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

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

Parent Directory Parent Directory | Revision Log Revision Log


Revision 567 - (show annotations) (download)
Sat Nov 6 17:10:00 2010 UTC (3 years, 9 months ago) by ph10
File MIME type: text/plain
File size: 363403 byte(s)
Test for ridiculous values of starting offsets; tidy UTF-8 code.

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. Neither has
6 the pcredemo program. There are separate text files for the pcregrep and
7 pcretest commands.
8 -----------------------------------------------------------------------------
9
10
11 PCRE(3) PCRE(3)
12
13
14 NAME
15 PCRE - Perl-compatible regular expressions
16
17
18 INTRODUCTION
19
20 The PCRE library is a set of functions that implement regular expres-
21 sion pattern matching using the same syntax and semantics as Perl, with
22 just a few differences. Some features that appeared in Python and PCRE
23 before they appeared in Perl are also available using the Python syn-
24 tax, there is some support for one or two .NET and Oniguruma syntax
25 items, and there is an option for requesting some minor changes that
26 give better JavaScript compatibility.
27
28 The current implementation of PCRE corresponds approximately with Perl
29 5.10/5.11, including support for UTF-8 encoded strings and Unicode gen-
30 eral category properties. However, UTF-8 and Unicode support has to be
31 explicitly enabled; it is not the default. The Unicode tables corre-
32 spond to Unicode release 5.2.0.
33
34 In addition to the Perl-compatible matching function, PCRE contains an
35 alternative function that matches the same compiled patterns in a dif-
36 ferent way. In certain circumstances, the alternative function has some
37 advantages. For a discussion of the two matching algorithms, see the
38 pcrematching page.
39
40 PCRE is written in C and released as a C library. A number of people
41 have written wrappers and interfaces of various kinds. In particular,
42 Google Inc. have provided a comprehensive C++ wrapper. This is now
43 included as part of the PCRE distribution. The pcrecpp page has details
44 of this interface. Other people's contributions can be found in the
45 Contrib directory at the primary FTP site, which is:
46
47 ftp://ftp.csx.cam.ac.uk/pub/software/programming/pcre
48
49 Details of exactly which Perl regular expression features are and are
50 not supported by PCRE are given in separate documents. See the pcrepat-
51 tern and pcrecompat pages. There is a syntax summary in the pcresyntax
52 page.
53
54 Some features of PCRE can be included, excluded, or changed when the
55 library is built. The pcre_config() function makes it possible for a
56 client to discover which features are available. The features them-
57 selves are described in the pcrebuild page. Documentation about build-
58 ing PCRE for various operating systems can be found in the README and
59 NON-UNIX-USE files in the source distribution.
60
61 The library contains a number of undocumented internal functions and
62 data tables that are used by more than one of the exported external
63 functions, but which are not intended for use by external callers.
64 Their names all begin with "_pcre_", which hopefully will not provoke
65 any name clashes. In some environments, it is possible to control which
66 external symbols are exported when a shared library is built, and in
67 these cases the undocumented symbols are not exported.
68
69
70 USER DOCUMENTATION
71
72 The user documentation for PCRE comprises a number of different sec-
73 tions. In the "man" format, each of these is a separate "man page". In
74 the HTML format, each is a separate page, linked from the index page.
75 In the plain text format, all the sections, except the pcredemo sec-
76 tion, are concatenated, for ease of searching. The sections are as fol-
77 lows:
78
79 pcre this document
80 pcre-config show PCRE installation configuration information
81 pcreapi details of PCRE's native C API
82 pcrebuild options for building PCRE
83 pcrecallout details of the callout feature
84 pcrecompat discussion of Perl compatibility
85 pcrecpp details of the C++ wrapper
86 pcredemo a demonstration C program that uses PCRE
87 pcregrep description of the pcregrep command
88 pcrematching discussion of the two matching algorithms
89 pcrepartial details of the partial matching facility
90 pcrepattern syntax and semantics of supported
91 regular expressions
92 pcreperform discussion of performance issues
93 pcreposix the POSIX-compatible C API
94 pcreprecompile details of saving and re-using precompiled patterns
95 pcresample discussion of the pcredemo program
96 pcrestack discussion of stack usage
97 pcresyntax quick syntax reference
98 pcretest description of the pcretest testing command
99
100 In addition, in the "man" and HTML formats, there is a short page for
101 each C library function, listing its arguments and results.
102
103
104 LIMITATIONS
105
106 There are some size limitations in PCRE but it is hoped that they will
107 never in practice be relevant.
108
109 The maximum length of a compiled pattern is 65539 (sic) bytes if PCRE
110 is compiled with the default internal linkage size of 2. If you want to
111 process regular expressions that are truly enormous, you can compile
112 PCRE with an internal linkage size of 3 or 4 (see the README file in
113 the source distribution and the pcrebuild documentation for details).
114 In these cases the limit is substantially larger. However, the speed
115 of execution is slower.
116
117 All values in repeating quantifiers must be less than 65536.
118
119 There is no limit to the number of parenthesized subpatterns, but there
120 can be no more than 65535 capturing subpatterns.
121
122 The maximum length of name for a named subpattern is 32 characters, and
123 the maximum number of named subpatterns is 10000.
124
125 The maximum length of a subject string is the largest positive number
126 that an integer variable can hold. However, when using the traditional
127 matching function, PCRE uses recursion to handle subpatterns and indef-
128 inite repetition. This means that the available stack space may limit
129 the size of a subject string that can be processed by certain patterns.
130 For a discussion of stack issues, see the pcrestack documentation.
131
132
133 UTF-8 AND UNICODE PROPERTY SUPPORT
134
135 From release 3.3, PCRE has had some support for character strings
136 encoded in the UTF-8 format. For release 4.0 this was greatly extended
137 to cover most common requirements, and in release 5.0 additional sup-
138 port for Unicode general category properties was added.
139
140 In order process UTF-8 strings, you must build PCRE to include UTF-8
141 support in the code, and, in addition, you must call pcre_compile()
142 with the PCRE_UTF8 option flag, or the pattern must start with the
143 sequence (*UTF8). When either of these is the case, both the pattern
144 and any subject strings that are matched against it are treated as
145 UTF-8 strings instead of strings of 1-byte characters.
146
147 If you compile PCRE with UTF-8 support, but do not use it at run time,
148 the library will be a bit bigger, but the additional run time overhead
149 is limited to testing the PCRE_UTF8 flag occasionally, so should not be
150 very big.
151
152 If PCRE is built with Unicode character property support (which implies
153 UTF-8 support), the escape sequences \p{..}, \P{..}, and \X are sup-
154 ported. The available properties that can be tested are limited to the
155 general category properties such as Lu for an upper case letter or Nd
156 for a decimal number, the Unicode script names such as Arabic or Han,
157 and the derived properties Any and L&. A full list is given in the
158 pcrepattern documentation. Only the short names for properties are sup-
159 ported. For example, \p{L} matches a letter. Its Perl synonym, \p{Let-
160 ter}, is not supported. Furthermore, in Perl, many properties may
161 optionally be prefixed by "Is", for compatibility with Perl 5.6. PCRE
162 does not support this.
163
164 Validity of UTF-8 strings
165
166 When you set the PCRE_UTF8 flag, the strings passed as patterns and
167 subjects are (by default) checked for validity on entry to the relevant
168 functions. From release 7.3 of PCRE, the check is according the rules
169 of RFC 3629, which are themselves derived from the Unicode specifica-
170 tion. Earlier releases of PCRE followed the rules of RFC 2279, which
171 allows the full range of 31-bit values (0 to 0x7FFFFFFF). The current
172 check allows only values in the range U+0 to U+10FFFF, excluding U+D800
173 to U+DFFF.
174
175 The excluded code points are the "Low Surrogate Area" of Unicode, of
176 which the Unicode Standard says this: "The Low Surrogate Area does not
177 contain any character assignments, consequently no character code
178 charts or namelists are provided for this area. Surrogates are reserved
179 for use with UTF-16 and then must be used in pairs." The code points
180 that are encoded by UTF-16 pairs are available as independent code
181 points in the UTF-8 encoding. (In other words, the whole surrogate
182 thing is a fudge for UTF-16 which unfortunately messes up UTF-8.)
183
184 If an invalid UTF-8 string is passed to PCRE, an error return
185 (PCRE_ERROR_BADUTF8) is given. In some situations, you may already know
186 that your strings are valid, and therefore want to skip these checks in
187 order to improve performance. If you set the PCRE_NO_UTF8_CHECK flag at
188 compile time or at run time, PCRE assumes that the pattern or subject
189 it is given (respectively) contains only valid UTF-8 codes. In this
190 case, it does not diagnose an invalid UTF-8 string.
191
192 If you pass an invalid UTF-8 string when PCRE_NO_UTF8_CHECK is set,
193 what happens depends on why the string is invalid. If the string con-
194 forms to the "old" definition of UTF-8 (RFC 2279), it is processed as a
195 string of characters in the range 0 to 0x7FFFFFFF. In other words,
196 apart from the initial validity test, PCRE (when in UTF-8 mode) handles
197 strings according to the more liberal rules of RFC 2279. However, if
198 the string does not even conform to RFC 2279, the result is undefined.
199 Your program may crash.
200
201 If you want to process strings of values in the full range 0 to
202 0x7FFFFFFF, encoded in a UTF-8-like manner as per the old RFC, you can
203 set PCRE_NO_UTF8_CHECK to bypass the more restrictive test. However, in
204 this situation, you will have to apply your own validity check.
205
206 General comments about UTF-8 mode
207
208 1. An unbraced hexadecimal escape sequence (such as \xb3) matches a
209 two-byte UTF-8 character if the value is greater than 127.
210
211 2. Octal numbers up to \777 are recognized, and match two-byte UTF-8
212 characters for values greater than \177.
213
214 3. Repeat quantifiers apply to complete UTF-8 characters, not to indi-
215 vidual bytes, for example: \x{100}{3}.
216
217 4. The dot metacharacter matches one UTF-8 character instead of a sin-
218 gle byte.
219
220 5. The escape sequence \C can be used to match a single byte in UTF-8
221 mode, but its use can lead to some strange effects. This facility is
222 not available in the alternative matching function, pcre_dfa_exec().
223
224 6. The character escapes \b, \B, \d, \D, \s, \S, \w, and \W correctly
225 test characters of any code value, but, by default, the characters that
226 PCRE recognizes as digits, spaces, or word characters remain the same
227 set as before, all with values less than 256. This remains true even
228 when PCRE is built to include Unicode property support, because to do
229 otherwise would slow down PCRE in many common cases. Note in particular
230 that this applies to \b and \B, because they are defined in terms of \w
231 and \W. If you really want to test for a wider sense of, say, "digit",
232 you can use explicit Unicode property tests such as \p{Nd}. Alterna-
233 tively, if you set the PCRE_UCP option, the way that the character
234 escapes work is changed so that Unicode properties are used to deter-
235 mine which characters match. There are more details in the section on
236 generic character types in the pcrepattern documentation.
237
238 7. Similarly, characters that match the POSIX named character classes
239 are all low-valued characters, unless the PCRE_UCP option is set.
240
241 8. However, the Perl 5.10 horizontal and vertical whitespace matching
242 escapes (\h, \H, \v, and \V) do match all the appropriate Unicode char-
243 acters, whether or not PCRE_UCP is set.
244
245 9. Case-insensitive matching applies only to characters whose values
246 are less than 128, unless PCRE is built with Unicode property support.
247 Even when Unicode property support is available, PCRE still uses its
248 own character tables when checking the case of low-valued characters,
249 so as not to degrade performance. The Unicode property information is
250 used only for characters with higher values. Even when Unicode property
251 support is available, PCRE supports case-insensitive matching only when
252 there is a one-to-one mapping between a letter's cases. There are a
253 small number of many-to-one mappings in Unicode; these are not sup-
254 ported by PCRE.
255
256
257 AUTHOR
258
259 Philip Hazel
260 University Computing Service
261 Cambridge CB2 3QH, England.
262
263 Putting an actual email address here seems to have been a spam magnet,
264 so I've taken it away. If you want to email me, use my two initials,
265 followed by the two digits 10, at the domain cam.ac.uk.
266
267
268 REVISION
269
270 Last updated: 22 October 2010
271 Copyright (c) 1997-2010 University of Cambridge.
272 ------------------------------------------------------------------------------
273
274
275 PCREBUILD(3) PCREBUILD(3)
276
277
278 NAME
279 PCRE - Perl-compatible regular expressions
280
281
282 PCRE BUILD-TIME OPTIONS
283
284 This document describes the optional features of PCRE that can be
285 selected when the library is compiled. It assumes use of the configure
286 script, where the optional features are selected or deselected by pro-
287 viding options to configure before running the make command. However,
288 the same options can be selected in both Unix-like and non-Unix-like
289 environments using the GUI facility of cmake-gui if you are using CMake
290 instead of configure to build PCRE.
291
292 There is a lot more information about building PCRE in non-Unix-like
293 environments in the file called NON_UNIX_USE, which is part of the PCRE
294 distribution. You should consult this file as well as the README file
295 if you are building in a non-Unix-like environment.
296
297 The complete list of options for configure (which includes the standard
298 ones such as the selection of the installation directory) can be
299 obtained by running
300
301 ./configure --help
302
303 The following sections include descriptions of options whose names
304 begin with --enable or --disable. These settings specify changes to the
305 defaults for the configure command. Because of the way that configure
306 works, --enable and --disable always come in pairs, so the complemen-
307 tary option always exists as well, but as it specifies the default, it
308 is not described.
309
310
311 C++ SUPPORT
312
313 By default, the configure script will search for a C++ compiler and C++
314 header files. If it finds them, it automatically builds the C++ wrapper
315 library for PCRE. You can disable this by adding
316
317 --disable-cpp
318
319 to the configure command.
320
321
322 UTF-8 SUPPORT
323
324 To build PCRE with support for UTF-8 Unicode character strings, add
325
326 --enable-utf8
327
328 to the configure command. Of itself, this does not make PCRE treat
329 strings as UTF-8. As well as compiling PCRE with this option, you also
330 have have to set the PCRE_UTF8 option when you call the pcre_compile()
331 or pcre_compile2() functions.
332
333 If you set --enable-utf8 when compiling in an EBCDIC environment, PCRE
334 expects its input to be either ASCII or UTF-8 (depending on the runtime
335 option). It is not possible to support both EBCDIC and UTF-8 codes in
336 the same version of the library. Consequently, --enable-utf8 and
337 --enable-ebcdic are mutually exclusive.
338
339
340 UNICODE CHARACTER PROPERTY SUPPORT
341
342 UTF-8 support allows PCRE to process character values greater than 255
343 in the strings that it handles. On its own, however, it does not pro-
344 vide any facilities for accessing the properties of such characters. If
345 you want to be able to use the pattern escapes \P, \p, and \X, which
346 refer to Unicode character properties, you must add
347
348 --enable-unicode-properties
349
350 to the configure command. This implies UTF-8 support, even if you have
351 not explicitly requested it.
352
353 Including Unicode property support adds around 30K of tables to the
354 PCRE library. Only the general category properties such as Lu and Nd
355 are supported. Details are given in the pcrepattern documentation.
356
357
358 CODE VALUE OF NEWLINE
359
360 By default, PCRE interprets the linefeed (LF) character as indicating
361 the end of a line. This is the normal newline character on Unix-like
362 systems. You can compile PCRE to use carriage return (CR) instead, by
363 adding
364
365 --enable-newline-is-cr
366
367 to the configure command. There is also a --enable-newline-is-lf
368 option, which explicitly specifies linefeed as the newline character.
369
370 Alternatively, you can specify that line endings are to be indicated by
371 the two character sequence CRLF. If you want this, add
372
373 --enable-newline-is-crlf
374
375 to the configure command. There is a fourth option, specified by
376
377 --enable-newline-is-anycrlf
378
379 which causes PCRE to recognize any of the three sequences CR, LF, or
380 CRLF as indicating a line ending. Finally, a fifth option, specified by
381
382 --enable-newline-is-any
383
384 causes PCRE to recognize any Unicode newline sequence.
385
386 Whatever line ending convention is selected when PCRE is built can be
387 overridden when the library functions are called. At build time it is
388 conventional to use the standard for your operating system.
389
390
391 WHAT \R MATCHES
392
393 By default, the sequence \R in a pattern matches any Unicode newline
394 sequence, whatever has been selected as the line ending sequence. If
395 you specify
396
397 --enable-bsr-anycrlf
398
399 the default is changed so that \R matches only CR, LF, or CRLF. What-
400 ever is selected when PCRE is built can be overridden when the library
401 functions are called.
402
403
404 BUILDING SHARED AND STATIC LIBRARIES
405
406 The PCRE building process uses libtool to build both shared and static
407 Unix libraries by default. You can suppress one of these by adding one
408 of
409
410 --disable-shared
411 --disable-static
412
413 to the configure command, as required.
414
415
416 POSIX MALLOC USAGE
417
418 When PCRE is called through the POSIX interface (see the pcreposix doc-
419 umentation), additional working storage is required for holding the
420 pointers to capturing substrings, because PCRE requires three integers
421 per substring, whereas the POSIX interface provides only two. If the
422 number of expected substrings is small, the wrapper function uses space
423 on the stack, because this is faster than using malloc() for each call.
424 The default threshold above which the stack is no longer used is 10; it
425 can be changed by adding a setting such as
426
427 --with-posix-malloc-threshold=20
428
429 to the configure command.
430
431
432 HANDLING VERY LARGE PATTERNS
433
434 Within a compiled pattern, offset values are used to point from one
435 part to another (for example, from an opening parenthesis to an alter-
436 nation metacharacter). By default, two-byte values are used for these
437 offsets, leading to a maximum size for a compiled pattern of around
438 64K. This is sufficient to handle all but the most gigantic patterns.
439 Nevertheless, some people do want to process truyl enormous patterns,
440 so it is possible to compile PCRE to use three-byte or four-byte off-
441 sets by adding a setting such as
442
443 --with-link-size=3
444
445 to the configure command. The value given must be 2, 3, or 4. Using
446 longer offsets slows down the operation of PCRE because it has to load
447 additional bytes when handling them.
448
449
450 AVOIDING EXCESSIVE STACK USAGE
451
452 When matching with the pcre_exec() function, PCRE implements backtrack-
453 ing by making recursive calls to an internal function called match().
454 In environments where the size of the stack is limited, this can se-
455 verely limit PCRE's operation. (The Unix environment does not usually
456 suffer from this problem, but it may sometimes be necessary to increase
457 the maximum stack size. There is a discussion in the pcrestack docu-
458 mentation.) An alternative approach to recursion that uses memory from
459 the heap to remember data, instead of using recursive function calls,
460 has been implemented to work round the problem of limited stack size.
461 If you want to build a version of PCRE that works this way, add
462
463 --disable-stack-for-recursion
464
465 to the configure command. With this configuration, PCRE will use the
466 pcre_stack_malloc and pcre_stack_free variables to call memory manage-
467 ment functions. By default these point to malloc() and free(), but you
468 can replace the pointers so that your own functions are used instead.
469
470 Separate functions are provided rather than using pcre_malloc and
471 pcre_free because the usage is very predictable: the block sizes
472 requested are always the same, and the blocks are always freed in
473 reverse order. A calling program might be able to implement optimized
474 functions that perform better than malloc() and free(). PCRE runs
475 noticeably more slowly when built in this way. This option affects only
476 the pcre_exec() function; it is not relevant for pcre_dfa_exec().
477
478
479 LIMITING PCRE RESOURCE USAGE
480
481 Internally, PCRE has a function called match(), which it calls repeat-
482 edly (sometimes recursively) when matching a pattern with the
483 pcre_exec() function. By controlling the maximum number of times this
484 function may be called during a single matching operation, a limit can
485 be placed on the resources used by a single call to pcre_exec(). The
486 limit can be changed at run time, as described in the pcreapi documen-
487 tation. The default is 10 million, but this can be changed by adding a
488 setting such as
489
490 --with-match-limit=500000
491
492 to the configure command. This setting has no effect on the
493 pcre_dfa_exec() matching function.
494
495 In some environments it is desirable to limit the depth of recursive
496 calls of match() more strictly than the total number of calls, in order
497 to restrict the maximum amount of stack (or heap, if --disable-stack-
498 for-recursion is specified) that is used. A second limit controls this;
499 it defaults to the value that is set for --with-match-limit, which
500 imposes no additional constraints. However, you can set a lower limit
501 by adding, for example,
502
503 --with-match-limit-recursion=10000
504
505 to the configure command. This value can also be overridden at run
506 time.
507
508
509 CREATING CHARACTER TABLES AT BUILD TIME
510
511 PCRE uses fixed tables for processing characters whose code values are
512 less than 256. By default, PCRE is built with a set of tables that are
513 distributed in the file pcre_chartables.c.dist. These tables are for
514 ASCII codes only. If you add
515
516 --enable-rebuild-chartables
517
518 to the configure command, the distributed tables are no longer used.
519 Instead, a program called dftables is compiled and run. This outputs
520 the source for new set of tables, created in the default locale of your
521 C runtime system. (This method of replacing the tables does not work if
522 you are cross compiling, because dftables is run on the local host. If
523 you need to create alternative tables when cross compiling, you will
524 have to do so "by hand".)
525
526
527 USING EBCDIC CODE
528
529 PCRE assumes by default that it will run in an environment where the
530 character code is ASCII (or Unicode, which is a superset of ASCII).
531 This is the case for most computer operating systems. PCRE can, how-
532 ever, be compiled to run in an EBCDIC environment by adding
533
534 --enable-ebcdic
535
536 to the configure command. This setting implies --enable-rebuild-charta-
537 bles. You should only use it if you know that you are in an EBCDIC
538 environment (for example, an IBM mainframe operating system). The
539 --enable-ebcdic option is incompatible with --enable-utf8.
540
541
542 PCREGREP OPTIONS FOR COMPRESSED FILE SUPPORT
543
544 By default, pcregrep reads all files as plain text. You can build it so
545 that it recognizes files whose names end in .gz or .bz2, and reads them
546 with libz or libbz2, respectively, by adding one or both of
547
548 --enable-pcregrep-libz
549 --enable-pcregrep-libbz2
550
551 to the configure command. These options naturally require that the rel-
552 evant libraries are installed on your system. Configuration will fail
553 if they are not.
554
555
556 PCRETEST OPTION FOR LIBREADLINE SUPPORT
557
558 If you add
559
560 --enable-pcretest-libreadline
561
562 to the configure command, pcretest is linked with the libreadline
563 library, and when its input is from a terminal, it reads it using the
564 readline() function. This provides line-editing and history facilities.
565 Note that libreadline is GPL-licensed, so if you distribute a binary of
566 pcretest linked in this way, there may be licensing issues.
567
568 Setting this option causes the -lreadline option to be added to the
569 pcretest build. In many operating environments with a sytem-installed
570 libreadline this is sufficient. However, in some environments (e.g. if
571 an unmodified distribution version of readline is in use), some extra
572 configuration may be necessary. The INSTALL file for libreadline says
573 this:
574
575 "Readline uses the termcap functions, but does not link with the
576 termcap or curses library itself, allowing applications which link
577 with readline the to choose an appropriate library."
578
579 If your environment has not been set up so that an appropriate library
580 is automatically included, you may need to add something like
581
582 LIBS="-ncurses"
583
584 immediately before the configure command.
585
586
587 SEE ALSO
588
589 pcreapi(3), pcre_config(3).
590
591
592 AUTHOR
593
594 Philip Hazel
595 University Computing Service
596 Cambridge CB2 3QH, England.
597
598
599 REVISION
600
601 Last updated: 29 September 2009
602 Copyright (c) 1997-2009 University of Cambridge.
603 ------------------------------------------------------------------------------
604
605
606 PCREMATCHING(3) PCREMATCHING(3)
607
608
609 NAME
610 PCRE - Perl-compatible regular expressions
611
612
613 PCRE MATCHING ALGORITHMS
614
615 This document describes the two different algorithms that are available
616 in PCRE for matching a compiled regular expression against a given sub-
617 ject string. The "standard" algorithm is the one provided by the
618 pcre_exec() function. This works in the same was as Perl's matching
619 function, and provides a Perl-compatible matching operation.
620
621 An alternative algorithm is provided by the pcre_dfa_exec() function;
622 this operates in a different way, and is not Perl-compatible. It has
623 advantages and disadvantages compared with the standard algorithm, and
624 these are described below.
625
626 When there is only one possible way in which a given subject string can
627 match a pattern, the two algorithms give the same answer. A difference
628 arises, however, when there are multiple possibilities. For example, if
629 the pattern
630
631 ^<.*>
632
633 is matched against the string
634
635 <something> <something else> <something further>
636
637 there are three possible answers. The standard algorithm finds only one
638 of them, whereas the alternative algorithm finds all three.
639
640
641 REGULAR EXPRESSIONS AS TREES
642
643 The set of strings that are matched by a regular expression can be rep-
644 resented as a tree structure. An unlimited repetition in the pattern
645 makes the tree of infinite size, but it is still a tree. Matching the
646 pattern to a given subject string (from a given starting point) can be
647 thought of as a search of the tree. There are two ways to search a
648 tree: depth-first and breadth-first, and these correspond to the two
649 matching algorithms provided by PCRE.
650
651
652 THE STANDARD MATCHING ALGORITHM
653
654 In the terminology of Jeffrey Friedl's book "Mastering Regular Expres-
655 sions", the standard algorithm is an "NFA algorithm". It conducts a
656 depth-first search of the pattern tree. That is, it proceeds along a
657 single path through the tree, checking that the subject matches what is
658 required. When there is a mismatch, the algorithm tries any alterna-
659 tives at the current point, and if they all fail, it backs up to the
660 previous branch point in the tree, and tries the next alternative
661 branch at that level. This often involves backing up (moving to the
662 left) in the subject string as well. The order in which repetition
663 branches are tried is controlled by the greedy or ungreedy nature of
664 the quantifier.
665
666 If a leaf node is reached, a matching string has been found, and at
667 that point the algorithm stops. Thus, if there is more than one possi-
668 ble match, this algorithm returns the first one that it finds. Whether
669 this is the shortest, the longest, or some intermediate length depends
670 on the way the greedy and ungreedy repetition quantifiers are specified
671 in the pattern.
672
673 Because it ends up with a single path through the tree, it is rela-
674 tively straightforward for this algorithm to keep track of the sub-
675 strings that are matched by portions of the pattern in parentheses.
676 This provides support for capturing parentheses and back references.
677
678
679 THE ALTERNATIVE MATCHING ALGORITHM
680
681 This algorithm conducts a breadth-first search of the tree. Starting
682 from the first matching point in the subject, it scans the subject
683 string from left to right, once, character by character, and as it does
684 this, it remembers all the paths through the tree that represent valid
685 matches. In Friedl's terminology, this is a kind of "DFA algorithm",
686 though it is not implemented as a traditional finite state machine (it
687 keeps multiple states active simultaneously).
688
689 Although the general principle of this matching algorithm is that it
690 scans the subject string only once, without backtracking, there is one
691 exception: when a lookaround assertion is encountered, the characters
692 following or preceding the current point have to be independently
693 inspected.
694
695 The scan continues until either the end of the subject is reached, or
696 there are no more unterminated paths. At this point, terminated paths
697 represent the different matching possibilities (if there are none, the
698 match has failed). Thus, if there is more than one possible match,
699 this algorithm finds all of them, and in particular, it finds the long-
700 est. There is an option to stop the algorithm after the first match
701 (which is necessarily the shortest) is found.
702
703 Note that all the matches that are found start at the same point in the
704 subject. If the pattern
705
706 cat(er(pillar)?)
707
708 is matched against the string "the caterpillar catchment", the result
709 will be the three strings "cat", "cater", and "caterpillar" that start
710 at the fourth character of the subject. The algorithm does not automat-
711 ically move on to find matches that start at later positions.
712
713 There are a number of features of PCRE regular expressions that are not
714 supported by the alternative matching algorithm. They are as follows:
715
716 1. Because the algorithm finds all possible matches, the greedy or
717 ungreedy nature of repetition quantifiers is not relevant. Greedy and
718 ungreedy quantifiers are treated in exactly the same way. However, pos-
719 sessive quantifiers can make a difference when what follows could also
720 match what is quantified, for example in a pattern like this:
721
722 ^a++\w!
723
724 This pattern matches "aaab!" but not "aaa!", which would be matched by
725 a non-possessive quantifier. Similarly, if an atomic group is present,
726 it is matched as if it were a standalone pattern at the current point,
727 and the longest match is then "locked in" for the rest of the overall
728 pattern.
729
730 2. When dealing with multiple paths through the tree simultaneously, it
731 is not straightforward to keep track of captured substrings for the
732 different matching possibilities, and PCRE's implementation of this
733 algorithm does not attempt to do this. This means that no captured sub-
734 strings are available.
735
736 3. Because no substrings are captured, back references within the pat-
737 tern are not supported, and cause errors if encountered.
738
739 4. For the same reason, conditional expressions that use a backrefer-
740 ence as the condition or test for a specific group recursion are not
741 supported.
742
743 5. Because many paths through the tree may be active, the \K escape
744 sequence, which resets the start of the match when encountered (but may
745 be on some paths and not on others), is not supported. It causes an
746 error if encountered.
747
748 6. Callouts are supported, but the value of the capture_top field is
749 always 1, and the value of the capture_last field is always -1.
750
751 7. The \C escape sequence, which (in the standard algorithm) matches a
752 single byte, even in UTF-8 mode, is not supported because the alterna-
753 tive algorithm moves through the subject string one character at a
754 time, for all active paths through the tree.
755
756 8. Except for (*FAIL), the backtracking control verbs such as (*PRUNE)
757 are not supported. (*FAIL) is supported, and behaves like a failing
758 negative assertion.
759
760
761 ADVANTAGES OF THE ALTERNATIVE ALGORITHM
762
763 Using the alternative matching algorithm provides the following advan-
764 tages:
765
766 1. All possible matches (at a single point in the subject) are automat-
767 ically found, and in particular, the longest match is found. To find
768 more than one match using the standard algorithm, you have to do kludgy
769 things with callouts.
770
771 2. Because the alternative algorithm scans the subject string just
772 once, and never needs to backtrack, it is possible to pass very long
773 subject strings to the matching function in several pieces, checking
774 for partial matching each time. It is possible to do multi-segment
775 matching using pcre_exec() (by retaining partially matched substrings),
776 but it is more complicated. The pcrepartial documentation gives details
777 of partial matching and discusses multi-segment matching.
778
779
780 DISADVANTAGES OF THE ALTERNATIVE ALGORITHM
781
782 The alternative algorithm suffers from a number of disadvantages:
783
784 1. It is substantially slower than the standard algorithm. This is
785 partly because it has to search for all possible matches, but is also
786 because it is less susceptible to optimization.
787
788 2. Capturing parentheses and back references are not supported.
789
790 3. Although atomic groups are supported, their use does not provide the
791 performance advantage that it does for the standard algorithm.
