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

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

Parent Directory Parent Directory | Revision Log Revision Log


Revision 227 - (show annotations)
Tue Aug 21 15:00:15 2007 UTC (12 years ago) by ph10
File MIME type: text/plain
File size: 297671 byte(s)
Add (*CR) etc.
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. They are all selected, or dese-
275 lected, by providing options to the configure script that is run before
276 the make command. The complete list of options for configure (which
277 includes the standard ones such as the selection of the installation
278 directory) can be obtained by running
279
280 ./configure --help
281
282 The following sections include descriptions of options whose names
283 begin with --enable or --disable. These settings specify changes to the
284 defaults for the configure command. Because of the way that configure
285 works, --enable and --disable always come in pairs, so the complemen-
286 tary option always exists as well, but as it specifies the default, it
287 is not described.
288
289
290 C++ SUPPORT
291
292 By default, the configure script will search for a C++ compiler and C++
293 header files. If it finds them, it automatically builds the C++ wrapper
294 library for PCRE. You can disable this by adding
295
296 --disable-cpp
297
298 to the configure command.
299
300
301 UTF-8 SUPPORT
302
303 To build PCRE with support for UTF-8 character strings, add
304
305 --enable-utf8
306
307 to the configure command. Of itself, this does not make PCRE treat
308 strings as UTF-8. As well as compiling PCRE with this option, you also
309 have have to set the PCRE_UTF8 option when you call the pcre_compile()
310 function.
311
312
313 UNICODE CHARACTER PROPERTY SUPPORT
314
315 UTF-8 support allows PCRE to process character values greater than 255
316 in the strings that it handles. On its own, however, it does not pro-
317 vide any facilities for accessing the properties of such characters. If
318 you want to be able to use the pattern escapes \P, \p, and \X, which
319 refer to Unicode character properties, you must add
320
321 --enable-unicode-properties
322
323 to the configure command. This implies UTF-8 support, even if you have
324 not explicitly requested it.
325
326 Including Unicode property support adds around 30K of tables to the
327 PCRE library. Only the general category properties such as Lu and Nd
328 are supported. Details are given in the pcrepattern documentation.
329
330
331 CODE VALUE OF NEWLINE
332
333 By default, PCRE interprets character 10 (linefeed, LF) as indicating
334 the end of a line. This is the normal newline character on Unix-like
335 systems. You can compile PCRE to use character 13 (carriage return, CR)
336 instead, by adding
337
338 --enable-newline-is-cr
339
340 to the configure command. There is also a --enable-newline-is-lf
341 option, which explicitly specifies linefeed as the newline character.
342
343 Alternatively, you can specify that line endings are to be indicated by
344 the two character sequence CRLF. If you want this, add
345
346 --enable-newline-is-crlf
347
348 to the configure command. There is a fourth option, specified by
349
350 --enable-newline-is-anycrlf
351
352 which causes PCRE to recognize any of the three sequences CR, LF, or
353 CRLF as indicating a line ending. Finally, a fifth option, specified by
354
355 --enable-newline-is-any
356
357 causes PCRE to recognize any Unicode newline sequence.
358
359 Whatever line ending convention is selected when PCRE is built can be
360 overridden when the library functions are called. At build time it is
361 conventional to use the standard for your operating system.
362
363
364 BUILDING SHARED AND STATIC LIBRARIES
365
366 The PCRE building process uses libtool to build both shared and static
367 Unix libraries by default. You can suppress one of these by adding one
368 of
369
370 --disable-shared
371 --disable-static
372
373 to the configure command, as required.
374
375
376 POSIX MALLOC USAGE
377
378 When PCRE is called through the POSIX interface (see the pcreposix doc-
379 umentation), additional working storage is required for holding the
380 pointers to capturing substrings, because PCRE requires three integers
381 per substring, whereas the POSIX interface provides only two. If the
382 number of expected substrings is small, the wrapper function uses space
383 on the stack, because this is faster than using malloc() for each call.
384 The default threshold above which the stack is no longer used is 10; it
385 can be changed by adding a setting such as
386
387 --with-posix-malloc-threshold=20
388
389 to the configure command.
390
391
392 HANDLING VERY LARGE PATTERNS
393
394 Within a compiled pattern, offset values are used to point from one
395 part to another (for example, from an opening parenthesis to an alter-
396 nation metacharacter). By default, two-byte values are used for these
397 offsets, leading to a maximum size for a compiled pattern of around
398 64K. This is sufficient to handle all but the most gigantic patterns.
399 Nevertheless, some people do want to process enormous patterns, so it
400 is possible to compile PCRE to use three-byte or four-byte offsets by
401 adding a setting such as
402
403 --with-link-size=3
404
405 to the configure command. The value given must be 2, 3, or 4. Using
406 longer offsets slows down the operation of PCRE because it has to load
407 additional bytes when handling them.
408
409
410 AVOIDING EXCESSIVE STACK USAGE
411
412 When matching with the pcre_exec() function, PCRE implements backtrack-
413 ing by making recursive calls to an internal function called match().
414 In environments where the size of the stack is limited, this can se-
415 verely limit PCRE's operation. (The Unix environment does not usually
416 suffer from this problem, but it may sometimes be necessary to increase
417 the maximum stack size. There is a discussion in the pcrestack docu-
418 mentation.) An alternative approach to recursion that uses memory from
419 the heap to remember data, instead of using recursive function calls,
420 has been implemented to work round the problem of limited stack size.
