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

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

Parent Directory Parent Directory | Revision Log Revision Log


Revision 172 - (show annotations)
Tue Jun 5 10:40:13 2007 UTC (12 years, 3 months ago) by ph10
File MIME type: text/plain
File size: 271268 byte(s)
Error occurred while calculating annotation data.
Drastically reduce workspace used for alternatives in groups; also some 
trailing space removals for a test release.
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. Because many paths through the tree may be active, the \K escape
614 sequence, which resets the start of the match when encountered (but may
615 be on some paths and not on others), is not supported. It causes an
616 error if encountered.
617
618 6. Callouts are supported, but the value of the capture_top field is
619 always 1, and the value of the capture_last field is always -1.
620
621 7. The \C escape sequence, which (in the standard algorithm) matches a
622 single byte, even in UTF-8 mode, is not supported because the alterna-
623 tive algorithm moves through the subject string one character at a
624 time, for all active paths through the tree.
625
626
627 ADVANTAGES OF THE ALTERNATIVE ALGORITHM
628
629 Using the alternative matching algorithm provides the following advan-
630 tages:
631
632 1. All possible matches (at a single point in the subject) are automat-
633 ically found, and in particular, the longest match is found. To find
634 more than one match using the standard algorithm, you have to do kludgy
635 things with callouts.
636
637 2. There is much better support for partial matching. The restrictions
638 on the content of the pattern that apply when using the standard algo-
639 rithm for partial matching do not apply to the alternative algorithm.
640 For non-anchored patterns, the starting position of a partial match is
641 available.
642
643 3. Because the alternative algorithm scans the subject string just
644 once, and never needs to backtrack, it is possible to pass very long
645 subject strings to the matching function in several pieces, checking
646 for partial matching each time.
647
648
649 DISADVANTAGES OF THE ALTERNATIVE ALGORITHM
650
651 The alternative algorithm suffers from a number of disadvantages:
652
653 1. It is substantially slower than the standard algorithm. This is
654 partly because it has to search for all possible matches, but is also
655 because it is less susceptible to optimization.
656
657 2. Capturing parentheses and back references are not supported.
658
659 3. Although atomic groups are supported, their use does not provide the
660 performance advantage that it does for the standard algorithm.
661
662
663 AUTHOR
664
665 Philip Hazel
666 University Computing Service
667 Cambridge CB2 3QH, England.
668
669
670 REVISION
671
672 Last updated: 29 May 2007
673 Copyright (c) 1997-2007 University of Cambridge.
674 ------------------------------------------------------------------------------
675
676
677 PCREAPI(3) PCREAPI(3)
678
679
680 NAME
681 PCRE - Perl-compatible regular expressions
682
683
684 PCRE NATIVE API
685
686 #include <pcre.h>
687
688 pcre *pcre_compile(const char *pattern, int options,
689 const char **errptr, int *erroffset,
690 const unsigned char *tableptr);
691
692 pcre *pcre_compile2(const char *pattern, int options,
693 int *errorcodeptr,
694 const char **errptr, int *erroffset,
695 const unsigned char *tableptr);
696
697 pcre_extra *pcre_study(const pcre *code, int options,
698 const char **errptr);
699
700 int pcre_exec(const pcre *code, const pcre_extra *extra,
701 const char *subject, int length, int startoffset,
702 int options, int *ovector, int ovecsize);
703
704 int pcre_dfa_exec(const pcre *code, const pcre_extra *extra,
705 const char *subject, int length, int startoffset,
706 int options, int *ovector, int ovecsize,
707 int *workspace, int wscount);
708
709 int pcre_copy_named_substring(const pcre *code,
710 const char *subject, int *ovector,
711 int stringcount, const char *stringname,
712 char *buffer, int buffersize);
713
714 int pcre_copy_substring(const char *subject, int *ovector,
715 int stringcount, int stringnumber, char *buffer,
716 int buffersize);
717
718 int pcre_get_named_substring(const pcre *code,
719 const char *subject, int *ovector,
720 int stringcount, const char *stringname,
721 const char **stringptr);
722
723 int pcre_get_stringnumber(const pcre *code,
724 const char *name);
725
726 int pcre_get_stringtable_entries(const pcre *code,
727 const char *name, char **first, char **last);
728
729 int pcre_get_substring(const char *subject, int *ovector,
730 int stringcount, int stringnumber,
731 const char **stringptr);
732
733 int pcre_get_substring_list(const char *subject,
734 int *ovector, int stringcount, const char ***listptr);
735
736 void pcre_free_substring(const char *stringptr);
737
738 void pcre_free_substring_list(const char **stringptr);
739
740 const unsigned char *pcre_maketables(void);
741
742 int pcre_fullinfo(const pcre *code, const pcre_extra *extra,
743 int what, void *where);
744
745 int pcre_info(const pcre *code, int *optptr, int *firstcharptr);
746
747 int pcre_refcount(pcre *code, int adjust);
748
749 int pcre_config(int what, void *where);
750
751 char *pcre_version(void);
752
753 void *(*pcre_malloc)(size_t);
754
755 void (*pcre_free)(void *);
756
757 void *(*pcre_stack_malloc)(size_t);
758
759 void (*pcre_stack_free)(void *);
760
761 int (*pcre_callout)(pcre_callout_block *);
762
763
764 PCRE API OVERVIEW
765
766 PCRE has its own native API, which is described in this document. There
767 are also some wrapper functions that correspond to the POSIX regular
768 expression API. These are described in the pcreposix documentation.
769 Both of these APIs define a set of C function calls. A C++ wrapper is
770 distributed with PCRE. It is documented in the pcrecpp page.
771
772 The native API C function prototypes are defined in the header file
773 pcre.h, and on Unix systems the library itself is called libpcre. It
774 can normally be accessed by adding -lpcre to the command for linking an
775 application that uses PCRE. The header file defines the macros
776 PCRE_MAJOR and PCRE_MINOR to contain the major and minor release num-
777 bers for the library. Applications can use these to include support
778 for different releases of PCRE.
779
780 The functions pcre_compile(), pcre_compile2(), pcre_study(), and
781 pcre_exec() are used for compiling and matching regular expressions in
782 a Perl-compatible manner. A sample program that demonstrates the sim-
783 plest way of using them is provided in the file called pcredemo.c in
784 the source distribution. The pcresample documentation describes how to
785 run it.
786
787 A second matching function, pcre_dfa_exec(), which is not Perl-compati-
788 ble, is also provided. This uses a different algorithm for the match-
789 ing. The alternative algorithm finds all possible matches (at a given
790 point in the subject), and scans the subject just once. However, this
791 algorithm does not return captured substrings. A description of the two
792 matching algorithms and their advantages and disadvantages is given in
793 the pcrematching documentation.
794
795 In addition to the main compiling and matching functions, there are
796 convenience functions for extracting captured substrings from a subject
797 string that is matched by pcre_exec(). They are:
798
799 pcre_copy_substring()
800 pcre_copy_named_substring()
801 pcre_get_substring()
802 pcre_get_named_substring()
803 pcre_get_substring_list()
804 pcre_get_stringnumber()
805 pcre_get_stringtable_entries()
806
807 pcre_free_substring() and pcre_free_substring_list() are also provided,
808 to free the memory used for extracted strings.
809
810 The function pcre_maketables() is used to build a set of character
811 tables in the current locale for passing to pcre_compile(),
812 pcre_exec(), or pcre_dfa_exec(). This is an optional facility that is
813 provided for specialist use. Most commonly, no special tables are
814 passed, in which case internal tables that are generated when PCRE is
815 built are used.
816
817 The function pcre_fullinfo() is used to find out information about a
818 compiled pattern; pcre_info() is an obsolete version that returns only
819 some of the available information, but is retained for backwards com-
820 patibility. The function pcre_version() returns a pointer to a string
821 containing the version of PCRE and its date of release.
822
823 The function pcre_refcount() maintains a reference count in a data
824 block containing a compiled pattern. This is provided for the benefit
825 of object-oriented applications.
826
827 The global variables pcre_malloc and pcre_free initially contain the
828 entry points of the standard malloc() and free() functions, respec-
829 tively. PCRE calls the memory management functions via these variables,
830 so a calling program can replace them if it wishes to intercept the
831 calls. This should be done before calling any PCRE functions.
832
833 The global variables pcre_stack_malloc and pcre_stack_free are also
834 indirections to memory management functions. These special functions
835 are used only when PCRE is compiled to use the heap for remembering
836 data, instead of recursive function calls, when running the pcre_exec()
837 function. See the pcrebuild documentation for details of how to do
838 this. It is a non-standard way of building PCRE, for use in environ-
839 ments that have limited stacks. Because of the greater use of memory
840 management, it runs more slowly. Separate functions are provided so
841 that special-purpose external code can be used for this case. When
842 used, these functions are always called in a stack-like manner (last
843 obtained, first freed), and always for memory blocks of the same size.
844 There is a discussion about PCRE's stack usage in the pcrestack docu-
845 mentation.
846
847 The global variable pcre_callout initially contains NULL. It can be set
848 by the caller to a "callout" function, which PCRE will then call at
849 specified points during a matching operation. Details are given in the
850 pcrecallout documentation.
851
852
853 NEWLINES
854
855 PCRE supports five different conventions for indicating line breaks in
856 strings: a single CR (carriage return) character, a single LF (line-
857 feed) character, the two-character sequence CRLF, any of the three pre-
858 ceding, or any Unicode newline sequence. The Unicode newline sequences
859 are the three just mentioned, plus the single characters VT (vertical
860 tab, U+000B), FF (formfeed, U+000C), NEL (next line, U+0085), LS (line
861 separator, U+2028), and PS (paragraph separator, U+2029).
862
863 Each of the first three conventions is used by at least one operating
864 system as its standard newline sequence. When PCRE is built, a default
865 can be specified. The default default is LF, which is the Unix stan-
866 dard. When PCRE is run, the default can be overridden, either when a
867 pattern is compiled, or when it is matched.
868
869 In the PCRE documentation the word "newline" is used to mean "the char-
870 acter or pair of characters that indicate a line break". The choice of
871 newline convention affects the handling of the dot, circumflex, and
872 dollar metacharacters, the handling of #-comments in /x mode, and, when
873 CRLF is a recognized line ending sequence, the match position advance-
874 ment for a non-anchored pattern. The choice of newline convention does
875 not affect the interpretation of the \n or \r escape sequences.
876
877
878 MULTITHREADING
879
880 The PCRE functions can be used in multi-threading applications, with
881 the proviso that the memory management functions pointed to by
882 pcre_malloc, pcre_free, pcre_stack_malloc, and pcre_stack_free, and the
883 callout function pointed to by pcre_callout, are shared by all threads.
884
885 The compiled form of a regular expression is not altered during match-
886 ing, so the same compiled pattern can safely be used by several threads
887 at once.
888
889
890 SAVING PRECOMPILED PATTERNS FOR LATER USE
891
892 The compiled form of a regular expression can be saved and re-used at a
893 later time, possibly by a different program, and even on a host other
894 than the one on which it was compiled. Details are given in the
895 pcreprecompile documentation. However, compiling a regular expression
896 with one version of PCRE for use with a different version is not guar-
897 anteed to work and may cause crashes.
898
899
900 CHECKING BUILD-TIME OPTIONS
901
902 int pcre_config(int what, void *where);
903
904 The function pcre_config() makes it possible for a PCRE client to dis-
905 cover which optional features have been compiled into the PCRE library.
906 The pcrebuild documentation has more details about these optional fea-
907 tures.
908
909 The first argument for pcre_config() is an integer, specifying which
910 information is required; the second argument is a pointer to a variable
911 into which the information is placed. The following information is
912 available:
913
914 PCRE_CONFIG_UTF8
915
916 The output is an integer that is set to one if UTF-8 support is avail-
917 able; otherwise it is set to zero.
918
919 PCRE_CONFIG_UNICODE_PROPERTIES
920
921 The output is an integer that is set to one if support for Unicode
922 character properties is available; otherwise it is set to zero.
923
924 PCRE_CONFIG_NEWLINE
925
926 The output is an integer whose value specifies the default character
927 sequence that is recognized as meaning "newline". The four values that
928 are supported are: 10 for LF, 13 for CR, 3338 for CRLF, -2 for ANYCRLF,
929 and -1 for ANY. The default should normally be the standard sequence
930 for your operating system.
931
932 PCRE_CONFIG_LINK_SIZE
933
934 The output is an integer that contains the number of bytes used for
935 internal linkage in compiled regular expressions. The value is 2, 3, or
936 4. Larger values allow larger regular expressions to be compiled, at
937 the expense of slower matching. The default value of 2 is sufficient
938 for all but the most massive patterns, since it allows the compiled
939 pattern to be up to 64K in size.
940
941 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
942
943 The output is an integer that contains the threshold above which the
944 POSIX interface uses malloc() for output vectors. Further details are
945 given in the pcreposix documentation.
946
947 PCRE_CONFIG_MATCH_LIMIT
948
949 The output is an integer that gives the default limit for the number of
950 internal matching function calls in a pcre_exec() execution. Further
951 details are given with pcre_exec() below.
952
953 PCRE_CONFIG_MATCH_LIMIT_RECURSION
954
955 The output is an integer that gives the default limit for the depth of
956 recursion when calling the internal matching function in a pcre_exec()
957 execution. Further details are given with pcre_exec() below.
958
959 PCRE_CONFIG_STACKRECURSE
960
961 The output is an integer that is set to one if internal recursion when
962 running pcre_exec() is implemented by recursive function calls that use
963 the stack to remember their state. This is the usual way that PCRE is
964 compiled. The output is zero if PCRE was compiled to use blocks of data
965 on the heap instead of recursive function calls. In this case,
966 pcre_stack_malloc and pcre_stack_free are called to manage memory
967 blocks on the heap, thus avoiding the use of the stack.
968
969
970 COMPILING A PATTERN
971
972 pcre *pcre_compile(const char *pattern, int options,
973 const char **errptr, int *erroffset,
974 const unsigned char *tableptr);
975
976 pcre *pcre_compile2(const char *pattern, int options,
977 int *errorcodeptr,
978 const char **errptr, int *erroffset,
979 const unsigned char *tableptr);
980
981 Either of the functions pcre_compile() or pcre_compile2() can be called
982 to compile a pattern into an internal form. The only difference between
983 the two interfaces is that pcre_compile2() has an additional argument,
984 errorcodeptr, via which a numerical error code can be returned.
985
986 The pattern is a C string terminated by a binary zero, and is passed in
987 the pattern argument. A pointer to a single block of memory that is
988 obtained via pcre_malloc is returned. This contains the compiled code
989 and related data. The pcre type is defined for the returned block; this
990 is a typedef for a structure whose contents are not externally defined.
991 It is up to the caller to free the memory (via pcre_free) when it is no
992 longer required.
993
994 Although the compiled code of a PCRE regex is relocatable, that is, it
995 does not depend on memory location, the complete pcre data block is not
996 fully relocatable, because it may contain a copy of the tableptr argu-
997 ment, which is an address (see below).
998
999 The options argument contains various bit settings that affect the com-
1000 pilation. It should be zero if no options are required. The available
1001 options are described below. Some of them, in particular, those that
1002 are compatible with Perl, can also be set and unset from within the
1003 pattern (see the detailed description in the pcrepattern documenta-
1004 tion). For these options, the contents of the options argument speci-
1005 fies their initial settings at the start of compilation and execution.
1006 The PCRE_ANCHORED and PCRE_NEWLINE_xxx options can be set at the time
1007 of matching as well as at compile time.
1008
1009 If errptr is NULL, pcre_compile() returns NULL immediately. Otherwise,
1010 if compilation of a pattern fails, pcre_compile() returns NULL, and
1011 sets the variable pointed to by errptr to point to a textual error mes-
1012 sage. This is a static string that is part of the library. You must not
1013 try to free it. The offset from the start of the pattern to the charac-
1014 ter where the error was discovered is placed in the variable pointed to
1015 by erroffset, which must not be NULL. If it is, an immediate error is
1016 given.
1017
1018 If pcre_compile2() is used instead of pcre_compile(), and the error-
1019 codeptr argument is not NULL, a non-zero error code number is returned
1020 via this argument in the event of an error. This is in addition to the
1021 textual error message. Error codes and messages are listed below.
1022
1023 If the final argument, tableptr, is NULL, PCRE uses a default set of
1024 character tables that are built when PCRE is compiled, using the
1025 default C locale. Otherwise, tableptr must be an address that is the
1026 result of a call to pcre_maketables(). This value is stored with the
1027 compiled pattern, and used again by pcre_exec(), unless another table
1028 pointer is passed to it. For more discussion, see the section on locale
1029 support below.
1030
1031 This code fragment shows a typical straightforward call to pcre_com-
1032 pile():
1033
1034 pcre *re;
1035 const char *error;
1036 int erroffset;
1037 re = pcre_compile(
1038 "^A.*Z", /* the pattern */
1039 0, /* default options */
1040 &error, /* for error message */
1041 &erroffset, /* for error offset */
1042 NULL); /* use default character tables */
1043
1044 The following names for option bits are defined in the pcre.h header
1045 file:
1046
1047 PCRE_ANCHORED
1048
1049 If this bit is set, the pattern is forced to be "anchored", that is, it
1050 is constrained to match only at the first matching point in the string
1051 that is being searched (the "subject string"). This effect can also be
1052 achieved by appropriate constructs in the pattern itself, which is the
1053 only way to do it in Perl.
1054
1055 PCRE_AUTO_CALLOUT
1056
1057 If this bit is set, pcre_compile() automatically inserts callout items,
1058 all with number 255, before each pattern item. For discussion of the
1059 callout facility, see the pcrecallout documentation.
1060
1061 PCRE_CASELESS
1062
1063 If this bit is set, letters in the pattern match both upper and lower
1064 case letters. It is equivalent to Perl's /i option, and it can be
1065 changed within a pattern by a (?i) option setting. In UTF-8 mode, PCRE
1066 always understands the concept of case for characters whose values are
1067 less than 128, so caseless matching is always possible. For characters
1068 with higher values, the concept of case is supported if PCRE is com-
1069 piled with Unicode property support, but not otherwise. If you want to
1070 use caseless matching for characters 128 and above, you must ensure
1071 that PCRE is compiled with Unicode property support as well as with
1072 UTF-8 support.
1073
1074 PCRE_DOLLAR_ENDONLY
1075
1076 If this bit is set, a dollar metacharacter in the pattern matches only
1077 at the end of the subject string. Without this option, a dollar also
1078 matches immediately before a newline at the end of the string (but not
1079 before any other newlines). The PCRE_DOLLAR_ENDONLY option is ignored
1080 if PCRE_MULTILINE is set. There is no equivalent to this option in
1081 Perl, and no way to set it within a pattern.
1082
1083 PCRE_DOTALL
1084
1085 If this bit is set, a dot metacharater in the pattern matches all char-
1086 acters, including those that indicate newline. Without it, a dot does
1087 not match when the current position is at a newline. This option is
1088 equivalent to Perl's /s option, and it can be changed within a pattern
1089 by a (?s) option setting. A negative class such as [^a] always matches
1090 newline characters, independent of the setting of this option.
1091
1092 PCRE_DUPNAMES
1093
1094 If this bit is set, names used to identify capturing subpatterns need
1095 not be unique. This can be helpful for certain types of pattern when it
1096 is known that only one instance of the named subpattern can ever be
1097 matched. There are more details of named subpatterns below; see also
1098 the pcrepattern documentation.
1099
1100 PCRE_EXTENDED
1101
1102 If this bit is set, whitespace data characters in the pattern are
1103 totally ignored except when escaped or inside a character class. White-
1104 space does not include the VT character (code 11). In addition, charac-
1105 ters between an unescaped # outside a character class and the next new-
1106 line, inclusive, are also ignored. This is equivalent to Perl's /x
1107 option, and it can be changed within a pattern by a (?x) option set-
1108 ting.
1109
1110 This option makes it possible to include comments inside complicated
1111 patterns. Note, however, that this applies only to data characters.
1112 Whitespace characters may never appear within special character
1113 sequences in a pattern, for example within the sequence (?( which
1114 introduces a conditional subpattern.
1115
1116 PCRE_EXTRA
1117
1118 This option was invented in order to turn on additional functionality
1119 of PCRE that is incompatible with Perl, but it is currently of very
1120 little use. When set, any backslash in a pattern that is followed by a
1121 letter that has no special meaning causes an error, thus reserving
1122 these combinations for future expansion. By default, as in Perl, a
1123 backslash followed by a letter with no special meaning is treated as a
1124 literal. (Perl can, however, be persuaded to give a warning for this.)
1125 There are at present no other features controlled by this option. It
1126 can also be set by a (?X) option setting within a pattern.
1127
1128 PCRE_FIRSTLINE
1129
1130 If this option is set, an unanchored pattern is required to match
1131 before or at the first newline in the subject string, though the
1132 matched text may continue over the newline.
1133
1134 PCRE_MULTILINE
1135
1136 By default, PCRE treats the subject string as consisting of a single
1137 line of characters (even if it actually contains newlines). The "start
1138 of line" metacharacter (^) matches only at the start of the string,
1139 while the "end of line" metacharacter ($) matches only at the end of
1140 the string, or before a terminating newline (unless PCRE_DOLLAR_ENDONLY
1141 is set). This is the same as Perl.
1142
1143 When PCRE_MULTILINE it is set, the "start of line" and "end of line"
1144 constructs match immediately following or immediately before internal
1145 newlines in the subject string, respectively, as well as at the very
1146 start and end. This is equivalent to Perl's /m option, and it can be
1147 changed within a pattern by a (?m) option setting. If there are no new-
1148 lines in a subject string, or no occurrences of ^ or $ in a pattern,
1149 setting PCRE_MULTILINE has no effect.
1150
1151 PCRE_NEWLINE_CR
1152 PCRE_NEWLINE_LF
1153 PCRE_NEWLINE_CRLF
1154 PCRE_NEWLINE_ANYCRLF
1155 PCRE_NEWLINE_ANY
1156
1157 These options override the default newline definition that was chosen
1158 when PCRE was built. Setting the first or the second specifies that a
1159 newline is indicated by a single character (CR or LF, respectively).
1160 Setting PCRE_NEWLINE_CRLF specifies that a newline is indicated by the
1161 two-character CRLF sequence. Setting PCRE_NEWLINE_ANYCRLF specifies
1162 that any of the three preceding sequences should be recognized. Setting
1163 PCRE_NEWLINE_ANY specifies that any Unicode newline sequence should be
1164 recognized. The Unicode newline sequences are the three just mentioned,
1165 plus the single characters VT (vertical tab, U+000B), FF (formfeed,
1166 U+000C), NEL (next line, U+0085), LS (line separator, U+2028), and PS
1167 (paragraph separator, U+2029). The last two are recognized only in
1168 UTF-8 mode.
1169
1170 The newline setting in the options word uses three bits that are
1171 treated as a number, giving eight possibilities. Currently only six are
1172 used (default plus the five values above). This means that if you set
1173 more than one newline option, the combination may or may not be sensi-
1174 ble. For example, PCRE_NEWLINE_CR with PCRE_NEWLINE_LF is equivalent to
1175 PCRE_NEWLINE_CRLF, but other combinations may yield unused numbers and
1176 cause an error.
1177
1178 The only time that a line break is specially recognized when compiling
1179 a pattern is if PCRE_EXTENDED is set, and an unescaped # outside a
1180 character class is encountered. This indicates a comment that lasts
1181 until after the next line break sequence. In other circumstances, line
1182 break sequences are treated as literal data, except that in
1183 PCRE_EXTENDED mode, both CR and LF are treated as whitespace characters
1184 and are therefore ignored.
1185
1186 The newline option that is set at compile time becomes the default that
1187 is used for pcre_exec() and pcre_dfa_exec(), but it can be overridden.
1188
1189 PCRE_NO_AUTO_CAPTURE
1190
1191 If this option is set, it disables the use of numbered capturing paren-
1192 theses in the pattern. Any opening parenthesis that is not followed by
1193 ? behaves as if it were followed by ?: but named parentheses can still
1194 be used for capturing (and they acquire numbers in the usual way).
1195 There is no equivalent of this option in Perl.
1196
1197 PCRE_UNGREEDY
1198
1199 This option inverts the "greediness" of the quantifiers so that they
1200 are not greedy by default, but become greedy if followed by "?". It is
1201 not compatible with Perl. It can also be set by a (?U) option setting
1202 within the pattern.
1203
1204 PCRE_UTF8
1205
1206 This option causes PCRE to regard both the pattern and the subject as
1207 strings of UTF-8 characters instead of single-byte character strings.
1208 However, it is available only when PCRE is built to include UTF-8 sup-
1209 port. If not, the use of this option provokes an error. Details of how
1210 this option changes the behaviour of PCRE are given in the section on
1211 UTF-8 support in the main pcre page.
1212
1213 PCRE_NO_UTF8_CHECK
1214
1215 When PCRE_UTF8 is set, the validity of the pattern as a UTF-8 string is
1216 automatically checked. If an invalid UTF-8 sequence of bytes is found,
1217 pcre_compile() returns an error. If you already know that your pattern
1218 is valid, and you want to skip this check for performance reasons, you
1219 can set the PCRE_NO_UTF8_CHECK option. When it is set, the effect of
1220 passing an invalid UTF-8 string as a pattern is undefined. It may cause
1221 your program to crash. Note that this option can also be passed to
1222 pcre_exec() and pcre_dfa_exec(), to suppress the UTF-8 validity check-
1223 ing of subject strings.
1224
1225
1226 COMPILATION ERROR CODES
1227
1228 The following table lists the error codes than may be returned by
1229 pcre_compile2(), along with the error messages that may be returned by
1230 both compiling functions. As PCRE has developed, some error codes have
1231 fallen out of use. To avoid confusion, they have not been re-used.
1232
1233 0 no error
1234 1 \ at end of pattern
1235 2 \c at end of pattern
1236 3 unrecognized character follows \
1237 4 numbers out of order in {} quantifier
1238 5 number too big in {} quantifier
1239 6 missing terminating ] for character class
1240 7 invalid escape sequence in character class
1241 8 range out of order in character class
1242 9 nothing to repeat
1243 10 [this code is not in use]
1244 11 internal error: unexpected repeat
1245 12 unrecognized character after (?
