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

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

Parent Directory Parent Directory | Revision Log Revision Log


Revision 286 - (show annotations) (download)
Mon Dec 17 14:46:11 2007 UTC (6 years, 7 months ago) by ph10
File MIME type: text/plain
File size: 303098 byte(s)
Add .gz and .bz2 optional support to pcregrep.

1 -----------------------------------------------------------------------------
2 This file contains a concatenation of the PCRE man pages, converted to plain
3 text format for ease of searching with a text editor, or for use on systems
4 that do not have a man page processor. The small individual files that give
5 synopses of each function in the library have not been included. There are
6 separate text files for the pcregrep and pcretest commands.
7 -----------------------------------------------------------------------------
8
9
10 PCRE(3) PCRE(3)
11
12
13 NAME
14 PCRE - Perl-compatible regular expressions
15
16
17 INTRODUCTION
18
19 The PCRE library is a set of functions that implement regular expres-
20 sion pattern matching using the same syntax and semantics as Perl, with
21 just a few differences. (Certain features that appeared in Python and
22 PCRE before they appeared in Perl are also available using the Python
23 syntax.)
24
25 The current implementation of PCRE (release 7.x) corresponds approxi-
26 mately with Perl 5.10, including support for UTF-8 encoded strings and
27 Unicode general category properties. However, UTF-8 and Unicode support
28 has to be explicitly enabled; it is not the default. The Unicode tables
29 correspond to Unicode release 5.0.0.
30
31 In addition to the Perl-compatible matching function, PCRE contains an
32 alternative matching function that matches the same compiled patterns
33 in a different way. In certain circumstances, the alternative function
34 has some advantages. For a discussion of the two matching algorithms,
35 see the pcrematching page.
36
37 PCRE is written in C and released as a C library. A number of people
38 have written wrappers and interfaces of various kinds. In particular,
39 Google Inc. have provided a comprehensive C++ wrapper. This is now
40 included as part of the PCRE distribution. The pcrecpp page has details
41 of this interface. Other people's contributions can be found in the
42 Contrib directory at the primary FTP site, which is:
43
44 ftp://ftp.csx.cam.ac.uk/pub/software/programming/pcre
45
46 Details of exactly which Perl regular expression features are and are
47 not supported by PCRE are given in separate documents. See the pcrepat-
48 tern and pcrecompat pages. There is a syntax summary in the pcresyntax
49 page.
50
51 Some features of PCRE can be included, excluded, or changed when the
52 library is built. The pcre_config() function makes it possible for a
53 client to discover which features are available. The features them-
54 selves are described in the pcrebuild page. Documentation about build-
55 ing PCRE for various operating systems can be found in the README file
56 in the source distribution.
57
58 The library contains a number of undocumented internal functions and
59 data tables that are used by more than one of the exported external
60 functions, but which are not intended for use by external callers.
61 Their names all begin with "_pcre_", which hopefully will not provoke
62 any name clashes. In some environments, it is possible to control which
63 external symbols are exported when a shared library is built, and in
64 these cases the undocumented symbols are not exported.
65
66
67 USER DOCUMENTATION
68
69 The user documentation for PCRE comprises a number of different sec-
70 tions. In the "man" format, each of these is a separate "man page". In
71 the HTML format, each is a separate page, linked from the index page.
72 In the plain text format, all the sections are concatenated, for ease
73 of searching. The sections are as follows:
74
75 pcre this document
76 pcre-config show PCRE installation configuration information
77 pcreapi details of PCRE's native C API
78 pcrebuild options for building PCRE
79 pcrecallout details of the callout feature
80 pcrecompat discussion of Perl compatibility
81 pcrecpp details of the C++ wrapper
82 pcregrep description of the pcregrep command
83 pcrematching discussion of the two matching algorithms
84 pcrepartial details of the partial matching facility
85 pcrepattern syntax and semantics of supported
86 regular expressions
87 pcresyntax quick syntax reference
88 pcreperform discussion of performance issues
89 pcreposix the POSIX-compatible C API
90 pcreprecompile details of saving and re-using precompiled patterns
91 pcresample discussion of the sample program
92 pcrestack discussion of stack usage
93 pcretest description of the pcretest testing command
94
95 In addition, in the "man" and HTML formats, there is a short page for
96 each C library function, listing its arguments and results.
97
98
99 LIMITATIONS
100
101 There are some size limitations in PCRE but it is hoped that they will
102 never in practice be relevant.
103
104 The maximum length of a compiled pattern is 65539 (sic) bytes if PCRE
105 is compiled with the default internal linkage size of 2. If you want to
106 process regular expressions that are truly enormous, you can compile
107 PCRE with an internal linkage size of 3 or 4 (see the README file in
108 the source distribution and the pcrebuild documentation for details).
109 In these cases the limit is substantially larger. However, the speed
110 of execution is slower.
111
112 All values in repeating quantifiers must be less than 65536.
113
114 There is no limit to the number of parenthesized subpatterns, but there
115 can be no more than 65535 capturing subpatterns.
116
117 The maximum length of name for a named subpattern is 32 characters, and
118 the maximum number of named subpatterns is 10000.
119
120 The maximum length of a subject string is the largest positive number
121 that an integer variable can hold. However, when using the traditional
122 matching function, PCRE uses recursion to handle subpatterns and indef-
123 inite repetition. This means that the available stack space may limit
124 the size of a subject string that can be processed by certain patterns.
125 For a discussion of stack issues, see the pcrestack documentation.
126
127
128 UTF-8 AND UNICODE PROPERTY SUPPORT
129
130 From release 3.3, PCRE has had some support for character strings
131 encoded in the UTF-8 format. For release 4.0 this was greatly extended
132 to cover most common requirements, and in release 5.0 additional sup-
133 port for Unicode general category properties was added.
134
135 In order process UTF-8 strings, you must build PCRE to include UTF-8
136 support in the code, and, in addition, you must call pcre_compile()
137 with the PCRE_UTF8 option flag. When you do this, both the pattern and
138 any subject strings that are matched against it are treated as UTF-8
139 strings instead of just strings of bytes.
140
141 If you compile PCRE with UTF-8 support, but do not use it at run time,
142 the library will be a bit bigger, but the additional run time overhead
143 is limited to testing the PCRE_UTF8 flag occasionally, so should not be
144 very big.
145
146 If PCRE is built with Unicode character property support (which implies
147 UTF-8 support), the escape sequences \p{..}, \P{..}, and \X are sup-
148 ported. The available properties that can be tested are limited to the
149 general category properties such as Lu for an upper case letter or Nd
150 for a decimal number, the Unicode script names such as Arabic or Han,
151 and the derived properties Any and L&. A full list is given in the
152 pcrepattern documentation. Only the short names for properties are sup-
153 ported. For example, \p{L} matches a letter. Its Perl synonym, \p{Let-
154 ter}, is not supported. Furthermore, in Perl, many properties may
155 optionally be prefixed by "Is", for compatibility with Perl 5.6. PCRE
156 does not support this.
157
158 Validity of UTF-8 strings
159
160 When you set the PCRE_UTF8 flag, the strings passed as patterns and
161 subjects are (by default) checked for validity on entry to the relevant
162 functions. From release 7.3 of PCRE, the check is according the rules
163 of RFC 3629, which are themselves derived from the Unicode specifica-
164 tion. Earlier releases of PCRE followed the rules of RFC 2279, which
165 allows the full range of 31-bit values (0 to 0x7FFFFFFF). The current
166 check allows only values in the range U+0 to U+10FFFF, excluding U+D800
167 to U+DFFF.
168
169 The excluded code points are the "Low Surrogate Area" of Unicode, of
170 which the Unicode Standard says this: "The Low Surrogate Area does not
171 contain any character assignments, consequently no character code
172 charts or namelists are provided for this area. Surrogates are reserved
173 for use with UTF-16 and then must be used in pairs." The code points
174 that are encoded by UTF-16 pairs are available as independent code
175 points in the UTF-8 encoding. (In other words, the whole surrogate
176 thing is a fudge for UTF-16 which unfortunately messes up UTF-8.)
177
178 If an invalid UTF-8 string is passed to PCRE, an error return
179 (PCRE_ERROR_BADUTF8) is given. In some situations, you may already know
180 that your strings are valid, and therefore want to skip these checks in
181 order to improve performance. If you set the PCRE_NO_UTF8_CHECK flag at
182 compile time or at run time, PCRE assumes that the pattern or subject
183 it is given (respectively) contains only valid UTF-8 codes. In this
184 case, it does not diagnose an invalid UTF-8 string.
185
186 If you pass an invalid UTF-8 string when PCRE_NO_UTF8_CHECK is set,
187 what happens depends on why the string is invalid. If the string con-
188 forms to the "old" definition of UTF-8 (RFC 2279), it is processed as a
189 string of characters in the range 0 to 0x7FFFFFFF. In other words,
190 apart from the initial validity test, PCRE (when in UTF-8 mode) handles
191 strings according to the more liberal rules of RFC 2279. However, if
192 the string does not even conform to RFC 2279, the result is undefined.
193 Your program may crash.
194
195 If you want to process strings of values in the full range 0 to
196 0x7FFFFFFF, encoded in a UTF-8-like manner as per the old RFC, you can
197 set PCRE_NO_UTF8_CHECK to bypass the more restrictive test. However, in
198 this situation, you will have to apply your own validity check.
199
200 General comments about UTF-8 mode
201
202 1. An unbraced hexadecimal escape sequence (such as \xb3) matches a
203 two-byte UTF-8 character if the value is greater than 127.
204
205 2. Octal numbers up to \777 are recognized, and match two-byte UTF-8
206 characters for values greater than \177.
207
208 3. Repeat quantifiers apply to complete UTF-8 characters, not to indi-
209 vidual bytes, for example: \x{100}{3}.
210
211 4. The dot metacharacter matches one UTF-8 character instead of a sin-
212 gle byte.
213
214 5. The escape sequence \C can be used to match a single byte in UTF-8
215 mode, but its use can lead to some strange effects. This facility is
216 not available in the alternative matching function, pcre_dfa_exec().
217
218 6. The character escapes \b, \B, \d, \D, \s, \S, \w, and \W correctly
219 test characters of any code value, but the characters that PCRE recog-
220 nizes as digits, spaces, or word characters remain the same set as
221 before, all with values less than 256. This remains true even when PCRE
222 includes Unicode property support, because to do otherwise would slow
223 down PCRE in many common cases. If you really want to test for a wider
224 sense of, say, "digit", you must use Unicode property tests such as
225 \p{Nd}.
226
227 7. Similarly, characters that match the POSIX named character classes
228 are all low-valued characters.
229
230 8. However, the Perl 5.10 horizontal and vertical whitespace matching
231 escapes (\h, \H, \v, and \V) do match all the appropriate Unicode char-
232 acters.
233
234 9. Case-insensitive matching applies only to characters whose values
235 are less than 128, unless PCRE is built with Unicode property support.
236 Even when Unicode property support is available, PCRE still uses its
237 own character tables when checking the case of low-valued characters,
238 so as not to degrade performance. The Unicode property information is
239 used only for characters with higher values. Even when Unicode property
240 support is available, PCRE supports case-insensitive matching only when
241 there is a one-to-one mapping between a letter's cases. There are a
242 small number of many-to-one mappings in Unicode; these are not sup-
243 ported by PCRE.
244
245
246 AUTHOR
247
248 Philip Hazel
249 University Computing Service
250 Cambridge CB2 3QH, England.
251
252 Putting an actual email address here seems to have been a spam magnet,
253 so I've taken it away. If you want to email me, use my two initials,
254 followed by the two digits 10, at the domain cam.ac.uk.
255
256
257 REVISION
258
259 Last updated: 09 August 2007
260 Copyright (c) 1997-2007 University of Cambridge.
261 ------------------------------------------------------------------------------
262
263
264 PCREBUILD(3) PCREBUILD(3)
265
266
267 NAME
268 PCRE - Perl-compatible regular expressions
269
270
271 PCRE BUILD-TIME OPTIONS
272
273 This document describes the optional features of PCRE that can be
274 selected when the library is compiled. It assumes use of the configure
275 script, where the optional features are selected or deselected by pro-
276 viding options to configure before running the make command. However,
277 the same options can be selected in both Unix-like and non-Unix-like
278 environments using the GUI facility of CMakeSetup if you are using
279 CMake instead of configure to build PCRE.
280
281 The complete list of options for configure (which includes the standard
282 ones such as the selection of the installation directory) can be
283 obtained by running
284
285 ./configure --help
286
287 The following sections include descriptions of options whose names
288 begin with --enable or --disable. These settings specify changes to the
289 defaults for the configure command. Because of the way that configure
290 works, --enable and --disable always come in pairs, so the complemen-
291 tary option always exists as well, but as it specifies the default, it
292 is not described.
293
294
295 C++ SUPPORT
296
297 By default, the configure script will search for a C++ compiler and C++
298 header files. If it finds them, it automatically builds the C++ wrapper
299 library for PCRE. You can disable this by adding
300
301 --disable-cpp
302
303 to the configure command.
304
305
306 UTF-8 SUPPORT
307
308 To build PCRE with support for UTF-8 character strings, add
309
310 --enable-utf8
311
312 to the configure command. Of itself, this does not make PCRE treat
313 strings as UTF-8. As well as compiling PCRE with this option, you also
314 have have to set the PCRE_UTF8 option when you call the pcre_compile()
315 function.
316
317
318 UNICODE CHARACTER PROPERTY SUPPORT
319
320 UTF-8 support allows PCRE to process character values greater than 255
321 in the strings that it handles. On its own, however, it does not pro-
322 vide any facilities for accessing the properties of such characters. If
323 you want to be able to use the pattern escapes \P, \p, and \X, which
324 refer to Unicode character properties, you must add
325
326 --enable-unicode-properties
327
328 to the configure command. This implies UTF-8 support, even if you have
329 not explicitly requested it.
330
331 Including Unicode property support adds around 30K of tables to the
332 PCRE library. Only the general category properties such as Lu and Nd
333 are supported. Details are given in the pcrepattern documentation.
334
335
336 CODE VALUE OF NEWLINE
337
338 By default, PCRE interprets character 10 (linefeed, LF) as indicating
339 the end of a line. This is the normal newline character on Unix-like
340 systems. You can compile PCRE to use character 13 (carriage return, CR)
341 instead, by adding
342
343 --enable-newline-is-cr
344
345 to the configure command. There is also a --enable-newline-is-lf
346 option, which explicitly specifies linefeed as the newline character.
347
348 Alternatively, you can specify that line endings are to be indicated by
349 the two character sequence CRLF. If you want this, add
350
351 --enable-newline-is-crlf
352
353 to the configure command. There is a fourth option, specified by
354
355 --enable-newline-is-anycrlf
356
357 which causes PCRE to recognize any of the three sequences CR, LF, or
358 CRLF as indicating a line ending. Finally, a fifth option, specified by
359
360 --enable-newline-is-any
361
362 causes PCRE to recognize any Unicode newline sequence.
363
364 Whatever line ending convention is selected when PCRE is built can be
365 overridden when the library functions are called. At build time it is
366 conventional to use the standard for your operating system.
367
368
369 WHAT \R MATCHES
370
371 By default, the sequence \R in a pattern matches any Unicode newline
372 sequence, whatever has been selected as the line ending sequence. If
373 you specify
374
375 --enable-bsr-anycrlf
376
377 the default is changed so that \R matches only CR, LF, or CRLF. What-
378 ever is selected when PCRE is built can be overridden when the library
379 functions are called.
380
381
382 BUILDING SHARED AND STATIC LIBRARIES
383
384 The PCRE building process uses libtool to build both shared and static
385 Unix libraries by default. You can suppress one of these by adding one
386 of
387
388 --disable-shared
389 --disable-static
390
391 to the configure command, as required.
392
393
394 POSIX MALLOC USAGE
395
396 When PCRE is called through the POSIX interface (see the pcreposix doc-
397 umentation), additional working storage is required for holding the
398 pointers to capturing substrings, because PCRE requires three integers
399 per substring, whereas the POSIX interface provides only two. If the
400 number of expected substrings is small, the wrapper function uses space
401 on the stack, because this is faster than using malloc() for each call.
402 The default threshold above which the stack is no longer used is 10; it
403 can be changed by adding a setting such as
404
405 --with-posix-malloc-threshold=20
406
407 to the configure command.
408
409
410 HANDLING VERY LARGE PATTERNS
411
412 Within a compiled pattern, offset values are used to point from one
413 part to another (for example, from an opening parenthesis to an alter-
414 nation metacharacter). By default, two-byte values are used for these
415 offsets, leading to a maximum size for a compiled pattern of around
416 64K. This is sufficient to handle all but the most gigantic patterns.
417 Nevertheless, some people do want to process enormous patterns, so it
418 is possible to compile PCRE to use three-byte or four-byte offsets by
419 adding a setting such as
420
421 --with-link-size=3
422
423 to the configure command. The value given must be 2, 3, or 4. Using
424 longer offsets slows down the operation of PCRE because it has to load
425 additional bytes when handling them.
426
427
428 AVOIDING EXCESSIVE STACK USAGE
429
430 When matching with the pcre_exec() function, PCRE implements backtrack-
431 ing by making recursive calls to an internal function called match().
432 In environments where the size of the stack is limited, this can se-
433 verely limit PCRE's operation. (The Unix environment does not usually
434 suffer from this problem, but it may sometimes be necessary to increase
435 the maximum stack size. There is a discussion in the pcrestack docu-
436 mentation.) An alternative approach to recursion that uses memory from
437 the heap to remember data, instead of using recursive function calls,
438 has been implemented to work round the problem of limited stack size.
439 If you want to build a version of PCRE that works this way, add
440
441 --disable-stack-for-recursion
442
443 to the configure command. With this configuration, PCRE will use the
444 pcre_stack_malloc and pcre_stack_free variables to call memory manage-
445 ment functions. By default these point to malloc() and free(), but you
446 can replace the pointers so that your own functions are used.
447
448 Separate functions are provided rather than using pcre_malloc and
449 pcre_free because the usage is very predictable: the block sizes
450 requested are always the same, and the blocks are always freed in
451 reverse order. A calling program might be able to implement optimized
452 functions that perform better than malloc() and free(). PCRE runs
453 noticeably more slowly when built in this way. This option affects only
454 the pcre_exec() function; it is not relevant for the the
455 pcre_dfa_exec() function.
456
457
458 LIMITING PCRE RESOURCE USAGE
459
460 Internally, PCRE has a function called match(), which it calls repeat-
461 edly (sometimes recursively) when matching a pattern with the
462 pcre_exec() function. By controlling the maximum number of times this
463 function may be called during a single matching operation, a limit can
464 be placed on the resources used by a single call to pcre_exec(). The
465 limit can be changed at run time, as described in the pcreapi documen-
466 tation. The default is 10 million, but this can be changed by adding a
467 setting such as
468
469 --with-match-limit=500000
470
471 to the configure command. This setting has no effect on the
472 pcre_dfa_exec() matching function.
473
474 In some environments it is desirable to limit the depth of recursive
475 calls of match() more strictly than the total number of calls, in order
476 to restrict the maximum amount of stack (or heap, if --disable-stack-
477 for-recursion is specified) that is used. A second limit controls this;
478 it defaults to the value that is set for --with-match-limit, which
479 imposes no additional constraints. However, you can set a lower limit
480 by adding, for example,
481
482 --with-match-limit-recursion=10000
483
484 to the configure command. This value can also be overridden at run
485 time.
486
487
488 CREATING CHARACTER TABLES AT BUILD TIME
489
490 PCRE uses fixed tables for processing characters whose code values are
491 less than 256. By default, PCRE is built with a set of tables that are
492 distributed in the file pcre_chartables.c.dist. These tables are for
493 ASCII codes only. If you add
494
495 --enable-rebuild-chartables
496
497 to the configure command, the distributed tables are no longer used.
498 Instead, a program called dftables is compiled and run. This outputs
499 the source for new set of tables, created in the default locale of your
500 C runtime system. (This method of replacing the tables does not work if
501 you are cross compiling, because dftables is run on the local host. If
502 you need to create alternative tables when cross compiling, you will
503 have to do so "by hand".)
504
505
506 USING EBCDIC CODE
507
508 PCRE assumes by default that it will run in an environment where the
509 character code is ASCII (or Unicode, which is a superset of ASCII).
510 This is the case for most computer operating systems. PCRE can, how-
511 ever, be compiled to run in an EBCDIC environment by adding
512
513 --enable-ebcdic
514
515 to the configure command. This setting implies --enable-rebuild-charta-
516 bles. You should only use it if you know that you are in an EBCDIC
517 environment (for example, an IBM mainframe operating system).
518
519
520 PCREGREP OPTIONS FOR COMPRESSED FILE SUPPORT
521
522 By default, pcregrep reads all files as plain text. You can build it so
523 that it recognizes files whose names end in .gz or .bz2, and reads them
524 with libz or libbz2, respectively, by adding one or both of
525
526 --enable-pcregrep-libz
527 --enable-pcregrep-libbz2
528
529 to the configure command. These options naturally require that the rel-
530 evant libraries are installed on your system. Configuration will fail
531 if they are not.
532
533
534 SEE ALSO
535
536 pcreapi(3), pcre_config(3).
537
538
539 AUTHOR
540
541 Philip Hazel
542 University Computing Service
543 Cambridge CB2 3QH, England.
544
545
546 REVISION
547
548 Last updated: 17 December 2007
549 Copyright (c) 1997-2007 University of Cambridge.
550 ------------------------------------------------------------------------------
551
552
553 PCREMATCHING(3) PCREMATCHING(3)
554
555
556 NAME
557 PCRE - Perl-compatible regular expressions
558
559
560 PCRE MATCHING ALGORITHMS
561
562 This document describes the two different algorithms that are available
563 in PCRE for matching a compiled regular expression against a given sub-
564 ject string. The "standard" algorithm is the one provided by the
565 pcre_exec() function. This works in the same was as Perl's matching
566 function, and provides a Perl-compatible matching operation.
567
568 An alternative algorithm is provided by the pcre_dfa_exec() function;
569 this operates in a different way, and is not Perl-compatible. It has
570 advantages and disadvantages compared with the standard algorithm, and
571 these are described below.
572
573 When there is only one possible way in which a given subject string can
574 match a pattern, the two algorithms give the same answer. A difference
575 arises, however, when there are multiple possibilities. For example, if
576 the pattern
577
578 ^<.*>
579
580 is matched against the string
581
582 <something> <something else> <something further>
583
584 there are three possible answers. The standard algorithm finds only one
585 of them, whereas the alternative algorithm finds all three.
586
587
588 REGULAR EXPRESSIONS AS TREES
589
590 The set of strings that are matched by a regular expression can be rep-
591 resented as a tree structure. An unlimited repetition in the pattern
592 makes the tree of infinite size, but it is still a tree. Matching the
593 pattern to a given subject string (from a given starting point) can be
594 thought of as a search of the tree. There are two ways to search a
595 tree: depth-first and breadth-first, and these correspond to the two
596 matching algorithms provided by PCRE.
597
598
599 THE STANDARD MATCHING ALGORITHM
600
601 In the terminology of Jeffrey Friedl's book "Mastering Regular Expres-
602 sions", the standard algorithm is an "NFA algorithm". It conducts a
603 depth-first search of the pattern tree. That is, it proceeds along a
604 single path through the tree, checking that the subject matches what is
605 required. When there is a mismatch, the algorithm tries any alterna-
606 tives at the current point, and if they all fail, it backs up to the
607 previous branch point in the tree, and tries the next alternative
608 branch at that level. This often involves backing up (moving to the
609 left) in the subject string as well. The order in which repetition
610 branches are tried is controlled by the greedy or ungreedy nature of
611 the quantifier.
612
613 If a leaf node is reached, a matching string has been found, and at
614 that point the algorithm stops. Thus, if there is more than one possi-
615 ble match, this algorithm returns the first one that it finds. Whether
616 this is the shortest, the longest, or some intermediate length depends
617 on the way the greedy and ungreedy repetition quantifiers are specified
618 in the pattern.
619
620 Because it ends up with a single path through the tree, it is rela-
621 tively straightforward for this algorithm to keep track of the sub-
622 strings that are matched by portions of the pattern in parentheses.
623 This provides support for capturing parentheses and back references.
624
625
626 THE ALTERNATIVE MATCHING ALGORITHM
627
628 This algorithm conducts a breadth-first search of the tree. Starting
629 from the first matching point in the subject, it scans the subject
630 string from left to right, once, character by character, and as it does
631 this, it remembers all the paths through the tree that represent valid
632 matches. In Friedl's terminology, this is a kind of "DFA algorithm",
633 though it is not implemented as a traditional finite state machine (it
634 keeps multiple states active simultaneously).
635
636 The scan continues until either the end of the subject is reached, or
637 there are no more unterminated paths. At this point, terminated paths
638 represent the different matching possibilities (if there are none, the
639 match has failed). Thus, if there is more than one possible match,
640 this algorithm finds all of them, and in particular, it finds the long-
641 est. In PCRE, there is an option to stop the algorithm after the first
642 match (which is necessarily the shortest) has been found.
643
644 Note that all the matches that are found start at the same point in the
645 subject. If the pattern
646
647 cat(er(pillar)?)
648
649 is matched against the string "the caterpillar catchment", the result
650 will be the three strings "cat", "cater", and "caterpillar" that start
651 at the fourth character of the subject. The algorithm does not automat-
652 ically move on to find matches that start at later positions.
653
654 There are a number of features of PCRE regular expressions that are not
655 supported by the alternative matching algorithm. They are as follows:
656
657 1. Because the algorithm finds all possible matches, the greedy or
658 ungreedy nature of repetition quantifiers is not relevant. Greedy and
659 ungreedy quantifiers are treated in exactly the same way. However, pos-
660 sessive quantifiers can make a difference when what follows could also
661 match what is quantified, for example in a pattern like this:
662
663 ^a++\w!
664
665 This pattern matches "aaab!" but not "aaa!", which would be matched by
666 a non-possessive quantifier. Similarly, if an atomic group is present,
667 it is matched as if it were a standalone pattern at the current point,
668 and the longest match is then "locked in" for the rest of the overall
669 pattern.
670
671 2. When dealing with multiple paths through the tree simultaneously, it
672 is not straightforward to keep track of captured substrings for the
673 different matching possibilities, and PCRE's implementation of this
674 algorithm does not attempt to do this. This means that no captured sub-
675 strings are available.
676
677 3. Because no substrings are captured, back references within the pat-
678 tern are not supported, and cause errors if encountered.
679
680 4. For the same reason, conditional expressions that use a backrefer-
681 ence as the condition or test for a specific group recursion are not
682 supported.
683
684 5. Because many paths through the tree may be active, the \K escape
685 sequence, which resets the start of the match when encountered (but may
686 be on some paths and not on others), is not supported. It causes an
687 error if encountered.
688
689 6. Callouts are supported, but the value of the capture_top field is
690 always 1, and the value of the capture_last field is always -1.
691
692 7. The \C escape sequence, which (in the standard algorithm) matches a
693 single byte, even in UTF-8 mode, is not supported because the alterna-
694 tive algorithm moves through the subject string one character at a
695 time, for all active paths through the tree.
696
697 8. None of the backtracking control verbs such as (*PRUNE) are sup-
698 ported.
699
700
701 ADVANTAGES OF THE ALTERNATIVE ALGORITHM
702
703 Using the alternative matching algorithm provides the following advan-
704 tages:
705
706 1. All possible matches (at a single point in the subject) are automat-
707 ically found, and in particular, the longest match is found. To find
708 more than one match using the standard algorithm, you have to do kludgy
709 things with callouts.
710
711 2. There is much better support for partial matching. The restrictions
712 on the content of the pattern that apply when using the standard algo-
713 rithm for partial matching do not apply to the alternative algorithm.
714 For non-anchored patterns, the starting position of a partial match is
715 available.
716
717 3. Because the alternative algorithm scans the subject string just
718 once, and never needs to backtrack, it is possible to pass very long
719 subject strings to the matching function in several pieces, checking
720 for partial matching each time.
721
722
723 DISADVANTAGES OF THE ALTERNATIVE ALGORITHM
724
725 The alternative algorithm suffers from a number of disadvantages:
726
727 1. It is substantially slower than the standard algorithm. This is
728 partly because it has to search for all possible matches, but is also
729 because it is less susceptible to optimization.
730
731 2. Capturing parentheses and back references are not supported.
732
733 3. Although atomic groups are supported, their use does not provide the
734 performance advantage that it does for the standard algorithm.
735
736
737 AUTHOR
738
739 Philip Hazel
740 University Computing Service
741 Cambridge CB2 3QH, England.
742
743
744 REVISION
745
746 Last updated: 08 August 2007
747 Copyright (c) 1997-2007 University of Cambridge.
748 ------------------------------------------------------------------------------
749
750
751 PCREAPI(3) PCREAPI(3)
752
753
754 NAME
755 PCRE - Perl-compatible regular expressions
756
757
758 PCRE NATIVE API
759
760 #include <pcre.h>
761
762 pcre *pcre_compile(const char *pattern, int options,
763 const char **errptr, int *erroffset,
764 const unsigned char *tableptr);
765
766 pcre *pcre_compile2(const char *pattern, int options,
767 int *errorcodeptr,
768 const char **errptr, int *erroffset,
769 const unsigned char *tableptr);
770
771 pcre_extra *pcre_study(const pcre *code, int options,
772 const char **errptr);
773
774 int pcre_exec(const pcre *code, const pcre_extra *extra,
775 const char *subject, int length, int startoffset,
776 int options, int *ovector, int ovecsize);
777
778 int pcre_dfa_exec(const pcre *code, const pcre_extra *extra,
779 const char *subject, int length, int startoffset,
780 int options, int *ovector, int ovecsize,
781 int *workspace, int wscount);
782
783 int pcre_copy_named_substring(const pcre *code,
784 const char *subject, int *ovector,
785 int stringcount, const char *stringname,
786 char *buffer, int buffersize);
787
788 int pcre_copy_substring(const char *subject, int *ovector,
789 int stringcount, int stringnumber, char *buffer,
790 int buffersize);
791
792 int pcre_get_named_substring(const pcre *code,
793 const char *subject, int *ovector,
794 int stringcount, const char *stringname,
795 const char **stringptr);
796
797 int pcre_get_stringnumber(const pcre *code,
798 const char *name);
799
800 int pcre_get_stringtable_entries(const pcre *code,
801 const char *name, char **first, char **last);
802
803 int pcre_get_substring(const char *subject, int *ovector,
804 int stringcount, int stringnumber,
805 const char **stringptr);
806
807 int pcre_get_substring_list(const char *subject,
808 int *ovector, int stringcount, const char ***listptr);
809
810 void pcre_free_substring(const char *stringptr);
811
812 void pcre_free_substring_list(const char **stringptr);
813
814 const unsigned char *pcre_maketables(void);
815
816 int pcre_fullinfo(const pcre *code, const pcre_extra *extra,
817 int what, void *where);
818
819 int pcre_info(const pcre *code, int *optptr, int *firstcharptr);
820
821 int pcre_refcount(pcre *code, int adjust);
822
823 int pcre_config(int what, void *where);
824
825 char *pcre_version(void);
826
827 void *(*pcre_malloc)(size_t);
828
829 void (*pcre_free)(void *);
830
831 void *(*pcre_stack_malloc)(size_t);
832
833 void (*pcre_stack_free)(void *);
834
835 int (*pcre_callout)(pcre_callout_block *);
836
837
838 PCRE API OVERVIEW
839
840 PCRE has its own native API, which is described in this document. There
841 are also some wrapper functions that correspond to the POSIX regular
842 expression API. These are described in the pcreposix documentation.
