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

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

Parent Directory Parent Directory | Revision Log Revision Log


Revision 289 - (show annotations)
Sun Dec 23 12:17:20 2007 UTC (11 years, 9 months ago) by ph10
File MIME type: text/plain
File size: 303583 byte(s)
Error occurred while calculating annotation data.
Tidies for 7.5-RC1 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. There is a syntax summary in the pcresyntax
49 page.
50
51 Some features of PCRE can be included, excluded, or changed when the
52 library is built. The pcre_config() function makes it possible for a
53 client to discover which features are available. The features them-
54 selves are described in the pcrebuild page. Documentation about build-
55 ing PCRE for various operating systems can be found in the README file
56 in the source distribution.
57
58 The library contains a number of undocumented internal functions and
59 data tables that are used by more than one of the exported external
60 functions, but which are not intended for use by external callers.
61 Their names all begin with "_pcre_", which hopefully will not provoke
62 any name clashes. In some environments, it is possible to control which
63 external symbols are exported when a shared library is built, and in
64 these cases the undocumented symbols are not exported.
65
66
67 USER DOCUMENTATION
68
69 The user documentation for PCRE comprises a number of different sec-
70 tions. In the "man" format, each of these is a separate "man page". In
71 the HTML format, each is a separate page, linked from the index page.
72 In the plain text format, all the sections are concatenated, for ease
73 of searching. The sections are as follows:
74
75 pcre this document
76 pcre-config show PCRE installation configuration information
77 pcreapi details of PCRE's native C API
78 pcrebuild options for building PCRE
79 pcrecallout details of the callout feature
80 pcrecompat discussion of Perl compatibility
81 pcrecpp details of the C++ wrapper
82 pcregrep description of the pcregrep command
83 pcrematching discussion of the two matching algorithms
84 pcrepartial details of the partial matching facility
85 pcrepattern syntax and semantics of supported
86 regular expressions
87 pcresyntax quick syntax reference
88 pcreperform discussion of performance issues
89 pcreposix the POSIX-compatible C API
90 pcreprecompile details of saving and re-using precompiled patterns
91 pcresample discussion of the sample program
92 pcrestack discussion of stack usage
93 pcretest description of the pcretest testing command
94
95 In addition, in the "man" and HTML formats, there is a short page for
96 each C library function, listing its arguments and results.
97
98
99 LIMITATIONS
100
101 There are some size limitations in PCRE but it is hoped that they will
102 never in practice be relevant.
103
104 The maximum length of a compiled pattern is 65539 (sic) bytes if PCRE
105 is compiled with the default internal linkage size of 2. If you want to
106 process regular expressions that are truly enormous, you can compile
107 PCRE with an internal linkage size of 3 or 4 (see the README file in
108 the source distribution and the pcrebuild documentation for details).
109 In these cases the limit is substantially larger. However, the speed
110 of execution is slower.
111
112 All values in repeating quantifiers must be less than 65536.
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 Validity of UTF-8 strings
159
160 When you set the PCRE_UTF8 flag, the strings passed as patterns and
161 subjects are (by default) checked for validity on entry to the relevant
162 functions. From release 7.3 of PCRE, the check is according the rules
163 of RFC 3629, which are themselves derived from the Unicode specifica-
164 tion. Earlier releases of PCRE followed the rules of RFC 2279, which
165 allows the full range of 31-bit values (0 to 0x7FFFFFFF). The current
166 check allows only values in the range U+0 to U+10FFFF, excluding U+D800
167 to U+DFFF.
168
169 The excluded code points are the "Low Surrogate Area" of Unicode, of
170 which the Unicode Standard says this: "The Low Surrogate Area does not
171 contain any character assignments, consequently no character code
172 charts or namelists are provided for this area. Surrogates are reserved
173 for use with UTF-16 and then must be used in pairs." The code points
174 that are encoded by UTF-16 pairs are available as independent code
175 points in the UTF-8 encoding. (In other words, the whole surrogate
176 thing is a fudge for UTF-16 which unfortunately messes up UTF-8.)
177
178 If an invalid UTF-8 string is passed to PCRE, an error return
179 (PCRE_ERROR_BADUTF8) is given. In some situations, you may already know
180 that your strings are valid, and therefore want to skip these checks in
181 order to improve performance. If you set the PCRE_NO_UTF8_CHECK flag at
182 compile time or at run time, PCRE assumes that the pattern or subject
183 it is given (respectively) contains only valid UTF-8 codes. In this
184 case, it does not diagnose an invalid UTF-8 string.
185
186 If you pass an invalid UTF-8 string when PCRE_NO_UTF8_CHECK is set,
187 what happens depends on why the string is invalid. If the string con-
188 forms to the "old" definition of UTF-8 (RFC 2279), it is processed as a
189 string of characters in the range 0 to 0x7FFFFFFF. In other words,
190 apart from the initial validity test, PCRE (when in UTF-8 mode) handles
191 strings according to the more liberal rules of RFC 2279. However, if
192 the string does not even conform to RFC 2279, the result is undefined.
193 Your program may crash.
194
195 If you want to process strings of values in the full range 0 to
196 0x7FFFFFFF, encoded in a UTF-8-like manner as per the old RFC, you can
197 set PCRE_NO_UTF8_CHECK to bypass the more restrictive test. However, in
198 this situation, you will have to apply your own validity check.
199
200 General comments about UTF-8 mode
201
202 1. An unbraced hexadecimal escape sequence (such as \xb3) matches a
203 two-byte UTF-8 character if the value is greater than 127.
204
205 2. Octal numbers up to \777 are recognized, and match two-byte UTF-8
206 characters for values greater than \177.
207
208 3. Repeat quantifiers apply to complete UTF-8 characters, not to indi-
209 vidual bytes, for example: \x{100}{3}.
210
211 4. The dot metacharacter matches one UTF-8 character instead of a sin-
212 gle byte.
213
214 5. The escape sequence \C can be used to match a single byte in UTF-8
215 mode, but its use can lead to some strange effects. This facility is
216 not available in the alternative matching function, pcre_dfa_exec().
217
218 6. The character escapes \b, \B, \d, \D, \s, \S, \w, and \W correctly
219 test characters of any code value, but the characters that PCRE recog-
220 nizes as digits, spaces, or word characters remain the same set as
221 before, all with values less than 256. This remains true even when PCRE
222 includes Unicode property support, because to do otherwise would slow
223 down PCRE in many common cases. If you really want to test for a wider
224 sense of, say, "digit", you must use Unicode property tests such as
225 \p{Nd}.
226
227 7. Similarly, characters that match the POSIX named character classes
228 are all low-valued characters.
229
230 8. However, the Perl 5.10 horizontal and vertical whitespace matching
231 escapes (\h, \H, \v, and \V) do match all the appropriate Unicode char-
232 acters.
233
234 9. Case-insensitive matching applies only to characters whose values
235 are less than 128, unless PCRE is built with Unicode property support.
236 Even when Unicode property support is available, PCRE still uses its
237 own character tables when checking the case of low-valued characters,
238 so as not to degrade performance. The Unicode property information is
239 used only for characters with higher values. Even when Unicode property
240 support is available, PCRE supports case-insensitive matching only when
241 there is a one-to-one mapping between a letter's cases. There are a
242 small number of many-to-one mappings in Unicode; these are not sup-
243 ported by PCRE.
244
245
246 AUTHOR
247
248 Philip Hazel
249 University Computing Service
250 Cambridge CB2 3QH, England.
251
252 Putting an actual email address here seems to have been a spam magnet,
253 so I've taken it away. If you want to email me, use my two initials,
254 followed by the two digits 10, at the domain cam.ac.uk.
255
256
257 REVISION
258
259 Last updated: 09 August 2007
260 Copyright (c) 1997-2007 University of Cambridge.
261 ------------------------------------------------------------------------------
262
263
264 PCREBUILD(3) PCREBUILD(3)
265
266
267 NAME
268 PCRE - Perl-compatible regular expressions
269
270
271 PCRE BUILD-TIME OPTIONS
272
273 This document describes the optional features of PCRE that can be
274 selected when the library is compiled. It assumes use of the configure
275 script, where the optional features are selected or deselected by pro-
276 viding options to configure before running the make command. However,
277 the same options can be selected in both Unix-like and non-Unix-like
278 environments using the GUI facility of CMakeSetup if you are using
279 CMake instead of configure to build PCRE.
280
281 The complete list of options for configure (which includes the standard
282 ones such as the selection of the installation directory) can be
283 obtained by running
284
285 ./configure --help
286
287 The following sections include descriptions of options whose names
288 begin with --enable or --disable. These settings specify changes to the
289 defaults for the configure command. Because of the way that configure
290 works, --enable and --disable always come in pairs, so the complemen-
291 tary option always exists as well, but as it specifies the default, it
292 is not described.
293
294
295 C++ SUPPORT
296
297 By default, the configure script will search for a C++ compiler and C++
298 header files. If it finds them, it automatically builds the C++ wrapper
299 library for PCRE. You can disable this by adding
300
301 --disable-cpp
302
303 to the configure command.
304
305
306 UTF-8 SUPPORT
307
308 To build PCRE with support for UTF-8 character strings, add
309
310 --enable-utf8
311
312 to the configure command. Of itself, this does not make PCRE treat
313 strings as UTF-8. As well as compiling PCRE with this option, you also
314 have have to set the PCRE_UTF8 option when you call the pcre_compile()
315 function.
316
317
318 UNICODE CHARACTER PROPERTY SUPPORT
319
320 UTF-8 support allows PCRE to process character values greater than 255
321 in the strings that it handles. On its own, however, it does not pro-
322 vide any facilities for accessing the properties of such characters. If
323 you want to be able to use the pattern escapes \P, \p, and \X, which
324 refer to Unicode character properties, you must add
325
326 --enable-unicode-properties
327
328 to the configure command. This implies UTF-8 support, even if you have
329 not explicitly requested it.
330
331 Including Unicode property support adds around 30K of tables to the
332 PCRE library. Only the general category properties such as Lu and Nd
333 are supported. Details are given in the pcrepattern documentation.
334
335
336 CODE VALUE OF NEWLINE
337
338 By default, PCRE interprets character 10 (linefeed, LF) as indicating
339 the end of a line. This is the normal newline character on Unix-like
340 systems. You can compile PCRE to use character 13 (carriage return, CR)
341 instead, by adding
342
343 --enable-newline-is-cr
344
345 to the configure command. There is also a --enable-newline-is-lf
346 option, which explicitly specifies linefeed as the newline character.
347
348 Alternatively, you can specify that line endings are to be indicated by
349 the two character sequence CRLF. If you want this, add
350
351 --enable-newline-is-crlf
352
353 to the configure command. There is a fourth option, specified by
354
355 --enable-newline-is-anycrlf
356
357 which causes PCRE to recognize any of the three sequences CR, LF, or
358 CRLF as indicating a line ending. Finally, a fifth option, specified by
359
360 --enable-newline-is-any
361
362 causes PCRE to recognize any Unicode newline sequence.
363
364 Whatever line ending convention is selected when PCRE is built can be
365 overridden when the library functions are called. At build time it is
366 conventional to use the standard for your operating system.
367
368
369 WHAT \R MATCHES
370
371 By default, the sequence \R in a pattern matches any Unicode newline
372 sequence, whatever has been selected as the line ending sequence. If
373 you specify
374
375 --enable-bsr-anycrlf
376
377 the default is changed so that \R matches only CR, LF, or CRLF. What-
378 ever is selected when PCRE is built can be overridden when the library
379 functions are called.
380
381
382 BUILDING SHARED AND STATIC LIBRARIES
383
384 The PCRE building process uses libtool to build both shared and static
385 Unix libraries by default. You can suppress one of these by adding one
386 of
387
388 --disable-shared
389 --disable-static
390
391 to the configure command, as required.
392
393
394 POSIX MALLOC USAGE
395
396 When PCRE is called through the POSIX interface (see the pcreposix doc-
397 umentation), additional working storage is required for holding the
398 pointers to capturing substrings, because PCRE requires three integers
399 per substring, whereas the POSIX interface provides only two. If the
400 number of expected substrings is small, the wrapper function uses space
401 on the stack, because this is faster than using malloc() for each call.
402 The default threshold above which the stack is no longer used is 10; it
403 can be changed by adding a setting such as
404
405 --with-posix-malloc-threshold=20
406
407 to the configure command.
408
409
410 HANDLING VERY LARGE PATTERNS
411
412 Within a compiled pattern, offset values are used to point from one
413 part to another (for example, from an opening parenthesis to an alter-
414 nation metacharacter). By default, two-byte values are used for these
415 offsets, leading to a maximum size for a compiled pattern of around
416 64K. This is sufficient to handle all but the most gigantic patterns.
417 Nevertheless, some people do want to process enormous patterns, so it
418 is possible to compile PCRE to use three-byte or four-byte offsets by
419 adding a setting such as
420
421 --with-link-size=3
422
423 to the configure command. The value given must be 2, 3, or 4. Using
424 longer offsets slows down the operation of PCRE because it has to load
425 additional bytes when handling them.
426
427
428 AVOIDING EXCESSIVE STACK USAGE
429
430 When matching with the pcre_exec() function, PCRE implements backtrack-
431 ing by making recursive calls to an internal function called match().
432 In environments where the size of the stack is limited, this can se-
433 verely limit PCRE's operation. (The Unix environment does not usually
434 suffer from this problem, but it may sometimes be necessary to increase
435 the maximum stack size. There is a discussion in the pcrestack docu-
436 mentation.) An alternative approach to recursion that uses memory from
437 the heap to remember data, instead of using recursive function calls,
438 has been implemented to work round the problem of limited stack size.
439 If you want to build a version of PCRE that works this way, add
440
441 --disable-stack-for-recursion
442
443 to the configure command. With this configuration, PCRE will use the
444 pcre_stack_malloc and pcre_stack_free variables to call memory manage-
445 ment functions. By default these point to malloc() and free(), but you
446 can replace the pointers so that your own functions are used.
447
448 Separate functions are provided rather than using pcre_malloc and
449 pcre_free because the usage is very predictable: the block sizes
450 requested are always the same, and the blocks are always freed in
451 reverse order. A calling program might be able to implement optimized
452 functions that perform better than malloc() and free(). PCRE runs
453 noticeably more slowly when built in this way. This option affects only
454 the pcre_exec() function; it is not relevant for the the
455 pcre_dfa_exec() function.
456
457
458 LIMITING PCRE RESOURCE USAGE
459
460 Internally, PCRE has a function called match(), which it calls repeat-
461 edly (sometimes recursively) when matching a pattern with the
462 pcre_exec() function. By controlling the maximum number of times this
463 function may be called during a single matching operation, a limit can
464 be placed on the resources used by a single call to pcre_exec(). The
465 limit can be changed at run time, as described in the pcreapi documen-
466 tation. The default is 10 million, but this can be changed by adding a
467 setting such as
468
469 --with-match-limit=500000
470
471 to the configure command. This setting has no effect on the
472 pcre_dfa_exec() matching function.
473
474 In some environments it is desirable to limit the depth of recursive
475 calls of match() more strictly than the total number of calls, in order
476 to restrict the maximum amount of stack (or heap, if --disable-stack-
477 for-recursion is specified) that is used. A second limit controls this;
478 it defaults to the value that is set for --with-match-limit, which
479 imposes no additional constraints. However, you can set a lower limit
480 by adding, for example,
481
482 --with-match-limit-recursion=10000
483
484 to the configure command. This value can also be overridden at run
485 time.
486
487
488 CREATING CHARACTER TABLES AT BUILD TIME
489
490 PCRE uses fixed tables for processing characters whose code values are
491 less than 256. By default, PCRE is built with a set of tables that are
492 distributed in the file pcre_chartables.c.dist. These tables are for
493 ASCII codes only. If you add
494
495 --enable-rebuild-chartables
496
497 to the configure command, the distributed tables are no longer used.
498 Instead, a program called dftables is compiled and run. This outputs
499 the source for new set of tables, created in the default locale of your
500 C runtime system. (This method of replacing the tables does not work if
501 you are cross compiling, because dftables is run on the local host. If
502 you need to create alternative tables when cross compiling, you will
503 have to do so "by hand".)
504
505
506 USING EBCDIC CODE
507
508 PCRE assumes by default that it will run in an environment where the
509 character code is ASCII (or Unicode, which is a superset of ASCII).
510 This is the case for most computer operating systems. PCRE can, how-
511 ever, be compiled to run in an EBCDIC environment by adding
512
513 --enable-ebcdic
514
515 to the configure command. This setting implies --enable-rebuild-charta-
516 bles. You should only use it if you know that you are in an EBCDIC
517 environment (for example, an IBM mainframe operating system).
518
519
520 PCREGREP OPTIONS FOR COMPRESSED FILE SUPPORT
521
522 By default, pcregrep reads all files as plain text. You can build it so
523 that it recognizes files whose names end in .gz or .bz2, and reads them
524 with libz or libbz2, respectively, by adding one or both of
525
526 --enable-pcregrep-libz
527 --enable-pcregrep-libbz2
528
529 to the configure command. These options naturally require that the rel-
530 evant libraries are installed on your system. Configuration will fail
531 if they are not.
532
533
534 PCRETEST OPTION FOR LIBREADLINE SUPPORT
535
536 If you add
537
538 --enable-pcretest-libreadline
539
540 to the configure command, pcretest is linked with the libreadline
541 library, and when its input is from a terminal, it reads it using the
542 readline() function. This provides line-editing and history facilities.
543 Note that libreadline is GPL-licenced, so if you distribute a binary of
544 pcretest linked in this way, there may be licensing issues.
545
546
547 SEE ALSO
548
549 pcreapi(3), pcre_config(3).
550
551
552 AUTHOR
553
554 Philip Hazel
555 University Computing Service
556 Cambridge CB2 3QH, England.
557
558
559 REVISION
560
561 Last updated: 18 December 2007
562 Copyright (c) 1997-2007 University of Cambridge.
563 ------------------------------------------------------------------------------
564
565
566 PCREMATCHING(3) PCREMATCHING(3)
567
568
569 NAME
570 PCRE - Perl-compatible regular expressions
571
572
573 PCRE MATCHING ALGORITHMS
574
575 This document describes the two different algorithms that are available
576 in PCRE for matching a compiled regular expression against a given sub-
577 ject string. The "standard" algorithm is the one provided by the
578 pcre_exec() function. This works in the same was as Perl's matching
579 function, and provides a Perl-compatible matching operation.
580
581 An alternative algorithm is provided by the pcre_dfa_exec() function;
582 this operates in a different way, and is not Perl-compatible. It has
583 advantages and disadvantages compared with the standard algorithm, and
584 these are described below.
585
586 When there is only one possible way in which a given subject string can
587 match a pattern, the two algorithms give the same answer. A difference
588 arises, however, when there are multiple possibilities. For example, if
589 the pattern
590
591 ^<.*>
592
593 is matched against the string
594
595 <something> <something else> <something further>
596
597 there are three possible answers. The standard algorithm finds only one
598 of them, whereas the alternative algorithm finds all three.
599
600
601 REGULAR EXPRESSIONS AS TREES
602
603 The set of strings that are matched by a regular expression can be rep-
604 resented as a tree structure. An unlimited repetition in the pattern
605 makes the tree of infinite size, but it is still a tree. Matching the
606 pattern to a given subject string (from a given starting point) can be
607 thought of as a search of the tree. There are two ways to search a
608 tree: depth-first and breadth-first, and these correspond to the two
609 matching algorithms provided by PCRE.
610
611
612 THE STANDARD MATCHING ALGORITHM
613
614 In the terminology of Jeffrey Friedl's book "Mastering Regular Expres-
615 sions", the standard algorithm is an "NFA algorithm". It conducts a
616 depth-first search of the pattern tree. That is, it proceeds along a
617 single path through the tree, checking that the subject matches what is
618 required. When there is a mismatch, the algorithm tries any alterna-
619 tives at the current point, and if they all fail, it backs up to the
620 previous branch point in the tree, and tries the next alternative
621 branch at that level. This often involves backing up (moving to the
622 left) in the subject string as well. The order in which repetition
623 branches are tried is controlled by the greedy or ungreedy nature of
624 the quantifier.
625
626 If a leaf node is reached, a matching string has been found, and at
627 that point the algorithm stops. Thus, if there is more than one possi-
628 ble match, this algorithm returns the first one that it finds. Whether
629 this is the shortest, the longest, or some intermediate length depends
630 on the way the greedy and ungreedy repetition quantifiers are specified
631 in the pattern.
632
633 Because it ends up with a single path through the tree, it is rela-
634 tively straightforward for this algorithm to keep track of the sub-
635 strings that are matched by portions of the pattern in parentheses.
636 This provides support for capturing parentheses and back references.
637
638
639 THE ALTERNATIVE MATCHING ALGORITHM
640
641 This algorithm conducts a breadth-first search of the tree. Starting
642 from the first matching point in the subject, it scans the subject
643 string from left to right, once, character by character, and as it does
644 this, it remembers all the paths through the tree that represent valid
645 matches. In Friedl's terminology, this is a kind of "DFA algorithm",
646 though it is not implemented as a traditional finite state machine (it
647 keeps multiple states active simultaneously).
648
649 The scan continues until either the end of the subject is reached, or
650 there are no more unterminated paths. At this point, terminated paths
651 represent the different matching possibilities (if there are none, the
652 match has failed). Thus, if there is more than one possible match,
653 this algorithm finds all of them, and in particular, it finds the long-
654 est. In PCRE, there is an option to stop the algorithm after the first
655 match (which is necessarily the shortest) has been found.
656
657 Note that all the matches that are found start at the same point in the
658 subject. If the pattern
659
660 cat(er(pillar)?)
661
662 is matched against the string "the caterpillar catchment", the result
663 will be the three strings "cat", "cater", and "caterpillar" that start
664 at the fourth character of the subject. The algorithm does not automat-
665 ically move on to find matches that start at later positions.
666
667 There are a number of features of PCRE regular expressions that are not
668 supported by the alternative matching algorithm. They are as follows:
669
670 1. Because the algorithm finds all possible matches, the greedy or
671 ungreedy nature of repetition quantifiers is not relevant. Greedy and
672 ungreedy quantifiers are treated in exactly the same way. However, pos-
673 sessive quantifiers can make a difference when what follows could also
674 match what is quantified, for example in a pattern like this:
675
676 ^a++\w!
677
678 This pattern matches "aaab!" but not "aaa!", which would be matched by
679 a non-possessive quantifier. Similarly, if an atomic group is present,
680 it is matched as if it were a standalone pattern at the current point,
681 and the longest match is then "locked in" for the rest of the overall
682 pattern.
683
684 2. When dealing with multiple paths through the tree simultaneously, it
685 is not straightforward to keep track of captured substrings for the
686 different matching possibilities, and PCRE's implementation of this
687 algorithm does not attempt to do this. This means that no captured sub-
688 strings are available.
689
690 3. Because no substrings are captured, back references within the pat-
691 tern are not supported, and cause errors if encountered.
692
693 4. For the same reason, conditional expressions that use a backrefer-
694 ence as the condition or test for a specific group recursion are not
695 supported.
696
697 5. Because many paths through the tree may be active, the \K escape
698 sequence, which resets the start of the match when encountered (but may
699 be on some paths and not on others), is not supported. It causes an
700 error if encountered.
701
702 6. Callouts are supported, but the value of the capture_top field is
703 always 1, and the value of the capture_last field is always -1.
704
705 7. The \C escape sequence, which (in the standard algorithm) matches a
706 single byte, even in UTF-8 mode, is not supported because the alterna-
707 tive algorithm moves through the subject string one character at a
708 time, for all active paths through the tree.
709
710 8. None of the backtracking control verbs such as (*PRUNE) are sup-
711 ported.
712
713
714 ADVANTAGES OF THE ALTERNATIVE ALGORITHM
715
716 Using the alternative matching algorithm provides the following advan-
717 tages:
718
719 1. All possible matches (at a single point in the subject) are automat-
720 ically found, and in particular, the longest match is found. To find
721 more than one match using the standard algorithm, you have to do kludgy
722 things with callouts.
723
724 2. There is much better support for partial matching. The restrictions
725 on the content of the pattern that apply when using the standard algo-
726 rithm for partial matching do not apply to the alternative algorithm.
727 For non-anchored patterns, the starting position of a partial match is
728 available.
729
730 3. Because the alternative algorithm scans the subject string just
731 once, and never needs to backtrack, it is possible to pass very long
732 subject strings to the matching function in several pieces, checking
733 for partial matching each time.
734
735
736 DISADVANTAGES OF THE ALTERNATIVE ALGORITHM
737
738 The alternative algorithm suffers from a number of disadvantages:
739
740 1. It is substantially slower than the standard algorithm. This is
741 partly because it has to search for all possible matches, but is also
742 because it is less susceptible to optimization.
743
744 2. Capturing parentheses and back references are not supported.
745
746 3. Although atomic groups are supported, their use does not provide the
747 performance advantage that it does for the standard algorithm.
748
749
750 AUTHOR
751
752 Philip Hazel
753 University Computing Service
754 Cambridge CB2 3QH, England.
755
756
757 REVISION
758
759 Last updated: 08 August 2007
760 Copyright (c) 1997-2007 University of Cambridge.
761 ------------------------------------------------------------------------------
762
763
764 PCREAPI(3) PCREAPI(3)
765
766
767 NAME
768 PCRE - Perl-compatible regular expressions
769
770
771 PCRE NATIVE API
772
773 #include <pcre.h>
774
775 pcre *pcre_compile(const char *pattern, int options,
776 const char **errptr, int *erroffset,
777 const unsigned char *tableptr);
778
779 pcre *pcre_compile2(const char *pattern, int options,
780 int *errorcodeptr,
781 const char **errptr, int *erroffset,
782 const unsigned char *tableptr);
783
784 pcre_extra *pcre_study(const pcre *code, int options,
785 const char **errptr);
786
787 int pcre_exec(const pcre *code, const pcre_extra *extra,
788 const char *subject, int length, int startoffset,
789 int options, int *ovector, int ovecsize);
790
791 int pcre_dfa_exec(const pcre *code, const pcre_extra *extra,
792 const char *subject, int length, int startoffset,
793 int options, int *ovector, int ovecsize,
794 int *workspace, int wscount);
795
796 int pcre_copy_named_substring(const pcre *code,
797 const char *subject, int *ovector,
798 int stringcount, const char *stringname,
799 char *buffer, int buffersize);
800
801 int pcre_copy_substring(const char *subject, int *ovector,
802 int stringcount, int stringnumber, char *buffer,
803 int buffersize);
804
805 int pcre_get_named_substring(const pcre *code,
806 const char *subject, int *ovector,
807 int stringcount, const char *stringname,
808 const char **stringptr);
809
810 int pcre_get_stringnumber(const pcre *code,
811 const char *name);
812
813 int pcre_get_stringtable_entries(const pcre *code,
814 const char *name, char **first, char **last);
815
816 int pcre_get_substring(const char *subject, int *ovector,
817 int stringcount, int stringnumber,
818 const char **stringptr);
819
820 int pcre_get_substring_list(const char *subject,
821 int *ovector, int stringcount, const char ***listptr);
822
823 void pcre_free_substring(const char *stringptr);
824
825 void pcre_free_substring_list(const char **stringptr);
826
827 const unsigned char *pcre_maketables(void);
828
829 int pcre_fullinfo(const pcre *code, const pcre_extra *extra,
830 int what, void *where);
831
832 int pcre_info(const pcre *code, int *optptr, int *firstcharptr);
833
834 int pcre_refcount(pcre *code, int adjust);
835
836 int pcre_config(int what, void *where);
837
838 char *pcre_version(void);
839
840 void *(*pcre_malloc)(size_t);
841
842 void (*pcre_free)(void *);
843
844 void *(*pcre_stack_malloc)(size_t);
845
846 void (*pcre_stack_free)(void *);
847
848 int (*pcre_callout)(pcre_callout_block *);
849
850
851 PCRE API OVERVIEW
852
853 PCRE has its own native API, which is described in this document. There
854 are also some wrapper functions that correspond to the POSIX regular
855 expression API. These are described in the pcreposix documentation.
856 Both of these APIs define a set of C function calls. A C++ wrapper is
857 distributed with PCRE. It is documented in the pcrecpp page.
858
859 The native API C function prototypes are defined in the header file
860 pcre.h, and on Unix systems the library itself is called libpcre. It
861 can normally be accessed by adding -lpcre to the command for linking an
862 application that uses PCRE. The header file defines the macros
863 PCRE_MAJOR and PCRE_MINOR to contain the major and minor release num-
864 bers for the library. Applications can use these to include support
865 for different releases of PCRE.
866
867 The functions pcre_compile(), pcre_compile2(), pcre_study(), and
868 pcre_exec() are used for compiling and matching regular expressions in
869 a Perl-compatible manner. A sample program that demonstrates the sim-
870 plest way of using them is provided in the file called pcredemo.c in
871 the source distribution. The pcresample documentation describes how to
872 run it.
873
874 A second matching function, pcre_dfa_exec(), which is not Perl-compati-
875 ble, is also provided. This uses a different algorithm for the match-
876 ing. The alternative algorithm finds all possible matches (at a given
877 point in the subject), and scans the subject just once. However, this
878 algorithm does not return captured substrings. A description of the two
879 matching algorithms and their advantages and disadvantages is given in
880 the pcrematching documentation.
881
882 In addition to the main compiling and matching functions, there are
883 convenience functions for extracting captured substrings from a subject
884 string that is matched by pcre_exec(). They are:
885
886 pcre_copy_substring()
887 pcre_copy_named_substring()
888 pcre_get_substring()
889 pcre_get_named_substring()
890 pcre_get_substring_list()
891 pcre_get_stringnumber()
892 pcre_get_stringtable_entries()
893
894 pcre_free_substring() and pcre_free_substring_list() are also provided,
895 to free the memory used for extracted strings.
896
897 The function pcre_maketables() is used to build a set of character
898 tables in the current locale for passing to pcre_compile(),
899 pcre_exec(), or pcre_dfa_exec(). This is an optional facility that is
900 provided for specialist use. Most commonly, no special tables are
901 passed, in which case internal tables that are generated when PCRE is
902 built are used.
903
904 The function pcre_fullinfo() is used to find out information about a
905 compiled pattern; pcre_info() is an obsolete version that returns only
906 some of the available information, but is retained for backwards com-
907 patibility. The function pcre_version() returns a pointer to a string
908 containing the version of PCRE and its date of release.
909
910 The function pcre_refcount() maintains a reference count in a data
911 block containing a compiled pattern. This is provided for the benefit
912 of object-oriented applications.
913
914 The global variables pcre_malloc and pcre_free initially contain the
915 entry points of the standard malloc() and free() functions, respec-
916 tively. PCRE calls the memory management functions via these variables,
917 so a calling program can replace them if it wishes to intercept the
918 calls. This should be done before calling any PCRE functions.