1246 13 POSIX named classes are supported only within a class
1247 14 missing )
1248 15 reference to non-existent subpattern
1249 16 erroffset passed as NULL
1250 17 unknown option bit(s) set
1251 18 missing ) after comment
1252 19 [this code is not in use]
1253 20 regular expression too large
1254 21 failed to get memory
1255 22 unmatched parentheses
1256 23 internal error: code overflow
1257 24 unrecognized character after (?<
1258 25 lookbehind assertion is not fixed length
1259 26 malformed number or name after (?(
1260 27 conditional group contains more than two branches
1261 28 assertion expected after (?(
1262 29 (?R or (?digits must be followed by )
1263 30 unknown POSIX class name
1264 31 POSIX collating elements are not supported
1265 32 this version of PCRE is not compiled with PCRE_UTF8 support
1266 33 [this code is not in use]
1267 34 character value in \x{...} sequence is too large
1268 35 invalid condition (?(0)
1269 36 \C not allowed in lookbehind assertion
1270 37 PCRE does not support \L, \l, \N, \U, or \u
1271 38 number after (?C is > 255
1272 39 closing ) for (?C expected
1273 40 recursive call could loop indefinitely
1274 41 unrecognized character after (?P
1275 42 syntax error in subpattern name (missing terminator)
1276 43 two named subpatterns have the same name
1277 44 invalid UTF-8 string
1278 45 support for \P, \p, and \X has not been compiled
1279 46 malformed \P or \p sequence
1280 47 unknown property name after \P or \p
1281 48 subpattern name is too long (maximum 32 characters)
1282 49 too many named subpatterns (maximum 10,000)
1283 50 repeated subpattern is too long
1284 51 octal value is greater than \377 (not in UTF-8 mode)
1285 52 internal error: overran compiling workspace
1286 53 internal error: previously-checked referenced subpattern not
1287 found
1288 54 DEFINE group contains more than one branch
1289 55 repeating a DEFINE group is not allowed
1290 56 inconsistent NEWLINE options"
1291
1292
1293 STUDYING A PATTERN
1294
1295 pcre_extra *pcre_study(const pcre *code, int options
1296 const char **errptr);
1297
1298 If a compiled pattern is going to be used several times, it is worth
1299 spending more time analyzing it in order to speed up the time taken for
1300 matching. The function pcre_study() takes a pointer to a compiled pat-
1301 tern as its first argument. If studying the pattern produces additional
1302 information that will help speed up matching, pcre_study() returns a
1303 pointer to a pcre_extra block, in which the study_data field points to
1304 the results of the study.
1305
1306 The returned value from pcre_study() can be passed directly to
1307 pcre_exec(). However, a pcre_extra block also contains other fields
1308 that can be set by the caller before the block is passed; these are
1309 described below in the section on matching a pattern.
1310
1311 If studying the pattern does not produce any additional information
1312 pcre_study() returns NULL. In that circumstance, if the calling program
1313 wants to pass any of the other fields to pcre_exec(), it must set up
1314 its own pcre_extra block.
1315
1316 The second argument of pcre_study() contains option bits. At present,
1317 no options are defined, and this argument should always be zero.
1318
1319 The third argument for pcre_study() is a pointer for an error message.
1320 If studying succeeds (even if no data is returned), the variable it
1321 points to is set to NULL. Otherwise it is set to point to a textual
1322 error message. This is a static string that is part of the library. You
1323 must not try to free it. You should test the error pointer for NULL
1324 after calling pcre_study(), to be sure that it has run successfully.
1325
1326 This is a typical call to pcre_study():
1327
1328 pcre_extra *pe;
1329 pe = pcre_study(
1330 re, /* result of pcre_compile() */
1331 0, /* no options exist */
1332 &error); /* set to NULL or points to a message */
1333
1334 At present, studying a pattern is useful only for non-anchored patterns
1335 that do not have a single fixed starting character. A bitmap of possi-
1336 ble starting bytes is created.
1337
1338
1339 LOCALE SUPPORT
1340
1341 PCRE handles caseless matching, and determines whether characters are
1342 letters, digits, or whatever, by reference to a set of tables, indexed
1343 by character value. When running in UTF-8 mode, this applies only to
1344 characters with codes less than 128. Higher-valued codes never match
1345 escapes such as \w or \d, but can be tested with \p if PCRE is built
1346 with Unicode character property support. The use of locales with Uni-
1347 code is discouraged. If you are handling characters with codes greater
1348 than 128, you should either use UTF-8 and Unicode, or use locales, but
1349 not try to mix the two.
1350
1351 PCRE contains an internal set of tables that are used when the final
1352 argument of pcre_compile() is NULL. These are sufficient for many
1353 applications. Normally, the internal tables recognize only ASCII char-
1354 acters. However, when PCRE is built, it is possible to cause the inter-
1355 nal tables to be rebuilt in the default "C" locale of the local system,
1356 which may cause them to be different.
1357
1358 The internal tables can always be overridden by tables supplied by the
1359 application that calls PCRE. These may be created in a different locale
1360 from the default. As more and more applications change to using Uni-
1361 code, the need for this locale support is expected to die away.
1362
1363 External tables are built by calling the pcre_maketables() function,
1364 which has no arguments, in the relevant locale. The result can then be
1365 passed to pcre_compile() or pcre_exec() as often as necessary. For
1366 example, to build and use tables that are appropriate for the French
1367 locale (where accented characters with values greater than 128 are
1368 treated as letters), the following code could be used:
1369
1370 setlocale(LC_CTYPE, "fr_FR");
1371 tables = pcre_maketables();
1372 re = pcre_compile(..., tables);
1373
1374 The locale name "fr_FR" is used on Linux and other Unix-like systems;
1375 if you are using Windows, the name for the French locale is "french".
1376
1377 When pcre_maketables() runs, the tables are built in memory that is
1378 obtained via pcre_malloc. It is the caller's responsibility to ensure
1379 that the memory containing the tables remains available for as long as
1380 it is needed.
1381
1382 The pointer that is passed to pcre_compile() is saved with the compiled
1383 pattern, and the same tables are used via this pointer by pcre_study()
1384 and normally also by pcre_exec(). Thus, by default, for any single pat-
1385 tern, compilation, studying and matching all happen in the same locale,
1386 but different patterns can be compiled in different locales.
1387
1388 It is possible to pass a table pointer or NULL (indicating the use of
1389 the internal tables) to pcre_exec(). Although not intended for this
1390 purpose, this facility could be used to match a pattern in a different
1391 locale from the one in which it was compiled. Passing table pointers at
1392 run time is discussed below in the section on matching a pattern.
1393
1394
1395 INFORMATION ABOUT A PATTERN
1396
1397 int pcre_fullinfo(const pcre *code, const pcre_extra *extra,
1398 int what, void *where);
1399
1400 The pcre_fullinfo() function returns information about a compiled pat-
1401 tern. It replaces the obsolete pcre_info() function, which is neverthe-
1402 less retained for backwards compability (and is documented below).
1403
1404 The first argument for pcre_fullinfo() is a pointer to the compiled
1405 pattern. The second argument is the result of pcre_study(), or NULL if
1406 the pattern was not studied. The third argument specifies which piece
1407 of information is required, and the fourth argument is a pointer to a
1408 variable to receive the data. The yield of the function is zero for
1409 success, or one of the following negative numbers:
1410
1411 PCRE_ERROR_NULL the argument code was NULL
1412 the argument where was NULL
1413 PCRE_ERROR_BADMAGIC the "magic number" was not found
1414 PCRE_ERROR_BADOPTION the value of what was invalid
1415
1416 The "magic number" is placed at the start of each compiled pattern as
1417 an simple check against passing an arbitrary memory pointer. Here is a
1418 typical call of pcre_fullinfo(), to obtain the length of the compiled
1419 pattern:
1420
1421 int rc;
1422 size_t length;
1423 rc = pcre_fullinfo(
1424 re, /* result of pcre_compile() */
1425 pe, /* result of pcre_study(), or NULL */
1426 PCRE_INFO_SIZE, /* what is required */
1427 &length); /* where to put the data */
1428
1429 The possible values for the third argument are defined in pcre.h, and
1430 are as follows:
1431
1432 PCRE_INFO_BACKREFMAX
1433
1434 Return the number of the highest back reference in the pattern. The
1435 fourth argument should point to an int variable. Zero is returned if
1436 there are no back references.
1437
1438 PCRE_INFO_CAPTURECOUNT
1439
1440 Return the number of capturing subpatterns in the pattern. The fourth
1441 argument should point to an int variable.
1442
1443 PCRE_INFO_DEFAULT_TABLES
1444
1445 Return a pointer to the internal default character tables within PCRE.
1446 The fourth argument should point to an unsigned char * variable. This
1447 information call is provided for internal use by the pcre_study() func-
1448 tion. External callers can cause PCRE to use its internal tables by
1449 passing a NULL table pointer.
1450
1451 PCRE_INFO_FIRSTBYTE
1452
1453 Return information about the first byte of any matched string, for a
1454 non-anchored pattern. The fourth argument should point to an int vari-
1455 able. (This option used to be called PCRE_INFO_FIRSTCHAR; the old name
1456 is still recognized for backwards compatibility.)
1457
1458 If there is a fixed first byte, for example, from a pattern such as
1459 (cat|cow|coyote), its value is returned. Otherwise, if either
1460
1461 (a) the pattern was compiled with the PCRE_MULTILINE option, and every
1462 branch starts with "^", or
1463
1464 (b) every branch of the pattern starts with ".*" and PCRE_DOTALL is not
1465 set (if it were set, the pattern would be anchored),
1466
1467 -1 is returned, indicating that the pattern matches only at the start
1468 of a subject string or after any newline within the string. Otherwise
1469 -2 is returned. For anchored patterns, -2 is returned.
1470
1471 PCRE_INFO_FIRSTTABLE
1472
1473 If the pattern was studied, and this resulted in the construction of a
1474 256-bit table indicating a fixed set of bytes for the first byte in any
1475 matching string, a pointer to the table is returned. Otherwise NULL is
1476 returned. The fourth argument should point to an unsigned char * vari-
1477 able.
1478
1479 PCRE_INFO_JCHANGED
1480
1481 Return 1 if the (?J) option setting is used in the pattern, otherwise
1482 0. The fourth argument should point to an int variable. The (?J) inter-
1483 nal option setting changes the local PCRE_DUPNAMES value.
1484
1485 PCRE_INFO_LASTLITERAL
1486
1487 Return the value of the rightmost literal byte that must exist in any
1488 matched string, other than at its start, if such a byte has been
1489 recorded. The fourth argument should point to an int variable. If there
1490 is no such byte, -1 is returned. For anchored patterns, a last literal
1491 byte is recorded only if it follows something of variable length. For
1492 example, for the pattern /^a\d+z\d+/ the returned value is "z", but for
1493 /^a\dz\d/ the returned value is -1.
1494
1495 PCRE_INFO_NAMECOUNT
1496 PCRE_INFO_NAMEENTRYSIZE
1497 PCRE_INFO_NAMETABLE
1498
1499 PCRE supports the use of named as well as numbered capturing parenthe-
1500 ses. The names are just an additional way of identifying the parenthe-
1501 ses, which still acquire numbers. Several convenience functions such as
1502 pcre_get_named_substring() are provided for extracting captured sub-
1503 strings by name. It is also possible to extract the data directly, by
1504 first converting the name to a number in order to access the correct
1505 pointers in the output vector (described with pcre_exec() below). To do
1506 the conversion, you need to use the name-to-number map, which is
1507 described by these three values.
1508
1509 The map consists of a number of fixed-size entries. PCRE_INFO_NAMECOUNT
1510 gives the number of entries, and PCRE_INFO_NAMEENTRYSIZE gives the size
1511 of each entry; both of these return an int value. The entry size
1512 depends on the length of the longest name. PCRE_INFO_NAMETABLE returns
1513 a pointer to the first entry of the table (a pointer to char). The
1514 first two bytes of each entry are the number of the capturing parenthe-
1515 sis, most significant byte first. The rest of the entry is the corre-
1516 sponding name, zero terminated. The names are in alphabetical order.
1517 When PCRE_DUPNAMES is set, duplicate names are in order of their paren-
1518 theses numbers. For example, consider the following pattern (assume
1519 PCRE_EXTENDED is set, so white space - including newlines - is
1520 ignored):
1521
1522 (?<date> (?<year>(\d\d)?\d\d) -
1523 (?<month>\d\d) - (?<day>\d\d) )
1524
1525 There are four named subpatterns, so the table has four entries, and
1526 each entry in the table is eight bytes long. The table is as follows,
1527 with non-printing bytes shows in hexadecimal, and undefined bytes shown
1528 as ??:
1529
1530 00 01 d a t e 00 ??
1531 00 05 d a y 00 ?? ??
1532 00 04 m o n t h 00
1533 00 02 y e a r 00 ??
1534
1535 When writing code to extract data from named subpatterns using the
1536 name-to-number map, remember that the length of the entries is likely
1537 to be different for each compiled pattern.
1538
1539 PCRE_INFO_OKPARTIAL
1540
1541 Return 1 if the pattern can be used for partial matching, otherwise 0.
1542 The fourth argument should point to an int variable. The pcrepartial
1543 documentation lists the restrictions that apply to patterns when par-
1544 tial matching is used.
1545
1546 PCRE_INFO_OPTIONS
1547
1548 Return a copy of the options with which the pattern was compiled. The
1549 fourth argument should point to an unsigned long int variable. These
1550 option bits are those specified in the call to pcre_compile(), modified
1551 by any top-level option settings within the pattern itself.
1552
1553 A pattern is automatically anchored by PCRE if all of its top-level
1554 alternatives begin with one of the following:
1555
1556 ^ unless PCRE_MULTILINE is set
1557 \A always
1558 \G always
1559 .* if PCRE_DOTALL is set and there are no back
1560 references to the subpattern in which .* appears
1561
1562 For such patterns, the PCRE_ANCHORED bit is set in the options returned
1563 by pcre_fullinfo().
1564
1565 PCRE_INFO_SIZE
1566
1567 Return the size of the compiled pattern, that is, the value that was
1568 passed as the argument to pcre_malloc() when PCRE was getting memory in
1569 which to place the compiled data. The fourth argument should point to a
1570 size_t variable.
1571
1572 PCRE_INFO_STUDYSIZE
1573
1574 Return the size of the data block pointed to by the study_data field in
1575 a pcre_extra block. That is, it is the value that was passed to
1576 pcre_malloc() when PCRE was getting memory into which to place the data
1577 created by pcre_study(). The fourth argument should point to a size_t
1578 variable.
1579
1580
1581 OBSOLETE INFO FUNCTION
1582
1583 int pcre_info(const pcre *code, int *optptr, int *firstcharptr);
1584
1585 The pcre_info() function is now obsolete because its interface is too
1586 restrictive to return all the available data about a compiled pattern.
1587 New programs should use pcre_fullinfo() instead. The yield of
1588 pcre_info() is the number of capturing subpatterns, or one of the fol-
1589 lowing negative numbers:
1590
1591 PCRE_ERROR_NULL the argument code was NULL
1592 PCRE_ERROR_BADMAGIC the "magic number" was not found
1593
1594 If the optptr argument is not NULL, a copy of the options with which
1595 the pattern was compiled is placed in the integer it points to (see
1596 PCRE_INFO_OPTIONS above).
1597
1598 If the pattern is not anchored and the firstcharptr argument is not
1599 NULL, it is used to pass back information about the first character of
1600 any matched string (see PCRE_INFO_FIRSTBYTE above).
1601
1602
1603 REFERENCE COUNTS
1604
1605 int pcre_refcount(pcre *code, int adjust);
1606
1607 The pcre_refcount() function is used to maintain a reference count in
1608 the data block that contains a compiled pattern. It is provided for the
1609 benefit of applications that operate in an object-oriented manner,
1610 where different parts of the application may be using the same compiled
1611 pattern, but you want to free the block when they are all done.
1612
1613 When a pattern is compiled, the reference count field is initialized to
1614 zero. It is changed only by calling this function, whose action is to
1615 add the adjust value (which may be positive or negative) to it. The
1616 yield of the function is the new value. However, the value of the count
1617 is constrained to lie between 0 and 65535, inclusive. If the new value
1618 is outside these limits, it is forced to the appropriate limit value.
1619
1620 Except when it is zero, the reference count is not correctly preserved
1621 if a pattern is compiled on one host and then transferred to a host
1622 whose byte-order is different. (This seems a highly unlikely scenario.)
1623
1624
1625 MATCHING A PATTERN: THE TRADITIONAL FUNCTION
1626
1627 int pcre_exec(const pcre *code, const pcre_extra *extra,
1628 const char *subject, int length, int startoffset,
1629 int options, int *ovector, int ovecsize);
1630
1631 The function pcre_exec() is called to match a subject string against a
1632 compiled pattern, which is passed in the code argument. If the pattern
1633 has been studied, the result of the study should be passed in the extra
1634 argument. This function is the main matching facility of the library,
1635 and it operates in a Perl-like manner. For specialist use there is also
1636 an alternative matching function, which is described below in the sec-
1637 tion about the pcre_dfa_exec() function.
1638
1639 In most applications, the pattern will have been compiled (and option-
1640 ally studied) in the same process that calls pcre_exec(). However, it
1641 is possible to save compiled patterns and study data, and then use them
1642 later in different processes, possibly even on different hosts. For a
1643 discussion about this, see the pcreprecompile documentation.
1644
1645 Here is an example of a simple call to pcre_exec():
1646
1647 int rc;
1648 int ovector[30];
1649 rc = pcre_exec(
1650 re, /* result of pcre_compile() */
1651 NULL, /* we didn't study the pattern */
1652 "some string", /* the subject string */
1653 11, /* the length of the subject string */
1654 0, /* start at offset 0 in the subject */
1655 0, /* default options */
1656 ovector, /* vector of integers for substring information */
1657 30); /* number of elements (NOT size in bytes) */
1658
1659 Extra data for pcre_exec()
1660
1661 If the extra argument is not NULL, it must point to a pcre_extra data
1662 block. The pcre_study() function returns such a block (when it doesn't
1663 return NULL), but you can also create one for yourself, and pass addi-
1664 tional information in it. The pcre_extra block contains the following
1665 fields (not necessarily in this order):
1666
1667 unsigned long int flags;
1668 void *study_data;
1669 unsigned long int match_limit;
1670 unsigned long int match_limit_recursion;
1671 void *callout_data;
1672 const unsigned char *tables;
1673
1674 The flags field is a bitmap that specifies which of the other fields
1675 are set. The flag bits are:
1676
1677 PCRE_EXTRA_STUDY_DATA
1678 PCRE_EXTRA_MATCH_LIMIT
1679 PCRE_EXTRA_MATCH_LIMIT_RECURSION
1680 PCRE_EXTRA_CALLOUT_DATA
1681 PCRE_EXTRA_TABLES
1682
1683 Other flag bits should be set to zero. The study_data field is set in
1684 the pcre_extra block that is returned by pcre_study(), together with
1685 the appropriate flag bit. You should not set this yourself, but you may
1686 add to the block by setting the other fields and their corresponding
1687 flag bits.
1688
1689 The match_limit field provides a means of preventing PCRE from using up
1690 a vast amount of resources when running patterns that are not going to
1691 match, but which have a very large number of possibilities in their
1692 search trees. The classic example is the use of nested unlimited
1693 repeats.
1694
1695 Internally, PCRE uses a function called match() which it calls repeat-
1696 edly (sometimes recursively). The limit set by match_limit is imposed
1697 on the number of times this function is called during a match, which
1698 has the effect of limiting the amount of backtracking that can take
1699 place. For patterns that are not anchored, the count restarts from zero
1700 for each position in the subject string.
1701
1702 The default value for the limit can be set when PCRE is built; the
1703 default default is 10 million, which handles all but the most extreme
1704 cases. You can override the default by suppling pcre_exec() with a
1705 pcre_extra block in which match_limit is set, and
1706 PCRE_EXTRA_MATCH_LIMIT is set in the flags field. If the limit is
1707 exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
1708
1709 The match_limit_recursion field is similar to match_limit, but instead
1710 of limiting the total number of times that match() is called, it limits
1711 the depth of recursion. The recursion depth is a smaller number than
1712 the total number of calls, because not all calls to match() are recur-
1713 sive. This limit is of use only if it is set smaller than match_limit.
1714
1715 Limiting the recursion depth limits the amount of stack that can be
1716 used, or, when PCRE has been compiled to use memory on the heap instead
1717 of the stack, the amount of heap memory that can be used.
1718
1719 The default value for match_limit_recursion can be set when PCRE is
1720 built; the default default is the same value as the default for
1721 match_limit. You can override the default by suppling pcre_exec() with
1722 a pcre_extra block in which match_limit_recursion is set, and
1723 PCRE_EXTRA_MATCH_LIMIT_RECURSION is set in the flags field. If the
1724 limit is exceeded, pcre_exec() returns PCRE_ERROR_RECURSIONLIMIT.
1725
1726 The pcre_callout field is used in conjunction with the "callout" fea-
1727 ture, which is described in the pcrecallout documentation.
1728
1729 The tables field is used to pass a character tables pointer to
1730 pcre_exec(); this overrides the value that is stored with the compiled
1731 pattern. A non-NULL value is stored with the compiled pattern only if
1732 custom tables were supplied to pcre_compile() via its tableptr argu-
1733 ment. If NULL is passed to pcre_exec() using this mechanism, it forces
1734 PCRE's internal tables to be used. This facility is helpful when re-
1735 using patterns that have been saved after compiling with an external
1736 set of tables, because the external tables might be at a different
1737 address when pcre_exec() is called. See the pcreprecompile documenta-
1738 tion for a discussion of saving compiled patterns for later use.
1739
1740 Option bits for pcre_exec()
1741
1742 The unused bits of the options argument for pcre_exec() must be zero.
1743 The only bits that may be set are PCRE_ANCHORED, PCRE_NEWLINE_xxx,
1744 PCRE_NOTBOL, PCRE_NOTEOL, PCRE_NOTEMPTY, PCRE_NO_UTF8_CHECK and
1745 PCRE_PARTIAL.
1746
1747 PCRE_ANCHORED
1748
1749 The PCRE_ANCHORED option limits pcre_exec() to matching at the first
1750 matching position. If a pattern was compiled with PCRE_ANCHORED, or
1751 turned out to be anchored by virtue of its contents, it cannot be made
1752 unachored at matching time.
1753
1754 PCRE_NEWLINE_CR
1755 PCRE_NEWLINE_LF
1756 PCRE_NEWLINE_CRLF
1757 PCRE_NEWLINE_ANYCRLF
1758 PCRE_NEWLINE_ANY
1759
1760 These options override the newline definition that was chosen or
1761 defaulted when the pattern was compiled. For details, see the descrip-
1762 tion of pcre_compile() above. During matching, the newline choice
1763 affects the behaviour of the dot, circumflex, and dollar metacharac-
1764 ters. It may also alter the way the match position is advanced after a
1765 match failure for an unanchored pattern. When PCRE_NEWLINE_CRLF,
1766 PCRE_NEWLINE_ANYCRLF, or PCRE_NEWLINE_ANY is set, and a match attempt
1767 fails when the current position is at a CRLF sequence, the match posi-
1768 tion is advanced by two characters instead of one, in other words, to
1769 after the CRLF.
1770
1771 PCRE_NOTBOL
1772
1773 This option specifies that first character of the subject string is not
1774 the beginning of a line, so the circumflex metacharacter should not
1775 match before it. Setting this without PCRE_MULTILINE (at compile time)
1776 causes circumflex never to match. This option affects only the behav-
1777 iour of the circumflex metacharacter. It does not affect \A.
1778
1779 PCRE_NOTEOL
1780
1781 This option specifies that the end of the subject string is not the end
1782 of a line, so the dollar metacharacter should not match it nor (except
1783 in multiline mode) a newline immediately before it. Setting this with-
1784 out PCRE_MULTILINE (at compile time) causes dollar never to match. This
1785 option affects only the behaviour of the dollar metacharacter. It does
1786 not affect \Z or \z.
1787
1788 PCRE_NOTEMPTY
1789
1790 An empty string is not considered to be a valid match if this option is
1791 set. If there are alternatives in the pattern, they are tried. If all
1792 the alternatives match the empty string, the entire match fails. For
1793 example, if the pattern
1794
1795 a?b?
1796
1797 is applied to a string not beginning with "a" or "b", it matches the
1798 empty string at the start of the subject. With PCRE_NOTEMPTY set, this
1799 match is not valid, so PCRE searches further into the string for occur-
1800 rences of "a" or "b".
1801
1802 Perl has no direct equivalent of PCRE_NOTEMPTY, but it does make a spe-
1803 cial case of a pattern match of the empty string within its split()
1804 function, and when using the /g modifier. It is possible to emulate
1805 Perl's behaviour after matching a null string by first trying the match
1806 again at the same offset with PCRE_NOTEMPTY and PCRE_ANCHORED, and then
1807 if that fails by advancing the starting offset (see below) and trying
1808 an ordinary match again. There is some code that demonstrates how to do
1809 this in the pcredemo.c sample program.
1810
1811 PCRE_NO_UTF8_CHECK
1812
1813 When PCRE_UTF8 is set at compile time, the validity of the subject as a
1814 UTF-8 string is automatically checked when pcre_exec() is subsequently
1815 called. The value of startoffset is also checked to ensure that it
1816 points to the start of a UTF-8 character. If an invalid UTF-8 sequence
1817 of bytes is found, pcre_exec() returns the error PCRE_ERROR_BADUTF8. If
1818 startoffset contains an invalid value, PCRE_ERROR_BADUTF8_OFFSET is
1819 returned.
1820
1821 If you already know that your subject is valid, and you want to skip
1822 these checks for performance reasons, you can set the
1823 PCRE_NO_UTF8_CHECK option when calling pcre_exec(). You might want to
1824 do this for the second and subsequent calls to pcre_exec() if you are
1825 making repeated calls to find all the matches in a single subject
1826 string. However, you should be sure that the value of startoffset
1827 points to the start of a UTF-8 character. When PCRE_NO_UTF8_CHECK is
1828 set, the effect of passing an invalid UTF-8 string as a subject, or a
1829 value of startoffset that does not point to the start of a UTF-8 char-
1830 acter, is undefined. Your program may crash.
1831
1832 PCRE_PARTIAL
1833
1834 This option turns on the partial matching feature. If the subject
1835 string fails to match the pattern, but at some point during the match-
1836 ing process the end of the subject was reached (that is, the subject
1837 partially matches the pattern and the failure to match occurred only
1838 because there were not enough subject characters), pcre_exec() returns
1839 PCRE_ERROR_PARTIAL instead of PCRE_ERROR_NOMATCH. When PCRE_PARTIAL is
1840 used, there are restrictions on what may appear in the pattern. These
1841 are discussed in the pcrepartial documentation.
1842
1843 The string to be matched by pcre_exec()
1844
1845 The subject string is passed to pcre_exec() as a pointer in subject, a
1846 length in length, and a starting byte offset in startoffset. In UTF-8
1847 mode, the byte offset must point to the start of a UTF-8 character.
1848 Unlike the pattern string, the subject may contain binary zero bytes.
1849 When the starting offset is zero, the search for a match starts at the
1850 beginning of the subject, and this is by far the most common case.
