -lpcreposix to the
4916 command for linking an application that uses them. Because the POSIX
4917 functions call the native ones, it is also necessary to add -lpcre.
4918
4919 I have implemented only those option bits that can be reasonably mapped
4920 to PCRE native options. In addition, the option REG_EXTENDED is defined
4921 with the value zero. This has no effect, but since programs that are
4922 written to the POSIX interface often use it, this makes it easier to
4923 slot in PCRE as a replacement library. Other POSIX options are not even
4924 defined.
4925
4926 When PCRE is called via these functions, it is only the API that is
4927 POSIX-like in style. The syntax and semantics of the regular expres-
4928 sions themselves are still those of Perl, subject to the setting of
4929 various PCRE options, as described below. "POSIX-like in style" means
4930 that the API approximates to the POSIX definition; it is not fully
4931 POSIX-compatible, and in multi-byte encoding domains it is probably
4932 even less compatible.
4933
4934 The header for these functions is supplied as pcreposix.h to avoid any
4935 potential clash with other POSIX libraries. It can, of course, be
4936 renamed or aliased as regex.h, which is the "correct" name. It provides
4937 two structure types, regex_t for compiled internal forms, and reg-
4938 match_t for returning captured substrings. It also defines some con-
4939 stants whose names start with "REG_"; these are used for setting
4940 options and identifying error codes.
4941
4942
4943 COMPILING A PATTERN
4944
4945 The function regcomp() is called to compile a pattern into an internal
4946 form. The pattern is a C string terminated by a binary zero, and is
4947 passed in the argument pattern. The preg argument is a pointer to a
4948 regex_t structure that is used as a base for storing information about
4949 the compiled regular expression.
4950
4951 The argument cflags is either zero, or contains one or more of the bits
4952 defined by the following macros:
4953
4954 REG_DOTALL
4955
4956 The PCRE_DOTALL option is set when the regular expression is passed for
4957 compilation to the native function. Note that REG_DOTALL is not part of
4958 the POSIX standard.
4959
4960 REG_ICASE
4961
4962 The PCRE_CASELESS option is set when the regular expression is passed
4963 for compilation to the native function.
4964
4965 REG_NEWLINE
4966
4967 The PCRE_MULTILINE option is set when the regular expression is passed
4968 for compilation to the native function. Note that this does not mimic
4969 the defined POSIX behaviour for REG_NEWLINE (see the following sec-
4970 tion).
4971
4972 REG_NOSUB
4973
4974 The PCRE_NO_AUTO_CAPTURE option is set when the regular expression is
4975 passed for compilation to the native function. In addition, when a pat-
4976 tern that is compiled with this flag is passed to regexec() for match-
4977 ing, the nmatch and pmatch arguments are ignored, and no captured
4978 strings are returned.
4979
4980 REG_UTF8
4981
4982 The PCRE_UTF8 option is set when the regular expression is passed for
4983 compilation to the native function. This causes the pattern itself and
4984 all data strings used for matching it to be treated as UTF-8 strings.
4985 Note that REG_UTF8 is not part of the POSIX standard.
4986
4987 In the absence of these flags, no options are passed to the native
4988 function. This means the the regex is compiled with PCRE default
4989 semantics. In particular, the way it handles newline characters in the
4990 subject string is the Perl way, not the POSIX way. Note that setting
4991 PCRE_MULTILINE has only some of the effects specified for REG_NEWLINE.
4992 It does not affect the way newlines are matched by . (they aren't) or
4993 by a negative class such as [^a] (they are).
4994
4995 The yield of regcomp() is zero on success, and non-zero otherwise. The
4996 preg structure is filled in on success, and one member of the structure
4997 is public: re_nsub contains the number of capturing subpatterns in the
4998 regular expression. Various error codes are defined in the header file.