919
920 The global variables pcre_stack_malloc and pcre_stack_free are also
921 indirections to memory management functions. These special functions
922 are used only when PCRE is compiled to use the heap for remembering
923 data, instead of recursive function calls, when running the pcre_exec()
924 function. See the pcrebuild documentation for details of how to do
925 this. It is a non-standard way of building PCRE, for use in environ-
926 ments that have limited stacks. Because of the greater use of memory
927 management, it runs more slowly. Separate functions are provided so
928 that special-purpose external code can be used for this case. When
929 used, these functions are always called in a stack-like manner (last
930 obtained, first freed), and always for memory blocks of the same size.
931 There is a discussion about PCRE's stack usage in the pcrestack docu-
932 mentation.
933
934 The global variable pcre_callout initially contains NULL. It can be set
935 by the caller to a "callout" function, which PCRE will then call at
936 specified points during a matching operation. Details are given in the
937 pcrecallout documentation.
938
939
940 NEWLINES
941
942 PCRE supports five different conventions for indicating line breaks in
943 strings: a single CR (carriage return) character, a single LF (line-
944 feed) character, the two-character sequence CRLF, any of the three pre-
945 ceding, or any Unicode newline sequence. The Unicode newline sequences
946 are the three just mentioned, plus the single characters VT (vertical
947 tab, U+000B), FF (formfeed, U+000C), NEL (next line, U+0085), LS (line
948 separator, U+2028), and PS (paragraph separator, U+2029).
949
950 Each of the first three conventions is used by at least one operating
951 system as its standard newline sequence. When PCRE is built, a default
952 can be specified. The default default is LF, which is the Unix stan-
953 dard. When PCRE is run, the default can be overridden, either when a
954 pattern is compiled, or when it is matched.
955
956 At compile time, the newline convention can be specified by the options
957 argument of pcre_compile(), or it can be specified by special text at
958 the start of the pattern itself; this overrides any other settings. See
959 the pcrepattern page for details of the special character sequences.
960
961 In the PCRE documentation the word "newline" is used to mean "the char-
962 acter or pair of characters that indicate a line break". The choice of
963 newline convention affects the handling of the dot, circumflex, and
964 dollar metacharacters, the handling of #-comments in /x mode, and, when
965 CRLF is a recognized line ending sequence, the match position advance-
966 ment for a non-anchored pattern. There is more detail about this in the
967 section on pcre_exec() options below.
968
969 The choice of newline convention does not affect the interpretation of
970 the \n or \r escape sequences, nor does it affect what \R matches,
971 which is controlled in a similar way, but by separate options.
972
973
974 MULTITHREADING
975
976 The PCRE functions can be used in multi-threading applications, with
977 the proviso that the memory management functions pointed to by
978 pcre_malloc, pcre_free, pcre_stack_malloc, and pcre_stack_free, and the
979 callout function pointed to by pcre_callout, are shared by all threads.
980
981 The compiled form of a regular expression is not altered during match-
982 ing, so the same compiled pattern can safely be used by several threads
983 at once.
984
985
986 SAVING PRECOMPILED PATTERNS FOR LATER USE
987
988 The compiled form of a regular expression can be saved and re-used at a
989 later time, possibly by a different program, and even on a host other
990 than the one on which it was compiled. Details are given in the
991 pcreprecompile documentation. However, compiling a regular expression
992 with one version of PCRE for use with a different version is not guar-
993 anteed to work and may cause crashes.
994
995
996 CHECKING BUILD-TIME OPTIONS
997
998 int pcre_config(int what, void *where);
999
1000 The function pcre_config() makes it possible for a PCRE client to dis-
1001 cover which optional features have been compiled into the PCRE library.
1002 The pcrebuild documentation has more details about these optional fea-
1003 tures.
1004
1005 The first argument for pcre_config() is an integer, specifying which
1006 information is required; the second argument is a pointer to a variable
1007 into which the information is placed. The following information is
1008 available:
1009
1010 PCRE_CONFIG_UTF8
1011
1012 The output is an integer that is set to one if UTF-8 support is avail-
1013 able; otherwise it is set to zero.
1014
1015 PCRE_CONFIG_UNICODE_PROPERTIES
1016
1017 The output is an integer that is set to one if support for Unicode
1018 character properties is available; otherwise it is set to zero.
1019
1020 PCRE_CONFIG_NEWLINE
1021
1022 The output is an integer whose value specifies the default character
1023 sequence that is recognized as meaning "newline". The four values that
1024 are supported are: 10 for LF, 13 for CR, 3338 for CRLF, -2 for ANYCRLF,
1025 and -1 for ANY. The default should normally be the standard sequence
1026 for your operating system.
1027
1028 PCRE_CONFIG_BSR
1029
1030 The output is an integer whose value indicates what character sequences
1031 the \R escape sequence matches by default. A value of 0 means that \R
1032 matches any Unicode line ending sequence; a value of 1 means that \R
1033 matches only CR, LF, or CRLF. The default can be overridden when a pat-
1034 tern is compiled or matched.
1035
1036 PCRE_CONFIG_LINK_SIZE
1037
1038 The output is an integer that contains the number of bytes used for
1039 internal linkage in compiled regular expressions. The value is 2, 3, or
1040 4. Larger values allow larger regular expressions to be compiled, at
1041 the expense of slower matching. The default value of 2 is sufficient
1042 for all but the most massive patterns, since it allows the compiled
1043 pattern to be up to 64K in size.
1044
1045 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
1046
1047 The output is an integer that contains the threshold above which the
1048 POSIX interface uses malloc() for output vectors. Further details are
1049 given in the pcreposix documentation.
1050
1051 PCRE_CONFIG_MATCH_LIMIT
1052
1053 The output is an integer that gives the default limit for the number of
1054 internal matching function calls in a pcre_exec() execution. Further
1055 details are given with pcre_exec() below.
1056
1057 PCRE_CONFIG_MATCH_LIMIT_RECURSION
1058
1059 The output is an integer that gives the default limit for the depth of
1060 recursion when calling the internal matching function in a pcre_exec()
1061 execution. Further details are given with pcre_exec() below.
1062
1063 PCRE_CONFIG_STACKRECURSE
1064
1065 The output is an integer that is set to one if internal recursion when
1066 running pcre_exec() is implemented by recursive function calls that use
1067 the stack to remember their state. This is the usual way that PCRE is
1068 compiled. The output is zero if PCRE was compiled to use blocks of data
1069 on the heap instead of recursive function calls. In this case,
1070 pcre_stack_malloc and pcre_stack_free are called to manage memory
1071 blocks on the heap, thus avoiding the use of the stack.
1072
1073
1074 COMPILING A PATTERN
1075
1076 pcre *pcre_compile(const char *pattern, int options,
1077 const char **errptr, int *erroffset,
1078 const unsigned char *tableptr);
1079
1080 pcre *pcre_compile2(const char *pattern, int options,
1081 int *errorcodeptr,
1082 const char **errptr, int *erroffset,
1083 const unsigned char *tableptr);
1084
1085 Either of the functions pcre_compile() or pcre_compile2() can be called
1086 to compile a pattern into an internal form. The only difference between
1087 the two interfaces is that pcre_compile2() has an additional argument,
1088 errorcodeptr, via which a numerical error code can be returned.
1089
1090 The pattern is a C string terminated by a binary zero, and is passed in
1091 the pattern argument. A pointer to a single block of memory that is
1092 obtained via pcre_malloc is returned. This contains the compiled code
1093 and related data. The pcre type is defined for the returned block; this
1094 is a typedef for a structure whose contents are not externally defined.
1095 It is up to the caller to free the memory (via pcre_free) when it is no
1096 longer required.
1097
1098 Although the compiled code of a PCRE regex is relocatable, that is, it
1099 does not depend on memory location, the complete pcre data block is not
1100 fully relocatable, because it may contain a copy of the tableptr argu-
1101 ment, which is an address (see below).
1102
1103 The options argument contains various bit settings that affect the com-
1104 pilation. It should be zero if no options are required. The available
1105 options are described below. Some of them, in particular, those that
1106 are compatible with Perl, can also be set and unset from within the
1107 pattern (see the detailed description in the pcrepattern documenta-
1108 tion). For these options, the contents of the options argument speci-
1109 fies their initial settings at the start of compilation and execution.
1110 The PCRE_ANCHORED and PCRE_NEWLINE_xxx options can be set at the time
1111 of matching as well as at compile time.
1112
1113 If errptr is NULL, pcre_compile() returns NULL immediately. Otherwise,
1114 if compilation of a pattern fails, pcre_compile() returns NULL, and
1115 sets the variable pointed to by errptr to point to a textual error mes-
1116 sage. This is a static string that is part of the library. You must not
1117 try to free it. The offset from the start of the pattern to the charac-
1118 ter where the error was discovered is placed in the variable pointed to
1119 by erroffset, which must not be NULL. If it is, an immediate error is
1120 given.
1121
1122 If pcre_compile2() is used instead of pcre_compile(), and the error-
1123 codeptr argument is not NULL, a non-zero error code number is returned
1124 via this argument in the event of an error. This is in addition to the
1125 textual error message. Error codes and messages are listed below.
1126
1127 If the final argument, tableptr, is NULL, PCRE uses a default set of
1128 character tables that are built when PCRE is compiled, using the
1129 default C locale. Otherwise, tableptr must be an address that is the
1130 result of a call to pcre_maketables(). This value is stored with the
1131 compiled pattern, and used again by pcre_exec(), unless another table
1132 pointer is passed to it. For more discussion, see the section on locale
1133 support below.
1134
1135 This code fragment shows a typical straightforward call to pcre_com-
1136 pile():
1137
1138 pcre *re;
1139 const char *error;
1140 int erroffset;
1141 re = pcre_compile(
1142 "^A.*Z", /* the pattern */
1143 0, /* default options */
1144 &error, /* for error message */
1145 &erroffset, /* for error offset */
1146 NULL); /* use default character tables */
1147
1148 The following names for option bits are defined in the pcre.h header
1149 file:
1150
1151 PCRE_ANCHORED
1152
1153 If this bit is set, the pattern is forced to be "anchored", that is, it
1154 is constrained to match only at the first matching point in the string
1155 that is being searched (the "subject string"). This effect can also be
1156 achieved by appropriate constructs in the pattern itself, which is the
1157 only way to do it in Perl.
1158
1159 PCRE_AUTO_CALLOUT
1160
1161 If this bit is set, pcre_compile() automatically inserts callout items,
1162 all with number 255, before each pattern item. For discussion of the
1163 callout facility, see the pcrecallout documentation.
1164
1165 PCRE_BSR_ANYCRLF
1166 PCRE_BSR_UNICODE
1167
1168 These options (which are mutually exclusive) control what the \R escape
1169 sequence matches. The choice is either to match only CR, LF, or CRLF,
1170 or to match any Unicode newline sequence. The default is specified when
1171 PCRE is built. It can be overridden from within the pattern, or by set-
1172 ting an option when a compiled pattern is matched.
1173
1174 PCRE_CASELESS
1175
1176 If this bit is set, letters in the pattern match both upper and lower
1177 case letters. It is equivalent to Perl's /i option, and it can be
1178 changed within a pattern by a (?i) option setting. In UTF-8 mode, PCRE
1179 always understands the concept of case for characters whose values are
1180 less than 128, so caseless matching is always possible. For characters
1181 with higher values, the concept of case is supported if PCRE is com-
1182 piled with Unicode property support, but not otherwise. If you want to
1183 use caseless matching for characters 128 and above, you must ensure
1184 that PCRE is compiled with Unicode property support as well as with
1185 UTF-8 support.
1186
1187 PCRE_DOLLAR_ENDONLY
1188
1189 If this bit is set, a dollar metacharacter in the pattern matches only
1190 at the end of the subject string. Without this option, a dollar also
1191 matches immediately before a newline at the end of the string (but not
1192 before any other newlines). The PCRE_DOLLAR_ENDONLY option is ignored
1193 if PCRE_MULTILINE is set. There is no equivalent to this option in
1194 Perl, and no way to set it within a pattern.
1195
1196 PCRE_DOTALL
1197
1198 If this bit is set, a dot metacharater in the pattern matches all char-
1199 acters, including those that indicate newline. Without it, a dot does
1200 not match when the current position is at a newline. This option is
1201 equivalent to Perl's /s option, and it can be changed within a pattern
1202 by a (?s) option setting. A negative class such as [^a] always matches
1203 newline characters, independent of the setting of this option.
1204
1205 PCRE_DUPNAMES
1206
1207 If this bit is set, names used to identify capturing subpatterns need
1208 not be unique. This can be helpful for certain types of pattern when it
1209 is known that only one instance of the named subpattern can ever be
1210 matched. There are more details of named subpatterns below; see also
1211 the pcrepattern documentation.
1212
1213 PCRE_EXTENDED
1214
1215 If this bit is set, whitespace data characters in the pattern are
1216 totally ignored except when escaped or inside a character class. White-
1217 space does not include the VT character (code 11). In addition, charac-
1218 ters between an unescaped # outside a character class and the next new-
1219 line, inclusive, are also ignored. This is equivalent to Perl's /x
1220 option, and it can be changed within a pattern by a (?x) option set-
1221 ting.
1222
1223 This option makes it possible to include comments inside complicated
1224 patterns. Note, however, that this applies only to data characters.
1225 Whitespace characters may never appear within special character
1226 sequences in a pattern, for example within the sequence (?( which
1227 introduces a conditional subpattern.
1228
1229 PCRE_EXTRA
1230
1231 This option was invented in order to turn on additional functionality
1232 of PCRE that is incompatible with Perl, but it is currently of very
1233 little use. When set, any backslash in a pattern that is followed by a
1234 letter that has no special meaning causes an error, thus reserving
1235 these combinations for future expansion. By default, as in Perl, a
1236 backslash followed by a letter with no special meaning is treated as a
1237 literal. (Perl can, however, be persuaded to give a warning for this.)
1238 There are at present no other features controlled by this option. It
1239 can also be set by a (?X) option setting within a pattern.
1240
1241 PCRE_FIRSTLINE
1242
1243 If this option is set, an unanchored pattern is required to match
1244 before or at the first newline in the subject string, though the
1245 matched text may continue over the newline.
1246
1247 PCRE_MULTILINE
1248
1249 By default, PCRE treats the subject string as consisting of a single
1250 line of characters (even if it actually contains newlines). The "start
1251 of line" metacharacter (^) matches only at the start of the string,
1252 while the "end of line" metacharacter ($) matches only at the end of
1253 the string, or before a terminating newline (unless PCRE_DOLLAR_ENDONLY
1254 is set). This is the same as Perl.
1255
1256 When PCRE_MULTILINE it is set, the "start of line" and "end of line"
1257 constructs match immediately following or immediately before internal
1258 newlines in the subject string, respectively, as well as at the very
1259 start and end. This is equivalent to Perl's /m option, and it can be
1260 changed within a pattern by a (?m) option setting. If there are no new-
1261 lines in a subject string, or no occurrences of ^ or $ in a pattern,
1262 setting PCRE_MULTILINE has no effect.
1263
1264 PCRE_NEWLINE_CR
1265 PCRE_NEWLINE_LF
1266 PCRE_NEWLINE_CRLF
1267 PCRE_NEWLINE_ANYCRLF
1268 PCRE_NEWLINE_ANY
1269
1270 These options override the default newline definition that was chosen
1271 when PCRE was built. Setting the first or the second specifies that a
1272 newline is indicated by a single character (CR or LF, respectively).
1273 Setting PCRE_NEWLINE_CRLF specifies that a newline is indicated by the
1274 two-character CRLF sequence. Setting PCRE_NEWLINE_ANYCRLF specifies
1275 that any of the three preceding sequences should be recognized. Setting
1276 PCRE_NEWLINE_ANY specifies that any Unicode newline sequence should be
1277 recognized. The Unicode newline sequences are the three just mentioned,
1278 plus the single characters VT (vertical tab, U+000B), FF (formfeed,
1279 U+000C), NEL (next line, U+0085), LS (line separator, U+2028), and PS
1280 (paragraph separator, U+2029). The last two are recognized only in
1281 UTF-8 mode.
1282
1283 The newline setting in the options word uses three bits that are
1284 treated as a number, giving eight possibilities. Currently only six are
1285 used (default plus the five values above). This means that if you set
1286 more than one newline option, the combination may or may not be sensi-
1287 ble. For example, PCRE_NEWLINE_CR with PCRE_NEWLINE_LF is equivalent to
1288 PCRE_NEWLINE_CRLF, but other combinations may yield unused numbers and
1289 cause an error.
1290
1291 The only time that a line break is specially recognized when compiling
1292 a pattern is if PCRE_EXTENDED is set, and an unescaped # outside a
1293 character class is encountered. This indicates a comment that lasts
1294 until after the next line break sequence. In other circumstances, line
1295 break sequences are treated as literal data, except that in
1296 PCRE_EXTENDED mode, both CR and LF are treated as whitespace characters
1297 and are therefore ignored.
1298
1299 The newline option that is set at compile time becomes the default that
1300 is used for pcre_exec() and pcre_dfa_exec(), but it can be overridden.
1301
1302 PCRE_NO_AUTO_CAPTURE
1303
1304 If this option is set, it disables the use of numbered capturing paren-
1305 theses in the pattern. Any opening parenthesis that is not followed by
1306 ? behaves as if it were followed by ?: but named parentheses can still
1307 be used for capturing (and they acquire numbers in the usual way).
1308 There is no equivalent of this option in Perl.
1309
1310 PCRE_UNGREEDY
1311
1312 This option inverts the "greediness" of the quantifiers so that they
1313 are not greedy by default, but become greedy if followed by "?". It is
1314 not compatible with Perl. It can also be set by a (?U) option setting
1315 within the pattern.
1316
1317 PCRE_UTF8
1318
1319 This option causes PCRE to regard both the pattern and the subject as
1320 strings of UTF-8 characters instead of single-byte character strings.
1321 However, it is available only when PCRE is built to include UTF-8 sup-
1322 port. If not, the use of this option provokes an error. Details of how
1323 this option changes the behaviour of PCRE are given in the section on
1324 UTF-8 support in the main pcre page.
1325
1326 PCRE_NO_UTF8_CHECK
1327
1328 When PCRE_UTF8 is set, the validity of the pattern as a UTF-8 string is
1329 automatically checked. There is a discussion about the validity of
1330 UTF-8 strings in the main pcre page. If an invalid UTF-8 sequence of
1331 bytes is found, pcre_compile() returns an error. If you already know
1332 that your pattern is valid, and you want to skip this check for perfor-
1333 mance reasons, you can set the PCRE_NO_UTF8_CHECK option. When it is
1334 set, the effect of passing an invalid UTF-8 string as a pattern is
1335 undefined. It may cause your program to crash. Note that this option
1336 can also be passed to pcre_exec() and pcre_dfa_exec(), to suppress the
1337 UTF-8 validity checking of subject strings.
1338
1339
1340 COMPILATION ERROR CODES
1341
1342 The following table lists the error codes than may be returned by
1343 pcre_compile2(), along with the error messages that may be returned by
1344 both compiling functions. As PCRE has developed, some error codes have
1345 fallen out of use. To avoid confusion, they have not been re-used.
1346
1347 0 no error
1348 1 \ at end of pattern
1349 2 \c at end of pattern
1350 3 unrecognized character follows \
1351 4 numbers out of order in {} quantifier
1352 5 number too big in {} quantifier
1353 6 missing terminating ] for character class
1354 7 invalid escape sequence in character class
1355 8 range out of order in character class
1356 9 nothing to repeat
1357 10 [this code is not in use]
1358 11 internal error: unexpected repeat
1359 12 unrecognized character after (?
1360 13 POSIX named classes are supported only within a class
1361 14 missing )
1362 15 reference to non-existent subpattern
1363 16 erroffset passed as NULL
1364 17 unknown option bit(s) set
1365 18 missing ) after comment
1366 19 [this code is not in use]
1367 20 regular expression too large
1368 21 failed to get memory
1369 22 unmatched parentheses
1370 23 internal error: code overflow
1371 24 unrecognized character after (?<
1372 25 lookbehind assertion is not fixed length
1373 26 malformed number or name after (?(
1374 27 conditional group contains more than two branches
1375 28 assertion expected after (?(
1376 29 (?R or (?[+-]digits must be followed by )
1377 30 unknown POSIX class name
1378 31 POSIX collating elements are not supported
1379 32 this version of PCRE is not compiled with PCRE_UTF8 support
1380 33 [this code is not in use]
1381 34 character value in \x{...} sequence is too large
1382 35 invalid condition (?(0)
1383 36 \C not allowed in lookbehind assertion
1384 37 PCRE does not support \L, \l, \N, \U, or \u
1385 38 number after (?C is > 255
1386 39 closing ) for (?C expected
1387 40 recursive call could loop indefinitely
1388 41 unrecognized character after (?P
1389 42 syntax error in subpattern name (missing terminator)
1390 43 two named subpatterns have the same name
1391 44 invalid UTF-8 string
1392 45 support for \P, \p, and \X has not been compiled
1393 46 malformed \P or \p sequence
1394 47 unknown property name after \P or \p
1395 48 subpattern name is too long (maximum 32 characters)
1396 49 too many named subpatterns (maximum 10,000)
1397 50 [this code is not in use]
1398 51 octal value is greater than \377 (not in UTF-8 mode)
1399 52 internal error: overran compiling workspace
1400 53 internal error: previously-checked referenced subpattern not
1401 found
1402 54 DEFINE group contains more than one branch
1403 55 repeating a DEFINE group is not allowed
1404 56 inconsistent NEWLINE options
1405 57 \g is not followed by a braced name or an optionally braced
1406 non-zero number
1407 58 (?+ or (?- or (?(+ or (?(- must be followed by a non-zero number
1408
1409
1410 STUDYING A PATTERN
1411
1412 pcre_extra *pcre_study(const pcre *code, int options
1413 const char **errptr);
1414
1415 If a compiled pattern is going to be used several times, it is worth
1416 spending more time analyzing it in order to speed up the time taken for
1417 matching. The function pcre_study() takes a pointer to a compiled pat-
1418 tern as its first argument. If studying the pattern produces additional
1419 information that will help speed up matching, pcre_study() returns a
1420 pointer to a pcre_extra block, in which the study_data field points to
1421 the results of the study.
1422
1423 The returned value from pcre_study() can be passed directly to
1424 pcre_exec(). However, a pcre_extra block also contains other fields
1425 that can be set by the caller before the block is passed; these are
1426 described below in the section on matching a pattern.
1427
1428 If studying the pattern does not produce any additional information
1429 pcre_study() returns NULL. In that circumstance, if the calling program
1430 wants to pass any of the other fields to pcre_exec(), it must set up
1431 its own pcre_extra block.
1432
1433 The second argument of pcre_study() contains option bits. At present,
1434 no options are defined, and this argument should always be zero.
1435
1436 The third argument for pcre_study() is a pointer for an error message.
1437 If studying succeeds (even if no data is returned), the variable it
1438 points to is set to NULL. Otherwise it is set to point to a textual
1439 error message. This is a static string that is part of the library. You
1440 must not try to free it. You should test the error pointer for NULL
1441 after calling pcre_study(), to be sure that it has run successfully.
1442
1443 This is a typical call to pcre_study():
1444
1445 pcre_extra *pe;
1446 pe = pcre_study(
1447 re, /* result of pcre_compile() */
1448 0, /* no options exist */
1449 &error); /* set to NULL or points to a message */
1450
1451 At present, studying a pattern is useful only for non-anchored patterns
1452 that do not have a single fixed starting character. A bitmap of possi-
1453 ble starting bytes is created.
1454
1455
1456 LOCALE SUPPORT
1457
1458 PCRE handles caseless matching, and determines whether characters are
1459 letters, digits, or whatever, by reference to a set of tables, indexed
1460 by character value. When running in UTF-8 mode, this applies only to
1461 characters with codes less than 128. Higher-valued codes never match
1462 escapes such as \w or \d, but can be tested with \p if PCRE is built
1463 with Unicode character property support. The use of locales with Uni-
1464 code is discouraged. If you are handling characters with codes greater
1465 than 128, you should either use UTF-8 and Unicode, or use locales, but
1466 not try to mix the two.
1467
1468 PCRE contains an internal set of tables that are used when the final
1469 argument of pcre_compile() is NULL. These are sufficient for many
1470 applications. Normally, the internal tables recognize only ASCII char-
1471 acters. However, when PCRE is built, it is possible to cause the inter-
1472 nal tables to be rebuilt in the default "C" locale of the local system,
1473 which may cause them to be different.
1474
1475 The internal tables can always be overridden by tables supplied by the
1476 application that calls PCRE. These may be created in a different locale
1477 from the default. As more and more applications change to using Uni-
1478 code, the need for this locale support is expected to die away.
1479
1480 External tables are built by calling the pcre_maketables() function,
1481 which has no arguments, in the relevant locale. The result can then be
1482 passed to pcre_compile() or pcre_exec() as often as necessary. For
1483 example, to build and use tables that are appropriate for the French
1484 locale (where accented characters with values greater than 128 are
1485 treated as letters), the following code could be used:
1486
1487 setlocale(LC_CTYPE, "fr_FR");
1488 tables = pcre_maketables();
1489 re = pcre_compile(..., tables);
1490
1491 The locale name "fr_FR" is used on Linux and other Unix-like systems;
1492 if you are using Windows, the name for the French locale is "french".
1493
1494 When pcre_maketables() runs, the tables are built in memory that is
1495 obtained via pcre_malloc. It is the caller's responsibility to ensure
1496 that the memory containing the tables remains available for as long as
1497 it is needed.
1498
1499 The pointer that is passed to pcre_compile() is saved with the compiled
1500 pattern, and the same tables are used via this pointer by pcre_study()
1501 and normally also by pcre_exec(). Thus, by default, for any single pat-
1502 tern, compilation, studying and matching all happen in the same locale,
1503 but different patterns can be compiled in different locales.
1504
1505 It is possible to pass a table pointer or NULL (indicating the use of
1506 the internal tables) to pcre_exec(). Although not intended for this
1507 purpose, this facility could be used to match a pattern in a different
1508 locale from the one in which it was compiled. Passing table pointers at
1509 run time is discussed below in the section on matching a pattern.
1510
1511
1512 INFORMATION ABOUT A PATTERN
1513
1514 int pcre_fullinfo(const pcre *code, const pcre_extra *extra,
1515 int what, void *where);
1516
1517 The pcre_fullinfo() function returns information about a compiled pat-
1518 tern. It replaces the obsolete pcre_info() function, which is neverthe-
1519 less retained for backwards compability (and is documented below).
1520
1521 The first argument for pcre_fullinfo() is a pointer to the compiled
1522 pattern. The second argument is the result of pcre_study(), or NULL if
1523 the pattern was not studied. The third argument specifies which piece
1524 of information is required, and the fourth argument is a pointer to a
1525 variable to receive the data. The yield of the function is zero for
1526 success, or one of the following negative numbers:
1527
1528 PCRE_ERROR_NULL the argument code was NULL
1529 the argument where was NULL
1530 PCRE_ERROR_BADMAGIC the "magic number" was not found
1531 PCRE_ERROR_BADOPTION the value of what was invalid
1532
1533 The "magic number" is placed at the start of each compiled pattern as
1534 an simple check against passing an arbitrary memory pointer. Here is a
1535 typical call of pcre_fullinfo(), to obtain the length of the compiled
1536 pattern:
1537
1538 int rc;
1539 size_t length;
1540 rc = pcre_fullinfo(
1541 re, /* result of pcre_compile() */
1542 pe, /* result of pcre_study(), or NULL */
1543 PCRE_INFO_SIZE, /* what is required */
1544 &length); /* where to put the data */
1545
1546 The possible values for the third argument are defined in pcre.h, and
1547 are as follows:
1548
1549 PCRE_INFO_BACKREFMAX
1550
1551 Return the number of the highest back reference in the pattern. The
1552 fourth argument should point to an int variable. Zero is returned if
1553 there are no back references.
1554
1555 PCRE_INFO_CAPTURECOUNT
1556
1557 Return the number of capturing subpatterns in the pattern. The fourth
1558 argument should point to an int variable.
1559
1560 PCRE_INFO_DEFAULT_TABLES
1561
1562 Return a pointer to the internal default character tables within PCRE.
1563 The fourth argument should point to an unsigned char * variable. This
1564 information call is provided for internal use by the pcre_study() func-
1565 tion. External callers can cause PCRE to use its internal tables by
1566 passing a NULL table pointer.
1567
1568 PCRE_INFO_FIRSTBYTE
1569
1570 Return information about the first byte of any matched string, for a
1571 non-anchored pattern. The fourth argument should point to an int vari-
1572 able. (This option used to be called PCRE_INFO_FIRSTCHAR; the old name
1573 is still recognized for backwards compatibility.)
1574
1575 If there is a fixed first byte, for example, from a pattern such as
1576 (cat|cow|coyote), its value is returned. Otherwise, if either
1577
1578 (a) the pattern was compiled with the PCRE_MULTILINE option, and every
1579 branch starts with "^", or
1580
1581 (b) every branch of the pattern starts with ".*" and PCRE_DOTALL is not
1582 set (if it were set, the pattern would be anchored),
1583
1584 -1 is returned, indicating that the pattern matches only at the start
1585 of a subject string or after any newline within the string. Otherwise
1586 -2 is returned. For anchored patterns, -2 is returned.
1587
1588 PCRE_INFO_FIRSTTABLE
1589
1590 If the pattern was studied, and this resulted in the construction of a
1591 256-bit table indicating a fixed set of bytes for the first byte in any
1592 matching string, a pointer to the table is returned. Otherwise NULL is
1593 returned. The fourth argument should point to an unsigned char * vari-
1594 able.
1595
1596 PCRE_INFO_HASCRORLF
1597
1598 Return 1 if the pattern contains any explicit matches for CR or LF
1599 characters, otherwise 0. The fourth argument should point to an int
1600 variable. An explicit match is either a literal CR or LF character, or
1601 \r or \n.
1602
1603 PCRE_INFO_JCHANGED
1604
1605 Return 1 if the (?J) or (?-J) option setting is used in the pattern,
1606 otherwise 0. The fourth argument should point to an int variable. (?J)
1607 and (?-J) set and unset the local PCRE_DUPNAMES option, respectively.
1608
1609 PCRE_INFO_LASTLITERAL
1610
1611 Return the value of the rightmost literal byte that must exist in any
1612 matched string, other than at its start, if such a byte has been
1613 recorded. The fourth argument should point to an int variable. If there
1614 is no such byte, -1 is returned. For anchored patterns, a last literal
1615 byte is recorded only if it follows something of variable length. For
1616 example, for the pattern /^a\d+z\d+/ the returned value is "z", but for
1617 /^a\dz\d/ the returned value is -1.