792
793
794 AUTHOR
795
796 Philip Hazel
797 University Computing Service
798 Cambridge CB2 3QH, England.
799
800
801 REVISION
802
803 Last updated: 22 October 2010
804 Copyright (c) 1997-2010 University of Cambridge.
805 ------------------------------------------------------------------------------
806
807
808 PCREAPI(3) PCREAPI(3)
809
810
811 NAME
812 PCRE - Perl-compatible regular expressions
813
814
815 PCRE NATIVE API
816
817 #include <pcre.h>
818
819 pcre *pcre_compile(const char *pattern, int options,
820 const char **errptr, int *erroffset,
821 const unsigned char *tableptr);
822
823 pcre *pcre_compile2(const char *pattern, int options,
824 int *errorcodeptr,
825 const char **errptr, int *erroffset,
826 const unsigned char *tableptr);
827
828 pcre_extra *pcre_study(const pcre *code, int options,
829 const char **errptr);
830
831 int pcre_exec(const pcre *code, const pcre_extra *extra,
832 const char *subject, int length, int startoffset,
833 int options, int *ovector, int ovecsize);
834
835 int pcre_dfa_exec(const pcre *code, const pcre_extra *extra,
836 const char *subject, int length, int startoffset,
837 int options, int *ovector, int ovecsize,
838 int *workspace, int wscount);
839
840 int pcre_copy_named_substring(const pcre *code,
841 const char *subject, int *ovector,
842 int stringcount, const char *stringname,
843 char *buffer, int buffersize);
844
845 int pcre_copy_substring(const char *subject, int *ovector,
846 int stringcount, int stringnumber, char *buffer,
847 int buffersize);
848
849 int pcre_get_named_substring(const pcre *code,
850 const char *subject, int *ovector,
851 int stringcount, const char *stringname,
852 const char **stringptr);
853
854 int pcre_get_stringnumber(const pcre *code,
855 const char *name);
856
857 int pcre_get_stringtable_entries(const pcre *code,
858 const char *name, char **first, char **last);
859
860 int pcre_get_substring(const char *subject, int *ovector,
861 int stringcount, int stringnumber,
862 const char **stringptr);
863
864 int pcre_get_substring_list(const char *subject,
865 int *ovector, int stringcount, const char ***listptr);
866
867 void pcre_free_substring(const char *stringptr);
868
869 void pcre_free_substring_list(const char **stringptr);
870
871 const unsigned char *pcre_maketables(void);
872
873 int pcre_fullinfo(const pcre *code, const pcre_extra *extra,
874 int what, void *where);
875
876 int pcre_info(const pcre *code, int *optptr, int *firstcharptr);
877
878 int pcre_refcount(pcre *code, int adjust);
879
880 int pcre_config(int what, void *where);
881
882 char *pcre_version(void);
883
884 void *(*pcre_malloc)(size_t);
885
886 void (*pcre_free)(void *);
887
888 void *(*pcre_stack_malloc)(size_t);
889
890 void (*pcre_stack_free)(void *);
891
892 int (*pcre_callout)(pcre_callout_block *);
893
894
895 PCRE API OVERVIEW
896
897 PCRE has its own native API, which is described in this document. There
898 are also some wrapper functions that correspond to the POSIX regular
899 expression API. These are described in the pcreposix documentation.
900 Both of these APIs define a set of C function calls. A C++ wrapper is
901 distributed with PCRE. It is documented in the pcrecpp page.
902
903 The native API C function prototypes are defined in the header file
904 pcre.h, and on Unix systems the library itself is called libpcre. It
905 can normally be accessed by adding -lpcre to the command for linking an
906 application that uses PCRE. The header file defines the macros
907 PCRE_MAJOR and PCRE_MINOR to contain the major and minor release num-
908 bers for the library. Applications can use these to include support
909 for different releases of PCRE.
910
911 In a Windows environment, if you want to statically link an application
912 program against a non-dll pcre.a file, you must define PCRE_STATIC
913 before including pcre.h or pcrecpp.h, because otherwise the pcre_mal-
914 loc() and pcre_free() exported functions will be declared
915 __declspec(dllimport), with unwanted results.
916
917 The functions pcre_compile(), pcre_compile2(), pcre_study(), and
918 pcre_exec() are used for compiling and matching regular expressions in
919 a Perl-compatible manner. A sample program that demonstrates the sim-
920 plest way of using them is provided in the file called pcredemo.c in
921 the PCRE source distribution. A listing of this program is given in the
922 pcredemo documentation, and the pcresample documentation describes how
923 to compile and run it.
924
925 A second matching function, pcre_dfa_exec(), which is not Perl-compati-
926 ble, is also provided. This uses a different algorithm for the match-
927 ing. The alternative algorithm finds all possible matches (at a given
928 point in the subject), and scans the subject just once (unless there
929 are lookbehind assertions). However, this algorithm does not return
930 captured substrings. A description of the two matching algorithms and
931 their advantages and disadvantages is given in the pcrematching docu-
932 mentation.
933
934 In addition to the main compiling and matching functions, there are
935 convenience functions for extracting captured substrings from a subject
936 string that is matched by pcre_exec(). They are:
937
938 pcre_copy_substring()
939 pcre_copy_named_substring()
940 pcre_get_substring()
941 pcre_get_named_substring()
942 pcre_get_substring_list()
943 pcre_get_stringnumber()
944 pcre_get_stringtable_entries()
945
946 pcre_free_substring() and pcre_free_substring_list() are also provided,
947 to free the memory used for extracted strings.
948
949 The function pcre_maketables() is used to build a set of character
950 tables in the current locale for passing to pcre_compile(),
951 pcre_exec(), or pcre_dfa_exec(). This is an optional facility that is
952 provided for specialist use. Most commonly, no special tables are
953 passed, in which case internal tables that are generated when PCRE is
954 built are used.
955
956 The function pcre_fullinfo() is used to find out information about a
957 compiled pattern; pcre_info() is an obsolete version that returns only
958 some of the available information, but is retained for backwards com-
959 patibility. The function pcre_version() returns a pointer to a string
960 containing the version of PCRE and its date of release.
961
962 The function pcre_refcount() maintains a reference count in a data
963 block containing a compiled pattern. This is provided for the benefit
964 of object-oriented applications.
965
966 The global variables pcre_malloc and pcre_free initially contain the
967 entry points of the standard malloc() and free() functions, respec-
968 tively. PCRE calls the memory management functions via these variables,
969 so a calling program can replace them if it wishes to intercept the
970 calls. This should be done before calling any PCRE functions.
971
972 The global variables pcre_stack_malloc and pcre_stack_free are also
973 indirections to memory management functions. These special functions
974 are used only when PCRE is compiled to use the heap for remembering
975 data, instead of recursive function calls, when running the pcre_exec()
976 function. See the pcrebuild documentation for details of how to do
977 this. It is a non-standard way of building PCRE, for use in environ-
978 ments that have limited stacks. Because of the greater use of memory
979 management, it runs more slowly. Separate functions are provided so
980 that special-purpose external code can be used for this case. When
981 used, these functions are always called in a stack-like manner (last
982 obtained, first freed), and always for memory blocks of the same size.
983 There is a discussion about PCRE's stack usage in the pcrestack docu-
984 mentation.
985
986 The global variable pcre_callout initially contains NULL. It can be set
987 by the caller to a "callout" function, which PCRE will then call at
988 specified points during a matching operation. Details are given in the
989 pcrecallout documentation.
990
991
992 NEWLINES
993
994 PCRE supports five different conventions for indicating line breaks in
995 strings: a single CR (carriage return) character, a single LF (line-
996 feed) character, the two-character sequence CRLF, any of the three pre-
997 ceding, or any Unicode newline sequence. The Unicode newline sequences
998 are the three just mentioned, plus the single characters VT (vertical
999 tab, U+000B), FF (formfeed, U+000C), NEL (next line, U+0085), LS (line
1000 separator, U+2028), and PS (paragraph separator, U+2029).
1001
1002 Each of the first three conventions is used by at least one operating
1003 system as its standard newline sequence. When PCRE is built, a default
1004 can be specified. The default default is LF, which is the Unix stan-
1005 dard. When PCRE is run, the default can be overridden, either when a
1006 pattern is compiled, or when it is matched.
1007
1008 At compile time, the newline convention can be specified by the options
1009 argument of pcre_compile(), or it can be specified by special text at
1010 the start of the pattern itself; this overrides any other settings. See
1011 the pcrepattern page for details of the special character sequences.
1012
1013 In the PCRE documentation the word "newline" is used to mean "the char-
1014 acter or pair of characters that indicate a line break". The choice of
1015 newline convention affects the handling of the dot, circumflex, and
1016 dollar metacharacters, the handling of #-comments in /x mode, and, when
1017 CRLF is a recognized line ending sequence, the match position advance-
1018 ment for a non-anchored pattern. There is more detail about this in the
1019 section on pcre_exec() options below.
1020
1021 The choice of newline convention does not affect the interpretation of
1022 the \n or \r escape sequences, nor does it affect what \R matches,
1023 which is controlled in a similar way, but by separate options.
1024
1025
1026 MULTITHREADING
1027
1028 The PCRE functions can be used in multi-threading applications, with
1029 the proviso that the memory management functions pointed to by
1030 pcre_malloc, pcre_free, pcre_stack_malloc, and pcre_stack_free, and the
1031 callout function pointed to by pcre_callout, are shared by all threads.
1032
1033 The compiled form of a regular expression is not altered during match-
1034 ing, so the same compiled pattern can safely be used by several threads
1035 at once.
1036
1037
1038 SAVING PRECOMPILED PATTERNS FOR LATER USE
1039
1040 The compiled form of a regular expression can be saved and re-used at a
1041 later time, possibly by a different program, and even on a host other
1042 than the one on which it was compiled. Details are given in the
1043 pcreprecompile documentation. However, compiling a regular expression
1044 with one version of PCRE for use with a different version is not guar-
1045 anteed to work and may cause crashes.
1046
1047
1048 CHECKING BUILD-TIME OPTIONS
1049
1050 int pcre_config(int what, void *where);
1051
1052 The function pcre_config() makes it possible for a PCRE client to dis-
1053 cover which optional features have been compiled into the PCRE library.
1054 The pcrebuild documentation has more details about these optional fea-
1055 tures.
1056
1057 The first argument for pcre_config() is an integer, specifying which
1058 information is required; the second argument is a pointer to a variable
1059 into which the information is placed. The following information is
1060 available:
1061
1062 PCRE_CONFIG_UTF8
1063
1064 The output is an integer that is set to one if UTF-8 support is avail-
1065 able; otherwise it is set to zero.
1066
1067 PCRE_CONFIG_UNICODE_PROPERTIES
1068
1069 The output is an integer that is set to one if support for Unicode
1070 character properties is available; otherwise it is set to zero.
1071
1072 PCRE_CONFIG_NEWLINE
1073
1074 The output is an integer whose value specifies the default character
1075 sequence that is recognized as meaning "newline". The four values that
1076 are supported are: 10 for LF, 13 for CR, 3338 for CRLF, -2 for ANYCRLF,
1077 and -1 for ANY. Though they are derived from ASCII, the same values
1078 are returned in EBCDIC environments. The default should normally corre-
1079 spond to the standard sequence for your operating system.
1080
1081 PCRE_CONFIG_BSR
1082
1083 The output is an integer whose value indicates what character sequences
1084 the \R escape sequence matches by default. A value of 0 means that \R
1085 matches any Unicode line ending sequence; a value of 1 means that \R
1086 matches only CR, LF, or CRLF. The default can be overridden when a pat-
1087 tern is compiled or matched.
1088
1089 PCRE_CONFIG_LINK_SIZE
1090
1091 The output is an integer that contains the number of bytes used for
1092 internal linkage in compiled regular expressions. The value is 2, 3, or
1093 4. Larger values allow larger regular expressions to be compiled, at
1094 the expense of slower matching. The default value of 2 is sufficient
1095 for all but the most massive patterns, since it allows the compiled
1096 pattern to be up to 64K in size.
1097
1098 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
1099
1100 The output is an integer that contains the threshold above which the
1101 POSIX interface uses malloc() for output vectors. Further details are
1102 given in the pcreposix documentation.
1103
1104 PCRE_CONFIG_MATCH_LIMIT
1105
1106 The output is a long integer that gives the default limit for the num-
1107 ber of internal matching function calls in a pcre_exec() execution.
1108 Further details are given with pcre_exec() below.
1109
1110 PCRE_CONFIG_MATCH_LIMIT_RECURSION
1111
1112 The output is a long integer that gives the default limit for the depth
1113 of recursion when calling the internal matching function in a
1114 pcre_exec() execution. Further details are given with pcre_exec()
1115 below.
1116
1117 PCRE_CONFIG_STACKRECURSE
1118
1119 The output is an integer that is set to one if internal recursion when
1120 running pcre_exec() is implemented by recursive function calls that use
1121 the stack to remember their state. This is the usual way that PCRE is
1122 compiled. The output is zero if PCRE was compiled to use blocks of data
1123 on the heap instead of recursive function calls. In this case,
1124 pcre_stack_malloc and pcre_stack_free are called to manage memory
1125 blocks on the heap, thus avoiding the use of the stack.
1126
1127
1128 COMPILING A PATTERN
1129
1130 pcre *pcre_compile(const char *pattern, int options,
1131 const char **errptr, int *erroffset,
1132 const unsigned char *tableptr);
1133
1134 pcre *pcre_compile2(const char *pattern, int options,
1135 int *errorcodeptr,
1136 const char **errptr, int *erroffset,
1137 const unsigned char *tableptr);
1138
1139 Either of the functions pcre_compile() or pcre_compile2() can be called
1140 to compile a pattern into an internal form. The only difference between
1141 the two interfaces is that pcre_compile2() has an additional argument,
1142 errorcodeptr, via which a numerical error code can be returned. To
1143 avoid too much repetition, we refer just to pcre_compile() below, but
1144 the information applies equally to pcre_compile2().
1145
1146 The pattern is a C string terminated by a binary zero, and is passed in
1147 the pattern argument. A pointer to a single block of memory that is
1148 obtained via pcre_malloc is returned. This contains the compiled code
1149 and related data. The pcre type is defined for the returned block; this
1150 is a typedef for a structure whose contents are not externally defined.
1151 It is up to the caller to free the memory (via pcre_free) when it is no
1152 longer required.
1153
1154 Although the compiled code of a PCRE regex is relocatable, that is, it
1155 does not depend on memory location, the complete pcre data block is not
1156 fully relocatable, because it may contain a copy of the tableptr argu-
1157 ment, which is an address (see below).
1158
1159 The options argument contains various bit settings that affect the com-
1160 pilation. It should be zero if no options are required. The available
1161 options are described below. Some of them (in particular, those that
1162 are compatible with Perl, but some others as well) can also be set and
1163 unset from within the pattern (see the detailed description in the
1164 pcrepattern documentation). For those options that can be different in
1165 different parts of the pattern, the contents of the options argument
1166 specifies their settings at the start of compilation and execution. The
1167 PCRE_ANCHORED, PCRE_BSR_xxx, and PCRE_NEWLINE_xxx options can be set at
1168 the time of matching as well as at compile time.
1169
1170 If errptr is NULL, pcre_compile() returns NULL immediately. Otherwise,
1171 if compilation of a pattern fails, pcre_compile() returns NULL, and
1172 sets the variable pointed to by errptr to point to a textual error mes-
1173 sage. This is a static string that is part of the library. You must not
1174 try to free it. The byte offset from the start of the pattern to the
1175 character that was being processed when the error was discovered is
1176 placed in the variable pointed to by erroffset, which must not be NULL.
1177 If it is, an immediate error is given. Some errors are not detected
1178 until checks are carried out when the whole pattern has been scanned;
1179 in this case the offset is set to the end of the pattern.
1180
1181 If pcre_compile2() is used instead of pcre_compile(), and the error-
1182 codeptr argument is not NULL, a non-zero error code number is returned
1183 via this argument in the event of an error. This is in addition to the
1184 textual error message. Error codes and messages are listed below.
1185
1186 If the final argument, tableptr, is NULL, PCRE uses a default set of
1187 character tables that are built when PCRE is compiled, using the
1188 default C locale. Otherwise, tableptr must be an address that is the
1189 result of a call to pcre_maketables(). This value is stored with the
1190 compiled pattern, and used again by pcre_exec(), unless another table
1191 pointer is passed to it. For more discussion, see the section on locale
1192 support below.
1193
1194 This code fragment shows a typical straightforward call to pcre_com-
1195 pile():
1196
1197 pcre *re;
1198 const char *error;
1199 int erroffset;
1200 re = pcre_compile(
1201 "^A.*Z", /* the pattern */
1202 0, /* default options */
1203 &error, /* for error message */
1204 &erroffset, /* for error offset */
1205 NULL); /* use default character tables */
1206
1207 The following names for option bits are defined in the pcre.h header
1208 file:
1209
1210 PCRE_ANCHORED
1211
1212 If this bit is set, the pattern is forced to be "anchored", that is, it
1213 is constrained to match only at the first matching point in the string
1214 that is being searched (the "subject string"). This effect can also be
1215 achieved by appropriate constructs in the pattern itself, which is the
1216 only way to do it in Perl.
1217
1218 PCRE_AUTO_CALLOUT
1219
1220 If this bit is set, pcre_compile() automatically inserts callout items,
1221 all with number 255, before each pattern item. For discussion of the
1222 callout facility, see the pcrecallout documentation.
1223
1224 PCRE_BSR_ANYCRLF
1225 PCRE_BSR_UNICODE
1226
1227 These options (which are mutually exclusive) control what the \R escape
1228 sequence matches. The choice is either to match only CR, LF, or CRLF,
1229 or to match any Unicode newline sequence. The default is specified when
1230 PCRE is built. It can be overridden from within the pattern, or by set-
1231 ting an option when a compiled pattern is matched.
1232
1233 PCRE_CASELESS
1234
1235 If this bit is set, letters in the pattern match both upper and lower
1236 case letters. It is equivalent to Perl's /i option, and it can be
1237 changed within a pattern by a (?i) option setting. In UTF-8 mode, PCRE
1238 always understands the concept of case for characters whose values are
1239 less than 128, so caseless matching is always possible. For characters
1240 with higher values, the concept of case is supported if PCRE is com-
1241 piled with Unicode property support, but not otherwise. If you want to
1242 use caseless matching for characters 128 and above, you must ensure
1243 that PCRE is compiled with Unicode property support as well as with
1244 UTF-8 support.
1245
1246 PCRE_DOLLAR_ENDONLY
1247
1248 If this bit is set, a dollar metacharacter in the pattern matches only
1249 at the end of the subject string. Without this option, a dollar also
1250 matches immediately before a newline at the end of the string (but not
1251 before any other newlines). The PCRE_DOLLAR_ENDONLY option is ignored
1252 if PCRE_MULTILINE is set. There is no equivalent to this option in
1253 Perl, and no way to set it within a pattern.
1254
1255 PCRE_DOTALL
1256
1257 If this bit is set, a dot metacharater in the pattern matches all char-
1258 acters, including those that indicate newline. Without it, a dot does
1259 not match when the current position is at a newline. This option is
1260 equivalent to Perl's /s option, and it can be changed within a pattern
1261 by a (?s) option setting. A negative class such as [^a] always matches
1262 newline characters, independent of the setting of this option.
1263
1264 PCRE_DUPNAMES
1265
1266 If this bit is set, names used to identify capturing subpatterns need
1267 not be unique. This can be helpful for certain types of pattern when it
1268 is known that only one instance of the named subpattern can ever be
1269 matched. There are more details of named subpatterns below; see also
1270 the pcrepattern documentation.
1271
1272 PCRE_EXTENDED
1273
1274 If this bit is set, whitespace data characters in the pattern are
1275 totally ignored except when escaped or inside a character class. White-
1276 space does not include the VT character (code 11). In addition, charac-
1277 ters between an unescaped # outside a character class and the next new-
1278 line, inclusive, are also ignored. This is equivalent to Perl's /x
1279 option, and it can be changed within a pattern by a (?x) option set-
1280 ting.
1281
1282 This option makes it possible to include comments inside complicated
1283 patterns. Note, however, that this applies only to data characters.
1284 Whitespace characters may never appear within special character
1285 sequences in a pattern, for example within the sequence (?( which
1286 introduces a conditional subpattern.
1287
1288 PCRE_EXTRA
1289
1290 This option was invented in order to turn on additional functionality
1291 of PCRE that is incompatible with Perl, but it is currently of very
1292 little use. When set, any backslash in a pattern that is followed by a
1293 letter that has no special meaning causes an error, thus reserving
1294 these combinations for future expansion. By default, as in Perl, a
1295 backslash followed by a letter with no special meaning is treated as a
1296 literal. (Perl can, however, be persuaded to give an error for this, by
1297 running it with the -w option.) There are at present no other features
1298 controlled by this option. It can also be set by a (?X) option setting
1299 within a pattern.
1300
1301 PCRE_FIRSTLINE
1302
1303 If this option is set, an unanchored pattern is required to match
1304 before or at the first newline in the subject string, though the
1305 matched text may continue over the newline.
1306
1307 PCRE_JAVASCRIPT_COMPAT
1308
1309 If this option is set, PCRE's behaviour is changed in some ways so that
1310 it is compatible with JavaScript rather than Perl. The changes are as
1311 follows:
1312
1313 (1) A lone closing square bracket in a pattern causes a compile-time
1314 error, because this is illegal in JavaScript (by default it is treated
1315 as a data character). Thus, the pattern AB]CD becomes illegal when this
1316 option is set.
1317
1318 (2) At run time, a back reference to an unset subpattern group matches
1319 an empty string (by default this causes the current matching alterna-
1320 tive to fail). A pattern such as (\1)(a) succeeds when this option is
1321 set (assuming it can find an "a" in the subject), whereas it fails by
1322 default, for Perl compatibility.
1323
1324 PCRE_MULTILINE
1325
1326 By default, PCRE treats the subject string as consisting of a single
1327 line of characters (even if it actually contains newlines). The "start
1328 of line" metacharacter (^) matches only at the start of the string,
1329 while the "end of line" metacharacter ($) matches only at the end of
1330 the string, or before a terminating newline (unless PCRE_DOLLAR_ENDONLY
1331 is set). This is the same as Perl.
1332
1333 When PCRE_MULTILINE it is set, the "start of line" and "end of line"
1334 constructs match immediately following or immediately before internal
1335 newlines in the subject string, respectively, as well as at the very
1336 start and end. This is equivalent to Perl's /m option, and it can be
1337 changed within a pattern by a (?m) option setting. If there are no new-
1338 lines in a subject string, or no occurrences of ^ or $ in a pattern,
1339 setting PCRE_MULTILINE has no effect.
1340
1341 PCRE_NEWLINE_CR
1342 PCRE_NEWLINE_LF
1343 PCRE_NEWLINE_CRLF
1344 PCRE_NEWLINE_ANYCRLF
1345 PCRE_NEWLINE_ANY
1346
1347 These options override the default newline definition that was chosen
1348 when PCRE was built. Setting the first or the second specifies that a
1349 newline is indicated by a single character (CR or LF, respectively).
1350 Setting PCRE_NEWLINE_CRLF specifies that a newline is indicated by the
1351 two-character CRLF sequence. Setting PCRE_NEWLINE_ANYCRLF specifies
1352 that any of the three preceding sequences should be recognized. Setting
1353 PCRE_NEWLINE_ANY specifies that any Unicode newline sequence should be
1354 recognized. The Unicode newline sequences are the three just mentioned,
1355 plus the single characters VT (vertical tab, U+000B), FF (formfeed,
1356 U+000C), NEL (next line, U+0085), LS (line separator, U+2028), and PS
1357 (paragraph separator, U+2029). The last two are recognized only in
1358 UTF-8 mode.
1359
1360 The newline setting in the options word uses three bits that are
1361 treated as a number, giving eight possibilities. Currently only six are
1362 used (default plus the five values above). This means that if you set
1363 more than one newline option, the combination may or may not be sensi-
1364 ble. For example, PCRE_NEWLINE_CR with PCRE_NEWLINE_LF is equivalent to
1365 PCRE_NEWLINE_CRLF, but other combinations may yield unused numbers and
1366 cause an error.
1367
1368 The only time that a line break is specially recognized when compiling
1369 a pattern is if PCRE_EXTENDED is set, and an unescaped # outside a
1370 character class is encountered. This indicates a comment that lasts
1371 until after the next line break sequence. In other circumstances, line
1372 break sequences are treated as literal data, except that in
1373 PCRE_EXTENDED mode, both CR and LF are treated as whitespace characters
1374 and are therefore ignored.
1375
1376 The newline option that is set at compile time becomes the default that
1377 is used for pcre_exec() and pcre_dfa_exec(), but it can be overridden.
1378
1379 PCRE_NO_AUTO_CAPTURE
1380
1381 If this option is set, it disables the use of numbered capturing paren-
1382 theses in the pattern. Any opening parenthesis that is not followed by
1383 ? behaves as if it were followed by ?: but named parentheses can still
1384 be used for capturing (and they acquire numbers in the usual way).
1385 There is no equivalent of this option in Perl.
1386
1387 PCRE_UCP
1388
1389 This option changes the way PCRE processes \b, \d, \s, \w, and some of
1390 the POSIX character classes. By default, only ASCII characters are rec-
1391 ognized, but if PCRE_UCP is set, Unicode properties are used instead to
1392 classify characters. More details are given in the section on generic
1393 character types in the pcrepattern page. If you set PCRE_UCP, matching
1394 one of the items it affects takes much longer. The option is available
1395 only if PCRE has been compiled with Unicode property support.
1396
1397 PCRE_UNGREEDY
1398
1399 This option inverts the "greediness" of the quantifiers so that they
1400 are not greedy by default, but become greedy if followed by "?". It is
1401 not compatible with Perl. It can also be set by a (?U) option setting
1402 within the pattern.
1403
1404 PCRE_UTF8
1405
1406 This option causes PCRE to regard both the pattern and the subject as
1407 strings of UTF-8 characters instead of single-byte character strings.
1408 However, it is available only when PCRE is built to include UTF-8 sup-
1409 port. If not, the use of this option provokes an error. Details of how
1410 this option changes the behaviour of PCRE are given in the section on
1411 UTF-8 support in the main pcre page.
1412
1413 PCRE_NO_UTF8_CHECK
1414
1415 When PCRE_UTF8 is set, the validity of the pattern as a UTF-8 string is
1416 automatically checked. There is a discussion about the validity of
1417 UTF-8 strings in the main pcre page. If an invalid UTF-8 sequence of
1418 bytes is found, pcre_compile() returns an error. If you already know
1419 that your pattern is valid, and you want to skip this check for perfor-
1420 mance reasons, you can set the PCRE_NO_UTF8_CHECK option. When it is
1421 set, the effect of passing an invalid UTF-8 string as a pattern is
1422 undefined. It may cause your program to crash. Note that this option
1423 can also be passed to pcre_exec() and pcre_dfa_exec(), to suppress the
1424 UTF-8 validity checking of subject strings.
1425
1426
1427 COMPILATION ERROR CODES
1428
1429 The following table lists the error codes than may be returned by
1430 pcre_compile2(), along with the error messages that may be returned by
1431 both compiling functions. As PCRE has developed, some error codes have
1432 fallen out of use. To avoid confusion, they have not been re-used.
1433
1434 0 no error
1435 1 \ at end of pattern
1436 2 \c at end of pattern
1437 3 unrecognized character follows \
1438 4 numbers out of order in {} quantifier
1439 5 number too big in {} quantifier
1440 6 missing terminating ] for character class
1441 7 invalid escape sequence in character class
1442 8 range out of order in character class
1443 9 nothing to repeat
1444 10 [this code is not in use]
1445 11 internal error: unexpected repeat
1446 12 unrecognized character after (? or (?-
1447 13 POSIX named classes are supported only within a class
1448 14 missing )
1449 15 reference to non-existent subpattern
1450 16 erroffset passed as NULL
1451 17 unknown option bit(s) set
1452 18 missing ) after comment
1453 19 [this code is not in use]
1454 20 regular expression is too large
1455 21 failed to get memory
1456 22 unmatched parentheses
1457 23 internal error: code overflow
1458 24 unrecognized character after (?<
1459 25 lookbehind assertion is not fixed length
1460 26 malformed number or name after (?(
1461 27 conditional group contains more than two branches
1462 28 assertion expected after (?(
1463 29 (?R or (?[+-]digits must be followed by )
1464 30 unknown POSIX class name
1465 31 POSIX collating elements are not supported
1466 32 this version of PCRE is not compiled with PCRE_UTF8 support
1467 33 [this code is not in use]
1468 34 character value in \x{...} sequence is too large
1469 35 invalid condition (?(0)
1470 36 \C not allowed in lookbehind assertion
1471 37 PCRE does not support \L, \l, \N, \U, or \u
1472 38 number after (?C is > 255
1473 39 closing ) for (?C expected
1474 40 recursive call could loop indefinitely
1475 41 unrecognized character after (?P
1476 42 syntax error in subpattern name (missing terminator)
1477 43 two named subpatterns have the same name
1478 44 invalid UTF-8 string
1479 45 support for \P, \p, and \X has not been compiled
1480 46 malformed \P or \p sequence
1481 47 unknown property name after \P or \p
1482 48 subpattern name is too long (maximum 32 characters)
1483 49 too many named subpatterns (maximum 10000)
1484 50 [this code is not in use]
1485 51 octal value is greater than \377 (not in UTF-8 mode)
1486 52 internal error: overran compiling workspace
1487 53 internal error: previously-checked referenced subpattern
1488 not found
1489 54 DEFINE group contains more than one branch
1490 55 repeating a DEFINE group is not allowed
1491 56 inconsistent NEWLINE options
1492 57 \g is not followed by a braced, angle-bracketed, or quoted
1493 name/number or by a plain number
1494 58 a numbered reference must not be zero
1495 59 an argument is not allowed for (*ACCEPT), (*FAIL), or (*COMMIT)
1496 60 (*VERB) not recognized
1497 61 number is too big
1498 62 subpattern name expected
1499 63 digit expected after (?+
1500 64 ] is an invalid data character in JavaScript compatibility mode
1501 65 different names for subpatterns of the same number are
1502 not allowed
1503 66 (*MARK) must have an argument
1504 67 this version of PCRE is not compiled with PCRE_UCP support
1505
1506 The numbers 32 and 10000 in errors 48 and 49 are defaults; different
1507 values may be used if the limits were changed when PCRE was built.
1508
1509
1510 STUDYING A PATTERN
1511
1512 pcre_extra *pcre_study(const pcre *code, int options
1513 const char **errptr);
1514
1515 If a compiled pattern is going to be used several times, it is worth
1516 spending more time analyzing it in order to speed up the time taken for
1517 matching. The function pcre_study() takes a pointer to a compiled pat-
1518 tern as its first argument. If studying the pattern produces additional
1519 information that will help speed up matching, pcre_study() returns a
1520 pointer to a pcre_extra block, in which the study_data field points to
1521 the results of the study.
1522
1523 The returned value from pcre_study() can be passed directly to
1524 pcre_exec() or pcre_dfa_exec(). However, a pcre_extra block also con-
1525 tains other fields that can be set by the caller before the block is
1526 passed; these are described below in the section on matching a pattern.
1527
1528 If studying the pattern does not produce any useful information,
1529 pcre_study() returns NULL. In that circumstance, if the calling program
1530 wants to pass any of the other fields to pcre_exec() or
1531 pcre_dfa_exec(), it must set up its own pcre_extra block.
1532
1533 The second argument of pcre_study() contains option bits. At present,
1534 no options are defined, and this argument should always be zero.
1535
1536 The third argument for pcre_study() is a pointer for an error message.
1537 If studying succeeds (even if no data is returned), the variable it
1538 points to is set to NULL. Otherwise it is set to point to a textual
1539 error message. This is a static string that is part of the library. You
1540 must not try to free it. You should test the error pointer for NULL
1541 after calling pcre_study(), to be sure that it has run successfully.
1542
1543 This is a typical call to pcre_study():
1544
1545 pcre_extra *pe;
1546 pe = pcre_study(
1547 re, /* result of pcre_compile() */
1548 0, /* no options exist */
1549 &error); /* set to NULL or points to a message */
1550
1551 Studying a pattern does two things: first, a lower bound for the length
1552 of subject string that is needed to match the pattern is computed. This
1553 does not mean that there are any strings of that length that match, but
1554 it does guarantee that no shorter strings match. The value is used by
1555 pcre_exec() and pcre_dfa_exec() to avoid wasting time by trying to
1556 match strings that are shorter than the lower bound. You can find out
1557 the value in a calling program via the pcre_fullinfo() function.
1558
1559 Studying a pattern is also useful for non-anchored patterns that do not
1560 have a single fixed starting character. A bitmap of possible starting
1561 bytes is created. This speeds up finding a position in the subject at
1562 which to start matching.
1563
1564 The two optimizations just described can be disabled by setting the
1565 PCRE_NO_START_OPTIMIZE option when calling pcre_exec() or
1566 pcre_dfa_exec(). You might want to do this if your pattern contains
1567 callouts, or make use of (*MARK), and you make use of these in cases
1568 where matching fails. See the discussion of PCRE_NO_START_OPTIMIZE
1569 below.