843 Both of these APIs define a set of C function calls. A C++ wrapper is
844 distributed with PCRE. It is documented in the pcrecpp page.
845
846 The native API C function prototypes are defined in the header file
847 pcre.h, and on Unix systems the library itself is called libpcre. It
848 can normally be accessed by adding -lpcre to the command for linking an
849 application that uses PCRE. The header file defines the macros
850 PCRE_MAJOR and PCRE_MINOR to contain the major and minor release num-
851 bers for the library. Applications can use these to include support
852 for different releases of PCRE.
853
854 The functions pcre_compile(), pcre_compile2(), pcre_study(), and
855 pcre_exec() are used for compiling and matching regular expressions in
856 a Perl-compatible manner. A sample program that demonstrates the sim-
857 plest way of using them is provided in the file called pcredemo.c in
858 the source distribution. The pcresample documentation describes how to
859 run it.
860
861 A second matching function, pcre_dfa_exec(), which is not Perl-compati-
862 ble, is also provided. This uses a different algorithm for the match-
863 ing. The alternative algorithm finds all possible matches (at a given
864 point in the subject), and scans the subject just once. However, this
865 algorithm does not return captured substrings. A description of the two
866 matching algorithms and their advantages and disadvantages is given in
867 the pcrematching documentation.
868
869 In addition to the main compiling and matching functions, there are
870 convenience functions for extracting captured substrings from a subject
871 string that is matched by pcre_exec(). They are:
872
873 pcre_copy_substring()
874 pcre_copy_named_substring()
875 pcre_get_substring()
876 pcre_get_named_substring()
877 pcre_get_substring_list()
878 pcre_get_stringnumber()
879 pcre_get_stringtable_entries()
880
881 pcre_free_substring() and pcre_free_substring_list() are also provided,
882 to free the memory used for extracted strings.
883
884 The function pcre_maketables() is used to build a set of character
885 tables in the current locale for passing to pcre_compile(),
886 pcre_exec(), or pcre_dfa_exec(). This is an optional facility that is
887 provided for specialist use. Most commonly, no special tables are
888 passed, in which case internal tables that are generated when PCRE is
889 built are used.
890
891 The function pcre_fullinfo() is used to find out information about a
892 compiled pattern; pcre_info() is an obsolete version that returns only
893 some of the available information, but is retained for backwards com-
894 patibility. The function pcre_version() returns a pointer to a string
895 containing the version of PCRE and its date of release.
896
897 The function pcre_refcount() maintains a reference count in a data
898 block containing a compiled pattern. This is provided for the benefit
899 of object-oriented applications.
900
901 The global variables pcre_malloc and pcre_free initially contain the
902 entry points of the standard malloc() and free() functions, respec-
903 tively. PCRE calls the memory management functions via these variables,
904 so a calling program can replace them if it wishes to intercept the
905 calls. This should be done before calling any PCRE functions.
906
907 The global variables pcre_stack_malloc and pcre_stack_free are also
908 indirections to memory management functions. These special functions
909 are used only when PCRE is compiled to use the heap for remembering
910 data, instead of recursive function calls, when running the pcre_exec()
911 function. See the pcrebuild documentation for details of how to do
912 this. It is a non-standard way of building PCRE, for use in environ-
913 ments that have limited stacks. Because of the greater use of memory
914 management, it runs more slowly. Separate functions are provided so
915 that special-purpose external code can be used for this case. When
916 used, these functions are always called in a stack-like manner (last
917 obtained, first freed), and always for memory blocks of the same size.
918 There is a discussion about PCRE's stack usage in the pcrestack docu-
919 mentation.
920
921 The global variable pcre_callout initially contains NULL. It can be set
922 by the caller to a "callout" function, which PCRE will then call at
923 specified points during a matching operation. Details are given in the
924 pcrecallout documentation.
925
926
927 NEWLINES
928
929 PCRE supports five different conventions for indicating line breaks in
930 strings: a single CR (carriage return) character, a single LF (line-
931 feed) character, the two-character sequence CRLF, any of the three pre-
932 ceding, or any Unicode newline sequence. The Unicode newline sequences
933 are the three just mentioned, plus the single characters VT (vertical
934 tab, U+000B), FF (formfeed, U+000C), NEL (next line, U+0085), LS (line
935 separator, U+2028), and PS (paragraph separator, U+2029).
936
937 Each of the first three conventions is used by at least one operating
938 system as its standard newline sequence. When PCRE is built, a default
939 can be specified. The default default is LF, which is the Unix stan-
940 dard. When PCRE is run, the default can be overridden, either when a
941 pattern is compiled, or when it is matched.
942
943 At compile time, the newline convention can be specified by the options
944 argument of pcre_compile(), or it can be specified by special text at
945 the start of the pattern itself; this overrides any other settings. See
946 the pcrepattern page for details of the special character sequences.
947
948 In the PCRE documentation the word "newline" is used to mean "the char-
949 acter or pair of characters that indicate a line break". The choice of
950 newline convention affects the handling of the dot, circumflex, and
951 dollar metacharacters, the handling of #-comments in /x mode, and, when
952 CRLF is a recognized line ending sequence, the match position advance-
953 ment for a non-anchored pattern. There is more detail about this in the
954 section on pcre_exec() options below.
955
956 The choice of newline convention does not affect the interpretation of
957 the \n or \r escape sequences, nor does it affect what \R matches,
958 which is controlled in a similar way, but by separate options.
959
960
961 MULTITHREADING
962
963 The PCRE functions can be used in multi-threading applications, with
964 the proviso that the memory management functions pointed to by
965 pcre_malloc, pcre_free, pcre_stack_malloc, and pcre_stack_free, and the
966 callout function pointed to by pcre_callout, are shared by all threads.
967
968 The compiled form of a regular expression is not altered during match-
969 ing, so the same compiled pattern can safely be used by several threads
970 at once.
971
972
973 SAVING PRECOMPILED PATTERNS FOR LATER USE
974
975 The compiled form of a regular expression can be saved and re-used at a
976 later time, possibly by a different program, and even on a host other
977 than the one on which it was compiled. Details are given in the
978 pcreprecompile documentation. However, compiling a regular expression
979 with one version of PCRE for use with a different version is not guar-
980 anteed to work and may cause crashes.
981
982
983 CHECKING BUILD-TIME OPTIONS
984
985 int pcre_config(int what, void *where);
986
987 The function pcre_config() makes it possible for a PCRE client to dis-
988 cover which optional features have been compiled into the PCRE library.
989 The pcrebuild documentation has more details about these optional fea-
990 tures.
991
992 The first argument for pcre_config() is an integer, specifying which
993 information is required; the second argument is a pointer to a variable
994 into which the information is placed. The following information is
995 available:
996
997 PCRE_CONFIG_UTF8
998
999 The output is an integer that is set to one if UTF-8 support is avail-
1000 able; otherwise it is set to zero.
1001
1002 PCRE_CONFIG_UNICODE_PROPERTIES
1003
1004 The output is an integer that is set to one if support for Unicode
1005 character properties is available; otherwise it is set to zero.
1006
1007 PCRE_CONFIG_NEWLINE
1008
1009 The output is an integer whose value specifies the default character
1010 sequence that is recognized as meaning "newline". The four values that
1011 are supported are: 10 for LF, 13 for CR, 3338 for CRLF, -2 for ANYCRLF,
1012 and -1 for ANY. The default should normally be the standard sequence
1013 for your operating system.
1014
1015 PCRE_CONFIG_BSR
1016
1017 The output is an integer whose value indicates what character sequences
1018 the \R escape sequence matches by default. A value of 0 means that \R
1019 matches any Unicode line ending sequence; a value of 1 means that \R
1020 matches only CR, LF, or CRLF. The default can be overridden when a pat-
1021 tern is compiled or matched.
1022
1023 PCRE_CONFIG_LINK_SIZE
1024
1025 The output is an integer that contains the number of bytes used for
1026 internal linkage in compiled regular expressions. The value is 2, 3, or
1027 4. Larger values allow larger regular expressions to be compiled, at
1028 the expense of slower matching. The default value of 2 is sufficient
1029 for all but the most massive patterns, since it allows the compiled
1030 pattern to be up to 64K in size.
1031
1032 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
1033
1034 The output is an integer that contains the threshold above which the
1035 POSIX interface uses malloc() for output vectors. Further details are
1036 given in the pcreposix documentation.
1037
1038 PCRE_CONFIG_MATCH_LIMIT
1039
1040 The output is an integer that gives the default limit for the number of
1041 internal matching function calls in a pcre_exec() execution. Further
1042 details are given with pcre_exec() below.
1043
1044 PCRE_CONFIG_MATCH_LIMIT_RECURSION
1045
1046 The output is an integer that gives the default limit for the depth of
1047 recursion when calling the internal matching function in a pcre_exec()
1048 execution. Further details are given with pcre_exec() below.
1049
1050 PCRE_CONFIG_STACKRECURSE
1051
1052 The output is an integer that is set to one if internal recursion when
1053 running pcre_exec() is implemented by recursive function calls that use
1054 the stack to remember their state. This is the usual way that PCRE is
1055 compiled. The output is zero if PCRE was compiled to use blocks of data
1056 on the heap instead of recursive function calls. In this case,
1057 pcre_stack_malloc and pcre_stack_free are called to manage memory
1058 blocks on the heap, thus avoiding the use of the stack.
1059
1060
1061 COMPILING A PATTERN
1062
1063 pcre *pcre_compile(const char *pattern, int options,
1064 const char **errptr, int *erroffset,
1065 const unsigned char *tableptr);
1066
1067 pcre *pcre_compile2(const char *pattern, int options,
1068 int *errorcodeptr,
1069 const char **errptr, int *erroffset,
1070 const unsigned char *tableptr);
1071
1072 Either of the functions pcre_compile() or pcre_compile2() can be called
1073 to compile a pattern into an internal form. The only difference between
1074 the two interfaces is that pcre_compile2() has an additional argument,
1075 errorcodeptr, via which a numerical error code can be returned.
1076
1077 The pattern is a C string terminated by a binary zero, and is passed in
1078 the pattern argument. A pointer to a single block of memory that is
1079 obtained via pcre_malloc is returned. This contains the compiled code
1080 and related data. The pcre type is defined for the returned block; this
1081 is a typedef for a structure whose contents are not externally defined.
1082 It is up to the caller to free the memory (via pcre_free) when it is no
1083 longer required.
1084
1085 Although the compiled code of a PCRE regex is relocatable, that is, it
1086 does not depend on memory location, the complete pcre data block is not
1087 fully relocatable, because it may contain a copy of the tableptr argu-
1088 ment, which is an address (see below).
1089
1090 The options argument contains various bit settings that affect the com-
1091 pilation. It should be zero if no options are required. The available
1092 options are described below. Some of them, in particular, those that
1093 are compatible with Perl, can also be set and unset from within the
1094 pattern (see the detailed description in the pcrepattern documenta-
1095 tion). For these options, the contents of the options argument speci-
1096 fies their initial settings at the start of compilation and execution.
1097 The PCRE_ANCHORED and PCRE_NEWLINE_xxx options can be set at the time
1098 of matching as well as at compile time.
1099
1100 If errptr is NULL, pcre_compile() returns NULL immediately. Otherwise,
1101 if compilation of a pattern fails, pcre_compile() returns NULL, and
1102 sets the variable pointed to by errptr to point to a textual error mes-
1103 sage. This is a static string that is part of the library. You must not
1104 try to free it. The offset from the start of the pattern to the charac-
1105 ter where the error was discovered is placed in the variable pointed to
1106 by erroffset, which must not be NULL. If it is, an immediate error is
1107 given.
1108
1109 If pcre_compile2() is used instead of pcre_compile(), and the error-
1110 codeptr argument is not NULL, a non-zero error code number is returned
1111 via this argument in the event of an error. This is in addition to the
1112 textual error message. Error codes and messages are listed below.
1113
1114 If the final argument, tableptr, is NULL, PCRE uses a default set of
1115 character tables that are built when PCRE is compiled, using the
1116 default C locale. Otherwise, tableptr must be an address that is the
1117 result of a call to pcre_maketables(). This value is stored with the
1118 compiled pattern, and used again by pcre_exec(), unless another table
1119 pointer is passed to it. For more discussion, see the section on locale
1120 support below.
1121
1122 This code fragment shows a typical straightforward call to pcre_com-
1123 pile():
1124
1125 pcre *re;
1126 const char *error;
1127 int erroffset;
1128 re = pcre_compile(
1129 "^A.*Z", /* the pattern */
1130 0, /* default options */
1131 &error, /* for error message */
1132 &erroffset, /* for error offset */
1133 NULL); /* use default character tables */
1134
1135 The following names for option bits are defined in the pcre.h header
1136 file:
1137
1138 PCRE_ANCHORED
1139
1140 If this bit is set, the pattern is forced to be "anchored", that is, it
1141 is constrained to match only at the first matching point in the string
1142 that is being searched (the "subject string"). This effect can also be
1143 achieved by appropriate constructs in the pattern itself, which is the
1144 only way to do it in Perl.
1145
1146 PCRE_AUTO_CALLOUT
1147
1148 If this bit is set, pcre_compile() automatically inserts callout items,
1149 all with number 255, before each pattern item. For discussion of the
1150 callout facility, see the pcrecallout documentation.
1151
1152 PCRE_BSR_ANYCRLF
1153 PCRE_BSR_UNICODE
1154
1155 These options (which are mutually exclusive) control what the \R escape
1156 sequence matches. The choice is either to match only CR, LF, or CRLF,
1157 or to match any Unicode newline sequence. The default is specified when
1158 PCRE is built. It can be overridden from within the pattern, or by set-
1159 ting an option when a compiled pattern is matched.
1160
1161 PCRE_CASELESS
1162
1163 If this bit is set, letters in the pattern match both upper and lower
1164 case letters. It is equivalent to Perl's /i option, and it can be
1165 changed within a pattern by a (?i) option setting. In UTF-8 mode, PCRE
1166 always understands the concept of case for characters whose values are
1167 less than 128, so caseless matching is always possible. For characters
1168 with higher values, the concept of case is supported if PCRE is com-
1169 piled with Unicode property support, but not otherwise. If you want to
1170 use caseless matching for characters 128 and above, you must ensure
1171 that PCRE is compiled with Unicode property support as well as with
1172 UTF-8 support.
1173
1174 PCRE_DOLLAR_ENDONLY
1175
1176 If this bit is set, a dollar metacharacter in the pattern matches only
1177 at the end of the subject string. Without this option, a dollar also
1178 matches immediately before a newline at the end of the string (but not
1179 before any other newlines). The PCRE_DOLLAR_ENDONLY option is ignored
1180 if PCRE_MULTILINE is set. There is no equivalent to this option in
1181 Perl, and no way to set it within a pattern.
1182
1183 PCRE_DOTALL
1184
1185 If this bit is set, a dot metacharater in the pattern matches all char-
1186 acters, including those that indicate newline. Without it, a dot does
1187 not match when the current position is at a newline. This option is
1188 equivalent to Perl's /s option, and it can be changed within a pattern
1189 by a (?s) option setting. A negative class such as [^a] always matches
1190 newline characters, independent of the setting of this option.
1191
1192 PCRE_DUPNAMES
1193
1194 If this bit is set, names used to identify capturing subpatterns need
1195 not be unique. This can be helpful for certain types of pattern when it
1196 is known that only one instance of the named subpattern can ever be
1197 matched. There are more details of named subpatterns below; see also
1198 the pcrepattern documentation.
1199
1200 PCRE_EXTENDED
1201
1202 If this bit is set, whitespace data characters in the pattern are
1203 totally ignored except when escaped or inside a character class. White-
1204 space does not include the VT character (code 11). In addition, charac-
1205 ters between an unescaped # outside a character class and the next new-
1206 line, inclusive, are also ignored. This is equivalent to Perl's /x
1207 option, and it can be changed within a pattern by a (?x) option set-
1208 ting.
1209
1210 This option makes it possible to include comments inside complicated
1211 patterns. Note, however, that this applies only to data characters.
1212 Whitespace characters may never appear within special character
1213 sequences in a pattern, for example within the sequence (?( which
1214 introduces a conditional subpattern.
1215
1216 PCRE_EXTRA
1217
1218 This option was invented in order to turn on additional functionality
1219 of PCRE that is incompatible with Perl, but it is currently of very
1220 little use. When set, any backslash in a pattern that is followed by a
1221 letter that has no special meaning causes an error, thus reserving
1222 these combinations for future expansion. By default, as in Perl, a
1223 backslash followed by a letter with no special meaning is treated as a
1224 literal. (Perl can, however, be persuaded to give a warning for this.)
1225 There are at present no other features controlled by this option. It
1226 can also be set by a (?X) option setting within a pattern.
1227
1228 PCRE_FIRSTLINE
1229
1230 If this option is set, an unanchored pattern is required to match
1231 before or at the first newline in the subject string, though the
1232 matched text may continue over the newline.
1233
1234 PCRE_MULTILINE
1235
1236 By default, PCRE treats the subject string as consisting of a single
1237 line of characters (even if it actually contains newlines). The "start
1238 of line" metacharacter (^) matches only at the start of the string,
1239 while the "end of line" metacharacter ($) matches only at the end of
1240 the string, or before a terminating newline (unless PCRE_DOLLAR_ENDONLY
1241 is set). This is the same as Perl.
1242
1243 When PCRE_MULTILINE it is set, the "start of line" and "end of line"
1244 constructs match immediately following or immediately before internal
1245 newlines in the subject string, respectively, as well as at the very
1246 start and end. This is equivalent to Perl's /m option, and it can be
1247 changed within a pattern by a (?m) option setting. If there are no new-
1248 lines in a subject string, or no occurrences of ^ or $ in a pattern,
1249 setting PCRE_MULTILINE has no effect.
1250
1251 PCRE_NEWLINE_CR
1252 PCRE_NEWLINE_LF
1253 PCRE_NEWLINE_CRLF
1254 PCRE_NEWLINE_ANYCRLF
1255 PCRE_NEWLINE_ANY
1256
1257 These options override the default newline definition that was chosen
1258 when PCRE was built. Setting the first or the second specifies that a
1259 newline is indicated by a single character (CR or LF, respectively).
1260 Setting PCRE_NEWLINE_CRLF specifies that a newline is indicated by the
1261 two-character CRLF sequence. Setting PCRE_NEWLINE_ANYCRLF specifies
1262 that any of the three preceding sequences should be recognized. Setting
1263 PCRE_NEWLINE_ANY specifies that any Unicode newline sequence should be
1264 recognized. The Unicode newline sequences are the three just mentioned,
1265 plus the single characters VT (vertical tab, U+000B), FF (formfeed,
1266 U+000C), NEL (next line, U+0085), LS (line separator, U+2028), and PS
1267 (paragraph separator, U+2029). The last two are recognized only in
1268 UTF-8 mode.
1269
1270 The newline setting in the options word uses three bits that are
1271 treated as a number, giving eight possibilities. Currently only six are
1272 used (default plus the five values above). This means that if you set
1273 more than one newline option, the combination may or may not be sensi-
1274 ble. For example, PCRE_NEWLINE_CR with PCRE_NEWLINE_LF is equivalent to
1275 PCRE_NEWLINE_CRLF, but other combinations may yield unused numbers and
1276 cause an error.
1277
1278 The only time that a line break is specially recognized when compiling
1279 a pattern is if PCRE_EXTENDED is set, and an unescaped # outside a
1280 character class is encountered. This indicates a comment that lasts
1281 until after the next line break sequence. In other circumstances, line
1282 break sequences are treated as literal data, except that in
1283 PCRE_EXTENDED mode, both CR and LF are treated as whitespace characters
1284 and are therefore ignored.
1285
1286 The newline option that is set at compile time becomes the default that
1287 is used for pcre_exec() and pcre_dfa_exec(), but it can be overridden.
1288
1289 PCRE_NO_AUTO_CAPTURE
1290
1291 If this option is set, it disables the use of numbered capturing paren-
1292 theses in the pattern. Any opening parenthesis that is not followed by
1293 ? behaves as if it were followed by ?: but named parentheses can still
1294 be used for capturing (and they acquire numbers in the usual way).
1295 There is no equivalent of this option in Perl.
1296
1297 PCRE_UNGREEDY
1298
1299 This option inverts the "greediness" of the quantifiers so that they
1300 are not greedy by default, but become greedy if followed by "?". It is
1301 not compatible with Perl. It can also be set by a (?U) option setting
1302 within the pattern.
1303
1304 PCRE_UTF8
1305
1306 This option causes PCRE to regard both the pattern and the subject as
1307 strings of UTF-8 characters instead of single-byte character strings.
1308 However, it is available only when PCRE is built to include UTF-8 sup-
1309 port. If not, the use of this option provokes an error. Details of how
1310 this option changes the behaviour of PCRE are given in the section on
1311 UTF-8 support in the main pcre page.
1312
1313 PCRE_NO_UTF8_CHECK
1314
1315 When PCRE_UTF8 is set, the validity of the pattern as a UTF-8 string is
1316 automatically checked. There is a discussion about the validity of
1317 UTF-8 strings in the main pcre page. If an invalid UTF-8 sequence of
1318 bytes is found, pcre_compile() returns an error. If you already know
1319 that your pattern is valid, and you want to skip this check for perfor-
1320 mance reasons, you can set the PCRE_NO_UTF8_CHECK option. When it is
1321 set, the effect of passing an invalid UTF-8 string as a pattern is
1322 undefined. It may cause your program to crash. Note that this option
1323 can also be passed to pcre_exec() and pcre_dfa_exec(), to suppress the
1324 UTF-8 validity checking of subject strings.
1325
1326
1327 COMPILATION ERROR CODES
1328
1329 The following table lists the error codes than may be returned by
1330 pcre_compile2(), along with the error messages that may be returned by
1331 both compiling functions. As PCRE has developed, some error codes have
1332 fallen out of use. To avoid confusion, they have not been re-used.
1333
1334 0 no error
1335 1 \ at end of pattern
1336 2 \c at end of pattern
1337 3 unrecognized character follows \
1338 4 numbers out of order in {} quantifier
1339 5 number too big in {} quantifier
1340 6 missing terminating ] for character class
1341 7 invalid escape sequence in character class
1342 8 range out of order in character class
1343 9 nothing to repeat
1344 10 [this code is not in use]
1345 11 internal error: unexpected repeat
1346 12 unrecognized character after (?
1347 13 POSIX named classes are supported only within a class
1348 14 missing )
1349 15 reference to non-existent subpattern
1350 16 erroffset passed as NULL
1351 17 unknown option bit(s) set
1352 18 missing ) after comment
1353 19 [this code is not in use]
1354 20 regular expression too large
1355 21 failed to get memory
1356 22 unmatched parentheses
1357 23 internal error: code overflow
1358 24 unrecognized character after (?<
1359 25 lookbehind assertion is not fixed length
1360 26 malformed number or name after (?(
1361 27 conditional group contains more than two branches
1362 28 assertion expected after (?(
1363 29 (?R or (?[+-]digits must be followed by )
1364 30 unknown POSIX class name
1365 31 POSIX collating elements are not supported
1366 32 this version of PCRE is not compiled with PCRE_UTF8 support
1367 33 [this code is not in use]
1368 34 character value in \x{...} sequence is too large
1369 35 invalid condition (?(0)
1370 36 \C not allowed in lookbehind assertion
1371 37 PCRE does not support \L, \l, \N, \U, or \u
1372 38 number after (?C is > 255
1373 39 closing ) for (?C expected
1374 40 recursive call could loop indefinitely
1375 41 unrecognized character after (?P
1376 42 syntax error in subpattern name (missing terminator)
1377 43 two named subpatterns have the same name
1378 44 invalid UTF-8 string
1379 45 support for \P, \p, and \X has not been compiled
1380 46 malformed \P or \p sequence
1381 47 unknown property name after \P or \p
1382 48 subpattern name is too long (maximum 32 characters)
1383 49 too many named subpatterns (maximum 10,000)
1384 50 [this code is not in use]
1385 51 octal value is greater than \377 (not in UTF-8 mode)
1386 52 internal error: overran compiling workspace
1387 53 internal error: previously-checked referenced subpattern not
1388 found
1389 54 DEFINE group contains more than one branch
1390 55 repeating a DEFINE group is not allowed
1391 56 inconsistent NEWLINE options
1392 57 \g is not followed by a braced name or an optionally braced
1393 non-zero number
1394 58 (?+ or (?- or (?(+ or (?(- must be followed by a non-zero number
1395
1396
1397 STUDYING A PATTERN
1398
1399 pcre_extra *pcre_study(const pcre *code, int options
1400 const char **errptr);
1401
1402 If a compiled pattern is going to be used several times, it is worth
1403 spending more time analyzing it in order to speed up the time taken for
1404 matching. The function pcre_study() takes a pointer to a compiled pat-
1405 tern as its first argument. If studying the pattern produces additional
1406 information that will help speed up matching, pcre_study() returns a
1407 pointer to a pcre_extra block, in which the study_data field points to
1408 the results of the study.
1409
1410 The returned value from pcre_study() can be passed directly to
1411 pcre_exec(). However, a pcre_extra block also contains other fields
1412 that can be set by the caller before the block is passed; these are
1413 described below in the section on matching a pattern.
1414
1415 If studying the pattern does not produce any additional information
1416 pcre_study() returns NULL. In that circumstance, if the calling program
1417 wants to pass any of the other fields to pcre_exec(), it must set up
1418 its own pcre_extra block.
1419
1420 The second argument of pcre_study() contains option bits. At present,
1421 no options are defined, and this argument should always be zero.
1422
1423 The third argument for pcre_study() is a pointer for an error message.
1424 If studying succeeds (even if no data is returned), the variable it
1425 points to is set to NULL. Otherwise it is set to point to a textual
1426 error message. This is a static string that is part of the library. You
1427 must not try to free it. You should test the error pointer for NULL
1428 after calling pcre_study(), to be sure that it has run successfully.
1429
1430 This is a typical call to pcre_study():
1431
1432 pcre_extra *pe;
1433 pe = pcre_study(
1434 re, /* result of pcre_compile() */
1435 0, /* no options exist */
1436 &error); /* set to NULL or points to a message */
1437
1438 At present, studying a pattern is useful only for non-anchored patterns
1439 that do not have a single fixed starting character. A bitmap of possi-
1440 ble starting bytes is created.
1441
1442
1443 LOCALE SUPPORT
1444
1445 PCRE handles caseless matching, and determines whether characters are
1446 letters, digits, or whatever, by reference to a set of tables, indexed
1447 by character value. When running in UTF-8 mode, this applies only to
1448 characters with codes less than 128. Higher-valued codes never match
1449 escapes such as \w or \d, but can be tested with \p if PCRE is built
1450 with Unicode character property support. The use of locales with Uni-
1451 code is discouraged. If you are handling characters with codes greater
1452 than 128, you should either use UTF-8 and Unicode, or use locales, but
1453 not try to mix the two.
1454
1455 PCRE contains an internal set of tables that are used when the final
1456 argument of pcre_compile() is NULL. These are sufficient for many
1457 applications. Normally, the internal tables recognize only ASCII char-
1458 acters. However, when PCRE is built, it is possible to cause the inter-
1459 nal tables to be rebuilt in the default "C" locale of the local system,
1460 which may cause them to be different.
1461
1462 The internal tables can always be overridden by tables supplied by the
1463 application that calls PCRE. These may be created in a different locale
1464 from the default. As more and more applications change to using Uni-
1465 code, the need for this locale support is expected to die away.
1466
1467 External tables are built by calling the pcre_maketables() function,
1468 which has no arguments, in the relevant locale. The result can then be
1469 passed to pcre_compile() or pcre_exec() as often as necessary. For
1470 example, to build and use tables that are appropriate for the French
1471 locale (where accented characters with values greater than 128 are
1472 treated as letters), the following code could be used:
1473
1474 setlocale(LC_CTYPE, "fr_FR");
1475 tables = pcre_maketables();
1476 re = pcre_compile(..., tables);
1477
1478 The locale name "fr_FR" is used on Linux and other Unix-like systems;
1479 if you are using Windows, the name for the French locale is "french".
1480
1481 When pcre_maketables() runs, the tables are built in memory that is
1482 obtained via pcre_malloc. It is the caller's responsibility to ensure
1483 that the memory containing the tables remains available for as long as
1484 it is needed.
1485
1486 The pointer that is passed to pcre_compile() is saved with the compiled
1487 pattern, and the same tables are used via this pointer by pcre_study()
1488 and normally also by pcre_exec(). Thus, by default, for any single pat-
1489 tern, compilation, studying and matching all happen in the same locale,
1490 but different patterns can be compiled in different locales.
1491
1492 It is possible to pass a table pointer or NULL (indicating the use of
1493 the internal tables) to pcre_exec(). Although not intended for this
1494 purpose, this facility could be used to match a pattern in a different
1495 locale from the one in which it was compiled. Passing table pointers at
1496 run time is discussed below in the section on matching a pattern.
1497
1498
1499 INFORMATION ABOUT A PATTERN
1500
1501 int pcre_fullinfo(const pcre *code, const pcre_extra *extra,
1502 int what, void *where);
1503
1504 The pcre_fullinfo() function returns information about a compiled pat-
1505 tern. It replaces the obsolete pcre_info() function, which is neverthe-
1506 less retained for backwards compability (and is documented below).
1507
1508 The first argument for pcre_fullinfo() is a pointer to the compiled
1509 pattern. The second argument is the result of pcre_study(), or NULL if
1510 the pattern was not studied. The third argument specifies which piece
1511 of information is required, and the fourth argument is a pointer to a
1512 variable to receive the data. The yield of the function is zero for
1513 success, or one of the following negative numbers:
1514
1515 PCRE_ERROR_NULL the argument code was NULL
1516 the argument where was NULL
1517 PCRE_ERROR_BADMAGIC the "magic number" was not found
1518 PCRE_ERROR_BADOPTION the value of what was invalid
1519
1520 The "magic number" is placed at the start of each compiled pattern as
1521 an simple check against passing an arbitrary memory pointer. Here is a
1522 typical call of pcre_fullinfo(), to obtain the length of the compiled
1523 pattern:
1524
1525 int rc;
1526 size_t length;
1527 rc = pcre_fullinfo(
1528 re, /* result of pcre_compile() */
1529 pe, /* result of pcre_study(), or NULL */
1530 PCRE_INFO_SIZE, /* what is required */
1531 &length); /* where to put the data */
1532
1533 The possible values for the third argument are defined in pcre.h, and
1534 are as follows:
1535
1536 PCRE_INFO_BACKREFMAX
1537
1538 Return the number of the highest back reference in the pattern. The
1539 fourth argument should point to an int variable. Zero is returned if
1540 there are no back references.
1541
1542 PCRE_INFO_CAPTURECOUNT
1543
1544 Return the number of capturing subpatterns in the pattern. The fourth
1545 argument should point to an int variable.
1546
1547 PCRE_INFO_DEFAULT_TABLES
1548
1549 Return a pointer to the internal default character tables within PCRE.
1550 The fourth argument should point to an unsigned char * variable. This
1551 information call is provided for internal use by the pcre_study() func-
1552 tion. External callers can cause PCRE to use its internal tables by
1553 passing a NULL table pointer.