1618
1619 PCRE_INFO_NAMECOUNT
1620 PCRE_INFO_NAMEENTRYSIZE
1621 PCRE_INFO_NAMETABLE
1622
1623 PCRE supports the use of named as well as numbered capturing parenthe-
1624 ses. The names are just an additional way of identifying the parenthe-
1625 ses, which still acquire numbers. Several convenience functions such as
1626 pcre_get_named_substring() are provided for extracting captured sub-
1627 strings by name. It is also possible to extract the data directly, by
1628 first converting the name to a number in order to access the correct
1629 pointers in the output vector (described with pcre_exec() below). To do
1630 the conversion, you need to use the name-to-number map, which is
1631 described by these three values.
1632
1633 The map consists of a number of fixed-size entries. PCRE_INFO_NAMECOUNT
1634 gives the number of entries, and PCRE_INFO_NAMEENTRYSIZE gives the size
1635 of each entry; both of these return an int value. The entry size
1636 depends on the length of the longest name. PCRE_INFO_NAMETABLE returns
1637 a pointer to the first entry of the table (a pointer to char). The
1638 first two bytes of each entry are the number of the capturing parenthe-
1639 sis, most significant byte first. The rest of the entry is the corre-
1640 sponding name, zero terminated. The names are in alphabetical order.
1641 When PCRE_DUPNAMES is set, duplicate names are in order of their paren-
1642 theses numbers. For example, consider the following pattern (assume
1643 PCRE_EXTENDED is set, so white space - including newlines - is
1644 ignored):
1645
1646 (?<date> (?<year>(\d\d)?\d\d) -
1647 (?<month>\d\d) - (?<day>\d\d) )
1648
1649 There are four named subpatterns, so the table has four entries, and
1650 each entry in the table is eight bytes long. The table is as follows,
1651 with non-printing bytes shows in hexadecimal, and undefined bytes shown
1652 as ??:
1653
1654 00 01 d a t e 00 ??
1655 00 05 d a y 00 ?? ??
1656 00 04 m o n t h 00
1657 00 02 y e a r 00 ??
1658
1659 When writing code to extract data from named subpatterns using the
1660 name-to-number map, remember that the length of the entries is likely
1661 to be different for each compiled pattern.
1662
1663 PCRE_INFO_OKPARTIAL
1664
1665 Return 1 if the pattern can be used for partial matching, otherwise 0.
1666 The fourth argument should point to an int variable. The pcrepartial
1667 documentation lists the restrictions that apply to patterns when par-
1668 tial matching is used.
1669
1670 PCRE_INFO_OPTIONS
1671
1672 Return a copy of the options with which the pattern was compiled. The
1673 fourth argument should point to an unsigned long int variable. These
1674 option bits are those specified in the call to pcre_compile(), modified
1675 by any top-level option settings at the start of the pattern itself. In
1676 other words, they are the options that will be in force when matching
1677 starts. For example, if the pattern /(?im)abc(?-i)d/ is compiled with
1678 the PCRE_EXTENDED option, the result is PCRE_CASELESS, PCRE_MULTILINE,
1679 and PCRE_EXTENDED.
1680
1681 A pattern is automatically anchored by PCRE if all of its top-level
1682 alternatives begin with one of the following:
1683
1684 ^ unless PCRE_MULTILINE is set
1685 \A always
1686 \G always
1687 .* if PCRE_DOTALL is set and there are no back
1688 references to the subpattern in which .* appears
1689
1690 For such patterns, the PCRE_ANCHORED bit is set in the options returned
1691 by pcre_fullinfo().
1692
1693 PCRE_INFO_SIZE
1694
1695 Return the size of the compiled pattern, that is, the value that was
1696 passed as the argument to pcre_malloc() when PCRE was getting memory in
1697 which to place the compiled data. The fourth argument should point to a
1698 size_t variable.
1699
1700 PCRE_INFO_STUDYSIZE
1701
1702 Return the size of the data block pointed to by the study_data field in
1703 a pcre_extra block. That is, it is the value that was passed to
1704 pcre_malloc() when PCRE was getting memory into which to place the data
1705 created by pcre_study(). The fourth argument should point to a size_t
1706 variable.
1707
1708
1709 OBSOLETE INFO FUNCTION
1710
1711 int pcre_info(const pcre *code, int *optptr, int *firstcharptr);
1712
1713 The pcre_info() function is now obsolete because its interface is too
1714 restrictive to return all the available data about a compiled pattern.
1715 New programs should use pcre_fullinfo() instead. The yield of
1716 pcre_info() is the number of capturing subpatterns, or one of the fol-
1717 lowing negative numbers:
1718
1719 PCRE_ERROR_NULL the argument code was NULL
1720 PCRE_ERROR_BADMAGIC the "magic number" was not found
1721
1722 If the optptr argument is not NULL, a copy of the options with which
1723 the pattern was compiled is placed in the integer it points to (see
1724 PCRE_INFO_OPTIONS above).
1725
1726 If the pattern is not anchored and the firstcharptr argument is not
1727 NULL, it is used to pass back information about the first character of
1728 any matched string (see PCRE_INFO_FIRSTBYTE above).
1729
1730
1731 REFERENCE COUNTS
1732
1733 int pcre_refcount(pcre *code, int adjust);
1734
1735 The pcre_refcount() function is used to maintain a reference count in
1736 the data block that contains a compiled pattern. It is provided for the
1737 benefit of applications that operate in an object-oriented manner,
1738 where different parts of the application may be using the same compiled
1739 pattern, but you want to free the block when they are all done.
1740
1741 When a pattern is compiled, the reference count field is initialized to
1742 zero. It is changed only by calling this function, whose action is to
1743 add the adjust value (which may be positive or negative) to it. The
1744 yield of the function is the new value. However, the value of the count
1745 is constrained to lie between 0 and 65535, inclusive. If the new value
1746 is outside these limits, it is forced to the appropriate limit value.
1747
1748 Except when it is zero, the reference count is not correctly preserved
1749 if a pattern is compiled on one host and then transferred to a host
1750 whose byte-order is different. (This seems a highly unlikely scenario.)
1751
1752
1753 MATCHING A PATTERN: THE TRADITIONAL FUNCTION
1754
1755 int pcre_exec(const pcre *code, const pcre_extra *extra,
1756 const char *subject, int length, int startoffset,
1757 int options, int *ovector, int ovecsize);
1758
1759 The function pcre_exec() is called to match a subject string against a
1760 compiled pattern, which is passed in the code argument. If the pattern
1761 has been studied, the result of the study should be passed in the extra
1762 argument. This function is the main matching facility of the library,
1763 and it operates in a Perl-like manner. For specialist use there is also
1764 an alternative matching function, which is described below in the sec-
1765 tion about the pcre_dfa_exec() function.
1766
1767 In most applications, the pattern will have been compiled (and option-
1768 ally studied) in the same process that calls pcre_exec(). However, it
1769 is possible to save compiled patterns and study data, and then use them
1770 later in different processes, possibly even on different hosts. For a
1771 discussion about this, see the pcreprecompile documentation.
1772
1773 Here is an example of a simple call to pcre_exec():
1774
1775 int rc;
1776 int ovector[30];
1777 rc = pcre_exec(
1778 re, /* result of pcre_compile() */
1779 NULL, /* we didn't study the pattern */
1780 "some string", /* the subject string */
1781 11, /* the length of the subject string */
1782 0, /* start at offset 0 in the subject */
1783 0, /* default options */
1784 ovector, /* vector of integers for substring information */
1785 30); /* number of elements (NOT size in bytes) */
1786
1787 Extra data for pcre_exec()
1788
1789 If the extra argument is not NULL, it must point to a pcre_extra data
1790 block. The pcre_study() function returns such a block (when it doesn't
1791 return NULL), but you can also create one for yourself, and pass addi-
1792 tional information in it. The pcre_extra block contains the following
1793 fields (not necessarily in this order):
1794
1795 unsigned long int flags;
1796 void *study_data;
1797 unsigned long int match_limit;
1798 unsigned long int match_limit_recursion;
1799 void *callout_data;
1800 const unsigned char *tables;
1801
1802 The flags field is a bitmap that specifies which of the other fields
1803 are set. The flag bits are:
1804
1805 PCRE_EXTRA_STUDY_DATA
1806 PCRE_EXTRA_MATCH_LIMIT
1807 PCRE_EXTRA_MATCH_LIMIT_RECURSION
1808 PCRE_EXTRA_CALLOUT_DATA
1809 PCRE_EXTRA_TABLES
1810
1811 Other flag bits should be set to zero. The study_data field is set in
1812 the pcre_extra block that is returned by pcre_study(), together with
1813 the appropriate flag bit. You should not set this yourself, but you may
1814 add to the block by setting the other fields and their corresponding
1815 flag bits.
1816
1817 The match_limit field provides a means of preventing PCRE from using up
1818 a vast amount of resources when running patterns that are not going to
1819 match, but which have a very large number of possibilities in their
1820 search trees. The classic example is the use of nested unlimited
1821 repeats.
1822
1823 Internally, PCRE uses a function called match() which it calls repeat-
1824 edly (sometimes recursively). The limit set by match_limit is imposed
1825 on the number of times this function is called during a match, which
1826 has the effect of limiting the amount of backtracking that can take
1827 place. For patterns that are not anchored, the count restarts from zero
1828 for each position in the subject string.
1829
1830 The default value for the limit can be set when PCRE is built; the
1831 default default is 10 million, which handles all but the most extreme
1832 cases. You can override the default by suppling pcre_exec() with a
1833 pcre_extra block in which match_limit is set, and
1834 PCRE_EXTRA_MATCH_LIMIT is set in the flags field. If the limit is
1835 exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
1836
1837 The match_limit_recursion field is similar to match_limit, but instead
1838 of limiting the total number of times that match() is called, it limits
1839 the depth of recursion. The recursion depth is a smaller number than
1840 the total number of calls, because not all calls to match() are recur-
1841 sive. This limit is of use only if it is set smaller than match_limit.
1842
1843 Limiting the recursion depth limits the amount of stack that can be
1844 used, or, when PCRE has been compiled to use memory on the heap instead
1845 of the stack, the amount of heap memory that can be used.
1846
1847 The default value for match_limit_recursion can be set when PCRE is
1848 built; the default default is the same value as the default for
1849 match_limit. You can override the default by suppling pcre_exec() with
1850 a pcre_extra block in which match_limit_recursion is set, and
1851 PCRE_EXTRA_MATCH_LIMIT_RECURSION is set in the flags field. If the
1852 limit is exceeded, pcre_exec() returns PCRE_ERROR_RECURSIONLIMIT.
1853
1854 The pcre_callout field is used in conjunction with the "callout" fea-
1855 ture, which is described in the pcrecallout documentation.
1856
1857 The tables field is used to pass a character tables pointer to
1858 pcre_exec(); this overrides the value that is stored with the compiled
1859 pattern. A non-NULL value is stored with the compiled pattern only if
1860 custom tables were supplied to pcre_compile() via its tableptr argu-
1861 ment. If NULL is passed to pcre_exec() using this mechanism, it forces
1862 PCRE's internal tables to be used. This facility is helpful when re-
1863 using patterns that have been saved after compiling with an external
1864 set of tables, because the external tables might be at a different
1865 address when pcre_exec() is called. See the pcreprecompile documenta-
1866 tion for a discussion of saving compiled patterns for later use.
1867
1868 Option bits for pcre_exec()
1869
1870 The unused bits of the options argument for pcre_exec() must be zero.
1871 The only bits that may be set are PCRE_ANCHORED, PCRE_NEWLINE_xxx,
1872 PCRE_NOTBOL, PCRE_NOTEOL, PCRE_NOTEMPTY, PCRE_NO_UTF8_CHECK and
1873 PCRE_PARTIAL.
1874
1875 PCRE_ANCHORED
1876
1877 The PCRE_ANCHORED option limits pcre_exec() to matching at the first
1878 matching position. If a pattern was compiled with PCRE_ANCHORED, or
1879 turned out to be anchored by virtue of its contents, it cannot be made
1880 unachored at matching time.
1881
1882 PCRE_BSR_ANYCRLF
1883 PCRE_BSR_UNICODE
1884
1885 These options (which are mutually exclusive) control what the \R escape
1886 sequence matches. The choice is either to match only CR, LF, or CRLF,
1887 or to match any Unicode newline sequence. These options override the
1888 choice that was made or defaulted when the pattern was compiled.
1889
1890 PCRE_NEWLINE_CR
1891 PCRE_NEWLINE_LF
1892 PCRE_NEWLINE_CRLF
1893 PCRE_NEWLINE_ANYCRLF
1894 PCRE_NEWLINE_ANY
1895
1896 These options override the newline definition that was chosen or
1897 defaulted when the pattern was compiled. For details, see the descrip-
1898 tion of pcre_compile() above. During matching, the newline choice
1899 affects the behaviour of the dot, circumflex, and dollar metacharac-
1900 ters. It may also alter the way the match position is advanced after a
1901 match failure for an unanchored pattern.
1902
1903 When PCRE_NEWLINE_CRLF, PCRE_NEWLINE_ANYCRLF, or PCRE_NEWLINE_ANY is
1904 set, and a match attempt for an unanchored pattern fails when the cur-
1905 rent position is at a CRLF sequence, and the pattern contains no
1906 explicit matches for CR or LF characters, the match position is
1907 advanced by two characters instead of one, in other words, to after the
1908 CRLF.
1909
1910 The above rule is a compromise that makes the most common cases work as
1911 expected. For example, if the pattern is .+A (and the PCRE_DOTALL
1912 option is not set), it does not match the string "\r\nA" because, after
1913 failing at the start, it skips both the CR and the LF before retrying.
1914 However, the pattern [\r\n]A does match that string, because it con-
1915 tains an explicit CR or LF reference, and so advances only by one char-
1916 acter after the first failure.
1917
1918 An explicit match for CR of LF is either a literal appearance of one of
1919 those characters, or one of the \r or \n escape sequences. Implicit
1920 matches such as [^X] do not count, nor does \s (which includes CR and
1921 LF in the characters that it matches).
1922
1923 Notwithstanding the above, anomalous effects may still occur when CRLF
1924 is a valid newline sequence and explicit \r or \n escapes appear in the
1925 pattern.
1926
1927 PCRE_NOTBOL
1928
1929 This option specifies that first character of the subject string is not
1930 the beginning of a line, so the circumflex metacharacter should not
1931 match before it. Setting this without PCRE_MULTILINE (at compile time)
1932 causes circumflex never to match. This option affects only the behav-
1933 iour of the circumflex metacharacter. It does not affect \A.
1934
1935 PCRE_NOTEOL
1936
1937 This option specifies that the end of the subject string is not the end
1938 of a line, so the dollar metacharacter should not match it nor (except
1939 in multiline mode) a newline immediately before it. Setting this with-
1940 out PCRE_MULTILINE (at compile time) causes dollar never to match. This
1941 option affects only the behaviour of the dollar metacharacter. It does
1942 not affect \Z or \z.
1943
1944 PCRE_NOTEMPTY
1945
1946 An empty string is not considered to be a valid match if this option is
1947 set. If there are alternatives in the pattern, they are tried. If all
1948 the alternatives match the empty string, the entire match fails. For
1949 example, if the pattern
1950
1951 a?b?
1952
1953 is applied to a string not beginning with "a" or "b", it matches the
1954 empty string at the start of the subject. With PCRE_NOTEMPTY set, this
1955 match is not valid, so PCRE searches further into the string for occur-
1956 rences of "a" or "b".
1957
1958 Perl has no direct equivalent of PCRE_NOTEMPTY, but it does make a spe-
1959 cial case of a pattern match of the empty string within its split()
1960 function, and when using the /g modifier. It is possible to emulate
1961 Perl's behaviour after matching a null string by first trying the match
1962 again at the same offset with PCRE_NOTEMPTY and PCRE_ANCHORED, and then
1963 if that fails by advancing the starting offset (see below) and trying
1964 an ordinary match again. There is some code that demonstrates how to do
1965 this in the pcredemo.c sample program.
1966
1967 PCRE_NO_UTF8_CHECK
1968
1969 When PCRE_UTF8 is set at compile time, the validity of the subject as a
1970 UTF-8 string is automatically checked when pcre_exec() is subsequently
1971 called. The value of startoffset is also checked to ensure that it
1972 points to the start of a UTF-8 character. There is a discussion about
1973 the validity of UTF-8 strings in the section on UTF-8 support in the
1974 main pcre page. If an invalid UTF-8 sequence of bytes is found,
1975 pcre_exec() returns the error PCRE_ERROR_BADUTF8. If startoffset con-
1976 tains an invalid value, PCRE_ERROR_BADUTF8_OFFSET is returned.
1977
1978 If you already know that your subject is valid, and you want to skip
1979 these checks for performance reasons, you can set the
1980 PCRE_NO_UTF8_CHECK option when calling pcre_exec(). You might want to
1981 do this for the second and subsequent calls to pcre_exec() if you are
1982 making repeated calls to find all the matches in a single subject
1983 string. However, you should be sure that the value of startoffset
1984 points to the start of a UTF-8 character. When PCRE_NO_UTF8_CHECK is
1985 set, the effect of passing an invalid UTF-8 string as a subject, or a
1986 value of startoffset that does not point to the start of a UTF-8 char-
1987 acter, is undefined. Your program may crash.
1988
1989 PCRE_PARTIAL
1990
1991 This option turns on the partial matching feature. If the subject
1992 string fails to match the pattern, but at some point during the match-
1993 ing process the end of the subject was reached (that is, the subject
1994 partially matches the pattern and the failure to match occurred only
1995 because there were not enough subject characters), pcre_exec() returns
1996 PCRE_ERROR_PARTIAL instead of PCRE_ERROR_NOMATCH. When PCRE_PARTIAL is
1997 used, there are restrictions on what may appear in the pattern. These
1998 are discussed in the pcrepartial documentation.
1999
2000 The string to be matched by pcre_exec()
2001
2002 The subject string is passed to pcre_exec() as a pointer in subject, a
2003 length in length, and a starting byte offset in startoffset. In UTF-8
2004 mode, the byte offset must point to the start of a UTF-8 character.
2005 Unlike the pattern string, the subject may contain binary zero bytes.
2006 When the starting offset is zero, the search for a match starts at the
2007 beginning of the subject, and this is by far the most common case.
2008
2009 A non-zero starting offset is useful when searching for another match
2010 in the same subject by calling pcre_exec() again after a previous suc-
2011 cess. Setting startoffset differs from just passing over a shortened
2012 string and setting PCRE_NOTBOL in the case of a pattern that begins
2013 with any kind of lookbehind. For example, consider the pattern
2014
2015 \Biss\B
2016
2017 which finds occurrences of "iss" in the middle of words. (\B matches
2018 only if the current position in the subject is not a word boundary.)
2019 When applied to the string "Mississipi" the first call to pcre_exec()
2020 finds the first occurrence. If pcre_exec() is called again with just
2021 the remainder of the subject, namely "issipi", it does not match,
2022 because \B is always false at the start of the subject, which is deemed
2023 to be a word boundary. However, if pcre_exec() is passed the entire
2024 string again, but with startoffset set to 4, it finds the second occur-
2025 rence of "iss" because it is able to look behind the starting point to
2026 discover that it is preceded by a letter.
2027
2028 If a non-zero starting offset is passed when the pattern is anchored,
2029 one attempt to match at the given offset is made. This can only succeed
2030 if the pattern does not require the match to be at the start of the
2031 subject.
2032
2033 How pcre_exec() returns captured substrings
2034
2035 In general, a pattern matches a certain portion of the subject, and in
2036 addition, further substrings from the subject may be picked out by
2037 parts of the pattern. Following the usage in Jeffrey Friedl's book,
2038 this is called "capturing" in what follows, and the phrase "capturing
2039 subpattern" is used for a fragment of a pattern that picks out a sub-
2040 string. PCRE supports several other kinds of parenthesized subpattern
2041 that do not cause substrings to be captured.
2042
2043 Captured substrings are returned to the caller via a vector of integer
2044 offsets whose address is passed in ovector. The number of elements in
2045 the vector is passed in ovecsize, which must be a non-negative number.
2046 Note: this argument is NOT the size of ovector in bytes.
2047
2048 The first two-thirds of the vector is used to pass back captured sub-
2049 strings, each substring using a pair of integers. The remaining third
2050 of the vector is used as workspace by pcre_exec() while matching cap-
2051 turing subpatterns, and is not available for passing back information.
2052 The length passed in ovecsize should always be a multiple of three. If
2053 it is not, it is rounded down.
2054
2055 When a match is successful, information about captured substrings is
2056 returned in pairs of integers, starting at the beginning of ovector,
2057 and continuing up to two-thirds of its length at the most. The first
2058 element of a pair is set to the offset of the first character in a sub-
2059 string, and the second is set to the offset of the first character
2060 after the end of a substring. The first pair, ovector[0] and ovec-
2061 tor[1], identify the portion of the subject string matched by the
2062 entire pattern. The next pair is used for the first capturing subpat-
2063 tern, and so on. The value returned by pcre_exec() is one more than the
2064 highest numbered pair that has been set. For example, if two substrings
2065 have been captured, the returned value is 3. If there are no capturing
2066 subpatterns, the return value from a successful match is 1, indicating
2067 that just the first pair of offsets has been set.
2068
2069 If a capturing subpattern is matched repeatedly, it is the last portion
2070 of the string that it matched that is returned.
2071
2072 If the vector is too small to hold all the captured substring offsets,
2073 it is used as far as possible (up to two-thirds of its length), and the
2074 function returns a value of zero. In particular, if the substring off-
2075 sets are not of interest, pcre_exec() may be called with ovector passed
2076 as NULL and ovecsize as zero. However, if the pattern contains back
2077 references and the ovector is not big enough to remember the related
2078 substrings, PCRE has to get additional memory for use during matching.
2079 Thus it is usually advisable to supply an ovector.
2080
2081 The pcre_info() function can be used to find out how many capturing
2082 subpatterns there are in a compiled pattern. The smallest size for
2083 ovector that will allow for n captured substrings, in addition to the
2084 offsets of the substring matched by the whole pattern, is (n+1)*3.
2085
2086 It is possible for capturing subpattern number n+1 to match some part
2087 of the subject when subpattern n has not been used at all. For example,
2088 if the string "abc" is matched against the pattern (a|(z))(bc) the
2089 return from the function is 4, and subpatterns 1 and 3 are matched, but
2090 2 is not. When this happens, both values in the offset pairs corre-
2091 sponding to unused subpatterns are set to -1.
2092
2093 Offset values that correspond to unused subpatterns at the end of the
2094 expression are also set to -1. For example, if the string "abc" is
2095 matched against the pattern (abc)(x(yz)?)? subpatterns 2 and 3 are not
2096 matched. The return from the function is 2, because the highest used
2097 capturing subpattern number is 1. However, you can refer to the offsets
2098 for the second and third capturing subpatterns if you wish (assuming
2099 the vector is large enough, of course).
2100
2101 Some convenience functions are provided for extracting the captured
2102 substrings as separate strings. These are described below.
2103
2104 Error return values from pcre_exec()
2105
2106 If pcre_exec() fails, it returns a negative number. The following are
2107 defined in the header file:
2108
2109 PCRE_ERROR_NOMATCH (-1)
2110
2111 The subject string did not match the pattern.
2112
2113 PCRE_ERROR_NULL (-2)
2114
2115 Either code or subject was passed as NULL, or ovector was NULL and
2116 ovecsize was not zero.
2117
2118 PCRE_ERROR_BADOPTION (-3)
2119
2120 An unrecognized bit was set in the options argument.
2121
2122 PCRE_ERROR_BADMAGIC (-4)
2123
2124 PCRE stores a 4-byte "magic number" at the start of the compiled code,
2125 to catch the case when it is passed a junk pointer and to detect when a
2126 pattern that was compiled in an environment of one endianness is run in
2127 an environment with the other endianness. This is the error that PCRE
2128 gives when the magic number is not present.
2129
2130 PCRE_ERROR_UNKNOWN_OPCODE (-5)
2131
2132 While running the pattern match, an unknown item was encountered in the
2133 compiled pattern. This error could be caused by a bug in PCRE or by
2134 overwriting of the compiled pattern.
2135
2136 PCRE_ERROR_NOMEMORY (-6)
2137
2138 If a pattern contains back references, but the ovector that is passed
2139 to pcre_exec() is not big enough to remember the referenced substrings,
2140 PCRE gets a block of memory at the start of matching to use for this
2141 purpose. If the call via pcre_malloc() fails, this error is given. The
2142 memory is automatically freed at the end of matching.
2143
2144 PCRE_ERROR_NOSUBSTRING (-7)
2145
2146 This error is used by the pcre_copy_substring(), pcre_get_substring(),
2147 and pcre_get_substring_list() functions (see below). It is never
2148 returned by pcre_exec().
2149
2150 PCRE_ERROR_MATCHLIMIT (-8)
2151
2152 The backtracking limit, as specified by the match_limit field in a
2153 pcre_extra structure (or defaulted) was reached. See the description
2154 above.
2155
2156 PCRE_ERROR_CALLOUT (-9)
2157
2158 This error is never generated by pcre_exec() itself. It is provided for
2159 use by callout functions that want to yield a distinctive error code.
2160 See the pcrecallout documentation for details.
2161
2162 PCRE_ERROR_BADUTF8 (-10)
2163
2164 A string that contains an invalid UTF-8 byte sequence was passed as a
2165 subject.
2166
2167 PCRE_ERROR_BADUTF8_OFFSET (-11)
2168
2169 The UTF-8 byte sequence that was passed as a subject was valid, but the
2170 value of startoffset did not point to the beginning of a UTF-8 charac-
2171 ter.
2172
2173 PCRE_ERROR_PARTIAL (-12)
2174
2175 The subject string did not match, but it did match partially. See the
2176 pcrepartial documentation for details of partial matching.
2177
2178 PCRE_ERROR_BADPARTIAL (-13)
2179
2180 The PCRE_PARTIAL option was used with a compiled pattern containing
2181 items that are not supported for partial matching. See the pcrepartial
2182 documentation for details of partial matching.
2183
2184 PCRE_ERROR_INTERNAL (-14)
2185
2186 An unexpected internal error has occurred. This error could be caused
2187 by a bug in PCRE or by overwriting of the compiled pattern.
2188
2189 PCRE_ERROR_BADCOUNT (-15)
2190
2191 This error is given if the value of the ovecsize argument is negative.
2192
2193 PCRE_ERROR_RECURSIONLIMIT (-21)
2194
2195 The internal recursion limit, as specified by the match_limit_recursion
2196 field in a pcre_extra structure (or defaulted) was reached. See the
2197 description above.
2198
2199 PCRE_ERROR_BADNEWLINE (-23)
2200
2201 An invalid combination of PCRE_NEWLINE_xxx options was given.
2202
2203 Error numbers -16 to -20 and -22 are not used by pcre_exec().
2204
2205
2206 EXTRACTING CAPTURED SUBSTRINGS BY NUMBER
2207
2208 int pcre_copy_substring(const char *subject, int *ovector,
2209 int stringcount, int stringnumber, char *buffer,
2210 int buffersize);
2211
2212 int pcre_get_substring(const char *subject, int *ovector,
2213 int stringcount, int stringnumber,
2214 const char **stringptr);
2215
2216 int pcre_get_substring_list(const char *subject,
2217 int *ovector, int stringcount, const char ***listptr);
2218
2219 Captured substrings can be accessed directly by using the offsets
2220 returned by pcre_exec() in ovector. For convenience, the functions
2221 pcre_copy_substring(), pcre_get_substring(), and pcre_get_sub-
2222 string_list() are provided for extracting captured substrings as new,
2223 separate, zero-terminated strings. These functions identify substrings
2224 by number. The next section describes functions for extracting named
2225 substrings.
2226
2227 A substring that contains a binary zero is correctly extracted and has
2228 a further zero added on the end, but the result is not, of course, a C
2229 string. However, you can process such a string by referring to the
2230 length that is returned by pcre_copy_substring() and pcre_get_sub-
2231 string(). Unfortunately, the interface to pcre_get_substring_list() is
2232 not adequate for handling strings containing binary zeros, because the
2233 end of the final string is not independently indicated.
2234
2235 The first three arguments are the same for all three of these func-
2236 tions: subject is the subject string that has just been successfully
2237 matched, ovector is a pointer to the vector of integer offsets that was
2238 passed to pcre_exec(), and stringcount is the number of substrings that
2239 were captured by the match, including the substring that matched the
2240 entire regular expression. This is the value returned by pcre_exec() if
2241 it is greater than zero. If pcre_exec() returned zero, indicating that
2242 it ran out of space in ovector, the value passed as stringcount should
2243 be the number of elements in the vector divided by three.
2244
2245 The functions pcre_copy_substring() and pcre_get_substring() extract a
2246 single substring, whose number is given as stringnumber. A value of
2247 zero extracts the substring that matched the entire pattern, whereas
2248 higher values extract the captured substrings. For pcre_copy_sub-
2249 string(), the string is placed in buffer, whose length is given by
2250 buffersize, while for pcre_get_substring() a new block of memory is
2251 obtained via pcre_malloc, and its address is returned via stringptr.
2252 The yield of the function is the length of the string, not including
2253 the terminating zero, or one of these error codes:
2254
2255 PCRE_ERROR_NOMEMORY (-6)
2256
2257 The buffer was too small for pcre_copy_substring(), or the attempt to
2258 get memory failed for pcre_get_substring().
2259
2260 PCRE_ERROR_NOSUBSTRING (-7)
2261
2262 There is no substring whose number is stringnumber.
2263
2264 The pcre_get_substring_list() function extracts all available sub-
2265 strings and builds a list of pointers to them. All this is done in a
2266 single block of memory that is obtained via pcre_malloc. The address of
2267 the memory block is returned via listptr, which is also the start of
2268 the list of string pointers. The end of the list is marked by a NULL
2269 pointer. The yield of the function is zero if all went well, or the
2270 error code
2271
2272 PCRE_ERROR_NOMEMORY (-6)
2273
2274 if the attempt to get the memory block failed.
2275
2276 When any of these functions encounter a substring that is unset, which
2277 can happen when capturing subpattern number n+1 matches some part of
2278 the subject, but subpattern n has not been used at all, they return an
2279 empty string. This can be distinguished from a genuine zero-length sub-
2280 string by inspecting the appropriate offset in ovector, which is nega-
2281 tive for unset substrings.
2282
2283 The two convenience functions pcre_free_substring() and pcre_free_sub-
2284 string_list() can be used to free the memory returned by a previous
2285 call of pcre_get_substring() or pcre_get_substring_list(), respec-
2286 tively. They do nothing more than call the function pointed to by
2287 pcre_free, which of course could be called directly from a C program.
2288 However, PCRE is used in some situations where it is linked via a spe-
2289 cial interface to another programming language that cannot use
2290 pcre_free directly; it is for these cases that the functions are pro-
2291 vided.
2292
2293
2294 EXTRACTING CAPTURED SUBSTRINGS BY NAME
2295
2296 int pcre_get_stringnumber(const pcre *code,
2297 const char *name);
2298
2299 int pcre_copy_named_substring(const pcre *code,
2300 const char *subject, int *ovector,
2301 int stringcount, const char *stringname,
2302 char *buffer, int buffersize);
2303
2304 int pcre_get_named_substring(const pcre *code,
2305 const char *subject, int *ovector,
2306 int stringcount, const char *stringname,
2307 const char **stringptr);
2308
2309 To extract a substring by name, you first have to find associated num-
2310 ber. For example, for this pattern
2311
2312 (a+)b(?<xxx>\d+)...