1570
1571
1572 LOCALE SUPPORT
1573
1574 PCRE handles caseless matching, and determines whether characters are
1575 letters, digits, or whatever, by reference to a set of tables, indexed
1576 by character value. When running in UTF-8 mode, this applies only to
1577 characters with codes less than 128. By default, higher-valued codes
1578 never match escapes such as \w or \d, but they can be tested with \p if
1579 PCRE is built with Unicode character property support. Alternatively,
1580 the PCRE_UCP option can be set at compile time; this causes \w and
1581 friends to use Unicode property support instead of built-in tables. The
1582 use of locales with Unicode is discouraged. If you are handling charac-
1583 ters with codes greater than 128, you should either use UTF-8 and Uni-
1584 code, or use locales, but not try to mix the two.
1585
1586 PCRE contains an internal set of tables that are used when the final
1587 argument of pcre_compile() is NULL. These are sufficient for many
1588 applications. Normally, the internal tables recognize only ASCII char-
1589 acters. However, when PCRE is built, it is possible to cause the inter-
1590 nal tables to be rebuilt in the default "C" locale of the local system,
1591 which may cause them to be different.
1592
1593 The internal tables can always be overridden by tables supplied by the
1594 application that calls PCRE. These may be created in a different locale
1595 from the default. As more and more applications change to using Uni-
1596 code, the need for this locale support is expected to die away.
1597
1598 External tables are built by calling the pcre_maketables() function,
1599 which has no arguments, in the relevant locale. The result can then be
1600 passed to pcre_compile() or pcre_exec() as often as necessary. For
1601 example, to build and use tables that are appropriate for the French
1602 locale (where accented characters with values greater than 128 are
1603 treated as letters), the following code could be used:
1604
1605 setlocale(LC_CTYPE, "fr_FR");
1606 tables = pcre_maketables();
1607 re = pcre_compile(..., tables);
1608
1609 The locale name "fr_FR" is used on Linux and other Unix-like systems;
1610 if you are using Windows, the name for the French locale is "french".
1611
1612 When pcre_maketables() runs, the tables are built in memory that is
1613 obtained via pcre_malloc. It is the caller's responsibility to ensure
1614 that the memory containing the tables remains available for as long as
1615 it is needed.
1616
1617 The pointer that is passed to pcre_compile() is saved with the compiled
1618 pattern, and the same tables are used via this pointer by pcre_study()
1619 and normally also by pcre_exec(). Thus, by default, for any single pat-
1620 tern, compilation, studying and matching all happen in the same locale,
1621 but different patterns can be compiled in different locales.
1622
1623 It is possible to pass a table pointer or NULL (indicating the use of
1624 the internal tables) to pcre_exec(). Although not intended for this
1625 purpose, this facility could be used to match a pattern in a different
1626 locale from the one in which it was compiled. Passing table pointers at
1627 run time is discussed below in the section on matching a pattern.
1628
1629
1630 INFORMATION ABOUT A PATTERN
1631
1632 int pcre_fullinfo(const pcre *code, const pcre_extra *extra,
1633 int what, void *where);
1634
1635 The pcre_fullinfo() function returns information about a compiled pat-
1636 tern. It replaces the obsolete pcre_info() function, which is neverthe-
1637 less retained for backwards compability (and is documented below).
1638
1639 The first argument for pcre_fullinfo() is a pointer to the compiled
1640 pattern. The second argument is the result of pcre_study(), or NULL if
1641 the pattern was not studied. The third argument specifies which piece
1642 of information is required, and the fourth argument is a pointer to a
1643 variable to receive the data. The yield of the function is zero for
1644 success, or one of the following negative numbers:
1645
1646 PCRE_ERROR_NULL the argument code was NULL
1647 the argument where was NULL
1648 PCRE_ERROR_BADMAGIC the "magic number" was not found
1649 PCRE_ERROR_BADOPTION the value of what was invalid
1650
1651 The "magic number" is placed at the start of each compiled pattern as
1652 an simple check against passing an arbitrary memory pointer. Here is a
1653 typical call of pcre_fullinfo(), to obtain the length of the compiled
1654 pattern:
1655
1656 int rc;
1657 size_t length;
1658 rc = pcre_fullinfo(
1659 re, /* result of pcre_compile() */
1660 pe, /* result of pcre_study(), or NULL */
1661 PCRE_INFO_SIZE, /* what is required */
1662 &length); /* where to put the data */
1663
1664 The possible values for the third argument are defined in pcre.h, and
1665 are as follows:
1666
1667 PCRE_INFO_BACKREFMAX
1668
1669 Return the number of the highest back reference in the pattern. The
1670 fourth argument should point to an int variable. Zero is returned if
1671 there are no back references.
1672
1673 PCRE_INFO_CAPTURECOUNT
1674
1675 Return the number of capturing subpatterns in the pattern. The fourth
1676 argument should point to an int variable.
1677
1678 PCRE_INFO_DEFAULT_TABLES
1679
1680 Return a pointer to the internal default character tables within PCRE.
1681 The fourth argument should point to an unsigned char * variable. This
1682 information call is provided for internal use by the pcre_study() func-
1683 tion. External callers can cause PCRE to use its internal tables by
1684 passing a NULL table pointer.
1685
1686 PCRE_INFO_FIRSTBYTE
1687
1688 Return information about the first byte of any matched string, for a
1689 non-anchored pattern. The fourth argument should point to an int vari-
1690 able. (This option used to be called PCRE_INFO_FIRSTCHAR; the old name
1691 is still recognized for backwards compatibility.)
1692
1693 If there is a fixed first byte, for example, from a pattern such as
1694 (cat|cow|coyote), its value is returned. Otherwise, if either
1695
1696 (a) the pattern was compiled with the PCRE_MULTILINE option, and every
1697 branch starts with "^", or
1698
1699 (b) every branch of the pattern starts with ".*" and PCRE_DOTALL is not
1700 set (if it were set, the pattern would be anchored),
1701
1702 -1 is returned, indicating that the pattern matches only at the start
1703 of a subject string or after any newline within the string. Otherwise
1704 -2 is returned. For anchored patterns, -2 is returned.
1705
1706 PCRE_INFO_FIRSTTABLE
1707
1708 If the pattern was studied, and this resulted in the construction of a
1709 256-bit table indicating a fixed set of bytes for the first byte in any
1710 matching string, a pointer to the table is returned. Otherwise NULL is
1711 returned. The fourth argument should point to an unsigned char * vari-
1712 able.
1713
1714 PCRE_INFO_HASCRORLF
1715
1716 Return 1 if the pattern contains any explicit matches for CR or LF
1717 characters, otherwise 0. The fourth argument should point to an int
1718 variable. An explicit match is either a literal CR or LF character, or
1719 \r or \n.
1720
1721 PCRE_INFO_JCHANGED
1722
1723 Return 1 if the (?J) or (?-J) option setting is used in the pattern,
1724 otherwise 0. The fourth argument should point to an int variable. (?J)
1725 and (?-J) set and unset the local PCRE_DUPNAMES option, respectively.
1726
1727 PCRE_INFO_LASTLITERAL
1728
1729 Return the value of the rightmost literal byte that must exist in any
1730 matched string, other than at its start, if such a byte has been
1731 recorded. The fourth argument should point to an int variable. If there
1732 is no such byte, -1 is returned. For anchored patterns, a last literal
1733 byte is recorded only if it follows something of variable length. For
1734 example, for the pattern /^a\d+z\d+/ the returned value is "z", but for
1735 /^a\dz\d/ the returned value is -1.
1736
1737 PCRE_INFO_MINLENGTH
1738
1739 If the pattern was studied and a minimum length for matching subject
1740 strings was computed, its value is returned. Otherwise the returned
1741 value is -1. The value is a number of characters, not bytes (this may
1742 be relevant in UTF-8 mode). The fourth argument should point to an int
1743 variable. A non-negative value is a lower bound to the length of any
1744 matching string. There may not be any strings of that length that do
1745 actually match, but every string that does match is at least that long.
1746
1747 PCRE_INFO_NAMECOUNT
1748 PCRE_INFO_NAMEENTRYSIZE
1749 PCRE_INFO_NAMETABLE
1750
1751 PCRE supports the use of named as well as numbered capturing parenthe-
1752 ses. The names are just an additional way of identifying the parenthe-
1753 ses, which still acquire numbers. Several convenience functions such as
1754 pcre_get_named_substring() are provided for extracting captured sub-
1755 strings by name. It is also possible to extract the data directly, by
1756 first converting the name to a number in order to access the correct
1757 pointers in the output vector (described with pcre_exec() below). To do
1758 the conversion, you need to use the name-to-number map, which is
1759 described by these three values.
1760
1761 The map consists of a number of fixed-size entries. PCRE_INFO_NAMECOUNT
1762 gives the number of entries, and PCRE_INFO_NAMEENTRYSIZE gives the size
1763 of each entry; both of these return an int value. The entry size
1764 depends on the length of the longest name. PCRE_INFO_NAMETABLE returns
1765 a pointer to the first entry of the table (a pointer to char). The
1766 first two bytes of each entry are the number of the capturing parenthe-
1767 sis, most significant byte first. The rest of the entry is the corre-
1768 sponding name, zero terminated.
1769
1770 The names are in alphabetical order. Duplicate names may appear if (?|
1771 is used to create multiple groups with the same number, as described in
1772 the section on duplicate subpattern numbers in the pcrepattern page.
1773 Duplicate names for subpatterns with different numbers are permitted
1774 only if PCRE_DUPNAMES is set. In all cases of duplicate names, they
1775 appear in the table in the order in which they were found in the pat-
1776 tern. In the absence of (?| this is the order of increasing number;
1777 when (?| is used this is not necessarily the case because later subpat-
1778 terns may have lower numbers.
1779
1780 As a simple example of the name/number table, consider the following
1781 pattern (assume PCRE_EXTENDED is set, so white space - including new-
1782 lines - is ignored):
1783
1784 (?<date> (?<year>(\d\d)?\d\d) -
1785 (?<month>\d\d) - (?<day>\d\d) )
1786
1787 There are four named subpatterns, so the table has four entries, and
1788 each entry in the table is eight bytes long. The table is as follows,
1789 with non-printing bytes shows in hexadecimal, and undefined bytes shown
1790 as ??:
1791
1792 00 01 d a t e 00 ??
1793 00 05 d a y 00 ?? ??
1794 00 04 m o n t h 00
1795 00 02 y e a r 00 ??
1796
1797 When writing code to extract data from named subpatterns using the
1798 name-to-number map, remember that the length of the entries is likely
1799 to be different for each compiled pattern.
1800
1801 PCRE_INFO_OKPARTIAL
1802
1803 Return 1 if the pattern can be used for partial matching with
1804 pcre_exec(), otherwise 0. The fourth argument should point to an int
1805 variable. From release 8.00, this always returns 1, because the
1806 restrictions that previously applied to partial matching have been
1807 lifted. The pcrepartial documentation gives details of partial match-
1808 ing.
1809
1810 PCRE_INFO_OPTIONS
1811
1812 Return a copy of the options with which the pattern was compiled. The
1813 fourth argument should point to an unsigned long int variable. These
1814 option bits are those specified in the call to pcre_compile(), modified
1815 by any top-level option settings at the start of the pattern itself. In
1816 other words, they are the options that will be in force when matching
1817 starts. For example, if the pattern /(?im)abc(?-i)d/ is compiled with
1818 the PCRE_EXTENDED option, the result is PCRE_CASELESS, PCRE_MULTILINE,
1819 and PCRE_EXTENDED.
1820
1821 A pattern is automatically anchored by PCRE if all of its top-level
1822 alternatives begin with one of the following:
1823
1824 ^ unless PCRE_MULTILINE is set
1825 \A always
1826 \G always
1827 .* if PCRE_DOTALL is set and there are no back
1828 references to the subpattern in which .* appears
1829
1830 For such patterns, the PCRE_ANCHORED bit is set in the options returned
1831 by pcre_fullinfo().
1832
1833 PCRE_INFO_SIZE
1834
1835 Return the size of the compiled pattern, that is, the value that was
1836 passed as the argument to pcre_malloc() when PCRE was getting memory in
1837 which to place the compiled data. The fourth argument should point to a
1838 size_t variable.
1839
1840 PCRE_INFO_STUDYSIZE
1841
1842 Return the size of the data block pointed to by the study_data field in
1843 a pcre_extra block. That is, it is the value that was passed to
1844 pcre_malloc() when PCRE was getting memory into which to place the data
1845 created by pcre_study(). If pcre_extra is NULL, or there is no study
1846 data, zero is returned. The fourth argument should point to a size_t
1847 variable.
1848
1849
1850 OBSOLETE INFO FUNCTION
1851
1852 int pcre_info(const pcre *code, int *optptr, int *firstcharptr);
1853
1854 The pcre_info() function is now obsolete because its interface is too
1855 restrictive to return all the available data about a compiled pattern.
1856 New programs should use pcre_fullinfo() instead. The yield of
1857 pcre_info() is the number of capturing subpatterns, or one of the fol-
1858 lowing negative numbers:
1859
1860 PCRE_ERROR_NULL the argument code was NULL
1861 PCRE_ERROR_BADMAGIC the "magic number" was not found
1862
1863 If the optptr argument is not NULL, a copy of the options with which
1864 the pattern was compiled is placed in the integer it points to (see
1865 PCRE_INFO_OPTIONS above).
1866
1867 If the pattern is not anchored and the firstcharptr argument is not
1868 NULL, it is used to pass back information about the first character of
1869 any matched string (see PCRE_INFO_FIRSTBYTE above).
1870
1871
1872 REFERENCE COUNTS
1873
1874 int pcre_refcount(pcre *code, int adjust);
1875
1876 The pcre_refcount() function is used to maintain a reference count in
1877 the data block that contains a compiled pattern. It is provided for the
1878 benefit of applications that operate in an object-oriented manner,
1879 where different parts of the application may be using the same compiled
1880 pattern, but you want to free the block when they are all done.
1881
1882 When a pattern is compiled, the reference count field is initialized to
1883 zero. It is changed only by calling this function, whose action is to
1884 add the adjust value (which may be positive or negative) to it. The
1885 yield of the function is the new value. However, the value of the count
1886 is constrained to lie between 0 and 65535, inclusive. If the new value
1887 is outside these limits, it is forced to the appropriate limit value.
1888
1889 Except when it is zero, the reference count is not correctly preserved
1890 if a pattern is compiled on one host and then transferred to a host
1891 whose byte-order is different. (This seems a highly unlikely scenario.)
1892
1893
1894 MATCHING A PATTERN: THE TRADITIONAL FUNCTION
1895
1896 int pcre_exec(const pcre *code, const pcre_extra *extra,
1897 const char *subject, int length, int startoffset,
1898 int options, int *ovector, int ovecsize);
1899
1900 The function pcre_exec() is called to match a subject string against a
1901 compiled pattern, which is passed in the code argument. If the pattern
1902 was studied, the result of the study should be passed in the extra
1903 argument. This function is the main matching facility of the library,
1904 and it operates in a Perl-like manner. For specialist use there is also
1905 an alternative matching function, which is described below in the sec-
1906 tion about the pcre_dfa_exec() function.
1907
1908 In most applications, the pattern will have been compiled (and option-
1909 ally studied) in the same process that calls pcre_exec(). However, it
1910 is possible to save compiled patterns and study data, and then use them
1911 later in different processes, possibly even on different hosts. For a
1912 discussion about this, see the pcreprecompile documentation.
1913
1914 Here is an example of a simple call to pcre_exec():
1915
1916 int rc;
1917 int ovector[30];
1918 rc = pcre_exec(
1919 re, /* result of pcre_compile() */
1920 NULL, /* we didn't study the pattern */
1921 "some string", /* the subject string */
1922 11, /* the length of the subject string */
1923 0, /* start at offset 0 in the subject */
1924 0, /* default options */
1925 ovector, /* vector of integers for substring information */
1926 30); /* number of elements (NOT size in bytes) */
1927
1928 Extra data for pcre_exec()
1929
1930 If the extra argument is not NULL, it must point to a pcre_extra data
1931 block. The pcre_study() function returns such a block (when it doesn't
1932 return NULL), but you can also create one for yourself, and pass addi-
1933 tional information in it. The pcre_extra block contains the following
1934 fields (not necessarily in this order):
1935
1936 unsigned long int flags;
1937 void *study_data;
1938 unsigned long int match_limit;
1939 unsigned long int match_limit_recursion;
1940 void *callout_data;
1941 const unsigned char *tables;
1942 unsigned char **mark;
1943
1944 The flags field is a bitmap that specifies which of the other fields
1945 are set. The flag bits are:
1946
1947 PCRE_EXTRA_STUDY_DATA
1948 PCRE_EXTRA_MATCH_LIMIT
1949 PCRE_EXTRA_MATCH_LIMIT_RECURSION
1950 PCRE_EXTRA_CALLOUT_DATA
1951 PCRE_EXTRA_TABLES
1952 PCRE_EXTRA_MARK
1953
1954 Other flag bits should be set to zero. The study_data field is set in
1955 the pcre_extra block that is returned by pcre_study(), together with
1956 the appropriate flag bit. You should not set this yourself, but you may
1957 add to the block by setting the other fields and their corresponding
1958 flag bits.
1959
1960 The match_limit field provides a means of preventing PCRE from using up
1961 a vast amount of resources when running patterns that are not going to
1962 match, but which have a very large number of possibilities in their
1963 search trees. The classic example is a pattern that uses nested unlim-
1964 ited repeats.
1965
1966 Internally, PCRE uses a function called match() which it calls repeat-
1967 edly (sometimes recursively). The limit set by match_limit is imposed
1968 on the number of times this function is called during a match, which
1969 has the effect of limiting the amount of backtracking that can take
1970 place. For patterns that are not anchored, the count restarts from zero
1971 for each position in the subject string.
1972
1973 The default value for the limit can be set when PCRE is built; the
1974 default default is 10 million, which handles all but the most extreme
1975 cases. You can override the default by suppling pcre_exec() with a
1976 pcre_extra block in which match_limit is set, and
1977 PCRE_EXTRA_MATCH_LIMIT is set in the flags field. If the limit is
1978 exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
1979
1980 The match_limit_recursion field is similar to match_limit, but instead
1981 of limiting the total number of times that match() is called, it limits
1982 the depth of recursion. The recursion depth is a smaller number than
1983 the total number of calls, because not all calls to match() are recur-
1984 sive. This limit is of use only if it is set smaller than match_limit.
1985
1986 Limiting the recursion depth limits the amount of stack that can be
1987 used, or, when PCRE has been compiled to use memory on the heap instead
1988 of the stack, the amount of heap memory that can be used.
1989
1990 The default value for match_limit_recursion can be set when PCRE is
1991 built; the default default is the same value as the default for
1992 match_limit. You can override the default by suppling pcre_exec() with
1993 a pcre_extra block in which match_limit_recursion is set, and
1994 PCRE_EXTRA_MATCH_LIMIT_RECURSION is set in the flags field. If the
1995 limit is exceeded, pcre_exec() returns PCRE_ERROR_RECURSIONLIMIT.
1996
1997 The callout_data field is used in conjunction with the "callout" fea-
1998 ture, and is described in the pcrecallout documentation.
1999
2000 The tables field is used to pass a character tables pointer to
2001 pcre_exec(); this overrides the value that is stored with the compiled
2002 pattern. A non-NULL value is stored with the compiled pattern only if
2003 custom tables were supplied to pcre_compile() via its tableptr argu-
2004 ment. If NULL is passed to pcre_exec() using this mechanism, it forces
2005 PCRE's internal tables to be used. This facility is helpful when re-
2006 using patterns that have been saved after compiling with an external
2007 set of tables, because the external tables might be at a different
2008 address when pcre_exec() is called. See the pcreprecompile documenta-
2009 tion for a discussion of saving compiled patterns for later use.
2010
2011 If PCRE_EXTRA_MARK is set in the flags field, the mark field must be
2012 set to point to a char * variable. If the pattern contains any back-
2013 tracking control verbs such as (*MARK:NAME), and the execution ends up
2014 with a name to pass back, a pointer to the name string (zero termi-
2015 nated) is placed in the variable pointed to by the mark field. The
2016 names are within the compiled pattern; if you wish to retain such a
2017 name you must copy it before freeing the memory of a compiled pattern.
2018 If there is no name to pass back, the variable pointed to by the mark
2019 field set to NULL. For details of the backtracking control verbs, see
2020 the section entitled "Backtracking control" in the pcrepattern documen-
2021 tation.
2022
2023 Option bits for pcre_exec()
2024
2025 The unused bits of the options argument for pcre_exec() must be zero.
2026 The only bits that may be set are PCRE_ANCHORED, PCRE_NEWLINE_xxx,
2027 PCRE_NOTBOL, PCRE_NOTEOL, PCRE_NOTEMPTY, PCRE_NOTEMPTY_ATSTART,
2028 PCRE_NO_START_OPTIMIZE, PCRE_NO_UTF8_CHECK, PCRE_PARTIAL_SOFT, and
2029 PCRE_PARTIAL_HARD.
2030
2031 PCRE_ANCHORED
2032
2033 The PCRE_ANCHORED option limits pcre_exec() to matching at the first
2034 matching position. If a pattern was compiled with PCRE_ANCHORED, or
2035 turned out to be anchored by virtue of its contents, it cannot be made
2036 unachored at matching time.
2037
2038 PCRE_BSR_ANYCRLF
2039 PCRE_BSR_UNICODE
2040
2041 These options (which are mutually exclusive) control what the \R escape
2042 sequence matches. The choice is either to match only CR, LF, or CRLF,
2043 or to match any Unicode newline sequence. These options override the
2044 choice that was made or defaulted when the pattern was compiled.
2045
2046 PCRE_NEWLINE_CR
2047 PCRE_NEWLINE_LF
2048 PCRE_NEWLINE_CRLF
2049 PCRE_NEWLINE_ANYCRLF
2050 PCRE_NEWLINE_ANY
2051
2052 These options override the newline definition that was chosen or
2053 defaulted when the pattern was compiled. For details, see the descrip-
2054 tion of pcre_compile() above. During matching, the newline choice
2055 affects the behaviour of the dot, circumflex, and dollar metacharac-
2056 ters. It may also alter the way the match position is advanced after a
2057 match failure for an unanchored pattern.
2058
2059 When PCRE_NEWLINE_CRLF, PCRE_NEWLINE_ANYCRLF, or PCRE_NEWLINE_ANY is
2060 set, and a match attempt for an unanchored pattern fails when the cur-
2061 rent position is at a CRLF sequence, and the pattern contains no
2062 explicit matches for CR or LF characters, the match position is
2063 advanced by two characters instead of one, in other words, to after the
2064 CRLF.
2065
2066 The above rule is a compromise that makes the most common cases work as
2067 expected. For example, if the pattern is .+A (and the PCRE_DOTALL
2068 option is not set), it does not match the string "\r\nA" because, after
2069 failing at the start, it skips both the CR and the LF before retrying.
2070 However, the pattern [\r\n]A does match that string, because it con-
2071 tains an explicit CR or LF reference, and so advances only by one char-
2072 acter after the first failure.
2073
2074 An explicit match for CR of LF is either a literal appearance of one of
2075 those characters, or one of the \r or \n escape sequences. Implicit
2076 matches such as [^X] do not count, nor does \s (which includes CR and
2077 LF in the characters that it matches).
2078
2079 Notwithstanding the above, anomalous effects may still occur when CRLF
2080 is a valid newline sequence and explicit \r or \n escapes appear in the
2081 pattern.
2082
2083 PCRE_NOTBOL
2084
2085 This option specifies that first character of the subject string is not
2086 the beginning of a line, so the circumflex metacharacter should not
2087 match before it. Setting this without PCRE_MULTILINE (at compile time)
2088 causes circumflex never to match. This option affects only the behav-
2089 iour of the circumflex metacharacter. It does not affect \A.
2090
2091 PCRE_NOTEOL
2092
2093 This option specifies that the end of the subject string is not the end
2094 of a line, so the dollar metacharacter should not match it nor (except
2095 in multiline mode) a newline immediately before it. Setting this with-
2096 out PCRE_MULTILINE (at compile time) causes dollar never to match. This
2097 option affects only the behaviour of the dollar metacharacter. It does
2098 not affect \Z or \z.
2099
2100 PCRE_NOTEMPTY
2101
2102 An empty string is not considered to be a valid match if this option is
2103 set. If there are alternatives in the pattern, they are tried. If all
2104 the alternatives match the empty string, the entire match fails. For
2105 example, if the pattern
2106
2107 a?b?
2108
2109 is applied to a string not beginning with "a" or "b", it matches an
2110 empty string at the start of the subject. With PCRE_NOTEMPTY set, this
2111 match is not valid, so PCRE searches further into the string for occur-
2112 rences of "a" or "b".
2113
2114 PCRE_NOTEMPTY_ATSTART
2115
2116 This is like PCRE_NOTEMPTY, except that an empty string match that is
2117 not at the start of the subject is permitted. If the pattern is
2118 anchored, such a match can occur only if the pattern contains \K.
2119
2120 Perl has no direct equivalent of PCRE_NOTEMPTY or
2121 PCRE_NOTEMPTY_ATSTART, but it does make a special case of a pattern
2122 match of the empty string within its split() function, and when using
2123 the /g modifier. It is possible to emulate Perl's behaviour after
2124 matching a null string by first trying the match again at the same off-
2125 set with PCRE_NOTEMPTY_ATSTART and PCRE_ANCHORED, and then if that
2126 fails, by advancing the starting offset (see below) and trying an ordi-
2127 nary match again. There is some code that demonstrates how to do this
2128 in the pcredemo sample program. In the most general case, you have to
2129 check to see if the newline convention recognizes CRLF as a newline,
2130 and if so, and the current character is CR followed by LF, advance the
2131 starting offset by two characters instead of one.
2132
2133 PCRE_NO_START_OPTIMIZE
2134
2135 There are a number of optimizations that pcre_exec() uses at the start
2136 of a match, in order to speed up the process. For example, if it is
2137 known that an unanchored match must start with a specific character, it
2138 searches the subject for that character, and fails immediately if it
2139 cannot find it, without actually running the main matching function.
2140 This means that a special item such as (*COMMIT) at the start of a pat-
2141 tern is not considered until after a suitable starting point for the
2142 match has been found. When callouts or (*MARK) items are in use, these
2143 "start-up" optimizations can cause them to be skipped if the pattern is
2144 never actually used. The start-up optimizations are in effect a pre-
2145 scan of the subject that takes place before the pattern is run.
2146
2147 The PCRE_NO_START_OPTIMIZE option disables the start-up optimizations,
2148 possibly causing performance to suffer, but ensuring that in cases
2149 where the result is "no match", the callouts do occur, and that items
2150 such as (*COMMIT) and (*MARK) are considered at every possible starting
2151 position in the subject string. Setting PCRE_NO_START_OPTIMIZE can
2152 change the outcome of a matching operation. Consider the pattern
2153
2154 (*COMMIT)ABC
2155
2156 When this is compiled, PCRE records the fact that a match must start
2157 with the character "A". Suppose the subject string is "DEFABC". The
2158 start-up optimization scans along the subject, finds "A" and runs the
2159 first match attempt from there. The (*COMMIT) item means that the pat-
2160 tern must match the current starting position, which in this case, it
2161 does. However, if the same match is run with PCRE_NO_START_OPTIMIZE
2162 set, the initial scan along the subject string does not happen. The
2163 first match attempt is run starting from "D" and when this fails,
2164 (*COMMIT) prevents any further matches being tried, so the overall
2165 result is "no match". If the pattern is studied, more start-up opti-
2166 mizations may be used. For example, a minimum length for the subject
2167 may be recorded. Consider the pattern
2168
2169 (*MARK:A)(X|Y)
2170
2171 The minimum length for a match is one character. If the subject is
2172 "ABC", there will be attempts to match "ABC", "BC", "C", and then
2173 finally an empty string. If the pattern is studied, the final attempt
2174 does not take place, because PCRE knows that the subject is too short,
2175 and so the (*MARK) is never encountered. In this case, studying the
2176 pattern does not affect the overall match result, which is still "no
2177 match", but it does affect the auxiliary information that is returned.
2178
2179 PCRE_NO_UTF8_CHECK
2180
2181 When PCRE_UTF8 is set at compile time, the validity of the subject as a
2182 UTF-8 string is automatically checked when pcre_exec() is subsequently
2183 called. The value of startoffset is also checked to ensure that it
2184 points to the start of a UTF-8 character. There is a discussion about
2185 the validity of UTF-8 strings in the section on UTF-8 support in the
2186 main pcre page. If an invalid UTF-8 sequence of bytes is found,
2187 pcre_exec() returns the error PCRE_ERROR_BADUTF8. If startoffset con-
2188 tains a value that does not point to the start of a UTF-8 character (or
2189 to the end of the subject), PCRE_ERROR_BADUTF8_OFFSET is returned.
2190
2191 If you already know that your subject is valid, and you want to skip
2192 these checks for performance reasons, you can set the
2193 PCRE_NO_UTF8_CHECK option when calling pcre_exec(). You might want to
2194 do this for the second and subsequent calls to pcre_exec() if you are
2195 making repeated calls to find all the matches in a single subject
2196 string. However, you should be sure that the value of startoffset
2197 points to the start of a UTF-8 character (or the end of the subject).
2198 When PCRE_NO_UTF8_CHECK is set, the effect of passing an invalid UTF-8
2199 string as a subject or an invalid value of startoffset is undefined.
2200 Your program may crash.
2201
2202 PCRE_PARTIAL_HARD
2203 PCRE_PARTIAL_SOFT
2204
2205 These options turn on the partial matching feature. For backwards com-
2206 patibility, PCRE_PARTIAL is a synonym for PCRE_PARTIAL_SOFT. A partial
2207 match occurs if the end of the subject string is reached successfully,
2208 but there are not enough subject characters to complete the match. If
2209 this happens when PCRE_PARTIAL_SOFT (but not PCRE_PARTIAL_HARD) is set,
2210 matching continues by testing any remaining alternatives. Only if no
2211 complete match can be found is PCRE_ERROR_PARTIAL returned instead of
2212 PCRE_ERROR_NOMATCH. In other words, PCRE_PARTIAL_SOFT says that the
2213 caller is prepared to handle a partial match, but only if no complete
2214 match can be found.
2215
2216 If PCRE_PARTIAL_HARD is set, it overrides PCRE_PARTIAL_SOFT. In this
2217 case, if a partial match is found, pcre_exec() immediately returns
2218 PCRE_ERROR_PARTIAL, without considering any other alternatives. In
2219 other words, when PCRE_PARTIAL_HARD is set, a partial match is consid-
2220 ered to be more important that an alternative complete match.
2221
2222 In both cases, the portion of the string that was inspected when the
2223 partial match was found is set as the first matching string. There is a
2224 more detailed discussion of partial and multi-segment matching, with
2225 examples, in the pcrepartial documentation.
2226
2227 The string to be matched by pcre_exec()
2228
2229 The subject string is passed to pcre_exec() as a pointer in subject, a
2230 length (in bytes) in length, and a starting byte offset in startoffset.
2231 If this is negative or greater than the length of the subject,
2232 pcre_exec() returns PCRE_ERROR_BADOFFSET.
2233
2234 In UTF-8 mode, the byte offset must point to the start of a UTF-8 char-
2235 acter (or the end of the subject). Unlike the pattern string, the sub-
2236 ject may contain binary zero bytes. When the starting offset is zero,
2237 the search for a match starts at the beginning of the subject, and this
2238 is by far the most common case.
2239
2240 A non-zero starting offset is useful when searching for another match
2241 in the same subject by calling pcre_exec() again after a previous suc-
2242 cess. Setting startoffset differs from just passing over a shortened
2243 string and setting PCRE_NOTBOL in the case of a pattern that begins
2244 with any kind of lookbehind. For example, consider the pattern
2245
2246 \Biss\B
2247
2248 which finds occurrences of "iss" in the middle of words. (\B matches
2249 only if the current position in the subject is not a word boundary.)
2250 When applied to the string "Mississipi" the first call to pcre_exec()
2251 finds the first occurrence. If pcre_exec() is called again with just
2252 the remainder of the subject, namely "issipi", it does not match,
2253 because \B is always false at the start of the subject, which is deemed
2254 to be a word boundary. However, if pcre_exec() is passed the entire
2255 string again, but with startoffset set to 4, it finds the second occur-
2256 rence of "iss" because it is able to look behind the starting point to
2257 discover that it is preceded by a letter.