1554
1555 PCRE_INFO_FIRSTBYTE
1556
1557 Return information about the first byte of any matched string, for a
1558 non-anchored pattern. The fourth argument should point to an int vari-
1559 able. (This option used to be called PCRE_INFO_FIRSTCHAR; the old name
1560 is still recognized for backwards compatibility.)
1561
1562 If there is a fixed first byte, for example, from a pattern such as
1563 (cat|cow|coyote), its value is returned. Otherwise, if either
1564
1565 (a) the pattern was compiled with the PCRE_MULTILINE option, and every
1566 branch starts with "^", or
1567
1568 (b) every branch of the pattern starts with ".*" and PCRE_DOTALL is not
1569 set (if it were set, the pattern would be anchored),
1570
1571 -1 is returned, indicating that the pattern matches only at the start
1572 of a subject string or after any newline within the string. Otherwise
1573 -2 is returned. For anchored patterns, -2 is returned.
1574
1575 PCRE_INFO_FIRSTTABLE
1576
1577 If the pattern was studied, and this resulted in the construction of a
1578 256-bit table indicating a fixed set of bytes for the first byte in any
1579 matching string, a pointer to the table is returned. Otherwise NULL is
1580 returned. The fourth argument should point to an unsigned char * vari-
1581 able.
1582
1583 PCRE_INFO_HASCRORLF
1584
1585 Return 1 if the pattern contains any explicit matches for CR or LF
1586 characters, otherwise 0. The fourth argument should point to an int
1587 variable. An explicit match is either a literal CR or LF character, or
1588 \r or \n.
1589
1590 PCRE_INFO_JCHANGED
1591
1592 Return 1 if the (?J) or (?-J) option setting is used in the pattern,
1593 otherwise 0. The fourth argument should point to an int variable. (?J)
1594 and (?-J) set and unset the local PCRE_DUPNAMES option, respectively.
1595
1596 PCRE_INFO_LASTLITERAL
1597
1598 Return the value of the rightmost literal byte that must exist in any
1599 matched string, other than at its start, if such a byte has been
1600 recorded. The fourth argument should point to an int variable. If there
1601 is no such byte, -1 is returned. For anchored patterns, a last literal
1602 byte is recorded only if it follows something of variable length. For
1603 example, for the pattern /^a\d+z\d+/ the returned value is "z", but for
1604 /^a\dz\d/ the returned value is -1.
1605
1606 PCRE_INFO_NAMECOUNT
1607 PCRE_INFO_NAMEENTRYSIZE
1608 PCRE_INFO_NAMETABLE
1609
1610 PCRE supports the use of named as well as numbered capturing parenthe-
1611 ses. The names are just an additional way of identifying the parenthe-
1612 ses, which still acquire numbers. Several convenience functions such as
1613 pcre_get_named_substring() are provided for extracting captured sub-
1614 strings by name. It is also possible to extract the data directly, by
1615 first converting the name to a number in order to access the correct
1616 pointers in the output vector (described with pcre_exec() below). To do
1617 the conversion, you need to use the name-to-number map, which is
1618 described by these three values.
1619
1620 The map consists of a number of fixed-size entries. PCRE_INFO_NAMECOUNT
1621 gives the number of entries, and PCRE_INFO_NAMEENTRYSIZE gives the size
1622 of each entry; both of these return an int value. The entry size
1623 depends on the length of the longest name. PCRE_INFO_NAMETABLE returns
1624 a pointer to the first entry of the table (a pointer to char). The
1625 first two bytes of each entry are the number of the capturing parenthe-
1626 sis, most significant byte first. The rest of the entry is the corre-
1627 sponding name, zero terminated. The names are in alphabetical order.
1628 When PCRE_DUPNAMES is set, duplicate names are in order of their paren-
1629 theses numbers. For example, consider the following pattern (assume
1630 PCRE_EXTENDED is set, so white space - including newlines - is
1631 ignored):
1632
1633 (?<date> (?<year>(\d\d)?\d\d) -
1634 (?<month>\d\d) - (?<day>\d\d) )
1635
1636 There are four named subpatterns, so the table has four entries, and
1637 each entry in the table is eight bytes long. The table is as follows,
1638 with non-printing bytes shows in hexadecimal, and undefined bytes shown
1639 as ??:
1640
1641 00 01 d a t e 00 ??
1642 00 05 d a y 00 ?? ??
1643 00 04 m o n t h 00
1644 00 02 y e a r 00 ??
1645
1646 When writing code to extract data from named subpatterns using the
1647 name-to-number map, remember that the length of the entries is likely
1648 to be different for each compiled pattern.
1649
1650 PCRE_INFO_OKPARTIAL
1651
1652 Return 1 if the pattern can be used for partial matching, otherwise 0.
1653 The fourth argument should point to an int variable. The pcrepartial
1654 documentation lists the restrictions that apply to patterns when par-
1655 tial matching is used.
1656
1657 PCRE_INFO_OPTIONS
1658
1659 Return a copy of the options with which the pattern was compiled. The
1660 fourth argument should point to an unsigned long int variable. These
1661 option bits are those specified in the call to pcre_compile(), modified
1662 by any top-level option settings at the start of the pattern itself. In
1663 other words, they are the options that will be in force when matching
1664 starts. For example, if the pattern /(?im)abc(?-i)d/ is compiled with
1665 the PCRE_EXTENDED option, the result is PCRE_CASELESS, PCRE_MULTILINE,
1666 and PCRE_EXTENDED.
1667
1668 A pattern is automatically anchored by PCRE if all of its top-level
1669 alternatives begin with one of the following:
1670
1671 ^ unless PCRE_MULTILINE is set
1672 \A always
1673 \G always
1674 .* if PCRE_DOTALL is set and there are no back
1675 references to the subpattern in which .* appears
1676
1677 For such patterns, the PCRE_ANCHORED bit is set in the options returned
1678 by pcre_fullinfo().
1679
1680 PCRE_INFO_SIZE
1681
1682 Return the size of the compiled pattern, that is, the value that was
1683 passed as the argument to pcre_malloc() when PCRE was getting memory in
1684 which to place the compiled data. The fourth argument should point to a
1685 size_t variable.
1686
1687 PCRE_INFO_STUDYSIZE
1688
1689 Return the size of the data block pointed to by the study_data field in
1690 a pcre_extra block. That is, it is the value that was passed to
1691 pcre_malloc() when PCRE was getting memory into which to place the data
1692 created by pcre_study(). The fourth argument should point to a size_t
1693 variable.
1694
1695
1696 OBSOLETE INFO FUNCTION
1697
1698 int pcre_info(const pcre *code, int *optptr, int *firstcharptr);
1699
1700 The pcre_info() function is now obsolete because its interface is too
1701 restrictive to return all the available data about a compiled pattern.
1702 New programs should use pcre_fullinfo() instead. The yield of
1703 pcre_info() is the number of capturing subpatterns, or one of the fol-
1704 lowing negative numbers:
1705
1706 PCRE_ERROR_NULL the argument code was NULL
1707 PCRE_ERROR_BADMAGIC the "magic number" was not found
1708
1709 If the optptr argument is not NULL, a copy of the options with which
1710 the pattern was compiled is placed in the integer it points to (see
1711 PCRE_INFO_OPTIONS above).
1712
1713 If the pattern is not anchored and the firstcharptr argument is not
1714 NULL, it is used to pass back information about the first character of
1715 any matched string (see PCRE_INFO_FIRSTBYTE above).
1716
1717
1718 REFERENCE COUNTS
1719
1720 int pcre_refcount(pcre *code, int adjust);
1721
1722 The pcre_refcount() function is used to maintain a reference count in
1723 the data block that contains a compiled pattern. It is provided for the
1724 benefit of applications that operate in an object-oriented manner,
1725 where different parts of the application may be using the same compiled
1726 pattern, but you want to free the block when they are all done.
1727
1728 When a pattern is compiled, the reference count field is initialized to
1729 zero. It is changed only by calling this function, whose action is to
1730 add the adjust value (which may be positive or negative) to it. The
1731 yield of the function is the new value. However, the value of the count
1732 is constrained to lie between 0 and 65535, inclusive. If the new value
1733 is outside these limits, it is forced to the appropriate limit value.
1734
1735 Except when it is zero, the reference count is not correctly preserved
1736 if a pattern is compiled on one host and then transferred to a host
1737 whose byte-order is different. (This seems a highly unlikely scenario.)
1738
1739
1740 MATCHING A PATTERN: THE TRADITIONAL FUNCTION
1741
1742 int pcre_exec(const pcre *code, const pcre_extra *extra,
1743 const char *subject, int length, int startoffset,
1744 int options, int *ovector, int ovecsize);
1745
1746 The function pcre_exec() is called to match a subject string against a
1747 compiled pattern, which is passed in the code argument. If the pattern
1748 has been studied, the result of the study should be passed in the extra
1749 argument. This function is the main matching facility of the library,
1750 and it operates in a Perl-like manner. For specialist use there is also
1751 an alternative matching function, which is described below in the sec-
1752 tion about the pcre_dfa_exec() function.
1753
1754 In most applications, the pattern will have been compiled (and option-
1755 ally studied) in the same process that calls pcre_exec(). However, it
1756 is possible to save compiled patterns and study data, and then use them
1757 later in different processes, possibly even on different hosts. For a
1758 discussion about this, see the pcreprecompile documentation.
1759
1760 Here is an example of a simple call to pcre_exec():
1761
1762 int rc;
1763 int ovector[30];
1764 rc = pcre_exec(
1765 re, /* result of pcre_compile() */
1766 NULL, /* we didn't study the pattern */
1767 "some string", /* the subject string */
1768 11, /* the length of the subject string */
1769 0, /* start at offset 0 in the subject */
1770 0, /* default options */
1771 ovector, /* vector of integers for substring information */
1772 30); /* number of elements (NOT size in bytes) */
1773
1774 Extra data for pcre_exec()
1775
1776 If the extra argument is not NULL, it must point to a pcre_extra data
1777 block. The pcre_study() function returns such a block (when it doesn't
1778 return NULL), but you can also create one for yourself, and pass addi-
1779 tional information in it. The pcre_extra block contains the following
1780 fields (not necessarily in this order):
1781
1782 unsigned long int flags;
1783 void *study_data;
1784 unsigned long int match_limit;
1785 unsigned long int match_limit_recursion;
1786 void *callout_data;
1787 const unsigned char *tables;
1788
1789 The flags field is a bitmap that specifies which of the other fields
1790 are set. The flag bits are:
1791
1792 PCRE_EXTRA_STUDY_DATA
1793 PCRE_EXTRA_MATCH_LIMIT
1794 PCRE_EXTRA_MATCH_LIMIT_RECURSION
1795 PCRE_EXTRA_CALLOUT_DATA
1796 PCRE_EXTRA_TABLES
1797
1798 Other flag bits should be set to zero. The study_data field is set in
1799 the pcre_extra block that is returned by pcre_study(), together with
1800 the appropriate flag bit. You should not set this yourself, but you may
1801 add to the block by setting the other fields and their corresponding
1802 flag bits.
1803
1804 The match_limit field provides a means of preventing PCRE from using up
1805 a vast amount of resources when running patterns that are not going to
1806 match, but which have a very large number of possibilities in their
1807 search trees. The classic example is the use of nested unlimited
1808 repeats.
1809
1810 Internally, PCRE uses a function called match() which it calls repeat-
1811 edly (sometimes recursively). The limit set by match_limit is imposed
1812 on the number of times this function is called during a match, which
1813 has the effect of limiting the amount of backtracking that can take
1814 place. For patterns that are not anchored, the count restarts from zero
1815 for each position in the subject string.
1816
1817 The default value for the limit can be set when PCRE is built; the
1818 default default is 10 million, which handles all but the most extreme
1819 cases. You can override the default by suppling pcre_exec() with a
1820 pcre_extra block in which match_limit is set, and
1821 PCRE_EXTRA_MATCH_LIMIT is set in the flags field. If the limit is
1822 exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
1823
1824 The match_limit_recursion field is similar to match_limit, but instead
1825 of limiting the total number of times that match() is called, it limits
1826 the depth of recursion. The recursion depth is a smaller number than
1827 the total number of calls, because not all calls to match() are recur-
1828 sive. This limit is of use only if it is set smaller than match_limit.
1829
1830 Limiting the recursion depth limits the amount of stack that can be
1831 used, or, when PCRE has been compiled to use memory on the heap instead
1832 of the stack, the amount of heap memory that can be used.
1833
1834 The default value for match_limit_recursion can be set when PCRE is
1835 built; the default default is the same value as the default for
1836 match_limit. You can override the default by suppling pcre_exec() with
1837 a pcre_extra block in which match_limit_recursion is set, and
1838 PCRE_EXTRA_MATCH_LIMIT_RECURSION is set in the flags field. If the
1839 limit is exceeded, pcre_exec() returns PCRE_ERROR_RECURSIONLIMIT.
1840
1841 The pcre_callout field is used in conjunction with the "callout" fea-
1842 ture, which is described in the pcrecallout documentation.
1843
1844 The tables field is used to pass a character tables pointer to
1845 pcre_exec(); this overrides the value that is stored with the compiled
1846 pattern. A non-NULL value is stored with the compiled pattern only if
1847 custom tables were supplied to pcre_compile() via its tableptr argu-
1848 ment. If NULL is passed to pcre_exec() using this mechanism, it forces
1849 PCRE's internal tables to be used. This facility is helpful when re-
1850 using patterns that have been saved after compiling with an external
1851 set of tables, because the external tables might be at a different
1852 address when pcre_exec() is called. See the pcreprecompile documenta-
1853 tion for a discussion of saving compiled patterns for later use.
1854
1855 Option bits for pcre_exec()
1856
1857 The unused bits of the options argument for pcre_exec() must be zero.
1858 The only bits that may be set are PCRE_ANCHORED, PCRE_NEWLINE_xxx,
1859 PCRE_NOTBOL, PCRE_NOTEOL, PCRE_NOTEMPTY, PCRE_NO_UTF8_CHECK and
1860 PCRE_PARTIAL.
1861
1862 PCRE_ANCHORED
1863
1864 The PCRE_ANCHORED option limits pcre_exec() to matching at the first
1865 matching position. If a pattern was compiled with PCRE_ANCHORED, or
1866 turned out to be anchored by virtue of its contents, it cannot be made
1867 unachored at matching time.
1868
1869 PCRE_BSR_ANYCRLF
1870 PCRE_BSR_UNICODE
1871
1872 These options (which are mutually exclusive) control what the \R escape
1873 sequence matches. The choice is either to match only CR, LF, or CRLF,
1874 or to match any Unicode newline sequence. These options override the
1875 choice that was made or defaulted when the pattern was compiled.
1876
1877 PCRE_NEWLINE_CR
1878 PCRE_NEWLINE_LF
1879 PCRE_NEWLINE_CRLF
1880 PCRE_NEWLINE_ANYCRLF
1881 PCRE_NEWLINE_ANY
1882
1883 These options override the newline definition that was chosen or
1884 defaulted when the pattern was compiled. For details, see the descrip-
1885 tion of pcre_compile() above. During matching, the newline choice
1886 affects the behaviour of the dot, circumflex, and dollar metacharac-
1887 ters. It may also alter the way the match position is advanced after a
1888 match failure for an unanchored pattern.
1889
1890 When PCRE_NEWLINE_CRLF, PCRE_NEWLINE_ANYCRLF, or PCRE_NEWLINE_ANY is
1891 set, and a match attempt for an unanchored pattern fails when the cur-
1892 rent position is at a CRLF sequence, and the pattern contains no
1893 explicit matches for CR or LF characters, the match position is
1894 advanced by two characters instead of one, in other words, to after the
1895 CRLF.
1896
1897 The above rule is a compromise that makes the most common cases work as
1898 expected. For example, if the pattern is .+A (and the PCRE_DOTALL
1899 option is not set), it does not match the string "\r\nA" because, after
1900 failing at the start, it skips both the CR and the LF before retrying.
1901 However, the pattern [\r\n]A does match that string, because it con-
1902 tains an explicit CR or LF reference, and so advances only by one char-
1903 acter after the first failure.
1904
1905 An explicit match for CR of LF is either a literal appearance of one of
1906 those characters, or one of the \r or \n escape sequences. Implicit
1907 matches such as [^X] do not count, nor does \s (which includes CR and
1908 LF in the characters that it matches).
1909
1910 Notwithstanding the above, anomalous effects may still occur when CRLF
1911 is a valid newline sequence and explicit \r or \n escapes appear in the
1912 pattern.
1913
1914 PCRE_NOTBOL
1915
1916 This option specifies that first character of the subject string is not
1917 the beginning of a line, so the circumflex metacharacter should not
1918 match before it. Setting this without PCRE_MULTILINE (at compile time)
1919 causes circumflex never to match. This option affects only the behav-
1920 iour of the circumflex metacharacter. It does not affect \A.
1921
1922 PCRE_NOTEOL
1923
1924 This option specifies that the end of the subject string is not the end
1925 of a line, so the dollar metacharacter should not match it nor (except
1926 in multiline mode) a newline immediately before it. Setting this with-
1927 out PCRE_MULTILINE (at compile time) causes dollar never to match. This
1928 option affects only the behaviour of the dollar metacharacter. It does
1929 not affect \Z or \z.
1930
1931 PCRE_NOTEMPTY
1932
1933 An empty string is not considered to be a valid match if this option is
1934 set. If there are alternatives in the pattern, they are tried. If all
1935 the alternatives match the empty string, the entire match fails. For
1936 example, if the pattern
1937
1938 a?b?
1939
1940 is applied to a string not beginning with "a" or "b", it matches the
1941 empty string at the start of the subject. With PCRE_NOTEMPTY set, this
1942 match is not valid, so PCRE searches further into the string for occur-
1943 rences of "a" or "b".
1944
1945 Perl has no direct equivalent of PCRE_NOTEMPTY, but it does make a spe-
1946 cial case of a pattern match of the empty string within its split()
1947 function, and when using the /g modifier. It is possible to emulate
1948 Perl's behaviour after matching a null string by first trying the match
1949 again at the same offset with PCRE_NOTEMPTY and PCRE_ANCHORED, and then
1950 if that fails by advancing the starting offset (see below) and trying
1951 an ordinary match again. There is some code that demonstrates how to do
1952 this in the pcredemo.c sample program.
1953
1954 PCRE_NO_UTF8_CHECK
1955
1956 When PCRE_UTF8 is set at compile time, the validity of the subject as a
1957 UTF-8 string is automatically checked when pcre_exec() is subsequently
1958 called. The value of startoffset is also checked to ensure that it
1959 points to the start of a UTF-8 character. There is a discussion about
1960 the validity of UTF-8 strings in the section on UTF-8 support in the
1961 main pcre page. If an invalid UTF-8 sequence of bytes is found,
1962 pcre_exec() returns the error PCRE_ERROR_BADUTF8. If startoffset con-
1963 tains an invalid value, PCRE_ERROR_BADUTF8_OFFSET is returned.
1964
1965 If you already know that your subject is valid, and you want to skip
1966 these checks for performance reasons, you can set the
1967 PCRE_NO_UTF8_CHECK option when calling pcre_exec(). You might want to
1968 do this for the second and subsequent calls to pcre_exec() if you are
1969 making repeated calls to find all the matches in a single subject
1970 string. However, you should be sure that the value of startoffset
1971 points to the start of a UTF-8 character. When PCRE_NO_UTF8_CHECK is
1972 set, the effect of passing an invalid UTF-8 string as a subject, or a
1973 value of startoffset that does not point to the start of a UTF-8 char-
1974 acter, is undefined. Your program may crash.
1975
1976 PCRE_PARTIAL
1977
1978 This option turns on the partial matching feature. If the subject
1979 string fails to match the pattern, but at some point during the match-
1980 ing process the end of the subject was reached (that is, the subject
1981 partially matches the pattern and the failure to match occurred only
1982 because there were not enough subject characters), pcre_exec() returns
1983 PCRE_ERROR_PARTIAL instead of PCRE_ERROR_NOMATCH. When PCRE_PARTIAL is
1984 used, there are restrictions on what may appear in the pattern. These
1985 are discussed in the pcrepartial documentation.
1986
1987 The string to be matched by pcre_exec()
1988
1989 The subject string is passed to pcre_exec() as a pointer in subject, a
1990 length in length, and a starting byte offset in startoffset. In UTF-8
1991 mode, the byte offset must point to the start of a UTF-8 character.
1992 Unlike the pattern string, the subject may contain binary zero bytes.
1993 When the starting offset is zero, the search for a match starts at the
1994 beginning of the subject, and this is by far the most common case.
1995
1996 A non-zero starting offset is useful when searching for another match
1997 in the same subject by calling pcre_exec() again after a previous suc-
1998 cess. Setting startoffset differs from just passing over a shortened
1999 string and setting PCRE_NOTBOL in the case of a pattern that begins
2000 with any kind of lookbehind. For example, consider the pattern
2001
2002 \Biss\B
2003
2004 which finds occurrences of "iss" in the middle of words. (\B matches
2005 only if the current position in the subject is not a word boundary.)
2006 When applied to the string "Mississipi" the first call to pcre_exec()
2007 finds the first occurrence. If pcre_exec() is called again with just
2008 the remainder of the subject, namely "issipi", it does not match,
2009 because \B is always false at the start of the subject, which is deemed
2010 to be a word boundary. However, if pcre_exec() is passed the entire
2011 string again, but with startoffset set to 4, it finds the second occur-
2012 rence of "iss" because it is able to look behind the starting point to
2013 discover that it is preceded by a letter.
2014
2015 If a non-zero starting offset is passed when the pattern is anchored,
2016 one attempt to match at the given offset is made. This can only succeed
2017 if the pattern does not require the match to be at the start of the
2018 subject.
2019
2020 How pcre_exec() returns captured substrings
2021
2022 In general, a pattern matches a certain portion of the subject, and in
2023 addition, further substrings from the subject may be picked out by
2024 parts of the pattern. Following the usage in Jeffrey Friedl's book,
2025 this is called "capturing" in what follows, and the phrase "capturing
2026 subpattern" is used for a fragment of a pattern that picks out a sub-
2027 string. PCRE supports several other kinds of parenthesized subpattern
2028 that do not cause substrings to be captured.
2029
2030 Captured substrings are returned to the caller via a vector of integer
2031 offsets whose address is passed in ovector. The number of elements in
2032 the vector is passed in ovecsize, which must be a non-negative number.
2033 Note: this argument is NOT the size of ovector in bytes.
2034
2035 The first two-thirds of the vector is used to pass back captured sub-
2036 strings, each substring using a pair of integers. The remaining third
2037 of the vector is used as workspace by pcre_exec() while matching cap-
2038 turing subpatterns, and is not available for passing back information.
2039 The length passed in ovecsize should always be a multiple of three. If
2040 it is not, it is rounded down.
2041
2042 When a match is successful, information about captured substrings is
2043 returned in pairs of integers, starting at the beginning of ovector,
2044 and continuing up to two-thirds of its length at the most. The first
2045 element of a pair is set to the offset of the first character in a sub-
2046 string, and the second is set to the offset of the first character
2047 after the end of a substring. The first pair, ovector[0] and ovec-
2048 tor[1], identify the portion of the subject string matched by the
2049 entire pattern. The next pair is used for the first capturing subpat-
2050 tern, and so on. The value returned by pcre_exec() is one more than the
2051 highest numbered pair that has been set. For example, if two substrings
2052 have been captured, the returned value is 3. If there are no capturing
2053 subpatterns, the return value from a successful match is 1, indicating
2054 that just the first pair of offsets has been set.
2055
2056 If a capturing subpattern is matched repeatedly, it is the last portion
2057 of the string that it matched that is returned.
2058
2059 If the vector is too small to hold all the captured substring offsets,
2060 it is used as far as possible (up to two-thirds of its length), and the
2061 function returns a value of zero. In particular, if the substring off-
2062 sets are not of interest, pcre_exec() may be called with ovector passed
2063 as NULL and ovecsize as zero. However, if the pattern contains back
2064 references and the ovector is not big enough to remember the related
2065 substrings, PCRE has to get additional memory for use during matching.
2066 Thus it is usually advisable to supply an ovector.
2067
2068 The pcre_info() function can be used to find out how many capturing
2069 subpatterns there are in a compiled pattern. The smallest size for
2070 ovector that will allow for n captured substrings, in addition to the
2071 offsets of the substring matched by the whole pattern, is (n+1)*3.
2072
2073 It is possible for capturing subpattern number n+1 to match some part
2074 of the subject when subpattern n has not been used at all. For example,
2075 if the string "abc" is matched against the pattern (a|(z))(bc) the
2076 return from the function is 4, and subpatterns 1 and 3 are matched, but
2077 2 is not. When this happens, both values in the offset pairs corre-
2078 sponding to unused subpatterns are set to -1.
2079
2080 Offset values that correspond to unused subpatterns at the end of the
2081 expression are also set to -1. For example, if the string "abc" is
2082 matched against the pattern (abc)(x(yz)?)? subpatterns 2 and 3 are not
2083 matched. The return from the function is 2, because the highest used
2084 capturing subpattern number is 1. However, you can refer to the offsets
2085 for the second and third capturing subpatterns if you wish (assuming
2086 the vector is large enough, of course).
2087
2088 Some convenience functions are provided for extracting the captured
2089 substrings as separate strings. These are described below.
2090
2091 Error return values from pcre_exec()
2092
2093 If pcre_exec() fails, it returns a negative number. The following are
2094 defined in the header file:
2095
2096 PCRE_ERROR_NOMATCH (-1)
2097
2098 The subject string did not match the pattern.
2099
2100 PCRE_ERROR_NULL (-2)
2101
2102 Either code or subject was passed as NULL, or ovector was NULL and
2103 ovecsize was not zero.
2104
2105 PCRE_ERROR_BADOPTION (-3)
2106
2107 An unrecognized bit was set in the options argument.
2108
2109 PCRE_ERROR_BADMAGIC (-4)
2110
2111 PCRE stores a 4-byte "magic number" at the start of the compiled code,
2112 to catch the case when it is passed a junk pointer and to detect when a
2113 pattern that was compiled in an environment of one endianness is run in
2114 an environment with the other endianness. This is the error that PCRE
2115 gives when the magic number is not present.
2116
2117 PCRE_ERROR_UNKNOWN_OPCODE (-5)
2118
2119 While running the pattern match, an unknown item was encountered in the
2120 compiled pattern. This error could be caused by a bug in PCRE or by
2121 overwriting of the compiled pattern.
2122
2123 PCRE_ERROR_NOMEMORY (-6)
2124
2125 If a pattern contains back references, but the ovector that is passed
2126 to pcre_exec() is not big enough to remember the referenced substrings,
2127 PCRE gets a block of memory at the start of matching to use for this
2128 purpose. If the call via pcre_malloc() fails, this error is given. The
2129 memory is automatically freed at the end of matching.
2130
2131 PCRE_ERROR_NOSUBSTRING (-7)
2132
2133 This error is used by the pcre_copy_substring(), pcre_get_substring(),
2134 and pcre_get_substring_list() functions (see below). It is never
2135 returned by pcre_exec().
2136
2137 PCRE_ERROR_MATCHLIMIT (-8)
2138
2139 The backtracking limit, as specified by the match_limit field in a
2140 pcre_extra structure (or defaulted) was reached. See the description
2141 above.
2142
2143 PCRE_ERROR_CALLOUT (-9)
2144
2145 This error is never generated by pcre_exec() itself. It is provided for
2146 use by callout functions that want to yield a distinctive error code.
2147 See the pcrecallout documentation for details.
2148
2149 PCRE_ERROR_BADUTF8 (-10)
2150
2151 A string that contains an invalid UTF-8 byte sequence was passed as a
2152 subject.
2153
2154 PCRE_ERROR_BADUTF8_OFFSET (-11)
2155
2156 The UTF-8 byte sequence that was passed as a subject was valid, but the
2157 value of startoffset did not point to the beginning of a UTF-8 charac-
2158 ter.
2159
2160 PCRE_ERROR_PARTIAL (-12)
2161
2162 The subject string did not match, but it did match partially. See the
2163 pcrepartial documentation for details of partial matching.
2164
2165 PCRE_ERROR_BADPARTIAL (-13)
2166
2167 The PCRE_PARTIAL option was used with a compiled pattern containing
2168 items that are not supported for partial matching. See the pcrepartial
2169 documentation for details of partial matching.
2170
2171 PCRE_ERROR_INTERNAL (-14)
2172
2173 An unexpected internal error has occurred. This error could be caused
2174 by a bug in PCRE or by overwriting of the compiled pattern.
2175
2176 PCRE_ERROR_BADCOUNT (-15)
2177
2178 This error is given if the value of the ovecsize argument is negative.
2179
2180 PCRE_ERROR_RECURSIONLIMIT (-21)
2181
2182 The internal recursion limit, as specified by the match_limit_recursion
2183 field in a pcre_extra structure (or defaulted) was reached. See the
2184 description above.
2185
2186 PCRE_ERROR_BADNEWLINE (-23)
2187
2188 An invalid combination of PCRE_NEWLINE_xxx options was given.
2189
2190 Error numbers -16 to -20 and -22 are not used by pcre_exec().
2191
2192
2193 EXTRACTING CAPTURED SUBSTRINGS BY NUMBER
2194
2195 int pcre_copy_substring(const char *subject, int *ovector,
2196 int stringcount, int stringnumber, char *buffer,
2197 int buffersize);
2198
2199 int pcre_get_substring(const char *subject, int *ovector,
2200 int stringcount, int stringnumber,
2201 const char **stringptr);
2202
2203 int pcre_get_substring_list(const char *subject,
2204 int *ovector, int stringcount, const char ***listptr);
2205
2206 Captured substrings can be accessed directly by using the offsets
2207 returned by pcre_exec() in ovector. For convenience, the functions
2208 pcre_copy_substring(), pcre_get_substring(), and pcre_get_sub-
2209 string_list() are provided for extracting captured substrings as new,
2210 separate, zero-terminated strings. These functions identify substrings
2211 by number. The next section describes functions for extracting named
2212 substrings.
2213
2214 A substring that contains a binary zero is correctly extracted and has
2215 a further zero added on the end, but the result is not, of course, a C
2216 string. However, you can process such a string by referring to the
2217 length that is returned by pcre_copy_substring() and pcre_get_sub-
2218 string(). Unfortunately, the interface to pcre_get_substring_list() is
2219 not adequate for handling strings containing binary zeros, because the
2220 end of the final string is not independently indicated.
2221
2222 The first three arguments are the same for all three of these func-
2223 tions: subject is the subject string that has just been successfully
2224 matched, ovector is a pointer to the vector of integer offsets that was
2225 passed to pcre_exec(), and stringcount is the number of substrings that
2226 were captured by the match, including the substring that matched the
2227 entire regular expression. This is the value returned by pcre_exec() if
2228 it is greater than zero. If pcre_exec() returned zero, indicating that
2229 it ran out of space in ovector, the value passed as stringcount should
2230 be the number of elements in the vector divided by three.
2231
2232 The functions pcre_copy_substring() and pcre_get_substring() extract a
2233 single substring, whose number is given as stringnumber. A value of
2234 zero extracts the substring that matched the entire pattern, whereas
2235 higher values extract the captured substrings. For pcre_copy_sub-
2236 string(), the string is placed in buffer, whose length is given by
2237 buffersize, while for pcre_get_substring() a new block of memory is
2238 obtained via pcre_malloc, and its address is returned via stringptr.