2313
2314 the number of the subpattern called "xxx" is 2. If the name is known to
2315 be unique (PCRE_DUPNAMES was not set), you can find the number from the
2316 name by calling pcre_get_stringnumber(). The first argument is the com-
2317 piled pattern, and the second is the name. The yield of the function is
2318 the subpattern number, or PCRE_ERROR_NOSUBSTRING (-7) if there is no
2319 subpattern of that name.
2320
2321 Given the number, you can extract the substring directly, or use one of
2322 the functions described in the previous section. For convenience, there
2323 are also two functions that do the whole job.
2324
2325 Most of the arguments of pcre_copy_named_substring() and
2326 pcre_get_named_substring() are the same as those for the similarly
2327 named functions that extract by number. As these are described in the
2328 previous section, they are not re-described here. There are just two
2329 differences:
2330
2331 First, instead of a substring number, a substring name is given. Sec-
2332 ond, there is an extra argument, given at the start, which is a pointer
2333 to the compiled pattern. This is needed in order to gain access to the
2334 name-to-number translation table.
2335
2336 These functions call pcre_get_stringnumber(), and if it succeeds, they
2337 then call pcre_copy_substring() or pcre_get_substring(), as appropri-
2338 ate. NOTE: If PCRE_DUPNAMES is set and there are duplicate names, the
2339 behaviour may not be what you want (see the next section).
2340
2341
2342 DUPLICATE SUBPATTERN NAMES
2343
2344 int pcre_get_stringtable_entries(const pcre *code,
2345 const char *name, char **first, char **last);
2346
2347 When a pattern is compiled with the PCRE_DUPNAMES option, names for
2348 subpatterns are not required to be unique. Normally, patterns with
2349 duplicate names are such that in any one match, only one of the named
2350 subpatterns participates. An example is shown in the pcrepattern docu-
2351 mentation.
2352
2353 When duplicates are present, pcre_copy_named_substring() and
2354 pcre_get_named_substring() return the first substring corresponding to
2355 the given name that is set. If none are set, PCRE_ERROR_NOSUBSTRING
2356 (-7) is returned; no data is returned. The pcre_get_stringnumber()
2357 function returns one of the numbers that are associated with the name,
2358 but it is not defined which it is.
2359
2360 If you want to get full details of all captured substrings for a given
2361 name, you must use the pcre_get_stringtable_entries() function. The
2362 first argument is the compiled pattern, and the second is the name. The
2363 third and fourth are pointers to variables which are updated by the
2364 function. After it has run, they point to the first and last entries in
2365 the name-to-number table for the given name. The function itself
2366 returns the length of each entry, or PCRE_ERROR_NOSUBSTRING (-7) if
2367 there are none. The format of the table is described above in the sec-
2368 tion entitled Information about a pattern. Given all the relevant
2369 entries for the name, you can extract each of their numbers, and hence
2370 the captured data, if any.
2371
2372
2373 FINDING ALL POSSIBLE MATCHES
2374
2375 The traditional matching function uses a similar algorithm to Perl,
2376 which stops when it finds the first match, starting at a given point in
2377 the subject. If you want to find all possible matches, or the longest
2378 possible match, consider using the alternative matching function (see
2379 below) instead. If you cannot use the alternative function, but still
2380 need to find all possible matches, you can kludge it up by making use
2381 of the callout facility, which is described in the pcrecallout documen-
2382 tation.
2383
2384 What you have to do is to insert a callout right at the end of the pat-
2385 tern. When your callout function is called, extract and save the cur-
2386 rent matched substring. Then return 1, which forces pcre_exec() to
2387 backtrack and try other alternatives. Ultimately, when it runs out of
2388 matches, pcre_exec() will yield PCRE_ERROR_NOMATCH.
2389
2390
2391 MATCHING A PATTERN: THE ALTERNATIVE FUNCTION
2392
2393 int pcre_dfa_exec(const pcre *code, const pcre_extra *extra,
2394 const char *subject, int length, int startoffset,
2395 int options, int *ovector, int ovecsize,
2396 int *workspace, int wscount);
2397
2398 The function pcre_dfa_exec() is called to match a subject string
2399 against a compiled pattern, using a matching algorithm that scans the
2400 subject string just once, and does not backtrack. This has different
2401 characteristics to the normal algorithm, and is not compatible with
2402 Perl. Some of the features of PCRE patterns are not supported. Never-
2403 theless, there are times when this kind of matching can be useful. For
2404 a discussion of the two matching algorithms, see the pcrematching docu-
2405 mentation.
2406
2407 The arguments for the pcre_dfa_exec() function are the same as for
2408 pcre_exec(), plus two extras. The ovector argument is used in a differ-
2409 ent way, and this is described below. The other common arguments are
2410 used in the same way as for pcre_exec(), so their description is not
2411 repeated here.
2412
2413 The two additional arguments provide workspace for the function. The
2414 workspace vector should contain at least 20 elements. It is used for
2415 keeping track of multiple paths through the pattern tree. More
2416 workspace will be needed for patterns and subjects where there are a
2417 lot of potential matches.
2418
2419 Here is an example of a simple call to pcre_dfa_exec():
2420
2421 int rc;
2422 int ovector[10];
2423 int wspace[20];
2424 rc = pcre_dfa_exec(
2425 re, /* result of pcre_compile() */
2426 NULL, /* we didn't study the pattern */
2427 "some string", /* the subject string */
2428 11, /* the length of the subject string */
2429 0, /* start at offset 0 in the subject */
2430 0, /* default options */
2431 ovector, /* vector of integers for substring information */
2432 10, /* number of elements (NOT size in bytes) */
2433 wspace, /* working space vector */
2434 20); /* number of elements (NOT size in bytes) */
2435
2436 Option bits for pcre_dfa_exec()
2437
2438 The unused bits of the options argument for pcre_dfa_exec() must be
2439 zero. The only bits that may be set are PCRE_ANCHORED, PCRE_NEW-
2440 LINE_xxx, PCRE_NOTBOL, PCRE_NOTEOL, PCRE_NOTEMPTY, PCRE_NO_UTF8_CHECK,
2441 PCRE_PARTIAL, PCRE_DFA_SHORTEST, and PCRE_DFA_RESTART. All but the last
2442 three of these are the same as for pcre_exec(), so their description is
2443 not repeated here.
2444
2445 PCRE_PARTIAL
2446
2447 This has the same general effect as it does for pcre_exec(), but the
2448 details are slightly different. When PCRE_PARTIAL is set for
2449 pcre_dfa_exec(), the return code PCRE_ERROR_NOMATCH is converted into
2450 PCRE_ERROR_PARTIAL if the end of the subject is reached, there have
2451 been no complete matches, but there is still at least one matching pos-
2452 sibility. The portion of the string that provided the partial match is
2453 set as the first matching string.
2454
2455 PCRE_DFA_SHORTEST
2456
2457 Setting the PCRE_DFA_SHORTEST option causes the matching algorithm to
2458 stop as soon as it has found one match. Because of the way the alterna-
2459 tive algorithm works, this is necessarily the shortest possible match
2460 at the first possible matching point in the subject string.
2461
2462 PCRE_DFA_RESTART
2463
2464 When pcre_dfa_exec() is called with the PCRE_PARTIAL option, and
2465 returns a partial match, it is possible to call it again, with addi-
2466 tional subject characters, and have it continue with the same match.
2467 The PCRE_DFA_RESTART option requests this action; when it is set, the
2468 workspace and wscount options must reference the same vector as before
2469 because data about the match so far is left in them after a partial
2470 match. There is more discussion of this facility in the pcrepartial
2471 documentation.
2472
2473 Successful returns from pcre_dfa_exec()
2474
2475 When pcre_dfa_exec() succeeds, it may have matched more than one sub-
2476 string in the subject. Note, however, that all the matches from one run
2477 of the function start at the same point in the subject. The shorter
2478 matches are all initial substrings of the longer matches. For example,
2479 if the pattern
2480
2481 <.*>
2482
2483 is matched against the string
2484
2485 This is <something> <something else> <something further> no more
2486
2487 the three matched strings are
2488
2489 <something>
2490 <something> <something else>
2491 <something> <something else> <something further>
2492
2493 On success, the yield of the function is a number greater than zero,
2494 which is the number of matched substrings. The substrings themselves
2495 are returned in ovector. Each string uses two elements; the first is
2496 the offset to the start, and the second is the offset to the end. In
2497 fact, all the strings have the same start offset. (Space could have
2498 been saved by giving this only once, but it was decided to retain some
2499 compatibility with the way pcre_exec() returns data, even though the
2500 meaning of the strings is different.)
2501
2502 The strings are returned in reverse order of length; that is, the long-
2503 est matching string is given first. If there were too many matches to
2504 fit into ovector, the yield of the function is zero, and the vector is
2505 filled with the longest matches.
2506
2507 Error returns from pcre_dfa_exec()
2508
2509 The pcre_dfa_exec() function returns a negative number when it fails.
2510 Many of the errors are the same as for pcre_exec(), and these are
2511 described above. There are in addition the following errors that are
2512 specific to pcre_dfa_exec():
2513
2514 PCRE_ERROR_DFA_UITEM (-16)
2515
2516 This return is given if pcre_dfa_exec() encounters an item in the pat-
2517 tern that it does not support, for instance, the use of \C or a back
2518 reference.
2519
2520 PCRE_ERROR_DFA_UCOND (-17)
2521
2522 This return is given if pcre_dfa_exec() encounters a condition item
2523 that uses a back reference for the condition, or a test for recursion
2524 in a specific group. These are not supported.
2525
2526 PCRE_ERROR_DFA_UMLIMIT (-18)
2527
2528 This return is given if pcre_dfa_exec() is called with an extra block
2529 that contains a setting of the match_limit field. This is not supported
2530 (it is meaningless).
2531
2532 PCRE_ERROR_DFA_WSSIZE (-19)
2533
2534 This return is given if pcre_dfa_exec() runs out of space in the
2535 workspace vector.
2536
2537 PCRE_ERROR_DFA_RECURSE (-20)
2538
2539 When a recursive subpattern is processed, the matching function calls
2540 itself recursively, using private vectors for ovector and workspace.
2541 This error is given if the output vector is not large enough. This
2542 should be extremely rare, as a vector of size 1000 is used.
2543
2544
2545 SEE ALSO
2546
2547 pcrebuild(3), pcrecallout(3), pcrecpp(3)(3), pcrematching(3), pcrepar-
2548 tial(3), pcreposix(3), pcreprecompile(3), pcresample(3), pcrestack(3).
2549
2550
2551 AUTHOR
2552
2553 Philip Hazel
2554 University Computing Service
2555 Cambridge CB2 3QH, England.
2556
2557
2558 REVISION
2559
2560 Last updated: 27 November 2007
2561 Copyright (c) 1997-2007 University of Cambridge.
2562 ------------------------------------------------------------------------------
2563
2564
2565 PCRECALLOUT(3) PCRECALLOUT(3)
2566
2567
2568 NAME
2569 PCRE - Perl-compatible regular expressions
2570
2571
2572 PCRE CALLOUTS
2573
2574 int (*pcre_callout)(pcre_callout_block *);
2575
2576 PCRE provides a feature called "callout", which is a means of temporar-
2577 ily passing control to the caller of PCRE in the middle of pattern
2578 matching. The caller of PCRE provides an external function by putting
2579 its entry point in the global variable pcre_callout. By default, this
2580 variable contains NULL, which disables all calling out.
2581
2582 Within a regular expression, (?C) indicates the points at which the
2583 external function is to be called. Different callout points can be
2584 identified by putting a number less than 256 after the letter C. The
2585 default value is zero. For example, this pattern has two callout
2586 points:
2587
2588 (?C1)abc(?C2)def
2589
2590 If the PCRE_AUTO_CALLOUT option bit is set when pcre_compile() is
2591 called, PCRE automatically inserts callouts, all with number 255,
2592 before each item in the pattern. For example, if PCRE_AUTO_CALLOUT is
2593 used with the pattern
2594
2595 A(\d{2}|--)
2596
2597 it is processed as if it were
2598
2599 (?C255)A(?C255)((?C255)\d{2}(?C255)|(?C255)-(?C255)-(?C255))(?C255)
2600
2601 Notice that there is a callout before and after each parenthesis and
2602 alternation bar. Automatic callouts can be used for tracking the
2603 progress of pattern matching. The pcretest command has an option that
2604 sets automatic callouts; when it is used, the output indicates how the
2605 pattern is matched. This is useful information when you are trying to
2606 optimize the performance of a particular pattern.
2607
2608
2609 MISSING CALLOUTS
2610
2611 You should be aware that, because of optimizations in the way PCRE
2612 matches patterns, callouts sometimes do not happen. For example, if the
2613 pattern is
2614
2615 ab(?C4)cd
2616
2617 PCRE knows that any matching string must contain the letter "d". If the
2618 subject string is "abyz", the lack of "d" means that matching doesn't
2619 ever start, and the callout is never reached. However, with "abyd",
2620 though the result is still no match, the callout is obeyed.
2621
2622
2623 THE CALLOUT INTERFACE
2624
2625 During matching, when PCRE reaches a callout point, the external func-
2626 tion defined by pcre_callout is called (if it is set). This applies to
2627 both the pcre_exec() and the pcre_dfa_exec() matching functions. The
2628 only argument to the callout function is a pointer to a pcre_callout
2629 block. This structure contains the following fields:
2630
2631 int version;
2632 int callout_number;
2633 int *offset_vector;
2634 const char *subject;
2635 int subject_length;
2636 int start_match;
2637 int current_position;
2638 int capture_top;
2639 int capture_last;
2640 void *callout_data;
2641 int pattern_position;
2642 int next_item_length;
2643
2644 The version field is an integer containing the version number of the
2645 block format. The initial version was 0; the current version is 1. The
2646 version number will change again in future if additional fields are
2647 added, but the intention is never to remove any of the existing fields.
2648
2649 The callout_number field contains the number of the callout, as com-
2650 piled into the pattern (that is, the number after ?C for manual call-
2651 outs, and 255 for automatically generated callouts).
2652
2653 The offset_vector field is a pointer to the vector of offsets that was
2654 passed by the caller to pcre_exec() or pcre_dfa_exec(). When
2655 pcre_exec() is used, the contents can be inspected in order to extract
2656 substrings that have been matched so far, in the same way as for
2657 extracting substrings after a match has completed. For pcre_dfa_exec()
2658 this field is not useful.
2659
2660 The subject and subject_length fields contain copies of the values that
2661 were passed to pcre_exec().
2662
2663 The start_match field normally contains the offset within the subject
2664 at which the current match attempt started. However, if the escape
2665 sequence \K has been encountered, this value is changed to reflect the
2666 modified starting point. If the pattern is not anchored, the callout
2667 function may be called several times from the same point in the pattern
2668 for different starting points in the subject.
2669
2670 The current_position field contains the offset within the subject of
2671 the current match pointer.
2672
2673 When the pcre_exec() function is used, the capture_top field contains
2674 one more than the number of the highest numbered captured substring so
2675 far. If no substrings have been captured, the value of capture_top is
2676 one. This is always the case when pcre_dfa_exec() is used, because it
2677 does not support captured substrings.
2678
2679 The capture_last field contains the number of the most recently cap-
2680 tured substring. If no substrings have been captured, its value is -1.
2681 This is always the case when pcre_dfa_exec() is used.
2682
2683 The callout_data field contains a value that is passed to pcre_exec()
2684 or pcre_dfa_exec() specifically so that it can be passed back in call-
2685 outs. It is passed in the pcre_callout field of the pcre_extra data
2686 structure. If no such data was passed, the value of callout_data in a
2687 pcre_callout block is NULL. There is a description of the pcre_extra
2688 structure in the pcreapi documentation.
2689
2690 The pattern_position field is present from version 1 of the pcre_call-
2691 out structure. It contains the offset to the next item to be matched in
2692 the pattern string.
2693
2694 The next_item_length field is present from version 1 of the pcre_call-
2695 out structure. It contains the length of the next item to be matched in
2696 the pattern string. When the callout immediately precedes an alterna-
2697 tion bar, a closing parenthesis, or the end of the pattern, the length
2698 is zero. When the callout precedes an opening parenthesis, the length
2699 is that of the entire subpattern.
2700
2701 The pattern_position and next_item_length fields are intended to help
2702 in distinguishing between different automatic callouts, which all have
2703 the same callout number. However, they are set for all callouts.
2704
2705
2706 RETURN VALUES
2707
2708 The external callout function returns an integer to PCRE. If the value
2709 is zero, matching proceeds as normal. If the value is greater than
2710 zero, matching fails at the current point, but the testing of other
2711 matching possibilities goes ahead, just as if a lookahead assertion had
2712 failed. If the value is less than zero, the match is abandoned, and
2713 pcre_exec() (or pcre_dfa_exec()) returns the negative value.
2714
2715 Negative values should normally be chosen from the set of
2716 PCRE_ERROR_xxx values. In particular, PCRE_ERROR_NOMATCH forces a stan-
2717 dard "no match" failure. The error number PCRE_ERROR_CALLOUT is
2718 reserved for use by callout functions; it will never be used by PCRE
2719 itself.
2720
2721
2722 AUTHOR
2723
2724 Philip Hazel
2725 University Computing Service
2726 Cambridge CB2 3QH, England.
2727
2728
2729 REVISION
2730
2731 Last updated: 29 May 2007
2732 Copyright (c) 1997-2007 University of Cambridge.
2733 ------------------------------------------------------------------------------
2734
2735
2736 PCRECOMPAT(3) PCRECOMPAT(3)
2737
2738
2739 NAME
2740 PCRE - Perl-compatible regular expressions
2741
2742
2743 DIFFERENCES BETWEEN PCRE AND PERL
2744
2745 This document describes the differences in the ways that PCRE and Perl
2746 handle regular expressions. The differences described here are mainly
2747 with respect to Perl 5.8, though PCRE versions 7.0 and later contain
2748 some features that are expected to be in the forthcoming Perl 5.10.
2749
2750 1. PCRE has only a subset of Perl's UTF-8 and Unicode support. Details
2751 of what it does have are given in the section on UTF-8 support in the
2752 main pcre page.
2753
2754 2. PCRE does not allow repeat quantifiers on lookahead assertions. Perl
2755 permits them, but they do not mean what you might think. For example,
2756 (?!a){3} does not assert that the next three characters are not "a". It
2757 just asserts that the next character is not "a" three times.
2758
2759 3. Capturing subpatterns that occur inside negative lookahead asser-
2760 tions are counted, but their entries in the offsets vector are never
2761 set. Perl sets its numerical variables from any such patterns that are
2762 matched before the assertion fails to match something (thereby succeed-
2763 ing), but only if the negative lookahead assertion contains just one
2764 branch.
2765
2766 4. Though binary zero characters are supported in the subject string,
2767 they are not allowed in a pattern string because it is passed as a nor-
2768 mal C string, terminated by zero. The escape sequence \0 can be used in
2769 the pattern to represent a binary zero.
2770
2771 5. The following Perl escape sequences are not supported: \l, \u, \L,
2772 \U, and \N. In fact these are implemented by Perl's general string-han-
2773 dling and are not part of its pattern matching engine. If any of these
2774 are encountered by PCRE, an error is generated.
2775
2776 6. The Perl escape sequences \p, \P, and \X are supported only if PCRE
2777 is built with Unicode character property support. The properties that
2778 can be tested with \p and \P are limited to the general category prop-
2779 erties such as Lu and Nd, script names such as Greek or Han, and the
2780 derived properties Any and L&.
2781
2782 7. PCRE does support the \Q...\E escape for quoting substrings. Charac-
2783 ters in between are treated as literals. This is slightly different
2784 from Perl in that $ and @ are also handled as literals inside the
2785 quotes. In Perl, they cause variable interpolation (but of course PCRE
2786 does not have variables). Note the following examples:
2787
2788 Pattern PCRE matches Perl matches
2789
2790 \Qabc$xyz\E abc$xyz abc followed by the
2791 contents of $xyz
2792 \Qabc\$xyz\E abc\$xyz abc\$xyz
2793 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
2794
2795 The \Q...\E sequence is recognized both inside and outside character
2796 classes.
2797
2798 8. Fairly obviously, PCRE does not support the (?{code}) and (??{code})
2799 constructions. However, there is support for recursive patterns. This
2800 is not available in Perl 5.8, but will be in Perl 5.10. Also, the PCRE
2801 "callout" feature allows an external function to be called during pat-
2802 tern matching. See the pcrecallout documentation for details.
2803
2804 9. Subpatterns that are called recursively or as "subroutines" are
2805 always treated as atomic groups in PCRE. This is like Python, but
2806 unlike Perl.
2807
2808 10. There are some differences that are concerned with the settings of
2809 captured strings when part of a pattern is repeated. For example,
2810 matching "aba" against the pattern /^(a(b)?)+$/ in Perl leaves $2
2811 unset, but in PCRE it is set to "b".
2812
2813 11. PCRE does support Perl 5.10's backtracking verbs (*ACCEPT),
2814 (*FAIL), (*F), (*COMMIT), (*PRUNE), (*SKIP), and (*THEN), but only in
2815 the forms without an argument. PCRE does not support (*MARK). If
2816 (*ACCEPT) is within capturing parentheses, PCRE does not set that cap-
2817 ture group; this is different to Perl.
2818
2819 12. PCRE provides some extensions to the Perl regular expression facil-
2820 ities. Perl 5.10 will include new features that are not in earlier
2821 versions, some of which (such as named parentheses) have been in PCRE
2822 for some time. This list is with respect to Perl 5.10:
2823
2824 (a) Although lookbehind assertions must match fixed length strings,
2825 each alternative branch of a lookbehind assertion can match a different
2826 length of string. Perl requires them all to have the same length.
2827
2828 (b) If PCRE_DOLLAR_ENDONLY is set and PCRE_MULTILINE is not set, the $
2829 meta-character matches only at the very end of the string.
2830
2831 (c) If PCRE_EXTRA is set, a backslash followed by a letter with no spe-
2832 cial meaning is faulted. Otherwise, like Perl, the backslash is quietly
2833 ignored. (Perl can be made to issue a warning.)
2834
2835 (d) If PCRE_UNGREEDY is set, the greediness of the repetition quanti-
2836 fiers is inverted, that is, by default they are not greedy, but if fol-
2837 lowed by a question mark they are.
2838
2839 (e) PCRE_ANCHORED can be used at matching time to force a pattern to be
2840 tried only at the first matching position in the subject string.
2841
2842 (f) The PCRE_NOTBOL, PCRE_NOTEOL, PCRE_NOTEMPTY, and PCRE_NO_AUTO_CAP-
2843 TURE options for pcre_exec() have no Perl equivalents.
2844
2845 (g) The \R escape sequence can be restricted to match only CR, LF, or
2846 CRLF by the PCRE_BSR_ANYCRLF option.
2847
2848 (h) The callout facility is PCRE-specific.
2849
2850 (i) The partial matching facility is PCRE-specific.
2851
2852 (j) Patterns compiled by PCRE can be saved and re-used at a later time,
2853 even on different hosts that have the other endianness.
2854
2855 (k) The alternative matching function (pcre_dfa_exec()) matches in a
2856 different way and is not Perl-compatible.
2857
2858 (l) PCRE recognizes some special sequences such as (*CR) at the start
2859 of a pattern that set overall options that cannot be changed within the
2860 pattern.
2861
2862
2863 AUTHOR
2864
2865 Philip Hazel
2866 University Computing Service
2867 Cambridge CB2 3QH, England.
2868
2869
2870 REVISION
2871
2872 Last updated: 11 September 2007
2873 Copyright (c) 1997-2007 University of Cambridge.
2874 ------------------------------------------------------------------------------
2875
2876
2877 PCREPATTERN(3) PCREPATTERN(3)
2878
2879
2880 NAME
2881 PCRE - Perl-compatible regular expressions
2882
2883
2884 PCRE REGULAR EXPRESSION DETAILS
2885
2886 The syntax and semantics of the regular expressions that are supported
2887 by PCRE are described in detail below. There is a quick-reference syn-
2888 tax summary in the pcresyntax page. Perl's regular expressions are
2889 described in its own documentation, and regular expressions in general
2890 are covered in a number of books, some of which have copious examples.
2891 Jeffrey Friedl's "Mastering Regular Expressions", published by
2892 O'Reilly, covers regular expressions in great detail. This description
2893 of PCRE's regular expressions is intended as reference material.
2894
2895 The original operation of PCRE was on strings of one-byte characters.
2896 However, there is now also support for UTF-8 character strings. To use
2897 this, you must build PCRE to include UTF-8 support, and then call
2898 pcre_compile() with the PCRE_UTF8 option. How this affects pattern
2899 matching is mentioned in several places below. There is also a summary
2900 of UTF-8 features in the section on UTF-8 support in the main pcre
2901 page.
2902
2903 The remainder of this document discusses the patterns that are sup-
2904 ported by PCRE when its main matching function, pcre_exec(), is used.
2905 From release 6.0, PCRE offers a second matching function,
2906 pcre_dfa_exec(), which matches using a different algorithm that is not
2907 Perl-compatible. Some of the features discussed below are not available
2908 when pcre_dfa_exec() is used. The advantages and disadvantages of the
2909 alternative function, and how it differs from the normal function, are
2910 discussed in the pcrematching page.
2911
2912
2913 NEWLINE CONVENTIONS
2914
2915 PCRE supports five different conventions for indicating line breaks in
2916 strings: a single CR (carriage return) character, a single LF (line-
2917 feed) character, the two-character sequence CRLF, any of the three pre-
2918 ceding, or any Unicode newline sequence. The pcreapi page has further
2919 discussion about newlines, and shows how to set the newline convention
2920 in the options arguments for the compiling and matching functions.
2921
2922 It is also possible to specify a newline convention by starting a pat-
2923 tern string with one of the following five sequences:
2924
2925 (*CR) carriage return
2926 (*LF) linefeed
2927 (*CRLF) carriage return, followed by linefeed
2928 (*ANYCRLF) any of the three above
2929 (*ANY) all Unicode newline sequences
2930
2931 These override the default and the options given to pcre_compile(). For
2932 example, on a Unix system where LF is the default newline sequence, the
2933 pattern
2934
2935 (*CR)a.b
2936
2937 changes the convention to CR. That pattern matches "a\nb" because LF is
2938 no longer a newline. Note that these special settings, which are not
2939 Perl-compatible, are recognized only at the very start of a pattern,
2940 and that they must be in upper case. If more than one of them is
2941 present, the last one is used.
2942
2943 The newline convention does not affect what the \R escape sequence
2944 matches. By default, this is any Unicode newline sequence, for Perl
2945 compatibility. However, this can be changed; see the description of \R
2946 in the section entitled "Newline sequences" below. A change of \R set-
2947 ting can be combined with a change of newline convention.
2948
2949
2950 CHARACTERS AND METACHARACTERS
2951
2952 A regular expression is a pattern that is matched against a subject
2953 string from left to right. Most characters stand for themselves in a
2954 pattern, and match the corresponding characters in the subject. As a
2955 trivial example, the pattern
2956
2957 The quick brown fox
2958
2959 matches a portion of a subject string that is identical to itself. When
2960 caseless matching is specified (the PCRE_CASELESS option), letters are
2961 matched independently of case. In UTF-8 mode, PCRE always understands
2962 the concept of case for characters whose values are less than 128, so
2963 caseless matching is always possible. For characters with higher val-
2964 ues, the concept of case is supported if PCRE is compiled with Unicode
2965 property support, but not otherwise. If you want to use caseless
2966 matching for characters 128 and above, you must ensure that PCRE is
2967 compiled with Unicode property support as well as with UTF-8 support.
2968
2969 The power of regular expressions comes from the ability to include
2970 alternatives and repetitions in the pattern. These are encoded in the
2971 pattern by the use of metacharacters, which do not stand for themselves
2972 but instead are interpreted in some special way.
2973
2974 There are two different sets of metacharacters: those that are recog-
2975 nized anywhere in the pattern except within square brackets, and those
2976 that are recognized within square brackets. Outside square brackets,
2977 the metacharacters are as follows:
2978
2979 \ general escape character with several uses
2980 ^ assert start of string (or line, in multiline mode)
2981 $ assert end of string (or line, in multiline mode)
2982 . match any character except newline (by default)
2983 [ start character class definition
2984 | start of alternative branch
2985 ( start subpattern
2986 ) end subpattern
2987 ? extends the meaning of (
2988 also 0 or 1 quantifier
2989 also quantifier minimizer
2990 * 0 or more quantifier
2991 + 1 or more quantifier
2992 also "possessive quantifier"
2993 { start min/max quantifier
2994
2995 Part of a pattern that is in square brackets is called a "character
2996 class". In a character class the only metacharacters are:
2997
2998 \ general escape character
2999 ^ negate the class, but only if the first character
3000 - indicates character range
3001 [ POSIX character class (only if followed by POSIX
3002 syntax)
3003 ] terminates the character class
3004
3005 The following sections describe the use of each of the metacharacters.
3006
3007
3008 BACKSLASH
3009
3010 The backslash character has several uses. Firstly, if it is followed by
3011 a non-alphanumeric character, it takes away any special meaning that
3012 character may have. This use of backslash as an escape character
3013 applies both inside and outside character classes.
3014
3015 For example, if you want to match a * character, you write \* in the
3016 pattern. This escaping action applies whether or not the following
3017 character would otherwise be interpreted as a metacharacter, so it is
3018 always safe to precede a non-alphanumeric with backslash to specify
3019 that it stands for itself. In particular, if you want to match a back-
3020 slash, you write \\.
3021
3022 If a pattern is compiled with the PCRE_EXTENDED option, whitespace in
3023 the pattern (other than in a character class) and characters between a
3024 # outside a character class and the next newline are ignored. An escap-
3025 ing backslash can be used to include a whitespace or # character as
3026 part of the pattern.
3027
3028 If you want to remove the special meaning from a sequence of charac-
3029 ters, you can do so by putting them between \Q and \E. This is differ-
3030 ent from Perl in that $ and @ are handled as literals in \Q...\E
3031 sequences in PCRE, whereas in Perl, $ and @ cause variable interpola-
3032 tion. Note the following examples:
3033
3034 Pattern PCRE matches Perl matches
3035
3036 \Qabc$xyz\E abc$xyz abc followed by the
3037 contents of $xyz
3038 \Qabc\$xyz\E abc\$xyz abc\$xyz
3039 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3040
3041 The \Q...\E sequence is recognized both inside and outside character
3042 classes.
