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

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

Parent Directory Parent Directory | Revision Log Revision Log


Revision 185 - (show annotations)
Tue Jun 19 13:39:46 2007 UTC (12 years, 3 months ago) by ph10
File MIME type: text/plain
File size: 275544 byte(s)
Documentation final tidies for 7.2 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. However, the Perl 5.10 horizontal and vertical whitespace matching
201 escapes (\h, \H, \v, and \V) do match all the appropriate Unicode char-
202 acters.
203
204 10. Case-insensitive matching applies only to characters whose values
205 are less than 128, unless PCRE is built with Unicode property support.
206 Even when Unicode property support is available, PCRE still uses its
207 own character tables when checking the case of low-valued characters,
208 so as not to degrade performance. The Unicode property information is
209 used only for characters with higher values. Even when Unicode property
210 support is available, PCRE supports case-insensitive matching only when
211 there is a one-to-one mapping between a letter's cases. There are a
212 small number of many-to-one mappings in Unicode; these are not sup-
213 ported by PCRE.
214
215
216 AUTHOR
217
218 Philip Hazel
219 University Computing Service
220 Cambridge CB2 3QH, England.
221
222 Putting an actual email address here seems to have been a spam magnet,
223 so I've taken it away. If you want to email me, use my two initials,
224 followed by the two digits 10, at the domain cam.ac.uk.
225
226
227 REVISION
228
229 Last updated: 13 June 2007
230 Copyright (c) 1997-2007 University of Cambridge.
231 ------------------------------------------------------------------------------
232
233
234 PCREBUILD(3) PCREBUILD(3)
235
236
237 NAME
238 PCRE - Perl-compatible regular expressions
239
240
241 PCRE BUILD-TIME OPTIONS
242
243 This document describes the optional features of PCRE that can be
244 selected when the library is compiled. They are all selected, or dese-
245 lected, by providing options to the configure script that is run before
246 the make command. The complete list of options for configure (which
247 includes the standard ones such as the selection of the installation
248 directory) can be obtained by running
249
250 ./configure --help
251
252 The following sections include descriptions of options whose names
253 begin with --enable or --disable. These settings specify changes to the
254 defaults for the configure command. Because of the way that configure
255 works, --enable and --disable always come in pairs, so the complemen-
256 tary option always exists as well, but as it specifies the default, it
257 is not described.
258
259
260 C++ SUPPORT
261
262 By default, the configure script will search for a C++ compiler and C++
263 header files. If it finds them, it automatically builds the C++ wrapper
264 library for PCRE. You can disable this by adding
265
266 --disable-cpp
267
268 to the configure command.
269
270
271 UTF-8 SUPPORT
272
273 To build PCRE with support for UTF-8 character strings, add
274
275 --enable-utf8
276
277 to the configure command. Of itself, this does not make PCRE treat
278 strings as UTF-8. As well as compiling PCRE with this option, you also
279 have have to set the PCRE_UTF8 option when you call the pcre_compile()
280 function.
281
282
283 UNICODE CHARACTER PROPERTY SUPPORT
284
285 UTF-8 support allows PCRE to process character values greater than 255
286 in the strings that it handles. On its own, however, it does not pro-
287 vide any facilities for accessing the properties of such characters. If
288 you want to be able to use the pattern escapes \P, \p, and \X, which
289 refer to Unicode character properties, you must add
290
291 --enable-unicode-properties
292
293 to the configure command. This implies UTF-8 support, even if you have
294 not explicitly requested it.
295
296 Including Unicode property support adds around 30K of tables to the
297 PCRE library. Only the general category properties such as Lu and Nd
298 are supported. Details are given in the pcrepattern documentation.
299
300
301 CODE VALUE OF NEWLINE
302
303 By default, PCRE interprets character 10 (linefeed, LF) as indicating
304 the end of a line. This is the normal newline character on Unix-like
305 systems. You can compile PCRE to use character 13 (carriage return, CR)
306 instead, by adding
307
308 --enable-newline-is-cr
309
310 to the configure command. There is also a --enable-newline-is-lf
311 option, which explicitly specifies linefeed as the newline character.
312
313 Alternatively, you can specify that line endings are to be indicated by
314 the two character sequence CRLF. If you want this, add
315
316 --enable-newline-is-crlf
317
318 to the configure command. There is a fourth option, specified by
319
320 --enable-newline-is-anycrlf
321
322 which causes PCRE to recognize any of the three sequences CR, LF, or
323 CRLF as indicating a line ending. Finally, a fifth option, specified by
324
325 --enable-newline-is-any
326
327 causes PCRE to recognize any Unicode newline sequence.
328
329 Whatever line ending convention is selected when PCRE is built can be
330 overridden when the library functions are called. At build time it is
331 conventional to use the standard for your operating system.
332
333
334 BUILDING SHARED AND STATIC LIBRARIES
335
336 The PCRE building process uses libtool to build both shared and static
337 Unix libraries by default. You can suppress one of these by adding one
338 of
339
340 --disable-shared
341 --disable-static
342
343 to the configure command, as required.
344
345
346 POSIX MALLOC USAGE
347
348 When PCRE is called through the POSIX interface (see the pcreposix doc-
349 umentation), additional working storage is required for holding the
350 pointers to capturing substrings, because PCRE requires three integers
351 per substring, whereas the POSIX interface provides only two. If the
352 number of expected substrings is small, the wrapper function uses space
353 on the stack, because this is faster than using malloc() for each call.
354 The default threshold above which the stack is no longer used is 10; it
355 can be changed by adding a setting such as
356
357 --with-posix-malloc-threshold=20
358
359 to the configure command.
360
361
362 HANDLING VERY LARGE PATTERNS
363
364 Within a compiled pattern, offset values are used to point from one
365 part to another (for example, from an opening parenthesis to an alter-
366 nation metacharacter). By default, two-byte values are used for these
367 offsets, leading to a maximum size for a compiled pattern of around
368 64K. This is sufficient to handle all but the most gigantic patterns.
369 Nevertheless, some people do want to process enormous patterns, so it
370 is possible to compile PCRE to use three-byte or four-byte offsets by
371 adding a setting such as
372
373 --with-link-size=3
374
375 to the configure command. The value given must be 2, 3, or 4. Using
376 longer offsets slows down the operation of PCRE because it has to load
377 additional bytes when handling them.
378
379
380 AVOIDING EXCESSIVE STACK USAGE
381
382 When matching with the pcre_exec() function, PCRE implements backtrack-
383 ing by making recursive calls to an internal function called match().
384 In environments where the size of the stack is limited, this can se-
385 verely limit PCRE's operation. (The Unix environment does not usually
386 suffer from this problem, but it may sometimes be necessary to increase
387 the maximum stack size. There is a discussion in the pcrestack docu-
388 mentation.) An alternative approach to recursion that uses memory from
389 the heap to remember data, instead of using recursive function calls,
390 has been implemented to work round the problem of limited stack size.
391 If you want to build a version of PCRE that works this way, add
392
393 --disable-stack-for-recursion
394
395 to the configure command. With this configuration, PCRE will use the
396 pcre_stack_malloc and pcre_stack_free variables to call memory manage-
397 ment functions. By default these point to malloc() and free(), but you
398 can replace the pointers so that your own functions are used.
399
400 Separate functions are provided rather than using pcre_malloc and
401 pcre_free because the usage is very predictable: the block sizes
402 requested are always the same, and the blocks are always freed in
403 reverse order. A calling program might be able to implement optimized
404 functions that perform better than malloc() and free(). PCRE runs
405 noticeably more slowly when built in this way. This option affects only
406 the pcre_exec() function; it is not relevant for the the
407 pcre_dfa_exec() function.
408
409
410 LIMITING PCRE RESOURCE USAGE
411
412 Internally, PCRE has a function called match(), which it calls repeat-
413 edly (sometimes recursively) when matching a pattern with the
414 pcre_exec() function. By controlling the maximum number of times this
415 function may be called during a single matching operation, a limit can
416 be placed on the resources used by a single call to pcre_exec(). The
417 limit can be changed at run time, as described in the pcreapi documen-
418 tation. The default is 10 million, but this can be changed by adding a
419 setting such as
420
421 --with-match-limit=500000
422
423 to the configure command. This setting has no effect on the
424 pcre_dfa_exec() matching function.
425
426 In some environments it is desirable to limit the depth of recursive
427 calls of match() more strictly than the total number of calls, in order
428 to restrict the maximum amount of stack (or heap, if --disable-stack-
429 for-recursion is specified) that is used. A second limit controls this;
430 it defaults to the value that is set for --with-match-limit, which
431 imposes no additional constraints. However, you can set a lower limit
432 by adding, for example,
433
434 --with-match-limit-recursion=10000
435
436 to the configure command. This value can also be overridden at run
437 time.
438
439
440 CREATING CHARACTER TABLES AT BUILD TIME
441
442 PCRE uses fixed tables for processing characters whose code values are
443 less than 256. By default, PCRE is built with a set of tables that are
444 distributed in the file pcre_chartables.c.dist. These tables are for
445 ASCII codes only. If you add
446
447 --enable-rebuild-chartables
448
449 to the configure command, the distributed tables are no longer used.
450 Instead, a program called dftables is compiled and run. This outputs
451 the source for new set of tables, created in the default locale of your
452 C runtime system. (This method of replacing the tables does not work if
453 you are cross compiling, because dftables is run on the local host. If
454 you need to create alternative tables when cross compiling, you will
455 have to do so "by hand".)
456
457
458 USING EBCDIC CODE
459
460 PCRE assumes by default that it will run in an environment where the
461 character code is ASCII (or Unicode, which is a superset of ASCII).
462 PCRE can, however, be compiled to run in an EBCDIC environment by
463 adding
464
465 --enable-ebcdic
466
467 to the configure command. This setting implies --enable-rebuild-charta-
468 bles.
469
470
471 SEE ALSO
472
473 pcreapi(3), pcre_config(3).
474
475
476 AUTHOR
477
478 Philip Hazel
479 University Computing Service
480 Cambridge CB2 3QH, England.
481
482
483 REVISION
484
485 Last updated: 05 June 2007
486 Copyright (c) 1997-2007 University of Cambridge.
487 ------------------------------------------------------------------------------
488
489
490 PCREMATCHING(3) PCREMATCHING(3)
491
492
493 NAME
494 PCRE - Perl-compatible regular expressions
495
496
497 PCRE MATCHING ALGORITHMS
498
499 This document describes the two different algorithms that are available
500 in PCRE for matching a compiled regular expression against a given sub-
501 ject string. The "standard" algorithm is the one provided by the
502 pcre_exec() function. This works in the same was as Perl's matching
503 function, and provides a Perl-compatible matching operation.
504
505 An alternative algorithm is provided by the pcre_dfa_exec() function;
506 this operates in a different way, and is not Perl-compatible. It has
507 advantages and disadvantages compared with the standard algorithm, and
508 these are described below.
509
510 When there is only one possible way in which a given subject string can
511 match a pattern, the two algorithms give the same answer. A difference
512 arises, however, when there are multiple possibilities. For example, if
513 the pattern
514
515 ^<.*>
516
517 is matched against the string
518
519 <something> <something else> <something further>
520
521 there are three possible answers. The standard algorithm finds only one
522 of them, whereas the alternative algorithm finds all three.
523
524
525 REGULAR EXPRESSIONS AS TREES
526
527 The set of strings that are matched by a regular expression can be rep-
528 resented as a tree structure. An unlimited repetition in the pattern
529 makes the tree of infinite size, but it is still a tree. Matching the
530 pattern to a given subject string (from a given starting point) can be
531 thought of as a search of the tree. There are two ways to search a
532 tree: depth-first and breadth-first, and these correspond to the two
533 matching algorithms provided by PCRE.
534
535
536 THE STANDARD MATCHING ALGORITHM
537
538 In the terminology of Jeffrey Friedl's book "Mastering Regular Expres-
539 sions", the standard algorithm is an "NFA algorithm". It conducts a
540 depth-first search of the pattern tree. That is, it proceeds along a
541 single path through the tree, checking that the subject matches what is
542 required. When there is a mismatch, the algorithm tries any alterna-
543 tives at the current point, and if they all fail, it backs up to the
544 previous branch point in the tree, and tries the next alternative
545 branch at that level. This often involves backing up (moving to the
546 left) in the subject string as well. The order in which repetition
547 branches are tried is controlled by the greedy or ungreedy nature of
548 the quantifier.
549
550 If a leaf node is reached, a matching string has been found, and at
551 that point the algorithm stops. Thus, if there is more than one possi-
552 ble match, this algorithm returns the first one that it finds. Whether
553 this is the shortest, the longest, or some intermediate length depends
554 on the way the greedy and ungreedy repetition quantifiers are specified
555 in the pattern.
556
557 Because it ends up with a single path through the tree, it is rela-
558 tively straightforward for this algorithm to keep track of the sub-
559 strings that are matched by portions of the pattern in parentheses.
560 This provides support for capturing parentheses and back references.
561
562
563 THE ALTERNATIVE MATCHING ALGORITHM
564
565 This algorithm conducts a breadth-first search of the tree. Starting
566 from the first matching point in the subject, it scans the subject
567 string from left to right, once, character by character, and as it does
568 this, it remembers all the paths through the tree that represent valid
569 matches. In Friedl's terminology, this is a kind of "DFA algorithm",
570 though it is not implemented as a traditional finite state machine (it
571 keeps multiple states active simultaneously).
572
573 The scan continues until either the end of the subject is reached, or
574 there are no more unterminated paths. At this point, terminated paths
575 represent the different matching possibilities (if there are none, the
576 match has failed). Thus, if there is more than one possible match,
577 this algorithm finds all of them, and in particular, it finds the long-
578 est. In PCRE, there is an option to stop the algorithm after the first
579 match (which is necessarily the shortest) has been found.
580
581 Note that all the matches that are found start at the same point in the
582 subject. If the pattern
583
584 cat(er(pillar)?)
585
586 is matched against the string "the caterpillar catchment", the result
587 will be the three strings "cat", "cater", and "caterpillar" that start
588 at the fourth character of the subject. The algorithm does not automat-
589 ically move on to find matches that start at later positions.
590
591 There are a number of features of PCRE regular expressions that are not
592 supported by the alternative matching algorithm. They are as follows:
593
594 1. Because the algorithm finds all possible matches, the greedy or
595 ungreedy nature of repetition quantifiers is not relevant. Greedy and
596 ungreedy quantifiers are treated in exactly the same way. However, pos-
597 sessive quantifiers can make a difference when what follows could also
598 match what is quantified, for example in a pattern like this:
599
600 ^a++\w!
601
602 This pattern matches "aaab!" but not "aaa!", which would be matched by
603 a non-possessive quantifier. Similarly, if an atomic group is present,
604 it is matched as if it were a standalone pattern at the current point,
605 and the longest match is then "locked in" for the rest of the overall
606 pattern.
607
608 2. When dealing with multiple paths through the tree simultaneously, it
609 is not straightforward to keep track of captured substrings for the
610 different matching possibilities, and PCRE's implementation of this
611 algorithm does not attempt to do this. This means that no captured sub-
612 strings are available.
613
614 3. Because no substrings are captured, back references within the pat-
615 tern are not supported, and cause errors if encountered.
616
617 4. For the same reason, conditional expressions that use a backrefer-
618 ence as the condition or test for a specific group recursion are not
619 supported.
620
621 5. Because many paths through the tree may be active, the \K escape
622 sequence, which resets the start of the match when encountered (but may
623 be on some paths and not on others), is not supported. It causes an
624 error if encountered.
625
626 6. Callouts are supported, but the value of the capture_top field is
627 always 1, and the value of the capture_last field is always -1.
628
629 7. The \C escape sequence, which (in the standard algorithm) matches a
630 single byte, even in UTF-8 mode, is not supported because the alterna-
631 tive algorithm moves through the subject string one character at a
632 time, for all active paths through the tree.
633
634
635 ADVANTAGES OF THE ALTERNATIVE ALGORITHM
636
637 Using the alternative matching algorithm provides the following advan-
638 tages:
639
640 1. All possible matches (at a single point in the subject) are automat-
641 ically found, and in particular, the longest match is found. To find
642 more than one match using the standard algorithm, you have to do kludgy
643 things with callouts.
644
645 2. There is much better support for partial matching. The restrictions
646 on the content of the pattern that apply when using the standard algo-
647 rithm for partial matching do not apply to the alternative algorithm.
648 For non-anchored patterns, the starting position of a partial match is
649 available.
650
651 3. Because the alternative algorithm scans the subject string just
652 once, and never needs to backtrack, it is possible to pass very long
653 subject strings to the matching function in several pieces, checking
654 for partial matching each time.
655
656
657 DISADVANTAGES OF THE ALTERNATIVE ALGORITHM
658
659 The alternative algorithm suffers from a number of disadvantages:
660
661 1. It is substantially slower than the standard algorithm. This is
662 partly because it has to search for all possible matches, but is also
663 because it is less susceptible to optimization.
664
665 2. Capturing parentheses and back references are not supported.
666
667 3. Although atomic groups are supported, their use does not provide the
668 performance advantage that it does for the standard algorithm.
669
670
671 AUTHOR
672
673 Philip Hazel
674 University Computing Service
675 Cambridge CB2 3QH, England.
676
677
678 REVISION
679
680 Last updated: 29 May 2007
681 Copyright (c) 1997-2007 University of Cambridge.
682 ------------------------------------------------------------------------------
683
684
685 PCREAPI(3) PCREAPI(3)
686
687
688 NAME
689 PCRE - Perl-compatible regular expressions
690
691
692 PCRE NATIVE API
693
694 #include <pcre.h>
695
696 pcre *pcre_compile(const char *pattern, int options,
697 const char **errptr, int *erroffset,
698 const unsigned char *tableptr);
699
700 pcre *pcre_compile2(const char *pattern, int options,
701 int *errorcodeptr,
702 const char **errptr, int *erroffset,
703 const unsigned char *tableptr);
704
705 pcre_extra *pcre_study(const pcre *code, int options,
706 const char **errptr);
707
708 int pcre_exec(const pcre *code, const pcre_extra *extra,
709 const char *subject, int length, int startoffset,
710 int options, int *ovector, int ovecsize);
711
712 int pcre_dfa_exec(const pcre *code, const pcre_extra *extra,
713 const char *subject, int length, int startoffset,
714 int options, int *ovector, int ovecsize,
715 int *workspace, int wscount);
716
717 int pcre_copy_named_substring(const pcre *code,
718 const char *subject, int *ovector,
719 int stringcount, const char *stringname,
720 char *buffer, int buffersize);
721
722 int pcre_copy_substring(const char *subject, int *ovector,
723 int stringcount, int stringnumber, char *buffer,
724 int buffersize);
725
726 int pcre_get_named_substring(const pcre *code,
727 const char *subject, int *ovector,
728 int stringcount, const char *stringname,
729 const char **stringptr);
730
731 int pcre_get_stringnumber(const pcre *code,
732 const char *name);
733
734 int pcre_get_stringtable_entries(const pcre *code,
735 const char *name, char **first, char **last);
736
737 int pcre_get_substring(const char *subject, int *ovector,
738 int stringcount, int stringnumber,
739 const char **stringptr);
740
741 int pcre_get_substring_list(const char *subject,
742 int *ovector, int stringcount, const char ***listptr);
743
744 void pcre_free_substring(const char *stringptr);
745
746 void pcre_free_substring_list(const char **stringptr);
747
748 const unsigned char *pcre_maketables(void);
749
750 int pcre_fullinfo(const pcre *code, const pcre_extra *extra,
751 int what, void *where);
752
753 int pcre_info(const pcre *code, int *optptr, int *firstcharptr);
754
755 int pcre_refcount(pcre *code, int adjust);
756
757 int pcre_config(int what, void *where);
758
759 char *pcre_version(void);
760
761 void *(*pcre_malloc)(size_t);
762
763 void (*pcre_free)(void *);
764
765 void *(*pcre_stack_malloc)(size_t);
766
767 void (*pcre_stack_free)(void *);
768
769 int (*pcre_callout)(pcre_callout_block *);
770
771
772 PCRE API OVERVIEW
773
774 PCRE has its own native API, which is described in this document. There
775 are also some wrapper functions that correspond to the POSIX regular
776 expression API. These are described in the pcreposix documentation.
777 Both of these APIs define a set of C function calls. A C++ wrapper is
778 distributed with PCRE. It is documented in the pcrecpp page.
779
780 The native API C function prototypes are defined in the header file
781 pcre.h, and on Unix systems the library itself is called libpcre. It
782 can normally be accessed by adding -lpcre to the command for linking an
783 application that uses PCRE. The header file defines the macros
784 PCRE_MAJOR and PCRE_MINOR to contain the major and minor release num-
785 bers for the library. Applications can use these to include support
786 for different releases of PCRE.
787
788 The functions pcre_compile(), pcre_compile2(), pcre_study(), and
789 pcre_exec() are used for compiling and matching regular expressions in
790 a Perl-compatible manner. A sample program that demonstrates the sim-
791 plest way of using them is provided in the file called pcredemo.c in
792 the source distribution. The pcresample documentation describes how to
793 run it.
794
795 A second matching function, pcre_dfa_exec(), which is not Perl-compati-
796 ble, is also provided. This uses a different algorithm for the match-
797 ing. The alternative algorithm finds all possible matches (at a given
798 point in the subject), and scans the subject just once. However, this
799 algorithm does not return captured substrings. A description of the two
800 matching algorithms and their advantages and disadvantages is given in
801 the pcrematching documentation.
802
803 In addition to the main compiling and matching functions, there are
804 convenience functions for extracting captured substrings from a subject
805 string that is matched by pcre_exec(). They are:
806
807 pcre_copy_substring()
808 pcre_copy_named_substring()
809 pcre_get_substring()
810 pcre_get_named_substring()
811 pcre_get_substring_list()
812 pcre_get_stringnumber()
813 pcre_get_stringtable_entries()
814
815 pcre_free_substring() and pcre_free_substring_list() are also provided,
816 to free the memory used for extracted strings.
817
818 The function pcre_maketables() is used to build a set of character
819 tables in the current locale for passing to pcre_compile(),
820 pcre_exec(), or pcre_dfa_exec(). This is an optional facility that is
821 provided for specialist use. Most commonly, no special tables are
822 passed, in which case internal tables that are generated when PCRE is
823 built are used.
824
825 The function pcre_fullinfo() is used to find out information about a
826 compiled pattern; pcre_info() is an obsolete version that returns only
827 some of the available information, but is retained for backwards com-
828 patibility. The function pcre_version() returns a pointer to a string
829 containing the version of PCRE and its date of release.
830
831 The function pcre_refcount() maintains a reference count in a data
832 block containing a compiled pattern. This is provided for the benefit
833 of object-oriented applications.
834
835 The global variables pcre_malloc and pcre_free initially contain the
836 entry points of the standard malloc() and free() functions, respec-
837 tively. PCRE calls the memory management functions via these variables,
838 so a calling program can replace them if it wishes to intercept the
839 calls. This should be done before calling any PCRE functions.
840
841 The global variables pcre_stack_malloc and pcre_stack_free are also
842 indirections to memory management functions. These special functions
843 are used only when PCRE is compiled to use the heap for remembering
844 data, instead of recursive function calls, when running the pcre_exec()
845 function. See the pcrebuild documentation for details of how to do
846 this. It is a non-standard way of building PCRE, for use in environ-
847 ments that have limited stacks. Because of the greater use of memory
848 management, it runs more slowly. Separate functions are provided so
849 that special-purpose external code can be used for this case. When
850 used, these functions are always called in a stack-like manner (last
851 obtained, first freed), and always for memory blocks of the same size.
852 There is a discussion about PCRE's stack usage in the pcrestack docu-
853 mentation.
854
855 The global variable pcre_callout initially contains NULL. It can be set
856 by the caller to a "callout" function, which PCRE will then call at
857 specified points during a matching operation. Details are given in the
858 pcrecallout documentation.
859
860
861 NEWLINES
862
863 PCRE supports five different conventions for indicating line breaks in
864 strings: a single CR (carriage return) character, a single LF (line-
865 feed) character, the two-character sequence CRLF, any of the three pre-
866 ceding, or any Unicode newline sequence. The Unicode newline sequences
867 are the three just mentioned, plus the single characters VT (vertical
868 tab, U+000B), FF (formfeed, U+000C), NEL (next line, U+0085), LS (line
869 separator, U+2028), and PS (paragraph separator, U+2029).
870
871 Each of the first three conventions is used by at least one operating
872 system as its standard newline sequence. When PCRE is built, a default
873 can be specified. The default default is LF, which is the Unix stan-
874 dard. When PCRE is run, the default can be overridden, either when a
875 pattern is compiled, or when it is matched.
876
877 In the PCRE documentation the word "newline" is used to mean "the char-
878 acter or pair of characters that indicate a line break". The choice of
879 newline convention affects the handling of the dot, circumflex, and
880 dollar metacharacters, the handling of #-comments in /x mode, and, when
881 CRLF is a recognized line ending sequence, the match position advance-
882 ment for a non-anchored pattern. The choice of newline convention does
883 not affect the interpretation of the \n or \r escape sequences.
884
885
886 MULTITHREADING
887
888 The PCRE functions can be used in multi-threading applications, with
889 the proviso that the memory management functions pointed to by
890 pcre_malloc, pcre_free, pcre_stack_malloc, and pcre_stack_free, and the
891 callout function pointed to by pcre_callout, are shared by all threads.
892
893 The compiled form of a regular expression is not altered during match-
894 ing, so the same compiled pattern can safely be used by several threads
895 at once.
896
897
898 SAVING PRECOMPILED PATTERNS FOR LATER USE
899
900 The compiled form of a regular expression can be saved and re-used at a
901 later time, possibly by a different program, and even on a host other
902 than the one on which it was compiled. Details are given in the
903 pcreprecompile documentation. However, compiling a regular expression
904 with one version of PCRE for use with a different version is not guar-
905 anteed to work and may cause crashes.
906
907
908 CHECKING BUILD-TIME OPTIONS
909
910 int pcre_config(int what, void *where);
911
912 The function pcre_config() makes it possible for a PCRE client to dis-
913 cover which optional features have been compiled into the PCRE library.
914 The pcrebuild documentation has more details about these optional fea-
915 tures.
916
917 The first argument for pcre_config() is an integer, specifying which
918 information is required; the second argument is a pointer to a variable
919 into which the information is placed. The following information is
920 available:
921
922 PCRE_CONFIG_UTF8
923
924 The output is an integer that is set to one if UTF-8 support is avail-
925 able; otherwise it is set to zero.
926
927 PCRE_CONFIG_UNICODE_PROPERTIES
928
929 The output is an integer that is set to one if support for Unicode
930 character properties is available; otherwise it is set to zero.
931
932 PCRE_CONFIG_NEWLINE
933
934 The output is an integer whose value specifies the default character
935 sequence that is recognized as meaning "newline". The four values that
936 are supported are: 10 for LF, 13 for CR, 3338 for CRLF, -2 for ANYCRLF,
937 and -1 for ANY. The default should normally be the standard sequence
938 for your operating system.
939
940 PCRE_CONFIG_LINK_SIZE
941
942 The output is an integer that contains the number of bytes used for
943 internal linkage in compiled regular expressions. The value is 2, 3, or
944 4. Larger values allow larger regular expressions to be compiled, at
945 the expense of slower matching. The default value of 2 is sufficient
946 for all but the most massive patterns, since it allows the compiled
947 pattern to be up to 64K in size.
948
949 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
950
951 The output is an integer that contains the threshold above which the
952 POSIX interface uses malloc() for output vectors. Further details are
953 given in the pcreposix documentation.
954
955 PCRE_CONFIG_MATCH_LIMIT
956
957 The output is an integer that gives the default limit for the number of
958 internal matching function calls in a pcre_exec() execution. Further
959 details are given with pcre_exec() below.
960
961 PCRE_CONFIG_MATCH_LIMIT_RECURSION
962
963 The output is an integer that gives the default limit for the depth of
964 recursion when calling the internal matching function in a pcre_exec()
965 execution. Further details are given with pcre_exec() below.
966
967 PCRE_CONFIG_STACKRECURSE
968
969 The output is an integer that is set to one if internal recursion when
970 running pcre_exec() is implemented by recursive function calls that use
971 the stack to remember their state. This is the usual way that PCRE is
972 compiled. The output is zero if PCRE was compiled to use blocks of data
973 on the heap instead of recursive function calls. In this case,
974 pcre_stack_malloc and pcre_stack_free are called to manage memory
975 blocks on the heap, thus avoiding the use of the stack.
976
977
978 COMPILING A PATTERN
979
980 pcre *pcre_compile(const char *pattern, int options,
981 const char **errptr, int *erroffset,
982 const unsigned char *tableptr);
983
984 pcre *pcre_compile2(const char *pattern, int options,
985 int *errorcodeptr,
986 const char **errptr, int *erroffset,
987 const unsigned char *tableptr);
988
989 Either of the functions pcre_compile() or pcre_compile2() can be called
990 to compile a pattern into an internal form. The only difference between
991 the two interfaces is that pcre_compile2() has an additional argument,
992 errorcodeptr, via which a numerical error code can be returned.