2258
2259 Finding all the matches in a subject is tricky when the pattern can
2260 match an empty string. It is possible to emulate Perl's /g behaviour by
2261 first trying the match again at the same offset, with the
2262 PCRE_NOTEMPTY_ATSTART and PCRE_ANCHORED options, and then if that
2263 fails, advancing the starting offset and trying an ordinary match
2264 again. There is some code that demonstrates how to do this in the pcre-
2265 demo sample program. In the most general case, you have to check to see
2266 if the newline convention recognizes CRLF as a newline, and if so, and
2267 the current character is CR followed by LF, advance the starting offset
2268 by two characters instead of one.
2269
2270 If a non-zero starting offset is passed when the pattern is anchored,
2271 one attempt to match at the given offset is made. This can only succeed
2272 if the pattern does not require the match to be at the start of the
2273 subject.
2274
2275 How pcre_exec() returns captured substrings
2276
2277 In general, a pattern matches a certain portion of the subject, and in
2278 addition, further substrings from the subject may be picked out by
2279 parts of the pattern. Following the usage in Jeffrey Friedl's book,
2280 this is called "capturing" in what follows, and the phrase "capturing
2281 subpattern" is used for a fragment of a pattern that picks out a sub-
2282 string. PCRE supports several other kinds of parenthesized subpattern
2283 that do not cause substrings to be captured.
2284
2285 Captured substrings are returned to the caller via a vector of integers
2286 whose address is passed in ovector. The number of elements in the vec-
2287 tor is passed in ovecsize, which must be a non-negative number. Note:
2288 this argument is NOT the size of ovector in bytes.
2289
2290 The first two-thirds of the vector is used to pass back captured sub-
2291 strings, each substring using a pair of integers. The remaining third
2292 of the vector is used as workspace by pcre_exec() while matching cap-
2293 turing subpatterns, and is not available for passing back information.
2294 The number passed in ovecsize should always be a multiple of three. If
2295 it is not, it is rounded down.
2296
2297 When a match is successful, information about captured substrings is
2298 returned in pairs of integers, starting at the beginning of ovector,
2299 and continuing up to two-thirds of its length at the most. The first
2300 element of each pair is set to the byte offset of the first character
2301 in a substring, and the second is set to the byte offset of the first
2302 character after the end of a substring. Note: these values are always
2303 byte offsets, even in UTF-8 mode. They are not character counts.
2304
2305 The first pair of integers, ovector[0] and ovector[1], identify the
2306 portion of the subject string matched by the entire pattern. The next
2307 pair is used for the first capturing subpattern, and so on. The value
2308 returned by pcre_exec() is one more than the highest numbered pair that
2309 has been set. For example, if two substrings have been captured, the
2310 returned value is 3. If there are no capturing subpatterns, the return
2311 value from a successful match is 1, indicating that just the first pair
2312 of offsets has been set.
2313
2314 If a capturing subpattern is matched repeatedly, it is the last portion
2315 of the string that it matched that is returned.
2316
2317 If the vector is too small to hold all the captured substring offsets,
2318 it is used as far as possible (up to two-thirds of its length), and the
2319 function returns a value of zero. If the substring offsets are not of
2320 interest, pcre_exec() may be called with ovector passed as NULL and
2321 ovecsize as zero. However, if the pattern contains back references and
2322 the ovector is not big enough to remember the related substrings, PCRE
2323 has to get additional memory for use during matching. Thus it is usu-
2324 ally advisable to supply an ovector.
2325
2326 The pcre_fullinfo() function can be used to find out how many capturing
2327 subpatterns there are in a compiled pattern. The smallest size for
2328 ovector that will allow for n captured substrings, in addition to the
2329 offsets of the substring matched by the whole pattern, is (n+1)*3.
2330
2331 It is possible for capturing subpattern number n+1 to match some part
2332 of the subject when subpattern n has not been used at all. For example,
2333 if the string "abc" is matched against the pattern (a|(z))(bc) the
2334 return from the function is 4, and subpatterns 1 and 3 are matched, but
2335 2 is not. When this happens, both values in the offset pairs corre-
2336 sponding to unused subpatterns are set to -1.
2337
2338 Offset values that correspond to unused subpatterns at the end of the
2339 expression are also set to -1. For example, if the string "abc" is
2340 matched against the pattern (abc)(x(yz)?)? subpatterns 2 and 3 are not
2341 matched. The return from the function is 2, because the highest used
2342 capturing subpattern number is 1. However, you can refer to the offsets
2343 for the second and third capturing subpatterns if you wish (assuming
2344 the vector is large enough, of course).
2345
2346 Some convenience functions are provided for extracting the captured
2347 substrings as separate strings. These are described below.
2348
2349 Error return values from pcre_exec()
2350
2351 If pcre_exec() fails, it returns a negative number. The following are
2352 defined in the header file:
2353
2354 PCRE_ERROR_NOMATCH (-1)
2355
2356 The subject string did not match the pattern.
2357
2358 PCRE_ERROR_NULL (-2)
2359
2360 Either code or subject was passed as NULL, or ovector was NULL and
2361 ovecsize was not zero.
2362
2363 PCRE_ERROR_BADOPTION (-3)
2364
2365 An unrecognized bit was set in the options argument.
2366
2367 PCRE_ERROR_BADMAGIC (-4)
2368
2369 PCRE stores a 4-byte "magic number" at the start of the compiled code,
2370 to catch the case when it is passed a junk pointer and to detect when a
2371 pattern that was compiled in an environment of one endianness is run in
2372 an environment with the other endianness. This is the error that PCRE
2373 gives when the magic number is not present.
2374
2375 PCRE_ERROR_UNKNOWN_OPCODE (-5)
2376
2377 While running the pattern match, an unknown item was encountered in the
2378 compiled pattern. This error could be caused by a bug in PCRE or by
2379 overwriting of the compiled pattern.
2380
2381 PCRE_ERROR_NOMEMORY (-6)
2382
2383 If a pattern contains back references, but the ovector that is passed
2384 to pcre_exec() is not big enough to remember the referenced substrings,
2385 PCRE gets a block of memory at the start of matching to use for this
2386 purpose. If the call via pcre_malloc() fails, this error is given. The
2387 memory is automatically freed at the end of matching.
2388
2389 This error is also given if pcre_stack_malloc() fails in pcre_exec().
2390 This can happen only when PCRE has been compiled with --disable-stack-
2391 for-recursion.
2392
2393 PCRE_ERROR_NOSUBSTRING (-7)
2394
2395 This error is used by the pcre_copy_substring(), pcre_get_substring(),
2396 and pcre_get_substring_list() functions (see below). It is never
2397 returned by pcre_exec().
2398
2399 PCRE_ERROR_MATCHLIMIT (-8)
2400
2401 The backtracking limit, as specified by the match_limit field in a
2402 pcre_extra structure (or defaulted) was reached. See the description
2403 above.
2404
2405 PCRE_ERROR_CALLOUT (-9)
2406
2407 This error is never generated by pcre_exec() itself. It is provided for
2408 use by callout functions that want to yield a distinctive error code.
2409 See the pcrecallout documentation for details.
2410
2411 PCRE_ERROR_BADUTF8 (-10)
2412
2413 A string that contains an invalid UTF-8 byte sequence was passed as a
2414 subject.
2415
2416 PCRE_ERROR_BADUTF8_OFFSET (-11)
2417
2418 The UTF-8 byte sequence that was passed as a subject was valid, but the
2419 value of startoffset did not point to the beginning of a UTF-8 charac-
2420 ter.
2421
2422 PCRE_ERROR_PARTIAL (-12)
2423
2424 The subject string did not match, but it did match partially. See the
2425 pcrepartial documentation for details of partial matching.
2426
2427 PCRE_ERROR_BADPARTIAL (-13)
2428
2429 This code is no longer in use. It was formerly returned when the
2430 PCRE_PARTIAL option was used with a compiled pattern containing items
2431 that were not supported for partial matching. From release 8.00
2432 onwards, there are no restrictions on partial matching.
2433
2434 PCRE_ERROR_INTERNAL (-14)
2435
2436 An unexpected internal error has occurred. This error could be caused
2437 by a bug in PCRE or by overwriting of the compiled pattern.
2438
2439 PCRE_ERROR_BADCOUNT (-15)
2440
2441 This error is given if the value of the ovecsize argument is negative.
2442
2443 PCRE_ERROR_RECURSIONLIMIT (-21)
2444
2445 The internal recursion limit, as specified by the match_limit_recursion
2446 field in a pcre_extra structure (or defaulted) was reached. See the
2447 description above.
2448
2449 PCRE_ERROR_BADNEWLINE (-23)
2450
2451 An invalid combination of PCRE_NEWLINE_xxx options was given.
2452
2453 PCRE_ERROR_BADOFFSET (-24)
2454
2455 The value of startoffset was negative or greater than the length of the
2456 subject, that is, the value in length.
2457
2458 Error numbers -16 to -20 and -22 are not used by pcre_exec().
2459
2460
2461 EXTRACTING CAPTURED SUBSTRINGS BY NUMBER
2462
2463 int pcre_copy_substring(const char *subject, int *ovector,
2464 int stringcount, int stringnumber, char *buffer,
2465 int buffersize);
2466
2467 int pcre_get_substring(const char *subject, int *ovector,
2468 int stringcount, int stringnumber,
2469 const char **stringptr);
2470
2471 int pcre_get_substring_list(const char *subject,
2472 int *ovector, int stringcount, const char ***listptr);
2473
2474 Captured substrings can be accessed directly by using the offsets
2475 returned by pcre_exec() in ovector. For convenience, the functions
2476 pcre_copy_substring(), pcre_get_substring(), and pcre_get_sub-
2477 string_list() are provided for extracting captured substrings as new,
2478 separate, zero-terminated strings. These functions identify substrings
2479 by number. The next section describes functions for extracting named
2480 substrings.
2481
2482 A substring that contains a binary zero is correctly extracted and has
2483 a further zero added on the end, but the result is not, of course, a C
2484 string. However, you can process such a string by referring to the
2485 length that is returned by pcre_copy_substring() and pcre_get_sub-
2486 string(). Unfortunately, the interface to pcre_get_substring_list() is
2487 not adequate for handling strings containing binary zeros, because the
2488 end of the final string is not independently indicated.
2489
2490 The first three arguments are the same for all three of these func-
2491 tions: subject is the subject string that has just been successfully
2492 matched, ovector is a pointer to the vector of integer offsets that was
2493 passed to pcre_exec(), and stringcount is the number of substrings that
2494 were captured by the match, including the substring that matched the
2495 entire regular expression. This is the value returned by pcre_exec() if
2496 it is greater than zero. If pcre_exec() returned zero, indicating that
2497 it ran out of space in ovector, the value passed as stringcount should
2498 be the number of elements in the vector divided by three.
2499
2500 The functions pcre_copy_substring() and pcre_get_substring() extract a
2501 single substring, whose number is given as stringnumber. A value of
2502 zero extracts the substring that matched the entire pattern, whereas
2503 higher values extract the captured substrings. For pcre_copy_sub-
2504 string(), the string is placed in buffer, whose length is given by
2505 buffersize, while for pcre_get_substring() a new block of memory is
2506 obtained via pcre_malloc, and its address is returned via stringptr.
2507 The yield of the function is the length of the string, not including
2508 the terminating zero, or one of these error codes:
2509
2510 PCRE_ERROR_NOMEMORY (-6)
2511
2512 The buffer was too small for pcre_copy_substring(), or the attempt to
2513 get memory failed for pcre_get_substring().
2514
2515 PCRE_ERROR_NOSUBSTRING (-7)
2516
2517 There is no substring whose number is stringnumber.
2518
2519 The pcre_get_substring_list() function extracts all available sub-
2520 strings and builds a list of pointers to them. All this is done in a
2521 single block of memory that is obtained via pcre_malloc. The address of
2522 the memory block is returned via listptr, which is also the start of
2523 the list of string pointers. The end of the list is marked by a NULL
2524 pointer. The yield of the function is zero if all went well, or the
2525 error code
2526
2527 PCRE_ERROR_NOMEMORY (-6)
2528
2529 if the attempt to get the memory block failed.
2530
2531 When any of these functions encounter a substring that is unset, which
2532 can happen when capturing subpattern number n+1 matches some part of
2533 the subject, but subpattern n has not been used at all, they return an
2534 empty string. This can be distinguished from a genuine zero-length sub-
2535 string by inspecting the appropriate offset in ovector, which is nega-
2536 tive for unset substrings.
2537
2538 The two convenience functions pcre_free_substring() and pcre_free_sub-
2539 string_list() can be used to free the memory returned by a previous
2540 call of pcre_get_substring() or pcre_get_substring_list(), respec-
2541 tively. They do nothing more than call the function pointed to by
2542 pcre_free, which of course could be called directly from a C program.
2543 However, PCRE is used in some situations where it is linked via a spe-
2544 cial interface to another programming language that cannot use
2545 pcre_free directly; it is for these cases that the functions are pro-
2546 vided.
2547
2548
2549 EXTRACTING CAPTURED SUBSTRINGS BY NAME
2550
2551 int pcre_get_stringnumber(const pcre *code,
2552 const char *name);
2553
2554 int pcre_copy_named_substring(const pcre *code,
2555 const char *subject, int *ovector,
2556 int stringcount, const char *stringname,
2557 char *buffer, int buffersize);
2558
2559 int pcre_get_named_substring(const pcre *code,
2560 const char *subject, int *ovector,
2561 int stringcount, const char *stringname,
2562 const char **stringptr);
2563
2564 To extract a substring by name, you first have to find associated num-
2565 ber. For example, for this pattern
2566
2567 (a+)b(?<xxx>\d+)...
2568
2569 the number of the subpattern called "xxx" is 2. If the name is known to
2570 be unique (PCRE_DUPNAMES was not set), you can find the number from the
2571 name by calling pcre_get_stringnumber(). The first argument is the com-
2572 piled pattern, and the second is the name. The yield of the function is
2573 the subpattern number, or PCRE_ERROR_NOSUBSTRING (-7) if there is no
2574 subpattern of that name.
2575
2576 Given the number, you can extract the substring directly, or use one of
2577 the functions described in the previous section. For convenience, there
2578 are also two functions that do the whole job.
2579
2580 Most of the arguments of pcre_copy_named_substring() and
2581 pcre_get_named_substring() are the same as those for the similarly
2582 named functions that extract by number. As these are described in the
2583 previous section, they are not re-described here. There are just two
2584 differences:
2585
2586 First, instead of a substring number, a substring name is given. Sec-
2587 ond, there is an extra argument, given at the start, which is a pointer
2588 to the compiled pattern. This is needed in order to gain access to the
2589 name-to-number translation table.
2590
2591 These functions call pcre_get_stringnumber(), and if it succeeds, they
2592 then call pcre_copy_substring() or pcre_get_substring(), as appropri-
2593 ate. NOTE: If PCRE_DUPNAMES is set and there are duplicate names, the
2594 behaviour may not be what you want (see the next section).
2595
2596 Warning: If the pattern uses the (?| feature to set up multiple subpat-
2597 terns with the same number, as described in the section on duplicate
2598 subpattern numbers in the pcrepattern page, you cannot use names to
2599 distinguish the different subpatterns, because names are not included
2600 in the compiled code. The matching process uses only numbers. For this
2601 reason, the use of different names for subpatterns of the same number
2602 causes an error at compile time.
2603
2604
2605 DUPLICATE SUBPATTERN NAMES
2606
2607 int pcre_get_stringtable_entries(const pcre *code,
2608 const char *name, char **first, char **last);
2609
2610 When a pattern is compiled with the PCRE_DUPNAMES option, names for
2611 subpatterns are not required to be unique. (Duplicate names are always
2612 allowed for subpatterns with the same number, created by using the (?|
2613 feature. Indeed, if such subpatterns are named, they are required to
2614 use the same names.)
2615
2616 Normally, patterns with duplicate names are such that in any one match,
2617 only one of the named subpatterns participates. An example is shown in
2618 the pcrepattern documentation.
2619
2620 When duplicates are present, pcre_copy_named_substring() and
2621 pcre_get_named_substring() return the first substring corresponding to
2622 the given name that is set. If none are set, PCRE_ERROR_NOSUBSTRING
2623 (-7) is returned; no data is returned. The pcre_get_stringnumber()
2624 function returns one of the numbers that are associated with the name,
2625 but it is not defined which it is.
2626
2627 If you want to get full details of all captured substrings for a given
2628 name, you must use the pcre_get_stringtable_entries() function. The
2629 first argument is the compiled pattern, and the second is the name. The
2630 third and fourth are pointers to variables which are updated by the
2631 function. After it has run, they point to the first and last entries in
2632 the name-to-number table for the given name. The function itself
2633 returns the length of each entry, or PCRE_ERROR_NOSUBSTRING (-7) if
2634 there are none. The format of the table is described above in the sec-
2635 tion entitled Information about a pattern. Given all the relevant
2636 entries for the name, you can extract each of their numbers, and hence
2637 the captured data, if any.
2638
2639
2640 FINDING ALL POSSIBLE MATCHES
2641
2642 The traditional matching function uses a similar algorithm to Perl,
2643 which stops when it finds the first match, starting at a given point in
2644 the subject. If you want to find all possible matches, or the longest
2645 possible match, consider using the alternative matching function (see
2646 below) instead. If you cannot use the alternative function, but still
2647 need to find all possible matches, you can kludge it up by making use
2648 of the callout facility, which is described in the pcrecallout documen-
2649 tation.
2650
2651 What you have to do is to insert a callout right at the end of the pat-
2652 tern. When your callout function is called, extract and save the cur-
2653 rent matched substring. Then return 1, which forces pcre_exec() to
2654 backtrack and try other alternatives. Ultimately, when it runs out of
2655 matches, pcre_exec() will yield PCRE_ERROR_NOMATCH.
2656
2657
2658 MATCHING A PATTERN: THE ALTERNATIVE FUNCTION
2659
2660 int pcre_dfa_exec(const pcre *code, const pcre_extra *extra,
2661 const char *subject, int length, int startoffset,
2662 int options, int *ovector, int ovecsize,
2663 int *workspace, int wscount);
2664
2665 The function pcre_dfa_exec() is called to match a subject string
2666 against a compiled pattern, using a matching algorithm that scans the
2667 subject string just once, and does not backtrack. This has different
2668 characteristics to the normal algorithm, and is not compatible with
2669 Perl. Some of the features of PCRE patterns are not supported. Never-
2670 theless, there are times when this kind of matching can be useful. For
2671 a discussion of the two matching algorithms, and a list of features
2672 that pcre_dfa_exec() does not support, see the pcrematching documenta-
2673 tion.
2674
2675 The arguments for the pcre_dfa_exec() function are the same as for
2676 pcre_exec(), plus two extras. The ovector argument is used in a differ-
2677 ent way, and this is described below. The other common arguments are
2678 used in the same way as for pcre_exec(), so their description is not
2679 repeated here.
2680
2681 The two additional arguments provide workspace for the function. The
2682 workspace vector should contain at least 20 elements. It is used for
2683 keeping track of multiple paths through the pattern tree. More
2684 workspace will be needed for patterns and subjects where there are a
2685 lot of potential matches.
2686
2687 Here is an example of a simple call to pcre_dfa_exec():
2688
2689 int rc;
2690 int ovector[10];
2691 int wspace[20];
2692 rc = pcre_dfa_exec(
2693 re, /* result of pcre_compile() */
2694 NULL, /* we didn't study the pattern */
2695 "some string", /* the subject string */
2696 11, /* the length of the subject string */
2697 0, /* start at offset 0 in the subject */
2698 0, /* default options */
2699 ovector, /* vector of integers for substring information */
2700 10, /* number of elements (NOT size in bytes) */
2701 wspace, /* working space vector */
2702 20); /* number of elements (NOT size in bytes) */
2703
2704 Option bits for pcre_dfa_exec()
2705
2706 The unused bits of the options argument for pcre_dfa_exec() must be
2707 zero. The only bits that may be set are PCRE_ANCHORED, PCRE_NEW-
2708 LINE_xxx, PCRE_NOTBOL, PCRE_NOTEOL, PCRE_NOTEMPTY,
2709 PCRE_NOTEMPTY_ATSTART, PCRE_NO_UTF8_CHECK, PCRE_BSR_ANYCRLF,
2710 PCRE_BSR_UNICODE, PCRE_NO_START_OPTIMIZE, PCRE_PARTIAL_HARD, PCRE_PAR-
2711 TIAL_SOFT, PCRE_DFA_SHORTEST, and PCRE_DFA_RESTART. All but the last
2712 four of these are exactly the same as for pcre_exec(), so their
2713 description is not repeated here.
2714
2715 PCRE_PARTIAL_HARD
2716 PCRE_PARTIAL_SOFT
2717
2718 These have the same general effect as they do for pcre_exec(), but the
2719 details are slightly different. When PCRE_PARTIAL_HARD is set for
2720 pcre_dfa_exec(), it returns PCRE_ERROR_PARTIAL if the end of the sub-
2721 ject is reached and there is still at least one matching possibility
2722 that requires additional characters. This happens even if some complete
2723 matches have also been found. When PCRE_PARTIAL_SOFT is set, the return
2724 code PCRE_ERROR_NOMATCH is converted into PCRE_ERROR_PARTIAL if the end
2725 of the subject is reached, there have been no complete matches, but
2726 there is still at least one matching possibility. The portion of the
2727 string that was inspected when the longest partial match was found is
2728 set as the first matching string in both cases. There is a more
2729 detailed discussion of partial and multi-segment matching, with exam-
2730 ples, in the pcrepartial documentation.
2731
2732 PCRE_DFA_SHORTEST
2733
2734 Setting the PCRE_DFA_SHORTEST option causes the matching algorithm to
2735 stop as soon as it has found one match. Because of the way the alterna-
2736 tive algorithm works, this is necessarily the shortest possible match
2737 at the first possible matching point in the subject string.
2738
2739 PCRE_DFA_RESTART
2740
2741 When pcre_dfa_exec() returns a partial match, it is possible to call it
2742 again, with additional subject characters, and have it continue with
2743 the same match. The PCRE_DFA_RESTART option requests this action; when
2744 it is set, the workspace and wscount options must reference the same
2745 vector as before because data about the match so far is left in them
2746 after a partial match. There is more discussion of this facility in the
2747 pcrepartial documentation.
2748
2749 Successful returns from pcre_dfa_exec()
2750
2751 When pcre_dfa_exec() succeeds, it may have matched more than one sub-
2752 string in the subject. Note, however, that all the matches from one run
2753 of the function start at the same point in the subject. The shorter
2754 matches are all initial substrings of the longer matches. For example,
2755 if the pattern
2756
2757 <.*>
2758
2759 is matched against the string
2760
2761 This is <something> <something else> <something further> no more
2762
2763 the three matched strings are
2764
2765 <something>
2766 <something> <something else>
2767 <something> <something else> <something further>
2768
2769 On success, the yield of the function is a number greater than zero,
2770 which is the number of matched substrings. The substrings themselves
2771 are returned in ovector. Each string uses two elements; the first is
2772 the offset to the start, and the second is the offset to the end. In
2773 fact, all the strings have the same start offset. (Space could have
2774 been saved by giving this only once, but it was decided to retain some
2775 compatibility with the way pcre_exec() returns data, even though the
2776 meaning of the strings is different.)
2777
2778 The strings are returned in reverse order of length; that is, the long-
2779 est matching string is given first. If there were too many matches to
2780 fit into ovector, the yield of the function is zero, and the vector is
2781 filled with the longest matches.
2782
2783 Error returns from pcre_dfa_exec()
2784
2785 The pcre_dfa_exec() function returns a negative number when it fails.
2786 Many of the errors are the same as for pcre_exec(), and these are
2787 described above. There are in addition the following errors that are
2788 specific to pcre_dfa_exec():
2789
2790 PCRE_ERROR_DFA_UITEM (-16)
2791
2792 This return is given if pcre_dfa_exec() encounters an item in the pat-
2793 tern that it does not support, for instance, the use of \C or a back
2794 reference.
2795
2796 PCRE_ERROR_DFA_UCOND (-17)
2797
2798 This return is given if pcre_dfa_exec() encounters a condition item
2799 that uses a back reference for the condition, or a test for recursion
2800 in a specific group. These are not supported.
2801
2802 PCRE_ERROR_DFA_UMLIMIT (-18)
2803
2804 This return is given if pcre_dfa_exec() is called with an extra block
2805 that contains a setting of the match_limit field. This is not supported
2806 (it is meaningless).
2807
2808 PCRE_ERROR_DFA_WSSIZE (-19)
2809
2810 This return is given if pcre_dfa_exec() runs out of space in the
2811 workspace vector.
2812
2813 PCRE_ERROR_DFA_RECURSE (-20)
2814
2815 When a recursive subpattern is processed, the matching function calls
2816 itself recursively, using private vectors for ovector and workspace.
2817 This error is given if the output vector is not large enough. This
2818 should be extremely rare, as a vector of size 1000 is used.
2819
2820
2821 SEE ALSO
2822
2823 pcrebuild(3), pcrecallout(3), pcrecpp(3)(3), pcrematching(3), pcrepar-
2824 tial(3), pcreposix(3), pcreprecompile(3), pcresample(3), pcrestack(3).
2825
2826
2827 AUTHOR
2828
2829 Philip Hazel
2830 University Computing Service
2831 Cambridge CB2 3QH, England.
2832
2833
2834 REVISION
2835
2836 Last updated: 06 November 2010
2837 Copyright (c) 1997-2010 University of Cambridge.
2838 ------------------------------------------------------------------------------
2839
2840
2841 PCRECALLOUT(3) PCRECALLOUT(3)
2842
2843
2844 NAME
2845 PCRE - Perl-compatible regular expressions
2846
2847
2848 PCRE CALLOUTS
2849
2850 int (*pcre_callout)(pcre_callout_block *);
2851
2852 PCRE provides a feature called "callout", which is a means of temporar-
2853 ily passing control to the caller of PCRE in the middle of pattern
2854 matching. The caller of PCRE provides an external function by putting
2855 its entry point in the global variable pcre_callout. By default, this
2856 variable contains NULL, which disables all calling out.
2857
2858 Within a regular expression, (?C) indicates the points at which the
2859 external function is to be called. Different callout points can be
2860 identified by putting a number less than 256 after the letter C. The
2861 default value is zero. For example, this pattern has two callout
2862 points:
2863
2864 (?C1)abc(?C2)def
2865
2866 If the PCRE_AUTO_CALLOUT option bit is set when pcre_compile() or
2867 pcre_compile2() is called, PCRE automatically inserts callouts, all
2868 with number 255, before each item in the pattern. For example, if
2869 PCRE_AUTO_CALLOUT is used with the pattern
2870
2871 A(\d{2}|--)
2872
2873 it is processed as if it were
2874
2875 (?C255)A(?C255)((?C255)\d{2}(?C255)|(?C255)-(?C255)-(?C255))(?C255)
2876
2877 Notice that there is a callout before and after each parenthesis and
2878 alternation bar. Automatic callouts can be used for tracking the
2879 progress of pattern matching. The pcretest command has an option that
2880 sets automatic callouts; when it is used, the output indicates how the
2881 pattern is matched. This is useful information when you are trying to
2882 optimize the performance of a particular pattern.
2883
2884
2885 MISSING CALLOUTS
2886
2887 You should be aware that, because of optimizations in the way PCRE
2888 matches patterns by default, callouts sometimes do not happen. For
2889 example, if the pattern is
2890
2891 ab(?C4)cd
2892
2893 PCRE knows that any matching string must contain the letter "d". If the
2894 subject string is "abyz", the lack of "d" means that matching doesn't
2895 ever start, and the callout is never reached. However, with "abyd",
2896 though the result is still no match, the callout is obeyed.
2897
2898 If the pattern is studied, PCRE knows the minimum length of a matching
2899 string, and will immediately give a "no match" return without actually
2900 running a match if the subject is not long enough, or, for unanchored
2901 patterns, if it has been scanned far enough.
2902
2903 You can disable these optimizations by passing the PCRE_NO_START_OPTI-
2904 MIZE option to pcre_exec() or pcre_dfa_exec(). This slows down the
2905 matching process, but does ensure that callouts such as the example
2906 above are obeyed.
2907
2908
2909 THE CALLOUT INTERFACE
2910
2911 During matching, when PCRE reaches a callout point, the external func-
2912 tion defined by pcre_callout is called (if it is set). This applies to
2913 both the pcre_exec() and the pcre_dfa_exec() matching functions. The
2914 only argument to the callout function is a pointer to a pcre_callout
2915 block. This structure contains the following fields:
2916
2917 int version;
2918 int callout_number;
2919 int *offset_vector;
2920 const char *subject;
2921 int subject_length;
2922 int start_match;
2923 int current_position;
2924 int capture_top;
2925 int capture_last;
2926 void *callout_data;
2927 int pattern_position;
2928 int next_item_length;
2929
2930 The version field is an integer containing the version number of the
2931 block format. The initial version was 0; the current version is 1. The
2932 version number will change again in future if additional fields are
2933 added, but the intention is never to remove any of the existing fields.
2934
2935 The callout_number field contains the number of the callout, as com-
2936 piled into the pattern (that is, the number after ?C for manual call-
2937 outs, and 255 for automatically generated callouts).
2938
2939 The offset_vector field is a pointer to the vector of offsets that was
2940 passed by the caller to pcre_exec() or pcre_dfa_exec(). When
2941 pcre_exec() is used, the contents can be inspected in order to extract
2942 substrings that have been matched so far, in the same way as for
2943 extracting substrings after a match has completed. For pcre_dfa_exec()
2944 this field is not useful.
2945
2946 The subject and subject_length fields contain copies of the values that
2947 were passed to pcre_exec().
2948
2949 The start_match field normally contains the offset within the subject
2950 at which the current match attempt started. However, if the escape
2951 sequence \K has been encountered, this value is changed to reflect the
2952 modified starting point. If the pattern is not anchored, the callout
2953 function may be called several times from the same point in the pattern
2954 for different starting points in the subject.
2955
2956 The current_position field contains the offset within the subject of
2957 the current match pointer.
2958
2959 When the pcre_exec() function is used, the capture_top field contains
2960 one more than the number of the highest numbered captured substring so
2961 far. If no substrings have been captured, the value of capture_top is
2962 one. This is always the case when pcre_dfa_exec() is used, because it
2963 does not support captured substrings.
2964
2965 The capture_last field contains the number of the most recently cap-
2966 tured substring. If no substrings have been captured, its value is -1.
2967 This is always the case when pcre_dfa_exec() is used.
2968
2969 The callout_data field contains a value that is passed to pcre_exec()
2970 or pcre_dfa_exec() specifically so that it can be passed back in call-
2971 outs. It is passed in the pcre_callout field of the pcre_extra data
2972 structure. If no such data was passed, the value of callout_data in a
2973 pcre_callout block is NULL. There is a description of the pcre_extra
2974 structure in the pcreapi documentation.
2975
2976 The pattern_position field is present from version 1 of the pcre_call-
2977 out structure. It contains the offset to the next item to be matched in
2978 the pattern string.
2979
2980 The next_item_length field is present from version 1 of the pcre_call-
2981 out structure. It contains the length of the next item to be matched in
2982 the pattern string. When the callout immediately precedes an alterna-
2983 tion bar, a closing parenthesis, or the end of the pattern, the length
2984 is zero. When the callout precedes an opening parenthesis, the length
2985 is that of the entire subpattern.
2986
2987 The pattern_position and next_item_length fields are intended to help
2988 in distinguishing between different automatic callouts, which all have
2989 the same callout number. However, they are set for all callouts.
2990
2991
2992 RETURN VALUES
2993
2994 The external callout function returns an integer to PCRE. If the value
2995 is zero, matching proceeds as normal. If the value is greater than
2996 zero, matching fails at the current point, but the testing of other
2997 matching possibilities goes ahead, just as if a lookahead assertion had
2998 failed. If the value is less than zero, the match is abandoned, and
2999 pcre_exec() or pcre_dfa_exec() returns the negative value.
3000
3001 Negative values should normally be chosen from the set of
3002 PCRE_ERROR_xxx values. In particular, PCRE_ERROR_NOMATCH forces a stan-
3003 dard "no match" failure. The error number PCRE_ERROR_CALLOUT is
3004 reserved for use by callout functions; it will never be used by PCRE
3005 itself.
