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

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

Parent Directory Parent Directory | Revision Log Revision Log


Revision 518 - (show annotations) (download)
Tue May 18 15:47:01 2010 UTC (4 years, 5 months ago) by ph10
File MIME type: text/plain
File size: 348560 byte(s)
Added PCRE_UCP and related stuff to make \w etc use Unicode properties.

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 that this
230 also applies to \b, because it is defined in terms of \w and \W. If you
231 really want to test for a wider sense of, say, "digit", you can use
232 explicit Unicode property tests such as \p{Nd}. Alternatively, if you
233 set the PCRE_UCP option, the way that the character escapes work is
234 changed so that Unicode properties are used to determine which charac-
235 ters match. There are more details in the section on generic character
236 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: 12 May 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. The pcrepartial documentation gives
775 details of partial matching.
776
777
778 DISADVANTAGES OF THE ALTERNATIVE ALGORITHM
779
780 The alternative algorithm suffers from a number of disadvantages:
781
782 1. It is substantially slower than the standard algorithm. This is
783 partly because it has to search for all possible matches, but is also
784 because it is less susceptible to optimization.
785
786 2. Capturing parentheses and back references are not supported.
787
788 3. Although atomic groups are supported, their use does not provide the
789 performance advantage that it does for the standard algorithm.
790
791
792 AUTHOR
793
794 Philip Hazel
795 University Computing Service
796 Cambridge CB2 3QH, England.
797
798
799 REVISION
800
801 Last updated: 29 September 2009
802 Copyright (c) 1997-2009 University of Cambridge.
803 ------------------------------------------------------------------------------
804
805
806 PCREAPI(3) PCREAPI(3)
807
808
809 NAME
810 PCRE - Perl-compatible regular expressions
811
812
813 PCRE NATIVE API
814
815 #include <pcre.h>
816
817 pcre *pcre_compile(const char *pattern, int options,
818 const char **errptr, int *erroffset,
819 const unsigned char *tableptr);
820
821 pcre *pcre_compile2(const char *pattern, int options,
822 int *errorcodeptr,
823 const char **errptr, int *erroffset,
824 const unsigned char *tableptr);
825
826 pcre_extra *pcre_study(const pcre *code, int options,
827 const char **errptr);
828
829 int pcre_exec(const pcre *code, const pcre_extra *extra,
830 const char *subject, int length, int startoffset,
831 int options, int *ovector, int ovecsize);
832
833 int pcre_dfa_exec(const pcre *code, const pcre_extra *extra,
834 const char *subject, int length, int startoffset,
835 int options, int *ovector, int ovecsize,
836 int *workspace, int wscount);
837
838 int pcre_copy_named_substring(const pcre *code,
839 const char *subject, int *ovector,
840 int stringcount, const char *stringname,
841 char *buffer, int buffersize);
842
843 int pcre_copy_substring(const char *subject, int *ovector,
844 int stringcount, int stringnumber, char *buffer,
845 int buffersize);
846
847 int pcre_get_named_substring(const pcre *code,
848 const char *subject, int *ovector,
849 int stringcount, const char *stringname,
850 const char **stringptr);
851
852 int pcre_get_stringnumber(const pcre *code,
853 const char *name);
854
855 int pcre_get_stringtable_entries(const pcre *code,
856 const char *name, char **first, char **last);
857
858 int pcre_get_substring(const char *subject, int *ovector,
859 int stringcount, int stringnumber,
860 const char **stringptr);
861
862 int pcre_get_substring_list(const char *subject,
863 int *ovector, int stringcount, const char ***listptr);
864
865 void pcre_free_substring(const char *stringptr);
866
867 void pcre_free_substring_list(const char **stringptr);
868
869 const unsigned char *pcre_maketables(void);
870
871 int pcre_fullinfo(const pcre *code, const pcre_extra *extra,
872 int what, void *where);
873
874 int pcre_info(const pcre *code, int *optptr, int *firstcharptr);
875
876 int pcre_refcount(pcre *code, int adjust);
877
878 int pcre_config(int what, void *where);
879
880 char *pcre_version(void);
881
882 void *(*pcre_malloc)(size_t);
883
884 void (*pcre_free)(void *);
885
886 void *(*pcre_stack_malloc)(size_t);
887
888 void (*pcre_stack_free)(void *);
889
890 int (*pcre_callout)(pcre_callout_block *);
891
892
893 PCRE API OVERVIEW
894
895 PCRE has its own native API, which is described in this document. There
896 are also some wrapper functions that correspond to the POSIX regular
897 expression API. These are described in the pcreposix documentation.
898 Both of these APIs define a set of C function calls. A C++ wrapper is
899 distributed with PCRE. It is documented in the pcrecpp page.
900
901 The native API C function prototypes are defined in the header file
902 pcre.h, and on Unix systems the library itself is called libpcre. It
903 can normally be accessed by adding -lpcre to the command for linking an
904 application that uses PCRE. The header file defines the macros
905 PCRE_MAJOR and PCRE_MINOR to contain the major and minor release num-
906 bers for the library. Applications can use these to include support
907 for different releases of PCRE.
908
909 The functions pcre_compile(), pcre_compile2(), pcre_study(), and
910 pcre_exec() are used for compiling and matching regular expressions in
911 a Perl-compatible manner. A sample program that demonstrates the sim-
912 plest way of using them is provided in the file called pcredemo.c in
913 the PCRE source distribution. A listing of this program is given in the
914 pcredemo documentation, and the pcresample documentation describes how
915 to compile and run it.
916
917 A second matching function, pcre_dfa_exec(), which is not Perl-compati-
918 ble, is also provided. This uses a different algorithm for the match-
919 ing. The alternative algorithm finds all possible matches (at a given
920 point in the subject), and scans the subject just once (unless there
921 are lookbehind assertions). However, this algorithm does not return
922 captured substrings. A description of the two matching algorithms and
923 their advantages and disadvantages is given in the pcrematching docu-
924 mentation.
925
926 In addition to the main compiling and matching functions, there are
927 convenience functions for extracting captured substrings from a subject
928 string that is matched by pcre_exec(). They are:
929
930 pcre_copy_substring()
931 pcre_copy_named_substring()
932 pcre_get_substring()
933 pcre_get_named_substring()
934 pcre_get_substring_list()
935 pcre_get_stringnumber()
936 pcre_get_stringtable_entries()
937
938 pcre_free_substring() and pcre_free_substring_list() are also provided,
939 to free the memory used for extracted strings.
940
941 The function pcre_maketables() is used to build a set of character
942 tables in the current locale for passing to pcre_compile(),
943 pcre_exec(), or pcre_dfa_exec(). This is an optional facility that is
944 provided for specialist use. Most commonly, no special tables are
945 passed, in which case internal tables that are generated when PCRE is
946 built are used.
947
948 The function pcre_fullinfo() is used to find out information about a
949 compiled pattern; pcre_info() is an obsolete version that returns only
950 some of the available information, but is retained for backwards com-
951 patibility. The function pcre_version() returns a pointer to a string
952 containing the version of PCRE and its date of release.
953
954 The function pcre_refcount() maintains a reference count in a data
955 block containing a compiled pattern. This is provided for the benefit
956 of object-oriented applications.
957
958 The global variables pcre_malloc and pcre_free initially contain the
959 entry points of the standard malloc() and free() functions, respec-
960 tively. PCRE calls the memory management functions via these variables,
961 so a calling program can replace them if it wishes to intercept the
962 calls. This should be done before calling any PCRE functions.
963
964 The global variables pcre_stack_malloc and pcre_stack_free are also
965 indirections to memory management functions. These special functions
966 are used only when PCRE is compiled to use the heap for remembering
967 data, instead of recursive function calls, when running the pcre_exec()
968 function. See the pcrebuild documentation for details of how to do
969 this. It is a non-standard way of building PCRE, for use in environ-
970 ments that have limited stacks. Because of the greater use of memory
971 management, it runs more slowly. Separate functions are provided so
972 that special-purpose external code can be used for this case. When
973 used, these functions are always called in a stack-like manner (last
974 obtained, first freed), and always for memory blocks of the same size.
975 There is a discussion about PCRE's stack usage in the pcrestack docu-
976 mentation.
977
978 The global variable pcre_callout initially contains NULL. It can be set
979 by the caller to a "callout" function, which PCRE will then call at
980 specified points during a matching operation. Details are given in the
981 pcrecallout documentation.
982
983
984 NEWLINES
985
986 PCRE supports five different conventions for indicating line breaks in
987 strings: a single CR (carriage return) character, a single LF (line-
988 feed) character, the two-character sequence CRLF, any of the three pre-
989 ceding, or any Unicode newline sequence. The Unicode newline sequences
990 are the three just mentioned, plus the single characters VT (vertical
991 tab, U+000B), FF (formfeed, U+000C), NEL (next line, U+0085), LS (line
992 separator, U+2028), and PS (paragraph separator, U+2029).
993
994 Each of the first three conventions is used by at least one operating
995 system as its standard newline sequence. When PCRE is built, a default
996 can be specified. The default default is LF, which is the Unix stan-
997 dard. When PCRE is run, the default can be overridden, either when a
998 pattern is compiled, or when it is matched.
999
1000 At compile time, the newline convention can be specified by the options
1001 argument of pcre_compile(), or it can be specified by special text at
1002 the start of the pattern itself; this overrides any other settings. See
1003 the pcrepattern page for details of the special character sequences.
1004
1005 In the PCRE documentation the word "newline" is used to mean "the char-
1006 acter or pair of characters that indicate a line break". The choice of
1007 newline convention affects the handling of the dot, circumflex, and
1008 dollar metacharacters, the handling of #-comments in /x mode, and, when
1009 CRLF is a recognized line ending sequence, the match position advance-
1010 ment for a non-anchored pattern. There is more detail about this in the
1011 section on pcre_exec() options below.
1012
1013 The choice of newline convention does not affect the interpretation of
1014 the \n or \r escape sequences, nor does it affect what \R matches,
1015 which is controlled in a similar way, but by separate options.
1016
1017
1018 MULTITHREADING
1019
1020 The PCRE functions can be used in multi-threading applications, with
1021 the proviso that the memory management functions pointed to by
1022 pcre_malloc, pcre_free, pcre_stack_malloc, and pcre_stack_free, and the
1023 callout function pointed to by pcre_callout, are shared by all threads.
1024
1025 The compiled form of a regular expression is not altered during match-
1026 ing, so the same compiled pattern can safely be used by several threads
1027 at once.
1028
1029
1030 SAVING PRECOMPILED PATTERNS FOR LATER USE
1031
1032 The compiled form of a regular expression can be saved and re-used at a
1033 later time, possibly by a different program, and even on a host other
1034 than the one on which it was compiled. Details are given in the
1035 pcreprecompile documentation. However, compiling a regular expression
1036 with one version of PCRE for use with a different version is not guar-
1037 anteed to work and may cause crashes.
1038
1039
1040 CHECKING BUILD-TIME OPTIONS
1041
1042 int pcre_config(int what, void *where);
1043
1044 The function pcre_config() makes it possible for a PCRE client to dis-
1045 cover which optional features have been compiled into the PCRE library.
1046 The pcrebuild documentation has more details about these optional fea-
1047 tures.
1048
1049 The first argument for pcre_config() is an integer, specifying which
1050 information is required; the second argument is a pointer to a variable
1051 into which the information is placed. The following information is
1052 available:
1053
1054 PCRE_CONFIG_UTF8
1055
1056 The output is an integer that is set to one if UTF-8 support is avail-
1057 able; otherwise it is set to zero.
1058
1059 PCRE_CONFIG_UNICODE_PROPERTIES
1060
1061 The output is an integer that is set to one if support for Unicode
1062 character properties is available; otherwise it is set to zero.
1063
1064 PCRE_CONFIG_NEWLINE
1065
1066 The output is an integer whose value specifies the default character
1067 sequence that is recognized as meaning "newline". The four values that
1068 are supported are: 10 for LF, 13 for CR, 3338 for CRLF, -2 for ANYCRLF,
1069 and -1 for ANY. Though they are derived from ASCII, the same values
1070 are returned in EBCDIC environments. The default should normally corre-
1071 spond to the standard sequence for your operating system.
1072
1073 PCRE_CONFIG_BSR
1074
1075 The output is an integer whose value indicates what character sequences
1076 the \R escape sequence matches by default. A value of 0 means that \R
1077 matches any Unicode line ending sequence; a value of 1 means that \R
1078 matches only CR, LF, or CRLF. The default can be overridden when a pat-
1079 tern is compiled or matched.
1080
1081 PCRE_CONFIG_LINK_SIZE
1082
1083 The output is an integer that contains the number of bytes used for
1084 internal linkage in compiled regular expressions. The value is 2, 3, or
1085 4. Larger values allow larger regular expressions to be compiled, at
1086 the expense of slower matching. The default value of 2 is sufficient
1087 for all but the most massive patterns, since it allows the compiled
1088 pattern to be up to 64K in size.
1089
1090 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
1091
1092 The output is an integer that contains the threshold above which the
1093 POSIX interface uses malloc() for output vectors. Further details are
1094 given in the pcreposix documentation.
1095
1096 PCRE_CONFIG_MATCH_LIMIT
1097
1098 The output is a long integer that gives the default limit for the num-
1099 ber of internal matching function calls in a pcre_exec() execution.
1100 Further details are given with pcre_exec() below.
1101
1102 PCRE_CONFIG_MATCH_LIMIT_RECURSION
1103
1104 The output is a long integer that gives the default limit for the depth
1105 of recursion when calling the internal matching function in a
1106 pcre_exec() execution. Further details are given with pcre_exec()
1107 below.
1108
1109 PCRE_CONFIG_STACKRECURSE
1110
1111 The output is an integer that is set to one if internal recursion when
1112 running pcre_exec() is implemented by recursive function calls that use
1113 the stack to remember their state. This is the usual way that PCRE is
1114 compiled. The output is zero if PCRE was compiled to use blocks of data
1115 on the heap instead of recursive function calls. In this case,
1116 pcre_stack_malloc and pcre_stack_free are called to manage memory
1117 blocks on the heap, thus avoiding the use of the stack.
1118
1119
1120 COMPILING A PATTERN
1121
1122 pcre *pcre_compile(const char *pattern, int options,
1123 const char **errptr, int *erroffset,
1124 const unsigned char *tableptr);
1125
1126 pcre *pcre_compile2(const char *pattern, int options,
1127 int *errorcodeptr,
1128 const char **errptr, int *erroffset,
1129 const unsigned char *tableptr);
1130
1131 Either of the functions pcre_compile() or pcre_compile2() can be called
1132 to compile a pattern into an internal form. The only difference between
1133 the two interfaces is that pcre_compile2() has an additional argument,
1134 errorcodeptr, via which a numerical error code can be returned. To
1135 avoid too much repetition, we refer just to pcre_compile() below, but
1136 the information applies equally to pcre_compile2().
1137
1138 The pattern is a C string terminated by a binary zero, and is passed in
1139 the pattern argument. A pointer to a single block of memory that is
1140 obtained via pcre_malloc is returned. This contains the compiled code
1141 and related data. The pcre type is defined for the returned block; this
1142 is a typedef for a structure whose contents are not externally defined.
1143 It is up to the caller to free the memory (via pcre_free) when it is no
1144 longer required.
1145
1146 Although the compiled code of a PCRE regex is relocatable, that is, it
1147 does not depend on memory location, the complete pcre data block is not
1148 fully relocatable, because it may contain a copy of the tableptr argu-
1149 ment, which is an address (see below).
1150
1151 The options argument contains various bit settings that affect the com-
1152 pilation. It should be zero if no options are required. The available
1153 options are described below. Some of them (in particular, those that
1154 are compatible with Perl, but some others as well) can also be set and
1155 unset from within the pattern (see the detailed description in the
1156 pcrepattern documentation). For those options that can be different in
1157 different parts of the pattern, the contents of the options argument
1158 specifies their settings at the start of compilation and execution. The
1159 PCRE_ANCHORED, PCRE_BSR_xxx, and PCRE_NEWLINE_xxx options can be set at
1160 the time of matching as well as at compile time.
1161
1162 If errptr is NULL, pcre_compile() returns NULL immediately. Otherwise,
1163 if compilation of a pattern fails, pcre_compile() returns NULL, and
1164 sets the variable pointed to by errptr to point to a textual error mes-
1165 sage. This is a static string that is part of the library. You must not
1166 try to free it. The byte offset from the start of the pattern to the
1167 character that was being processed when the error was discovered is
1168 placed in the variable pointed to by erroffset, which must not be NULL.
1169 If it is, an immediate error is given. Some errors are not detected
1170 until checks are carried out when the whole pattern has been scanned;
1171 in this case the offset is set to the end of the pattern.
1172
1173 If pcre_compile2() is used instead of pcre_compile(), and the error-
1174 codeptr argument is not NULL, a non-zero error code number is returned
1175 via this argument in the event of an error. This is in addition to the
1176 textual error message. Error codes and messages are listed below.
1177
1178 If the final argument, tableptr, is NULL, PCRE uses a default set of
1179 character tables that are built when PCRE is compiled, using the
1180 default C locale. Otherwise, tableptr must be an address that is the
1181 result of a call to pcre_maketables(). This value is stored with the
1182 compiled pattern, and used again by pcre_exec(), unless another table
1183 pointer is passed to it. For more discussion, see the section on locale
1184 support below.
1185
1186 This code fragment shows a typical straightforward call to pcre_com-
1187 pile():
1188
1189 pcre *re;
1190 const char *error;
1191 int erroffset;
1192 re = pcre_compile(
1193 "^A.*Z", /* the pattern */
1194 0, /* default options */
1195 &error, /* for error message */
1196 &erroffset, /* for error offset */
1197 NULL); /* use default character tables */
1198
1199 The following names for option bits are defined in the pcre.h header
1200 file:
1201
1202 PCRE_ANCHORED
1203
1204 If this bit is set, the pattern is forced to be "anchored", that is, it
1205 is constrained to match only at the first matching point in the string
1206 that is being searched (the "subject string"). This effect can also be
1207 achieved by appropriate constructs in the pattern itself, which is the
1208 only way to do it in Perl.
1209
1210 PCRE_AUTO_CALLOUT
1211
1212 If this bit is set, pcre_compile() automatically inserts callout items,
1213 all with number 255, before each pattern item. For discussion of the
1214 callout facility, see the pcrecallout documentation.
1215
1216 PCRE_BSR_ANYCRLF
1217 PCRE_BSR_UNICODE
1218
1219 These options (which are mutually exclusive) control what the \R escape
1220 sequence matches. The choice is either to match only CR, LF, or CRLF,
1221 or to match any Unicode newline sequence. The default is specified when
1222 PCRE is built. It can be overridden from within the pattern, or by set-
1223 ting an option when a compiled pattern is matched.
1224
1225 PCRE_CASELESS
1226
1227 If this bit is set, letters in the pattern match both upper and lower
1228 case letters. It is equivalent to Perl's /i option, and it can be
1229 changed within a pattern by a (?i) option setting. In UTF-8 mode, PCRE
1230 always understands the concept of case for characters whose values are
1231 less than 128, so caseless matching is always possible. For characters
1232 with higher values, the concept of case is supported if PCRE is com-
1233 piled with Unicode property support, but not otherwise. If you want to
1234 use caseless matching for characters 128 and above, you must ensure
1235 that PCRE is compiled with Unicode property support as well as with
1236 UTF-8 support.
1237
1238 PCRE_DOLLAR_ENDONLY
1239
1240 If this bit is set, a dollar metacharacter in the pattern matches only
1241 at the end of the subject string. Without this option, a dollar also
1242 matches immediately before a newline at the end of the string (but not
1243 before any other newlines). The PCRE_DOLLAR_ENDONLY option is ignored
1244 if PCRE_MULTILINE is set. There is no equivalent to this option in
1245 Perl, and no way to set it within a pattern.
1246
1247 PCRE_DOTALL
1248
1249 If this bit is set, a dot metacharater in the pattern matches all char-
1250 acters, including those that indicate newline. Without it, a dot does
1251 not match when the current position is at a newline. This option is
1252 equivalent to Perl's /s option, and it can be changed within a pattern
1253 by a (?s) option setting. A negative class such as [^a] always matches
1254 newline characters, independent of the setting of this option.
1255
1256 PCRE_DUPNAMES
1257
1258 If this bit is set, names used to identify capturing subpatterns need
1259 not be unique. This can be helpful for certain types of pattern when it
1260 is known that only one instance of the named subpattern can ever be
1261 matched. There are more details of named subpatterns below; see also
1262 the pcrepattern documentation.
1263
1264 PCRE_EXTENDED
1265
1266 If this bit is set, whitespace data characters in the pattern are
1267 totally ignored except when escaped or inside a character class. White-
1268 space does not include the VT character (code 11). In addition, charac-
1269 ters between an unescaped # outside a character class and the next new-
1270 line, inclusive, are also ignored. This is equivalent to Perl's /x
1271 option, and it can be changed within a pattern by a (?x) option set-
1272 ting.
1273
1274 This option makes it possible to include comments inside complicated
1275 patterns. Note, however, that this applies only to data characters.
1276 Whitespace characters may never appear within special character
1277 sequences in a pattern, for example within the sequence (?( which
1278 introduces a conditional subpattern.
1279
1280 PCRE_EXTRA
1281
1282 This option was invented in order to turn on additional functionality
1283 of PCRE that is incompatible with Perl, but it is currently of very
1284 little use. When set, any backslash in a pattern that is followed by a
1285 letter that has no special meaning causes an error, thus reserving
1286 these combinations for future expansion. By default, as in Perl, a
1287 backslash followed by a letter with no special meaning is treated as a
1288 literal. (Perl can, however, be persuaded to give an error for this, by
1289 running it with the -w option.) There are at present no other features
1290 controlled by this option. It can also be set by a (?X) option setting
1291 within a pattern.
1292
1293 PCRE_FIRSTLINE
1294
1295 If this option is set, an unanchored pattern is required to match
1296 before or at the first newline in the subject string, though the
1297 matched text may continue over the newline.
1298
1299 PCRE_JAVASCRIPT_COMPAT
1300
1301 If this option is set, PCRE's behaviour is changed in some ways so that
1302 it is compatible with JavaScript rather than Perl. The changes are as
1303 follows:
1304
1305 (1) A lone closing square bracket in a pattern causes a compile-time
1306 error, because this is illegal in JavaScript (by default it is treated
1307 as a data character). Thus, the pattern AB]CD becomes illegal when this
1308 option is set.
1309
1310 (2) At run time, a back reference to an unset subpattern group matches
1311 an empty string (by default this causes the current matching alterna-
1312 tive to fail). A pattern such as (\1)(a) succeeds when this option is
1313 set (assuming it can find an "a" in the subject), whereas it fails by
1314 default, for Perl compatibility.
1315
1316 PCRE_MULTILINE
1317
1318 By default, PCRE treats the subject string as consisting of a single
1319 line of characters (even if it actually contains newlines). The "start
1320 of line" metacharacter (^) matches only at the start of the string,
1321 while the "end of line" metacharacter ($) matches only at the end of
1322 the string, or before a terminating newline (unless PCRE_DOLLAR_ENDONLY
1323 is set). This is the same as Perl.
1324
1325 When PCRE_MULTILINE it is set, the "start of line" and "end of line"
1326 constructs match immediately following or immediately before internal
1327 newlines in the subject string, respectively, as well as at the very
1328 start and end. This is equivalent to Perl's /m option, and it can be
1329 changed within a pattern by a (?m) option setting. If there are no new-
1330 lines in a subject string, or no occurrences of ^ or $ in a pattern,
1331 setting PCRE_MULTILINE has no effect.
1332
1333 PCRE_NEWLINE_CR
1334 PCRE_NEWLINE_LF
1335 PCRE_NEWLINE_CRLF
1336 PCRE_NEWLINE_ANYCRLF
1337 PCRE_NEWLINE_ANY
1338
1339 These options override the default newline definition that was chosen
1340 when PCRE was built. Setting the first or the second specifies that a
1341 newline is indicated by a single character (CR or LF, respectively).
1342 Setting PCRE_NEWLINE_CRLF specifies that a newline is indicated by the
1343 two-character CRLF sequence. Setting PCRE_NEWLINE_ANYCRLF specifies
1344 that any of the three preceding sequences should be recognized. Setting
1345 PCRE_NEWLINE_ANY specifies that any Unicode newline sequence should be
1346 recognized. The Unicode newline sequences are the three just mentioned,
1347 plus the single characters VT (vertical tab, U+000B), FF (formfeed,
1348 U+000C), NEL (next line, U+0085), LS (line separator, U+2028), and PS
1349 (paragraph separator, U+2029). The last two are recognized only in
1350 UTF-8 mode.
1351
1352 The newline setting in the options word uses three bits that are
1353 treated as a number, giving eight possibilities. Currently only six are
1354 used (default plus the five values above). This means that if you set
1355 more than one newline option, the combination may or may not be sensi-
1356 ble. For example, PCRE_NEWLINE_CR with PCRE_NEWLINE_LF is equivalent to
1357 PCRE_NEWLINE_CRLF, but other combinations may yield unused numbers and
1358 cause an error.
1359
1360 The only time that a line break is specially recognized when compiling
1361 a pattern is if PCRE_EXTENDED is set, and an unescaped # outside a
1362 character class is encountered. This indicates a comment that lasts
1363 until after the next line break sequence. In other circumstances, line
1364 break sequences are treated as literal data, except that in
1365 PCRE_EXTENDED mode, both CR and LF are treated as whitespace characters
1366 and are therefore ignored.
1367
1368 The newline option that is set at compile time becomes the default that
1369 is used for pcre_exec() and pcre_dfa_exec(), but it can be overridden.
1370
1371 PCRE_NO_AUTO_CAPTURE
1372
1373 If this option is set, it disables the use of numbered capturing paren-
1374 theses in the pattern. Any opening parenthesis that is not followed by
1375 ? behaves as if it were followed by ?: but named parentheses can still
1376 be used for capturing (and they acquire numbers in the usual way).
1377 There is no equivalent of this option in Perl.
1378
1379 PCRE_UNGREEDY
1380
1381 This option inverts the "greediness" of the quantifiers so that they
1382 are not greedy by default, but become greedy if followed by "?". It is
1383 not compatible with Perl. It can also be set by a (?U) option setting
1384 within the pattern.
1385
1386 PCRE_UTF8
1387
1388 This option causes PCRE to regard both the pattern and the subject as
1389 strings of UTF-8 characters instead of single-byte character strings.
1390 However, it is available only when PCRE is built to include UTF-8 sup-
1391 port. If not, the use of this option provokes an error. Details of how
1392 this option changes the behaviour of PCRE are given in the section on
1393 UTF-8 support in the main pcre page.
1394
1395 PCRE_NO_UTF8_CHECK
1396
1397 When PCRE_UTF8 is set, the validity of the pattern as a UTF-8 string is
1398 automatically checked. There is a discussion about the validity of
1399 UTF-8 strings in the main pcre page. If an invalid UTF-8 sequence of
1400 bytes is found, pcre_compile() returns an error. If you already know
1401 that your pattern is valid, and you want to skip this check for perfor-
1402 mance reasons, you can set the PCRE_NO_UTF8_CHECK option. When it is
1403 set, the effect of passing an invalid UTF-8 string as a pattern is
1404 undefined. It may cause your program to crash. Note that this option
1405 can also be passed to pcre_exec() and pcre_dfa_exec(), to suppress the
1406 UTF-8 validity checking of subject strings.
1407
1408
1409 COMPILATION ERROR CODES
1410
1411 The following table lists the error codes than may be returned by
1412 pcre_compile2(), along with the error messages that may be returned by
1413 both compiling functions. As PCRE has developed, some error codes have
1414 fallen out of use. To avoid confusion, they have not been re-used.