2239 The yield of the function is the length of the string, not including
2240 the terminating zero, or one of these error codes:
2241
2242 PCRE_ERROR_NOMEMORY (-6)
2243
2244 The buffer was too small for pcre_copy_substring(), or the attempt to
2245 get memory failed for pcre_get_substring().
2246
2247 PCRE_ERROR_NOSUBSTRING (-7)
2248
2249 There is no substring whose number is stringnumber.
2250
2251 The pcre_get_substring_list() function extracts all available sub-
2252 strings and builds a list of pointers to them. All this is done in a
2253 single block of memory that is obtained via pcre_malloc. The address of
2254 the memory block is returned via listptr, which is also the start of
2255 the list of string pointers. The end of the list is marked by a NULL
2256 pointer. The yield of the function is zero if all went well, or the
2257 error code
2258
2259 PCRE_ERROR_NOMEMORY (-6)
2260
2261 if the attempt to get the memory block failed.
2262
2263 When any of these functions encounter a substring that is unset, which
2264 can happen when capturing subpattern number n+1 matches some part of
2265 the subject, but subpattern n has not been used at all, they return an
2266 empty string. This can be distinguished from a genuine zero-length sub-
2267 string by inspecting the appropriate offset in ovector, which is nega-
2268 tive for unset substrings.
2269
2270 The two convenience functions pcre_free_substring() and pcre_free_sub-
2271 string_list() can be used to free the memory returned by a previous
2272 call of pcre_get_substring() or pcre_get_substring_list(), respec-
2273 tively. They do nothing more than call the function pointed to by
2274 pcre_free, which of course could be called directly from a C program.
2275 However, PCRE is used in some situations where it is linked via a spe-
2276 cial interface to another programming language that cannot use
2277 pcre_free directly; it is for these cases that the functions are pro-
2278 vided.
2279
2280
2281 EXTRACTING CAPTURED SUBSTRINGS BY NAME
2282
2283 int pcre_get_stringnumber(const pcre *code,
2284 const char *name);
2285
2286 int pcre_copy_named_substring(const pcre *code,
2287 const char *subject, int *ovector,
2288 int stringcount, const char *stringname,
2289 char *buffer, int buffersize);
2290
2291 int pcre_get_named_substring(const pcre *code,
2292 const char *subject, int *ovector,
2293 int stringcount, const char *stringname,
2294 const char **stringptr);
2295
2296 To extract a substring by name, you first have to find associated num-
2297 ber. For example, for this pattern
2298
2299 (a+)b(?<xxx>\d+)...
2300
2301 the number of the subpattern called "xxx" is 2. If the name is known to
2302 be unique (PCRE_DUPNAMES was not set), you can find the number from the
2303 name by calling pcre_get_stringnumber(). The first argument is the com-
2304 piled pattern, and the second is the name. The yield of the function is
2305 the subpattern number, or PCRE_ERROR_NOSUBSTRING (-7) if there is no
2306 subpattern of that name.
2307
2308 Given the number, you can extract the substring directly, or use one of
2309 the functions described in the previous section. For convenience, there
2310 are also two functions that do the whole job.
2311
2312 Most of the arguments of pcre_copy_named_substring() and
2313 pcre_get_named_substring() are the same as those for the similarly
2314 named functions that extract by number. As these are described in the
2315 previous section, they are not re-described here. There are just two
2316 differences:
2317
2318 First, instead of a substring number, a substring name is given. Sec-
2319 ond, there is an extra argument, given at the start, which is a pointer
2320 to the compiled pattern. This is needed in order to gain access to the
2321 name-to-number translation table.
2322
2323 These functions call pcre_get_stringnumber(), and if it succeeds, they
2324 then call pcre_copy_substring() or pcre_get_substring(), as appropri-
2325 ate. NOTE: If PCRE_DUPNAMES is set and there are duplicate names, the
2326 behaviour may not be what you want (see the next section).
2327
2328
2329 DUPLICATE SUBPATTERN NAMES
2330
2331 int pcre_get_stringtable_entries(const pcre *code,
2332 const char *name, char **first, char **last);
2333
2334 When a pattern is compiled with the PCRE_DUPNAMES option, names for
2335 subpatterns are not required to be unique. Normally, patterns with
2336 duplicate names are such that in any one match, only one of the named
2337 subpatterns participates. An example is shown in the pcrepattern docu-
2338 mentation.
2339
2340 When duplicates are present, pcre_copy_named_substring() and
2341 pcre_get_named_substring() return the first substring corresponding to
2342 the given name that is set. If none are set, PCRE_ERROR_NOSUBSTRING
2343 (-7) is returned; no data is returned. The pcre_get_stringnumber()
2344 function returns one of the numbers that are associated with the name,
2345 but it is not defined which it is.
2346
2347 If you want to get full details of all captured substrings for a given
2348 name, you must use the pcre_get_stringtable_entries() function. The
2349 first argument is the compiled pattern, and the second is the name. The
2350 third and fourth are pointers to variables which are updated by the
2351 function. After it has run, they point to the first and last entries in
2352 the name-to-number table for the given name. The function itself
2353 returns the length of each entry, or PCRE_ERROR_NOSUBSTRING (-7) if
2354 there are none. The format of the table is described above in the sec-
2355 tion entitled Information about a pattern. Given all the relevant
2356 entries for the name, you can extract each of their numbers, and hence
2357 the captured data, if any.
2358
2359
2360 FINDING ALL POSSIBLE MATCHES
2361
2362 The traditional matching function uses a similar algorithm to Perl,
2363 which stops when it finds the first match, starting at a given point in
2364 the subject. If you want to find all possible matches, or the longest
2365 possible match, consider using the alternative matching function (see
2366 below) instead. If you cannot use the alternative function, but still
2367 need to find all possible matches, you can kludge it up by making use
2368 of the callout facility, which is described in the pcrecallout documen-
2369 tation.
2370
2371 What you have to do is to insert a callout right at the end of the pat-
2372 tern. When your callout function is called, extract and save the cur-
2373 rent matched substring. Then return 1, which forces pcre_exec() to
2374 backtrack and try other alternatives. Ultimately, when it runs out of
2375 matches, pcre_exec() will yield PCRE_ERROR_NOMATCH.
2376
2377
2378 MATCHING A PATTERN: THE ALTERNATIVE FUNCTION
2379
2380 int pcre_dfa_exec(const pcre *code, const pcre_extra *extra,
2381 const char *subject, int length, int startoffset,
2382 int options, int *ovector, int ovecsize,
2383 int *workspace, int wscount);
2384
2385 The function pcre_dfa_exec() is called to match a subject string
2386 against a compiled pattern, using a matching algorithm that scans the
2387 subject string just once, and does not backtrack. This has different
2388 characteristics to the normal algorithm, and is not compatible with
2389 Perl. Some of the features of PCRE patterns are not supported. Never-
2390 theless, there are times when this kind of matching can be useful. For
2391 a discussion of the two matching algorithms, see the pcrematching docu-
2392 mentation.
2393
2394 The arguments for the pcre_dfa_exec() function are the same as for
2395 pcre_exec(), plus two extras. The ovector argument is used in a differ-
2396 ent way, and this is described below. The other common arguments are
2397 used in the same way as for pcre_exec(), so their description is not
2398 repeated here.
2399
2400 The two additional arguments provide workspace for the function. The
2401 workspace vector should contain at least 20 elements. It is used for
2402 keeping track of multiple paths through the pattern tree. More
2403 workspace will be needed for patterns and subjects where there are a
2404 lot of potential matches.
2405
2406 Here is an example of a simple call to pcre_dfa_exec():
2407
2408 int rc;
2409 int ovector[10];
2410 int wspace[20];
2411 rc = pcre_dfa_exec(
2412 re, /* result of pcre_compile() */
2413 NULL, /* we didn't study the pattern */
2414 "some string", /* the subject string */
2415 11, /* the length of the subject string */
2416 0, /* start at offset 0 in the subject */
2417 0, /* default options */
2418 ovector, /* vector of integers for substring information */
2419 10, /* number of elements (NOT size in bytes) */
2420 wspace, /* working space vector */
2421 20); /* number of elements (NOT size in bytes) */
2422
2423 Option bits for pcre_dfa_exec()
2424
2425 The unused bits of the options argument for pcre_dfa_exec() must be
2426 zero. The only bits that may be set are PCRE_ANCHORED, PCRE_NEW-
2427 LINE_xxx, PCRE_NOTBOL, PCRE_NOTEOL, PCRE_NOTEMPTY, PCRE_NO_UTF8_CHECK,
2428 PCRE_PARTIAL, PCRE_DFA_SHORTEST, and PCRE_DFA_RESTART. All but the last
2429 three of these are the same as for pcre_exec(), so their description is
2430 not repeated here.
2431
2432 PCRE_PARTIAL
2433
2434 This has the same general effect as it does for pcre_exec(), but the
2435 details are slightly different. When PCRE_PARTIAL is set for
2436 pcre_dfa_exec(), the return code PCRE_ERROR_NOMATCH is converted into
2437 PCRE_ERROR_PARTIAL if the end of the subject is reached, there have
2438 been no complete matches, but there is still at least one matching pos-
2439 sibility. The portion of the string that provided the partial match is
2440 set as the first matching string.
2441
2442 PCRE_DFA_SHORTEST
2443
2444 Setting the PCRE_DFA_SHORTEST option causes the matching algorithm to
2445 stop as soon as it has found one match. Because of the way the alterna-
2446 tive algorithm works, this is necessarily the shortest possible match
2447 at the first possible matching point in the subject string.
2448
2449 PCRE_DFA_RESTART
2450
2451 When pcre_dfa_exec() is called with the PCRE_PARTIAL option, and
2452 returns a partial match, it is possible to call it again, with addi-
2453 tional subject characters, and have it continue with the same match.
2454 The PCRE_DFA_RESTART option requests this action; when it is set, the
2455 workspace and wscount options must reference the same vector as before
2456 because data about the match so far is left in them after a partial
2457 match. There is more discussion of this facility in the pcrepartial
2458 documentation.
2459
2460 Successful returns from pcre_dfa_exec()
2461
2462 When pcre_dfa_exec() succeeds, it may have matched more than one sub-
2463 string in the subject. Note, however, that all the matches from one run
2464 of the function start at the same point in the subject. The shorter
2465 matches are all initial substrings of the longer matches. For example,
2466 if the pattern
2467
2468 <.*>
2469
2470 is matched against the string
2471
2472 This is <something> <something else> <something further> no more
2473
2474 the three matched strings are
2475
2476 <something>
2477 <something> <something else>
2478 <something> <something else> <something further>
2479
2480 On success, the yield of the function is a number greater than zero,
2481 which is the number of matched substrings. The substrings themselves
2482 are returned in ovector. Each string uses two elements; the first is
2483 the offset to the start, and the second is the offset to the end. In
2484 fact, all the strings have the same start offset. (Space could have
2485 been saved by giving this only once, but it was decided to retain some
2486 compatibility with the way pcre_exec() returns data, even though the
2487 meaning of the strings is different.)
2488
2489 The strings are returned in reverse order of length; that is, the long-
2490 est matching string is given first. If there were too many matches to
2491 fit into ovector, the yield of the function is zero, and the vector is
2492 filled with the longest matches.
2493
2494 Error returns from pcre_dfa_exec()
2495
2496 The pcre_dfa_exec() function returns a negative number when it fails.
2497 Many of the errors are the same as for pcre_exec(), and these are
2498 described above. There are in addition the following errors that are
2499 specific to pcre_dfa_exec():
2500
2501 PCRE_ERROR_DFA_UITEM (-16)
2502
2503 This return is given if pcre_dfa_exec() encounters an item in the pat-
2504 tern that it does not support, for instance, the use of \C or a back
2505 reference.
2506
2507 PCRE_ERROR_DFA_UCOND (-17)
2508
2509 This return is given if pcre_dfa_exec() encounters a condition item
2510 that uses a back reference for the condition, or a test for recursion
2511 in a specific group. These are not supported.
2512
2513 PCRE_ERROR_DFA_UMLIMIT (-18)
2514
2515 This return is given if pcre_dfa_exec() is called with an extra block
2516 that contains a setting of the match_limit field. This is not supported
2517 (it is meaningless).
2518
2519 PCRE_ERROR_DFA_WSSIZE (-19)
2520
2521 This return is given if pcre_dfa_exec() runs out of space in the
2522 workspace vector.
2523
2524 PCRE_ERROR_DFA_RECURSE (-20)
2525
2526 When a recursive subpattern is processed, the matching function calls
2527 itself recursively, using private vectors for ovector and workspace.
2528 This error is given if the output vector is not large enough. This
2529 should be extremely rare, as a vector of size 1000 is used.
2530
2531
2532 SEE ALSO
2533
2534 pcrebuild(3), pcrecallout(3), pcrecpp(3)(3), pcrematching(3), pcrepar-
2535 tial(3), pcreposix(3), pcreprecompile(3), pcresample(3), pcrestack(3).
2536
2537
2538 AUTHOR
2539
2540 Philip Hazel
2541 University Computing Service
2542 Cambridge CB2 3QH, England.
2543
2544
2545 REVISION
2546
2547 Last updated: 27 November 2007
2548 Copyright (c) 1997-2007 University of Cambridge.
2549 ------------------------------------------------------------------------------
2550
2551
2552 PCRECALLOUT(3) PCRECALLOUT(3)
2553
2554
2555 NAME
2556 PCRE - Perl-compatible regular expressions
2557
2558
2559 PCRE CALLOUTS
2560
2561 int (*pcre_callout)(pcre_callout_block *);
2562
2563 PCRE provides a feature called "callout", which is a means of temporar-
2564 ily passing control to the caller of PCRE in the middle of pattern
2565 matching. The caller of PCRE provides an external function by putting
2566 its entry point in the global variable pcre_callout. By default, this
2567 variable contains NULL, which disables all calling out.
2568
2569 Within a regular expression, (?C) indicates the points at which the
2570 external function is to be called. Different callout points can be
2571 identified by putting a number less than 256 after the letter C. The
2572 default value is zero. For example, this pattern has two callout
2573 points:
2574
2575 (?C1)abc(?C2)def
2576
2577 If the PCRE_AUTO_CALLOUT option bit is set when pcre_compile() is
2578 called, PCRE automatically inserts callouts, all with number 255,
2579 before each item in the pattern. For example, if PCRE_AUTO_CALLOUT is
2580 used with the pattern
2581
2582 A(\d{2}|--)
2583
2584 it is processed as if it were
2585
2586 (?C255)A(?C255)((?C255)\d{2}(?C255)|(?C255)-(?C255)-(?C255))(?C255)
2587
2588 Notice that there is a callout before and after each parenthesis and
2589 alternation bar. Automatic callouts can be used for tracking the
2590 progress of pattern matching. The pcretest command has an option that
2591 sets automatic callouts; when it is used, the output indicates how the
2592 pattern is matched. This is useful information when you are trying to
2593 optimize the performance of a particular pattern.
2594
2595
2596 MISSING CALLOUTS
2597
2598 You should be aware that, because of optimizations in the way PCRE
2599 matches patterns, callouts sometimes do not happen. For example, if the
2600 pattern is
2601
2602 ab(?C4)cd
2603
2604 PCRE knows that any matching string must contain the letter "d". If the
2605 subject string is "abyz", the lack of "d" means that matching doesn't
2606 ever start, and the callout is never reached. However, with "abyd",
2607 though the result is still no match, the callout is obeyed.
2608
2609
2610 THE CALLOUT INTERFACE
2611
2612 During matching, when PCRE reaches a callout point, the external func-
2613 tion defined by pcre_callout is called (if it is set). This applies to
2614 both the pcre_exec() and the pcre_dfa_exec() matching functions. The
2615 only argument to the callout function is a pointer to a pcre_callout
2616 block. This structure contains the following fields:
2617
2618 int version;
2619 int callout_number;
2620 int *offset_vector;
2621 const char *subject;
2622 int subject_length;
2623 int start_match;
2624 int current_position;
2625 int capture_top;
2626 int capture_last;
2627 void *callout_data;
2628 int pattern_position;
2629 int next_item_length;
2630
2631 The version field is an integer containing the version number of the
2632 block format. The initial version was 0; the current version is 1. The
2633 version number will change again in future if additional fields are
2634 added, but the intention is never to remove any of the existing fields.
2635
2636 The callout_number field contains the number of the callout, as com-
2637 piled into the pattern (that is, the number after ?C for manual call-
2638 outs, and 255 for automatically generated callouts).
2639
2640 The offset_vector field is a pointer to the vector of offsets that was
2641 passed by the caller to pcre_exec() or pcre_dfa_exec(). When
2642 pcre_exec() is used, the contents can be inspected in order to extract
2643 substrings that have been matched so far, in the same way as for
2644 extracting substrings after a match has completed. For pcre_dfa_exec()
2645 this field is not useful.
2646
2647 The subject and subject_length fields contain copies of the values that
2648 were passed to pcre_exec().
2649
2650 The start_match field normally contains the offset within the subject
2651 at which the current match attempt started. However, if the escape
2652 sequence \K has been encountered, this value is changed to reflect the
2653 modified starting point. If the pattern is not anchored, the callout
2654 function may be called several times from the same point in the pattern
2655 for different starting points in the subject.
2656
2657 The current_position field contains the offset within the subject of
2658 the current match pointer.
2659
2660 When the pcre_exec() function is used, the capture_top field contains
2661 one more than the number of the highest numbered captured substring so
2662 far. If no substrings have been captured, the value of capture_top is
2663 one. This is always the case when pcre_dfa_exec() is used, because it
2664 does not support captured substrings.
2665
2666 The capture_last field contains the number of the most recently cap-
2667 tured substring. If no substrings have been captured, its value is -1.
2668 This is always the case when pcre_dfa_exec() is used.
2669
2670 The callout_data field contains a value that is passed to pcre_exec()
2671 or pcre_dfa_exec() specifically so that it can be passed back in call-
2672 outs. It is passed in the pcre_callout field of the pcre_extra data
2673 structure. If no such data was passed, the value of callout_data in a
2674 pcre_callout block is NULL. There is a description of the pcre_extra
2675 structure in the pcreapi documentation.
2676
2677 The pattern_position field is present from version 1 of the pcre_call-
2678 out structure. It contains the offset to the next item to be matched in
2679 the pattern string.
2680
2681 The next_item_length field is present from version 1 of the pcre_call-
2682 out structure. It contains the length of the next item to be matched in
2683 the pattern string. When the callout immediately precedes an alterna-
2684 tion bar, a closing parenthesis, or the end of the pattern, the length
2685 is zero. When the callout precedes an opening parenthesis, the length
2686 is that of the entire subpattern.
2687
2688 The pattern_position and next_item_length fields are intended to help
2689 in distinguishing between different automatic callouts, which all have
2690 the same callout number. However, they are set for all callouts.
2691
2692
2693 RETURN VALUES
2694
2695 The external callout function returns an integer to PCRE. If the value
2696 is zero, matching proceeds as normal. If the value is greater than
2697 zero, matching fails at the current point, but the testing of other
2698 matching possibilities goes ahead, just as if a lookahead assertion had
2699 failed. If the value is less than zero, the match is abandoned, and
2700 pcre_exec() (or pcre_dfa_exec()) returns the negative value.
2701
2702 Negative values should normally be chosen from the set of
2703 PCRE_ERROR_xxx values. In particular, PCRE_ERROR_NOMATCH forces a stan-
2704 dard "no match" failure. The error number PCRE_ERROR_CALLOUT is
2705 reserved for use by callout functions; it will never be used by PCRE
2706 itself.
2707
2708
2709 AUTHOR
2710
2711 Philip Hazel
2712 University Computing Service
2713 Cambridge CB2 3QH, England.
2714
2715
2716 REVISION
2717
2718 Last updated: 29 May 2007
2719 Copyright (c) 1997-2007 University of Cambridge.
2720 ------------------------------------------------------------------------------
2721
2722
2723 PCRECOMPAT(3) PCRECOMPAT(3)
2724
2725
2726 NAME
2727 PCRE - Perl-compatible regular expressions
2728
2729
2730 DIFFERENCES BETWEEN PCRE AND PERL
2731
2732 This document describes the differences in the ways that PCRE and Perl
2733 handle regular expressions. The differences described here are mainly
2734 with respect to Perl 5.8, though PCRE versions 7.0 and later contain
2735 some features that are expected to be in the forthcoming Perl 5.10.
2736
2737 1. PCRE has only a subset of Perl's UTF-8 and Unicode support. Details
2738 of what it does have are given in the section on UTF-8 support in the
2739 main pcre page.
2740
2741 2. PCRE does not allow repeat quantifiers on lookahead assertions. Perl
2742 permits them, but they do not mean what you might think. For example,
2743 (?!a){3} does not assert that the next three characters are not "a". It
2744 just asserts that the next character is not "a" three times.
2745
2746 3. Capturing subpatterns that occur inside negative lookahead asser-
2747 tions are counted, but their entries in the offsets vector are never
2748 set. Perl sets its numerical variables from any such patterns that are
2749 matched before the assertion fails to match something (thereby succeed-
2750 ing), but only if the negative lookahead assertion contains just one
2751 branch.
2752
2753 4. Though binary zero characters are supported in the subject string,
2754 they are not allowed in a pattern string because it is passed as a nor-
2755 mal C string, terminated by zero. The escape sequence \0 can be used in
2756 the pattern to represent a binary zero.
2757
2758 5. The following Perl escape sequences are not supported: \l, \u, \L,
2759 \U, and \N. In fact these are implemented by Perl's general string-han-
2760 dling and are not part of its pattern matching engine. If any of these
2761 are encountered by PCRE, an error is generated.
2762
2763 6. The Perl escape sequences \p, \P, and \X are supported only if PCRE
2764 is built with Unicode character property support. The properties that
2765 can be tested with \p and \P are limited to the general category prop-
2766 erties such as Lu and Nd, script names such as Greek or Han, and the
2767 derived properties Any and L&.
2768
2769 7. PCRE does support the \Q...\E escape for quoting substrings. Charac-
2770 ters in between are treated as literals. This is slightly different
2771 from Perl in that $ and @ are also handled as literals inside the
2772 quotes. In Perl, they cause variable interpolation (but of course PCRE
2773 does not have variables). Note the following examples:
2774
2775 Pattern PCRE matches Perl matches
2776
2777 \Qabc$xyz\E abc$xyz abc followed by the
2778 contents of $xyz
2779 \Qabc\$xyz\E abc\$xyz abc\$xyz
2780 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
2781
2782 The \Q...\E sequence is recognized both inside and outside character
2783 classes.
2784
2785 8. Fairly obviously, PCRE does not support the (?{code}) and (??{code})
2786 constructions. However, there is support for recursive patterns. This
2787 is not available in Perl 5.8, but will be in Perl 5.10. Also, the PCRE
2788 "callout" feature allows an external function to be called during pat-
2789 tern matching. See the pcrecallout documentation for details.
2790
2791 9. Subpatterns that are called recursively or as "subroutines" are
2792 always treated as atomic groups in PCRE. This is like Python, but
2793 unlike Perl.
2794
2795 10. There are some differences that are concerned with the settings of
2796 captured strings when part of a pattern is repeated. For example,
2797 matching "aba" against the pattern /^(a(b)?)+$/ in Perl leaves $2
2798 unset, but in PCRE it is set to "b".
2799
2800 11. PCRE does support Perl 5.10's backtracking verbs (*ACCEPT),
2801 (*FAIL), (*F), (*COMMIT), (*PRUNE), (*SKIP), and (*THEN), but only in
2802 the forms without an argument. PCRE does not support (*MARK). If
2803 (*ACCEPT) is within capturing parentheses, PCRE does not set that cap-
2804 ture group; this is different to Perl.
2805
2806 12. PCRE provides some extensions to the Perl regular expression facil-
2807 ities. Perl 5.10 will include new features that are not in earlier
2808 versions, some of which (such as named parentheses) have been in PCRE
2809 for some time. This list is with respect to Perl 5.10:
2810
2811 (a) Although lookbehind assertions must match fixed length strings,
2812 each alternative branch of a lookbehind assertion can match a different
2813 length of string. Perl requires them all to have the same length.
2814
2815 (b) If PCRE_DOLLAR_ENDONLY is set and PCRE_MULTILINE is not set, the $
2816 meta-character matches only at the very end of the string.
2817
2818 (c) If PCRE_EXTRA is set, a backslash followed by a letter with no spe-
2819 cial meaning is faulted. Otherwise, like Perl, the backslash is quietly
2820 ignored. (Perl can be made to issue a warning.)
2821
2822 (d) If PCRE_UNGREEDY is set, the greediness of the repetition quanti-
2823 fiers is inverted, that is, by default they are not greedy, but if fol-
2824 lowed by a question mark they are.
2825
2826 (e) PCRE_ANCHORED can be used at matching time to force a pattern to be
2827 tried only at the first matching position in the subject string.
2828
2829 (f) The PCRE_NOTBOL, PCRE_NOTEOL, PCRE_NOTEMPTY, and PCRE_NO_AUTO_CAP-
2830 TURE options for pcre_exec() have no Perl equivalents.
2831
2832 (g) The \R escape sequence can be restricted to match only CR, LF, or
2833 CRLF by the PCRE_BSR_ANYCRLF option.
2834
2835 (h) The callout facility is PCRE-specific.
2836
2837 (i) The partial matching facility is PCRE-specific.
2838
2839 (j) Patterns compiled by PCRE can be saved and re-used at a later time,
2840 even on different hosts that have the other endianness.
2841
2842 (k) The alternative matching function (pcre_dfa_exec()) matches in a
2843 different way and is not Perl-compatible.
2844
2845 (l) PCRE recognizes some special sequences such as (*CR) at the start
2846 of a pattern that set overall options that cannot be changed within the
2847 pattern.
2848
2849
2850 AUTHOR
2851
2852 Philip Hazel
2853 University Computing Service
2854 Cambridge CB2 3QH, England.
2855
2856
2857 REVISION
2858
2859 Last updated: 11 September 2007
2860 Copyright (c) 1997-2007 University of Cambridge.
2861 ------------------------------------------------------------------------------
2862
2863
2864 PCREPATTERN(3) PCREPATTERN(3)
2865
2866
2867 NAME
2868 PCRE - Perl-compatible regular expressions
2869
2870
2871 PCRE REGULAR EXPRESSION DETAILS
2872
2873 The syntax and semantics of the regular expressions that are supported
2874 by PCRE are described in detail below. There is a quick-reference syn-
2875 tax summary in the pcresyntax page. Perl's regular expressions are
2876 described in its own documentation, and regular expressions in general
2877 are covered in a number of books, some of which have copious examples.
2878 Jeffrey Friedl's "Mastering Regular Expressions", published by
2879 O'Reilly, covers regular expressions in great detail. This description
2880 of PCRE's regular expressions is intended as reference material.
2881
2882 The original operation of PCRE was on strings of one-byte characters.
2883 However, there is now also support for UTF-8 character strings. To use
2884 this, you must build PCRE to include UTF-8 support, and then call
2885 pcre_compile() with the PCRE_UTF8 option. How this affects pattern
2886 matching is mentioned in several places below. There is also a summary
2887 of UTF-8 features in the section on UTF-8 support in the main pcre
2888 page.
2889
2890 The remainder of this document discusses the patterns that are sup-
2891 ported by PCRE when its main matching function, pcre_exec(), is used.
2892 From release 6.0, PCRE offers a second matching function,
2893 pcre_dfa_exec(), which matches using a different algorithm that is not
2894 Perl-compatible. Some of the features discussed below are not available
2895 when pcre_dfa_exec() is used. The advantages and disadvantages of the
2896 alternative function, and how it differs from the normal function, are
2897 discussed in the pcrematching page.
2898
2899
2900 NEWLINE CONVENTIONS
2901
2902 PCRE supports five different conventions for indicating line breaks in
2903 strings: a single CR (carriage return) character, a single LF (line-
2904 feed) character, the two-character sequence CRLF, any of the three pre-
2905 ceding, or any Unicode newline sequence. The pcreapi page has further
2906 discussion about newlines, and shows how to set the newline convention
2907 in the options arguments for the compiling and matching functions.
2908
2909 It is also possible to specify a newline convention by starting a pat-
2910 tern string with one of the following five sequences:
2911
2912 (*CR) carriage return
2913 (*LF) linefeed
2914 (*CRLF) carriage return, followed by linefeed
2915 (*ANYCRLF) any of the three above
2916 (*ANY) all Unicode newline sequences
2917
2918 These override the default and the options given to pcre_compile(). For
2919 example, on a Unix system where LF is the default newline sequence, the
2920 pattern
2921
2922 (*CR)a.b
2923
2924 changes the convention to CR. That pattern matches "a\nb" because LF is
2925 no longer a newline. Note that these special settings, which are not
2926 Perl-compatible, are recognized only at the very start of a pattern,
2927 and that they must be in upper case. If more than one of them is
2928 present, the last one is used.
2929
2930 The newline convention does not affect what the \R escape sequence
2931 matches. By default, this is any Unicode newline sequence, for Perl
2932 compatibility. However, this can be changed; see the description of \R
2933 in the section entitled "Newline sequences" below. A change of \R set-
2934 ting can be combined with a change of newline convention.
2935
2936
2937 CHARACTERS AND METACHARACTERS
2938
2939 A regular expression is a pattern that is matched against a subject
2940 string from left to right. Most characters stand for themselves in a
2941 pattern, and match the corresponding characters in the subject. As a
2942 trivial example, the pattern
2943
2944 The quick brown fox
2945
2946 matches a portion of a subject string that is identical to itself. When
2947 caseless matching is specified (the PCRE_CASELESS option), letters are
2948 matched independently of case. In UTF-8 mode, PCRE always understands
2949 the concept of case for characters whose values are less than 128, so
2950 caseless matching is always possible. For characters with higher val-
2951 ues, the concept of case is supported if PCRE is compiled with Unicode
2952 property support, but not otherwise. If you want to use caseless
2953 matching for characters 128 and above, you must ensure that PCRE is
2954 compiled with Unicode property support as well as with UTF-8 support.
2955
2956 The power of regular expressions comes from the ability to include
2957 alternatives and repetitions in the pattern. These are encoded in the
2958 pattern by the use of metacharacters, which do not stand for themselves
2959 but instead are interpreted in some special way.
2960
2961 There are two different sets of metacharacters: those that are recog-
2962 nized anywhere in the pattern except within square brackets, and those
2963 that are recognized within square brackets. Outside square brackets,
2964 the metacharacters are as follows:
2965
2966 \ general escape character with several uses
2967 ^ assert start of string (or line, in multiline mode)
2968 $ assert end of string (or line, in multiline mode)
2969 . match any character except newline (by default)
2970 [ start character class definition
2971 | start of alternative branch
2972 ( start subpattern
2973 ) end subpattern
2974 ? extends the meaning of (
2975 also 0 or 1 quantifier
2976 also quantifier minimizer
2977 * 0 or more quantifier
2978 + 1 or more quantifier
2979 also "possessive quantifier"
2980 { start min/max quantifier
2981
2982 Part of a pattern that is in square brackets is called a "character
2983 class". In a character class the only metacharacters are:
2984
2985 \ general escape character
2986 ^ negate the class, but only if the first character
2987 - indicates character range
2988 [ POSIX character class (only if followed by POSIX
2989 syntax)
2990 ] terminates the character class
2991
2992 The following sections describe the use of each of the metacharacters.
2993
2994
2995 BACKSLASH
2996
2997 The backslash character has several uses. Firstly, if it is followed by
2998 a non-alphanumeric character, it takes away any special meaning that
2999 character may have. This use of backslash as an escape character
3000 applies both inside and outside character classes.