993
994 The pattern is a C string terminated by a binary zero, and is passed in
995 the pattern argument. A pointer to a single block of memory that is
996 obtained via pcre_malloc is returned. This contains the compiled code
997 and related data. The pcre type is defined for the returned block; this
998 is a typedef for a structure whose contents are not externally defined.
999 It is up to the caller to free the memory (via pcre_free) when it is no
1000 longer required.
1001
1002 Although the compiled code of a PCRE regex is relocatable, that is, it
1003 does not depend on memory location, the complete pcre data block is not
1004 fully relocatable, because it may contain a copy of the tableptr argu-
1005 ment, which is an address (see below).
1006
1007 The options argument contains various bit settings that affect the com-
1008 pilation. It should be zero if no options are required. The available
1009 options are described below. Some of them, in particular, those that
1010 are compatible with Perl, can also be set and unset from within the
1011 pattern (see the detailed description in the pcrepattern documenta-
1012 tion). For these options, the contents of the options argument speci-
1013 fies their initial settings at the start of compilation and execution.
1014 The PCRE_ANCHORED and PCRE_NEWLINE_xxx options can be set at the time
1015 of matching as well as at compile time.
1016
1017 If errptr is NULL, pcre_compile() returns NULL immediately. Otherwise,
1018 if compilation of a pattern fails, pcre_compile() returns NULL, and
1019 sets the variable pointed to by errptr to point to a textual error mes-
1020 sage. This is a static string that is part of the library. You must not
1021 try to free it. The offset from the start of the pattern to the charac-
1022 ter where the error was discovered is placed in the variable pointed to
1023 by erroffset, which must not be NULL. If it is, an immediate error is
1024 given.
1025
1026 If pcre_compile2() is used instead of pcre_compile(), and the error-
1027 codeptr argument is not NULL, a non-zero error code number is returned
1028 via this argument in the event of an error. This is in addition to the
1029 textual error message. Error codes and messages are listed below.
1030
1031 If the final argument, tableptr, is NULL, PCRE uses a default set of
1032 character tables that are built when PCRE is compiled, using the
1033 default C locale. Otherwise, tableptr must be an address that is the
1034 result of a call to pcre_maketables(). This value is stored with the
1035 compiled pattern, and used again by pcre_exec(), unless another table
1036 pointer is passed to it. For more discussion, see the section on locale
1037 support below.
1038
1039 This code fragment shows a typical straightforward call to pcre_com-
1040 pile():
1041
1042 pcre *re;
1043 const char *error;
1044 int erroffset;
1045 re = pcre_compile(
1046 "^A.*Z", /* the pattern */
1047 0, /* default options */
1048 &error, /* for error message */
1049 &erroffset, /* for error offset */
1050 NULL); /* use default character tables */
1051
1052 The following names for option bits are defined in the pcre.h header
1053 file:
1054
1055 PCRE_ANCHORED
1056
1057 If this bit is set, the pattern is forced to be "anchored", that is, it
1058 is constrained to match only at the first matching point in the string
1059 that is being searched (the "subject string"). This effect can also be
1060 achieved by appropriate constructs in the pattern itself, which is the
1061 only way to do it in Perl.
1062
1063 PCRE_AUTO_CALLOUT
1064
1065 If this bit is set, pcre_compile() automatically inserts callout items,
1066 all with number 255, before each pattern item. For discussion of the
1067 callout facility, see the pcrecallout documentation.
1068
1069 PCRE_CASELESS
1070
1071 If this bit is set, letters in the pattern match both upper and lower
1072 case letters. It is equivalent to Perl's /i option, and it can be
1073 changed within a pattern by a (?i) option setting. In UTF-8 mode, PCRE
1074 always understands the concept of case for characters whose values are
1075 less than 128, so caseless matching is always possible. For characters
1076 with higher values, the concept of case is supported if PCRE is com-
1077 piled with Unicode property support, but not otherwise. If you want to
1078 use caseless matching for characters 128 and above, you must ensure
1079 that PCRE is compiled with Unicode property support as well as with
1080 UTF-8 support.
1081
1082 PCRE_DOLLAR_ENDONLY
1083
1084 If this bit is set, a dollar metacharacter in the pattern matches only
1085 at the end of the subject string. Without this option, a dollar also
1086 matches immediately before a newline at the end of the string (but not
1087 before any other newlines). The PCRE_DOLLAR_ENDONLY option is ignored
1088 if PCRE_MULTILINE is set. There is no equivalent to this option in
1089 Perl, and no way to set it within a pattern.
1090
1091 PCRE_DOTALL
1092
1093 If this bit is set, a dot metacharater in the pattern matches all char-
1094 acters, including those that indicate newline. Without it, a dot does
1095 not match when the current position is at a newline. This option is
1096 equivalent to Perl's /s option, and it can be changed within a pattern
1097 by a (?s) option setting. A negative class such as [^a] always matches
1098 newline characters, independent of the setting of this option.
1099
1100 PCRE_DUPNAMES
1101
1102 If this bit is set, names used to identify capturing subpatterns need
1103 not be unique. This can be helpful for certain types of pattern when it
1104 is known that only one instance of the named subpattern can ever be
1105 matched. There are more details of named subpatterns below; see also
1106 the pcrepattern documentation.
1107
1108 PCRE_EXTENDED
1109
1110 If this bit is set, whitespace data characters in the pattern are
1111 totally ignored except when escaped or inside a character class. White-
1112 space does not include the VT character (code 11). In addition, charac-
1113 ters between an unescaped # outside a character class and the next new-
1114 line, inclusive, are also ignored. This is equivalent to Perl's /x
1115 option, and it can be changed within a pattern by a (?x) option set-
1116 ting.
1117
1118 This option makes it possible to include comments inside complicated
1119 patterns. Note, however, that this applies only to data characters.
1120 Whitespace characters may never appear within special character
1121 sequences in a pattern, for example within the sequence (?( which
1122 introduces a conditional subpattern.
1123
1124 PCRE_EXTRA
1125
1126 This option was invented in order to turn on additional functionality
1127 of PCRE that is incompatible with Perl, but it is currently of very
1128 little use. When set, any backslash in a pattern that is followed by a
1129 letter that has no special meaning causes an error, thus reserving
1130 these combinations for future expansion. By default, as in Perl, a
1131 backslash followed by a letter with no special meaning is treated as a
1132 literal. (Perl can, however, be persuaded to give a warning for this.)
1133 There are at present no other features controlled by this option. It
1134 can also be set by a (?X) option setting within a pattern.
1135
1136 PCRE_FIRSTLINE
1137
1138 If this option is set, an unanchored pattern is required to match
1139 before or at the first newline in the subject string, though the
1140 matched text may continue over the newline.
1141
1142 PCRE_MULTILINE
1143
1144 By default, PCRE treats the subject string as consisting of a single
1145 line of characters (even if it actually contains newlines). The "start
1146 of line" metacharacter (^) matches only at the start of the string,
1147 while the "end of line" metacharacter ($) matches only at the end of
1148 the string, or before a terminating newline (unless PCRE_DOLLAR_ENDONLY
1149 is set). This is the same as Perl.
1150
1151 When PCRE_MULTILINE it is set, the "start of line" and "end of line"
1152 constructs match immediately following or immediately before internal
1153 newlines in the subject string, respectively, as well as at the very
1154 start and end. This is equivalent to Perl's /m option, and it can be
1155 changed within a pattern by a (?m) option setting. If there are no new-
1156 lines in a subject string, or no occurrences of ^ or $ in a pattern,
1157 setting PCRE_MULTILINE has no effect.
1158
1159 PCRE_NEWLINE_CR
1160 PCRE_NEWLINE_LF
1161 PCRE_NEWLINE_CRLF
1162 PCRE_NEWLINE_ANYCRLF
1163 PCRE_NEWLINE_ANY
1164
1165 These options override the default newline definition that was chosen
1166 when PCRE was built. Setting the first or the second specifies that a
1167 newline is indicated by a single character (CR or LF, respectively).
1168 Setting PCRE_NEWLINE_CRLF specifies that a newline is indicated by the
1169 two-character CRLF sequence. Setting PCRE_NEWLINE_ANYCRLF specifies
1170 that any of the three preceding sequences should be recognized. Setting
1171 PCRE_NEWLINE_ANY specifies that any Unicode newline sequence should be
1172 recognized. The Unicode newline sequences are the three just mentioned,
1173 plus the single characters VT (vertical tab, U+000B), FF (formfeed,
1174 U+000C), NEL (next line, U+0085), LS (line separator, U+2028), and PS
1175 (paragraph separator, U+2029). The last two are recognized only in
1176 UTF-8 mode.
1177
1178 The newline setting in the options word uses three bits that are
1179 treated as a number, giving eight possibilities. Currently only six are
1180 used (default plus the five values above). This means that if you set
1181 more than one newline option, the combination may or may not be sensi-
1182 ble. For example, PCRE_NEWLINE_CR with PCRE_NEWLINE_LF is equivalent to
1183 PCRE_NEWLINE_CRLF, but other combinations may yield unused numbers and
1184 cause an error.
1185
1186 The only time that a line break is specially recognized when compiling
1187 a pattern is if PCRE_EXTENDED is set, and an unescaped # outside a
1188 character class is encountered. This indicates a comment that lasts
1189 until after the next line break sequence. In other circumstances, line
1190 break sequences are treated as literal data, except that in
1191 PCRE_EXTENDED mode, both CR and LF are treated as whitespace characters
1192 and are therefore ignored.
1193
1194 The newline option that is set at compile time becomes the default that
1195 is used for pcre_exec() and pcre_dfa_exec(), but it can be overridden.
1196
1197 PCRE_NO_AUTO_CAPTURE
1198
1199 If this option is set, it disables the use of numbered capturing paren-
1200 theses in the pattern. Any opening parenthesis that is not followed by
1201 ? behaves as if it were followed by ?: but named parentheses can still
1202 be used for capturing (and they acquire numbers in the usual way).
1203 There is no equivalent of this option in Perl.
1204
1205 PCRE_UNGREEDY
1206
1207 This option inverts the "greediness" of the quantifiers so that they
1208 are not greedy by default, but become greedy if followed by "?". It is
1209 not compatible with Perl. It can also be set by a (?U) option setting
1210 within the pattern.
1211
1212 PCRE_UTF8
1213
1214 This option causes PCRE to regard both the pattern and the subject as
1215 strings of UTF-8 characters instead of single-byte character strings.
1216 However, it is available only when PCRE is built to include UTF-8 sup-
1217 port. If not, the use of this option provokes an error. Details of how
1218 this option changes the behaviour of PCRE are given in the section on
1219 UTF-8 support in the main pcre page.
1220
1221 PCRE_NO_UTF8_CHECK
1222
1223 When PCRE_UTF8 is set, the validity of the pattern as a UTF-8 string is
1224 automatically checked. If an invalid UTF-8 sequence of bytes is found,
1225 pcre_compile() returns an error. If you already know that your pattern
1226 is valid, and you want to skip this check for performance reasons, you
1227 can set the PCRE_NO_UTF8_CHECK option. When it is set, the effect of
1228 passing an invalid UTF-8 string as a pattern is undefined. It may cause
1229 your program to crash. Note that this option can also be passed to
1230 pcre_exec() and pcre_dfa_exec(), to suppress the UTF-8 validity check-
1231 ing of subject strings.
1232
1233
1234 COMPILATION ERROR CODES
1235
1236 The following table lists the error codes than may be returned by
1237 pcre_compile2(), along with the error messages that may be returned by
1238 both compiling functions. As PCRE has developed, some error codes have
1239 fallen out of use. To avoid confusion, they have not been re-used.
1240
1241 0 no error
1242 1 \ at end of pattern
1243 2 \c at end of pattern
1244 3 unrecognized character follows \
1245 4 numbers out of order in {} quantifier
1246 5 number too big in {} quantifier
1247 6 missing terminating ] for character class
1248 7 invalid escape sequence in character class
1249 8 range out of order in character class
1250 9 nothing to repeat
1251 10 [this code is not in use]
1252 11 internal error: unexpected repeat
1253 12 unrecognized character after (?
1254 13 POSIX named classes are supported only within a class
1255 14 missing )
1256 15 reference to non-existent subpattern
1257 16 erroffset passed as NULL
1258 17 unknown option bit(s) set
1259 18 missing ) after comment
1260 19 [this code is not in use]
1261 20 regular expression too large
1262 21 failed to get memory
1263 22 unmatched parentheses
1264 23 internal error: code overflow
1265 24 unrecognized character after (?<
1266 25 lookbehind assertion is not fixed length
1267 26 malformed number or name after (?(
1268 27 conditional group contains more than two branches
1269 28 assertion expected after (?(
1270 29 (?R or (?[+-]digits must be followed by )
1271 30 unknown POSIX class name
1272 31 POSIX collating elements are not supported
1273 32 this version of PCRE is not compiled with PCRE_UTF8 support
1274 33 [this code is not in use]
1275 34 character value in \x{...} sequence is too large
1276 35 invalid condition (?(0)
1277 36 \C not allowed in lookbehind assertion
1278 37 PCRE does not support \L, \l, \N, \U, or \u
1279 38 number after (?C is > 255
1280 39 closing ) for (?C expected
1281 40 recursive call could loop indefinitely
1282 41 unrecognized character after (?P
1283 42 syntax error in subpattern name (missing terminator)
1284 43 two named subpatterns have the same name
1285 44 invalid UTF-8 string
1286 45 support for \P, \p, and \X has not been compiled
1287 46 malformed \P or \p sequence
1288 47 unknown property name after \P or \p
1289 48 subpattern name is too long (maximum 32 characters)
1290 49 too many named subpatterns (maximum 10,000)
1291 50 repeated subpattern is too long
1292 51 octal value is greater than \377 (not in UTF-8 mode)
1293 52 internal error: overran compiling workspace
1294 53 internal error: previously-checked referenced subpattern not
1295 found
1296 54 DEFINE group contains more than one branch
1297 55 repeating a DEFINE group is not allowed
1298 56 inconsistent NEWLINE options"
1299 57 \g is not followed by a braced name or an optionally braced
1300 non-zero number
1301 58 (?+ or (?- or (?(+ or (?(- must be followed by a non-zero number
1302
1303
1304 STUDYING A PATTERN
1305
1306 pcre_extra *pcre_study(const pcre *code, int options
1307 const char **errptr);
1308
1309 If a compiled pattern is going to be used several times, it is worth
1310 spending more time analyzing it in order to speed up the time taken for
1311 matching. The function pcre_study() takes a pointer to a compiled pat-
1312 tern as its first argument. If studying the pattern produces additional
1313 information that will help speed up matching, pcre_study() returns a
1314 pointer to a pcre_extra block, in which the study_data field points to
1315 the results of the study.
1316
1317 The returned value from pcre_study() can be passed directly to
1318 pcre_exec(). However, a pcre_extra block also contains other fields
1319 that can be set by the caller before the block is passed; these are
1320 described below in the section on matching a pattern.
1321
1322 If studying the pattern does not produce any additional information
1323 pcre_study() returns NULL. In that circumstance, if the calling program
1324 wants to pass any of the other fields to pcre_exec(), it must set up
1325 its own pcre_extra block.
1326
1327 The second argument of pcre_study() contains option bits. At present,
1328 no options are defined, and this argument should always be zero.
1329
1330 The third argument for pcre_study() is a pointer for an error message.
1331 If studying succeeds (even if no data is returned), the variable it
1332 points to is set to NULL. Otherwise it is set to point to a textual
1333 error message. This is a static string that is part of the library. You
1334 must not try to free it. You should test the error pointer for NULL
1335 after calling pcre_study(), to be sure that it has run successfully.
1336
1337 This is a typical call to pcre_study():
1338
1339 pcre_extra *pe;
1340 pe = pcre_study(
1341 re, /* result of pcre_compile() */
1342 0, /* no options exist */
1343 &error); /* set to NULL or points to a message */
1344
1345 At present, studying a pattern is useful only for non-anchored patterns
1346 that do not have a single fixed starting character. A bitmap of possi-
1347 ble starting bytes is created.
1348
1349
1350 LOCALE SUPPORT
1351
1352 PCRE handles caseless matching, and determines whether characters are
1353 letters, digits, or whatever, by reference to a set of tables, indexed
1354 by character value. When running in UTF-8 mode, this applies only to
1355 characters with codes less than 128. Higher-valued codes never match
1356 escapes such as \w or \d, but can be tested with \p if PCRE is built
1357 with Unicode character property support. The use of locales with Uni-
1358 code is discouraged. If you are handling characters with codes greater
1359 than 128, you should either use UTF-8 and Unicode, or use locales, but
1360 not try to mix the two.
1361
1362 PCRE contains an internal set of tables that are used when the final
1363 argument of pcre_compile() is NULL. These are sufficient for many
1364 applications. Normally, the internal tables recognize only ASCII char-
1365 acters. However, when PCRE is built, it is possible to cause the inter-
1366 nal tables to be rebuilt in the default "C" locale of the local system,
1367 which may cause them to be different.
1368
1369 The internal tables can always be overridden by tables supplied by the
1370 application that calls PCRE. These may be created in a different locale
1371 from the default. As more and more applications change to using Uni-
1372 code, the need for this locale support is expected to die away.
1373
1374 External tables are built by calling the pcre_maketables() function,
1375 which has no arguments, in the relevant locale. The result can then be
1376 passed to pcre_compile() or pcre_exec() as often as necessary. For
1377 example, to build and use tables that are appropriate for the French
1378 locale (where accented characters with values greater than 128 are
1379 treated as letters), the following code could be used:
1380
1381 setlocale(LC_CTYPE, "fr_FR");
1382 tables = pcre_maketables();
1383 re = pcre_compile(..., tables);
1384
1385 The locale name "fr_FR" is used on Linux and other Unix-like systems;
1386 if you are using Windows, the name for the French locale is "french".
1387
1388 When pcre_maketables() runs, the tables are built in memory that is
1389 obtained via pcre_malloc. It is the caller's responsibility to ensure
1390 that the memory containing the tables remains available for as long as
1391 it is needed.
1392
1393 The pointer that is passed to pcre_compile() is saved with the compiled
1394 pattern, and the same tables are used via this pointer by pcre_study()
1395 and normally also by pcre_exec(). Thus, by default, for any single pat-
1396 tern, compilation, studying and matching all happen in the same locale,
1397 but different patterns can be compiled in different locales.
1398
1399 It is possible to pass a table pointer or NULL (indicating the use of
1400 the internal tables) to pcre_exec(). Although not intended for this
1401 purpose, this facility could be used to match a pattern in a different
1402 locale from the one in which it was compiled. Passing table pointers at
1403 run time is discussed below in the section on matching a pattern.
1404
1405
1406 INFORMATION ABOUT A PATTERN
1407
1408 int pcre_fullinfo(const pcre *code, const pcre_extra *extra,
1409 int what, void *where);
1410
1411 The pcre_fullinfo() function returns information about a compiled pat-
1412 tern. It replaces the obsolete pcre_info() function, which is neverthe-
1413 less retained for backwards compability (and is documented below).
1414
1415 The first argument for pcre_fullinfo() is a pointer to the compiled
1416 pattern. The second argument is the result of pcre_study(), or NULL if
1417 the pattern was not studied. The third argument specifies which piece
1418 of information is required, and the fourth argument is a pointer to a
1419 variable to receive the data. The yield of the function is zero for
1420 success, or one of the following negative numbers:
1421
1422 PCRE_ERROR_NULL the argument code was NULL
1423 the argument where was NULL
1424 PCRE_ERROR_BADMAGIC the "magic number" was not found
1425 PCRE_ERROR_BADOPTION the value of what was invalid
1426
1427 The "magic number" is placed at the start of each compiled pattern as
1428 an simple check against passing an arbitrary memory pointer. Here is a
1429 typical call of pcre_fullinfo(), to obtain the length of the compiled
1430 pattern:
1431
1432 int rc;
1433 size_t length;
1434 rc = pcre_fullinfo(
1435 re, /* result of pcre_compile() */
1436 pe, /* result of pcre_study(), or NULL */
1437 PCRE_INFO_SIZE, /* what is required */
1438 &length); /* where to put the data */
1439
1440 The possible values for the third argument are defined in pcre.h, and
1441 are as follows:
1442
1443 PCRE_INFO_BACKREFMAX
1444
1445 Return the number of the highest back reference in the pattern. The
1446 fourth argument should point to an int variable. Zero is returned if
1447 there are no back references.
1448
1449 PCRE_INFO_CAPTURECOUNT
1450
1451 Return the number of capturing subpatterns in the pattern. The fourth
1452 argument should point to an int variable.
1453
1454 PCRE_INFO_DEFAULT_TABLES
1455
1456 Return a pointer to the internal default character tables within PCRE.
1457 The fourth argument should point to an unsigned char * variable. This
1458 information call is provided for internal use by the pcre_study() func-
1459 tion. External callers can cause PCRE to use its internal tables by
1460 passing a NULL table pointer.
1461
1462 PCRE_INFO_FIRSTBYTE
1463
1464 Return information about the first byte of any matched string, for a
1465 non-anchored pattern. The fourth argument should point to an int vari-
1466 able. (This option used to be called PCRE_INFO_FIRSTCHAR; the old name
1467 is still recognized for backwards compatibility.)
1468
1469 If there is a fixed first byte, for example, from a pattern such as
1470 (cat|cow|coyote), its value is returned. Otherwise, if either
1471
1472 (a) the pattern was compiled with the PCRE_MULTILINE option, and every
1473 branch starts with "^", or
1474
1475 (b) every branch of the pattern starts with ".*" and PCRE_DOTALL is not
1476 set (if it were set, the pattern would be anchored),
1477
1478 -1 is returned, indicating that the pattern matches only at the start
1479 of a subject string or after any newline within the string. Otherwise
1480 -2 is returned. For anchored patterns, -2 is returned.
1481
1482 PCRE_INFO_FIRSTTABLE
1483
1484 If the pattern was studied, and this resulted in the construction of a
1485 256-bit table indicating a fixed set of bytes for the first byte in any
1486 matching string, a pointer to the table is returned. Otherwise NULL is
1487 returned. The fourth argument should point to an unsigned char * vari-
1488 able.
1489
1490 PCRE_INFO_JCHANGED
1491
1492 Return 1 if the (?J) option setting is used in the pattern, otherwise
1493 0. The fourth argument should point to an int variable. The (?J) inter-
1494 nal option setting changes the local PCRE_DUPNAMES option.
1495
1496 PCRE_INFO_LASTLITERAL
1497
1498 Return the value of the rightmost literal byte that must exist in any
1499 matched string, other than at its start, if such a byte has been
1500 recorded. The fourth argument should point to an int variable. If there
1501 is no such byte, -1 is returned. For anchored patterns, a last literal
1502 byte is recorded only if it follows something of variable length. For
1503 example, for the pattern /^a\d+z\d+/ the returned value is "z", but for
1504 /^a\dz\d/ the returned value is -1.
1505
1506 PCRE_INFO_NAMECOUNT
1507 PCRE_INFO_NAMEENTRYSIZE
1508 PCRE_INFO_NAMETABLE
1509
1510 PCRE supports the use of named as well as numbered capturing parenthe-
1511 ses. The names are just an additional way of identifying the parenthe-
1512 ses, which still acquire numbers. Several convenience functions such as
1513 pcre_get_named_substring() are provided for extracting captured sub-
1514 strings by name. It is also possible to extract the data directly, by
1515 first converting the name to a number in order to access the correct
1516 pointers in the output vector (described with pcre_exec() below). To do
1517 the conversion, you need to use the name-to-number map, which is
1518 described by these three values.
1519
1520 The map consists of a number of fixed-size entries. PCRE_INFO_NAMECOUNT
1521 gives the number of entries, and PCRE_INFO_NAMEENTRYSIZE gives the size
1522 of each entry; both of these return an int value. The entry size
1523 depends on the length of the longest name. PCRE_INFO_NAMETABLE returns
1524 a pointer to the first entry of the table (a pointer to char). The
1525 first two bytes of each entry are the number of the capturing parenthe-
1526 sis, most significant byte first. The rest of the entry is the corre-
1527 sponding name, zero terminated. The names are in alphabetical order.
1528 When PCRE_DUPNAMES is set, duplicate names are in order of their paren-
1529 theses numbers. For example, consider the following pattern (assume
1530 PCRE_EXTENDED is set, so white space - including newlines - is
1531 ignored):
1532
1533 (?<date> (?<year>(\d\d)?\d\d) -
1534 (?<month>\d\d) - (?<day>\d\d) )
1535
1536 There are four named subpatterns, so the table has four entries, and
1537 each entry in the table is eight bytes long. The table is as follows,
1538 with non-printing bytes shows in hexadecimal, and undefined bytes shown
1539 as ??:
1540
1541 00 01 d a t e 00 ??
1542 00 05 d a y 00 ?? ??
1543 00 04 m o n t h 00
1544 00 02 y e a r 00 ??
1545
1546 When writing code to extract data from named subpatterns using the
1547 name-to-number map, remember that the length of the entries is likely
1548 to be different for each compiled pattern.
1549
1550 PCRE_INFO_OKPARTIAL
1551
1552 Return 1 if the pattern can be used for partial matching, otherwise 0.
1553 The fourth argument should point to an int variable. The pcrepartial
1554 documentation lists the restrictions that apply to patterns when par-
1555 tial matching is used.
1556
1557 PCRE_INFO_OPTIONS
1558
1559 Return a copy of the options with which the pattern was compiled. The
1560 fourth argument should point to an unsigned long int variable. These
1561 option bits are those specified in the call to pcre_compile(), modified
1562 by any top-level option settings within the pattern itself.
1563
1564 A pattern is automatically anchored by PCRE if all of its top-level
1565 alternatives begin with one of the following:
1566
1567 ^ unless PCRE_MULTILINE is set
1568 \A always
1569 \G always
1570 .* if PCRE_DOTALL is set and there are no back
1571 references to the subpattern in which .* appears
1572
1573 For such patterns, the PCRE_ANCHORED bit is set in the options returned
1574 by pcre_fullinfo().
1575
1576 PCRE_INFO_SIZE
1577
1578 Return the size of the compiled pattern, that is, the value that was
1579 passed as the argument to pcre_malloc() when PCRE was getting memory in
1580 which to place the compiled data. The fourth argument should point to a
1581 size_t variable.
1582
1583 PCRE_INFO_STUDYSIZE
1584
1585 Return the size of the data block pointed to by the study_data field in
1586 a pcre_extra block. That is, it is the value that was passed to
1587 pcre_malloc() when PCRE was getting memory into which to place the data
1588 created by pcre_study(). The fourth argument should point to a size_t
1589 variable.
1590
1591
1592 OBSOLETE INFO FUNCTION
1593
1594 int pcre_info(const pcre *code, int *optptr, int *firstcharptr);
1595
1596 The pcre_info() function is now obsolete because its interface is too
1597 restrictive to return all the available data about a compiled pattern.
1598 New programs should use pcre_fullinfo() instead. The yield of
1599 pcre_info() is the number of capturing subpatterns, or one of the fol-
1600 lowing negative numbers:
1601
1602 PCRE_ERROR_NULL the argument code was NULL
1603 PCRE_ERROR_BADMAGIC the "magic number" was not found
1604
1605 If the optptr argument is not NULL, a copy of the options with which
1606 the pattern was compiled is placed in the integer it points to (see
1607 PCRE_INFO_OPTIONS above).
1608
1609 If the pattern is not anchored and the firstcharptr argument is not
1610 NULL, it is used to pass back information about the first character of
1611 any matched string (see PCRE_INFO_FIRSTBYTE above).
1612
1613
1614 REFERENCE COUNTS
1615
1616 int pcre_refcount(pcre *code, int adjust);
1617
1618 The pcre_refcount() function is used to maintain a reference count in
1619 the data block that contains a compiled pattern. It is provided for the
1620 benefit of applications that operate in an object-oriented manner,
1621 where different parts of the application may be using the same compiled
1622 pattern, but you want to free the block when they are all done.
1623
1624 When a pattern is compiled, the reference count field is initialized to
1625 zero. It is changed only by calling this function, whose action is to
1626 add the adjust value (which may be positive or negative) to it. The
1627 yield of the function is the new value. However, the value of the count
1628 is constrained to lie between 0 and 65535, inclusive. If the new value
1629 is outside these limits, it is forced to the appropriate limit value.
1630
1631 Except when it is zero, the reference count is not correctly preserved
1632 if a pattern is compiled on one host and then transferred to a host
1633 whose byte-order is different. (This seems a highly unlikely scenario.)
1634
1635
1636 MATCHING A PATTERN: THE TRADITIONAL FUNCTION
1637
1638 int pcre_exec(const pcre *code, const pcre_extra *extra,
1639 const char *subject, int length, int startoffset,
1640 int options, int *ovector, int ovecsize);
1641
1642 The function pcre_exec() is called to match a subject string against a
1643 compiled pattern, which is passed in the code argument. If the pattern
1644 has been studied, the result of the study should be passed in the extra
1645 argument. This function is the main matching facility of the library,
1646 and it operates in a Perl-like manner. For specialist use there is also
1647 an alternative matching function, which is described below in the sec-
1648 tion about the pcre_dfa_exec() function.