1415
1416 0 no error
1417 1 \ at end of pattern
1418 2 \c at end of pattern
1419 3 unrecognized character follows \
1420 4 numbers out of order in {} quantifier
1421 5 number too big in {} quantifier
1422 6 missing terminating ] for character class
1423 7 invalid escape sequence in character class
1424 8 range out of order in character class
1425 9 nothing to repeat
1426 10 [this code is not in use]
1427 11 internal error: unexpected repeat
1428 12 unrecognized character after (? or (?-
1429 13 POSIX named classes are supported only within a class
1430 14 missing )
1431 15 reference to non-existent subpattern
1432 16 erroffset passed as NULL
1433 17 unknown option bit(s) set
1434 18 missing ) after comment
1435 19 [this code is not in use]
1436 20 regular expression is too large
1437 21 failed to get memory
1438 22 unmatched parentheses
1439 23 internal error: code overflow
1440 24 unrecognized character after (?<
1441 25 lookbehind assertion is not fixed length
1442 26 malformed number or name after (?(
1443 27 conditional group contains more than two branches
1444 28 assertion expected after (?(
1445 29 (?R or (?[+-]digits must be followed by )
1446 30 unknown POSIX class name
1447 31 POSIX collating elements are not supported
1448 32 this version of PCRE is not compiled with PCRE_UTF8 support
1449 33 [this code is not in use]
1450 34 character value in \x{...} sequence is too large
1451 35 invalid condition (?(0)
1452 36 \C not allowed in lookbehind assertion
1453 37 PCRE does not support \L, \l, \N, \U, or \u
1454 38 number after (?C is > 255
1455 39 closing ) for (?C expected
1456 40 recursive call could loop indefinitely
1457 41 unrecognized character after (?P
1458 42 syntax error in subpattern name (missing terminator)
1459 43 two named subpatterns have the same name
1460 44 invalid UTF-8 string
1461 45 support for \P, \p, and \X has not been compiled
1462 46 malformed \P or \p sequence
1463 47 unknown property name after \P or \p
1464 48 subpattern name is too long (maximum 32 characters)
1465 49 too many named subpatterns (maximum 10000)
1466 50 [this code is not in use]
1467 51 octal value is greater than \377 (not in UTF-8 mode)
1468 52 internal error: overran compiling workspace
1469 53 internal error: previously-checked referenced subpattern not
1470 found
1471 54 DEFINE group contains more than one branch
1472 55 repeating a DEFINE group is not allowed
1473 56 inconsistent NEWLINE options
1474 57 \g is not followed by a braced, angle-bracketed, or quoted
1475 name/number or by a plain number
1476 58 a numbered reference must not be zero
1477 59 an argument is not allowed for (*ACCEPT), (*FAIL), or (*COMMIT)
1478 60 (*VERB) not recognized
1479 61 number is too big
1480 62 subpattern name expected
1481 63 digit expected after (?+
1482 64 ] is an invalid data character in JavaScript compatibility mode
1483 65 different names for subpatterns of the same number are not
1484 allowed
1485 66 (*MARK) must have an argument
1486
1487 The numbers 32 and 10000 in errors 48 and 49 are defaults; different
1488 values may be used if the limits were changed when PCRE was built.
1489
1490
1491 STUDYING A PATTERN
1492
1493 pcre_extra *pcre_study(const pcre *code, int options
1494 const char **errptr);
1495
1496 If a compiled pattern is going to be used several times, it is worth
1497 spending more time analyzing it in order to speed up the time taken for
1498 matching. The function pcre_study() takes a pointer to a compiled pat-
1499 tern as its first argument. If studying the pattern produces additional
1500 information that will help speed up matching, pcre_study() returns a
1501 pointer to a pcre_extra block, in which the study_data field points to
1502 the results of the study.
1503
1504 The returned value from pcre_study() can be passed directly to
1505 pcre_exec() or pcre_dfa_exec(). However, a pcre_extra block also con-
1506 tains other fields that can be set by the caller before the block is
1507 passed; these are described below in the section on matching a pattern.
1508
1509 If studying the pattern does not produce any useful information,
1510 pcre_study() returns NULL. In that circumstance, if the calling program
1511 wants to pass any of the other fields to pcre_exec() or
1512 pcre_dfa_exec(), it must set up its own pcre_extra block.
1513
1514 The second argument of pcre_study() contains option bits. At present,
1515 no options are defined, and this argument should always be zero.
1516
1517 The third argument for pcre_study() is a pointer for an error message.
1518 If studying succeeds (even if no data is returned), the variable it
1519 points to is set to NULL. Otherwise it is set to point to a textual
1520 error message. This is a static string that is part of the library. You
1521 must not try to free it. You should test the error pointer for NULL
1522 after calling pcre_study(), to be sure that it has run successfully.
1523
1524 This is a typical call to pcre_study():
1525
1526 pcre_extra *pe;
1527 pe = pcre_study(
1528 re, /* result of pcre_compile() */
1529 0, /* no options exist */
1530 &error); /* set to NULL or points to a message */
1531
1532 Studying a pattern does two things: first, a lower bound for the length
1533 of subject string that is needed to match the pattern is computed. This
1534 does not mean that there are any strings of that length that match, but
1535 it does guarantee that no shorter strings match. The value is used by
1536 pcre_exec() and pcre_dfa_exec() to avoid wasting time by trying to
1537 match strings that are shorter than the lower bound. You can find out
1538 the value in a calling program via the pcre_fullinfo() function.
1539
1540 Studying a pattern is also useful for non-anchored patterns that do not
1541 have a single fixed starting character. A bitmap of possible starting
1542 bytes is created. This speeds up finding a position in the subject at
1543 which to start matching.
1544
1545
1546 LOCALE SUPPORT
1547
1548 PCRE handles caseless matching, and determines whether characters are
1549 letters, digits, or whatever, by reference to a set of tables, indexed
1550 by character value. When running in UTF-8 mode, this applies only to
1551 characters with codes less than 128. Higher-valued codes never match
1552 escapes such as \w or \d, but can be tested with \p if PCRE is built
1553 with Unicode character property support. The use of locales with Uni-
1554 code is discouraged. If you are handling characters with codes greater
1555 than 128, you should either use UTF-8 and Unicode, or use locales, but
1556 not try to mix the two.
1557
1558 PCRE contains an internal set of tables that are used when the final
1559 argument of pcre_compile() is NULL. These are sufficient for many
1560 applications. Normally, the internal tables recognize only ASCII char-
1561 acters. However, when PCRE is built, it is possible to cause the inter-
1562 nal tables to be rebuilt in the default "C" locale of the local system,
1563 which may cause them to be different.
1564
1565 The internal tables can always be overridden by tables supplied by the
1566 application that calls PCRE. These may be created in a different locale
1567 from the default. As more and more applications change to using Uni-
1568 code, the need for this locale support is expected to die away.
1569
1570 External tables are built by calling the pcre_maketables() function,
1571 which has no arguments, in the relevant locale. The result can then be
1572 passed to pcre_compile() or pcre_exec() as often as necessary. For
1573 example, to build and use tables that are appropriate for the French
1574 locale (where accented characters with values greater than 128 are
1575 treated as letters), the following code could be used:
1576
1577 setlocale(LC_CTYPE, "fr_FR");
1578 tables = pcre_maketables();
1579 re = pcre_compile(..., tables);
1580
1581 The locale name "fr_FR" is used on Linux and other Unix-like systems;
1582 if you are using Windows, the name for the French locale is "french".
1583
1584 When pcre_maketables() runs, the tables are built in memory that is
1585 obtained via pcre_malloc. It is the caller's responsibility to ensure
1586 that the memory containing the tables remains available for as long as
1587 it is needed.
1588
1589 The pointer that is passed to pcre_compile() is saved with the compiled
1590 pattern, and the same tables are used via this pointer by pcre_study()
1591 and normally also by pcre_exec(). Thus, by default, for any single pat-
1592 tern, compilation, studying and matching all happen in the same locale,
1593 but different patterns can be compiled in different locales.
1594
1595 It is possible to pass a table pointer or NULL (indicating the use of
1596 the internal tables) to pcre_exec(). Although not intended for this
1597 purpose, this facility could be used to match a pattern in a different
1598 locale from the one in which it was compiled. Passing table pointers at
1599 run time is discussed below in the section on matching a pattern.
1600
1601
1602 INFORMATION ABOUT A PATTERN
1603
1604 int pcre_fullinfo(const pcre *code, const pcre_extra *extra,
1605 int what, void *where);
1606
1607 The pcre_fullinfo() function returns information about a compiled pat-
1608 tern. It replaces the obsolete pcre_info() function, which is neverthe-
1609 less retained for backwards compability (and is documented below).
1610
1611 The first argument for pcre_fullinfo() is a pointer to the compiled
1612 pattern. The second argument is the result of pcre_study(), or NULL if
1613 the pattern was not studied. The third argument specifies which piece
1614 of information is required, and the fourth argument is a pointer to a
1615 variable to receive the data. The yield of the function is zero for
1616 success, or one of the following negative numbers:
1617
1618 PCRE_ERROR_NULL the argument code was NULL
1619 the argument where was NULL
1620 PCRE_ERROR_BADMAGIC the "magic number" was not found
1621 PCRE_ERROR_BADOPTION the value of what was invalid
1622
1623 The "magic number" is placed at the start of each compiled pattern as
1624 an simple check against passing an arbitrary memory pointer. Here is a
1625 typical call of pcre_fullinfo(), to obtain the length of the compiled
1626 pattern:
1627
1628 int rc;
1629 size_t length;
1630 rc = pcre_fullinfo(
1631 re, /* result of pcre_compile() */
1632 pe, /* result of pcre_study(), or NULL */
1633 PCRE_INFO_SIZE, /* what is required */
1634 &length); /* where to put the data */
1635
1636 The possible values for the third argument are defined in pcre.h, and
1637 are as follows:
1638
1639 PCRE_INFO_BACKREFMAX
1640
1641 Return the number of the highest back reference in the pattern. The
1642 fourth argument should point to an int variable. Zero is returned if
1643 there are no back references.
1644
1645 PCRE_INFO_CAPTURECOUNT
1646
1647 Return the number of capturing subpatterns in the pattern. The fourth
1648 argument should point to an int variable.
1649
1650 PCRE_INFO_DEFAULT_TABLES
1651
1652 Return a pointer to the internal default character tables within PCRE.
1653 The fourth argument should point to an unsigned char * variable. This
1654 information call is provided for internal use by the pcre_study() func-
1655 tion. External callers can cause PCRE to use its internal tables by
1656 passing a NULL table pointer.
1657
1658 PCRE_INFO_FIRSTBYTE
1659
1660 Return information about the first byte of any matched string, for a
1661 non-anchored pattern. The fourth argument should point to an int vari-
1662 able. (This option used to be called PCRE_INFO_FIRSTCHAR; the old name
1663 is still recognized for backwards compatibility.)
1664
1665 If there is a fixed first byte, for example, from a pattern such as
1666 (cat|cow|coyote), its value is returned. Otherwise, if either
1667
1668 (a) the pattern was compiled with the PCRE_MULTILINE option, and every
1669 branch starts with "^", or
1670
1671 (b) every branch of the pattern starts with ".*" and PCRE_DOTALL is not
1672 set (if it were set, the pattern would be anchored),
1673
1674 -1 is returned, indicating that the pattern matches only at the start
1675 of a subject string or after any newline within the string. Otherwise
1676 -2 is returned. For anchored patterns, -2 is returned.
1677
1678 PCRE_INFO_FIRSTTABLE
1679
1680 If the pattern was studied, and this resulted in the construction of a
1681 256-bit table indicating a fixed set of bytes for the first byte in any
1682 matching string, a pointer to the table is returned. Otherwise NULL is
1683 returned. The fourth argument should point to an unsigned char * vari-
1684 able.
1685
1686 PCRE_INFO_HASCRORLF
1687
1688 Return 1 if the pattern contains any explicit matches for CR or LF
1689 characters, otherwise 0. The fourth argument should point to an int
1690 variable. An explicit match is either a literal CR or LF character, or
1691 \r or \n.
1692
1693 PCRE_INFO_JCHANGED
1694
1695 Return 1 if the (?J) or (?-J) option setting is used in the pattern,
1696 otherwise 0. The fourth argument should point to an int variable. (?J)
1697 and (?-J) set and unset the local PCRE_DUPNAMES option, respectively.
1698
1699 PCRE_INFO_LASTLITERAL
1700
1701 Return the value of the rightmost literal byte that must exist in any
1702 matched string, other than at its start, if such a byte has been
1703 recorded. The fourth argument should point to an int variable. If there
1704 is no such byte, -1 is returned. For anchored patterns, a last literal
1705 byte is recorded only if it follows something of variable length. For
1706 example, for the pattern /^a\d+z\d+/ the returned value is "z", but for
1707 /^a\dz\d/ the returned value is -1.
1708
1709 PCRE_INFO_MINLENGTH
1710
1711 If the pattern was studied and a minimum length for matching subject
1712 strings was computed, its value is returned. Otherwise the returned
1713 value is -1. The value is a number of characters, not bytes (this may
1714 be relevant in UTF-8 mode). The fourth argument should point to an int
1715 variable. A non-negative value is a lower bound to the length of any
1716 matching string. There may not be any strings of that length that do
1717 actually match, but every string that does match is at least that long.
1718
1719 PCRE_INFO_NAMECOUNT
1720 PCRE_INFO_NAMEENTRYSIZE
1721 PCRE_INFO_NAMETABLE
1722
1723 PCRE supports the use of named as well as numbered capturing parenthe-
1724 ses. The names are just an additional way of identifying the parenthe-
1725 ses, which still acquire numbers. Several convenience functions such as
1726 pcre_get_named_substring() are provided for extracting captured sub-
1727 strings by name. It is also possible to extract the data directly, by
1728 first converting the name to a number in order to access the correct
1729 pointers in the output vector (described with pcre_exec() below). To do
1730 the conversion, you need to use the name-to-number map, which is
1731 described by these three values.
1732
1733 The map consists of a number of fixed-size entries. PCRE_INFO_NAMECOUNT
1734 gives the number of entries, and PCRE_INFO_NAMEENTRYSIZE gives the size
1735 of each entry; both of these return an int value. The entry size
1736 depends on the length of the longest name. PCRE_INFO_NAMETABLE returns
1737 a pointer to the first entry of the table (a pointer to char). The
1738 first two bytes of each entry are the number of the capturing parenthe-
1739 sis, most significant byte first. The rest of the entry is the corre-
1740 sponding name, zero terminated.
1741
1742 The names are in alphabetical order. Duplicate names may appear if (?|
1743 is used to create multiple groups with the same number, as described in
1744 the section on duplicate subpattern numbers in the pcrepattern page.
1745 Duplicate names for subpatterns with different numbers are permitted
1746 only if PCRE_DUPNAMES is set. In all cases of duplicate names, they
1747 appear in the table in the order in which they were found in the pat-
1748 tern. In the absence of (?| this is the order of increasing number;
1749 when (?| is used this is not necessarily the case because later subpat-
1750 terns may have lower numbers.
1751
1752 As a simple example of the name/number table, consider the following
1753 pattern (assume PCRE_EXTENDED is set, so white space - including new-
1754 lines - is ignored):
1755
1756 (?<date> (?<year>(\d\d)?\d\d) -
1757 (?<month>\d\d) - (?<day>\d\d) )
1758
1759 There are four named subpatterns, so the table has four entries, and
1760 each entry in the table is eight bytes long. The table is as follows,
1761 with non-printing bytes shows in hexadecimal, and undefined bytes shown
1762 as ??:
1763
1764 00 01 d a t e 00 ??
1765 00 05 d a y 00 ?? ??
1766 00 04 m o n t h 00
1767 00 02 y e a r 00 ??
1768
1769 When writing code to extract data from named subpatterns using the
1770 name-to-number map, remember that the length of the entries is likely
1771 to be different for each compiled pattern.
1772
1773 PCRE_INFO_OKPARTIAL
1774
1775 Return 1 if the pattern can be used for partial matching with
1776 pcre_exec(), otherwise 0. The fourth argument should point to an int
1777 variable. From release 8.00, this always returns 1, because the
1778 restrictions that previously applied to partial matching have been
1779 lifted. The pcrepartial documentation gives details of partial match-
1780 ing.
1781
1782 PCRE_INFO_OPTIONS
1783
1784 Return a copy of the options with which the pattern was compiled. The
1785 fourth argument should point to an unsigned long int variable. These
1786 option bits are those specified in the call to pcre_compile(), modified
1787 by any top-level option settings at the start of the pattern itself. In
1788 other words, they are the options that will be in force when matching
1789 starts. For example, if the pattern /(?im)abc(?-i)d/ is compiled with
1790 the PCRE_EXTENDED option, the result is PCRE_CASELESS, PCRE_MULTILINE,
1791 and PCRE_EXTENDED.
1792
1793 A pattern is automatically anchored by PCRE if all of its top-level
1794 alternatives begin with one of the following:
1795
1796 ^ unless PCRE_MULTILINE is set
1797 \A always
1798 \G always
1799 .* if PCRE_DOTALL is set and there are no back
1800 references to the subpattern in which .* appears
1801
1802 For such patterns, the PCRE_ANCHORED bit is set in the options returned
1803 by pcre_fullinfo().
1804
1805 PCRE_INFO_SIZE
1806
1807 Return the size of the compiled pattern, that is, the value that was
1808 passed as the argument to pcre_malloc() when PCRE was getting memory in
1809 which to place the compiled data. The fourth argument should point to a
1810 size_t variable.
1811
1812 PCRE_INFO_STUDYSIZE
1813
1814 Return the size of the data block pointed to by the study_data field in
1815 a pcre_extra block. That is, it is the value that was passed to
1816 pcre_malloc() when PCRE was getting memory into which to place the data
1817 created by pcre_study(). If pcre_extra is NULL, or there is no study
1818 data, zero is returned. The fourth argument should point to a size_t
1819 variable.
1820
1821
1822 OBSOLETE INFO FUNCTION
1823
1824 int pcre_info(const pcre *code, int *optptr, int *firstcharptr);
1825
1826 The pcre_info() function is now obsolete because its interface is too
1827 restrictive to return all the available data about a compiled pattern.
1828 New programs should use pcre_fullinfo() instead. The yield of
1829 pcre_info() is the number of capturing subpatterns, or one of the fol-
1830 lowing negative numbers:
1831
1832 PCRE_ERROR_NULL the argument code was NULL
1833 PCRE_ERROR_BADMAGIC the "magic number" was not found
1834
1835 If the optptr argument is not NULL, a copy of the options with which
1836 the pattern was compiled is placed in the integer it points to (see
1837 PCRE_INFO_OPTIONS above).
1838
1839 If the pattern is not anchored and the firstcharptr argument is not
1840 NULL, it is used to pass back information about the first character of
1841 any matched string (see PCRE_INFO_FIRSTBYTE above).
1842
1843
1844 REFERENCE COUNTS
1845
1846 int pcre_refcount(pcre *code, int adjust);
1847
1848 The pcre_refcount() function is used to maintain a reference count in
1849 the data block that contains a compiled pattern. It is provided for the
1850 benefit of applications that operate in an object-oriented manner,
1851 where different parts of the application may be using the same compiled
1852 pattern, but you want to free the block when they are all done.
1853
1854 When a pattern is compiled, the reference count field is initialized to
1855 zero. It is changed only by calling this function, whose action is to
1856 add the adjust value (which may be positive or negative) to it. The
1857 yield of the function is the new value. However, the value of the count
1858 is constrained to lie between 0 and 65535, inclusive. If the new value
1859 is outside these limits, it is forced to the appropriate limit value.
1860
1861 Except when it is zero, the reference count is not correctly preserved
1862 if a pattern is compiled on one host and then transferred to a host
1863 whose byte-order is different. (This seems a highly unlikely scenario.)
1864
1865
1866 MATCHING A PATTERN: THE TRADITIONAL FUNCTION
1867
1868 int pcre_exec(const pcre *code, const pcre_extra *extra,
1869 const char *subject, int length, int startoffset,
1870 int options, int *ovector, int ovecsize);
1871
1872 The function pcre_exec() is called to match a subject string against a
1873 compiled pattern, which is passed in the code argument. If the pattern
1874 was studied, the result of the study should be passed in the extra
1875 argument. This function is the main matching facility of the library,
1876 and it operates in a Perl-like manner. For specialist use there is also
1877 an alternative matching function, which is described below in the sec-
1878 tion about the pcre_dfa_exec() function.
1879
1880 In most applications, the pattern will have been compiled (and option-
1881 ally studied) in the same process that calls pcre_exec(). However, it
1882 is possible to save compiled patterns and study data, and then use them
1883 later in different processes, possibly even on different hosts. For a
1884 discussion about this, see the pcreprecompile documentation.
1885
1886 Here is an example of a simple call to pcre_exec():
1887
1888 int rc;
1889 int ovector[30];
1890 rc = pcre_exec(
1891 re, /* result of pcre_compile() */
1892 NULL, /* we didn't study the pattern */
1893 "some string", /* the subject string */
1894 11, /* the length of the subject string */
1895 0, /* start at offset 0 in the subject */
1896 0, /* default options */
1897 ovector, /* vector of integers for substring information */
1898 30); /* number of elements (NOT size in bytes) */
1899
1900 Extra data for pcre_exec()
1901
1902 If the extra argument is not NULL, it must point to a pcre_extra data
1903 block. The pcre_study() function returns such a block (when it doesn't
1904 return NULL), but you can also create one for yourself, and pass addi-
1905 tional information in it. The pcre_extra block contains the following
1906 fields (not necessarily in this order):
1907
1908 unsigned long int flags;
1909 void *study_data;
1910 unsigned long int match_limit;
1911 unsigned long int match_limit_recursion;
1912 void *callout_data;
1913 const unsigned char *tables;
1914 unsigned char **mark;
1915
1916 The flags field is a bitmap that specifies which of the other fields
1917 are set. The flag bits are:
1918
1919 PCRE_EXTRA_STUDY_DATA
1920 PCRE_EXTRA_MATCH_LIMIT
1921 PCRE_EXTRA_MATCH_LIMIT_RECURSION
1922 PCRE_EXTRA_CALLOUT_DATA
1923 PCRE_EXTRA_TABLES
1924 PCRE_EXTRA_MARK
1925
1926 Other flag bits should be set to zero. The study_data field is set in
1927 the pcre_extra block that is returned by pcre_study(), together with
1928 the appropriate flag bit. You should not set this yourself, but you may
1929 add to the block by setting the other fields and their corresponding
1930 flag bits.
1931
1932 The match_limit field provides a means of preventing PCRE from using up
1933 a vast amount of resources when running patterns that are not going to
1934 match, but which have a very large number of possibilities in their
1935 search trees. The classic example is a pattern that uses nested unlim-
1936 ited repeats.
1937
1938 Internally, PCRE uses a function called match() which it calls repeat-
1939 edly (sometimes recursively). The limit set by match_limit is imposed
1940 on the number of times this function is called during a match, which
1941 has the effect of limiting the amount of backtracking that can take
1942 place. For patterns that are not anchored, the count restarts from zero
1943 for each position in the subject string.
1944
1945 The default value for the limit can be set when PCRE is built; the
1946 default default is 10 million, which handles all but the most extreme
1947 cases. You can override the default by suppling pcre_exec() with a
1948 pcre_extra block in which match_limit is set, and
1949 PCRE_EXTRA_MATCH_LIMIT is set in the flags field. If the limit is
1950 exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
1951
1952 The match_limit_recursion field is similar to match_limit, but instead
1953 of limiting the total number of times that match() is called, it limits
1954 the depth of recursion. The recursion depth is a smaller number than
1955 the total number of calls, because not all calls to match() are recur-
1956 sive. This limit is of use only if it is set smaller than match_limit.
1957
1958 Limiting the recursion depth limits the amount of stack that can be
1959 used, or, when PCRE has been compiled to use memory on the heap instead
1960 of the stack, the amount of heap memory that can be used.
1961
1962 The default value for match_limit_recursion can be set when PCRE is
1963 built; the default default is the same value as the default for
1964 match_limit. You can override the default by suppling pcre_exec() with
1965 a pcre_extra block in which match_limit_recursion is set, and
1966 PCRE_EXTRA_MATCH_LIMIT_RECURSION is set in the flags field. If the
1967 limit is exceeded, pcre_exec() returns PCRE_ERROR_RECURSIONLIMIT.
1968
1969 The callout_data field is used in conjunction with the "callout" fea-
1970 ture, and is described in the pcrecallout documentation.
1971
1972 The tables field is used to pass a character tables pointer to
1973 pcre_exec(); this overrides the value that is stored with the compiled
1974 pattern. A non-NULL value is stored with the compiled pattern only if
1975 custom tables were supplied to pcre_compile() via its tableptr argu-
1976 ment. If NULL is passed to pcre_exec() using this mechanism, it forces
1977 PCRE's internal tables to be used. This facility is helpful when re-
1978 using patterns that have been saved after compiling with an external
1979 set of tables, because the external tables might be at a different
1980 address when pcre_exec() is called. See the pcreprecompile documenta-
1981 tion for a discussion of saving compiled patterns for later use.
1982
1983 If PCRE_EXTRA_MARK is set in the flags field, the mark field must be
1984 set to point to a char * variable. If the pattern contains any back-
1985 tracking control verbs such as (*MARK:NAME), and the execution ends up
1986 with a name to pass back, a pointer to the name string (zero termi-
1987 nated) is placed in the variable pointed to by the mark field. The
1988 names are within the compiled pattern; if you wish to retain such a
1989 name you must copy it before freeing the memory of a compiled pattern.
1990 If there is no name to pass back, the variable pointed to by the mark
1991 field set to NULL. For details of the backtracking control verbs, see
1992 the section entitled "Backtracking control" in the pcrepattern documen-
1993 tation.
1994
1995 Option bits for pcre_exec()
1996
1997 The unused bits of the options argument for pcre_exec() must be zero.
1998 The only bits that may be set are PCRE_ANCHORED, PCRE_NEWLINE_xxx,
1999 PCRE_NOTBOL, PCRE_NOTEOL, PCRE_NOTEMPTY, PCRE_NOTEMPTY_ATSTART,
2000 PCRE_NO_START_OPTIMIZE, PCRE_NO_UTF8_CHECK, PCRE_PARTIAL_SOFT, and
2001 PCRE_PARTIAL_HARD.
2002
2003 PCRE_ANCHORED
2004
2005 The PCRE_ANCHORED option limits pcre_exec() to matching at the first
2006 matching position. If a pattern was compiled with PCRE_ANCHORED, or
2007 turned out to be anchored by virtue of its contents, it cannot be made
2008 unachored at matching time.
2009
2010 PCRE_BSR_ANYCRLF
2011 PCRE_BSR_UNICODE
2012
2013 These options (which are mutually exclusive) control what the \R escape
2014 sequence matches. The choice is either to match only CR, LF, or CRLF,
2015 or to match any Unicode newline sequence. These options override the
2016 choice that was made or defaulted when the pattern was compiled.
2017
2018 PCRE_NEWLINE_CR
2019 PCRE_NEWLINE_LF
2020 PCRE_NEWLINE_CRLF
2021 PCRE_NEWLINE_ANYCRLF
2022 PCRE_NEWLINE_ANY
2023
2024 These options override the newline definition that was chosen or
2025 defaulted when the pattern was compiled. For details, see the descrip-
2026 tion of pcre_compile() above. During matching, the newline choice
2027 affects the behaviour of the dot, circumflex, and dollar metacharac-
2028 ters. It may also alter the way the match position is advanced after a
2029 match failure for an unanchored pattern.
2030
2031 When PCRE_NEWLINE_CRLF, PCRE_NEWLINE_ANYCRLF, or PCRE_NEWLINE_ANY is
2032 set, and a match attempt for an unanchored pattern fails when the cur-
2033 rent position is at a CRLF sequence, and the pattern contains no
2034 explicit matches for CR or LF characters, the match position is
2035 advanced by two characters instead of one, in other words, to after the
2036 CRLF.
2037
2038 The above rule is a compromise that makes the most common cases work as
2039 expected. For example, if the pattern is .+A (and the PCRE_DOTALL
2040 option is not set), it does not match the string "\r\nA" because, after
2041 failing at the start, it skips both the CR and the LF before retrying.
2042 However, the pattern [\r\n]A does match that string, because it con-
2043 tains an explicit CR or LF reference, and so advances only by one char-
2044 acter after the first failure.
2045
2046 An explicit match for CR of LF is either a literal appearance of one of
2047 those characters, or one of the \r or \n escape sequences. Implicit
2048 matches such as [^X] do not count, nor does \s (which includes CR and
2049 LF in the characters that it matches).
2050
2051 Notwithstanding the above, anomalous effects may still occur when CRLF
2052 is a valid newline sequence and explicit \r or \n escapes appear in the
2053 pattern.