3006
3007
3008 AUTHOR
3009
3010 Philip Hazel
3011 University Computing Service
3012 Cambridge CB2 3QH, England.
3013
3014
3015 REVISION
3016
3017 Last updated: 29 September 2009
3018 Copyright (c) 1997-2009 University of Cambridge.
3019 ------------------------------------------------------------------------------
3020
3021
3022 PCRECOMPAT(3) PCRECOMPAT(3)
3023
3024
3025 NAME
3026 PCRE - Perl-compatible regular expressions
3027
3028
3029 DIFFERENCES BETWEEN PCRE AND PERL
3030
3031 This document describes the differences in the ways that PCRE and Perl
3032 handle regular expressions. The differences described here are with
3033 respect to Perl versions 5.10 and above.
3034
3035 1. PCRE has only a subset of Perl's UTF-8 and Unicode support. Details
3036 of what it does have are given in the section on UTF-8 support in the
3037 main pcre page.
3038
3039 2. PCRE does not allow repeat quantifiers on lookahead assertions. Perl
3040 permits them, but they do not mean what you might think. For example,
3041 (?!a){3} does not assert that the next three characters are not "a". It
3042 just asserts that the next character is not "a" three times.
3043
3044 3. Capturing subpatterns that occur inside negative lookahead asser-
3045 tions are counted, but their entries in the offsets vector are never
3046 set. Perl sets its numerical variables from any such patterns that are
3047 matched before the assertion fails to match something (thereby succeed-
3048 ing), but only if the negative lookahead assertion contains just one
3049 branch.
3050
3051 4. Though binary zero characters are supported in the subject string,
3052 they are not allowed in a pattern string because it is passed as a nor-
3053 mal C string, terminated by zero. The escape sequence \0 can be used in
3054 the pattern to represent a binary zero.
3055
3056 5. The following Perl escape sequences are not supported: \l, \u, \L,
3057 \U, and \N. In fact these are implemented by Perl's general string-han-
3058 dling and are not part of its pattern matching engine. If any of these
3059 are encountered by PCRE, an error is generated.
3060
3061 6. The Perl escape sequences \p, \P, and \X are supported only if PCRE
3062 is built with Unicode character property support. The properties that
3063 can be tested with \p and \P are limited to the general category prop-
3064 erties such as Lu and Nd, script names such as Greek or Han, and the
3065 derived properties Any and L&. PCRE does support the Cs (surrogate)
3066 property, which Perl does not; the Perl documentation says "Because
3067 Perl hides the need for the user to understand the internal representa-
3068 tion of Unicode characters, there is no need to implement the somewhat
3069 messy concept of surrogates."
3070
3071 7. PCRE does support the \Q...\E escape for quoting substrings. Charac-
3072 ters in between are treated as literals. This is slightly different
3073 from Perl in that $ and @ are also handled as literals inside the
3074 quotes. In Perl, they cause variable interpolation (but of course PCRE
3075 does not have variables). Note the following examples:
3076
3077 Pattern PCRE matches Perl matches
3078
3079 \Qabc$xyz\E abc$xyz abc followed by the
3080 contents of $xyz
3081 \Qabc\$xyz\E abc\$xyz abc\$xyz
3082 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3083
3084 The \Q...\E sequence is recognized both inside and outside character
3085 classes.
3086
3087 8. Fairly obviously, PCRE does not support the (?{code}) and (??{code})
3088 constructions. However, there is support for recursive patterns. This
3089 is not available in Perl 5.8, but it is in Perl 5.10. Also, the PCRE
3090 "callout" feature allows an external function to be called during pat-
3091 tern matching. See the pcrecallout documentation for details.
3092
3093 9. Subpatterns that are called recursively or as "subroutines" are
3094 always treated as atomic groups in PCRE. This is like Python, but
3095 unlike Perl. There is a discussion of an example that explains this in
3096 more detail in the section on recursion differences from Perl in the
3097 pcrepattern page.
3098
3099 10. There are some differences that are concerned with the settings of
3100 captured strings when part of a pattern is repeated. For example,
3101 matching "aba" against the pattern /^(a(b)?)+$/ in Perl leaves $2
3102 unset, but in PCRE it is set to "b".
3103
3104 11. PCRE's handling of duplicate subpattern numbers and duplicate sub-
3105 pattern names is not as general as Perl's. This is a consequence of the
3106 fact the PCRE works internally just with numbers, using an external ta-
3107 ble to translate between numbers and names. In particular, a pattern
3108 such as (?|(?<a>A)|(?<b)B), where the two capturing parentheses have
3109 the same number but different names, is not supported, and causes an
3110 error at compile time. If it were allowed, it would not be possible to
3111 distinguish which parentheses matched, because both names map to cap-
3112 turing subpattern number 1. To avoid this confusing situation, an error
3113 is given at compile time.
3114
3115 12. Perl recognizes comments in some places that PCRE doesn't, for
3116 example, between the ( and ? at the start of a subpattern.
3117
3118 13. PCRE provides some extensions to the Perl regular expression facil-
3119 ities. Perl 5.10 includes new features that are not in earlier ver-
3120 sions of Perl, some of which (such as named parentheses) have been in
3121 PCRE for some time. This list is with respect to Perl 5.10:
3122
3123 (a) Although lookbehind assertions in PCRE must match fixed length
3124 strings, each alternative branch of a lookbehind assertion can match a
3125 different length of string. Perl requires them all to have the same
3126 length.
3127
3128 (b) If PCRE_DOLLAR_ENDONLY is set and PCRE_MULTILINE is not set, the $
3129 meta-character matches only at the very end of the string.
3130
3131 (c) If PCRE_EXTRA is set, a backslash followed by a letter with no spe-
3132 cial meaning is faulted. Otherwise, like Perl, the backslash is quietly
3133 ignored. (Perl can be made to issue a warning.)
3134
3135 (d) If PCRE_UNGREEDY is set, the greediness of the repetition quanti-
3136 fiers is inverted, that is, by default they are not greedy, but if fol-
3137 lowed by a question mark they are.
3138
3139 (e) PCRE_ANCHORED can be used at matching time to force a pattern to be
3140 tried only at the first matching position in the subject string.
3141
3142 (f) The PCRE_NOTBOL, PCRE_NOTEOL, PCRE_NOTEMPTY, PCRE_NOTEMPTY_ATSTART,
3143 and PCRE_NO_AUTO_CAPTURE options for pcre_exec() have no Perl equiva-
3144 lents.
3145
3146 (g) The \R escape sequence can be restricted to match only CR, LF, or
3147 CRLF by the PCRE_BSR_ANYCRLF option.
3148
3149 (h) The callout facility is PCRE-specific.
3150
3151 (i) The partial matching facility is PCRE-specific.
3152
3153 (j) Patterns compiled by PCRE can be saved and re-used at a later time,
3154 even on different hosts that have the other endianness.
3155
3156 (k) The alternative matching function (pcre_dfa_exec()) matches in a
3157 different way and is not Perl-compatible.
3158
3159 (l) PCRE recognizes some special sequences such as (*CR) at the start
3160 of a pattern that set overall options that cannot be changed within the
3161 pattern.
3162
3163
3164 AUTHOR
3165
3166 Philip Hazel
3167 University Computing Service
3168 Cambridge CB2 3QH, England.
3169
3170
3171 REVISION
3172
3173 Last updated: 31 October 2010
3174 Copyright (c) 1997-2010 University of Cambridge.
3175 ------------------------------------------------------------------------------
3176
3177
3178 PCREPATTERN(3) PCREPATTERN(3)
3179
3180
3181 NAME
3182 PCRE - Perl-compatible regular expressions
3183
3184
3185 PCRE REGULAR EXPRESSION DETAILS
3186
3187 The syntax and semantics of the regular expressions that are supported
3188 by PCRE are described in detail below. There is a quick-reference syn-
3189 tax summary in the pcresyntax page. PCRE tries to match Perl syntax and
3190 semantics as closely as it can. PCRE also supports some alternative
3191 regular expression syntax (which does not conflict with the Perl syn-
3192 tax) in order to provide some compatibility with regular expressions in
3193 Python, .NET, and Oniguruma.
3194
3195 Perl's regular expressions are described in its own documentation, and
3196 regular expressions in general are covered in a number of books, some
3197 of which have copious examples. Jeffrey Friedl's "Mastering Regular
3198 Expressions", published by O'Reilly, covers regular expressions in
3199 great detail. This description of PCRE's regular expressions is
3200 intended as reference material.
3201
3202 The original operation of PCRE was on strings of one-byte characters.
3203 However, there is now also support for UTF-8 character strings. To use
3204 this, PCRE must be built to include UTF-8 support, and you must call
3205 pcre_compile() or pcre_compile2() with the PCRE_UTF8 option. There is
3206 also a special sequence that can be given at the start of a pattern:
3207
3208 (*UTF8)
3209
3210 Starting a pattern with this sequence is equivalent to setting the
3211 PCRE_UTF8 option. This feature is not Perl-compatible. How setting
3212 UTF-8 mode affects pattern matching is mentioned in several places
3213 below. There is also a summary of UTF-8 features in the section on
3214 UTF-8 support in the main pcre page.
3215
3216 Another special sequence that may appear at the start of a pattern or
3217 in combination with (*UTF8) is:
3218
3219 (*UCP)
3220
3221 This has the same effect as setting the PCRE_UCP option: it causes
3222 sequences such as \d and \w to use Unicode properties to determine
3223 character types, instead of recognizing only characters with codes less
3224 than 128 via a lookup table.
3225
3226 The remainder of this document discusses the patterns that are sup-
3227 ported by PCRE when its main matching function, pcre_exec(), is used.
3228 From release 6.0, PCRE offers a second matching function,
3229 pcre_dfa_exec(), which matches using a different algorithm that is not
3230 Perl-compatible. Some of the features discussed below are not available
3231 when pcre_dfa_exec() is used. The advantages and disadvantages of the
3232 alternative function, and how it differs from the normal function, are
3233 discussed in the pcrematching page.
3234
3235
3236 NEWLINE CONVENTIONS
3237
3238 PCRE supports five different conventions for indicating line breaks in
3239 strings: a single CR (carriage return) character, a single LF (line-
3240 feed) character, the two-character sequence CRLF, any of the three pre-
3241 ceding, or any Unicode newline sequence. The pcreapi page has further
3242 discussion about newlines, and shows how to set the newline convention
3243 in the options arguments for the compiling and matching functions.
3244
3245 It is also possible to specify a newline convention by starting a pat-
3246 tern string with one of the following five sequences:
3247
3248 (*CR) carriage return
3249 (*LF) linefeed
3250 (*CRLF) carriage return, followed by linefeed
3251 (*ANYCRLF) any of the three above
3252 (*ANY) all Unicode newline sequences
3253
3254 These override the default and the options given to pcre_compile() or
3255 pcre_compile2(). For example, on a Unix system where LF is the default
3256 newline sequence, the pattern
3257
3258 (*CR)a.b
3259
3260 changes the convention to CR. That pattern matches "a\nb" because LF is
3261 no longer a newline. Note that these special settings, which are not
3262 Perl-compatible, are recognized only at the very start of a pattern,
3263 and that they must be in upper case. If more than one of them is
3264 present, the last one is used.
3265
3266 The newline convention affects the interpretation of the dot metachar-
3267 acter when PCRE_DOTALL is not set, and also the behaviour of \N. How-
3268 ever, it does not affect what the \R escape sequence matches. By
3269 default, this is any Unicode newline sequence, for Perl compatibility.
3270 However, this can be changed; see the description of \R in the section
3271 entitled "Newline sequences" below. A change of \R setting can be com-
3272 bined with a change of newline convention.
3273
3274
3275 CHARACTERS AND METACHARACTERS
3276
3277 A regular expression is a pattern that is matched against a subject
3278 string from left to right. Most characters stand for themselves in a
3279 pattern, and match the corresponding characters in the subject. As a
3280 trivial example, the pattern
3281
3282 The quick brown fox
3283
3284 matches a portion of a subject string that is identical to itself. When
3285 caseless matching is specified (the PCRE_CASELESS option), letters are
3286 matched independently of case. In UTF-8 mode, PCRE always understands
3287 the concept of case for characters whose values are less than 128, so
3288 caseless matching is always possible. For characters with higher val-
3289 ues, the concept of case is supported if PCRE is compiled with Unicode
3290 property support, but not otherwise. If you want to use caseless
3291 matching for characters 128 and above, you must ensure that PCRE is
3292 compiled with Unicode property support as well as with UTF-8 support.
3293
3294 The power of regular expressions comes from the ability to include
3295 alternatives and repetitions in the pattern. These are encoded in the
3296 pattern by the use of metacharacters, which do not stand for themselves
3297 but instead are interpreted in some special way.
3298
3299 There are two different sets of metacharacters: those that are recog-
3300 nized anywhere in the pattern except within square brackets, and those
3301 that are recognized within square brackets. Outside square brackets,
3302 the metacharacters are as follows:
3303
3304 \ general escape character with several uses
3305 ^ assert start of string (or line, in multiline mode)
3306 $ assert end of string (or line, in multiline mode)
3307 . match any character except newline (by default)
3308 [ start character class definition
3309 | start of alternative branch
3310 ( start subpattern
3311 ) end subpattern
3312 ? extends the meaning of (
3313 also 0 or 1 quantifier
3314 also quantifier minimizer
3315 * 0 or more quantifier
3316 + 1 or more quantifier
3317 also "possessive quantifier"
3318 { start min/max quantifier
3319
3320 Part of a pattern that is in square brackets is called a "character
3321 class". In a character class the only metacharacters are:
3322
3323 \ general escape character
3324 ^ negate the class, but only if the first character
3325 - indicates character range
3326 [ POSIX character class (only if followed by POSIX
3327 syntax)
3328 ] terminates the character class
3329
3330 The following sections describe the use of each of the metacharacters.
3331
3332
3333 BACKSLASH
3334
3335 The backslash character has several uses. Firstly, if it is followed by
3336 a non-alphanumeric character, it takes away any special meaning that
3337 character may have. This use of backslash as an escape character
3338 applies both inside and outside character classes.
3339
3340 For example, if you want to match a * character, you write \* in the
3341 pattern. This escaping action applies whether or not the following
3342 character would otherwise be interpreted as a metacharacter, so it is
3343 always safe to precede a non-alphanumeric with backslash to specify
3344 that it stands for itself. In particular, if you want to match a back-
3345 slash, you write \\.
3346
3347 If a pattern is compiled with the PCRE_EXTENDED option, whitespace in
3348 the pattern (other than in a character class) and characters between a
3349 # outside a character class and the next newline are ignored. An escap-
3350 ing backslash can be used to include a whitespace or # character as
3351 part of the pattern.
3352
3353 If you want to remove the special meaning from a sequence of charac-
3354 ters, you can do so by putting them between \Q and \E. This is differ-
3355 ent from Perl in that $ and @ are handled as literals in \Q...\E
3356 sequences in PCRE, whereas in Perl, $ and @ cause variable interpola-
3357 tion. Note the following examples:
3358
3359 Pattern PCRE matches Perl matches
3360
3361 \Qabc$xyz\E abc$xyz abc followed by the
3362 contents of $xyz
3363 \Qabc\$xyz\E abc\$xyz abc\$xyz
3364 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3365
3366 The \Q...\E sequence is recognized both inside and outside character
3367 classes. An isolated \E that is not preceded by \Q is ignored.
3368
3369 Non-printing characters
3370
3371 A second use of backslash provides a way of encoding non-printing char-
3372 acters in patterns in a visible manner. There is no restriction on the
3373 appearance of non-printing characters, apart from the binary zero that
3374 terminates a pattern, but when a pattern is being prepared by text
3375 editing, it is often easier to use one of the following escape
3376 sequences than the binary character it represents:
3377
3378 \a alarm, that is, the BEL character (hex 07)
3379 \cx "control-x", where x is any character
3380 \e escape (hex 1B)
3381 \f formfeed (hex 0C)
3382 \n linefeed (hex 0A)
3383 \r carriage return (hex 0D)
3384 \t tab (hex 09)
3385 \ddd character with octal code ddd, or back reference
3386 \xhh character with hex code hh
3387 \x{hhh..} character with hex code hhh..
3388
3389 The precise effect of \cx is as follows: if x is a lower case letter,
3390 it is converted to upper case. Then bit 6 of the character (hex 40) is
3391 inverted. Thus \cz becomes hex 1A, but \c{ becomes hex 3B, while \c;
3392 becomes hex 7B.
3393
3394 After \x, from zero to two hexadecimal digits are read (letters can be
3395 in upper or lower case). Any number of hexadecimal digits may appear
3396 between \x{ and }, but the value of the character code must be less
3397 than 256 in non-UTF-8 mode, and less than 2**31 in UTF-8 mode. That is,
3398 the maximum value in hexadecimal is 7FFFFFFF. Note that this is bigger
3399 than the largest Unicode code point, which is 10FFFF.
3400
3401 If characters other than hexadecimal digits appear between \x{ and },
3402 or if there is no terminating }, this form of escape is not recognized.
3403 Instead, the initial \x will be interpreted as a basic hexadecimal
3404 escape, with no following digits, giving a character whose value is
3405 zero.
3406
3407 Characters whose value is less than 256 can be defined by either of the
3408 two syntaxes for \x. There is no difference in the way they are han-
3409 dled. For example, \xdc is exactly the same as \x{dc}.
3410
3411 After \0 up to two further octal digits are read. If there are fewer
3412 than two digits, just those that are present are used. Thus the
3413 sequence \0\x\07 specifies two binary zeros followed by a BEL character
3414 (code value 7). Make sure you supply two digits after the initial zero
3415 if the pattern character that follows is itself an octal digit.
3416
3417 The handling of a backslash followed by a digit other than 0 is compli-
3418 cated. Outside a character class, PCRE reads it and any following dig-
3419 its as a decimal number. If the number is less than 10, or if there
3420 have been at least that many previous capturing left parentheses in the
3421 expression, the entire sequence is taken as a back reference. A
3422 description of how this works is given later, following the discussion
3423 of parenthesized subpatterns.
3424
3425 Inside a character class, or if the decimal number is greater than 9
3426 and there have not been that many capturing subpatterns, PCRE re-reads
3427 up to three octal digits following the backslash, and uses them to gen-
3428 erate a data character. Any subsequent digits stand for themselves. In
3429 non-UTF-8 mode, the value of a character specified in octal must be
3430 less than \400. In UTF-8 mode, values up to \777 are permitted. For
3431 example:
3432
3433 \040 is another way of writing a space
3434 \40 is the same, provided there are fewer than 40
3435 previous capturing subpatterns
3436 \7 is always a back reference
3437 \11 might be a back reference, or another way of
3438 writing a tab
3439 \011 is always a tab
3440 \0113 is a tab followed by the character "3"
3441 \113 might be a back reference, otherwise the
3442 character with octal code 113
3443 \377 might be a back reference, otherwise
3444 the byte consisting entirely of 1 bits
3445 \81 is either a back reference, or a binary zero
3446 followed by the two characters "8" and "1"
3447
3448 Note that octal values of 100 or greater must not be introduced by a
3449 leading zero, because no more than three octal digits are ever read.
3450
3451 All the sequences that define a single character value can be used both
3452 inside and outside character classes. In addition, inside a character
3453 class, the sequence \b is interpreted as the backspace character (hex
3454 08). The sequences \B, \N, \R, and \X are not special inside a charac-
3455 ter class. Like any other unrecognized escape sequences, they are
3456 treated as the literal characters "B", "N", "R", and "X" by default,
3457 but cause an error if the PCRE_EXTRA option is set. Outside a character
3458 class, these sequences have different meanings.
3459
3460 Absolute and relative back references
3461
3462 The sequence \g followed by an unsigned or a negative number, option-
3463 ally enclosed in braces, is an absolute or relative back reference. A
3464 named back reference can be coded as \g{name}. Back references are dis-
3465 cussed later, following the discussion of parenthesized subpatterns.
3466
3467 Absolute and relative subroutine calls
3468
3469 For compatibility with Oniguruma, the non-Perl syntax \g followed by a
3470 name or a number enclosed either in angle brackets or single quotes, is
3471 an alternative syntax for referencing a subpattern as a "subroutine".
3472 Details are discussed later. Note that \g{...} (Perl syntax) and
3473 \g<...> (Oniguruma syntax) are not synonymous. The former is a back
3474 reference; the latter is a subroutine call.
3475
3476 Generic character types
3477
3478 Another use of backslash is for specifying generic character types:
3479
3480 \d any decimal digit
3481 \D any character that is not a decimal digit
3482 \h any horizontal whitespace character
3483 \H any character that is not a horizontal whitespace character
3484 \s any whitespace character
3485 \S any character that is not a whitespace character
3486 \v any vertical whitespace character
3487 \V any character that is not a vertical whitespace character
3488 \w any "word" character
3489 \W any "non-word" character
3490
3491 There is also the single sequence \N, which matches a non-newline char-
3492 acter. This is the same as the "." metacharacter when PCRE_DOTALL is
3493 not set.
3494
3495 Each pair of lower and upper case escape sequences partitions the com-
3496 plete set of characters into two disjoint sets. Any given character
3497 matches one, and only one, of each pair. The sequences can appear both
3498 inside and outside character classes. They each match one character of
3499 the appropriate type. If the current matching point is at the end of
3500 the subject string, all of them fail, because there is no character to
3501 match.
3502
3503 For compatibility with Perl, \s does not match the VT character (code
3504 11). This makes it different from the the POSIX "space" class. The \s
3505 characters are HT (9), LF (10), FF (12), CR (13), and space (32). If
3506 "use locale;" is included in a Perl script, \s may match the VT charac-
3507 ter. In PCRE, it never does.
3508
3509 A "word" character is an underscore or any character that is a letter
3510 or digit. By default, the definition of letters and digits is con-
3511 trolled by PCRE's low-valued character tables, and may vary if locale-
3512 specific matching is taking place (see "Locale support" in the pcreapi
3513 page). For example, in a French locale such as "fr_FR" in Unix-like
3514 systems, or "french" in Windows, some character codes greater than 128
3515 are used for accented letters, and these are then matched by \w. The
3516 use of locales with Unicode is discouraged.
3517
3518 By default, in UTF-8 mode, characters with values greater than 128
3519 never match \d, \s, or \w, and always match \D, \S, and \W. These
3520 sequences retain their original meanings from before UTF-8 support was
3521 available, mainly for efficiency reasons. However, if PCRE is compiled
3522 with Unicode property support, and the PCRE_UCP option is set, the be-
3523 haviour is changed so that Unicode properties are used to determine
3524 character types, as follows:
3525
3526 \d any character that \p{Nd} matches (decimal digit)
3527 \s any character that \p{Z} matches, plus HT, LF, FF, CR
3528 \w any character that \p{L} or \p{N} matches, plus underscore
3529
3530 The upper case escapes match the inverse sets of characters. Note that
3531 \d matches only decimal digits, whereas \w matches any Unicode digit,
3532 as well as any Unicode letter, and underscore. Note also that PCRE_UCP
3533 affects \b, and \B because they are defined in terms of \w and \W.
3534 Matching these sequences is noticeably slower when PCRE_UCP is set.
3535
3536 The sequences \h, \H, \v, and \V are Perl 5.10 features. In contrast to
3537 the other sequences, which match only ASCII characters by default,
3538 these always match certain high-valued codepoints in UTF-8 mode,
3539 whether or not PCRE_UCP is set. The horizontal space characters are:
3540
3541 U+0009 Horizontal tab
3542 U+0020 Space
3543 U+00A0 Non-break space
3544 U+1680 Ogham space mark
3545 U+180E Mongolian vowel separator
3546 U+2000 En quad
3547 U+2001 Em quad
3548 U+2002 En space
3549 U+2003 Em space
3550 U+2004 Three-per-em space
3551 U+2005 Four-per-em space
3552 U+2006 Six-per-em space
3553 U+2007 Figure space
3554 U+2008 Punctuation space
3555 U+2009 Thin space
3556 U+200A Hair space
3557 U+202F Narrow no-break space
3558 U+205F Medium mathematical space
3559 U+3000 Ideographic space
3560
3561 The vertical space characters are:
3562
3563 U+000A Linefeed
3564 U+000B Vertical tab
3565 U+000C Formfeed
3566 U+000D Carriage return
3567 U+0085 Next line
3568 U+2028 Line separator
3569 U+2029 Paragraph separator
3570
3571 Newline sequences
3572
3573 Outside a character class, by default, the escape sequence \R matches
3574 any Unicode newline sequence. This is a Perl 5.10 feature. In non-UTF-8
3575 mode \R is equivalent to the following:
3576
3577 (?>\r\n|\n|\x0b|\f|\r|\x85)
3578
3579 This is an example of an "atomic group", details of which are given
3580 below. This particular group matches either the two-character sequence
3581 CR followed by LF, or one of the single characters LF (linefeed,
3582 U+000A), VT (vertical tab, U+000B), FF (formfeed, U+000C), CR (carriage
3583 return, U+000D), or NEL (next line, U+0085). The two-character sequence
3584 is treated as a single unit that cannot be split.
3585
3586 In UTF-8 mode, two additional characters whose codepoints are greater
3587 than 255 are added: LS (line separator, U+2028) and PS (paragraph sepa-
3588 rator, U+2029). Unicode character property support is not needed for
3589 these characters to be recognized.
3590
3591 It is possible to restrict \R to match only CR, LF, or CRLF (instead of
3592 the complete set of Unicode line endings) by setting the option
3593 PCRE_BSR_ANYCRLF either at compile time or when the pattern is matched.
3594 (BSR is an abbrevation for "backslash R".) This can be made the default
3595 when PCRE is built; if this is the case, the other behaviour can be
3596 requested via the PCRE_BSR_UNICODE option. It is also possible to
3597 specify these settings by starting a pattern string with one of the
3598 following sequences:
3599
3600 (*BSR_ANYCRLF) CR, LF, or CRLF only
3601 (*BSR_UNICODE) any Unicode newline sequence
3602
3603 These override the default and the options given to pcre_compile() or
3604 pcre_compile2(), but they can be overridden by options given to
3605 pcre_exec() or pcre_dfa_exec(). Note that these special settings, which
3606 are not Perl-compatible, are recognized only at the very start of a
3607 pattern, and that they must be in upper case. If more than one of them
3608 is present, the last one is used. They can be combined with a change of
3609 newline convention; for example, a pattern can start with:
3610
3611 (*ANY)(*BSR_ANYCRLF)
3612
3613 They can also be combined with the (*UTF8) or (*UCP) special sequences.
3614 Inside a character class, \R is treated as an unrecognized escape
3615 sequence, and so matches the letter "R" by default, but causes an error
3616 if PCRE_EXTRA is set.
3617
3618 Unicode character properties
3619
3620 When PCRE is built with Unicode character property support, three addi-
3621 tional escape sequences that match characters with specific properties
3622 are available. When not in UTF-8 mode, these sequences are of course
3623 limited to testing characters whose codepoints are less than 256, but
3624 they do work in this mode. The extra escape sequences are:
3625
3626 \p{xx} a character with the xx property
3627 \P{xx} a character without the xx property
3628 \X an extended Unicode sequence
3629
3630 The property names represented by xx above are limited to the Unicode
3631 script names, the general category properties, "Any", which matches any
3632 character (including newline), and some special PCRE properties
3633 (described in the next section). Other Perl properties such as "InMu-
3634 sicalSymbols" are not currently supported by PCRE. Note that \P{Any}
3635 does not match any characters, so always causes a match failure.
3636
3637 Sets of Unicode characters are defined as belonging to certain scripts.
3638 A character from one of these sets can be matched using a script name.
3639 For example:
3640
3641 \p{Greek}
3642 \P{Han}
3643
3644 Those that are not part of an identified script are lumped together as
3645 "Common". The current list of scripts is:
3646
3647 Arabic, Armenian, Avestan, Balinese, Bamum, Bengali, Bopomofo, Braille,
3648 Buginese, Buhid, Canadian_Aboriginal, Carian, Cham, Cherokee, Common,
3649 Coptic, Cuneiform, Cypriot, Cyrillic, Deseret, Devanagari, Egyp-
3650 tian_Hieroglyphs, Ethiopic, Georgian, Glagolitic, Gothic, Greek,
3651 Gujarati, Gurmukhi, Han, Hangul, Hanunoo, Hebrew, Hiragana, Impe-
3652 rial_Aramaic, Inherited, Inscriptional_Pahlavi, Inscriptional_Parthian,
3653 Javanese, Kaithi, Kannada, Katakana, Kayah_Li, Kharoshthi, Khmer, Lao,
3654 Latin, Lepcha, Limbu, Linear_B, Lisu, Lycian, Lydian, Malayalam,
3655 Meetei_Mayek, Mongolian, Myanmar, New_Tai_Lue, Nko, Ogham, Old_Italic,
3656 Old_Persian, Old_South_Arabian, Old_Turkic, Ol_Chiki, Oriya, Osmanya,
3657 Phags_Pa, Phoenician, Rejang, Runic, Samaritan, Saurashtra, Shavian,
3658 Sinhala, Sundanese, Syloti_Nagri, Syriac, Tagalog, Tagbanwa, Tai_Le,
3659 Tai_Tham, Tai_Viet, Tamil, Telugu, Thaana, Thai, Tibetan, Tifinagh,
3660 Ugaritic, Vai, Yi.
3661
3662 Each character has exactly one Unicode general category property, spec-
3663 ified by a two-letter abbreviation. For compatibility with Perl, nega-
3664 tion can be specified by including a circumflex between the opening
3665 brace and the property name. For example, \p{^Lu} is the same as
3666 \P{Lu}.
3667
3668 If only one letter is specified with \p or \P, it includes all the gen-
3669 eral category properties that start with that letter. In this case, in
3670 the absence of negation, the curly brackets in the escape sequence are
3671 optional; these two examples have the same effect:
3672
3673 \p{L}
3674 \pL
3675
3676 The following general category property codes are supported:
3677
3678 C Other
3679 Cc Control
3680 Cf Format
3681 Cn Unassigned
3682 Co Private use
3683 Cs Surrogate
3684
3685 L Letter
3686 Ll Lower case letter
3687 Lm Modifier letter
3688 Lo Other letter
3689 Lt Title case letter
3690 Lu Upper case letter
3691
3692 M Mark
3693 Mc Spacing mark
3694 Me Enclosing mark
3695 Mn Non-spacing mark
3696
3697 N Number
3698 Nd Decimal number
3699 Nl Letter number
3700 No Other number
3701
3702 P Punctuation
3703 Pc Connector punctuation
3704 Pd Dash punctuation
3705 Pe Close punctuation
3706 Pf Final punctuation
3707 Pi Initial punctuation
3708 Po Other punctuation
3709 Ps Open punctuation
3710
3711 S Symbol
3712 Sc Currency symbol
3713 Sk Modifier symbol
3714 Sm Mathematical symbol
3715 So Other symbol
3716
3717 Z Separator
3718 Zl Line separator
3719 Zp Paragraph separator
3720 Zs Space separator
3721
3722 The special property L& is also supported: it matches a character that
3723 has the Lu, Ll, or Lt property, in other words, a letter that is not
3724 classified as a modifier or "other".
3725
3726 The Cs (Surrogate) property applies only to characters in the range
3727 U+D800 to U+DFFF. Such characters are not valid in UTF-8 strings (see
3728 RFC 3629) and so cannot be tested by PCRE, unless UTF-8 validity check-
3729 ing has been turned off (see the discussion of PCRE_NO_UTF8_CHECK in
3730 the pcreapi page). Perl does not support the Cs property.
3731
3732 The long synonyms for property names that Perl supports (such as
3733 \p{Letter}) are not supported by PCRE, nor is it permitted to prefix
3734 any of these properties with "Is".
3735
3736 No character that is in the Unicode table has the Cn (unassigned) prop-
3737 erty. Instead, this property is assumed for any code point that is not
3738 in the Unicode table.
3739
3740 Specifying caseless matching does not affect these escape sequences.
3741 For example, \p{Lu} always matches only upper case letters.
3742
3743 The \X escape matches any number of Unicode characters that form an
3744 extended Unicode sequence. \X is equivalent to
3745
3746 (?>\PM\pM*)
3747
3748 That is, it matches a character without the "mark" property, followed
3749 by zero or more characters with the "mark" property, and treats the
3750 sequence as an atomic group (see below). Characters with the "mark"
3751 property are typically accents that affect the preceding character.