1851
1852 A non-zero starting offset is useful when searching for another match
1853 in the same subject by calling pcre_exec() again after a previous suc-
1854 cess. Setting startoffset differs from just passing over a shortened
1855 string and setting PCRE_NOTBOL in the case of a pattern that begins
1856 with any kind of lookbehind. For example, consider the pattern
1857
1858 \Biss\B
1859
1860 which finds occurrences of "iss" in the middle of words. (\B matches
1861 only if the current position in the subject is not a word boundary.)
1862 When applied to the string "Mississipi" the first call to pcre_exec()
1863 finds the first occurrence. If pcre_exec() is called again with just
1864 the remainder of the subject, namely "issipi", it does not match,
1865 because \B is always false at the start of the subject, which is deemed
1866 to be a word boundary. However, if pcre_exec() is passed the entire
1867 string again, but with startoffset set to 4, it finds the second occur-
1868 rence of "iss" because it is able to look behind the starting point to
1869 discover that it is preceded by a letter.
1870
1871 If a non-zero starting offset is passed when the pattern is anchored,
1872 one attempt to match at the given offset is made. This can only succeed
1873 if the pattern does not require the match to be at the start of the
1874 subject.
1875
1876 How pcre_exec() returns captured substrings
1877
1878 In general, a pattern matches a certain portion of the subject, and in
1879 addition, further substrings from the subject may be picked out by
1880 parts of the pattern. Following the usage in Jeffrey Friedl's book,
1881 this is called "capturing" in what follows, and the phrase "capturing
1882 subpattern" is used for a fragment of a pattern that picks out a sub-
1883 string. PCRE supports several other kinds of parenthesized subpattern
1884 that do not cause substrings to be captured.
1885
1886 Captured substrings are returned to the caller via a vector of integer
1887 offsets whose address is passed in ovector. The number of elements in
1888 the vector is passed in ovecsize, which must be a non-negative number.
1889 Note: this argument is NOT the size of ovector in bytes.
1890
1891 The first two-thirds of the vector is used to pass back captured sub-
1892 strings, each substring using a pair of integers. The remaining third
1893 of the vector is used as workspace by pcre_exec() while matching cap-
1894 turing subpatterns, and is not available for passing back information.
1895 The length passed in ovecsize should always be a multiple of three. If
1896 it is not, it is rounded down.
1897
1898 When a match is successful, information about captured substrings is
1899 returned in pairs of integers, starting at the beginning of ovector,
1900 and continuing up to two-thirds of its length at the most. The first
1901 element of a pair is set to the offset of the first character in a sub-
1902 string, and the second is set to the offset of the first character
1903 after the end of a substring. The first pair, ovector[0] and ovec-
1904 tor[1], identify the portion of the subject string matched by the
1905 entire pattern. The next pair is used for the first capturing subpat-
1906 tern, and so on. The value returned by pcre_exec() is one more than the
1907 highest numbered pair that has been set. For example, if two substrings
1908 have been captured, the returned value is 3. If there are no capturing
1909 subpatterns, the return value from a successful match is 1, indicating
1910 that just the first pair of offsets has been set.
1911
1912 If a capturing subpattern is matched repeatedly, it is the last portion
1913 of the string that it matched that is returned.
1914
1915 If the vector is too small to hold all the captured substring offsets,
1916 it is used as far as possible (up to two-thirds of its length), and the
1917 function returns a value of zero. In particular, if the substring off-
1918 sets are not of interest, pcre_exec() may be called with ovector passed
1919 as NULL and ovecsize as zero. However, if the pattern contains back
1920 references and the ovector is not big enough to remember the related
1921 substrings, PCRE has to get additional memory for use during matching.
1922 Thus it is usually advisable to supply an ovector.
1923
1924 The pcre_info() function can be used to find out how many capturing
1925 subpatterns there are in a compiled pattern. The smallest size for
1926 ovector that will allow for n captured substrings, in addition to the
1927 offsets of the substring matched by the whole pattern, is (n+1)*3.
1928
1929 It is possible for capturing subpattern number n+1 to match some part
1930 of the subject when subpattern n has not been used at all. For example,
1931 if the string "abc" is matched against the pattern (a|(z))(bc) the
1932 return from the function is 4, and subpatterns 1 and 3 are matched, but
1933 2 is not. When this happens, both values in the offset pairs corre-
1934 sponding to unused subpatterns are set to -1.
1935
1936 Offset values that correspond to unused subpatterns at the end of the
1937 expression are also set to -1. For example, if the string "abc" is
1938 matched against the pattern (abc)(x(yz)?)? subpatterns 2 and 3 are not
1939 matched. The return from the function is 2, because the highest used
1940 capturing subpattern number is 1. However, you can refer to the offsets
1941 for the second and third capturing subpatterns if you wish (assuming
1942 the vector is large enough, of course).
1943
1944 Some convenience functions are provided for extracting the captured
1945 substrings as separate strings. These are described below.
1946
1947 Error return values from pcre_exec()
1948
1949 If pcre_exec() fails, it returns a negative number. The following are
1950 defined in the header file:
1951
1952 PCRE_ERROR_NOMATCH (-1)
1953
1954 The subject string did not match the pattern.
1955
1956 PCRE_ERROR_NULL (-2)
1957
1958 Either code or subject was passed as NULL, or ovector was NULL and
1959 ovecsize was not zero.
1960
1961 PCRE_ERROR_BADOPTION (-3)
1962
1963 An unrecognized bit was set in the options argument.
1964
1965 PCRE_ERROR_BADMAGIC (-4)
1966
1967 PCRE stores a 4-byte "magic number" at the start of the compiled code,
1968 to catch the case when it is passed a junk pointer and to detect when a
1969 pattern that was compiled in an environment of one endianness is run in
1970 an environment with the other endianness. This is the error that PCRE
1971 gives when the magic number is not present.
1972
1973 PCRE_ERROR_UNKNOWN_OPCODE (-5)
1974
1975 While running the pattern match, an unknown item was encountered in the
1976 compiled pattern. This error could be caused by a bug in PCRE or by
1977 overwriting of the compiled pattern.
1978
1979 PCRE_ERROR_NOMEMORY (-6)
1980
1981 If a pattern contains back references, but the ovector that is passed
1982 to pcre_exec() is not big enough to remember the referenced substrings,
1983 PCRE gets a block of memory at the start of matching to use for this
1984 purpose. If the call via pcre_malloc() fails, this error is given. The
1985 memory is automatically freed at the end of matching.
1986
1987 PCRE_ERROR_NOSUBSTRING (-7)
1988
1989 This error is used by the pcre_copy_substring(), pcre_get_substring(),
1990 and pcre_get_substring_list() functions (see below). It is never
1991 returned by pcre_exec().
1992
1993 PCRE_ERROR_MATCHLIMIT (-8)
1994
1995 The backtracking limit, as specified by the match_limit field in a
1996 pcre_extra structure (or defaulted) was reached. See the description
1997 above.
1998
1999 PCRE_ERROR_CALLOUT (-9)
2000
2001 This error is never generated by pcre_exec() itself. It is provided for
2002 use by callout functions that want to yield a distinctive error code.
2003 See the pcrecallout documentation for details.
2004
2005 PCRE_ERROR_BADUTF8 (-10)
2006
2007 A string that contains an invalid UTF-8 byte sequence was passed as a
2008 subject.
2009
2010 PCRE_ERROR_BADUTF8_OFFSET (-11)
2011
2012 The UTF-8 byte sequence that was passed as a subject was valid, but the
2013 value of startoffset did not point to the beginning of a UTF-8 charac-
2014 ter.
2015
2016 PCRE_ERROR_PARTIAL (-12)
2017
2018 The subject string did not match, but it did match partially. See the
2019 pcrepartial documentation for details of partial matching.
2020
2021 PCRE_ERROR_BADPARTIAL (-13)
2022
2023 The PCRE_PARTIAL option was used with a compiled pattern containing
2024 items that are not supported for partial matching. See the pcrepartial
2025 documentation for details of partial matching.
2026
2027 PCRE_ERROR_INTERNAL (-14)
2028
2029 An unexpected internal error has occurred. This error could be caused
2030 by a bug in PCRE or by overwriting of the compiled pattern.
2031
2032 PCRE_ERROR_BADCOUNT (-15)
2033
2034 This error is given if the value of the ovecsize argument is negative.
2035
2036 PCRE_ERROR_RECURSIONLIMIT (-21)
2037
2038 The internal recursion limit, as specified by the match_limit_recursion
2039 field in a pcre_extra structure (or defaulted) was reached. See the
2040 description above.
2041
2042 PCRE_ERROR_NULLWSLIMIT (-22)
2043
2044 When a group that can match an empty substring is repeated with an
2045 unbounded upper limit, the subject position at the start of the group
2046 must be remembered, so that a test for an empty string can be made when
2047 the end of the group is reached. Some workspace is required for this;
2048 if it runs out, this error is given.
2049
2050 PCRE_ERROR_BADNEWLINE (-23)
2051
2052 An invalid combination of PCRE_NEWLINE_xxx options was given.
2053
2054 Error numbers -16 to -20 are not used by pcre_exec().
2055
2056
2057 EXTRACTING CAPTURED SUBSTRINGS BY NUMBER
2058
2059 int pcre_copy_substring(const char *subject, int *ovector,
2060 int stringcount, int stringnumber, char *buffer,
2061 int buffersize);
2062
2063 int pcre_get_substring(const char *subject, int *ovector,
2064 int stringcount, int stringnumber,
2065 const char **stringptr);
2066
2067 int pcre_get_substring_list(const char *subject,
2068 int *ovector, int stringcount, const char ***listptr);
2069
2070 Captured substrings can be accessed directly by using the offsets
2071 returned by pcre_exec() in ovector. For convenience, the functions
2072 pcre_copy_substring(), pcre_get_substring(), and pcre_get_sub-
2073 string_list() are provided for extracting captured substrings as new,
2074 separate, zero-terminated strings. These functions identify substrings
2075 by number. The next section describes functions for extracting named
2076 substrings.
2077
2078 A substring that contains a binary zero is correctly extracted and has
2079 a further zero added on the end, but the result is not, of course, a C
2080 string. However, you can process such a string by referring to the
2081 length that is returned by pcre_copy_substring() and pcre_get_sub-
2082 string(). Unfortunately, the interface to pcre_get_substring_list() is
2083 not adequate for handling strings containing binary zeros, because the
2084 end of the final string is not independently indicated.
2085
2086 The first three arguments are the same for all three of these func-
2087 tions: subject is the subject string that has just been successfully
2088 matched, ovector is a pointer to the vector of integer offsets that was
2089 passed to pcre_exec(), and stringcount is the number of substrings that
2090 were captured by the match, including the substring that matched the
2091 entire regular expression. This is the value returned by pcre_exec() if
2092 it is greater than zero. If pcre_exec() returned zero, indicating that
2093 it ran out of space in ovector, the value passed as stringcount should
2094 be the number of elements in the vector divided by three.
2095
2096 The functions pcre_copy_substring() and pcre_get_substring() extract a
2097 single substring, whose number is given as stringnumber. A value of
2098 zero extracts the substring that matched the entire pattern, whereas
2099 higher values extract the captured substrings. For pcre_copy_sub-
2100 string(), the string is placed in buffer, whose length is given by
2101 buffersize, while for pcre_get_substring() a new block of memory is
2102 obtained via pcre_malloc, and its address is returned via stringptr.
2103 The yield of the function is the length of the string, not including
2104 the terminating zero, or one of these error codes:
2105
2106 PCRE_ERROR_NOMEMORY (-6)
2107
2108 The buffer was too small for pcre_copy_substring(), or the attempt to
2109 get memory failed for pcre_get_substring().
2110
2111 PCRE_ERROR_NOSUBSTRING (-7)
2112
2113 There is no substring whose number is stringnumber.
2114
2115 The pcre_get_substring_list() function extracts all available sub-
2116 strings and builds a list of pointers to them. All this is done in a
2117 single block of memory that is obtained via pcre_malloc. The address of
2118 the memory block is returned via listptr, which is also the start of
2119 the list of string pointers. The end of the list is marked by a NULL
2120 pointer. The yield of the function is zero if all went well, or the
2121 error code
2122
2123 PCRE_ERROR_NOMEMORY (-6)
2124
2125 if the attempt to get the memory block failed.
2126
2127 When any of these functions encounter a substring that is unset, which
2128 can happen when capturing subpattern number n+1 matches some part of
2129 the subject, but subpattern n has not been used at all, they return an
2130 empty string. This can be distinguished from a genuine zero-length sub-
2131 string by inspecting the appropriate offset in ovector, which is nega-
2132 tive for unset substrings.
2133
2134 The two convenience functions pcre_free_substring() and pcre_free_sub-
2135 string_list() can be used to free the memory returned by a previous
2136 call of pcre_get_substring() or pcre_get_substring_list(), respec-
2137 tively. They do nothing more than call the function pointed to by
2138 pcre_free, which of course could be called directly from a C program.
2139 However, PCRE is used in some situations where it is linked via a spe-
2140 cial interface to another programming language that cannot use
2141 pcre_free directly; it is for these cases that the functions are pro-
2142 vided.
2143
2144
2145 EXTRACTING CAPTURED SUBSTRINGS BY NAME
2146
2147 int pcre_get_stringnumber(const pcre *code,
2148 const char *name);
2149
2150 int pcre_copy_named_substring(const pcre *code,
2151 const char *subject, int *ovector,
2152 int stringcount, const char *stringname,
2153 char *buffer, int buffersize);
2154
2155 int pcre_get_named_substring(const pcre *code,
2156 const char *subject, int *ovector,
2157 int stringcount, const char *stringname,
2158 const char **stringptr);
2159
2160 To extract a substring by name, you first have to find associated num-
2161 ber. For example, for this pattern
2162
2163 (a+)b(?<xxx>\d+)...
2164
2165 the number of the subpattern called "xxx" is 2. If the name is known to
2166 be unique (PCRE_DUPNAMES was not set), you can find the number from the
2167 name by calling pcre_get_stringnumber(). The first argument is the com-
2168 piled pattern, and the second is the name. The yield of the function is
2169 the subpattern number, or PCRE_ERROR_NOSUBSTRING (-7) if there is no
2170 subpattern of that name.
2171
2172 Given the number, you can extract the substring directly, or use one of
2173 the functions described in the previous section. For convenience, there
2174 are also two functions that do the whole job.
2175
2176 Most of the arguments of pcre_copy_named_substring() and
2177 pcre_get_named_substring() are the same as those for the similarly
2178 named functions that extract by number. As these are described in the
2179 previous section, they are not re-described here. There are just two
2180 differences:
2181
2182 First, instead of a substring number, a substring name is given. Sec-
2183 ond, there is an extra argument, given at the start, which is a pointer
2184 to the compiled pattern. This is needed in order to gain access to the
2185 name-to-number translation table.
2186
2187 These functions call pcre_get_stringnumber(), and if it succeeds, they
2188 then call pcre_copy_substring() or pcre_get_substring(), as appropri-
2189 ate. NOTE: If PCRE_DUPNAMES is set and there are duplicate names, the
2190 behaviour may not be what you want (see the next section).
2191
2192
2193 DUPLICATE SUBPATTERN NAMES
2194
2195 int pcre_get_stringtable_entries(const pcre *code,
2196 const char *name, char **first, char **last);
2197
2198 When a pattern is compiled with the PCRE_DUPNAMES option, names for
2199 subpatterns are not required to be unique. Normally, patterns with
2200 duplicate names are such that in any one match, only one of the named
2201 subpatterns participates. An example is shown in the pcrepattern docu-
2202 mentation. When duplicates are present, pcre_copy_named_substring() and
2203 pcre_get_named_substring() return the first substring corresponding to
2204 the given name that is set. If none are set, an empty string is
2205 returned. The pcre_get_stringnumber() function returns one of the num-
2206 bers that are associated with the name, but it is not defined which it
2207 is.
2208
2209 If you want to get full details of all captured substrings for a given
2210 name, you must use the pcre_get_stringtable_entries() function. The
2211 first argument is the compiled pattern, and the second is the name. The
2212 third and fourth are pointers to variables which are updated by the
2213 function. After it has run, they point to the first and last entries in
2214 the name-to-number table for the given name. The function itself
2215 returns the length of each entry, or PCRE_ERROR_NOSUBSTRING (-7) if
2216 there are none. The format of the table is described above in the sec-
2217 tion entitled Information about a pattern. Given all the relevant
2218 entries for the name, you can extract each of their numbers, and hence
2219 the captured data, if any.
2220
2221
2222 FINDING ALL POSSIBLE MATCHES
2223
2224 The traditional matching function uses a similar algorithm to Perl,
2225 which stops when it finds the first match, starting at a given point in
2226 the subject. If you want to find all possible matches, or the longest
2227 possible match, consider using the alternative matching function (see
2228 below) instead. If you cannot use the alternative function, but still
2229 need to find all possible matches, you can kludge it up by making use
2230 of the callout facility, which is described in the pcrecallout documen-
2231 tation.
2232
2233 What you have to do is to insert a callout right at the end of the pat-
2234 tern. When your callout function is called, extract and save the cur-
2235 rent matched substring. Then return 1, which forces pcre_exec() to
2236 backtrack and try other alternatives. Ultimately, when it runs out of
2237 matches, pcre_exec() will yield PCRE_ERROR_NOMATCH.
2238
2239
2240 MATCHING A PATTERN: THE ALTERNATIVE FUNCTION
2241
2242 int pcre_dfa_exec(const pcre *code, const pcre_extra *extra,
2243 const char *subject, int length, int startoffset,
2244 int options, int *ovector, int ovecsize,
2245 int *workspace, int wscount);
2246
2247 The function pcre_dfa_exec() is called to match a subject string
2248 against a compiled pattern, using a matching algorithm that scans the
2249 subject string just once, and does not backtrack. This has different
2250 characteristics to the normal algorithm, and is not compatible with
2251 Perl. Some of the features of PCRE patterns are not supported. Never-
2252 theless, there are times when this kind of matching can be useful. For
2253 a discussion of the two matching algorithms, see the pcrematching docu-
2254 mentation.
2255
2256 The arguments for the pcre_dfa_exec() function are the same as for
2257 pcre_exec(), plus two extras. The ovector argument is used in a differ-
2258 ent way, and this is described below. The other common arguments are
2259 used in the same way as for pcre_exec(), so their description is not
2260 repeated here.
2261
2262 The two additional arguments provide workspace for the function. The
2263 workspace vector should contain at least 20 elements. It is used for
2264 keeping track of multiple paths through the pattern tree. More
2265 workspace will be needed for patterns and subjects where there are a
2266 lot of potential matches.
2267
2268 Here is an example of a simple call to pcre_dfa_exec():
2269
2270 int rc;
2271 int ovector[10];
2272 int wspace[20];
2273 rc = pcre_dfa_exec(
2274 re, /* result of pcre_compile() */
2275 NULL, /* we didn't study the pattern */
2276 "some string", /* the subject string */
2277 11, /* the length of the subject string */
2278 0, /* start at offset 0 in the subject */
2279 0, /* default options */
2280 ovector, /* vector of integers for substring information */
2281 10, /* number of elements (NOT size in bytes) */
2282 wspace, /* working space vector */
2283 20); /* number of elements (NOT size in bytes) */
2284
2285 Option bits for pcre_dfa_exec()
2286
2287 The unused bits of the options argument for pcre_dfa_exec() must be
2288 zero. The only bits that may be set are PCRE_ANCHORED, PCRE_NEW-
2289 LINE_xxx, PCRE_NOTBOL, PCRE_NOTEOL, PCRE_NOTEMPTY, PCRE_NO_UTF8_CHECK,
2290 PCRE_PARTIAL, PCRE_DFA_SHORTEST, and PCRE_DFA_RESTART. All but the last
2291 three of these are the same as for pcre_exec(), so their description is
2292 not repeated here.
2293
2294 PCRE_PARTIAL
2295
2296 This has the same general effect as it does for pcre_exec(), but the
2297 details are slightly different. When PCRE_PARTIAL is set for
2298 pcre_dfa_exec(), the return code PCRE_ERROR_NOMATCH is converted into
2299 PCRE_ERROR_PARTIAL if the end of the subject is reached, there have
2300 been no complete matches, but there is still at least one matching pos-
2301 sibility. The portion of the string that provided the partial match is
2302 set as the first matching string.
2303
2304 PCRE_DFA_SHORTEST
2305
2306 Setting the PCRE_DFA_SHORTEST option causes the matching algorithm to
2307 stop as soon as it has found one match. Because of the way the alterna-
2308 tive algorithm works, this is necessarily the shortest possible match
2309 at the first possible matching point in the subject string.
2310
2311 PCRE_DFA_RESTART
2312
2313 When pcre_dfa_exec() is called with the PCRE_PARTIAL option, and
2314 returns a partial match, it is possible to call it again, with addi-
2315 tional subject characters, and have it continue with the same match.
2316 The PCRE_DFA_RESTART option requests this action; when it is set, the
2317 workspace and wscount options must reference the same vector as before
2318 because data about the match so far is left in them after a partial
2319 match. There is more discussion of this facility in the pcrepartial
2320 documentation.
2321
2322 Successful returns from pcre_dfa_exec()
2323
2324 When pcre_dfa_exec() succeeds, it may have matched more than one sub-
2325 string in the subject. Note, however, that all the matches from one run
2326 of the function start at the same point in the subject. The shorter
2327 matches are all initial substrings of the longer matches. For example,
2328 if the pattern
2329
2330 <.*>
2331
2332 is matched against the string
2333
2334 This is <something> <something else> <something further> no more
2335
2336 the three matched strings are
2337
2338 <something>
2339 <something> <something else>
2340 <something> <something else> <something further>
2341
2342 On success, the yield of the function is a number greater than zero,
2343 which is the number of matched substrings. The substrings themselves
2344 are returned in ovector. Each string uses two elements; the first is
2345 the offset to the start, and the second is the offset to the end. In
2346 fact, all the strings have the same start offset. (Space could have
2347 been saved by giving this only once, but it was decided to retain some
2348 compatibility with the way pcre_exec() returns data, even though the
2349 meaning of the strings is different.)
2350
2351 The strings are returned in reverse order of length; that is, the long-
2352 est matching string is given first. If there were too many matches to
2353 fit into ovector, the yield of the function is zero, and the vector is
2354 filled with the longest matches.
2355
2356 Error returns from pcre_dfa_exec()
2357
2358 The pcre_dfa_exec() function returns a negative number when it fails.
2359 Many of the errors are the same as for pcre_exec(), and these are
2360 described above. There are in addition the following errors that are
2361 specific to pcre_dfa_exec():
2362
2363 PCRE_ERROR_DFA_UITEM (-16)
2364
2365 This return is given if pcre_dfa_exec() encounters an item in the pat-
2366 tern that it does not support, for instance, the use of \C or a back
2367 reference.
2368
2369 PCRE_ERROR_DFA_UCOND (-17)
2370
2371 This return is given if pcre_dfa_exec() encounters a condition item
2372 that uses a back reference for the condition, or a test for recursion
2373 in a specific group. These are not supported.
2374
2375 PCRE_ERROR_DFA_UMLIMIT (-18)
2376
2377 This return is given if pcre_dfa_exec() is called with an extra block
2378 that contains a setting of the match_limit field. This is not supported
2379 (it is meaningless).
2380
2381 PCRE_ERROR_DFA_WSSIZE (-19)
2382
2383 This return is given if pcre_dfa_exec() runs out of space in the
2384 workspace vector.
2385
2386 PCRE_ERROR_DFA_RECURSE (-20)
2387
2388 When a recursive subpattern is processed, the matching function calls
2389 itself recursively, using private vectors for ovector and workspace.
2390 This error is given if the output vector is not large enough. This
2391 should be extremely rare, as a vector of size 1000 is used.
2392
2393
2394 SEE ALSO
2395
2396 pcrebuild(3), pcrecallout(3), pcrecpp(3)(3), pcrematching(3), pcrepar-
2397 tial(3), pcreposix(3), pcreprecompile(3), pcresample(3), pcrestack(3).
2398
2399
2400 AUTHOR
2401
2402 Philip Hazel
2403 University Computing Service
2404 Cambridge CB2 3QH, England.
2405
2406
2407 REVISION
2408
2409 Last updated: 04 June 2007
2410 Copyright (c) 1997-2007 University of Cambridge.
2411 ------------------------------------------------------------------------------
2412
2413
2414 PCRECALLOUT(3) PCRECALLOUT(3)
2415
2416
2417 NAME
2418 PCRE - Perl-compatible regular expressions
2419
2420
2421 PCRE CALLOUTS
2422
2423 int (*pcre_callout)(pcre_callout_block *);
2424
2425 PCRE provides a feature called "callout", which is a means of temporar-
2426 ily passing control to the caller of PCRE in the middle of pattern
2427 matching. The caller of PCRE provides an external function by putting
2428 its entry point in the global variable pcre_callout. By default, this
2429 variable contains NULL, which disables all calling out.
2430
2431 Within a regular expression, (?C) indicates the points at which the
2432 external function is to be called. Different callout points can be
2433 identified by putting a number less than 256 after the letter C. The
2434 default value is zero. For example, this pattern has two callout
2435 points:
2436
2437 (?C1)abc(?C2)def
2438
2439 If the PCRE_AUTO_CALLOUT option bit is set when pcre_compile() is
2440 called, PCRE automatically inserts callouts, all with number 255,
2441 before each item in the pattern. For example, if PCRE_AUTO_CALLOUT is
2442 used with the pattern
2443
2444 A(\d{2}|--)
2445
2446 it is processed as if it were
2447
2448 (?C255)A(?C255)((?C255)\d{2}(?C255)|(?C255)-(?C255)-(?C255))(?C255)
2449
2450 Notice that there is a callout before and after each parenthesis and
2451 alternation bar. Automatic callouts can be used for tracking the
2452 progress of pattern matching. The pcretest command has an option that
2453 sets automatic callouts; when it is used, the output indicates how the
2454 pattern is matched. This is useful information when you are trying to
2455 optimize the performance of a particular pattern.
2456
2457
2458 MISSING CALLOUTS
2459
2460 You should be aware that, because of optimizations in the way PCRE
2461 matches patterns, callouts sometimes do not happen. For example, if the
2462 pattern is
2463
2464 ab(?C4)cd
2465
2466 PCRE knows that any matching string must contain the letter "d". If the
2467 subject string is "abyz", the lack of "d" means that matching doesn't
2468 ever start, and the callout is never reached. However, with "abyd",
2469 though the result is still no match, the callout is obeyed.