4999
5000
5001 MATCHING NEWLINE CHARACTERS
5002
5003 This area is not simple, because POSIX and Perl take different views of
5004 things. It is not possible to get PCRE to obey POSIX semantics, but
5005 then PCRE was never intended to be a POSIX engine. The following table
5006 lists the different possibilities for matching newline characters in
5007 PCRE:
5008
5009 Default Change with
5010
5011 . matches newline no PCRE_DOTALL
5012 newline matches [^a] yes not changeable
5013 $ matches \n at end yes PCRE_DOLLARENDONLY
5014 $ matches \n in middle no PCRE_MULTILINE
5015 ^ matches \n in middle no PCRE_MULTILINE
5016
5017 This is the equivalent table for POSIX:
5018
5019 Default Change with
5020
5021 . matches newline yes REG_NEWLINE
5022 newline matches [^a] yes REG_NEWLINE
5023 $ matches \n at end no REG_NEWLINE
5024 $ matches \n in middle no REG_NEWLINE
5025 ^ matches \n in middle no REG_NEWLINE
5026
5027 PCRE's behaviour is the same as Perl's, except that there is no equiva-
5028 lent for PCRE_DOLLAR_ENDONLY in Perl. In both PCRE and Perl, there is
5029 no way to stop newline from matching [^a].
5030
5031 The default POSIX newline handling can be obtained by setting
5032 PCRE_DOTALL and PCRE_DOLLAR_ENDONLY, but there is no way to make PCRE
5033 behave exactly as for the REG_NEWLINE action.
5034
5035
5036 MATCHING A PATTERN
5037
5038 The function regexec() is called to match a compiled pattern preg
5039 against a given string, which is terminated by a zero byte, subject to
5040 the options in eflags. These can be:
5041
5042 REG_NOTBOL
5043
5044 The PCRE_NOTBOL option is set when calling the underlying PCRE matching
5045 function.
5046
5047 REG_NOTEOL
5048
5049 The PCRE_NOTEOL option is set when calling the underlying PCRE matching
5050 function.
5051
5052 If the pattern was compiled with the REG_NOSUB flag, no data about any
5053 matched strings is returned. The nmatch and pmatch arguments of
5054 regexec() are ignored.
5055
5056 Otherwise,the portion of the string that was matched, and also any cap-
5057 tured substrings, are returned via the pmatch argument, which points to
5058 an array of nmatch structures of type regmatch_t, containing the mem-
5059 bers rm_so and rm_eo. These contain the offset to the first character
5060 of each substring and the offset to the first character after the end
5061 of each substring, respectively. The 0th element of the vector relates
5062 to the entire portion of string that was matched; subsequent elements
5063 relate to the capturing subpatterns of the regular expression. Unused
5064 entries in the array have both structure members set to -1.
5065
5066 A successful match yields a zero return; various error codes are
5067 defined in the header file, of which REG_NOMATCH is the "expected"
5068 failure code.
5069
5070
5071 ERROR MESSAGES
5072
5073 The regerror() function maps a non-zero errorcode from either regcomp()
5074 or regexec() to a printable message. If preg is not NULL, the error
5075 should have arisen from the use of that structure. A message terminated
5076 by a binary zero is placed in errbuf. The length of the message,
5077 including the zero, is limited to errbuf_size. The yield of the func-
5078 tion is the size of buffer needed to hold the whole message.
5079
5080
5081 MEMORY USAGE
5082
5083 Compiling a regular expression causes memory to be allocated and asso-
5084 ciated with the preg structure. The function regfree() frees all such
5085 memory, after which preg may no longer be used as a compiled expres-
5086 sion.
5087
5088
5089 AUTHOR
5090
5091 Philip Hazel
5092 University Computing Service
5093 Cambridge CB2 3QH, England.
5094
5095
5096 REVISION
5097
5098 Last updated: 06 March 2007
5099 Copyright (c) 1997-2007 University of Cambridge.
5100 ------------------------------------------------------------------------------
5101
5102
5103 PCRECPP(3) PCRECPP(3)
5104
5105
5106 NAME
5107 PCRE - Perl-compatible regular expressions.
5108
5109
5110 SYNOPSIS OF C++ WRAPPER
5111
5112 #include <pcrecpp.h>
5113
5114
5115 DESCRIPTION
5116
5117 The C++ wrapper for PCRE was provided by Google Inc. Some additional
5118 functionality was added by Giuseppe Maxia. This brief man page was con-
5119 structed from the notes in the pcrecpp.h file, which should be con-
5120 sulted for further details.
5121
5122
5123 MATCHING INTERFACE
5124
5125 The "FullMatch" operation checks that supplied text matches a supplied
5126 pattern exactly. If pointer arguments are supplied, it copies matched
5127 sub-strings that match sub-patterns into them.