3043
3044 Non-printing characters
3045
3046 A second use of backslash provides a way of encoding non-printing char-
3047 acters in patterns in a visible manner. There is no restriction on the
3048 appearance of non-printing characters, apart from the binary zero that
3049 terminates a pattern, but when a pattern is being prepared by text
3050 editing, it is usually easier to use one of the following escape
3051 sequences than the binary character it represents:
3052
3053 \a alarm, that is, the BEL character (hex 07)
3054 \cx "control-x", where x is any character
3055 \e escape (hex 1B)
3056 \f formfeed (hex 0C)
3057 \n linefeed (hex 0A)
3058 \r carriage return (hex 0D)
3059 \t tab (hex 09)
3060 \ddd character with octal code ddd, or backreference
3061 \xhh character with hex code hh
3062 \x{hhh..} character with hex code hhh..
3063
3064 The precise effect of \cx is as follows: if x is a lower case letter,
3065 it is converted to upper case. Then bit 6 of the character (hex 40) is
3066 inverted. Thus \cz becomes hex 1A, but \c{ becomes hex 3B, while \c;
3067 becomes hex 7B.
3068
3069 After \x, from zero to two hexadecimal digits are read (letters can be
3070 in upper or lower case). Any number of hexadecimal digits may appear
3071 between \x{ and }, but the value of the character code must be less
3072 than 256 in non-UTF-8 mode, and less than 2**31 in UTF-8 mode. That is,
3073 the maximum value in hexadecimal is 7FFFFFFF. Note that this is bigger
3074 than the largest Unicode code point, which is 10FFFF.
3075
3076 If characters other than hexadecimal digits appear between \x{ and },
3077 or if there is no terminating }, this form of escape is not recognized.
3078 Instead, the initial \x will be interpreted as a basic hexadecimal
3079 escape, with no following digits, giving a character whose value is
3080 zero.
3081
3082 Characters whose value is less than 256 can be defined by either of the
3083 two syntaxes for \x. There is no difference in the way they are han-
3084 dled. For example, \xdc is exactly the same as \x{dc}.
3085
3086 After \0 up to two further octal digits are read. If there are fewer
3087 than two digits, just those that are present are used. Thus the
3088 sequence \0\x\07 specifies two binary zeros followed by a BEL character
3089 (code value 7). Make sure you supply two digits after the initial zero
3090 if the pattern character that follows is itself an octal digit.
3091
3092 The handling of a backslash followed by a digit other than 0 is compli-
3093 cated. Outside a character class, PCRE reads it and any following dig-
3094 its as a decimal number. If the number is less than 10, or if there
3095 have been at least that many previous capturing left parentheses in the
3096 expression, the entire sequence is taken as a back reference. A
3097 description of how this works is given later, following the discussion
3098 of parenthesized subpatterns.
3099
3100 Inside a character class, or if the decimal number is greater than 9
3101 and there have not been that many capturing subpatterns, PCRE re-reads
3102 up to three octal digits following the backslash, and uses them to gen-
3103 erate a data character. Any subsequent digits stand for themselves. In
3104 non-UTF-8 mode, the value of a character specified in octal must be
3105 less than \400. In UTF-8 mode, values up to \777 are permitted. For
3106 example:
3107
3108 \040 is another way of writing a space
3109 \40 is the same, provided there are fewer than 40
3110 previous capturing subpatterns
3111 \7 is always a back reference
3112 \11 might be a back reference, or another way of
3113 writing a tab
3114 \011 is always a tab
3115 \0113 is a tab followed by the character "3"
3116 \113 might be a back reference, otherwise the
3117 character with octal code 113
3118 \377 might be a back reference, otherwise
3119 the byte consisting entirely of 1 bits
3120 \81 is either a back reference, or a binary zero
3121 followed by the two characters "8" and "1"
3122
3123 Note that octal values of 100 or greater must not be introduced by a
3124 leading zero, because no more than three octal digits are ever read.
3125
3126 All the sequences that define a single character value can be used both
3127 inside and outside character classes. In addition, inside a character
3128 class, the sequence \b is interpreted as the backspace character (hex
3129 08), and the sequences \R and \X are interpreted as the characters "R"
3130 and "X", respectively. Outside a character class, these sequences have
3131 different meanings (see below).
3132
3133 Absolute and relative back references
3134
3135 The sequence \g followed by an unsigned or a negative number, option-
3136 ally enclosed in braces, is an absolute or relative back reference. A
3137 named back reference can be coded as \g{name}. Back references are dis-
3138 cussed later, following the discussion of parenthesized subpatterns.
3139
3140 Generic character types
3141
3142 Another use of backslash is for specifying generic character types. The
3143 following are always recognized:
3144
3145 \d any decimal digit
3146 \D any character that is not a decimal digit
3147 \h any horizontal whitespace character
3148 \H any character that is not a horizontal whitespace character
3149 \s any whitespace character
3150 \S any character that is not a whitespace character
3151 \v any vertical whitespace character
3152 \V any character that is not a vertical whitespace character
3153 \w any "word" character
3154 \W any "non-word" character
3155
3156 Each pair of escape sequences partitions the complete set of characters
3157 into two disjoint sets. Any given character matches one, and only one,
3158 of each pair.
3159
3160 These character type sequences can appear both inside and outside char-
3161 acter classes. They each match one character of the appropriate type.
3162 If the current matching point is at the end of the subject string, all
3163 of them fail, since there is no character to match.
3164
3165 For compatibility with Perl, \s does not match the VT character (code
3166 11). This makes it different from the the POSIX "space" class. The \s
3167 characters are HT (9), LF (10), FF (12), CR (13), and space (32). If
3168 "use locale;" is included in a Perl script, \s may match the VT charac-
3169 ter. In PCRE, it never does.
3170
3171 In UTF-8 mode, characters with values greater than 128 never match \d,
3172 \s, or \w, and always match \D, \S, and \W. This is true even when Uni-
3173 code character property support is available. These sequences retain
3174 their original meanings from before UTF-8 support was available, mainly
3175 for efficiency reasons.
3176
3177 The sequences \h, \H, \v, and \V are Perl 5.10 features. In contrast to
3178 the other sequences, these do match certain high-valued codepoints in
3179 UTF-8 mode. The horizontal space characters are:
3180
3181 U+0009 Horizontal tab
3182 U+0020 Space
3183 U+00A0 Non-break space
3184 U+1680 Ogham space mark
3185 U+180E Mongolian vowel separator
3186 U+2000 En quad
3187 U+2001 Em quad
3188 U+2002 En space
3189 U+2003 Em space
3190 U+2004 Three-per-em space
3191 U+2005 Four-per-em space
3192 U+2006 Six-per-em space
3193 U+2007 Figure space
3194 U+2008 Punctuation space
3195 U+2009 Thin space
3196 U+200A Hair space
3197 U+202F Narrow no-break space
3198 U+205F Medium mathematical space
3199 U+3000 Ideographic space
3200
3201 The vertical space characters are:
3202
3203 U+000A Linefeed
3204 U+000B Vertical tab
3205 U+000C Formfeed
3206 U+000D Carriage return
3207 U+0085 Next line
3208 U+2028 Line separator
3209 U+2029 Paragraph separator
3210
3211 A "word" character is an underscore or any character less than 256 that
3212 is a letter or digit. The definition of letters and digits is con-
3213 trolled by PCRE's low-valued character tables, and may vary if locale-
3214 specific matching is taking place (see "Locale support" in the pcreapi
3215 page). For example, in a French locale such as "fr_FR" in Unix-like
3216 systems, or "french" in Windows, some character codes greater than 128
3217 are used for accented letters, and these are matched by \w. The use of
3218 locales with Unicode is discouraged.
3219
3220 Newline sequences
3221
3222 Outside a character class, by default, the escape sequence \R matches
3223 any Unicode newline sequence. This is a Perl 5.10 feature. In non-UTF-8
3224 mode \R is equivalent to the following:
3225
3226 (?>\r\n|\n|\x0b|\f|\r|\x85)
3227
3228 This is an example of an "atomic group", details of which are given
3229 below. This particular group matches either the two-character sequence
3230 CR followed by LF, or one of the single characters LF (linefeed,
3231 U+000A), VT (vertical tab, U+000B), FF (formfeed, U+000C), CR (carriage
3232 return, U+000D), or NEL (next line, U+0085). The two-character sequence
3233 is treated as a single unit that cannot be split.
3234
3235 In UTF-8 mode, two additional characters whose codepoints are greater
3236 than 255 are added: LS (line separator, U+2028) and PS (paragraph sepa-
3237 rator, U+2029). Unicode character property support is not needed for
3238 these characters to be recognized.
3239
3240 It is possible to restrict \R to match only CR, LF, or CRLF (instead of
3241 the complete set of Unicode line endings) by setting the option
3242 PCRE_BSR_ANYCRLF either at compile time or when the pattern is matched.
3243 (BSR is an abbrevation for "backslash R".) This can be made the default
3244 when PCRE is built; if this is the case, the other behaviour can be
3245 requested via the PCRE_BSR_UNICODE option. It is also possible to
3246 specify these settings by starting a pattern string with one of the
3247 following sequences:
3248
3249 (*BSR_ANYCRLF) CR, LF, or CRLF only
3250 (*BSR_UNICODE) any Unicode newline sequence
3251
3252 These override the default and the options given to pcre_compile(), but
3253 they can be overridden by options given to pcre_exec(). Note that these
3254 special settings, which are not Perl-compatible, are recognized only at
3255 the very start of a pattern, and that they must be in upper case. If
3256 more than one of them is present, the last one is used. They can be
3257 combined with a change of newline convention, for example, a pattern
3258 can start with:
3259
3260 (*ANY)(*BSR_ANYCRLF)
3261
3262 Inside a character class, \R matches the letter "R".
3263
3264 Unicode character properties
3265
3266 When PCRE is built with Unicode character property support, three addi-
3267 tional escape sequences that match characters with specific properties
3268 are available. When not in UTF-8 mode, these sequences are of course
3269 limited to testing characters whose codepoints are less than 256, but
3270 they do work in this mode. The extra escape sequences are:
3271
3272 \p{xx} a character with the xx property
3273 \P{xx} a character without the xx property
3274 \X an extended Unicode sequence
3275
3276 The property names represented by xx above are limited to the Unicode
3277 script names, the general category properties, and "Any", which matches
3278 any character (including newline). Other properties such as "InMusical-
3279 Symbols" are not currently supported by PCRE. Note that \P{Any} does
3280 not match any characters, so always causes a match failure.
3281
3282 Sets of Unicode characters are defined as belonging to certain scripts.
3283 A character from one of these sets can be matched using a script name.
3284 For example:
3285
3286 \p{Greek}
3287 \P{Han}
3288
3289 Those that are not part of an identified script are lumped together as
3290 "Common". The current list of scripts is:
3291
3292 Arabic, Armenian, Balinese, Bengali, Bopomofo, Braille, Buginese,
3293 Buhid, Canadian_Aboriginal, Cherokee, Common, Coptic, Cuneiform,
3294 Cypriot, Cyrillic, Deseret, Devanagari, Ethiopic, Georgian, Glagolitic,
3295 Gothic, Greek, Gujarati, Gurmukhi, Han, Hangul, Hanunoo, Hebrew, Hira-
3296 gana, Inherited, Kannada, Katakana, Kharoshthi, Khmer, Lao, Latin,
3297 Limbu, Linear_B, Malayalam, Mongolian, Myanmar, New_Tai_Lue, Nko,
3298 Ogham, Old_Italic, Old_Persian, Oriya, Osmanya, Phags_Pa, Phoenician,
3299 Runic, Shavian, Sinhala, Syloti_Nagri, Syriac, Tagalog, Tagbanwa,
3300 Tai_Le, Tamil, Telugu, Thaana, Thai, Tibetan, Tifinagh, Ugaritic, Yi.
3301
3302 Each character has exactly one general category property, specified by
3303 a two-letter abbreviation. For compatibility with Perl, negation can be
3304 specified by including a circumflex between the opening brace and the
3305 property name. For example, \p{^Lu} is the same as \P{Lu}.
3306
3307 If only one letter is specified with \p or \P, it includes all the gen-
3308 eral category properties that start with that letter. In this case, in
3309 the absence of negation, the curly brackets in the escape sequence are
3310 optional; these two examples have the same effect:
3311
3312 \p{L}
3313 \pL
3314
3315 The following general category property codes are supported:
3316
3317 C Other
3318 Cc Control
3319 Cf Format
3320 Cn Unassigned
3321 Co Private use
3322 Cs Surrogate
3323
3324 L Letter
3325 Ll Lower case letter
3326 Lm Modifier letter
3327 Lo Other letter
3328 Lt Title case letter
3329 Lu Upper case letter
3330
3331 M Mark
3332 Mc Spacing mark
3333 Me Enclosing mark
3334 Mn Non-spacing mark
3335
3336 N Number
3337 Nd Decimal number
3338 Nl Letter number
3339 No Other number
3340
3341 P Punctuation
3342 Pc Connector punctuation
3343 Pd Dash punctuation
3344 Pe Close punctuation
3345 Pf Final punctuation
3346 Pi Initial punctuation
3347 Po Other punctuation
3348 Ps Open punctuation
3349
3350 S Symbol
3351 Sc Currency symbol
3352 Sk Modifier symbol
3353 Sm Mathematical symbol
3354 So Other symbol
3355
3356 Z Separator
3357 Zl Line separator
3358 Zp Paragraph separator
3359 Zs Space separator
3360
3361 The special property L& is also supported: it matches a character that
3362 has the Lu, Ll, or Lt property, in other words, a letter that is not
3363 classified as a modifier or "other".
3364
3365 The Cs (Surrogate) property applies only to characters in the range
3366 U+D800 to U+DFFF. Such characters are not valid in UTF-8 strings (see
3367 RFC 3629) and so cannot be tested by PCRE, unless UTF-8 validity check-
3368 ing has been turned off (see the discussion of PCRE_NO_UTF8_CHECK in
3369 the pcreapi page).
3370
3371 The long synonyms for these properties that Perl supports (such as
3372 \p{Letter}) are not supported by PCRE, nor is it permitted to prefix
3373 any of these properties with "Is".
3374
3375 No character that is in the Unicode table has the Cn (unassigned) prop-
3376 erty. Instead, this property is assumed for any code point that is not
3377 in the Unicode table.
3378
3379 Specifying caseless matching does not affect these escape sequences.
3380 For example, \p{Lu} always matches only upper case letters.
3381
3382 The \X escape matches any number of Unicode characters that form an
3383 extended Unicode sequence. \X is equivalent to
3384
3385 (?>\PM\pM*)
3386
3387 That is, it matches a character without the "mark" property, followed
3388 by zero or more characters with the "mark" property, and treats the
3389 sequence as an atomic group (see below). Characters with the "mark"
3390 property are typically accents that affect the preceding character.
3391 None of them have codepoints less than 256, so in non-UTF-8 mode \X
3392 matches any one character.
3393
3394 Matching characters by Unicode property is not fast, because PCRE has
3395 to search a structure that contains data for over fifteen thousand
3396 characters. That is why the traditional escape sequences such as \d and
3397 \w do not use Unicode properties in PCRE.
3398
3399 Resetting the match start
3400
3401 The escape sequence \K, which is a Perl 5.10 feature, causes any previ-
3402 ously matched characters not to be included in the final matched
3403 sequence. For example, the pattern:
3404
3405 foo\Kbar
3406
3407 matches "foobar", but reports that it has matched "bar". This feature
3408 is similar to a lookbehind assertion (described below). However, in
3409 this case, the part of the subject before the real match does not have
3410 to be of fixed length, as lookbehind assertions do. The use of \K does
3411 not interfere with the setting of captured substrings. For example,
3412 when the pattern
3413
3414 (foo)\Kbar
3415
3416 matches "foobar", the first substring is still set to "foo".
3417
3418 Simple assertions
3419
3420 The final use of backslash is for certain simple assertions. An asser-
3421 tion specifies a condition that has to be met at a particular point in
3422 a match, without consuming any characters from the subject string. The
3423 use of subpatterns for more complicated assertions is described below.
3424 The backslashed assertions are:
3425
3426 \b matches at a word boundary
3427 \B matches when not at a word boundary
3428 \A matches at the start of the subject
3429 \Z matches at the end of the subject
3430 also matches before a newline at the end of the subject
3431 \z matches only at the end of the subject
3432 \G matches at the first matching position in the subject
3433
3434 These assertions may not appear in character classes (but note that \b
3435 has a different meaning, namely the backspace character, inside a char-
3436 acter class).
3437
3438 A word boundary is a position in the subject string where the current
3439 character and the previous character do not both match \w or \W (i.e.
3440 one matches \w and the other matches \W), or the start or end of the
3441 string if the first or last character matches \w, respectively.
3442
3443 The \A, \Z, and \z assertions differ from the traditional circumflex
3444 and dollar (described in the next section) in that they only ever match
3445 at the very start and end of the subject string, whatever options are
3446 set. Thus, they are independent of multiline mode. These three asser-
3447 tions are not affected by the PCRE_NOTBOL or PCRE_NOTEOL options, which
3448 affect only the behaviour of the circumflex and dollar metacharacters.
3449 However, if the startoffset argument of pcre_exec() is non-zero, indi-
3450 cating that matching is to start at a point other than the beginning of
3451 the subject, \A can never match. The difference between \Z and \z is
3452 that \Z matches before a newline at the end of the string as well as at
3453 the very end, whereas \z matches only at the end.
3454
3455 The \G assertion is true only when the current matching position is at
3456 the start point of the match, as specified by the startoffset argument
3457 of pcre_exec(). It differs from \A when the value of startoffset is
3458 non-zero. By calling pcre_exec() multiple times with appropriate argu-
3459 ments, you can mimic Perl's /g option, and it is in this kind of imple-
3460 mentation where \G can be useful.
3461
3462 Note, however, that PCRE's interpretation of \G, as the start of the
3463 current match, is subtly different from Perl's, which defines it as the
3464 end of the previous match. In Perl, these can be different when the
3465 previously matched string was empty. Because PCRE does just one match
3466 at a time, it cannot reproduce this behaviour.
3467
3468 If all the alternatives of a pattern begin with \G, the expression is
3469 anchored to the starting match position, and the "anchored" flag is set
3470 in the compiled regular expression.
3471
3472
3473 CIRCUMFLEX AND DOLLAR
3474
3475 Outside a character class, in the default matching mode, the circumflex
3476 character is an assertion that is true only if the current matching
3477 point is at the start of the subject string. If the startoffset argu-
3478 ment of pcre_exec() is non-zero, circumflex can never match if the
3479 PCRE_MULTILINE option is unset. Inside a character class, circumflex
3480 has an entirely different meaning (see below).
3481
3482 Circumflex need not be the first character of the pattern if a number
3483 of alternatives are involved, but it should be the first thing in each
3484 alternative in which it appears if the pattern is ever to match that
3485 branch. If all possible alternatives start with a circumflex, that is,
3486 if the pattern is constrained to match only at the start of the sub-
3487 ject, it is said to be an "anchored" pattern. (There are also other
3488 constructs that can cause a pattern to be anchored.)
3489
3490 A dollar character is an assertion that is true only if the current
3491 matching point is at the end of the subject string, or immediately
3492 before a newline at the end of the string (by default). Dollar need not
3493 be the last character of the pattern if a number of alternatives are
3494 involved, but it should be the last item in any branch in which it
3495 appears. Dollar has no special meaning in a character class.
3496
3497 The meaning of dollar can be changed so that it matches only at the
3498 very end of the string, by setting the PCRE_DOLLAR_ENDONLY option at
3499 compile time. This does not affect the \Z assertion.
3500
3501 The meanings of the circumflex and dollar characters are changed if the
3502 PCRE_MULTILINE option is set. When this is the case, a circumflex
3503 matches immediately after internal newlines as well as at the start of
3504 the subject string. It does not match after a newline that ends the
3505 string. A dollar matches before any newlines in the string, as well as
3506 at the very end, when PCRE_MULTILINE is set. When newline is specified
3507 as the two-character sequence CRLF, isolated CR and LF characters do
3508 not indicate newlines.
3509
3510 For example, the pattern /^abc$/ matches the subject string "def\nabc"
3511 (where \n represents a newline) in multiline mode, but not otherwise.
3512 Consequently, patterns that are anchored in single line mode because
3513 all branches start with ^ are not anchored in multiline mode, and a
3514 match for circumflex is possible when the startoffset argument of
3515 pcre_exec() is non-zero. The PCRE_DOLLAR_ENDONLY option is ignored if
3516 PCRE_MULTILINE is set.
3517
3518 Note that the sequences \A, \Z, and \z can be used to match the start
3519 and end of the subject in both modes, and if all branches of a pattern
3520 start with \A it is always anchored, whether or not PCRE_MULTILINE is
3521 set.
3522
3523
3524 FULL STOP (PERIOD, DOT)
3525
3526 Outside a character class, a dot in the pattern matches any one charac-
3527 ter in the subject string except (by default) a character that signi-
3528 fies the end of a line. In UTF-8 mode, the matched character may be
3529 more than one byte long.
3530
3531 When a line ending is defined as a single character, dot never matches
3532 that character; when the two-character sequence CRLF is used, dot does
3533 not match CR if it is immediately followed by LF, but otherwise it
3534 matches all characters (including isolated CRs and LFs). When any Uni-
3535 code line endings are being recognized, dot does not match CR or LF or
3536 any of the other line ending characters.
3537
3538 The behaviour of dot with regard to newlines can be changed. If the
3539 PCRE_DOTALL option is set, a dot matches any one character, without
3540 exception. If the two-character sequence CRLF is present in the subject
3541 string, it takes two dots to match it.
3542
3543 The handling of dot is entirely independent of the handling of circum-
3544 flex and dollar, the only relationship being that they both involve
3545 newlines. Dot has no special meaning in a character class.
3546
3547
3548 MATCHING A SINGLE BYTE
3549
3550 Outside a character class, the escape sequence \C matches any one byte,
3551 both in and out of UTF-8 mode. Unlike a dot, it always matches any
3552 line-ending characters. The feature is provided in Perl in order to
3553 match individual bytes in UTF-8 mode. Because it breaks up UTF-8 char-
3554 acters into individual bytes, what remains in the string may be a mal-
3555 formed UTF-8 string. For this reason, the \C escape sequence is best
3556 avoided.
3557
3558 PCRE does not allow \C to appear in lookbehind assertions (described
3559 below), because in UTF-8 mode this would make it impossible to calcu-
3560 late the length of the lookbehind.
3561
3562
3563 SQUARE BRACKETS AND CHARACTER CLASSES
3564
3565 An opening square bracket introduces a character class, terminated by a
3566 closing square bracket. A closing square bracket on its own is not spe-
3567 cial. If a closing square bracket is required as a member of the class,
3568 it should be the first data character in the class (after an initial
3569 circumflex, if present) or escaped with a backslash.
3570
3571 A character class matches a single character in the subject. In UTF-8
3572 mode, the character may occupy more than one byte. A matched character
3573 must be in the set of characters defined by the class, unless the first
3574 character in the class definition is a circumflex, in which case the
3575 subject character must not be in the set defined by the class. If a
3576 circumflex is actually required as a member of the class, ensure it is
3577 not the first character, or escape it with a backslash.
3578
3579 For example, the character class [aeiou] matches any lower case vowel,
3580 while [^aeiou] matches any character that is not a lower case vowel.
3581 Note that a circumflex is just a convenient notation for specifying the
3582 characters that are in the class by enumerating those that are not. A
3583 class that starts with a circumflex is not an assertion: it still con-
3584 sumes a character from the subject string, and therefore it fails if
3585 the current pointer is at the end of the string.
3586
3587 In UTF-8 mode, characters with values greater than 255 can be included
3588 in a class as a literal string of bytes, or by using the \x{ escaping
3589 mechanism.
3590
3591 When caseless matching is set, any letters in a class represent both
3592 their upper case and lower case versions, so for example, a caseless
3593 [aeiou] matches "A" as well as "a", and a caseless [^aeiou] does not
3594 match "A", whereas a caseful version would. In UTF-8 mode, PCRE always
3595 understands the concept of case for characters whose values are less
3596 than 128, so caseless matching is always possible. For characters with
3597 higher values, the concept of case is supported if PCRE is compiled
3598 with Unicode property support, but not otherwise. If you want to use
3599 caseless matching for characters 128 and above, you must ensure that
3600 PCRE is compiled with Unicode property support as well as with UTF-8
3601 support.
3602
3603 Characters that might indicate line breaks are never treated in any
3604 special way when matching character classes, whatever line-ending
3605 sequence is in use, and whatever setting of the PCRE_DOTALL and
3606 PCRE_MULTILINE options is used. A class such as [^a] always matches one
3607 of these characters.
3608
3609 The minus (hyphen) character can be used to specify a range of charac-
3610 ters in a character class. For example, [d-m] matches any letter
3611 between d and m, inclusive. If a minus character is required in a
3612 class, it must be escaped with a backslash or appear in a position
3613 where it cannot be interpreted as indicating a range, typically as the
3614 first or last character in the class.
3615
3616 It is not possible to have the literal character "]" as the end charac-
3617 ter of a range. A pattern such as [W-]46] is interpreted as a class of
3618 two characters ("W" and "-") followed by a literal string "46]", so it
3619 would match "W46]" or "-46]". However, if the "]" is escaped with a
3620 backslash it is interpreted as the end of range, so [W-\]46] is inter-
3621 preted as a class containing a range followed by two other characters.
3622 The octal or hexadecimal representation of "]" can also be used to end
3623 a range.
3624
3625 Ranges operate in the collating sequence of character values. They can
3626 also be used for characters specified numerically, for example
3627 [\000-\037]. In UTF-8 mode, ranges can include characters whose values
3628 are greater than 255, for example [\x{100}-\x{2ff}].
3629
3630 If a range that includes letters is used when caseless matching is set,
3631 it matches the letters in either case. For example, [W-c] is equivalent
3632 to [][\\^_`wxyzabc], matched caselessly, and in non-UTF-8 mode, if
3633 character tables for a French locale are in use, [\xc8-\xcb] matches
3634 accented E characters in both cases. In UTF-8 mode, PCRE supports the
3635 concept of case for characters with values greater than 128 only when
3636 it is compiled with Unicode property support.
3637
3638 The character types \d, \D, \p, \P, \s, \S, \w, and \W may also appear
3639 in a character class, and add the characters that they match to the
3640 class. For example, [\dABCDEF] matches any hexadecimal digit. A circum-
3641 flex can conveniently be used with the upper case character types to
3642 specify a more restricted set of characters than the matching lower
3643 case type. For example, the class [^\W_] matches any letter or digit,
3644 but not underscore.
3645
3646 The only metacharacters that are recognized in character classes are
3647 backslash, hyphen (only where it can be interpreted as specifying a
3648 range), circumflex (only at the start), opening square bracket (only
3649 when it can be interpreted as introducing a POSIX class name - see the
3650 next section), and the terminating closing square bracket. However,
3651 escaping other non-alphanumeric characters does no harm.
3652
3653
3654 POSIX CHARACTER CLASSES
3655
3656 Perl supports the POSIX notation for character classes. This uses names
3657 enclosed by [: and :] within the enclosing square brackets. PCRE also
3658 supports this notation. For example,
3659
3660 [01[:alpha:]%]
3661
3662 matches "0", "1", any alphabetic character, or "%". The supported class
3663 names are
3664
3665 alnum letters and digits
3666 alpha letters
3667 ascii character codes 0 - 127
3668 blank space or tab only
3669 cntrl control characters
3670 digit decimal digits (same as \d)
3671 graph printing characters, excluding space
3672 lower lower case letters
3673 print printing characters, including space
3674 punct printing characters, excluding letters and digits
3675 space white space (not quite the same as \s)
3676 upper upper case letters
3677 word "word" characters (same as \w)
3678 xdigit hexadecimal digits
3679
3680 The "space" characters are HT (9), LF (10), VT (11), FF (12), CR (13),
3681 and space (32). Notice that this list includes the VT character (code
3682 11). This makes "space" different to \s, which does not include VT (for
3683 Perl compatibility).
3684
3685 The name "word" is a Perl extension, and "blank" is a GNU extension
3686 from Perl 5.8. Another Perl extension is negation, which is indicated
3687 by a ^ character after the colon. For example,
3688
3689 [12[:^digit:]]
3690
3691 matches "1", "2", or any non-digit. PCRE (and Perl) also recognize the
3692 POSIX syntax [.ch.] and [=ch=] where "ch" is a "collating element", but
3693 these are not supported, and an error is given if they are encountered.
3694
3695 In UTF-8 mode, characters with values greater than 128 do not match any
3696 of the POSIX character classes.
3697
3698
3699 VERTICAL BAR
3700
3701 Vertical bar characters are used to separate alternative patterns. For
3702 example, the pattern
3703
3704 gilbert|sullivan
3705
3706 matches either "gilbert" or "sullivan". Any number of alternatives may
3707 appear, and an empty alternative is permitted (matching the empty
3708 string). The matching process tries each alternative in turn, from left
3709 to right, and the first one that succeeds is used. If the alternatives
3710 are within a subpattern (defined below), "succeeds" means matching the
3711 rest of the main pattern as well as the alternative in the subpattern.
3712
3713
3714 INTERNAL OPTION SETTING
3715
3716 The settings of the PCRE_CASELESS, PCRE_MULTILINE, PCRE_DOTALL, and
3717 PCRE_EXTENDED options (which are Perl-compatible) can be changed from
3718 within the pattern by a sequence of Perl option letters enclosed
3719 between "(?" and ")". The option letters are
3720
3721 i for PCRE_CASELESS
3722 m for PCRE_MULTILINE
3723 s for PCRE_DOTALL
3724 x for PCRE_EXTENDED
3725
3726 For example, (?im) sets caseless, multiline matching. It is also possi-
3727 ble to unset these options by preceding the letter with a hyphen, and a
3728 combined setting and unsetting such as (?im-sx), which sets PCRE_CASE-
3729 LESS and PCRE_MULTILINE while unsetting PCRE_DOTALL and PCRE_EXTENDED,
3730 is also permitted. If a letter appears both before and after the
3731 hyphen, the option is unset.
3732
3733 The PCRE-specific options PCRE_DUPNAMES, PCRE_UNGREEDY, and PCRE_EXTRA
3734 can be changed in the same way as the Perl-compatible options by using
3735 the characters J, U and X respectively.
3736
3737 When an option change occurs at top level (that is, not inside subpat-
3738 tern parentheses), the change applies to the remainder of the pattern
3739 that follows. If the change is placed right at the start of a pattern,
3740 PCRE extracts it into the global options (and it will therefore show up
3741 in data extracted by the pcre_fullinfo() function).
3742
3743 An option change within a subpattern (see below for a description of
3744 subpatterns) affects only that part of the current pattern that follows
3745 it, so
3746
3747 (a(?i)b)c
3748
3749 matches abc and aBc and no other strings (assuming PCRE_CASELESS is not
3750 used). By this means, options can be made to have different settings
3751 in different parts of the pattern. Any changes made in one alternative
3752 do carry on into subsequent branches within the same subpattern. For
3753 example,
3754
3755 (a(?i)b|c)
3756
3757 matches "ab", "aB", "c", and "C", even though when matching "C" the
3758 first branch is abandoned before the option setting. This is because
3759 the effects of option settings happen at compile time. There would be
3760 some very weird behaviour otherwise.