421 If you want to build a version of PCRE that works this way, add
422
423 --disable-stack-for-recursion
424
425 to the configure command. With this configuration, PCRE will use the
426 pcre_stack_malloc and pcre_stack_free variables to call memory manage-
427 ment functions. By default these point to malloc() and free(), but you
428 can replace the pointers so that your own functions are used.
429
430 Separate functions are provided rather than using pcre_malloc and
431 pcre_free because the usage is very predictable: the block sizes
432 requested are always the same, and the blocks are always freed in
433 reverse order. A calling program might be able to implement optimized
434 functions that perform better than malloc() and free(). PCRE runs
435 noticeably more slowly when built in this way. This option affects only
436 the pcre_exec() function; it is not relevant for the the
437 pcre_dfa_exec() function.
438
439
440 LIMITING PCRE RESOURCE USAGE
441
442 Internally, PCRE has a function called match(), which it calls repeat-
443 edly (sometimes recursively) when matching a pattern with the
444 pcre_exec() function. By controlling the maximum number of times this
445 function may be called during a single matching operation, a limit can
446 be placed on the resources used by a single call to pcre_exec(). The
447 limit can be changed at run time, as described in the pcreapi documen-
448 tation. The default is 10 million, but this can be changed by adding a
449 setting such as
450
451 --with-match-limit=500000
452
453 to the configure command. This setting has no effect on the
454 pcre_dfa_exec() matching function.
455
456 In some environments it is desirable to limit the depth of recursive
457 calls of match() more strictly than the total number of calls, in order
458 to restrict the maximum amount of stack (or heap, if --disable-stack-
459 for-recursion is specified) that is used. A second limit controls this;
460 it defaults to the value that is set for --with-match-limit, which
461 imposes no additional constraints. However, you can set a lower limit
462 by adding, for example,
463
464 --with-match-limit-recursion=10000
465
466 to the configure command. This value can also be overridden at run
467 time.
468
469
470 CREATING CHARACTER TABLES AT BUILD TIME
471
472 PCRE uses fixed tables for processing characters whose code values are
473 less than 256. By default, PCRE is built with a set of tables that are
474 distributed in the file pcre_chartables.c.dist. These tables are for
475 ASCII codes only. If you add
476
477 --enable-rebuild-chartables
478
479 to the configure command, the distributed tables are no longer used.
480 Instead, a program called dftables is compiled and run. This outputs
481 the source for new set of tables, created in the default locale of your
482 C runtime system. (This method of replacing the tables does not work if
483 you are cross compiling, because dftables is run on the local host. If
484 you need to create alternative tables when cross compiling, you will
485 have to do so "by hand".)
486
487
488 USING EBCDIC CODE
489
490 PCRE assumes by default that it will run in an environment where the
491 character code is ASCII (or Unicode, which is a superset of ASCII).
492 This is the case for most computer operating systems. PCRE can, how-
493 ever, be compiled to run in an EBCDIC environment by adding
494
495 --enable-ebcdic
496
497 to the configure command. This setting implies --enable-rebuild-charta-
498 bles. You should only use it if you know that you are in an EBCDIC
499 environment (for example, an IBM mainframe operating system).
500
501
502 SEE ALSO
503
504 pcreapi(3), pcre_config(3).
505
506
507 AUTHOR
508
509 Philip Hazel
510 University Computing Service
511 Cambridge CB2 3QH, England.
512
513
514 REVISION
515
516 Last updated: 30 July 2007
517 Copyright (c) 1997-2007 University of Cambridge.
518 ------------------------------------------------------------------------------
519
520
521 PCREMATCHING(3) PCREMATCHING(3)
522
523
524 NAME
525 PCRE - Perl-compatible regular expressions
526
527
528 PCRE MATCHING ALGORITHMS
529
530 This document describes the two different algorithms that are available
531 in PCRE for matching a compiled regular expression against a given sub-
532 ject string. The "standard" algorithm is the one provided by the
533 pcre_exec() function. This works in the same was as Perl's matching
534 function, and provides a Perl-compatible matching operation.
535
536 An alternative algorithm is provided by the pcre_dfa_exec() function;
537 this operates in a different way, and is not Perl-compatible. It has
538 advantages and disadvantages compared with the standard algorithm, and
539 these are described below.
540
541 When there is only one possible way in which a given subject string can
542 match a pattern, the two algorithms give the same answer. A difference
543 arises, however, when there are multiple possibilities. For example, if
544 the pattern
545
546 ^<.*>
547
548 is matched against the string
549
550 <something> <something else> <something further>
551
552 there are three possible answers. The standard algorithm finds only one
553 of them, whereas the alternative algorithm finds all three.
554
555
556 REGULAR EXPRESSIONS AS TREES
557
558 The set of strings that are matched by a regular expression can be rep-
559 resented as a tree structure. An unlimited repetition in the pattern
560 makes the tree of infinite size, but it is still a tree. Matching the
561 pattern to a given subject string (from a given starting point) can be
562 thought of as a search of the tree. There are two ways to search a
563 tree: depth-first and breadth-first, and these correspond to the two
564 matching algorithms provided by PCRE.
565
566
567 THE STANDARD MATCHING ALGORITHM
568
569 In the terminology of Jeffrey Friedl's book "Mastering Regular Expres-
570 sions", the standard algorithm is an "NFA algorithm". It conducts a
571 depth-first search of the pattern tree. That is, it proceeds along a
572 single path through the tree, checking that the subject matches what is
573 required. When there is a mismatch, the algorithm tries any alterna-
574 tives at the current point, and if they all fail, it backs up to the
575 previous branch point in the tree, and tries the next alternative
576 branch at that level. This often involves backing up (moving to the
577 left) in the subject string as well. The order in which repetition
578 branches are tried is controlled by the greedy or ungreedy nature of
579 the quantifier.