3001
3002 For example, if you want to match a * character, you write \* in the
3003 pattern. This escaping action applies whether or not the following
3004 character would otherwise be interpreted as a metacharacter, so it is
3005 always safe to precede a non-alphanumeric with backslash to specify
3006 that it stands for itself. In particular, if you want to match a back-
3007 slash, you write \\.
3008
3009 If a pattern is compiled with the PCRE_EXTENDED option, whitespace in
3010 the pattern (other than in a character class) and characters between a
3011 # outside a character class and the next newline are ignored. An escap-
3012 ing backslash can be used to include a whitespace or # character as
3013 part of the pattern.
3014
3015 If you want to remove the special meaning from a sequence of charac-
3016 ters, you can do so by putting them between \Q and \E. This is differ-
3017 ent from Perl in that $ and @ are handled as literals in \Q...\E
3018 sequences in PCRE, whereas in Perl, $ and @ cause variable interpola-
3019 tion. Note the following examples:
3020
3021 Pattern PCRE matches Perl matches
3022
3023 \Qabc$xyz\E abc$xyz abc followed by the
3024 contents of $xyz
3025 \Qabc\$xyz\E abc\$xyz abc\$xyz
3026 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3027
3028 The \Q...\E sequence is recognized both inside and outside character
3029 classes.
3030
3031 Non-printing characters
3032
3033 A second use of backslash provides a way of encoding non-printing char-
3034 acters in patterns in a visible manner. There is no restriction on the
3035 appearance of non-printing characters, apart from the binary zero that
3036 terminates a pattern, but when a pattern is being prepared by text
3037 editing, it is usually easier to use one of the following escape
3038 sequences than the binary character it represents:
3039
3040 \a alarm, that is, the BEL character (hex 07)
3041 \cx "control-x", where x is any character
3042 \e escape (hex 1B)
3043 \f formfeed (hex 0C)
3044 \n linefeed (hex 0A)
3045 \r carriage return (hex 0D)
3046 \t tab (hex 09)
3047 \ddd character with octal code ddd, or backreference
3048 \xhh character with hex code hh
3049 \x{hhh..} character with hex code hhh..
3050
3051 The precise effect of \cx is as follows: if x is a lower case letter,
3052 it is converted to upper case. Then bit 6 of the character (hex 40) is
3053 inverted. Thus \cz becomes hex 1A, but \c{ becomes hex 3B, while \c;
3054 becomes hex 7B.
3055
3056 After \x, from zero to two hexadecimal digits are read (letters can be
3057 in upper or lower case). Any number of hexadecimal digits may appear
3058 between \x{ and }, but the value of the character code must be less
3059 than 256 in non-UTF-8 mode, and less than 2**31 in UTF-8 mode. That is,
3060 the maximum value in hexadecimal is 7FFFFFFF. Note that this is bigger
3061 than the largest Unicode code point, which is 10FFFF.
3062
3063 If characters other than hexadecimal digits appear between \x{ and },
3064 or if there is no terminating }, this form of escape is not recognized.
3065 Instead, the initial \x will be interpreted as a basic hexadecimal
3066 escape, with no following digits, giving a character whose value is
3067 zero.
3068
3069 Characters whose value is less than 256 can be defined by either of the
3070 two syntaxes for \x. There is no difference in the way they are han-
3071 dled. For example, \xdc is exactly the same as \x{dc}.
3072
3073 After \0 up to two further octal digits are read. If there are fewer
3074 than two digits, just those that are present are used. Thus the
3075 sequence \0\x\07 specifies two binary zeros followed by a BEL character
3076 (code value 7). Make sure you supply two digits after the initial zero
3077 if the pattern character that follows is itself an octal digit.
3078
3079 The handling of a backslash followed by a digit other than 0 is compli-
3080 cated. Outside a character class, PCRE reads it and any following dig-
3081 its as a decimal number. If the number is less than 10, or if there
3082 have been at least that many previous capturing left parentheses in the
3083 expression, the entire sequence is taken as a back reference. A
3084 description of how this works is given later, following the discussion
3085 of parenthesized subpatterns.
3086
3087 Inside a character class, or if the decimal number is greater than 9
3088 and there have not been that many capturing subpatterns, PCRE re-reads
3089 up to three octal digits following the backslash, and uses them to gen-
3090 erate a data character. Any subsequent digits stand for themselves. In
3091 non-UTF-8 mode, the value of a character specified in octal must be
3092 less than \400. In UTF-8 mode, values up to \777 are permitted. For
3093 example:
3094
3095 \040 is another way of writing a space
3096 \40 is the same, provided there are fewer than 40
3097 previous capturing subpatterns
3098 \7 is always a back reference
3099 \11 might be a back reference, or another way of
3100 writing a tab
3101 \011 is always a tab
3102 \0113 is a tab followed by the character "3"
3103 \113 might be a back reference, otherwise the
3104 character with octal code 113
3105 \377 might be a back reference, otherwise
3106 the byte consisting entirely of 1 bits
3107 \81 is either a back reference, or a binary zero
3108 followed by the two characters "8" and "1"
3109
3110 Note that octal values of 100 or greater must not be introduced by a
3111 leading zero, because no more than three octal digits are ever read.
3112
3113 All the sequences that define a single character value can be used both
3114 inside and outside character classes. In addition, inside a character
3115 class, the sequence \b is interpreted as the backspace character (hex
3116 08), and the sequences \R and \X are interpreted as the characters "R"
3117 and "X", respectively. Outside a character class, these sequences have
3118 different meanings (see below).
3119
3120 Absolute and relative back references
3121
3122 The sequence \g followed by an unsigned or a negative number, option-
3123 ally enclosed in braces, is an absolute or relative back reference. A
3124 named back reference can be coded as \g{name}. Back references are dis-
3125 cussed later, following the discussion of parenthesized subpatterns.
3126
3127 Generic character types
3128
3129 Another use of backslash is for specifying generic character types. The
3130 following are always recognized:
3131
3132 \d any decimal digit
3133 \D any character that is not a decimal digit
3134 \h any horizontal whitespace character
3135 \H any character that is not a horizontal whitespace character
3136 \s any whitespace character
3137 \S any character that is not a whitespace character
3138 \v any vertical whitespace character
3139 \V any character that is not a vertical whitespace character
3140 \w any "word" character
3141 \W any "non-word" character
3142
3143 Each pair of escape sequences partitions the complete set of characters
3144 into two disjoint sets. Any given character matches one, and only one,
3145 of each pair.
3146
3147 These character type sequences can appear both inside and outside char-
3148 acter classes. They each match one character of the appropriate type.
3149 If the current matching point is at the end of the subject string, all
3150 of them fail, since there is no character to match.
3151
3152 For compatibility with Perl, \s does not match the VT character (code
3153 11). This makes it different from the the POSIX "space" class. The \s
3154 characters are HT (9), LF (10), FF (12), CR (13), and space (32). If
3155 "use locale;" is included in a Perl script, \s may match the VT charac-
3156 ter. In PCRE, it never does.
3157
3158 In UTF-8 mode, characters with values greater than 128 never match \d,
3159 \s, or \w, and always match \D, \S, and \W. This is true even when Uni-
3160 code character property support is available. These sequences retain
3161 their original meanings from before UTF-8 support was available, mainly
3162 for efficiency reasons.
3163
3164 The sequences \h, \H, \v, and \V are Perl 5.10 features. In contrast to
3165 the other sequences, these do match certain high-valued codepoints in
3166 UTF-8 mode. The horizontal space characters are:
3167
3168 U+0009 Horizontal tab
3169 U+0020 Space
3170 U+00A0 Non-break space
3171 U+1680 Ogham space mark
3172 U+180E Mongolian vowel separator
3173 U+2000 En quad
3174 U+2001 Em quad
3175 U+2002 En space
3176 U+2003 Em space
3177 U+2004 Three-per-em space
3178 U+2005 Four-per-em space
3179 U+2006 Six-per-em space
3180 U+2007 Figure space
3181 U+2008 Punctuation space
3182 U+2009 Thin space
3183 U+200A Hair space
3184 U+202F Narrow no-break space
3185 U+205F Medium mathematical space
3186 U+3000 Ideographic space
3187
3188 The vertical space characters are:
3189
3190 U+000A Linefeed
3191 U+000B Vertical tab
3192 U+000C Formfeed
3193 U+000D Carriage return
3194 U+0085 Next line
3195 U+2028 Line separator
3196 U+2029 Paragraph separator
3197
3198 A "word" character is an underscore or any character less than 256 that
3199 is a letter or digit. The definition of letters and digits is con-
3200 trolled by PCRE's low-valued character tables, and may vary if locale-
3201 specific matching is taking place (see "Locale support" in the pcreapi
3202 page). For example, in a French locale such as "fr_FR" in Unix-like
3203 systems, or "french" in Windows, some character codes greater than 128
3204 are used for accented letters, and these are matched by \w. The use of
3205 locales with Unicode is discouraged.
3206
3207 Newline sequences
3208
3209 Outside a character class, by default, the escape sequence \R matches
3210 any Unicode newline sequence. This is a Perl 5.10 feature. In non-UTF-8
3211 mode \R is equivalent to the following:
3212
3213 (?>\r\n|\n|\x0b|\f|\r|\x85)
3214
3215 This is an example of an "atomic group", details of which are given
3216 below. This particular group matches either the two-character sequence
3217 CR followed by LF, or one of the single characters LF (linefeed,
3218 U+000A), VT (vertical tab, U+000B), FF (formfeed, U+000C), CR (carriage
3219 return, U+000D), or NEL (next line, U+0085). The two-character sequence
3220 is treated as a single unit that cannot be split.
3221
3222 In UTF-8 mode, two additional characters whose codepoints are greater
3223 than 255 are added: LS (line separator, U+2028) and PS (paragraph sepa-
3224 rator, U+2029). Unicode character property support is not needed for
3225 these characters to be recognized.
3226
3227 It is possible to restrict \R to match only CR, LF, or CRLF (instead of
3228 the complete set of Unicode line endings) by setting the option
3229 PCRE_BSR_ANYCRLF either at compile time or when the pattern is matched.
3230 (BSR is an abbrevation for "backslash R".) This can be made the default
3231 when PCRE is built; if this is the case, the other behaviour can be
3232 requested via the PCRE_BSR_UNICODE option. It is also possible to
3233 specify these settings by starting a pattern string with one of the
3234 following sequences:
3235
3236 (*BSR_ANYCRLF) CR, LF, or CRLF only
3237 (*BSR_UNICODE) any Unicode newline sequence
3238
3239 These override the default and the options given to pcre_compile(), but
3240 they can be overridden by options given to pcre_exec(). Note that these
3241 special settings, which are not Perl-compatible, are recognized only at
3242 the very start of a pattern, and that they must be in upper case. If
3243 more than one of them is present, the last one is used. They can be
3244 combined with a change of newline convention, for example, a pattern
3245 can start with:
3246
3247 (*ANY)(*BSR_ANYCRLF)
3248
3249 Inside a character class, \R matches the letter "R".
3250
3251 Unicode character properties
3252
3253 When PCRE is built with Unicode character property support, three addi-
3254 tional escape sequences that match characters with specific properties
3255 are available. When not in UTF-8 mode, these sequences are of course
3256 limited to testing characters whose codepoints are less than 256, but
3257 they do work in this mode. The extra escape sequences are:
3258
3259 \p{xx} a character with the xx property
3260 \P{xx} a character without the xx property
3261 \X an extended Unicode sequence
3262
3263 The property names represented by xx above are limited to the Unicode
3264 script names, the general category properties, and "Any", which matches
3265 any character (including newline). Other properties such as "InMusical-
3266 Symbols" are not currently supported by PCRE. Note that \P{Any} does
3267 not match any characters, so always causes a match failure.
3268
3269 Sets of Unicode characters are defined as belonging to certain scripts.
3270 A character from one of these sets can be matched using a script name.
3271 For example:
3272
3273 \p{Greek}
3274 \P{Han}
3275
3276 Those that are not part of an identified script are lumped together as
3277 "Common". The current list of scripts is:
3278
3279 Arabic, Armenian, Balinese, Bengali, Bopomofo, Braille, Buginese,
3280 Buhid, Canadian_Aboriginal, Cherokee, Common, Coptic, Cuneiform,
3281 Cypriot, Cyrillic, Deseret, Devanagari, Ethiopic, Georgian, Glagolitic,
3282 Gothic, Greek, Gujarati, Gurmukhi, Han, Hangul, Hanunoo, Hebrew, Hira-
3283 gana, Inherited, Kannada, Katakana, Kharoshthi, Khmer, Lao, Latin,
3284 Limbu, Linear_B, Malayalam, Mongolian, Myanmar, New_Tai_Lue, Nko,
3285 Ogham, Old_Italic, Old_Persian, Oriya, Osmanya, Phags_Pa, Phoenician,
3286 Runic, Shavian, Sinhala, Syloti_Nagri, Syriac, Tagalog, Tagbanwa,
3287 Tai_Le, Tamil, Telugu, Thaana, Thai, Tibetan, Tifinagh, Ugaritic, Yi.
3288
3289 Each character has exactly one general category property, specified by
3290 a two-letter abbreviation. For compatibility with Perl, negation can be
3291 specified by including a circumflex between the opening brace and the
3292 property name. For example, \p{^Lu} is the same as \P{Lu}.
3293
3294 If only one letter is specified with \p or \P, it includes all the gen-
3295 eral category properties that start with that letter. In this case, in
3296 the absence of negation, the curly brackets in the escape sequence are
3297 optional; these two examples have the same effect:
3298
3299 \p{L}
3300 \pL
3301
3302 The following general category property codes are supported:
3303
3304 C Other
3305 Cc Control
3306 Cf Format
3307 Cn Unassigned
3308 Co Private use
3309 Cs Surrogate
3310
3311 L Letter
3312 Ll Lower case letter
3313 Lm Modifier letter
3314 Lo Other letter
3315 Lt Title case letter
3316 Lu Upper case letter
3317
3318 M Mark
3319 Mc Spacing mark
3320 Me Enclosing mark
3321 Mn Non-spacing mark
3322
3323 N Number
3324 Nd Decimal number
3325 Nl Letter number
3326 No Other number
3327
3328 P Punctuation
3329 Pc Connector punctuation
3330 Pd Dash punctuation
3331 Pe Close punctuation
3332 Pf Final punctuation
3333 Pi Initial punctuation
3334 Po Other punctuation
3335 Ps Open punctuation
3336
3337 S Symbol
3338 Sc Currency symbol
3339 Sk Modifier symbol
3340 Sm Mathematical symbol
3341 So Other symbol
3342
3343 Z Separator
3344 Zl Line separator
3345 Zp Paragraph separator
3346 Zs Space separator
3347
3348 The special property L& is also supported: it matches a character that
3349 has the Lu, Ll, or Lt property, in other words, a letter that is not
3350 classified as a modifier or "other".
3351
3352 The Cs (Surrogate) property applies only to characters in the range
3353 U+D800 to U+DFFF. Such characters are not valid in UTF-8 strings (see
3354 RFC 3629) and so cannot be tested by PCRE, unless UTF-8 validity check-
3355 ing has been turned off (see the discussion of PCRE_NO_UTF8_CHECK in
3356 the pcreapi page).
3357
3358 The long synonyms for these properties that Perl supports (such as
3359 \p{Letter}) are not supported by PCRE, nor is it permitted to prefix
3360 any of these properties with "Is".
3361
3362 No character that is in the Unicode table has the Cn (unassigned) prop-
3363 erty. Instead, this property is assumed for any code point that is not
3364 in the Unicode table.
3365
3366 Specifying caseless matching does not affect these escape sequences.
3367 For example, \p{Lu} always matches only upper case letters.
3368
3369 The \X escape matches any number of Unicode characters that form an
3370 extended Unicode sequence. \X is equivalent to
3371
3372 (?>\PM\pM*)
3373
3374 That is, it matches a character without the "mark" property, followed
3375 by zero or more characters with the "mark" property, and treats the
3376 sequence as an atomic group (see below). Characters with the "mark"
3377 property are typically accents that affect the preceding character.
3378 None of them have codepoints less than 256, so in non-UTF-8 mode \X
3379 matches any one character.
3380
3381 Matching characters by Unicode property is not fast, because PCRE has
3382 to search a structure that contains data for over fifteen thousand
3383 characters. That is why the traditional escape sequences such as \d and
3384 \w do not use Unicode properties in PCRE.
3385
3386 Resetting the match start
3387
3388 The escape sequence \K, which is a Perl 5.10 feature, causes any previ-
3389 ously matched characters not to be included in the final matched
3390 sequence. For example, the pattern:
3391
3392 foo\Kbar
3393
3394 matches "foobar", but reports that it has matched "bar". This feature
3395 is similar to a lookbehind assertion (described below). However, in
3396 this case, the part of the subject before the real match does not have
3397 to be of fixed length, as lookbehind assertions do. The use of \K does
3398 not interfere with the setting of captured substrings. For example,
3399 when the pattern
3400
3401 (foo)\Kbar
3402
3403 matches "foobar", the first substring is still set to "foo".
3404
3405 Simple assertions
3406
3407 The final use of backslash is for certain simple assertions. An asser-
3408 tion specifies a condition that has to be met at a particular point in
3409 a match, without consuming any characters from the subject string. The
3410 use of subpatterns for more complicated assertions is described below.
3411 The backslashed assertions are:
3412
3413 \b matches at a word boundary
3414 \B matches when not at a word boundary
3415 \A matches at the start of the subject
3416 \Z matches at the end of the subject
3417 also matches before a newline at the end of the subject
3418 \z matches only at the end of the subject
3419 \G matches at the first matching position in the subject
3420
3421 These assertions may not appear in character classes (but note that \b
3422 has a different meaning, namely the backspace character, inside a char-
3423 acter class).
3424
3425 A word boundary is a position in the subject string where the current
3426 character and the previous character do not both match \w or \W (i.e.
3427 one matches \w and the other matches \W), or the start or end of the
3428 string if the first or last character matches \w, respectively.
3429
3430 The \A, \Z, and \z assertions differ from the traditional circumflex
3431 and dollar (described in the next section) in that they only ever match
3432 at the very start and end of the subject string, whatever options are
3433 set. Thus, they are independent of multiline mode. These three asser-
3434 tions are not affected by the PCRE_NOTBOL or PCRE_NOTEOL options, which
3435 affect only the behaviour of the circumflex and dollar metacharacters.
3436 However, if the startoffset argument of pcre_exec() is non-zero, indi-
3437 cating that matching is to start at a point other than the beginning of
3438 the subject, \A can never match. The difference between \Z and \z is
3439 that \Z matches before a newline at the end of the string as well as at
3440 the very end, whereas \z matches only at the end.
3441
3442 The \G assertion is true only when the current matching position is at
3443 the start point of the match, as specified by the startoffset argument
3444 of pcre_exec(). It differs from \A when the value of startoffset is
3445 non-zero. By calling pcre_exec() multiple times with appropriate argu-
3446 ments, you can mimic Perl's /g option, and it is in this kind of imple-
3447 mentation where \G can be useful.
3448
3449 Note, however, that PCRE's interpretation of \G, as the start of the
3450 current match, is subtly different from Perl's, which defines it as the
3451 end of the previous match. In Perl, these can be different when the
3452 previously matched string was empty. Because PCRE does just one match
3453 at a time, it cannot reproduce this behaviour.
3454
3455 If all the alternatives of a pattern begin with \G, the expression is
3456 anchored to the starting match position, and the "anchored" flag is set
3457 in the compiled regular expression.
3458
3459
3460 CIRCUMFLEX AND DOLLAR
3461
3462 Outside a character class, in the default matching mode, the circumflex
3463 character is an assertion that is true only if the current matching
3464 point is at the start of the subject string. If the startoffset argu-
3465 ment of pcre_exec() is non-zero, circumflex can never match if the
3466 PCRE_MULTILINE option is unset. Inside a character class, circumflex
3467 has an entirely different meaning (see below).
3468
3469 Circumflex need not be the first character of the pattern if a number
3470 of alternatives are involved, but it should be the first thing in each
3471 alternative in which it appears if the pattern is ever to match that
3472 branch. If all possible alternatives start with a circumflex, that is,
3473 if the pattern is constrained to match only at the start of the sub-
3474 ject, it is said to be an "anchored" pattern. (There are also other
3475 constructs that can cause a pattern to be anchored.)
3476
3477 A dollar character is an assertion that is true only if the current
3478 matching point is at the end of the subject string, or immediately
3479 before a newline at the end of the string (by default). Dollar need not
3480 be the last character of the pattern if a number of alternatives are
3481 involved, but it should be the last item in any branch in which it
3482 appears. Dollar has no special meaning in a character class.
3483
3484 The meaning of dollar can be changed so that it matches only at the
3485 very end of the string, by setting the PCRE_DOLLAR_ENDONLY option at
3486 compile time. This does not affect the \Z assertion.
3487
3488 The meanings of the circumflex and dollar characters are changed if the
3489 PCRE_MULTILINE option is set. When this is the case, a circumflex
3490 matches immediately after internal newlines as well as at the start of
3491 the subject string. It does not match after a newline that ends the
3492 string. A dollar matches before any newlines in the string, as well as
3493 at the very end, when PCRE_MULTILINE is set. When newline is specified
3494 as the two-character sequence CRLF, isolated CR and LF characters do
3495 not indicate newlines.
3496
3497 For example, the pattern /^abc$/ matches the subject string "def\nabc"
3498 (where \n represents a newline) in multiline mode, but not otherwise.
3499 Consequently, patterns that are anchored in single line mode because
3500 all branches start with ^ are not anchored in multiline mode, and a
3501 match for circumflex is possible when the startoffset argument of
3502 pcre_exec() is non-zero. The PCRE_DOLLAR_ENDONLY option is ignored if
3503 PCRE_MULTILINE is set.
3504
3505 Note that the sequences \A, \Z, and \z can be used to match the start
3506 and end of the subject in both modes, and if all branches of a pattern
3507 start with \A it is always anchored, whether or not PCRE_MULTILINE is
3508 set.
3509
3510
3511 FULL STOP (PERIOD, DOT)
3512
3513 Outside a character class, a dot in the pattern matches any one charac-
3514 ter in the subject string except (by default) a character that signi-
3515 fies the end of a line. In UTF-8 mode, the matched character may be
3516 more than one byte long.
3517
3518 When a line ending is defined as a single character, dot never matches
3519 that character; when the two-character sequence CRLF is used, dot does
3520 not match CR if it is immediately followed by LF, but otherwise it
3521 matches all characters (including isolated CRs and LFs). When any Uni-
3522 code line endings are being recognized, dot does not match CR or LF or
3523 any of the other line ending characters.
3524
3525 The behaviour of dot with regard to newlines can be changed. If the
3526 PCRE_DOTALL option is set, a dot matches any one character, without
3527 exception. If the two-character sequence CRLF is present in the subject
3528 string, it takes two dots to match it.
3529
3530 The handling of dot is entirely independent of the handling of circum-
3531 flex and dollar, the only relationship being that they both involve
3532 newlines. Dot has no special meaning in a character class.
3533
3534
3535 MATCHING A SINGLE BYTE
3536
3537 Outside a character class, the escape sequence \C matches any one byte,
3538 both in and out of UTF-8 mode. Unlike a dot, it always matches any
3539 line-ending characters. The feature is provided in Perl in order to
3540 match individual bytes in UTF-8 mode. Because it breaks up UTF-8 char-
3541 acters into individual bytes, what remains in the string may be a mal-
3542 formed UTF-8 string. For this reason, the \C escape sequence is best
3543 avoided.
3544
3545 PCRE does not allow \C to appear in lookbehind assertions (described
3546 below), because in UTF-8 mode this would make it impossible to calcu-
3547 late the length of the lookbehind.
3548
3549
3550 SQUARE BRACKETS AND CHARACTER CLASSES
3551
3552 An opening square bracket introduces a character class, terminated by a
3553 closing square bracket. A closing square bracket on its own is not spe-
3554 cial. If a closing square bracket is required as a member of the class,
3555 it should be the first data character in the class (after an initial
3556 circumflex, if present) or escaped with a backslash.
3557
3558 A character class matches a single character in the subject. In UTF-8
3559 mode, the character may occupy more than one byte. A matched character
3560 must be in the set of characters defined by the class, unless the first
3561 character in the class definition is a circumflex, in which case the
3562 subject character must not be in the set defined by the class. If a
3563 circumflex is actually required as a member of the class, ensure it is
3564 not the first character, or escape it with a backslash.
3565
3566 For example, the character class [aeiou] matches any lower case vowel,
3567 while [^aeiou] matches any character that is not a lower case vowel.
3568 Note that a circumflex is just a convenient notation for specifying the
3569 characters that are in the class by enumerating those that are not. A
3570 class that starts with a circumflex is not an assertion: it still con-
3571 sumes a character from the subject string, and therefore it fails if
3572 the current pointer is at the end of the string.
3573
3574 In UTF-8 mode, characters with values greater than 255 can be included
3575 in a class as a literal string of bytes, or by using the \x{ escaping
3576 mechanism.
3577
3578 When caseless matching is set, any letters in a class represent both
3579 their upper case and lower case versions, so for example, a caseless
3580 [aeiou] matches "A" as well as "a", and a caseless [^aeiou] does not
3581 match "A", whereas a caseful version would. In UTF-8 mode, PCRE always
3582 understands the concept of case for characters whose values are less
3583 than 128, so caseless matching is always possible. For characters with
3584 higher values, the concept of case is supported if PCRE is compiled
3585 with Unicode property support, but not otherwise. If you want to use
3586 caseless matching for characters 128 and above, you must ensure that
3587 PCRE is compiled with Unicode property support as well as with UTF-8
3588 support.
3589
3590 Characters that might indicate line breaks are never treated in any
3591 special way when matching character classes, whatever line-ending
3592 sequence is in use, and whatever setting of the PCRE_DOTALL and
3593 PCRE_MULTILINE options is used. A class such as [^a] always matches one
3594 of these characters.
3595
3596 The minus (hyphen) character can be used to specify a range of charac-
3597 ters in a character class. For example, [d-m] matches any letter
3598 between d and m, inclusive. If a minus character is required in a
3599 class, it must be escaped with a backslash or appear in a position
3600 where it cannot be interpreted as indicating a range, typically as the
3601 first or last character in the class.
3602
3603 It is not possible to have the literal character "]" as the end charac-
3604 ter of a range. A pattern such as [W-]46] is interpreted as a class of
3605 two characters ("W" and "-") followed by a literal string "46]", so it
3606 would match "W46]" or "-46]". However, if the "]" is escaped with a
3607 backslash it is interpreted as the end of range, so [W-\]46] is inter-
3608 preted as a class containing a range followed by two other characters.
3609 The octal or hexadecimal representation of "]" can also be used to end
3610 a range.
3611
3612 Ranges operate in the collating sequence of character values. They can
3613 also be used for characters specified numerically, for example
3614 [\000-\037]. In UTF-8 mode, ranges can include characters whose values
3615 are greater than 255, for example [\x{100}-\x{2ff}].
3616
3617 If a range that includes letters is used when caseless matching is set,
3618 it matches the letters in either case. For example, [W-c] is equivalent
3619 to [][\\^_`wxyzabc], matched caselessly, and in non-UTF-8 mode, if
3620 character tables for a French locale are in use, [\xc8-\xcb] matches
3621 accented E characters in both cases. In UTF-8 mode, PCRE supports the
3622 concept of case for characters with values greater than 128 only when
3623 it is compiled with Unicode property support.
3624
3625 The character types \d, \D, \p, \P, \s, \S, \w, and \W may also appear
3626 in a character class, and add the characters that they match to the
3627 class. For example, [\dABCDEF] matches any hexadecimal digit. A circum-
3628 flex can conveniently be used with the upper case character types to
3629 specify a more restricted set of characters than the matching lower
3630 case type. For example, the class [^\W_] matches any letter or digit,
3631 but not underscore.
3632
3633 The only metacharacters that are recognized in character classes are
3634 backslash, hyphen (only where it can be interpreted as specifying a
3635 range), circumflex (only at the start), opening square bracket (only
3636 when it can be interpreted as introducing a POSIX class name - see the
3637 next section), and the terminating closing square bracket. However,
3638 escaping other non-alphanumeric characters does no harm.
3639
3640
3641 POSIX CHARACTER CLASSES
3642
3643 Perl supports the POSIX notation for character classes. This uses names
3644 enclosed by [: and :] within the enclosing square brackets. PCRE also
3645 supports this notation. For example,
3646
3647 [01[:alpha:]%]
3648
3649 matches "0", "1", any alphabetic character, or "%". The supported class
3650 names are
3651
3652 alnum letters and digits
3653 alpha letters
3654 ascii character codes 0 - 127
3655 blank space or tab only
3656 cntrl control characters
3657 digit decimal digits (same as \d)
3658 graph printing characters, excluding space
3659 lower lower case letters
3660 print printing characters, including space
3661 punct printing characters, excluding letters and digits
3662 space white space (not quite the same as \s)
3663 upper upper case letters
3664 word "word" characters (same as \w)
3665 xdigit hexadecimal digits
3666
3667 The "space" characters are HT (9), LF (10), VT (11), FF (12), CR (13),
3668 and space (32). Notice that this list includes the VT character (code
3669 11). This makes "space" different to \s, which does not include VT (for
3670 Perl compatibility).
3671
3672 The name "word" is a Perl extension, and "blank" is a GNU extension
3673 from Perl 5.8. Another Perl extension is negation, which is indicated
3674 by a ^ character after the colon. For example,
3675
3676 [12[:^digit:]]
3677
3678 matches "1", "2", or any non-digit. PCRE (and Perl) also recognize the
3679 POSIX syntax [.ch.] and [=ch=] where "ch" is a "collating element", but
3680 these are not supported, and an error is given if they are encountered.
3681
3682 In UTF-8 mode, characters with values greater than 128 do not match any
3683 of the POSIX character classes.
3684
3685
3686 VERTICAL BAR
3687
3688 Vertical bar characters are used to separate alternative patterns. For
3689 example, the pattern
3690
3691 gilbert|sullivan
3692
3693 matches either "gilbert" or "sullivan". Any number of alternatives may
3694 appear, and an empty alternative is permitted (matching the empty
3695 string). The matching process tries each alternative in turn, from left
3696 to right, and the first one that succeeds is used. If the alternatives
3697 are within a subpattern (defined below), "succeeds" means matching the
3698 rest of the main pattern as well as the alternative in the subpattern.
3699
3700
3701 INTERNAL OPTION SETTING
3702
3703 The settings of the PCRE_CASELESS, PCRE_MULTILINE, PCRE_DOTALL, and
3704 PCRE_EXTENDED options (which are Perl-compatible) can be changed from
3705 within the pattern by a sequence of Perl option letters enclosed
3706 between "(?" and ")". The option letters are
3707
3708 i for PCRE_CASELESS
3709 m for PCRE_MULTILINE
3710 s for PCRE_DOTALL
3711 x for PCRE_EXTENDED
3712
3713 For example, (?im) sets caseless, multiline matching. It is also possi-
3714 ble to unset these options by preceding the letter with a hyphen, and a
3715 combined setting and unsetting such as (?im-sx), which sets PCRE_CASE-
3716 LESS and PCRE_MULTILINE while unsetting PCRE_DOTALL and PCRE_EXTENDED,
3717 is also permitted. If a letter appears both before and after the
3718 hyphen, the option is unset.