1649
1650 In most applications, the pattern will have been compiled (and option-
1651 ally studied) in the same process that calls pcre_exec(). However, it
1652 is possible to save compiled patterns and study data, and then use them
1653 later in different processes, possibly even on different hosts. For a
1654 discussion about this, see the pcreprecompile documentation.
1655
1656 Here is an example of a simple call to pcre_exec():
1657
1658 int rc;
1659 int ovector[30];
1660 rc = pcre_exec(
1661 re, /* result of pcre_compile() */
1662 NULL, /* we didn't study the pattern */
1663 "some string", /* the subject string */
1664 11, /* the length of the subject string */
1665 0, /* start at offset 0 in the subject */
1666 0, /* default options */
1667 ovector, /* vector of integers for substring information */
1668 30); /* number of elements (NOT size in bytes) */
1669
1670 Extra data for pcre_exec()
1671
1672 If the extra argument is not NULL, it must point to a pcre_extra data
1673 block. The pcre_study() function returns such a block (when it doesn't
1674 return NULL), but you can also create one for yourself, and pass addi-
1675 tional information in it. The pcre_extra block contains the following
1676 fields (not necessarily in this order):
1677
1678 unsigned long int flags;
1679 void *study_data;
1680 unsigned long int match_limit;
1681 unsigned long int match_limit_recursion;
1682 void *callout_data;
1683 const unsigned char *tables;
1684
1685 The flags field is a bitmap that specifies which of the other fields
1686 are set. The flag bits are:
1687
1688 PCRE_EXTRA_STUDY_DATA
1689 PCRE_EXTRA_MATCH_LIMIT
1690 PCRE_EXTRA_MATCH_LIMIT_RECURSION
1691 PCRE_EXTRA_CALLOUT_DATA
1692 PCRE_EXTRA_TABLES
1693
1694 Other flag bits should be set to zero. The study_data field is set in
1695 the pcre_extra block that is returned by pcre_study(), together with
1696 the appropriate flag bit. You should not set this yourself, but you may
1697 add to the block by setting the other fields and their corresponding
1698 flag bits.
1699
1700 The match_limit field provides a means of preventing PCRE from using up
1701 a vast amount of resources when running patterns that are not going to
1702 match, but which have a very large number of possibilities in their
1703 search trees. The classic example is the use of nested unlimited
1704 repeats.
1705
1706 Internally, PCRE uses a function called match() which it calls repeat-
1707 edly (sometimes recursively). The limit set by match_limit is imposed
1708 on the number of times this function is called during a match, which
1709 has the effect of limiting the amount of backtracking that can take
1710 place. For patterns that are not anchored, the count restarts from zero
1711 for each position in the subject string.
1712
1713 The default value for the limit can be set when PCRE is built; the
1714 default default is 10 million, which handles all but the most extreme
1715 cases. You can override the default by suppling pcre_exec() with a
1716 pcre_extra block in which match_limit is set, and
1717 PCRE_EXTRA_MATCH_LIMIT is set in the flags field. If the limit is
1718 exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
1719
1720 The match_limit_recursion field is similar to match_limit, but instead
1721 of limiting the total number of times that match() is called, it limits
1722 the depth of recursion. The recursion depth is a smaller number than
1723 the total number of calls, because not all calls to match() are recur-
1724 sive. This limit is of use only if it is set smaller than match_limit.
1725
1726 Limiting the recursion depth limits the amount of stack that can be
1727 used, or, when PCRE has been compiled to use memory on the heap instead
1728 of the stack, the amount of heap memory that can be used.
1729
1730 The default value for match_limit_recursion can be set when PCRE is
1731 built; the default default is the same value as the default for
1732 match_limit. You can override the default by suppling pcre_exec() with
1733 a pcre_extra block in which match_limit_recursion is set, and
1734 PCRE_EXTRA_MATCH_LIMIT_RECURSION is set in the flags field. If the
1735 limit is exceeded, pcre_exec() returns PCRE_ERROR_RECURSIONLIMIT.
1736
1737 The pcre_callout field is used in conjunction with the "callout" fea-
1738 ture, which is described in the pcrecallout documentation.
1739
1740 The tables field is used to pass a character tables pointer to
1741 pcre_exec(); this overrides the value that is stored with the compiled
1742 pattern. A non-NULL value is stored with the compiled pattern only if
1743 custom tables were supplied to pcre_compile() via its tableptr argu-
1744 ment. If NULL is passed to pcre_exec() using this mechanism, it forces
1745 PCRE's internal tables to be used. This facility is helpful when re-
1746 using patterns that have been saved after compiling with an external
1747 set of tables, because the external tables might be at a different
1748 address when pcre_exec() is called. See the pcreprecompile documenta-
1749 tion for a discussion of saving compiled patterns for later use.
1750
1751 Option bits for pcre_exec()
1752
1753 The unused bits of the options argument for pcre_exec() must be zero.
1754 The only bits that may be set are PCRE_ANCHORED, PCRE_NEWLINE_xxx,
1755 PCRE_NOTBOL, PCRE_NOTEOL, PCRE_NOTEMPTY, PCRE_NO_UTF8_CHECK and
1756 PCRE_PARTIAL.
1757
1758 PCRE_ANCHORED
1759
1760 The PCRE_ANCHORED option limits pcre_exec() to matching at the first
1761 matching position. If a pattern was compiled with PCRE_ANCHORED, or
1762 turned out to be anchored by virtue of its contents, it cannot be made
1763 unachored at matching time.
1764
1765 PCRE_NEWLINE_CR
1766 PCRE_NEWLINE_LF
1767 PCRE_NEWLINE_CRLF
1768 PCRE_NEWLINE_ANYCRLF
1769 PCRE_NEWLINE_ANY
1770
1771 These options override the newline definition that was chosen or
1772 defaulted when the pattern was compiled. For details, see the descrip-
1773 tion of pcre_compile() above. During matching, the newline choice
1774 affects the behaviour of the dot, circumflex, and dollar metacharac-
1775 ters. It may also alter the way the match position is advanced after a
1776 match failure for an unanchored pattern. When PCRE_NEWLINE_CRLF,
1777 PCRE_NEWLINE_ANYCRLF, or PCRE_NEWLINE_ANY is set, and a match attempt
1778 fails when the current position is at a CRLF sequence, the match posi-
1779 tion is advanced by two characters instead of one, in other words, to
1780 after the CRLF.
1781
1782 PCRE_NOTBOL
1783
1784 This option specifies that first character of the subject string is not
1785 the beginning of a line, so the circumflex metacharacter should not
1786 match before it. Setting this without PCRE_MULTILINE (at compile time)
1787 causes circumflex never to match. This option affects only the behav-
1788 iour of the circumflex metacharacter. It does not affect \A.
1789
1790 PCRE_NOTEOL
1791
1792 This option specifies that the end of the subject string is not the end
1793 of a line, so the dollar metacharacter should not match it nor (except
1794 in multiline mode) a newline immediately before it. Setting this with-
1795 out PCRE_MULTILINE (at compile time) causes dollar never to match. This
1796 option affects only the behaviour of the dollar metacharacter. It does
1797 not affect \Z or \z.
1798
1799 PCRE_NOTEMPTY
1800
1801 An empty string is not considered to be a valid match if this option is
1802 set. If there are alternatives in the pattern, they are tried. If all
1803 the alternatives match the empty string, the entire match fails. For
1804 example, if the pattern
1805
1806 a?b?
1807
1808 is applied to a string not beginning with "a" or "b", it matches the
1809 empty string at the start of the subject. With PCRE_NOTEMPTY set, this
1810 match is not valid, so PCRE searches further into the string for occur-
1811 rences of "a" or "b".
1812
1813 Perl has no direct equivalent of PCRE_NOTEMPTY, but it does make a spe-
1814 cial case of a pattern match of the empty string within its split()
1815 function, and when using the /g modifier. It is possible to emulate
1816 Perl's behaviour after matching a null string by first trying the match
1817 again at the same offset with PCRE_NOTEMPTY and PCRE_ANCHORED, and then
1818 if that fails by advancing the starting offset (see below) and trying
1819 an ordinary match again. There is some code that demonstrates how to do
1820 this in the pcredemo.c sample program.
1821
1822 PCRE_NO_UTF8_CHECK
1823
1824 When PCRE_UTF8 is set at compile time, the validity of the subject as a
1825 UTF-8 string is automatically checked when pcre_exec() is subsequently
1826 called. The value of startoffset is also checked to ensure that it
1827 points to the start of a UTF-8 character. If an invalid UTF-8 sequence
1828 of bytes is found, pcre_exec() returns the error PCRE_ERROR_BADUTF8. If
1829 startoffset contains an invalid value, PCRE_ERROR_BADUTF8_OFFSET is
1830 returned.
1831
1832 If you already know that your subject is valid, and you want to skip
1833 these checks for performance reasons, you can set the
1834 PCRE_NO_UTF8_CHECK option when calling pcre_exec(). You might want to
1835 do this for the second and subsequent calls to pcre_exec() if you are
1836 making repeated calls to find all the matches in a single subject
1837 string. However, you should be sure that the value of startoffset
1838 points to the start of a UTF-8 character. When PCRE_NO_UTF8_CHECK is
1839 set, the effect of passing an invalid UTF-8 string as a subject, or a
1840 value of startoffset that does not point to the start of a UTF-8 char-
1841 acter, is undefined. Your program may crash.
1842
1843 PCRE_PARTIAL
1844
1845 This option turns on the partial matching feature. If the subject
1846 string fails to match the pattern, but at some point during the match-
1847 ing process the end of the subject was reached (that is, the subject
1848 partially matches the pattern and the failure to match occurred only
1849 because there were not enough subject characters), pcre_exec() returns
1850 PCRE_ERROR_PARTIAL instead of PCRE_ERROR_NOMATCH. When PCRE_PARTIAL is
1851 used, there are restrictions on what may appear in the pattern. These
1852 are discussed in the pcrepartial documentation.
1853
1854 The string to be matched by pcre_exec()
1855
1856 The subject string is passed to pcre_exec() as a pointer in subject, a
1857 length in length, and a starting byte offset in startoffset. In UTF-8
1858 mode, the byte offset must point to the start of a UTF-8 character.
1859 Unlike the pattern string, the subject may contain binary zero bytes.
1860 When the starting offset is zero, the search for a match starts at the
1861 beginning of the subject, and this is by far the most common case.
1862
1863 A non-zero starting offset is useful when searching for another match
1864 in the same subject by calling pcre_exec() again after a previous suc-
1865 cess. Setting startoffset differs from just passing over a shortened
1866 string and setting PCRE_NOTBOL in the case of a pattern that begins
1867 with any kind of lookbehind. For example, consider the pattern
1868
1869 \Biss\B
1870
1871 which finds occurrences of "iss" in the middle of words. (\B matches
1872 only if the current position in the subject is not a word boundary.)
1873 When applied to the string "Mississipi" the first call to pcre_exec()
1874 finds the first occurrence. If pcre_exec() is called again with just
1875 the remainder of the subject, namely "issipi", it does not match,
1876 because \B is always false at the start of the subject, which is deemed
1877 to be a word boundary. However, if pcre_exec() is passed the entire
1878 string again, but with startoffset set to 4, it finds the second occur-
1879 rence of "iss" because it is able to look behind the starting point to
1880 discover that it is preceded by a letter.
1881
1882 If a non-zero starting offset is passed when the pattern is anchored,
1883 one attempt to match at the given offset is made. This can only succeed
1884 if the pattern does not require the match to be at the start of the
1885 subject.
1886
1887 How pcre_exec() returns captured substrings
1888
1889 In general, a pattern matches a certain portion of the subject, and in
1890 addition, further substrings from the subject may be picked out by
1891 parts of the pattern. Following the usage in Jeffrey Friedl's book,
1892 this is called "capturing" in what follows, and the phrase "capturing
1893 subpattern" is used for a fragment of a pattern that picks out a sub-
1894 string. PCRE supports several other kinds of parenthesized subpattern
1895 that do not cause substrings to be captured.
1896
1897 Captured substrings are returned to the caller via a vector of integer
1898 offsets whose address is passed in ovector. The number of elements in
1899 the vector is passed in ovecsize, which must be a non-negative number.
1900 Note: this argument is NOT the size of ovector in bytes.
1901
1902 The first two-thirds of the vector is used to pass back captured sub-
1903 strings, each substring using a pair of integers. The remaining third
1904 of the vector is used as workspace by pcre_exec() while matching cap-
1905 turing subpatterns, and is not available for passing back information.
1906 The length passed in ovecsize should always be a multiple of three. If
1907 it is not, it is rounded down.
1908
1909 When a match is successful, information about captured substrings is
1910 returned in pairs of integers, starting at the beginning of ovector,
1911 and continuing up to two-thirds of its length at the most. The first
1912 element of a pair is set to the offset of the first character in a sub-
1913 string, and the second is set to the offset of the first character
1914 after the end of a substring. The first pair, ovector[0] and ovec-
1915 tor[1], identify the portion of the subject string matched by the
1916 entire pattern. The next pair is used for the first capturing subpat-
1917 tern, and so on. The value returned by pcre_exec() is one more than the
1918 highest numbered pair that has been set. For example, if two substrings
1919 have been captured, the returned value is 3. If there are no capturing
1920 subpatterns, the return value from a successful match is 1, indicating
1921 that just the first pair of offsets has been set.
1922
1923 If a capturing subpattern is matched repeatedly, it is the last portion
1924 of the string that it matched that is returned.
1925
1926 If the vector is too small to hold all the captured substring offsets,
1927 it is used as far as possible (up to two-thirds of its length), and the
1928 function returns a value of zero. In particular, if the substring off-
1929 sets are not of interest, pcre_exec() may be called with ovector passed
1930 as NULL and ovecsize as zero. However, if the pattern contains back
1931 references and the ovector is not big enough to remember the related
1932 substrings, PCRE has to get additional memory for use during matching.
1933 Thus it is usually advisable to supply an ovector.
1934
1935 The pcre_info() function can be used to find out how many capturing
1936 subpatterns there are in a compiled pattern. The smallest size for
1937 ovector that will allow for n captured substrings, in addition to the
1938 offsets of the substring matched by the whole pattern, is (n+1)*3.
1939
1940 It is possible for capturing subpattern number n+1 to match some part
1941 of the subject when subpattern n has not been used at all. For example,
1942 if the string "abc" is matched against the pattern (a|(z))(bc) the
1943 return from the function is 4, and subpatterns 1 and 3 are matched, but
1944 2 is not. When this happens, both values in the offset pairs corre-
1945 sponding to unused subpatterns are set to -1.
1946
1947 Offset values that correspond to unused subpatterns at the end of the
1948 expression are also set to -1. For example, if the string "abc" is
1949 matched against the pattern (abc)(x(yz)?)? subpatterns 2 and 3 are not
1950 matched. The return from the function is 2, because the highest used
1951 capturing subpattern number is 1. However, you can refer to the offsets
1952 for the second and third capturing subpatterns if you wish (assuming
1953 the vector is large enough, of course).
1954
1955 Some convenience functions are provided for extracting the captured
1956 substrings as separate strings. These are described below.
1957
1958 Error return values from pcre_exec()
1959
1960 If pcre_exec() fails, it returns a negative number. The following are
1961 defined in the header file:
1962
1963 PCRE_ERROR_NOMATCH (-1)
1964
1965 The subject string did not match the pattern.
1966
1967 PCRE_ERROR_NULL (-2)
1968
1969 Either code or subject was passed as NULL, or ovector was NULL and
1970 ovecsize was not zero.
1971
1972 PCRE_ERROR_BADOPTION (-3)
1973
1974 An unrecognized bit was set in the options argument.
1975
1976 PCRE_ERROR_BADMAGIC (-4)
1977
1978 PCRE stores a 4-byte "magic number" at the start of the compiled code,
1979 to catch the case when it is passed a junk pointer and to detect when a
1980 pattern that was compiled in an environment of one endianness is run in
1981 an environment with the other endianness. This is the error that PCRE
1982 gives when the magic number is not present.
1983
1984 PCRE_ERROR_UNKNOWN_OPCODE (-5)
1985
1986 While running the pattern match, an unknown item was encountered in the
1987 compiled pattern. This error could be caused by a bug in PCRE or by
1988 overwriting of the compiled pattern.
1989
1990 PCRE_ERROR_NOMEMORY (-6)
1991
1992 If a pattern contains back references, but the ovector that is passed
1993 to pcre_exec() is not big enough to remember the referenced substrings,
1994 PCRE gets a block of memory at the start of matching to use for this
1995 purpose. If the call via pcre_malloc() fails, this error is given. The
1996 memory is automatically freed at the end of matching.
1997
1998 PCRE_ERROR_NOSUBSTRING (-7)
1999
2000 This error is used by the pcre_copy_substring(), pcre_get_substring(),
2001 and pcre_get_substring_list() functions (see below). It is never
2002 returned by pcre_exec().
2003
2004 PCRE_ERROR_MATCHLIMIT (-8)
2005
2006 The backtracking limit, as specified by the match_limit field in a
2007 pcre_extra structure (or defaulted) was reached. See the description
2008 above.
2009
2010 PCRE_ERROR_CALLOUT (-9)
2011
2012 This error is never generated by pcre_exec() itself. It is provided for
2013 use by callout functions that want to yield a distinctive error code.
2014 See the pcrecallout documentation for details.
2015
2016 PCRE_ERROR_BADUTF8 (-10)
2017
2018 A string that contains an invalid UTF-8 byte sequence was passed as a
2019 subject.
2020
2021 PCRE_ERROR_BADUTF8_OFFSET (-11)
2022
2023 The UTF-8 byte sequence that was passed as a subject was valid, but the
2024 value of startoffset did not point to the beginning of a UTF-8 charac-
2025 ter.
2026
2027 PCRE_ERROR_PARTIAL (-12)
2028
2029 The subject string did not match, but it did match partially. See the
2030 pcrepartial documentation for details of partial matching.
2031
2032 PCRE_ERROR_BADPARTIAL (-13)
2033
2034 The PCRE_PARTIAL option was used with a compiled pattern containing
2035 items that are not supported for partial matching. See the pcrepartial
2036 documentation for details of partial matching.
2037
2038 PCRE_ERROR_INTERNAL (-14)
2039
2040 An unexpected internal error has occurred. This error could be caused
2041 by a bug in PCRE or by overwriting of the compiled pattern.
2042
2043 PCRE_ERROR_BADCOUNT (-15)
2044
2045 This error is given if the value of the ovecsize argument is negative.
2046
2047 PCRE_ERROR_RECURSIONLIMIT (-21)
2048
2049 The internal recursion limit, as specified by the match_limit_recursion
2050 field in a pcre_extra structure (or defaulted) was reached. See the
2051 description above.
2052
2053 PCRE_ERROR_NULLWSLIMIT (-22)
2054
2055 When a group that can match an empty substring is repeated with an
2056 unbounded upper limit, the subject position at the start of the group
2057 must be remembered, so that a test for an empty string can be made when
2058 the end of the group is reached. Some workspace is required for this;
2059 if it runs out, this error is given.
2060
2061 PCRE_ERROR_BADNEWLINE (-23)
2062
2063 An invalid combination of PCRE_NEWLINE_xxx options was given.
2064
2065 Error numbers -16 to -20 are not used by pcre_exec().
2066
2067
2068 EXTRACTING CAPTURED SUBSTRINGS BY NUMBER
2069
2070 int pcre_copy_substring(const char *subject, int *ovector,
2071 int stringcount, int stringnumber, char *buffer,
2072 int buffersize);
2073
2074 int pcre_get_substring(const char *subject, int *ovector,
2075 int stringcount, int stringnumber,
2076 const char **stringptr);
2077
2078 int pcre_get_substring_list(const char *subject,
2079 int *ovector, int stringcount, const char ***listptr);
2080
2081 Captured substrings can be accessed directly by using the offsets
2082 returned by pcre_exec() in ovector. For convenience, the functions
2083 pcre_copy_substring(), pcre_get_substring(), and pcre_get_sub-
2084 string_list() are provided for extracting captured substrings as new,
2085 separate, zero-terminated strings. These functions identify substrings
2086 by number. The next section describes functions for extracting named
2087 substrings.
2088
2089 A substring that contains a binary zero is correctly extracted and has
2090 a further zero added on the end, but the result is not, of course, a C
2091 string. However, you can process such a string by referring to the
2092 length that is returned by pcre_copy_substring() and pcre_get_sub-
2093 string(). Unfortunately, the interface to pcre_get_substring_list() is
2094 not adequate for handling strings containing binary zeros, because the
2095 end of the final string is not independently indicated.
2096
2097 The first three arguments are the same for all three of these func-
2098 tions: subject is the subject string that has just been successfully
2099 matched, ovector is a pointer to the vector of integer offsets that was
2100 passed to pcre_exec(), and stringcount is the number of substrings that
2101 were captured by the match, including the substring that matched the
2102 entire regular expression. This is the value returned by pcre_exec() if
2103 it is greater than zero. If pcre_exec() returned zero, indicating that
2104 it ran out of space in ovector, the value passed as stringcount should
2105 be the number of elements in the vector divided by three.
2106
2107 The functions pcre_copy_substring() and pcre_get_substring() extract a
2108 single substring, whose number is given as stringnumber. A value of
2109 zero extracts the substring that matched the entire pattern, whereas
2110 higher values extract the captured substrings. For pcre_copy_sub-
2111 string(), the string is placed in buffer, whose length is given by
2112 buffersize, while for pcre_get_substring() a new block of memory is
2113 obtained via pcre_malloc, and its address is returned via stringptr.
2114 The yield of the function is the length of the string, not including
2115 the terminating zero, or one of these error codes:
2116
2117 PCRE_ERROR_NOMEMORY (-6)
2118
2119 The buffer was too small for pcre_copy_substring(), or the attempt to
2120 get memory failed for pcre_get_substring().
2121
2122 PCRE_ERROR_NOSUBSTRING (-7)
2123
2124 There is no substring whose number is stringnumber.
2125
2126 The pcre_get_substring_list() function extracts all available sub-
2127 strings and builds a list of pointers to them. All this is done in a
2128 single block of memory that is obtained via pcre_malloc. The address of
2129 the memory block is returned via listptr, which is also the start of
2130 the list of string pointers. The end of the list is marked by a NULL
2131 pointer. The yield of the function is zero if all went well, or the
2132 error code
2133
2134 PCRE_ERROR_NOMEMORY (-6)
2135
2136 if the attempt to get the memory block failed.
2137
2138 When any of these functions encounter a substring that is unset, which
2139 can happen when capturing subpattern number n+1 matches some part of
2140 the subject, but subpattern n has not been used at all, they return an
2141 empty string. This can be distinguished from a genuine zero-length sub-
2142 string by inspecting the appropriate offset in ovector, which is nega-
2143 tive for unset substrings.
2144
2145 The two convenience functions pcre_free_substring() and pcre_free_sub-
2146 string_list() can be used to free the memory returned by a previous
2147 call of pcre_get_substring() or pcre_get_substring_list(), respec-
2148 tively. They do nothing more than call the function pointed to by
2149 pcre_free, which of course could be called directly from a C program.
2150 However, PCRE is used in some situations where it is linked via a spe-
2151 cial interface to another programming language that cannot use
2152 pcre_free directly; it is for these cases that the functions are pro-
2153 vided.
2154
2155
2156 EXTRACTING CAPTURED SUBSTRINGS BY NAME
2157
2158 int pcre_get_stringnumber(const pcre *code,
2159 const char *name);
2160
2161 int pcre_copy_named_substring(const pcre *code,
2162 const char *subject, int *ovector,
2163 int stringcount, const char *stringname,
2164 char *buffer, int buffersize);
2165
2166 int pcre_get_named_substring(const pcre *code,
2167 const char *subject, int *ovector,
2168 int stringcount, const char *stringname,
2169 const char **stringptr);
2170
2171 To extract a substring by name, you first have to find associated num-
2172 ber. For example, for this pattern
2173
2174 (a+)b(?<xxx>\d+)...
2175
2176 the number of the subpattern called "xxx" is 2. If the name is known to
2177 be unique (PCRE_DUPNAMES was not set), you can find the number from the
2178 name by calling pcre_get_stringnumber(). The first argument is the com-
2179 piled pattern, and the second is the name. The yield of the function is
2180 the subpattern number, or PCRE_ERROR_NOSUBSTRING (-7) if there is no
2181 subpattern of that name.
2182
2183 Given the number, you can extract the substring directly, or use one of
2184 the functions described in the previous section. For convenience, there
2185 are also two functions that do the whole job.
2186
2187 Most of the arguments of pcre_copy_named_substring() and
2188 pcre_get_named_substring() are the same as those for the similarly
2189 named functions that extract by number. As these are described in the
2190 previous section, they are not re-described here. There are just two
2191 differences:
2192
2193 First, instead of a substring number, a substring name is given. Sec-
2194 ond, there is an extra argument, given at the start, which is a pointer
2195 to the compiled pattern. This is needed in order to gain access to the
2196 name-to-number translation table.
2197
2198 These functions call pcre_get_stringnumber(), and if it succeeds, they
2199 then call pcre_copy_substring() or pcre_get_substring(), as appropri-
2200 ate. NOTE: If PCRE_DUPNAMES is set and there are duplicate names, the
2201 behaviour may not be what you want (see the next section).
2202
2203
2204 DUPLICATE SUBPATTERN NAMES
2205
2206 int pcre_get_stringtable_entries(const pcre *code,
2207 const char *name, char **first, char **last);
2208
2209 When a pattern is compiled with the PCRE_DUPNAMES option, names for
2210 subpatterns are not required to be unique. Normally, patterns with
2211 duplicate names are such that in any one match, only one of the named
2212 subpatterns participates. An example is shown in the pcrepattern docu-
2213 mentation. When duplicates are present, pcre_copy_named_substring() and
2214 pcre_get_named_substring() return the first substring corresponding to
2215 the given name that is set. If none are set, an empty string is
2216 returned. The pcre_get_stringnumber() function returns one of the num-
2217 bers that are associated with the name, but it is not defined which it
2218 is.
2219
2220 If you want to get full details of all captured substrings for a given
2221 name, you must use the pcre_get_stringtable_entries() function. The
2222 first argument is the compiled pattern, and the second is the name. The
2223 third and fourth are pointers to variables which are updated by the
2224 function. After it has run, they point to the first and last entries in
2225 the name-to-number table for the given name. The function itself
2226 returns the length of each entry, or PCRE_ERROR_NOSUBSTRING (-7) if
2227 there are none. The format of the table is described above in the sec-
2228 tion entitled Information about a pattern. Given all the relevant
2229 entries for the name, you can extract each of their numbers, and hence
2230 the captured data, if any.
2231
2232
2233 FINDING ALL POSSIBLE MATCHES
2234
2235 The traditional matching function uses a similar algorithm to Perl,
2236 which stops when it finds the first match, starting at a given point in
2237 the subject. If you want to find all possible matches, or the longest
2238 possible match, consider using the alternative matching function (see
2239 below) instead. If you cannot use the alternative function, but still
2240 need to find all possible matches, you can kludge it up by making use
2241 of the callout facility, which is described in the pcrecallout documen-
2242 tation.
2243
2244 What you have to do is to insert a callout right at the end of the pat-
2245 tern. When your callout function is called, extract and save the cur-
2246 rent matched substring. Then return 1, which forces pcre_exec() to
2247 backtrack and try other alternatives. Ultimately, when it runs out of
2248 matches, pcre_exec() will yield PCRE_ERROR_NOMATCH.
2249
2250
2251 MATCHING A PATTERN: THE ALTERNATIVE FUNCTION
2252
2253 int pcre_dfa_exec(const pcre *code, const pcre_extra *extra,
2254 const char *subject, int length, int startoffset,
2255 int options, int *ovector, int ovecsize,
2256 int *workspace, int wscount);
2257
2258 The function pcre_dfa_exec() is called to match a subject string
2259 against a compiled pattern, using a matching algorithm that scans the
2260 subject string just once, and does not backtrack. This has different
2261 characteristics to the normal algorithm, and is not compatible with
2262 Perl. Some of the features of PCRE patterns are not supported. Never-
2263 theless, there are times when this kind of matching can be useful. For
2264 a discussion of the two matching algorithms, see the pcrematching docu-
2265 mentation.
2266
2267 The arguments for the pcre_dfa_exec() function are the same as for
2268 pcre_exec(), plus two extras. The ovector argument is used in a differ-
2269 ent way, and this is described below. The other common arguments are
2270 used in the same way as for pcre_exec(), so their description is not
2271 repeated here.
2272
2273 The two additional arguments provide workspace for the function. The
2274 workspace vector should contain at least 20 elements. It is used for
2275 keeping track of multiple paths through the pattern tree. More
2276 workspace will be needed for patterns and subjects where there are a
2277 lot of potential matches.