580
581 If a leaf node is reached, a matching string has been found, and at
582 that point the algorithm stops. Thus, if there is more than one possi-
583 ble match, this algorithm returns the first one that it finds. Whether
584 this is the shortest, the longest, or some intermediate length depends
585 on the way the greedy and ungreedy repetition quantifiers are specified
586 in the pattern.
587
588 Because it ends up with a single path through the tree, it is rela-
589 tively straightforward for this algorithm to keep track of the sub-
590 strings that are matched by portions of the pattern in parentheses.
591 This provides support for capturing parentheses and back references.
592
593
594 THE ALTERNATIVE MATCHING ALGORITHM
595
596 This algorithm conducts a breadth-first search of the tree. Starting
597 from the first matching point in the subject, it scans the subject
598 string from left to right, once, character by character, and as it does
599 this, it remembers all the paths through the tree that represent valid
600 matches. In Friedl's terminology, this is a kind of "DFA algorithm",
601 though it is not implemented as a traditional finite state machine (it
602 keeps multiple states active simultaneously).
603
604 The scan continues until either the end of the subject is reached, or
605 there are no more unterminated paths. At this point, terminated paths
606 represent the different matching possibilities (if there are none, the
607 match has failed). Thus, if there is more than one possible match,
608 this algorithm finds all of them, and in particular, it finds the long-
609 est. In PCRE, there is an option to stop the algorithm after the first
610 match (which is necessarily the shortest) has been found.
611
612 Note that all the matches that are found start at the same point in the
613 subject. If the pattern
614
615 cat(er(pillar)?)
616
617 is matched against the string "the caterpillar catchment", the result
618 will be the three strings "cat", "cater", and "caterpillar" that start
619 at the fourth character of the subject. The algorithm does not automat-
620 ically move on to find matches that start at later positions.
621
622 There are a number of features of PCRE regular expressions that are not
623 supported by the alternative matching algorithm. They are as follows:
624
625 1. Because the algorithm finds all possible matches, the greedy or
626 ungreedy nature of repetition quantifiers is not relevant. Greedy and
627 ungreedy quantifiers are treated in exactly the same way. However, pos-
628 sessive quantifiers can make a difference when what follows could also
629 match what is quantified, for example in a pattern like this:
630
631 ^a++\w!
632
633 This pattern matches "aaab!" but not "aaa!", which would be matched by
634 a non-possessive quantifier. Similarly, if an atomic group is present,
635 it is matched as if it were a standalone pattern at the current point,
636 and the longest match is then "locked in" for the rest of the overall
637 pattern.
638
639 2. When dealing with multiple paths through the tree simultaneously, it
640 is not straightforward to keep track of captured substrings for the
641 different matching possibilities, and PCRE's implementation of this
642 algorithm does not attempt to do this. This means that no captured sub-
643 strings are available.
644
645 3. Because no substrings are captured, back references within the pat-
646 tern are not supported, and cause errors if encountered.
647
648 4. For the same reason, conditional expressions that use a backrefer-
649 ence as the condition or test for a specific group recursion are not
650 supported.
651
652 5. Because many paths through the tree may be active, the \K escape
653 sequence, which resets the start of the match when encountered (but may
654 be on some paths and not on others), is not supported. It causes an
655 error if encountered.
656
657 6. Callouts are supported, but the value of the capture_top field is
658 always 1, and the value of the capture_last field is always -1.
659
660 7. The \C escape sequence, which (in the standard algorithm) matches a
661 single byte, even in UTF-8 mode, is not supported because the alterna-
662 tive algorithm moves through the subject string one character at a
663 time, for all active paths through the tree.
664
665 8. None of the backtracking control verbs such as (*PRUNE) are sup-
666 ported.
667
668
669 ADVANTAGES OF THE ALTERNATIVE ALGORITHM
670
671 Using the alternative matching algorithm provides the following advan-
672 tages:
673
674 1. All possible matches (at a single point in the subject) are automat-
675 ically found, and in particular, the longest match is found. To find
676 more than one match using the standard algorithm, you have to do kludgy
677 things with callouts.
678
679 2. There is much better support for partial matching. The restrictions
680 on the content of the pattern that apply when using the standard algo-
681 rithm for partial matching do not apply to the alternative algorithm.
682 For non-anchored patterns, the starting position of a partial match is
683 available.
684
685 3. Because the alternative algorithm scans the subject string just
686 once, and never needs to backtrack, it is possible to pass very long
687 subject strings to the matching function in several pieces, checking
688 for partial matching each time.
689
690
691 DISADVANTAGES OF THE ALTERNATIVE ALGORITHM
692
693 The alternative algorithm suffers from a number of disadvantages:
694
695 1. It is substantially slower than the standard algorithm. This is
696 partly because it has to search for all possible matches, but is also
697 because it is less susceptible to optimization.
698
699 2. Capturing parentheses and back references are not supported.
700
701 3. Although atomic groups are supported, their use does not provide the
702 performance advantage that it does for the standard algorithm.