3719
3720 The PCRE-specific options PCRE_DUPNAMES, PCRE_UNGREEDY, and PCRE_EXTRA
3721 can be changed in the same way as the Perl-compatible options by using
3722 the characters J, U and X respectively.
3723
3724 When an option change occurs at top level (that is, not inside subpat-
3725 tern parentheses), the change applies to the remainder of the pattern
3726 that follows. If the change is placed right at the start of a pattern,
3727 PCRE extracts it into the global options (and it will therefore show up
3728 in data extracted by the pcre_fullinfo() function).
3729
3730 An option change within a subpattern (see below for a description of
3731 subpatterns) affects only that part of the current pattern that follows
3732 it, so
3733
3734 (a(?i)b)c
3735
3736 matches abc and aBc and no other strings (assuming PCRE_CASELESS is not
3737 used). By this means, options can be made to have different settings
3738 in different parts of the pattern. Any changes made in one alternative
3739 do carry on into subsequent branches within the same subpattern. For
3740 example,
3741
3742 (a(?i)b|c)
3743
3744 matches "ab", "aB", "c", and "C", even though when matching "C" the
3745 first branch is abandoned before the option setting. This is because
3746 the effects of option settings happen at compile time. There would be
3747 some very weird behaviour otherwise.
3748
3749 Note: There are other PCRE-specific options that can be set by the
3750 application when the compile or match functions are called. In some
3751 cases the pattern can contain special leading sequences to override
3752 what the application has set or what has been defaulted. Details are
3753 given in the section entitled "Newline sequences" above.
3754
3755
3756 SUBPATTERNS
3757
3758 Subpatterns are delimited by parentheses (round brackets), which can be
3759 nested. Turning part of a pattern into a subpattern does two things:
3760
3761 1. It localizes a set of alternatives. For example, the pattern
3762
3763 cat(aract|erpillar|)
3764
3765 matches one of the words "cat", "cataract", or "caterpillar". Without
3766 the parentheses, it would match "cataract", "erpillar" or an empty
3767 string.
3768
3769 2. It sets up the subpattern as a capturing subpattern. This means
3770 that, when the whole pattern matches, that portion of the subject
3771 string that matched the subpattern is passed back to the caller via the
3772 ovector argument of pcre_exec(). Opening parentheses are counted from
3773 left to right (starting from 1) to obtain numbers for the capturing
3774 subpatterns.
3775
3776 For example, if the string "the red king" is matched against the pat-
3777 tern
3778
3779 the ((red|white) (king|queen))
3780
3781 the captured substrings are "red king", "red", and "king", and are num-
3782 bered 1, 2, and 3, respectively.
3783
3784 The fact that plain parentheses fulfil two functions is not always
3785 helpful. There are often times when a grouping subpattern is required
3786 without a capturing requirement. If an opening parenthesis is followed
3787 by a question mark and a colon, the subpattern does not do any captur-
3788 ing, and is not counted when computing the number of any subsequent
3789 capturing subpatterns. For example, if the string "the white queen" is
3790 matched against the pattern
3791
3792 the ((?:red|white) (king|queen))
3793
3794 the captured substrings are "white queen" and "queen", and are numbered
3795 1 and 2. The maximum number of capturing subpatterns is 65535.
3796
3797 As a convenient shorthand, if any option settings are required at the
3798 start of a non-capturing subpattern, the option letters may appear
3799 between the "?" and the ":". Thus the two patterns
3800
3801 (?i:saturday|sunday)
3802 (?:(?i)saturday|sunday)
3803
3804 match exactly the same set of strings. Because alternative branches are
3805 tried from left to right, and options are not reset until the end of
3806 the subpattern is reached, an option setting in one branch does affect
3807 subsequent branches, so the above patterns match "SUNDAY" as well as
3808 "Saturday".
3809
3810
3811 DUPLICATE SUBPATTERN NUMBERS
3812
3813 Perl 5.10 introduced a feature whereby each alternative in a subpattern
3814 uses the same numbers for its capturing parentheses. Such a subpattern
3815 starts with (?| and is itself a non-capturing subpattern. For example,
3816 consider this pattern:
3817
3818 (?|(Sat)ur|(Sun))day
3819
3820 Because the two alternatives are inside a (?| group, both sets of cap-
3821 turing parentheses are numbered one. Thus, when the pattern matches,
3822 you can look at captured substring number one, whichever alternative
3823 matched. This construct is useful when you want to capture part, but
3824 not all, of one of a number of alternatives. Inside a (?| group, paren-
3825 theses are numbered as usual, but the number is reset at the start of
3826 each branch. The numbers of any capturing buffers that follow the sub-
3827 pattern start after the highest number used in any branch. The follow-
3828 ing example is taken from the Perl documentation. The numbers under-
3829 neath show in which buffer the captured content will be stored.
3830
3831 # before ---------------branch-reset----------- after
3832 / ( a ) (?| x ( y ) z | (p (q) r) | (t) u (v) ) ( z ) /x
3833 # 1 2 2 3 2 3 4
3834
3835 A backreference or a recursive call to a numbered subpattern always
3836 refers to the first one in the pattern with the given number.
3837
3838 An alternative approach to using this "branch reset" feature is to use
3839 duplicate named subpatterns, as described in the next section.
3840
3841
3842 NAMED SUBPATTERNS
3843
3844 Identifying capturing parentheses by number is simple, but it can be
3845 very hard to keep track of the numbers in complicated regular expres-
3846 sions. Furthermore, if an expression is modified, the numbers may
3847 change. To help with this difficulty, PCRE supports the naming of sub-
3848 patterns. This feature was not added to Perl until release 5.10. Python
3849 had the feature earlier, and PCRE introduced it at release 4.0, using
3850 the Python syntax. PCRE now supports both the Perl and the Python syn-
3851 tax.
3852
3853 In PCRE, a subpattern can be named in one of three ways: (?<name>...)
3854 or (?'name'...) as in Perl, or (?P<name>...) as in Python. References
3855 to capturing parentheses from other parts of the pattern, such as back-
3856 references, recursion, and conditions, can be made by name as well as
3857 by number.
3858
3859 Names consist of up to 32 alphanumeric characters and underscores.
3860 Named capturing parentheses are still allocated numbers as well as
3861 names, exactly as if the names were not present. The PCRE API provides
3862 function calls for extracting the name-to-number translation table from
3863 a compiled pattern. There is also a convenience function for extracting
3864 a captured substring by name.
3865
3866 By default, a name must be unique within a pattern, but it is possible
3867 to relax this constraint by setting the PCRE_DUPNAMES option at compile
3868 time. This can be useful for patterns where only one instance of the
3869 named parentheses can match. Suppose you want to match the name of a
3870 weekday, either as a 3-letter abbreviation or as the full name, and in
3871 both cases you want to extract the abbreviation. This pattern (ignoring
3872 the line breaks) does the job:
3873
3874 (?<DN>Mon|Fri|Sun)(?:day)?|
3875 (?<DN>Tue)(?:sday)?|
3876 (?<DN>Wed)(?:nesday)?|
3877 (?<DN>Thu)(?:rsday)?|
3878 (?<DN>Sat)(?:urday)?
3879
3880 There are five capturing substrings, but only one is ever set after a
3881 match. (An alternative way of solving this problem is to use a "branch
3882 reset" subpattern, as described in the previous section.)
3883
3884 The convenience function for extracting the data by name returns the
3885 substring for the first (and in this example, the only) subpattern of
3886 that name that matched. This saves searching to find which numbered
3887 subpattern it was. If you make a reference to a non-unique named sub-
3888 pattern from elsewhere in the pattern, the one that corresponds to the
3889 lowest number is used. For further details of the interfaces for han-
3890 dling named subpatterns, see the pcreapi documentation.
3891
3892
3893 REPETITION
3894
3895 Repetition is specified by quantifiers, which can follow any of the
3896 following items:
3897
3898 a literal data character
3899 the dot metacharacter
3900 the \C escape sequence
3901 the \X escape sequence (in UTF-8 mode with Unicode properties)
3902 the \R escape sequence
3903 an escape such as \d that matches a single character
3904 a character class
3905 a back reference (see next section)
3906 a parenthesized subpattern (unless it is an assertion)
3907
3908 The general repetition quantifier specifies a minimum and maximum num-
3909 ber of permitted matches, by giving the two numbers in curly brackets
3910 (braces), separated by a comma. The numbers must be less than 65536,
3911 and the first must be less than or equal to the second. For example:
3912
3913 z{2,4}
3914
3915 matches "zz", "zzz", or "zzzz". A closing brace on its own is not a
3916 special character. If the second number is omitted, but the comma is
3917 present, there is no upper limit; if the second number and the comma
3918 are both omitted, the quantifier specifies an exact number of required
3919 matches. Thus
3920
3921 [aeiou]{3,}
3922
3923 matches at least 3 successive vowels, but may match many more, while
3924
3925 \d{8}
3926
3927 matches exactly 8 digits. An opening curly bracket that appears in a
3928 position where a quantifier is not allowed, or one that does not match
3929 the syntax of a quantifier, is taken as a literal character. For exam-
3930 ple, {,6} is not a quantifier, but a literal string of four characters.
3931
3932 In UTF-8 mode, quantifiers apply to UTF-8 characters rather than to
3933 individual bytes. Thus, for example, \x{100}{2} matches two UTF-8 char-
3934 acters, each of which is represented by a two-byte sequence. Similarly,
3935 when Unicode property support is available, \X{3} matches three Unicode
3936 extended sequences, each of which may be several bytes long (and they
3937 may be of different lengths).
3938
3939 The quantifier {0} is permitted, causing the expression to behave as if
3940 the previous item and the quantifier were not present.
3941
3942 For convenience, the three most common quantifiers have single-charac-
3943 ter abbreviations:
3944
3945 * is equivalent to {0,}
3946 + is equivalent to {1,}
3947 ? is equivalent to {0,1}
3948
3949 It is possible to construct infinite loops by following a subpattern
3950 that can match no characters with a quantifier that has no upper limit,
3951 for example:
3952
3953 (a?)*
3954
3955 Earlier versions of Perl and PCRE used to give an error at compile time
3956 for such patterns. However, because there are cases where this can be
3957 useful, such patterns are now accepted, but if any repetition of the
3958 subpattern does in fact match no characters, the loop is forcibly bro-
3959 ken.
3960
3961 By default, the quantifiers are "greedy", that is, they match as much
3962 as possible (up to the maximum number of permitted times), without
3963 causing the rest of the pattern to fail. The classic example of where
3964 this gives problems is in trying to match comments in C programs. These
3965 appear between /* and */ and within the comment, individual * and /
3966 characters may appear. An attempt to match C comments by applying the
3967 pattern
3968
3969 /\*.*\*/
3970
3971 to the string
3972
3973 /* first comment */ not comment /* second comment */
3974
3975 fails, because it matches the entire string owing to the greediness of
3976 the .* item.
3977
3978 However, if a quantifier is followed by a question mark, it ceases to
3979 be greedy, and instead matches the minimum number of times possible, so
3980 the pattern
3981
3982 /\*.*?\*/
3983
3984 does the right thing with the C comments. The meaning of the various
3985 quantifiers is not otherwise changed, just the preferred number of
3986 matches. Do not confuse this use of question mark with its use as a
3987 quantifier in its own right. Because it has two uses, it can sometimes
3988 appear doubled, as in
3989
3990 \d??\d
3991
3992 which matches one digit by preference, but can match two if that is the
3993 only way the rest of the pattern matches.
3994
3995 If the PCRE_UNGREEDY option is set (an option that is not available in
3996 Perl), the quantifiers are not greedy by default, but individual ones
3997 can be made greedy by following them with a question mark. In other
3998 words, it inverts the default behaviour.
3999
4000 When a parenthesized subpattern is quantified with a minimum repeat
4001 count that is greater than 1 or with a limited maximum, more memory is
4002 required for the compiled pattern, in proportion to the size of the
4003 minimum or maximum.
4004
4005 If a pattern starts with .* or .{0,} and the PCRE_DOTALL option (equiv-
4006 alent to Perl's /s) is set, thus allowing the dot to match newlines,
4007 the pattern is implicitly anchored, because whatever follows will be
4008 tried against every character position in the subject string, so there
4009 is no point in retrying the overall match at any position after the
4010 first. PCRE normally treats such a pattern as though it were preceded
4011 by \A.
4012
4013 In cases where it is known that the subject string contains no new-
4014 lines, it is worth setting PCRE_DOTALL in order to obtain this opti-
4015 mization, or alternatively using ^ to indicate anchoring explicitly.
4016
4017 However, there is one situation where the optimization cannot be used.
4018 When .* is inside capturing parentheses that are the subject of a
4019 backreference elsewhere in the pattern, a match at the start may fail
4020 where a later one succeeds. Consider, for example:
4021
4022 (.*)abc\1
4023
4024 If the subject is "xyz123abc123" the match point is the fourth charac-
4025 ter. For this reason, such a pattern is not implicitly anchored.
4026
4027 When a capturing subpattern is repeated, the value captured is the sub-
4028 string that matched the final iteration. For example, after
4029
4030 (tweedle[dume]{3}\s*)+
4031
4032 has matched "tweedledum tweedledee" the value of the captured substring
4033 is "tweedledee". However, if there are nested capturing subpatterns,
4034 the corresponding captured values may have been set in previous itera-
4035 tions. For example, after
4036
4037 /(a|(b))+/
4038
4039 matches "aba" the value of the second captured substring is "b".
4040
4041
4042 ATOMIC GROUPING AND POSSESSIVE QUANTIFIERS
4043
4044 With both maximizing ("greedy") and minimizing ("ungreedy" or "lazy")
4045 repetition, failure of what follows normally causes the repeated item
4046 to be re-evaluated to see if a different number of repeats allows the
4047 rest of the pattern to match. Sometimes it is useful to prevent this,
4048 either to change the nature of the match, or to cause it fail earlier
4049 than it otherwise might, when the author of the pattern knows there is
4050 no point in carrying on.
4051
4052 Consider, for example, the pattern \d+foo when applied to the subject
4053 line
4054
4055 123456bar
4056
4057 After matching all 6 digits and then failing to match "foo", the normal
4058 action of the matcher is to try again with only 5 digits matching the
4059 \d+ item, and then with 4, and so on, before ultimately failing.
4060 "Atomic grouping" (a term taken from Jeffrey Friedl's book) provides
4061 the means for specifying that once a subpattern has matched, it is not
4062 to be re-evaluated in this way.
4063
4064 If we use atomic grouping for the previous example, the matcher gives
4065 up immediately on failing to match "foo" the first time. The notation
4066 is a kind of special parenthesis, starting with (?> as in this example:
4067
4068 (?>\d+)foo
4069
4070 This kind of parenthesis "locks up" the part of the pattern it con-
4071 tains once it has matched, and a failure further into the pattern is
4072 prevented from backtracking into it. Backtracking past it to previous
4073 items, however, works as normal.
4074
4075 An alternative description is that a subpattern of this type matches
4076 the string of characters that an identical standalone pattern would
4077 match, if anchored at the current point in the subject string.
4078
4079 Atomic grouping subpatterns are not capturing subpatterns. Simple cases
4080 such as the above example can be thought of as a maximizing repeat that
4081 must swallow everything it can. So, while both \d+ and \d+? are pre-
4082 pared to adjust the number of digits they match in order to make the
4083 rest of the pattern match, (?>\d+) can only match an entire sequence of
4084 digits.
4085
4086 Atomic groups in general can of course contain arbitrarily complicated
4087 subpatterns, and can be nested. However, when the subpattern for an
4088 atomic group is just a single repeated item, as in the example above, a
4089 simpler notation, called a "possessive quantifier" can be used. This
4090 consists of an additional + character following a quantifier. Using
4091 this notation, the previous example can be rewritten as
4092
4093 \d++foo
4094
4095 Note that a possessive quantifier can be used with an entire group, for
4096 example:
4097
4098 (abc|xyz){2,3}+
4099
4100 Possessive quantifiers are always greedy; the setting of the
4101 PCRE_UNGREEDY option is ignored. They are a convenient notation for the
4102 simpler forms of atomic group. However, there is no difference in the
4103 meaning of a possessive quantifier and the equivalent atomic group,
4104 though there may be a performance difference; possessive quantifiers
4105 should be slightly faster.
4106
4107 The possessive quantifier syntax is an extension to the Perl 5.8 syn-
4108 tax. Jeffrey Friedl originated the idea (and the name) in the first
4109 edition of his book. Mike McCloskey liked it, so implemented it when he
4110 built Sun's Java package, and PCRE copied it from there. It ultimately
4111 found its way into Perl at release 5.10.
4112
4113 PCRE has an optimization that automatically "possessifies" certain sim-
4114 ple pattern constructs. For example, the sequence A+B is treated as
4115 A++B because there is no point in backtracking into a sequence of A's
4116 when B must follow.
4117
4118 When a pattern contains an unlimited repeat inside a subpattern that
4119 can itself be repeated an unlimited number of times, the use of an
4120 atomic group is the only way to avoid some failing matches taking a
4121 very long time indeed. The pattern
4122
4123 (\D+|<\d+>)*[!?]
4124
4125 matches an unlimited number of substrings that either consist of non-
4126 digits, or digits enclosed in <>, followed by either ! or ?. When it
4127 matches, it runs quickly. However, if it is applied to
4128
4129 aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
4130
4131 it takes a long time before reporting failure. This is because the
4132 string can be divided between the internal \D+ repeat and the external
4133 * repeat in a large number of ways, and all have to be tried. (The
4134 example uses [!?] rather than a single character at the end, because
4135 both PCRE and Perl have an optimization that allows for fast failure
4136 when a single character is used. They remember the last single charac-
4137 ter that is required for a match, and fail early if it is not present
4138 in the string.) If the pattern is changed so that it uses an atomic
4139 group, like this:
4140
4141 ((?>\D+)|<\d+>)*[!?]
4142
4143 sequences of non-digits cannot be broken, and failure happens quickly.
4144
4145
4146 BACK REFERENCES
4147
4148 Outside a character class, a backslash followed by a digit greater than
4149 0 (and possibly further digits) is a back reference to a capturing sub-
4150 pattern earlier (that is, to its left) in the pattern, provided there
4151 have been that many previous capturing left parentheses.
4152
4153 However, if the decimal number following the backslash is less than 10,
4154 it is always taken as a back reference, and causes an error only if
4155 there are not that many capturing left parentheses in the entire pat-
4156 tern. In other words, the parentheses that are referenced need not be
4157 to the left of the reference for numbers less than 10. A "forward back
4158 reference" of this type can make sense when a repetition is involved
4159 and the subpattern to the right has participated in an earlier itera-
4160 tion.
4161
4162 It is not possible to have a numerical "forward back reference" to a
4163 subpattern whose number is 10 or more using this syntax because a
4164 sequence such as \50 is interpreted as a character defined in octal.
4165 See the subsection entitled "Non-printing characters" above for further
4166 details of the handling of digits following a backslash. There is no
4167 such problem when named parentheses are used. A back reference to any
4168 subpattern is possible using named parentheses (see below).
4169
4170 Another way of avoiding the ambiguity inherent in the use of digits
4171 following a backslash is to use the \g escape sequence, which is a fea-
4172 ture introduced in Perl 5.10. This escape must be followed by an
4173 unsigned number or a negative number, optionally enclosed in braces.
4174 These examples are all identical:
4175
4176 (ring), \1
4177 (ring), \g1
4178 (ring), \g{1}
4179
4180 An unsigned number specifies an absolute reference without the ambigu-
4181 ity that is present in the older syntax. It is also useful when literal
4182 digits follow the reference. A negative number is a relative reference.
4183 Consider this example:
4184
4185 (abc(def)ghi)\g{-1}
4186
4187 The sequence \g{-1} is a reference to the most recently started captur-
4188 ing subpattern before \g, that is, is it equivalent to \2. Similarly,
4189 \g{-2} would be equivalent to \1. The use of relative references can be
4190 helpful in long patterns, and also in patterns that are created by
4191 joining together fragments that contain references within themselves.
4192
4193 A back reference matches whatever actually matched the capturing sub-
4194 pattern in the current subject string, rather than anything matching
4195 the subpattern itself (see "Subpatterns as subroutines" below for a way
4196 of doing that). So the pattern
4197
4198 (sens|respons)e and \1ibility
4199
4200 matches "sense and sensibility" and "response and responsibility", but
4201 not "sense and responsibility". If caseful matching is in force at the
4202 time of the back reference, the case of letters is relevant. For exam-
4203 ple,
4204
4205 ((?i)rah)\s+\1
4206
4207 matches "rah rah" and "RAH RAH", but not "RAH rah", even though the
4208 original capturing subpattern is matched caselessly.
4209
4210 There are several different ways of writing back references to named
4211 subpatterns. The .NET syntax \k{name} and the Perl syntax \k<name> or
4212 \k'name' are supported, as is the Python syntax (?P=name). Perl 5.10's
4213 unified back reference syntax, in which \g can be used for both numeric
4214 and named references, is also supported. We could rewrite the above
4215 example in any of the following ways:
4216
4217 (?<p1>(?i)rah)\s+\k<p1>
4218 (?'p1'(?i)rah)\s+\k{p1}
4219 (?P<p1>(?i)rah)\s+(?P=p1)
4220 (?<p1>(?i)rah)\s+\g{p1}
4221
4222 A subpattern that is referenced by name may appear in the pattern
4223 before or after the reference.
4224
4225 There may be more than one back reference to the same subpattern. If a
4226 subpattern has not actually been used in a particular match, any back
4227 references to it always fail. For example, the pattern
4228
4229 (a|(bc))\2
4230
4231 always fails if it starts to match "a" rather than "bc". Because there
4232 may be many capturing parentheses in a pattern, all digits following
4233 the backslash are taken as part of a potential back reference number.
4234 If the pattern continues with a digit character, some delimiter must be
4235 used to terminate the back reference. If the PCRE_EXTENDED option is
4236 set, this can be whitespace. Otherwise an empty comment (see "Com-
4237 ments" below) can be used.
4238
4239 A back reference that occurs inside the parentheses to which it refers
4240 fails when the subpattern is first used, so, for example, (a\1) never
4241 matches. However, such references can be useful inside repeated sub-
4242 patterns. For example, the pattern
4243
4244 (a|b\1)+
4245
4246 matches any number of "a"s and also "aba", "ababbaa" etc. At each iter-
4247 ation of the subpattern, the back reference matches the character
4248 string corresponding to the previous iteration. In order for this to
4249 work, the pattern must be such that the first iteration does not need
4250 to match the back reference. This can be done using alternation, as in
4251 the example above, or by a quantifier with a minimum of zero.
4252
4253
4254 ASSERTIONS
4255
4256 An assertion is a test on the characters following or preceding the
4257 current matching point that does not actually consume any characters.
4258 The simple assertions coded as \b, \B, \A, \G, \Z, \z, ^ and $ are
4259 described above.
4260
4261 More complicated assertions are coded as subpatterns. There are two
4262 kinds: those that look ahead of the current position in the subject
4263 string, and those that look behind it. An assertion subpattern is
4264 matched in the normal way, except that it does not cause the current
4265 matching position to be changed.
4266
4267 Assertion subpatterns are not capturing subpatterns, and may not be
4268 repeated, because it makes no sense to assert the same thing several
4269 times. If any kind of assertion contains capturing subpatterns within
4270 it, these are counted for the purposes of numbering the capturing sub-
4271 patterns in the whole pattern. However, substring capturing is carried
4272 out only for positive assertions, because it does not make sense for
4273 negative assertions.
4274
4275 Lookahead assertions
4276
4277 Lookahead assertions start with (?= for positive assertions and (?! for
4278 negative assertions. For example,
4279
4280 \w+(?=;)
4281
4282 matches a word followed by a semicolon, but does not include the semi-
4283 colon in the match, and
4284
4285 foo(?!bar)
4286
4287 matches any occurrence of "foo" that is not followed by "bar". Note
4288 that the apparently similar pattern
4289
4290 (?!foo)bar
4291
4292 does not find an occurrence of "bar" that is preceded by something
4293 other than "foo"; it finds any occurrence of "bar" whatsoever, because
4294 the assertion (?!foo) is always true when the next three characters are
4295 "bar". A lookbehind assertion is needed to achieve the other effect.
4296
4297 If you want to force a matching failure at some point in a pattern, the
4298 most convenient way to do it is with (?!) because an empty string
4299 always matches, so an assertion that requires there not to be an empty
4300 string must always fail.
4301
4302 Lookbehind assertions
4303
4304 Lookbehind assertions start with (?<= for positive assertions and (?<!
4305 for negative assertions. For example,
4306
4307 (?<!foo)bar
4308
4309 does find an occurrence of "bar" that is not preceded by "foo". The
4310 contents of a lookbehind assertion are restricted such that all the
4311 strings it matches must have a fixed length. However, if there are sev-
4312 eral top-level alternatives, they do not all have to have the same
4313 fixed length. Thus
4314
4315 (?<=bullock|donkey)
4316
4317 is permitted, but
4318
4319 (?<!dogs?|cats?)
4320
4321 causes an error at compile time. Branches that match different length
4322 strings are permitted only at the top level of a lookbehind assertion.
4323 This is an extension compared with Perl (at least for 5.8), which
4324 requires all branches to match the same length of string. An assertion
4325 such as
4326
4327 (?<=ab(c|de))
4328
4329 is not permitted, because its single top-level branch can match two
4330 different lengths, but it is acceptable if rewritten to use two top-
4331 level branches:
4332
4333 (?<=abc|abde)
4334
4335 In some cases, the Perl 5.10 escape sequence \K (see above) can be used
4336 instead of a lookbehind assertion; this is not restricted to a fixed-
4337 length.
4338
4339 The implementation of lookbehind assertions is, for each alternative,
4340 to temporarily move the current position back by the fixed length and
4341 then try to match. If there are insufficient characters before the cur-
4342 rent position, the assertion fails.
4343
4344 PCRE does not allow the \C escape (which matches a single byte in UTF-8
4345 mode) to appear in lookbehind assertions, because it makes it impossi-
4346 ble to calculate the length of the lookbehind. The \X and \R escapes,
4347 which can match different numbers of bytes, are also not permitted.
4348
4349 Possessive quantifiers can be used in conjunction with lookbehind
4350 assertions to specify efficient matching at the end of the subject
4351 string. Consider a simple pattern such as
4352
4353 abcd$
4354
4355 when applied to a long string that does not match. Because matching
4356 proceeds from left to right, PCRE will look for each "a" in the subject
4357 and then see if what follows matches the rest of the pattern. If the
4358 pattern is specified as
4359
4360 ^.*abcd$
4361
4362 the initial .* matches the entire string at first, but when this fails
4363 (because there is no following "a"), it backtracks to match all but the
4364 last character, then all but the last two characters, and so on. Once
4365 again the search for "a" covers the entire string, from right to left,
4366 so we are no better off. However, if the pattern is written as
4367
4368 ^.*+(?<=abcd)
4369
4370 there can be no backtracking for the .*+ item; it can match only the
4371 entire string. The subsequent lookbehind assertion does a single test
4372 on the last four characters. If it fails, the match fails immediately.
4373 For long strings, this approach makes a significant difference to the
4374 processing time.
4375
4376 Using multiple assertions
4377
4378 Several assertions (of any sort) may occur in succession. For example,
4379
4380 (?<=\d{3})(?<!999)foo
4381
4382 matches "foo" preceded by three digits that are not "999". Notice that
4383 each of the assertions is applied independently at the same point in
4384 the subject string. First there is a check that the previous three
4385 characters are all digits, and then there is a check that the same
4386 three characters are not "999". This pattern does not match "foo" pre-
4387 ceded by six characters, the first of which are digits and the last
4388 three of which are not "999". For example, it doesn't match "123abc-
4389 foo". A pattern to do that is
4390
4391 (?<=\d{3}...)(?<!999)foo
4392
4393 This time the first assertion looks at the preceding six characters,
4394 checking that the first three are digits, and then the second assertion
4395 checks that the preceding three characters are not "999".
4396
4397 Assertions can be nested in any combination. For example,
4398
4399 (?<=(?<!foo)bar)baz
4400
4401 matches an occurrence of "baz" that is preceded by "bar" which in turn
4402 is not preceded by "foo", while
4403
4404 (?<=\d{3}(?!999)...)foo
4405
4406 is another pattern that matches "foo" preceded by three digits and any
4407 three characters that are not "999".
4408
4409
4410 CONDITIONAL SUBPATTERNS
4411
4412 It is possible to cause the matching process to obey a subpattern con-
4413 ditionally or to choose between two alternative subpatterns, depending
4414 on the result of an assertion, or whether a previous capturing subpat-
4415 tern matched or not. The two possible forms of conditional subpattern
4416 are
4417
4418 (?(condition)yes-pattern)
4419 (?(condition)yes-pattern|no-pattern)
4420
4421 If the condition is satisfied, the yes-pattern is used; otherwise the
4422 no-pattern (if present) is used. If there are more than two alterna-
4423 tives in the subpattern, a compile-time error occurs.
4424
4425 There are four kinds of condition: references to subpatterns, refer-
4426 ences to recursion, a pseudo-condition called DEFINE, and assertions.
4427
4428 Checking for a used subpattern by number
4429
4430 If the text between the parentheses consists of a sequence of digits,
4431 the condition is true if the capturing subpattern of that number has
4432 previously matched. An alternative notation is to precede the digits
4433 with a plus or minus sign. In this case, the subpattern number is rela-
4434 tive rather than absolute. The most recently opened parentheses can be
4435 referenced by (?(-1), the next most recent by (?(-2), and so on. In
4436 looping constructs it can also make sense to refer to subsequent groups
4437 with constructs such as (?(+2).
4438
4439 Consider the following pattern, which contains non-significant white
4440 space to make it more readable (assume the PCRE_EXTENDED option) and to
4441 divide it into three parts for ease of discussion:
4442
4443 ( \( )? [^()]+ (?(1) \) )
4444
4445 The first part matches an optional opening parenthesis, and if that
4446 character is present, sets it as the first captured substring. The sec-
4447 ond part matches one or more characters that are not parentheses. The
4448 third part is a conditional subpattern that tests whether the first set
4449 of parentheses matched or not. If they did, that is, if subject started
4450 with an opening parenthesis, the condition is true, and so the yes-pat-
4451 tern is executed and a closing parenthesis is required. Otherwise,
4452 since no-pattern is not present, the subpattern matches nothing. In
4453 other words, this pattern matches a sequence of non-parentheses,
4454 optionally enclosed in parentheses.
4455
4456 If you were embedding this pattern in a larger one, you could use a
4457 relative reference:
4458
4459 ...other stuff... ( \( )? [^()]+ (?(-1) \) ) ...
4460
4461 This makes the fragment independent of the parentheses in the larger
4462 pattern.
4463
4464 Checking for a used subpattern by name
4465
4466 Perl uses the syntax (?(<name>)...) or (?('name')...) to test for a
4467 used subpattern by name. For compatibility with earlier versions of
4468 PCRE, which had this facility before Perl, the syntax (?(name)...) is
4469 also recognized. However, there is a possible ambiguity with this syn-
4470 tax, because subpattern names may consist entirely of digits. PCRE
4471 looks first for a named subpattern; if it cannot find one and the name
4472 consists entirely of digits, PCRE looks for a subpattern of that num-
4473 ber, which must be greater than zero. Using subpattern names that con-
4474 sist entirely of digits is not recommended.