2278
2279 Here is an example of a simple call to pcre_dfa_exec():
2280
2281 int rc;
2282 int ovector[10];
2283 int wspace[20];
2284 rc = pcre_dfa_exec(
2285 re, /* result of pcre_compile() */
2286 NULL, /* we didn't study the pattern */
2287 "some string", /* the subject string */
2288 11, /* the length of the subject string */
2289 0, /* start at offset 0 in the subject */
2290 0, /* default options */
2291 ovector, /* vector of integers for substring information */
2292 10, /* number of elements (NOT size in bytes) */
2293 wspace, /* working space vector */
2294 20); /* number of elements (NOT size in bytes) */
2295
2296 Option bits for pcre_dfa_exec()
2297
2298 The unused bits of the options argument for pcre_dfa_exec() must be
2299 zero. The only bits that may be set are PCRE_ANCHORED, PCRE_NEW-
2300 LINE_xxx, PCRE_NOTBOL, PCRE_NOTEOL, PCRE_NOTEMPTY, PCRE_NO_UTF8_CHECK,
2301 PCRE_PARTIAL, PCRE_DFA_SHORTEST, and PCRE_DFA_RESTART. All but the last
2302 three of these are the same as for pcre_exec(), so their description is
2303 not repeated here.
2304
2305 PCRE_PARTIAL
2306
2307 This has the same general effect as it does for pcre_exec(), but the
2308 details are slightly different. When PCRE_PARTIAL is set for
2309 pcre_dfa_exec(), the return code PCRE_ERROR_NOMATCH is converted into
2310 PCRE_ERROR_PARTIAL if the end of the subject is reached, there have
2311 been no complete matches, but there is still at least one matching pos-
2312 sibility. The portion of the string that provided the partial match is
2313 set as the first matching string.
2314
2315 PCRE_DFA_SHORTEST
2316
2317 Setting the PCRE_DFA_SHORTEST option causes the matching algorithm to
2318 stop as soon as it has found one match. Because of the way the alterna-
2319 tive algorithm works, this is necessarily the shortest possible match
2320 at the first possible matching point in the subject string.
2321
2322 PCRE_DFA_RESTART
2323
2324 When pcre_dfa_exec() is called with the PCRE_PARTIAL option, and
2325 returns a partial match, it is possible to call it again, with addi-
2326 tional subject characters, and have it continue with the same match.
2327 The PCRE_DFA_RESTART option requests this action; when it is set, the
2328 workspace and wscount options must reference the same vector as before
2329 because data about the match so far is left in them after a partial
2330 match. There is more discussion of this facility in the pcrepartial
2331 documentation.
2332
2333 Successful returns from pcre_dfa_exec()
2334
2335 When pcre_dfa_exec() succeeds, it may have matched more than one sub-
2336 string in the subject. Note, however, that all the matches from one run
2337 of the function start at the same point in the subject. The shorter
2338 matches are all initial substrings of the longer matches. For example,
2339 if the pattern
2340
2341 <.*>
2342
2343 is matched against the string
2344
2345 This is <something> <something else> <something further> no more
2346
2347 the three matched strings are
2348
2349 <something>
2350 <something> <something else>
2351 <something> <something else> <something further>
2352
2353 On success, the yield of the function is a number greater than zero,
2354 which is the number of matched substrings. The substrings themselves
2355 are returned in ovector. Each string uses two elements; the first is
2356 the offset to the start, and the second is the offset to the end. In
2357 fact, all the strings have the same start offset. (Space could have
2358 been saved by giving this only once, but it was decided to retain some
2359 compatibility with the way pcre_exec() returns data, even though the
2360 meaning of the strings is different.)
2361
2362 The strings are returned in reverse order of length; that is, the long-
2363 est matching string is given first. If there were too many matches to
2364 fit into ovector, the yield of the function is zero, and the vector is
2365 filled with the longest matches.
2366
2367 Error returns from pcre_dfa_exec()
2368
2369 The pcre_dfa_exec() function returns a negative number when it fails.
2370 Many of the errors are the same as for pcre_exec(), and these are
2371 described above. There are in addition the following errors that are
2372 specific to pcre_dfa_exec():
2373
2374 PCRE_ERROR_DFA_UITEM (-16)
2375
2376 This return is given if pcre_dfa_exec() encounters an item in the pat-
2377 tern that it does not support, for instance, the use of \C or a back
2378 reference.
2379
2380 PCRE_ERROR_DFA_UCOND (-17)
2381
2382 This return is given if pcre_dfa_exec() encounters a condition item
2383 that uses a back reference for the condition, or a test for recursion
2384 in a specific group. These are not supported.
2385
2386 PCRE_ERROR_DFA_UMLIMIT (-18)
2387
2388 This return is given if pcre_dfa_exec() is called with an extra block
2389 that contains a setting of the match_limit field. This is not supported
2390 (it is meaningless).
2391
2392 PCRE_ERROR_DFA_WSSIZE (-19)
2393
2394 This return is given if pcre_dfa_exec() runs out of space in the
2395 workspace vector.
2396
2397 PCRE_ERROR_DFA_RECURSE (-20)
2398
2399 When a recursive subpattern is processed, the matching function calls
2400 itself recursively, using private vectors for ovector and workspace.
2401 This error is given if the output vector is not large enough. This
2402 should be extremely rare, as a vector of size 1000 is used.
2403
2404
2405 SEE ALSO
2406
2407 pcrebuild(3), pcrecallout(3), pcrecpp(3)(3), pcrematching(3), pcrepar-
2408 tial(3), pcreposix(3), pcreprecompile(3), pcresample(3), pcrestack(3).
2409
2410
2411 AUTHOR
2412
2413 Philip Hazel
2414 University Computing Service
2415 Cambridge CB2 3QH, England.
2416
2417
2418 REVISION
2419
2420 Last updated: 13 June 2007
2421 Copyright (c) 1997-2007 University of Cambridge.
2422 ------------------------------------------------------------------------------
2423
2424
2425 PCRECALLOUT(3) PCRECALLOUT(3)
2426
2427
2428 NAME
2429 PCRE - Perl-compatible regular expressions
2430
2431
2432 PCRE CALLOUTS
2433
2434 int (*pcre_callout)(pcre_callout_block *);
2435
2436 PCRE provides a feature called "callout", which is a means of temporar-
2437 ily passing control to the caller of PCRE in the middle of pattern
2438 matching. The caller of PCRE provides an external function by putting
2439 its entry point in the global variable pcre_callout. By default, this
2440 variable contains NULL, which disables all calling out.
2441
2442 Within a regular expression, (?C) indicates the points at which the
2443 external function is to be called. Different callout points can be
2444 identified by putting a number less than 256 after the letter C. The
2445 default value is zero. For example, this pattern has two callout
2446 points:
2447
2448 (?C1)abc(?C2)def
2449
2450 If the PCRE_AUTO_CALLOUT option bit is set when pcre_compile() is
2451 called, PCRE automatically inserts callouts, all with number 255,
2452 before each item in the pattern. For example, if PCRE_AUTO_CALLOUT is
2453 used with the pattern
2454
2455 A(\d{2}|--)
2456
2457 it is processed as if it were
2458
2459 (?C255)A(?C255)((?C255)\d{2}(?C255)|(?C255)-(?C255)-(?C255))(?C255)
2460
2461 Notice that there is a callout before and after each parenthesis and
2462 alternation bar. Automatic callouts can be used for tracking the
2463 progress of pattern matching. The pcretest command has an option that
2464 sets automatic callouts; when it is used, the output indicates how the
2465 pattern is matched. This is useful information when you are trying to
2466 optimize the performance of a particular pattern.
2467
2468
2469 MISSING CALLOUTS
2470
2471 You should be aware that, because of optimizations in the way PCRE
2472 matches patterns, callouts sometimes do not happen. For example, if the
2473 pattern is
2474
2475 ab(?C4)cd
2476
2477 PCRE knows that any matching string must contain the letter "d". If the
2478 subject string is "abyz", the lack of "d" means that matching doesn't
2479 ever start, and the callout is never reached. However, with "abyd",
2480 though the result is still no match, the callout is obeyed.
2481
2482
2483 THE CALLOUT INTERFACE
2484
2485 During matching, when PCRE reaches a callout point, the external func-
2486 tion defined by pcre_callout is called (if it is set). This applies to
2487 both the pcre_exec() and the pcre_dfa_exec() matching functions. The
2488 only argument to the callout function is a pointer to a pcre_callout
2489 block. This structure contains the following fields:
2490
2491 int version;
2492 int callout_number;
2493 int *offset_vector;
2494 const char *subject;
2495 int subject_length;
2496 int start_match;
2497 int current_position;
2498 int capture_top;
2499 int capture_last;
2500 void *callout_data;
2501 int pattern_position;
2502 int next_item_length;
2503
2504 The version field is an integer containing the version number of the
2505 block format. The initial version was 0; the current version is 1. The
2506 version number will change again in future if additional fields are
2507 added, but the intention is never to remove any of the existing fields.
2508
2509 The callout_number field contains the number of the callout, as com-
2510 piled into the pattern (that is, the number after ?C for manual call-
2511 outs, and 255 for automatically generated callouts).
2512
2513 The offset_vector field is a pointer to the vector of offsets that was
2514 passed by the caller to pcre_exec() or pcre_dfa_exec(). When
2515 pcre_exec() is used, the contents can be inspected in order to extract
2516 substrings that have been matched so far, in the same way as for
2517 extracting substrings after a match has completed. For pcre_dfa_exec()
2518 this field is not useful.
2519
2520 The subject and subject_length fields contain copies of the values that
2521 were passed to pcre_exec().
2522
2523 The start_match field normally contains the offset within the subject
2524 at which the current match attempt started. However, if the escape
2525 sequence \K has been encountered, this value is changed to reflect the
2526 modified starting point. If the pattern is not anchored, the callout
2527 function may be called several times from the same point in the pattern
2528 for different starting points in the subject.
2529
2530 The current_position field contains the offset within the subject of
2531 the current match pointer.
2532
2533 When the pcre_exec() function is used, the capture_top field contains
2534 one more than the number of the highest numbered captured substring so
2535 far. If no substrings have been captured, the value of capture_top is
2536 one. This is always the case when pcre_dfa_exec() is used, because it
2537 does not support captured substrings.
2538
2539 The capture_last field contains the number of the most recently cap-
2540 tured substring. If no substrings have been captured, its value is -1.
2541 This is always the case when pcre_dfa_exec() is used.
2542
2543 The callout_data field contains a value that is passed to pcre_exec()
2544 or pcre_dfa_exec() specifically so that it can be passed back in call-
2545 outs. It is passed in the pcre_callout field of the pcre_extra data
2546 structure. If no such data was passed, the value of callout_data in a
2547 pcre_callout block is NULL. There is a description of the pcre_extra
2548 structure in the pcreapi documentation.
2549
2550 The pattern_position field is present from version 1 of the pcre_call-
2551 out structure. It contains the offset to the next item to be matched in
2552 the pattern string.
2553
2554 The next_item_length field is present from version 1 of the pcre_call-
2555 out structure. It contains the length of the next item to be matched in
2556 the pattern string. When the callout immediately precedes an alterna-
2557 tion bar, a closing parenthesis, or the end of the pattern, the length
2558 is zero. When the callout precedes an opening parenthesis, the length
2559 is that of the entire subpattern.
2560
2561 The pattern_position and next_item_length fields are intended to help
2562 in distinguishing between different automatic callouts, which all have
2563 the same callout number. However, they are set for all callouts.
2564
2565
2566 RETURN VALUES
2567
2568 The external callout function returns an integer to PCRE. If the value
2569 is zero, matching proceeds as normal. If the value is greater than
2570 zero, matching fails at the current point, but the testing of other
2571 matching possibilities goes ahead, just as if a lookahead assertion had
2572 failed. If the value is less than zero, the match is abandoned, and
2573 pcre_exec() (or pcre_dfa_exec()) returns the negative value.
2574
2575 Negative values should normally be chosen from the set of
2576 PCRE_ERROR_xxx values. In particular, PCRE_ERROR_NOMATCH forces a stan-
2577 dard "no match" failure. The error number PCRE_ERROR_CALLOUT is
2578 reserved for use by callout functions; it will never be used by PCRE
2579 itself.
2580
2581
2582 AUTHOR
2583
2584 Philip Hazel
2585 University Computing Service
2586 Cambridge CB2 3QH, England.
2587
2588
2589 REVISION
2590
2591 Last updated: 29 May 2007
2592 Copyright (c) 1997-2007 University of Cambridge.
2593 ------------------------------------------------------------------------------
2594
2595
2596 PCRECOMPAT(3) PCRECOMPAT(3)
2597
2598
2599 NAME
2600 PCRE - Perl-compatible regular expressions
2601
2602
2603 DIFFERENCES BETWEEN PCRE AND PERL
2604
2605 This document describes the differences in the ways that PCRE and Perl
2606 handle regular expressions. The differences described here are mainly
2607 with respect to Perl 5.8, though PCRE versions 7.0 and later contain
2608 some features that are expected to be in the forthcoming Perl 5.10.
2609
2610 1. PCRE has only a subset of Perl's UTF-8 and Unicode support. Details
2611 of what it does have are given in the section on UTF-8 support in the
2612 main pcre page.
2613
2614 2. PCRE does not allow repeat quantifiers on lookahead assertions. Perl
2615 permits them, but they do not mean what you might think. For example,
2616 (?!a){3} does not assert that the next three characters are not "a". It
2617 just asserts that the next character is not "a" three times.
2618
2619 3. Capturing subpatterns that occur inside negative lookahead asser-
2620 tions are counted, but their entries in the offsets vector are never
2621 set. Perl sets its numerical variables from any such patterns that are
2622 matched before the assertion fails to match something (thereby succeed-
2623 ing), but only if the negative lookahead assertion contains just one
2624 branch.
2625
2626 4. Though binary zero characters are supported in the subject string,
2627 they are not allowed in a pattern string because it is passed as a nor-
2628 mal C string, terminated by zero. The escape sequence \0 can be used in
2629 the pattern to represent a binary zero.
2630
2631 5. The following Perl escape sequences are not supported: \l, \u, \L,
2632 \U, and \N. In fact these are implemented by Perl's general string-han-
2633 dling and are not part of its pattern matching engine. If any of these
2634 are encountered by PCRE, an error is generated.
2635
2636 6. The Perl escape sequences \p, \P, and \X are supported only if PCRE
2637 is built with Unicode character property support. The properties that
2638 can be tested with \p and \P are limited to the general category prop-
2639 erties such as Lu and Nd, script names such as Greek or Han, and the
2640 derived properties Any and L&.
2641
2642 7. PCRE does support the \Q...\E escape for quoting substrings. Charac-
2643 ters in between are treated as literals. This is slightly different
2644 from Perl in that $ and @ are also handled as literals inside the
2645 quotes. In Perl, they cause variable interpolation (but of course PCRE
2646 does not have variables). Note the following examples:
2647
2648 Pattern PCRE matches Perl matches
2649
2650 \Qabc$xyz\E abc$xyz abc followed by the
2651 contents of $xyz
2652 \Qabc\$xyz\E abc\$xyz abc\$xyz
2653 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
2654
2655 The \Q...\E sequence is recognized both inside and outside character
2656 classes.
2657
2658 8. Fairly obviously, PCRE does not support the (?{code}) and (??{code})
2659 constructions. However, there is support for recursive patterns. This
2660 is not available in Perl 5.8, but will be in Perl 5.10. Also, the PCRE
2661 "callout" feature allows an external function to be called during pat-
2662 tern matching. See the pcrecallout documentation for details.
2663
2664 9. Subpatterns that are called recursively or as "subroutines" are
2665 always treated as atomic groups in PCRE. This is like Python, but
2666 unlike Perl.
2667
2668 10. There are some differences that are concerned with the settings of
2669 captured strings when part of a pattern is repeated. For example,
2670 matching "aba" against the pattern /^(a(b)?)+$/ in Perl leaves $2
2671 unset, but in PCRE it is set to "b".
2672
2673 11. PCRE provides some extensions to the Perl regular expression facil-
2674 ities. Perl 5.10 will include new features that are not in earlier
2675 versions, some of which (such as named parentheses) have been in PCRE
2676 for some time. This list is with respect to Perl 5.10:
2677
2678 (a) Although lookbehind assertions must match fixed length strings,
2679 each alternative branch of a lookbehind assertion can match a different
2680 length of string. Perl requires them all to have the same length.
2681
2682 (b) If PCRE_DOLLAR_ENDONLY is set and PCRE_MULTILINE is not set, the $
2683 meta-character matches only at the very end of the string.
2684
2685 (c) If PCRE_EXTRA is set, a backslash followed by a letter with no spe-
2686 cial meaning is faulted. Otherwise, like Perl, the backslash is quietly
2687 ignored. (Perl can be made to issue a warning.)
2688
2689 (d) If PCRE_UNGREEDY is set, the greediness of the repetition quanti-
2690 fiers is inverted, that is, by default they are not greedy, but if fol-
2691 lowed by a question mark they are.
2692
2693 (e) PCRE_ANCHORED can be used at matching time to force a pattern to be
2694 tried only at the first matching position in the subject string.
2695
2696 (f) The PCRE_NOTBOL, PCRE_NOTEOL, PCRE_NOTEMPTY, and PCRE_NO_AUTO_CAP-
2697 TURE options for pcre_exec() have no Perl equivalents.
2698
2699 (g) The callout facility is PCRE-specific.
2700
2701 (h) The partial matching facility is PCRE-specific.
2702
2703 (i) Patterns compiled by PCRE can be saved and re-used at a later time,
2704 even on different hosts that have the other endianness.
2705
2706 (j) The alternative matching function (pcre_dfa_exec()) matches in a
2707 different way and is not Perl-compatible.
2708
2709
2710 AUTHOR
2711
2712 Philip Hazel
2713 University Computing Service
2714 Cambridge CB2 3QH, England.
2715
2716
2717 REVISION
2718
2719 Last updated: 13 June 2007
2720 Copyright (c) 1997-2007 University of Cambridge.
2721 ------------------------------------------------------------------------------
2722
2723
2724 PCREPATTERN(3) PCREPATTERN(3)
2725
2726
2727 NAME
2728 PCRE - Perl-compatible regular expressions
2729
2730
2731 PCRE REGULAR EXPRESSION DETAILS
2732
2733 The syntax and semantics of the regular expressions supported by PCRE
2734 are described below. Regular expressions are also described in the Perl
2735 documentation and in a number of books, some of which have copious
2736 examples. Jeffrey Friedl's "Mastering Regular Expressions", published
2737 by O'Reilly, covers regular expressions in great detail. This descrip-
2738 tion of PCRE's regular expressions is intended as reference material.
2739
2740 The original operation of PCRE was on strings of one-byte characters.
2741 However, there is now also support for UTF-8 character strings. To use
2742 this, you must build PCRE to include UTF-8 support, and then call
2743 pcre_compile() with the PCRE_UTF8 option. How this affects pattern
2744 matching is mentioned in several places below. There is also a summary
2745 of UTF-8 features in the section on UTF-8 support in the main pcre
2746 page.
2747
2748 The remainder of this document discusses the patterns that are sup-
2749 ported by PCRE when its main matching function, pcre_exec(), is used.
2750 From release 6.0, PCRE offers a second matching function,
2751 pcre_dfa_exec(), which matches using a different algorithm that is not
2752 Perl-compatible. Some of the features discussed below are not available
2753 when pcre_dfa_exec() is used. The advantages and disadvantages of the
2754 alternative function, and how it differs from the normal function, are
2755 discussed in the pcrematching page.
2756
2757
2758 CHARACTERS AND METACHARACTERS
2759
2760 A regular expression is a pattern that is matched against a subject
2761 string from left to right. Most characters stand for themselves in a
2762 pattern, and match the corresponding characters in the subject. As a
2763 trivial example, the pattern
2764
2765 The quick brown fox
2766
2767 matches a portion of a subject string that is identical to itself. When
2768 caseless matching is specified (the PCRE_CASELESS option), letters are
2769 matched independently of case. In UTF-8 mode, PCRE always understands
2770 the concept of case for characters whose values are less than 128, so
2771 caseless matching is always possible. For characters with higher val-
2772 ues, the concept of case is supported if PCRE is compiled with Unicode
2773 property support, but not otherwise. If you want to use caseless
2774 matching for characters 128 and above, you must ensure that PCRE is
2775 compiled with Unicode property support as well as with UTF-8 support.
2776
2777 The power of regular expressions comes from the ability to include
2778 alternatives and repetitions in the pattern. These are encoded in the
2779 pattern by the use of metacharacters, which do not stand for themselves
2780 but instead are interpreted in some special way.
2781
2782 There are two different sets of metacharacters: those that are recog-
2783 nized anywhere in the pattern except within square brackets, and those
2784 that are recognized within square brackets. Outside square brackets,
2785 the metacharacters are as follows:
2786
2787 \ general escape character with several uses
2788 ^ assert start of string (or line, in multiline mode)
2789 $ assert end of string (or line, in multiline mode)
2790 . match any character except newline (by default)
2791 [ start character class definition
2792 | start of alternative branch
2793 ( start subpattern
2794 ) end subpattern
2795 ? extends the meaning of (
2796 also 0 or 1 quantifier
2797 also quantifier minimizer
2798 * 0 or more quantifier
2799 + 1 or more quantifier
2800 also "possessive quantifier"
2801 { start min/max quantifier
2802
2803 Part of a pattern that is in square brackets is called a "character
2804 class". In a character class the only metacharacters are:
2805
2806 \ general escape character
2807 ^ negate the class, but only if the first character
2808 - indicates character range
2809 [ POSIX character class (only if followed by POSIX
2810 syntax)
2811 ] terminates the character class
2812
2813 The following sections describe the use of each of the metacharacters.
2814
2815
2816 BACKSLASH
2817
2818 The backslash character has several uses. Firstly, if it is followed by
2819 a non-alphanumeric character, it takes away any special meaning that
2820 character may have. This use of backslash as an escape character
2821 applies both inside and outside character classes.
2822
2823 For example, if you want to match a * character, you write \* in the
2824 pattern. This escaping action applies whether or not the following
2825 character would otherwise be interpreted as a metacharacter, so it is
2826 always safe to precede a non-alphanumeric with backslash to specify
2827 that it stands for itself. In particular, if you want to match a back-
2828 slash, you write \\.
2829
2830 If a pattern is compiled with the PCRE_EXTENDED option, whitespace in
2831 the pattern (other than in a character class) and characters between a
2832 # outside a character class and the next newline are ignored. An escap-
2833 ing backslash can be used to include a whitespace or # character as
2834 part of the pattern.
2835
2836 If you want to remove the special meaning from a sequence of charac-
2837 ters, you can do so by putting them between \Q and \E. This is differ-
2838 ent from Perl in that $ and @ are handled as literals in \Q...\E
2839 sequences in PCRE, whereas in Perl, $ and @ cause variable interpola-
2840 tion. Note the following examples:
2841
2842 Pattern PCRE matches Perl matches
2843
2844 \Qabc$xyz\E abc$xyz abc followed by the
2845 contents of $xyz
2846 \Qabc\$xyz\E abc\$xyz abc\$xyz
2847 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
2848
2849 The \Q...\E sequence is recognized both inside and outside character
2850 classes.
2851
2852 Non-printing characters
2853
2854 A second use of backslash provides a way of encoding non-printing char-
2855 acters in patterns in a visible manner. There is no restriction on the
2856 appearance of non-printing characters, apart from the binary zero that
2857 terminates a pattern, but when a pattern is being prepared by text
2858 editing, it is usually easier to use one of the following escape
2859 sequences than the binary character it represents:
2860
2861 \a alarm, that is, the BEL character (hex 07)
2862 \cx "control-x", where x is any character
2863 \e escape (hex 1B)
2864 \f formfeed (hex 0C)
2865 \n newline (hex 0A)
2866 \r carriage return (hex 0D)
2867 \t tab (hex 09)
2868 \ddd character with octal code ddd, or backreference
2869 \xhh character with hex code hh
2870 \x{hhh..} character with hex code hhh..
2871
2872 The precise effect of \cx is as follows: if x is a lower case letter,
2873 it is converted to upper case. Then bit 6 of the character (hex 40) is
2874 inverted. Thus \cz becomes hex 1A, but \c{ becomes hex 3B, while \c;
2875 becomes hex 7B.
2876
2877 After \x, from zero to two hexadecimal digits are read (letters can be
2878 in upper or lower case). Any number of hexadecimal digits may appear
2879 between \x{ and }, but the value of the character code must be less
2880 than 256 in non-UTF-8 mode, and less than 2**31 in UTF-8 mode (that is,
2881 the maximum hexadecimal value is 7FFFFFFF). If characters other than
2882 hexadecimal digits appear between \x{ and }, or if there is no termi-
2883 nating }, this form of escape is not recognized. Instead, the initial
2884 \x will be interpreted as a basic hexadecimal escape, with no following
2885 digits, giving a character whose value is zero.
2886
2887 Characters whose value is less than 256 can be defined by either of the
2888 two syntaxes for \x. There is no difference in the way they are han-
2889 dled. For example, \xdc is exactly the same as \x{dc}.
2890
2891 After \0 up to two further octal digits are read. If there are fewer
2892 than two digits, just those that are present are used. Thus the
2893 sequence \0\x\07 specifies two binary zeros followed by a BEL character
2894 (code value 7). Make sure you supply two digits after the initial zero
2895 if the pattern character that follows is itself an octal digit.
2896
2897 The handling of a backslash followed by a digit other than 0 is compli-
2898 cated. Outside a character class, PCRE reads it and any following dig-
2899 its as a decimal number. If the number is less than 10, or if there
2900 have been at least that many previous capturing left parentheses in the
2901 expression, the entire sequence is taken as a back reference. A
2902 description of how this works is given later, following the discussion
2903 of parenthesized subpatterns.
2904
2905 Inside a character class, or if the decimal number is greater than 9
2906 and there have not been that many capturing subpatterns, PCRE re-reads
2907 up to three octal digits following the backslash, and uses them to gen-
2908 erate a data character. Any subsequent digits stand for themselves. In
2909 non-UTF-8 mode, the value of a character specified in octal must be
2910 less than \400. In UTF-8 mode, values up to \777 are permitted. For
2911 example:
2912
2913 \040 is another way of writing a space
2914 \40 is the same, provided there are fewer than 40
2915 previous capturing subpatterns
2916 \7 is always a back reference
2917 \11 might be a back reference, or another way of
2918 writing a tab
2919 \011 is always a tab
2920 \0113 is a tab followed by the character "3"
2921 \113 might be a back reference, otherwise the
2922 character with octal code 113
2923 \377 might be a back reference, otherwise
2924 the byte consisting entirely of 1 bits
2925 \81 is either a back reference, or a binary zero
2926 followed by the two characters "8" and "1"
2927
2928 Note that octal values of 100 or greater must not be introduced by a
2929 leading zero, because no more than three octal digits are ever read.
2930
2931 All the sequences that define a single character value can be used both
2932 inside and outside character classes. In addition, inside a character
2933 class, the sequence \b is interpreted as the backspace character (hex
2934 08), and the sequences \R and \X are interpreted as the characters "R"
2935 and "X", respectively. Outside a character class, these sequences have
2936 different meanings (see below).
2937
2938 Absolute and relative back references
2939
2940 The sequence \g followed by a positive or negative number, optionally
2941 enclosed in braces, is an absolute or relative back reference. A named
2942 back reference can be coded as \g{name}. Back references are discussed
2943 later, following the discussion of parenthesized subpatterns.
2944
2945 Generic character types
2946
2947 Another use of backslash is for specifying generic character types. The
2948 following are always recognized:
2949
2950 \d any decimal digit
2951 \D any character that is not a decimal digit
2952 \h any horizontal whitespace character
2953 \H any character that is not a horizontal whitespace character
2954 \s any whitespace character
2955 \S any character that is not a whitespace character
2956 \v any vertical whitespace character
2957 \V any character that is not a vertical whitespace character
2958 \w any "word" character
2959 \W any "non-word" character
2960
2961 Each pair of escape sequences partitions the complete set of characters
2962 into two disjoint sets. Any given character matches one, and only one,
2963 of each pair.
2964
2965 These character type sequences can appear both inside and outside char-
2966 acter classes. They each match one character of the appropriate type.
2967 If the current matching point is at the end of the subject string, all
2968 of them fail, since there is no character to match.
2969
2970 For compatibility with Perl, \s does not match the VT character (code
2971 11). This makes it different from the the POSIX "space" class. The \s
2972 characters are HT (9), LF (10), FF (12), CR (13), and space (32). If
2973 "use locale;" is included in a Perl script, \s may match the VT charac-
2974 ter. In PCRE, it never does.
2975
2976 In UTF-8 mode, characters with values greater than 128 never match \d,
2977 \s, or \w, and always match \D, \S, and \W. This is true even when Uni-
2978 code character property support is available. These sequences retain
2979 their original meanings from before UTF-8 support was available, mainly
2980 for efficiency reasons.