5128
5129 Example: successful match
5130 pcrecpp::RE re("h.*o");
5131 re.FullMatch("hello");
5132
5133 Example: unsuccessful match (requires full match):
5134 pcrecpp::RE re("e");
5135 !re.FullMatch("hello");
5136
5137 Example: creating a temporary RE object:
5138 pcrecpp::RE("h.*o").FullMatch("hello");
5139
5140 You can pass in a "const char*" or a "string" for "text". The examples
5141 below tend to use a const char*. You can, as in the different examples
5142 above, store the RE object explicitly in a variable or use a temporary
5143 RE object. The examples below use one mode or the other arbitrarily.
5144 Either could correctly be used for any of these examples.
5145
5146 You must supply extra pointer arguments to extract matched subpieces.
5147
5148 Example: extracts "ruby" into "s" and 1234 into "i"
5149 int i;
5150 string s;
5151 pcrecpp::RE re("(\\w+):(\\d+)");
5152 re.FullMatch("ruby:1234", &s, &i);
5153
5154 Example: does not try to extract any extra sub-patterns
5155 re.FullMatch("ruby:1234", &s);
5156
5157 Example: does not try to extract into NULL
5158 re.FullMatch("ruby:1234", NULL, &i);
5159
5160 Example: integer overflow causes failure
5161 !re.FullMatch("ruby:1234567891234", NULL, &i);
5162
5163 Example: fails because there aren't enough sub-patterns:
5164 !pcrecpp::RE("\\w+:\\d+").FullMatch("ruby:1234", &s);
5165
5166 Example: fails because string cannot be stored in integer
5167 !pcrecpp::RE("(.*)").FullMatch("ruby", &i);
5168
5169 The provided pointer arguments can be pointers to any scalar numeric
5170 type, or one of:
5171
5172 string (matched piece is copied to string)
5173 StringPiece (StringPiece is mutated to point to matched piece)
5174 T (where "bool T::ParseFrom(const char*, int)" exists)
5175 NULL (the corresponding matched sub-pattern is not copied)
5176
5177 The function returns true iff all of the following conditions are sat-
5178 isfied:
5179
5180 a. "text" matches "pattern" exactly;
5181
5182 b. The number of matched sub-patterns is >= number of supplied
5183 pointers;
5184
5185 c. The "i"th argument has a suitable type for holding the
5186 string captured as the "i"th sub-pattern. If you pass in
5187 NULL for the "i"th argument, or pass fewer arguments than
5188 number of sub-patterns, "i"th captured sub-pattern is
5189 ignored.
5190
5191 CAVEAT: An optional sub-pattern that does not exist in the matched
5192 string is assigned the empty string. Therefore, the following will
5193 return false (because the empty string is not a valid number):
5194
5195 int number;
5196 pcrecpp::RE::FullMatch("abc", "[a-z]+(\\d+)?", &number);
5197
5198 The matching interface supports at most 16 arguments per call. If you
5199 need more, consider using the more general interface
5200 pcrecpp::RE::DoMatch. See pcrecpp.h for the signature for DoMatch.
5201
5202
5203 QUOTING METACHARACTERS
5204
5205 You can use the "QuoteMeta" operation to insert backslashes before all
5206 potentially meaningful characters in a string. The returned string,
5207 used as a regular expression, will exactly match the original string.
5208
5209 Example:
5210 string quoted = RE::QuoteMeta(unquoted);
5211
5212 Note that it's legal to escape a character even if it has no special
5213 meaning in a regular expression -- so this function does that. (This
5214 also makes it identical to the perl function of the same name; see
5215 "perldoc -f quotemeta".) For example, "1.5-2.0?" becomes
5216 "1\.5\-2\.0\?".
5217
5218
5219 PARTIAL MATCHES
5220
5221 You can use the "PartialMatch" operation when you want the pattern to
5222 match any substring of the text.