703
704
705 AUTHOR
706
707 Philip Hazel
708 University Computing Service
709 Cambridge CB2 3QH, England.
710
711
712 REVISION
713
714 Last updated: 08 August 2007
715 Copyright (c) 1997-2007 University of Cambridge.
716 ------------------------------------------------------------------------------
717
718
719 PCREAPI(3) PCREAPI(3)
720
721
722 NAME
723 PCRE - Perl-compatible regular expressions
724
725
726 PCRE NATIVE API
727
728 #include <pcre.h>
729
730 pcre *pcre_compile(const char *pattern, int options,
731 const char **errptr, int *erroffset,
732 const unsigned char *tableptr);
733
734 pcre *pcre_compile2(const char *pattern, int options,
735 int *errorcodeptr,
736 const char **errptr, int *erroffset,
737 const unsigned char *tableptr);
738
739 pcre_extra *pcre_study(const pcre *code, int options,
740 const char **errptr);
741
742 int pcre_exec(const pcre *code, const pcre_extra *extra,
743 const char *subject, int length, int startoffset,
744 int options, int *ovector, int ovecsize);
745
746 int pcre_dfa_exec(const pcre *code, const pcre_extra *extra,
747 const char *subject, int length, int startoffset,
748 int options, int *ovector, int ovecsize,
749 int *workspace, int wscount);
750
751 int pcre_copy_named_substring(const pcre *code,
752 const char *subject, int *ovector,
753 int stringcount, const char *stringname,
754 char *buffer, int buffersize);
755
756 int pcre_copy_substring(const char *subject, int *ovector,
757 int stringcount, int stringnumber, char *buffer,
758 int buffersize);
759
760 int pcre_get_named_substring(const pcre *code,
761 const char *subject, int *ovector,
762 int stringcount, const char *stringname,
763 const char **stringptr);
764
765 int pcre_get_stringnumber(const pcre *code,
766 const char *name);
767
768 int pcre_get_stringtable_entries(const pcre *code,
769 const char *name, char **first, char **last);
770
771 int pcre_get_substring(const char *subject, int *ovector,
772 int stringcount, int stringnumber,
773 const char **stringptr);
774
775 int pcre_get_substring_list(const char *subject,
776 int *ovector, int stringcount, const char ***listptr);
777
778 void pcre_free_substring(const char *stringptr);
779
780 void pcre_free_substring_list(const char **stringptr);
781
782 const unsigned char *pcre_maketables(void);
783
784 int pcre_fullinfo(const pcre *code, const pcre_extra *extra,
785 int what, void *where);
786
787 int pcre_info(const pcre *code, int *optptr, int *firstcharptr);
788
789 int pcre_refcount(pcre *code, int adjust);
790
791 int pcre_config(int what, void *where);
792
793 char *pcre_version(void);
794
795 void *(*pcre_malloc)(size_t);
796
797 void (*pcre_free)(void *);
798
799 void *(*pcre_stack_malloc)(size_t);
800
801 void (*pcre_stack_free)(void *);
802
803 int (*pcre_callout)(pcre_callout_block *);
804
805
806 PCRE API OVERVIEW
807
808 PCRE has its own native API, which is described in this document. There
809 are also some wrapper functions that correspond to the POSIX regular
810 expression API. These are described in the pcreposix documentation.
811 Both of these APIs define a set of C function calls. A C++ wrapper is
812 distributed with PCRE. It is documented in the pcrecpp page.
813
814 The native API C function prototypes are defined in the header file
815 pcre.h, and on Unix systems the library itself is called libpcre. It
816 can normally be accessed by adding -lpcre to the command for linking an
817 application that uses PCRE. The header file defines the macros
818 PCRE_MAJOR and PCRE_MINOR to contain the major and minor release num-
819 bers for the library. Applications can use these to include support
820 for different releases of PCRE.
821
822 The functions pcre_compile(), pcre_compile2(), pcre_study(), and
823 pcre_exec() are used for compiling and matching regular expressions in
824 a Perl-compatible manner. A sample program that demonstrates the sim-
825 plest way of using them is provided in the file called pcredemo.c in
826 the source distribution. The pcresample documentation describes how to
827 run it.
828
829 A second matching function, pcre_dfa_exec(), which is not Perl-compati-
830 ble, is also provided. This uses a different algorithm for the match-
831 ing. The alternative algorithm finds all possible matches (at a given
832 point in the subject), and scans the subject just once. However, this
833 algorithm does not return captured substrings. A description of the two
834 matching algorithms and their advantages and disadvantages is given in
835 the pcrematching documentation.
836
837 In addition to the main compiling and matching functions, there are
838 convenience functions for extracting captured substrings from a subject
839 string that is matched by pcre_exec(). They are:
840
841 pcre_copy_substring()
842 pcre_copy_named_substring()
843 pcre_get_substring()
844 pcre_get_named_substring()
845 pcre_get_substring_list()
846 pcre_get_stringnumber()
847 pcre_get_stringtable_entries()
848
849 pcre_free_substring() and pcre_free_substring_list() are also provided,
850 to free the memory used for extracted strings.
851
852 The function pcre_maketables() is used to build a set of character
853 tables in the current locale for passing to pcre_compile(),
854 pcre_exec(), or pcre_dfa_exec(). This is an optional facility that is
855 provided for specialist use. Most commonly, no special tables are
856 passed, in which case internal tables that are generated when PCRE is
857 built are used.
858
859 The function pcre_fullinfo() is used to find out information about a
860 compiled pattern; pcre_info() is an obsolete version that returns only
861 some of the available information, but is retained for backwards com-
862 patibility. The function pcre_version() returns a pointer to a string
863 containing the version of PCRE and its date of release.
864
865 The function pcre_refcount() maintains a reference count in a data
866 block containing a compiled pattern. This is provided for the benefit
867 of object-oriented applications.