2981
2982 The sequences \h, \H, \v, and \V are Perl 5.10 features. In contrast to
2983 the other sequences, these do match certain high-valued codepoints in
2984 UTF-8 mode. The horizontal space characters are:
2985
2986 U+0009 Horizontal tab
2987 U+0020 Space
2988 U+00A0 Non-break space
2989 U+1680 Ogham space mark
2990 U+180E Mongolian vowel separator
2991 U+2000 En quad
2992 U+2001 Em quad
2993 U+2002 En space
2994 U+2003 Em space
2995 U+2004 Three-per-em space
2996 U+2005 Four-per-em space
2997 U+2006 Six-per-em space
2998 U+2007 Figure space
2999 U+2008 Punctuation space
3000 U+2009 Thin space
3001 U+200A Hair space
3002 U+202F Narrow no-break space
3003 U+205F Medium mathematical space
3004 U+3000 Ideographic space
3005
3006 The vertical space characters are:
3007
3008 U+000A Linefeed
3009 U+000B Vertical tab
3010 U+000C Formfeed
3011 U+000D Carriage return
3012 U+0085 Next line
3013 U+2028 Line separator
3014 U+2029 Paragraph separator
3015
3016 A "word" character is an underscore or any character less than 256 that
3017 is a letter or digit. The definition of letters and digits is con-
3018 trolled by PCRE's low-valued character tables, and may vary if locale-
3019 specific matching is taking place (see "Locale support" in the pcreapi
3020 page). For example, in a French locale such as "fr_FR" in Unix-like
3021 systems, or "french" in Windows, some character codes greater than 128
3022 are used for accented letters, and these are matched by \w. The use of
3023 locales with Unicode is discouraged.
3024
3025 Newline sequences
3026
3027 Outside a character class, the escape sequence \R matches any Unicode
3028 newline sequence. This is a Perl 5.10 feature. In non-UTF-8 mode \R is
3029 equivalent to the following:
3030
3031 (?>\r\n|\n|\x0b|\f|\r|\x85)
3032
3033 This is an example of an "atomic group", details of which are given
3034 below. This particular group matches either the two-character sequence
3035 CR followed by LF, or one of the single characters LF (linefeed,
3036 U+000A), VT (vertical tab, U+000B), FF (formfeed, U+000C), CR (carriage
3037 return, U+000D), or NEL (next line, U+0085). The two-character sequence
3038 is treated as a single unit that cannot be split.
3039
3040 In UTF-8 mode, two additional characters whose codepoints are greater
3041 than 255 are added: LS (line separator, U+2028) and PS (paragraph sepa-
3042 rator, U+2029). Unicode character property support is not needed for
3043 these characters to be recognized.
3044
3045 Inside a character class, \R matches the letter "R".
3046
3047 Unicode character properties
3048
3049 When PCRE is built with Unicode character property support, three addi-
3050 tional escape sequences that match characters with specific properties
3051 are available. When not in UTF-8 mode, these sequences are of course
3052 limited to testing characters whose codepoints are less than 256, but
3053 they do work in this mode. The extra escape sequences are:
3054
3055 \p{xx} a character with the xx property
3056 \P{xx} a character without the xx property
3057 \X an extended Unicode sequence
3058
3059 The property names represented by xx above are limited to the Unicode
3060 script names, the general category properties, and "Any", which matches
3061 any character (including newline). Other properties such as "InMusical-
3062 Symbols" are not currently supported by PCRE. Note that \P{Any} does
3063 not match any characters, so always causes a match failure.
3064
3065 Sets of Unicode characters are defined as belonging to certain scripts.
3066 A character from one of these sets can be matched using a script name.
3067 For example:
3068
3069 \p{Greek}
3070 \P{Han}
3071
3072 Those that are not part of an identified script are lumped together as
3073 "Common". The current list of scripts is:
3074
3075 Arabic, Armenian, Balinese, Bengali, Bopomofo, Braille, Buginese,
3076 Buhid, Canadian_Aboriginal, Cherokee, Common, Coptic, Cuneiform,
3077 Cypriot, Cyrillic, Deseret, Devanagari, Ethiopic, Georgian, Glagolitic,
3078 Gothic, Greek, Gujarati, Gurmukhi, Han, Hangul, Hanunoo, Hebrew, Hira-
3079 gana, Inherited, Kannada, Katakana, Kharoshthi, Khmer, Lao, Latin,
3080 Limbu, Linear_B, Malayalam, Mongolian, Myanmar, New_Tai_Lue, Nko,
3081 Ogham, Old_Italic, Old_Persian, Oriya, Osmanya, Phags_Pa, Phoenician,
3082 Runic, Shavian, Sinhala, Syloti_Nagri, Syriac, Tagalog, Tagbanwa,
3083 Tai_Le, Tamil, Telugu, Thaana, Thai, Tibetan, Tifinagh, Ugaritic, Yi.
3084
3085 Each character has exactly one general category property, specified by
3086 a two-letter abbreviation. For compatibility with Perl, negation can be
3087 specified by including a circumflex between the opening brace and the
3088 property name. For example, \p{^Lu} is the same as \P{Lu}.
3089
3090 If only one letter is specified with \p or \P, it includes all the gen-
3091 eral category properties that start with that letter. In this case, in
3092 the absence of negation, the curly brackets in the escape sequence are
3093 optional; these two examples have the same effect:
3094
3095 \p{L}
3096 \pL
3097
3098 The following general category property codes are supported:
3099
3100 C Other
3101 Cc Control
3102 Cf Format
3103 Cn Unassigned
3104 Co Private use
3105 Cs Surrogate
3106
3107 L Letter
3108 Ll Lower case letter
3109 Lm Modifier letter
3110 Lo Other letter
3111 Lt Title case letter
3112 Lu Upper case letter
3113
3114 M Mark
3115 Mc Spacing mark
3116 Me Enclosing mark
3117 Mn Non-spacing mark
3118
3119 N Number
3120 Nd Decimal number
3121 Nl Letter number
3122 No Other number
3123
3124 P Punctuation
3125 Pc Connector punctuation
3126 Pd Dash punctuation
3127 Pe Close punctuation
3128 Pf Final punctuation
3129 Pi Initial punctuation
3130 Po Other punctuation
3131 Ps Open punctuation
3132
3133 S Symbol
3134 Sc Currency symbol
3135 Sk Modifier symbol
3136 Sm Mathematical symbol
3137 So Other symbol
3138
3139 Z Separator
3140 Zl Line separator
3141 Zp Paragraph separator
3142 Zs Space separator
3143
3144 The special property L& is also supported: it matches a character that
3145 has the Lu, Ll, or Lt property, in other words, a letter that is not
3146 classified as a modifier or "other".
3147
3148 The long synonyms for these properties that Perl supports (such as
3149 \p{Letter}) are not supported by PCRE, nor is it permitted to prefix
3150 any of these properties with "Is".
3151
3152 No character that is in the Unicode table has the Cn (unassigned) prop-
3153 erty. Instead, this property is assumed for any code point that is not
3154 in the Unicode table.
3155
3156 Specifying caseless matching does not affect these escape sequences.
3157 For example, \p{Lu} always matches only upper case letters.
3158
3159 The \X escape matches any number of Unicode characters that form an
3160 extended Unicode sequence. \X is equivalent to
3161
3162 (?>\PM\pM*)
3163
3164 That is, it matches a character without the "mark" property, followed
3165 by zero or more characters with the "mark" property, and treats the
3166 sequence as an atomic group (see below). Characters with the "mark"
3167 property are typically accents that affect the preceding character.
3168 None of them have codepoints less than 256, so in non-UTF-8 mode \X
3169 matches any one character.
3170
3171 Matching characters by Unicode property is not fast, because PCRE has
3172 to search a structure that contains data for over fifteen thousand
3173 characters. That is why the traditional escape sequences such as \d and
3174 \w do not use Unicode properties in PCRE.
3175
3176 Resetting the match start
3177
3178 The escape sequence \K, which is a Perl 5.10 feature, causes any previ-
3179 ously matched characters not to be included in the final matched
3180 sequence. For example, the pattern:
3181
3182 foo\Kbar
3183
3184 matches "foobar", but reports that it has matched "bar". This feature
3185 is similar to a lookbehind assertion (described below). However, in
3186 this case, the part of the subject before the real match does not have
3187 to be of fixed length, as lookbehind assertions do. The use of \K does
3188 not interfere with the setting of captured substrings. For example,
3189 when the pattern
3190
3191 (foo)\Kbar
3192
3193 matches "foobar", the first substring is still set to "foo".
3194
3195 Simple assertions
3196
3197 The final use of backslash is for certain simple assertions. An asser-
3198 tion specifies a condition that has to be met at a particular point in
3199 a match, without consuming any characters from the subject string. The
3200 use of subpatterns for more complicated assertions is described below.
3201 The backslashed assertions are:
3202
3203 \b matches at a word boundary
3204 \B matches when not at a word boundary
3205 \A matches at the start of the subject
3206 \Z matches at the end of the subject
3207 also matches before a newline at the end of the subject
3208 \z matches only at the end of the subject
3209 \G matches at the first matching position in the subject
3210
3211 These assertions may not appear in character classes (but note that \b
3212 has a different meaning, namely the backspace character, inside a char-
3213 acter class).
3214
3215 A word boundary is a position in the subject string where the current
3216 character and the previous character do not both match \w or \W (i.e.
3217 one matches \w and the other matches \W), or the start or end of the
3218 string if the first or last character matches \w, respectively.
3219
3220 The \A, \Z, and \z assertions differ from the traditional circumflex
3221 and dollar (described in the next section) in that they only ever match
3222 at the very start and end of the subject string, whatever options are
3223 set. Thus, they are independent of multiline mode. These three asser-
3224 tions are not affected by the PCRE_NOTBOL or PCRE_NOTEOL options, which
3225 affect only the behaviour of the circumflex and dollar metacharacters.
3226 However, if the startoffset argument of pcre_exec() is non-zero, indi-
3227 cating that matching is to start at a point other than the beginning of
3228 the subject, \A can never match. The difference between \Z and \z is
3229 that \Z matches before a newline at the end of the string as well as at
3230 the very end, whereas \z matches only at the end.
3231
3232 The \G assertion is true only when the current matching position is at
3233 the start point of the match, as specified by the startoffset argument
3234 of pcre_exec(). It differs from \A when the value of startoffset is
3235 non-zero. By calling pcre_exec() multiple times with appropriate argu-
3236 ments, you can mimic Perl's /g option, and it is in this kind of imple-
3237 mentation where \G can be useful.
3238
3239 Note, however, that PCRE's interpretation of \G, as the start of the
3240 current match, is subtly different from Perl's, which defines it as the
3241 end of the previous match. In Perl, these can be different when the
3242 previously matched string was empty. Because PCRE does just one match
3243 at a time, it cannot reproduce this behaviour.
3244
3245 If all the alternatives of a pattern begin with \G, the expression is
3246 anchored to the starting match position, and the "anchored" flag is set
3247 in the compiled regular expression.
3248
3249
3250 CIRCUMFLEX AND DOLLAR
3251
3252 Outside a character class, in the default matching mode, the circumflex
3253 character is an assertion that is true only if the current matching
3254 point is at the start of the subject string. If the startoffset argu-
3255 ment of pcre_exec() is non-zero, circumflex can never match if the
3256 PCRE_MULTILINE option is unset. Inside a character class, circumflex
3257 has an entirely different meaning (see below).
3258
3259 Circumflex need not be the first character of the pattern if a number
3260 of alternatives are involved, but it should be the first thing in each
3261 alternative in which it appears if the pattern is ever to match that
3262 branch. If all possible alternatives start with a circumflex, that is,
3263 if the pattern is constrained to match only at the start of the sub-
3264 ject, it is said to be an "anchored" pattern. (There are also other
3265 constructs that can cause a pattern to be anchored.)
3266
3267 A dollar character is an assertion that is true only if the current
3268 matching point is at the end of the subject string, or immediately
3269 before a newline at the end of the string (by default). Dollar need not
3270 be the last character of the pattern if a number of alternatives are
3271 involved, but it should be the last item in any branch in which it
3272 appears. Dollar has no special meaning in a character class.
3273
3274 The meaning of dollar can be changed so that it matches only at the
3275 very end of the string, by setting the PCRE_DOLLAR_ENDONLY option at
3276 compile time. This does not affect the \Z assertion.
3277
3278 The meanings of the circumflex and dollar characters are changed if the
3279 PCRE_MULTILINE option is set. When this is the case, a circumflex
3280 matches immediately after internal newlines as well as at the start of
3281 the subject string. It does not match after a newline that ends the
3282 string. A dollar matches before any newlines in the string, as well as
3283 at the very end, when PCRE_MULTILINE is set. When newline is specified
3284 as the two-character sequence CRLF, isolated CR and LF characters do
3285 not indicate newlines.
3286
3287 For example, the pattern /^abc$/ matches the subject string "def\nabc"
3288 (where \n represents a newline) in multiline mode, but not otherwise.
3289 Consequently, patterns that are anchored in single line mode because
3290 all branches start with ^ are not anchored in multiline mode, and a
3291 match for circumflex is possible when the startoffset argument of
3292 pcre_exec() is non-zero. The PCRE_DOLLAR_ENDONLY option is ignored if
3293 PCRE_MULTILINE is set.
3294
3295 Note that the sequences \A, \Z, and \z can be used to match the start
3296 and end of the subject in both modes, and if all branches of a pattern
3297 start with \A it is always anchored, whether or not PCRE_MULTILINE is
3298 set.
3299
3300
3301 FULL STOP (PERIOD, DOT)
3302
3303 Outside a character class, a dot in the pattern matches any one charac-
3304 ter in the subject string except (by default) a character that signi-
3305 fies the end of a line. In UTF-8 mode, the matched character may be
3306 more than one byte long.
3307
3308 When a line ending is defined as a single character, dot never matches
3309 that character; when the two-character sequence CRLF is used, dot does
3310 not match CR if it is immediately followed by LF, but otherwise it
3311 matches all characters (including isolated CRs and LFs). When any Uni-
3312 code line endings are being recognized, dot does not match CR or LF or
3313 any of the other line ending characters.
3314
3315 The behaviour of dot with regard to newlines can be changed. If the
3316 PCRE_DOTALL option is set, a dot matches any one character, without
3317 exception. If the two-character sequence CRLF is present in the subject
3318 string, it takes two dots to match it.
3319
3320 The handling of dot is entirely independent of the handling of circum-
3321 flex and dollar, the only relationship being that they both involve
3322 newlines. Dot has no special meaning in a character class.
3323
3324
3325 MATCHING A SINGLE BYTE
3326
3327 Outside a character class, the escape sequence \C matches any one byte,
3328 both in and out of UTF-8 mode. Unlike a dot, it always matches any
3329 line-ending characters. The feature is provided in Perl in order to
3330 match individual bytes in UTF-8 mode. Because it breaks up UTF-8 char-
3331 acters into individual bytes, what remains in the string may be a mal-
3332 formed UTF-8 string. For this reason, the \C escape sequence is best
3333 avoided.
3334
3335 PCRE does not allow \C to appear in lookbehind assertions (described
3336 below), because in UTF-8 mode this would make it impossible to calcu-
3337 late the length of the lookbehind.
3338
3339
3340 SQUARE BRACKETS AND CHARACTER CLASSES
3341
3342 An opening square bracket introduces a character class, terminated by a
3343 closing square bracket. A closing square bracket on its own is not spe-
3344 cial. If a closing square bracket is required as a member of the class,
3345 it should be the first data character in the class (after an initial
3346 circumflex, if present) or escaped with a backslash.
3347
3348 A character class matches a single character in the subject. In UTF-8
3349 mode, the character may occupy more than one byte. A matched character
3350 must be in the set of characters defined by the class, unless the first
3351 character in the class definition is a circumflex, in which case the
3352 subject character must not be in the set defined by the class. If a
3353 circumflex is actually required as a member of the class, ensure it is
3354 not the first character, or escape it with a backslash.
3355
3356 For example, the character class [aeiou] matches any lower case vowel,
3357 while [^aeiou] matches any character that is not a lower case vowel.
3358 Note that a circumflex is just a convenient notation for specifying the
3359 characters that are in the class by enumerating those that are not. A
3360 class that starts with a circumflex is not an assertion: it still con-
3361 sumes a character from the subject string, and therefore it fails if
3362 the current pointer is at the end of the string.
3363
3364 In UTF-8 mode, characters with values greater than 255 can be included
3365 in a class as a literal string of bytes, or by using the \x{ escaping
3366 mechanism.
3367
3368 When caseless matching is set, any letters in a class represent both
3369 their upper case and lower case versions, so for example, a caseless
3370 [aeiou] matches "A" as well as "a", and a caseless [^aeiou] does not
3371 match "A", whereas a caseful version would. In UTF-8 mode, PCRE always
3372 understands the concept of case for characters whose values are less
3373 than 128, so caseless matching is always possible. For characters with
3374 higher values, the concept of case is supported if PCRE is compiled
3375 with Unicode property support, but not otherwise. If you want to use
3376 caseless matching for characters 128 and above, you must ensure that
3377 PCRE is compiled with Unicode property support as well as with UTF-8
3378 support.
3379
3380 Characters that might indicate line breaks are never treated in any
3381 special way when matching character classes, whatever line-ending
3382 sequence is in use, and whatever setting of the PCRE_DOTALL and
3383 PCRE_MULTILINE options is used. A class such as [^a] always matches one
3384 of these characters.
3385
3386 The minus (hyphen) character can be used to specify a range of charac-
3387 ters in a character class. For example, [d-m] matches any letter
3388 between d and m, inclusive. If a minus character is required in a
3389 class, it must be escaped with a backslash or appear in a position
3390 where it cannot be interpreted as indicating a range, typically as the
3391 first or last character in the class.
3392
3393 It is not possible to have the literal character "]" as the end charac-
3394 ter of a range. A pattern such as [W-]46] is interpreted as a class of
3395 two characters ("W" and "-") followed by a literal string "46]", so it
3396 would match "W46]" or "-46]". However, if the "]" is escaped with a
3397 backslash it is interpreted as the end of range, so [W-\]46] is inter-
3398 preted as a class containing a range followed by two other characters.
3399 The octal or hexadecimal representation of "]" can also be used to end
3400 a range.
3401
3402 Ranges operate in the collating sequence of character values. They can
3403 also be used for characters specified numerically, for example
3404 [\000-\037]. In UTF-8 mode, ranges can include characters whose values
3405 are greater than 255, for example [\x{100}-\x{2ff}].
3406
3407 If a range that includes letters is used when caseless matching is set,
3408 it matches the letters in either case. For example, [W-c] is equivalent
3409 to [][\\^_`wxyzabc], matched caselessly, and in non-UTF-8 mode, if
3410 character tables for a French locale are in use, [\xc8-\xcb] matches
3411 accented E characters in both cases. In UTF-8 mode, PCRE supports the
3412 concept of case for characters with values greater than 128 only when
3413 it is compiled with Unicode property support.
3414
3415 The character types \d, \D, \p, \P, \s, \S, \w, and \W may also appear
3416 in a character class, and add the characters that they match to the
3417 class. For example, [\dABCDEF] matches any hexadecimal digit. A circum-
3418 flex can conveniently be used with the upper case character types to
3419 specify a more restricted set of characters than the matching lower
3420 case type. For example, the class [^\W_] matches any letter or digit,
3421 but not underscore.
3422
3423 The only metacharacters that are recognized in character classes are
3424 backslash, hyphen (only where it can be interpreted as specifying a
3425 range), circumflex (only at the start), opening square bracket (only
3426 when it can be interpreted as introducing a POSIX class name - see the
3427 next section), and the terminating closing square bracket. However,
3428 escaping other non-alphanumeric characters does no harm.
3429
3430
3431 POSIX CHARACTER CLASSES
3432
3433 Perl supports the POSIX notation for character classes. This uses names
3434 enclosed by [: and :] within the enclosing square brackets. PCRE also
3435 supports this notation. For example,
3436
3437 [01[:alpha:]%]
3438
3439 matches "0", "1", any alphabetic character, or "%". The supported class
3440 names are
3441
3442 alnum letters and digits
3443 alpha letters
3444 ascii character codes 0 - 127
3445 blank space or tab only
3446 cntrl control characters
3447 digit decimal digits (same as \d)
3448 graph printing characters, excluding space
3449 lower lower case letters
3450 print printing characters, including space
3451 punct printing characters, excluding letters and digits
3452 space white space (not quite the same as \s)
3453 upper upper case letters
3454 word "word" characters (same as \w)
3455 xdigit hexadecimal digits
3456
3457 The "space" characters are HT (9), LF (10), VT (11), FF (12), CR (13),
3458 and space (32). Notice that this list includes the VT character (code
3459 11). This makes "space" different to \s, which does not include VT (for
3460 Perl compatibility).
3461
3462 The name "word" is a Perl extension, and "blank" is a GNU extension
3463 from Perl 5.8. Another Perl extension is negation, which is indicated
3464 by a ^ character after the colon. For example,
3465
3466 [12[:^digit:]]
3467
3468 matches "1", "2", or any non-digit. PCRE (and Perl) also recognize the
3469 POSIX syntax [.ch.] and [=ch=] where "ch" is a "collating element", but
3470 these are not supported, and an error is given if they are encountered.
3471
3472 In UTF-8 mode, characters with values greater than 128 do not match any
3473 of the POSIX character classes.
3474
3475
3476 VERTICAL BAR
3477
3478 Vertical bar characters are used to separate alternative patterns. For
3479 example, the pattern
3480
3481 gilbert|sullivan
3482
3483 matches either "gilbert" or "sullivan". Any number of alternatives may
3484 appear, and an empty alternative is permitted (matching the empty
3485 string). The matching process tries each alternative in turn, from left
3486 to right, and the first one that succeeds is used. If the alternatives
3487 are within a subpattern (defined below), "succeeds" means matching the
3488 rest of the main pattern as well as the alternative in the subpattern.
3489
3490
3491 INTERNAL OPTION SETTING
3492
3493 The settings of the PCRE_CASELESS, PCRE_MULTILINE, PCRE_DOTALL, and
3494 PCRE_EXTENDED options can be changed from within the pattern by a
3495 sequence of Perl option letters enclosed between "(?" and ")". The
3496 option letters are
3497
3498 i for PCRE_CASELESS
3499 m for PCRE_MULTILINE
3500 s for PCRE_DOTALL
3501 x for PCRE_EXTENDED
3502
3503 For example, (?im) sets caseless, multiline matching. It is also possi-
3504 ble to unset these options by preceding the letter with a hyphen, and a
3505 combined setting and unsetting such as (?im-sx), which sets PCRE_CASE-
3506 LESS and PCRE_MULTILINE while unsetting PCRE_DOTALL and PCRE_EXTENDED,
3507 is also permitted. If a letter appears both before and after the
3508 hyphen, the option is unset.
3509
3510 When an option change occurs at top level (that is, not inside subpat-
3511 tern parentheses), the change applies to the remainder of the pattern
3512 that follows. If the change is placed right at the start of a pattern,
3513 PCRE extracts it into the global options (and it will therefore show up
3514 in data extracted by the pcre_fullinfo() function).
3515
3516 An option change within a subpattern (see below for a description of
3517 subpatterns) affects only that part of the current pattern that follows
3518 it, so
3519
3520 (a(?i)b)c
3521
3522 matches abc and aBc and no other strings (assuming PCRE_CASELESS is not
3523 used). By this means, options can be made to have different settings
3524 in different parts of the pattern. Any changes made in one alternative
3525 do carry on into subsequent branches within the same subpattern. For
3526 example,
3527
3528 (a(?i)b|c)
3529
3530 matches "ab", "aB", "c", and "C", even though when matching "C" the
3531 first branch is abandoned before the option setting. This is because
3532 the effects of option settings happen at compile time. There would be
3533 some very weird behaviour otherwise.
3534
3535 The PCRE-specific options PCRE_DUPNAMES, PCRE_UNGREEDY, and PCRE_EXTRA
3536 can be changed in the same way as the Perl-compatible options by using
3537 the characters J, U and X respectively.
3538
3539
3540 SUBPATTERNS
3541
3542 Subpatterns are delimited by parentheses (round brackets), which can be
3543 nested. Turning part of a pattern into a subpattern does two things:
3544
3545 1. It localizes a set of alternatives. For example, the pattern
3546
3547 cat(aract|erpillar|)
3548
3549 matches one of the words "cat", "cataract", or "caterpillar". Without
3550 the parentheses, it would match "cataract", "erpillar" or an empty
3551 string.
3552
3553 2. It sets up the subpattern as a capturing subpattern. This means
3554 that, when the whole pattern matches, that portion of the subject
3555 string that matched the subpattern is passed back to the caller via the
3556 ovector argument of pcre_exec(). Opening parentheses are counted from
3557 left to right (starting from 1) to obtain numbers for the capturing
3558 subpatterns.
3559
3560 For example, if the string "the red king" is matched against the pat-
3561 tern
3562
3563 the ((red|white) (king|queen))
3564
3565 the captured substrings are "red king", "red", and "king", and are num-
3566 bered 1, 2, and 3, respectively.
3567
3568 The fact that plain parentheses fulfil two functions is not always
3569 helpful. There are often times when a grouping subpattern is required
3570 without a capturing requirement. If an opening parenthesis is followed
3571 by a question mark and a colon, the subpattern does not do any captur-
3572 ing, and is not counted when computing the number of any subsequent
3573 capturing subpatterns. For example, if the string "the white queen" is
3574 matched against the pattern
3575
3576 the ((?:red|white) (king|queen))
3577
3578 the captured substrings are "white queen" and "queen", and are numbered
3579 1 and 2. The maximum number of capturing subpatterns is 65535.
3580
3581 As a convenient shorthand, if any option settings are required at the
3582 start of a non-capturing subpattern, the option letters may appear
3583 between the "?" and the ":". Thus the two patterns
3584
3585 (?i:saturday|sunday)
3586 (?:(?i)saturday|sunday)
3587
3588 match exactly the same set of strings. Because alternative branches are
3589 tried from left to right, and options are not reset until the end of
3590 the subpattern is reached, an option setting in one branch does affect
3591 subsequent branches, so the above patterns match "SUNDAY" as well as
3592 "Saturday".
3593
3594
3595 DUPLICATE SUBPATTERN NUMBERS
3596
3597 Perl 5.10 introduced a feature whereby each alternative in a subpattern
3598 uses the same numbers for its capturing parentheses. Such a subpattern
3599 starts with (?| and is itself a non-capturing subpattern. For example,
3600 consider this pattern:
3601
3602 (?|(Sat)ur|(Sun))day
3603
3604 Because the two alternatives are inside a (?| group, both sets of cap-
3605 turing parentheses are numbered one. Thus, when the pattern matches,
3606 you can look at captured substring number one, whichever alternative
3607 matched. This construct is useful when you want to capture part, but
3608 not all, of one of a number of alternatives. Inside a (?| group, paren-
3609 theses are numbered as usual, but the number is reset at the start of
3610 each branch. The numbers of any capturing buffers that follow the sub-
3611 pattern start after the highest number used in any branch. The follow-
3612 ing example is taken from the Perl documentation. The numbers under-
3613 neath show in which buffer the captured content will be stored.
3614
3615 # before ---------------branch-reset----------- after
3616 / ( a ) (?| x ( y ) z | (p (q) r) | (t) u (v) ) ( z ) /x
3617 # 1 2 2 3 2 3 4
3618
3619 A backreference or a recursive call to a numbered subpattern always
3620 refers to the first one in the pattern with the given number.
3621
3622 An alternative approach to using this "branch reset" feature is to use
3623 duplicate named subpatterns, as described in the next section.
3624
3625
3626 NAMED SUBPATTERNS
3627
3628 Identifying capturing parentheses by number is simple, but it can be
3629 very hard to keep track of the numbers in complicated regular expres-
3630 sions. Furthermore, if an expression is modified, the numbers may
3631 change. To help with this difficulty, PCRE supports the naming of sub-
3632 patterns. This feature was not added to Perl until release 5.10. Python
3633 had the feature earlier, and PCRE introduced it at release 4.0, using
3634 the Python syntax. PCRE now supports both the Perl and the Python syn-
3635 tax.
3636
3637 In PCRE, a subpattern can be named in one of three ways: (?<name>...)
3638 or (?'name'...) as in Perl, or (?P<name>...) as in Python. References
3639 to capturing parentheses from other parts of the pattern, such as back-
3640 references, recursion, and conditions, can be made by name as well as
3641 by number.
3642
3643 Names consist of up to 32 alphanumeric characters and underscores.
3644 Named capturing parentheses are still allocated numbers as well as
3645 names, exactly as if the names were not present. The PCRE API provides
3646 function calls for extracting the name-to-number translation table from
3647 a compiled pattern. There is also a convenience function for extracting
3648 a captured substring by name.
3649
3650 By default, a name must be unique within a pattern, but it is possible
3651 to relax this constraint by setting the PCRE_DUPNAMES option at compile
3652 time. This can be useful for patterns where only one instance of the
3653 named parentheses can match. Suppose you want to match the name of a
3654 weekday, either as a 3-letter abbreviation or as the full name, and in
3655 both cases you want to extract the abbreviation. This pattern (ignoring
3656 the line breaks) does the job:
3657
3658 (?<DN>Mon|Fri|Sun)(?:day)?|
3659 (?<DN>Tue)(?:sday)?|
3660 (?<DN>Wed)(?:nesday)?|
3661 (?<DN>Thu)(?:rsday)?|
3662 (?<DN>Sat)(?:urday)?