5223
5224 Example: simple search for a string:
5225 pcrecpp::RE("ell").PartialMatch("hello");
5226
5227 Example: find first number in a string:
5228 int number;
5229 pcrecpp::RE re("(\\d+)");
5230 re.PartialMatch("x*100 + 20", &number);
5231 assert(number == 100);
5232
5233
5234 UTF-8 AND THE MATCHING INTERFACE
5235
5236 By default, pattern and text are plain text, one byte per character.
5237 The UTF8 flag, passed to the constructor, causes both pattern and
5238 string to be treated as UTF-8 text, still a byte stream but potentially
5239 multiple bytes per character. In practice, the text is likelier to be
5240 UTF-8 than the pattern, but the match returned may depend on the UTF8
5241 flag, so always use it when matching UTF8 text. For example, "." will
5242 match one byte normally but with UTF8 set may match up to three bytes
5243 of a multi-byte character.
5244
5245 Example:
5246 pcrecpp::RE_Options options;
5247 options.set_utf8();
5248 pcrecpp::RE re(utf8_pattern, options);
5249 re.FullMatch(utf8_string);
5250
5251 Example: using the convenience function UTF8():
5252 pcrecpp::RE re(utf8_pattern, pcrecpp::UTF8());
5253 re.FullMatch(utf8_string);
5254
5255 NOTE: The UTF8 flag is ignored if pcre was not configured with the
5256 --enable-utf8 flag.
5257
5258
5259 PASSING MODIFIERS TO THE REGULAR EXPRESSION ENGINE
5260
5261 PCRE defines some modifiers to change the behavior of the regular
5262 expression engine. The C++ wrapper defines an auxiliary class,
5263 RE_Options, as a vehicle to pass such modifiers to a RE class. Cur-
5264 rently, the following modifiers are supported:
5265
5266 modifier description Perl corresponding
5267
5268 PCRE_CASELESS case insensitive match /i
5269 PCRE_MULTILINE multiple lines match /m
5270 PCRE_DOTALL dot matches newlines /s
5271 PCRE_DOLLAR_ENDONLY $ matches only at end N/A
5272 PCRE_EXTRA strict escape parsing N/A
5273 PCRE_EXTENDED ignore whitespaces /x
5274 PCRE_UTF8 handles UTF8 chars built-in
5275 PCRE_UNGREEDY reverses * and *? N/A
5276 PCRE_NO_AUTO_CAPTURE disables capturing parens N/A (*)
5277
5278 (*) Both Perl and PCRE allow non capturing parentheses by means of the
5279 "?:" modifier within the pattern itself. e.g. (?:ab|cd) does not cap-
5280 ture, while (ab|cd) does.
5281
5282 For a full account on how each modifier works, please check the PCRE
5283 API reference page.
5284
5285 For each modifier, there are two member functions whose name is made
5286 out of the modifier in lowercase, without the "PCRE_" prefix. For
5287 instance, PCRE_CASELESS is handled by
5288
5289 bool caseless()
5290
5291 which returns true if the modifier is set, and
5292
5293 RE_Options & set_caseless(bool)
5294
5295 which sets or unsets the modifier. Moreover, PCRE_EXTRA_MATCH_LIMIT can
5296 be accessed through the set_match_limit() and match_limit() member
5297 functions. Setting match_limit to a non-zero value will limit the exe-
5298 cution of pcre to keep it from doing bad things like blowing the stack
5299 or taking an eternity to return a result. A value of 5000 is good
5300 enough to stop stack blowup in a 2MB thread stack. Setting match_limit
5301 to zero disables match limiting. Alternatively, you can call
5302 match_limit_recursion() which uses PCRE_EXTRA_MATCH_LIMIT_RECURSION to
5303 limit how much PCRE recurses. match_limit() limits the number of
5304 matches PCRE does; match_limit_recursion() limits the depth of internal
5305 recursion, and therefore the amount of stack that is used.
5306
5307 Normally, to pass one or more modifiers to a RE class, you declare a
5308 RE_Options object, set the appropriate options, and pass this object to
5309 a RE constructor. Example:
5310
5311 RE_options opt;
5312 opt.set_caseless(true);
5313 if (RE("HELLO", opt).PartialMatch("hello world")) ...
5314
5315 RE_options has two constructors. The default constructor takes no argu-
5316 ments and creates a set of flags that are off by default. The optional
5317 parameter option_flags is to facilitate transfer of legacy code from C
5318 programs. This lets you do
5319
5320 RE(pattern,
5321 RE_Options(PCRE_CASELESS|PCRE_MULTILINE)).PartialMatch(str);
5322
5323 However, new code is better off doing
5324
5325 RE(pattern,
5326 RE_Options().set_caseless(true).set_multiline(true))
5327 .PartialMatch(str);
5328
5329 If you are going to pass one of the most used modifiers, there are some
5330 convenience functions that return a RE_Options class with the appropri-
5331 ate modifier already set: CASELESS(), UTF8(), MULTILINE(), DOTALL(),
5332 and EXTENDED().