2054
2055 PCRE_NOTBOL
2056
2057 This option specifies that first character of the subject string is not
2058 the beginning of a line, so the circumflex metacharacter should not
2059 match before it. Setting this without PCRE_MULTILINE (at compile time)
2060 causes circumflex never to match. This option affects only the behav-
2061 iour of the circumflex metacharacter. It does not affect \A.
2062
2063 PCRE_NOTEOL
2064
2065 This option specifies that the end of the subject string is not the end
2066 of a line, so the dollar metacharacter should not match it nor (except
2067 in multiline mode) a newline immediately before it. Setting this with-
2068 out PCRE_MULTILINE (at compile time) causes dollar never to match. This
2069 option affects only the behaviour of the dollar metacharacter. It does
2070 not affect \Z or \z.
2071
2072 PCRE_NOTEMPTY
2073
2074 An empty string is not considered to be a valid match if this option is
2075 set. If there are alternatives in the pattern, they are tried. If all
2076 the alternatives match the empty string, the entire match fails. For
2077 example, if the pattern
2078
2079 a?b?
2080
2081 is applied to a string not beginning with "a" or "b", it matches an
2082 empty string at the start of the subject. With PCRE_NOTEMPTY set, this
2083 match is not valid, so PCRE searches further into the string for occur-
2084 rences of "a" or "b".
2085
2086 PCRE_NOTEMPTY_ATSTART
2087
2088 This is like PCRE_NOTEMPTY, except that an empty string match that is
2089 not at the start of the subject is permitted. If the pattern is
2090 anchored, such a match can occur only if the pattern contains \K.
2091
2092 Perl has no direct equivalent of PCRE_NOTEMPTY or
2093 PCRE_NOTEMPTY_ATSTART, but it does make a special case of a pattern
2094 match of the empty string within its split() function, and when using
2095 the /g modifier. It is possible to emulate Perl's behaviour after
2096 matching a null string by first trying the match again at the same off-
2097 set with PCRE_NOTEMPTY_ATSTART and PCRE_ANCHORED, and then if that
2098 fails, by advancing the starting offset (see below) and trying an ordi-
2099 nary match again. There is some code that demonstrates how to do this
2100 in the pcredemo sample program.
2101
2102 PCRE_NO_START_OPTIMIZE
2103
2104 There are a number of optimizations that pcre_exec() uses at the start
2105 of a match, in order to speed up the process. For example, if it is
2106 known that a match must start with a specific character, it searches
2107 the subject for that character, and fails immediately if it cannot find
2108 it, without actually running the main matching function. When callouts
2109 are in use, these optimizations can cause them to be skipped. This
2110 option disables the "start-up" optimizations, causing performance to
2111 suffer, but ensuring that the callouts do occur.
2112
2113 PCRE_NO_UTF8_CHECK
2114
2115 When PCRE_UTF8 is set at compile time, the validity of the subject as a
2116 UTF-8 string is automatically checked when pcre_exec() is subsequently
2117 called. The value of startoffset is also checked to ensure that it
2118 points to the start of a UTF-8 character. There is a discussion about
2119 the validity of UTF-8 strings in the section on UTF-8 support in the
2120 main pcre page. If an invalid UTF-8 sequence of bytes is found,
2121 pcre_exec() returns the error PCRE_ERROR_BADUTF8. If startoffset con-
2122 tains an invalid value, PCRE_ERROR_BADUTF8_OFFSET is returned.
2123
2124 If you already know that your subject is valid, and you want to skip
2125 these checks for performance reasons, you can set the
2126 PCRE_NO_UTF8_CHECK option when calling pcre_exec(). You might want to
2127 do this for the second and subsequent calls to pcre_exec() if you are
2128 making repeated calls to find all the matches in a single subject
2129 string. However, you should be sure that the value of startoffset
2130 points to the start of a UTF-8 character. When PCRE_NO_UTF8_CHECK is
2131 set, the effect of passing an invalid UTF-8 string as a subject, or a
2132 value of startoffset that does not point to the start of a UTF-8 char-
2133 acter, is undefined. Your program may crash.
2134
2135 PCRE_PARTIAL_HARD
2136 PCRE_PARTIAL_SOFT
2137
2138 These options turn on the partial matching feature. For backwards com-
2139 patibility, PCRE_PARTIAL is a synonym for PCRE_PARTIAL_SOFT. A partial
2140 match occurs if the end of the subject string is reached successfully,
2141 but there are not enough subject characters to complete the match. If
2142 this happens when PCRE_PARTIAL_HARD is set, pcre_exec() immediately
2143 returns PCRE_ERROR_PARTIAL. Otherwise, if PCRE_PARTIAL_SOFT is set,
2144 matching continues by testing any other alternatives. Only if they all
2145 fail is PCRE_ERROR_PARTIAL returned (instead of PCRE_ERROR_NOMATCH).
2146 The portion of the string that was inspected when the partial match was
2147 found is set as the first matching string. There is a more detailed
2148 discussion in the pcrepartial documentation.
2149
2150 The string to be matched by pcre_exec()
2151
2152 The subject string is passed to pcre_exec() as a pointer in subject, a
2153 length (in bytes) in length, and a starting byte offset in startoffset.
2154 In UTF-8 mode, the byte offset must point to the start of a UTF-8 char-
2155 acter. Unlike the pattern string, the subject may contain binary zero
2156 bytes. When the starting offset is zero, the search for a match starts
2157 at the beginning of the subject, and this is by far the most common
2158 case.
2159
2160 A non-zero starting offset is useful when searching for another match
2161 in the same subject by calling pcre_exec() again after a previous suc-
2162 cess. Setting startoffset differs from just passing over a shortened
2163 string and setting PCRE_NOTBOL in the case of a pattern that begins
2164 with any kind of lookbehind. For example, consider the pattern
2165
2166 \Biss\B
2167
2168 which finds occurrences of "iss" in the middle of words. (\B matches
2169 only if the current position in the subject is not a word boundary.)
2170 When applied to the string "Mississipi" the first call to pcre_exec()
2171 finds the first occurrence. If pcre_exec() is called again with just
2172 the remainder of the subject, namely "issipi", it does not match,
2173 because \B is always false at the start of the subject, which is deemed
2174 to be a word boundary. However, if pcre_exec() is passed the entire
2175 string again, but with startoffset set to 4, it finds the second occur-
2176 rence of "iss" because it is able to look behind the starting point to
2177 discover that it is preceded by a letter.
2178
2179 If a non-zero starting offset is passed when the pattern is anchored,
2180 one attempt to match at the given offset is made. This can only succeed
2181 if the pattern does not require the match to be at the start of the
2182 subject.
2183
2184 How pcre_exec() returns captured substrings
2185
2186 In general, a pattern matches a certain portion of the subject, and in
2187 addition, further substrings from the subject may be picked out by
2188 parts of the pattern. Following the usage in Jeffrey Friedl's book,
2189 this is called "capturing" in what follows, and the phrase "capturing
2190 subpattern" is used for a fragment of a pattern that picks out a sub-
2191 string. PCRE supports several other kinds of parenthesized subpattern
2192 that do not cause substrings to be captured.
2193
2194 Captured substrings are returned to the caller via a vector of integers
2195 whose address is passed in ovector. The number of elements in the vec-
2196 tor is passed in ovecsize, which must be a non-negative number. Note:
2197 this argument is NOT the size of ovector in bytes.
2198
2199 The first two-thirds of the vector is used to pass back captured sub-
2200 strings, each substring using a pair of integers. The remaining third
2201 of the vector is used as workspace by pcre_exec() while matching cap-
2202 turing subpatterns, and is not available for passing back information.
2203 The number passed in ovecsize should always be a multiple of three. If
2204 it is not, it is rounded down.
2205
2206 When a match is successful, information about captured substrings is
2207 returned in pairs of integers, starting at the beginning of ovector,
2208 and continuing up to two-thirds of its length at the most. The first
2209 element of each pair is set to the byte offset of the first character
2210 in a substring, and the second is set to the byte offset of the first
2211 character after the end of a substring. Note: these values are always
2212 byte offsets, even in UTF-8 mode. They are not character counts.
2213
2214 The first pair of integers, ovector[0] and ovector[1], identify the
2215 portion of the subject string matched by the entire pattern. The next
2216 pair is used for the first capturing subpattern, and so on. The value
2217 returned by pcre_exec() is one more than the highest numbered pair that
2218 has been set. For example, if two substrings have been captured, the
2219 returned value is 3. If there are no capturing subpatterns, the return
2220 value from a successful match is 1, indicating that just the first pair
2221 of offsets has been set.
2222
2223 If a capturing subpattern is matched repeatedly, it is the last portion
2224 of the string that it matched that is returned.
2225
2226 If the vector is too small to hold all the captured substring offsets,
2227 it is used as far as possible (up to two-thirds of its length), and the
2228 function returns a value of zero. If the substring offsets are not of
2229 interest, pcre_exec() may be called with ovector passed as NULL and
2230 ovecsize as zero. However, if the pattern contains back references and
2231 the ovector is not big enough to remember the related substrings, PCRE
2232 has to get additional memory for use during matching. Thus it is usu-
2233 ally advisable to supply an ovector.
2234
2235 The pcre_fullinfo() function can be used to find out how many capturing
2236 subpatterns there are in a compiled pattern. The smallest size for
2237 ovector that will allow for n captured substrings, in addition to the
2238 offsets of the substring matched by the whole pattern, is (n+1)*3.
2239
2240 It is possible for capturing subpattern number n+1 to match some part
2241 of the subject when subpattern n has not been used at all. For example,
2242 if the string "abc" is matched against the pattern (a|(z))(bc) the
2243 return from the function is 4, and subpatterns 1 and 3 are matched, but
2244 2 is not. When this happens, both values in the offset pairs corre-
2245 sponding to unused subpatterns are set to -1.
2246
2247 Offset values that correspond to unused subpatterns at the end of the
2248 expression are also set to -1. For example, if the string "abc" is
2249 matched against the pattern (abc)(x(yz)?)? subpatterns 2 and 3 are not
2250 matched. The return from the function is 2, because the highest used
2251 capturing subpattern number is 1. However, you can refer to the offsets
2252 for the second and third capturing subpatterns if you wish (assuming
2253 the vector is large enough, of course).
2254
2255 Some convenience functions are provided for extracting the captured
2256 substrings as separate strings. These are described below.
2257
2258 Error return values from pcre_exec()
2259
2260 If pcre_exec() fails, it returns a negative number. The following are
2261 defined in the header file:
2262
2263 PCRE_ERROR_NOMATCH (-1)
2264
2265 The subject string did not match the pattern.
2266
2267 PCRE_ERROR_NULL (-2)
2268
2269 Either code or subject was passed as NULL, or ovector was NULL and
2270 ovecsize was not zero.
2271
2272 PCRE_ERROR_BADOPTION (-3)
2273
2274 An unrecognized bit was set in the options argument.
2275
2276 PCRE_ERROR_BADMAGIC (-4)
2277
2278 PCRE stores a 4-byte "magic number" at the start of the compiled code,
2279 to catch the case when it is passed a junk pointer and to detect when a
2280 pattern that was compiled in an environment of one endianness is run in
2281 an environment with the other endianness. This is the error that PCRE
2282 gives when the magic number is not present.
2283
2284 PCRE_ERROR_UNKNOWN_OPCODE (-5)
2285
2286 While running the pattern match, an unknown item was encountered in the
2287 compiled pattern. This error could be caused by a bug in PCRE or by
2288 overwriting of the compiled pattern.
2289
2290 PCRE_ERROR_NOMEMORY (-6)
2291
2292 If a pattern contains back references, but the ovector that is passed
2293 to pcre_exec() is not big enough to remember the referenced substrings,
2294 PCRE gets a block of memory at the start of matching to use for this
2295 purpose. If the call via pcre_malloc() fails, this error is given. The
2296 memory is automatically freed at the end of matching.
2297
2298 PCRE_ERROR_NOSUBSTRING (-7)
2299
2300 This error is used by the pcre_copy_substring(), pcre_get_substring(),
2301 and pcre_get_substring_list() functions (see below). It is never
2302 returned by pcre_exec().
2303
2304 PCRE_ERROR_MATCHLIMIT (-8)
2305
2306 The backtracking limit, as specified by the match_limit field in a
2307 pcre_extra structure (or defaulted) was reached. See the description
2308 above.
2309
2310 PCRE_ERROR_CALLOUT (-9)
2311
2312 This error is never generated by pcre_exec() itself. It is provided for
2313 use by callout functions that want to yield a distinctive error code.
2314 See the pcrecallout documentation for details.
2315
2316 PCRE_ERROR_BADUTF8 (-10)
2317
2318 A string that contains an invalid UTF-8 byte sequence was passed as a
2319 subject.
2320
2321 PCRE_ERROR_BADUTF8_OFFSET (-11)
2322
2323 The UTF-8 byte sequence that was passed as a subject was valid, but the
2324 value of startoffset did not point to the beginning of a UTF-8 charac-
2325 ter.
2326
2327 PCRE_ERROR_PARTIAL (-12)
2328
2329 The subject string did not match, but it did match partially. See the
2330 pcrepartial documentation for details of partial matching.
2331
2332 PCRE_ERROR_BADPARTIAL (-13)
2333
2334 This code is no longer in use. It was formerly returned when the
2335 PCRE_PARTIAL option was used with a compiled pattern containing items
2336 that were not supported for partial matching. From release 8.00
2337 onwards, there are no restrictions on partial matching.
2338
2339 PCRE_ERROR_INTERNAL (-14)
2340
2341 An unexpected internal error has occurred. This error could be caused
2342 by a bug in PCRE or by overwriting of the compiled pattern.
2343
2344 PCRE_ERROR_BADCOUNT (-15)
2345
2346 This error is given if the value of the ovecsize argument is negative.
2347
2348 PCRE_ERROR_RECURSIONLIMIT (-21)
2349
2350 The internal recursion limit, as specified by the match_limit_recursion
2351 field in a pcre_extra structure (or defaulted) was reached. See the
2352 description above.
2353
2354 PCRE_ERROR_BADNEWLINE (-23)
2355
2356 An invalid combination of PCRE_NEWLINE_xxx options was given.
2357
2358 Error numbers -16 to -20 and -22 are not used by pcre_exec().
2359
2360
2361 EXTRACTING CAPTURED SUBSTRINGS BY NUMBER
2362
2363 int pcre_copy_substring(const char *subject, int *ovector,
2364 int stringcount, int stringnumber, char *buffer,
2365 int buffersize);
2366
2367 int pcre_get_substring(const char *subject, int *ovector,
2368 int stringcount, int stringnumber,
2369 const char **stringptr);
2370
2371 int pcre_get_substring_list(const char *subject,
2372 int *ovector, int stringcount, const char ***listptr);
2373
2374 Captured substrings can be accessed directly by using the offsets
2375 returned by pcre_exec() in ovector. For convenience, the functions
2376 pcre_copy_substring(), pcre_get_substring(), and pcre_get_sub-
2377 string_list() are provided for extracting captured substrings as new,
2378 separate, zero-terminated strings. These functions identify substrings
2379 by number. The next section describes functions for extracting named
2380 substrings.
2381
2382 A substring that contains a binary zero is correctly extracted and has
2383 a further zero added on the end, but the result is not, of course, a C
2384 string. However, you can process such a string by referring to the
2385 length that is returned by pcre_copy_substring() and pcre_get_sub-
2386 string(). Unfortunately, the interface to pcre_get_substring_list() is
2387 not adequate for handling strings containing binary zeros, because the
2388 end of the final string is not independently indicated.
2389
2390 The first three arguments are the same for all three of these func-
2391 tions: subject is the subject string that has just been successfully
2392 matched, ovector is a pointer to the vector of integer offsets that was
2393 passed to pcre_exec(), and stringcount is the number of substrings that
2394 were captured by the match, including the substring that matched the
2395 entire regular expression. This is the value returned by pcre_exec() if
2396 it is greater than zero. If pcre_exec() returned zero, indicating that
2397 it ran out of space in ovector, the value passed as stringcount should
2398 be the number of elements in the vector divided by three.
2399
2400 The functions pcre_copy_substring() and pcre_get_substring() extract a
2401 single substring, whose number is given as stringnumber. A value of
2402 zero extracts the substring that matched the entire pattern, whereas
2403 higher values extract the captured substrings. For pcre_copy_sub-
2404 string(), the string is placed in buffer, whose length is given by
2405 buffersize, while for pcre_get_substring() a new block of memory is
2406 obtained via pcre_malloc, and its address is returned via stringptr.
2407 The yield of the function is the length of the string, not including
2408 the terminating zero, or one of these error codes:
2409
2410 PCRE_ERROR_NOMEMORY (-6)
2411
2412 The buffer was too small for pcre_copy_substring(), or the attempt to
2413 get memory failed for pcre_get_substring().
2414
2415 PCRE_ERROR_NOSUBSTRING (-7)
2416
2417 There is no substring whose number is stringnumber.
2418
2419 The pcre_get_substring_list() function extracts all available sub-
2420 strings and builds a list of pointers to them. All this is done in a
2421 single block of memory that is obtained via pcre_malloc. The address of
2422 the memory block is returned via listptr, which is also the start of
2423 the list of string pointers. The end of the list is marked by a NULL
2424 pointer. The yield of the function is zero if all went well, or the
2425 error code
2426
2427 PCRE_ERROR_NOMEMORY (-6)
2428
2429 if the attempt to get the memory block failed.
2430
2431 When any of these functions encounter a substring that is unset, which
2432 can happen when capturing subpattern number n+1 matches some part of
2433 the subject, but subpattern n has not been used at all, they return an
2434 empty string. This can be distinguished from a genuine zero-length sub-
2435 string by inspecting the appropriate offset in ovector, which is nega-
2436 tive for unset substrings.
2437
2438 The two convenience functions pcre_free_substring() and pcre_free_sub-
2439 string_list() can be used to free the memory returned by a previous
2440 call of pcre_get_substring() or pcre_get_substring_list(), respec-
2441 tively. They do nothing more than call the function pointed to by
2442 pcre_free, which of course could be called directly from a C program.
2443 However, PCRE is used in some situations where it is linked via a spe-
2444 cial interface to another programming language that cannot use
2445 pcre_free directly; it is for these cases that the functions are pro-
2446 vided.
2447
2448
2449 EXTRACTING CAPTURED SUBSTRINGS BY NAME
2450
2451 int pcre_get_stringnumber(const pcre *code,
2452 const char *name);
2453
2454 int pcre_copy_named_substring(const pcre *code,
2455 const char *subject, int *ovector,
2456 int stringcount, const char *stringname,
2457 char *buffer, int buffersize);
2458
2459 int pcre_get_named_substring(const pcre *code,
2460 const char *subject, int *ovector,
2461 int stringcount, const char *stringname,
2462 const char **stringptr);
2463
2464 To extract a substring by name, you first have to find associated num-
2465 ber. For example, for this pattern
2466
2467 (a+)b(?<xxx>\d+)...
2468
2469 the number of the subpattern called "xxx" is 2. If the name is known to
2470 be unique (PCRE_DUPNAMES was not set), you can find the number from the
2471 name by calling pcre_get_stringnumber(). The first argument is the com-
2472 piled pattern, and the second is the name. The yield of the function is
2473 the subpattern number, or PCRE_ERROR_NOSUBSTRING (-7) if there is no
2474 subpattern of that name.
2475
2476 Given the number, you can extract the substring directly, or use one of
2477 the functions described in the previous section. For convenience, there
2478 are also two functions that do the whole job.
2479
2480 Most of the arguments of pcre_copy_named_substring() and
2481 pcre_get_named_substring() are the same as those for the similarly
2482 named functions that extract by number. As these are described in the
2483 previous section, they are not re-described here. There are just two
2484 differences:
2485
2486 First, instead of a substring number, a substring name is given. Sec-
2487 ond, there is an extra argument, given at the start, which is a pointer
2488 to the compiled pattern. This is needed in order to gain access to the
2489 name-to-number translation table.
2490
2491 These functions call pcre_get_stringnumber(), and if it succeeds, they
2492 then call pcre_copy_substring() or pcre_get_substring(), as appropri-
2493 ate. NOTE: If PCRE_DUPNAMES is set and there are duplicate names, the
2494 behaviour may not be what you want (see the next section).
2495
2496 Warning: If the pattern uses the (?| feature to set up multiple subpat-
2497 terns with the same number, as described in the section on duplicate
2498 subpattern numbers in the pcrepattern page, you cannot use names to
2499 distinguish the different subpatterns, because names are not included
2500 in the compiled code. The matching process uses only numbers. For this
2501 reason, the use of different names for subpatterns of the same number
2502 causes an error at compile time.
2503
2504
2505 DUPLICATE SUBPATTERN NAMES
2506
2507 int pcre_get_stringtable_entries(const pcre *code,
2508 const char *name, char **first, char **last);
2509
2510 When a pattern is compiled with the PCRE_DUPNAMES option, names for
2511 subpatterns are not required to be unique. (Duplicate names are always
2512 allowed for subpatterns with the same number, created by using the (?|
2513 feature. Indeed, if such subpatterns are named, they are required to
2514 use the same names.)
2515
2516 Normally, patterns with duplicate names are such that in any one match,
2517 only one of the named subpatterns participates. An example is shown in
2518 the pcrepattern documentation.
2519
2520 When duplicates are present, pcre_copy_named_substring() and
2521 pcre_get_named_substring() return the first substring corresponding to
2522 the given name that is set. If none are set, PCRE_ERROR_NOSUBSTRING
2523 (-7) is returned; no data is returned. The pcre_get_stringnumber()
2524 function returns one of the numbers that are associated with the name,
2525 but it is not defined which it is.
2526
2527 If you want to get full details of all captured substrings for a given
2528 name, you must use the pcre_get_stringtable_entries() function. The
2529 first argument is the compiled pattern, and the second is the name. The
2530 third and fourth are pointers to variables which are updated by the
2531 function. After it has run, they point to the first and last entries in
2532 the name-to-number table for the given name. The function itself
2533 returns the length of each entry, or PCRE_ERROR_NOSUBSTRING (-7) if
2534 there are none. The format of the table is described above in the sec-
2535 tion entitled Information about a pattern. Given all the relevant
2536 entries for the name, you can extract each of their numbers, and hence
2537 the captured data, if any.
2538
2539
2540 FINDING ALL POSSIBLE MATCHES
2541
2542 The traditional matching function uses a similar algorithm to Perl,
2543 which stops when it finds the first match, starting at a given point in
2544 the subject. If you want to find all possible matches, or the longest
2545 possible match, consider using the alternative matching function (see
2546 below) instead. If you cannot use the alternative function, but still
2547 need to find all possible matches, you can kludge it up by making use
2548 of the callout facility, which is described in the pcrecallout documen-
2549 tation.
2550
2551 What you have to do is to insert a callout right at the end of the pat-
2552 tern. When your callout function is called, extract and save the cur-
2553 rent matched substring. Then return 1, which forces pcre_exec() to
2554 backtrack and try other alternatives. Ultimately, when it runs out of
2555 matches, pcre_exec() will yield PCRE_ERROR_NOMATCH.
2556
2557
2558 MATCHING A PATTERN: THE ALTERNATIVE FUNCTION
2559
2560 int pcre_dfa_exec(const pcre *code, const pcre_extra *extra,
2561 const char *subject, int length, int startoffset,
2562 int options, int *ovector, int ovecsize,
2563 int *workspace, int wscount);
2564
2565 The function pcre_dfa_exec() is called to match a subject string
2566 against a compiled pattern, using a matching algorithm that scans the
2567 subject string just once, and does not backtrack. This has different
2568 characteristics to the normal algorithm, and is not compatible with
2569 Perl. Some of the features of PCRE patterns are not supported. Never-
2570 theless, there are times when this kind of matching can be useful. For
2571 a discussion of the two matching algorithms, and a list of features
2572 that pcre_dfa_exec() does not support, see the pcrematching documenta-
2573 tion.
2574
2575 The arguments for the pcre_dfa_exec() function are the same as for
2576 pcre_exec(), plus two extras. The ovector argument is used in a differ-
2577 ent way, and this is described below. The other common arguments are
2578 used in the same way as for pcre_exec(), so their description is not
2579 repeated here.
2580
2581 The two additional arguments provide workspace for the function. The
2582 workspace vector should contain at least 20 elements. It is used for
2583 keeping track of multiple paths through the pattern tree. More
2584 workspace will be needed for patterns and subjects where there are a
2585 lot of potential matches.
2586
2587 Here is an example of a simple call to pcre_dfa_exec():
2588
2589 int rc;
2590 int ovector[10];
2591 int wspace[20];
2592 rc = pcre_dfa_exec(
2593 re, /* result of pcre_compile() */
2594 NULL, /* we didn't study the pattern */
2595 "some string", /* the subject string */
2596 11, /* the length of the subject string */
2597 0, /* start at offset 0 in the subject */
2598 0, /* default options */
2599 ovector, /* vector of integers for substring information */
2600 10, /* number of elements (NOT size in bytes) */
2601 wspace, /* working space vector */
2602 20); /* number of elements (NOT size in bytes) */
2603
2604 Option bits for pcre_dfa_exec()
2605
2606 The unused bits of the options argument for pcre_dfa_exec() must be
2607 zero. The only bits that may be set are PCRE_ANCHORED, PCRE_NEW-
2608 LINE_xxx, PCRE_NOTBOL, PCRE_NOTEOL, PCRE_NOTEMPTY,
2609 PCRE_NOTEMPTY_ATSTART, PCRE_NO_UTF8_CHECK, PCRE_PARTIAL_HARD, PCRE_PAR-
2610 TIAL_SOFT, PCRE_DFA_SHORTEST, and PCRE_DFA_RESTART. All but the last
2611 four of these are exactly the same as for pcre_exec(), so their
2612 description is not repeated here.
2613
2614 PCRE_PARTIAL_HARD
2615 PCRE_PARTIAL_SOFT
2616
2617 These have the same general effect as they do for pcre_exec(), but the
2618 details are slightly different. When PCRE_PARTIAL_HARD is set for
2619 pcre_dfa_exec(), it returns PCRE_ERROR_PARTIAL if the end of the sub-
2620 ject is reached and there is still at least one matching possibility
2621 that requires additional characters. This happens even if some complete
2622 matches have also been found. When PCRE_PARTIAL_SOFT is set, the return
2623 code PCRE_ERROR_NOMATCH is converted into PCRE_ERROR_PARTIAL if the end
2624 of the subject is reached, there have been no complete matches, but
2625 there is still at least one matching possibility. The portion of the
2626 string that was inspected when the longest partial match was found is
2627 set as the first matching string in both cases.
2628
2629 PCRE_DFA_SHORTEST
2630
2631 Setting the PCRE_DFA_SHORTEST option causes the matching algorithm to
2632 stop as soon as it has found one match. Because of the way the alterna-
2633 tive algorithm works, this is necessarily the shortest possible match
2634 at the first possible matching point in the subject string.
2635
2636 PCRE_DFA_RESTART
2637
2638 When pcre_dfa_exec() returns a partial match, it is possible to call it
2639 again, with additional subject characters, and have it continue with
2640 the same match. The PCRE_DFA_RESTART option requests this action; when
2641 it is set, the workspace and wscount options must reference the same
2642 vector as before because data about the match so far is left in them
2643 after a partial match. There is more discussion of this facility in the
2644 pcrepartial documentation.
2645
2646 Successful returns from pcre_dfa_exec()
2647
2648 When pcre_dfa_exec() succeeds, it may have matched more than one sub-
2649 string in the subject. Note, however, that all the matches from one run
2650 of the function start at the same point in the subject. The shorter
2651 matches are all initial substrings of the longer matches. For example,
2652 if the pattern
2653
2654 <.*>
2655
2656 is matched against the string
2657
2658 This is <something> <something else> <something further> no more
2659
2660 the three matched strings are
2661
2662 <something>
2663 <something> <something else>
2664 <something> <something else> <something further>
2665
2666 On success, the yield of the function is a number greater than zero,
2667 which is the number of matched substrings. The substrings themselves
2668 are returned in ovector. Each string uses two elements; the first is
2669 the offset to the start, and the second is the offset to the end. In
2670 fact, all the strings have the same start offset. (Space could have
2671 been saved by giving this only once, but it was decided to retain some
2672 compatibility with the way pcre_exec() returns data, even though the
2673 meaning of the strings is different.)