3663
3664 There are five capturing substrings, but only one is ever set after a
3665 match. (An alternative way of solving this problem is to use a "branch
3666 reset" subpattern, as described in the previous section.)
3667
3668 The convenience function for extracting the data by name returns the
3669 substring for the first (and in this example, the only) subpattern of
3670 that name that matched. This saves searching to find which numbered
3671 subpattern it was. If you make a reference to a non-unique named sub-
3672 pattern from elsewhere in the pattern, the one that corresponds to the
3673 lowest number is used. For further details of the interfaces for han-
3674 dling named subpatterns, see the pcreapi documentation.
3675
3676
3677 REPETITION
3678
3679 Repetition is specified by quantifiers, which can follow any of the
3680 following items:
3681
3682 a literal data character
3683 the dot metacharacter
3684 the \C escape sequence
3685 the \X escape sequence (in UTF-8 mode with Unicode properties)
3686 the \R escape sequence
3687 an escape such as \d that matches a single character
3688 a character class
3689 a back reference (see next section)
3690 a parenthesized subpattern (unless it is an assertion)
3691
3692 The general repetition quantifier specifies a minimum and maximum num-
3693 ber of permitted matches, by giving the two numbers in curly brackets
3694 (braces), separated by a comma. The numbers must be less than 65536,
3695 and the first must be less than or equal to the second. For example:
3696
3697 z{2,4}
3698
3699 matches "zz", "zzz", or "zzzz". A closing brace on its own is not a
3700 special character. If the second number is omitted, but the comma is
3701 present, there is no upper limit; if the second number and the comma
3702 are both omitted, the quantifier specifies an exact number of required
3703 matches. Thus
3704
3705 [aeiou]{3,}
3706
3707 matches at least 3 successive vowels, but may match many more, while
3708
3709 \d{8}
3710
3711 matches exactly 8 digits. An opening curly bracket that appears in a
3712 position where a quantifier is not allowed, or one that does not match
3713 the syntax of a quantifier, is taken as a literal character. For exam-
3714 ple, {,6} is not a quantifier, but a literal string of four characters.
3715
3716 In UTF-8 mode, quantifiers apply to UTF-8 characters rather than to
3717 individual bytes. Thus, for example, \x{100}{2} matches two UTF-8 char-
3718 acters, each of which is represented by a two-byte sequence. Similarly,
3719 when Unicode property support is available, \X{3} matches three Unicode
3720 extended sequences, each of which may be several bytes long (and they
3721 may be of different lengths).
3722
3723 The quantifier {0} is permitted, causing the expression to behave as if
3724 the previous item and the quantifier were not present.
3725
3726 For convenience, the three most common quantifiers have single-charac-
3727 ter abbreviations:
3728
3729 * is equivalent to {0,}
3730 + is equivalent to {1,}
3731 ? is equivalent to {0,1}
3732
3733 It is possible to construct infinite loops by following a subpattern
3734 that can match no characters with a quantifier that has no upper limit,
3735 for example:
3736
3737 (a?)*
3738
3739 Earlier versions of Perl and PCRE used to give an error at compile time
3740 for such patterns. However, because there are cases where this can be
3741 useful, such patterns are now accepted, but if any repetition of the
3742 subpattern does in fact match no characters, the loop is forcibly bro-
3743 ken.
3744
3745 By default, the quantifiers are "greedy", that is, they match as much
3746 as possible (up to the maximum number of permitted times), without
3747 causing the rest of the pattern to fail. The classic example of where
3748 this gives problems is in trying to match comments in C programs. These
3749 appear between /* and */ and within the comment, individual * and /
3750 characters may appear. An attempt to match C comments by applying the
3751 pattern
3752
3753 /\*.*\*/
3754
3755 to the string
3756
3757 /* first comment */ not comment /* second comment */
3758
3759 fails, because it matches the entire string owing to the greediness of
3760 the .* item.
3761
3762 However, if a quantifier is followed by a question mark, it ceases to
3763 be greedy, and instead matches the minimum number of times possible, so
3764 the pattern
3765
3766 /\*.*?\*/
3767
3768 does the right thing with the C comments. The meaning of the various
3769 quantifiers is not otherwise changed, just the preferred number of
3770 matches. Do not confuse this use of question mark with its use as a
3771 quantifier in its own right. Because it has two uses, it can sometimes
3772 appear doubled, as in
3773
3774 \d??\d
3775
3776 which matches one digit by preference, but can match two if that is the
3777 only way the rest of the pattern matches.
3778
3779 If the PCRE_UNGREEDY option is set (an option that is not available in
3780 Perl), the quantifiers are not greedy by default, but individual ones
3781 can be made greedy by following them with a question mark. In other
3782 words, it inverts the default behaviour.
3783
3784 When a parenthesized subpattern is quantified with a minimum repeat
3785 count that is greater than 1 or with a limited maximum, more memory is
3786 required for the compiled pattern, in proportion to the size of the
3787 minimum or maximum.
3788
3789 If a pattern starts with .* or .{0,} and the PCRE_DOTALL option (equiv-
3790 alent to Perl's /s) is set, thus allowing the dot to match newlines,
3791 the pattern is implicitly anchored, because whatever follows will be
3792 tried against every character position in the subject string, so there
3793 is no point in retrying the overall match at any position after the
3794 first. PCRE normally treats such a pattern as though it were preceded
3795 by \A.
3796
3797 In cases where it is known that the subject string contains no new-
3798 lines, it is worth setting PCRE_DOTALL in order to obtain this opti-
3799 mization, or alternatively using ^ to indicate anchoring explicitly.
3800
3801 However, there is one situation where the optimization cannot be used.
3802 When .* is inside capturing parentheses that are the subject of a
3803 backreference elsewhere in the pattern, a match at the start may fail
3804 where a later one succeeds. Consider, for example:
3805
3806 (.*)abc\1
3807
3808 If the subject is "xyz123abc123" the match point is the fourth charac-
3809 ter. For this reason, such a pattern is not implicitly anchored.
3810
3811 When a capturing subpattern is repeated, the value captured is the sub-
3812 string that matched the final iteration. For example, after
3813
3814 (tweedle[dume]{3}\s*)+
3815
3816 has matched "tweedledum tweedledee" the value of the captured substring
3817 is "tweedledee". However, if there are nested capturing subpatterns,
3818 the corresponding captured values may have been set in previous itera-
3819 tions. For example, after
3820
3821 /(a|(b))+/
3822
3823 matches "aba" the value of the second captured substring is "b".
3824
3825
3826 ATOMIC GROUPING AND POSSESSIVE QUANTIFIERS
3827
3828 With both maximizing ("greedy") and minimizing ("ungreedy" or "lazy")
3829 repetition, failure of what follows normally causes the repeated item
3830 to be re-evaluated to see if a different number of repeats allows the
3831 rest of the pattern to match. Sometimes it is useful to prevent this,
3832 either to change the nature of the match, or to cause it fail earlier
3833 than it otherwise might, when the author of the pattern knows there is
3834 no point in carrying on.
3835
3836 Consider, for example, the pattern \d+foo when applied to the subject
3837 line
3838
3839 123456bar
3840
3841 After matching all 6 digits and then failing to match "foo", the normal
3842 action of the matcher is to try again with only 5 digits matching the
3843 \d+ item, and then with 4, and so on, before ultimately failing.
3844 "Atomic grouping" (a term taken from Jeffrey Friedl's book) provides
3845 the means for specifying that once a subpattern has matched, it is not
3846 to be re-evaluated in this way.
3847
3848 If we use atomic grouping for the previous example, the matcher gives
3849 up immediately on failing to match "foo" the first time. The notation
3850 is a kind of special parenthesis, starting with (?> as in this example:
3851
3852 (?>\d+)foo
3853
3854 This kind of parenthesis "locks up" the part of the pattern it con-
3855 tains once it has matched, and a failure further into the pattern is
3856 prevented from backtracking into it. Backtracking past it to previous
3857 items, however, works as normal.
3858
3859 An alternative description is that a subpattern of this type matches
3860 the string of characters that an identical standalone pattern would
3861 match, if anchored at the current point in the subject string.
3862
3863 Atomic grouping subpatterns are not capturing subpatterns. Simple cases
3864 such as the above example can be thought of as a maximizing repeat that
3865 must swallow everything it can. So, while both \d+ and \d+? are pre-
3866 pared to adjust the number of digits they match in order to make the
3867 rest of the pattern match, (?>\d+) can only match an entire sequence of
3868 digits.
3869
3870 Atomic groups in general can of course contain arbitrarily complicated
3871 subpatterns, and can be nested. However, when the subpattern for an
3872 atomic group is just a single repeated item, as in the example above, a
3873 simpler notation, called a "possessive quantifier" can be used. This
3874 consists of an additional + character following a quantifier. Using
3875 this notation, the previous example can be rewritten as
3876
3877 \d++foo
3878
3879 Possessive quantifiers are always greedy; the setting of the
3880 PCRE_UNGREEDY option is ignored. They are a convenient notation for the
3881 simpler forms of atomic group. However, there is no difference in the
3882 meaning of a possessive quantifier and the equivalent atomic group,
3883 though there may be a performance difference; possessive quantifiers
3884 should be slightly faster.
3885
3886 The possessive quantifier syntax is an extension to the Perl 5.8 syn-
3887 tax. Jeffrey Friedl originated the idea (and the name) in the first
3888 edition of his book. Mike McCloskey liked it, so implemented it when he
3889 built Sun's Java package, and PCRE copied it from there. It ultimately
3890 found its way into Perl at release 5.10.
3891
3892 PCRE has an optimization that automatically "possessifies" certain sim-
3893 ple pattern constructs. For example, the sequence A+B is treated as
3894 A++B because there is no point in backtracking into a sequence of A's
3895 when B must follow.
3896
3897 When a pattern contains an unlimited repeat inside a subpattern that
3898 can itself be repeated an unlimited number of times, the use of an
3899 atomic group is the only way to avoid some failing matches taking a
3900 very long time indeed. The pattern
3901
3902 (\D+|<\d+>)*[!?]
3903
3904 matches an unlimited number of substrings that either consist of non-
3905 digits, or digits enclosed in <>, followed by either ! or ?. When it
3906 matches, it runs quickly. However, if it is applied to
3907
3908 aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
3909
3910 it takes a long time before reporting failure. This is because the
3911 string can be divided between the internal \D+ repeat and the external
3912 * repeat in a large number of ways, and all have to be tried. (The
3913 example uses [!?] rather than a single character at the end, because
3914 both PCRE and Perl have an optimization that allows for fast failure
3915 when a single character is used. They remember the last single charac-
3916 ter that is required for a match, and fail early if it is not present
3917 in the string.) If the pattern is changed so that it uses an atomic
3918 group, like this:
3919
3920 ((?>\D+)|<\d+>)*[!?]
3921
3922 sequences of non-digits cannot be broken, and failure happens quickly.
3923
3924
3925 BACK REFERENCES
3926
3927 Outside a character class, a backslash followed by a digit greater than
3928 0 (and possibly further digits) is a back reference to a capturing sub-
3929 pattern earlier (that is, to its left) in the pattern, provided there
3930 have been that many previous capturing left parentheses.
3931
3932 However, if the decimal number following the backslash is less than 10,
3933 it is always taken as a back reference, and causes an error only if
3934 there are not that many capturing left parentheses in the entire pat-
3935 tern. In other words, the parentheses that are referenced need not be
3936 to the left of the reference for numbers less than 10. A "forward back
3937 reference" of this type can make sense when a repetition is involved
3938 and the subpattern to the right has participated in an earlier itera-
3939 tion.
3940
3941 It is not possible to have a numerical "forward back reference" to a
3942 subpattern whose number is 10 or more using this syntax because a
3943 sequence such as \50 is interpreted as a character defined in octal.
3944 See the subsection entitled "Non-printing characters" above for further
3945 details of the handling of digits following a backslash. There is no
3946 such problem when named parentheses are used. A back reference to any
3947 subpattern is possible using named parentheses (see below).
3948
3949 Another way of avoiding the ambiguity inherent in the use of digits
3950 following a backslash is to use the \g escape sequence, which is a fea-
3951 ture introduced in Perl 5.10. This escape must be followed by a posi-
3952 tive or a negative number, optionally enclosed in braces. These exam-
3953 ples are all identical:
3954
3955 (ring), \1
3956 (ring), \g1
3957 (ring), \g{1}
3958
3959 A positive number specifies an absolute reference without the ambiguity
3960 that is present in the older syntax. It is also useful when literal
3961 digits follow the reference. A negative number is a relative reference.
3962 Consider this example:
3963
3964 (abc(def)ghi)\g{-1}
3965
3966 The sequence \g{-1} is a reference to the most recently started captur-
3967 ing subpattern before \g, that is, is it equivalent to \2. Similarly,
3968 \g{-2} would be equivalent to \1. The use of relative references can be
3969 helpful in long patterns, and also in patterns that are created by
3970 joining together fragments that contain references within themselves.
3971
3972 A back reference matches whatever actually matched the capturing sub-
3973 pattern in the current subject string, rather than anything matching
3974 the subpattern itself (see "Subpatterns as subroutines" below for a way
3975 of doing that). So the pattern
3976
3977 (sens|respons)e and \1ibility
3978
3979 matches "sense and sensibility" and "response and responsibility", but
3980 not "sense and responsibility". If caseful matching is in force at the
3981 time of the back reference, the case of letters is relevant. For exam-
3982 ple,
3983
3984 ((?i)rah)\s+\1
3985
3986 matches "rah rah" and "RAH RAH", but not "RAH rah", even though the
3987 original capturing subpattern is matched caselessly.
3988
3989 There are several different ways of writing back references to named
3990 subpatterns. The .NET syntax \k{name} and the Perl syntax \k<name> or
3991 \k'name' are supported, as is the Python syntax (?P=name). Perl 5.10's
3992 unified back reference syntax, in which \g can be used for both numeric
3993 and named references, is also supported. We could rewrite the above
3994 example in any of the following ways:
3995
3996 (?<p1>(?i)rah)\s+\k<p1>
3997 (?'p1'(?i)rah)\s+\k{p1}
3998 (?P<p1>(?i)rah)\s+(?P=p1)
3999 (?<p1>(?i)rah)\s+\g{p1}
4000
4001 A subpattern that is referenced by name may appear in the pattern
4002 before or after the reference.
4003
4004 There may be more than one back reference to the same subpattern. If a
4005 subpattern has not actually been used in a particular match, any back
4006 references to it always fail. For example, the pattern
4007
4008 (a|(bc))\2
4009
4010 always fails if it starts to match "a" rather than "bc". Because there
4011 may be many capturing parentheses in a pattern, all digits following
4012 the backslash are taken as part of a potential back reference number.
4013 If the pattern continues with a digit character, some delimiter must be
4014 used to terminate the back reference. If the PCRE_EXTENDED option is
4015 set, this can be whitespace. Otherwise an empty comment (see "Com-
4016 ments" below) can be used.
4017
4018 A back reference that occurs inside the parentheses to which it refers
4019 fails when the subpattern is first used, so, for example, (a\1) never
4020 matches. However, such references can be useful inside repeated sub-
4021 patterns. For example, the pattern
4022
4023 (a|b\1)+
4024
4025 matches any number of "a"s and also "aba", "ababbaa" etc. At each iter-
4026 ation of the subpattern, the back reference matches the character
4027 string corresponding to the previous iteration. In order for this to
4028 work, the pattern must be such that the first iteration does not need
4029 to match the back reference. This can be done using alternation, as in
4030 the example above, or by a quantifier with a minimum of zero.
4031
4032
4033 ASSERTIONS
4034
4035 An assertion is a test on the characters following or preceding the
4036 current matching point that does not actually consume any characters.
4037 The simple assertions coded as \b, \B, \A, \G, \Z, \z, ^ and $ are
4038 described above.
4039
4040 More complicated assertions are coded as subpatterns. There are two
4041 kinds: those that look ahead of the current position in the subject
4042 string, and those that look behind it. An assertion subpattern is
4043 matched in the normal way, except that it does not cause the current
4044 matching position to be changed.
4045
4046 Assertion subpatterns are not capturing subpatterns, and may not be
4047 repeated, because it makes no sense to assert the same thing several
4048 times. If any kind of assertion contains capturing subpatterns within
4049 it, these are counted for the purposes of numbering the capturing sub-
4050 patterns in the whole pattern. However, substring capturing is carried
4051 out only for positive assertions, because it does not make sense for
4052 negative assertions.
4053
4054 Lookahead assertions
4055
4056 Lookahead assertions start with (?= for positive assertions and (?! for
4057 negative assertions. For example,
4058
4059 \w+(?=;)
4060
4061 matches a word followed by a semicolon, but does not include the semi-
4062 colon in the match, and
4063
4064 foo(?!bar)
4065
4066 matches any occurrence of "foo" that is not followed by "bar". Note
4067 that the apparently similar pattern
4068
4069 (?!foo)bar
4070
4071 does not find an occurrence of "bar" that is preceded by something
4072 other than "foo"; it finds any occurrence of "bar" whatsoever, because
4073 the assertion (?!foo) is always true when the next three characters are
4074 "bar". A lookbehind assertion is needed to achieve the other effect.
4075
4076 If you want to force a matching failure at some point in a pattern, the
4077 most convenient way to do it is with (?!) because an empty string
4078 always matches, so an assertion that requires there not to be an empty
4079 string must always fail.
4080
4081 Lookbehind assertions
4082
4083 Lookbehind assertions start with (?<= for positive assertions and (?<!
4084 for negative assertions. For example,
4085
4086 (?<!foo)bar
4087
4088 does find an occurrence of "bar" that is not preceded by "foo". The
4089 contents of a lookbehind assertion are restricted such that all the
4090 strings it matches must have a fixed length. However, if there are sev-
4091 eral top-level alternatives, they do not all have to have the same
4092 fixed length. Thus
4093
4094 (?<=bullock|donkey)
4095
4096 is permitted, but
4097
4098 (?<!dogs?|cats?)
4099
4100 causes an error at compile time. Branches that match different length
4101 strings are permitted only at the top level of a lookbehind assertion.
4102 This is an extension compared with Perl (at least for 5.8), which
4103 requires all branches to match the same length of string. An assertion
4104 such as
4105
4106 (?<=ab(c|de))
4107
4108 is not permitted, because its single top-level branch can match two
4109 different lengths, but it is acceptable if rewritten to use two top-
4110 level branches:
4111
4112 (?<=abc|abde)
4113
4114 In some cases, the Perl 5.10 escape sequence \K (see above) can be used
4115 instead of a lookbehind assertion; this is not restricted to a fixed-
4116 length.
4117
4118 The implementation of lookbehind assertions is, for each alternative,
4119 to temporarily move the current position back by the fixed length and
4120 then try to match. If there are insufficient characters before the cur-
4121 rent position, the assertion fails.
4122
4123 PCRE does not allow the \C escape (which matches a single byte in UTF-8
4124 mode) to appear in lookbehind assertions, because it makes it impossi-
4125 ble to calculate the length of the lookbehind. The \X and \R escapes,
4126 which can match different numbers of bytes, are also not permitted.
4127
4128 Possessive quantifiers can be used in conjunction with lookbehind
4129 assertions to specify efficient matching at the end of the subject
4130 string. Consider a simple pattern such as
4131
4132 abcd$
4133
4134 when applied to a long string that does not match. Because matching
4135 proceeds from left to right, PCRE will look for each "a" in the subject
4136 and then see if what follows matches the rest of the pattern. If the
4137 pattern is specified as
4138
4139 ^.*abcd$
4140
4141 the initial .* matches the entire string at first, but when this fails
4142 (because there is no following "a"), it backtracks to match all but the
4143 last character, then all but the last two characters, and so on. Once
4144 again the search for "a" covers the entire string, from right to left,
4145 so we are no better off. However, if the pattern is written as
4146
4147 ^.*+(?<=abcd)
4148
4149 there can be no backtracking for the .*+ item; it can match only the
4150 entire string. The subsequent lookbehind assertion does a single test
4151 on the last four characters. If it fails, the match fails immediately.
4152 For long strings, this approach makes a significant difference to the
4153 processing time.
4154
4155 Using multiple assertions
4156
4157 Several assertions (of any sort) may occur in succession. For example,
4158
4159 (?<=\d{3})(?<!999)foo
4160
4161 matches "foo" preceded by three digits that are not "999". Notice that
4162 each of the assertions is applied independently at the same point in
4163 the subject string. First there is a check that the previous three
4164 characters are all digits, and then there is a check that the same
4165 three characters are not "999". This pattern does not match "foo" pre-
4166 ceded by six characters, the first of which are digits and the last
4167 three of which are not "999". For example, it doesn't match "123abc-
4168 foo". A pattern to do that is
4169
4170 (?<=\d{3}...)(?<!999)foo
4171
4172 This time the first assertion looks at the preceding six characters,
4173 checking that the first three are digits, and then the second assertion
4174 checks that the preceding three characters are not "999".
4175
4176 Assertions can be nested in any combination. For example,
4177
4178 (?<=(?<!foo)bar)baz
4179
4180 matches an occurrence of "baz" that is preceded by "bar" which in turn
4181 is not preceded by "foo", while
4182
4183 (?<=\d{3}(?!999)...)foo
4184
4185 is another pattern that matches "foo" preceded by three digits and any
4186 three characters that are not "999".
4187
4188
4189 CONDITIONAL SUBPATTERNS
4190
4191 It is possible to cause the matching process to obey a subpattern con-
4192 ditionally or to choose between two alternative subpatterns, depending
4193 on the result of an assertion, or whether a previous capturing subpat-
4194 tern matched or not. The two possible forms of conditional subpattern
4195 are
4196
4197 (?(condition)yes-pattern)
4198 (?(condition)yes-pattern|no-pattern)
4199
4200 If the condition is satisfied, the yes-pattern is used; otherwise the
4201 no-pattern (if present) is used. If there are more than two alterna-
4202 tives in the subpattern, a compile-time error occurs.
4203
4204 There are four kinds of condition: references to subpatterns, refer-
4205 ences to recursion, a pseudo-condition called DEFINE, and assertions.
4206
4207 Checking for a used subpattern by number
4208
4209 If the text between the parentheses consists of a sequence of digits,
4210 the condition is true if the capturing subpattern of that number has
4211 previously matched. An alternative notation is to precede the digits
4212 with a plus or minus sign. In this case, the subpattern number is rela-
4213 tive rather than absolute. The most recently opened parentheses can be
4214 referenced by (?(-1), the next most recent by (?(-2), and so on. In
4215 looping constructs it can also make sense to refer to subsequent groups
4216 with constructs such as (?(+2).
4217
4218 Consider the following pattern, which contains non-significant white
4219 space to make it more readable (assume the PCRE_EXTENDED option) and to
4220 divide it into three parts for ease of discussion:
4221
4222 ( \( )? [^()]+ (?(1) \) )
4223
4224 The first part matches an optional opening parenthesis, and if that
4225 character is present, sets it as the first captured substring. The sec-
4226 ond part matches one or more characters that are not parentheses. The
4227 third part is a conditional subpattern that tests whether the first set
4228 of parentheses matched or not. If they did, that is, if subject started
4229 with an opening parenthesis, the condition is true, and so the yes-pat-
4230 tern is executed and a closing parenthesis is required. Otherwise,
4231 since no-pattern is not present, the subpattern matches nothing. In
4232 other words, this pattern matches a sequence of non-parentheses,
4233 optionally enclosed in parentheses.
4234
4235 If you were embedding this pattern in a larger one, you could use a
4236 relative reference:
4237
4238 ...other stuff... ( \( )? [^()]+ (?(-1) \) ) ...
4239
4240 This makes the fragment independent of the parentheses in the larger
4241 pattern.
4242
4243 Checking for a used subpattern by name
4244
4245 Perl uses the syntax (?(<name>)...) or (?('name')...) to test for a
4246 used subpattern by name. For compatibility with earlier versions of
4247 PCRE, which had this facility before Perl, the syntax (?(name)...) is
4248 also recognized. However, there is a possible ambiguity with this syn-
4249 tax, because subpattern names may consist entirely of digits. PCRE
4250 looks first for a named subpattern; if it cannot find one and the name
4251 consists entirely of digits, PCRE looks for a subpattern of that num-
4252 ber, which must be greater than zero. Using subpattern names that con-
4253 sist entirely of digits is not recommended.
4254
4255 Rewriting the above example to use a named subpattern gives this:
4256
4257 (?<OPEN> \( )? [^()]+ (?(<OPEN>) \) )
4258
4259
4260 Checking for pattern recursion
4261
4262 If the condition is the string (R), and there is no subpattern with the
4263 name R, the condition is true if a recursive call to the whole pattern
4264 or any subpattern has been made. If digits or a name preceded by amper-
4265 sand follow the letter R, for example:
4266
4267 (?(R3)...) or (?(R&name)...)
4268
4269 the condition is true if the most recent recursion is into the subpat-
4270 tern whose number or name is given. This condition does not check the
4271 entire recursion stack.
4272
4273 At "top level", all these recursion test conditions are false. Recur-
4274 sive patterns are described below.
4275
4276 Defining subpatterns for use by reference only
4277
4278 If the condition is the string (DEFINE), and there is no subpattern
4279 with the name DEFINE, the condition is always false. In this case,
4280 there may be only one alternative in the subpattern. It is always
4281 skipped if control reaches this point in the pattern; the idea of
4282 DEFINE is that it can be used to define "subroutines" that can be ref-
4283 erenced from elsewhere. (The use of "subroutines" is described below.)
4284 For example, a pattern to match an IPv4 address could be written like
4285 this (ignore whitespace and line breaks):
4286
4287 (?(DEFINE) (?<byte> 2[0-4]\d | 25[0-5] | 1\d\d | [1-9]?\d) )
4288 \b (?&byte) (\.(?&byte)){3} \b
4289
4290 The first part of the pattern is a DEFINE group inside which a another
4291 group named "byte" is defined. This matches an individual component of
4292 an IPv4 address (a number less than 256). When matching takes place,
4293 this part of the pattern is skipped because DEFINE acts like a false
4294 condition.
4295
4296 The rest of the pattern uses references to the named group to match the
4297 four dot-separated components of an IPv4 address, insisting on a word
4298 boundary at each end.
4299
4300 Assertion conditions
4301
4302 If the condition is not in any of the above formats, it must be an
4303 assertion. This may be a positive or negative lookahead or lookbehind
4304 assertion. Consider this pattern, again containing non-significant
4305 white space, and with the two alternatives on the second line:
4306
4307 (?(?=[^a-z]*[a-z])
4308 \d{2}-[a-z]{3}-\d{2} | \d{2}-\d{2}-\d{2} )
4309
4310 The condition is a positive lookahead assertion that matches an
4311 optional sequence of non-letters followed by a letter. In other words,
4312 it tests for the presence of at least one letter in the subject. If a
4313 letter is found, the subject is matched against the first alternative;
4314 otherwise it is matched against the second. This pattern matches
4315 strings in one of the two forms dd-aaa-dd or dd-dd-dd, where aaa are
4316 letters and dd are digits.
4317
4318
4319 COMMENTS
4320
4321 The sequence (?# marks the start of a comment that continues up to the
4322 next closing parenthesis. Nested parentheses are not permitted. The
4323 characters that make up a comment play no part in the pattern matching
4324 at all.
4325
4326 If the PCRE_EXTENDED option is set, an unescaped # character outside a
4327 character class introduces a comment that continues to immediately
4328 after the next newline in the pattern.
4329
4330
4331 RECURSIVE PATTERNS
4332
4333 Consider the problem of matching a string in parentheses, allowing for
4334 unlimited nested parentheses. Without the use of recursion, the best
4335 that can be done is to use a pattern that matches up to some fixed
4336 depth of nesting. It is not possible to handle an arbitrary nesting
4337 depth.
4338
4339 For some time, Perl has provided a facility that allows regular expres-
4340 sions to recurse (amongst other things). It does this by interpolating
4341 Perl code in the expression at run time, and the code can refer to the
4342 expression itself. A Perl pattern using code interpolation to solve the
4343 parentheses problem can be created like this:
4344
4345 $re = qr{\( (?: (?>[^()]+) | (?p{$re}) )* \)}x;
4346
4347 The (?p{...}) item interpolates Perl code at run time, and in this case
4348 refers recursively to the pattern in which it appears.
4349
4350 Obviously, PCRE cannot support the interpolation of Perl code. Instead,
4351 it supports special syntax for recursion of the entire pattern, and
4352 also for individual subpattern recursion. After its introduction in
4353 PCRE and Python, this kind of recursion was introduced into Perl at
4354 release 5.10.
4355
4356 A special item that consists of (? followed by a number greater than
4357 zero and a closing parenthesis is a recursive call of the subpattern of
4358 the given number, provided that it occurs inside that subpattern. (If
4359 not, it is a "subroutine" call, which is described in the next sec-
4360 tion.) The special item (?R) or (?0) is a recursive call of the entire
4361 regular expression.