868
869 The global variables pcre_malloc and pcre_free initially contain the
870 entry points of the standard malloc() and free() functions, respec-
871 tively. PCRE calls the memory management functions via these variables,
872 so a calling program can replace them if it wishes to intercept the
873 calls. This should be done before calling any PCRE functions.
874
875 The global variables pcre_stack_malloc and pcre_stack_free are also
876 indirections to memory management functions. These special functions
877 are used only when PCRE is compiled to use the heap for remembering
878 data, instead of recursive function calls, when running the pcre_exec()
879 function. See the pcrebuild documentation for details of how to do
880 this. It is a non-standard way of building PCRE, for use in environ-
881 ments that have limited stacks. Because of the greater use of memory
882 management, it runs more slowly. Separate functions are provided so
883 that special-purpose external code can be used for this case. When
884 used, these functions are always called in a stack-like manner (last
885 obtained, first freed), and always for memory blocks of the same size.
886 There is a discussion about PCRE's stack usage in the pcrestack docu-
887 mentation.
888
889 The global variable pcre_callout initially contains NULL. It can be set
890 by the caller to a "callout" function, which PCRE will then call at
891 specified points during a matching operation. Details are given in the
892 pcrecallout documentation.
893
894
895 NEWLINES
896
897 PCRE supports five different conventions for indicating line breaks in
898 strings: a single CR (carriage return) character, a single LF (line-
899 feed) character, the two-character sequence CRLF, any of the three pre-
900 ceding, or any Unicode newline sequence. The Unicode newline sequences
901 are the three just mentioned, plus the single characters VT (vertical
902 tab, U+000B), FF (formfeed, U+000C), NEL (next line, U+0085), LS (line
903 separator, U+2028), and PS (paragraph separator, U+2029).
904
905 Each of the first three conventions is used by at least one operating
906 system as its standard newline sequence. When PCRE is built, a default
907 can be specified. The default default is LF, which is the Unix stan-
908 dard. When PCRE is run, the default can be overridden, either when a
909 pattern is compiled, or when it is matched.
910
911 At compile time, the newline convention can be specified by the options
912 argument of pcre_compile(), or it can be specified by special text at
913 the start of the pattern itself; this overrides any other settings. See
914 the pcrepattern page for details of the special character sequences.
915
916 In the PCRE documentation the word "newline" is used to mean "the char-
917 acter or pair of characters that indicate a line break". The choice of
918 newline convention affects the handling of the dot, circumflex, and
919 dollar metacharacters, the handling of #-comments in /x mode, and, when
920 CRLF is a recognized line ending sequence, the match position advance-
921 ment for a non-anchored pattern. There is more detail about this in the
922 section on pcre_exec() options below. The choice of newline convention
923 does not affect the interpretation of the \n or \r escape sequences.
924
925
926 MULTITHREADING
927
928 The PCRE functions can be used in multi-threading applications, with
929 the proviso that the memory management functions pointed to by
930 pcre_malloc, pcre_free, pcre_stack_malloc, and pcre_stack_free, and the
931 callout function pointed to by pcre_callout, are shared by all threads.
932
933 The compiled form of a regular expression is not altered during match-
934 ing, so the same compiled pattern can safely be used by several threads
935 at once.
936
937
938 SAVING PRECOMPILED PATTERNS FOR LATER USE
939
940 The compiled form of a regular expression can be saved and re-used at a
941 later time, possibly by a different program, and even on a host other
942 than the one on which it was compiled. Details are given in the
943 pcreprecompile documentation. However, compiling a regular expression
944 with one version of PCRE for use with a different version is not guar-
945 anteed to work and may cause crashes.
946
947
948 CHECKING BUILD-TIME OPTIONS
949
950 int pcre_config(int what, void *where);
951
952 The function pcre_config() makes it possible for a PCRE client to dis-
953 cover which optional features have been compiled into the PCRE library.
954 The pcrebuild documentation has more details about these optional fea-
955 tures.
956
957 The first argument for pcre_config() is an integer, specifying which
958 information is required; the second argument is a pointer to a variable
959 into which the information is placed. The following information is
960 available:
961
962 PCRE_CONFIG_UTF8
963
964 The output is an integer that is set to one if UTF-8 support is avail-
965 able; otherwise it is set to zero.
966
967 PCRE_CONFIG_UNICODE_PROPERTIES
968
969 The output is an integer that is set to one if support for Unicode
970 character properties is available; otherwise it is set to zero.
971
972 PCRE_CONFIG_NEWLINE
973
974 The output is an integer whose value specifies the default character
975 sequence that is recognized as meaning "newline". The four values that
976 are supported are: 10 for LF, 13 for CR, 3338 for CRLF, -2 for ANYCRLF,
977 and -1 for ANY. The default should normally be the standard sequence
978 for your operating system.
979
980 PCRE_CONFIG_LINK_SIZE
981
982 The output is an integer that contains the number of bytes used for
983 internal linkage in compiled regular expressions. The value is 2, 3, or
984 4. Larger values allow larger regular expressions to be compiled, at
985 the expense of slower matching. The default value of 2 is sufficient
986 for all but the most massive patterns, since it allows the compiled
987 pattern to be up to 64K in size.
988
989 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
990
991 The output is an integer that contains the threshold above which the
992 POSIX interface uses malloc() for output vectors. Further details are
993 given in the pcreposix documentation.
994
995 PCRE_CONFIG_MATCH_LIMIT
996
997 The output is an integer that gives the default limit for the number of
998 internal matching function calls in a pcre_exec() execution. Further
999 details are given with pcre_exec() below.