2470
2471
2472 THE CALLOUT INTERFACE
2473
2474 During matching, when PCRE reaches a callout point, the external func-
2475 tion defined by pcre_callout is called (if it is set). This applies to
2476 both the pcre_exec() and the pcre_dfa_exec() matching functions. The
2477 only argument to the callout function is a pointer to a pcre_callout
2478 block. This structure contains the following fields:
2479
2480 int version;
2481 int callout_number;
2482 int *offset_vector;
2483 const char *subject;
2484 int subject_length;
2485 int start_match;
2486 int current_position;
2487 int capture_top;
2488 int capture_last;
2489 void *callout_data;
2490 int pattern_position;
2491 int next_item_length;
2492
2493 The version field is an integer containing the version number of the
2494 block format. The initial version was 0; the current version is 1. The
2495 version number will change again in future if additional fields are
2496 added, but the intention is never to remove any of the existing fields.
2497
2498 The callout_number field contains the number of the callout, as com-
2499 piled into the pattern (that is, the number after ?C for manual call-
2500 outs, and 255 for automatically generated callouts).
2501
2502 The offset_vector field is a pointer to the vector of offsets that was
2503 passed by the caller to pcre_exec() or pcre_dfa_exec(). When
2504 pcre_exec() is used, the contents can be inspected in order to extract
2505 substrings that have been matched so far, in the same way as for
2506 extracting substrings after a match has completed. For pcre_dfa_exec()
2507 this field is not useful.
2508
2509 The subject and subject_length fields contain copies of the values that
2510 were passed to pcre_exec().
2511
2512 The start_match field normally contains the offset within the subject
2513 at which the current match attempt started. However, if the escape
2514 sequence \K has been encountered, this value is changed to reflect the
2515 modified starting point. If the pattern is not anchored, the callout
2516 function may be called several times from the same point in the pattern
2517 for different starting points in the subject.
2518
2519 The current_position field contains the offset within the subject of
2520 the current match pointer.
2521
2522 When the pcre_exec() function is used, the capture_top field contains
2523 one more than the number of the highest numbered captured substring so
2524 far. If no substrings have been captured, the value of capture_top is
2525 one. This is always the case when pcre_dfa_exec() is used, because it
2526 does not support captured substrings.
2527
2528 The capture_last field contains the number of the most recently cap-
2529 tured substring. If no substrings have been captured, its value is -1.
2530 This is always the case when pcre_dfa_exec() is used.
2531
2532 The callout_data field contains a value that is passed to pcre_exec()
2533 or pcre_dfa_exec() specifically so that it can be passed back in call-
2534 outs. It is passed in the pcre_callout field of the pcre_extra data
2535 structure. If no such data was passed, the value of callout_data in a
2536 pcre_callout block is NULL. There is a description of the pcre_extra
2537 structure in the pcreapi documentation.
2538
2539 The pattern_position field is present from version 1 of the pcre_call-
2540 out structure. It contains the offset to the next item to be matched in
2541 the pattern string.
2542
2543 The next_item_length field is present from version 1 of the pcre_call-
2544 out structure. It contains the length of the next item to be matched in
2545 the pattern string. When the callout immediately precedes an alterna-
2546 tion bar, a closing parenthesis, or the end of the pattern, the length
2547 is zero. When the callout precedes an opening parenthesis, the length
2548 is that of the entire subpattern.
2549
2550 The pattern_position and next_item_length fields are intended to help
2551 in distinguishing between different automatic callouts, which all have
2552 the same callout number. However, they are set for all callouts.
2553
2554
2555 RETURN VALUES
2556
2557 The external callout function returns an integer to PCRE. If the value
2558 is zero, matching proceeds as normal. If the value is greater than
2559 zero, matching fails at the current point, but the testing of other
2560 matching possibilities goes ahead, just as if a lookahead assertion had
2561 failed. If the value is less than zero, the match is abandoned, and
2562 pcre_exec() (or pcre_dfa_exec()) returns the negative value.
2563
2564 Negative values should normally be chosen from the set of
2565 PCRE_ERROR_xxx values. In particular, PCRE_ERROR_NOMATCH forces a stan-
2566 dard "no match" failure. The error number PCRE_ERROR_CALLOUT is
2567 reserved for use by callout functions; it will never be used by PCRE
2568 itself.
2569
2570
2571 AUTHOR
2572
2573 Philip Hazel
2574 University Computing Service
2575 Cambridge CB2 3QH, England.
2576
2577
2578 REVISION
2579
2580 Last updated: 29 May 2007
2581 Copyright (c) 1997-2007 University of Cambridge.
2582 ------------------------------------------------------------------------------
2583
2584
2585 PCRECOMPAT(3) PCRECOMPAT(3)
2586
2587
2588 NAME
2589 PCRE - Perl-compatible regular expressions
2590
2591
2592 DIFFERENCES BETWEEN PCRE AND PERL
2593
2594 This document describes the differences in the ways that PCRE and Perl
2595 handle regular expressions. The differences described here are mainly
2596 with respect to Perl 5.8, though PCRE version 7.0 contains some fea-
2597 tures that are expected to be in the forthcoming Perl 5.10.
2598
2599 1. PCRE has only a subset of Perl's UTF-8 and Unicode support. Details
2600 of what it does have are given in the section on UTF-8 support in the
2601 main pcre page.
2602
2603 2. PCRE does not allow repeat quantifiers on lookahead assertions. Perl
2604 permits them, but they do not mean what you might think. For example,
2605 (?!a){3} does not assert that the next three characters are not "a". It
2606 just asserts that the next character is not "a" three times.
2607
2608 3. Capturing subpatterns that occur inside negative lookahead asser-
2609 tions are counted, but their entries in the offsets vector are never
2610 set. Perl sets its numerical variables from any such patterns that are
2611 matched before the assertion fails to match something (thereby succeed-
2612 ing), but only if the negative lookahead assertion contains just one
2613 branch.
2614
2615 4. Though binary zero characters are supported in the subject string,
2616 they are not allowed in a pattern string because it is passed as a nor-
2617 mal C string, terminated by zero. The escape sequence \0 can be used in
2618 the pattern to represent a binary zero.
2619
2620 5. The following Perl escape sequences are not supported: \l, \u, \L,
2621 \U, and \N. In fact these are implemented by Perl's general string-han-
2622 dling and are not part of its pattern matching engine. If any of these
2623 are encountered by PCRE, an error is generated.
2624
2625 6. The Perl escape sequences \p, \P, and \X are supported only if PCRE
2626 is built with Unicode character property support. The properties that
2627 can be tested with \p and \P are limited to the general category prop-
2628 erties such as Lu and Nd, script names such as Greek or Han, and the
2629 derived properties Any and L&.
2630
2631 7. PCRE does support the \Q...\E escape for quoting substrings. Charac-
2632 ters in between are treated as literals. This is slightly different
2633 from Perl in that $ and @ are also handled as literals inside the
2634 quotes. In Perl, they cause variable interpolation (but of course PCRE
2635 does not have variables). Note the following examples:
2636
2637 Pattern PCRE matches Perl matches
2638
2639 \Qabc$xyz\E abc$xyz abc followed by the
2640 contents of $xyz
2641 \Qabc\$xyz\E abc\$xyz abc\$xyz
2642 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
2643
2644 The \Q...\E sequence is recognized both inside and outside character
2645 classes.
2646
2647 8. Fairly obviously, PCRE does not support the (?{code}) and (??{code})
2648 constructions. However, there is support for recursive patterns. This
2649 is not available in Perl 5.8, but will be in Perl 5.10. Also, the PCRE
2650 "callout" feature allows an external function to be called during pat-
2651 tern matching. See the pcrecallout documentation for details.
2652
2653 9. Subpatterns that are called recursively or as "subroutines" are
2654 always treated as atomic groups in PCRE. This is like Python, but
2655 unlike Perl.
2656
2657 10. There are some differences that are concerned with the settings of
2658 captured strings when part of a pattern is repeated. For example,
2659 matching "aba" against the pattern /^(a(b)?)+$/ in Perl leaves $2
2660 unset, but in PCRE it is set to "b".
2661
2662 11. PCRE provides some extensions to the Perl regular expression facil-
2663 ities. Perl 5.10 will include new features that are not in earlier
2664 versions, some of which (such as named parentheses) have been in PCRE
2665 for some time. This list is with respect to Perl 5.10:
2666
2667 (a) Although lookbehind assertions must match fixed length strings,
2668 each alternative branch of a lookbehind assertion can match a different
2669 length of string. Perl requires them all to have the same length.
2670
2671 (b) If PCRE_DOLLAR_ENDONLY is set and PCRE_MULTILINE is not set, the $
2672 meta-character matches only at the very end of the string.
2673
2674 (c) If PCRE_EXTRA is set, a backslash followed by a letter with no spe-
2675 cial meaning is faulted. Otherwise, like Perl, the backslash is
2676 ignored. (Perl can be made to issue a warning.)
2677
2678 (d) If PCRE_UNGREEDY is set, the greediness of the repetition quanti-
2679 fiers is inverted, that is, by default they are not greedy, but if fol-
2680 lowed by a question mark they are.
2681
2682 (e) PCRE_ANCHORED can be used at matching time to force a pattern to be
2683 tried only at the first matching position in the subject string.
2684
2685 (f) The PCRE_NOTBOL, PCRE_NOTEOL, PCRE_NOTEMPTY, and PCRE_NO_AUTO_CAP-
2686 TURE options for pcre_exec() have no Perl equivalents.
2687
2688 (g) The callout facility is PCRE-specific.
2689
2690 (h) The partial matching facility is PCRE-specific.
2691
2692 (i) Patterns compiled by PCRE can be saved and re-used at a later time,
2693 even on different hosts that have the other endianness.
2694
2695 (j) The alternative matching function (pcre_dfa_exec()) matches in a
2696 different way and is not Perl-compatible.
2697
2698
2699 AUTHOR
2700
2701 Philip Hazel
2702 University Computing Service
2703 Cambridge CB2 3QH, England.
2704
2705
2706 REVISION
2707
2708 Last updated: 06 March 2007
2709 Copyright (c) 1997-2007 University of Cambridge.
2710 ------------------------------------------------------------------------------
2711
2712
2713 PCREPATTERN(3) PCREPATTERN(3)
2714
2715
2716 NAME
2717 PCRE - Perl-compatible regular expressions
2718
2719
2720 PCRE REGULAR EXPRESSION DETAILS
2721
2722 The syntax and semantics of the regular expressions supported by PCRE
2723 are described below. Regular expressions are also described in the Perl
2724 documentation and in a number of books, some of which have copious
2725 examples. Jeffrey Friedl's "Mastering Regular Expressions", published
2726 by O'Reilly, covers regular expressions in great detail. This descrip-
2727 tion of PCRE's regular expressions is intended as reference material.
2728
2729 The original operation of PCRE was on strings of one-byte characters.
2730 However, there is now also support for UTF-8 character strings. To use
2731 this, you must build PCRE to include UTF-8 support, and then call
2732 pcre_compile() with the PCRE_UTF8 option. How this affects pattern
2733 matching is mentioned in several places below. There is also a summary
2734 of UTF-8 features in the section on UTF-8 support in the main pcre
2735 page.
2736
2737 The remainder of this document discusses the patterns that are sup-
2738 ported by PCRE when its main matching function, pcre_exec(), is used.
2739 From release 6.0, PCRE offers a second matching function,
2740 pcre_dfa_exec(), which matches using a different algorithm that is not
2741 Perl-compatible. Some of the features discussed below are not available
2742 when pcre_dfa_exec() is used. The advantages and disadvantages of the
2743 alternative function, and how it differs from the normal function, are
2744 discussed in the pcrematching page.
2745
2746
2747 CHARACTERS AND METACHARACTERS
2748
2749 A regular expression is a pattern that is matched against a subject
2750 string from left to right. Most characters stand for themselves in a
2751 pattern, and match the corresponding characters in the subject. As a
2752 trivial example, the pattern
2753
2754 The quick brown fox
2755
2756 matches a portion of a subject string that is identical to itself. When
2757 caseless matching is specified (the PCRE_CASELESS option), letters are
2758 matched independently of case. In UTF-8 mode, PCRE always understands
2759 the concept of case for characters whose values are less than 128, so
2760 caseless matching is always possible. For characters with higher val-
2761 ues, the concept of case is supported if PCRE is compiled with Unicode
2762 property support, but not otherwise. If you want to use caseless
2763 matching for characters 128 and above, you must ensure that PCRE is
2764 compiled with Unicode property support as well as with UTF-8 support.
2765
2766 The power of regular expressions comes from the ability to include
2767 alternatives and repetitions in the pattern. These are encoded in the
2768 pattern by the use of metacharacters, which do not stand for themselves
2769 but instead are interpreted in some special way.
2770
2771 There are two different sets of metacharacters: those that are recog-
2772 nized anywhere in the pattern except within square brackets, and those
2773 that are recognized within square brackets. Outside square brackets,
2774 the metacharacters are as follows:
2775
2776 \ general escape character with several uses
2777 ^ assert start of string (or line, in multiline mode)
2778 $ assert end of string (or line, in multiline mode)
2779 . match any character except newline (by default)
2780 [ start character class definition
2781 | start of alternative branch
2782 ( start subpattern
2783 ) end subpattern
2784 ? extends the meaning of (
2785 also 0 or 1 quantifier
2786 also quantifier minimizer
2787 * 0 or more quantifier
2788 + 1 or more quantifier
2789 also "possessive quantifier"
2790 { start min/max quantifier
2791
2792 Part of a pattern that is in square brackets is called a "character
2793 class". In a character class the only metacharacters are:
2794
2795 \ general escape character
2796 ^ negate the class, but only if the first character
2797 - indicates character range
2798 [ POSIX character class (only if followed by POSIX
2799 syntax)
2800 ] terminates the character class
2801
2802 The following sections describe the use of each of the metacharacters.
2803
2804
2805 BACKSLASH
2806
2807 The backslash character has several uses. Firstly, if it is followed by
2808 a non-alphanumeric character, it takes away any special meaning that
2809 character may have. This use of backslash as an escape character
2810 applies both inside and outside character classes.
2811
2812 For example, if you want to match a * character, you write \* in the
2813 pattern. This escaping action applies whether or not the following
2814 character would otherwise be interpreted as a metacharacter, so it is
2815 always safe to precede a non-alphanumeric with backslash to specify
2816 that it stands for itself. In particular, if you want to match a back-
2817 slash, you write \\.
2818
2819 If a pattern is compiled with the PCRE_EXTENDED option, whitespace in
2820 the pattern (other than in a character class) and characters between a
2821 # outside a character class and the next newline are ignored. An escap-
2822 ing backslash can be used to include a whitespace or # character as
2823 part of the pattern.
2824
2825 If you want to remove the special meaning from a sequence of charac-
2826 ters, you can do so by putting them between \Q and \E. This is differ-
2827 ent from Perl in that $ and @ are handled as literals in \Q...\E
2828 sequences in PCRE, whereas in Perl, $ and @ cause variable interpola-
2829 tion. Note the following examples:
2830
2831 Pattern PCRE matches Perl matches
2832
2833 \Qabc$xyz\E abc$xyz abc followed by the
2834 contents of $xyz
2835 \Qabc\$xyz\E abc\$xyz abc\$xyz
2836 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
2837
2838 The \Q...\E sequence is recognized both inside and outside character
2839 classes.
2840
2841 Non-printing characters
2842
2843 A second use of backslash provides a way of encoding non-printing char-
2844 acters in patterns in a visible manner. There is no restriction on the
2845 appearance of non-printing characters, apart from the binary zero that
2846 terminates a pattern, but when a pattern is being prepared by text
2847 editing, it is usually easier to use one of the following escape
2848 sequences than the binary character it represents:
2849
2850 \a alarm, that is, the BEL character (hex 07)
2851 \cx "control-x", where x is any character
2852 \e escape (hex 1B)
2853 \f formfeed (hex 0C)
2854 \n newline (hex 0A)
2855 \r carriage return (hex 0D)
2856 \t tab (hex 09)
2857 \ddd character with octal code ddd, or backreference
2858 \xhh character with hex code hh
2859 \x{hhh..} character with hex code hhh..
2860
2861 The precise effect of \cx is as follows: if x is a lower case letter,
2862 it is converted to upper case. Then bit 6 of the character (hex 40) is
2863 inverted. Thus \cz becomes hex 1A, but \c{ becomes hex 3B, while \c;
2864 becomes hex 7B.
2865
2866 After \x, from zero to two hexadecimal digits are read (letters can be
2867 in upper or lower case). Any number of hexadecimal digits may appear
2868 between \x{ and }, but the value of the character code must be less
2869 than 256 in non-UTF-8 mode, and less than 2**31 in UTF-8 mode (that is,
2870 the maximum hexadecimal value is 7FFFFFFF). If characters other than
2871 hexadecimal digits appear between \x{ and }, or if there is no termi-
2872 nating }, this form of escape is not recognized. Instead, the initial
2873 \x will be interpreted as a basic hexadecimal escape, with no following
2874 digits, giving a character whose value is zero.
2875
2876 Characters whose value is less than 256 can be defined by either of the
2877 two syntaxes for \x. There is no difference in the way they are han-
2878 dled. For example, \xdc is exactly the same as \x{dc}.
2879
2880 After \0 up to two further octal digits are read. If there are fewer
2881 than two digits, just those that are present are used. Thus the
2882 sequence \0\x\07 specifies two binary zeros followed by a BEL character
2883 (code value 7). Make sure you supply two digits after the initial zero
2884 if the pattern character that follows is itself an octal digit.
2885
2886 The handling of a backslash followed by a digit other than 0 is compli-
2887 cated. Outside a character class, PCRE reads it and any following dig-
2888 its as a decimal number. If the number is less than 10, or if there
2889 have been at least that many previous capturing left parentheses in the
2890 expression, the entire sequence is taken as a back reference. A
2891 description of how this works is given later, following the discussion
2892 of parenthesized subpatterns.
2893
2894 Inside a character class, or if the decimal number is greater than 9
2895 and there have not been that many capturing subpatterns, PCRE re-reads
2896 up to three octal digits following the backslash, and uses them to gen-
2897 erate a data character. Any subsequent digits stand for themselves. In
2898 non-UTF-8 mode, the value of a character specified in octal must be
2899 less than \400. In UTF-8 mode, values up to \777 are permitted. For
2900 example:
2901
2902 \040 is another way of writing a space
2903 \40 is the same, provided there are fewer than 40
2904 previous capturing subpatterns
2905 \7 is always a back reference
2906 \11 might be a back reference, or another way of
2907 writing a tab
2908 \011 is always a tab
2909 \0113 is a tab followed by the character "3"
2910 \113 might be a back reference, otherwise the
2911 character with octal code 113
2912 \377 might be a back reference, otherwise
2913 the byte consisting entirely of 1 bits
2914 \81 is either a back reference, or a binary zero
2915 followed by the two characters "8" and "1"
2916
2917 Note that octal values of 100 or greater must not be introduced by a
2918 leading zero, because no more than three octal digits are ever read.
2919
2920 All the sequences that define a single character value can be used both
2921 inside and outside character classes. In addition, inside a character
2922 class, the sequence \b is interpreted as the backspace character (hex
2923 08), and the sequences \R and \X are interpreted as the characters "R"
2924 and "X", respectively. Outside a character class, these sequences have
2925 different meanings (see below).
2926
2927 Absolute and relative back references
2928
2929 The sequence \g followed by a positive or negative number, optionally
2930 enclosed in braces, is an absolute or relative back reference. A named
2931 back reference can be coded as \g{name}. Back references are discussed
2932 later, following the discussion of parenthesized subpatterns.
2933
2934 Generic character types
2935
2936 Another use of backslash is for specifying generic character types. The
2937 following are always recognized:
2938
2939 \d any decimal digit
2940 \D any character that is not a decimal digit
2941 \s any whitespace character
2942 \S any character that is not a whitespace character
2943 \w any "word" character
2944 \W any "non-word" character
2945
2946 Each pair of escape sequences partitions the complete set of characters
2947 into two disjoint sets. Any given character matches one, and only one,
2948 of each pair.
2949
2950 These character type sequences can appear both inside and outside char-
2951 acter classes. They each match one character of the appropriate type.
2952 If the current matching point is at the end of the subject string, all
2953 of them fail, since there is no character to match.
2954
2955 For compatibility with Perl, \s does not match the VT character (code
2956 11). This makes it different from the the POSIX "space" class. The \s
2957 characters are HT (9), LF (10), FF (12), CR (13), and space (32). (If
2958 "use locale;" is included in a Perl script, \s may match the VT charac-
2959 ter. In PCRE, it never does.)
2960
2961 A "word" character is an underscore or any character less than 256 that
2962 is a letter or digit. The definition of letters and digits is con-
2963 trolled by PCRE's low-valued character tables, and may vary if locale-
2964 specific matching is taking place (see "Locale support" in the pcreapi
2965 page). For example, in a French locale such as "fr_FR" in Unix-like
2966 systems, or "french" in Windows, some character codes greater than 128
2967 are used for accented letters, and these are matched by \w.
2968
2969 In UTF-8 mode, characters with values greater than 128 never match \d,
2970 \s, or \w, and always match \D, \S, and \W. This is true even when Uni-
2971 code character property support is available. The use of locales with
2972 Unicode is discouraged.
2973
2974 Newline sequences
2975
2976 Outside a character class, the escape sequence \R matches any Unicode
2977 newline sequence. This is an extension to Perl. In non-UTF-8 mode \R is
2978 equivalent to the following:
2979
2980 (?>\r\n|\n|\x0b|\f|\r|\x85)
2981
2982 This is an example of an "atomic group", details of which are given
2983 below. This particular group matches either the two-character sequence
2984 CR followed by LF, or one of the single characters LF (linefeed,
2985 U+000A), VT (vertical tab, U+000B), FF (formfeed, U+000C), CR (carriage
2986 return, U+000D), or NEL (next line, U+0085). The two-character sequence
2987 is treated as a single unit that cannot be split.
2988
2989 In UTF-8 mode, two additional characters whose codepoints are greater
2990 than 255 are added: LS (line separator, U+2028) and PS (paragraph sepa-
2991 rator, U+2029). Unicode character property support is not needed for
2992 these characters to be recognized.
2993
2994 Inside a character class, \R matches the letter "R".
2995
2996 Unicode character properties
2997
2998 When PCRE is built with Unicode character property support, three addi-
2999 tional escape sequences to match character properties are available
3000 when UTF-8 mode is selected. They are:
3001
3002 \p{xx} a character with the xx property
3003 \P{xx} a character without the xx property
3004 \X an extended Unicode sequence
3005
3006 The property names represented by xx above are limited to the Unicode
3007 script names, the general category properties, and "Any", which matches
3008 any character (including newline). Other properties such as "InMusical-
3009 Symbols" are not currently supported by PCRE. Note that \P{Any} does
3010 not match any characters, so always causes a match failure.
3011
3012 Sets of Unicode characters are defined as belonging to certain scripts.
3013 A character from one of these sets can be matched using a script name.
3014 For example:
3015
3016 \p{Greek}
3017 \P{Han}
3018
3019 Those that are not part of an identified script are lumped together as
3020 "Common". The current list of scripts is:
3021
3022 Arabic, Armenian, Balinese, Bengali, Bopomofo, Braille, Buginese,
3023 Buhid, Canadian_Aboriginal, Cherokee, Common, Coptic, Cuneiform,
3024 Cypriot, Cyrillic, Deseret, Devanagari, Ethiopic, Georgian, Glagolitic,
3025 Gothic, Greek, Gujarati, Gurmukhi, Han, Hangul, Hanunoo, Hebrew, Hira-
3026 gana, Inherited, Kannada, Katakana, Kharoshthi, Khmer, Lao, Latin,
3027 Limbu, Linear_B, Malayalam, Mongolian, Myanmar, New_Tai_Lue, Nko,
3028 Ogham, Old_Italic, Old_Persian, Oriya, Osmanya, Phags_Pa, Phoenician,
3029 Runic, Shavian, Sinhala, Syloti_Nagri, Syriac, Tagalog, Tagbanwa,
3030 Tai_Le, Tamil, Telugu, Thaana, Thai, Tibetan, Tifinagh, Ugaritic, Yi.
3031
3032 Each character has exactly one general category property, specified by
3033 a two-letter abbreviation. For compatibility with Perl, negation can be
3034 specified by including a circumflex between the opening brace and the
3035 property name. For example, \p{^Lu} is the same as \P{Lu}.
3036
3037 If only one letter is specified with \p or \P, it includes all the gen-
3038 eral category properties that start with that letter. In this case, in
3039 the absence of negation, the curly brackets in the escape sequence are
3040 optional; these two examples have the same effect:
3041
3042 \p{L}
3043 \pL
3044
3045 The following general category property codes are supported:
3046
3047 C Other
3048 Cc Control
3049 Cf Format
3050 Cn Unassigned
3051 Co Private use
3052 Cs Surrogate
3053
3054 L Letter
3055 Ll Lower case letter
3056 Lm Modifier letter
3057 Lo Other letter
3058 Lt Title case letter
3059 Lu Upper case letter
3060
3061 M Mark
3062 Mc Spacing mark
3063 Me Enclosing mark
3064 Mn Non-spacing mark
3065
3066 N Number
3067 Nd Decimal number
3068 Nl Letter number
3069 No Other number
3070
3071 P Punctuation
3072 Pc Connector punctuation
3073 Pd Dash punctuation
3074 Pe Close punctuation
3075 Pf Final punctuation
3076 Pi Initial punctuation
3077 Po Other punctuation
3078 Ps Open punctuation
3079
3080 S Symbol
3081 Sc Currency symbol
3082 Sk Modifier symbol
3083 Sm Mathematical symbol
3084 So Other symbol
3085
3086 Z Separator
3087 Zl Line separator
3088 Zp Paragraph separator
3089 Zs Space separator
3090
3091 The special property L& is also supported: it matches a character that
3092 has the Lu, Ll, or Lt property, in other words, a letter that is not
3093 classified as a modifier or "other".
3094
3095 The long synonyms for these properties that Perl supports (such as
3096 \p{Letter}) are not supported by PCRE, nor is it permitted to prefix
3097 any of these properties with "Is".
3098
3099 No character that is in the Unicode table has the Cn (unassigned) prop-
3100 erty. Instead, this property is assumed for any code point that is not
3101 in the Unicode table.
3102
3103 Specifying caseless matching does not affect these escape sequences.
3104 For example, \p{Lu} always matches only upper case letters.
3105
3106 The \X escape matches any number of Unicode characters that form an
3107 extended Unicode sequence. \X is equivalent to
3108
3109 (?>\PM\pM*)
3110
3111 That is, it matches a character without the "mark" property, followed
3112 by zero or more characters with the "mark" property, and treats the
3113 sequence as an atomic group (see below). Characters with the "mark"
3114 property are typically accents that affect the preceding character.
3115
3116 Matching characters by Unicode property is not fast, because PCRE has
3117 to search a structure that contains data for over fifteen thousand
3118 characters. That is why the traditional escape sequences such as \d and
3119 \w do not use Unicode properties in PCRE.