3761
3762 Note: There are other PCRE-specific options that can be set by the
3763 application when the compile or match functions are called. In some
3764 cases the pattern can contain special leading sequences to override
3765 what the application has set or what has been defaulted. Details are
3766 given in the section entitled "Newline sequences" above.
3767
3768
3769 SUBPATTERNS
3770
3771 Subpatterns are delimited by parentheses (round brackets), which can be
3772 nested. Turning part of a pattern into a subpattern does two things:
3773
3774 1. It localizes a set of alternatives. For example, the pattern
3775
3776 cat(aract|erpillar|)
3777
3778 matches one of the words "cat", "cataract", or "caterpillar". Without
3779 the parentheses, it would match "cataract", "erpillar" or an empty
3780 string.
3781
3782 2. It sets up the subpattern as a capturing subpattern. This means
3783 that, when the whole pattern matches, that portion of the subject
3784 string that matched the subpattern is passed back to the caller via the
3785 ovector argument of pcre_exec(). Opening parentheses are counted from
3786 left to right (starting from 1) to obtain numbers for the capturing
3787 subpatterns.
3788
3789 For example, if the string "the red king" is matched against the pat-
3790 tern
3791
3792 the ((red|white) (king|queen))
3793
3794 the captured substrings are "red king", "red", and "king", and are num-
3795 bered 1, 2, and 3, respectively.
3796
3797 The fact that plain parentheses fulfil two functions is not always
3798 helpful. There are often times when a grouping subpattern is required
3799 without a capturing requirement. If an opening parenthesis is followed
3800 by a question mark and a colon, the subpattern does not do any captur-
3801 ing, and is not counted when computing the number of any subsequent
3802 capturing subpatterns. For example, if the string "the white queen" is
3803 matched against the pattern
3804
3805 the ((?:red|white) (king|queen))
3806
3807 the captured substrings are "white queen" and "queen", and are numbered
3808 1 and 2. The maximum number of capturing subpatterns is 65535.
3809
3810 As a convenient shorthand, if any option settings are required at the
3811 start of a non-capturing subpattern, the option letters may appear
3812 between the "?" and the ":". Thus the two patterns
3813
3814 (?i:saturday|sunday)
3815 (?:(?i)saturday|sunday)
3816
3817 match exactly the same set of strings. Because alternative branches are
3818 tried from left to right, and options are not reset until the end of
3819 the subpattern is reached, an option setting in one branch does affect
3820 subsequent branches, so the above patterns match "SUNDAY" as well as
3821 "Saturday".
3822
3823
3824 DUPLICATE SUBPATTERN NUMBERS
3825
3826 Perl 5.10 introduced a feature whereby each alternative in a subpattern
3827 uses the same numbers for its capturing parentheses. Such a subpattern
3828 starts with (?| and is itself a non-capturing subpattern. For example,
3829 consider this pattern:
3830
3831 (?|(Sat)ur|(Sun))day
3832
3833 Because the two alternatives are inside a (?| group, both sets of cap-
3834 turing parentheses are numbered one. Thus, when the pattern matches,
3835 you can look at captured substring number one, whichever alternative
3836 matched. This construct is useful when you want to capture part, but
3837 not all, of one of a number of alternatives. Inside a (?| group, paren-
3838 theses are numbered as usual, but the number is reset at the start of
3839 each branch. The numbers of any capturing buffers that follow the sub-
3840 pattern start after the highest number used in any branch. The follow-
3841 ing example is taken from the Perl documentation. The numbers under-
3842 neath show in which buffer the captured content will be stored.
3843
3844 # before ---------------branch-reset----------- after
3845 / ( a ) (?| x ( y ) z | (p (q) r) | (t) u (v) ) ( z ) /x
3846 # 1 2 2 3 2 3 4
3847
3848 A backreference or a recursive call to a numbered subpattern always
3849 refers to the first one in the pattern with the given number.
3850
3851 An alternative approach to using this "branch reset" feature is to use
3852 duplicate named subpatterns, as described in the next section.
3853
3854
3855 NAMED SUBPATTERNS
3856
3857 Identifying capturing parentheses by number is simple, but it can be
3858 very hard to keep track of the numbers in complicated regular expres-
3859 sions. Furthermore, if an expression is modified, the numbers may
3860 change. To help with this difficulty, PCRE supports the naming of sub-
3861 patterns. This feature was not added to Perl until release 5.10. Python
3862 had the feature earlier, and PCRE introduced it at release 4.0, using
3863 the Python syntax. PCRE now supports both the Perl and the Python syn-
3864 tax.
3865
3866 In PCRE, a subpattern can be named in one of three ways: (?<name>...)
3867 or (?'name'...) as in Perl, or (?P<name>...) as in Python. References
3868 to capturing parentheses from other parts of the pattern, such as back-
3869 references, recursion, and conditions, can be made by name as well as
3870 by number.
3871
3872 Names consist of up to 32 alphanumeric characters and underscores.
3873 Named capturing parentheses are still allocated numbers as well as
3874 names, exactly as if the names were not present. The PCRE API provides
3875 function calls for extracting the name-to-number translation table from
3876 a compiled pattern. There is also a convenience function for extracting
3877 a captured substring by name.
3878
3879 By default, a name must be unique within a pattern, but it is possible
3880 to relax this constraint by setting the PCRE_DUPNAMES option at compile
3881 time. This can be useful for patterns where only one instance of the
3882 named parentheses can match. Suppose you want to match the name of a
3883 weekday, either as a 3-letter abbreviation or as the full name, and in
3884 both cases you want to extract the abbreviation. This pattern (ignoring
3885 the line breaks) does the job:
3886
3887 (?<DN>Mon|Fri|Sun)(?:day)?|
3888 (?<DN>Tue)(?:sday)?|
3889 (?<DN>Wed)(?:nesday)?|
3890 (?<DN>Thu)(?:rsday)?|
3891 (?<DN>Sat)(?:urday)?
3892
3893 There are five capturing substrings, but only one is ever set after a
3894 match. (An alternative way of solving this problem is to use a "branch
3895 reset" subpattern, as described in the previous section.)
3896
3897 The convenience function for extracting the data by name returns the
3898 substring for the first (and in this example, the only) subpattern of
3899 that name that matched. This saves searching to find which numbered
3900 subpattern it was. If you make a reference to a non-unique named sub-
3901 pattern from elsewhere in the pattern, the one that corresponds to the
3902 lowest number is used. For further details of the interfaces for han-
3903 dling named subpatterns, see the pcreapi documentation.
3904
3905
3906 REPETITION
3907
3908 Repetition is specified by quantifiers, which can follow any of the
3909 following items:
3910
3911 a literal data character
3912 the dot metacharacter
3913 the \C escape sequence
3914 the \X escape sequence (in UTF-8 mode with Unicode properties)
3915 the \R escape sequence
3916 an escape such as \d that matches a single character
3917 a character class
3918 a back reference (see next section)
3919 a parenthesized subpattern (unless it is an assertion)
3920
3921 The general repetition quantifier specifies a minimum and maximum num-
3922 ber of permitted matches, by giving the two numbers in curly brackets
3923 (braces), separated by a comma. The numbers must be less than 65536,
3924 and the first must be less than or equal to the second. For example:
3925
3926 z{2,4}
3927
3928 matches "zz", "zzz", or "zzzz". A closing brace on its own is not a
3929 special character. If the second number is omitted, but the comma is
3930 present, there is no upper limit; if the second number and the comma
3931 are both omitted, the quantifier specifies an exact number of required
3932 matches. Thus
3933
3934 [aeiou]{3,}
3935
3936 matches at least 3 successive vowels, but may match many more, while
3937
3938 \d{8}
3939
3940 matches exactly 8 digits. An opening curly bracket that appears in a
3941 position where a quantifier is not allowed, or one that does not match
3942 the syntax of a quantifier, is taken as a literal character. For exam-
3943 ple, {,6} is not a quantifier, but a literal string of four characters.
3944
3945 In UTF-8 mode, quantifiers apply to UTF-8 characters rather than to
3946 individual bytes. Thus, for example, \x{100}{2} matches two UTF-8 char-
3947 acters, each of which is represented by a two-byte sequence. Similarly,
3948 when Unicode property support is available, \X{3} matches three Unicode
3949 extended sequences, each of which may be several bytes long (and they
3950 may be of different lengths).
3951
3952 The quantifier {0} is permitted, causing the expression to behave as if
3953 the previous item and the quantifier were not present.
3954
3955 For convenience, the three most common quantifiers have single-charac-
3956 ter abbreviations:
3957
3958 * is equivalent to {0,}
3959 + is equivalent to {1,}
3960 ? is equivalent to {0,1}
3961
3962 It is possible to construct infinite loops by following a subpattern
3963 that can match no characters with a quantifier that has no upper limit,
3964 for example:
3965
3966 (a?)*
3967
3968 Earlier versions of Perl and PCRE used to give an error at compile time
3969 for such patterns. However, because there are cases where this can be
3970 useful, such patterns are now accepted, but if any repetition of the
3971 subpattern does in fact match no characters, the loop is forcibly bro-
3972 ken.
3973
3974 By default, the quantifiers are "greedy", that is, they match as much
3975 as possible (up to the maximum number of permitted times), without
3976 causing the rest of the pattern to fail. The classic example of where
3977 this gives problems is in trying to match comments in C programs. These
3978 appear between /* and */ and within the comment, individual * and /
3979 characters may appear. An attempt to match C comments by applying the
3980 pattern
3981
3982 /\*.*\*/
3983
3984 to the string
3985
3986 /* first comment */ not comment /* second comment */
3987
3988 fails, because it matches the entire string owing to the greediness of
3989 the .* item.
3990
3991 However, if a quantifier is followed by a question mark, it ceases to
3992 be greedy, and instead matches the minimum number of times possible, so
3993 the pattern
3994
3995 /\*.*?\*/
3996
3997 does the right thing with the C comments. The meaning of the various
3998 quantifiers is not otherwise changed, just the preferred number of
3999 matches. Do not confuse this use of question mark with its use as a
4000 quantifier in its own right. Because it has two uses, it can sometimes
4001 appear doubled, as in
4002
4003 \d??\d
4004
4005 which matches one digit by preference, but can match two if that is the
4006 only way the rest of the pattern matches.
4007
4008 If the PCRE_UNGREEDY option is set (an option that is not available in
4009 Perl), the quantifiers are not greedy by default, but individual ones
4010 can be made greedy by following them with a question mark. In other
4011 words, it inverts the default behaviour.
4012
4013 When a parenthesized subpattern is quantified with a minimum repeat
4014 count that is greater than 1 or with a limited maximum, more memory is
4015 required for the compiled pattern, in proportion to the size of the
4016 minimum or maximum.
4017
4018 If a pattern starts with .* or .{0,} and the PCRE_DOTALL option (equiv-
4019 alent to Perl's /s) is set, thus allowing the dot to match newlines,
4020 the pattern is implicitly anchored, because whatever follows will be
4021 tried against every character position in the subject string, so there
4022 is no point in retrying the overall match at any position after the
4023 first. PCRE normally treats such a pattern as though it were preceded
4024 by \A.
4025
4026 In cases where it is known that the subject string contains no new-
4027 lines, it is worth setting PCRE_DOTALL in order to obtain this opti-
4028 mization, or alternatively using ^ to indicate anchoring explicitly.
4029
4030 However, there is one situation where the optimization cannot be used.
4031 When .* is inside capturing parentheses that are the subject of a
4032 backreference elsewhere in the pattern, a match at the start may fail
4033 where a later one succeeds. Consider, for example:
4034
4035 (.*)abc\1
4036
4037 If the subject is "xyz123abc123" the match point is the fourth charac-
4038 ter. For this reason, such a pattern is not implicitly anchored.
4039
4040 When a capturing subpattern is repeated, the value captured is the sub-
4041 string that matched the final iteration. For example, after
4042
4043 (tweedle[dume]{3}\s*)+
4044
4045 has matched "tweedledum tweedledee" the value of the captured substring
4046 is "tweedledee". However, if there are nested capturing subpatterns,
4047 the corresponding captured values may have been set in previous itera-
4048 tions. For example, after
4049
4050 /(a|(b))+/
4051
4052 matches "aba" the value of the second captured substring is "b".
4053
4054
4055 ATOMIC GROUPING AND POSSESSIVE QUANTIFIERS
4056
4057 With both maximizing ("greedy") and minimizing ("ungreedy" or "lazy")
4058 repetition, failure of what follows normally causes the repeated item
4059 to be re-evaluated to see if a different number of repeats allows the
4060 rest of the pattern to match. Sometimes it is useful to prevent this,
4061 either to change the nature of the match, or to cause it fail earlier
4062 than it otherwise might, when the author of the pattern knows there is
4063 no point in carrying on.
4064
4065 Consider, for example, the pattern \d+foo when applied to the subject
4066 line
4067
4068 123456bar
4069
4070 After matching all 6 digits and then failing to match "foo", the normal
4071 action of the matcher is to try again with only 5 digits matching the
4072 \d+ item, and then with 4, and so on, before ultimately failing.
4073 "Atomic grouping" (a term taken from Jeffrey Friedl's book) provides
4074 the means for specifying that once a subpattern has matched, it is not
4075 to be re-evaluated in this way.
4076
4077 If we use atomic grouping for the previous example, the matcher gives
4078 up immediately on failing to match "foo" the first time. The notation
4079 is a kind of special parenthesis, starting with (?> as in this example:
4080
4081 (?>\d+)foo
4082
4083 This kind of parenthesis "locks up" the part of the pattern it con-
4084 tains once it has matched, and a failure further into the pattern is
4085 prevented from backtracking into it. Backtracking past it to previous
4086 items, however, works as normal.
4087
4088 An alternative description is that a subpattern of this type matches
4089 the string of characters that an identical standalone pattern would
4090 match, if anchored at the current point in the subject string.
4091
4092 Atomic grouping subpatterns are not capturing subpatterns. Simple cases
4093 such as the above example can be thought of as a maximizing repeat that
4094 must swallow everything it can. So, while both \d+ and \d+? are pre-
4095 pared to adjust the number of digits they match in order to make the
4096 rest of the pattern match, (?>\d+) can only match an entire sequence of
4097 digits.
4098
4099 Atomic groups in general can of course contain arbitrarily complicated
4100 subpatterns, and can be nested. However, when the subpattern for an
4101 atomic group is just a single repeated item, as in the example above, a
4102 simpler notation, called a "possessive quantifier" can be used. This
4103 consists of an additional + character following a quantifier. Using
4104 this notation, the previous example can be rewritten as
4105
4106 \d++foo
4107
4108 Note that a possessive quantifier can be used with an entire group, for
4109 example:
4110
4111 (abc|xyz){2,3}+
4112
4113 Possessive quantifiers are always greedy; the setting of the
4114 PCRE_UNGREEDY option is ignored. They are a convenient notation for the
4115 simpler forms of atomic group. However, there is no difference in the
4116 meaning of a possessive quantifier and the equivalent atomic group,
4117 though there may be a performance difference; possessive quantifiers
4118 should be slightly faster.
4119
4120 The possessive quantifier syntax is an extension to the Perl 5.8 syn-
4121 tax. Jeffrey Friedl originated the idea (and the name) in the first
4122 edition of his book. Mike McCloskey liked it, so implemented it when he
4123 built Sun's Java package, and PCRE copied it from there. It ultimately
4124 found its way into Perl at release 5.10.
4125
4126 PCRE has an optimization that automatically "possessifies" certain sim-
4127 ple pattern constructs. For example, the sequence A+B is treated as
4128 A++B because there is no point in backtracking into a sequence of A's
4129 when B must follow.
4130
4131 When a pattern contains an unlimited repeat inside a subpattern that
4132 can itself be repeated an unlimited number of times, the use of an
4133 atomic group is the only way to avoid some failing matches taking a
4134 very long time indeed. The pattern
4135
4136 (\D+|<\d+>)*[!?]
4137
4138 matches an unlimited number of substrings that either consist of non-
4139 digits, or digits enclosed in <>, followed by either ! or ?. When it
4140 matches, it runs quickly. However, if it is applied to
4141
4142 aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
4143
4144 it takes a long time before reporting failure. This is because the
4145 string can be divided between the internal \D+ repeat and the external
4146 * repeat in a large number of ways, and all have to be tried. (The
4147 example uses [!?] rather than a single character at the end, because
4148 both PCRE and Perl have an optimization that allows for fast failure
4149 when a single character is used. They remember the last single charac-
4150 ter that is required for a match, and fail early if it is not present
4151 in the string.) If the pattern is changed so that it uses an atomic
4152 group, like this:
4153
4154 ((?>\D+)|<\d+>)*[!?]
4155
4156 sequences of non-digits cannot be broken, and failure happens quickly.
4157
4158
4159 BACK REFERENCES
4160
4161 Outside a character class, a backslash followed by a digit greater than
4162 0 (and possibly further digits) is a back reference to a capturing sub-
4163 pattern earlier (that is, to its left) in the pattern, provided there
4164 have been that many previous capturing left parentheses.
4165
4166 However, if the decimal number following the backslash is less than 10,
4167 it is always taken as a back reference, and causes an error only if
4168 there are not that many capturing left parentheses in the entire pat-
4169 tern. In other words, the parentheses that are referenced need not be
4170 to the left of the reference for numbers less than 10. A "forward back
4171 reference" of this type can make sense when a repetition is involved
4172 and the subpattern to the right has participated in an earlier itera-
4173 tion.
4174
4175 It is not possible to have a numerical "forward back reference" to a
4176 subpattern whose number is 10 or more using this syntax because a
4177 sequence such as \50 is interpreted as a character defined in octal.
4178 See the subsection entitled "Non-printing characters" above for further
4179 details of the handling of digits following a backslash. There is no
4180 such problem when named parentheses are used. A back reference to any
4181 subpattern is possible using named parentheses (see below).
4182
4183 Another way of avoiding the ambiguity inherent in the use of digits
4184 following a backslash is to use the \g escape sequence, which is a fea-
4185 ture introduced in Perl 5.10. This escape must be followed by an
4186 unsigned number or a negative number, optionally enclosed in braces.
4187 These examples are all identical:
4188
4189 (ring), \1
4190 (ring), \g1
4191 (ring), \g{1}
4192
4193 An unsigned number specifies an absolute reference without the ambigu-
4194 ity that is present in the older syntax. It is also useful when literal
4195 digits follow the reference. A negative number is a relative reference.
4196 Consider this example:
4197
4198 (abc(def)ghi)\g{-1}
4199
4200 The sequence \g{-1} is a reference to the most recently started captur-
4201 ing subpattern before \g, that is, is it equivalent to \2. Similarly,
4202 \g{-2} would be equivalent to \1. The use of relative references can be
4203 helpful in long patterns, and also in patterns that are created by
4204 joining together fragments that contain references within themselves.
4205
4206 A back reference matches whatever actually matched the capturing sub-
4207 pattern in the current subject string, rather than anything matching
4208 the subpattern itself (see "Subpatterns as subroutines" below for a way
4209 of doing that). So the pattern
4210
4211 (sens|respons)e and \1ibility
4212
4213 matches "sense and sensibility" and "response and responsibility", but
4214 not "sense and responsibility". If caseful matching is in force at the
4215 time of the back reference, the case of letters is relevant. For exam-
4216 ple,
4217
4218 ((?i)rah)\s+\1
4219
4220 matches "rah rah" and "RAH RAH", but not "RAH rah", even though the
4221 original capturing subpattern is matched caselessly.
4222
4223 There are several different ways of writing back references to named
4224 subpatterns. The .NET syntax \k{name} and the Perl syntax \k<name> or
4225 \k'name' are supported, as is the Python syntax (?P=name). Perl 5.10's
4226 unified back reference syntax, in which \g can be used for both numeric
4227 and named references, is also supported. We could rewrite the above
4228 example in any of the following ways:
4229
4230 (?<p1>(?i)rah)\s+\k<p1>
4231 (?'p1'(?i)rah)\s+\k{p1}
4232 (?P<p1>(?i)rah)\s+(?P=p1)
4233 (?<p1>(?i)rah)\s+\g{p1}
4234
4235 A subpattern that is referenced by name may appear in the pattern
4236 before or after the reference.
4237
4238 There may be more than one back reference to the same subpattern. If a
4239 subpattern has not actually been used in a particular match, any back
4240 references to it always fail. For example, the pattern
4241
4242 (a|(bc))\2
4243
4244 always fails if it starts to match "a" rather than "bc". Because there
4245 may be many capturing parentheses in a pattern, all digits following
4246 the backslash are taken as part of a potential back reference number.
4247 If the pattern continues with a digit character, some delimiter must be
4248 used to terminate the back reference. If the PCRE_EXTENDED option is
4249 set, this can be whitespace. Otherwise an empty comment (see "Com-
4250 ments" below) can be used.
4251
4252 A back reference that occurs inside the parentheses to which it refers
4253 fails when the subpattern is first used, so, for example, (a\1) never
4254 matches. However, such references can be useful inside repeated sub-
4255 patterns. For example, the pattern
4256
4257 (a|b\1)+
4258
4259 matches any number of "a"s and also "aba", "ababbaa" etc. At each iter-
4260 ation of the subpattern, the back reference matches the character
4261 string corresponding to the previous iteration. In order for this to
4262 work, the pattern must be such that the first iteration does not need
4263 to match the back reference. This can be done using alternation, as in
4264 the example above, or by a quantifier with a minimum of zero.
4265
4266
4267 ASSERTIONS
4268
4269 An assertion is a test on the characters following or preceding the
4270 current matching point that does not actually consume any characters.
4271 The simple assertions coded as \b, \B, \A, \G, \Z, \z, ^ and $ are
4272 described above.
4273
4274 More complicated assertions are coded as subpatterns. There are two
4275 kinds: those that look ahead of the current position in the subject
4276 string, and those that look behind it. An assertion subpattern is
4277 matched in the normal way, except that it does not cause the current
4278 matching position to be changed.
4279
4280 Assertion subpatterns are not capturing subpatterns, and may not be
4281 repeated, because it makes no sense to assert the same thing several
4282 times. If any kind of assertion contains capturing subpatterns within
4283 it, these are counted for the purposes of numbering the capturing sub-
4284 patterns in the whole pattern. However, substring capturing is carried
4285 out only for positive assertions, because it does not make sense for
4286 negative assertions.
4287
4288 Lookahead assertions
4289
4290 Lookahead assertions start with (?= for positive assertions and (?! for
4291 negative assertions. For example,
4292
4293 \w+(?=;)
4294
4295 matches a word followed by a semicolon, but does not include the semi-
4296 colon in the match, and
4297
4298 foo(?!bar)
4299
4300 matches any occurrence of "foo" that is not followed by "bar". Note
4301 that the apparently similar pattern
4302
4303 (?!foo)bar
4304
4305 does not find an occurrence of "bar" that is preceded by something
4306 other than "foo"; it finds any occurrence of "bar" whatsoever, because
4307 the assertion (?!foo) is always true when the next three characters are
4308 "bar". A lookbehind assertion is needed to achieve the other effect.
4309
4310 If you want to force a matching failure at some point in a pattern, the
4311 most convenient way to do it is with (?!) because an empty string
4312 always matches, so an assertion that requires there not to be an empty
4313 string must always fail.
4314
4315 Lookbehind assertions
4316
4317 Lookbehind assertions start with (?<= for positive assertions and (?<!
4318 for negative assertions. For example,
4319
4320 (?<!foo)bar
4321
4322 does find an occurrence of "bar" that is not preceded by "foo". The
4323 contents of a lookbehind assertion are restricted such that all the
4324 strings it matches must have a fixed length. However, if there are sev-
4325 eral top-level alternatives, they do not all have to have the same
4326 fixed length. Thus
4327
4328 (?<=bullock|donkey)
4329
4330 is permitted, but
4331
4332 (?<!dogs?|cats?)
4333
4334 causes an error at compile time. Branches that match different length
4335 strings are permitted only at the top level of a lookbehind assertion.
4336 This is an extension compared with Perl (at least for 5.8), which
4337 requires all branches to match the same length of string. An assertion
4338 such as
4339
4340 (?<=ab(c|de))
4341
4342 is not permitted, because its single top-level branch can match two
4343 different lengths, but it is acceptable if rewritten to use two top-
4344 level branches:
4345
4346 (?<=abc|abde)
4347
4348 In some cases, the Perl 5.10 escape sequence \K (see above) can be used
4349 instead of a lookbehind assertion; this is not restricted to a fixed-
4350 length.
4351
4352 The implementation of lookbehind assertions is, for each alternative,
4353 to temporarily move the current position back by the fixed length and
4354 then try to match. If there are insufficient characters before the cur-
4355 rent position, the assertion fails.
4356
4357 PCRE does not allow the \C escape (which matches a single byte in UTF-8
4358 mode) to appear in lookbehind assertions, because it makes it impossi-
4359 ble to calculate the length of the lookbehind. The \X and \R escapes,
4360 which can match different numbers of bytes, are also not permitted.
4361
4362 Possessive quantifiers can be used in conjunction with lookbehind
4363 assertions to specify efficient matching at the end of the subject
4364 string. Consider a simple pattern such as
4365
4366 abcd$
4367
4368 when applied to a long string that does not match. Because matching
4369 proceeds from left to right, PCRE will look for each "a" in the subject
4370 and then see if what follows matches the rest of the pattern. If the
4371 pattern is specified as
4372
4373 ^.*abcd$
4374
4375 the initial .* matches the entire string at first, but when this fails
4376 (because there is no following "a"), it backtracks to match all but the
4377 last character, then all but the last two characters, and so on. Once
4378 again the search for "a" covers the entire string, from right to left,
4379 so we are no better off. However, if the pattern is written as
4380
4381 ^.*+(?<=abcd)
4382
4383 there can be no backtracking for the .*+ item; it can match only the
4384 entire string. The subsequent lookbehind assertion does a single test
4385 on the last four characters. If it fails, the match fails immediately.
4386 For long strings, this approach makes a significant difference to the
4387 processing time.
4388
4389 Using multiple assertions
4390
4391 Several assertions (of any sort) may occur in succession. For example,
4392
4393 (?<=\d{3})(?<!999)foo
4394
4395 matches "foo" preceded by three digits that are not "999". Notice that
4396 each of the assertions is applied independently at the same point in
4397 the subject string. First there is a check that the previous three
4398 characters are all digits, and then there is a check that the same
4399 three characters are not "999". This pattern does not match "foo" pre-
4400 ceded by six characters, the first of which are digits and the last
4401 three of which are not "999". For example, it doesn't match "123abc-
4402 foo". A pattern to do that is
4403
4404 (?<=\d{3}...)(?<!999)foo
4405
4406 This time the first assertion looks at the preceding six characters,
4407 checking that the first three are digits, and then the second assertion
4408 checks that the preceding three characters are not "999".
4409
4410 Assertions can be nested in any combination. For example,
4411
4412 (?<=(?<!foo)bar)baz
4413
4414 matches an occurrence of "baz" that is preceded by "bar" which in turn
4415 is not preceded by "foo", while
4416
4417 (?<=\d{3}(?!999)...)foo
4418
4419 is another pattern that matches "foo" preceded by three digits and any
4420 three characters that are not "999".
4421
4422
4423 CONDITIONAL SUBPATTERNS
4424
4425 It is possible to cause the matching process to obey a subpattern con-
4426 ditionally or to choose between two alternative subpatterns, depending
4427 on the result of an assertion, or whether a previous capturing subpat-
4428 tern matched or not. The two possible forms of conditional subpattern
4429 are
4430
4431 (?(condition)yes-pattern)
4432 (?(condition)yes-pattern|no-pattern)
4433
4434 If the condition is satisfied, the yes-pattern is used; otherwise the
4435 no-pattern (if present) is used. If there are more than two alterna-
4436 tives in the subpattern, a compile-time error occurs.
4437
4438 There are four kinds of condition: references to subpatterns, refer-
4439 ences to recursion, a pseudo-condition called DEFINE, and assertions.
4440
4441 Checking for a used subpattern by number
4442
4443 If the text between the parentheses consists of a sequence of digits,
4444 the condition is true if the capturing subpattern of that number has
4445 previously matched. An alternative notation is to precede the digits
4446 with a plus or minus sign. In this case, the subpattern number is rela-
4447 tive rather than absolute. The most recently opened parentheses can be
4448 referenced by (?(-1), the next most recent by (?(-2), and so on. In
4449 looping constructs it can also make sense to refer to subsequent groups
4450 with constructs such as (?(+2).
4451
4452 Consider the following pattern, which contains non-significant white
4453 space to make it more readable (assume the PCRE_EXTENDED option) and to
4454 divide it into three parts for ease of discussion:
4455
4456 ( \( )? [^()]+ (?(1) \) )
4457
4458 The first part matches an optional opening parenthesis, and if that
4459 character is present, sets it as the first captured substring. The sec-
4460 ond part matches one or more characters that are not parentheses. The
4461 third part is a conditional subpattern that tests whether the first set
4462 of parentheses matched or not. If they did, that is, if subject started
4463 with an opening parenthesis, the condition is true, and so the yes-pat-
4464 tern is executed and a closing parenthesis is required. Otherwise,
4465 since no-pattern is not present, the subpattern matches nothing. In
4466 other words, this pattern matches a sequence of non-parentheses,
4467 optionally enclosed in parentheses.
4468
4469 If you were embedding this pattern in a larger one, you could use a
4470 relative reference:
4471
4472 ...other stuff... ( \( )? [^()]+ (?(-1) \) ) ...
4473
4474 This makes the fragment independent of the parentheses in the larger
4475 pattern.
4476
4477 Checking for a used subpattern by name
4478
4479 Perl uses the syntax (?(<name>)...) or (?('name')...) to test for a
4480 used subpattern by name. For compatibility with earlier versions of
4481 PCRE, which had this facility before Perl, the syntax (?(name)...) is
4482 also recognized. However, there is a possible ambiguity with this syn-
4483 tax, because subpattern names may consist entirely of digits. PCRE
4484 looks first for a named subpattern; if it cannot find one and the name
4485 consists entirely of digits, PCRE looks for a subpattern of that num-
4486 ber, which must be greater than zero. Using subpattern names that con-
4487 sist entirely of digits is not recommended.
4488
4489 Rewriting the above example to use a named subpattern gives this:
4490
4491 (?<OPEN> \( )? [^()]+ (?(<OPEN>) \) )
4492
4493
4494 Checking for pattern recursion
4495
4496 If the condition is the string (R), and there is no subpattern with the
4497 name R, the condition is true if a recursive call to the whole pattern
4498 or any subpattern has been made. If digits or a name preceded by amper-
4499 sand follow the letter R, for example:
4500
4501 (?(R3)...) or (?(R&name)...)
4502
4503 the condition is true if the most recent recursion is into the subpat-
4504 tern whose number or name is given. This condition does not check the
4505 entire recursion stack.