1000
1001 PCRE_CONFIG_MATCH_LIMIT_RECURSION
1002
1003 The output is an integer that gives the default limit for the depth of
1004 recursion when calling the internal matching function in a pcre_exec()
1005 execution. Further details are given with pcre_exec() below.
1006
1007 PCRE_CONFIG_STACKRECURSE
1008
1009 The output is an integer that is set to one if internal recursion when
1010 running pcre_exec() is implemented by recursive function calls that use
1011 the stack to remember their state. This is the usual way that PCRE is
1012 compiled. The output is zero if PCRE was compiled to use blocks of data
1013 on the heap instead of recursive function calls. In this case,
1014 pcre_stack_malloc and pcre_stack_free are called to manage memory
1015 blocks on the heap, thus avoiding the use of the stack.
1016
1017
1018 COMPILING A PATTERN
1019
1020 pcre *pcre_compile(const char *pattern, int options,
1021 const char **errptr, int *erroffset,
1022 const unsigned char *tableptr);
1023
1024 pcre *pcre_compile2(const char *pattern, int options,
1025 int *errorcodeptr,
1026 const char **errptr, int *erroffset,
1027 const unsigned char *tableptr);
1028
1029 Either of the functions pcre_compile() or pcre_compile2() can be called
1030 to compile a pattern into an internal form. The only difference between
1031 the two interfaces is that pcre_compile2() has an additional argument,
1032 errorcodeptr, via which a numerical error code can be returned.
1033
1034 The pattern is a C string terminated by a binary zero, and is passed in
1035 the pattern argument. A pointer to a single block of memory that is
1036 obtained via pcre_malloc is returned. This contains the compiled code
1037 and related data. The pcre type is defined for the returned block; this
1038 is a typedef for a structure whose contents are not externally defined.
1039 It is up to the caller to free the memory (via pcre_free) when it is no
1040 longer required.
1041
1042 Although the compiled code of a PCRE regex is relocatable, that is, it
1043 does not depend on memory location, the complete pcre data block is not
1044 fully relocatable, because it may contain a copy of the tableptr argu-
1045 ment, which is an address (see below).
1046
1047 The options argument contains various bit settings that affect the com-
1048 pilation. It should be zero if no options are required. The available
1049 options are described below. Some of them, in particular, those that
1050 are compatible with Perl, can also be set and unset from within the
1051 pattern (see the detailed description in the pcrepattern documenta-
1052 tion). For these options, the contents of the options argument speci-
1053 fies their initial settings at the start of compilation and execution.
1054 The PCRE_ANCHORED and PCRE_NEWLINE_xxx options can be set at the time
1055 of matching as well as at compile time.
1056
1057 If errptr is NULL, pcre_compile() returns NULL immediately. Otherwise,
1058 if compilation of a pattern fails, pcre_compile() returns NULL, and
1059 sets the variable pointed to by errptr to point to a textual error mes-
1060 sage. This is a static string that is part of the library. You must not
1061 try to free it. The offset from the start of the pattern to the charac-
1062 ter where the error was discovered is placed in the variable pointed to
1063 by erroffset, which must not be NULL. If it is, an immediate error is
1064 given.
1065
1066 If pcre_compile2() is used instead of pcre_compile(), and the error-
1067 codeptr argument is not NULL, a non-zero error code number is returned
1068 via this argument in the event of an error. This is in addition to the
1069 textual error message. Error codes and messages are listed below.
1070
1071 If the final argument, tableptr, is NULL, PCRE uses a default set of
1072 character tables that are built when PCRE is compiled, using the
1073 default C locale. Otherwise, tableptr must be an address that is the
1074 result of a call to pcre_maketables(). This value is stored with the
1075 compiled pattern, and used again by pcre_exec(), unless another table
1076 pointer is passed to it. For more discussion, see the section on locale
1077 support below.
1078
1079 This code fragment shows a typical straightforward call to pcre_com-
1080 pile():
1081
1082 pcre *re;
1083 const char *error;
1084 int erroffset;
1085 re = pcre_compile(
1086 "^A.*Z", /* the pattern */
1087 0, /* default options */
1088 &error, /* for error message */
1089 &erroffset, /* for error offset */
1090 NULL); /* use default character tables */
1091
1092 The following names for option bits are defined in the pcre.h header
1093 file:
1094
1095 PCRE_ANCHORED
1096
1097 If this bit is set, the pattern is forced to be "anchored", that is, it
1098 is constrained to match only at the first matching point in the string
1099 that is being searched (the "subject string"). This effect can also be
1100 achieved by appropriate constructs in the pattern itself, which is the
1101 only way to do it in Perl.
1102
1103 PCRE_AUTO_CALLOUT
1104
1105 If this bit is set, pcre_compile() automatically inserts callout items,
1106 all with number 255, before each pattern item. For discussion of the
1107 callout facility, see the pcrecallout documentation.