3120
3121 Resetting the match start
3122
3123 The escape sequence \K, which is a Perl 5.10 feature, causes any previ-
3124 ously matched characters not to be included in the final matched
3125 sequence. For example, the pattern:
3126
3127 foo\Kbar
3128
3129 matches "foobar", but reports that it has matched "bar". This feature
3130 is similar to a lookbehind assertion (described below). However, in
3131 this case, the part of the subject before the real match does not have
3132 to be of fixed length, as lookbehind assertions do. The use of \K does
3133 not interfere with the setting of captured substrings. For example,
3134 when the pattern
3135
3136 (foo)\Kbar
3137
3138 matches "foobar", the first substring is still set to "foo".
3139
3140 Simple assertions
3141
3142 The final use of backslash is for certain simple assertions. An asser-
3143 tion specifies a condition that has to be met at a particular point in
3144 a match, without consuming any characters from the subject string. The
3145 use of subpatterns for more complicated assertions is described below.
3146 The backslashed assertions are:
3147
3148 \b matches at a word boundary
3149 \B matches when not at a word boundary
3150 \A matches at the start of the subject
3151 \Z matches at the end of the subject
3152 also matches before a newline at the end of the subject
3153 \z matches only at the end of the subject
3154 \G matches at the first matching position in the subject
3155
3156 These assertions may not appear in character classes (but note that \b
3157 has a different meaning, namely the backspace character, inside a char-
3158 acter class).
3159
3160 A word boundary is a position in the subject string where the current
3161 character and the previous character do not both match \w or \W (i.e.
3162 one matches \w and the other matches \W), or the start or end of the
3163 string if the first or last character matches \w, respectively.
3164
3165 The \A, \Z, and \z assertions differ from the traditional circumflex
3166 and dollar (described in the next section) in that they only ever match
3167 at the very start and end of the subject string, whatever options are
3168 set. Thus, they are independent of multiline mode. These three asser-
3169 tions are not affected by the PCRE_NOTBOL or PCRE_NOTEOL options, which
3170 affect only the behaviour of the circumflex and dollar metacharacters.
3171 However, if the startoffset argument of pcre_exec() is non-zero, indi-
3172 cating that matching is to start at a point other than the beginning of
3173 the subject, \A can never match. The difference between \Z and \z is
3174 that \Z matches before a newline at the end of the string as well as at
3175 the very end, whereas \z matches only at the end.
3176
3177 The \G assertion is true only when the current matching position is at
3178 the start point of the match, as specified by the startoffset argument
3179 of pcre_exec(). It differs from \A when the value of startoffset is
3180 non-zero. By calling pcre_exec() multiple times with appropriate argu-
3181 ments, you can mimic Perl's /g option, and it is in this kind of imple-
3182 mentation where \G can be useful.
3183
3184 Note, however, that PCRE's interpretation of \G, as the start of the
3185 current match, is subtly different from Perl's, which defines it as the
3186 end of the previous match. In Perl, these can be different when the
3187 previously matched string was empty. Because PCRE does just one match
3188 at a time, it cannot reproduce this behaviour.
3189
3190 If all the alternatives of a pattern begin with \G, the expression is
3191 anchored to the starting match position, and the "anchored" flag is set
3192 in the compiled regular expression.
3193
3194
3195 CIRCUMFLEX AND DOLLAR
3196
3197 Outside a character class, in the default matching mode, the circumflex
3198 character is an assertion that is true only if the current matching
3199 point is at the start of the subject string. If the startoffset argu-
3200 ment of pcre_exec() is non-zero, circumflex can never match if the
3201 PCRE_MULTILINE option is unset. Inside a character class, circumflex
3202 has an entirely different meaning (see below).
3203
3204 Circumflex need not be the first character of the pattern if a number
3205 of alternatives are involved, but it should be the first thing in each
3206 alternative in which it appears if the pattern is ever to match that
3207 branch. If all possible alternatives start with a circumflex, that is,
3208 if the pattern is constrained to match only at the start of the sub-
3209 ject, it is said to be an "anchored" pattern. (There are also other
3210 constructs that can cause a pattern to be anchored.)
3211
3212 A dollar character is an assertion that is true only if the current
3213 matching point is at the end of the subject string, or immediately
3214 before a newline at the end of the string (by default). Dollar need not
3215 be the last character of the pattern if a number of alternatives are
3216 involved, but it should be the last item in any branch in which it
3217 appears. Dollar has no special meaning in a character class.
3218
3219 The meaning of dollar can be changed so that it matches only at the
3220 very end of the string, by setting the PCRE_DOLLAR_ENDONLY option at
3221 compile time. This does not affect the \Z assertion.
3222
3223 The meanings of the circumflex and dollar characters are changed if the
3224 PCRE_MULTILINE option is set. When this is the case, a circumflex
3225 matches immediately after internal newlines as well as at the start of
3226 the subject string. It does not match after a newline that ends the
3227 string. A dollar matches before any newlines in the string, as well as
3228 at the very end, when PCRE_MULTILINE is set. When newline is specified
3229 as the two-character sequence CRLF, isolated CR and LF characters do
3230 not indicate newlines.
3231
3232 For example, the pattern /^abc$/ matches the subject string "def\nabc"
3233 (where \n represents a newline) in multiline mode, but not otherwise.
3234 Consequently, patterns that are anchored in single line mode because
3235 all branches start with ^ are not anchored in multiline mode, and a
3236 match for circumflex is possible when the startoffset argument of
3237 pcre_exec() is non-zero. The PCRE_DOLLAR_ENDONLY option is ignored if
3238 PCRE_MULTILINE is set.
3239
3240 Note that the sequences \A, \Z, and \z can be used to match the start
3241 and end of the subject in both modes, and if all branches of a pattern
3242 start with \A it is always anchored, whether or not PCRE_MULTILINE is
3243 set.
3244
3245
3246 FULL STOP (PERIOD, DOT)
3247
3248 Outside a character class, a dot in the pattern matches any one charac-
3249 ter in the subject string except (by default) a character that signi-
3250 fies the end of a line. In UTF-8 mode, the matched character may be
3251 more than one byte long.
3252
3253 When a line ending is defined as a single character, dot never matches
3254 that character; when the two-character sequence CRLF is used, dot does
3255 not match CR if it is immediately followed by LF, but otherwise it
3256 matches all characters (including isolated CRs and LFs). When any Uni-
3257 code line endings are being recognized, dot does not match CR or LF or
3258 any of the other line ending characters.
3259
3260 The behaviour of dot with regard to newlines can be changed. If the
3261 PCRE_DOTALL option is set, a dot matches any one character, without
3262 exception. If the two-character sequence CRLF is present in the subject
3263 string, it takes two dots to match it.
3264
3265 The handling of dot is entirely independent of the handling of circum-
3266 flex and dollar, the only relationship being that they both involve
3267 newlines. Dot has no special meaning in a character class.
3268
3269
3270 MATCHING A SINGLE BYTE
3271
3272 Outside a character class, the escape sequence \C matches any one byte,
3273 both in and out of UTF-8 mode. Unlike a dot, it always matches any
3274 line-ending characters. The feature is provided in Perl in order to
3275 match individual bytes in UTF-8 mode. Because it breaks up UTF-8 char-
3276 acters into individual bytes, what remains in the string may be a mal-
3277 formed UTF-8 string. For this reason, the \C escape sequence is best
3278 avoided.
3279
3280 PCRE does not allow \C to appear in lookbehind assertions (described
3281 below), because in UTF-8 mode this would make it impossible to calcu-
3282 late the length of the lookbehind.
3283
3284
3285 SQUARE BRACKETS AND CHARACTER CLASSES
3286
3287 An opening square bracket introduces a character class, terminated by a
3288 closing square bracket. A closing square bracket on its own is not spe-
3289 cial. If a closing square bracket is required as a member of the class,
3290 it should be the first data character in the class (after an initial
3291 circumflex, if present) or escaped with a backslash.
3292
3293 A character class matches a single character in the subject. In UTF-8
3294 mode, the character may occupy more than one byte. A matched character
3295 must be in the set of characters defined by the class, unless the first
3296 character in the class definition is a circumflex, in which case the
3297 subject character must not be in the set defined by the class. If a
3298 circumflex is actually required as a member of the class, ensure it is
3299 not the first character, or escape it with a backslash.
3300
3301 For example, the character class [aeiou] matches any lower case vowel,
3302 while [^aeiou] matches any character that is not a lower case vowel.
3303 Note that a circumflex is just a convenient notation for specifying the
3304 characters that are in the class by enumerating those that are not. A
3305 class that starts with a circumflex is not an assertion: it still con-
3306 sumes a character from the subject string, and therefore it fails if
3307 the current pointer is at the end of the string.
3308
3309 In UTF-8 mode, characters with values greater than 255 can be included
3310 in a class as a literal string of bytes, or by using the \x{ escaping
3311 mechanism.
3312
3313 When caseless matching is set, any letters in a class represent both
3314 their upper case and lower case versions, so for example, a caseless
3315 [aeiou] matches "A" as well as "a", and a caseless [^aeiou] does not
3316 match "A", whereas a caseful version would. In UTF-8 mode, PCRE always
3317 understands the concept of case for characters whose values are less
3318 than 128, so caseless matching is always possible. For characters with
3319 higher values, the concept of case is supported if PCRE is compiled
3320 with Unicode property support, but not otherwise. If you want to use
3321 caseless matching for characters 128 and above, you must ensure that
3322 PCRE is compiled with Unicode property support as well as with UTF-8
3323 support.
3324
3325 Characters that might indicate line breaks are never treated in any
3326 special way when matching character classes, whatever line-ending
3327 sequence is in use, and whatever setting of the PCRE_DOTALL and
3328 PCRE_MULTILINE options is used. A class such as [^a] always matches one
3329 of these characters.
3330
3331 The minus (hyphen) character can be used to specify a range of charac-
3332 ters in a character class. For example, [d-m] matches any letter
3333 between d and m, inclusive. If a minus character is required in a
3334 class, it must be escaped with a backslash or appear in a position
3335 where it cannot be interpreted as indicating a range, typically as the
3336 first or last character in the class.
3337
3338 It is not possible to have the literal character "]" as the end charac-
3339 ter of a range. A pattern such as [W-]46] is interpreted as a class of
3340 two characters ("W" and "-") followed by a literal string "46]", so it
3341 would match "W46]" or "-46]". However, if the "]" is escaped with a
3342 backslash it is interpreted as the end of range, so [W-\]46] is inter-
3343 preted as a class containing a range followed by two other characters.
3344 The octal or hexadecimal representation of "]" can also be used to end
3345 a range.
3346
3347 Ranges operate in the collating sequence of character values. They can
3348 also be used for characters specified numerically, for example
3349 [\000-\037]. In UTF-8 mode, ranges can include characters whose values
3350 are greater than 255, for example [\x{100}-\x{2ff}].
3351
3352 If a range that includes letters is used when caseless matching is set,
3353 it matches the letters in either case. For example, [W-c] is equivalent
3354 to [][\\^_`wxyzabc], matched caselessly, and in non-UTF-8 mode, if
3355 character tables for a French locale are in use, [\xc8-\xcb] matches
3356 accented E characters in both cases. In UTF-8 mode, PCRE supports the
3357 concept of case for characters with values greater than 128 only when
3358 it is compiled with Unicode property support.
3359
3360 The character types \d, \D, \p, \P, \s, \S, \w, and \W may also appear
3361 in a character class, and add the characters that they match to the
3362 class. For example, [\dABCDEF] matches any hexadecimal digit. A circum-
3363 flex can conveniently be used with the upper case character types to
3364 specify a more restricted set of characters than the matching lower
3365 case type. For example, the class [^\W_] matches any letter or digit,
3366 but not underscore.
3367
3368 The only metacharacters that are recognized in character classes are
3369 backslash, hyphen (only where it can be interpreted as specifying a
3370 range), circumflex (only at the start), opening square bracket (only
3371 when it can be interpreted as introducing a POSIX class name - see the
3372 next section), and the terminating closing square bracket. However,
3373 escaping other non-alphanumeric characters does no harm.
3374
3375
3376 POSIX CHARACTER CLASSES
3377
3378 Perl supports the POSIX notation for character classes. This uses names
3379 enclosed by [: and :] within the enclosing square brackets. PCRE also
3380 supports this notation. For example,
3381
3382 [01[:alpha:]%]
3383
3384 matches "0", "1", any alphabetic character, or "%". The supported class
3385 names are
3386
3387 alnum letters and digits
3388 alpha letters
3389 ascii character codes 0 - 127
3390 blank space or tab only
3391 cntrl control characters
3392 digit decimal digits (same as \d)
3393 graph printing characters, excluding space
3394 lower lower case letters
3395 print printing characters, including space
3396 punct printing characters, excluding letters and digits
3397 space white space (not quite the same as \s)
3398 upper upper case letters
3399 word "word" characters (same as \w)
3400 xdigit hexadecimal digits
3401
3402 The "space" characters are HT (9), LF (10), VT (11), FF (12), CR (13),
3403 and space (32). Notice that this list includes the VT character (code
3404 11). This makes "space" different to \s, which does not include VT (for
3405 Perl compatibility).
3406
3407 The name "word" is a Perl extension, and "blank" is a GNU extension
3408 from Perl 5.8. Another Perl extension is negation, which is indicated
3409 by a ^ character after the colon. For example,
3410
3411 [12[:^digit:]]
3412
3413 matches "1", "2", or any non-digit. PCRE (and Perl) also recognize the
3414 POSIX syntax [.ch.] and [=ch=] where "ch" is a "collating element", but
3415 these are not supported, and an error is given if they are encountered.
3416
3417 In UTF-8 mode, characters with values greater than 128 do not match any
3418 of the POSIX character classes.
3419
3420
3421 VERTICAL BAR
3422
3423 Vertical bar characters are used to separate alternative patterns. For
3424 example, the pattern
3425
3426 gilbert|sullivan
3427
3428 matches either "gilbert" or "sullivan". Any number of alternatives may
3429 appear, and an empty alternative is permitted (matching the empty
3430 string). The matching process tries each alternative in turn, from left
3431 to right, and the first one that succeeds is used. If the alternatives
3432 are within a subpattern (defined below), "succeeds" means matching the
3433 rest of the main pattern as well as the alternative in the subpattern.
3434
3435
3436 INTERNAL OPTION SETTING
3437
3438 The settings of the PCRE_CASELESS, PCRE_MULTILINE, PCRE_DOTALL, and
3439 PCRE_EXTENDED options can be changed from within the pattern by a
3440 sequence of Perl option letters enclosed between "(?" and ")". The
3441 option letters are
3442
3443 i for PCRE_CASELESS
3444 m for PCRE_MULTILINE
3445 s for PCRE_DOTALL
3446 x for PCRE_EXTENDED
3447
3448 For example, (?im) sets caseless, multiline matching. It is also possi-
3449 ble to unset these options by preceding the letter with a hyphen, and a
3450 combined setting and unsetting such as (?im-sx), which sets PCRE_CASE-
3451 LESS and PCRE_MULTILINE while unsetting PCRE_DOTALL and PCRE_EXTENDED,
3452 is also permitted. If a letter appears both before and after the
3453 hyphen, the option is unset.
3454
3455 When an option change occurs at top level (that is, not inside subpat-
3456 tern parentheses), the change applies to the remainder of the pattern
3457 that follows. If the change is placed right at the start of a pattern,
3458 PCRE extracts it into the global options (and it will therefore show up
3459 in data extracted by the pcre_fullinfo() function).
3460
3461 An option change within a subpattern (see below for a description of
3462 subpatterns) affects only that part of the current pattern that follows
3463 it, so
3464
3465 (a(?i)b)c
3466
3467 matches abc and aBc and no other strings (assuming PCRE_CASELESS is not
3468 used). By this means, options can be made to have different settings
3469 in different parts of the pattern. Any changes made in one alternative
3470 do carry on into subsequent branches within the same subpattern. For
3471 example,
3472
3473 (a(?i)b|c)
3474
3475 matches "ab", "aB", "c", and "C", even though when matching "C" the
3476 first branch is abandoned before the option setting. This is because
3477 the effects of option settings happen at compile time. There would be
3478 some very weird behaviour otherwise.
3479
3480 The PCRE-specific options PCRE_DUPNAMES, PCRE_UNGREEDY, and PCRE_EXTRA
3481 can be changed in the same way as the Perl-compatible options by using
3482 the characters J, U and X respectively.
3483
3484
3485 SUBPATTERNS
3486
3487 Subpatterns are delimited by parentheses (round brackets), which can be
3488 nested. Turning part of a pattern into a subpattern does two things:
3489
3490 1. It localizes a set of alternatives. For example, the pattern
3491
3492 cat(aract|erpillar|)
3493
3494 matches one of the words "cat", "cataract", or "caterpillar". Without
3495 the parentheses, it would match "cataract", "erpillar" or an empty
3496 string.
3497
3498 2. It sets up the subpattern as a capturing subpattern. This means
3499 that, when the whole pattern matches, that portion of the subject
3500 string that matched the subpattern is passed back to the caller via the
3501 ovector argument of pcre_exec(). Opening parentheses are counted from
3502 left to right (starting from 1) to obtain numbers for the capturing
3503 subpatterns.
3504
3505 For example, if the string "the red king" is matched against the pat-
3506 tern
3507
3508 the ((red|white) (king|queen))
3509
3510 the captured substrings are "red king", "red", and "king", and are num-
3511 bered 1, 2, and 3, respectively.
3512
3513 The fact that plain parentheses fulfil two functions is not always
3514 helpful. There are often times when a grouping subpattern is required
3515 without a capturing requirement. If an opening parenthesis is followed
3516 by a question mark and a colon, the subpattern does not do any captur-
3517 ing, and is not counted when computing the number of any subsequent
3518 capturing subpatterns. For example, if the string "the white queen" is
3519 matched against the pattern
3520
3521 the ((?:red|white) (king|queen))
3522
3523 the captured substrings are "white queen" and "queen", and are numbered
3524 1 and 2. The maximum number of capturing subpatterns is 65535.
3525
3526 As a convenient shorthand, if any option settings are required at the
3527 start of a non-capturing subpattern, the option letters may appear
3528 between the "?" and the ":". Thus the two patterns
3529
3530 (?i:saturday|sunday)
3531 (?:(?i)saturday|sunday)
3532
3533 match exactly the same set of strings. Because alternative branches are
3534 tried from left to right, and options are not reset until the end of
3535 the subpattern is reached, an option setting in one branch does affect
3536 subsequent branches, so the above patterns match "SUNDAY" as well as
3537 "Saturday".
3538
3539
3540 NAMED SUBPATTERNS
3541
3542 Identifying capturing parentheses by number is simple, but it can be
3543 very hard to keep track of the numbers in complicated regular expres-
3544 sions. Furthermore, if an expression is modified, the numbers may
3545 change. To help with this difficulty, PCRE supports the naming of sub-
3546 patterns. This feature was not added to Perl until release 5.10. Python
3547 had the feature earlier, and PCRE introduced it at release 4.0, using
3548 the Python syntax. PCRE now supports both the Perl and the Python syn-
3549 tax.
3550
3551 In PCRE, a subpattern can be named in one of three ways: (?<name>...)
3552 or (?'name'...) as in Perl, or (?P<name>...) as in Python. References
3553 to capturing parentheses from other parts of the pattern, such as back-
3554 references, recursion, and conditions, can be made by name as well as
3555 by number.
3556
3557 Names consist of up to 32 alphanumeric characters and underscores.
3558 Named capturing parentheses are still allocated numbers as well as
3559 names, exactly as if the names were not present. The PCRE API provides
3560 function calls for extracting the name-to-number translation table from
3561 a compiled pattern. There is also a convenience function for extracting
3562 a captured substring by name.
3563
3564 By default, a name must be unique within a pattern, but it is possible
3565 to relax this constraint by setting the PCRE_DUPNAMES option at compile
3566 time. This can be useful for patterns where only one instance of the
3567 named parentheses can match. Suppose you want to match the name of a
3568 weekday, either as a 3-letter abbreviation or as the full name, and in
3569 both cases you want to extract the abbreviation. This pattern (ignoring
3570 the line breaks) does the job:
3571
3572 (?<DN>Mon|Fri|Sun)(?:day)?|
3573 (?<DN>Tue)(?:sday)?|
3574 (?<DN>Wed)(?:nesday)?|
3575 (?<DN>Thu)(?:rsday)?|
3576 (?<DN>Sat)(?:urday)?
3577
3578 There are five capturing substrings, but only one is ever set after a
3579 match. The convenience function for extracting the data by name
3580 returns the substring for the first (and in this example, the only)
3581 subpattern of that name that matched. This saves searching to find
3582 which numbered subpattern it was. If you make a reference to a non-
3583 unique named subpattern from elsewhere in the pattern, the one that
3584 corresponds to the lowest number is used. For further details of the
3585 interfaces for handling named subpatterns, see the pcreapi documenta-
3586 tion.
3587
3588
3589 REPETITION
3590
3591 Repetition is specified by quantifiers, which can follow any of the
3592 following items:
3593
3594 a literal data character
3595 the dot metacharacter
3596 the \C escape sequence
3597 the \X escape sequence (in UTF-8 mode with Unicode properties)
3598 the \R escape sequence
3599 an escape such as \d that matches a single character
3600 a character class
3601 a back reference (see next section)
3602 a parenthesized subpattern (unless it is an assertion)
3603
3604 The general repetition quantifier specifies a minimum and maximum num-
3605 ber of permitted matches, by giving the two numbers in curly brackets
3606 (braces), separated by a comma. The numbers must be less than 65536,
3607 and the first must be less than or equal to the second. For example:
3608
3609 z{2,4}
3610
3611 matches "zz", "zzz", or "zzzz". A closing brace on its own is not a
3612 special character. If the second number is omitted, but the comma is
3613 present, there is no upper limit; if the second number and the comma
3614 are both omitted, the quantifier specifies an exact number of required
3615 matches. Thus
3616
3617 [aeiou]{3,}
3618
3619 matches at least 3 successive vowels, but may match many more, while
3620
3621 \d{8}
3622
3623 matches exactly 8 digits. An opening curly bracket that appears in a
3624 position where a quantifier is not allowed, or one that does not match
3625 the syntax of a quantifier, is taken as a literal character. For exam-
3626 ple, {,6} is not a quantifier, but a literal string of four characters.
3627
3628 In UTF-8 mode, quantifiers apply to UTF-8 characters rather than to
3629 individual bytes. Thus, for example, \x{100}{2} matches two UTF-8 char-
3630 acters, each of which is represented by a two-byte sequence. Similarly,
3631 when Unicode property support is available, \X{3} matches three Unicode
3632 extended sequences, each of which may be several bytes long (and they
3633 may be of different lengths).
3634
3635 The quantifier {0} is permitted, causing the expression to behave as if
3636 the previous item and the quantifier were not present.
3637
3638 For convenience, the three most common quantifiers have single-charac-
3639 ter abbreviations:
3640
3641 * is equivalent to {0,}
3642 + is equivalent to {1,}
3643 ? is equivalent to {0,1}
3644
3645 It is possible to construct infinite loops by following a subpattern
3646 that can match no characters with a quantifier that has no upper limit,
3647 for example:
3648
3649 (a?)*
3650
3651 Earlier versions of Perl and PCRE used to give an error at compile time
3652 for such patterns. However, because there are cases where this can be
3653 useful, such patterns are now accepted, but if any repetition of the
3654 subpattern does in fact match no characters, the loop is forcibly bro-
3655 ken.
3656
3657 By default, the quantifiers are "greedy", that is, they match as much
3658 as possible (up to the maximum number of permitted times), without
3659 causing the rest of the pattern to fail. The classic example of where
3660 this gives problems is in trying to match comments in C programs. These
3661 appear between /* and */ and within the comment, individual * and /
3662 characters may appear. An attempt to match C comments by applying the
3663 pattern
3664
3665 /\*.*\*/
3666
3667 to the string
3668
3669 /* first comment */ not comment /* second comment */
3670
3671 fails, because it matches the entire string owing to the greediness of
3672 the .* item.
3673
3674 However, if a quantifier is followed by a question mark, it ceases to
3675 be greedy, and instead matches the minimum number of times possible, so
3676 the pattern
3677
3678 /\*.*?\*/
3679
3680 does the right thing with the C comments. The meaning of the various
3681 quantifiers is not otherwise changed, just the preferred number of
3682 matches. Do not confuse this use of question mark with its use as a
3683 quantifier in its own right. Because it has two uses, it can sometimes
3684 appear doubled, as in
3685
3686 \d??\d
3687
3688 which matches one digit by preference, but can match two if that is the
3689 only way the rest of the pattern matches.
3690
3691 If the PCRE_UNGREEDY option is set (an option that is not available in
3692 Perl), the quantifiers are not greedy by default, but individual ones
3693 can be made greedy by following them with a question mark. In other
3694 words, it inverts the default behaviour.
3695
3696 When a parenthesized subpattern is quantified with a minimum repeat
3697 count that is greater than 1 or with a limited maximum, more memory is
3698 required for the compiled pattern, in proportion to the size of the
3699 minimum or maximum.
3700
3701 If a pattern starts with .* or .{0,} and the PCRE_DOTALL option (equiv-
3702 alent to Perl's /s) is set, thus allowing the dot to match newlines,
3703 the pattern is implicitly anchored, because whatever follows will be
3704 tried against every character position in the subject string, so there
3705 is no point in retrying the overall match at any position after the
3706 first. PCRE normally treats such a pattern as though it were preceded
3707 by \A.
3708
3709 In cases where it is known that the subject string contains no new-
3710 lines, it is worth setting PCRE_DOTALL in order to obtain this opti-
3711 mization, or alternatively using ^ to indicate anchoring explicitly.
3712
3713 However, there is one situation where the optimization cannot be used.
3714 When .* is inside capturing parentheses that are the subject of a
3715 backreference elsewhere in the pattern, a match at the start may fail
3716 where a later one succeeds. Consider, for example:
3717
3718 (.*)abc\1
3719
3720 If the subject is "xyz123abc123" the match point is the fourth charac-
3721 ter. For this reason, such a pattern is not implicitly anchored.
3722
3723 When a capturing subpattern is repeated, the value captured is the sub-
3724 string that matched the final iteration. For example, after
3725
3726 (tweedle[dume]{3}\s*)+
3727
3728 has matched "tweedledum tweedledee" the value of the captured substring
3729 is "tweedledee". However, if there are nested capturing subpatterns,
3730 the corresponding captured values may have been set in previous itera-
3731 tions. For example, after
3732
3733 /(a|(b))+/
3734
3735 matches "aba" the value of the second captured substring is "b".