4475
4476 Rewriting the above example to use a named subpattern gives this:
4477
4478 (?<OPEN> \( )? [^()]+ (?(<OPEN>) \) )
4479
4480
4481 Checking for pattern recursion
4482
4483 If the condition is the string (R), and there is no subpattern with the
4484 name R, the condition is true if a recursive call to the whole pattern
4485 or any subpattern has been made. If digits or a name preceded by amper-
4486 sand follow the letter R, for example:
4487
4488 (?(R3)...) or (?(R&name)...)
4489
4490 the condition is true if the most recent recursion is into the subpat-
4491 tern whose number or name is given. This condition does not check the
4492 entire recursion stack.
4493
4494 At "top level", all these recursion test conditions are false. Recur-
4495 sive patterns are described below.
4496
4497 Defining subpatterns for use by reference only
4498
4499 If the condition is the string (DEFINE), and there is no subpattern
4500 with the name DEFINE, the condition is always false. In this case,
4501 there may be only one alternative in the subpattern. It is always
4502 skipped if control reaches this point in the pattern; the idea of
4503 DEFINE is that it can be used to define "subroutines" that can be ref-
4504 erenced from elsewhere. (The use of "subroutines" is described below.)
4505 For example, a pattern to match an IPv4 address could be written like
4506 this (ignore whitespace and line breaks):
4507
4508 (?(DEFINE) (?<byte> 2[0-4]\d | 25[0-5] | 1\d\d | [1-9]?\d) )
4509 \b (?&byte) (\.(?&byte)){3} \b
4510
4511 The first part of the pattern is a DEFINE group inside which a another
4512 group named "byte" is defined. This matches an individual component of
4513 an IPv4 address (a number less than 256). When matching takes place,
4514 this part of the pattern is skipped because DEFINE acts like a false
4515 condition.
4516
4517 The rest of the pattern uses references to the named group to match the
4518 four dot-separated components of an IPv4 address, insisting on a word
4519 boundary at each end.
4520
4521 Assertion conditions
4522
4523 If the condition is not in any of the above formats, it must be an
4524 assertion. This may be a positive or negative lookahead or lookbehind
4525 assertion. Consider this pattern, again containing non-significant
4526 white space, and with the two alternatives on the second line:
4527
4528 (?(?=[^a-z]*[a-z])
4529 \d{2}-[a-z]{3}-\d{2} | \d{2}-\d{2}-\d{2} )
4530
4531 The condition is a positive lookahead assertion that matches an
4532 optional sequence of non-letters followed by a letter. In other words,
4533 it tests for the presence of at least one letter in the subject. If a
4534 letter is found, the subject is matched against the first alternative;
4535 otherwise it is matched against the second. This pattern matches
4536 strings in one of the two forms dd-aaa-dd or dd-dd-dd, where aaa are
4537 letters and dd are digits.
4538
4539
4540 COMMENTS
4541
4542 The sequence (?# marks the start of a comment that continues up to the
4543 next closing parenthesis. Nested parentheses are not permitted. The
4544 characters that make up a comment play no part in the pattern matching
4545 at all.
4546
4547 If the PCRE_EXTENDED option is set, an unescaped # character outside a
4548 character class introduces a comment that continues to immediately
4549 after the next newline in the pattern.
4550
4551
4552 RECURSIVE PATTERNS
4553
4554 Consider the problem of matching a string in parentheses, allowing for
4555 unlimited nested parentheses. Without the use of recursion, the best
4556 that can be done is to use a pattern that matches up to some fixed
4557 depth of nesting. It is not possible to handle an arbitrary nesting
4558 depth.
4559
4560 For some time, Perl has provided a facility that allows regular expres-
4561 sions to recurse (amongst other things). It does this by interpolating
4562 Perl code in the expression at run time, and the code can refer to the
4563 expression itself. A Perl pattern using code interpolation to solve the
4564 parentheses problem can be created like this:
4565
4566 $re = qr{\( (?: (?>[^()]+) | (?p{$re}) )* \)}x;
4567
4568 The (?p{...}) item interpolates Perl code at run time, and in this case
4569 refers recursively to the pattern in which it appears.
4570
4571 Obviously, PCRE cannot support the interpolation of Perl code. Instead,
4572 it supports special syntax for recursion of the entire pattern, and
4573 also for individual subpattern recursion. After its introduction in
4574 PCRE and Python, this kind of recursion was introduced into Perl at
4575 release 5.10.
4576
4577 A special item that consists of (? followed by a number greater than
4578 zero and a closing parenthesis is a recursive call of the subpattern of
4579 the given number, provided that it occurs inside that subpattern. (If
4580 not, it is a "subroutine" call, which is described in the next sec-
4581 tion.) The special item (?R) or (?0) is a recursive call of the entire
4582 regular expression.
4583
4584 In PCRE (like Python, but unlike Perl), a recursive subpattern call is
4585 always treated as an atomic group. That is, once it has matched some of
4586 the subject string, it is never re-entered, even if it contains untried
4587 alternatives and there is a subsequent matching failure.
4588
4589 This PCRE pattern solves the nested parentheses problem (assume the
4590 PCRE_EXTENDED option is set so that white space is ignored):
4591
4592 \( ( (?>[^()]+) | (?R) )* \)
4593
4594 First it matches an opening parenthesis. Then it matches any number of
4595 substrings which can either be a sequence of non-parentheses, or a
4596 recursive match of the pattern itself (that is, a correctly parenthe-
4597 sized substring). Finally there is a closing parenthesis.
4598
4599 If this were part of a larger pattern, you would not want to recurse
4600 the entire pattern, so instead you could use this:
4601
4602 ( \( ( (?>[^()]+) | (?1) )* \) )
4603
4604 We have put the pattern into parentheses, and caused the recursion to
4605 refer to them instead of the whole pattern.
4606
4607 In a larger pattern, keeping track of parenthesis numbers can be
4608 tricky. This is made easier by the use of relative references. (A Perl
4609 5.10 feature.) Instead of (?1) in the pattern above you can write
4610 (?-2) to refer to the second most recently opened parentheses preceding
4611 the recursion. In other words, a negative number counts capturing
4612 parentheses leftwards from the point at which it is encountered.
4613
4614 It is also possible to refer to subsequently opened parentheses, by
4615 writing references such as (?+2). However, these cannot be recursive
4616 because the reference is not inside the parentheses that are refer-
4617 enced. They are always "subroutine" calls, as described in the next
4618 section.
4619
4620 An alternative approach is to use named parentheses instead. The Perl
4621 syntax for this is (?&name); PCRE's earlier syntax (?P>name) is also
4622 supported. We could rewrite the above example as follows:
4623
4624 (?<pn> \( ( (?>[^()]+) | (?&pn) )* \) )
4625
4626 If there is more than one subpattern with the same name, the earliest
4627 one is used.
4628
4629 This particular example pattern that we have been looking at contains
4630 nested unlimited repeats, and so the use of atomic grouping for match-
4631 ing strings of non-parentheses is important when applying the pattern
4632 to strings that do not match. For example, when this pattern is applied
4633 to
4634
4635 (aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa()
4636
4637 it yields "no match" quickly. However, if atomic grouping is not used,
4638 the match runs for a very long time indeed because there are so many
4639 different ways the + and * repeats can carve up the subject, and all
4640 have to be tested before failure can be reported.
4641
4642 At the end of a match, the values set for any capturing subpatterns are
4643 those from the outermost level of the recursion at which the subpattern
4644 value is set. If you want to obtain intermediate values, a callout
4645 function can be used (see below and the pcrecallout documentation). If
4646 the pattern above is matched against
4647
4648 (ab(cd)ef)
4649
4650 the value for the capturing parentheses is "ef", which is the last
4651 value taken on at the top level. If additional parentheses are added,
4652 giving
4653
4654 \( ( ( (?>[^()]+) | (?R) )* ) \)
4655 ^ ^
4656 ^ ^
4657
4658 the string they capture is "ab(cd)ef", the contents of the top level
4659 parentheses. If there are more than 15 capturing parentheses in a pat-
4660 tern, PCRE has to obtain extra memory to store data during a recursion,
4661 which it does by using pcre_malloc, freeing it via pcre_free after-
4662 wards. If no memory can be obtained, the match fails with the
4663 PCRE_ERROR_NOMEMORY error.
4664
4665 Do not confuse the (?R) item with the condition (R), which tests for
4666 recursion. Consider this pattern, which matches text in angle brack-
4667 ets, allowing for arbitrary nesting. Only digits are allowed in nested
4668 brackets (that is, when recursing), whereas any characters are permit-
4669 ted at the outer level.
4670
4671 < (?: (?(R) \d++ | [^<>]*+) | (?R)) * >
4672
4673 In this pattern, (?(R) is the start of a conditional subpattern, with
4674 two different alternatives for the recursive and non-recursive cases.
4675 The (?R) item is the actual recursive call.
4676
4677
4678 SUBPATTERNS AS SUBROUTINES
4679
4680 If the syntax for a recursive subpattern reference (either by number or
4681 by name) is used outside the parentheses to which it refers, it oper-
4682 ates like a subroutine in a programming language. The "called" subpat-
4683 tern may be defined before or after the reference. A numbered reference
4684 can be absolute or relative, as in these examples:
4685
4686 (...(absolute)...)...(?2)...
4687 (...(relative)...)...(?-1)...
4688 (...(?+1)...(relative)...
4689
4690 An earlier example pointed out that the pattern
4691
4692 (sens|respons)e and \1ibility
4693
4694 matches "sense and sensibility" and "response and responsibility", but
4695 not "sense and responsibility". If instead the pattern
4696
4697 (sens|respons)e and (?1)ibility
4698
4699 is used, it does match "sense and responsibility" as well as the other
4700 two strings. Another example is given in the discussion of DEFINE
4701 above.
4702
4703 Like recursive subpatterns, a "subroutine" call is always treated as an
4704 atomic group. That is, once it has matched some of the subject string,
4705 it is never re-entered, even if it contains untried alternatives and
4706 there is a subsequent matching failure.
4707
4708 When a subpattern is used as a subroutine, processing options such as
4709 case-independence are fixed when the subpattern is defined. They cannot
4710 be changed for different calls. For example, consider this pattern:
4711
4712 (abc)(?i:(?-1))
4713
4714 It matches "abcabc". It does not match "abcABC" because the change of
4715 processing option does not affect the called subpattern.
4716
4717
4718 CALLOUTS
4719
4720 Perl has a feature whereby using the sequence (?{...}) causes arbitrary
4721 Perl code to be obeyed in the middle of matching a regular expression.
4722 This makes it possible, amongst other things, to extract different sub-
4723 strings that match the same pair of parentheses when there is a repeti-
4724 tion.
4725
4726 PCRE provides a similar feature, but of course it cannot obey arbitrary
4727 Perl code. The feature is called "callout". The caller of PCRE provides
4728 an external function by putting its entry point in the global variable
4729 pcre_callout. By default, this variable contains NULL, which disables
4730 all calling out.
4731
4732 Within a regular expression, (?C) indicates the points at which the
4733 external function is to be called. If you want to identify different
4734 callout points, you can put a number less than 256 after the letter C.
4735 The default value is zero. For example, this pattern has two callout
4736 points:
4737
4738 (?C1)abc(?C2)def
4739
4740 If the PCRE_AUTO_CALLOUT flag is passed to pcre_compile(), callouts are
4741 automatically installed before each item in the pattern. They are all
4742 numbered 255.
4743
4744 During matching, when PCRE reaches a callout point (and pcre_callout is
4745 set), the external function is called. It is provided with the number
4746 of the callout, the position in the pattern, and, optionally, one item
4747 of data originally supplied by the caller of pcre_exec(). The callout
4748 function may cause matching to proceed, to backtrack, or to fail alto-
4749 gether. A complete description of the interface to the callout function
4750 is given in the pcrecallout documentation.
4751
4752
4753 BACKTRACKING CONTROL
4754
4755 Perl 5.10 introduced a number of "Special Backtracking Control Verbs",
4756 which are described in the Perl documentation as "experimental and sub-
4757 ject to change or removal in a future version of Perl". It goes on to
4758 say: "Their usage in production code should be noted to avoid problems
4759 during upgrades." The same remarks apply to the PCRE features described
4760 in this section.
4761
4762 Since these verbs are specifically related to backtracking, they can be
4763 used only when the pattern is to be matched using pcre_exec(), which
4764 uses a backtracking algorithm. They cause an error if encountered by
4765 pcre_dfa_exec().
4766
4767 The new verbs make use of what was previously invalid syntax: an open-
4768 ing parenthesis followed by an asterisk. In Perl, they are generally of
4769 the form (*VERB:ARG) but PCRE does not support the use of arguments, so
4770 its general form is just (*VERB). Any number of these verbs may occur
4771 in a pattern. There are two kinds:
4772
4773 Verbs that act immediately
4774
4775 The following verbs act as soon as they are encountered:
4776
4777 (*ACCEPT)
4778
4779 This verb causes the match to end successfully, skipping the remainder
4780 of the pattern. When inside a recursion, only the innermost pattern is
4781 ended immediately. PCRE differs from Perl in what happens if the
4782 (*ACCEPT) is inside capturing parentheses. In Perl, the data so far is
4783 captured: in PCRE no data is captured. For example:
4784
4785 A(A|B(*ACCEPT)|C)D
4786
4787 This matches "AB", "AAD", or "ACD", but when it matches "AB", no data
4788 is captured.
4789
4790 (*FAIL) or (*F)
4791
4792 This verb causes the match to fail, forcing backtracking to occur. It
4793 is equivalent to (?!) but easier to read. The Perl documentation notes
4794 that it is probably useful only when combined with (?{}) or (??{}).
4795 Those are, of course, Perl features that are not present in PCRE. The
4796 nearest equivalent is the callout feature, as for example in this pat-
4797 tern:
4798
4799 a+(?C)(*FAIL)
4800
4801 A match with the string "aaaa" always fails, but the callout is taken
4802 before each backtrack happens (in this example, 10 times).
4803
4804 Verbs that act after backtracking
4805
4806 The following verbs do nothing when they are encountered. Matching con-
4807 tinues with what follows, but if there is no subsequent match, a fail-
4808 ure is forced. The verbs differ in exactly what kind of failure
4809 occurs.
4810
4811 (*COMMIT)
4812
4813 This verb causes the whole match to fail outright if the rest of the
4814 pattern does not match. Even if the pattern is unanchored, no further
4815 attempts to find a match by advancing the start point take place. Once
4816 (*COMMIT) has been passed, pcre_exec() is committed to finding a match
4817 at the current starting point, or not at all. For example:
4818
4819 a+(*COMMIT)b
4820
4821 This matches "xxaab" but not "aacaab". It can be thought of as a kind
4822 of dynamic anchor, or "I've started, so I must finish."
4823
4824 (*PRUNE)
4825
4826 This verb causes the match to fail at the current position if the rest
4827 of the pattern does not match. If the pattern is unanchored, the normal
4828 "bumpalong" advance to the next starting character then happens. Back-
4829 tracking can occur as usual to the left of (*PRUNE), or when matching
4830 to the right of (*PRUNE), but if there is no match to the right, back-
4831 tracking cannot cross (*PRUNE). In simple cases, the use of (*PRUNE)
4832 is just an alternative to an atomic group or possessive quantifier, but
4833 there are some uses of (*PRUNE) that cannot be expressed in any other
4834 way.
4835
4836 (*SKIP)
4837
4838 This verb is like (*PRUNE), except that if the pattern is unanchored,
4839 the "bumpalong" advance is not to the next character, but to the posi-
4840 tion in the subject where (*SKIP) was encountered. (*SKIP) signifies
4841 that whatever text was matched leading up to it cannot be part of a
4842 successful match. Consider:
4843
4844 a+(*SKIP)b
4845
4846 If the subject is "aaaac...", after the first match attempt fails
4847 (starting at the first character in the string), the starting point
4848 skips on to start the next attempt at "c". Note that a possessive quan-
4849 tifer does not have the same effect in this example; although it would
4850 suppress backtracking during the first match attempt, the second
4851 attempt would start at the second character instead of skipping on to
4852 "c".
4853
4854 (*THEN)
4855
4856 This verb causes a skip to the next alternation if the rest of the pat-
4857 tern does not match. That is, it cancels pending backtracking, but only
4858 within the current alternation. Its name comes from the observation
4859 that it can be used for a pattern-based if-then-else block:
4860
4861 ( COND1 (*THEN) FOO | COND2 (*THEN) BAR | COND3 (*THEN) BAZ ) ...
4862
4863 If the COND1 pattern matches, FOO is tried (and possibly further items
4864 after the end of the group if FOO succeeds); on failure the matcher
4865 skips to the second alternative and tries COND2, without backtracking
4866 into COND1. If (*THEN) is used outside of any alternation, it acts
4867 exactly like (*PRUNE).
4868
4869
4870 SEE ALSO
4871
4872 pcreapi(3), pcrecallout(3), pcrematching(3), pcre(3).
4873
4874
4875 AUTHOR
4876
4877 Philip Hazel
4878 University Computing Service
4879 Cambridge CB2 3QH, England.
4880
4881
4882 REVISION
4883
4884 Last updated: 17 September 2007
4885 Copyright (c) 1997-2007 University of Cambridge.
4886 ------------------------------------------------------------------------------
4887
4888
4889 PCRESYNTAX(3) PCRESYNTAX(3)
4890
4891
4892 NAME
4893 PCRE - Perl-compatible regular expressions
4894
4895
4896 PCRE REGULAR EXPRESSION SYNTAX SUMMARY
4897
4898 The full syntax and semantics of the regular expressions that are sup-
4899 ported by PCRE are described in the pcrepattern documentation. This
4900 document contains just a quick-reference summary of the syntax.
4901
4902
4903 QUOTING
4904
4905 \x where x is non-alphanumeric is a literal x
4906 \Q...\E treat enclosed characters as literal
4907
4908
4909 CHARACTERS
4910
4911 \a alarm, that is, the BEL character (hex 07)
4912 \cx "control-x", where x is any character
4913 \e escape (hex 1B)
4914 \f formfeed (hex 0C)
4915 \n newline (hex 0A)
4916 \r carriage return (hex 0D)
4917 \t tab (hex 09)
4918 \ddd character with octal code ddd, or backreference
4919 \xhh character with hex code hh
4920 \x{hhh..} character with hex code hhh..
4921
4922
4923 CHARACTER TYPES
4924
4925 . any character except newline;
4926 in dotall mode, any character whatsoever
4927 \C one byte, even in UTF-8 mode (best avoided)
4928 \d a decimal digit
4929 \D a character that is not a decimal digit
4930 \h a horizontal whitespace character
4931 \H a character that is not a horizontal whitespace character
4932 \p{xx} a character with the xx property
4933 \P{xx} a character without the xx property
4934 \R a newline sequence
4935 \s a whitespace character
4936 \S a character that is not a whitespace character
4937 \v a vertical whitespace character
4938 \V a character that is not a vertical whitespace character
4939 \w a "word" character
4940 \W a "non-word" character
4941 \X an extended Unicode sequence
4942
4943 In PCRE, \d, \D, \s, \S, \w, and \W recognize only ASCII characters.
4944
4945
4946 GENERAL CATEGORY PROPERTY CODES FOR \p and \P
4947
4948 C Other
4949 Cc Control
4950 Cf Format
4951 Cn Unassigned
4952 Co Private use
4953 Cs Surrogate
4954
4955 L Letter
4956 Ll Lower case letter
4957 Lm Modifier letter
4958 Lo Other letter
4959 Lt Title case letter
4960 Lu Upper case letter
4961 L& Ll, Lu, or Lt
4962
4963 M Mark
4964 Mc Spacing mark
4965 Me Enclosing mark
4966 Mn Non-spacing mark
4967
4968 N Number
4969 Nd Decimal number
4970 Nl Letter number
4971 No Other number
4972
4973 P Punctuation
4974 Pc Connector punctuation
4975 Pd Dash punctuation
4976 Pe Close punctuation
4977 Pf Final punctuation
4978 Pi Initial punctuation
4979 Po Other punctuation
4980 Ps Open punctuation
4981
4982 S Symbol
4983 Sc Currency symbol
4984 Sk Modifier symbol
4985 Sm Mathematical symbol
4986 So Other symbol
4987
4988 Z Separator
4989 Zl Line separator
4990 Zp Paragraph separator
4991 Zs Space separator
4992
4993
4994 SCRIPT NAMES FOR \p AND \P
4995
4996 Arabic, Armenian, Balinese, Bengali, Bopomofo, Braille, Buginese,
4997 Buhid, Canadian_Aboriginal, Cherokee, Common, Coptic, Cuneiform,
4998 Cypriot, Cyrillic, Deseret, Devanagari, Ethiopic, Georgian, Glagolitic,
4999 Gothic, Greek, Gujarati, Gurmukhi, Han, Hangul, Hanunoo, Hebrew, Hira-
5000 gana, Inherited, Kannada, Katakana, Kharoshthi, Khmer, Lao, Latin,
5001 Limbu, Linear_B, Malayalam, Mongolian, Myanmar, New_Tai_Lue, Nko,
5002 Ogham, Old_Italic, Old_Persian, Oriya, Osmanya, Phags_Pa, Phoenician,
5003 Runic, Shavian, Sinhala, Syloti_Nagri, Syriac, Tagalog, Tagbanwa,
5004 Tai_Le, Tamil, Telugu, Thaana, Thai, Tibetan, Tifinagh, Ugaritic, Yi.
5005
5006
5007 CHARACTER CLASSES
5008
5009 [...] positive character class
5010 [^...] negative character class
5011 [x-y] range (can be used for hex characters)
5012 [[:xxx:]] positive POSIX named set
5013 [[:^xxx:]] negative POSIX named set
5014
5015 alnum alphanumeric
5016 alpha alphabetic
5017 ascii 0-127
5018 blank space or tab
5019 cntrl control character
5020 digit decimal digit
5021 graph printing, excluding space
5022 lower lower case letter
5023 print printing, including space
5024 punct printing, excluding alphanumeric
5025 space whitespace
5026 upper upper case letter
5027 word same as \w
5028 xdigit hexadecimal digit
5029
5030 In PCRE, POSIX character set names recognize only ASCII characters. You
5031 can use \Q...\E inside a character class.
5032
5033
5034 QUANTIFIERS
5035
5036 ? 0 or 1, greedy
5037 ?+ 0 or 1, possessive
5038 ?? 0 or 1, lazy
5039 * 0 or more, greedy
5040 *+ 0 or more, possessive
5041 *? 0 or more, lazy
5042 + 1 or more, greedy
5043 ++ 1 or more, possessive
5044 +? 1 or more, lazy
5045 {n} exactly n
5046 {n,m} at least n, no more than m, greedy
5047 {n,m}+ at least n, no more than m, possessive
5048 {n,m}? at least n, no more than m, lazy
5049 {n,} n or more, greedy
5050 {n,}+ n or more, possessive
5051 {n,}? n or more, lazy
5052
5053
5054 ANCHORS AND SIMPLE ASSERTIONS
5055
5056 \b word boundary
5057 \B not a word boundary
5058 ^ start of subject
5059 also after internal newline in multiline mode
5060 \A start of subject
5061 $ end of subject
5062 also before newline at end of subject
5063 also before internal newline in multiline mode
5064 \Z end of subject
5065 also before newline at end of subject
5066 \z end of subject
5067 \G first matching position in subject
5068
5069
5070 MATCH POINT RESET
5071
5072 \K reset start of match
5073
5074
5075 ALTERNATION
5076
5077 expr|expr|expr...
5078
5079
5080 CAPTURING
5081
5082 (...) capturing group
5083 (?<name>...) named capturing group (Perl)
5084 (?'name'...) named capturing group (Perl)
5085 (?P<name>...) named capturing group (Python)
5086 (?:...) non-capturing group
5087 (?|...) non-capturing group; reset group numbers for
5088 capturing groups in each alternative
5089
5090
5091 ATOMIC GROUPS
5092
5093 (?>...) atomic, non-capturing group
5094
5095
5096 COMMENT
5097
5098 (?#....) comment (not nestable)
5099
5100
5101 OPTION SETTING
5102
5103 (?i) caseless
5104 (?J) allow duplicate names
5105 (?m) multiline
5106 (?s) single line (dotall)
5107 (?U) default ungreedy (lazy)
5108 (?x) extended (ignore white space)
5109 (?-...) unset option(s)
5110
5111
5112 LOOKAHEAD AND LOOKBEHIND ASSERTIONS
5113
5114 (?=...) positive look ahead
5115 (?!...) negative look ahead
5116 (?<=...) positive look behind
5117 (?<!...) negative look behind
5118
5119 Each top-level branch of a look behind must be of a fixed length.
5120
5121
5122 BACKREFERENCES
5123
5124 \n reference by number (can be ambiguous)
5125 \gn reference by number
5126 \g{n} reference by number
5127 \g{-n} relative reference by number
5128 \k<name> reference by name (Perl)
5129 \k'name' reference by name (Perl)
5130 \g{name} reference by name (Perl)
5131 \k{name} reference by name (.NET)
5132 (?P=name) reference by name (Python)
5133
5134
5135 SUBROUTINE REFERENCES (POSSIBLY RECURSIVE)
5136
5137 (?R) recurse whole pattern
5138 (?n) call subpattern by absolute number
5139 (?+n) call subpattern by relative number
5140 (?-n) call subpattern by relative number
5141 (?&name) call subpattern by name (Perl)
5142 (?P>name) call subpattern by name (Python)
5143
5144
5145 CONDITIONAL PATTERNS
5146
5147 (?(condition)yes-pattern)
5148 (?(condition)yes-pattern|no-pattern)
5149
5150 (?(n)... absolute reference condition
5151 (?(+n)... relative reference condition
5152 (?(-n)... relative reference condition
5153 (?(<name>)... named reference condition (Perl)
5154 (?('name')... named reference condition (Perl)
5155 (?(name)... named reference condition (PCRE)
5156 (?(R)... overall recursion condition
5157 (?(Rn)... specific group recursion condition
5158 (?(R&name)... specific recursion condition
5159 (?(DEFINE)... define subpattern for reference
5160 (?(assert)... assertion condition
5161
5162
5163 BACKTRACKING CONTROL
5164
5165 The following act immediately they are reached:
5166
5167 (*ACCEPT) force successful match
5168 (*FAIL) force backtrack; synonym (*F)
5169
5170 The following act only when a subsequent match failure causes a back-
5171 track to reach them. They all force a match failure, but they differ in
5172 what happens afterwards. Those that advance the start-of-match point do
5173 so only if the pattern is not anchored.
5174
5175 (*COMMIT) overall failure, no advance of starting point
5176 (*PRUNE) advance to next starting character
5177 (*SKIP) advance start to current matching position
5178 (*THEN) local failure, backtrack to next alternation
5179
5180
5181 NEWLINE CONVENTIONS
5182
5183 These are recognized only at the very start of the pattern or after a
5184 (*BSR_...) option.
5185
5186 (*CR)
5187 (*LF)
5188 (*CRLF)
5189 (*ANYCRLF)
5190 (*ANY)
5191
5192
5193 WHAT \R MATCHES
5194
5195 These are recognized only at the very start of the pattern or after a
5196 (*...) option that sets the newline convention.
5197
5198 (*BSR_ANYCRLF)
5199 (*BSR_UNICODE)
5200
5201
5202 CALLOUTS
5203
5204 (?C) callout
5205 (?Cn) callout with data n
5206
5207
5208 SEE ALSO
5209
5210 pcrepattern(3), pcreapi(3), pcrecallout(3), pcrematching(3), pcre(3).
5211
5212
5213 AUTHOR
5214
5215 Philip Hazel
5216 University Computing Service
5217 Cambridge CB2 3QH, England.
5218
5219
5220 REVISION
5221
5222 Last updated: 14 November 2007
5223 Copyright (c) 1997-2007 University of Cambridge.
5224 ------------------------------------------------------------------------------
5225
5226
5227 PCREPARTIAL(3) PCREPARTIAL(3)
5228
5229
5230 NAME
5231 PCRE - Perl-compatible regular expressions
5232
5233
5234 PARTIAL MATCHING IN PCRE
5235
5236 In normal use of PCRE, if the subject string that is passed to
5237 pcre_exec() or pcre_dfa_exec() matches as far as it goes, but is too
5238 short to match the entire pattern, PCRE_ERROR_NOMATCH is returned.
5239 There are circumstances where it might be helpful to distinguish this
5240 case from other cases in which there is no match.
5241
5242 Consider, for example, an application where a human is required to type
5243 in data for a field with specific formatting requirements. An example
5244 might be a date in the form ddmmmyy, defined by this pattern:
5245
5246 ^\d?\d(jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)\d\d$
5247
5248 If the application sees the user's keystrokes one by one, and can check
5249 that what has been typed so far is potentially valid, it is able to
5250 raise an error as soon as a mistake is made, possibly beeping and not
5251 reflecting the character that has been typed. This immediate feedback
5252 is likely to be a better user interface than a check that is delayed
5253 until the entire string has been entered.
5254
5255 PCRE supports the concept of partial matching by means of the PCRE_PAR-
5256 TIAL option, which can be set when calling pcre_exec() or
5257 pcre_dfa_exec(). When this flag is set for pcre_exec(), the return code
5258 PCRE_ERROR_NOMATCH is converted into PCRE_ERROR_PARTIAL if at any time
5259 during the matching process the last part of the subject string matched
5260 part of the pattern. Unfortunately, for non-anchored matching, it is
5261 not possible to obtain the position of the start of the partial match.
5262 No captured data is set when PCRE_ERROR_PARTIAL is returned.
5263
5264 When PCRE_PARTIAL is set for pcre_dfa_exec(), the return code
5265 PCRE_ERROR_NOMATCH is converted into PCRE_ERROR_PARTIAL if the end of
5266 the subject is reached, there have been no complete matches, but there
5267 is still at least one matching possibility. The portion of the string
5268 that provided the partial match is set as the first matching string.
5269
5270 Using PCRE_PARTIAL disables one of PCRE's optimizations. PCRE remembers
5271 the last literal byte in a pattern, and abandons matching immediately
5272 if such a byte is not present in the subject string. This optimization
5273 cannot be used for a subject string that might match only partially.
5274
5275
5276 RESTRICTED PATTERNS FOR PCRE_PARTIAL
5277
5278 Because of the way certain internal optimizations are implemented in
5279 the pcre_exec() function, the PCRE_PARTIAL option cannot be used with
5280 all patterns. These restrictions do not apply when pcre_dfa_exec() is
5281 used. For pcre_exec(), repeated single characters such as
5282
5283 a{2,4}
5284
5285 and repeated single metasequences such as
5286
5287 \d+
5288
5289 are not permitted if the maximum number of occurrences is greater than
5290 one. Optional items such as \d? (where the maximum is one) are permit-
5291 ted. Quantifiers with any values are permitted after parentheses, so
5292 the invalid examples above can be coded thus:
5293
5294 (a){2,4}
5295 (\d)+
5296
5297 These constructions run more slowly, but for the kinds of application
5298 that are envisaged for this facility, this is not felt to be a major
5299 restriction.