1108
1109 PCRE_CASELESS
1110
1111 If this bit is set, letters in the pattern match both upper and lower
1112 case letters. It is equivalent to Perl's /i option, and it can be
1113 changed within a pattern by a (?i) option setting. In UTF-8 mode, PCRE
1114 always understands the concept of case for characters whose values are
1115 less than 128, so caseless matching is always possible. For characters
1116 with higher values, the concept of case is supported if PCRE is com-
1117 piled with Unicode property support, but not otherwise. If you want to
1118 use caseless matching for characters 128 and above, you must ensure
1119 that PCRE is compiled with Unicode property support as well as with
1120 UTF-8 support.
1121
1122 PCRE_DOLLAR_ENDONLY
1123
1124 If this bit is set, a dollar metacharacter in the pattern matches only
1125 at the end of the subject string. Without this option, a dollar also
1126 matches immediately before a newline at the end of the string (but not
1127 before any other newlines). The PCRE_DOLLAR_ENDONLY option is ignored
1128 if PCRE_MULTILINE is set. There is no equivalent to this option in
1129 Perl, and no way to set it within a pattern.
1130
1131 PCRE_DOTALL
1132
1133 If this bit is set, a dot metacharater in the pattern matches all char-
1134 acters, including those that indicate newline. Without it, a dot does
1135 not match when the current position is at a newline. This option is
1136 equivalent to Perl's /s option, and it can be changed within a pattern
1137 by a (?s) option setting. A negative class such as [^a] always matches
1138 newline characters, independent of the setting of this option.
1139
1140 PCRE_DUPNAMES
1141
1142 If this bit is set, names used to identify capturing subpatterns need
1143 not be unique. This can be helpful for certain types of pattern when it
1144 is known that only one instance of the named subpattern can ever be
1145 matched. There are more details of named subpatterns below; see also
1146 the pcrepattern documentation.
1147
1148 PCRE_EXTENDED
1149
1150 If this bit is set, whitespace data characters in the pattern are
1151 totally ignored except when escaped or inside a character class. White-
1152 space does not include the VT character (code 11). In addition, charac-
1153 ters between an unescaped # outside a character class and the next new-
1154 line, inclusive, are also ignored. This is equivalent to Perl's /x
1155 option, and it can be changed within a pattern by a (?x) option set-
1156 ting.
1157
1158 This option makes it possible to include comments inside complicated
1159 patterns. Note, however, that this applies only to data characters.
1160 Whitespace characters may never appear within special character
1161 sequences in a pattern, for example within the sequence (?( which
1162 introduces a conditional subpattern.
1163
1164 PCRE_EXTRA
1165
1166 This option was invented in order to turn on additional functionality
1167 of PCRE that is incompatible with Perl, but it is currently of very
1168 little use. When set, any backslash in a pattern that is followed by a
1169 letter that has no special meaning causes an error, thus reserving
1170 these combinations for future expansion. By default, as in Perl, a
1171 backslash followed by a letter with no special meaning is treated as a
1172 literal. (Perl can, however, be persuaded to give a warning for this.)
1173 There are at present no other features controlled by this option. It
1174 can also be set by a (?X) option setting within a pattern.
1175
1176 PCRE_FIRSTLINE
1177
1178 If this option is set, an unanchored pattern is required to match
1179 before or at the first newline in the subject string, though the
1180 matched text may continue over the newline.
1181
1182 PCRE_MULTILINE
1183
1184 By default, PCRE treats the subject string as consisting of a single
1185 line of characters (even if it actually contains newlines). The "start
1186 of line" metacharacter (^) matches only at the start of the string,
1187 while the "end of line" metacharacter ($) matches only at the end of
1188 the string, or before a terminating newline (unless PCRE_DOLLAR_ENDONLY
1189 is set). This is the same as Perl.
1190
1191 When PCRE_MULTILINE it is set, the "start of line" and "end of line"
1192 constructs match immediately following or immediately before internal
1193 newlines in the subject string, respectively, as well as at the very
1194 start and end. This is equivalent to Perl's /m option, and it can be
1195 changed within a pattern by a (?m) option setting. If there are no new-
1196 lines in a subject string, or no occurrences of ^ or $ in a pattern,
1197 setting PCRE_MULTILINE has no effect.
1198
1199 PCRE_NEWLINE_CR
1200 PCRE_NEWLINE_LF
1201 PCRE_NEWLINE_CRLF
1202 PCRE_NEWLINE_ANYCRLF
1203 PCRE_NEWLINE_ANY
1204
1205 These options override the default newline definition that was chosen
1206 when PCRE was built. Setting the first or the second specifies that a
1207 newline is indicated by a single character (CR or LF, respectively).
1208 Setting PCRE_NEWLINE_CRLF specifies that a newline is indicated by the
1209 two-character CRLF sequence. Setting PCRE_NEWLINE_ANYCRLF specifies
1210 that any of the three preceding sequences should be recognized. Setting
1211 PCRE_NEWLINE_ANY specifies that any Unicode newline sequence should be
1212 recognized. The Unicode newline sequences are the three just mentioned,
1213 plus the single characters VT (vertical tab, U+000B), FF (formfeed,
1214 U+000C), NEL (next line, U+0085), LS (line separator, U+2028), and PS
1215 (paragraph separator, U+2029). The last two are recognized only in
1216 UTF-8 mode.
1217
1218 The newline setting in the options word uses three bits that are
1219 treated as a number, giving eight possibilities. Currently only six are
1220 used (default plus the five values above). This means that if you set
1221 more than one newline option, the combination may or may not be sensi-
1222 ble. For example, PCRE_NEWLINE_CR with PCRE_NEWLINE_LF is equivalent to
1223 PCRE_NEWLINE_CRLF, but other combinations may yield unused numbers and
1224 cause an error.
1225
1226 The only time that a line break is specially recognized when compiling
1227 a pattern is if PCRE_EXTENDED is set, and an unescaped # outside a
1228 character class is encountered. This indicates a comment that lasts
1229 until after the next line break sequence. In other circumstances, line
1230 break sequences are treated as literal data, except that in
1231 PCRE_EXTENDED mode, both CR and LF are treated as whitespace characters
1232 and are therefore ignored.