5333
5334 If you need to set several options at once, and you don't want to go
5335 through the pains of declaring a RE_Options object and setting several
5336 options, there is a parallel method that give you such ability on the
5337 fly. You can concatenate several set_xxxxx() member functions, since
5338 each of them returns a reference to its class object. For example, to
5339 pass PCRE_CASELESS, PCRE_EXTENDED, and PCRE_MULTILINE to a RE with one
5340 statement, you may write:
5341
5342 RE(" ^ xyz \\s+ .* blah$",
5343 RE_Options()
5344 .set_caseless(true)
5345 .set_extended(true)
5346 .set_multiline(true)).PartialMatch(sometext);
5347
5348
5349 SCANNING TEXT INCREMENTALLY
5350
5351 The "Consume" operation may be useful if you want to repeatedly match
5352 regular expressions at the front of a string and skip over them as they
5353 match. This requires use of the "StringPiece" type, which represents a
5354 sub-range of a real string. Like RE, StringPiece is defined in the
5355 pcrecpp namespace.
5356
5357 Example: read lines of the form "var = value" from a string.
5358 string contents = ...; // Fill string somehow
5359 pcrecpp::StringPiece input(contents); // Wrap in a StringPiece
5360
5361 string var;
5362 int value;
5363 pcrecpp::RE re("(\\w+) = (\\d+)\n");
5364 while (re.Consume(&input, &var, &value)) {
5365 ...;
5366 }
5367
5368 Each successful call to "Consume" will set "var/value", and also
5369 advance "input" so it points past the matched text.
5370
5371 The "FindAndConsume" operation is similar to "Consume" but does not
5372 anchor your match at the beginning of the string. For example, you
5373 could extract all words from a string by repeatedly calling
5374
5375 pcrecpp::RE("(\\w+)").FindAndConsume(&input, &word)
5376
5377
5378 PARSING HEX/OCTAL/C-RADIX NUMBERS
5379
5380 By default, if you pass a pointer to a numeric value, the corresponding
5381 text is interpreted as a base-10 number. You can instead wrap the
5382 pointer with a call to one of the operators Hex(), Octal(), or CRadix()
5383 to interpret the text in another base. The CRadix operator interprets
5384 C-style "0" (base-8) and "0x" (base-16) prefixes, but defaults to
5385 base-10.
5386
5387 Example:
5388 int a, b, c, d;
5389 pcrecpp::RE re("(.*) (.*) (.*) (.*)");
5390 re.FullMatch("100 40 0100 0x40",
5391 pcrecpp::Octal(&a), pcrecpp::Hex(&b),
5392 pcrecpp::CRadix(&c), pcrecpp::CRadix(&d));
5393
5394 will leave 64 in a, b, c, and d.
5395
5396
5397 REPLACING PARTS OF STRINGS
5398
5399 You can replace the first match of "pattern" in "str" with "rewrite".
5400 Within "rewrite", backslash-escaped digits (\1 to \9) can be used to
5401 insert text matching corresponding parenthesized group from the pat-
5402 tern. \0 in "rewrite" refers to the entire matching text. For example:
5403
5404 string s = "yabba dabba doo";
5405 pcrecpp::RE("b+").Replace("d", &s);
5406
5407 will leave "s" containing "yada dabba doo". The result is true if the
5408 pattern matches and a replacement occurs, false otherwise.
5409
5410 GlobalReplace is like Replace except that it replaces all occurrences
5411 of the pattern in the string with the rewrite. Replacements are not
5412 subject to re-matching. For example:
5413
5414 string s = "yabba dabba doo";
5415 pcrecpp::RE("b+").GlobalReplace("d", &s);
5416
5417 will leave "s" containing "yada dada doo". It returns the number of
5418 replacements made.
5419
5420 Extract is like Replace, except that if the pattern matches, "rewrite"
5421 is copied into "out" (an additional argument) with substitutions. The
5422 non-matching portions of "text" are ignored. Returns true iff a match
5423 occurred and the extraction happened successfully; if no match occurs,
5424 the string is left unaffected.