4506
4507 At "top level", all these recursion test conditions are false. Recur-
4508 sive patterns are described below.
4509
4510 Defining subpatterns for use by reference only
4511
4512 If the condition is the string (DEFINE), and there is no subpattern
4513 with the name DEFINE, the condition is always false. In this case,
4514 there may be only one alternative in the subpattern. It is always
4515 skipped if control reaches this point in the pattern; the idea of
4516 DEFINE is that it can be used to define "subroutines" that can be ref-
4517 erenced from elsewhere. (The use of "subroutines" is described below.)
4518 For example, a pattern to match an IPv4 address could be written like
4519 this (ignore whitespace and line breaks):
4520
4521 (?(DEFINE) (?<byte> 2[0-4]\d | 25[0-5] | 1\d\d | [1-9]?\d) )
4522 \b (?&byte) (\.(?&byte)){3} \b
4523
4524 The first part of the pattern is a DEFINE group inside which a another
4525 group named "byte" is defined. This matches an individual component of
4526 an IPv4 address (a number less than 256). When matching takes place,
4527 this part of the pattern is skipped because DEFINE acts like a false
4528 condition.
4529
4530 The rest of the pattern uses references to the named group to match the
4531 four dot-separated components of an IPv4 address, insisting on a word
4532 boundary at each end.
4533
4534 Assertion conditions
4535
4536 If the condition is not in any of the above formats, it must be an
4537 assertion. This may be a positive or negative lookahead or lookbehind
4538 assertion. Consider this pattern, again containing non-significant
4539 white space, and with the two alternatives on the second line:
4540
4541 (?(?=[^a-z]*[a-z])
4542 \d{2}-[a-z]{3}-\d{2} | \d{2}-\d{2}-\d{2} )
4543
4544 The condition is a positive lookahead assertion that matches an
4545 optional sequence of non-letters followed by a letter. In other words,
4546 it tests for the presence of at least one letter in the subject. If a
4547 letter is found, the subject is matched against the first alternative;
4548 otherwise it is matched against the second. This pattern matches
4549 strings in one of the two forms dd-aaa-dd or dd-dd-dd, where aaa are
4550 letters and dd are digits.
4551
4552
4553 COMMENTS
4554
4555 The sequence (?# marks the start of a comment that continues up to the
4556 next closing parenthesis. Nested parentheses are not permitted. The
4557 characters that make up a comment play no part in the pattern matching
4558 at all.
4559
4560 If the PCRE_EXTENDED option is set, an unescaped # character outside a
4561 character class introduces a comment that continues to immediately
4562 after the next newline in the pattern.
4563
4564
4565 RECURSIVE PATTERNS
4566
4567 Consider the problem of matching a string in parentheses, allowing for
4568 unlimited nested parentheses. Without the use of recursion, the best
4569 that can be done is to use a pattern that matches up to some fixed
4570 depth of nesting. It is not possible to handle an arbitrary nesting
4571 depth.
4572
4573 For some time, Perl has provided a facility that allows regular expres-
4574 sions to recurse (amongst other things). It does this by interpolating
4575 Perl code in the expression at run time, and the code can refer to the
4576 expression itself. A Perl pattern using code interpolation to solve the
4577 parentheses problem can be created like this:
4578
4579 $re = qr{\( (?: (?>[^()]+) | (?p{$re}) )* \)}x;
4580
4581 The (?p{...}) item interpolates Perl code at run time, and in this case
4582 refers recursively to the pattern in which it appears.
4583
4584 Obviously, PCRE cannot support the interpolation of Perl code. Instead,
4585 it supports special syntax for recursion of the entire pattern, and
4586 also for individual subpattern recursion. After its introduction in
4587 PCRE and Python, this kind of recursion was introduced into Perl at
4588 release 5.10.
4589
4590 A special item that consists of (? followed by a number greater than
4591 zero and a closing parenthesis is a recursive call of the subpattern of
4592 the given number, provided that it occurs inside that subpattern. (If
4593 not, it is a "subroutine" call, which is described in the next sec-
4594 tion.) The special item (?R) or (?0) is a recursive call of the entire
4595 regular expression.
4596
4597 In PCRE (like Python, but unlike Perl), a recursive subpattern call is
4598 always treated as an atomic group. That is, once it has matched some of
4599 the subject string, it is never re-entered, even if it contains untried
4600 alternatives and there is a subsequent matching failure.
4601
4602 This PCRE pattern solves the nested parentheses problem (assume the
4603 PCRE_EXTENDED option is set so that white space is ignored):
4604
4605 \( ( (?>[^()]+) | (?R) )* \)
4606
4607 First it matches an opening parenthesis. Then it matches any number of
4608 substrings which can either be a sequence of non-parentheses, or a
4609 recursive match of the pattern itself (that is, a correctly parenthe-
4610 sized substring). Finally there is a closing parenthesis.
4611
4612 If this were part of a larger pattern, you would not want to recurse
4613 the entire pattern, so instead you could use this:
4614
4615 ( \( ( (?>[^()]+) | (?1) )* \) )
4616
4617 We have put the pattern into parentheses, and caused the recursion to
4618 refer to them instead of the whole pattern.
4619
4620 In a larger pattern, keeping track of parenthesis numbers can be
4621 tricky. This is made easier by the use of relative references. (A Perl
4622 5.10 feature.) Instead of (?1) in the pattern above you can write
4623 (?-2) to refer to the second most recently opened parentheses preceding
4624 the recursion. In other words, a negative number counts capturing
4625 parentheses leftwards from the point at which it is encountered.
4626
4627 It is also possible to refer to subsequently opened parentheses, by
4628 writing references such as (?+2). However, these cannot be recursive
4629 because the reference is not inside the parentheses that are refer-
4630 enced. They are always "subroutine" calls, as described in the next
4631 section.
4632
4633 An alternative approach is to use named parentheses instead. The Perl
4634 syntax for this is (?&name); PCRE's earlier syntax (?P>name) is also
4635 supported. We could rewrite the above example as follows:
4636
4637 (?<pn> \( ( (?>[^()]+) | (?&pn) )* \) )
4638
4639 If there is more than one subpattern with the same name, the earliest
4640 one is used.
4641
4642 This particular example pattern that we have been looking at contains
4643 nested unlimited repeats, and so the use of atomic grouping for match-
4644 ing strings of non-parentheses is important when applying the pattern
4645 to strings that do not match. For example, when this pattern is applied
4646 to
4647
4648 (aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa()
4649
4650 it yields "no match" quickly. However, if atomic grouping is not used,
4651 the match runs for a very long time indeed because there are so many
4652 different ways the + and * repeats can carve up the subject, and all
4653 have to be tested before failure can be reported.
4654
4655 At the end of a match, the values set for any capturing subpatterns are
4656 those from the outermost level of the recursion at which the subpattern
4657 value is set. If you want to obtain intermediate values, a callout
4658 function can be used (see below and the pcrecallout documentation). If
4659 the pattern above is matched against
4660
4661 (ab(cd)ef)
4662
4663 the value for the capturing parentheses is "ef", which is the last
4664 value taken on at the top level. If additional parentheses are added,
4665 giving
4666
4667 \( ( ( (?>[^()]+) | (?R) )* ) \)
4668 ^ ^
4669 ^ ^
4670
4671 the string they capture is "ab(cd)ef", the contents of the top level
4672 parentheses. If there are more than 15 capturing parentheses in a pat-
4673 tern, PCRE has to obtain extra memory to store data during a recursion,
4674 which it does by using pcre_malloc, freeing it via pcre_free after-
4675 wards. If no memory can be obtained, the match fails with the
4676 PCRE_ERROR_NOMEMORY error.
4677
4678 Do not confuse the (?R) item with the condition (R), which tests for
4679 recursion. Consider this pattern, which matches text in angle brack-
4680 ets, allowing for arbitrary nesting. Only digits are allowed in nested
4681 brackets (that is, when recursing), whereas any characters are permit-
4682 ted at the outer level.
4683
4684 < (?: (?(R) \d++ | [^<>]*+) | (?R)) * >
4685
4686 In this pattern, (?(R) is the start of a conditional subpattern, with
4687 two different alternatives for the recursive and non-recursive cases.
4688 The (?R) item is the actual recursive call.
4689
4690
4691 SUBPATTERNS AS SUBROUTINES
4692
4693 If the syntax for a recursive subpattern reference (either by number or
4694 by name) is used outside the parentheses to which it refers, it oper-
4695 ates like a subroutine in a programming language. The "called" subpat-
4696 tern may be defined before or after the reference. A numbered reference
4697 can be absolute or relative, as in these examples:
4698
4699 (...(absolute)...)...(?2)...
4700 (...(relative)...)...(?-1)...
4701 (...(?+1)...(relative)...
4702
4703 An earlier example pointed out that the pattern
4704
4705 (sens|respons)e and \1ibility
4706
4707 matches "sense and sensibility" and "response and responsibility", but
4708 not "sense and responsibility". If instead the pattern
4709
4710 (sens|respons)e and (?1)ibility
4711
4712 is used, it does match "sense and responsibility" as well as the other
4713 two strings. Another example is given in the discussion of DEFINE
4714 above.
4715
4716 Like recursive subpatterns, a "subroutine" call is always treated as an
4717 atomic group. That is, once it has matched some of the subject string,
4718 it is never re-entered, even if it contains untried alternatives and
4719 there is a subsequent matching failure.
4720
4721 When a subpattern is used as a subroutine, processing options such as
4722 case-independence are fixed when the subpattern is defined. They cannot
4723 be changed for different calls. For example, consider this pattern:
4724
4725 (abc)(?i:(?-1))
4726
4727 It matches "abcabc". It does not match "abcABC" because the change of
4728 processing option does not affect the called subpattern.
4729
4730
4731 CALLOUTS
4732
4733 Perl has a feature whereby using the sequence (?{...}) causes arbitrary
4734 Perl code to be obeyed in the middle of matching a regular expression.
4735 This makes it possible, amongst other things, to extract different sub-
4736 strings that match the same pair of parentheses when there is a repeti-
4737 tion.
4738
4739 PCRE provides a similar feature, but of course it cannot obey arbitrary
4740 Perl code. The feature is called "callout". The caller of PCRE provides
4741 an external function by putting its entry point in the global variable
4742 pcre_callout. By default, this variable contains NULL, which disables
4743 all calling out.
4744
4745 Within a regular expression, (?C) indicates the points at which the
4746 external function is to be called. If you want to identify different
4747 callout points, you can put a number less than 256 after the letter C.
4748 The default value is zero. For example, this pattern has two callout
4749 points:
4750
4751 (?C1)abc(?C2)def
4752
4753 If the PCRE_AUTO_CALLOUT flag is passed to pcre_compile(), callouts are
4754 automatically installed before each item in the pattern. They are all
4755 numbered 255.
4756
4757 During matching, when PCRE reaches a callout point (and pcre_callout is
4758 set), the external function is called. It is provided with the number
4759 of the callout, the position in the pattern, and, optionally, one item
4760 of data originally supplied by the caller of pcre_exec(). The callout
4761 function may cause matching to proceed, to backtrack, or to fail alto-
4762 gether. A complete description of the interface to the callout function
4763 is given in the pcrecallout documentation.
4764
4765
4766 BACKTRACKING CONTROL
4767
4768 Perl 5.10 introduced a number of "Special Backtracking Control Verbs",
4769 which are described in the Perl documentation as "experimental and sub-
4770 ject to change or removal in a future version of Perl". It goes on to
4771 say: "Their usage in production code should be noted to avoid problems
4772 during upgrades." The same remarks apply to the PCRE features described
4773 in this section.
4774
4775 Since these verbs are specifically related to backtracking, they can be
4776 used only when the pattern is to be matched using pcre_exec(), which
4777 uses a backtracking algorithm. They cause an error if encountered by
4778 pcre_dfa_exec().
4779
4780 The new verbs make use of what was previously invalid syntax: an open-
4781 ing parenthesis followed by an asterisk. In Perl, they are generally of
4782 the form (*VERB:ARG) but PCRE does not support the use of arguments, so
4783 its general form is just (*VERB). Any number of these verbs may occur
4784 in a pattern. There are two kinds:
4785
4786 Verbs that act immediately
4787
4788 The following verbs act as soon as they are encountered:
4789
4790 (*ACCEPT)
4791
4792 This verb causes the match to end successfully, skipping the remainder
4793 of the pattern. When inside a recursion, only the innermost pattern is
4794 ended immediately. PCRE differs from Perl in what happens if the
4795 (*ACCEPT) is inside capturing parentheses. In Perl, the data so far is
4796 captured: in PCRE no data is captured. For example:
4797
4798 A(A|B(*ACCEPT)|C)D
4799
4800 This matches "AB", "AAD", or "ACD", but when it matches "AB", no data
4801 is captured.
4802
4803 (*FAIL) or (*F)
4804
4805 This verb causes the match to fail, forcing backtracking to occur. It
4806 is equivalent to (?!) but easier to read. The Perl documentation notes
4807 that it is probably useful only when combined with (?{}) or (??{}).
4808 Those are, of course, Perl features that are not present in PCRE. The
4809 nearest equivalent is the callout feature, as for example in this pat-
4810 tern:
4811
4812 a+(?C)(*FAIL)
4813
4814 A match with the string "aaaa" always fails, but the callout is taken
4815 before each backtrack happens (in this example, 10 times).
4816
4817 Verbs that act after backtracking
4818
4819 The following verbs do nothing when they are encountered. Matching con-
4820 tinues with what follows, but if there is no subsequent match, a fail-
4821 ure is forced. The verbs differ in exactly what kind of failure
4822 occurs.
4823
4824 (*COMMIT)
4825
4826 This verb causes the whole match to fail outright if the rest of the
4827 pattern does not match. Even if the pattern is unanchored, no further
4828 attempts to find a match by advancing the start point take place. Once
4829 (*COMMIT) has been passed, pcre_exec() is committed to finding a match
4830 at the current starting point, or not at all. For example:
4831
4832 a+(*COMMIT)b
4833
4834 This matches "xxaab" but not "aacaab". It can be thought of as a kind
4835 of dynamic anchor, or "I've started, so I must finish."
4836
4837 (*PRUNE)
4838
4839 This verb causes the match to fail at the current position if the rest
4840 of the pattern does not match. If the pattern is unanchored, the normal
4841 "bumpalong" advance to the next starting character then happens. Back-
4842 tracking can occur as usual to the left of (*PRUNE), or when matching
4843 to the right of (*PRUNE), but if there is no match to the right, back-
4844 tracking cannot cross (*PRUNE). In simple cases, the use of (*PRUNE)
4845 is just an alternative to an atomic group or possessive quantifier, but
4846 there are some uses of (*PRUNE) that cannot be expressed in any other
4847 way.
4848
4849 (*SKIP)
4850
4851 This verb is like (*PRUNE), except that if the pattern is unanchored,
4852 the "bumpalong" advance is not to the next character, but to the posi-
4853 tion in the subject where (*SKIP) was encountered. (*SKIP) signifies
4854 that whatever text was matched leading up to it cannot be part of a
4855 successful match. Consider:
4856
4857 a+(*SKIP)b
4858
4859 If the subject is "aaaac...", after the first match attempt fails
4860 (starting at the first character in the string), the starting point
4861 skips on to start the next attempt at "c". Note that a possessive quan-
4862 tifer does not have the same effect in this example; although it would
4863 suppress backtracking during the first match attempt, the second
4864 attempt would start at the second character instead of skipping on to
4865 "c".
4866
4867 (*THEN)
4868
4869 This verb causes a skip to the next alternation if the rest of the pat-
4870 tern does not match. That is, it cancels pending backtracking, but only
4871 within the current alternation. Its name comes from the observation
4872 that it can be used for a pattern-based if-then-else block:
4873
4874 ( COND1 (*THEN) FOO | COND2 (*THEN) BAR | COND3 (*THEN) BAZ ) ...
4875
4876 If the COND1 pattern matches, FOO is tried (and possibly further items
4877 after the end of the group if FOO succeeds); on failure the matcher
4878 skips to the second alternative and tries COND2, without backtracking
4879 into COND1. If (*THEN) is used outside of any alternation, it acts
4880 exactly like (*PRUNE).
4881
4882
4883 SEE ALSO
4884
4885 pcreapi(3), pcrecallout(3), pcrematching(3), pcre(3).
4886
4887
4888 AUTHOR
4889
4890 Philip Hazel
4891 University Computing Service
4892 Cambridge CB2 3QH, England.
4893
4894
4895 REVISION
4896
4897 Last updated: 17 September 2007
4898 Copyright (c) 1997-2007 University of Cambridge.
4899 ------------------------------------------------------------------------------
4900
4901
4902 PCRESYNTAX(3) PCRESYNTAX(3)
4903
4904
4905 NAME
4906 PCRE - Perl-compatible regular expressions
4907
4908
4909 PCRE REGULAR EXPRESSION SYNTAX SUMMARY
4910
4911 The full syntax and semantics of the regular expressions that are sup-
4912 ported by PCRE are described in the pcrepattern documentation. This
4913 document contains just a quick-reference summary of the syntax.
4914
4915
4916 QUOTING
4917
4918 \x where x is non-alphanumeric is a literal x
4919 \Q...\E treat enclosed characters as literal
4920
4921
4922 CHARACTERS
4923
4924 \a alarm, that is, the BEL character (hex 07)
4925 \cx "control-x", where x is any character
4926 \e escape (hex 1B)
4927 \f formfeed (hex 0C)
4928 \n newline (hex 0A)
4929 \r carriage return (hex 0D)
4930 \t tab (hex 09)
4931 \ddd character with octal code ddd, or backreference
4932 \xhh character with hex code hh
4933 \x{hhh..} character with hex code hhh..
4934
4935
4936 CHARACTER TYPES
4937
4938 . any character except newline;
4939 in dotall mode, any character whatsoever
4940 \C one byte, even in UTF-8 mode (best avoided)
4941 \d a decimal digit
4942 \D a character that is not a decimal digit
4943 \h a horizontal whitespace character
4944 \H a character that is not a horizontal whitespace character
4945 \p{xx} a character with the xx property
4946 \P{xx} a character without the xx property
4947 \R a newline sequence
4948 \s a whitespace character
4949 \S a character that is not a whitespace character
4950 \v a vertical whitespace character
4951 \V a character that is not a vertical whitespace character
4952 \w a "word" character
4953 \W a "non-word" character
4954 \X an extended Unicode sequence
4955
4956 In PCRE, \d, \D, \s, \S, \w, and \W recognize only ASCII characters.
4957
4958
4959 GENERAL CATEGORY PROPERTY CODES FOR \p and \P
4960
4961 C Other
4962 Cc Control
4963 Cf Format
4964 Cn Unassigned
4965 Co Private use
4966 Cs Surrogate
4967
4968 L Letter
4969 Ll Lower case letter
4970 Lm Modifier letter
4971 Lo Other letter
4972 Lt Title case letter
4973 Lu Upper case letter
4974 L& Ll, Lu, or Lt
4975
4976 M Mark
4977 Mc Spacing mark
4978 Me Enclosing mark
4979 Mn Non-spacing mark
4980
4981 N Number
4982 Nd Decimal number
4983 Nl Letter number
4984 No Other number
4985
4986 P Punctuation
4987 Pc Connector punctuation
4988 Pd Dash punctuation
4989 Pe Close punctuation
4990 Pf Final punctuation
4991 Pi Initial punctuation
4992 Po Other punctuation
4993 Ps Open punctuation
4994
4995 S Symbol
4996 Sc Currency symbol
4997 Sk Modifier symbol
4998 Sm Mathematical symbol
4999 So Other symbol
5000
5001 Z Separator
5002 Zl Line separator
5003 Zp Paragraph separator
5004 Zs Space separator
5005
5006
5007 SCRIPT NAMES FOR \p AND \P
5008
5009 Arabic, Armenian, Balinese, Bengali, Bopomofo, Braille, Buginese,
5010 Buhid, Canadian_Aboriginal, Cherokee, Common, Coptic, Cuneiform,
5011 Cypriot, Cyrillic, Deseret, Devanagari, Ethiopic, Georgian, Glagolitic,
5012 Gothic, Greek, Gujarati, Gurmukhi, Han, Hangul, Hanunoo, Hebrew, Hira-
5013 gana, Inherited, Kannada, Katakana, Kharoshthi, Khmer, Lao, Latin,
5014 Limbu, Linear_B, Malayalam, Mongolian, Myanmar, New_Tai_Lue, Nko,
5015 Ogham, Old_Italic, Old_Persian, Oriya, Osmanya, Phags_Pa, Phoenician,
5016 Runic, Shavian, Sinhala, Syloti_Nagri, Syriac, Tagalog, Tagbanwa,
5017 Tai_Le, Tamil, Telugu, Thaana, Thai, Tibetan, Tifinagh, Ugaritic, Yi.
5018
5019
5020 CHARACTER CLASSES
5021
5022 [...] positive character class
5023 [^...] negative character class
5024 [x-y] range (can be used for hex characters)
5025 [[:xxx:]] positive POSIX named set
5026 [[:^xxx:]] negative POSIX named set
5027
5028 alnum alphanumeric
5029 alpha alphabetic
5030 ascii 0-127
5031 blank space or tab
5032 cntrl control character
5033 digit decimal digit
5034 graph printing, excluding space
5035 lower lower case letter
5036 print printing, including space
5037 punct printing, excluding alphanumeric
5038 space whitespace
5039 upper upper case letter
5040 word same as \w
5041 xdigit hexadecimal digit
5042
5043 In PCRE, POSIX character set names recognize only ASCII characters. You
5044 can use \Q...\E inside a character class.
5045
5046
5047 QUANTIFIERS
5048
5049 ? 0 or 1, greedy
5050 ?+ 0 or 1, possessive
5051 ?? 0 or 1, lazy
5052 * 0 or more, greedy
5053 *+ 0 or more, possessive
5054 *? 0 or more, lazy
5055 + 1 or more, greedy
5056 ++ 1 or more, possessive
5057 +? 1 or more, lazy
5058 {n} exactly n
5059 {n,m} at least n, no more than m, greedy
5060 {n,m}+ at least n, no more than m, possessive
5061 {n,m}? at least n, no more than m, lazy
5062 {n,} n or more, greedy
5063 {n,}+ n or more, possessive
5064 {n,}? n or more, lazy
5065
5066
5067 ANCHORS AND SIMPLE ASSERTIONS
5068
5069 \b word boundary
5070 \B not a word boundary
5071 ^ start of subject
5072 also after internal newline in multiline mode
5073 \A start of subject
5074 $ end of subject
5075 also before newline at end of subject
5076 also before internal newline in multiline mode
5077 \Z end of subject
5078 also before newline at end of subject
5079 \z end of subject
5080 \G first matching position in subject
5081
5082
5083 MATCH POINT RESET
5084
5085 \K reset start of match
5086
5087
5088 ALTERNATION
5089
5090 expr|expr|expr...
5091
5092
5093 CAPTURING
5094
5095 (...) capturing group
5096 (?<name>...) named capturing group (Perl)
5097 (?'name'...) named capturing group (Perl)
5098 (?P<name>...) named capturing group (Python)
5099 (?:...) non-capturing group
5100 (?|...) non-capturing group; reset group numbers for
5101 capturing groups in each alternative
5102
5103
5104 ATOMIC GROUPS
5105
5106 (?>...) atomic, non-capturing group
5107
5108
5109 COMMENT
5110
5111 (?#....) comment (not nestable)
5112
5113
5114 OPTION SETTING
5115
5116 (?i) caseless
5117 (?J) allow duplicate names
5118 (?m) multiline
5119 (?s) single line (dotall)
5120 (?U) default ungreedy (lazy)
5121 (?x) extended (ignore white space)
5122 (?-...) unset option(s)
5123
5124
5125 LOOKAHEAD AND LOOKBEHIND ASSERTIONS
5126
5127 (?=...) positive look ahead
5128 (?!...) negative look ahead
5129 (?<=...) positive look behind
5130 (?<!...) negative look behind
5131
5132 Each top-level branch of a look behind must be of a fixed length.
5133
5134
5135 BACKREFERENCES
5136
5137 \n reference by number (can be ambiguous)
5138 \gn reference by number
5139 \g{n} reference by number
5140 \g{-n} relative reference by number
5141 \k<name> reference by name (Perl)
5142 \k'name' reference by name (Perl)
5143 \g{name} reference by name (Perl)
5144 \k{name} reference by name (.NET)
5145 (?P=name) reference by name (Python)
5146
5147
5148 SUBROUTINE REFERENCES (POSSIBLY RECURSIVE)
5149
5150 (?R) recurse whole pattern
5151 (?n) call subpattern by absolute number
5152 (?+n) call subpattern by relative number
5153 (?-n) call subpattern by relative number
5154 (?&name) call subpattern by name (Perl)
5155 (?P>name) call subpattern by name (Python)
5156
5157
5158 CONDITIONAL PATTERNS
5159
5160 (?(condition)yes-pattern)
5161 (?(condition)yes-pattern|no-pattern)
5162
5163 (?(n)... absolute reference condition
5164 (?(+n)... relative reference condition
5165 (?(-n)... relative reference condition
5166 (?(<name>)... named reference condition (Perl)
5167 (?('name')... named reference condition (Perl)
5168 (?(name)... named reference condition (PCRE)
5169 (?(R)... overall recursion condition
5170 (?(Rn)... specific group recursion condition
5171 (?(R&name)... specific recursion condition
5172 (?(DEFINE)... define subpattern for reference
5173 (?(assert)... assertion condition
5174
5175
5176 BACKTRACKING CONTROL
5177
5178 The following act immediately they are reached:
5179
5180 (*ACCEPT) force successful match
5181 (*FAIL) force backtrack; synonym (*F)
5182
5183 The following act only when a subsequent match failure causes a back-
5184 track to reach them. They all force a match failure, but they differ in
5185 what happens afterwards. Those that advance the start-of-match point do
5186 so only if the pattern is not anchored.
5187
5188 (*COMMIT) overall failure, no advance of starting point
5189 (*PRUNE) advance to next starting character
5190 (*SKIP) advance start to current matching position
5191 (*THEN) local failure, backtrack to next alternation
5192
5193
5194 NEWLINE CONVENTIONS
5195
5196 These are recognized only at the very start of the pattern or after a
5197 (*BSR_...) option.
5198
5199 (*CR)
5200 (*LF)
5201 (*CRLF)
5202 (*ANYCRLF)
5203 (*ANY)
5204
5205
5206 WHAT \R MATCHES
5207
5208 These are recognized only at the very start of the pattern or after a
5209 (*...) option that sets the newline convention.
5210
5211 (*BSR_ANYCRLF)
5212 (*BSR_UNICODE)
5213
5214
5215 CALLOUTS
5216
5217 (?C) callout
5218 (?Cn) callout with data n
5219
5220
5221 SEE ALSO
5222
5223 pcrepattern(3), pcreapi(3), pcrecallout(3), pcrematching(3), pcre(3).
5224
5225
5226 AUTHOR
5227
5228 Philip Hazel
5229 University Computing Service
5230 Cambridge CB2 3QH, England.
5231
5232
5233 REVISION
5234
5235 Last updated: 14 November 2007
5236 Copyright (c) 1997-2007 University of Cambridge.
5237 ------------------------------------------------------------------------------
5238
5239
5240 PCREPARTIAL(3) PCREPARTIAL(3)
5241
5242
5243 NAME
5244 PCRE - Perl-compatible regular expressions
5245
5246
5247 PARTIAL MATCHING IN PCRE
5248
5249 In normal use of PCRE, if the subject string that is passed to
5250 pcre_exec() or pcre_dfa_exec() matches as far as it goes, but is too
5251 short to match the entire pattern, PCRE_ERROR_NOMATCH is returned.
5252 There are circumstances where it might be helpful to distinguish this
5253 case from other cases in which there is no match.
5254
5255 Consider, for example, an application where a human is required to type
5256 in data for a field with specific formatting requirements. An example
5257 might be a date in the form ddmmmyy, defined by this pattern:
5258
5259 ^\d?\d(jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)\d\d$
5260
5261 If the application sees the user's keystrokes one by one, and can check
5262 that what has been typed so far is potentially valid, it is able to
5263 raise an error as soon as a mistake is made, possibly beeping and not
5264 reflecting the character that has been typed. This immediate feedback
5265 is likely to be a better user interface than a check that is delayed
5266 until the entire string has been entered.
5267
5268 PCRE supports the concept of partial matching by means of the PCRE_PAR-
5269 TIAL option, which can be set when calling pcre_exec() or
5270 pcre_dfa_exec(). When this flag is set for pcre_exec(), the return code
5271 PCRE_ERROR_NOMATCH is converted into PCRE_ERROR_PARTIAL if at any time
5272 during the matching process the last part of the subject string matched
5273 part of the pattern. Unfortunately, for non-anchored matching, it is
5274 not possible to obtain the position of the start of the partial match.
5275 No captured data is set when PCRE_ERROR_PARTIAL is returned.
5276
5277 When PCRE_PARTIAL is set for pcre_dfa_exec(), the return code
5278 PCRE_ERROR_NOMATCH is converted into PCRE_ERROR_PARTIAL if the end of
5279 the subject is reached, there have been no complete matches, but there
5280 is still at least one matching possibility. The portion of the string
5281 that provided the partial match is set as the first matching string.
5282
5283 Using PCRE_PARTIAL disables one of PCRE's optimizations. PCRE remembers
5284 the last literal byte in a pattern, and abandons matching immediately
5285 if such a byte is not present in the subject string. This optimization
5286 cannot be used for a subject string that might match only partially.
5287
5288
5289 RESTRICTED PATTERNS FOR PCRE_PARTIAL
5290
5291 Because of the way certain internal optimizations are implemented in
5292 the pcre_exec() function, the PCRE_PARTIAL option cannot be used with
5293 all patterns. These restrictions do not apply when pcre_dfa_exec() is
5294 used. For pcre_exec(), repeated single characters such as
5295
5296 a{2,4}
5297
5298 and repeated single metasequences such as
5299
5300 \d+
5301
5302 are not permitted if the maximum number of occurrences is greater than
5303 one. Optional items such as \d? (where the maximum is one) are permit-
5304 ted. Quantifiers with any values are permitted after parentheses, so
5305 the invalid examples above can be coded thus:
5306
5307 (a){2,4}
5308 (\d)+
5309
5310 These constructions run more slowly, but for the kinds of application
5311 that are envisaged for this facility, this is not felt to be a major
5312 restriction.
5313
5314 If PCRE_PARTIAL is set for a pattern that does not conform to the
5315 restrictions, pcre_exec() returns the error code PCRE_ERROR_BADPARTIAL
5316 (-13). You can use the PCRE_INFO_OKPARTIAL call to pcre_fullinfo() to
5317 find out if a compiled pattern can be used for partial matching.