4362
4363 In PCRE (like Python, but unlike Perl), a recursive subpattern call is
4364 always treated as an atomic group. That is, once it has matched some of
4365 the subject string, it is never re-entered, even if it contains untried
4366 alternatives and there is a subsequent matching failure.
4367
4368 This PCRE pattern solves the nested parentheses problem (assume the
4369 PCRE_EXTENDED option is set so that white space is ignored):
4370
4371 \( ( (?>[^()]+) | (?R) )* \)
4372
4373 First it matches an opening parenthesis. Then it matches any number of
4374 substrings which can either be a sequence of non-parentheses, or a
4375 recursive match of the pattern itself (that is, a correctly parenthe-
4376 sized substring). Finally there is a closing parenthesis.
4377
4378 If this were part of a larger pattern, you would not want to recurse
4379 the entire pattern, so instead you could use this:
4380
4381 ( \( ( (?>[^()]+) | (?1) )* \) )
4382
4383 We have put the pattern into parentheses, and caused the recursion to
4384 refer to them instead of the whole pattern.
4385
4386 In a larger pattern, keeping track of parenthesis numbers can be
4387 tricky. This is made easier by the use of relative references. (A Perl
4388 5.10 feature.) Instead of (?1) in the pattern above you can write
4389 (?-2) to refer to the second most recently opened parentheses preceding
4390 the recursion. In other words, a negative number counts capturing
4391 parentheses leftwards from the point at which it is encountered.
4392
4393 It is also possible to refer to subsequently opened parentheses, by
4394 writing references such as (?+2). However, these cannot be recursive
4395 because the reference is not inside the parentheses that are refer-
4396 enced. They are always "subroutine" calls, as described in the next
4397 section.
4398
4399 An alternative approach is to use named parentheses instead. The Perl
4400 syntax for this is (?&name); PCRE's earlier syntax (?P>name) is also
4401 supported. We could rewrite the above example as follows:
4402
4403 (?<pn> \( ( (?>[^()]+) | (?&pn) )* \) )
4404
4405 If there is more than one subpattern with the same name, the earliest
4406 one is used.
4407
4408 This particular example pattern that we have been looking at contains
4409 nested unlimited repeats, and so the use of atomic grouping for match-
4410 ing strings of non-parentheses is important when applying the pattern
4411 to strings that do not match. For example, when this pattern is applied
4412 to
4413
4414 (aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa()
4415
4416 it yields "no match" quickly. However, if atomic grouping is not used,
4417 the match runs for a very long time indeed because there are so many
4418 different ways the + and * repeats can carve up the subject, and all
4419 have to be tested before failure can be reported.
4420
4421 At the end of a match, the values set for any capturing subpatterns are
4422 those from the outermost level of the recursion at which the subpattern
4423 value is set. If you want to obtain intermediate values, a callout
4424 function can be used (see below and the pcrecallout documentation). If
4425 the pattern above is matched against
4426
4427 (ab(cd)ef)
4428
4429 the value for the capturing parentheses is "ef", which is the last
4430 value taken on at the top level. If additional parentheses are added,
4431 giving
4432
4433 \( ( ( (?>[^()]+) | (?R) )* ) \)
4434 ^ ^
4435 ^ ^
4436
4437 the string they capture is "ab(cd)ef", the contents of the top level
4438 parentheses. If there are more than 15 capturing parentheses in a pat-
4439 tern, PCRE has to obtain extra memory to store data during a recursion,
4440 which it does by using pcre_malloc, freeing it via pcre_free after-
4441 wards. If no memory can be obtained, the match fails with the
4442 PCRE_ERROR_NOMEMORY error.
4443
4444 Do not confuse the (?R) item with the condition (R), which tests for
4445 recursion. Consider this pattern, which matches text in angle brack-
4446 ets, allowing for arbitrary nesting. Only digits are allowed in nested
4447 brackets (that is, when recursing), whereas any characters are permit-
4448 ted at the outer level.
4449
4450 < (?: (?(R) \d++ | [^<>]*+) | (?R)) * >
4451
4452 In this pattern, (?(R) is the start of a conditional subpattern, with
4453 two different alternatives for the recursive and non-recursive cases.
4454 The (?R) item is the actual recursive call.
4455
4456
4457 SUBPATTERNS AS SUBROUTINES
4458
4459 If the syntax for a recursive subpattern reference (either by number or
4460 by name) is used outside the parentheses to which it refers, it oper-
4461 ates like a subroutine in a programming language. The "called" subpat-
4462 tern may be defined before or after the reference. A numbered reference
4463 can be absolute or relative, as in these examples:
4464
4465 (...(absolute)...)...(?2)...
4466 (...(relative)...)...(?-1)...
4467 (...(?+1)...(relative)...
4468
4469 An earlier example pointed out that the pattern
4470
4471 (sens|respons)e and \1ibility
4472
4473 matches "sense and sensibility" and "response and responsibility", but
4474 not "sense and responsibility". If instead the pattern
4475
4476 (sens|respons)e and (?1)ibility
4477
4478 is used, it does match "sense and responsibility" as well as the other
4479 two strings. Another example is given in the discussion of DEFINE
4480 above.
4481
4482 Like recursive subpatterns, a "subroutine" call is always treated as an
4483 atomic group. That is, once it has matched some of the subject string,
4484 it is never re-entered, even if it contains untried alternatives and
4485 there is a subsequent matching failure.
4486
4487 When a subpattern is used as a subroutine, processing options such as
4488 case-independence are fixed when the subpattern is defined. They cannot
4489 be changed for different calls. For example, consider this pattern:
4490
4491 (abc)(?i:(?-1))
4492
4493 It matches "abcabc". It does not match "abcABC" because the change of
4494 processing option does not affect the called subpattern.
4495
4496
4497 CALLOUTS
4498
4499 Perl has a feature whereby using the sequence (?{...}) causes arbitrary
4500 Perl code to be obeyed in the middle of matching a regular expression.
4501 This makes it possible, amongst other things, to extract different sub-
4502 strings that match the same pair of parentheses when there is a repeti-
4503 tion.
4504
4505 PCRE provides a similar feature, but of course it cannot obey arbitrary
4506 Perl code. The feature is called "callout". The caller of PCRE provides
4507 an external function by putting its entry point in the global variable
4508 pcre_callout. By default, this variable contains NULL, which disables
4509 all calling out.
4510
4511 Within a regular expression, (?C) indicates the points at which the
4512 external function is to be called. If you want to identify different
4513 callout points, you can put a number less than 256 after the letter C.
4514 The default value is zero. For example, this pattern has two callout
4515 points:
4516
4517 (?C1)abc(?C2)def
4518
4519 If the PCRE_AUTO_CALLOUT flag is passed to pcre_compile(), callouts are
4520 automatically installed before each item in the pattern. They are all
4521 numbered 255.
4522
4523 During matching, when PCRE reaches a callout point (and pcre_callout is
4524 set), the external function is called. It is provided with the number
4525 of the callout, the position in the pattern, and, optionally, one item
4526 of data originally supplied by the caller of pcre_exec(). The callout
4527 function may cause matching to proceed, to backtrack, or to fail alto-
4528 gether. A complete description of the interface to the callout function
4529 is given in the pcrecallout documentation.
4530
4531
4532 SEE ALSO
4533
4534 pcreapi(3), pcrecallout(3), pcrematching(3), pcre(3).
4535
4536
4537 AUTHOR
4538
4539 Philip Hazel
4540 University Computing Service
4541 Cambridge CB2 3QH, England.
4542
4543
4544 REVISION
4545
4546 Last updated: 19 June 2007
4547 Copyright (c) 1997-2007 University of Cambridge.
4548 ------------------------------------------------------------------------------
4549
4550
4551 PCREPARTIAL(3) PCREPARTIAL(3)
4552
4553
4554 NAME
4555 PCRE - Perl-compatible regular expressions
4556
4557
4558 PARTIAL MATCHING IN PCRE
4559
4560 In normal use of PCRE, if the subject string that is passed to
4561 pcre_exec() or pcre_dfa_exec() matches as far as it goes, but is too
4562 short to match the entire pattern, PCRE_ERROR_NOMATCH is returned.
4563 There are circumstances where it might be helpful to distinguish this
4564 case from other cases in which there is no match.
4565
4566 Consider, for example, an application where a human is required to type
4567 in data for a field with specific formatting requirements. An example
4568 might be a date in the form ddmmmyy, defined by this pattern:
4569
4570 ^\d?\d(jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)\d\d$
4571
4572 If the application sees the user's keystrokes one by one, and can check
4573 that what has been typed so far is potentially valid, it is able to
4574 raise an error as soon as a mistake is made, possibly beeping and not
4575 reflecting the character that has been typed. This immediate feedback
4576 is likely to be a better user interface than a check that is delayed
4577 until the entire string has been entered.
4578
4579 PCRE supports the concept of partial matching by means of the PCRE_PAR-
4580 TIAL option, which can be set when calling pcre_exec() or
4581 pcre_dfa_exec(). When this flag is set for pcre_exec(), the return code
4582 PCRE_ERROR_NOMATCH is converted into PCRE_ERROR_PARTIAL if at any time
4583 during the matching process the last part of the subject string matched
4584 part of the pattern. Unfortunately, for non-anchored matching, it is
4585 not possible to obtain the position of the start of the partial match.
4586 No captured data is set when PCRE_ERROR_PARTIAL is returned.
4587
4588 When PCRE_PARTIAL is set for pcre_dfa_exec(), the return code
4589 PCRE_ERROR_NOMATCH is converted into PCRE_ERROR_PARTIAL if the end of
4590 the subject is reached, there have been no complete matches, but there
4591 is still at least one matching possibility. The portion of the string
4592 that provided the partial match is set as the first matching string.
4593
4594 Using PCRE_PARTIAL disables one of PCRE's optimizations. PCRE remembers
4595 the last literal byte in a pattern, and abandons matching immediately
4596 if such a byte is not present in the subject string. This optimization
4597 cannot be used for a subject string that might match only partially.
4598
4599
4600 RESTRICTED PATTERNS FOR PCRE_PARTIAL
4601
4602 Because of the way certain internal optimizations are implemented in
4603 the pcre_exec() function, the PCRE_PARTIAL option cannot be used with
4604 all patterns. These restrictions do not apply when pcre_dfa_exec() is
4605 used. For pcre_exec(), repeated single characters such as
4606
4607 a{2,4}
4608
4609 and repeated single metasequences such as
4610
4611 \d+
4612
4613 are not permitted if the maximum number of occurrences is greater than
4614 one. Optional items such as \d? (where the maximum is one) are permit-
4615 ted. Quantifiers with any values are permitted after parentheses, so
4616 the invalid examples above can be coded thus:
4617
4618 (a){2,4}
4619 (\d)+
4620
4621 These constructions run more slowly, but for the kinds of application
4622 that are envisaged for this facility, this is not felt to be a major
4623 restriction.
4624
4625 If PCRE_PARTIAL is set for a pattern that does not conform to the
4626 restrictions, pcre_exec() returns the error code PCRE_ERROR_BADPARTIAL
4627 (-13). You can use the PCRE_INFO_OKPARTIAL call to pcre_fullinfo() to
4628 find out if a compiled pattern can be used for partial matching.
4629
4630
4631 EXAMPLE OF PARTIAL MATCHING USING PCRETEST
4632
4633 If the escape sequence \P is present in a pcretest data line, the
4634 PCRE_PARTIAL flag is used for the match. Here is a run of pcretest that
4635 uses the date example quoted above:
4636
4637 re> /^\d?\d(jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)\d\d$/
4638 data> 25jun04\P
4639 0: 25jun04
4640 1: jun
4641 data> 25dec3\P
4642 Partial match
4643 data> 3ju\P
4644 Partial match
4645 data> 3juj\P
4646 No match
4647 data> j\P
4648 No match
4649
4650 The first data string is matched completely, so pcretest shows the
4651 matched substrings. The remaining four strings do not match the com-
4652 plete pattern, but the first two are partial matches. The same test,
4653 using pcre_dfa_exec() matching (by means of the \D escape sequence),
4654 produces the following output:
4655
4656 re> /^\d?\d(jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)\d\d$/
4657 data> 25jun04\P\D
4658 0: 25jun04
4659 data> 23dec3\P\D
4660 Partial match: 23dec3
4661 data> 3ju\P\D
4662 Partial match: 3ju
4663 data> 3juj\P\D
4664 No match
4665 data> j\P\D
4666 No match
4667
4668 Notice that in this case the portion of the string that was matched is
4669 made available.
4670
4671
4672 MULTI-SEGMENT MATCHING WITH pcre_dfa_exec()
4673
4674 When a partial match has been found using pcre_dfa_exec(), it is possi-
4675 ble to continue the match by providing additional subject data and
4676 calling pcre_dfa_exec() again with the same compiled regular expres-
4677 sion, this time setting the PCRE_DFA_RESTART option. You must also pass
4678 the same working space as before, because this is where details of the
4679 previous partial match are stored. Here is an example using pcretest,
4680 using the \R escape sequence to set the PCRE_DFA_RESTART option (\P and
4681 \D are as above):
4682
4683 re> /^\d?\d(jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)\d\d$/
4684 data> 23ja\P\D
4685 Partial match: 23ja
4686 data> n05\R\D
4687 0: n05
4688
4689 The first call has "23ja" as the subject, and requests partial match-
4690 ing; the second call has "n05" as the subject for the continued
4691 (restarted) match. Notice that when the match is complete, only the
4692 last part is shown; PCRE does not retain the previously partially-
4693 matched string. It is up to the calling program to do that if it needs
4694 to.
4695
4696 You can set PCRE_PARTIAL with PCRE_DFA_RESTART to continue partial
4697 matching over multiple segments. This facility can be used to pass very
4698 long subject strings to pcre_dfa_exec(). However, some care is needed
4699 for certain types of pattern.
4700
4701 1. If the pattern contains tests for the beginning or end of a line,
4702 you need to pass the PCRE_NOTBOL or PCRE_NOTEOL options, as appropri-
4703 ate, when the subject string for any call does not contain the begin-
4704 ning or end of a line.
4705
4706 2. If the pattern contains backward assertions (including \b or \B),
4707 you need to arrange for some overlap in the subject strings to allow
4708 for this. For example, you could pass the subject in chunks that are
4709 500 bytes long, but in a buffer of 700 bytes, with the starting offset
4710 set to 200 and the previous 200 bytes at the start of the buffer.
4711
4712 3. Matching a subject string that is split into multiple segments does
4713 not always produce exactly the same result as matching over one single
4714 long string. The difference arises when there are multiple matching
4715 possibilities, because a partial match result is given only when there
4716 are no completed matches in a call to pcre_dfa_exec(). This means that
4717 as soon as the shortest match has been found, continuation to a new
4718 subject segment is no longer possible. Consider this pcretest example:
4719
4720 re> /dog(sbody)?/
4721 data> do\P\D
4722 Partial match: do
4723 data> gsb\R\P\D
4724 0: g
4725 data> dogsbody\D
4726 0: dogsbody
4727 1: dog
4728
4729 The pattern matches the words "dog" or "dogsbody". When the subject is
4730 presented in several parts ("do" and "gsb" being the first two) the
4731 match stops when "dog" has been found, and it is not possible to con-
4732 tinue. On the other hand, if "dogsbody" is presented as a single
4733 string, both matches are found.
4734
4735 Because of this phenomenon, it does not usually make sense to end a
4736 pattern that is going to be matched in this way with a variable repeat.
4737
4738 4. Patterns that contain alternatives at the top level which do not all
4739 start with the same pattern item may not work as expected. For example,
4740 consider this pattern:
4741
4742 1234|3789
4743
4744 If the first part of the subject is "ABC123", a partial match of the
4745 first alternative is found at offset 3. There is no partial match for
4746 the second alternative, because such a match does not start at the same
4747 point in the subject string. Attempting to continue with the string
4748 "789" does not yield a match because only those alternatives that match
4749 at one point in the subject are remembered. The problem arises because
4750 the start of the second alternative matches within the first alterna-
4751 tive. There is no problem with anchored patterns or patterns such as:
4752
4753 1234|ABCD
4754
4755 where no string can be a partial match for both alternatives.
4756
4757
4758 AUTHOR
4759
4760 Philip Hazel
4761 University Computing Service
4762 Cambridge CB2 3QH, England.
4763
4764
4765 REVISION
4766
4767 Last updated: 04 June 2007
4768 Copyright (c) 1997-2007 University of Cambridge.
4769 ------------------------------------------------------------------------------
4770
4771
4772 PCREPRECOMPILE(3) PCREPRECOMPILE(3)
4773
4774
4775 NAME
4776 PCRE - Perl-compatible regular expressions
4777
4778
4779 SAVING AND RE-USING PRECOMPILED PCRE PATTERNS
4780
4781 If you are running an application that uses a large number of regular
4782 expression patterns, it may be useful to store them in a precompiled
4783 form instead of having to compile them every time the application is
4784 run. If you are not using any private character tables (see the
4785 pcre_maketables() documentation), this is relatively straightforward.
4786 If you are using private tables, it is a little bit more complicated.
4787
4788 If you save compiled patterns to a file, you can copy them to a differ-
4789 ent host and run them there. This works even if the new host has the
4790 opposite endianness to the one on which the patterns were compiled.
4791 There may be a small performance penalty, but it should be insignifi-
4792 cant. However, compiling regular expressions with one version of PCRE
4793 for use with a different version is not guaranteed to work and may
4794 cause crashes.
4795
4796
4797 SAVING A COMPILED PATTERN
4798 The value returned by pcre_compile() points to a single block of memory
4799 that holds the compiled pattern and associated data. You can find the
4800 length of this block in bytes by calling pcre_fullinfo() with an argu-
4801 ment of PCRE_INFO_SIZE. You can then save the data in any appropriate
4802 manner. Here is sample code that compiles a pattern and writes it to a
4803 file. It assumes that the variable fd refers to a file that is open for
4804 output:
4805
4806 int erroroffset, rc, size;
4807 char *error;
4808 pcre *re;
4809
4810 re = pcre_compile("my pattern", 0, &error, &erroroffset, NULL);
4811 if (re == NULL) { ... handle errors ... }
4812 rc = pcre_fullinfo(re, NULL, PCRE_INFO_SIZE, &size);
4813 if (rc < 0) { ... handle errors ... }
4814 rc = fwrite(re, 1, size, fd);
4815 if (rc != size) { ... handle errors ... }
4816
4817 In this example, the bytes that comprise the compiled pattern are
4818 copied exactly. Note that this is binary data that may contain any of
4819 the 256 possible byte values. On systems that make a distinction
4820 between binary and non-binary data, be sure that the file is opened for
4821 binary output.
4822
4823 If you want to write more than one pattern to a file, you will have to
4824 devise a way of separating them. For binary data, preceding each pat-
4825 tern with its length is probably the most straightforward approach.
4826 Another possibility is to write out the data in hexadecimal instead of
4827 binary, one pattern to a line.
4828
4829 Saving compiled patterns in a file is only one possible way of storing
4830 them for later use. They could equally well be saved in a database, or
4831 in the memory of some daemon process that passes them via sockets to
4832 the processes that want them.
4833
4834 If the pattern has been studied, it is also possible to save the study
4835 data in a similar way to the compiled pattern itself. When studying
4836 generates additional information, pcre_study() returns a pointer to a
4837 pcre_extra data block. Its format is defined in the section on matching
4838 a pattern in the pcreapi documentation. The study_data field points to
4839 the binary study data, and this is what you must save (not the
4840 pcre_extra block itself). The length of the study data can be obtained
4841 by calling pcre_fullinfo() with an argument of PCRE_INFO_STUDYSIZE.
4842 Remember to check that pcre_study() did return a non-NULL value before
4843 trying to save the study data.
4844
4845
4846 RE-USING A PRECOMPILED PATTERN
4847
4848 Re-using a precompiled pattern is straightforward. Having reloaded it
4849 into main memory, you pass its pointer to pcre_exec() or
4850 pcre_dfa_exec() in the usual way. This should work even on another
4851 host, and even if that host has the opposite endianness to the one
4852 where the pattern was compiled.
4853
4854 However, if you passed a pointer to custom character tables when the
4855 pattern was compiled (the tableptr argument of pcre_compile()), you
4856 must now pass a similar pointer to pcre_exec() or pcre_dfa_exec(),
4857 because the value saved with the compiled pattern will obviously be
4858 nonsense. A field in a pcre_extra() block is used to pass this data, as
4859 described in the section on matching a pattern in the pcreapi documen-
4860 tation.
4861
4862 If you did not provide custom character tables when the pattern was
4863 compiled, the pointer in the compiled pattern is NULL, which causes
4864 pcre_exec() to use PCRE's internal tables. Thus, you do not need to
4865 take any special action at run time in this case.
4866
4867 If you saved study data with the compiled pattern, you need to create
4868 your own pcre_extra data block and set the study_data field to point to
4869 the reloaded study data. You must also set the PCRE_EXTRA_STUDY_DATA
4870 bit in the flags field to indicate that study data is present. Then
4871 pass the pcre_extra block to pcre_exec() or pcre_dfa_exec() in the
4872 usual way.
4873
4874
4875 COMPATIBILITY WITH DIFFERENT PCRE RELEASES
4876
4877 In general, it is safest to recompile all saved patterns when you
4878 update to a new PCRE release, though not all updates actually require
4879 this. Recompiling is definitely needed for release 7.2.
4880
4881
4882 AUTHOR
4883
4884 Philip Hazel
4885 University Computing Service
4886 Cambridge CB2 3QH, England.
4887
4888
4889 REVISION
4890
4891 Last updated: 13 June 2007
4892 Copyright (c) 1997-2007 University of Cambridge.
4893 ------------------------------------------------------------------------------
4894
4895
4896 PCREPERFORM(3) PCREPERFORM(3)
4897
4898
4899 NAME
4900 PCRE - Perl-compatible regular expressions
4901
4902
4903 PCRE PERFORMANCE
4904
4905 Two aspects of performance are discussed below: memory usage and pro-
4906 cessing time. The way you express your pattern as a regular expression
4907 can affect both of them.
4908
4909
4910 MEMORY USAGE
4911
4912 Patterns are compiled by PCRE into a reasonably efficient byte code, so
4913 that most simple patterns do not use much memory. However, there is one
4914 case where memory usage can be unexpectedly large. When a parenthesized
4915 subpattern has a quantifier with a minimum greater than 1 and/or a lim-
4916 ited maximum, the whole subpattern is repeated in the compiled code.
4917 For example, the pattern
4918
4919 (abc|def){2,4}
4920
4921 is compiled as if it were
4922
4923 (abc|def)(abc|def)((abc|def)(abc|def)?)?
4924
4925 (Technical aside: It is done this way so that backtrack points within
4926 each of the repetitions can be independently maintained.)
4927
4928 For regular expressions whose quantifiers use only small numbers, this
4929 is not usually a problem. However, if the numbers are large, and par-
4930 ticularly if such repetitions are nested, the memory usage can become
4931 an embarrassment. For example, the very simple pattern
4932
4933 ((ab){1,1000}c){1,3}
4934
4935 uses 51K bytes when compiled. When PCRE is compiled with its default
4936 internal pointer size of two bytes, the size limit on a compiled pat-
4937 tern is 64K, and this is reached with the above pattern if the outer
4938 repetition is increased from 3 to 4. PCRE can be compiled to use larger
4939 internal pointers and thus handle larger compiled patterns, but it is
4940 better to try to rewrite your pattern to use less memory if you can.
4941
4942 One way of reducing the memory usage for such patterns is to make use
4943 of PCRE's "subroutine" facility. Re-writing the above pattern as
4944
4945 ((ab)(?2){0,999}c)(?1){0,2}
4946
4947 reduces the memory requirements to 18K, and indeed it remains under 20K
4948 even with the outer repetition increased to 100. However, this pattern
4949 is not exactly equivalent, because the "subroutine" calls are treated
4950 as atomic groups into which there can be no backtracking if there is a
4951 subsequent matching failure. Therefore, PCRE cannot do this kind of
4952 rewriting automatically. Furthermore, there is a noticeable loss of
4953 speed when executing the modified pattern. Nevertheless, if the atomic
4954 grouping is not a problem and the loss of speed is acceptable, this
4955 kind of rewriting will allow you to process patterns that PCRE cannot
4956 otherwise handle.
4957
4958
4959 PROCESSING TIME
4960
4961 Certain items in regular expression patterns are processed more effi-
4962 ciently than others. It is more efficient to use a character class like
4963 [aeiou] than a set of single-character alternatives such as
4964 (a|e|i|o|u). In general, the simplest construction that provides the
4965 required behaviour is usually the most efficient. Jeffrey Friedl's book
4966 contains a lot of useful general discussion about optimizing regular
4967 expressions for efficient performance. This document contains a few
4968 observations about PCRE.
4969
4970 Using Unicode character properties (the \p, \P, and \X escapes) is
4971 slow, because PCRE has to scan a structure that contains data for over
4972 fifteen thousand characters whenever it needs a character's property.
4973 If you can find an alternative pattern that does not use character
4974 properties, it will probably be faster.
4975
4976 When a pattern begins with .* not in parentheses, or in parentheses
4977 that are not the subject of a backreference, and the PCRE_DOTALL option
4978 is set, the pattern is implicitly anchored by PCRE, since it can match
4979 only at the start of a subject string. However, if PCRE_DOTALL is not
4980 set, PCRE cannot make this optimization, because the . metacharacter
4981 does not then match a newline, and if the subject string contains new-
4982 lines, the pattern may match from the character immediately following
4983 one of them instead of from the very start. For example, the pattern
4984
4985 .*second
4986
4987 matches the subject "first\nand second" (where \n stands for a newline
4988 character), with the match starting at the seventh character. In order
4989 to do this, PCRE has to retry the match starting after every newline in
4990 the subject.
4991
4992 If you are using such a pattern with subject strings that do not con-
4993 tain newlines, the best performance is obtained by setting PCRE_DOTALL,
4994 or starting the pattern with ^.* or ^.*? to indicate explicit anchor-
4995 ing. That saves PCRE from having to scan along the subject looking for
4996 a newline to restart at.
4997
4998 Beware of patterns that contain nested indefinite repeats. These can
4999 take a long time to run when applied to a string that does not match.
5000 Consider the pattern fragment
5001
5002 ^(a+)*
5003
5004 This can match "aaaa" in 16 different ways, and this number increases
5005 very rapidly as the string gets longer. (The * repeat can match 0, 1,
5006 2, 3, or 4 times, and for each of those cases other than 0 or 4, the +
5007 repeats can match different numbers of times.) When the remainder of
5008 the pattern is such that the entire match is going to fail, PCRE has in
5009 principle to try every possible variation, and this can take an
5010 extremely long time, even for relatively short strings.
5011
5012 An optimization catches some of the more simple cases such as
5013
5014 (a+)*b
5015
5016 where a literal character follows. Before embarking on the standard
5017 matching procedure, PCRE checks that there is a "b" later in the sub-
5018 ject string, and if there is not, it fails the match immediately. How-
5019 ever, when there is no following literal this optimization cannot be
5020 used. You can see the difference by comparing the behaviour of
5021
5022 (a+)*\d
5023
5024 with the pattern above. The former gives a failure almost instantly
5025 when applied to a whole line of "a" characters, whereas the latter
5026 takes an appreciable time with strings longer than about 20 characters.
5027
5028 In many cases, the solution to this kind of performance issue is to use
5029 an atomic group or a possessive quantifier.
5030
5031
5032 AUTHOR
5033
5034 Philip Hazel
5035 University Computing Service
5036 Cambridge CB2 3QH, England.
5037
5038
5039 REVISION
5040
5041 Last updated: 06 March 2007
5042 Copyright (c) 1997-2007 University of Cambridge.
5043 ------------------------------------------------------------------------------
5044
5045
5046 PCREPOSIX(3) PCREPOSIX(3)
5047
5048
5049 NAME
5050 PCRE - Perl-compatible regular expressions.
5051
5052
5053 SYNOPSIS OF POSIX API
5054
5055 #include <pcreposix.h>
5056
5057 int regcomp(regex_t *preg, const char *pattern,
5058 int cflags);
5059
5060 int regexec(regex_t *preg, const char *string,
5061 size_t nmatch, regmatch_t pmatch[], int eflags);
5062
5063 size_t regerror(int errcode, const regex_t *preg,
5064 char *errbuf, size_t errbuf_size);
5065
5066 void regfree(regex_t *preg);
5067
5068
5069 DESCRIPTION
5070
5071 This set of functions provides a POSIX-style API to the PCRE regular
5072 expression package. See the pcreapi documentation for a description of
5073 PCRE's native API, which contains much additional functionality.
5074
5075 The functions described here are just wrapper functions that ultimately
5076 call the PCRE native API. Their prototypes are defined in the
5077 pcreposix.h header file, and on Unix systems the library itself is
5078 called pcreposix.a, so can be accessed by adding -lpcreposix to the
5079 command for linking an application that uses them. Because the POSIX
5080 functions call the native ones, it is also necessary to add -lpcre.