5425
5426
5427 AUTHOR
5428
5429 The C++ wrapper was contributed by Google Inc.
5430 Copyright (c) 2007 Google Inc.
5431
5432
5433 REVISION
5434
5435 Last updated: 06 March 2007
5436 ------------------------------------------------------------------------------
5437
5438
5439 PCRESAMPLE(3) PCRESAMPLE(3)
5440
5441
5442 NAME
5443 PCRE - Perl-compatible regular expressions
5444
5445
5446 PCRE SAMPLE PROGRAM
5447
5448 A simple, complete demonstration program, to get you started with using
5449 PCRE, is supplied in the file pcredemo.c in the PCRE distribution.
5450
5451 The program compiles the regular expression that is its first argument,
5452 and matches it against the subject string in its second argument. No
5453 PCRE options are set, and default character tables are used. If match-
5454 ing succeeds, the program outputs the portion of the subject that
5455 matched, together with the contents of any captured substrings.
5456
5457 If the -g option is given on the command line, the program then goes on
5458 to check for further matches of the same regular expression in the same
5459 subject string. The logic is a little bit tricky because of the possi-
5460 bility of matching an empty string. Comments in the code explain what
5461 is going on.
5462
5463 If PCRE is installed in the standard include and library directories
5464 for your system, you should be able to compile the demonstration pro-
5465 gram using this command:
5466
5467 gcc -o pcredemo pcredemo.c -lpcre
5468
5469 If PCRE is installed elsewhere, you may need to add additional options
5470 to the command line. For example, on a Unix-like system that has PCRE
5471 installed in /usr/local, you can compile the demonstration program
5472 using a command like this:
5473
5474 gcc -o pcredemo -I/usr/local/include pcredemo.c \
5475 -L/usr/local/lib -lpcre
5476
5477 Once you have compiled the demonstration program, you can run simple
5478 tests like this:
5479
5480 ./pcredemo 'cat|dog' 'the cat sat on the mat'
5481 ./pcredemo -g 'cat|dog' 'the dog sat on the cat'
5482
5483 Note that there is a much more comprehensive test program, called
5484 pcretest, which supports many more facilities for testing regular
5485 expressions and the PCRE library. The pcredemo program is provided as a
5486 simple coding example.
5487
5488 On some operating systems (e.g. Solaris), when PCRE is not installed in
5489 the standard library directory, you may get an error like this when you
5490 try to run pcredemo:
5491
5492 ld.so.1: a.out: fatal: libpcre.so.0: open failed: No such file or
5493 directory
5494
5495 This is caused by the way shared library support works on those sys-
5496 tems. You need to add
5497
5498 -R/usr/local/lib
5499
5500 (for example) to the compile command to get round this problem.
5501
5502
5503 AUTHOR
5504
5505 Philip Hazel
5506 University Computing Service
5507 Cambridge CB2 3QH, England.
5508
5509
5510 REVISION
5511
5512 Last updated: 06 March 2007
5513 Copyright (c) 1997-2007 University of Cambridge.
5514 ------------------------------------------------------------------------------
5515 PCRESTACK(3) PCRESTACK(3)
5516
5517
5518 NAME
5519 PCRE - Perl-compatible regular expressions
5520
5521
5522 PCRE DISCUSSION OF STACK USAGE
5523
5524 When you call pcre_exec(), it makes use of an internal function called
5525 match(). This calls itself recursively at branch points in the pattern,
5526 in order to remember the state of the match so that it can back up and
5527 try a different alternative if the first one fails. As matching pro-
5528 ceeds deeper and deeper into the tree of possibilities, the recursion
5529 depth increases.
5530
5531 Not all calls of match() increase the recursion depth; for an item such
5532 as a* it may be called several times at the same level, after matching
5533 different numbers of a's. Furthermore, in a number of cases where the
5534 result of the recursive call would immediately be passed back as the
5535 result of the current call (a "tail recursion"), the function is just
5536 restarted instead.
5537
5538 The pcre_dfa_exec() function operates in an entirely different way, and
5539 hardly uses recursion at all. The limit on its complexity is the amount
5540 of workspace it is given. The comments that follow do NOT apply to
5541 pcre_dfa_exec(); they are relevant only for pcre_exec().