3736
3737
3738 ATOMIC GROUPING AND POSSESSIVE QUANTIFIERS
3739
3740 With both maximizing ("greedy") and minimizing ("ungreedy" or "lazy")
3741 repetition, failure of what follows normally causes the repeated item
3742 to be re-evaluated to see if a different number of repeats allows the
3743 rest of the pattern to match. Sometimes it is useful to prevent this,
3744 either to change the nature of the match, or to cause it fail earlier
3745 than it otherwise might, when the author of the pattern knows there is
3746 no point in carrying on.
3747
3748 Consider, for example, the pattern \d+foo when applied to the subject
3749 line
3750
3751 123456bar
3752
3753 After matching all 6 digits and then failing to match "foo", the normal
3754 action of the matcher is to try again with only 5 digits matching the
3755 \d+ item, and then with 4, and so on, before ultimately failing.
3756 "Atomic grouping" (a term taken from Jeffrey Friedl's book) provides
3757 the means for specifying that once a subpattern has matched, it is not
3758 to be re-evaluated in this way.
3759
3760 If we use atomic grouping for the previous example, the matcher gives
3761 up immediately on failing to match "foo" the first time. The notation
3762 is a kind of special parenthesis, starting with (?> as in this example:
3763
3764 (?>\d+)foo
3765
3766 This kind of parenthesis "locks up" the part of the pattern it con-
3767 tains once it has matched, and a failure further into the pattern is
3768 prevented from backtracking into it. Backtracking past it to previous
3769 items, however, works as normal.
3770
3771 An alternative description is that a subpattern of this type matches
3772 the string of characters that an identical standalone pattern would
3773 match, if anchored at the current point in the subject string.
3774
3775 Atomic grouping subpatterns are not capturing subpatterns. Simple cases
3776 such as the above example can be thought of as a maximizing repeat that
3777 must swallow everything it can. So, while both \d+ and \d+? are pre-
3778 pared to adjust the number of digits they match in order to make the
3779 rest of the pattern match, (?>\d+) can only match an entire sequence of
3780 digits.
3781
3782 Atomic groups in general can of course contain arbitrarily complicated
3783 subpatterns, and can be nested. However, when the subpattern for an
3784 atomic group is just a single repeated item, as in the example above, a
3785 simpler notation, called a "possessive quantifier" can be used. This
3786 consists of an additional + character following a quantifier. Using
3787 this notation, the previous example can be rewritten as
3788
3789 \d++foo
3790
3791 Possessive quantifiers are always greedy; the setting of the
3792 PCRE_UNGREEDY option is ignored. They are a convenient notation for the
3793 simpler forms of atomic group. However, there is no difference in the
3794 meaning of a possessive quantifier and the equivalent atomic group,
3795 though there may be a performance difference; possessive quantifiers
3796 should be slightly faster.
3797
3798 The possessive quantifier syntax is an extension to the Perl 5.8 syn-
3799 tax. Jeffrey Friedl originated the idea (and the name) in the first
3800 edition of his book. Mike McCloskey liked it, so implemented it when he
3801 built Sun's Java package, and PCRE copied it from there. It ultimately
3802 found its way into Perl at release 5.10.
3803
3804 PCRE has an optimization that automatically "possessifies" certain sim-
3805 ple pattern constructs. For example, the sequence A+B is treated as
3806 A++B because there is no point in backtracking into a sequence of A's
3807 when B must follow.
3808
3809 When a pattern contains an unlimited repeat inside a subpattern that
3810 can itself be repeated an unlimited number of times, the use of an
3811 atomic group is the only way to avoid some failing matches taking a
3812 very long time indeed. The pattern
3813
3814 (\D+|<\d+>)*[!?]
3815
3816 matches an unlimited number of substrings that either consist of non-
3817 digits, or digits enclosed in <>, followed by either ! or ?. When it
3818 matches, it runs quickly. However, if it is applied to
3819
3820 aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
3821
3822 it takes a long time before reporting failure. This is because the
3823 string can be divided between the internal \D+ repeat and the external
3824 * repeat in a large number of ways, and all have to be tried. (The
3825 example uses [!?] rather than a single character at the end, because
3826 both PCRE and Perl have an optimization that allows for fast failure
3827 when a single character is used. They remember the last single charac-
3828 ter that is required for a match, and fail early if it is not present
3829 in the string.) If the pattern is changed so that it uses an atomic
3830 group, like this:
3831
3832 ((?>\D+)|<\d+>)*[!?]
3833
3834 sequences of non-digits cannot be broken, and failure happens quickly.
3835
3836
3837 BACK REFERENCES
3838
3839 Outside a character class, a backslash followed by a digit greater than
3840 0 (and possibly further digits) is a back reference to a capturing sub-
3841 pattern earlier (that is, to its left) in the pattern, provided there
3842 have been that many previous capturing left parentheses.
3843
3844 However, if the decimal number following the backslash is less than 10,
3845 it is always taken as a back reference, and causes an error only if
3846 there are not that many capturing left parentheses in the entire pat-
3847 tern. In other words, the parentheses that are referenced need not be
3848 to the left of the reference for numbers less than 10. A "forward back
3849 reference" of this type can make sense when a repetition is involved
3850 and the subpattern to the right has participated in an earlier itera-
3851 tion.
3852
3853 It is not possible to have a numerical "forward back reference" to a
3854 subpattern whose number is 10 or more using this syntax because a
3855 sequence such as \50 is interpreted as a character defined in octal.
3856 See the subsection entitled "Non-printing characters" above for further
3857 details of the handling of digits following a backslash. There is no
3858 such problem when named parentheses are used. A back reference to any
3859 subpattern is possible using named parentheses (see below).
3860
3861 Another way of avoiding the ambiguity inherent in the use of digits
3862 following a backslash is to use the \g escape sequence, which is a fea-
3863 ture introduced in Perl 5.10. This escape must be followed by a posi-
3864 tive or a negative number, optionally enclosed in braces. These exam-
3865 ples are all identical:
3866
3867 (ring), \1
3868 (ring), \g1
3869 (ring), \g{1}
3870
3871 A positive number specifies an absolute reference without the ambiguity
3872 that is present in the older syntax. It is also useful when literal
3873 digits follow the reference. A negative number is a relative reference.
3874 Consider this example:
3875
3876 (abc(def)ghi)\g{-1}
3877
3878 The sequence \g{-1} is a reference to the most recently started captur-
3879 ing subpattern before \g, that is, is it equivalent to \2. Similarly,
3880 \g{-2} would be equivalent to \1. The use of relative references can be
3881 helpful in long patterns, and also in patterns that are created by
3882 joining together fragments that contain references within themselves.
3883
3884 A back reference matches whatever actually matched the capturing sub-
3885 pattern in the current subject string, rather than anything matching
3886 the subpattern itself (see "Subpatterns as subroutines" below for a way
3887 of doing that). So the pattern
3888
3889 (sens|respons)e and \1ibility
3890
3891 matches "sense and sensibility" and "response and responsibility", but
3892 not "sense and responsibility". If caseful matching is in force at the
3893 time of the back reference, the case of letters is relevant. For exam-
3894 ple,
3895
3896 ((?i)rah)\s+\1
3897
3898 matches "rah rah" and "RAH RAH", but not "RAH rah", even though the
3899 original capturing subpattern is matched caselessly.
3900
3901 There are several different ways of writing back references to named
3902 subpatterns. The .NET syntax \k{name} and the Perl syntax \k<name> or
3903 \k'name' are supported, as is the Python syntax (?P=name). Perl 5.10's
3904 unified back reference syntax, in which \g can be used for both numeric
3905 and named references, is also supported. We could rewrite the above
3906 example in any of the following ways:
3907
3908 (?<p1>(?i)rah)\s+\k<p1>
3909 (?'p1'(?i)rah)\s+\k{p1}
3910 (?P<p1>(?i)rah)\s+(?P=p1)
3911 (?<p1>(?i)rah)\s+\g{p1}
3912
3913 A subpattern that is referenced by name may appear in the pattern
3914 before or after the reference.
3915
3916 There may be more than one back reference to the same subpattern. If a
3917 subpattern has not actually been used in a particular match, any back
3918 references to it always fail. For example, the pattern
3919
3920 (a|(bc))\2
3921
3922 always fails if it starts to match "a" rather than "bc". Because there
3923 may be many capturing parentheses in a pattern, all digits following
3924 the backslash are taken as part of a potential back reference number.
3925 If the pattern continues with a digit character, some delimiter must be
3926 used to terminate the back reference. If the PCRE_EXTENDED option is
3927 set, this can be whitespace. Otherwise an empty comment (see "Com-
3928 ments" below) can be used.
3929
3930 A back reference that occurs inside the parentheses to which it refers
3931 fails when the subpattern is first used, so, for example, (a\1) never
3932 matches. However, such references can be useful inside repeated sub-
3933 patterns. For example, the pattern
3934
3935 (a|b\1)+
3936
3937 matches any number of "a"s and also "aba", "ababbaa" etc. At each iter-
3938 ation of the subpattern, the back reference matches the character
3939 string corresponding to the previous iteration. In order for this to
3940 work, the pattern must be such that the first iteration does not need
3941 to match the back reference. This can be done using alternation, as in
3942 the example above, or by a quantifier with a minimum of zero.
3943
3944
3945 ASSERTIONS
3946
3947 An assertion is a test on the characters following or preceding the
3948 current matching point that does not actually consume any characters.
3949 The simple assertions coded as \b, \B, \A, \G, \Z, \z, ^ and $ are
3950 described above.
3951
3952 More complicated assertions are coded as subpatterns. There are two
3953 kinds: those that look ahead of the current position in the subject
3954 string, and those that look behind it. An assertion subpattern is
3955 matched in the normal way, except that it does not cause the current
3956 matching position to be changed.
3957
3958 Assertion subpatterns are not capturing subpatterns, and may not be
3959 repeated, because it makes no sense to assert the same thing several
3960 times. If any kind of assertion contains capturing subpatterns within
3961 it, these are counted for the purposes of numbering the capturing sub-
3962 patterns in the whole pattern. However, substring capturing is carried
3963 out only for positive assertions, because it does not make sense for
3964 negative assertions.
3965
3966 Lookahead assertions
3967
3968 Lookahead assertions start with (?= for positive assertions and (?! for
3969 negative assertions. For example,
3970
3971 \w+(?=;)
3972
3973 matches a word followed by a semicolon, but does not include the semi-
3974 colon in the match, and
3975
3976 foo(?!bar)
3977
3978 matches any occurrence of "foo" that is not followed by "bar". Note
3979 that the apparently similar pattern
3980
3981 (?!foo)bar
3982
3983 does not find an occurrence of "bar" that is preceded by something
3984 other than "foo"; it finds any occurrence of "bar" whatsoever, because
3985 the assertion (?!foo) is always true when the next three characters are
3986 "bar". A lookbehind assertion is needed to achieve the other effect.
3987
3988 If you want to force a matching failure at some point in a pattern, the
3989 most convenient way to do it is with (?!) because an empty string
3990 always matches, so an assertion that requires there not to be an empty
3991 string must always fail.
3992
3993 Lookbehind assertions
3994
3995 Lookbehind assertions start with (?<= for positive assertions and (?<!
3996 for negative assertions. For example,
3997
3998 (?<!foo)bar
3999
4000 does find an occurrence of "bar" that is not preceded by "foo". The
4001 contents of a lookbehind assertion are restricted such that all the
4002 strings it matches must have a fixed length. However, if there are sev-
4003 eral top-level alternatives, they do not all have to have the same
4004 fixed length. Thus
4005
4006 (?<=bullock|donkey)
4007
4008 is permitted, but
4009
4010 (?<!dogs?|cats?)
4011
4012 causes an error at compile time. Branches that match different length
4013 strings are permitted only at the top level of a lookbehind assertion.
4014 This is an extension compared with Perl (at least for 5.8), which
4015 requires all branches to match the same length of string. An assertion
4016 such as
4017
4018 (?<=ab(c|de))
4019
4020 is not permitted, because its single top-level branch can match two
4021 different lengths, but it is acceptable if rewritten to use two top-
4022 level branches:
4023
4024 (?<=abc|abde)
4025
4026 In some cases, the Perl 5.10 escape sequence \K (see above) can be used
4027 instead of a lookbehind assertion; this is not restricted to a fixed-
4028 length.
4029
4030 The implementation of lookbehind assertions is, for each alternative,
4031 to temporarily move the current position back by the fixed length and
4032 then try to match. If there are insufficient characters before the cur-
4033 rent position, the assertion fails.
4034
4035 PCRE does not allow the \C escape (which matches a single byte in UTF-8
4036 mode) to appear in lookbehind assertions, because it makes it impossi-
4037 ble to calculate the length of the lookbehind. The \X and \R escapes,
4038 which can match different numbers of bytes, are also not permitted.
4039
4040 Possessive quantifiers can be used in conjunction with lookbehind
4041 assertions to specify efficient matching at the end of the subject
4042 string. Consider a simple pattern such as
4043
4044 abcd$
4045
4046 when applied to a long string that does not match. Because matching
4047 proceeds from left to right, PCRE will look for each "a" in the subject
4048 and then see if what follows matches the rest of the pattern. If the
4049 pattern is specified as
4050
4051 ^.*abcd$
4052
4053 the initial .* matches the entire string at first, but when this fails
4054 (because there is no following "a"), it backtracks to match all but the
4055 last character, then all but the last two characters, and so on. Once
4056 again the search for "a" covers the entire string, from right to left,
4057 so we are no better off. However, if the pattern is written as
4058
4059 ^.*+(?<=abcd)
4060
4061 there can be no backtracking for the .*+ item; it can match only the
4062 entire string. The subsequent lookbehind assertion does a single test
4063 on the last four characters. If it fails, the match fails immediately.
4064 For long strings, this approach makes a significant difference to the
4065 processing time.
4066
4067 Using multiple assertions
4068
4069 Several assertions (of any sort) may occur in succession. For example,
4070
4071 (?<=\d{3})(?<!999)foo
4072
4073 matches "foo" preceded by three digits that are not "999". Notice that
4074 each of the assertions is applied independently at the same point in
4075 the subject string. First there is a check that the previous three
4076 characters are all digits, and then there is a check that the same
4077 three characters are not "999". This pattern does not match "foo" pre-
4078 ceded by six characters, the first of which are digits and the last
4079 three of which are not "999". For example, it doesn't match "123abc-
4080 foo". A pattern to do that is
4081
4082 (?<=\d{3}...)(?<!999)foo
4083
4084 This time the first assertion looks at the preceding six characters,
4085 checking that the first three are digits, and then the second assertion
4086 checks that the preceding three characters are not "999".
4087
4088 Assertions can be nested in any combination. For example,
4089
4090 (?<=(?<!foo)bar)baz
4091
4092 matches an occurrence of "baz" that is preceded by "bar" which in turn
4093 is not preceded by "foo", while
4094
4095 (?<=\d{3}(?!999)...)foo
4096
4097 is another pattern that matches "foo" preceded by three digits and any
4098 three characters that are not "999".
4099
4100
4101 CONDITIONAL SUBPATTERNS
4102
4103 It is possible to cause the matching process to obey a subpattern con-
4104 ditionally or to choose between two alternative subpatterns, depending
4105 on the result of an assertion, or whether a previous capturing subpat-
4106 tern matched or not. The two possible forms of conditional subpattern
4107 are
4108
4109 (?(condition)yes-pattern)
4110 (?(condition)yes-pattern|no-pattern)
4111
4112 If the condition is satisfied, the yes-pattern is used; otherwise the
4113 no-pattern (if present) is used. If there are more than two alterna-
4114 tives in the subpattern, a compile-time error occurs.
4115
4116 There are four kinds of condition: references to subpatterns, refer-
4117 ences to recursion, a pseudo-condition called DEFINE, and assertions.
4118
4119 Checking for a used subpattern by number
4120
4121 If the text between the parentheses consists of a sequence of digits,
4122 the condition is true if the capturing subpattern of that number has
4123 previously matched. An alternative notation is to precede the digits
4124 with a plus or minus sign. In this case, the subpattern number is rela-
4125 tive rather than absolute. The most recently opened parentheses can be
4126 referenced by (?(-1), the next most recent by (?(-2), and so on. In
4127 looping constructs it can also make sense to refer to subsequent groups
4128 with constructs such as (?(+2).
4129
4130 Consider the following pattern, which contains non-significant white
4131 space to make it more readable (assume the PCRE_EXTENDED option) and to
4132 divide it into three parts for ease of discussion:
4133
4134 ( \( )? [^()]+ (?(1) \) )
4135
4136 The first part matches an optional opening parenthesis, and if that
4137 character is present, sets it as the first captured substring. The sec-
4138 ond part matches one or more characters that are not parentheses. The
4139 third part is a conditional subpattern that tests whether the first set
4140 of parentheses matched or not. If they did, that is, if subject started
4141 with an opening parenthesis, the condition is true, and so the yes-pat-
4142 tern is executed and a closing parenthesis is required. Otherwise,
4143 since no-pattern is not present, the subpattern matches nothing. In
4144 other words, this pattern matches a sequence of non-parentheses,
4145 optionally enclosed in parentheses.
4146
4147 If you were embedding this pattern in a larger one, you could use a
4148 relative reference:
4149
4150 ...other stuff... ( \( )? [^()]+ (?(-1) \) ) ...
4151
4152 This makes the fragment independent of the parentheses in the larger
4153 pattern.
4154
4155 Checking for a used subpattern by name
4156
4157 Perl uses the syntax (?(<name>)...) or (?('name')...) to test for a
4158 used subpattern by name. For compatibility with earlier versions of
4159 PCRE, which had this facility before Perl, the syntax (?(name)...) is
4160 also recognized. However, there is a possible ambiguity with this syn-
4161 tax, because subpattern names may consist entirely of digits. PCRE
4162 looks first for a named subpattern; if it cannot find one and the name
4163 consists entirely of digits, PCRE looks for a subpattern of that num-
4164 ber, which must be greater than zero. Using subpattern names that con-
4165 sist entirely of digits is not recommended.
4166
4167 Rewriting the above example to use a named subpattern gives this:
4168
4169 (?<OPEN> \( )? [^()]+ (?(<OPEN>) \) )
4170
4171
4172 Checking for pattern recursion
4173
4174 If the condition is the string (R), and there is no subpattern with the
4175 name R, the condition is true if a recursive call to the whole pattern
4176 or any subpattern has been made. If digits or a name preceded by amper-
4177 sand follow the letter R, for example:
4178
4179 (?(R3)...) or (?(R&name)...)
4180
4181 the condition is true if the most recent recursion is into the subpat-
4182 tern whose number or name is given. This condition does not check the
4183 entire recursion stack.
4184
4185 At "top level", all these recursion test conditions are false. Recur-
4186 sive patterns are described below.
4187
4188 Defining subpatterns for use by reference only
4189
4190 If the condition is the string (DEFINE), and there is no subpattern
4191 with the name DEFINE, the condition is always false. In this case,
4192 there may be only one alternative in the subpattern. It is always
4193 skipped if control reaches this point in the pattern; the idea of
4194 DEFINE is that it can be used to define "subroutines" that can be ref-
4195 erenced from elsewhere. (The use of "subroutines" is described below.)
4196 For example, a pattern to match an IPv4 address could be written like
4197 this (ignore whitespace and line breaks):
4198
4199 (?(DEFINE) (?<byte> 2[0-4]\d | 25[0-5] | 1\d\d | [1-9]?\d) )
4200 \b (?&byte) (\.(?&byte)){3} \b
4201
4202 The first part of the pattern is a DEFINE group inside which a another
4203 group named "byte" is defined. This matches an individual component of
4204 an IPv4 address (a number less than 256). When matching takes place,
4205 this part of the pattern is skipped because DEFINE acts like a false
4206 condition.
4207
4208 The rest of the pattern uses references to the named group to match the
4209 four dot-separated components of an IPv4 address, insisting on a word
4210 boundary at each end.
4211
4212 Assertion conditions
4213
4214 If the condition is not in any of the above formats, it must be an
4215 assertion. This may be a positive or negative lookahead or lookbehind
4216 assertion. Consider this pattern, again containing non-significant
4217 white space, and with the two alternatives on the second line:
4218
4219 (?(?=[^a-z]*[a-z])
4220 \d{2}-[a-z]{3}-\d{2} | \d{2}-\d{2}-\d{2} )
4221
4222 The condition is a positive lookahead assertion that matches an
4223 optional sequence of non-letters followed by a letter. In other words,
4224 it tests for the presence of at least one letter in the subject. If a
4225 letter is found, the subject is matched against the first alternative;
4226 otherwise it is matched against the second. This pattern matches
4227 strings in one of the two forms dd-aaa-dd or dd-dd-dd, where aaa are
4228 letters and dd are digits.
4229
4230
4231 COMMENTS
4232
4233 The sequence (?# marks the start of a comment that continues up to the
4234 next closing parenthesis. Nested parentheses are not permitted. The
4235 characters that make up a comment play no part in the pattern matching
4236 at all.
4237
4238 If the PCRE_EXTENDED option is set, an unescaped # character outside a
4239 character class introduces a comment that continues to immediately
4240 after the next newline in the pattern.
4241
4242
4243 RECURSIVE PATTERNS
4244
4245 Consider the problem of matching a string in parentheses, allowing for
4246 unlimited nested parentheses. Without the use of recursion, the best
4247 that can be done is to use a pattern that matches up to some fixed
4248 depth of nesting. It is not possible to handle an arbitrary nesting
4249 depth.
4250
4251 For some time, Perl has provided a facility that allows regular expres-
4252 sions to recurse (amongst other things). It does this by interpolating
4253 Perl code in the expression at run time, and the code can refer to the
4254 expression itself. A Perl pattern using code interpolation to solve the
4255 parentheses problem can be created like this:
4256
4257 $re = qr{\( (?: (?>[^()]+) | (?p{$re}) )* \)}x;
4258
4259 The (?p{...}) item interpolates Perl code at run time, and in this case
4260 refers recursively to the pattern in which it appears.
4261
4262 Obviously, PCRE cannot support the interpolation of Perl code. Instead,
4263 it supports special syntax for recursion of the entire pattern, and
4264 also for individual subpattern recursion. After its introduction in
4265 PCRE and Python, this kind of recursion was introduced into Perl at
4266 release 5.10.
4267
4268 A special item that consists of (? followed by a number greater than
4269 zero and a closing parenthesis is a recursive call of the subpattern of
4270 the given number, provided that it occurs inside that subpattern. (If
4271 not, it is a "subroutine" call, which is described in the next sec-
4272 tion.) The special item (?R) or (?0) is a recursive call of the entire
4273 regular expression.
4274
4275 In PCRE (like Python, but unlike Perl), a recursive subpattern call is
4276 always treated as an atomic group. That is, once it has matched some of
4277 the subject string, it is never re-entered, even if it contains untried
4278 alternatives and there is a subsequent matching failure.
4279
4280 This PCRE pattern solves the nested parentheses problem (assume the
4281 PCRE_EXTENDED option is set so that white space is ignored):
4282
4283 \( ( (?>[^()]+) | (?R) )* \)
4284
4285 First it matches an opening parenthesis. Then it matches any number of
4286 substrings which can either be a sequence of non-parentheses, or a
4287 recursive match of the pattern itself (that is, a correctly parenthe-
4288 sized substring). Finally there is a closing parenthesis.
4289
4290 If this were part of a larger pattern, you would not want to recurse
4291 the entire pattern, so instead you could use this:
4292
4293 ( \( ( (?>[^()]+) | (?1) )* \) )
4294
4295 We have put the pattern into parentheses, and caused the recursion to
4296 refer to them instead of the whole pattern.
4297
4298 In a larger pattern, keeping track of parenthesis numbers can be
4299 tricky. This is made easier by the use of relative references. (A Perl
4300 5.10 feature.) Instead of (?1) in the pattern above you can write
4301 (?-2) to refer to the second most recently opened parentheses preceding
4302 the recursion. In other words, a negative number counts capturing
4303 parentheses leftwards from the point at which it is encountered.
4304
4305 It is also possible to refer to subsequently opened parentheses, by
4306 writing references such as (?+2). However, these cannot be recursive
4307 because the reference is not inside the parentheses that are refer-
4308 enced. They are always "subroutine" calls, as described in the next
4309 section.
4310
4311 An alternative approach is to use named parentheses instead. The Perl
4312 syntax for this is (?&name); PCRE's earlier syntax (?P>name) is also
4313 supported. We could rewrite the above example as follows:
4314
4315 (?<pn> \( ( (?>[^()]+) | (?&pn) )* \) )
4316
4317 If there is more than one subpattern with the same name, the earliest
4318 one is used.
4319
4320 This particular example pattern that we have been looking at contains
4321 nested unlimited repeats, and so the use of atomic grouping for match-
4322 ing strings of non-parentheses is important when applying the pattern
4323 to strings that do not match. For example, when this pattern is applied
4324 to
4325
4326 (aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa()
4327
4328 it yields "no match" quickly. However, if atomic grouping is not used,
4329 the match runs for a very long time indeed because there are so many
4330 different ways the + and * repeats can carve up the subject, and all
4331 have to be tested before failure can be reported.
4332
4333 At the end of a match, the values set for any capturing subpatterns are
4334 those from the outermost level of the recursion at which the subpattern
4335 value is set. If you want to obtain intermediate values, a callout
4336 function can be used (see below and the pcrecallout documentation). If
4337 the pattern above is matched against
4338
4339 (ab(cd)ef)
4340
4341 the value for the capturing parentheses is "ef", which is the last
4342 value taken on at the top level. If additional parentheses are added,
4343 giving
4344
4345 \( ( ( (?>[^()]+) | (?R) )* ) \)
4346 ^ ^
4347 ^ ^
4348
4349 the string they capture is "ab(cd)ef", the contents of the top level
4350 parentheses. If there are more than 15 capturing parentheses in a pat-
4351 tern, PCRE has to obtain extra memory to store data during a recursion,
4352 which it does by using pcre_malloc, freeing it via pcre_free after-
4353 wards. If no memory can be obtained, the match fails with the
4354 PCRE_ERROR_NOMEMORY error.
4355
4356 Do not confuse the (?R) item with the condition (R), which tests for
4357 recursion. Consider this pattern, which matches text in angle brack-
4358 ets, allowing for arbitrary nesting. Only digits are allowed in nested
4359 brackets (that is, when recursing), whereas any characters are permit-
4360 ted at the outer level.
4361
4362 < (?: (?(R) \d++ | [^<>]*+) | (?R)) * >
4363
4364 In this pattern, (?(R) is the start of a conditional subpattern, with
4365 two different alternatives for the recursive and non-recursive cases.
4366 The (?R) item is the actual recursive call.
4367
4368
4369 SUBPATTERNS AS SUBROUTINES
4370
4371 If the syntax for a recursive subpattern reference (either by number or
4372 by name) is used outside the parentheses to which it refers, it oper-
4373 ates like a subroutine in a programming language. The "called" subpat-
4374 tern may be defined before or after the reference. A numbered reference
4375 can be absolute or relative, as in these examples:
4376
4377 (...(absolute)...)...(?2)...