5081
5082 I have implemented only those option bits that can be reasonably mapped
5083 to PCRE native options. In addition, the option REG_EXTENDED is defined
5084 with the value zero. This has no effect, but since programs that are
5085 written to the POSIX interface often use it, this makes it easier to
5086 slot in PCRE as a replacement library. Other POSIX options are not even
5087 defined.
5088
5089 When PCRE is called via these functions, it is only the API that is
5090 POSIX-like in style. The syntax and semantics of the regular expres-
5091 sions themselves are still those of Perl, subject to the setting of
5092 various PCRE options, as described below. "POSIX-like in style" means
5093 that the API approximates to the POSIX definition; it is not fully
5094 POSIX-compatible, and in multi-byte encoding domains it is probably
5095 even less compatible.
5096
5097 The header for these functions is supplied as pcreposix.h to avoid any
5098 potential clash with other POSIX libraries. It can, of course, be
5099 renamed or aliased as regex.h, which is the "correct" name. It provides
5100 two structure types, regex_t for compiled internal forms, and reg-
5101 match_t for returning captured substrings. It also defines some con-
5102 stants whose names start with "REG_"; these are used for setting
5103 options and identifying error codes.
5104
5105
5106 COMPILING A PATTERN
5107
5108 The function regcomp() is called to compile a pattern into an internal
5109 form. The pattern is a C string terminated by a binary zero, and is
5110 passed in the argument pattern. The preg argument is a pointer to a
5111 regex_t structure that is used as a base for storing information about
5112 the compiled regular expression.
5113
5114 The argument cflags is either zero, or contains one or more of the bits
5115 defined by the following macros:
5116
5117 REG_DOTALL
5118
5119 The PCRE_DOTALL option is set when the regular expression is passed for
5120 compilation to the native function. Note that REG_DOTALL is not part of
5121 the POSIX standard.
5122
5123 REG_ICASE
5124
5125 The PCRE_CASELESS option is set when the regular expression is passed
5126 for compilation to the native function.
5127
5128 REG_NEWLINE
5129
5130 The PCRE_MULTILINE option is set when the regular expression is passed
5131 for compilation to the native function. Note that this does not mimic
5132 the defined POSIX behaviour for REG_NEWLINE (see the following sec-
5133 tion).
5134
5135 REG_NOSUB
5136
5137 The PCRE_NO_AUTO_CAPTURE option is set when the regular expression is
5138 passed for compilation to the native function. In addition, when a pat-
5139 tern that is compiled with this flag is passed to regexec() for match-
5140 ing, the nmatch and pmatch arguments are ignored, and no captured
5141 strings are returned.
5142
5143 REG_UTF8
5144
5145 The PCRE_UTF8 option is set when the regular expression is passed for
5146 compilation to the native function. This causes the pattern itself and
5147 all data strings used for matching it to be treated as UTF-8 strings.
5148 Note that REG_UTF8 is not part of the POSIX standard.
5149
5150 In the absence of these flags, no options are passed to the native
5151 function. This means the the regex is compiled with PCRE default
5152 semantics. In particular, the way it handles newline characters in the
5153 subject string is the Perl way, not the POSIX way. Note that setting
5154 PCRE_MULTILINE has only some of the effects specified for REG_NEWLINE.
5155 It does not affect the way newlines are matched by . (they aren't) or
5156 by a negative class such as [^a] (they are).
5157
5158 The yield of regcomp() is zero on success, and non-zero otherwise. The
5159 preg structure is filled in on success, and one member of the structure
5160 is public: re_nsub contains the number of capturing subpatterns in the
5161 regular expression. Various error codes are defined in the header file.
5162
5163
5164 MATCHING NEWLINE CHARACTERS
5165
5166 This area is not simple, because POSIX and Perl take different views of
5167 things. It is not possible to get PCRE to obey POSIX semantics, but
5168 then PCRE was never intended to be a POSIX engine. The following table
5169 lists the different possibilities for matching newline characters in
5170 PCRE:
5171
5172 Default Change with
5173
5174 . matches newline no PCRE_DOTALL
5175 newline matches [^a] yes not changeable
5176 $ matches \n at end yes PCRE_DOLLARENDONLY
5177 $ matches \n in middle no PCRE_MULTILINE
5178 ^ matches \n in middle no PCRE_MULTILINE
5179
5180 This is the equivalent table for POSIX:
5181
5182 Default Change with
5183
5184 . matches newline yes REG_NEWLINE
5185 newline matches [^a] yes REG_NEWLINE
5186 $ matches \n at end no REG_NEWLINE
5187 $ matches \n in middle no REG_NEWLINE
5188 ^ matches \n in middle no REG_NEWLINE
5189
5190 PCRE's behaviour is the same as Perl's, except that there is no equiva-
5191 lent for PCRE_DOLLAR_ENDONLY in Perl. In both PCRE and Perl, there is
5192 no way to stop newline from matching [^a].
5193
5194 The default POSIX newline handling can be obtained by setting
5195 PCRE_DOTALL and PCRE_DOLLAR_ENDONLY, but there is no way to make PCRE
5196 behave exactly as for the REG_NEWLINE action.
5197
5198
5199 MATCHING A PATTERN
5200
5201 The function regexec() is called to match a compiled pattern preg
5202 against a given string, which is terminated by a zero byte, subject to
5203 the options in eflags. These can be:
5204
5205 REG_NOTBOL
5206
5207 The PCRE_NOTBOL option is set when calling the underlying PCRE matching
5208 function.
5209
5210 REG_NOTEOL
5211
5212 The PCRE_NOTEOL option is set when calling the underlying PCRE matching
5213 function.
5214
5215 If the pattern was compiled with the REG_NOSUB flag, no data about any
5216 matched strings is returned. The nmatch and pmatch arguments of
5217 regexec() are ignored.
5218
5219 Otherwise,the portion of the string that was matched, and also any cap-
5220 tured substrings, are returned via the pmatch argument, which points to
5221 an array of nmatch structures of type regmatch_t, containing the mem-
5222 bers rm_so and rm_eo. These contain the offset to the first character
5223 of each substring and the offset to the first character after the end
5224 of each substring, respectively. The 0th element of the vector relates
5225 to the entire portion of string that was matched; subsequent elements
5226 relate to the capturing subpatterns of the regular expression. Unused
5227 entries in the array have both structure members set to -1.
5228
5229 A successful match yields a zero return; various error codes are
5230 defined in the header file, of which REG_NOMATCH is the "expected"
5231 failure code.
5232
5233
5234 ERROR MESSAGES
5235
5236 The regerror() function maps a non-zero errorcode from either regcomp()
5237 or regexec() to a printable message. If preg is not NULL, the error
5238 should have arisen from the use of that structure. A message terminated
5239 by a binary zero is placed in errbuf. The length of the message,
5240 including the zero, is limited to errbuf_size. The yield of the func-
5241 tion is the size of buffer needed to hold the whole message.
5242
5243
5244 MEMORY USAGE
5245
5246 Compiling a regular expression causes memory to be allocated and asso-
5247 ciated with the preg structure. The function regfree() frees all such
5248 memory, after which preg may no longer be used as a compiled expres-
5249 sion.
5250
5251
5252 AUTHOR
5253
5254 Philip Hazel
5255 University Computing Service
5256 Cambridge CB2 3QH, England.
5257
5258
5259 REVISION
5260
5261 Last updated: 06 March 2007
5262 Copyright (c) 1997-2007 University of Cambridge.
5263 ------------------------------------------------------------------------------
5264
5265
5266 PCRECPP(3) PCRECPP(3)
5267
5268
5269 NAME
5270 PCRE - Perl-compatible regular expressions.
5271
5272
5273 SYNOPSIS OF C++ WRAPPER
5274
5275 #include <pcrecpp.h>
5276
5277
5278 DESCRIPTION
5279
5280 The C++ wrapper for PCRE was provided by Google Inc. Some additional
5281 functionality was added by Giuseppe Maxia. This brief man page was con-
5282 structed from the notes in the pcrecpp.h file, which should be con-
5283 sulted for further details.
5284
5285
5286 MATCHING INTERFACE
5287
5288 The "FullMatch" operation checks that supplied text matches a supplied
5289 pattern exactly. If pointer arguments are supplied, it copies matched
5290 sub-strings that match sub-patterns into them.
5291
5292 Example: successful match
5293 pcrecpp::RE re("h.*o");
5294 re.FullMatch("hello");
5295
5296 Example: unsuccessful match (requires full match):
5297 pcrecpp::RE re("e");
5298 !re.FullMatch("hello");
5299
5300 Example: creating a temporary RE object:
5301 pcrecpp::RE("h.*o").FullMatch("hello");
5302
5303 You can pass in a "const char*" or a "string" for "text". The examples
5304 below tend to use a const char*. You can, as in the different examples
5305 above, store the RE object explicitly in a variable or use a temporary
5306 RE object. The examples below use one mode or the other arbitrarily.
5307 Either could correctly be used for any of these examples.
5308
5309 You must supply extra pointer arguments to extract matched subpieces.
5310
5311 Example: extracts "ruby" into "s" and 1234 into "i"
5312 int i;
5313 string s;
5314 pcrecpp::RE re("(\\w+):(\\d+)");
5315 re.FullMatch("ruby:1234", &s, &i);
5316
5317 Example: does not try to extract any extra sub-patterns
5318 re.FullMatch("ruby:1234", &s);
5319
5320 Example: does not try to extract into NULL
5321 re.FullMatch("ruby:1234", NULL, &i);
5322
5323 Example: integer overflow causes failure
5324 !re.FullMatch("ruby:1234567891234", NULL, &i);
5325
5326 Example: fails because there aren't enough sub-patterns:
5327 !pcrecpp::RE("\\w+:\\d+").FullMatch("ruby:1234", &s);
5328
5329 Example: fails because string cannot be stored in integer
5330 !pcrecpp::RE("(.*)").FullMatch("ruby", &i);
5331
5332 The provided pointer arguments can be pointers to any scalar numeric
5333 type, or one of:
5334
5335 string (matched piece is copied to string)
5336 StringPiece (StringPiece is mutated to point to matched piece)
5337 T (where "bool T::ParseFrom(const char*, int)" exists)
5338 NULL (the corresponding matched sub-pattern is not copied)
5339
5340 The function returns true iff all of the following conditions are sat-
5341 isfied:
5342
5343 a. "text" matches "pattern" exactly;
5344
5345 b. The number of matched sub-patterns is >= number of supplied
5346 pointers;
5347
5348 c. The "i"th argument has a suitable type for holding the
5349 string captured as the "i"th sub-pattern. If you pass in
5350 NULL for the "i"th argument, or pass fewer arguments than
5351 number of sub-patterns, "i"th captured sub-pattern is
5352 ignored.
5353
5354 CAVEAT: An optional sub-pattern that does not exist in the matched
5355 string is assigned the empty string. Therefore, the following will
5356 return false (because the empty string is not a valid number):
5357
5358 int number;
5359 pcrecpp::RE::FullMatch("abc", "[a-z]+(\\d+)?", &number);
5360
5361 The matching interface supports at most 16 arguments per call. If you
5362 need more, consider using the more general interface
5363 pcrecpp::RE::DoMatch. See pcrecpp.h for the signature for DoMatch.
5364
5365
5366 QUOTING METACHARACTERS
5367
5368 You can use the "QuoteMeta" operation to insert backslashes before all
5369 potentially meaningful characters in a string. The returned string,
5370 used as a regular expression, will exactly match the original string.
5371
5372 Example:
5373 string quoted = RE::QuoteMeta(unquoted);
5374
5375 Note that it's legal to escape a character even if it has no special
5376 meaning in a regular expression -- so this function does that. (This
5377 also makes it identical to the perl function of the same name; see
5378 "perldoc -f quotemeta".) For example, "1.5-2.0?" becomes
5379 "1\.5\-2\.0\?".
5380
5381
5382 PARTIAL MATCHES
5383
5384 You can use the "PartialMatch" operation when you want the pattern to
5385 match any substring of the text.
5386
5387 Example: simple search for a string:
5388 pcrecpp::RE("ell").PartialMatch("hello");
5389
5390 Example: find first number in a string:
5391 int number;
5392 pcrecpp::RE re("(\\d+)");
5393 re.PartialMatch("x*100 + 20", &number);
5394 assert(number == 100);
5395
5396
5397 UTF-8 AND THE MATCHING INTERFACE
5398
5399 By default, pattern and text are plain text, one byte per character.
5400 The UTF8 flag, passed to the constructor, causes both pattern and
5401 string to be treated as UTF-8 text, still a byte stream but potentially
5402 multiple bytes per character. In practice, the text is likelier to be
5403 UTF-8 than the pattern, but the match returned may depend on the UTF8
5404 flag, so always use it when matching UTF8 text. For example, "." will
5405 match one byte normally but with UTF8 set may match up to three bytes
5406 of a multi-byte character.
5407
5408 Example:
5409 pcrecpp::RE_Options options;
5410 options.set_utf8();
5411 pcrecpp::RE re(utf8_pattern, options);
5412 re.FullMatch(utf8_string);
5413
5414 Example: using the convenience function UTF8():
5415 pcrecpp::RE re(utf8_pattern, pcrecpp::UTF8());
5416 re.FullMatch(utf8_string);
5417
5418 NOTE: The UTF8 flag is ignored if pcre was not configured with the
5419 --enable-utf8 flag.
5420
5421
5422 PASSING MODIFIERS TO THE REGULAR EXPRESSION ENGINE
5423
5424 PCRE defines some modifiers to change the behavior of the regular
5425 expression engine. The C++ wrapper defines an auxiliary class,
5426 RE_Options, as a vehicle to pass such modifiers to a RE class. Cur-
5427 rently, the following modifiers are supported:
5428
5429 modifier description Perl corresponding
5430
5431 PCRE_CASELESS case insensitive match /i
5432 PCRE_MULTILINE multiple lines match /m
5433 PCRE_DOTALL dot matches newlines /s
5434 PCRE_DOLLAR_ENDONLY $ matches only at end N/A
5435 PCRE_EXTRA strict escape parsing N/A
5436 PCRE_EXTENDED ignore whitespaces /x
5437 PCRE_UTF8 handles UTF8 chars built-in
5438 PCRE_UNGREEDY reverses * and *? N/A
5439 PCRE_NO_AUTO_CAPTURE disables capturing parens N/A (*)
5440
5441 (*) Both Perl and PCRE allow non capturing parentheses by means of the
5442 "?:" modifier within the pattern itself. e.g. (?:ab|cd) does not cap-
5443 ture, while (ab|cd) does.
5444
5445 For a full account on how each modifier works, please check the PCRE
5446 API reference page.
5447
5448 For each modifier, there are two member functions whose name is made
5449 out of the modifier in lowercase, without the "PCRE_" prefix. For
5450 instance, PCRE_CASELESS is handled by
5451
5452 bool caseless()
5453
5454 which returns true if the modifier is set, and
5455
5456 RE_Options & set_caseless(bool)
5457
5458 which sets or unsets the modifier. Moreover, PCRE_EXTRA_MATCH_LIMIT can
5459 be accessed through the set_match_limit() and match_limit() member
5460 functions. Setting match_limit to a non-zero value will limit the exe-
5461 cution of pcre to keep it from doing bad things like blowing the stack
5462 or taking an eternity to return a result. A value of 5000 is good
5463 enough to stop stack blowup in a 2MB thread stack. Setting match_limit
5464 to zero disables match limiting. Alternatively, you can call
5465 match_limit_recursion() which uses PCRE_EXTRA_MATCH_LIMIT_RECURSION to
5466 limit how much PCRE recurses. match_limit() limits the number of
5467 matches PCRE does; match_limit_recursion() limits the depth of internal
5468 recursion, and therefore the amount of stack that is used.
5469
5470 Normally, to pass one or more modifiers to a RE class, you declare a
5471 RE_Options object, set the appropriate options, and pass this object to
5472 a RE constructor. Example:
5473
5474 RE_options opt;
5475 opt.set_caseless(true);
5476 if (RE("HELLO", opt).PartialMatch("hello world")) ...
5477
5478 RE_options has two constructors. The default constructor takes no argu-
5479 ments and creates a set of flags that are off by default. The optional
5480 parameter option_flags is to facilitate transfer of legacy code from C
5481 programs. This lets you do
5482
5483 RE(pattern,
5484 RE_Options(PCRE_CASELESS|PCRE_MULTILINE)).PartialMatch(str);
5485
5486 However, new code is better off doing
5487
5488 RE(pattern,
5489 RE_Options().set_caseless(true).set_multiline(true))
5490 .PartialMatch(str);
5491
5492 If you are going to pass one of the most used modifiers, there are some
5493 convenience functions that return a RE_Options class with the appropri-
5494 ate modifier already set: CASELESS(), UTF8(), MULTILINE(), DOTALL(),
5495 and EXTENDED().
5496
5497 If you need to set several options at once, and you don't want to go
5498 through the pains of declaring a RE_Options object and setting several
5499 options, there is a parallel method that give you such ability on the
5500 fly. You can concatenate several set_xxxxx() member functions, since
5501 each of them returns a reference to its class object. For example, to
5502 pass PCRE_CASELESS, PCRE_EXTENDED, and PCRE_MULTILINE to a RE with one
5503 statement, you may write:
5504
5505 RE(" ^ xyz \\s+ .* blah$",
5506 RE_Options()
5507 .set_caseless(true)
5508 .set_extended(true)
5509 .set_multiline(true)).PartialMatch(sometext);
5510
5511
5512 SCANNING TEXT INCREMENTALLY
5513
5514 The "Consume" operation may be useful if you want to repeatedly match
5515 regular expressions at the front of a string and skip over them as they
5516 match. This requires use of the "StringPiece" type, which represents a
5517 sub-range of a real string. Like RE, StringPiece is defined in the
5518 pcrecpp namespace.
5519
5520 Example: read lines of the form "var = value" from a string.
5521 string contents = ...; // Fill string somehow
5522 pcrecpp::StringPiece input(contents); // Wrap in a StringPiece
5523
5524 string var;
5525 int value;
5526 pcrecpp::RE re("(\\w+) = (\\d+)\n");
5527 while (re.Consume(&input, &var, &value)) {
5528 ...;
5529 }
5530
5531 Each successful call to "Consume" will set "var/value", and also
5532 advance "input" so it points past the matched text.
5533
5534 The "FindAndConsume" operation is similar to "Consume" but does not
5535 anchor your match at the beginning of the string. For example, you
5536 could extract all words from a string by repeatedly calling
5537
5538 pcrecpp::RE("(\\w+)").FindAndConsume(&input, &word)
5539
5540
5541 PARSING HEX/OCTAL/C-RADIX NUMBERS
5542
5543 By default, if you pass a pointer to a numeric value, the corresponding
5544 text is interpreted as a base-10 number. You can instead wrap the
5545 pointer with a call to one of the operators Hex(), Octal(), or CRadix()
5546 to interpret the text in another base. The CRadix operator interprets
5547 C-style "0" (base-8) and "0x" (base-16) prefixes, but defaults to
5548 base-10.
5549
5550 Example:
5551 int a, b, c, d;
5552 pcrecpp::RE re("(.*) (.*) (.*) (.*)");
5553 re.FullMatch("100 40 0100 0x40",
5554 pcrecpp::Octal(&a), pcrecpp::Hex(&b),
5555 pcrecpp::CRadix(&c), pcrecpp::CRadix(&d));
5556
5557 will leave 64 in a, b, c, and d.
5558
5559
5560 REPLACING PARTS OF STRINGS
5561
5562 You can replace the first match of "pattern" in "str" with "rewrite".
5563 Within "rewrite", backslash-escaped digits (\1 to \9) can be used to
5564 insert text matching corresponding parenthesized group from the pat-
5565 tern. \0 in "rewrite" refers to the entire matching text. For example:
5566
5567 string s = "yabba dabba doo";
5568 pcrecpp::RE("b+").Replace("d", &s);
5569
5570 will leave "s" containing "yada dabba doo". The result is true if the
5571 pattern matches and a replacement occurs, false otherwise.
5572
5573 GlobalReplace is like Replace except that it replaces all occurrences
5574 of the pattern in the string with the rewrite. Replacements are not
5575 subject to re-matching. For example:
5576
5577 string s = "yabba dabba doo";
5578 pcrecpp::RE("b+").GlobalReplace("d", &s);
5579
5580 will leave "s" containing "yada dada doo". It returns the number of
5581 replacements made.
5582
5583 Extract is like Replace, except that if the pattern matches, "rewrite"
5584 is copied into "out" (an additional argument) with substitutions. The
5585 non-matching portions of "text" are ignored. Returns true iff a match
5586 occurred and the extraction happened successfully; if no match occurs,
5587 the string is left unaffected.
5588
5589
5590 AUTHOR
5591
5592 The C++ wrapper was contributed by Google Inc.
5593 Copyright (c) 2007 Google Inc.
5594
5595
5596 REVISION
5597
5598 Last updated: 06 March 2007
5599 ------------------------------------------------------------------------------
5600
5601
5602 PCRESAMPLE(3) PCRESAMPLE(3)
5603
5604
5605 NAME
5606 PCRE - Perl-compatible regular expressions
5607
5608
5609 PCRE SAMPLE PROGRAM
5610
5611 A simple, complete demonstration program, to get you started with using
5612 PCRE, is supplied in the file pcredemo.c in the PCRE distribution.
5613
5614 The program compiles the regular expression that is its first argument,
5615 and matches it against the subject string in its second argument. No
5616 PCRE options are set, and default character tables are used. If match-
5617 ing succeeds, the program outputs the portion of the subject that
5618 matched, together with the contents of any captured substrings.
5619
5620 If the -g option is given on the command line, the program then goes on
5621 to check for further matches of the same regular expression in the same
5622 subject string. The logic is a little bit tricky because of the possi-
5623 bility of matching an empty string. Comments in the code explain what
5624 is going on.
5625
5626 The demonstration program is automatically built if you use "./config-
5627 ure;make" to build PCRE. Otherwise, if PCRE is installed in the stan-
5628 dard include and library directories for your system, you should be
5629 able to compile the demonstration program using this command:
5630
5631 gcc -o pcredemo pcredemo.c -lpcre
5632
5633 If PCRE is installed elsewhere, you may need to add additional options
5634 to the command line. For example, on a Unix-like system that has PCRE
5635 installed in /usr/local, you can compile the demonstration program
5636 using a command like this:
5637
5638 gcc -o pcredemo -I/usr/local/include pcredemo.c \
5639 -L/usr/local/lib -lpcre
5640
5641 Once you have compiled the demonstration program, you can run simple
5642 tests like this:
5643
5644 ./pcredemo 'cat|dog' 'the cat sat on the mat'
5645 ./pcredemo -g 'cat|dog' 'the dog sat on the cat'
5646
5647 Note that there is a much more comprehensive test program, called
5648 pcretest, which supports many more facilities for testing regular
5649 expressions and the PCRE library. The pcredemo program is provided as a
5650 simple coding example.
5651
5652 On some operating systems (e.g. Solaris), when PCRE is not installed in
5653 the standard library directory, you may get an error like this when you
5654 try to run pcredemo:
5655
5656 ld.so.1: a.out: fatal: libpcre.so.0: open failed: No such file or
5657 directory
5658
5659 This is caused by the way shared library support works on those sys-
5660 tems. You need to add
5661
5662 -R/usr/local/lib
5663
5664 (for example) to the compile command to get round this problem.
5665
5666
5667 AUTHOR
5668
5669 Philip Hazel
5670 University Computing Service
5671 Cambridge CB2 3QH, England.
5672
5673
5674 REVISION
5675
5676 Last updated: 13 June 2007
5677 Copyright (c) 1997-2007 University of Cambridge.
5678 ------------------------------------------------------------------------------
5679 PCRESTACK(3) PCRESTACK(3)
5680
5681
5682 NAME
5683 PCRE - Perl-compatible regular expressions
5684
5685
5686 PCRE DISCUSSION OF STACK USAGE
5687
5688 When you call pcre_exec(), it makes use of an internal function called
5689 match(). This calls itself recursively at branch points in the pattern,
5690 in order to remember the state of the match so that it can back up and
5691 try a different alternative if the first one fails. As matching pro-
5692 ceeds deeper and deeper into the tree of possibilities, the recursion
5693 depth increases.
5694
5695 Not all calls of match() increase the recursion depth; for an item such
5696 as a* it may be called several times at the same level, after matching
5697 different numbers of a's. Furthermore, in a number of cases where the
5698 result of the recursive call would immediately be passed back as the
5699 result of the current call (a "tail recursion"), the function is just
5700 restarted instead.
5701
5702 The pcre_dfa_exec() function operates in an entirely different way, and
5703 hardly uses recursion at all. The limit on its complexity is the amount
5704 of workspace it is given. The comments that follow do NOT apply to
5705 pcre_dfa_exec(); they are relevant only for pcre_exec().
5706
5707 You can set limits on the number of times that match() is called, both
5708 in total and recursively. If the limit is exceeded, an error occurs.
5709 For details, see the section on extra data for pcre_exec() in the
5710 pcreapi documentation.
5711
5712 Each time that match() is actually called recursively, it uses memory
5713 from the process stack. For certain kinds of pattern and data, very
5714 large amounts of stack may be needed, despite the recognition of "tail
5715 recursion". You can often reduce the amount of recursion, and there-
5716 fore the amount of stack used, by modifying the pattern that is being
5717 matched. Consider, for example, this pattern:
5718
5719 ([^<]|<(?!inet))+
5720
5721 It matches from wherever it starts until it encounters "<inet" or the
5722 end of the data, and is the kind of pattern that might be used when
5723 processing an XML file. Each iteration of the outer parentheses matches
5724 either one character that is not "<" or a "<" that is not followed by
5725 "inet". However, each time a parenthesis is processed, a recursion
5726 occurs, so this formulation uses a stack frame for each matched charac-
5727 ter. For a long string, a lot of stack is required. Consider now this
5728 rewritten pattern, which matches exactly the same strings:
5729
5730 ([^<]++|<(?!inet))+
5731
5732 This uses very much less stack, because runs of characters that do not
5733 contain "<" are "swallowed" in one item inside the parentheses. Recur-
5734 sion happens only when a "<" character that is not followed by "inet"
5735 is encountered (and we assume this is relatively rare). A possessive
5736 quantifier is used to stop any backtracking into the runs of non-"<"
5737 characters, but that is not related to stack usage.
5738
5739 This example shows that one way of avoiding stack problems when match-
5740 ing long subject strings is to write repeated parenthesized subpatterns
5741 to match more than one character whenever possible.
5742
5743 In environments where stack memory is constrained, you might want to
5744 compile PCRE to use heap memory instead of stack for remembering back-
5745 up points. This makes it run a lot more slowly, however. Details of how
5746 to do this are given in the pcrebuild documentation. When built in this
5747 way, instead of using the stack, PCRE obtains and frees memory by call-
5748 ing the functions that are pointed to by the pcre_stack_malloc and
5749 pcre_stack_free variables. By default, these point to malloc() and
5750 free(), but you can replace the pointers to cause PCRE to use your own
5751 functions. Since the block sizes are always the same, and are always
5752 freed in reverse order, it may be possible to implement customized mem-
5753 ory handlers that are more efficient than the standard functions.
5754
5755 In Unix-like environments, there is not often a problem with the stack
5756 unless very long strings are involved, though the default limit on
5757 stack size varies from system to system. Values from 8Mb to 64Mb are
5758 common. You can find your default limit by running the command:
5759
5760 ulimit -s
5761
5762 Unfortunately, the effect of running out of stack is often SIGSEGV,
5763 though sometimes a more explicit error message is given. You can nor-
5764 mally increase the limit on stack size by code such as this:
5765
5766 struct rlimit rlim;
5767 getrlimit(RLIMIT_STACK, &rlim);
5768 rlim.rlim_cur = 100*1024*1024;
5769 setrlimit(RLIMIT_STACK, &rlim);
5770
5771 This reads the current limits (soft and hard) using getrlimit(), then
5772 attempts to increase the soft limit to 100Mb using setrlimit(). You
5773 must do this before calling pcre_exec().
5774
5775 PCRE has an internal counter that can be used to limit the depth of
5776 recursion, and thus cause pcre_exec() to give an error code before it
5777 runs out of stack. By default, the limit is very large, and unlikely
5778 ever to operate. It can be changed when PCRE is built, and it can also
5779 be set when pcre_exec() is called. For details of these interfaces, see
5780 the pcrebuild and pcreapi documentation.
5781
5782 As a very rough rule of thumb, you should reckon on about 500 bytes per
5783 recursion. Thus, if you want to limit your stack usage to 8Mb, you
5784 should set the limit at 16000 recursions. A 64Mb stack, on the other
5785 hand, can support around 128000 recursions. The pcretest test program
5786 has a command line option (-S) that can be used to increase the size of
5787 its stack.
5788
5789
5790 AUTHOR
5791
5792 Philip Hazel
5793 University Computing Service
5794 Cambridge CB2 3QH, England.
5795
5796
5797 REVISION
5798
5799 Last updated: 05 June 2007
5800 Copyright (c) 1997-2007 University of Cambridge.
5801 ------------------------------------------------------------------------------
5802
5803

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5