5542
5543 You can set limits on the number of times that match() is called, both
5544 in total and recursively. If the limit is exceeded, an error occurs.
5545 For details, see the section on extra data for pcre_exec() in the
5546 pcreapi documentation.
5547
5548 Each time that match() is actually called recursively, it uses memory
5549 from the process stack. For certain kinds of pattern and data, very
5550 large amounts of stack may be needed, despite the recognition of "tail
5551 recursion". You can often reduce the amount of recursion, and there-
5552 fore the amount of stack used, by modifying the pattern that is being
5553 matched. Consider, for example, this pattern:
5554
5555 ([^<]|<(?!inet))+
5556
5557 It matches from wherever it starts until it encounters "<inet" or the
5558 end of the data, and is the kind of pattern that might be used when
5559 processing an XML file. Each iteration of the outer parentheses matches
5560 either one character that is not "<" or a "<" that is not followed by
5561 "inet". However, each time a parenthesis is processed, a recursion
5562 occurs, so this formulation uses a stack frame for each matched charac-
5563 ter. For a long string, a lot of stack is required. Consider now this
5564 rewritten pattern, which matches exactly the same strings:
5565
5566 ([^<]++|<(?!inet))+
5567
5568 This uses very much less stack, because runs of characters that do not
5569 contain "<" are "swallowed" in one item inside the parentheses. Recur-
5570 sion happens only when a "<" character that is not followed by "inet"
5571 is encountered (and we assume this is relatively rare). A possessive
5572 quantifier is used to stop any backtracking into the runs of non-"<"
5573 characters, but that is not related to stack usage.
5574
5575 This example shows that one way of avoiding stack problems when match-
5576 ing long subject strings is to write repeated parenthesized subpatterns
5577 to match more than one character whenever possible.
5578
5579 In environments where stack memory is constrained, you might want to
5580 compile PCRE to use heap memory instead of stack for remembering back-
5581 up points. This makes it run a lot more slowly, however. Details of how
5582 to do this are given in the pcrebuild documentation.
5583
5584 In Unix-like environments, there is not often a problem with the stack
5585 unless very long strings are involved, though the default limit on
5586 stack size varies from system to system. Values from 8Mb to 64Mb are
5587 common. You can find your default limit by running the command:
5588
5589 ulimit -s
5590
5591 Unfortunately, the effect of running out of stack is often SIGSEGV,
5592 though sometimes a more explicit error message is given. You can nor-
5593 mally increase the limit on stack size by code such as this:
5594
5595 struct rlimit rlim;
5596 getrlimit(RLIMIT_STACK, &rlim);
5597 rlim.rlim_cur = 100*1024*1024;
5598 setrlimit(RLIMIT_STACK, &rlim);
5599
5600 This reads the current limits (soft and hard) using getrlimit(), then
5601 attempts to increase the soft limit to 100Mb using setrlimit(). You
5602 must do this before calling pcre_exec().
5603
5604 PCRE has an internal counter that can be used to limit the depth of
5605 recursion, and thus cause pcre_exec() to give an error code before it
5606 runs out of stack. By default, the limit is very large, and unlikely
5607 ever to operate. It can be changed when PCRE is built, and it can also
5608 be set when pcre_exec() is called. For details of these interfaces, see
5609 the pcrebuild and pcreapi documentation.
5610
5611 As a very rough rule of thumb, you should reckon on about 500 bytes per
5612 recursion. Thus, if you want to limit your stack usage to 8Mb, you
5613 should set the limit at 16000 recursions. A 64Mb stack, on the other
5614 hand, can support around 128000 recursions. The pcretest test program
5615 has a command line option (-S) that can be used to increase the size of
5616 its stack.
5617
5618
5619 AUTHOR
5620
5621 Philip Hazel
5622 University Computing Service
5623 Cambridge CB2 3QH, England.
5624
5625
5626 REVISION
5627
5628 Last updated: 12 March 2007
5629 Copyright (c) 1997-2007 University of Cambridge.
5630 ------------------------------------------------------------------------------
5631
5632

Properties

Name Value
svn:eol-style native
svn:keywords "Author Date Id Revision Url"

  ViewVC Help
Powered by ViewVC 1.1.5