3752 None of them have codepoints less than 256, so in non-UTF-8 mode \X
3753 matches any one character.
3754
3755 Matching characters by Unicode property is not fast, because PCRE has
3756 to search a structure that contains data for over fifteen thousand
3757 characters. That is why the traditional escape sequences such as \d and
3758 \w do not use Unicode properties in PCRE by default, though you can
3759 make them do so by setting the PCRE_UCP option for pcre_compile() or by
3760 starting the pattern with (*UCP).
3761
3762 PCRE's additional properties
3763
3764 As well as the standard Unicode properties described in the previous
3765 section, PCRE supports four more that make it possible to convert tra-
3766 ditional escape sequences such as \w and \s and POSIX character classes
3767 to use Unicode properties. PCRE uses these non-standard, non-Perl prop-
3768 erties internally when PCRE_UCP is set. They are:
3769
3770 Xan Any alphanumeric character
3771 Xps Any POSIX space character
3772 Xsp Any Perl space character
3773 Xwd Any Perl "word" character
3774
3775 Xan matches characters that have either the L (letter) or the N (num-
3776 ber) property. Xps matches the characters tab, linefeed, vertical tab,
3777 formfeed, or carriage return, and any other character that has the Z
3778 (separator) property. Xsp is the same as Xps, except that vertical tab
3779 is excluded. Xwd matches the same characters as Xan, plus underscore.
3780
3781 Resetting the match start
3782
3783 The escape sequence \K, which is a Perl 5.10 feature, causes any previ-
3784 ously matched characters not to be included in the final matched
3785 sequence. For example, the pattern:
3786
3787 foo\Kbar
3788
3789 matches "foobar", but reports that it has matched "bar". This feature
3790 is similar to a lookbehind assertion (described below). However, in
3791 this case, the part of the subject before the real match does not have
3792 to be of fixed length, as lookbehind assertions do. The use of \K does
3793 not interfere with the setting of captured substrings. For example,
3794 when the pattern
3795
3796 (foo)\Kbar
3797
3798 matches "foobar", the first substring is still set to "foo".
3799
3800 Perl documents that the use of \K within assertions is "not well
3801 defined". In PCRE, \K is acted upon when it occurs inside positive
3802 assertions, but is ignored in negative assertions.
3803
3804 Simple assertions
3805
3806 The final use of backslash is for certain simple assertions. An asser-
3807 tion specifies a condition that has to be met at a particular point in
3808 a match, without consuming any characters from the subject string. The
3809 use of subpatterns for more complicated assertions is described below.
3810 The backslashed assertions are:
3811
3812 \b matches at a word boundary
3813 \B matches when not at a word boundary
3814 \A matches at the start of the subject
3815 \Z matches at the end of the subject
3816 also matches before a newline at the end of the subject
3817 \z matches only at the end of the subject
3818 \G matches at the first matching position in the subject
3819
3820 Inside a character class, \b has a different meaning; it matches the
3821 backspace character. If any other of these assertions appears in a
3822 character class, by default it matches the corresponding literal char-
3823 acter (for example, \B matches the letter B). However, if the
3824 PCRE_EXTRA option is set, an "invalid escape sequence" error is gener-
3825 ated instead.
3826
3827 A word boundary is a position in the subject string where the current
3828 character and the previous character do not both match \w or \W (i.e.
3829 one matches \w and the other matches \W), or the start or end of the
3830 string if the first or last character matches \w, respectively. In
3831 UTF-8 mode, the meanings of \w and \W can be changed by setting the
3832 PCRE_UCP option. When this is done, it also affects \b and \B. Neither
3833 PCRE nor Perl has a separate "start of word" or "end of word" metase-
3834 quence. However, whatever follows \b normally determines which it is.
3835 For example, the fragment \ba matches "a" at the start of a word.
3836
3837 The \A, \Z, and \z assertions differ from the traditional circumflex
3838 and dollar (described in the next section) in that they only ever match
3839 at the very start and end of the subject string, whatever options are
3840 set. Thus, they are independent of multiline mode. These three asser-
3841 tions are not affected by the PCRE_NOTBOL or PCRE_NOTEOL options, which
3842 affect only the behaviour of the circumflex and dollar metacharacters.
3843 However, if the startoffset argument of pcre_exec() is non-zero, indi-
3844 cating that matching is to start at a point other than the beginning of
3845 the subject, \A can never match. The difference between \Z and \z is
3846 that \Z matches before a newline at the end of the string as well as at
3847 the very end, whereas \z matches only at the end.
3848
3849 The \G assertion is true only when the current matching position is at
3850 the start point of the match, as specified by the startoffset argument
3851 of pcre_exec(). It differs from \A when the value of startoffset is
3852 non-zero. By calling pcre_exec() multiple times with appropriate argu-
3853 ments, you can mimic Perl's /g option, and it is in this kind of imple-
3854 mentation where \G can be useful.
3855
3856 Note, however, that PCRE's interpretation of \G, as the start of the
3857 current match, is subtly different from Perl's, which defines it as the
3858 end of the previous match. In Perl, these can be different when the
3859 previously matched string was empty. Because PCRE does just one match
3860 at a time, it cannot reproduce this behaviour.
3861
3862 If all the alternatives of a pattern begin with \G, the expression is
3863 anchored to the starting match position, and the "anchored" flag is set
3864 in the compiled regular expression.
3865
3866
3867 CIRCUMFLEX AND DOLLAR
3868
3869 Outside a character class, in the default matching mode, the circumflex
3870 character is an assertion that is true only if the current matching
3871 point is at the start of the subject string. If the startoffset argu-
3872 ment of pcre_exec() is non-zero, circumflex can never match if the
3873 PCRE_MULTILINE option is unset. Inside a character class, circumflex
3874 has an entirely different meaning (see below).
3875
3876 Circumflex need not be the first character of the pattern if a number
3877 of alternatives are involved, but it should be the first thing in each
3878 alternative in which it appears if the pattern is ever to match that
3879 branch. If all possible alternatives start with a circumflex, that is,
3880 if the pattern is constrained to match only at the start of the sub-
3881 ject, it is said to be an "anchored" pattern. (There are also other
3882 constructs that can cause a pattern to be anchored.)
3883
3884 A dollar character is an assertion that is true only if the current
3885 matching point is at the end of the subject string, or immediately
3886 before a newline at the end of the string (by default). Dollar need not
3887 be the last character of the pattern if a number of alternatives are
3888 involved, but it should be the last item in any branch in which it
3889 appears. Dollar has no special meaning in a character class.
3890
3891 The meaning of dollar can be changed so that it matches only at the
3892 very end of the string, by setting the PCRE_DOLLAR_ENDONLY option at
3893 compile time. This does not affect the \Z assertion.
3894
3895 The meanings of the circumflex and dollar characters are changed if the
3896 PCRE_MULTILINE option is set. When this is the case, a circumflex
3897 matches immediately after internal newlines as well as at the start of
3898 the subject string. It does not match after a newline that ends the
3899 string. A dollar matches before any newlines in the string, as well as
3900 at the very end, when PCRE_MULTILINE is set. When newline is specified
3901 as the two-character sequence CRLF, isolated CR and LF characters do
3902 not indicate newlines.
3903
3904 For example, the pattern /^abc$/ matches the subject string "def\nabc"
3905 (where \n represents a newline) in multiline mode, but not otherwise.
3906 Consequently, patterns that are anchored in single line mode because
3907 all branches start with ^ are not anchored in multiline mode, and a
3908 match for circumflex is possible when the startoffset argument of
3909 pcre_exec() is non-zero. The PCRE_DOLLAR_ENDONLY option is ignored if
3910 PCRE_MULTILINE is set.
3911
3912 Note that the sequences \A, \Z, and \z can be used to match the start
3913 and end of the subject in both modes, and if all branches of a pattern
3914 start with \A it is always anchored, whether or not PCRE_MULTILINE is
3915 set.
3916
3917
3918 FULL STOP (PERIOD, DOT) AND \N
3919
3920 Outside a character class, a dot in the pattern matches any one charac-
3921 ter in the subject string except (by default) a character that signi-
3922 fies the end of a line. In UTF-8 mode, the matched character may be
3923 more than one byte long.
3924
3925 When a line ending is defined as a single character, dot never matches
3926 that character; when the two-character sequence CRLF is used, dot does
3927 not match CR if it is immediately followed by LF, but otherwise it
3928 matches all characters (including isolated CRs and LFs). When any Uni-
3929 code line endings are being recognized, dot does not match CR or LF or
3930 any of the other line ending characters.
3931
3932 The behaviour of dot with regard to newlines can be changed. If the
3933 PCRE_DOTALL option is set, a dot matches any one character, without
3934 exception. If the two-character sequence CRLF is present in the subject
3935 string, it takes two dots to match it.
3936
3937 The handling of dot is entirely independent of the handling of circum-
3938 flex and dollar, the only relationship being that they both involve
3939 newlines. Dot has no special meaning in a character class.
3940
3941 The escape sequence \N always behaves as a dot does when PCRE_DOTALL is
3942 not set. In other words, it matches any one character except one that
3943 signifies the end of a line.
3944
3945
3946 MATCHING A SINGLE BYTE
3947
3948 Outside a character class, the escape sequence \C matches any one byte,
3949 both in and out of UTF-8 mode. Unlike a dot, it always matches any
3950 line-ending characters. The feature is provided in Perl in order to
3951 match individual bytes in UTF-8 mode. Because it breaks up UTF-8 char-
3952 acters into individual bytes, what remains in the string may be a mal-
3953 formed UTF-8 string. For this reason, the \C escape sequence is best
3954 avoided.
3955
3956 PCRE does not allow \C to appear in lookbehind assertions (described
3957 below), because in UTF-8 mode this would make it impossible to calcu-
3958 late the length of the lookbehind.
3959
3960
3961 SQUARE BRACKETS AND CHARACTER CLASSES
3962
3963 An opening square bracket introduces a character class, terminated by a
3964 closing square bracket. A closing square bracket on its own is not spe-
3965 cial by default. However, if the PCRE_JAVASCRIPT_COMPAT option is set,
3966 a lone closing square bracket causes a compile-time error. If a closing
3967 square bracket is required as a member of the class, it should be the
3968 first data character in the class (after an initial circumflex, if
3969 present) or escaped with a backslash.
3970
3971 A character class matches a single character in the subject. In UTF-8
3972 mode, the character may be more than one byte long. A matched character
3973 must be in the set of characters defined by the class, unless the first
3974 character in the class definition is a circumflex, in which case the
3975 subject character must not be in the set defined by the class. If a
3976 circumflex is actually required as a member of the class, ensure it is
3977 not the first character, or escape it with a backslash.
3978
3979 For example, the character class [aeiou] matches any lower case vowel,
3980 while [^aeiou] matches any character that is not a lower case vowel.
3981 Note that a circumflex is just a convenient notation for specifying the
3982 characters that are in the class by enumerating those that are not. A
3983 class that starts with a circumflex is not an assertion; it still con-
3984 sumes a character from the subject string, and therefore it fails if
3985 the current pointer is at the end of the string.
3986
3987 In UTF-8 mode, characters with values greater than 255 can be included
3988 in a class as a literal string of bytes, or by using the \x{ escaping
3989 mechanism.
3990
3991 When caseless matching is set, any letters in a class represent both
3992 their upper case and lower case versions, so for example, a caseless
3993 [aeiou] matches "A" as well as "a", and a caseless [^aeiou] does not
3994 match "A", whereas a caseful version would. In UTF-8 mode, PCRE always
3995 understands the concept of case for characters whose values are less
3996 than 128, so caseless matching is always possible. For characters with
3997 higher values, the concept of case is supported if PCRE is compiled
3998 with Unicode property support, but not otherwise. If you want to use
3999 caseless matching in UTF8-mode for characters 128 and above, you must
4000 ensure that PCRE is compiled with Unicode property support as well as
4001 with UTF-8 support.
4002
4003 Characters that might indicate line breaks are never treated in any
4004 special way when matching character classes, whatever line-ending
4005 sequence is in use, and whatever setting of the PCRE_DOTALL and
4006 PCRE_MULTILINE options is used. A class such as [^a] always matches one
4007 of these characters.
4008
4009 The minus (hyphen) character can be used to specify a range of charac-
4010 ters in a character class. For example, [d-m] matches any letter
4011 between d and m, inclusive. If a minus character is required in a
4012 class, it must be escaped with a backslash or appear in a position
4013 where it cannot be interpreted as indicating a range, typically as the
4014 first or last character in the class.
4015
4016 It is not possible to have the literal character "]" as the end charac-
4017 ter of a range. A pattern such as [W-]46] is interpreted as a class of
4018 two characters ("W" and "-") followed by a literal string "46]", so it
4019 would match "W46]" or "-46]". However, if the "]" is escaped with a
4020 backslash it is interpreted as the end of range, so [W-\]46] is inter-
4021 preted as a class containing a range followed by two other characters.
4022 The octal or hexadecimal representation of "]" can also be used to end
4023 a range.
4024
4025 Ranges operate in the collating sequence of character values. They can
4026 also be used for characters specified numerically, for example
4027 [\000-\037]. In UTF-8 mode, ranges can include characters whose values
4028 are greater than 255, for example [\x{100}-\x{2ff}].
4029
4030 If a range that includes letters is used when caseless matching is set,
4031 it matches the letters in either case. For example, [W-c] is equivalent
4032 to [][\\^_`wxyzabc], matched caselessly, and in non-UTF-8 mode, if
4033 character tables for a French locale are in use, [\xc8-\xcb] matches
4034 accented E characters in both cases. In UTF-8 mode, PCRE supports the
4035 concept of case for characters with values greater than 128 only when
4036 it is compiled with Unicode property support.
4037
4038 The character types \d, \D, \h, \H, \p, \P, \s, \S, \v, \V, \w, and \W
4039 may also appear in a character class, and add the characters that they
4040 match to the class. For example, [\dABCDEF] matches any hexadecimal
4041 digit. A circumflex can conveniently be used with the upper case char-
4042 acter types to specify a more restricted set of characters than the
4043 matching lower case type. For example, the class [^\W_] matches any
4044 letter or digit, but not underscore.
4045
4046 The only metacharacters that are recognized in character classes are
4047 backslash, hyphen (only where it can be interpreted as specifying a
4048 range), circumflex (only at the start), opening square bracket (only
4049 when it can be interpreted as introducing a POSIX class name - see the
4050 next section), and the terminating closing square bracket. However,
4051 escaping other non-alphanumeric characters does no harm.
4052
4053
4054 POSIX CHARACTER CLASSES
4055
4056 Perl supports the POSIX notation for character classes. This uses names
4057 enclosed by [: and :] within the enclosing square brackets. PCRE also
4058 supports this notation. For example,
4059
4060 [01[:alpha:]%]
4061
4062 matches "0", "1", any alphabetic character, or "%". The supported class
4063 names are:
4064
4065 alnum letters and digits
4066 alpha letters
4067 ascii character codes 0 - 127
4068 blank space or tab only
4069 cntrl control characters
4070 digit decimal digits (same as \d)
4071 graph printing characters, excluding space
4072 lower lower case letters
4073 print printing characters, including space
4074 punct printing characters, excluding letters and digits and space
4075 space white space (not quite the same as \s)
4076 upper upper case letters
4077 word "word" characters (same as \w)
4078 xdigit hexadecimal digits
4079
4080 The "space" characters are HT (9), LF (10), VT (11), FF (12), CR (13),
4081 and space (32). Notice that this list includes the VT character (code
4082 11). This makes "space" different to \s, which does not include VT (for
4083 Perl compatibility).
4084
4085 The name "word" is a Perl extension, and "blank" is a GNU extension
4086 from Perl 5.8. Another Perl extension is negation, which is indicated
4087 by a ^ character after the colon. For example,
4088
4089 [12[:^digit:]]
4090
4091 matches "1", "2", or any non-digit. PCRE (and Perl) also recognize the
4092 POSIX syntax [.ch.] and [=ch=] where "ch" is a "collating element", but
4093 these are not supported, and an error is given if they are encountered.
4094
4095 By default, in UTF-8 mode, characters with values greater than 128 do
4096 not match any of the POSIX character classes. However, if the PCRE_UCP
4097 option is passed to pcre_compile(), some of the classes are changed so
4098 that Unicode character properties are used. This is achieved by replac-
4099 ing the POSIX classes by other sequences, as follows:
4100
4101 [:alnum:] becomes \p{Xan}
4102 [:alpha:] becomes \p{L}
4103 [:blank:] becomes \h
4104 [:digit:] becomes \p{Nd}
4105 [:lower:] becomes \p{Ll}
4106 [:space:] becomes \p{Xps}
4107 [:upper:] becomes \p{Lu}
4108 [:word:] becomes \p{Xwd}
4109
4110 Negated versions, such as [:^alpha:] use \P instead of \p. The other
4111 POSIX classes are unchanged, and match only characters with code points
4112 less than 128.
4113
4114
4115 VERTICAL BAR
4116
4117 Vertical bar characters are used to separate alternative patterns. For
4118 example, the pattern
4119
4120 gilbert|sullivan
4121
4122 matches either "gilbert" or "sullivan". Any number of alternatives may
4123 appear, and an empty alternative is permitted (matching the empty
4124 string). The matching process tries each alternative in turn, from left
4125 to right, and the first one that succeeds is used. If the alternatives
4126 are within a subpattern (defined below), "succeeds" means matching the
4127 rest of the main pattern as well as the alternative in the subpattern.
4128
4129
4130 INTERNAL OPTION SETTING
4131
4132 The settings of the PCRE_CASELESS, PCRE_MULTILINE, PCRE_DOTALL, and
4133 PCRE_EXTENDED options (which are Perl-compatible) can be changed from
4134 within the pattern by a sequence of Perl option letters enclosed
4135 between "(?" and ")". The option letters are
4136
4137 i for PCRE_CASELESS
4138 m for PCRE_MULTILINE
4139 s for PCRE_DOTALL
4140 x for PCRE_EXTENDED
4141
4142 For example, (?im) sets caseless, multiline matching. It is also possi-
4143 ble to unset these options by preceding the letter with a hyphen, and a
4144 combined setting and unsetting such as (?im-sx), which sets PCRE_CASE-
4145 LESS and PCRE_MULTILINE while unsetting PCRE_DOTALL and PCRE_EXTENDED,
4146 is also permitted. If a letter appears both before and after the
4147 hyphen, the option is unset.
4148
4149 The PCRE-specific options PCRE_DUPNAMES, PCRE_UNGREEDY, and PCRE_EXTRA
4150 can be changed in the same way as the Perl-compatible options by using
4151 the characters J, U and X respectively.
4152
4153 When one of these option changes occurs at top level (that is, not
4154 inside subpattern parentheses), the change applies to the remainder of
4155 the pattern that follows. If the change is placed right at the start of
4156 a pattern, PCRE extracts it into the global options (and it will there-
4157 fore show up in data extracted by the pcre_fullinfo() function).
4158
4159 An option change within a subpattern (see below for a description of
4160 subpatterns) affects only that part of the current pattern that follows
4161 it, so
4162
4163 (a(?i)b)c
4164
4165 matches abc and aBc and no other strings (assuming PCRE_CASELESS is not
4166 used). By this means, options can be made to have different settings
4167 in different parts of the pattern. Any changes made in one alternative
4168 do carry on into subsequent branches within the same subpattern. For
4169 example,
4170
4171 (a(?i)b|c)
4172
4173 matches "ab", "aB", "c", and "C", even though when matching "C" the
4174 first branch is abandoned before the option setting. This is because
4175 the effects of option settings happen at compile time. There would be
4176 some very weird behaviour otherwise.
4177
4178 Note: There are other PCRE-specific options that can be set by the
4179 application when the compile or match functions are called. In some
4180 cases the pattern can contain special leading sequences such as (*CRLF)
4181 to override what the application has set or what has been defaulted.
4182 Details are given in the section entitled "Newline sequences" above.
4183 There are also the (*UTF8) and (*UCP) leading sequences that can be
4184 used to set UTF-8 and Unicode property modes; they are equivalent to
4185 setting the PCRE_UTF8 and the PCRE_UCP options, respectively.
4186
4187
4188 SUBPATTERNS
4189
4190 Subpatterns are delimited by parentheses (round brackets), which can be
4191 nested. Turning part of a pattern into a subpattern does two things:
4192
4193 1. It localizes a set of alternatives. For example, the pattern
4194
4195 cat(aract|erpillar|)
4196
4197 matches one of the words "cat", "cataract", or "caterpillar". Without
4198 the parentheses, it would match "cataract", "erpillar" or an empty
4199 string.
4200
4201 2. It sets up the subpattern as a capturing subpattern. This means
4202 that, when the whole pattern matches, that portion of the subject
4203 string that matched the subpattern is passed back to the caller via the
4204 ovector argument of pcre_exec(). Opening parentheses are counted from
4205 left to right (starting from 1) to obtain numbers for the capturing
4206 subpatterns.
4207
4208 For example, if the string "the red king" is matched against the pat-
4209 tern
4210
4211 the ((red|white) (king|queen))
4212
4213 the captured substrings are "red king", "red", and "king", and are num-
4214 bered 1, 2, and 3, respectively.
4215
4216 The fact that plain parentheses fulfil two functions is not always
4217 helpful. There are often times when a grouping subpattern is required
4218 without a capturing requirement. If an opening parenthesis is followed
4219 by a question mark and a colon, the subpattern does not do any captur-
4220 ing, and is not counted when computing the number of any subsequent
4221 capturing subpatterns. For example, if the string "the white queen" is
4222 matched against the pattern
4223
4224 the ((?:red|white) (king|queen))
4225
4226 the captured substrings are "white queen" and "queen", and are numbered
4227 1 and 2. The maximum number of capturing subpatterns is 65535.
4228
4229 As a convenient shorthand, if any option settings are required at the
4230 start of a non-capturing subpattern, the option letters may appear
4231 between the "?" and the ":". Thus the two patterns
4232
4233 (?i:saturday|sunday)
4234 (?:(?i)saturday|sunday)
4235
4236 match exactly the same set of strings. Because alternative branches are
4237 tried from left to right, and options are not reset until the end of
4238 the subpattern is reached, an option setting in one branch does affect
4239 subsequent branches, so the above patterns match "SUNDAY" as well as
4240 "Saturday".
4241
4242
4243 DUPLICATE SUBPATTERN NUMBERS
4244
4245 Perl 5.10 introduced a feature whereby each alternative in a subpattern
4246 uses the same numbers for its capturing parentheses. Such a subpattern
4247 starts with (?| and is itself a non-capturing subpattern. For example,
4248 consider this pattern:
4249
4250 (?|(Sat)ur|(Sun))day
4251
4252 Because the two alternatives are inside a (?| group, both sets of cap-
4253 turing parentheses are numbered one. Thus, when the pattern matches,
4254 you can look at captured substring number one, whichever alternative
4255 matched. This construct is useful when you want to capture part, but
4256 not all, of one of a number of alternatives. Inside a (?| group, paren-
4257 theses are numbered as usual, but the number is reset at the start of
4258 each branch. The numbers of any capturing buffers that follow the sub-
4259 pattern start after the highest number used in any branch. The follow-
4260 ing example is taken from the Perl documentation. The numbers under-
4261 neath show in which buffer the captured content will be stored.
4262
4263 # before ---------------branch-reset----------- after
4264 / ( a ) (?| x ( y ) z | (p (q) r) | (t) u (v) ) ( z ) /x
4265 # 1 2 2 3 2 3 4
4266
4267 A back reference to a numbered subpattern uses the most recent value
4268 that is set for that number by any subpattern. The following pattern
4269 matches "abcabc" or "defdef":
4270
4271 /(?|(abc)|(def))\1/
4272
4273 In contrast, a recursive or "subroutine" call to a numbered subpattern
4274 always refers to the first one in the pattern with the given number.
4275 The following pattern matches "abcabc" or "defabc":
4276
4277 /(?|(abc)|(def))(?1)/
4278
4279 If a condition test for a subpattern's having matched refers to a non-
4280 unique number, the test is true if any of the subpatterns of that num-
4281 ber have matched.
4282
4283 An alternative approach to using this "branch reset" feature is to use
4284 duplicate named subpatterns, as described in the next section.
4285
4286
4287 NAMED SUBPATTERNS
4288
4289 Identifying capturing parentheses by number is simple, but it can be
4290 very hard to keep track of the numbers in complicated regular expres-
4291 sions. Furthermore, if an expression is modified, the numbers may
4292 change. To help with this difficulty, PCRE supports the naming of sub-
4293 patterns. This feature was not added to Perl until release 5.10. Python
4294 had the feature earlier, and PCRE introduced it at release 4.0, using
4295 the Python syntax. PCRE now supports both the Perl and the Python syn-
4296 tax. Perl allows identically numbered subpatterns to have different
4297 names, but PCRE does not.
4298
4299 In PCRE, a subpattern can be named in one of three ways: (?<name>...)
4300 or (?'name'...) as in Perl, or (?P<name>...) as in Python. References
4301 to capturing parentheses from other parts of the pattern, such as back
4302 references, recursion, and conditions, can be made by name as well as
4303 by number.
4304
4305 Names consist of up to 32 alphanumeric characters and underscores.
4306 Named capturing parentheses are still allocated numbers as well as
4307 names, exactly as if the names were not present. The PCRE API provides
4308 function calls for extracting the name-to-number translation table from
4309 a compiled pattern. There is also a convenience function for extracting
4310 a captured substring by name.
4311
4312 By default, a name must be unique within a pattern, but it is possible
4313 to relax this constraint by setting the PCRE_DUPNAMES option at compile
4314 time. (Duplicate names are also always permitted for subpatterns with
4315 the same number, set up as described in the previous section.) Dupli-
4316 cate names can be useful for patterns where only one instance of the
4317 named parentheses can match. Suppose you want to match the name of a
4318 weekday, either as a 3-letter abbreviation or as the full name, and in
4319 both cases you want to extract the abbreviation. This pattern (ignoring
4320 the line breaks) does the job:
4321
4322 (?<DN>Mon|Fri|Sun)(?:day)?|
4323 (?<DN>Tue)(?:sday)?|
4324 (?<DN>Wed)(?:nesday)?|
4325 (?<DN>Thu)(?:rsday)?|
4326 (?<DN>Sat)(?:urday)?
4327
4328 There are five capturing substrings, but only one is ever set after a
4329 match. (An alternative way of solving this problem is to use a "branch
4330 reset" subpattern, as described in the previous section.)
4331
4332 The convenience function for extracting the data by name returns the
4333 substring for the first (and in this example, the only) subpattern of
4334 that name that matched. This saves searching to find which numbered
4335 subpattern it was.
4336
4337 If you make a back reference to a non-unique named subpattern from
4338 elsewhere in the pattern, the one that corresponds to the first occur-
4339 rence of the name is used. In the absence of duplicate numbers (see the
4340 previous section) this is the one with the lowest number. If you use a
4341 named reference in a condition test (see the section about conditions
4342 below), either to check whether a subpattern has matched, or to check
4343 for recursion, all subpatterns with the same name are tested. If the
4344 condition is true for any one of them, the overall condition is true.
4345 This is the same behaviour as testing by number. For further details of
4346 the interfaces for handling named subpatterns, see the pcreapi documen-
4347 tation.
4348
4349 Warning: You cannot use different names to distinguish between two sub-
4350 patterns with the same number because PCRE uses only the numbers when
4351 matching. For this reason, an error is given at compile time if differ-
4352 ent names are given to subpatterns with the same number. However, you
4353 can give the same name to subpatterns with the same number, even when
4354 PCRE_DUPNAMES is not set.
4355
4356
4357 REPETITION
4358
4359 Repetition is specified by quantifiers, which can follow any of the
4360 following items:
4361
4362 a literal data character
4363 the dot metacharacter
4364 the \C escape sequence
4365 the \X escape sequence (in UTF-8 mode with Unicode properties)
4366 the \R escape sequence
4367 an escape such as \d that matches a single character
4368 a character class
4369 a back reference (see next section)
4370 a parenthesized subpattern (unless it is an assertion)
4371 a recursive or "subroutine" call to a subpattern
4372
4373 The general repetition quantifier specifies a minimum and maximum num-
4374 ber of permitted matches, by giving the two numbers in curly brackets
4375 (braces), separated by a comma. The numbers must be less than 65536,
4376 and the first must be less than or equal to the second. For example:
4377
4378 z{2,4}
4379
4380 matches "zz", "zzz", or "zzzz". A closing brace on its own is not a
4381 special character. If the second number is omitted, but the comma is
4382 present, there is no upper limit; if the second number and the comma
4383 are both omitted, the quantifier specifies an exact number of required
4384 matches. Thus
4385
4386 [aeiou]{3,}
4387
4388 matches at least 3 successive vowels, but may match many more, while
4389
4390 \d{8}
4391
4392 matches exactly 8 digits. An opening curly bracket that appears in a
4393 position where a quantifier is not allowed, or one that does not match
4394 the syntax of a quantifier, is taken as a literal character. For exam-
4395 ple, {,6} is not a quantifier, but a literal string of four characters.
4396
4397 In UTF-8 mode, quantifiers apply to UTF-8 characters rather than to
4398 individual bytes. Thus, for example, \x{100}{2} matches two UTF-8 char-
4399 acters, each of which is represented by a two-byte sequence. Similarly,
4400 when Unicode property support is available, \X{3} matches three Unicode
4401 extended sequences, each of which may be several bytes long (and they
4402 may be of different lengths).
4403
4404 The quantifier {0} is permitted, causing the expression to behave as if
4405 the previous item and the quantifier were not present. This may be use-
4406 ful for subpatterns that are referenced as subroutines from elsewhere
4407 in the pattern. Items other than subpatterns that have a {0} quantifier
4408 are omitted from the compiled pattern.
4409
4410 For convenience, the three most common quantifiers have single-charac-
4411 ter abbreviations:
4412
4413 * is equivalent to {0,}
4414 + is equivalent to {1,}
4415 ? is equivalent to {0,1}
4416
4417 It is possible to construct infinite loops by following a subpattern
4418 that can match no characters with a quantifier that has no upper limit,
4419 for example:
4420
4421 (a?)*
4422
4423 Earlier versions of Perl and PCRE used to give an error at compile time
4424 for such patterns. However, because there are cases where this can be
4425 useful, such patterns are now accepted, but if any repetition of the
4426 subpattern does in fact match no characters, the loop is forcibly bro-
4427 ken.
4428
4429 By default, the quantifiers are "greedy", that is, they match as much
4430 as possible (up to the maximum number of permitted times), without
4431 causing the rest of the pattern to fail. The classic example of where
4432 this gives problems is in trying to match comments in C programs. These
4433 appear between /* and */ and within the comment, individual * and /
4434 characters may appear. An attempt to match C comments by applying the
4435 pattern
4436
4437 /\*.*\*/
4438
4439 to the string
4440
4441 /* first comment */ not comment /* second comment */
4442
4443 fails, because it matches the entire string owing to the greediness of
4444 the .* item.
4445
4446 However, if a quantifier is followed by a question mark, it ceases to
4447 be greedy, and instead matches the minimum number of times possible, so
4448 the pattern
4449
4450 /\*.*?\*/
4451
4452 does the right thing with the C comments. The meaning of the various
4453 quantifiers is not otherwise changed, just the preferred number of
4454 matches. Do not confuse this use of question mark with its use as a
4455 quantifier in its own right. Because it has two uses, it can sometimes
4456 appear doubled, as in
4457
4458 \d??\d
4459
4460 which matches one digit by preference, but can match two if that is the
4461 only way the rest of the pattern matches.
4462
4463 If the PCRE_UNGREEDY option is set (an option that is not available in
4464 Perl), the quantifiers are not greedy by default, but individual ones
4465 can be made greedy by following them with a question mark. In other
4466 words, it inverts the default behaviour.
4467
4468 When a parenthesized subpattern is quantified with a minimum repeat
4469 count that is greater than 1 or with a limited maximum, more memory is
4470 required for the compiled pattern, in proportion to the size of the
4471 minimum or maximum.
4472
4473 If a pattern starts with .* or .{0,} and the PCRE_DOTALL option (equiv-
4474 alent to Perl's /s) is set, thus allowing the dot to match newlines,
4475 the pattern is implicitly anchored, because whatever follows will be
4476 tried against every character position in the subject string, so there
4477 is no point in retrying the overall match at any position after the
4478 first. PCRE normally treats such a pattern as though it were preceded
4479 by \A.