4378 (...(relative)...)...(?-1)...
4379 (...(?+1)...(relative)...
4380
4381 An earlier example pointed out that the pattern
4382
4383 (sens|respons)e and \1ibility
4384
4385 matches "sense and sensibility" and "response and responsibility", but
4386 not "sense and responsibility". If instead the pattern
4387
4388 (sens|respons)e and (?1)ibility
4389
4390 is used, it does match "sense and responsibility" as well as the other
4391 two strings. Another example is given in the discussion of DEFINE
4392 above.
4393
4394 Like recursive subpatterns, a "subroutine" call is always treated as an
4395 atomic group. That is, once it has matched some of the subject string,
4396 it is never re-entered, even if it contains untried alternatives and
4397 there is a subsequent matching failure.
4398
4399 When a subpattern is used as a subroutine, processing options such as
4400 case-independence are fixed when the subpattern is defined. They cannot
4401 be changed for different calls. For example, consider this pattern:
4402
4403 (abc)(?i:(?-1))
4404
4405 It matches "abcabc". It does not match "abcABC" because the change of
4406 processing option does not affect the called subpattern.
4407
4408
4409 CALLOUTS
4410
4411 Perl has a feature whereby using the sequence (?{...}) causes arbitrary
4412 Perl code to be obeyed in the middle of matching a regular expression.
4413 This makes it possible, amongst other things, to extract different sub-
4414 strings that match the same pair of parentheses when there is a repeti-
4415 tion.
4416
4417 PCRE provides a similar feature, but of course it cannot obey arbitrary
4418 Perl code. The feature is called "callout". The caller of PCRE provides
4419 an external function by putting its entry point in the global variable
4420 pcre_callout. By default, this variable contains NULL, which disables
4421 all calling out.
4422
4423 Within a regular expression, (?C) indicates the points at which the
4424 external function is to be called. If you want to identify different
4425 callout points, you can put a number less than 256 after the letter C.
4426 The default value is zero. For example, this pattern has two callout
4427 points:
4428
4429 (?C1)abc(?C2)def
4430
4431 If the PCRE_AUTO_CALLOUT flag is passed to pcre_compile(), callouts are
4432 automatically installed before each item in the pattern. They are all
4433 numbered 255.
4434
4435 During matching, when PCRE reaches a callout point (and pcre_callout is
4436 set), the external function is called. It is provided with the number
4437 of the callout, the position in the pattern, and, optionally, one item
4438 of data originally supplied by the caller of pcre_exec(). The callout
4439 function may cause matching to proceed, to backtrack, or to fail alto-
4440 gether. A complete description of the interface to the callout function
4441 is given in the pcrecallout documentation.
4442
4443
4444 SEE ALSO
4445
4446 pcreapi(3), pcrecallout(3), pcrematching(3), pcre(3).
4447
4448
4449 AUTHOR
4450
4451 Philip Hazel
4452 University Computing Service
4453 Cambridge CB2 3QH, England.
4454
4455
4456 REVISION
4457
4458 Last updated: 29 May 2007
4459 Copyright (c) 1997-2007 University of Cambridge.
4460 ------------------------------------------------------------------------------
4461
4462
4463 PCREPARTIAL(3) PCREPARTIAL(3)
4464
4465
4466 NAME
4467 PCRE - Perl-compatible regular expressions
4468
4469
4470 PARTIAL MATCHING IN PCRE
4471
4472 In normal use of PCRE, if the subject string that is passed to
4473 pcre_exec() or pcre_dfa_exec() matches as far as it goes, but is too
4474 short to match the entire pattern, PCRE_ERROR_NOMATCH is returned.
4475 There are circumstances where it might be helpful to distinguish this
4476 case from other cases in which there is no match.
4477
4478 Consider, for example, an application where a human is required to type
4479 in data for a field with specific formatting requirements. An example
4480 might be a date in the form ddmmmyy, defined by this pattern:
4481
4482 ^\d?\d(jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)\d\d$
4483
4484 If the application sees the user's keystrokes one by one, and can check
4485 that what has been typed so far is potentially valid, it is able to
4486 raise an error as soon as a mistake is made, possibly beeping and not
4487 reflecting the character that has been typed. This immediate feedback
4488 is likely to be a better user interface than a check that is delayed
4489 until the entire string has been entered.
4490
4491 PCRE supports the concept of partial matching by means of the PCRE_PAR-
4492 TIAL option, which can be set when calling pcre_exec() or
4493 pcre_dfa_exec(). When this flag is set for pcre_exec(), the return code
4494 PCRE_ERROR_NOMATCH is converted into PCRE_ERROR_PARTIAL if at any time
4495 during the matching process the last part of the subject string matched
4496 part of the pattern. Unfortunately, for non-anchored matching, it is
4497 not possible to obtain the position of the start of the partial match.
4498 No captured data is set when PCRE_ERROR_PARTIAL is returned.
4499
4500 When PCRE_PARTIAL is set for pcre_dfa_exec(), the return code
4501 PCRE_ERROR_NOMATCH is converted into PCRE_ERROR_PARTIAL if the end of
4502 the subject is reached, there have been no complete matches, but there
4503 is still at least one matching possibility. The portion of the string
4504 that provided the partial match is set as the first matching string.
4505
4506 Using PCRE_PARTIAL disables one of PCRE's optimizations. PCRE remembers
4507 the last literal byte in a pattern, and abandons matching immediately
4508 if such a byte is not present in the subject string. This optimization
4509 cannot be used for a subject string that might match only partially.
4510
4511
4512 RESTRICTED PATTERNS FOR PCRE_PARTIAL
4513
4514 Because of the way certain internal optimizations are implemented in
4515 the pcre_exec() function, the PCRE_PARTIAL option cannot be used with
4516 all patterns. These restrictions do not apply when pcre_dfa_exec() is
4517 used. For pcre_exec(), repeated single characters such as
4518
4519 a{2,4}
4520
4521 and repeated single metasequences such as
4522
4523 \d+
4524
4525 are not permitted if the maximum number of occurrences is greater than
4526 one. Optional items such as \d? (where the maximum is one) are permit-
4527 ted. Quantifiers with any values are permitted after parentheses, so
4528 the invalid examples above can be coded thus:
4529
4530 (a){2,4}
4531 (\d)+
4532
4533 These constructions run more slowly, but for the kinds of application
4534 that are envisaged for this facility, this is not felt to be a major
4535 restriction.
4536
4537 If PCRE_PARTIAL is set for a pattern that does not conform to the
4538 restrictions, pcre_exec() returns the error code PCRE_ERROR_BADPARTIAL
4539 (-13). You can use the PCRE_INFO_OKPARTIAL call to pcre_fullinfo() to
4540 find out if a compiled pattern can be used for partial matching.
4541
4542
4543 EXAMPLE OF PARTIAL MATCHING USING PCRETEST
4544
4545 If the escape sequence \P is present in a pcretest data line, the
4546 PCRE_PARTIAL flag is used for the match. Here is a run of pcretest that
4547 uses the date example quoted above:
4548
4549 re> /^\d?\d(jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)\d\d$/
4550 data> 25jun04\P
4551 0: 25jun04
4552 1: jun
4553 data> 25dec3\P
4554 Partial match
4555 data> 3ju\P
4556 Partial match
4557 data> 3juj\P
4558 No match
4559 data> j\P
4560 No match
4561
4562 The first data string is matched completely, so pcretest shows the
4563 matched substrings. The remaining four strings do not match the com-
4564 plete pattern, but the first two are partial matches. The same test,
4565 using pcre_dfa_exec() matching (by means of the \D escape sequence),
4566 produces the following output:
4567
4568 re> /^\d?\d(jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)\d\d$/
4569 data> 25jun04\P\D
4570 0: 25jun04
4571 data> 23dec3\P\D
4572 Partial match: 23dec3
4573 data> 3ju\P\D
4574 Partial match: 3ju
4575 data> 3juj\P\D
4576 No match
4577 data> j\P\D
4578 No match
4579
4580 Notice that in this case the portion of the string that was matched is
4581 made available.
4582
4583
4584 MULTI-SEGMENT MATCHING WITH pcre_dfa_exec()
4585
4586 When a partial match has been found using pcre_dfa_exec(), it is possi-
4587 ble to continue the match by providing additional subject data and
4588 calling pcre_dfa_exec() again with the same compiled regular expres-
4589 sion, this time setting the PCRE_DFA_RESTART option. You must also pass
4590 the same working space as before, because this is where details of the
4591 previous partial match are stored. Here is an example using pcretest,
4592 using the \R escape sequence to set the PCRE_DFA_RESTART option (\P and
4593 \D are as above):
4594
4595 re> /^\d?\d(jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)\d\d$/
4596 data> 23ja\P\D
4597 Partial match: 23ja
4598 data> n05\R\D
4599 0: n05
4600
4601 The first call has "23ja" as the subject, and requests partial match-
4602 ing; the second call has "n05" as the subject for the continued
4603 (restarted) match. Notice that when the match is complete, only the
4604 last part is shown; PCRE does not retain the previously partially-
4605 matched string. It is up to the calling program to do that if it needs
4606 to.
4607
4608 You can set PCRE_PARTIAL with PCRE_DFA_RESTART to continue partial
4609 matching over multiple segments. This facility can be used to pass very
4610 long subject strings to pcre_dfa_exec(). However, some care is needed
4611 for certain types of pattern.
4612
4613 1. If the pattern contains tests for the beginning or end of a line,
4614 you need to pass the PCRE_NOTBOL or PCRE_NOTEOL options, as appropri-
4615 ate, when the subject string for any call does not contain the begin-
4616 ning or end of a line.
4617
4618 2. If the pattern contains backward assertions (including \b or \B),
4619 you need to arrange for some overlap in the subject strings to allow
4620 for this. For example, you could pass the subject in chunks that are
4621 500 bytes long, but in a buffer of 700 bytes, with the starting offset
4622 set to 200 and the previous 200 bytes at the start of the buffer.
4623
4624 3. Matching a subject string that is split into multiple segments does
4625 not always produce exactly the same result as matching over one single
4626 long string. The difference arises when there are multiple matching
4627 possibilities, because a partial match result is given only when there
4628 are no completed matches in a call to pcre_dfa_exec(). This means that
4629 as soon as the shortest match has been found, continuation to a new
4630 subject segment is no longer possible. Consider this pcretest example:
4631
4632 re> /dog(sbody)?/
4633 data> do\P\D
4634 Partial match: do
4635 data> gsb\R\P\D
4636 0: g
4637 data> dogsbody\D
4638 0: dogsbody
4639 1: dog
4640
4641 The pattern matches the words "dog" or "dogsbody". When the subject is
4642 presented in several parts ("do" and "gsb" being the first two) the
4643 match stops when "dog" has been found, and it is not possible to con-
4644 tinue. On the other hand, if "dogsbody" is presented as a single
4645 string, both matches are found.
4646
4647 Because of this phenomenon, it does not usually make sense to end a
4648 pattern that is going to be matched in this way with a variable repeat.
4649
4650 4. Patterns that contain alternatives at the top level which do not all
4651 start with the same pattern item may not work as expected. For example,
4652 consider this pattern:
4653
4654 1234|3789
4655
4656 If the first part of the subject is "ABC123", a partial match of the
4657 first alternative is found at offset 3. There is no partial match for
4658 the second alternative, because such a match does not start at the same
4659 point in the subject string. Attempting to continue with the string
4660 "789" does not yield a match because only those alternatives that match
4661 at one point in the subject are remembered. The problem arises because
4662 the start of the second alternative matches within the first alterna-
4663 tive. There is no problem with anchored patterns or patterns such as:
4664
4665 1234|ABCD
4666
4667 where no string can be a partial match for both alternatives.
4668
4669
4670 AUTHOR
4671
4672 Philip Hazel
4673 University Computing Service
4674 Cambridge CB2 3QH, England.
4675
4676
4677 REVISION
4678
4679 Last updated: 04 June 2007
4680 Copyright (c) 1997-2007 University of Cambridge.
4681 ------------------------------------------------------------------------------
4682
4683
4684 PCREPRECOMPILE(3) PCREPRECOMPILE(3)
4685
4686
4687 NAME
4688 PCRE - Perl-compatible regular expressions
4689
4690
4691 SAVING AND RE-USING PRECOMPILED PCRE PATTERNS
4692
4693 If you are running an application that uses a large number of regular
4694 expression patterns, it may be useful to store them in a precompiled
4695 form instead of having to compile them every time the application is
4696 run. If you are not using any private character tables (see the
4697 pcre_maketables() documentation), this is relatively straightforward.
4698 If you are using private tables, it is a little bit more complicated.
4699
4700 If you save compiled patterns to a file, you can copy them to a differ-
4701 ent host and run them there. This works even if the new host has the
4702 opposite endianness to the one on which the patterns were compiled.
4703 There may be a small performance penalty, but it should be insignifi-
4704 cant. However, compiling regular expressions with one version of PCRE
4705 for use with a different version is not guaranteed to work and may
4706 cause crashes.
4707
4708
4709 SAVING A COMPILED PATTERN
4710 The value returned by pcre_compile() points to a single block of memory
4711 that holds the compiled pattern and associated data. You can find the
4712 length of this block in bytes by calling pcre_fullinfo() with an argu-
4713 ment of PCRE_INFO_SIZE. You can then save the data in any appropriate
4714 manner. Here is sample code that compiles a pattern and writes it to a
4715 file. It assumes that the variable fd refers to a file that is open for
4716 output:
4717
4718 int erroroffset, rc, size;
4719 char *error;
4720 pcre *re;
4721
4722 re = pcre_compile("my pattern", 0, &error, &erroroffset, NULL);
4723 if (re == NULL) { ... handle errors ... }
4724 rc = pcre_fullinfo(re, NULL, PCRE_INFO_SIZE, &size);
4725 if (rc < 0) { ... handle errors ... }
4726 rc = fwrite(re, 1, size, fd);
4727 if (rc != size) { ... handle errors ... }
4728
4729 In this example, the bytes that comprise the compiled pattern are
4730 copied exactly. Note that this is binary data that may contain any of
4731 the 256 possible byte values. On systems that make a distinction
4732 between binary and non-binary data, be sure that the file is opened for
4733 binary output.
4734
4735 If you want to write more than one pattern to a file, you will have to
4736 devise a way of separating them. For binary data, preceding each pat-
4737 tern with its length is probably the most straightforward approach.
4738 Another possibility is to write out the data in hexadecimal instead of
4739 binary, one pattern to a line.
4740
4741 Saving compiled patterns in a file is only one possible way of storing
4742 them for later use. They could equally well be saved in a database, or
4743 in the memory of some daemon process that passes them via sockets to
4744 the processes that want them.
4745
4746 If the pattern has been studied, it is also possible to save the study
4747 data in a similar way to the compiled pattern itself. When studying
4748 generates additional information, pcre_study() returns a pointer to a
4749 pcre_extra data block. Its format is defined in the section on matching
4750 a pattern in the pcreapi documentation. The study_data field points to
4751 the binary study data, and this is what you must save (not the
4752 pcre_extra block itself). The length of the study data can be obtained
4753 by calling pcre_fullinfo() with an argument of PCRE_INFO_STUDYSIZE.
4754 Remember to check that pcre_study() did return a non-NULL value before
4755 trying to save the study data.
4756
4757
4758 RE-USING A PRECOMPILED PATTERN
4759
4760 Re-using a precompiled pattern is straightforward. Having reloaded it
4761 into main memory, you pass its pointer to pcre_exec() or
4762 pcre_dfa_exec() in the usual way. This should work even on another
4763 host, and even if that host has the opposite endianness to the one
4764 where the pattern was compiled.
4765
4766 However, if you passed a pointer to custom character tables when the
4767 pattern was compiled (the tableptr argument of pcre_compile()), you
4768 must now pass a similar pointer to pcre_exec() or pcre_dfa_exec(),
4769 because the value saved with the compiled pattern will obviously be
4770 nonsense. A field in a pcre_extra() block is used to pass this data, as
4771 described in the section on matching a pattern in the pcreapi documen-
4772 tation.
4773
4774 If you did not provide custom character tables when the pattern was
4775 compiled, the pointer in the compiled pattern is NULL, which causes
4776 pcre_exec() to use PCRE's internal tables. Thus, you do not need to
4777 take any special action at run time in this case.
4778
4779 If you saved study data with the compiled pattern, you need to create
4780 your own pcre_extra data block and set the study_data field to point to
4781 the reloaded study data. You must also set the PCRE_EXTRA_STUDY_DATA
4782 bit in the flags field to indicate that study data is present. Then
4783 pass the pcre_extra block to pcre_exec() or pcre_dfa_exec() in the
4784 usual way.
4785
4786
4787 COMPATIBILITY WITH DIFFERENT PCRE RELEASES
4788
4789 The layout of the control block that is at the start of the data that
4790 makes up a compiled pattern was changed for release 5.0. If you have
4791 any saved patterns that were compiled with previous releases (not a
4792 facility that was previously advertised), you will have to recompile
4793 them for release 5.0 and above.
4794
4795 If you have any saved patterns in UTF-8 mode that use \p or \P that
4796 were compiled with any release up to and including 6.4, you will have
4797 to recompile them for release 6.5 and above.
4798
4799 All saved patterns from earlier releases must be recompiled for release
4800 7.0 or higher, because there was an internal reorganization at that
4801 release.
4802
4803
4804 AUTHOR
4805
4806 Philip Hazel
4807 University Computing Service
4808 Cambridge CB2 3QH, England.
4809
4810
4811 REVISION
4812
4813 Last updated: 24 April 2007
4814 Copyright (c) 1997-2007 University of Cambridge.
4815 ------------------------------------------------------------------------------
4816
4817
4818 PCREPERFORM(3) PCREPERFORM(3)
4819
4820
4821 NAME
4822 PCRE - Perl-compatible regular expressions
4823
4824
4825 PCRE PERFORMANCE
4826
4827 Two aspects of performance are discussed below: memory usage and pro-
4828 cessing time. The way you express your pattern as a regular expression
4829 can affect both of them.
4830
4831
4832 MEMORY USAGE
4833
4834 Patterns are compiled by PCRE into a reasonably efficient byte code, so
4835 that most simple patterns do not use much memory. However, there is one
4836 case where memory usage can be unexpectedly large. When a parenthesized
4837 subpattern has a quantifier with a minimum greater than 1 and/or a lim-
4838 ited maximum, the whole subpattern is repeated in the compiled code.
4839 For example, the pattern
4840
4841 (abc|def){2,4}
4842
4843 is compiled as if it were
4844
4845 (abc|def)(abc|def)((abc|def)(abc|def)?)?
4846
4847 (Technical aside: It is done this way so that backtrack points within
4848 each of the repetitions can be independently maintained.)
4849
4850 For regular expressions whose quantifiers use only small numbers, this
4851 is not usually a problem. However, if the numbers are large, and par-
4852 ticularly if such repetitions are nested, the memory usage can become
4853 an embarrassment. For example, the very simple pattern
4854
4855 ((ab){1,1000}c){1,3}
4856
4857 uses 51K bytes when compiled. When PCRE is compiled with its default
4858 internal pointer size of two bytes, the size limit on a compiled pat-
4859 tern is 64K, and this is reached with the above pattern if the outer
4860 repetition is increased from 3 to 4. PCRE can be compiled to use larger
4861 internal pointers and thus handle larger compiled patterns, but it is
4862 better to try to rewrite your pattern to use less memory if you can.
4863
4864 One way of reducing the memory usage for such patterns is to make use
4865 of PCRE's "subroutine" facility. Re-writing the above pattern as
4866
4867 ((ab)(?2){0,999}c)(?1){0,2}
4868
4869 reduces the memory requirements to 18K, and indeed it remains under 20K
4870 even with the outer repetition increased to 100. However, this pattern
4871 is not exactly equivalent, because the "subroutine" calls are treated
4872 as atomic groups into which there can be no backtracking if there is a
4873 subsequent matching failure. Therefore, PCRE cannot do this kind of
4874 rewriting automatically. Furthermore, there is a noticeable loss of
4875 speed when executing the modified pattern. Nevertheless, if the atomic
4876 grouping is not a problem and the loss of speed is acceptable, this
4877 kind of rewriting will allow you to process patterns that PCRE cannot
4878 otherwise handle.
4879
4880
4881 PROCESSING TIME
4882
4883 Certain items in regular expression patterns are processed more effi-
4884 ciently than others. It is more efficient to use a character class like
4885 [aeiou] than a set of single-character alternatives such as
4886 (a|e|i|o|u). In general, the simplest construction that provides the
4887 required behaviour is usually the most efficient. Jeffrey Friedl's book
4888 contains a lot of useful general discussion about optimizing regular
4889 expressions for efficient performance. This document contains a few
4890 observations about PCRE.
4891
4892 Using Unicode character properties (the \p, \P, and \X escapes) is
4893 slow, because PCRE has to scan a structure that contains data for over
4894 fifteen thousand characters whenever it needs a character's property.
4895 If you can find an alternative pattern that does not use character
4896 properties, it will probably be faster.
4897
4898 When a pattern begins with .* not in parentheses, or in parentheses
4899 that are not the subject of a backreference, and the PCRE_DOTALL option
4900 is set, the pattern is implicitly anchored by PCRE, since it can match
4901 only at the start of a subject string. However, if PCRE_DOTALL is not
4902 set, PCRE cannot make this optimization, because the . metacharacter
4903 does not then match a newline, and if the subject string contains new-
4904 lines, the pattern may match from the character immediately following
4905 one of them instead of from the very start. For example, the pattern
4906
4907 .*second
4908
4909 matches the subject "first\nand second" (where \n stands for a newline
4910 character), with the match starting at the seventh character. In order
4911 to do this, PCRE has to retry the match starting after every newline in
4912 the subject.
4913
4914 If you are using such a pattern with subject strings that do not con-
4915 tain newlines, the best performance is obtained by setting PCRE_DOTALL,
4916 or starting the pattern with ^.* or ^.*? to indicate explicit anchor-
4917 ing. That saves PCRE from having to scan along the subject looking for
4918 a newline to restart at.
4919
4920 Beware of patterns that contain nested indefinite repeats. These can
4921 take a long time to run when applied to a string that does not match.
4922 Consider the pattern fragment
4923
4924 ^(a+)*
4925
4926 This can match "aaaa" in 16 different ways, and this number increases
4927 very rapidly as the string gets longer. (The * repeat can match 0, 1,
4928 2, 3, or 4 times, and for each of those cases other than 0 or 4, the +
4929 repeats can match different numbers of times.) When the remainder of
4930 the pattern is such that the entire match is going to fail, PCRE has in
4931 principle to try every possible variation, and this can take an
4932 extremely long time, even for relatively short strings.
4933
4934 An optimization catches some of the more simple cases such as
4935
4936 (a+)*b
4937
4938 where a literal character follows. Before embarking on the standard
4939 matching procedure, PCRE checks that there is a "b" later in the sub-
4940 ject string, and if there is not, it fails the match immediately. How-
4941 ever, when there is no following literal this optimization cannot be
4942 used. You can see the difference by comparing the behaviour of
4943
4944 (a+)*\d
4945
4946 with the pattern above. The former gives a failure almost instantly
4947 when applied to a whole line of "a" characters, whereas the latter
4948 takes an appreciable time with strings longer than about 20 characters.
4949
4950 In many cases, the solution to this kind of performance issue is to use
4951 an atomic group or a possessive quantifier.
4952
4953
4954 AUTHOR
4955
4956 Philip Hazel
4957 University Computing Service
4958 Cambridge CB2 3QH, England.
4959
4960
4961 REVISION
4962
4963 Last updated: 06 March 2007
4964 Copyright (c) 1997-2007 University of Cambridge.
4965 ------------------------------------------------------------------------------
4966
4967
4968 PCREPOSIX(3) PCREPOSIX(3)
4969
4970
4971 NAME
4972 PCRE - Perl-compatible regular expressions.
4973
4974
4975 SYNOPSIS OF POSIX API
4976
4977 #include <pcreposix.h>
4978
4979 int regcomp(regex_t *preg, const char *pattern,
4980 int cflags);
4981
4982 int regexec(regex_t *preg, const char *string,
4983 size_t nmatch, regmatch_t pmatch[], int eflags);
4984
4985 size_t regerror(int errcode, const regex_t *preg,
4986 char *errbuf, size_t errbuf_size);
4987
4988 void regfree(regex_t *preg);
4989
4990
4991 DESCRIPTION
4992
4993 This set of functions provides a POSIX-style API to the PCRE regular
4994 expression package. See the pcreapi documentation for a description of
4995 PCRE's native API, which contains much additional functionality.
4996
4997 The functions described here are just wrapper functions that ultimately
4998 call the PCRE native API. Their prototypes are defined in the
4999 pcreposix.h header file, and on Unix systems the library itself is
5000 called pcreposix.a, so can be accessed by adding -lpcreposix to the
5001 command for linking an application that uses them. Because the POSIX
5002 functions call the native ones, it is also necessary to add -lpcre.
5003
5004 I have implemented only those option bits that can be reasonably mapped
5005 to PCRE native options. In addition, the option REG_EXTENDED is defined
5006 with the value zero. This has no effect, but since programs that are
5007 written to the POSIX interface often use it, this makes it easier to
5008 slot in PCRE as a replacement library. Other POSIX options are not even
5009 defined.
5010
5011 When PCRE is called via these functions, it is only the API that is
5012 POSIX-like in style. The syntax and semantics of the regular expres-
5013 sions themselves are still those of Perl, subject to the setting of
5014 various PCRE options, as described below. "POSIX-like in style" means
5015 that the API approximates to the POSIX definition; it is not fully
5016 POSIX-compatible, and in multi-byte encoding domains it is probably
5017 even less compatible.
5018
5019 The header for these functions is supplied as pcreposix.h to avoid any
5020 potential clash with other POSIX libraries. It can, of course, be
5021 renamed or aliased as regex.h, which is the "correct" name. It provides
5022 two structure types, regex_t for compiled internal forms, and reg-
5023 match_t for returning captured substrings. It also defines some con-
5024 stants whose names start with "REG_"; these are used for setting
5025 options and identifying error codes.
5026
5027
5028 COMPILING A PATTERN
5029
5030 The function regcomp() is called to compile a pattern into an internal
5031 form. The pattern is a C string terminated by a binary zero, and is
5032 passed in the argument pattern. The preg argument is a pointer to a
5033 regex_t structure that is used as a base for storing information about
5034 the compiled regular expression.
5035
5036 The argument cflags is either zero, or contains one or more of the bits
5037 defined by the following macros:
5038
5039 REG_DOTALL
5040
5041 The PCRE_DOTALL option is set when the regular expression is passed for
5042 compilation to the native function. Note that REG_DOTALL is not part of
5043 the POSIX standard.