5300
5301 If PCRE_PARTIAL is set for a pattern that does not conform to the
5302 restrictions, pcre_exec() returns the error code PCRE_ERROR_BADPARTIAL
5303 (-13). You can use the PCRE_INFO_OKPARTIAL call to pcre_fullinfo() to
5304 find out if a compiled pattern can be used for partial matching.
5305
5306
5307 EXAMPLE OF PARTIAL MATCHING USING PCRETEST
5308
5309 If the escape sequence \P is present in a pcretest data line, the
5310 PCRE_PARTIAL flag is used for the match. Here is a run of pcretest that
5311 uses the date example quoted above:
5312
5313 re> /^\d?\d(jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)\d\d$/
5314 data> 25jun04\P
5315 0: 25jun04
5316 1: jun
5317 data> 25dec3\P
5318 Partial match
5319 data> 3ju\P
5320 Partial match
5321 data> 3juj\P
5322 No match
5323 data> j\P
5324 No match
5325
5326 The first data string is matched completely, so pcretest shows the
5327 matched substrings. The remaining four strings do not match the com-
5328 plete pattern, but the first two are partial matches. The same test,
5329 using pcre_dfa_exec() matching (by means of the \D escape sequence),
5330 produces the following output:
5331
5332 re> /^\d?\d(jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)\d\d$/
5333 data> 25jun04\P\D
5334 0: 25jun04
5335 data> 23dec3\P\D
5336 Partial match: 23dec3
5337 data> 3ju\P\D
5338 Partial match: 3ju
5339 data> 3juj\P\D
5340 No match
5341 data> j\P\D
5342 No match
5343
5344 Notice that in this case the portion of the string that was matched is
5345 made available.
5346
5347
5348 MULTI-SEGMENT MATCHING WITH pcre_dfa_exec()
5349
5350 When a partial match has been found using pcre_dfa_exec(), it is possi-
5351 ble to continue the match by providing additional subject data and
5352 calling pcre_dfa_exec() again with the same compiled regular expres-
5353 sion, this time setting the PCRE_DFA_RESTART option. You must also pass
5354 the same working space as before, because this is where details of the
5355 previous partial match are stored. Here is an example using pcretest,
5356 using the \R escape sequence to set the PCRE_DFA_RESTART option (\P and
5357 \D are as above):
5358
5359 re> /^\d?\d(jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)\d\d$/
5360 data> 23ja\P\D
5361 Partial match: 23ja
5362 data> n05\R\D
5363 0: n05
5364
5365 The first call has "23ja" as the subject, and requests partial match-
5366 ing; the second call has "n05" as the subject for the continued
5367 (restarted) match. Notice that when the match is complete, only the
5368 last part is shown; PCRE does not retain the previously partially-
5369 matched string. It is up to the calling program to do that if it needs
5370 to.
5371
5372 You can set PCRE_PARTIAL with PCRE_DFA_RESTART to continue partial
5373 matching over multiple segments. This facility can be used to pass very
5374 long subject strings to pcre_dfa_exec(). However, some care is needed
5375 for certain types of pattern.
5376
5377 1. If the pattern contains tests for the beginning or end of a line,
5378 you need to pass the PCRE_NOTBOL or PCRE_NOTEOL options, as appropri-
5379 ate, when the subject string for any call does not contain the begin-
5380 ning or end of a line.
5381
5382 2. If the pattern contains backward assertions (including \b or \B),
5383 you need to arrange for some overlap in the subject strings to allow
5384 for this. For example, you could pass the subject in chunks that are
5385 500 bytes long, but in a buffer of 700 bytes, with the starting offset
5386 set to 200 and the previous 200 bytes at the start of the buffer.
5387
5388 3. Matching a subject string that is split into multiple segments does
5389 not always produce exactly the same result as matching over one single
5390 long string. The difference arises when there are multiple matching
5391 possibilities, because a partial match result is given only when there
5392 are no completed matches in a call to pcre_dfa_exec(). This means that
5393 as soon as the shortest match has been found, continuation to a new
5394 subject segment is no longer possible. Consider this pcretest example:
5395
5396 re> /dog(sbody)?/
5397 data> do\P\D
5398 Partial match: do
5399 data> gsb\R\P\D
5400 0: g
5401 data> dogsbody\D
5402 0: dogsbody
5403 1: dog
5404
5405 The pattern matches the words "dog" or "dogsbody". When the subject is
5406 presented in several parts ("do" and "gsb" being the first two) the
5407 match stops when "dog" has been found, and it is not possible to con-
5408 tinue. On the other hand, if "dogsbody" is presented as a single
5409 string, both matches are found.
5410
5411 Because of this phenomenon, it does not usually make sense to end a
5412 pattern that is going to be matched in this way with a variable repeat.
5413
5414 4. Patterns that contain alternatives at the top level which do not all
5415 start with the same pattern item may not work as expected. For example,
5416 consider this pattern:
5417
5418 1234|3789
5419
5420 If the first part of the subject is "ABC123", a partial match of the
5421 first alternative is found at offset 3. There is no partial match for
5422 the second alternative, because such a match does not start at the same
5423 point in the subject string. Attempting to continue with the string
5424 "789" does not yield a match because only those alternatives that match
5425 at one point in the subject are remembered. The problem arises because
5426 the start of the second alternative matches within the first alterna-
5427 tive. There is no problem with anchored patterns or patterns such as:
5428
5429 1234|ABCD
5430
5431 where no string can be a partial match for both alternatives.
5432
5433
5434 AUTHOR
5435
5436 Philip Hazel
5437 University Computing Service
5438 Cambridge CB2 3QH, England.
5439
5440
5441 REVISION
5442
5443 Last updated: 04 June 2007
5444 Copyright (c) 1997-2007 University of Cambridge.
5445 ------------------------------------------------------------------------------
5446
5447
5448 PCREPRECOMPILE(3) PCREPRECOMPILE(3)
5449
5450
5451 NAME
5452 PCRE - Perl-compatible regular expressions
5453
5454
5455 SAVING AND RE-USING PRECOMPILED PCRE PATTERNS
5456
5457 If you are running an application that uses a large number of regular
5458 expression patterns, it may be useful to store them in a precompiled
5459 form instead of having to compile them every time the application is
5460 run. If you are not using any private character tables (see the
5461 pcre_maketables() documentation), this is relatively straightforward.
5462 If you are using private tables, it is a little bit more complicated.
5463
5464 If you save compiled patterns to a file, you can copy them to a differ-
5465 ent host and run them there. This works even if the new host has the
5466 opposite endianness to the one on which the patterns were compiled.
5467 There may be a small performance penalty, but it should be insignifi-
5468 cant. However, compiling regular expressions with one version of PCRE
5469 for use with a different version is not guaranteed to work and may
5470 cause crashes.
5471
5472
5473 SAVING A COMPILED PATTERN
5474 The value returned by pcre_compile() points to a single block of memory
5475 that holds the compiled pattern and associated data. You can find the
5476 length of this block in bytes by calling pcre_fullinfo() with an argu-
5477 ment of PCRE_INFO_SIZE. You can then save the data in any appropriate
5478 manner. Here is sample code that compiles a pattern and writes it to a
5479 file. It assumes that the variable fd refers to a file that is open for
5480 output:
5481
5482 int erroroffset, rc, size;
5483 char *error;
5484 pcre *re;
5485
5486 re = pcre_compile("my pattern", 0, &error, &erroroffset, NULL);
5487 if (re == NULL) { ... handle errors ... }
5488 rc = pcre_fullinfo(re, NULL, PCRE_INFO_SIZE, &size);
5489 if (rc < 0) { ... handle errors ... }
5490 rc = fwrite(re, 1, size, fd);
5491 if (rc != size) { ... handle errors ... }
5492
5493 In this example, the bytes that comprise the compiled pattern are
5494 copied exactly. Note that this is binary data that may contain any of
5495 the 256 possible byte values. On systems that make a distinction
5496 between binary and non-binary data, be sure that the file is opened for
5497 binary output.
5498
5499 If you want to write more than one pattern to a file, you will have to
5500 devise a way of separating them. For binary data, preceding each pat-
5501 tern with its length is probably the most straightforward approach.
5502 Another possibility is to write out the data in hexadecimal instead of
5503 binary, one pattern to a line.
5504
5505 Saving compiled patterns in a file is only one possible way of storing
5506 them for later use. They could equally well be saved in a database, or
5507 in the memory of some daemon process that passes them via sockets to
5508 the processes that want them.
5509
5510 If the pattern has been studied, it is also possible to save the study
5511 data in a similar way to the compiled pattern itself. When studying
5512 generates additional information, pcre_study() returns a pointer to a
5513 pcre_extra data block. Its format is defined in the section on matching
5514 a pattern in the pcreapi documentation. The study_data field points to
5515 the binary study data, and this is what you must save (not the
5516 pcre_extra block itself). The length of the study data can be obtained
5517 by calling pcre_fullinfo() with an argument of PCRE_INFO_STUDYSIZE.
5518 Remember to check that pcre_study() did return a non-NULL value before
5519 trying to save the study data.
5520
5521
5522 RE-USING A PRECOMPILED PATTERN
5523
5524 Re-using a precompiled pattern is straightforward. Having reloaded it
5525 into main memory, you pass its pointer to pcre_exec() or
5526 pcre_dfa_exec() in the usual way. This should work even on another
5527 host, and even if that host has the opposite endianness to the one
5528 where the pattern was compiled.
5529
5530 However, if you passed a pointer to custom character tables when the
5531 pattern was compiled (the tableptr argument of pcre_compile()), you
5532 must now pass a similar pointer to pcre_exec() or pcre_dfa_exec(),
5533 because the value saved with the compiled pattern will obviously be
5534 nonsense. A field in a pcre_extra() block is used to pass this data, as
5535 described in the section on matching a pattern in the pcreapi documen-
5536 tation.
5537
5538 If you did not provide custom character tables when the pattern was
5539 compiled, the pointer in the compiled pattern is NULL, which causes
5540 pcre_exec() to use PCRE's internal tables. Thus, you do not need to
5541 take any special action at run time in this case.
5542
5543 If you saved study data with the compiled pattern, you need to create
5544 your own pcre_extra data block and set the study_data field to point to
5545 the reloaded study data. You must also set the PCRE_EXTRA_STUDY_DATA
5546 bit in the flags field to indicate that study data is present. Then
5547 pass the pcre_extra block to pcre_exec() or pcre_dfa_exec() in the
5548 usual way.
5549
5550
5551 COMPATIBILITY WITH DIFFERENT PCRE RELEASES
5552
5553 In general, it is safest to recompile all saved patterns when you
5554 update to a new PCRE release, though not all updates actually require
5555 this. Recompiling is definitely needed for release 7.2.
5556
5557
5558 AUTHOR
5559
5560 Philip Hazel
5561 University Computing Service
5562 Cambridge CB2 3QH, England.
5563
5564
5565 REVISION
5566
5567 Last updated: 13 June 2007
5568 Copyright (c) 1997-2007 University of Cambridge.
5569 ------------------------------------------------------------------------------
5570
5571
5572 PCREPERFORM(3) PCREPERFORM(3)
5573
5574
5575 NAME
5576 PCRE - Perl-compatible regular expressions
5577
5578
5579 PCRE PERFORMANCE
5580
5581 Two aspects of performance are discussed below: memory usage and pro-
5582 cessing time. The way you express your pattern as a regular expression
5583 can affect both of them.
5584
5585
5586 MEMORY USAGE
5587
5588 Patterns are compiled by PCRE into a reasonably efficient byte code, so
5589 that most simple patterns do not use much memory. However, there is one
5590 case where memory usage can be unexpectedly large. When a parenthesized
5591 subpattern has a quantifier with a minimum greater than 1 and/or a lim-
5592 ited maximum, the whole subpattern is repeated in the compiled code.
5593 For example, the pattern
5594
5595 (abc|def){2,4}
5596
5597 is compiled as if it were
5598
5599 (abc|def)(abc|def)((abc|def)(abc|def)?)?
5600
5601 (Technical aside: It is done this way so that backtrack points within
5602 each of the repetitions can be independently maintained.)
5603
5604 For regular expressions whose quantifiers use only small numbers, this
5605 is not usually a problem. However, if the numbers are large, and par-
5606 ticularly if such repetitions are nested, the memory usage can become
5607 an embarrassment. For example, the very simple pattern
5608
5609 ((ab){1,1000}c){1,3}
5610
5611 uses 51K bytes when compiled. When PCRE is compiled with its default
5612 internal pointer size of two bytes, the size limit on a compiled pat-
5613 tern is 64K, and this is reached with the above pattern if the outer
5614 repetition is increased from 3 to 4. PCRE can be compiled to use larger
5615 internal pointers and thus handle larger compiled patterns, but it is
5616 better to try to rewrite your pattern to use less memory if you can.
5617
5618 One way of reducing the memory usage for such patterns is to make use
5619 of PCRE's "subroutine" facility. Re-writing the above pattern as
5620
5621 ((ab)(?2){0,999}c)(?1){0,2}
5622
5623 reduces the memory requirements to 18K, and indeed it remains under 20K
5624 even with the outer repetition increased to 100. However, this pattern
5625 is not exactly equivalent, because the "subroutine" calls are treated
5626 as atomic groups into which there can be no backtracking if there is a
5627 subsequent matching failure. Therefore, PCRE cannot do this kind of
5628 rewriting automatically. Furthermore, there is a noticeable loss of
5629 speed when executing the modified pattern. Nevertheless, if the atomic
5630 grouping is not a problem and the loss of speed is acceptable, this
5631 kind of rewriting will allow you to process patterns that PCRE cannot
5632 otherwise handle.
5633
5634
5635 PROCESSING TIME
5636
5637 Certain items in regular expression patterns are processed more effi-
5638 ciently than others. It is more efficient to use a character class like
5639 [aeiou] than a set of single-character alternatives such as
5640 (a|e|i|o|u). In general, the simplest construction that provides the
5641 required behaviour is usually the most efficient. Jeffrey Friedl's book
5642 contains a lot of useful general discussion about optimizing regular
5643 expressions for efficient performance. This document contains a few
5644 observations about PCRE.
5645
5646 Using Unicode character properties (the \p, \P, and \X escapes) is
5647 slow, because PCRE has to scan a structure that contains data for over
5648 fifteen thousand characters whenever it needs a character's property.
5649 If you can find an alternative pattern that does not use character
5650 properties, it will probably be faster.
5651
5652 When a pattern begins with .* not in parentheses, or in parentheses
5653 that are not the subject of a backreference, and the PCRE_DOTALL option
5654 is set, the pattern is implicitly anchored by PCRE, since it can match
5655 only at the start of a subject string. However, if PCRE_DOTALL is not
5656 set, PCRE cannot make this optimization, because the . metacharacter
5657 does not then match a newline, and if the subject string contains new-
5658 lines, the pattern may match from the character immediately following
5659 one of them instead of from the very start. For example, the pattern
5660
5661 .*second
5662
5663 matches the subject "first\nand second" (where \n stands for a newline
5664 character), with the match starting at the seventh character. In order
5665 to do this, PCRE has to retry the match starting after every newline in
5666 the subject.
5667
5668 If you are using such a pattern with subject strings that do not con-
5669 tain newlines, the best performance is obtained by setting PCRE_DOTALL,
5670 or starting the pattern with ^.* or ^.*? to indicate explicit anchor-
5671 ing. That saves PCRE from having to scan along the subject looking for
5672 a newline to restart at.
5673
5674 Beware of patterns that contain nested indefinite repeats. These can
5675 take a long time to run when applied to a string that does not match.
5676 Consider the pattern fragment
5677
5678 ^(a+)*
5679
5680 This can match "aaaa" in 16 different ways, and this number increases
5681 very rapidly as the string gets longer. (The * repeat can match 0, 1,
5682 2, 3, or 4 times, and for each of those cases other than 0 or 4, the +
5683 repeats can match different numbers of times.) When the remainder of
5684 the pattern is such that the entire match is going to fail, PCRE has in
5685 principle to try every possible variation, and this can take an
5686 extremely long time, even for relatively short strings.
5687
5688 An optimization catches some of the more simple cases such as
5689
5690 (a+)*b
5691
5692 where a literal character follows. Before embarking on the standard
5693 matching procedure, PCRE checks that there is a "b" later in the sub-
5694 ject string, and if there is not, it fails the match immediately. How-
5695 ever, when there is no following literal this optimization cannot be
5696 used. You can see the difference by comparing the behaviour of
5697
5698 (a+)*\d
5699
5700 with the pattern above. The former gives a failure almost instantly
5701 when applied to a whole line of "a" characters, whereas the latter
5702 takes an appreciable time with strings longer than about 20 characters.
5703
5704 In many cases, the solution to this kind of performance issue is to use
5705 an atomic group or a possessive quantifier.
5706
5707
5708 AUTHOR
5709
5710 Philip Hazel
5711 University Computing Service
5712 Cambridge CB2 3QH, England.
5713
5714
5715 REVISION
5716
5717 Last updated: 06 March 2007
5718 Copyright (c) 1997-2007 University of Cambridge.
5719 ------------------------------------------------------------------------------
5720
5721
5722 PCREPOSIX(3) PCREPOSIX(3)
5723
5724
5725 NAME
5726 PCRE - Perl-compatible regular expressions.
5727
5728
5729 SYNOPSIS OF POSIX API
5730
5731 #include <pcreposix.h>
5732
5733 int regcomp(regex_t *preg, const char *pattern,
5734 int cflags);
5735
5736 int regexec(regex_t *preg, const char *string,
5737 size_t nmatch, regmatch_t pmatch[], int eflags);
5738
5739 size_t regerror(int errcode, const regex_t *preg,
5740 char *errbuf, size_t errbuf_size);
5741
5742 void regfree(regex_t *preg);
5743
5744
5745 DESCRIPTION
5746
5747 This set of functions provides a POSIX-style API to the PCRE regular
5748 expression package. See the pcreapi documentation for a description of
5749 PCRE's native API, which contains much additional functionality.
5750
5751 The functions described here are just wrapper functions that ultimately
5752 call the PCRE native API. Their prototypes are defined in the
5753 pcreposix.h header file, and on Unix systems the library itself is
5754 called pcreposix.a, so can be accessed by adding -lpcreposix to the
5755 command for linking an application that uses them. Because the POSIX
5756 functions call the native ones, it is also necessary to add -lpcre.
5757
5758 I have implemented only those option bits that can be reasonably mapped
5759 to PCRE native options. In addition, the option REG_EXTENDED is defined
5760 with the value zero. This has no effect, but since programs that are
5761 written to the POSIX interface often use it, this makes it easier to
5762 slot in PCRE as a replacement library. Other POSIX options are not even
5763 defined.
5764
5765 When PCRE is called via these functions, it is only the API that is
5766 POSIX-like in style. The syntax and semantics of the regular expres-
5767 sions themselves are still those of Perl, subject to the setting of
5768 various PCRE options, as described below. "POSIX-like in style" means
5769 that the API approximates to the POSIX definition; it is not fully
5770 POSIX-compatible, and in multi-byte encoding domains it is probably
5771 even less compatible.
5772
5773 The header for these functions is supplied as pcreposix.h to avoid any
5774 potential clash with other POSIX libraries. It can, of course, be
5775 renamed or aliased as regex.h, which is the "correct" name. It provides
5776 two structure types, regex_t for compiled internal forms, and reg-
5777 match_t for returning captured substrings. It also defines some con-
5778 stants whose names start with "REG_"; these are used for setting
5779 options and identifying error codes.
5780
5781
5782 COMPILING A PATTERN
5783
5784 The function regcomp() is called to compile a pattern into an internal
5785 form. The pattern is a C string terminated by a binary zero, and is
5786 passed in the argument pattern. The preg argument is a pointer to a
5787 regex_t structure that is used as a base for storing information about
5788 the compiled regular expression.
5789
5790 The argument cflags is either zero, or contains one or more of the bits
5791 defined by the following macros:
5792
5793 REG_DOTALL
5794
5795 The PCRE_DOTALL option is set when the regular expression is passed for
5796 compilation to the native function. Note that REG_DOTALL is not part of
5797 the POSIX standard.
5798
5799 REG_ICASE
5800
5801 The PCRE_CASELESS option is set when the regular expression is passed
5802 for compilation to the native function.
5803
5804 REG_NEWLINE
5805
5806 The PCRE_MULTILINE option is set when the regular expression is passed
5807 for compilation to the native function. Note that this does not mimic
5808 the defined POSIX behaviour for REG_NEWLINE (see the following sec-
5809 tion).
5810
5811 REG_NOSUB
5812
5813 The PCRE_NO_AUTO_CAPTURE option is set when the regular expression is
5814 passed for compilation to the native function. In addition, when a pat-
5815 tern that is compiled with this flag is passed to regexec() for match-
5816 ing, the nmatch and pmatch arguments are ignored, and no captured
5817 strings are returned.
5818
5819 REG_UTF8
5820
5821 The PCRE_UTF8 option is set when the regular expression is passed for
5822 compilation to the native function. This causes the pattern itself and
5823 all data strings used for matching it to be treated as UTF-8 strings.
5824 Note that REG_UTF8 is not part of the POSIX standard.
5825
5826 In the absence of these flags, no options are passed to the native
5827 function. This means the the regex is compiled with PCRE default
5828 semantics. In particular, the way it handles newline characters in the
5829 subject string is the Perl way, not the POSIX way. Note that setting
5830 PCRE_MULTILINE has only some of the effects specified for REG_NEWLINE.
5831 It does not affect the way newlines are matched by . (they aren't) or
5832 by a negative class such as [^a] (they are).
5833
5834 The yield of regcomp() is zero on success, and non-zero otherwise. The
5835 preg structure is filled in on success, and one member of the structure
5836 is public: re_nsub contains the number of capturing subpatterns in the
5837 regular expression. Various error codes are defined in the header file.
5838
5839
5840 MATCHING NEWLINE CHARACTERS
5841
5842 This area is not simple, because POSIX and Perl take different views of
5843 things. It is not possible to get PCRE to obey POSIX semantics, but
5844 then PCRE was never intended to be a POSIX engine. The following table
5845 lists the different possibilities for matching newline characters in
5846 PCRE:
5847
5848 Default Change with
5849
5850 . matches newline no PCRE_DOTALL
5851 newline matches [^a] yes not changeable
5852 $ matches \n at end yes PCRE_DOLLARENDONLY
5853 $ matches \n in middle no PCRE_MULTILINE
5854 ^ matches \n in middle no PCRE_MULTILINE
5855
5856 This is the equivalent table for POSIX:
5857
5858 Default Change with
5859
5860 . matches newline yes REG_NEWLINE
5861 newline matches [^a] yes REG_NEWLINE
5862 $ matches \n at end no REG_NEWLINE
5863 $ matches \n in middle no REG_NEWLINE
5864 ^ matches \n in middle no REG_NEWLINE
5865
5866 PCRE's behaviour is the same as Perl's, except that there is no equiva-
5867 lent for PCRE_DOLLAR_ENDONLY in Perl. In both PCRE and Perl, there is
5868 no way to stop newline from matching [^a].
5869
5870 The default POSIX newline handling can be obtained by setting
5871 PCRE_DOTALL and PCRE_DOLLAR_ENDONLY, but there is no way to make PCRE
5872 behave exactly as for the REG_NEWLINE action.
5873
5874
5875 MATCHING A PATTERN
5876
5877 The function regexec() is called to match a compiled pattern preg
5878 against a given string, which is terminated by a zero byte, subject to
5879 the options in eflags. These can be:
5880
5881 REG_NOTBOL
5882
5883 The PCRE_NOTBOL option is set when calling the underlying PCRE matching
5884 function.
5885
5886 REG_NOTEOL
5887
5888 The PCRE_NOTEOL option is set when calling the underlying PCRE matching
5889 function.
5890
5891 If the pattern was compiled with the REG_NOSUB flag, no data about any
5892 matched strings is returned. The nmatch and pmatch arguments of
5893 regexec() are ignored.
5894
5895 Otherwise,the portion of the string that was matched, and also any cap-
5896 tured substrings, are returned via the pmatch argument, which points to
5897 an array of nmatch structures of type regmatch_t, containing the mem-
5898 bers rm_so and rm_eo. These contain the offset to the first character
5899 of each substring and the offset to the first character after the end
5900 of each substring, respectively. The 0th element of the vector relates
5901 to the entire portion of string that was matched; subsequent elements
5902 relate to the capturing subpatterns of the regular expression. Unused
5903 entries in the array have both structure members set to -1.
5904
5905 A successful match yields a zero return; various error codes are
5906 defined in the header file, of which REG_NOMATCH is the "expected"
5907 failure code.
5908
5909
5910 ERROR MESSAGES
5911
5912 The regerror() function maps a non-zero errorcode from either regcomp()
5913 or regexec() to a printable message. If preg is not NULL, the error
5914 should have arisen from the use of that structure. A message terminated
5915 by a binary zero is placed in errbuf. The length of the message,
5916 including the zero, is limited to errbuf_size. The yield of the func-
5917 tion is the size of buffer needed to hold the whole message.
5918
5919
5920 MEMORY USAGE
5921
5922 Compiling a regular expression causes memory to be allocated and asso-
5923 ciated with the preg structure. The function regfree() frees all such
5924 memory, after which preg may no longer be used as a compiled expres-
5925 sion.
5926
5927
5928 AUTHOR
5929
5930 Philip Hazel
5931 University Computing Service
5932 Cambridge CB2 3QH, England.
5933
5934
5935 REVISION
5936
5937 Last updated: 06 March 2007
5938 Copyright (c) 1997-2007 University of Cambridge.
5939 ------------------------------------------------------------------------------
5940
5941
5942 PCRECPP(3) PCRECPP(3)
5943
5944
5945 NAME
5946 PCRE - Perl-compatible regular expressions.
5947
5948
5949 SYNOPSIS OF C++ WRAPPER
5950
5951 #include <pcrecpp.h>
5952
5953
5954 DESCRIPTION
5955
5956 The C++ wrapper for PCRE was provided by Google Inc. Some additional
5957 functionality was added by Giuseppe Maxia. This brief man page was con-
5958 structed from the notes in the pcrecpp.h file, which should be con-
5959 sulted for further details.
5960
5961
5962 MATCHING INTERFACE
5963
5964 The "FullMatch" operation checks that supplied text matches a supplied
5965 pattern exactly. If pointer arguments are supplied, it copies matched
5966 sub-strings that match sub-patterns into them.
5967
5968 Example: successful match
5969 pcrecpp::RE re("h.*o");
5970 re.FullMatch("hello");
5971
5972 Example: unsuccessful match (requires full match):
5973 pcrecpp::RE re("e");
5974 !re.FullMatch("hello");
5975
5976 Example: creating a temporary RE object:
5977 pcrecpp::RE("h.*o").FullMatch("hello");
5978
5979 You can pass in a "const char*" or a "string" for "text". The examples
5980 below tend to use a const char*. You can, as in the different examples
5981 above, store the RE object explicitly in a variable or use a temporary
5982 RE object. The examples below use one mode or the other arbitrarily.
5983 Either could correctly be used for any of these examples.
5984
5985 You must supply extra pointer arguments to extract matched subpieces.
5986
5987 Example: extracts "ruby" into "s" and 1234 into "i"
5988 int i;
5989 string s;
5990 pcrecpp::RE re("(\\w+):(\\d+)");
5991 re.FullMatch("ruby:1234", &s, &i);
5992
5993 Example: does not try to extract any extra sub-patterns
5994 re.FullMatch("ruby:1234", &s);
5995
5996 Example: does not try to extract into NULL
5997 re.FullMatch("ruby:1234", NULL, &i);
5998
5999 Example: integer overflow causes failure
6000 !re.FullMatch("ruby:1234567891234", NULL, &i);
6001
6002 Example: fails because there aren't enough sub-patterns:
6003 !pcrecpp::RE("\\w+:\\d+").FullMatch("ruby:1234", &s);
6004
6005 Example: fails because string cannot be stored in integer
6006 !pcrecpp::RE("(.*)").FullMatch("ruby", &i);
6007
6008 The provided pointer arguments can be pointers to any scalar numeric
6009 type, or one of:
6010
6011 string (matched piece is copied to string)
6012 StringPiece (StringPiece is mutated to point to matched piece)
6013 T (where "bool T::ParseFrom(const char*, int)" exists)
6014 NULL (the corresponding matched sub-pattern is not copied)
6015
6016 The function returns true iff all of the following conditions are sat-
6017 isfied:
6018
6019 a. "text" matches "pattern" exactly;
6020
6021 b. The number of matched sub-patterns is >= number of supplied
6022 pointers;
6023
6024 c. The "i"th argument has a suitable type for holding the
6025 string captured as the "i"th sub-pattern. If you pass in
6026 void * NULL for the "i"th argument, or a non-void * NULL
6027 of the correct type, or pass fewer arguments than the
6028 number of sub-patterns, "i"th captured sub-pattern is
6029 ignored.
6030
6031 CAVEAT: An optional sub-pattern that does not exist in the matched
6032 string is assigned the empty string. Therefore, the following will
6033 return false (because the empty string is not a valid number):
6034
6035 int number;
6036 pcrecpp::RE::FullMatch("abc", "[a-z]+(\\d+)?", &number);
6037
6038 The matching interface supports at most 16 arguments per call. If you
6039 need more, consider using the more general interface
6040 pcrecpp::RE::DoMatch. See pcrecpp.h for the signature for DoMatch.
6041
6042
6043 QUOTING METACHARACTERS
6044
6045 You can use the "QuoteMeta" operation to insert backslashes before all
6046 potentially meaningful characters in a string. The returned string,
6047 used as a regular expression, will exactly match the original string.
6048
6049 Example:
6050 string quoted = RE::QuoteMeta(unquoted);
6051
6052 Note that it's legal to escape a character even if it has no special
6053 meaning in a regular expression -- so this function does that. (This
6054 also makes it identical to the perl function of the same name; see
6055 "perldoc -f quotemeta".) For example, "1.5-2.0?" becomes
6056 "1\.5\-2\.0\?".
6057
6058
6059 PARTIAL MATCHES
6060
6061 You can use the "PartialMatch" operation when you want the pattern to
6062 match any substring of the text.
6063
6064 Example: simple search for a string:
6065 pcrecpp::RE("ell").PartialMatch("hello");
6066
6067 Example: find first number in a string:
6068 int number;
6069 pcrecpp::RE re("(\\d+)");
6070 re.PartialMatch("x*100 + 20", &number);
6071 assert(number == 100);
6072
6073
6074 UTF-8 AND THE MATCHING INTERFACE
6075
6076 By default, pattern and text are plain text, one byte per character.
6077 The UTF8 flag, passed to the constructor, causes both pattern and
6078 string to be treated as UTF-8 text, still a byte stream but potentially
6079 multiple bytes per character. In practice, the text is likelier to be
6080 UTF-8 than the pattern, but the match returned may depend on the UTF8
6081 flag, so always use it when matching UTF8 text. For example, "." will
6082 match one byte normally but with UTF8 set may match up to three bytes
6083 of a multi-byte character.
6084
6085 Example:
6086 pcrecpp::RE_Options options;
6087 options.set_utf8();
6088 pcrecpp::RE re(utf8_pattern, options);
6089 re.FullMatch(utf8_string);
6090
6091 Example: using the convenience function UTF8():
6092 pcrecpp::RE re(utf8_pattern, pcrecpp::UTF8());
6093 re.FullMatch(utf8_string);
6094
6095 NOTE: The UTF8 flag is ignored if pcre was not configured with the
6096 --enable-utf8 flag.
6097
6098