4480
4481 In cases where it is known that the subject string contains no new-
4482 lines, it is worth setting PCRE_DOTALL in order to obtain this opti-
4483 mization, or alternatively using ^ to indicate anchoring explicitly.
4484
4485 However, there is one situation where the optimization cannot be used.
4486 When .* is inside capturing parentheses that are the subject of a back
4487 reference elsewhere in the pattern, a match at the start may fail where
4488 a later one succeeds. Consider, for example:
4489
4490 (.*)abc\1
4491
4492 If the subject is "xyz123abc123" the match point is the fourth charac-
4493 ter. For this reason, such a pattern is not implicitly anchored.
4494
4495 When a capturing subpattern is repeated, the value captured is the sub-
4496 string that matched the final iteration. For example, after
4497
4498 (tweedle[dume]{3}\s*)+
4499
4500 has matched "tweedledum tweedledee" the value of the captured substring
4501 is "tweedledee". However, if there are nested capturing subpatterns,
4502 the corresponding captured values may have been set in previous itera-
4503 tions. For example, after
4504
4505 /(a|(b))+/
4506
4507 matches "aba" the value of the second captured substring is "b".
4508
4509
4510 ATOMIC GROUPING AND POSSESSIVE QUANTIFIERS
4511
4512 With both maximizing ("greedy") and minimizing ("ungreedy" or "lazy")
4513 repetition, failure of what follows normally causes the repeated item
4514 to be re-evaluated to see if a different number of repeats allows the
4515 rest of the pattern to match. Sometimes it is useful to prevent this,
4516 either to change the nature of the match, or to cause it fail earlier
4517 than it otherwise might, when the author of the pattern knows there is
4518 no point in carrying on.
4519
4520 Consider, for example, the pattern \d+foo when applied to the subject
4521 line
4522
4523 123456bar
4524
4525 After matching all 6 digits and then failing to match "foo", the normal
4526 action of the matcher is to try again with only 5 digits matching the
4527 \d+ item, and then with 4, and so on, before ultimately failing.
4528 "Atomic grouping" (a term taken from Jeffrey Friedl's book) provides
4529 the means for specifying that once a subpattern has matched, it is not
4530 to be re-evaluated in this way.
4531
4532 If we use atomic grouping for the previous example, the matcher gives
4533 up immediately on failing to match "foo" the first time. The notation
4534 is a kind of special parenthesis, starting with (?> as in this example:
4535
4536 (?>\d+)foo
4537
4538 This kind of parenthesis "locks up" the part of the pattern it con-
4539 tains once it has matched, and a failure further into the pattern is
4540 prevented from backtracking into it. Backtracking past it to previous
4541 items, however, works as normal.
4542
4543 An alternative description is that a subpattern of this type matches
4544 the string of characters that an identical standalone pattern would
4545 match, if anchored at the current point in the subject string.
4546
4547 Atomic grouping subpatterns are not capturing subpatterns. Simple cases
4548 such as the above example can be thought of as a maximizing repeat that
4549 must swallow everything it can. So, while both \d+ and \d+? are pre-
4550 pared to adjust the number of digits they match in order to make the
4551 rest of the pattern match, (?>\d+) can only match an entire sequence of
4552 digits.
4553
4554 Atomic groups in general can of course contain arbitrarily complicated
4555 subpatterns, and can be nested. However, when the subpattern for an
4556 atomic group is just a single repeated item, as in the example above, a
4557 simpler notation, called a "possessive quantifier" can be used. This
4558 consists of an additional + character following a quantifier. Using
4559 this notation, the previous example can be rewritten as
4560
4561 \d++foo
4562
4563 Note that a possessive quantifier can be used with an entire group, for
4564 example:
4565
4566 (abc|xyz){2,3}+
4567
4568 Possessive quantifiers are always greedy; the setting of the
4569 PCRE_UNGREEDY option is ignored. They are a convenient notation for the
4570 simpler forms of atomic group. However, there is no difference in the
4571 meaning of a possessive quantifier and the equivalent atomic group,
4572 though there may be a performance difference; possessive quantifiers
4573 should be slightly faster.
4574
4575 The possessive quantifier syntax is an extension to the Perl 5.8 syn-
4576 tax. Jeffrey Friedl originated the idea (and the name) in the first
4577 edition of his book. Mike McCloskey liked it, so implemented it when he
4578 built Sun's Java package, and PCRE copied it from there. It ultimately
4579 found its way into Perl at release 5.10.
4580
4581 PCRE has an optimization that automatically "possessifies" certain sim-
4582 ple pattern constructs. For example, the sequence A+B is treated as
4583 A++B because there is no point in backtracking into a sequence of A's
4584 when B must follow.
4585
4586 When a pattern contains an unlimited repeat inside a subpattern that
4587 can itself be repeated an unlimited number of times, the use of an
4588 atomic group is the only way to avoid some failing matches taking a
4589 very long time indeed. The pattern
4590
4591 (\D+|<\d+>)*[!?]
4592
4593 matches an unlimited number of substrings that either consist of non-
4594 digits, or digits enclosed in <>, followed by either ! or ?. When it
4595 matches, it runs quickly. However, if it is applied to
4596
4597 aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
4598
4599 it takes a long time before reporting failure. This is because the
4600 string can be divided between the internal \D+ repeat and the external
4601 * repeat in a large number of ways, and all have to be tried. (The
4602 example uses [!?] rather than a single character at the end, because
4603 both PCRE and Perl have an optimization that allows for fast failure
4604 when a single character is used. They remember the last single charac-
4605 ter that is required for a match, and fail early if it is not present
4606 in the string.) If the pattern is changed so that it uses an atomic
4607 group, like this:
4608
4609 ((?>\D+)|<\d+>)*[!?]
4610
4611 sequences of non-digits cannot be broken, and failure happens quickly.
4612
4613
4614 BACK REFERENCES
4615
4616 Outside a character class, a backslash followed by a digit greater than
4617 0 (and possibly further digits) is a back reference to a capturing sub-
4618 pattern earlier (that is, to its left) in the pattern, provided there
4619 have been that many previous capturing left parentheses.
4620
4621 However, if the decimal number following the backslash is less than 10,
4622 it is always taken as a back reference, and causes an error only if
4623 there are not that many capturing left parentheses in the entire pat-
4624 tern. In other words, the parentheses that are referenced need not be
4625 to the left of the reference for numbers less than 10. A "forward back
4626 reference" of this type can make sense when a repetition is involved
4627 and the subpattern to the right has participated in an earlier itera-
4628 tion.
4629
4630 It is not possible to have a numerical "forward back reference" to a
4631 subpattern whose number is 10 or more using this syntax because a
4632 sequence such as \50 is interpreted as a character defined in octal.
4633 See the subsection entitled "Non-printing characters" above for further
4634 details of the handling of digits following a backslash. There is no
4635 such problem when named parentheses are used. A back reference to any
4636 subpattern is possible using named parentheses (see below).
4637
4638 Another way of avoiding the ambiguity inherent in the use of digits
4639 following a backslash is to use the \g escape sequence, which is a fea-
4640 ture introduced in Perl 5.10. This escape must be followed by an
4641 unsigned number or a negative number, optionally enclosed in braces.
4642 These examples are all identical:
4643
4644 (ring), \1
4645 (ring), \g1
4646 (ring), \g{1}
4647
4648 An unsigned number specifies an absolute reference without the ambigu-
4649 ity that is present in the older syntax. It is also useful when literal
4650 digits follow the reference. A negative number is a relative reference.
4651 Consider this example:
4652
4653 (abc(def)ghi)\g{-1}
4654
4655 The sequence \g{-1} is a reference to the most recently started captur-
4656 ing subpattern before \g, that is, is it equivalent to \2. Similarly,
4657 \g{-2} would be equivalent to \1. The use of relative references can be
4658 helpful in long patterns, and also in patterns that are created by
4659 joining together fragments that contain references within themselves.
4660
4661 A back reference matches whatever actually matched the capturing sub-
4662 pattern in the current subject string, rather than anything matching
4663 the subpattern itself (see "Subpatterns as subroutines" below for a way
4664 of doing that). So the pattern
4665
4666 (sens|respons)e and \1ibility
4667
4668 matches "sense and sensibility" and "response and responsibility", but
4669 not "sense and responsibility". If caseful matching is in force at the
4670 time of the back reference, the case of letters is relevant. For exam-
4671 ple,
4672
4673 ((?i)rah)\s+\1
4674
4675 matches "rah rah" and "RAH RAH", but not "RAH rah", even though the
4676 original capturing subpattern is matched caselessly.
4677
4678 There are several different ways of writing back references to named
4679 subpatterns. The .NET syntax \k{name} and the Perl syntax \k<name> or
4680 \k'name' are supported, as is the Python syntax (?P=name). Perl 5.10's
4681 unified back reference syntax, in which \g can be used for both numeric
4682 and named references, is also supported. We could rewrite the above
4683 example in any of the following ways:
4684
4685 (?<p1>(?i)rah)\s+\k<p1>
4686 (?'p1'(?i)rah)\s+\k{p1}
4687 (?P<p1>(?i)rah)\s+(?P=p1)
4688 (?<p1>(?i)rah)\s+\g{p1}
4689
4690 A subpattern that is referenced by name may appear in the pattern
4691 before or after the reference.
4692
4693 There may be more than one back reference to the same subpattern. If a
4694 subpattern has not actually been used in a particular match, any back
4695 references to it always fail by default. For example, the pattern
4696
4697 (a|(bc))\2
4698
4699 always fails if it starts to match "a" rather than "bc". However, if
4700 the PCRE_JAVASCRIPT_COMPAT option is set at compile time, a back refer-
4701 ence to an unset value matches an empty string.
4702
4703 Because there may be many capturing parentheses in a pattern, all dig-
4704 its following a backslash are taken as part of a potential back refer-
4705 ence number. If the pattern continues with a digit character, some
4706 delimiter must be used to terminate the back reference. If the
4707 PCRE_EXTENDED option is set, this can be whitespace. Otherwise, the \g{
4708 syntax or an empty comment (see "Comments" below) can be used.
4709
4710 Recursive back references
4711
4712 A back reference that occurs inside the parentheses to which it refers
4713 fails when the subpattern is first used, so, for example, (a\1) never
4714 matches. However, such references can be useful inside repeated sub-
4715 patterns. For example, the pattern
4716
4717 (a|b\1)+
4718
4719 matches any number of "a"s and also "aba", "ababbaa" etc. At each iter-
4720 ation of the subpattern, the back reference matches the character
4721 string corresponding to the previous iteration. In order for this to
4722 work, the pattern must be such that the first iteration does not need
4723 to match the back reference. This can be done using alternation, as in
4724 the example above, or by a quantifier with a minimum of zero.
4725
4726 Back references of this type cause the group that they reference to be
4727 treated as an atomic group. Once the whole group has been matched, a
4728 subsequent matching failure cannot cause backtracking into the middle
4729 of the group.
4730
4731
4732 ASSERTIONS
4733
4734 An assertion is a test on the characters following or preceding the
4735 current matching point that does not actually consume any characters.
4736 The simple assertions coded as \b, \B, \A, \G, \Z, \z, ^ and $ are
4737 described above.
4738
4739 More complicated assertions are coded as subpatterns. There are two
4740 kinds: those that look ahead of the current position in the subject
4741 string, and those that look behind it. An assertion subpattern is
4742 matched in the normal way, except that it does not cause the current
4743 matching position to be changed.
4744
4745 Assertion subpatterns are not capturing subpatterns, and may not be
4746 repeated, because it makes no sense to assert the same thing several
4747 times. If any kind of assertion contains capturing subpatterns within
4748 it, these are counted for the purposes of numbering the capturing sub-
4749 patterns in the whole pattern. However, substring capturing is carried
4750 out only for positive assertions, because it does not make sense for
4751 negative assertions.
4752
4753 Lookahead assertions
4754
4755 Lookahead assertions start with (?= for positive assertions and (?! for
4756 negative assertions. For example,
4757
4758 \w+(?=;)
4759
4760 matches a word followed by a semicolon, but does not include the semi-
4761 colon in the match, and
4762
4763 foo(?!bar)
4764
4765 matches any occurrence of "foo" that is not followed by "bar". Note
4766 that the apparently similar pattern
4767
4768 (?!foo)bar
4769
4770 does not find an occurrence of "bar" that is preceded by something
4771 other than "foo"; it finds any occurrence of "bar" whatsoever, because
4772 the assertion (?!foo) is always true when the next three characters are
4773 "bar". A lookbehind assertion is needed to achieve the other effect.
4774
4775 If you want to force a matching failure at some point in a pattern, the
4776 most convenient way to do it is with (?!) because an empty string
4777 always matches, so an assertion that requires there not to be an empty
4778 string must always fail. The Perl 5.10 backtracking control verb
4779 (*FAIL) or (*F) is essentially a synonym for (?!).
4780
4781 Lookbehind assertions
4782
4783 Lookbehind assertions start with (?<= for positive assertions and (?<!
4784 for negative assertions. For example,
4785
4786 (?<!foo)bar
4787
4788 does find an occurrence of "bar" that is not preceded by "foo". The
4789 contents of a lookbehind assertion are restricted such that all the
4790 strings it matches must have a fixed length. However, if there are sev-
4791 eral top-level alternatives, they do not all have to have the same
4792 fixed length. Thus
4793
4794 (?<=bullock|donkey)
4795
4796 is permitted, but
4797
4798 (?<!dogs?|cats?)
4799
4800 causes an error at compile time. Branches that match different length
4801 strings are permitted only at the top level of a lookbehind assertion.
4802 This is an extension compared with Perl (5.8 and 5.10), which requires
4803 all branches to match the same length of string. An assertion such as
4804
4805 (?<=ab(c|de))
4806
4807 is not permitted, because its single top-level branch can match two
4808 different lengths, but it is acceptable to PCRE if rewritten to use two
4809 top-level branches:
4810
4811 (?<=abc|abde)
4812
4813 In some cases, the Perl 5.10 escape sequence \K (see above) can be used
4814 instead of a lookbehind assertion to get round the fixed-length
4815 restriction.
4816
4817 The implementation of lookbehind assertions is, for each alternative,
4818 to temporarily move the current position back by the fixed length and
4819 then try to match. If there are insufficient characters before the cur-
4820 rent position, the assertion fails.
4821
4822 PCRE does not allow the \C escape (which matches a single byte in UTF-8
4823 mode) to appear in lookbehind assertions, because it makes it impossi-
4824 ble to calculate the length of the lookbehind. The \X and \R escapes,
4825 which can match different numbers of bytes, are also not permitted.
4826
4827 "Subroutine" calls (see below) such as (?2) or (?&X) are permitted in
4828 lookbehinds, as long as the subpattern matches a fixed-length string.
4829 Recursion, however, is not supported.
4830
4831 Possessive quantifiers can be used in conjunction with lookbehind
4832 assertions to specify efficient matching of fixed-length strings at the
4833 end of subject strings. Consider a simple pattern such as
4834
4835 abcd$
4836
4837 when applied to a long string that does not match. Because matching
4838 proceeds from left to right, PCRE will look for each "a" in the subject
4839 and then see if what follows matches the rest of the pattern. If the
4840 pattern is specified as
4841
4842 ^.*abcd$
4843
4844 the initial .* matches the entire string at first, but when this fails
4845 (because there is no following "a"), it backtracks to match all but the
4846 last character, then all but the last two characters, and so on. Once
4847 again the search for "a" covers the entire string, from right to left,
4848 so we are no better off. However, if the pattern is written as
4849
4850 ^.*+(?<=abcd)
4851
4852 there can be no backtracking for the .*+ item; it can match only the
4853 entire string. The subsequent lookbehind assertion does a single test
4854 on the last four characters. If it fails, the match fails immediately.
4855 For long strings, this approach makes a significant difference to the
4856 processing time.
4857
4858 Using multiple assertions
4859
4860 Several assertions (of any sort) may occur in succession. For example,
4861
4862 (?<=\d{3})(?<!999)foo
4863
4864 matches "foo" preceded by three digits that are not "999". Notice that
4865 each of the assertions is applied independently at the same point in
4866 the subject string. First there is a check that the previous three
4867 characters are all digits, and then there is a check that the same
4868 three characters are not "999". This pattern does not match "foo" pre-
4869 ceded by six characters, the first of which are digits and the last
4870 three of which are not "999". For example, it doesn't match "123abc-
4871 foo". A pattern to do that is
4872
4873 (?<=\d{3}...)(?<!999)foo
4874
4875 This time the first assertion looks at the preceding six characters,
4876 checking that the first three are digits, and then the second assertion
4877 checks that the preceding three characters are not "999".
4878
4879 Assertions can be nested in any combination. For example,
4880
4881 (?<=(?<!foo)bar)baz
4882
4883 matches an occurrence of "baz" that is preceded by "bar" which in turn
4884 is not preceded by "foo", while
4885
4886 (?<=\d{3}(?!999)...)foo
4887
4888 is another pattern that matches "foo" preceded by three digits and any
4889 three characters that are not "999".
4890
4891
4892 CONDITIONAL SUBPATTERNS
4893
4894 It is possible to cause the matching process to obey a subpattern con-
4895 ditionally or to choose between two alternative subpatterns, depending
4896 on the result of an assertion, or whether a specific capturing subpat-
4897 tern has already been matched. The two possible forms of conditional
4898 subpattern are:
4899
4900 (?(condition)yes-pattern)
4901 (?(condition)yes-pattern|no-pattern)
4902
4903 If the condition is satisfied, the yes-pattern is used; otherwise the
4904 no-pattern (if present) is used. If there are more than two alterna-
4905 tives in the subpattern, a compile-time error occurs. Each of the two
4906 alternatives may itself contain nested subpatterns of any form, includ-
4907 ing conditional subpatterns; the restriction to two alternatives
4908 applies only at the level of the condition. This pattern fragment is an
4909 example where the alternatives are complex:
4910
4911 (?(1) (A|B|C) | (D | (?(2)E|F) | E) )
4912
4913
4914 There are four kinds of condition: references to subpatterns, refer-
4915 ences to recursion, a pseudo-condition called DEFINE, and assertions.
4916
4917 Checking for a used subpattern by number
4918
4919 If the text between the parentheses consists of a sequence of digits,
4920 the condition is true if a capturing subpattern of that number has pre-
4921 viously matched. If there is more than one capturing subpattern with
4922 the same number (see the earlier section about duplicate subpattern
4923 numbers), the condition is true if any of them have been set. An alter-
4924 native notation is to precede the digits with a plus or minus sign. In
4925 this case, the subpattern number is relative rather than absolute. The
4926 most recently opened parentheses can be referenced by (?(-1), the next
4927 most recent by (?(-2), and so on. In looping constructs it can also
4928 make sense to refer to subsequent groups with constructs such as
4929 (?(+2).
4930
4931 Consider the following pattern, which contains non-significant white
4932 space to make it more readable (assume the PCRE_EXTENDED option) and to
4933 divide it into three parts for ease of discussion:
4934
4935 ( \( )? [^()]+ (?(1) \) )
4936
4937 The first part matches an optional opening parenthesis, and if that
4938 character is present, sets it as the first captured substring. The sec-
4939 ond part matches one or more characters that are not parentheses. The
4940 third part is a conditional subpattern that tests whether the first set
4941 of parentheses matched or not. If they did, that is, if subject started
4942 with an opening parenthesis, the condition is true, and so the yes-pat-
4943 tern is executed and a closing parenthesis is required. Otherwise,
4944 since no-pattern is not present, the subpattern matches nothing. In
4945 other words, this pattern matches a sequence of non-parentheses,
4946 optionally enclosed in parentheses.
4947
4948 If you were embedding this pattern in a larger one, you could use a
4949 relative reference:
4950
4951 ...other stuff... ( \( )? [^()]+ (?(-1) \) ) ...
4952
4953 This makes the fragment independent of the parentheses in the larger
4954 pattern.
4955
4956 Checking for a used subpattern by name
4957
4958 Perl uses the syntax (?(<name>)...) or (?('name')...) to test for a
4959 used subpattern by name. For compatibility with earlier versions of
4960 PCRE, which had this facility before Perl, the syntax (?(name)...) is
4961 also recognized. However, there is a possible ambiguity with this syn-
4962 tax, because subpattern names may consist entirely of digits. PCRE
4963 looks first for a named subpattern; if it cannot find one and the name
4964 consists entirely of digits, PCRE looks for a subpattern of that num-
4965 ber, which must be greater than zero. Using subpattern names that con-
4966 sist entirely of digits is not recommended.
4967
4968 Rewriting the above example to use a named subpattern gives this:
4969
4970 (?<OPEN> \( )? [^()]+ (?(<OPEN>) \) )
4971
4972 If the name used in a condition of this kind is a duplicate, the test
4973 is applied to all subpatterns of the same name, and is true if any one
4974 of them has matched.
4975
4976 Checking for pattern recursion
4977
4978 If the condition is the string (R), and there is no subpattern with the
4979 name R, the condition is true if a recursive call to the whole pattern
4980 or any subpattern has been made. If digits or a name preceded by amper-
4981 sand follow the letter R, for example:
4982
4983 (?(R3)...) or (?(R&name)...)
4984
4985 the condition is true if the most recent recursion is into a subpattern
4986 whose number or name is given. This condition does not check the entire
4987 recursion stack. If the name used in a condition of this kind is a
4988 duplicate, the test is applied to all subpatterns of the same name, and
4989 is true if any one of them is the most recent recursion.
4990
4991 At "top level", all these recursion test conditions are false. The
4992 syntax for recursive patterns is described below.
4993
4994 Defining subpatterns for use by reference only
4995
4996 If the condition is the string (DEFINE), and there is no subpattern
4997 with the name DEFINE, the condition is always false. In this case,
4998 there may be only one alternative in the subpattern. It is always
4999 skipped if control reaches this point in the pattern; the idea of
5000 DEFINE is that it can be used to define "subroutines" that can be ref-
5001 erenced from elsewhere. (The use of "subroutines" is described below.)
5002 For example, a pattern to match an IPv4 address could be written like
5003 this (ignore whitespace and line breaks):
5004
5005 (?(DEFINE) (?<byte> 2[0-4]\d | 25[0-5] | 1\d\d | [1-9]?\d) )
5006 \b (?&byte) (\.(?&byte)){3} \b
5007
5008 The first part of the pattern is a DEFINE group inside which a another
5009 group named "byte" is defined. This matches an individual component of
5010 an IPv4 address (a number less than 256). When matching takes place,
5011 this part of the pattern is skipped because DEFINE acts like a false
5012 condition. The rest of the pattern uses references to the named group
5013 to match the four dot-separated components of an IPv4 address, insist-
5014 ing on a word boundary at each end.
5015
5016 Assertion conditions
5017
5018 If the condition is not in any of the above formats, it must be an
5019 assertion. This may be a positive or negative lookahead or lookbehind
5020 assertion. Consider this pattern, again containing non-significant
5021 white space, and with the two alternatives on the second line:
5022
5023 (?(?=[^a-z]*[a-z])
5024 \d{2}-[a-z]{3}-\d{2} | \d{2}-\d{2}-\d{2} )
5025
5026 The condition is a positive lookahead assertion that matches an
5027 optional sequence of non-letters followed by a letter. In other words,
5028 it tests for the presence of at least one letter in the subject. If a
5029 letter is found, the subject is matched against the first alternative;
5030 otherwise it is matched against the second. This pattern matches
5031 strings in one of the two forms dd-aaa-dd or dd-dd-dd, where aaa are
5032 letters and dd are digits.
5033
5034
5035 COMMENTS
5036
5037 There are two ways of including comments in patterns that are processed
5038 by PCRE. In both cases, the start of the comment must not be in a char-
5039 acter class, nor in the middle of any other sequence of related charac-
5040 ters such as (?: or a subpattern name or number. The characters that
5041 make up a comment play no part in the pattern matching.
5042
5043 The sequence (?# marks the start of a comment that continues up to the
5044 next closing parenthesis. Nested parentheses are not permitted. If the
5045 PCRE_EXTENDED option is set, an unescaped # character also introduces a
5046 comment, which in this case continues to immediately after the next
5047 newline character or character sequence in the pattern. Which charac-
5048 ters are interpreted as newlines is controlled by the options passed to
5049 pcre_compile() or by a special sequence at the start of the pattern, as
5050 described in the section entitled "Newline conventions" above. Note
5051 that end of this type of comment is a literal newline sequence in the
5052 pattern; escape sequences that happen to represent a newline do not
5053 count. For example, consider this pattern when PCRE_EXTENDED is set,
5054 and the default newline convention is in force:
5055
5056 abc #comment \n still comment
5057
5058 On encountering the # character, pcre_compile() skips along, looking
5059 for a newline in the pattern. The sequence \n is still literal at this
5060 stage, so it does not terminate the comment. Only an actual character
5061 with the code value 0x0a (the default newline) does so.
5062
5063
5064 RECURSIVE PATTERNS
5065
5066 Consider the problem of matching a string in parentheses, allowing for
5067 unlimited nested parentheses. Without the use of recursion, the best
5068 that can be done is to use a pattern that matches up to some fixed
5069 depth of nesting. It is not possible to handle an arbitrary nesting
5070 depth.
5071
5072 For some time, Perl has provided a facility that allows regular expres-
5073 sions to recurse (amongst other things). It does this by interpolating
5074 Perl code in the expression at run time, and the code can refer to the
5075 expression itself. A Perl pattern using code interpolation to solve the
5076 parentheses problem can be created like this:
5077
5078 $re = qr{\( (?: (?>[^()]+) | (?p{$re}) )* \)}x;
5079
5080 The (?p{...}) item interpolates Perl code at run time, and in this case
5081 refers recursively to the pattern in which it appears.
5082
5083 Obviously, PCRE cannot support the interpolation of Perl code. Instead,
5084 it supports special syntax for recursion of the entire pattern, and
5085 also for individual subpattern recursion. After its introduction in
5086 PCRE and Python, this kind of recursion was subsequently introduced
5087 into Perl at release 5.10.
5088
5089 A special item that consists of (? followed by a number greater than
5090 zero and a closing parenthesis is a recursive call of the subpattern of
5091 the given number, provided that it occurs inside that subpattern. (If
5092 not, it is a "subroutine" call, which is described in the next sec-
5093 tion.) The special item (?R) or (?0) is a recursive call of the entire
5094 regular expression.
5095
5096 This PCRE pattern solves the nested parentheses problem (assume the
5097 PCRE_EXTENDED option is set so that white space is ignored):
5098
5099 \( ( [^()]++ | (?R) )* \)
5100
5101 First it matches an opening parenthesis. Then it matches any number of
5102 substrings which can either be a sequence of non-parentheses, or a
5103 recursive match of the pattern itself (that is, a correctly parenthe-
5104 sized substring). Finally there is a closing parenthesis. Note the use
5105 of a possessive quantifier to avoid backtracking into sequences of non-
5106 parentheses.
5107
5108 If this were part of a larger pattern, you would not want to recurse
5109 the entire pattern, so instead you could use this:
5110
5111 ( \( ( [^()]++ | (?1) )* \) )
5112
5113 We have put the pattern into parentheses, and caused the recursion to
5114 refer to them instead of the whole pattern.
5115
5116 In a larger pattern, keeping track of parenthesis numbers can be
5117 tricky. This is made easier by the use of relative references (a Perl
5118 5.10 feature). Instead of (?1) in the pattern above you can write
5119 (?-2) to refer to the second most recently opened parentheses preceding
5120 the recursion. In other words, a negative number counts capturing
5121 parentheses leftwards from the point at which it is encountered.
5122
5123 It is also possible to refer to subsequently opened parentheses, by
5124 writing references such as (?+2). However, these cannot be recursive
5125 because the reference is not inside the parentheses that are refer-
5126 enced. They are always "subroutine" calls, as described in the next
5127 section.
5128
5129 An alternative approach is to use named parentheses instead. The Perl
5130 syntax for this is (?&name); PCRE's earlier syntax (?P>name) is also
5131 supported. We could rewrite the above example as follows:
5132
5133 (?<pn> \( ( [^()]++ | (?&pn) )* \) )
5134
5135 If there is more than one subpattern with the same name, the earliest
5136 one is used.
5137
5138 This particular example pattern that we have been looking at contains
5139 nested unlimited repeats, and so the use of a possessive quantifier for
5140 matching strings of non-parentheses is important when applying the pat-
5141 tern to strings that do not match. For example, when this pattern is
5142 applied to
5143
5144 (aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa()
5145
5146 it yields "no match" quickly. However, if a possessive quantifier is
5147 not used, the match runs for a very long time indeed because there are
5148 so many different ways the + and * repeats can carve up the subject,
5149 and all have to be tested before failure can be reported.
5150
5151 At the end of a match, the values of capturing parentheses are those
5152 from the outermost level. If you want to obtain intermediate values, a
5153 callout function can be used (see below and the pcrecallout documenta-
5154 tion). If the pattern above is matched against
5155
5156 (ab(cd)ef)
5157
5158 the value for the inner capturing parentheses (numbered 2) is "ef",
5159 which is the last value taken on at the top level. If a capturing sub-
5160 pattern is not matched at the top level, its final value is unset, even
5161 if it is (temporarily) set at a deeper level.
5162
5163 If there are more than 15 capturing parentheses in a pattern, PCRE has
5164 to obtain extra memory to store data during a recursion, which it does
5165 by using pcre_malloc, freeing it via pcre_free afterwards. If no memory
5166 can be obtained, the match fails with the PCRE_ERROR_NOMEMORY error.
5167
5168 Do not confuse the (?R) item with the condition (R), which tests for
5169 recursion. Consider this pattern, which matches text in angle brack-
5170 ets, allowing for arbitrary nesting. Only digits are allowed in nested
5171 brackets (that is, when recursing), whereas any characters are permit-
5172 ted at the outer level.
5173
5174 < (?: (?(R) \d++ | [^<>]*+) | (?R)) * >
5175
5176 In this pattern, (?(R) is the start of a conditional subpattern, with
5177 two different alternatives for the recursive and non-recursive cases.
5178 The (?R) item is the actual recursive call.
5179
5180 Recursion difference from Perl
5181
5182 In PCRE (like Python, but unlike Perl), a recursive subpattern call is
5183 always treated as an atomic group. That is, once it has matched some of
5184 the subject string, it is never re-entered, even if it contains untried
5185 alternatives and there is a subsequent matching failure. This can be
5186 illustrated by the following pattern, which purports to match a palin-
5187 dromic string that contains an odd number of characters (for example,
5188 "a", "aba", "abcba", "abcdcba"):
5189
5190 ^(.|(.)(?1)\2)$
5191
5192 The idea is that it either matches a single character, or two identical
5193 characters surrounding a sub-palindrome. In Perl, this pattern works;
5194 in PCRE it does not if the pattern is longer than three characters.
5195 Consider the subject string "abcba":
5196
5197 At the top level, the first character is matched, but as it is not at
5198 the end of the string, the first alternative fails; the second alterna-
5199 tive is taken and the recursion kicks in. The recursive call to subpat-
5200 tern 1 successfully matches the next character ("b"). (Note that the
5201 beginning and end of line tests are not part of the recursion).
5202
5203 Back at the top level, the next character ("c") is compared with what
5204 subpattern 2 matched, which was "a". This fails. Because the recursion
5205 is treated as an atomic group, there are now no backtracking points,
5206 and so the entire match fails. (Perl is able, at this point, to re-
5207 enter the recursion and try the second alternative.) However, if the
5208 pattern is written with the alternatives in the other order, things are
5209 different:
5210
5211 ^((.)(?1)\2|.)$
5212
5213 This time, the recursing alternative is tried first, and continues to
5214 recurse until it runs out of characters, at which point the recursion
5215 fails. But this time we do have another alternative to try at the
5216 higher level. That is the big difference: in the previous case the
5217 remaining alternative is at a deeper recursion level, which PCRE cannot
5218 use.
5219
5220 To change the pattern so that matches all palindromic strings, not just
5221 those with an odd number of characters, it is tempting to change the
5222 pattern to this:
5223
5224 ^((.)(?1)\2|.?)$
5225
5226 Again, this works in Perl, but not in PCRE, and for the same reason.
5227 When a deeper recursion has matched a single character, it cannot be
5228 entered again in order to match an empty string. The solution is to
5229 separate the two cases, and write out the odd and even cases as alter-
5230 natives at the higher level:
5231
5232 ^(?:((.)(?1)\2|)|((.)(?3)\4|.))