5044
5045 REG_ICASE
5046
5047 The PCRE_CASELESS option is set when the regular expression is passed
5048 for compilation to the native function.
5049
5050 REG_NEWLINE
5051
5052 The PCRE_MULTILINE option is set when the regular expression is passed
5053 for compilation to the native function. Note that this does not mimic
5054 the defined POSIX behaviour for REG_NEWLINE (see the following sec-
5055 tion).
5056
5057 REG_NOSUB
5058
5059 The PCRE_NO_AUTO_CAPTURE option is set when the regular expression is
5060 passed for compilation to the native function. In addition, when a pat-
5061 tern that is compiled with this flag is passed to regexec() for match-
5062 ing, the nmatch and pmatch arguments are ignored, and no captured
5063 strings are returned.
5064
5065 REG_UTF8
5066
5067 The PCRE_UTF8 option is set when the regular expression is passed for
5068 compilation to the native function. This causes the pattern itself and
5069 all data strings used for matching it to be treated as UTF-8 strings.
5070 Note that REG_UTF8 is not part of the POSIX standard.
5071
5072 In the absence of these flags, no options are passed to the native
5073 function. This means the the regex is compiled with PCRE default
5074 semantics. In particular, the way it handles newline characters in the
5075 subject string is the Perl way, not the POSIX way. Note that setting
5076 PCRE_MULTILINE has only some of the effects specified for REG_NEWLINE.
5077 It does not affect the way newlines are matched by . (they aren't) or
5078 by a negative class such as [^a] (they are).
5079
5080 The yield of regcomp() is zero on success, and non-zero otherwise. The
5081 preg structure is filled in on success, and one member of the structure
5082 is public: re_nsub contains the number of capturing subpatterns in the
5083 regular expression. Various error codes are defined in the header file.
5084
5085
5086 MATCHING NEWLINE CHARACTERS
5087
5088 This area is not simple, because POSIX and Perl take different views of
5089 things. It is not possible to get PCRE to obey POSIX semantics, but
5090 then PCRE was never intended to be a POSIX engine. The following table
5091 lists the different possibilities for matching newline characters in
5092 PCRE:
5093
5094 Default Change with
5095
5096 . matches newline no PCRE_DOTALL
5097 newline matches [^a] yes not changeable
5098 $ matches \n at end yes PCRE_DOLLARENDONLY
5099 $ matches \n in middle no PCRE_MULTILINE
5100 ^ matches \n in middle no PCRE_MULTILINE
5101
5102 This is the equivalent table for POSIX:
5103
5104 Default Change with
5105
5106 . matches newline yes REG_NEWLINE
5107 newline matches [^a] yes REG_NEWLINE
5108 $ matches \n at end no REG_NEWLINE
5109 $ matches \n in middle no REG_NEWLINE
5110 ^ matches \n in middle no REG_NEWLINE
5111
5112 PCRE's behaviour is the same as Perl's, except that there is no equiva-
5113 lent for PCRE_DOLLAR_ENDONLY in Perl. In both PCRE and Perl, there is
5114 no way to stop newline from matching [^a].
5115
5116 The default POSIX newline handling can be obtained by setting
5117 PCRE_DOTALL and PCRE_DOLLAR_ENDONLY, but there is no way to make PCRE
5118 behave exactly as for the REG_NEWLINE action.
5119
5120
5121 MATCHING A PATTERN
5122
5123 The function regexec() is called to match a compiled pattern preg
5124 against a given string, which is terminated by a zero byte, subject to
5125 the options in eflags. These can be:
5126
5127 REG_NOTBOL
5128
5129 The PCRE_NOTBOL option is set when calling the underlying PCRE matching
5130 function.
5131
5132 REG_NOTEOL
5133
5134 The PCRE_NOTEOL option is set when calling the underlying PCRE matching
5135 function.
5136
5137 If the pattern was compiled with the REG_NOSUB flag, no data about any
5138 matched strings is returned. The nmatch and pmatch arguments of
5139 regexec() are ignored.
5140
5141 Otherwise,the portion of the string that was matched, and also any cap-
5142 tured substrings, are returned via the pmatch argument, which points to
5143 an array of nmatch structures of type regmatch_t, containing the mem-
5144 bers rm_so and rm_eo. These contain the offset to the first character
5145 of each substring and the offset to the first character after the end
5146 of each substring, respectively. The 0th element of the vector relates
5147 to the entire portion of string that was matched; subsequent elements
5148 relate to the capturing subpatterns of the regular expression. Unused
5149 entries in the array have both structure members set to -1.
5150
5151 A successful match yields a zero return; various error codes are
5152 defined in the header file, of which REG_NOMATCH is the "expected"
5153 failure code.
5154
5155
5156 ERROR MESSAGES
5157
5158 The regerror() function maps a non-zero errorcode from either regcomp()
5159 or regexec() to a printable message. If preg is not NULL, the error
5160 should have arisen from the use of that structure. A message terminated
5161 by a binary zero is placed in errbuf. The length of the message,
5162 including the zero, is limited to errbuf_size. The yield of the func-
5163 tion is the size of buffer needed to hold the whole message.
5164
5165
5166 MEMORY USAGE
5167
5168 Compiling a regular expression causes memory to be allocated and asso-
5169 ciated with the preg structure. The function regfree() frees all such
5170 memory, after which preg may no longer be used as a compiled expres-
5171 sion.
5172
5173
5174 AUTHOR
5175
5176 Philip Hazel
5177 University Computing Service
5178 Cambridge CB2 3QH, England.
5179
5180
5181 REVISION
5182
5183 Last updated: 06 March 2007
5184 Copyright (c) 1997-2007 University of Cambridge.
5185 ------------------------------------------------------------------------------
5186
5187
5188 PCRECPP(3) PCRECPP(3)
5189
5190
5191 NAME
5192 PCRE - Perl-compatible regular expressions.
5193
5194
5195 SYNOPSIS OF C++ WRAPPER
5196
5197 #include <pcrecpp.h>
5198
5199
5200 DESCRIPTION
5201
5202 The C++ wrapper for PCRE was provided by Google Inc. Some additional
5203 functionality was added by Giuseppe Maxia. This brief man page was con-
5204 structed from the notes in the pcrecpp.h file, which should be con-
5205 sulted for further details.
5206
5207
5208 MATCHING INTERFACE
5209
5210 The "FullMatch" operation checks that supplied text matches a supplied
5211 pattern exactly. If pointer arguments are supplied, it copies matched
5212 sub-strings that match sub-patterns into them.
5213
5214 Example: successful match
5215 pcrecpp::RE re("h.*o");
5216 re.FullMatch("hello");
5217
5218 Example: unsuccessful match (requires full match):
5219 pcrecpp::RE re("e");
5220 !re.FullMatch("hello");
5221
5222 Example: creating a temporary RE object:
5223 pcrecpp::RE("h.*o").FullMatch("hello");
5224
5225 You can pass in a "const char*" or a "string" for "text". The examples
5226 below tend to use a const char*. You can, as in the different examples
5227 above, store the RE object explicitly in a variable or use a temporary
5228 RE object. The examples below use one mode or the other arbitrarily.
5229 Either could correctly be used for any of these examples.
5230
5231 You must supply extra pointer arguments to extract matched subpieces.
5232
5233 Example: extracts "ruby" into "s" and 1234 into "i"
5234 int i;
5235 string s;
5236 pcrecpp::RE re("(\\w+):(\\d+)");
5237 re.FullMatch("ruby:1234", &s, &i);
5238
5239 Example: does not try to extract any extra sub-patterns
5240 re.FullMatch("ruby:1234", &s);
5241
5242 Example: does not try to extract into NULL
5243 re.FullMatch("ruby:1234", NULL, &i);
5244
5245 Example: integer overflow causes failure
5246 !re.FullMatch("ruby:1234567891234", NULL, &i);
5247
5248 Example: fails because there aren't enough sub-patterns:
5249 !pcrecpp::RE("\\w+:\\d+").FullMatch("ruby:1234", &s);
5250
5251 Example: fails because string cannot be stored in integer
5252 !pcrecpp::RE("(.*)").FullMatch("ruby", &i);
5253
5254 The provided pointer arguments can be pointers to any scalar numeric
5255 type, or one of:
5256
5257 string (matched piece is copied to string)
5258 StringPiece (StringPiece is mutated to point to matched piece)
5259 T (where "bool T::ParseFrom(const char*, int)" exists)
5260 NULL (the corresponding matched sub-pattern is not copied)
5261
5262 The function returns true iff all of the following conditions are sat-
5263 isfied:
5264
5265 a. "text" matches "pattern" exactly;
5266
5267 b. The number of matched sub-patterns is >= number of supplied
5268 pointers;
5269
5270 c. The "i"th argument has a suitable type for holding the
5271 string captured as the "i"th sub-pattern. If you pass in
5272 NULL for the "i"th argument, or pass fewer arguments than
5273 number of sub-patterns, "i"th captured sub-pattern is
5274 ignored.
5275
5276 CAVEAT: An optional sub-pattern that does not exist in the matched
5277 string is assigned the empty string. Therefore, the following will
5278 return false (because the empty string is not a valid number):
5279
5280 int number;
5281 pcrecpp::RE::FullMatch("abc", "[a-z]+(\\d+)?", &number);
5282
5283 The matching interface supports at most 16 arguments per call. If you
5284 need more, consider using the more general interface
5285 pcrecpp::RE::DoMatch. See pcrecpp.h for the signature for DoMatch.
5286
5287
5288 QUOTING METACHARACTERS
5289
5290 You can use the "QuoteMeta" operation to insert backslashes before all
5291 potentially meaningful characters in a string. The returned string,
5292 used as a regular expression, will exactly match the original string.
5293
5294 Example:
5295 string quoted = RE::QuoteMeta(unquoted);
5296
5297 Note that it's legal to escape a character even if it has no special
5298 meaning in a regular expression -- so this function does that. (This
5299 also makes it identical to the perl function of the same name; see
5300 "perldoc -f quotemeta".) For example, "1.5-2.0?" becomes
5301 "1\.5\-2\.0\?".
5302
5303
5304 PARTIAL MATCHES
5305
5306 You can use the "PartialMatch" operation when you want the pattern to
5307 match any substring of the text.
5308
5309 Example: simple search for a string:
5310 pcrecpp::RE("ell").PartialMatch("hello");
5311
5312 Example: find first number in a string:
5313 int number;
5314 pcrecpp::RE re("(\\d+)");
5315 re.PartialMatch("x*100 + 20", &number);
5316 assert(number == 100);
5317
5318
5319 UTF-8 AND THE MATCHING INTERFACE
5320
5321 By default, pattern and text are plain text, one byte per character.
5322 The UTF8 flag, passed to the constructor, causes both pattern and
5323 string to be treated as UTF-8 text, still a byte stream but potentially
5324 multiple bytes per character. In practice, the text is likelier to be
5325 UTF-8 than the pattern, but the match returned may depend on the UTF8
5326 flag, so always use it when matching UTF8 text. For example, "." will
5327 match one byte normally but with UTF8 set may match up to three bytes
5328 of a multi-byte character.
5329
5330 Example:
5331 pcrecpp::RE_Options options;
5332 options.set_utf8();
5333 pcrecpp::RE re(utf8_pattern, options);
5334 re.FullMatch(utf8_string);
5335
5336 Example: using the convenience function UTF8():
5337 pcrecpp::RE re(utf8_pattern, pcrecpp::UTF8());
5338 re.FullMatch(utf8_string);
5339
5340 NOTE: The UTF8 flag is ignored if pcre was not configured with the
5341 --enable-utf8 flag.
5342
5343
5344 PASSING MODIFIERS TO THE REGULAR EXPRESSION ENGINE
5345
5346 PCRE defines some modifiers to change the behavior of the regular
5347 expression engine. The C++ wrapper defines an auxiliary class,
5348 RE_Options, as a vehicle to pass such modifiers to a RE class. Cur-
5349 rently, the following modifiers are supported:
5350
5351 modifier description Perl corresponding
5352
5353 PCRE_CASELESS case insensitive match /i
5354 PCRE_MULTILINE multiple lines match /m
5355 PCRE_DOTALL dot matches newlines /s
5356 PCRE_DOLLAR_ENDONLY $ matches only at end N/A
5357 PCRE_EXTRA strict escape parsing N/A
5358 PCRE_EXTENDED ignore whitespaces /x
5359 PCRE_UTF8 handles UTF8 chars built-in
5360 PCRE_UNGREEDY reverses * and *? N/A
5361 PCRE_NO_AUTO_CAPTURE disables capturing parens N/A (*)
5362
5363 (*) Both Perl and PCRE allow non capturing parentheses by means of the
5364 "?:" modifier within the pattern itself. e.g. (?:ab|cd) does not cap-
5365 ture, while (ab|cd) does.
5366
5367 For a full account on how each modifier works, please check the PCRE
5368 API reference page.
5369
5370 For each modifier, there are two member functions whose name is made
5371 out of the modifier in lowercase, without the "PCRE_" prefix. For
5372 instance, PCRE_CASELESS is handled by
5373
5374 bool caseless()
5375
5376 which returns true if the modifier is set, and
5377
5378 RE_Options & set_caseless(bool)
5379
5380 which sets or unsets the modifier. Moreover, PCRE_EXTRA_MATCH_LIMIT can
5381 be accessed through the set_match_limit() and match_limit() member
5382 functions. Setting match_limit to a non-zero value will limit the exe-
5383 cution of pcre to keep it from doing bad things like blowing the stack
5384 or taking an eternity to return a result. A value of 5000 is good
5385 enough to stop stack blowup in a 2MB thread stack. Setting match_limit
5386 to zero disables match limiting. Alternatively, you can call
5387 match_limit_recursion() which uses PCRE_EXTRA_MATCH_LIMIT_RECURSION to
5388 limit how much PCRE recurses. match_limit() limits the number of
5389 matches PCRE does; match_limit_recursion() limits the depth of internal
5390 recursion, and therefore the amount of stack that is used.
5391
5392 Normally, to pass one or more modifiers to a RE class, you declare a
5393 RE_Options object, set the appropriate options, and pass this object to
5394 a RE constructor. Example:
5395
5396 RE_options opt;
5397 opt.set_caseless(true);
5398 if (RE("HELLO", opt).PartialMatch("hello world")) ...
5399
5400 RE_options has two constructors. The default constructor takes no argu-
5401 ments and creates a set of flags that are off by default. The optional
5402 parameter option_flags is to facilitate transfer of legacy code from C
5403 programs. This lets you do
5404
5405 RE(pattern,
5406 RE_Options(PCRE_CASELESS|PCRE_MULTILINE)).PartialMatch(str);
5407
5408 However, new code is better off doing
5409
5410 RE(pattern,
5411 RE_Options().set_caseless(true).set_multiline(true))
5412 .PartialMatch(str);
5413
5414 If you are going to pass one of the most used modifiers, there are some
5415 convenience functions that return a RE_Options class with the appropri-
5416 ate modifier already set: CASELESS(), UTF8(), MULTILINE(), DOTALL(),
5417 and EXTENDED().
5418
5419 If you need to set several options at once, and you don't want to go
5420 through the pains of declaring a RE_Options object and setting several
5421 options, there is a parallel method that give you such ability on the
5422 fly. You can concatenate several set_xxxxx() member functions, since
5423 each of them returns a reference to its class object. For example, to
5424 pass PCRE_CASELESS, PCRE_EXTENDED, and PCRE_MULTILINE to a RE with one
5425 statement, you may write:
5426
5427 RE(" ^ xyz \\s+ .* blah$",
5428 RE_Options()
5429 .set_caseless(true)
5430 .set_extended(true)
5431 .set_multiline(true)).PartialMatch(sometext);
5432
5433
5434 SCANNING TEXT INCREMENTALLY
5435
5436 The "Consume" operation may be useful if you want to repeatedly match
5437 regular expressions at the front of a string and skip over them as they
5438 match. This requires use of the "StringPiece" type, which represents a
5439 sub-range of a real string. Like RE, StringPiece is defined in the
5440 pcrecpp namespace.
5441
5442 Example: read lines of the form "var = value" from a string.
5443 string contents = ...; // Fill string somehow
5444 pcrecpp::StringPiece input(contents); // Wrap in a StringPiece
5445
5446 string var;
5447 int value;
5448 pcrecpp::RE re("(\\w+) = (\\d+)\n");
5449 while (re.Consume(&input, &var, &value)) {
5450 ...;
5451 }
5452
5453 Each successful call to "Consume" will set "var/value", and also
5454 advance "input" so it points past the matched text.
5455
5456 The "FindAndConsume" operation is similar to "Consume" but does not
5457 anchor your match at the beginning of the string. For example, you
5458 could extract all words from a string by repeatedly calling
5459
5460 pcrecpp::RE("(\\w+)").FindAndConsume(&input, &word)
5461
5462
5463 PARSING HEX/OCTAL/C-RADIX NUMBERS
5464
5465 By default, if you pass a pointer to a numeric value, the corresponding
5466 text is interpreted as a base-10 number. You can instead wrap the
5467 pointer with a call to one of the operators Hex(), Octal(), or CRadix()
5468 to interpret the text in another base. The CRadix operator interprets
5469 C-style "0" (base-8) and "0x" (base-16) prefixes, but defaults to
5470 base-10.
5471
5472 Example:
5473 int a, b, c, d;
5474 pcrecpp::RE re("(.*) (.*) (.*) (.*)");
5475 re.FullMatch("100 40 0100 0x40",
5476 pcrecpp::Octal(&a), pcrecpp::Hex(&b),
5477 pcrecpp::CRadix(&c), pcrecpp::CRadix(&d));
5478
5479 will leave 64 in a, b, c, and d.
5480
5481
5482 REPLACING PARTS OF STRINGS
5483
5484 You can replace the first match of "pattern" in "str" with "rewrite".
5485 Within "rewrite", backslash-escaped digits (\1 to \9) can be used to
5486 insert text matching corresponding parenthesized group from the pat-
5487 tern. \0 in "rewrite" refers to the entire matching text. For example:
5488
5489 string s = "yabba dabba doo";
5490 pcrecpp::RE("b+").Replace("d", &s);
5491
5492 will leave "s" containing "yada dabba doo". The result is true if the
5493 pattern matches and a replacement occurs, false otherwise.
5494
5495 GlobalReplace is like Replace except that it replaces all occurrences
5496 of the pattern in the string with the rewrite. Replacements are not
5497 subject to re-matching. For example:
5498
5499 string s = "yabba dabba doo";
5500 pcrecpp::RE("b+").GlobalReplace("d", &s);
5501
5502 will leave "s" containing "yada dada doo". It returns the number of
5503 replacements made.
5504
5505 Extract is like Replace, except that if the pattern matches, "rewrite"
5506 is copied into "out" (an additional argument) with substitutions. The
5507 non-matching portions of "text" are ignored. Returns true iff a match
5508 occurred and the extraction happened successfully; if no match occurs,
5509 the string is left unaffected.
5510
5511
5512 AUTHOR
5513
5514 The C++ wrapper was contributed by Google Inc.
5515 Copyright (c) 2007 Google Inc.
5516
5517
5518 REVISION
5519
5520 Last updated: 06 March 2007
5521 ------------------------------------------------------------------------------
5522
5523
5524 PCRESAMPLE(3) PCRESAMPLE(3)
5525
5526
5527 NAME
5528 PCRE - Perl-compatible regular expressions
5529
5530
5531 PCRE SAMPLE PROGRAM
5532
5533 A simple, complete demonstration program, to get you started with using
5534 PCRE, is supplied in the file pcredemo.c in the PCRE distribution.
5535
5536 The program compiles the regular expression that is its first argument,
5537 and matches it against the subject string in its second argument. No
5538 PCRE options are set, and default character tables are used. If match-
5539 ing succeeds, the program outputs the portion of the subject that
5540 matched, together with the contents of any captured substrings.
5541
5542 If the -g option is given on the command line, the program then goes on
5543 to check for further matches of the same regular expression in the same
5544 subject string. The logic is a little bit tricky because of the possi-
5545 bility of matching an empty string. Comments in the code explain what
5546 is going on.
5547
5548 If PCRE is installed in the standard include and library directories
5549 for your system, you should be able to compile the demonstration pro-
5550 gram using this command:
5551
5552 gcc -o pcredemo pcredemo.c -lpcre
5553
5554 If PCRE is installed elsewhere, you may need to add additional options
5555 to the command line. For example, on a Unix-like system that has PCRE
5556 installed in /usr/local, you can compile the demonstration program
5557 using a command like this:
5558
5559 gcc -o pcredemo -I/usr/local/include pcredemo.c \
5560 -L/usr/local/lib -lpcre
5561
5562 Once you have compiled the demonstration program, you can run simple
5563 tests like this:
5564
5565 ./pcredemo 'cat|dog' 'the cat sat on the mat'
5566 ./pcredemo -g 'cat|dog' 'the dog sat on the cat'
5567
5568 Note that there is a much more comprehensive test program, called
5569 pcretest, which supports many more facilities for testing regular
5570 expressions and the PCRE library. The pcredemo program is provided as a
5571 simple coding example.
5572
5573 On some operating systems (e.g. Solaris), when PCRE is not installed in
5574 the standard library directory, you may get an error like this when you
5575 try to run pcredemo:
5576
5577 ld.so.1: a.out: fatal: libpcre.so.0: open failed: No such file or
5578 directory
5579
5580 This is caused by the way shared library support works on those sys-
5581 tems. You need to add
5582
5583 -R/usr/local/lib
5584
5585 (for example) to the compile command to get round this problem.
5586
5587
5588 AUTHOR
5589
5590 Philip Hazel
5591 University Computing Service
5592 Cambridge CB2 3QH, England.
5593
5594
5595 REVISION
5596
5597 Last updated: 06 March 2007
5598 Copyright (c) 1997-2007 University of Cambridge.
5599 ------------------------------------------------------------------------------
5600 PCRESTACK(3) PCRESTACK(3)
5601
5602
5603 NAME
5604 PCRE - Perl-compatible regular expressions
5605
5606
5607 PCRE DISCUSSION OF STACK USAGE
5608
5609 When you call pcre_exec(), it makes use of an internal function called
5610 match(). This calls itself recursively at branch points in the pattern,
5611 in order to remember the state of the match so that it can back up and
5612 try a different alternative if the first one fails. As matching pro-
5613 ceeds deeper and deeper into the tree of possibilities, the recursion
5614 depth increases.
5615
5616 Not all calls of match() increase the recursion depth; for an item such
5617 as a* it may be called several times at the same level, after matching
5618 different numbers of a's. Furthermore, in a number of cases where the
5619 result of the recursive call would immediately be passed back as the
5620 result of the current call (a "tail recursion"), the function is just
5621 restarted instead.
5622
5623 The pcre_dfa_exec() function operates in an entirely different way, and
5624 hardly uses recursion at all. The limit on its complexity is the amount
5625 of workspace it is given. The comments that follow do NOT apply to
5626 pcre_dfa_exec(); they are relevant only for pcre_exec().
5627
5628 You can set limits on the number of times that match() is called, both
5629 in total and recursively. If the limit is exceeded, an error occurs.
5630 For details, see the section on extra data for pcre_exec() in the
5631 pcreapi documentation.
5632
5633 Each time that match() is actually called recursively, it uses memory
5634 from the process stack. For certain kinds of pattern and data, very
5635 large amounts of stack may be needed, despite the recognition of "tail
5636 recursion". You can often reduce the amount of recursion, and there-
5637 fore the amount of stack used, by modifying the pattern that is being
5638 matched. Consider, for example, this pattern:
5639
5640 ([^<]|<(?!inet))+
5641
5642 It matches from wherever it starts until it encounters "<inet" or the
5643 end of the data, and is the kind of pattern that might be used when
5644 processing an XML file. Each iteration of the outer parentheses matches
5645 either one character that is not "<" or a "<" that is not followed by
5646 "inet". However, each time a parenthesis is processed, a recursion
5647 occurs, so this formulation uses a stack frame for each matched charac-
5648 ter. For a long string, a lot of stack is required. Consider now this
5649 rewritten pattern, which matches exactly the same strings:
5650
5651 ([^<]++|<(?!inet))+
5652
5653 This uses very much less stack, because runs of characters that do not
5654 contain "<" are "swallowed" in one item inside the parentheses. Recur-
5655 sion happens only when a "<" character that is not followed by "inet"
5656 is encountered (and we assume this is relatively rare). A possessive
5657 quantifier is used to stop any backtracking into the runs of non-"<"
5658 characters, but that is not related to stack usage.
5659
5660 This example shows that one way of avoiding stack problems when match-
5661 ing long subject strings is to write repeated parenthesized subpatterns
5662 to match more than one character whenever possible.
5663
5664 In environments where stack memory is constrained, you might want to
5665 compile PCRE to use heap memory instead of stack for remembering back-
5666 up points. This makes it run a lot more slowly, however. Details of how
5667 to do this are given in the pcrebuild documentation.
5668
5669 In Unix-like environments, there is not often a problem with the stack
5670 unless very long strings are involved, though the default limit on
5671 stack size varies from system to system. Values from 8Mb to 64Mb are
5672 common. You can find your default limit by running the command:
5673
5674 ulimit -s
5675
5676 Unfortunately, the effect of running out of stack is often SIGSEGV,
5677 though sometimes a more explicit error message is given. You can nor-
5678 mally increase the limit on stack size by code such as this:
5679
5680 struct rlimit rlim;
5681 getrlimit(RLIMIT_STACK, &rlim);
5682 rlim.rlim_cur = 100*1024*1024;
5683 setrlimit(RLIMIT_STACK, &rlim);
5684
5685 This reads the current limits (soft and hard) using getrlimit(), then
5686 attempts to increase the soft limit to 100Mb using setrlimit(). You
5687 must do this before calling pcre_exec().
5688
5689 PCRE has an internal counter that can be used to limit the depth of
5690 recursion, and thus cause pcre_exec() to give an error code before it
5691 runs out of stack. By default, the limit is very large, and unlikely
5692 ever to operate. It can be changed when PCRE is built, and it can also
5693 be set when pcre_exec() is called. For details of these interfaces, see
5694 the pcrebuild and pcreapi documentation.
5695
5696 As a very rough rule of thumb, you should reckon on about 500 bytes per
5697 recursion. Thus, if you want to limit your stack usage to 8Mb, you
5698 should set the limit at 16000 recursions. A 64Mb stack, on the other
5699 hand, can support around 128000 recursions. The pcretest test program
5700 has a command line option (-S) that can be used to increase the size of
5701 its stack.
5702
5703
5704 AUTHOR
5705
5706 Philip Hazel
5707 University Computing Service
5708 Cambridge CB2 3QH, England.
5709
5710
5711 REVISION
5712
5713 Last updated: 12 March 2007
5714 Copyright (c) 1997-2007 University of Cambridge.
5715 ------------------------------------------------------------------------------
5716
5717

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5