2674
2675 The strings are returned in reverse order of length; that is, the long-
2676 est matching string is given first. If there were too many matches to
2677 fit into ovector, the yield of the function is zero, and the vector is
2678 filled with the longest matches.
2679
2680 Error returns from pcre_dfa_exec()
2681
2682 The pcre_dfa_exec() function returns a negative number when it fails.
2683 Many of the errors are the same as for pcre_exec(), and these are
2684 described above. There are in addition the following errors that are
2685 specific to pcre_dfa_exec():
2686
2687 PCRE_ERROR_DFA_UITEM (-16)
2688
2689 This return is given if pcre_dfa_exec() encounters an item in the pat-
2690 tern that it does not support, for instance, the use of \C or a back
2691 reference.
2692
2693 PCRE_ERROR_DFA_UCOND (-17)
2694
2695 This return is given if pcre_dfa_exec() encounters a condition item
2696 that uses a back reference for the condition, or a test for recursion
2697 in a specific group. These are not supported.
2698
2699 PCRE_ERROR_DFA_UMLIMIT (-18)
2700
2701 This return is given if pcre_dfa_exec() is called with an extra block
2702 that contains a setting of the match_limit field. This is not supported
2703 (it is meaningless).
2704
2705 PCRE_ERROR_DFA_WSSIZE (-19)
2706
2707 This return is given if pcre_dfa_exec() runs out of space in the
2708 workspace vector.
2709
2710 PCRE_ERROR_DFA_RECURSE (-20)
2711
2712 When a recursive subpattern is processed, the matching function calls
2713 itself recursively, using private vectors for ovector and workspace.
2714 This error is given if the output vector is not large enough. This
2715 should be extremely rare, as a vector of size 1000 is used.
2716
2717
2718 SEE ALSO
2719
2720 pcrebuild(3), pcrecallout(3), pcrecpp(3)(3), pcrematching(3), pcrepar-
2721 tial(3), pcreposix(3), pcreprecompile(3), pcresample(3), pcrestack(3).
2722
2723
2724 AUTHOR
2725
2726 Philip Hazel
2727 University Computing Service
2728 Cambridge CB2 3QH, England.
2729
2730
2731 REVISION
2732
2733 Last updated: 03 May 2010
2734 Copyright (c) 1997-2010 University of Cambridge.
2735 ------------------------------------------------------------------------------
2736
2737
2738 PCRECALLOUT(3) PCRECALLOUT(3)
2739
2740
2741 NAME
2742 PCRE - Perl-compatible regular expressions
2743
2744
2745 PCRE CALLOUTS
2746
2747 int (*pcre_callout)(pcre_callout_block *);
2748
2749 PCRE provides a feature called "callout", which is a means of temporar-
2750 ily passing control to the caller of PCRE in the middle of pattern
2751 matching. The caller of PCRE provides an external function by putting
2752 its entry point in the global variable pcre_callout. By default, this
2753 variable contains NULL, which disables all calling out.
2754
2755 Within a regular expression, (?C) indicates the points at which the
2756 external function is to be called. Different callout points can be
2757 identified by putting a number less than 256 after the letter C. The
2758 default value is zero. For example, this pattern has two callout
2759 points:
2760
2761 (?C1)abc(?C2)def
2762
2763 If the PCRE_AUTO_CALLOUT option bit is set when pcre_compile() or
2764 pcre_compile2() is called, PCRE automatically inserts callouts, all
2765 with number 255, before each item in the pattern. For example, if
2766 PCRE_AUTO_CALLOUT is used with the pattern
2767
2768 A(\d{2}|--)
2769
2770 it is processed as if it were
2771
2772 (?C255)A(?C255)((?C255)\d{2}(?C255)|(?C255)-(?C255)-(?C255))(?C255)
2773
2774 Notice that there is a callout before and after each parenthesis and
2775 alternation bar. Automatic callouts can be used for tracking the
2776 progress of pattern matching. The pcretest command has an option that
2777 sets automatic callouts; when it is used, the output indicates how the
2778 pattern is matched. This is useful information when you are trying to
2779 optimize the performance of a particular pattern.
2780
2781
2782 MISSING CALLOUTS
2783
2784 You should be aware that, because of optimizations in the way PCRE
2785 matches patterns by default, callouts sometimes do not happen. For
2786 example, if the pattern is
2787
2788 ab(?C4)cd
2789
2790 PCRE knows that any matching string must contain the letter "d". If the
2791 subject string is "abyz", the lack of "d" means that matching doesn't
2792 ever start, and the callout is never reached. However, with "abyd",
2793 though the result is still no match, the callout is obeyed.
2794
2795 If the pattern is studied, PCRE knows the minimum length of a matching
2796 string, and will immediately give a "no match" return without actually
2797 running a match if the subject is not long enough, or, for unanchored
2798 patterns, if it has been scanned far enough.
2799
2800 You can disable these optimizations by passing the PCRE_NO_START_OPTI-
2801 MIZE option to pcre_exec() or pcre_dfa_exec(). This slows down the
2802 matching process, but does ensure that callouts such as the example
2803 above are obeyed.
2804
2805
2806 THE CALLOUT INTERFACE
2807
2808 During matching, when PCRE reaches a callout point, the external func-
2809 tion defined by pcre_callout is called (if it is set). This applies to
2810 both the pcre_exec() and the pcre_dfa_exec() matching functions. The
2811 only argument to the callout function is a pointer to a pcre_callout
2812 block. This structure contains the following fields:
2813
2814 int version;
2815 int callout_number;
2816 int *offset_vector;
2817 const char *subject;
2818 int subject_length;
2819 int start_match;
2820 int current_position;
2821 int capture_top;
2822 int capture_last;
2823 void *callout_data;
2824 int pattern_position;
2825 int next_item_length;
2826
2827 The version field is an integer containing the version number of the
2828 block format. The initial version was 0; the current version is 1. The
2829 version number will change again in future if additional fields are
2830 added, but the intention is never to remove any of the existing fields.
2831
2832 The callout_number field contains the number of the callout, as com-
2833 piled into the pattern (that is, the number after ?C for manual call-
2834 outs, and 255 for automatically generated callouts).
2835
2836 The offset_vector field is a pointer to the vector of offsets that was
2837 passed by the caller to pcre_exec() or pcre_dfa_exec(). When
2838 pcre_exec() is used, the contents can be inspected in order to extract
2839 substrings that have been matched so far, in the same way as for
2840 extracting substrings after a match has completed. For pcre_dfa_exec()
2841 this field is not useful.
2842
2843 The subject and subject_length fields contain copies of the values that
2844 were passed to pcre_exec().
2845
2846 The start_match field normally contains the offset within the subject
2847 at which the current match attempt started. However, if the escape
2848 sequence \K has been encountered, this value is changed to reflect the
2849 modified starting point. If the pattern is not anchored, the callout
2850 function may be called several times from the same point in the pattern
2851 for different starting points in the subject.
2852
2853 The current_position field contains the offset within the subject of
2854 the current match pointer.
2855
2856 When the pcre_exec() function is used, the capture_top field contains
2857 one more than the number of the highest numbered captured substring so
2858 far. If no substrings have been captured, the value of capture_top is
2859 one. This is always the case when pcre_dfa_exec() is used, because it
2860 does not support captured substrings.
2861
2862 The capture_last field contains the number of the most recently cap-
2863 tured substring. If no substrings have been captured, its value is -1.
2864 This is always the case when pcre_dfa_exec() is used.
2865
2866 The callout_data field contains a value that is passed to pcre_exec()
2867 or pcre_dfa_exec() specifically so that it can be passed back in call-
2868 outs. It is passed in the pcre_callout field of the pcre_extra data
2869 structure. If no such data was passed, the value of callout_data in a
2870 pcre_callout block is NULL. There is a description of the pcre_extra
2871 structure in the pcreapi documentation.
2872
2873 The pattern_position field is present from version 1 of the pcre_call-
2874 out structure. It contains the offset to the next item to be matched in
2875 the pattern string.
2876
2877 The next_item_length field is present from version 1 of the pcre_call-
2878 out structure. It contains the length of the next item to be matched in
2879 the pattern string. When the callout immediately precedes an alterna-
2880 tion bar, a closing parenthesis, or the end of the pattern, the length
2881 is zero. When the callout precedes an opening parenthesis, the length
2882 is that of the entire subpattern.
2883
2884 The pattern_position and next_item_length fields are intended to help
2885 in distinguishing between different automatic callouts, which all have
2886 the same callout number. However, they are set for all callouts.
2887
2888
2889 RETURN VALUES
2890
2891 The external callout function returns an integer to PCRE. If the value
2892 is zero, matching proceeds as normal. If the value is greater than
2893 zero, matching fails at the current point, but the testing of other
2894 matching possibilities goes ahead, just as if a lookahead assertion had
2895 failed. If the value is less than zero, the match is abandoned, and
2896 pcre_exec() or pcre_dfa_exec() returns the negative value.
2897
2898 Negative values should normally be chosen from the set of
2899 PCRE_ERROR_xxx values. In particular, PCRE_ERROR_NOMATCH forces a stan-
2900 dard "no match" failure. The error number PCRE_ERROR_CALLOUT is
2901 reserved for use by callout functions; it will never be used by PCRE
2902 itself.
2903
2904
2905 AUTHOR
2906
2907 Philip Hazel
2908 University Computing Service
2909 Cambridge CB2 3QH, England.
2910
2911
2912 REVISION
2913
2914 Last updated: 29 September 2009
2915 Copyright (c) 1997-2009 University of Cambridge.
2916 ------------------------------------------------------------------------------
2917
2918
2919 PCRECOMPAT(3) PCRECOMPAT(3)
2920
2921
2922 NAME
2923 PCRE - Perl-compatible regular expressions
2924
2925
2926 DIFFERENCES BETWEEN PCRE AND PERL
2927
2928 This document describes the differences in the ways that PCRE and Perl
2929 handle regular expressions. The differences described here are with
2930 respect to Perl 5.10.
2931
2932 1. PCRE has only a subset of Perl's UTF-8 and Unicode support. Details
2933 of what it does have are given in the section on UTF-8 support in the
2934 main pcre page.
2935
2936 2. PCRE does not allow repeat quantifiers on lookahead assertions. Perl
2937 permits them, but they do not mean what you might think. For example,
2938 (?!a){3} does not assert that the next three characters are not "a". It
2939 just asserts that the next character is not "a" three times.
2940
2941 3. Capturing subpatterns that occur inside negative lookahead asser-
2942 tions are counted, but their entries in the offsets vector are never
2943 set. Perl sets its numerical variables from any such patterns that are
2944 matched before the assertion fails to match something (thereby succeed-
2945 ing), but only if the negative lookahead assertion contains just one
2946 branch.
2947
2948 4. Though binary zero characters are supported in the subject string,
2949 they are not allowed in a pattern string because it is passed as a nor-
2950 mal C string, terminated by zero. The escape sequence \0 can be used in
2951 the pattern to represent a binary zero.
2952
2953 5. The following Perl escape sequences are not supported: \l, \u, \L,
2954 \U, and \N. In fact these are implemented by Perl's general string-han-
2955 dling and are not part of its pattern matching engine. If any of these
2956 are encountered by PCRE, an error is generated.
2957
2958 6. The Perl escape sequences \p, \P, and \X are supported only if PCRE
2959 is built with Unicode character property support. The properties that
2960 can be tested with \p and \P are limited to the general category prop-
2961 erties such as Lu and Nd, script names such as Greek or Han, and the
2962 derived properties Any and L&. PCRE does support the Cs (surrogate)
2963 property, which Perl does not; the Perl documentation says "Because
2964 Perl hides the need for the user to understand the internal representa-
2965 tion of Unicode characters, there is no need to implement the somewhat
2966 messy concept of surrogates."
2967
2968 7. PCRE does support the \Q...\E escape for quoting substrings. Charac-
2969 ters in between are treated as literals. This is slightly different
2970 from Perl in that $ and @ are also handled as literals inside the
2971 quotes. In Perl, they cause variable interpolation (but of course PCRE
2972 does not have variables). Note the following examples:
2973
2974 Pattern PCRE matches Perl matches
2975
2976 \Qabc$xyz\E abc$xyz abc followed by the
2977 contents of $xyz
2978 \Qabc\$xyz\E abc\$xyz abc\$xyz
2979 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
2980
2981 The \Q...\E sequence is recognized both inside and outside character
2982 classes.
2983
2984 8. Fairly obviously, PCRE does not support the (?{code}) and (??{code})
2985 constructions. However, there is support for recursive patterns. This
2986 is not available in Perl 5.8, but it is in Perl 5.10. Also, the PCRE
2987 "callout" feature allows an external function to be called during pat-
2988 tern matching. See the pcrecallout documentation for details.
2989
2990 9. Subpatterns that are called recursively or as "subroutines" are
2991 always treated as atomic groups in PCRE. This is like Python, but
2992 unlike Perl. There is a discussion of an example that explains this in
2993 more detail in the section on recursion differences from Perl in the
2994 pcrepattern page.
2995
2996 10. There are some differences that are concerned with the settings of
2997 captured strings when part of a pattern is repeated. For example,
2998 matching "aba" against the pattern /^(a(b)?)+$/ in Perl leaves $2
2999 unset, but in PCRE it is set to "b".
3000
3001 11. PCRE does support Perl 5.10's backtracking verbs (*ACCEPT),
3002 (*FAIL), (*F), (*COMMIT), (*PRUNE), (*SKIP), and (*THEN), but only in
3003 the forms without an argument. PCRE does not support (*MARK).
3004
3005 12. PCRE's handling of duplicate subpattern numbers and duplicate sub-
3006 pattern names is not as general as Perl's. This is a consequence of the
3007 fact the PCRE works internally just with numbers, using an external ta-
3008 ble to translate between numbers and names. In particular, a pattern
3009 such as (?|(?<a>A)|(?<b)B), where the two capturing parentheses have
3010 the same number but different names, is not supported, and causes an
3011 error at compile time. If it were allowed, it would not be possible to
3012 distinguish which parentheses matched, because both names map to cap-
3013 turing subpattern number 1. To avoid this confusing situation, an error
3014 is given at compile time.
3015
3016 13. PCRE provides some extensions to the Perl regular expression facil-
3017 ities. Perl 5.10 includes new features that are not in earlier ver-
3018 sions of Perl, some of which (such as named parentheses) have been in
3019 PCRE for some time. This list is with respect to Perl 5.10:
3020
3021 (a) Although lookbehind assertions in PCRE must match fixed length
3022 strings, each alternative branch of a lookbehind assertion can match a
3023 different length of string. Perl requires them all to have the same
3024 length.
3025
3026 (b) If PCRE_DOLLAR_ENDONLY is set and PCRE_MULTILINE is not set, the $
3027 meta-character matches only at the very end of the string.
3028
3029 (c) If PCRE_EXTRA is set, a backslash followed by a letter with no spe-
3030 cial meaning is faulted. Otherwise, like Perl, the backslash is quietly
3031 ignored. (Perl can be made to issue a warning.)
3032
3033 (d) If PCRE_UNGREEDY is set, the greediness of the repetition quanti-
3034 fiers is inverted, that is, by default they are not greedy, but if fol-
3035 lowed by a question mark they are.
3036
3037 (e) PCRE_ANCHORED can be used at matching time to force a pattern to be
3038 tried only at the first matching position in the subject string.
3039
3040 (f) The PCRE_NOTBOL, PCRE_NOTEOL, PCRE_NOTEMPTY, PCRE_NOTEMPTY_ATSTART,
3041 and PCRE_NO_AUTO_CAPTURE options for pcre_exec() have no Perl equiva-
3042 lents.
3043
3044 (g) The \R escape sequence can be restricted to match only CR, LF, or
3045 CRLF by the PCRE_BSR_ANYCRLF option.
3046
3047 (h) The callout facility is PCRE-specific.
3048
3049 (i) The partial matching facility is PCRE-specific.
3050
3051 (j) Patterns compiled by PCRE can be saved and re-used at a later time,
3052 even on different hosts that have the other endianness.
3053
3054 (k) The alternative matching function (pcre_dfa_exec()) matches in a
3055 different way and is not Perl-compatible.
3056
3057 (l) PCRE recognizes some special sequences such as (*CR) at the start
3058 of a pattern that set overall options that cannot be changed within the
3059 pattern.
3060
3061
3062 AUTHOR
3063
3064 Philip Hazel
3065 University Computing Service
3066 Cambridge CB2 3QH, England.
3067
3068
3069 REVISION
3070
3071 Last updated: 04 October 2009
3072 Copyright (c) 1997-2009 University of Cambridge.
3073 ------------------------------------------------------------------------------
3074
3075
3076 PCREPATTERN(3) PCREPATTERN(3)
3077
3078
3079 NAME
3080 PCRE - Perl-compatible regular expressions
3081
3082
3083 PCRE REGULAR EXPRESSION DETAILS
3084
3085 The syntax and semantics of the regular expressions that are supported
3086 by PCRE are described in detail below. There is a quick-reference syn-
3087 tax summary in the pcresyntax page. PCRE tries to match Perl syntax and
3088 semantics as closely as it can. PCRE also supports some alternative
3089 regular expression syntax (which does not conflict with the Perl syn-
3090 tax) in order to provide some compatibility with regular expressions in
3091 Python, .NET, and Oniguruma.
3092
3093 Perl's regular expressions are described in its own documentation, and
3094 regular expressions in general are covered in a number of books, some
3095 of which have copious examples. Jeffrey Friedl's "Mastering Regular
3096 Expressions", published by O'Reilly, covers regular expressions in
3097 great detail. This description of PCRE's regular expressions is
3098 intended as reference material.
3099
3100 The original operation of PCRE was on strings of one-byte characters.
3101 However, there is now also support for UTF-8 character strings. To use
3102 this, PCRE must be built to include UTF-8 support, and you must call
3103 pcre_compile() or pcre_compile2() with the PCRE_UTF8 option. There is
3104 also a special sequence that can be given at the start of a pattern:
3105
3106 (*UTF8)
3107
3108 Starting a pattern with this sequence is equivalent to setting the
3109 PCRE_UTF8 option. This feature is not Perl-compatible. How setting
3110 UTF-8 mode affects pattern matching is mentioned in several places
3111 below. There is also a summary of UTF-8 features in the section on
3112 UTF-8 support in the main pcre page.
3113
3114 The remainder of this document discusses the patterns that are sup-
3115 ported by PCRE when its main matching function, pcre_exec(), is used.
3116 From release 6.0, PCRE offers a second matching function,
3117 pcre_dfa_exec(), which matches using a different algorithm that is not
3118 Perl-compatible. Some of the features discussed below are not available
3119 when pcre_dfa_exec() is used. The advantages and disadvantages of the
3120 alternative function, and how it differs from the normal function, are
3121 discussed in the pcrematching page.
3122
3123
3124 NEWLINE CONVENTIONS
3125
3126 PCRE supports five different conventions for indicating line breaks in
3127 strings: a single CR (carriage return) character, a single LF (line-
3128 feed) character, the two-character sequence CRLF, any of the three pre-
3129 ceding, or any Unicode newline sequence. The pcreapi page has further
3130 discussion about newlines, and shows how to set the newline convention
3131 in the options arguments for the compiling and matching functions.
3132
3133 It is also possible to specify a newline convention by starting a pat-
3134 tern string with one of the following five sequences:
3135
3136 (*CR) carriage return
3137 (*LF) linefeed
3138 (*CRLF) carriage return, followed by linefeed
3139 (*ANYCRLF) any of the three above
3140 (*ANY) all Unicode newline sequences
3141
3142 These override the default and the options given to pcre_compile() or
3143 pcre_compile2(). For example, on a Unix system where LF is the default
3144 newline sequence, the pattern
3145
3146 (*CR)a.b
3147
3148 changes the convention to CR. That pattern matches "a\nb" because LF is
3149 no longer a newline. Note that these special settings, which are not
3150 Perl-compatible, are recognized only at the very start of a pattern,
3151 and that they must be in upper case. If more than one of them is
3152 present, the last one is used.
3153
3154 The newline convention affects the interpretation of the dot metachar-
3155 acter when PCRE_DOTALL is not set, and also the behaviour of \N. How-
3156 ever, it does not affect what the \R escape sequence matches. By
3157 default, this is any Unicode newline sequence, for Perl compatibility.
3158 However, this can be changed; see the description of \R in the section
3159 entitled "Newline sequences" below. A change of \R setting can be com-
3160 bined with a change of newline convention.
3161
3162
3163 CHARACTERS AND METACHARACTERS
3164
3165 A regular expression is a pattern that is matched against a subject
3166 string from left to right. Most characters stand for themselves in a
3167 pattern, and match the corresponding characters in the subject. As a
3168 trivial example, the pattern
3169
3170 The quick brown fox
3171
3172 matches a portion of a subject string that is identical to itself. When
3173 caseless matching is specified (the PCRE_CASELESS option), letters are
3174 matched independently of case. In UTF-8 mode, PCRE always understands
3175 the concept of case for characters whose values are less than 128, so
3176 caseless matching is always possible. For characters with higher val-
3177 ues, the concept of case is supported if PCRE is compiled with Unicode
3178 property support, but not otherwise. If you want to use caseless
3179 matching for characters 128 and above, you must ensure that PCRE is
3180 compiled with Unicode property support as well as with UTF-8 support.
3181
3182 The power of regular expressions comes from the ability to include
3183 alternatives and repetitions in the pattern. These are encoded in the
3184 pattern by the use of metacharacters, which do not stand for themselves
3185 but instead are interpreted in some special way.
3186
3187 There are two different sets of metacharacters: those that are recog-
3188 nized anywhere in the pattern except within square brackets, and those
3189 that are recognized within square brackets. Outside square brackets,
3190 the metacharacters are as follows:
3191
3192 \ general escape character with several uses
3193 ^ assert start of string (or line, in multiline mode)
3194 $ assert end of string (or line, in multiline mode)
3195 . match any character except newline (by default)
3196 [ start character class definition
3197 | start of alternative branch
3198 ( start subpattern
3199 ) end subpattern
3200 ? extends the meaning of (
3201 also 0 or 1 quantifier
3202 also quantifier minimizer
3203 * 0 or more quantifier
3204 + 1 or more quantifier
3205 also "possessive quantifier"
3206 { start min/max quantifier
3207
3208 Part of a pattern that is in square brackets is called a "character
3209 class". In a character class the only metacharacters are:
3210
3211 \ general escape character
3212 ^ negate the class, but only if the first character
3213 - indicates character range
3214 [ POSIX character class (only if followed by POSIX
3215 syntax)
3216 ] terminates the character class
3217
3218 The following sections describe the use of each of the metacharacters.
3219
3220
3221 BACKSLASH
3222
3223 The backslash character has several uses. Firstly, if it is followed by
3224 a non-alphanumeric character, it takes away any special meaning that
3225 character may have. This use of backslash as an escape character
3226 applies both inside and outside character classes.
3227
3228 For example, if you want to match a * character, you write \* in the
3229 pattern. This escaping action applies whether or not the following
3230 character would otherwise be interpreted as a metacharacter, so it is
3231 always safe to precede a non-alphanumeric with backslash to specify
3232 that it stands for itself. In particular, if you want to match a back-
3233 slash, you write \\.
3234
3235 If a pattern is compiled with the PCRE_EXTENDED option, whitespace in
3236 the pattern (other than in a character class) and characters between a
3237 # outside a character class and the next newline are ignored. An escap-
3238 ing backslash can be used to include a whitespace or # character as
3239 part of the pattern.
3240
3241 If you want to remove the special meaning from a sequence of charac-
3242 ters, you can do so by putting them between \Q and \E. This is differ-
3243 ent from Perl in that $ and @ are handled as literals in \Q...\E
3244 sequences in PCRE, whereas in Perl, $ and @ cause variable interpola-
3245 tion. Note the following examples:
3246
3247 Pattern PCRE matches Perl matches
3248
3249 \Qabc$xyz\E abc$xyz abc followed by the
3250 contents of $xyz
3251 \Qabc\$xyz\E abc\$xyz abc\$xyz
3252 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3253
3254 The \Q...\E sequence is recognized both inside and outside character
3255 classes.
3256
3257 Non-printing characters
3258
3259 A second use of backslash provides a way of encoding non-printing char-
3260 acters in patterns in a visible manner. There is no restriction on the
3261 appearance of non-printing characters, apart from the binary zero that
3262 terminates a pattern, but when a pattern is being prepared by text
3263 editing, it is often easier to use one of the following escape
3264 sequences than the binary character it represents:
3265
3266 \a alarm, that is, the BEL character (hex 07)
3267 \cx "control-x", where x is any character
3268 \e escape (hex 1B)
3269 \f formfeed (hex 0C)
3270 \n linefeed (hex 0A)
3271 \r carriage return (hex 0D)
3272 \t tab (hex 09)
3273 \ddd character with octal code ddd, or back reference
3274 \xhh character with hex code hh
3275 \x{hhh..} character with hex code hhh..
3276
3277 The precise effect of \cx is as follows: if x is a lower case letter,
3278 it is converted to upper case. Then bit 6 of the character (hex 40) is
3279 inverted. Thus \cz becomes hex 1A, but \c{ becomes hex 3B, while \c;
3280 becomes hex 7B.
3281
3282 After \x, from zero to two hexadecimal digits are read (letters can be
3283 in upper or lower case). Any number of hexadecimal digits may appear
3284 between \x{ and }, but the value of the character code must be less
3285 than 256 in non-UTF-8 mode, and less than 2**31 in UTF-8 mode. That is,
3286 the maximum value in hexadecimal is 7FFFFFFF. Note that this is bigger
3287 than the largest Unicode code point, which is 10FFFF.
3288
3289 If characters other than hexadecimal digits appear between \x{ and },
3290 or if there is no terminating }, this form of escape is not recognized.
3291 Instead, the initial \x will be interpreted as a basic hexadecimal
3292 escape, with no following digits, giving a character whose value is
3293 zero.
3294
3295 Characters whose value is less than 256 can be defined by either of the
3296 two syntaxes for \x. There is no difference in the way they are han-
3297 dled. For example, \xdc is exactly the same as \x{dc}.
3298
3299 After \0 up to two further octal digits are read. If there are fewer
3300 than two digits, just those that are present are used. Thus the
3301 sequence \0\x\07 specifies two binary zeros followed by a BEL character
3302 (code value 7). Make sure you supply two digits after the initial zero
3303 if the pattern character that follows is itself an octal digit.
3304
3305 The handling of a backslash followed by a digit other than 0 is compli-
3306 cated. Outside a character class, PCRE reads it and any following dig-
3307 its as a decimal number. If the number is less than 10, or if there
3308 have been at least that many previous capturing left parentheses in the
3309 expression, the entire sequence is taken as a back reference. A
3310 description of how this works is given later, following the discussion
3311 of parenthesized subpatterns.
3312
3313 Inside a character class, or if the decimal number is greater than 9
3314 and there have not been that many capturing subpatterns, PCRE re-reads
3315 up to three octal digits following the backslash, and uses them to gen-
3316 erate a data character. Any subsequent digits stand for themselves. In
3317 non-UTF-8 mode, the value of a character specified in octal must be
3318 less than \400. In UTF-8 mode, values up to \777 are permitted. For
3319 example:
3320
3321 \040 is another way of writing a space
3322 \40 is the same, provided there are fewer than 40
3323 previous capturing subpatterns
3324 \7 is always a back reference
3325 \11 might be a back reference, or another way of
3326 writing a tab
3327 \011 is always a tab
3328 \0113 is a tab followed by the character "3"
3329 \113 might be a back reference, otherwise the
3330 character with octal code 113
3331 \377 might be a back reference, otherwise
3332 the byte consisting entirely of 1 bits
3333 \81 is either a back reference, or a binary zero
3334 followed by the two characters "8" and "1"
3335
3336 Note that octal values of 100 or greater must not be introduced by a
3337 leading zero, because no more than three octal digits are ever read.