5233
5234 If you want to match typical palindromic phrases, the pattern has to
5235 ignore all non-word characters, which can be done like this:
5236
5237 ^\W*+(?:((.)\W*+(?1)\W*+\2|)|((.)\W*+(?3)\W*+\4|\W*+.\W*+))\W*+$
5238
5239 If run with the PCRE_CASELESS option, this pattern matches phrases such
5240 as "A man, a plan, a canal: Panama!" and it works well in both PCRE and
5241 Perl. Note the use of the possessive quantifier *+ to avoid backtrack-
5242 ing into sequences of non-word characters. Without this, PCRE takes a
5243 great deal longer (ten times or more) to match typical phrases, and
5244 Perl takes so long that you think it has gone into a loop.
5245
5246 WARNING: The palindrome-matching patterns above work only if the sub-
5247 ject string does not start with a palindrome that is shorter than the
5248 entire string. For example, although "abcba" is correctly matched, if
5249 the subject is "ababa", PCRE finds the palindrome "aba" at the start,
5250 then fails at top level because the end of the string does not follow.
5251 Once again, it cannot jump back into the recursion to try other alter-
5252 natives, so the entire match fails.
5253
5254
5255 SUBPATTERNS AS SUBROUTINES
5256
5257 If the syntax for a recursive subpattern reference (either by number or
5258 by name) is used outside the parentheses to which it refers, it oper-
5259 ates like a subroutine in a programming language. The "called" subpat-
5260 tern may be defined before or after the reference. A numbered reference
5261 can be absolute or relative, as in these examples:
5262
5263 (...(absolute)...)...(?2)...
5264 (...(relative)...)...(?-1)...
5265 (...(?+1)...(relative)...
5266
5267 An earlier example pointed out that the pattern
5268
5269 (sens|respons)e and \1ibility
5270
5271 matches "sense and sensibility" and "response and responsibility", but
5272 not "sense and responsibility". If instead the pattern
5273
5274 (sens|respons)e and (?1)ibility
5275
5276 is used, it does match "sense and responsibility" as well as the other
5277 two strings. Another example is given in the discussion of DEFINE
5278 above.
5279
5280 Like recursive subpatterns, a subroutine call is always treated as an
5281 atomic group. That is, once it has matched some of the subject string,
5282 it is never re-entered, even if it contains untried alternatives and
5283 there is a subsequent matching failure. Any capturing parentheses that
5284 are set during the subroutine call revert to their previous values
5285 afterwards.
5286
5287 When a subpattern is used as a subroutine, processing options such as
5288 case-independence are fixed when the subpattern is defined. They cannot
5289 be changed for different calls. For example, consider this pattern:
5290
5291 (abc)(?i:(?-1))
5292
5293 It matches "abcabc". It does not match "abcABC" because the change of
5294 processing option does not affect the called subpattern.
5295
5296
5297 ONIGURUMA SUBROUTINE SYNTAX
5298
5299 For compatibility with Oniguruma, the non-Perl syntax \g followed by a
5300 name or a number enclosed either in angle brackets or single quotes, is
5301 an alternative syntax for referencing a subpattern as a subroutine,
5302 possibly recursively. Here are two of the examples used above, rewrit-
5303 ten using this syntax:
5304
5305 (?<pn> \( ( (?>[^()]+) | \g<pn> )* \) )
5306 (sens|respons)e and \g'1'ibility
5307
5308 PCRE supports an extension to Oniguruma: if a number is preceded by a
5309 plus or a minus sign it is taken as a relative reference. For example:
5310
5311 (abc)(?i:\g<-1>)
5312
5313 Note that \g{...} (Perl syntax) and \g<...> (Oniguruma syntax) are not
5314 synonymous. The former is a back reference; the latter is a subroutine
5315 call.
5316
5317
5318 CALLOUTS
5319
5320 Perl has a feature whereby using the sequence (?{...}) causes arbitrary
5321 Perl code to be obeyed in the middle of matching a regular expression.
5322 This makes it possible, amongst other things, to extract different sub-
5323 strings that match the same pair of parentheses when there is a repeti-
5324 tion.
5325
5326 PCRE provides a similar feature, but of course it cannot obey arbitrary
5327 Perl code. The feature is called "callout". The caller of PCRE provides
5328 an external function by putting its entry point in the global variable
5329 pcre_callout. By default, this variable contains NULL, which disables
5330 all calling out.
5331
5332 Within a regular expression, (?C) indicates the points at which the
5333 external function is to be called. If you want to identify different
5334 callout points, you can put a number less than 256 after the letter C.
5335 The default value is zero. For example, this pattern has two callout
5336 points:
5337
5338 (?C1)abc(?C2)def
5339
5340 If the PCRE_AUTO_CALLOUT flag is passed to pcre_compile(), callouts are
5341 automatically installed before each item in the pattern. They are all
5342 numbered 255.
5343
5344 During matching, when PCRE reaches a callout point (and pcre_callout is
5345 set), the external function is called. It is provided with the number
5346 of the callout, the position in the pattern, and, optionally, one item
5347 of data originally supplied by the caller of pcre_exec(). The callout
5348 function may cause matching to proceed, to backtrack, or to fail alto-
5349 gether. A complete description of the interface to the callout function
5350 is given in the pcrecallout documentation.
5351
5352
5353 BACKTRACKING CONTROL
5354
5355 Perl 5.10 introduced a number of "Special Backtracking Control Verbs",
5356 which are described in the Perl documentation as "experimental and sub-
5357 ject to change or removal in a future version of Perl". It goes on to
5358 say: "Their usage in production code should be noted to avoid problems
5359 during upgrades." The same remarks apply to the PCRE features described
5360 in this section.
5361
5362 Since these verbs are specifically related to backtracking, most of
5363 them can be used only when the pattern is to be matched using
5364 pcre_exec(), which uses a backtracking algorithm. With the exception of
5365 (*FAIL), which behaves like a failing negative assertion, they cause an
5366 error if encountered by pcre_dfa_exec().
5367
5368 If any of these verbs are used in an assertion or subroutine subpattern
5369 (including recursive subpatterns), their effect is confined to that
5370 subpattern; it does not extend to the surrounding pattern. Note that
5371 such subpatterns are processed as anchored at the point where they are
5372 tested.
5373
5374 The new verbs make use of what was previously invalid syntax: an open-
5375 ing parenthesis followed by an asterisk. They are generally of the form
5376 (*VERB) or (*VERB:NAME). Some may take either form, with differing be-
5377 haviour, depending on whether or not an argument is present. An name is
5378 a sequence of letters, digits, and underscores. If the name is empty,
5379 that is, if the closing parenthesis immediately follows the colon, the
5380 effect is as if the colon were not there. Any number of these verbs may
5381 occur in a pattern.
5382
5383 PCRE contains some optimizations that are used to speed up matching by
5384 running some checks at the start of each match attempt. For example, it
5385 may know the minimum length of matching subject, or that a particular
5386 character must be present. When one of these optimizations suppresses
5387 the running of a match, any included backtracking verbs will not, of
5388 course, be processed. You can suppress the start-of-match optimizations
5389 by setting the PCRE_NO_START_OPTIMIZE option when calling pcre_exec().
5390
5391 Verbs that act immediately
5392
5393 The following verbs act as soon as they are encountered. They may not
5394 be followed by a name.
5395
5396 (*ACCEPT)
5397
5398 This verb causes the match to end successfully, skipping the remainder
5399 of the pattern. When inside a recursion, only the innermost pattern is
5400 ended immediately. If (*ACCEPT) is inside capturing parentheses, the
5401 data so far is captured. (This feature was added to PCRE at release
5402 8.00.) For example:
5403
5404 A((?:A|B(*ACCEPT)|C)D)
5405
5406 This matches "AB", "AAD", or "ACD"; when it matches "AB", "B" is cap-
5407 tured by the outer parentheses.
5408
5409 (*FAIL) or (*F)
5410
5411 This verb causes the match to fail, forcing backtracking to occur. It
5412 is equivalent to (?!) but easier to read. The Perl documentation notes
5413 that it is probably useful only when combined with (?{}) or (??{}).
5414 Those are, of course, Perl features that are not present in PCRE. The
5415 nearest equivalent is the callout feature, as for example in this pat-
5416 tern:
5417
5418 a+(?C)(*FAIL)
5419
5420 A match with the string "aaaa" always fails, but the callout is taken
5421 before each backtrack happens (in this example, 10 times).
5422
5423 Recording which path was taken
5424
5425 There is one verb whose main purpose is to track how a match was
5426 arrived at, though it also has a secondary use in conjunction with
5427 advancing the match starting point (see (*SKIP) below).
5428
5429 (*MARK:NAME) or (*:NAME)
5430
5431 A name is always required with this verb. There may be as many
5432 instances of (*MARK) as you like in a pattern, and their names do not
5433 have to be unique.
5434
5435 When a match succeeds, the name of the last-encountered (*MARK) is
5436 passed back to the caller via the pcre_extra data structure, as
5437 described in the section on pcre_extra in the pcreapi documentation. No
5438 data is returned for a partial match. Here is an example of pcretest
5439 output, where the /K modifier requests the retrieval and outputting of
5440 (*MARK) data:
5441
5442 /X(*MARK:A)Y|X(*MARK:B)Z/K
5443 XY
5444 0: XY
5445 MK: A
5446 XZ
5447 0: XZ
5448 MK: B
5449
5450 The (*MARK) name is tagged with "MK:" in this output, and in this exam-
5451 ple it indicates which of the two alternatives matched. This is a more
5452 efficient way of obtaining this information than putting each alterna-
5453 tive in its own capturing parentheses.
5454
5455 A name may also be returned after a failed match if the final path
5456 through the pattern involves (*MARK). However, unless (*MARK) used in
5457 conjunction with (*COMMIT), this is unlikely to happen for an unan-
5458 chored pattern because, as the starting point for matching is advanced,
5459 the final check is often with an empty string, causing a failure before
5460 (*MARK) is reached. For example:
5461
5462 /X(*MARK:A)Y|X(*MARK:B)Z/K
5463 XP
5464 No match
5465
5466 There are three potential starting points for this match (starting with
5467 X, starting with P, and with an empty string). If the pattern is
5468 anchored, the result is different:
5469
5470 /^X(*MARK:A)Y|^X(*MARK:B)Z/K
5471 XP
5472 No match, mark = B
5473
5474 PCRE's start-of-match optimizations can also interfere with this. For
5475 example, if, as a result of a call to pcre_study(), it knows the mini-
5476 mum subject length for a match, a shorter subject will not be scanned
5477 at all.
5478
5479 Note that similar anomalies (though different in detail) exist in Perl,
5480 no doubt for the same reasons. The use of (*MARK) data after a failed
5481 match of an unanchored pattern is not recommended, unless (*COMMIT) is
5482 involved.
5483
5484 Verbs that act after backtracking
5485
5486 The following verbs do nothing when they are encountered. Matching con-
5487 tinues with what follows, but if there is no subsequent match, causing
5488 a backtrack to the verb, a failure is forced. That is, backtracking
5489 cannot pass to the left of the verb. However, when one of these verbs
5490 appears inside an atomic group, its effect is confined to that group,
5491 because once the group has been matched, there is never any backtrack-
5492 ing into it. In this situation, backtracking can "jump back" to the
5493 left of the entire atomic group. (Remember also, as stated above, that
5494 this localization also applies in subroutine calls and assertions.)
5495
5496 These verbs differ in exactly what kind of failure occurs when back-
5497 tracking reaches them.
5498
5499 (*COMMIT)
5500
5501 This verb, which may not be followed by a name, causes the whole match
5502 to fail outright if the rest of the pattern does not match. Even if the
5503 pattern is unanchored, no further attempts to find a match by advancing
5504 the starting point take place. Once (*COMMIT) has been passed,
5505 pcre_exec() is committed to finding a match at the current starting
5506 point, or not at all. For example:
5507
5508 a+(*COMMIT)b
5509
5510 This matches "xxaab" but not "aacaab". It can be thought of as a kind
5511 of dynamic anchor, or "I've started, so I must finish." The name of the
5512 most recently passed (*MARK) in the path is passed back when (*COMMIT)
5513 forces a match failure.
5514
5515 Note that (*COMMIT) at the start of a pattern is not the same as an
5516 anchor, unless PCRE's start-of-match optimizations are turned off, as
5517 shown in this pcretest example:
5518
5519 /(*COMMIT)abc/
5520 xyzabc
5521 0: abc
5522 xyzabc\Y
5523 No match
5524
5525 PCRE knows that any match must start with "a", so the optimization
5526 skips along the subject to "a" before running the first match attempt,
5527 which succeeds. When the optimization is disabled by the \Y escape in
5528 the second subject, the match starts at "x" and so the (*COMMIT) causes
5529 it to fail without trying any other starting points.
5530
5531 (*PRUNE) or (*PRUNE:NAME)
5532
5533 This verb causes the match to fail at the current starting position in
5534 the subject if the rest of the pattern does not match. If the pattern
5535 is unanchored, the normal "bumpalong" advance to the next starting
5536 character then happens. Backtracking can occur as usual to the left of
5537 (*PRUNE), before it is reached, or when matching to the right of
5538 (*PRUNE), but if there is no match to the right, backtracking cannot
5539 cross (*PRUNE). In simple cases, the use of (*PRUNE) is just an alter-
5540 native to an atomic group or possessive quantifier, but there are some
5541 uses of (*PRUNE) that cannot be expressed in any other way. The behav-
5542 iour of (*PRUNE:NAME) is the same as (*MARK:NAME)(*PRUNE) when the
5543 match fails completely; the name is passed back if this is the final
5544 attempt. (*PRUNE:NAME) does not pass back a name if the match suc-
5545 ceeds. In an anchored pattern (*PRUNE) has the same effect as (*COM-
5546 MIT).
5547
5548 (*SKIP)
5549
5550 This verb, when given without a name, is like (*PRUNE), except that if
5551 the pattern is unanchored, the "bumpalong" advance is not to the next
5552 character, but to the position in the subject where (*SKIP) was encoun-
5553 tered. (*SKIP) signifies that whatever text was matched leading up to
5554 it cannot be part of a successful match. Consider:
5555
5556 a+(*SKIP)b
5557
5558 If the subject is "aaaac...", after the first match attempt fails
5559 (starting at the first character in the string), the starting point
5560 skips on to start the next attempt at "c". Note that a possessive quan-
5561 tifer does not have the same effect as this example; although it would
5562 suppress backtracking during the first match attempt, the second
5563 attempt would start at the second character instead of skipping on to
5564 "c".
5565
5566 (*SKIP:NAME)
5567
5568 When (*SKIP) has an associated name, its behaviour is modified. If the
5569 following pattern fails to match, the previous path through the pattern
5570 is searched for the most recent (*MARK) that has the same name. If one
5571 is found, the "bumpalong" advance is to the subject position that cor-
5572 responds to that (*MARK) instead of to where (*SKIP) was encountered.
5573 If no (*MARK) with a matching name is found, normal "bumpalong" of one
5574 character happens (the (*SKIP) is ignored).
5575
5576 (*THEN) or (*THEN:NAME)
5577
5578 This verb causes a skip to the next alternation in the innermost
5579 enclosing group if the rest of the pattern does not match. That is, it
5580 cancels pending backtracking, but only within the current alternation.
5581 Its name comes from the observation that it can be used for a pattern-
5582 based if-then-else block:
5583
5584 ( COND1 (*THEN) FOO | COND2 (*THEN) BAR | COND3 (*THEN) BAZ ) ...
5585
5586 If the COND1 pattern matches, FOO is tried (and possibly further items
5587 after the end of the group if FOO succeeds); on failure the matcher
5588 skips to the second alternative and tries COND2, without backtracking
5589 into COND1. The behaviour of (*THEN:NAME) is exactly the same as
5590 (*MARK:NAME)(*THEN) if the overall match fails. If (*THEN) is not
5591 directly inside an alternation, it acts like (*PRUNE).
5592
5593 The above verbs provide four different "strengths" of control when sub-
5594 sequent matching fails. (*THEN) is the weakest, carrying on the match
5595 at the next alternation. (*PRUNE) comes next, failing the match at the
5596 current starting position, but allowing an advance to the next charac-
5597 ter (for an unanchored pattern). (*SKIP) is similar, except that the
5598 advance may be more than one character. (*COMMIT) is the strongest,
5599 causing the entire match to fail.
5600
5601 If more than one is present in a pattern, the "stongest" one wins. For
5602 example, consider this pattern, where A, B, etc. are complex pattern
5603 fragments:
5604
5605 (A(*COMMIT)B(*THEN)C|D)
5606
5607 Once A has matched, PCRE is committed to this match, at the current
5608 starting position. If subsequently B matches, but C does not, the nor-
5609 mal (*THEN) action of trying the next alternation (that is, D) does not
5610 happen because (*COMMIT) overrides.
5611
5612
5613 SEE ALSO
5614
5615 pcreapi(3), pcrecallout(3), pcrematching(3), pcresyntax(3), pcre(3).
5616
5617
5618 AUTHOR
5619
5620 Philip Hazel
5621 University Computing Service
5622 Cambridge CB2 3QH, England.
5623
5624
5625 REVISION
5626
5627 Last updated: 31 October 2010
5628 Copyright (c) 1997-2010 University of Cambridge.
5629 ------------------------------------------------------------------------------
5630
5631
5632 PCRESYNTAX(3) PCRESYNTAX(3)
5633
5634
5635 NAME
5636 PCRE - Perl-compatible regular expressions
5637
5638
5639 PCRE REGULAR EXPRESSION SYNTAX SUMMARY
5640
5641 The full syntax and semantics of the regular expressions that are sup-
5642 ported by PCRE are described in the pcrepattern documentation. This
5643 document contains just a quick-reference summary of the syntax.
5644
5645
5646 QUOTING
5647
5648 \x where x is non-alphanumeric is a literal x
5649 \Q...\E treat enclosed characters as literal
5650
5651
5652 CHARACTERS
5653
5654 \a alarm, that is, the BEL character (hex 07)
5655 \cx "control-x", where x is any character
5656 \e escape (hex 1B)
5657 \f formfeed (hex 0C)
5658 \n newline (hex 0A)
5659 \r carriage return (hex 0D)
5660 \t tab (hex 09)
5661 \ddd character with octal code ddd, or backreference
5662 \xhh character with hex code hh
5663 \x{hhh..} character with hex code hhh..
5664
5665
5666 CHARACTER TYPES
5667
5668 . any character except newline;
5669 in dotall mode, any character whatsoever
5670 \C one byte, even in UTF-8 mode (best avoided)
5671 \d a decimal digit
5672 \D a character that is not a decimal digit
5673 \h a horizontal whitespace character
5674 \H a character that is not a horizontal whitespace character
5675 \N a character that is not a newline
5676 \p{xx} a character with the xx property
5677 \P{xx} a character without the xx property
5678 \R a newline sequence
5679 \s a whitespace character
5680 \S a character that is not a whitespace character
5681 \v a vertical whitespace character
5682 \V a character that is not a vertical whitespace character
5683 \w a "word" character
5684 \W a "non-word" character
5685 \X an extended Unicode sequence
5686
5687 In PCRE, by default, \d, \D, \s, \S, \w, and \W recognize only ASCII
5688 characters, even in UTF-8 mode. However, this can be changed by setting
5689 the PCRE_UCP option.
5690
5691
5692 GENERAL CATEGORY PROPERTIES FOR \p and \P
5693
5694 C Other
5695 Cc Control
5696 Cf Format
5697 Cn Unassigned
5698 Co Private use
5699 Cs Surrogate
5700
5701 L Letter
5702 Ll Lower case letter
5703 Lm Modifier letter
5704 Lo Other letter
5705 Lt Title case letter
5706 Lu Upper case letter
5707 L& Ll, Lu, or Lt
5708
5709 M Mark
5710 Mc Spacing mark
5711 Me Enclosing mark
5712 Mn Non-spacing mark
5713
5714 N Number
5715 Nd Decimal number
5716 Nl Letter number
5717 No Other number
5718
5719 P Punctuation
5720 Pc Connector punctuation
5721 Pd Dash punctuation
5722 Pe Close punctuation
5723 Pf Final punctuation
5724 Pi Initial punctuation
5725 Po Other punctuation
5726 Ps Open punctuation
5727
5728 S Symbol
5729 Sc Currency symbol
5730 Sk Modifier symbol
5731 Sm Mathematical symbol
5732 So Other symbol
5733
5734 Z Separator
5735 Zl Line separator
5736 Zp Paragraph separator
5737 Zs Space separator
5738
5739
5740 PCRE SPECIAL CATEGORY PROPERTIES FOR \p and \P
5741
5742 Xan Alphanumeric: union of properties L and N
5743 Xps POSIX space: property Z or tab, NL, VT, FF, CR
5744 Xsp Perl space: property Z or tab, NL, FF, CR
5745 Xwd Perl word: property Xan or underscore
5746
5747
5748 SCRIPT NAMES FOR \p AND \P
5749
5750 Arabic, Armenian, Avestan, Balinese, Bamum, Bengali, Bopomofo, Braille,
5751 Buginese, Buhid, Canadian_Aboriginal, Carian, Cham, Cherokee, Common,
5752 Coptic, Cuneiform, Cypriot, Cyrillic, Deseret, Devanagari, Egyp-
5753 tian_Hieroglyphs, Ethiopic, Georgian, Glagolitic, Gothic, Greek,
5754 Gujarati, Gurmukhi, Han, Hangul, Hanunoo, Hebrew, Hiragana, Impe-
5755 rial_Aramaic, Inherited, Inscriptional_Pahlavi, Inscriptional_Parthian,
5756 Javanese, Kaithi, Kannada, Katakana, Kayah_Li, Kharoshthi, Khmer, Lao,
5757 Latin, Lepcha, Limbu, Linear_B, Lisu, Lycian, Lydian, Malayalam,
5758 Meetei_Mayek, Mongolian, Myanmar, New_Tai_Lue, Nko, Ogham, Old_Italic,
5759 Old_Persian, Old_South_Arabian, Old_Turkic, Ol_Chiki, Oriya, Osmanya,
5760 Phags_Pa, Phoenician, Rejang, Runic, Samaritan, Saurashtra, Shavian,
5761 Sinhala, Sundanese, Syloti_Nagri, Syriac, Tagalog, Tagbanwa, Tai_Le,
5762 Tai_Tham, Tai_Viet, Tamil, Telugu, Thaana, Thai, Tibetan, Tifinagh,
5763 Ugaritic, Vai, Yi.
5764
5765
5766 CHARACTER CLASSES
5767
5768 [...] positive character class
5769 [^...] negative character class
5770 [x-y] range (can be used for hex characters)
5771 [[:xxx:]] positive POSIX named set
5772 [[:^xxx:]] negative POSIX named set
5773
5774 alnum alphanumeric
5775 alpha alphabetic
5776 ascii 0-127
5777 blank space or tab
5778 cntrl control character
5779 digit decimal digit
5780 graph printing, excluding space
5781 lower lower case letter
5782 print printing, including space
5783 punct printing, excluding alphanumeric
5784 space whitespace
5785 upper upper case letter
5786 word same as \w
5787 xdigit hexadecimal digit
5788
5789 In PCRE, POSIX character set names recognize only ASCII characters by
5790 default, but some of them use Unicode properties if PCRE_UCP is set.
5791 You can use \Q...\E inside a character class.
5792
5793
5794 QUANTIFIERS
5795
5796 ? 0 or 1, greedy
5797 ?+ 0 or 1, possessive
5798 ?? 0 or 1, lazy
5799 * 0 or more, greedy
5800 *+ 0 or more, possessive
5801 *? 0 or more, lazy
5802 + 1 or more, greedy
5803 ++ 1 or more, possessive
5804 +? 1 or more, lazy
5805 {n} exactly n
5806 {n,m} at least n, no more than m, greedy
5807 {n,m}+ at least n, no more than m, possessive
5808 {n,m}? at least n, no more than m, lazy
5809 {n,} n or more, greedy
5810 {n,}+ n or more, possessive
5811 {n,}? n or more, lazy
5812
5813
5814 ANCHORS AND SIMPLE ASSERTIONS
5815
5816 \b word boundary
5817 \B not a word boundary
5818 ^ start of subject
5819 also after internal newline in multiline mode
5820 \A start of subject
5821 $ end of subject
5822 also before newline at end of subject
5823 also before internal newline in multiline mode
5824 \Z end of subject
5825 also before newline at end of subject
5826 \z end of subject
5827 \G first matching position in subject
5828
5829
5830 MATCH POINT RESET
5831
5832 \K reset start of match
5833
5834
5835 ALTERNATION
5836
5837 expr|expr|expr...
5838
5839
5840 CAPTURING
5841
5842 (...) capturing group
5843 (?<name>...) named capturing group (Perl)
5844 (?'name'...) named capturing group (Perl)
5845 (?P<name>...) named capturing group (Python)
5846 (?:...) non-capturing group
5847 (?|...) non-capturing group; reset group numbers for
5848 capturing groups in each alternative
5849
5850
5851 ATOMIC GROUPS
5852
5853 (?>...) atomic, non-capturing group
5854
5855
5856 COMMENT
5857
5858 (?#....) comment (not nestable)
5859
5860
5861 OPTION SETTING
5862
5863 (?i) caseless
5864 (?J) allow duplicate names
5865 (?m) multiline
5866 (?s) single line (dotall)
5867 (?U) default ungreedy (lazy)
5868 (?x) extended (ignore white space)
5869 (?-...) unset option(s)
5870
5871 The following are recognized only at the start of a pattern or after
5872 one of the newline-setting options with similar syntax:
5873
5874 (*UTF8) set UTF-8 mode (PCRE_UTF8)
5875 (*UCP) set PCRE_UCP (use Unicode properties for \d etc)
5876
5877
5878 LOOKAHEAD AND LOOKBEHIND ASSERTIONS
5879
5880 (?=...) positive look ahead
5881 (?!...) negative look ahead
5882 (?<=...) positive look behind
5883 (?<!...) negative look behind
5884
5885 Each top-level branch of a look behind must be of a fixed length.
5886
5887
5888 BACKREFERENCES
5889
5890 \n reference by number (can be ambiguous)
5891 \gn reference by number
5892 \g{n} reference by number
5893 \g{-n} relative reference by number
5894 \k<name> reference by name (Perl)
5895 \k'name' reference by name (Perl)
5896 \g{name} reference by name (Perl)
5897 \k{name} reference by name (.NET)
5898 (?P=name) reference by name (Python)
5899
5900
5901 SUBROUTINE REFERENCES (POSSIBLY RECURSIVE)
5902
5903 (?R) recurse whole pattern
5904 (?n) call subpattern by absolute number
5905 (?+n) call subpattern by relative number
5906 (?-n) call subpattern by relative number
5907 (?&name) call subpattern by name (Perl)
5908 (?P>name) call subpattern by name (Python)
5909 \g<name> call subpattern by name (Oniguruma)
5910 \g'name' call subpattern by name (Oniguruma)
5911 \g<n> call subpattern by absolute number (Oniguruma)
5912 \g'n' call subpattern by absolute number (Oniguruma)
5913 \g<+n> call subpattern by relative number (PCRE extension)
5914 \g'+n' call subpattern by relative number (PCRE extension)
5915 \g<-n> call subpattern by relative number (PCRE extension)
5916 \g'-n' call subpattern by relative number (PCRE extension)
5917
5918
5919 CONDITIONAL PATTERNS
5920
5921 (?(condition)yes-pattern)
5922 (?(condition)yes-pattern|no-pattern)
5923
5924 (?(n)... absolute reference condition
5925 (?(+n)... relative reference condition
5926 (?(-n)... relative reference condition
5927 (?(<name>)... named reference condition (Perl)
5928 (?('name')... named reference condition (Perl)
5929 (?(name)... named reference condition (PCRE)
5930 (?(R)... overall recursion condition
5931 (?(Rn)... specific group recursion condition
5932 (?(R&name)... specific recursion condition
5933 (?(DEFINE)... define subpattern for reference
5934 (?(assert)... assertion condition
5935
5936
5937 BACKTRACKING CONTROL
5938
5939 The following act immediately they are reached:
5940
5941 (*ACCEPT) force successful match
5942 (*FAIL) force backtrack; synonym (*F)
5943
5944 The following act only when a subsequent match failure causes a back-
5945 track to reach them. They all force a match failure, but they differ in
5946 what happens afterwards. Those that advance the start-of-match point do
5947 so only if the pattern is not anchored.
5948
5949 (*COMMIT) overall failure, no advance of starting point
5950 (*PRUNE) advance to next starting character
5951 (*SKIP) advance start to current matching position
5952 (*THEN) local failure, backtrack to next alternation
5953
5954
5955 NEWLINE CONVENTIONS
5956
5957 These are recognized only at the very start of the pattern or after a
5958 (*BSR_...) or (*UTF8) or (*UCP) option.
5959
5960 (*CR) carriage return only
5961 (*LF) linefeed only
5962 (*CRLF) carriage return followed by linefeed
5963 (*ANYCRLF) all three of the above
5964 (*ANY) any Unicode newline sequence
5965
5966
5967 WHAT \R MATCHES
5968
5969 These are recognized only at the very start of the pattern or after a
5970 (*...) option that sets the newline convention or UTF-8 or UCP mode.
5971
5972 (*BSR_ANYCRLF) CR, LF, or CRLF
5973 (*BSR_UNICODE) any Unicode newline sequence
5974
5975
5976 CALLOUTS
5977
5978 (?C) callout
5979 (?Cn) callout with data n
5980
5981
5982 SEE ALSO
5983
5984 pcrepattern(3), pcreapi(3), pcrecallout(3), pcrematching(3), pcre(3).
5985
5986
5987 AUTHOR
5988
5989 Philip Hazel
5990 University Computing Service
5991 Cambridge CB2 3QH, England.
5992
5993
5994 REVISION
5995
5996 Last updated: 12 May 2010
5997 Copyright (c) 1997-2010 University of Cambridge.
5998 ------------------------------------------------------------------------------
5999
6000
6001 PCREPARTIAL(3) PCREPARTIAL(3)
6002
6003
6004 NAME
6005 PCRE - Perl-compatible regular expressions
6006
6007
6008 PARTIAL MATCHING IN PCRE
6009
6010 In normal use of PCRE, if the subject string that is passed to
6011 pcre_exec() or pcre_dfa_exec() matches as far as it goes, but is too
6012 short to match the entire pattern, PCRE_ERROR_NOMATCH is returned.
6013 There are circumstances where it might be helpful to distinguish this
6014 case from other cases in which there is no match.
6015
6016 Consider, for example, an application where a human is required to type
6017 in data for a field with specific formatting requirements. An example
6018 might be a date in the form ddmmmyy, defined by this pattern:
6019
6020 ^\d?\d(jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)\d\d$
6021
6022 If the application sees the user's keystrokes one by one, and can check
6023 that what has been typed so far is potentially valid, it is able to
6024 raise an error as soon as a mistake is made, by beeping and not
6025 reflecting the character that has been typed, for example. This immedi-
6026 ate feedback is likely to be a better user interface than a check that
6027 is delayed until the entire string has been entered. Partial matching
6028 can also be useful when the subject string is very long and is not all
6029 available at once.
6030
6031 PCRE supports partial matching by means of the PCRE_PARTIAL_SOFT and
6032 PCRE_PARTIAL_HARD options, which can be set when calling pcre_exec() or
6033 pcre_dfa_exec(). For backwards compatibility, PCRE_PARTIAL is a synonym
6034 for PCRE_PARTIAL_SOFT. The essential difference between the two options
6035 is whether or not a partial match is preferred to an alternative com-
6036 plete match, though the details differ between the two matching func-
6037 tions. If both options are set, PCRE_PARTIAL_HARD takes precedence.
6038
6039 Setting a partial matching option disables two of PCRE's optimizations.
6040 PCRE remembers the last literal byte in a pattern, and abandons match-
6041 ing immediately if such a byte is not present in the subject string.
6042 This optimization cannot be used for a subject string that might match
6043 only partially. If the pattern was studied, PCRE knows the minimum