5318
5319
5320 EXAMPLE OF PARTIAL MATCHING USING PCRETEST
5321
5322 If the escape sequence \P is present in a pcretest data line, the
5323 PCRE_PARTIAL flag is used for the match. Here is a run of pcretest that
5324 uses the date example quoted above:
5325
5326 re> /^\d?\d(jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)\d\d$/
5327 data> 25jun04\P
5328 0: 25jun04
5329 1: jun
5330 data> 25dec3\P
5331 Partial match
5332 data> 3ju\P
5333 Partial match
5334 data> 3juj\P
5335 No match
5336 data> j\P
5337 No match
5338
5339 The first data string is matched completely, so pcretest shows the
5340 matched substrings. The remaining four strings do not match the com-
5341 plete pattern, but the first two are partial matches. The same test,
5342 using pcre_dfa_exec() matching (by means of the \D escape sequence),
5343 produces the following output:
5344
5345 re> /^\d?\d(jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)\d\d$/
5346 data> 25jun04\P\D
5347 0: 25jun04
5348 data> 23dec3\P\D
5349 Partial match: 23dec3
5350 data> 3ju\P\D
5351 Partial match: 3ju
5352 data> 3juj\P\D
5353 No match
5354 data> j\P\D
5355 No match
5356
5357 Notice that in this case the portion of the string that was matched is
5358 made available.
5359
5360
5361 MULTI-SEGMENT MATCHING WITH pcre_dfa_exec()
5362
5363 When a partial match has been found using pcre_dfa_exec(), it is possi-
5364 ble to continue the match by providing additional subject data and
5365 calling pcre_dfa_exec() again with the same compiled regular expres-
5366 sion, this time setting the PCRE_DFA_RESTART option. You must also pass
5367 the same working space as before, because this is where details of the
5368 previous partial match are stored. Here is an example using pcretest,
5369 using the \R escape sequence to set the PCRE_DFA_RESTART option (\P and
5370 \D are as above):
5371
5372 re> /^\d?\d(jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)\d\d$/
5373 data> 23ja\P\D
5374 Partial match: 23ja
5375 data> n05\R\D
5376 0: n05
5377
5378 The first call has "23ja" as the subject, and requests partial match-
5379 ing; the second call has "n05" as the subject for the continued
5380 (restarted) match. Notice that when the match is complete, only the
5381 last part is shown; PCRE does not retain the previously partially-
5382 matched string. It is up to the calling program to do that if it needs
5383 to.
5384
5385 You can set PCRE_PARTIAL with PCRE_DFA_RESTART to continue partial
5386 matching over multiple segments. This facility can be used to pass very
5387 long subject strings to pcre_dfa_exec(). However, some care is needed
5388 for certain types of pattern.
5389
5390 1. If the pattern contains tests for the beginning or end of a line,
5391 you need to pass the PCRE_NOTBOL or PCRE_NOTEOL options, as appropri-
5392 ate, when the subject string for any call does not contain the begin-
5393 ning or end of a line.
5394
5395 2. If the pattern contains backward assertions (including \b or \B),
5396 you need to arrange for some overlap in the subject strings to allow
5397 for this. For example, you could pass the subject in chunks that are
5398 500 bytes long, but in a buffer of 700 bytes, with the starting offset
5399 set to 200 and the previous 200 bytes at the start of the buffer.
5400
5401 3. Matching a subject string that is split into multiple segments does
5402 not always produce exactly the same result as matching over one single
5403 long string. The difference arises when there are multiple matching
5404 possibilities, because a partial match result is given only when there
5405 are no completed matches in a call to pcre_dfa_exec(). This means that
5406 as soon as the shortest match has been found, continuation to a new
5407 subject segment is no longer possible. Consider this pcretest example:
5408
5409 re> /dog(sbody)?/
5410 data> do\P\D
5411 Partial match: do
5412 data> gsb\R\P\D
5413 0: g
5414 data> dogsbody\D
5415 0: dogsbody
5416 1: dog
5417
5418 The pattern matches the words "dog" or "dogsbody". When the subject is
5419 presented in several parts ("do" and "gsb" being the first two) the
5420 match stops when "dog" has been found, and it is not possible to con-
5421 tinue. On the other hand, if "dogsbody" is presented as a single
5422 string, both matches are found.
5423
5424 Because of this phenomenon, it does not usually make sense to end a
5425 pattern that is going to be matched in this way with a variable repeat.
5426
5427 4. Patterns that contain alternatives at the top level which do not all
5428 start with the same pattern item may not work as expected. For example,
5429 consider this pattern:
5430
5431 1234|3789
5432
5433 If the first part of the subject is "ABC123", a partial match of the
5434 first alternative is found at offset 3. There is no partial match for
5435 the second alternative, because such a match does not start at the same
5436 point in the subject string. Attempting to continue with the string
5437 "789" does not yield a match because only those alternatives that match
5438 at one point in the subject are remembered. The problem arises because
5439 the start of the second alternative matches within the first alterna-
5440 tive. There is no problem with anchored patterns or patterns such as:
5441
5442 1234|ABCD
5443
5444 where no string can be a partial match for both alternatives.
5445
5446
5447 AUTHOR
5448
5449 Philip Hazel
5450 University Computing Service
5451 Cambridge CB2 3QH, England.
5452
5453
5454 REVISION
5455
5456 Last updated: 04 June 2007
5457 Copyright (c) 1997-2007 University of Cambridge.
5458 ------------------------------------------------------------------------------
5459
5460
5461 PCREPRECOMPILE(3) PCREPRECOMPILE(3)
5462
5463
5464 NAME
5465 PCRE - Perl-compatible regular expressions
5466
5467
5468 SAVING AND RE-USING PRECOMPILED PCRE PATTERNS
5469
5470 If you are running an application that uses a large number of regular
5471 expression patterns, it may be useful to store them in a precompiled
5472 form instead of having to compile them every time the application is
5473 run. If you are not using any private character tables (see the
5474 pcre_maketables() documentation), this is relatively straightforward.
5475 If you are using private tables, it is a little bit more complicated.
5476
5477 If you save compiled patterns to a file, you can copy them to a differ-
5478 ent host and run them there. This works even if the new host has the
5479 opposite endianness to the one on which the patterns were compiled.
5480 There may be a small performance penalty, but it should be insignifi-
5481 cant. However, compiling regular expressions with one version of PCRE
5482 for use with a different version is not guaranteed to work and may
5483 cause crashes.
5484
5485
5486 SAVING A COMPILED PATTERN
5487 The value returned by pcre_compile() points to a single block of memory
5488 that holds the compiled pattern and associated data. You can find the
5489 length of this block in bytes by calling pcre_fullinfo() with an argu-
5490 ment of PCRE_INFO_SIZE. You can then save the data in any appropriate
5491 manner. Here is sample code that compiles a pattern and writes it to a
5492 file. It assumes that the variable fd refers to a file that is open for
5493 output:
5494
5495 int erroroffset, rc, size;
5496 char *error;
5497 pcre *re;
5498
5499 re = pcre_compile("my pattern", 0, &error, &erroroffset, NULL);
5500 if (re == NULL) { ... handle errors ... }
5501 rc = pcre_fullinfo(re, NULL, PCRE_INFO_SIZE, &size);
5502 if (rc < 0) { ... handle errors ... }
5503 rc = fwrite(re, 1, size, fd);
5504 if (rc != size) { ... handle errors ... }
5505
5506 In this example, the bytes that comprise the compiled pattern are
5507 copied exactly. Note that this is binary data that may contain any of
5508 the 256 possible byte values. On systems that make a distinction
5509 between binary and non-binary data, be sure that the file is opened for
5510 binary output.
5511
5512 If you want to write more than one pattern to a file, you will have to
5513 devise a way of separating them. For binary data, preceding each pat-
5514 tern with its length is probably the most straightforward approach.
5515 Another possibility is to write out the data in hexadecimal instead of
5516 binary, one pattern to a line.
5517
5518 Saving compiled patterns in a file is only one possible way of storing
5519 them for later use. They could equally well be saved in a database, or
5520 in the memory of some daemon process that passes them via sockets to
5521 the processes that want them.
5522
5523 If the pattern has been studied, it is also possible to save the study
5524 data in a similar way to the compiled pattern itself. When studying
5525 generates additional information, pcre_study() returns a pointer to a
5526 pcre_extra data block. Its format is defined in the section on matching
5527 a pattern in the pcreapi documentation. The study_data field points to
5528 the binary study data, and this is what you must save (not the
5529 pcre_extra block itself). The length of the study data can be obtained
5530 by calling pcre_fullinfo() with an argument of PCRE_INFO_STUDYSIZE.
5531 Remember to check that pcre_study() did return a non-NULL value before
5532 trying to save the study data.
5533
5534
5535 RE-USING A PRECOMPILED PATTERN
5536
5537 Re-using a precompiled pattern is straightforward. Having reloaded it
5538 into main memory, you pass its pointer to pcre_exec() or
5539 pcre_dfa_exec() in the usual way. This should work even on another
5540 host, and even if that host has the opposite endianness to the one
5541 where the pattern was compiled.
5542
5543 However, if you passed a pointer to custom character tables when the
5544 pattern was compiled (the tableptr argument of pcre_compile()), you
5545 must now pass a similar pointer to pcre_exec() or pcre_dfa_exec(),
5546 because the value saved with the compiled pattern will obviously be
5547 nonsense. A field in a pcre_extra() block is used to pass this data, as
5548 described in the section on matching a pattern in the pcreapi documen-
5549 tation.
5550
5551 If you did not provide custom character tables when the pattern was
5552 compiled, the pointer in the compiled pattern is NULL, which causes
5553 pcre_exec() to use PCRE's internal tables. Thus, you do not need to
5554 take any special action at run time in this case.
5555
5556 If you saved study data with the compiled pattern, you need to create
5557 your own pcre_extra data block and set the study_data field to point to
5558 the reloaded study data. You must also set the PCRE_EXTRA_STUDY_DATA
5559 bit in the flags field to indicate that study data is present. Then
5560 pass the pcre_extra block to pcre_exec() or pcre_dfa_exec() in the
5561 usual way.
5562
5563
5564 COMPATIBILITY WITH DIFFERENT PCRE RELEASES
5565
5566 In general, it is safest to recompile all saved patterns when you
5567 update to a new PCRE release, though not all updates actually require
5568 this. Recompiling is definitely needed for release 7.2.
5569
5570
5571 AUTHOR
5572
5573 Philip Hazel
5574 University Computing Service
5575 Cambridge CB2 3QH, England.
5576
5577
5578 REVISION
5579
5580 Last updated: 13 June 2007
5581 Copyright (c) 1997-2007 University of Cambridge.
5582 ------------------------------------------------------------------------------
5583
5584
5585 PCREPERFORM(3) PCREPERFORM(3)
5586
5587
5588 NAME
5589 PCRE - Perl-compatible regular expressions
5590
5591
5592 PCRE PERFORMANCE
5593
5594 Two aspects of performance are discussed below: memory usage and pro-
5595 cessing time. The way you express your pattern as a regular expression
5596 can affect both of them.
5597
5598
5599 MEMORY USAGE
5600
5601 Patterns are compiled by PCRE into a reasonably efficient byte code, so
5602 that most simple patterns do not use much memory. However, there is one
5603 case where memory usage can be unexpectedly large. When a parenthesized
5604 subpattern has a quantifier with a minimum greater than 1 and/or a lim-
5605 ited maximum, the whole subpattern is repeated in the compiled code.
5606 For example, the pattern
5607
5608 (abc|def){2,4}
5609
5610 is compiled as if it were
5611
5612 (abc|def)(abc|def)((abc|def)(abc|def)?)?
5613
5614 (Technical aside: It is done this way so that backtrack points within
5615 each of the repetitions can be independently maintained.)
5616
5617 For regular expressions whose quantifiers use only small numbers, this
5618 is not usually a problem. However, if the numbers are large, and par-
5619 ticularly if such repetitions are nested, the memory usage can become
5620 an embarrassment. For example, the very simple pattern
5621
5622 ((ab){1,1000}c){1,3}
5623
5624 uses 51K bytes when compiled. When PCRE is compiled with its default
5625 internal pointer size of two bytes, the size limit on a compiled pat-
5626 tern is 64K, and this is reached with the above pattern if the outer
5627 repetition is increased from 3 to 4. PCRE can be compiled to use larger
5628 internal pointers and thus handle larger compiled patterns, but it is
5629 better to try to rewrite your pattern to use less memory if you can.
5630
5631 One way of reducing the memory usage for such patterns is to make use
5632 of PCRE's "subroutine" facility. Re-writing the above pattern as
5633
5634 ((ab)(?2){0,999}c)(?1){0,2}
5635
5636 reduces the memory requirements to 18K, and indeed it remains under 20K
5637 even with the outer repetition increased to 100. However, this pattern
5638 is not exactly equivalent, because the "subroutine" calls are treated
5639 as atomic groups into which there can be no backtracking if there is a
5640 subsequent matching failure. Therefore, PCRE cannot do this kind of
5641 rewriting automatically. Furthermore, there is a noticeable loss of
5642 speed when executing the modified pattern. Nevertheless, if the atomic
5643 grouping is not a problem and the loss of speed is acceptable, this
5644 kind of rewriting will allow you to process patterns that PCRE cannot
5645 otherwise handle.
5646
5647
5648 PROCESSING TIME
5649
5650 Certain items in regular expression patterns are processed more effi-
5651 ciently than others. It is more efficient to use a character class like
5652 [aeiou] than a set of single-character alternatives such as
5653 (a|e|i|o|u). In general, the simplest construction that provides the
5654 required behaviour is usually the most efficient. Jeffrey Friedl's book
5655 contains a lot of useful general discussion about optimizing regular
5656 expressions for efficient performance. This document contains a few
5657 observations about PCRE.
5658
5659 Using Unicode character properties (the \p, \P, and \X escapes) is
5660 slow, because PCRE has to scan a structure that contains data for over
5661 fifteen thousand characters whenever it needs a character's property.
5662 If you can find an alternative pattern that does not use character
5663 properties, it will probably be faster.
5664
5665 When a pattern begins with .* not in parentheses, or in parentheses
5666 that are not the subject of a backreference, and the PCRE_DOTALL option
5667 is set, the pattern is implicitly anchored by PCRE, since it can match
5668 only at the start of a subject string. However, if PCRE_DOTALL is not
5669 set, PCRE cannot make this optimization, because the . metacharacter
5670 does not then match a newline, and if the subject string contains new-
5671 lines, the pattern may match from the character immediately following
5672 one of them instead of from the very start. For example, the pattern
5673
5674 .*second
5675
5676 matches the subject "first\nand second" (where \n stands for a newline
5677 character), with the match starting at the seventh character. In order
5678 to do this, PCRE has to retry the match starting after every newline in
5679 the subject.
5680
5681 If you are using such a pattern with subject strings that do not con-
5682 tain newlines, the best performance is obtained by setting PCRE_DOTALL,
5683 or starting the pattern with ^.* or ^.*? to indicate explicit anchor-
5684 ing. That saves PCRE from having to scan along the subject looking for
5685 a newline to restart at.
5686
5687 Beware of patterns that contain nested indefinite repeats. These can
5688 take a long time to run when applied to a string that does not match.
5689 Consider the pattern fragment
5690
5691 ^(a+)*
5692
5693 This can match "aaaa" in 16 different ways, and this number increases
5694 very rapidly as the string gets longer. (The * repeat can match 0, 1,
5695 2, 3, or 4 times, and for each of those cases other than 0 or 4, the +
5696 repeats can match different numbers of times.) When the remainder of
5697 the pattern is such that the entire match is going to fail, PCRE has in
5698 principle to try every possible variation, and this can take an
5699 extremely long time, even for relatively short strings.
5700
5701 An optimization catches some of the more simple cases such as
5702
5703 (a+)*b
5704
5705 where a literal character follows. Before embarking on the standard
5706 matching procedure, PCRE checks that there is a "b" later in the sub-
5707 ject string, and if there is not, it fails the match immediately. How-
5708 ever, when there is no following literal this optimization cannot be
5709 used. You can see the difference by comparing the behaviour of
5710
5711 (a+)*\d
5712
5713 with the pattern above. The former gives a failure almost instantly
5714 when applied to a whole line of "a" characters, whereas the latter
5715 takes an appreciable time with strings longer than about 20 characters.
5716
5717 In many cases, the solution to this kind of performance issue is to use
5718 an atomic group or a possessive quantifier.
5719
5720
5721 AUTHOR
5722
5723 Philip Hazel
5724 University Computing Service
5725 Cambridge CB2 3QH, England.
5726
5727
5728 REVISION
5729
5730 Last updated: 06 March 2007
5731 Copyright (c) 1997-2007 University of Cambridge.
5732 ------------------------------------------------------------------------------
5733
5734
5735 PCREPOSIX(3) PCREPOSIX(3)
5736
5737
5738 NAME
5739 PCRE - Perl-compatible regular expressions.
5740
5741
5742 SYNOPSIS OF POSIX API
5743
5744 #include <pcreposix.h>
5745
5746 int regcomp(regex_t *preg, const char *pattern,
5747 int cflags);
5748
5749 int regexec(regex_t *preg, const char *string,
5750 size_t nmatch, regmatch_t pmatch[], int eflags);
5751
5752 size_t regerror(int errcode, const regex_t *preg,
5753 char *errbuf, size_t errbuf_size);
5754
5755 void regfree(regex_t *preg);
5756
5757
5758 DESCRIPTION
5759
5760 This set of functions provides a POSIX-style API to the PCRE regular
5761 expression package. See the pcreapi documentation for a description of
5762 PCRE's native API, which contains much additional functionality.
5763
5764 The functions described here are just wrapper functions that ultimately
5765 call the PCRE native API. Their prototypes are defined in the
5766 pcreposix.h header file, and on Unix systems the library itself is
5767 called pcreposix.a, so can be accessed by adding -lpcreposix to the
5768 command for linking an application that uses them. Because the POSIX
5769 functions call the native ones, it is also necessary to add -lpcre.
5770
5771 I have implemented only those option bits that can be reasonably mapped
5772 to PCRE native options. In addition, the option REG_EXTENDED is defined
5773 with the value zero. This has no effect, but since programs that are
5774 written to the POSIX interface often use it, this makes it easier to
5775 slot in PCRE as a replacement library. Other POSIX options are not even
5776 defined.
5777
5778 When PCRE is called via these functions, it is only the API that is
5779 POSIX-like in style. The syntax and semantics of the regular expres-
5780 sions themselves are still those of Perl, subject to the setting of
5781 various PCRE options, as described below. "POSIX-like in style" means
5782 that the API approximates to the POSIX definition; it is not fully
5783 POSIX-compatible, and in multi-byte encoding domains it is probably
5784 even less compatible.
5785
5786 The header for these functions is supplied as pcreposix.h to avoid any
5787 potential clash with other POSIX libraries. It can, of course, be
5788 renamed or aliased as regex.h, which is the "correct" name. It provides
5789 two structure types, regex_t for compiled internal forms, and reg-
5790 match_t for returning captured substrings. It also defines some con-
5791 stants whose names start with "REG_"; these are used for setting
5792 options and identifying error codes.
5793
5794
5795 COMPILING A PATTERN
5796
5797 The function regcomp() is called to compile a pattern into an internal
5798 form. The pattern is a C string terminated by a binary zero, and is
5799 passed in the argument pattern. The preg argument is a pointer to a
5800 regex_t structure that is used as a base for storing information about
5801 the compiled regular expression.
5802
5803 The argument cflags is either zero, or contains one or more of the bits
5804 defined by the following macros:
5805
5806 REG_DOTALL
5807
5808 The PCRE_DOTALL option is set when the regular expression is passed for
5809 compilation to the native function. Note that REG_DOTALL is not part of
5810 the POSIX standard.
5811
5812 REG_ICASE
5813
5814 The PCRE_CASELESS option is set when the regular expression is passed
5815 for compilation to the native function.
5816
5817 REG_NEWLINE
5818
5819 The PCRE_MULTILINE option is set when the regular expression is passed
5820 for compilation to the native function. Note that this does not mimic
5821 the defined POSIX behaviour for REG_NEWLINE (see the following sec-
5822 tion).
5823
5824 REG_NOSUB
5825
5826 The PCRE_NO_AUTO_CAPTURE option is set when the regular expression is
5827 passed for compilation to the native function. In addition, when a pat-
5828 tern that is compiled with this flag is passed to regexec() for match-
5829 ing, the nmatch and pmatch arguments are ignored, and no captured
5830 strings are returned.
5831
5832 REG_UTF8
5833
5834 The PCRE_UTF8 option is set when the regular expression is passed for
5835 compilation to the native function. This causes the pattern itself and
5836 all data strings used for matching it to be treated as UTF-8 strings.
5837 Note that REG_UTF8 is not part of the POSIX standard.
5838
5839 In the absence of these flags, no options are passed to the native
5840 function. This means the the regex is compiled with PCRE default
5841 semantics. In particular, the way it handles newline characters in the
5842 subject string is the Perl way, not the POSIX way. Note that setting
5843 PCRE_MULTILINE has only some of the effects specified for REG_NEWLINE.
5844 It does not affect the way newlines are matched by . (they aren't) or
5845 by a negative class such as [^a] (they are).
5846
5847 The yield of regcomp() is zero on success, and non-zero otherwise. The
5848 preg structure is filled in on success, and one member of the structure
5849 is public: re_nsub contains the number of capturing subpatterns in the
5850 regular expression. Various error codes are defined in the header file.
5851
5852
5853 MATCHING NEWLINE CHARACTERS
5854
5855 This area is not simple, because POSIX and Perl take different views of
5856 things. It is not possible to get PCRE to obey POSIX semantics, but
5857 then PCRE was never intended to be a POSIX engine. The following table
5858 lists the different possibilities for matching newline characters in
5859 PCRE:
5860
5861 Default Change with
5862
5863 . matches newline no PCRE_DOTALL
5864 newline matches [^a] yes not changeable
5865 $ matches \n at end yes PCRE_DOLLARENDONLY
5866 $ matches \n in middle no PCRE_MULTILINE
5867 ^ matches \n in middle no PCRE_MULTILINE
5868
5869 This is the equivalent table for POSIX:
5870
5871 Default Change with
5872
5873 . matches newline yes REG_NEWLINE
5874 newline matches [^a] yes REG_NEWLINE
5875 $ matches \n at end no REG_NEWLINE
5876 $ matches \n in middle no REG_NEWLINE
5877 ^ matches \n in middle no REG_NEWLINE
5878
5879 PCRE's behaviour is the same as Perl's, except that there is no equiva-
5880 lent for PCRE_DOLLAR_ENDONLY in Perl. In both PCRE and Perl, there is
5881 no way to stop newline from matching [^a].
5882
5883 The default POSIX newline handling can be obtained by setting
5884 PCRE_DOTALL and PCRE_DOLLAR_ENDONLY, but there is no way to make PCRE
5885 behave exactly as for the REG_NEWLINE action.
5886
5887
5888 MATCHING A PATTERN
5889
5890 The function regexec() is called to match a compiled pattern preg
5891 against a given string, which is terminated by a zero byte, subject to
5892 the options in eflags. These can be:
5893
5894 REG_NOTBOL
5895
5896 The PCRE_NOTBOL option is set when calling the underlying PCRE matching
5897 function.
5898
5899 REG_NOTEOL
5900
5901 The PCRE_NOTEOL option is set when calling the underlying PCRE matching
5902 function.
5903
5904 If the pattern was compiled with the REG_NOSUB flag, no data about any
5905 matched strings is returned. The nmatch and pmatch arguments of
5906 regexec() are ignored.
5907
5908 Otherwise,the portion of the string that was matched, and also any cap-
5909 tured substrings, are returned via the pmatch argument, which points to
5910 an array of nmatch structures of type regmatch_t, containing the mem-
5911 bers rm_so and rm_eo. These contain the offset to the first character
5912 of each substring and the offset to the first character after the end
5913 of each substring, respectively. The 0th element of the vector relates
5914 to the entire portion of string that was matched; subsequent elements
5915 relate to the capturing subpatterns of the regular expression. Unused
5916 entries in the array have both structure members set to -1.
5917
5918 A successful match yields a zero return; various error codes are
5919 defined in the header file, of which REG_NOMATCH is the "expected"
5920 failure code.
5921
5922
5923 ERROR MESSAGES
5924
5925 The regerror() function maps a non-zero errorcode from either regcomp()
5926 or regexec() to a printable message. If preg is not NULL, the error
5927 should have arisen from the use of that structure. A message terminated
5928 by a binary zero is placed in errbuf. The length of the message,
5929 including the zero, is limited to errbuf_size. The yield of the func-
5930 tion is the size of buffer needed to hold the whole message.
5931
5932
5933 MEMORY USAGE
5934
5935 Compiling a regular expression causes memory to be allocated and asso-
5936 ciated with the preg structure. The function regfree() frees all such
5937 memory, after which preg may no longer be used as a compiled expres-
5938 sion.
5939
5940
5941 AUTHOR
5942
5943 Philip Hazel
5944 University Computing Service
5945 Cambridge CB2 3QH, England.
5946
5947
5948 REVISION
5949
5950 Last updated: 06 March 2007
5951 Copyright (c) 1997-2007 University of Cambridge.
5952 ------------------------------------------------------------------------------
5953
5954
5955 PCRECPP(3) PCRECPP(3)
5956
5957
5958 NAME
5959 PCRE - Perl-compatible regular expressions.
5960
5961
5962 SYNOPSIS OF C++ WRAPPER
5963
5964 #include <pcrecpp.h>
5965
5966
5967 DESCRIPTION
5968
5969 The C++ wrapper for PCRE was provided by Google Inc. Some additional
5970 functionality was added by Giuseppe Maxia. This brief man page was con-
5971 structed from the notes in the pcrecpp.h file, which should be con-
5972 sulted for further details.
5973
5974
5975 MATCHING INTERFACE
5976
5977 The "FullMatch" operation checks that supplied text matches a supplied
5978 pattern exactly. If pointer arguments are supplied, it copies matched
5979 sub-strings that match sub-patterns into them.
5980
5981 Example: successful match
5982 pcrecpp::RE re("h.*o");
5983 re.FullMatch("hello");
5984
5985 Example: unsuccessful match (requires full match):
5986 pcrecpp::RE re("e");
5987 !re.FullMatch("hello");
5988
5989 Example: creating a temporary RE object:
5990 pcrecpp::RE("h.*o").FullMatch("hello");
5991
5992 You can pass in a "const char*" or a "string" for "text". The examples
5993 below tend to use a const char*. You can, as in the different examples
5994 above, store the RE object explicitly in a variable or use a temporary
5995 RE object. The examples below use one mode or the other arbitrarily.
5996 Either could correctly be used for any of these examples.
5997
5998 You must supply extra pointer arguments to extract matched subpieces.
5999
6000 Example: extracts "ruby" into "s" and 1234 into "i"
6001 int i;
6002 string s;
6003 pcrecpp::RE re("(\\w+):(\\d+)");
6004 re.FullMatch("ruby:1234", &s, &i);
6005
6006 Example: does not try to extract any extra sub-patterns
6007 re.FullMatch("ruby:1234", &s);
6008
6009 Example: does not try to extract into NULL
6010 re.FullMatch("ruby:1234", NULL, &i);
6011
6012 Example: integer overflow causes failure
6013 !re.FullMatch("ruby:1234567891234", NULL, &i);
6014
6015 Example: fails because there aren't enough sub-patterns:
6016 !pcrecpp::RE("\\w+:\\d+").FullMatch("ruby:1234", &s);
6017
6018 Example: fails because string cannot be stored in integer
6019 !pcrecpp::RE("(.*)").FullMatch("ruby", &i);
6020
6021 The provided pointer arguments can be pointers to any scalar numeric
6022 type, or one of:
6023
6024 string (matched piece is copied to string)
6025 StringPiece (StringPiece is mutated to point to matched piece)
6026 T (where "bool T::ParseFrom(const char*, int)" exists)
6027 NULL (the corresponding matched sub-pattern is not copied)
6028
6029 The function returns true iff all of the following conditions are sat-
6030 isfied:
6031
6032 a. "text" matches "pattern" exactly;
6033
6034 b. The number of matched sub-patterns is >= number of supplied
6035 pointers;
6036
6037 c. The "i"th argument has a suitable type for holding the
6038 string captured as the "i"th sub-pattern. If you pass in
6039 void * NULL for the "i"th argument, or a non-void * NULL
6040 of the correct type, or pass fewer arguments than the
6041 number of sub-patterns, "i"th captured sub-pattern is
6042 ignored.
6043
6044 CAVEAT: An optional sub-pattern that does not exist in the matched
6045 string is assigned the empty string. Therefore, the following will
6046 return false (because the empty string is not a valid number):
6047
6048 int number;
6049 pcrecpp::RE::FullMatch("abc", "[a-z]+(\\d+)?", &number);
6050
6051 The matching interface supports at most 16 arguments per call. If you
6052 need more, consider using the more general interface
6053 pcrecpp::RE::DoMatch. See pcrecpp.h for the signature for DoMatch.
6054
6055
6056 QUOTING METACHARACTERS
6057
6058 You can use the "QuoteMeta" operation to insert backslashes before all
6059 potentially meaningful characters in a string. The returned string,
6060 used as a regular expression, will exactly match the original string.
6061
6062 Example:
6063 string quoted = RE::QuoteMeta(unquoted);
6064
6065 Note that it's legal to escape a character even if it has no special
6066 meaning in a regular expression -- so this function does that. (This
6067 also makes it identical to the perl function of the same name; see
6068 "perldoc -f quotemeta".) For example, "1.5-2.0?" becomes
6069 "1\.5\-2\.0\?".
6070
6071
6072 PARTIAL MATCHES
6073
6074 You can use the "PartialMatch" operation when you want the pattern to
6075 match any substring of the text.
6076
6077 Example: simple search for a string:
6078 pcrecpp::RE("ell").PartialMatch("hello");
6079
6080 Example: find first number in a string:
6081 int number;
6082 pcrecpp::RE re("(\\d+)");
6083 re.PartialMatch("x*100 + 20", &number);
6084 assert(number == 100);
6085
6086
6087 UTF-8 AND THE MATCHING INTERFACE
6088
6089 By default, pattern and text are plain text, one byte per character.
6090 The UTF8 flag, passed to the constructor, causes both pattern and
6091 string to be treated as UTF-8 text, still a byte stream but potentially
6092 multiple bytes per character. In practice, the text is likelier to be
6093 UTF-8 than the pattern, but the match returned may depend on the UTF8
6094 flag, so always use it when matching UTF8 text. For example, "." will
6095 match one byte normally but with UTF8 set may match up to three bytes
6096 of a multi-byte character.
6097
6098 Example:
6099 pcrecpp::RE_Options options;
6100 options.set_utf8();
6101 pcrecpp::RE re(utf8_pattern, options);
6102 re.FullMatch(utf8_string);
6103
6104 Example: using the convenience function UTF8():
6105 pcrecpp::RE re(utf8_pattern, pcrecpp::UTF8());
6106 re.FullMatch(utf8_string);
6107
6108 NOTE: The UTF8 flag is ignored if pcre was not configured with the
6109 --enable-utf8 flag.
6110
6111
6112