3338
3339 All the sequences that define a single character value can be used both
3340 inside and outside character classes. In addition, inside a character
3341 class, the sequence \b is interpreted as the backspace character (hex
3342 08). The sequences \B, \N, \R, and \X are not special inside a charac-
3343 ter class. Like any other unrecognized escape sequences, they are
3344 treated as the literal characters "B", "N", "R", and "X" by default,
3345 but cause an error if the PCRE_EXTRA option is set. Outside a character
3346 class, these sequences have different meanings.
3347
3348 Absolute and relative back references
3349
3350 The sequence \g followed by an unsigned or a negative number, option-
3351 ally enclosed in braces, is an absolute or relative back reference. A
3352 named back reference can be coded as \g{name}. Back references are dis-
3353 cussed later, following the discussion of parenthesized subpatterns.
3354
3355 Absolute and relative subroutine calls
3356
3357 For compatibility with Oniguruma, the non-Perl syntax \g followed by a
3358 name or a number enclosed either in angle brackets or single quotes, is
3359 an alternative syntax for referencing a subpattern as a "subroutine".
3360 Details are discussed later. Note that \g{...} (Perl syntax) and
3361 \g<...> (Oniguruma syntax) are not synonymous. The former is a back
3362 reference; the latter is a subroutine call.
3363
3364 Generic character types
3365
3366 Another use of backslash is for specifying generic character types:
3367
3368 \d any decimal digit
3369 \D any character that is not a decimal digit
3370 \h any horizontal whitespace character
3371 \H any character that is not a horizontal whitespace character
3372 \s any whitespace character
3373 \S any character that is not a whitespace character
3374 \v any vertical whitespace character
3375 \V any character that is not a vertical whitespace character
3376 \w any "word" character
3377 \W any "non-word" character
3378
3379 There is also the single sequence \N, which matches a non-newline char-
3380 acter. This is the same as the "." metacharacter when PCRE_DOTALL is
3381 not set.
3382
3383 Each pair of lower and upper case escape sequences partitions the com-
3384 plete set of characters into two disjoint sets. Any given character
3385 matches one, and only one, of each pair.
3386
3387 These character type sequences can appear both inside and outside char-
3388 acter classes. They each match one character of the appropriate type.
3389 If the current matching point is at the end of the subject string, all
3390 of them fail, since there is no character to match.
3391
3392 For compatibility with Perl, \s does not match the VT character (code
3393 11). This makes it different from the the POSIX "space" class. The \s
3394 characters are HT (9), LF (10), FF (12), CR (13), and space (32). If
3395 "use locale;" is included in a Perl script, \s may match the VT charac-
3396 ter. In PCRE, it never does.
3397
3398 In UTF-8 mode, characters with values greater than 128 never match \d,
3399 \s, or \w, and always match \D, \S, and \W. This is true even when Uni-
3400 code character property support is available. These sequences retain
3401 their original meanings from before UTF-8 support was available, mainly
3402 for efficiency reasons. Note that this also affects \b, because it is
3403 defined in terms of \w and \W.
3404
3405 The sequences \h, \H, \v, and \V are Perl 5.10 features. In contrast to
3406 the other sequences, these do match certain high-valued codepoints in
3407 UTF-8 mode. The horizontal space characters are:
3408
3409 U+0009 Horizontal tab
3410 U+0020 Space
3411 U+00A0 Non-break space
3412 U+1680 Ogham space mark
3413 U+180E Mongolian vowel separator
3414 U+2000 En quad
3415 U+2001 Em quad
3416 U+2002 En space
3417 U+2003 Em space
3418 U+2004 Three-per-em space
3419 U+2005 Four-per-em space
3420 U+2006 Six-per-em space
3421 U+2007 Figure space
3422 U+2008 Punctuation space
3423 U+2009 Thin space
3424 U+200A Hair space
3425 U+202F Narrow no-break space
3426 U+205F Medium mathematical space
3427 U+3000 Ideographic space
3428
3429 The vertical space characters are:
3430
3431 U+000A Linefeed
3432 U+000B Vertical tab
3433 U+000C Formfeed
3434 U+000D Carriage return
3435 U+0085 Next line
3436 U+2028 Line separator
3437 U+2029 Paragraph separator
3438
3439 A "word" character is an underscore or any character less than 256 that
3440 is a letter or digit. The definition of letters and digits is con-
3441 trolled by PCRE's low-valued character tables, and may vary if locale-
3442 specific matching is taking place (see "Locale support" in the pcreapi
3443 page). For example, in a French locale such as "fr_FR" in Unix-like
3444 systems, or "french" in Windows, some character codes greater than 128
3445 are used for accented letters, and these are matched by \w. The use of
3446 locales with Unicode is discouraged.
3447
3448 Newline sequences
3449
3450 Outside a character class, by default, the escape sequence \R matches
3451 any Unicode newline sequence. This is a Perl 5.10 feature. In non-UTF-8
3452 mode \R is equivalent to the following:
3453
3454 (?>\r\n|\n|\x0b|\f|\r|\x85)
3455
3456 This is an example of an "atomic group", details of which are given
3457 below. This particular group matches either the two-character sequence
3458 CR followed by LF, or one of the single characters LF (linefeed,
3459 U+000A), VT (vertical tab, U+000B), FF (formfeed, U+000C), CR (carriage
3460 return, U+000D), or NEL (next line, U+0085). The two-character sequence
3461 is treated as a single unit that cannot be split.
3462
3463 In UTF-8 mode, two additional characters whose codepoints are greater
3464 than 255 are added: LS (line separator, U+2028) and PS (paragraph sepa-
3465 rator, U+2029). Unicode character property support is not needed for
3466 these characters to be recognized.
3467
3468 It is possible to restrict \R to match only CR, LF, or CRLF (instead of
3469 the complete set of Unicode line endings) by setting the option
3470 PCRE_BSR_ANYCRLF either at compile time or when the pattern is matched.
3471 (BSR is an abbrevation for "backslash R".) This can be made the default
3472 when PCRE is built; if this is the case, the other behaviour can be
3473 requested via the PCRE_BSR_UNICODE option. It is also possible to
3474 specify these settings by starting a pattern string with one of the
3475 following sequences:
3476
3477 (*BSR_ANYCRLF) CR, LF, or CRLF only
3478 (*BSR_UNICODE) any Unicode newline sequence
3479
3480 These override the default and the options given to pcre_compile() or
3481 pcre_compile2(), but they can be overridden by options given to
3482 pcre_exec() or pcre_dfa_exec(). Note that these special settings, which
3483 are not Perl-compatible, are recognized only at the very start of a
3484 pattern, and that they must be in upper case. If more than one of them
3485 is present, the last one is used. They can be combined with a change of
3486 newline convention, for example, a pattern can start with:
3487
3488 (*ANY)(*BSR_ANYCRLF)
3489
3490 Inside a character class, \R is treated as an unrecognized escape
3491 sequence, and so matches the letter "R" by default, but causes an error
3492 if PCRE_EXTRA is set.
3493
3494 Unicode character properties
3495
3496 When PCRE is built with Unicode character property support, three addi-
3497 tional escape sequences that match characters with specific properties
3498 are available. When not in UTF-8 mode, these sequences are of course
3499 limited to testing characters whose codepoints are less than 256, but
3500 they do work in this mode. The extra escape sequences are:
3501
3502 \p{xx} a character with the xx property
3503 \P{xx} a character without the xx property
3504 \X an extended Unicode sequence
3505
3506 The property names represented by xx above are limited to the Unicode
3507 script names, the general category properties, "Any", which matches any
3508 character (including newline), and some special PCRE properties
3509 (described in the next section). Other Perl properties such as "InMu-
3510 sicalSymbols" are not currently supported by PCRE. Note that \P{Any}
3511 does not match any characters, so always causes a match failure.
3512
3513 Sets of Unicode characters are defined as belonging to certain scripts.
3514 A character from one of these sets can be matched using a script name.
3515 For example:
3516
3517 \p{Greek}
3518 \P{Han}
3519
3520 Those that are not part of an identified script are lumped together as
3521 "Common". The current list of scripts is:
3522
3523 Arabic, Armenian, Avestan, Balinese, Bamum, Bengali, Bopomofo, Braille,
3524 Buginese, Buhid, Canadian_Aboriginal, Carian, Cham, Cherokee, Common,
3525 Coptic, Cuneiform, Cypriot, Cyrillic, Deseret, Devanagari, Egyp-
3526 tian_Hieroglyphs, Ethiopic, Georgian, Glagolitic, Gothic, Greek,
3527 Gujarati, Gurmukhi, Han, Hangul, Hanunoo, Hebrew, Hiragana, Impe-
3528 rial_Aramaic, Inherited, Inscriptional_Pahlavi, Inscriptional_Parthian,
3529 Javanese, Kaithi, Kannada, Katakana, Kayah_Li, Kharoshthi, Khmer, Lao,
3530 Latin, Lepcha, Limbu, Linear_B, Lisu, Lycian, Lydian, Malayalam,
3531 Meetei_Mayek, Mongolian, Myanmar, New_Tai_Lue, Nko, Ogham, Old_Italic,
3532 Old_Persian, Old_South_Arabian, Old_Turkic, Ol_Chiki, Oriya, Osmanya,
3533 Phags_Pa, Phoenician, Rejang, Runic, Samaritan, Saurashtra, Shavian,
3534 Sinhala, Sundanese, Syloti_Nagri, Syriac, Tagalog, Tagbanwa, Tai_Le,
3535 Tai_Tham, Tai_Viet, Tamil, Telugu, Thaana, Thai, Tibetan, Tifinagh,
3536 Ugaritic, Vai, Yi.
3537
3538 Each character has exactly one Unicode general category property, spec-
3539 ified by a two-letter abbreviation. For compatibility with Perl, nega-
3540 tion can be specified by including a circumflex between the opening
3541 brace and the property name. For example, \p{^Lu} is the same as
3542 \P{Lu}.
3543
3544 If only one letter is specified with \p or \P, it includes all the gen-
3545 eral category properties that start with that letter. In this case, in
3546 the absence of negation, the curly brackets in the escape sequence are
3547 optional; these two examples have the same effect:
3548
3549 \p{L}
3550 \pL
3551
3552 The following general category property codes are supported:
3553
3554 C Other
3555 Cc Control
3556 Cf Format
3557 Cn Unassigned
3558 Co Private use
3559 Cs Surrogate
3560
3561 L Letter
3562 Ll Lower case letter
3563 Lm Modifier letter
3564 Lo Other letter
3565 Lt Title case letter
3566 Lu Upper case letter
3567
3568 M Mark
3569 Mc Spacing mark
3570 Me Enclosing mark
3571 Mn Non-spacing mark
3572
3573 N Number
3574 Nd Decimal number
3575 Nl Letter number
3576 No Other number
3577
3578 P Punctuation
3579 Pc Connector punctuation
3580 Pd Dash punctuation
3581 Pe Close punctuation
3582 Pf Final punctuation
3583 Pi Initial punctuation
3584 Po Other punctuation
3585 Ps Open punctuation
3586
3587 S Symbol
3588 Sc Currency symbol
3589 Sk Modifier symbol
3590 Sm Mathematical symbol
3591 So Other symbol
3592
3593 Z Separator
3594 Zl Line separator
3595 Zp Paragraph separator
3596 Zs Space separator
3597
3598 The special property L& is also supported: it matches a character that
3599 has the Lu, Ll, or Lt property, in other words, a letter that is not
3600 classified as a modifier or "other".
3601
3602 The Cs (Surrogate) property applies only to characters in the range
3603 U+D800 to U+DFFF. Such characters are not valid in UTF-8 strings (see
3604 RFC 3629) and so cannot be tested by PCRE, unless UTF-8 validity check-
3605 ing has been turned off (see the discussion of PCRE_NO_UTF8_CHECK in
3606 the pcreapi page). Perl does not support the Cs property.
3607
3608 The long synonyms for property names that Perl supports (such as
3609 \p{Letter}) are not supported by PCRE, nor is it permitted to prefix
3610 any of these properties with "Is".
3611
3612 No character that is in the Unicode table has the Cn (unassigned) prop-
3613 erty. Instead, this property is assumed for any code point that is not
3614 in the Unicode table.
3615
3616 Specifying caseless matching does not affect these escape sequences.
3617 For example, \p{Lu} always matches only upper case letters.
3618
3619 The \X escape matches any number of Unicode characters that form an
3620 extended Unicode sequence. \X is equivalent to
3621
3622 (?>\PM\pM*)
3623
3624 That is, it matches a character without the "mark" property, followed
3625 by zero or more characters with the "mark" property, and treats the
3626 sequence as an atomic group (see below). Characters with the "mark"
3627 property are typically accents that affect the preceding character.
3628 None of them have codepoints less than 256, so in non-UTF-8 mode \X
3629 matches any one character.
3630
3631 Matching characters by Unicode property is not fast, because PCRE has
3632 to search a structure that contains data for over fifteen thousand
3633 characters. That is why the traditional escape sequences such as \d and
3634 \w do not use Unicode properties in PCRE.
3635
3636 PCRE's additional properties
3637
3638 As well as the standard Unicode properties described in the previous
3639 section, PCRE supports four more that make it possible to convert tra-
3640 ditional escape sequences such as \w and \s and POSIX character classes
3641 to use Unicode properties. These are:
3642
3643 Xan Any alphanumeric character
3644 Xps Any POSIX space character
3645 Xsp Any Perl space character
3646 Xwd Any Perl "word" character
3647
3648 Xan matches characters that have either the L (letter) or the N (num-
3649 ber) property. Xps matches the characters tab, linefeed, vertical tab,
3650 formfeed, or carriage return, and any other character that has the Z
3651 (separator) property. Xsp is the same as Xps, except that vertical tab
3652 is excluded. Xwd matches the same characters as Xan, plus underscore.
3653
3654 Resetting the match start
3655
3656 The escape sequence \K, which is a Perl 5.10 feature, causes any previ-
3657 ously matched characters not to be included in the final matched
3658 sequence. For example, the pattern:
3659
3660 foo\Kbar
3661
3662 matches "foobar", but reports that it has matched "bar". This feature
3663 is similar to a lookbehind assertion (described below). However, in
3664 this case, the part of the subject before the real match does not have
3665 to be of fixed length, as lookbehind assertions do. The use of \K does
3666 not interfere with the setting of captured substrings. For example,
3667 when the pattern
3668
3669 (foo)\Kbar
3670
3671 matches "foobar", the first substring is still set to "foo".
3672
3673 Perl documents that the use of \K within assertions is "not well
3674 defined". In PCRE, \K is acted upon when it occurs inside positive
3675 assertions, but is ignored in negative assertions.
3676
3677 Simple assertions
3678
3679 The final use of backslash is for certain simple assertions. An asser-
3680 tion specifies a condition that has to be met at a particular point in
3681 a match, without consuming any characters from the subject string. The
3682 use of subpatterns for more complicated assertions is described below.
3683 The backslashed assertions are:
3684
3685 \b matches at a word boundary
3686 \B matches when not at a word boundary
3687 \A matches at the start of the subject
3688 \Z matches at the end of the subject
3689 also matches before a newline at the end of the subject
3690 \z matches only at the end of the subject
3691 \G matches at the first matching position in the subject
3692
3693 Inside a character class, \b has a different meaning; it matches the
3694 backspace character. If any other of these assertions appears in a
3695 character class, by default it matches the corresponding literal char-
3696 acter (for example, \B matches the letter B). However, if the
3697 PCRE_EXTRA option is set, an "invalid escape sequence" error is gener-
3698 ated instead.
3699
3700 A word boundary is a position in the subject string where the current
3701 character and the previous character do not both match \w or \W (i.e.
3702 one matches \w and the other matches \W), or the start or end of the
3703 string if the first or last character matches \w, respectively. Neither
3704 PCRE nor Perl has a separte "start of word" or "end of word" metase-
3705 quence. However, whatever follows \b normally determines which it is.
3706 For example, the fragment \ba matches "a" at the start of a word.
3707
3708 The \A, \Z, and \z assertions differ from the traditional circumflex
3709 and dollar (described in the next section) in that they only ever match
3710 at the very start and end of the subject string, whatever options are
3711 set. Thus, they are independent of multiline mode. These three asser-
3712 tions are not affected by the PCRE_NOTBOL or PCRE_NOTEOL options, which
3713 affect only the behaviour of the circumflex and dollar metacharacters.
3714 However, if the startoffset argument of pcre_exec() is non-zero, indi-
3715 cating that matching is to start at a point other than the beginning of
3716 the subject, \A can never match. The difference between \Z and \z is
3717 that \Z matches before a newline at the end of the string as well as at
3718 the very end, whereas \z matches only at the end.
3719
3720 The \G assertion is true only when the current matching position is at
3721 the start point of the match, as specified by the startoffset argument
3722 of pcre_exec(). It differs from \A when the value of startoffset is
3723 non-zero. By calling pcre_exec() multiple times with appropriate argu-
3724 ments, you can mimic Perl's /g option, and it is in this kind of imple-
3725 mentation where \G can be useful.
3726
3727 Note, however, that PCRE's interpretation of \G, as the start of the
3728 current match, is subtly different from Perl's, which defines it as the
3729 end of the previous match. In Perl, these can be different when the
3730 previously matched string was empty. Because PCRE does just one match
3731 at a time, it cannot reproduce this behaviour.
3732
3733 If all the alternatives of a pattern begin with \G, the expression is
3734 anchored to the starting match position, and the "anchored" flag is set
3735 in the compiled regular expression.
3736
3737
3738 CIRCUMFLEX AND DOLLAR
3739
3740 Outside a character class, in the default matching mode, the circumflex
3741 character is an assertion that is true only if the current matching
3742 point is at the start of the subject string. If the startoffset argu-
3743 ment of pcre_exec() is non-zero, circumflex can never match if the
3744 PCRE_MULTILINE option is unset. Inside a character class, circumflex
3745 has an entirely different meaning (see below).
3746
3747 Circumflex need not be the first character of the pattern if a number
3748 of alternatives are involved, but it should be the first thing in each
3749 alternative in which it appears if the pattern is ever to match that
3750 branch. If all possible alternatives start with a circumflex, that is,
3751 if the pattern is constrained to match only at the start of the sub-
3752 ject, it is said to be an "anchored" pattern. (There are also other
3753 constructs that can cause a pattern to be anchored.)
3754
3755 A dollar character is an assertion that is true only if the current
3756 matching point is at the end of the subject string, or immediately
3757 before a newline at the end of the string (by default). Dollar need not
3758 be the last character of the pattern if a number of alternatives are
3759 involved, but it should be the last item in any branch in which it
3760 appears. Dollar has no special meaning in a character class.
3761
3762 The meaning of dollar can be changed so that it matches only at the
3763 very end of the string, by setting the PCRE_DOLLAR_ENDONLY option at
3764 compile time. This does not affect the \Z assertion.
3765
3766 The meanings of the circumflex and dollar characters are changed if the
3767 PCRE_MULTILINE option is set. When this is the case, a circumflex
3768 matches immediately after internal newlines as well as at the start of
3769 the subject string. It does not match after a newline that ends the
3770 string. A dollar matches before any newlines in the string, as well as
3771 at the very end, when PCRE_MULTILINE is set. When newline is specified
3772 as the two-character sequence CRLF, isolated CR and LF characters do
3773 not indicate newlines.
3774
3775 For example, the pattern /^abc$/ matches the subject string "def\nabc"
3776 (where \n represents a newline) in multiline mode, but not otherwise.
3777 Consequently, patterns that are anchored in single line mode because
3778 all branches start with ^ are not anchored in multiline mode, and a
3779 match for circumflex is possible when the startoffset argument of
3780 pcre_exec() is non-zero. The PCRE_DOLLAR_ENDONLY option is ignored if
3781 PCRE_MULTILINE is set.
3782
3783 Note that the sequences \A, \Z, and \z can be used to match the start
3784 and end of the subject in both modes, and if all branches of a pattern
3785 start with \A it is always anchored, whether or not PCRE_MULTILINE is
3786 set.
3787
3788
3789 FULL STOP (PERIOD, DOT) AND \N
3790
3791 Outside a character class, a dot in the pattern matches any one charac-
3792 ter in the subject string except (by default) a character that signi-
3793 fies the end of a line. In UTF-8 mode, the matched character may be
3794 more than one byte long.
3795
3796 When a line ending is defined as a single character, dot never matches
3797 that character; when the two-character sequence CRLF is used, dot does
3798 not match CR if it is immediately followed by LF, but otherwise it
3799 matches all characters (including isolated CRs and LFs). When any Uni-
3800 code line endings are being recognized, dot does not match CR or LF or
3801 any of the other line ending characters.
3802
3803 The behaviour of dot with regard to newlines can be changed. If the
3804 PCRE_DOTALL option is set, a dot matches any one character, without
3805 exception. If the two-character sequence CRLF is present in the subject
3806 string, it takes two dots to match it.
3807
3808 The handling of dot is entirely independent of the handling of circum-
3809 flex and dollar, the only relationship being that they both involve
3810 newlines. Dot has no special meaning in a character class.
3811
3812 The escape sequence \N always behaves as a dot does when PCRE_DOTALL is
3813 not set. In other words, it matches any one character except one that
3814 signifies the end of a line.
3815
3816
3817 MATCHING A SINGLE BYTE
3818
3819 Outside a character class, the escape sequence \C matches any one byte,
3820 both in and out of UTF-8 mode. Unlike a dot, it always matches any
3821 line-ending characters. The feature is provided in Perl in order to
3822 match individual bytes in UTF-8 mode. Because it breaks up UTF-8 char-
3823 acters into individual bytes, what remains in the string may be a mal-
3824 formed UTF-8 string. For this reason, the \C escape sequence is best
3825 avoided.
3826
3827 PCRE does not allow \C to appear in lookbehind assertions (described
3828 below), because in UTF-8 mode this would make it impossible to calcu-
3829 late the length of the lookbehind.
3830
3831
3832 SQUARE BRACKETS AND CHARACTER CLASSES
3833
3834 An opening square bracket introduces a character class, terminated by a
3835 closing square bracket. A closing square bracket on its own is not spe-
3836 cial by default. However, if the PCRE_JAVASCRIPT_COMPAT option is set,
3837 a lone closing square bracket causes a compile-time error. If a closing
3838 square bracket is required as a member of the class, it should be the
3839 first data character in the class (after an initial circumflex, if
3840 present) or escaped with a backslash.
3841
3842 A character class matches a single character in the subject. In UTF-8
3843 mode, the character may be more than one byte long. A matched character
3844 must be in the set of characters defined by the class, unless the first
3845 character in the class definition is a circumflex, in which case the
3846 subject character must not be in the set defined by the class. If a
3847 circumflex is actually required as a member of the class, ensure it is
3848 not the first character, or escape it with a backslash.
3849
3850 For example, the character class [aeiou] matches any lower case vowel,
3851 while [^aeiou] matches any character that is not a lower case vowel.
3852 Note that a circumflex is just a convenient notation for specifying the
3853 characters that are in the class by enumerating those that are not. A
3854 class that starts with a circumflex is not an assertion; it still con-
3855 sumes a character from the subject string, and therefore it fails if
3856 the current pointer is at the end of the string.
3857
3858 In UTF-8 mode, characters with values greater than 255 can be included
3859 in a class as a literal string of bytes, or by using the \x{ escaping
3860 mechanism.
3861
3862 When caseless matching is set, any letters in a class represent both
3863 their upper case and lower case versions, so for example, a caseless
3864 [aeiou] matches "A" as well as "a", and a caseless [^aeiou] does not
3865 match "A", whereas a caseful version would. In UTF-8 mode, PCRE always
3866 understands the concept of case for characters whose values are less
3867 than 128, so caseless matching is always possible. For characters with
3868 higher values, the concept of case is supported if PCRE is compiled
3869 with Unicode property support, but not otherwise. If you want to use
3870 caseless matching in UTF8-mode for characters 128 and above, you must
3871 ensure that PCRE is compiled with Unicode property support as well as
3872 with UTF-8 support.
3873
3874 Characters that might indicate line breaks are never treated in any
3875 special way when matching character classes, whatever line-ending
3876 sequence is in use, and whatever setting of the PCRE_DOTALL and
3877 PCRE_MULTILINE options is used. A class such as [^a] always matches one
3878 of these characters.
3879
3880 The minus (hyphen) character can be used to specify a range of charac-
3881 ters in a character class. For example, [d-m] matches any letter
3882 between d and m, inclusive. If a minus character is required in a
3883 class, it must be escaped with a backslash or appear in a position
3884 where it cannot be interpreted as indicating a range, typically as the
3885 first or last character in the class.
3886
3887 It is not possible to have the literal character "]" as the end charac-
3888 ter of a range. A pattern such as [W-]46] is interpreted as a class of
3889 two characters ("W" and "-") followed by a literal string "46]", so it
3890 would match "W46]" or "-46]". However, if the "]" is escaped with a
3891 backslash it is interpreted as the end of range, so [W-\]46] is inter-
3892 preted as a class containing a range followed by two other characters.
3893 The octal or hexadecimal representation of "]" can also be used to end
3894 a range.
3895
3896 Ranges operate in the collating sequence of character values. They can
3897 also be used for characters specified numerically, for example
3898 [\000-\037]. In UTF-8 mode, ranges can include characters whose values
3899 are greater than 255, for example [\x{100}-\x{2ff}].
3900
3901 If a range that includes letters is used when caseless matching is set,
3902 it matches the letters in either case. For example, [W-c] is equivalent
3903 to [][\\^_`wxyzabc], matched caselessly, and in non-UTF-8 mode, if
3904 character tables for a French locale are in use, [\xc8-\xcb] matches
3905 accented E characters in both cases. In UTF-8 mode, PCRE supports the
3906 concept of case for characters with values greater than 128 only when
3907 it is compiled with Unicode property support.
3908
3909 The character types \d, \D, \p, \P, \s, \S, \w, and \W may also appear
3910 in a character class, and add the characters that they match to the
3911 class. For example, [\dABCDEF] matches any hexadecimal digit. A circum-
3912 flex can conveniently be used with the upper case character types to
3913 specify a more restricted set of characters than the matching lower
3914 case type. For example, the class [^\W_] matches any letter or digit,
3915 but not underscore.
3916
3917 The only metacharacters that are recognized in character classes are
3918 backslash, hyphen (only where it can be interpreted as specifying a
3919 range), circumflex (only at the start), opening square bracket (only
3920 when it can be interpreted as introducing a POSIX class name - see the
3921 next section), and the terminating closing square bracket. However,
3922 escaping other non-alphanumeric characters does no harm.
3923
3924
3925 POSIX CHARACTER CLASSES
3926
3927 Perl supports the POSIX notation for character classes. This uses names
3928 enclosed by [: and :] within the enclosing square brackets. PCRE also
3929 supports this notation. For example,
3930
3931 [01[:alpha:]%]
3932
3933 matches "0", "1", any alphabetic character, or "%". The supported class
3934 names are
3935
3936 alnum letters and digits
3937 alpha letters
3938 ascii character codes 0 - 127
3939 blank space or tab only
3940 cntrl control characters
3941 digit decimal digits (same as \d)
3942 graph printing characters, excluding space
3943 lower lower case letters
3944 print printing characters, including space
3945 punct printing characters, excluding letters and digits
3946 space white space (not quite the same as \s)
3947 upper upper case letters
3948 word "word" characters (same as \w)
3949 xdigit hexadecimal digits
3950
3951 The "space" characters are HT (9), LF (10), VT (11), FF (12), CR (13),
3952 and space (32). Notice that this list includes the VT character (code
3953 11). This makes "space" different to \s, which does not include VT (for
3954 Perl compatibility).
3955
3956 The name "word" is a Perl extension, and "blank" is a GNU extension
3957 from Perl 5.8. Another Perl extension is negation, which is indicated
3958 by a ^ character after the colon. For example,
3959
3960 [12[:^digit:]]
3961
3962 matches "1", "2", or any non-digit. PCRE (and Perl) also recognize the
3963 POSIX syntax [.ch.] and [=ch=] where "ch" is a "collating element", but
3964 these are not supported, and an error is given if they are encountered.
3965
3966 In UTF-8 mode, characters with values greater than 128 do not match any
3967 of the POSIX character classes.
3968
3969
3970 VERTICAL BAR
3971
3972 Vertical bar characters are used to separate alternative patterns. For
3973 example, the pattern
3974
3975 gilbert|sullivan
3976
3977 matches either "gilbert" or "sullivan". Any number of alternatives may
3978 appear, and an empty alternative is permitted (matching the empty
3979 string). The matching process tries each alternative in turn, from left
3980 to right, and the first one that succeeds is used. If the alternatives
3981 are within a subpattern (defined below), "succeeds" means matching the
3982 rest of the main pattern as well as the alternative in the subpattern.
3983
3984
3985 INTERNAL OPTION SETTING
3986
3987 The settings of the PCRE_CASELESS, PCRE_MULTILINE, PCRE_DOTALL, and
3988 PCRE_EXTENDED options (which are Perl-compatible) can be changed from
3989 within the pattern by a sequence of Perl option letters enclosed
3990 between "(?" and ")". The option letters are
3991
3992 i for PCRE_CASELESS
3993 m for PCRE_MULTILINE
3994 s for PCRE_DOTALL
3995 x for PCRE_EXTENDED
3996
3997 For example, (?im) sets caseless, multiline matching. It is also possi-
3998 ble to unset these options by preceding the letter with a hyphen, and a
3999 combined setting and unsetting such as (?im-sx), which sets PCRE_CASE-
4000 LESS and PCRE_MULTILINE while unsetting PCRE_DOTALL and PCRE_EXTENDED,
4001 is also permitted. If a letter appears both before and after the
4002 hyphen, the option is unset.
4003
4004 The PCRE-specific options PCRE_DUPNAMES, PCRE_UNGREEDY, and PCRE_EXTRA
4005 can be changed in the same way as the Perl-compatible options by using
4006 the characters J, U and X respectively.
4007
4008 When one of these option changes occurs at top level (that is, not
4009 inside subpattern parentheses), the change applies to the remainder of
4010 the pattern that follows. If the change is placed right at the start of
4011 a pattern, PCRE extracts it into the global options (and it will there-
4012 fore show up in data extracted by the pcre_fullinfo() function).
4013
4014 An option change within a subpattern (see below for a description of
4015 subpatterns) affects only that part of the current pattern that follows
4016 it, so
4017
4018 (a(?i)b)c
4019
4020 matches abc and aBc and no other strings (assuming PCRE_CASELESS is not
4021 used). By this means, options can be made to have different settings
4022 in different parts of the pattern. Any changes made in one alternative
4023 do carry on into subsequent branches within the same subpattern. For
4024 example,
4025
4026 (a(?i)b|c)
4027
4028 matches "ab", "aB", "c", and "C", even though when matching "C" the
4029 first branch is abandoned before the option setting. This is because
4030 the effects of option settings happen at compile time. There would be
4031 some very weird behaviour otherwise.
4032
4033 Note: There are other PCRE-specific options that can be set by the
4034 application when the compile or match functions are called. In some
4035 cases the pattern can contain special leading sequences such as (*CRLF)
4036 to override what the application has set or what has been defaulted.
4037 Details are given in the section entitled "Newline sequences" above.
4038 There is also the (*UTF8) leading sequence that can be used to set
4039 UTF-8 mode; this is equivalent to setting the PCRE_UTF8 option.
4040
4041
4042 SUBPATTERNS
4043
4044 Subpatterns are delimited by parentheses (round brackets), which can be
4045 nested. Turning part of a pattern into a subpattern does two things:
4046
4047 1. It localizes a set of alternatives. For example, the pattern
4048
4049 cat(aract|erpillar|)
4050
4051 matches one of the words "cat", "cataract", or "caterpillar". Without
4052 the parentheses, it would match "cataract", "erpillar" or an empty
4053 string.
4054
4055 2. It sets up the subpattern as a capturing subpattern. This means
4056 that, when the whole pattern matches, that portion of the subject
4057 string that matched the subpattern is passed back to the caller via the
4058 ovector argument of pcre_exec(). Opening parentheses are counted from
4059 left to right (starting from 1) to obtain numbers for the capturing
4060 subpatterns.
4061
4062 For example, if the string "the red king" is matched against the pat-
4063 tern
4064
4065 the ((red|white) (king|queen))
4066
4067 the captured substrings are "red king", "red", and "king", and are num-
4068 bered 1, 2, and 3, respectively.
4069
4070 The fact that plain parentheses fulfil two functions is not always
4071 helpful. There are often times when a grouping subpattern is required
4072 without a capturing requirement. If an opening parenthesis is followed
4073 by a question mark and a colon, the subpattern does not do any captur-
4074 ing, and is not counted when computing the number of any subsequent
4075 capturing subpatterns. For example, if the string "the white queen" is
4076 matched against the pattern
4077
4078 the ((?:red|white) (king|queen))
4079
4080 the captured substrings are "white queen" and "queen", and are numbered
4081 1 and 2. The maximum number of capturing subpatterns is 65535.
4082
4083 As a convenient shorthand, if any option settings are required at the
4084 start of a non-capturing subpattern, the option letters may appear
4085 between the "?" and the ":". Thus the two patterns
4086
4087 (?i:saturday|sunday)
4088 (?:(?i)saturday|sunday)
4089
4090 match exactly the same set of strings. Because alternative branches are
4091 tried from left to right, and options are not reset until the end of
4092 the subpattern is reached, an option setting in one branch does affect
4093 subsequent branches, so the above patterns match "SUNDAY" as well as
4094 "Saturday".
4095
4096
4097 DUPLICATE SUBPATTERN NUMBERS
4098
4099 Perl 5.10 introduced a feature whereby each alternative in a subpattern
4100 uses the same numbers for its capturing parentheses. Such a subpattern
4101 starts with (?| and is itself a non-capturing subpattern. For example,
4102 consider this pattern:
4103
4104 (?|(Sat)ur|(Sun))day
4105
4106 Because the two alternatives are inside a (?| group, both sets of cap-
4107 turing parentheses are numbered one. Thus, when the pattern matches,
4108 you can look at captured substring number one, whichever alternative
4109 matched. This construct is useful when you want to capture part, but
4110 not all, of one of a number of alternatives. Inside a (?| group, paren-
4111 theses are numbered as usual, but the number is reset at the start of
4112 each branch. The numbers of any capturing buffers that follow the sub-
4113 pattern start after the highest number used in any branch. The follow-
4114 ing example is taken from the Perl documentation. The numbers under-
4115 neath show in which buffer the captured content will be stored.
4116
4117 # before ---------------branch-reset----------- after
4118 / ( a ) (?| x ( y ) z | (p (q) r) | (t) u (v) ) ( z ) /x
4119 # 1 2 2 3 2 3 4
4120
4121 A back reference to a numbered subpattern uses the most recent value
4122 that is set for that number by any subpattern. The following pattern
4123 matches "abcabc" or "defdef":
4124
4125 /(?|(abc)|(def))\1/
4126
4127 In contrast, a recursive or "subroutine" call to a numbered subpattern
4128 always refers to the first one in the pattern with the given number.
4129 The following pattern matches "abcabc" or "defabc":
4130
4131 /(?|(abc)|(def))(?1)/
4132
4133 If a condition test for a subpattern's having matched refers to a non-
4134 unique number, the test is true if any of the subpatterns of that num-
4135 ber have matched.
4136
4137 An alternative approach to using this "branch reset" feature is to use
4138 duplicate named subpatterns, as described in the next section.
4139
4140
4141 NAMED SUBPATTERNS
4142
4143 Identifying capturing parentheses by number is simple, but it can be
4144 very hard to keep track of the numbers in complicated regular expres-
4145 sions. Furthermore, if an expression is modified, the numbers may
4146 change. To help with this difficulty, PCRE supports the naming of sub-
4147 patterns. This feature was not added to Perl until release 5.10. Python
4148 had the feature earlier, and PCRE introduced it at release 4.0, using
4149 the Python syntax. PCRE now supports both the Perl and the Python syn-
4150 tax. Perl allows identically numbered subpatterns to have different
4151 names, but PCRE does not.
4152
4153 In PCRE, a subpattern can be named in one of three ways: (?<name>...)
4154 or (?'name'...) as in Perl, or (?P<name>...) as in Python. References
4155 to capturing parentheses from other parts of the pattern, such as back
4156 references, recursion, and conditions, can be made by name as well as
4157 by number.
4158
4159 Names consist of up to 32 alphanumeric characters and underscores.
4160 Named capturing parentheses are still allocated numbers as well as
4161 names, exactly as if the names were not present. The PCRE API provides
4162 function calls for extracting the name-to-number translation table from
4163 a compiled pattern. There is also a convenience function for extracting
4164 a captured substring by name.
4165
4166 By default, a name must be unique within a pattern, but it is possible
4167 to relax this constraint by setting the PCRE_DUPNAMES option at compile
4168 time. (Duplicate names are also always permitted for subpatterns with
4169 the same number, set up as described in the previous section.) Dupli-
4170 cate names can be useful for patterns where only one instance of the
4171 named parentheses can match. Suppose you want to match the name of a
4172 weekday, either as a 3-letter abbreviation or as the full name, and in
4173 both cases you want to extract the abbreviation. This pattern (ignoring
4174 the line breaks) does the job:
4175
4176 (?<DN>Mon|Fri|Sun)(?:day)?|
4177 (?<DN>Tue)(?:sday)?|
4178 (?<DN>Wed)(?:nesday)?|
4179 (?<DN>Thu)(?:rsday)?|
4180 (?<DN>Sat)(?:urday)?
4181
4182 There are five capturing substrings, but only one is ever set after a
4183 match. (An alternative way of solving this problem is to use a "branch
4184 reset" subpattern, as described in the previous section.)
4185
4186 The convenience function for extracting the data by name returns the
4187 substring for the first (and in this example, the only) subpattern of
4188 that name that matched. This saves searching to find which numbered
4189 subpattern it was.
4190
4191 If you make a back reference to a non-unique named subpattern from
4192 elsewhere in the pattern, the one that corresponds to the first occur-
4193 rence of the name is used. In the absence of duplicate numbers (see the
4194 previous section) this is the one with the lowest number. If you use a
4195 named reference in a condition test (see the section about conditions
4196 below), either to check whether a subpattern has matched, or to check
4197 for recursion, all subpatterns with the same name are tested. If the
4198 condition is true for any one of them, the overall condition is true.
4199 This is the same behaviour as testing by number. For further details of
4200 the interfaces for handling named subpatterns, see the pcreapi documen-
4201 tation.
4202
4203 Warning: You cannot use different names to distinguish between two sub-
4204 patterns with the same number because PCRE uses only the numbers when
4205 matching. For this reason, an error is given at compile time if differ-
4206 ent names are given to subpatterns with the same number. However, you
4207 can give the same name to subpatterns with the same number, even when
4208 PCRE_DUPNAMES is not set.
4209
4210
4211 REPETITION
4212
4213 Repetition is specified by quantifiers, which can follow any of the
4214 following items:
4215
4216 a literal data character
4217 the dot metacharacter
4218 the \C escape sequence
4219 the \X escape sequence (in UTF-8 mode with Unicode properties)
4220 the \R escape sequence
4221 an escape such as \d that matches a single character
4222 a character class
4223 a back reference (see next section)
4224 a parenthesized subpattern (unless it is an assertion)
4225 a recursive or "subroutine" call to a subpattern
4226
4227 The general repetition quantifier specifies a minimum and maximum num-
4228 ber of permitted matches, by giving the two numbers in curly brackets
4229 (braces), separated by a comma. The numbers must be less than 65536,
4230 and the first must be less than or equal to the second. For example:
4231
4232 z{2,4}
4233
4234 matches "zz", "zzz", or "zzzz". A closing brace on its own is not a
4235 special character. If the second number is omitted, but the comma is
4236 present, there is no upper limit; if the second number and the comma
4237 are both omitted, the quantifier specifies an exact number of required
4238 matches. Thus
4239
4240 [aeiou]{3,}
4241
4242 matches at least 3 successive vowels, but may match many more, while
4243
4244 \d{8}
4245
4246 matches exactly 8 digits. An opening curly bracket that appears in a
4247 position where a quantifier is not allowed, or one that does not match
4248 the syntax of a quantifier, is taken as a literal character. For exam-
4249 ple, {,6} is not a quantifier, but a literal string of four characters.
4250
4251 In UTF-8 mode, quantifiers apply to UTF-8 characters rather than to
4252 individual bytes. Thus, for example, \x{100}{2} matches two UTF-8 char-
4253 acters, each of which is represented by a two-byte sequence. Similarly,
4254 when Unicode property support is available, \X{3} matches three Unicode
4255 extended sequences, each of which may be several bytes long (and they
4256 may be of different lengths).
4257
4258 The quantifier {0} is permitted, causing the expression to behave as if
4259 the previous item and the quantifier were not present. This may be use-
4260 ful for subpatterns that are referenced as subroutines from elsewhere
4261 in the pattern. Items other than subpatterns that have a {0} quantifier
4262 are omitted from the compiled pattern.
4263
4264 For convenience, the three most common quantifiers have single-charac-
4265 ter abbreviations:
4266
4267 * is equivalent to {0,}
4268 + is equivalent to {1,}
4269 ? is equivalent to {0,1}
4270
4271 It is possible to construct infinite loops by following a subpattern
4272 that can match no characters with a quantifier that has no upper limit,
4273 for example:
4274
4275 (a?)*
4276
4277 Earlier versions of Perl and PCRE used to give an error at compile time
4278 for such patterns. However, because there are cases where this can be
4279 useful, such patterns are now accepted, but if any repetition of the
4280 subpattern does in fact match no characters, the loop is forcibly bro-
4281 ken.
4282
4283 By default, the quantifiers are "greedy", that is, they match as much
4284 as possible (up to the maximum number of permitted times), without
4285 causing the rest of the pattern to fail. The classic example of where
4286 this gives problems is in trying to match comments in C programs. These
4287 appear between /* and */ and within the comment, individual * and /
4288 characters may appear. An attempt to match C comments by applying the
4289 pattern
4290
4291 /\*.*\*/
4292
4293 to the string
4294
4295 /* first comment */ not comment /* second comment */
4296
4297 fails, because it matches the entire string owing to the greediness of
4298 the .* item.
4299
4300 However, if a quantifier is followed by a question mark, it ceases to
4301 be greedy, and instead matches the minimum number of times possible, so
4302 the pattern
4303
4304 /\*.*?\*/
4305
4306 does the right thing with the C comments. The meaning of the various
4307 quantifiers is not otherwise changed, just the preferred number of
4308 matches. Do not confuse this use of question mark with its use as a
4309 quantifier in its own right. Because it has two uses, it can sometimes
4310 appear doubled, as in
4311
4312 \d??\d
4313
4314 which matches one digit by preference, but can match two if that is the
4315 only way the rest of the pattern matches.
4316
4317 If the PCRE_UNGREEDY option is set (an option that is not available in
4318 Perl), the quantifiers are not greedy by default, but individual ones
4319 can be made greedy by following them with a question mark. In other
4320 words, it inverts the default behaviour.
4321
4322 When a parenthesized subpattern is quantified with a minimum repeat
4323 count that is greater than 1 or with a limited maximum, more memory is
4324 required for the compiled pattern, in proportion to the size of the
4325 minimum or maximum.
4326
4327 If a pattern starts with .* or .{0,} and the PCRE_DOTALL option (equiv-
4328 alent to Perl's /s) is set, thus allowing the dot to match newlines,
4329 the pattern is implicitly anchored, because whatever follows will be
4330 tried against every character position in the subject string, so there
4331 is no point in retrying the overall match at any position after the
4332 first. PCRE normally treats such a pattern as though it were preceded
4333 by \A.
4334
4335 In cases where it is known that the subject string contains no new-
4336 lines, it is worth setting PCRE_DOTALL in order to obtain this opti-
4337 mization, or alternatively using ^ to indicate anchoring explicitly.
4338
4339 However, there is one situation where the optimization cannot be used.
4340 When .* is inside capturing parentheses that are the subject of a back
4341 reference elsewhere in the pattern, a match at the start may fail where
4342 a later one succeeds. Consider, for example:
4343
4344 (.*)abc\1
4345
4346 If the subject is "xyz123abc123" the match point is the fourth charac-
4347 ter. For this reason, such a pattern is not implicitly anchored.
4348
4349 When a capturing subpattern is repeated, the value captured is the sub-
4350 string that matched the final iteration. For example, after
4351
4352 (tweedle[dume]{3}\s*)+
4353
4354 has matched "tweedledum tweedledee" the value of the captured substring
4355 is "tweedledee". However, if there are nested capturing subpatterns,
4356 the corresponding captured values may have been set in previous itera-
4357 tions. For example, after
4358
4359 /(a|(b))+/
4360
4361 matches "aba" the value of the second captured substring is "b".
4362
4363
4364 ATOMIC GROUPING AND POSSESSIVE QUANTIFIERS
4365
4366 With both maximizing ("greedy") and minimizing ("ungreedy" or "lazy")
4367 repetition, failure of what follows normally causes the repeated item
4368 to be re-evaluated to see if a different number of repeats allows the
4369 rest of the pattern to match. Sometimes it is useful to prevent this,
4370 either to change the nature of the match, or to cause it fail earlier
4371 than it otherwise might, when the author of the pattern knows there is
4372 no point in carrying on.
4373
4374 Consider, for example, the pattern \d+foo when applied to the subject
4375 line
4376
4377 123456bar
4378
4379 After matching all 6 digits and then failing to match "foo", the normal
4380 action of the matcher is to try again with only 5 digits matching the
4381 \d+ item, and then with 4, and so on, before ultimately failing.
4382 "Atomic grouping" (a term taken from Jeffrey Friedl's book) provides
4383 the means for specifying that once a subpattern has matched, it is not
4384 to be re-evaluated in this way.
4385
4386 If we use atomic grouping for the previous example, the matcher gives
4387 up immediately on failing to match "foo" the first time. The notation
4388 is a kind of special parenthesis, starting with (?> as in this example:
4389
4390 (?>\d+)foo
4391
4392 This kind of parenthesis "locks up" the part of the pattern it con-
4393 tains once it has matched, and a failure further into the pattern is
4394 prevented from backtracking into it. Backtracking past it to previous
4395 items, however, works as normal.
4396
4397 An alternative description is that a subpattern of this type matches
4398 the string of characters that an identical standalone pattern would
4399 match, if anchored at the current point in the subject string.
4400
4401 Atomic grouping subpatterns are not capturing subpatterns. Simple cases
4402 such as the above example can be thought of as a maximizing repeat that
4403 must swallow everything it can. So, while both \d+ and \d+? are pre-
4404 pared to adjust the number of digits they match in order to make the
4405 rest of the pattern match, (?>\d+) can only match an entire sequence of
4406 digits.
4407
4408 Atomic groups in general can of course contain arbitrarily complicated
4409 subpatterns, and can be nested. However, when the subpattern for an
4410 atomic group is just a single repeated item, as in the example above, a
4411 simpler notation, called a "possessive quantifier" can be used. This
4412 consists of an additional + character following a quantifier. Using
4413 this notation, the previous example can be rewritten as
4414
4415 \d++foo
4416
4417 Note that a possessive quantifier can be used with an entire group, for
4418 example:
4419
4420 (abc|xyz){2,3}+
4421
4422 Possessive quantifiers are always greedy; the setting of the
4423 PCRE_UNGREEDY option is ignored. They are a convenient notation for the
4424 simpler forms of atomic group. However, there is no difference in the
4425 meaning of a possessive quantifier and the equivalent atomic group,
4426 though there may be a performance difference; possessive quantifiers
4427 should be slightly faster.
4428
4429 The possessive quantifier syntax is an extension to the Perl 5.8 syn-
4430 tax. Jeffrey Friedl originated the idea (and the name) in the first
4431 edition of his book. Mike McCloskey liked it, so implemented it when he
4432 built Sun's Java package, and PCRE copied it from there. It ultimately
4433 found its way into Perl at release 5.10.
4434
4435 PCRE has an optimization that automatically "possessifies" certain sim-
4436 ple pattern constructs. For example, the sequence A+B is treated as
4437 A++B because there is no point in backtracking into a sequence of A's
4438 when B must follow.
4439
4440 When a pattern contains an unlimited repeat inside a subpattern that
4441 can itself be repeated an unlimited number of times, the use of an
4442 atomic group is the only way to avoid some failing matches taking a
4443 very long time indeed. The pattern
4444
4445 (\D+|<\d+>)*[!?]
4446
4447 matches an unlimited number of substrings that either consist of non-
4448 digits, or digits enclosed in <>, followed by either ! or ?. When it
4449 matches, it runs quickly. However, if it is applied to
4450
4451 aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
4452
4453 it takes a long time before reporting failure. This is because the
4454 string can be divided between the internal \D+ repeat and the external
4455 * repeat in a large number of ways, and all have to be tried. (The
4456 example uses [!?] rather than a single character at the end, because
4457 both PCRE and Perl have an optimization that allows for fast failure
4458 when a single character is used. They remember the last single charac-
4459 ter that is required for a match, and fail early if it is not present
4460 in the string.) If the pattern is changed so that it uses an atomic
4461 group, like this:
4462
4463 ((?>\D+)|<\d+>)*[!?]
4464
4465 sequences of non-digits cannot be broken, and failure happens quickly.
4466
4467
4468 BACK REFERENCES
4469
4470 Outside a character class, a backslash followed by a digit greater than
4471 0 (and possibly further digits) is a back reference to a capturing sub-
4472 pattern earlier (that is, to its left) in the pattern, provided there
4473 have been that many previous capturing left parentheses.
4474
4475 However, if the decimal number following the backslash is less than 10,
4476 it is always taken as a back reference, and causes an error only if
4477 there are not that many capturing left parentheses in the entire pat-
4478 tern. In other words, the parentheses that are referenced need not be
4479 to the left of the reference for numbers less than 10. A "forward back
4480 reference" of this type can make sense when a repetition is involved
4481 and the subpattern to the right has participated in an earlier itera-
4482 tion.
4483
4484 It is not possible to have a numerical "forward back reference" to a
4485 subpattern whose number is 10 or more using this syntax because a
4486 sequence such as \50 is interpreted as a character defined in octal.
4487 See the subsection entitled "Non-printing characters" above for further
4488 details of the handling of digits following a backslash. There is no
4489 such problem when named parentheses are used. A back reference to any
4490 subpattern is possible using named parentheses (see below).
4491
4492 Another way of avoiding the ambiguity inherent in the use of digits
4493 following a backslash is to use the \g escape sequence, which is a fea-
4494 ture introduced in Perl 5.10. This escape must be followed by an
4495 unsigned number or a negative number, optionally enclosed in braces.
4496 These examples are all identical:
4497
4498 (ring), \1
4499 (ring), \g1
4500 (ring), \g{1}
4501
4502 An unsigned number specifies an absolute reference without the ambigu-
4503 ity that is present in the older syntax. It is also useful when literal
4504 digits follow the reference. A negative number is a relative reference.
4505 Consider this example:
4506
4507 (abc(def)ghi)\g{-1}
4508
4509 The sequence \g{-1} is a reference to the most recently started captur-
4510 ing subpattern before \g, that is, is it equivalent to \2. Similarly,
4511 \g{-2} would be equivalent to \1. The use of relative references can be
4512 helpful in long patterns, and also in patterns that are created by
4513 joining together fragments that contain references within themselves.
4514
4515 A back reference matches whatever actually matched the capturing sub-
4516 pattern in the current subject string, rather than anything matching
4517 the subpattern itself (see "Subpatterns as subroutines" below for a way
4518 of doing that). So the pattern
4519
4520 (sens|respons)e and \1ibility
4521
4522 matches "sense and sensibility" and "response and responsibility", but
4523 not "sense and responsibility". If caseful matching is in force at the
4524 time of the back reference, the case of letters is relevant. For exam-
4525 ple,
4526
4527 ((?i)rah)\s+\1
4528
4529 matches "rah rah" and "RAH RAH", but not "RAH rah", even though the
4530 original capturing subpattern is matched caselessly.
4531
4532 There are several different ways of writing back references to named
4533 subpatterns. The .NET syntax \k{name} and the Perl syntax \k<name> or
4534 \k'name' are supported, as is the Python syntax (?P=name). Perl 5.10's
4535 unified back reference syntax, in which \g can be used for both numeric
4536 and named references, is also supported. We could rewrite the above
4537 example in any of the following ways:
4538
4539 (?<p1>(?i)rah)\s+\k<p1>
4540 (?'p1'(?i)rah)\s+\k{p1}
4541 (?P<p1>(?i)rah)\s+(?P=p1)
4542 (?<p1>(?i)rah)\s+\g{p1}
4543
4544 A subpattern that is referenced by name may appear in the pattern
4545 before or after the reference.
4546
4547 There may be more than one back reference to the same subpattern. If a
4548 subpattern has not actually been used in a particular match, any back
4549 references to it always fail by default. For example, the pattern
4550
4551 (a|(bc))\2
4552
4553 always fails if it starts to match "a" rather than "bc". However, if
4554 the PCRE_JAVASCRIPT_COMPAT option is set at compile time, a back refer-
4555 ence to an unset value matches an empty string.
4556
4557 Because there may be many capturing parentheses in a pattern, all dig-
4558 its following a backslash are taken as part of a potential back refer-
4559 ence number. If the pattern continues with a digit character, some
4560 delimiter must be used to terminate the back reference. If the
4561 PCRE_EXTENDED option is set, this can be whitespace. Otherwise, the \g{
4562 syntax or an empty comment (see "Comments" below) can be used.
4563
4564 Recursive back references
4565
4566 A back reference that occurs inside the parentheses to which it refers
4567 fails when the subpattern is first used, so, for example, (a\1) never
4568 matches. However, such references can be useful inside repeated sub-
4569 patterns. For example, the pattern
4570
4571 (a|b\1)+
4572
4573 matches any number of "a"s and also "aba", "ababbaa" etc. At each iter-
4574 ation of the subpattern, the back reference matches the character
4575 string corresponding to the previous iteration. In order for this to
4576 work, the pattern must be such that the first iteration does not need
4577 to match the back reference. This can be done using alternation, as in
4578 the example above, or by a quantifier with a minimum of zero.
4579
4580 Back references of this type cause the group that they reference to be
4581 treated as an atomic group. Once the whole group has been matched, a
4582 subsequent matching failure cannot cause backtracking into the middle
4583 of the group.
4584
4585
4586 ASSERTIONS
4587
4588 An assertion is a test on the characters following or preceding the
4589 current matching point that does not actually consume any characters.
4590 The simple assertions coded as \b, \B, \A, \G, \Z, \z, ^ and $ are
4591 described above.
4592
4593 More complicated assertions are coded as subpatterns. There are two
4594 kinds: those that look ahead of the current position in the subject
4595 string, and those that look behind it. An assertion subpattern is
4596 matched in the normal way, except that it does not cause the current
4597 matching position to be changed.
4598
4599 Assertion subpatterns are not capturing subpatterns, and may not be
4600 repeated, because it makes no sense to assert the same thing several
4601 times. If any kind of assertion contains capturing subpatterns within
4602 it, these are counted for the purposes of numbering the capturing sub-
4603 patterns in the whole pattern. However, substring capturing is carried
4604 out only for positive assertions, because it does not make sense for
4605 negative assertions.
4606
4607 Lookahead assertions
4608
4609 Lookahead assertions start with (?= for positive assertions and (?! for
4610 negative assertions. For example,
4611
4612 \w+(?=;)
4613
4614 matches a word followed by a semicolon, but does not include the semi-
4615 colon in the match, and
4616
4617 foo(?!bar)
4618
4619 matches any occurrence of "foo" that is not followed by "bar". Note
4620 that the apparently similar pattern
4621
4622 (?!foo)bar
4623
4624 does not find an occurrence of "bar" that is preceded by something
4625 other than "foo"; it finds any occurrence of "bar" whatsoever, because
4626 the assertion (?!foo) is always true when the next three characters are
4627 "bar". A lookbehind assertion is needed to achieve the other effect.
4628
4629 If you want to force a matching failure at some point in a pattern, the
4630 most convenient way to do it is with (?!) because an empty string
4631 always matches, so an assertion that requires there not to be an empty
4632 string must always fail. The Perl 5.10 backtracking control verb
4633 (*FAIL) or (*F) is essentially a synonym for (?!).
4634
4635 Lookbehind assertions
4636
4637 Lookbehind assertions start with (?<= for positive assertions and (?<!
4638 for negative assertions. For example,
4639
4640 (?<!foo)bar
4641
4642 does find an occurrence of "bar" that is not preceded by "foo". The
4643 contents of a lookbehind assertion are restricted such that all the
4644 strings it matches must have a fixed length. However, if there are sev-
4645 eral top-level alternatives, they do not all have to have the same
4646 fixed length. Thus
4647
4648 (?<=bullock|donkey)
4649
4650 is permitted, but
4651
4652 (?<!dogs?|cats?)
4653
4654 causes an error at compile time. Branches that match different length
4655 strings are permitted only at the top level of a lookbehind assertion.
4656 This is an extension compared with Perl (5.8 and 5.10), which requires
4657 all branches to match the same length of string. An assertion such as
4658
4659 (?<=ab(c|de))
4660
4661 is not permitted, because its single top-level branch can match two
4662 different lengths, but it is acceptable to PCRE if rewritten to use two
4663 top-level branches:
4664
4665 (?<=abc|abde)
4666
4667 In some cases, the Perl 5.10 escape sequence \K (see above) can be used
4668 instead of a lookbehind assertion to get round the fixed-length
4669 restriction.
4670
4671 The implementation of lookbehind assertions is, for each alternative,
4672 to temporarily move the current position back by the fixed length and
4673 then try to match. If there are insufficient characters before the cur-
4674 rent position, the assertion fails.
4675
4676 PCRE does not allow the \C escape (which matches a single byte in UTF-8
4677 mode) to appear in lookbehind assertions, because it makes it impossi-
4678 ble to calculate the length of the lookbehind. The \X and \R escapes,
4679 which can match different numbers of bytes, are also not permitted.
4680
4681 "Subroutine" calls (see below) such as (?2) or (?&X) are permitted in
4682 lookbehinds, as long as the subpattern matches a fixed-length string.
4683 Recursion, however, is not supported.
4684
4685 Possessive quantifiers can be used in conjunction with lookbehind
4686 assertions to specify efficient matching of fixed-length strings at the
4687 end of subject strings. Consider a simple pattern such as
4688
4689 abcd$
4690
4691 when applied to a long string that does not match. Because matching
4692 proceeds from left to right, PCRE will look for each "a" in the subject
4693 and then see if what follows matches the rest of the pattern. If the
4694 pattern is specified as
4695
4696 ^.*abcd$
4697
4698 the initial .* matches the entire string at first, but when this fails
4699 (because there is no following "a"), it backtracks to match all but the
4700 last character, then all but the last two characters, and so on. Once
4701 again the search for "a" covers the entire string, from right to left,
4702 so we are no better off. However, if the pattern is written as
4703
4704 ^.*+(?<=abcd)
4705
4706 there can be no backtracking for the .*+ item; it can match only the
4707 entire string. The subsequent lookbehind assertion does a single test
4708 on the last four characters. If it fails, the match fails immediately.
4709 For long strings, this approach makes a significant difference to the
4710 processing time.
4711
4712 Using multiple assertions
4713
4714 Several assertions (of any sort) may occur in succession. For example,
4715
4716 (?<=\d{3})(?<!999)foo
4717
4718 matches "foo" preceded by three digits that are not "999". Notice that
4719 each of the assertions is applied independently at the same point in
4720 the subject string. First there is a check that the previous three
4721 characters are all digits, and then there is a check that the same
4722 three characters are not "999". This pattern does not match "foo" pre-
4723 ceded by six characters, the first of which are digits and the last
4724 three of which are not "999". For example, it doesn't match "123abc-
4725 foo". A pattern to do that is
4726
4727 (?<=\d{3}...)(?<!999)foo
4728
4729 This time the first assertion looks at the preceding six characters,
4730 checking that the first three are digits, and then the second assertion
4731 checks that the preceding three characters are not "999".
4732
4733 Assertions can be nested in any combination. For example,
4734
4735 (?<=(?<!foo)bar)baz
4736
4737 matches an occurrence of "baz" that is preceded by "bar" which in turn
4738 is not preceded by "foo", while
4739
4740 (?<=\d{3}(?!999)...)foo
4741
4742 is another pattern that matches "foo" preceded by three digits and any
4743 three characters that are not "999".
4744
4745
4746 CONDITIONAL SUBPATTERNS
4747
4748 It is possible to cause the matching process to obey a subpattern con-
4749 ditionally or to choose between two alternative subpatterns, depending
4750 on the result of an assertion, or whether a specific capturing subpat-
4751 tern has already been matched. The two possible forms of conditional
4752 subpattern are:
4753
4754 (?(condition)yes-pattern)
4755 (?(condition)yes-pattern|no-pattern)
4756
4757 If the condition is satisfied, the yes-pattern is used; otherwise the
4758 no-pattern (if present) is used. If there are more than two alterna-
4759 tives in the subpattern, a compile-time error occurs.
4760
4761 There are four kinds of condition: references to subpatterns, refer-
4762 ences to recursion, a pseudo-condition called DEFINE, and assertions.
4763
4764 Checking for a used subpattern by number
4765
4766 If the text between the parentheses consists of a sequence of digits,
4767 the condition is true if a capturing subpattern of that number has pre-
4768 viously matched. If there is more than one capturing subpattern with
4769 the same number (see the earlier section about duplicate subpattern
4770 numbers), the condition is true if any of them have been set. An alter-
4771 native notation is to precede the digits with a plus or minus sign. In
4772 this case, the subpattern number is relative rather than absolute. The
4773 most recently opened parentheses can be referenced by (?(-1), the next
4774 most recent by (?(-2), and so on. In looping constructs it can also
4775 make sense to refer to subsequent groups with constructs such as
4776 (?(+2).
4777
4778 Consider the following pattern, which contains non-significant white
4779 space to make it more readable (assume the PCRE_EXTENDED option) and to
4780 divide it into three parts for ease of discussion:
4781
4782 ( \( )? [^()]+ (?(1) \) )
4783
4784 The first part matches an optional opening parenthesis, and if that
4785 character is present, sets it as the first captured substring. The sec-
4786 ond part matches one or more characters that are not parentheses. The
4787 third part is a conditional subpattern that tests whether the first set
4788 of parentheses matched or not. If they did, that is, if subject started
4789 with an opening parenthesis, the condition is true, and so the yes-pat-
4790 tern is executed and a closing parenthesis is required. Otherwise,
4791 since no-pattern is not present, the subpattern matches nothing. In
4792 other words, this pattern matches a sequence of non-parentheses,
4793 optionally enclosed in parentheses.
4794
4795 If you were embedding this pattern in a larger one, you could use a
4796 relative reference:
4797
4798 ...other stuff... ( \( )? [^()]+ (?(-1) \) ) ...
4799
4800 This makes the fragment independent of the parentheses in the larger
4801 pattern.
4802
4803 Checking for a used subpattern by name
4804
4805 Perl uses the syntax (?(<name>)...) or (?('name')...) to test for a
4806 used subpattern by name. For compatibility with earlier versions of
4807 PCRE, which had this facility before Perl, the syntax (?(name)...) is
4808 also recognized. However, there is a possible ambiguity with this syn-
4809 tax, because subpattern names may consist entirely of digits. PCRE
4810 looks first for a named subpattern; if it cannot find one and the name
4811 consists entirely of digits, PCRE looks for a subpattern of that num-
4812 ber, which must be greater than zero. Using subpattern names that con-
4813 sist entirely of digits is not recommended.
4814
4815 Rewriting the above example to use a named subpattern gives this:
4816
4817 (?<OPEN> \( )? [^()]+ (?(<OPEN>) \) )
4818
4819 If the name used in a condition of this kind is a duplicate, the test
4820 is applied to all subpatterns of the same name, and is true if any one
4821 of them has matched.
4822
4823 Checking for pattern recursion
4824
4825 If the condition is the string (R), and there is no subpattern with the
4826 name R, the condition is true if a recursive call to the whole pattern
4827 or any subpattern has been made. If digits or a name preceded by amper-
4828 sand follow the letter R, for example:
4829
4830 (?(R3)...) or (?(R&name)...)
4831
4832 the condition is true if the most recent recursion is into a subpattern
4833 whose number or name is given. This condition does not check the entire
4834 recursion stack. If the name used in a condition of this kind is a
4835 duplicate, the test is applied to all subpatterns of the same name, and
4836 is true if any one of them is the most recent recursion.
4837
4838 At "top level", all these recursion test conditions are false. The
4839 syntax for recursive patterns is described below.
4840
4841 Defining subpatterns for use by reference only
4842
4843 If the condition is the string (DEFINE), and there is no subpattern
4844 with the name DEFINE, the condition is always false. In this case,
4845 there may be only one alternative in the subpattern. It is always
4846 skipped if control reaches this point in the pattern; the idea of
4847 DEFINE is that it can be used to define "subroutines" that can be ref-
4848 erenced from elsewhere. (The use of "subroutines" is described below.)
4849 For example, a pattern to match an IPv4 address could be written like
4850 this (ignore whitespace and line breaks):
4851
4852 (?(DEFINE) (?<byte> 2[0-4]\d | 25[0-5] | 1\d\d | [1-9]?\d) )
4853 \b (?&byte) (\.(?&byte)){3} \b
4854
4855 The first part of the pattern is a DEFINE group inside which a another
4856 group named "byte" is defined. This matches an individual component of
4857 an IPv4 address (a number less than 256). When matching takes place,
4858 this part of the pattern is skipped because DEFINE acts like a false
4859 condition. The rest of the pattern uses references to the named group
4860 to match the four dot-separated components of an IPv4 address, insist-
4861 ing on a word boundary at each end.
4862
4863 Assertion conditions
4864
4865 If the condition is not in any of the above formats, it must be an
4866 assertion. This may be a positive or negative lookahead or lookbehind
4867 assertion. Consider this pattern, again containing non-significant
4868 white space, and with the two alternatives on the second line:
4869
4870 (?(?=[^a-z]*[a-z])
4871 \d{2}-[a-z]{3}-\d{2} | \d{2}-\d{2}-\d{2} )
4872
4873 The condition is a positive lookahead assertion that matches an
4874 optional sequence of non-letters followed by a letter. In other words,
4875 it tests for the presence of at least one letter in the subject. If a
4876 letter is found, the subject is matched against the first alternative;
4877 otherwise it is matched against the second. This pattern matches
4878 strings in one of the two forms dd-aaa-dd or dd-dd-dd, where aaa are
4879 letters and dd are digits.
4880
4881
4882 COMMENTS
4883
4884 The sequence (?# marks the start of a comment that continues up to the
4885 next closing parenthesis. Nested parentheses are not permitted. The
4886 characters that make up a comment play no part in the pattern matching
4887 at all.
4888
4889 If the PCRE_EXTENDED option is set, an unescaped # character outside a
4890 character class introduces a comment that continues to immediately
4891 after the next newline in the pattern.
4892
4893
4894 RECURSIVE PATTERNS
4895
4896 Consider the problem of matching a string in parentheses, allowing for
4897 unlimited nested parentheses. Without the use of recursion, the best
4898 that can be done is to use a pattern that matches up to some fixed
4899 depth of nesting. It is not possible to handle an arbitrary nesting
4900 depth.
4901
4902 For some time, Perl has provided a facility that allows regular expres-
4903 sions to recurse (amongst other things). It does this by interpolating
4904 Perl code in the expression at run time, and the code can refer to the
4905 expression itself. A Perl pattern using code interpolation to solve the
4906 parentheses problem can be created like this:
4907
4908 $re = qr{\( (?: (?>[^()]+) | (?p{$re}) )* \)}x;
4909
4910 The (?p{...}) item interpolates Perl code at run time, and in this case
4911 refers recursively to the pattern in which it appears.
4912
4913 Obviously, PCRE cannot support the interpolation of Perl code. Instead,
4914 it supports special syntax for recursion of the entire pattern, and
4915 also for individual subpattern recursion. After its introduction in
4916 PCRE and Python, this kind of recursion was subsequently introduced
4917 into Perl at release 5.10.
4918
4919 A special item that consists of (? followed by a number greater than
4920 zero and a closing parenthesis is a recursive call of the subpattern of
4921 the given number, provided that it occurs inside that subpattern. (If
4922 not, it is a "subroutine" call, which is described in the next sec-
4923 tion.) The special item (?R) or (?0) is a recursive call of the entire
4924 regular expression.
4925
4926 This PCRE pattern solves the nested parentheses problem (assume the
4927 PCRE_EXTENDED option is set so that white space is ignored):
4928
4929 \( ( [^()]++ | (?R) )* \)
4930
4931 First it matches an opening parenthesis. Then it matches any number of
4932 substrings which can either be a sequence of non-parentheses, or a
4933 recursive match of the pattern itself (that is, a correctly parenthe-
4934 sized substring). Finally there is a closing parenthesis. Note the use
4935 of a possessive quantifier to avoid backtracking into sequences of non-
4936 parentheses.
4937
4938 If this were part of a larger pattern, you would not want to recurse
4939 the entire pattern, so instead you could use this:
4940
4941 ( \( ( [^()]++ | (?1) )* \) )
4942
4943 We have put the pattern into parentheses, and caused the recursion to
4944 refer to them instead of the whole pattern.
4945
4946 In a larger pattern, keeping track of parenthesis numbers can be
4947 tricky. This is made easier by the use of relative references (a Perl
4948 5.10 feature). Instead of (?1) in the pattern above you can write
4949 (?-2) to refer to the second most recently opened parentheses preceding
4950 the recursion. In other words, a negative number counts capturing
4951 parentheses leftwards from the point at which it is encountered.
4952
4953 It is also possible to refer to subsequently opened parentheses, by
4954 writing references such as (?+2). However, these cannot be recursive
4955 because the reference is not inside the parentheses that are refer-
4956 enced. They are always "subroutine" calls, as described in the next
4957 section.
4958
4959 An alternative approach is to use named parentheses instead. The Perl
4960 syntax for this is (?&name); PCRE's earlier syntax (?P>name) is also
4961 supported. We could rewrite the above example as follows:
4962
4963 (?<pn> \( ( [^()]++ | (?&pn) )* \) )
4964
4965 If there is more than one subpattern with the same name, the earliest
4966 one is used.
4967
4968 This particular example pattern that we have been looking at contains
4969 nested unlimited repeats, and so the use of a possessive quantifier for
4970 matching strings of non-parentheses is important when applying the pat-
4971 tern to strings that do not match. For example, when this pattern is
4972 applied to
4973
4974 (aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa()
4975
4976 it yields "no match" quickly. However, if a possessive quantifier is
4977 not used, the match runs for a very long time indeed because there are
4978 so many different ways the + and * repeats can carve up the subject,
4979 and all have to be tested before failure can be reported.
4980
4981 At the end of a match, the values of capturing parentheses are those
4982 from the outermost level. If you want to obtain intermediate values, a
4983 callout function can be used (see below and the pcrecallout documenta-
4984 tion). If the pattern above is matched against
4985
4986 (ab(cd)ef)
4987
4988 the value for the inner capturing parentheses (numbered 2) is "ef",
4989 which is the last value taken on at the top level. If a capturing sub-
4990 pattern is not matched at the top level, its final value is unset, even
4991 if it is (temporarily) set at a deeper level.
4992
4993 If there are more than 15 capturing parentheses in a pattern, PCRE has
4994 to obtain extra memory to store data during a recursion, which it does
4995 by using pcre_malloc, freeing it via pcre_free afterwards. If no memory
4996 can be obtained, the match fails with the PCRE_ERROR_NOMEMORY error.
4997
4998 Do not confuse the (?R) item with the condition (R), which tests for
4999 recursion. Consider this pattern, which matches text in angle brack-
5000 ets, allowing for arbitrary nesting. Only digits are allowed in nested
5001 brackets (that is, when recursing), whereas any characters are permit-
5002 ted at the outer level.
5003
5004 < (?: (?(R) \d++ | [^<>]*+) | (?R)) * >
5005
5006 In this pattern, (?(R) is the start of a conditional subpattern, with
5007 two different alternatives for the recursive and non-recursive cases.
5008 The (?R) item is the actual recursive call.
5009
5010 Recursion difference from Perl
5011
5012 In PCRE (like Python, but unlike Perl), a recursive subpattern call is
5013 always treated as an atomic group. That is, once it has matched some of
5014 the subject string, it is never re-entered, even if it contains untried
5015 alternatives and there is a subsequent matching failure. This can be
5016 illustrated by the following pattern, which purports to match a palin-
5017 dromic string that contains an odd number of characters (for example,
5018 "a", "aba", "abcba", "abcdcba"):
5019
5020 ^(.|(.)(?1)\2)$
5021
5022 The idea is that it either matches a single character, or two identical
5023 characters surrounding a sub-palindrome. In Perl, this pattern works;
5024 in PCRE it does not if the pattern is longer than three characters.
5025 Consider the subject string "abcba":
5026
5027 At the top level, the first character is matched, but as it is not at
5028 the end of the string, the first alternative fails; the second alterna-
5029 tive is taken and the recursion kicks in. The recursive call to subpat-
5030 tern 1 successfully matches the next character ("b"). (Note that the
5031 beginning and end of line tests are not part of the recursion).
5032
5033 Back at the top level, the next character ("c") is compared with what
5034 subpattern 2 matched, which was "a". This fails. Because the recursion
5035 is treated as an atomic group, there are now no backtracking points,
5036 and so the entire match fails. (Perl is able, at this point, to re-
5037 enter the recursion and try the second alternative.) However, if the
5038 pattern is written with the alternatives in the other order, things are
5039 different:
5040
5041 ^((.)(?1)\2|.)$
5042
5043 This time, the recursing alternative is tried first, and continues to
5044 recurse until it runs out of characters, at which point the recursion
5045 fails. But this time we do have another alternative to try at the
5046 higher level. That is the big difference: in the previous case the
5047 remaining alternative is at a deeper recursion level, which PCRE cannot
5048 use.
5049
5050 To change the pattern so that matches all palindromic strings, not just
5051 those with an odd number of characters, it is tempting to change the
5052 pattern to this:
5053
5054 ^((.)(?1)\2|.?)$
5055
5056 Again, this works in Perl, but not in PCRE, and for the same reason.
5057 When a deeper recursion has matched a single character, it cannot be
5058 entered again in order to match an empty string. The solution is to
5059 separate the two cases, and write out the odd and even cases as alter-
5060 natives at the higher level:
5061
5062 ^(?:((.)(?1)\2|)|((.)(?3)\4|.))
5063
5064 If you want to match typical palindromic phrases, the pattern has to
5065 ignore all non-word characters, which can be done like this:
5066
5067 ^\W*+(?:((.)\W*+(?1)\W*+\2|)|((.)\W*+(?3)\W*+\4|\W*+.\W*+))\W*+$
5068
5069 If run with the PCRE_CASELESS option, this pattern matches phrases such
5070 as "A man, a plan, a canal: Panama!" and it works well in both PCRE and
5071 Perl. Note the use of the possessive quantifier *+ to avoid backtrack-
5072 ing into sequences of non-word characters. Without this, PCRE takes a
5073 great deal longer (ten times or more) to match typical phrases, and
5074 Perl takes so long that you think it has gone into a loop.
5075
5076 WARNING: The palindrome-matching patterns above work only if the sub-
5077 ject string does not start with a palindrome that is shorter than the
5078 entire string. For example, although "abcba" is correctly matched, if
5079 the subject is "ababa", PCRE finds the palindrome "aba" at the start,
5080 then fails at top level because the end of the string does not follow.
5081 Once again, it cannot jump back into the recursion to try other alter-
5082 natives, so the entire match fails.
5083
5084
5085 SUBPATTERNS AS SUBROUTINES
5086
5087 If the syntax for a recursive subpattern reference (either by number or
5088 by name) is used outside the parentheses to which it refers, it oper-
5089 ates like a subroutine in a programming language. The "called" subpat-
5090 tern may be defined before or after the reference. A numbered reference
5091 can be absolute or relative, as in these examples:
5092
5093 (...(absolute)...)...(?2)...
5094 (...(relative)...)...(?-1)...
5095 (...(?+1)...(relative)...
5096
5097 An earlier example pointed out that the pattern
5098
5099 (sens|respons)e and \1ibility
5100
5101 matches "sense and sensibility" and "response and responsibility", but
5102 not "sense and responsibility". If instead the pattern
5103
5104 (sens|respons)e and (?1)ibility
5105
5106 is used, it does match "sense and responsibility" as well as the other
5107 two strings. Another example is given in the discussion of DEFINE
5108 above.
5109
5110 Like recursive subpatterns, a subroutine call is always treated as an
5111 atomic group. That is, once it has matched some of the subject string,
5112 it is never re-entered, even if it contains untried alternatives and
5113 there is a subsequent matching failure. Any capturing parentheses that
5114 are set during the subroutine call revert to their previous values
5115 afterwards.
5116
5117 When a subpattern is used as a subroutine, processing options such as
5118 case-independence are fixed when the subpattern is defined. They cannot
5119 be changed for different calls. For example, consider this pattern:
5120
5121 (abc)(?i:(?-1))
5122
5123 It matches "abcabc". It does not match "abcABC" because the change of
5124 processing option does not affect the called subpattern.
5125
5126
5127 ONIGURUMA SUBROUTINE SYNTAX
5128
5129 For compatibility with Oniguruma, the non-Perl syntax \g followed by a
5130 name or a number enclosed either in angle brackets or single quotes, is
5131 an alternative syntax for referencing a subpattern as a subroutine,
5132 possibly recursively. Here are two of the examples used above, rewrit-
5133 ten using this syntax:
5134
5135 (?<pn> \( ( (?>[^()]+) | \g<pn> )* \) )
5136 (sens|respons)e and \g'1'ibility
5137
5138 PCRE supports an extension to Oniguruma: if a number is preceded by a
5139 plus or a minus sign it is taken as a relative reference. For example:
5140
5141 (abc)(?i:\g<-1>)
5142
5143 Note that \g{...} (Perl syntax) and \g<...> (Oniguruma syntax) are not
5144 synonymous. The former is a back reference; the latter is a subroutine
5145 call.
5146
5147
5148 CALLOUTS
5149
5150 Perl has a feature whereby using the sequence (?{...}) causes arbitrary
5151 Perl code to be obeyed in the middle of matching a regular expression.
5152 This makes it possible, amongst other things, to extract different sub-
5153 strings that match the same pair of parentheses when there is a repeti-
5154 tion.
5155
5156 PCRE provides a similar feature, but of course it cannot obey arbitrary
5157 Perl code. The feature is called "callout". The caller of PCRE provides
5158 an external function by putting its entry point in the global variable
5159 pcre_callout. By default, this variable contains NULL, which disables
5160 all calling out.
5161
5162 Within a regular expression, (?C) indicates the points at which the
5163 external function is to be called. If you want to identify different
5164 callout points, you can put a number less than 256 after the letter C.
5165 The default value is zero. For example, this pattern has two callout
5166 points:
5167
5168 (?C1)abc(?C2)def
5169
5170 If the PCRE_AUTO_CALLOUT flag is passed to pcre_compile(), callouts are
5171 automatically installed before each item in the pattern. They are all
5172 numbered 255.
5173
5174 During matching, when PCRE reaches a callout point (and pcre_callout is
5175 set), the external function is called. It is provided with the number
5176 of the callout, the position in the pattern, and, optionally, one item
5177 of data originally supplied by the caller of pcre_exec(). The callout
5178 function may cause matching to proceed, to backtrack, or to fail alto-
5179 gether. A complete description of the interface to the callout function
5180 is given in the pcrecallout documentation.
5181
5182
5183 BACKTRACKING CONTROL
5184
5185 Perl 5.10 introduced a number of "Special Backtracking Control Verbs",
5186 which are described in the Perl documentation as "experimental and sub-
5187 ject to change or removal in a future version of Perl". It goes on to
5188 say: "Their usage in production code should be noted to avoid problems
5189 during upgrades." The same remarks apply to the PCRE features described
5190 in this section.
5191
5192 Since these verbs are specifically related to backtracking, most of
5193 them can be used only when the pattern is to be matched using
5194 pcre_exec(), which uses a backtracking algorithm. With the exception of
5195 (*FAIL), which behaves like a failing negative assertion, they cause an
5196 error if encountered by pcre_dfa_exec().
5197
5198 If any of these verbs are used in an assertion or subroutine subpattern
5199 (including recursive subpatterns), their effect is confined to that
5200 subpattern; it does not extend to the surrounding pattern. Note that
5201 such subpatterns are processed as anchored at the point where they are
5202 tested.
5203
5204 The new verbs make use of what was previously invalid syntax: an open-
5205 ing parenthesis followed by an asterisk. They are generally of the form
5206 (*VERB) or (*VERB:NAME). Some may take either form, with differing be-
5207 haviour, depending on whether or not an argument is present. An name is
5208 a sequence of letters, digits, and underscores. If the name is empty,
5209 that is, if the closing parenthesis immediately follows the colon, the
5210 effect is as if the colon were not there. Any number of these verbs may
5211 occur in a pattern.
5212
5213 PCRE contains some optimizations that are used to speed up matching by
5214 running some checks at the start of each match attempt. For example, it
5215 may know the minimum length of matching subject, or that a particular
5216 character must be present. When one of these optimizations suppresses
5217 the running of a match, any included backtracking verbs will not, of
5218 course, be processed. You can suppress the start-of-match optimizations
5219 by setting the PCRE_NO_START_OPTIMIZE option when calling pcre_exec().
5220
5221 Verbs that act immediately
5222
5223 The following verbs act as soon as they are encountered. They may not
5224 be followed by a name.
5225
5226 (*ACCEPT)
5227
5228 This verb causes the match to end successfully, skipping the remainder
5229 of the pattern. When inside a recursion, only the innermost pattern is
5230 ended immediately. If (*ACCEPT) is inside capturing parentheses, the
5231 data so far is captured. (This feature was added to PCRE at release
5232 8.00.) For example:
5233
5234 A((?:A|B(*ACCEPT)|C)D)
5235
5236 This matches "AB", "AAD", or "ACD"; when it matches "AB", "B" is cap-
5237 tured by the outer parentheses.
5238
5239 (*FAIL) or (*F)
5240
5241 This verb causes the match to fail, forcing backtracking to occur. It
5242 is equivalent to (?!) but easier to read. The Perl documentation notes
5243 that it is probably useful only when combined with (?{}) or (??{}).
5244 Those are, of course, Perl features that are not present in PCRE. The
5245 nearest equivalent is the callout feature, as for example in this pat-
5246 tern:
5247
5248 a+(?C)(*FAIL)
5249
5250 A match with the string "aaaa" always fails, but the callout is taken
5251 before each backtrack happens (in this example, 10 times).
5252
5253 Recording which path was taken
5254
5255 There is one verb whose main purpose is to track how a match was
5256 arrived at, though it also has a secondary use in conjunction with
5257 advancing the match starting point (see (*SKIP) below).
5258
5259 (*MARK:NAME) or (*:NAME)
5260
5261 A name is always required with this verb. There may be as many
5262 instances of (*MARK) as you like in a pattern, and their names do not
5263 have to be unique.
5264
5265 When a match succeeds, the name of the last-encountered (*MARK) is
5266 passed back to the caller via the pcre_extra data structure, as
5267 described in the section on pcre_extra in the pcreapi documentation. No
5268 data is returned for a partial match. Here is an example of pcretest
5269 output, where the /K modifier requests the retrieval and outputting of
5270 (*MARK) data:
5271
5272 /X(*MARK:A)Y|X(*MARK:B)Z/K
5273 XY
5274 0: XY
5275 MK: A
5276 XZ
5277 0: XZ
5278 MK: B
5279
5280 The (*MARK) name is tagged with "MK:" in this output, and in this exam-
5281 ple it indicates which of the two alternatives matched. This is a more
5282 efficient way of obtaining this information than putting each alterna-
5283 tive in its own capturing parentheses.
5284
5285 A name may also be returned after a failed match if the final path
5286 through the pattern involves (*MARK). However, unless (*MARK) used in
5287 conjunction with (*COMMIT), this is unlikely to happen for an unan-
5288 chored pattern because, as the starting point for matching is advanced,
5289 the final check is often with an empty string, causing a failure before
5290 (*MARK) is reached. For example:
5291
5292 /X(*MARK:A)Y|X(*MARK:B)Z/K
5293 XP
5294 No match
5295
5296 There are three potential starting points for this match (starting with
5297 X, starting with P, and with an empty string). If the pattern is
5298 anchored, the result is different:
5299
5300 /^X(*MARK:A)Y|^X(*MARK:B)Z/K
5301 XP
5302 No match, mark = B
5303
5304 PCRE's start-of-match optimizations can also interfere with this. For
5305 example, if, as a result of a call to pcre_study(), it knows the mini-
5306 mum subject length for a match, a shorter subject will not be scanned
5307 at all.
5308
5309 Note that similar anomalies (though different in detail) exist in Perl,
5310 no doubt for the same reasons. The use of (*MARK) data after a failed
5311 match of an unanchored pattern is not recommended, unless (*COMMIT) is
5312 involved.
5313
5314 Verbs that act after backtracking
5315
5316 The following verbs do nothing when they are encountered. Matching con-
5317 tinues with what follows, but if there is no subsequent match, causing
5318 a backtrack to the verb, a failure is forced. That is, backtracking
5319 cannot pass to the left of the verb. However, when one of these verbs
5320 appears inside an atomic group, its effect is confined to that group,
5321 because once the group has been matched, there is never any backtrack-
5322 ing into it. In this situation, backtracking can "jump back" to the
5323 left of the entire atomic group. (Remember also, as stated above, that
5324 this localization also applies in subroutine calls and assertions.)
5325
5326 These verbs differ in exactly what kind of failure occurs when back-
5327 tracking reaches them.
5328
5329 (*COMMIT)
5330
5331 This verb, which may not be followed by a name, causes the whole match
5332 to fail outright if the rest of the pattern does not match. Even if the
5333 pattern is unanchored, no further attempts to find a match by advancing
5334 the starting point take place. Once (*COMMIT) has been passed,
5335 pcre_exec() is committed to finding a match at the current starting
5336 point, or not at all. For example:
5337
5338 a+(*COMMIT)b
5339
5340 This matches "xxaab" but not "aacaab". It can be thought of as a kind
5341 of dynamic anchor, or "I've started, so I must finish." The name of the
5342 most recently passed (*MARK) in the path is passed back when (*COMMIT)
5343 forces a match failure.
5344
5345 Note that (*COMMIT) at the start of a pattern is not the same as an
5346 anchor, unless PCRE's start-of-match optimizations are turned off, as
5347 shown in this pcretest example:
5348
5349 /(*COMMIT)abc/
5350 xyzabc
5351 0: abc
5352 xyzabc\Y
5353 No match
5354
5355 PCRE knows that any match must start with "a", so the optimization
5356 skips along the subject to "a" before running the first match attempt,
5357 which succeeds. When the optimization is disabled by the \Y escape in
5358 the second subject, the match starts at "x" and so the (*COMMIT) causes
5359 it to fail without trying any other starting points.
5360
5361 (*PRUNE) or (*PRUNE:NAME)
5362
5363 This verb causes the match to fail at the current starting position in
5364 the subject if the rest of the pattern does not match. If the pattern
5365 is unanchored, the normal "bumpalong" advance to the next starting
5366 character then happens. Backtracking can occur as usual to the left of
5367 (*PRUNE), before it is reached, or when matching to the right of
5368 (*PRUNE), but if there is no match to the right, backtracking cannot
5369 cross (*PRUNE). In simple cases, the use of (*PRUNE) is just an alter-
5370 native to an atomic group or possessive quantifier, but there are some
5371 uses of (*PRUNE) that cannot be expressed in any other way. The behav-
5372 iour of (*PRUNE:NAME) is the same as (*MARK:NAME)(*PRUNE) when the
5373 match fails completely; the name is passed back if this is the final
5374 attempt. (*PRUNE:NAME) does not pass back a name if the match suc-
5375 ceeds. In an anchored pattern (*PRUNE) has the same effect as (*COM-
5376 MIT).
5377
5378 (*SKIP)
5379
5380 This verb, when given without a name, is like (*PRUNE), except that if
5381 the pattern is unanchored, the "bumpalong" advance is not to the next
5382 character, but to the position in the subject where (*SKIP) was encoun-
5383 tered. (*SKIP) signifies that whatever text was matched leading up to
5384 it cannot be part of a successful match. Consider:
5385
5386 a+(*SKIP)b
5387
5388 If the subject is "aaaac...", after the first match attempt fails
5389 (starting at the first character in the string), the starting point
5390 skips on to start the next attempt at "c". Note that a possessive quan-
5391 tifer does not have the same effect as this example; although it would
5392 suppress backtracking during the first match attempt, the second
5393 attempt would start at the second character instead of skipping on to
5394 "c".
5395
5396 (*SKIP:NAME)
5397
5398 When (*SKIP) has an associated name, its behaviour is modified. If the
5399 following pattern fails to match, the previous path through the pattern
5400 is searched for the most recent (*MARK) that has the same name. If one
5401 is found, the "bumpalong" advance is to the subject position that cor-
5402 responds to that (*MARK) instead of to where (*SKIP) was encountered.
5403 If no (*MARK) with a matching name is found, normal "bumpalong" of one
5404 character happens (the (*SKIP) is ignored).
5405
5406 (*THEN) or (*THEN:NAME)
5407
5408 This verb causes a skip to the next alternation if the rest of the pat-
5409 tern does not match. That is, it cancels pending backtracking, but only
5410 within the current alternation. Its name comes from the observation
5411 that it can be used for a pattern-based if-then-else block:
5412
5413 ( COND1 (*THEN) FOO | COND2 (*THEN) BAR | COND3 (*THEN) BAZ ) ...
5414
5415 If the COND1 pattern matches, FOO is tried (and possibly further items
5416 after the end of the group if FOO succeeds); on failure the matcher
5417 skips to the second alternative and tries COND2, without backtracking
5418 into COND1. The behaviour of (*THEN:NAME) is exactly the same as
5419 (*MARK:NAME)(*THEN) if the overall match fails. If (*THEN) is not
5420 directly inside an alternation, it acts like (*PRUNE).
5421
5422
5423 SEE ALSO
5424
5425 pcreapi(3), pcrecallout(3), pcrematching(3), pcresyntax(3), pcre(3).
5426
5427
5428 AUTHOR
5429
5430 Philip Hazel
5431 University Computing Service
5432 Cambridge CB2 3QH, England.
5433
5434
5435 REVISION
5436
5437 Last updated: 05 May 2010
5438 Copyright (c) 1997-2010 University of Cambridge.
5439 ------------------------------------------------------------------------------
5440
5441
5442 PCRESYNTAX(3) PCRESYNTAX(3)
5443
5444
5445 NAME
5446 PCRE - Perl-compatible regular expressions
5447
5448
5449 PCRE REGULAR EXPRESSION SYNTAX SUMMARY
5450
5451 The full syntax and semantics of the regular expressions that are sup-
5452 ported by PCRE are described in the pcrepattern documentation. This
5453 document contains just a quick-reference summary of the syntax.
5454
5455
5456 QUOTING
5457
5458 \x where x is non-alphanumeric is a literal x
5459 \Q...\E treat enclosed characters as literal
5460
5461
5462 CHARACTERS
5463
5464 \a alarm, that is, the BEL character (hex 07)
5465 \cx "control-x", where x is any character
5466 \e escape (hex 1B)
5467 \f formfeed (hex 0C)
5468 \n newline (hex 0A)
5469 \r carriage return (hex 0D)
5470 \t tab (hex 09)
5471 \ddd character with octal code ddd, or backreference
5472 \xhh character with hex code hh
5473 \x{hhh..} character with hex code hhh..
5474
5475
5476 CHARACTER TYPES
5477
5478 . any character except newline;
5479 in dotall mode, any character whatsoever
5480 \C one byte, even in UTF-8 mode (best avoided)
5481 \d a decimal digit
5482 \D a character that is not a decimal digit
5483 \h a horizontal whitespace character
5484 \H a character that is not a horizontal whitespace character
5485 \N a character that is not a newline
5486 \p{xx} a character with the xx property
5487 \P{xx} a character without the xx property
5488 \R a newline sequence
5489 \s a whitespace character
5490 \S a character that is not a whitespace character
5491 \v a vertical whitespace character
5492 \V a character that is not a vertical whitespace character
5493 \w a "word" character
5494 \W a "non-word" character
5495 \X an extended Unicode sequence
5496
5497 In PCRE, \d, \D, \s, \S, \w, and \W recognize only ASCII characters.
5498
5499
5500 GENERAL CATEGORY PROPERTIES FOR \p and \P
5501
5502 C Other
5503 Cc Control
5504 Cf Format
5505 Cn Unassigned
5506 Co Private use
5507 Cs Surrogate
5508
5509 L Letter
5510 Ll Lower case letter
5511 Lm Modifier letter
5512 Lo Other letter
5513 Lt Title case letter
5514 Lu Upper case letter
5515 L& Ll, Lu, or Lt
5516
5517 M Mark
5518 Mc Spacing mark
5519 Me Enclosing mark
5520 Mn Non-spacing mark
5521
5522 N Number
5523 Nd Decimal number
5524 Nl Letter number
5525 No Other number
5526
5527 P Punctuation
5528 Pc Connector punctuation
5529 Pd Dash punctuation
5530 Pe Close punctuation
5531 Pf Final punctuation
5532 Pi Initial punctuation
5533 Po Other punctuation
5534 Ps Open punctuation
5535
5536 S Symbol
5537 Sc Currency symbol
5538 Sk Modifier symbol
5539 Sm Mathematical symbol
5540 So Other symbol
5541
5542 Z Separator
5543 Zl Line separator
5544 Zp Paragraph separator
5545 Zs Space separator
5546
5547
5548 PCRE SPECIAL CATEGORY PROPERTIES FOR \p and \P
5549
5550 Xan Alphanumeric: union of properties L and N
5551 Xps POSIX space: property Z or tab, NL, VT, FF, CR
5552 Xsp Perl space: property Z or tab, NL, FF, CR
5553 Xwd Perl word: property Xan or underscore
5554
5555
5556 SCRIPT NAMES FOR \p AND \P
5557
5558 Arabic, Armenian, Avestan, Balinese, Bamum, Bengali, Bopomofo, Braille,
5559 Buginese, Buhid, Canadian_Aboriginal, Carian, Cham, Cherokee, Common,
5560 Coptic, Cuneiform, Cypriot, Cyrillic, Deseret, Devanagari, Egyp-
5561 tian_Hieroglyphs, Ethiopic, Georgian, Glagolitic, Gothic, Greek,
5562 Gujarati, Gurmukhi, Han, Hangul, Hanunoo, Hebrew, Hiragana, Impe-
5563 rial_Aramaic, Inherited, Inscriptional_Pahlavi, Inscriptional_Parthian,
5564 Javanese, Kaithi, Kannada, Katakana, Kayah_Li, Kharoshthi, Khmer, Lao,
5565 Latin, Lepcha, Limbu, Linear_B, Lisu, Lycian, Lydian, Malayalam,
5566 Meetei_Mayek, Mongolian, Myanmar, New_Tai_Lue, Nko, Ogham, Old_Italic,
5567 Old_Persian, Old_South_Arabian, Old_Turkic, Ol_Chiki, Oriya, Osmanya,
5568 Phags_Pa, Phoenician, Rejang, Runic, Samaritan, Saurashtra, Shavian,
5569 Sinhala, Sundanese, Syloti_Nagri, Syriac, Tagalog, Tagbanwa, Tai_Le,
5570 Tai_Tham, Tai_Viet, Tamil, Telugu, Thaana, Thai, Tibetan, Tifinagh,
5571 Ugaritic, Vai, Yi.
5572
5573
5574 CHARACTER CLASSES
5575
5576 [...] positive character class
5577 [^...] negative character class
5578 [x-y] range (can be used for hex characters)
5579 [[:xxx:]] positive POSIX named set
5580 [[:^xxx:]] negative POSIX named set
5581
5582 alnum alphanumeric
5583 alpha alphabetic
5584 ascii 0-127
5585 blank space or tab
5586 cntrl control character
5587 digit decimal digit
5588 graph printing, excluding space
5589 lower lower case letter
5590 print printing, including space
5591 punct printing, excluding alphanumeric
5592 space whitespace
5593 upper upper case letter
5594 word same as \w
5595 xdigit hexadecimal digit
5596
5597 In PCRE, POSIX character set names recognize only ASCII characters. You
5598 can use \Q...\E inside a character class.
5599
5600
5601 QUANTIFIERS
5602
5603 ? 0 or 1, greedy
5604 ?+ 0 or 1, possessive
5605 ?? 0 or 1, lazy
5606 * 0 or more, greedy
5607 *+ 0 or more, possessive
5608 *? 0 or more, lazy
5609 + 1 or more, greedy
5610 ++ 1 or more, possessive
5611 +? 1 or more, lazy
5612 {n} exactly n
5613 {n,m} at least n, no more than m, greedy
5614 {n,m}+ at least n, no more than m, possessive
5615 {n,m}? at least n, no more than m, lazy
5616 {n,} n or more, greedy
5617 {n,}+ n or more, possessive
5618 {n,}? n or more, lazy
5619
5620
5621 ANCHORS AND SIMPLE ASSERTIONS
5622
5623 \b word boundary (only ASCII letters recognized)
5624 \B not a word boundary
5625 ^ start of subject
5626 also after internal newline in multiline mode
5627 \A start of subject
5628 $ end of subject
5629 also before newline at end of subject
5630 also before internal newline in multiline mode
5631 \Z end of subject
5632 also before newline at end of subject
5633 \z end of subject
5634 \G first matching position in subject
5635
5636
5637 MATCH POINT RESET
5638
5639 \K reset start of match
5640
5641
5642 ALTERNATION
5643
5644 expr|expr|expr...
5645
5646
5647 CAPTURING
5648
5649 (...) capturing group
5650 (?<name>...) named capturing group (Perl)
5651 (?'name'...) named capturing group (Perl)
5652 (?P<name>...) named capturing group (Python)
5653 (?:...) non-capturing group
5654 (?|...) non-capturing group; reset group numbers for
5655 capturing groups in each alternative
5656
5657
5658 ATOMIC GROUPS
5659
5660 (?>...) atomic, non-capturing group
5661
5662
5663 COMMENT
5664
5665 (?#....) comment (not nestable)
5666
5667
5668 OPTION SETTING
5669
5670 (?i) caseless
5671 (?J) allow duplicate names
5672 (?m) multiline
5673 (?s) single line (dotall)
5674 (?U) default ungreedy (lazy)
5675 (?x) extended (ignore white space)
5676 (?-...) unset option(s)
5677
5678 The following is recognized only at the start of a pattern or after one
5679 of the newline-setting options with similar syntax:
5680
5681 (*UTF8) set UTF-8 mode
5682
5683
5684 LOOKAHEAD AND LOOKBEHIND ASSERTIONS
5685
5686 (?=...) positive look ahead
5687 (?!...) negative look ahead
5688 (?<=...) positive look behind
5689 (?<!...) negative look behind
5690
5691 Each top-level branch of a look behind must be of a fixed length.
5692
5693
5694 BACKREFERENCES
5695
5696 \n reference by number (can be ambiguous)
5697 \gn reference by number
5698 \g{n} reference by number
5699 \g{-n} relative reference by number
5700 \k<name> reference by name (Perl)
5701 \k'name' reference by name (Perl)
5702 \g{name} reference by name (Perl)
5703 \k{name} reference by name (.NET)
5704 (?P=name) reference by name (Python)
5705
5706
5707 SUBROUTINE REFERENCES (POSSIBLY RECURSIVE)
5708
5709 (?R) recurse whole pattern
5710 (?n) call subpattern by absolute number
5711 (?+n) call subpattern by relative number
5712 (?-n) call subpattern by relative number
5713 (?&name) call subpattern by name (Perl)
5714 (?P>name) call subpattern by name (Python)
5715 \g<name> call subpattern by name (Oniguruma)
5716 \g'name' call subpattern by name (Oniguruma)
5717 \g<n> call subpattern by absolute number (Oniguruma)
5718 \g'n' call subpattern by absolute number (Oniguruma)
5719 \g<+n> call subpattern by relative number (PCRE extension)
5720 \g'+n' call subpattern by relative number (PCRE extension)
5721 \g<-n> call subpattern by relative number (PCRE extension)
5722 \g'-n' call subpattern by relative number (PCRE extension)
5723
5724
5725 CONDITIONAL PATTERNS
5726
5727 (?(condition)yes-pattern)
5728 (?(condition)yes-pattern|no-pattern)
5729
5730 (?(n)... absolute reference condition
5731 (?(+n)... relative reference condition
5732 (?(-n)... relative reference condition
5733 (?(<name>)... named reference condition (Perl)
5734 (?('name')... named reference condition (Perl)
5735 (?(name)... named reference condition (PCRE)
5736 (?(R)... overall recursion condition
5737 (?(Rn)... specific group recursion condition
5738 (?(R&name)... specific recursion condition
5739 (?(DEFINE)... define subpattern for reference
5740 (?(assert)... assertion condition
5741
5742
5743 BACKTRACKING CONTROL
5744
5745 The following act immediately they are reached:
5746
5747 (*ACCEPT) force successful match
5748 (*FAIL) force backtrack; synonym (*F)
5749
5750 The following act only when a subsequent match failure causes a back-
5751 track to reach them. They all force a match failure, but they differ in
5752 what happens afterwards. Those that advance the start-of-match point do
5753 so only if the pattern is not anchored.
5754
5755 (*COMMIT) overall failure, no advance of starting point
5756 (*PRUNE) advance to next starting character
5757 (*SKIP) advance start to current matching position
5758 (*THEN) local failure, backtrack to next alternation
5759
5760
5761 NEWLINE CONVENTIONS
5762
5763 These are recognized only at the very start of the pattern or after a
5764 (*BSR_...) or (*UTF8) option.
5765
5766 (*CR) carriage return only
5767 (*LF) linefeed only
5768 (*CRLF) carriage return followed by linefeed
5769 (*ANYCRLF) all three of the above
5770 (*ANY) any Unicode newline sequence
5771
5772
5773 WHAT \R MATCHES
5774
5775 These are recognized only at the very start of the pattern or after a
5776 (*...) option that sets the newline convention or UTF-8 mode.
5777
5778 (*BSR_ANYCRLF) CR, LF, or CRLF
5779 (*BSR_UNICODE) any Unicode newline sequence
5780
5781
5782 CALLOUTS
5783
5784 (?C) callout
5785 (?Cn) callout with data n
5786
5787
5788 SEE ALSO
5789
5790 pcrepattern(3), pcreapi(3), pcrecallout(3), pcrematching(3), pcre(3).
5791
5792
5793 AUTHOR
5794
5795 Philip Hazel
5796 University Computing Service
5797 Cambridge CB2 3QH, England.
5798
5799
5800 REVISION
5801
5802 Last updated: 05 May 2010
5803 Copyright (c) 1997-2010 University of Cambridge.
5804 ------------------------------------------------------------------------------
5805
5806
5807 PCREPARTIAL(3) PCREPARTIAL(3)
5808
5809
5810 NAME
5811 PCRE - Perl-compatible regular expressions
5812
5813
5814 PARTIAL MATCHING IN PCRE
5815
5816 In normal use of PCRE, if the subject string that is passed to
5817 pcre_exec() or pcre_dfa_exec() matches as far as it goes, but is too
5818 short to match the entire pattern, PCRE_ERROR_NOMATCH is returned.
5819 There are circumstances where it might be helpful to distinguish this
5820 case from other cases in which there is no match.
5821
5822 Consider, for example, an application where a human is required to type
5823 in data for a field with specific formatting requirements. An example
5824 might be a date in the form ddmmmyy, defined by this pattern:
5825
5826 ^\d?\d(jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)\d\d$
5827
5828 If the application sees the user's keystrokes one by one, and can check
5829 that what has been typed so far is potentially valid, it is able to
5830 raise an error as soon as a mistake is made, by beeping and not
5831 reflecting the character that has been typed, for example. This immedi-
5832 ate feedback is likely to be a better user interface than a check that
5833 is delayed until the entire string has been entered. Partial matching
5834 can also sometimes be useful when the subject string is very long and
5835 is not all available at once.
5836
5837 PCRE supports partial matching by means of the PCRE_PARTIAL_SOFT and
5838 PCRE_PARTIAL_HARD options, which can be set when calling pcre_exec() or
5839 pcre_dfa_exec(). For backwards compatibility, PCRE_PARTIAL is a synonym
5840 for PCRE_PARTIAL_SOFT. The essential difference between the two options
5841 is whether or not a partial match is preferred to an alternative com-
5842 plete match, though the details differ between the two matching func-
5843 tions. If both options are set, PCRE_PARTIAL_HARD takes precedence.
5844
5845 Setting a partial matching option disables two of PCRE's optimizations.
5846 PCRE remembers the last literal byte in a pattern, and abandons match-
5847 ing immediately if such a byte is not present in the subject string.
5848 This optimization cannot be used for a subject string that might match
5849 only partially. If the pattern was studied, PCRE knows the minimum
5850 length of a matching string, and does not bother to run the matching
5851 function on shorter strings. This optimization is also disabled for
5852 partial matching.
5853
5854
5855 PARTIAL MATCHING USING pcre_exec()
5856
5857 A partial match occurs during a call to pcre_exec() whenever the end of
5858 the subject string is reached successfully, but matching cannot con-
5859 tinue because more characters are needed. However, at least one charac-
5860 ter must have been matched. (In other words, a partial match can never
5861 be an empty string.)
5862
5863 If PCRE_PARTIAL_SOFT is set, the partial match is remembered, but
5864 matching continues as normal, and other alternatives in the pattern are
5865 tried. If no complete match can be found, pcre_exec() returns
5866 PCRE_ERROR_PARTIAL instead of PCRE_ERROR_NOMATCH. If there are at least
5867 two slots in the offsets vector, the first of them is set to the offset
5868 of the earliest character that was inspected when the partial match was
5869 found. For convenience, the second offset points to the end of the
5870 string so that a substring can easily be identified.
5871
5872 For the majority of patterns, the first offset identifies the start of
5873 the partially matched string. However, for patterns that contain look-
5874 behind assertions, or \K, or begin with \b or \B, earlier characters
5875 have been inspected while carrying out the match. For example:
5876
5877 /(?<=abc)123/
5878
5879 This pattern matches "123", but only if it is preceded by "abc". If the
5880 subject string is "xyzabc12", the offsets after a partial match are for
5881 the substring "abc12", because all these characters are needed if
5882 another match is tried with extra characters added.
5883
5884 If there is more than one partial match, the first one that was found
5885 provides the data that is returned. Consider this pattern:
5886
5887 /123\w+X|dogY/
5888
5889 If this is matched against the subject string "abc123dog", both alter-
5890 natives fail to match, but the end of the subject is reached during
5891 matching, so PCRE_ERROR_PARTIAL is returned instead of
5892 PCRE_ERROR_NOMATCH. The offsets are set to 3 and 9, identifying
5893 "123dog" as the first partial match that was found. (In this example,
5894 there are two partial matches, because "dog" on its own partially
5895 matches the second alternative.)
5896
5897 If PCRE_PARTIAL_HARD is set for pcre_exec(), it returns PCRE_ERROR_PAR-
5898 TIAL as soon as a partial match is found, without continuing to search
5899 for possible complete matches. The difference between the two options
5900 can be illustrated by a pattern such as:
5901
5902 /dog(sbody)?/
5903
5904 This matches either "dog" or "dogsbody", greedily (that is, it prefers
5905 the longer string if possible). If it is matched against the string
5906 "dog" with PCRE_PARTIAL_SOFT, it yields a complete match for "dog".
5907 However, if PCRE_PARTIAL_HARD is set, the result is PCRE_ERROR_PARTIAL.
5908 On the other hand, if the pattern is made ungreedy the result is dif-
5909 ferent:
5910
5911 /dog(sbody)??/
5912
5913 In this case the result is always a complete match because pcre_exec()
5914 finds that first, and it never continues after finding a match. It
5915 might be easier to follow this explanation by thinking of the two pat-
5916 terns like this:
5917
5918 /dog(sbody)?/ is the same as /dogsbody|dog/
5919 /dog(sbody)??/ is the same as /dog|dogsbody/
5920
5921 The second pattern will never match "dogsbody" when pcre_exec() is
5922 used, because it will always find the shorter match first.
5923
5924
5925 PARTIAL MATCHING USING pcre_dfa_exec()
5926
5927 The pcre_dfa_exec() function moves along the subject string character
5928 by character, without backtracking, searching for all possible matches
5929 simultaneously. If the end of the subject is reached before the end of
5930 the pattern, there is the possibility of a partial match, again pro-
5931 vided that at least one character has matched.
5932
5933 When PCRE_PARTIAL_SOFT is set, PCRE_ERROR_PARTIAL is returned only if
5934 there have been no complete matches. Otherwise, the complete matches
5935 are returned. However, if PCRE_PARTIAL_HARD is set, a partial match
5936 takes precedence over any complete matches. The portion of the string
5937 that was inspected when the longest partial match was found is set as
5938 the first matching string, provided there are at least two slots in the
5939 offsets vector.
5940
5941 Because pcre_dfa_exec() always searches for all possible matches, and
5942 there is no difference between greedy and ungreedy repetition, its be-
5943 haviour is different from pcre_exec when PCRE_PARTIAL_HARD is set. Con-
5944 sider the string "dog" matched against the ungreedy pattern shown
5945 above:
5946
5947 /dog(sbody)??/
5948
5949 Whereas pcre_exec() stops as soon as it finds the complete match for
5950 "dog", pcre_dfa_exec() also finds the partial match for "dogsbody", and
5951 so returns that when PCRE_PARTIAL_HARD is set.
5952
5953
5954 PARTIAL MATCHING AND WORD BOUNDARIES
5955
5956 If a pattern ends with one of sequences \b or \B, which test for word
5957 boundaries, partial matching with PCRE_PARTIAL_SOFT can give counter-
5958 intuitive results. Consider this pattern:
5959
5960 /\bcat\b/
5961
5962 This matches "cat", provided there is a word boundary at either end. If
5963 the subject string is "the cat", the comparison of the final "t" with a
5964 following character cannot take place, so a partial match is found.
5965 However, pcre_exec() carries on with normal matching, which matches \b
5966 at the end of the subject when the last character is a letter, thus
5967 finding a complete match. The result, therefore, is not PCRE_ERROR_PAR-
5968 TIAL. The same thing happens with pcre_dfa_exec(), because it also
5969 finds the complete match.
5970
5971 Using PCRE_PARTIAL_HARD in this case does yield PCRE_ERROR_PARTIAL,
5972 because then the partial match takes precedence.
5973
5974
5975 FORMERLY RESTRICTED PATTERNS
5976
5977 For releases of PCRE prior to 8.00, because of the way certain internal
5978 optimizations were implemented in the pcre_exec() function, the
5979 PCRE_PARTIAL option (predecessor of PCRE_PARTIAL_SOFT) could not be
5980 used with all patterns. From release 8.00 onwards, the restrictions no
5981 longer apply, and partial matching with pcre_exec() can be requested
5982 for any pattern.
5983
5984 Items that were formerly restricted were repeated single characters and
5985 repeated metasequences. If PCRE_PARTIAL was set for a pattern that did
5986 not conform to the restrictions, pcre_exec() returned the error code
5987 PCRE_ERROR_BADPARTIAL (-13). This error code is no longer in use. The
5988 PCRE_INFO_OKPARTIAL call to pcre_fullinfo() to find out if a compiled
5989 pattern can be used for partial matching now always returns 1.
5990
5991
5992 EXAMPLE OF PARTIAL MATCHING USING PCRETEST
5993
5994 If the escape sequence \P is present in a pcretest data line, the
5995 PCRE_PARTIAL_SOFT option is used for the match. Here is a run of
5996 pcretest that uses the date example quoted above:
5997
5998 re> /^\d?\d(jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)\d\d$/
5999 data> 25jun04\P
6000 0: 25jun04
6001 1: jun
6002 data> 25dec3\P
6003 Partial match: 23dec3
6004 data> 3ju\P
6005 Partial match: 3ju
6006 data> 3juj\P
6007 No match
6008 data> j\P
6009 No match
6010
6011 The first data string is matched completely, so pcretest shows the
6012