1233
1234 The newline option that is set at compile time becomes the default that
1235 is used for pcre_exec() and pcre_dfa_exec(), but it can be overridden.
1236
1237 PCRE_NO_AUTO_CAPTURE
1238
1239 If this option is set, it disables the use of numbered capturing paren-
1240 theses in the pattern. Any opening parenthesis that is not followed by
1241 ? behaves as if it were followed by ?: but named parentheses can still
1242 be used for capturing (and they acquire numbers in the usual way).
1243 There is no equivalent of this option in Perl.
1244
1245 PCRE_UNGREEDY
1246
1247 This option inverts the "greediness" of the quantifiers so that they
1248 are not greedy by default, but become greedy if followed by "?". It is
1249 not compatible with Perl. It can also be set by a (?U) option setting
1250 within the pattern.
1251
1252 PCRE_UTF8
1253
1254 This option causes PCRE to regard both the pattern and the subject as
1255 strings of UTF-8 characters instead of single-byte character strings.
1256 However, it is available only when PCRE is built to include UTF-8 sup-
1257 port. If not, the use of this option provokes an error. Details of how
1258 this option changes the behaviour of PCRE are given in the section on
1259 UTF-8 support in the main pcre page.
1260
1261 PCRE_NO_UTF8_CHECK
1262
1263 When PCRE_UTF8 is set, the validity of the pattern as a UTF-8 string is
1264 automatically checked. There is a discussion about the validity of
1265 UTF-8 strings in the main pcre page. If an invalid UTF-8 sequence of
1266 bytes is found, pcre_compile() returns an error. If you already know
1267 that your pattern is valid, and you want to skip this check for perfor-
1268 mance reasons, you can set the PCRE_NO_UTF8_CHECK option. When it is
1269 set, the effect of passing an invalid UTF-8 string as a pattern is
1270 undefined. It may cause your program to crash. Note that this option
1271 can also be passed to pcre_exec() and pcre_dfa_exec(), to suppress the
1272 UTF-8 validity checking of subject strings.
1273
1274
1275 COMPILATION ERROR CODES
1276
1277 The following table lists the error codes than may be returned by
1278 pcre_compile2(), along with the error messages that may be returned by
1279 both compiling functions. As PCRE has developed, some error codes have
1280 fallen out of use. To avoid confusion, they have not been re-used.
1281
1282 0 no error
1283 1 \ at end of pattern
1284 2 \c at end of pattern
1285 3 unrecognized character follows \
1286 4 numbers out of order in {} quantifier
1287 5 number too big in {} quantifier
1288 6 missing terminating ] for character class
1289 7 invalid escape sequence in character class
1290 8 range out of order in character class
1291 9 nothing to repeat
1292 10 [this code is not in use]
1293 11 internal error: unexpected repeat
1294 12 unrecognized character after (?
1295 13 POSIX named classes are supported only within a class
1296 14 missing )
1297 15 reference to non-existent subpattern
1298 16 erroffset passed as NULL
1299 17 unknown option bit(s) set
1300 18 missing ) after comment
1301 19 [this code is not in use]
1302 20 regular expression too large
1303 21 failed to get memory
1304 22 unmatched parentheses
1305 23 internal error: code overflow
1306 24 unrecognized character after (?<
1307 25 lookbehind assertion is not fixed length
1308 26 malformed number or name after (?(
1309 27 conditional group contains more than two branches
1310 28 assertion expected after (?(
1311 29 (?R or (?[+-]digits must be followed by )
1312 30 unknown POSIX class name
1313 31 POSIX collating elements are not supported
1314 32 this version of PCRE is not compiled with PCRE_UTF8 support
1315 33 [this code is not in use]
1316 34 character value in \x{...} sequence is too large
1317 35 invalid condition (?(0)
1318 36 \C not allowed in lookbehind assertion
1319 37 PCRE does not support \L, \l, \N, \U, or \u
1320 38 number after (?C is > 255
1321 39 closing ) for (?C expected
1322 40 recursive call could loop indefinitely
1323 41 unrecognized character after (?P
1324 42 syntax error in subpattern name (missing terminator)
1325 43 two named subpatterns have the same name
1326 44 invalid UTF-8 string
1327 45 support for \P, \p, and \X has not been compiled
1328 46 malformed \P or \p sequence
1329 47 unknown property name after \P or \p
1330 48 subpattern name is too long (maximum 32 characters)
1331 49 too many named subpatterns (maximum 10,000)
1332 50 [this code is not in use]
1333 51 octal value is greater than \377 (not in UTF-8 mode)
1334 52 internal error: overran compiling workspace
1335 53 internal error: previously-checked referenced subpattern not
1336 found
1337 54 DEFINE group contains more than one branch
1338 55 repeating a DEFINE group is not allowed
1339 56 inconsistent NEWLINE options"
1340 57 \g is not followed by a braced name or an optionally braced
1341 non-zero number
1342 58 (?+ or (?- or (?(+ or (?(- must be followed by a non-zero number
1343
1344
1345 STUDYING A PATTERN
1346
1347 pcre_extra *pcre_study(const pcre *code, int options
1348 const char **errptr);
1349
1350 If a compiled pattern is going to be used several times, it is worth
1351 spending more time analyzing it in order to speed up the time taken for
1352 matching. The function pcre_study() takes a pointer to a compiled pat-
1353 tern as its first argument. If studying the pattern produces additional
1354 information that will help speed up matching, pcre_study() returns a
1355 pointer to a pcre_extra block, in which the study_data field points to
1356 the results of the study.
1357
1358 The returned value from pcre_study() can be passed directly to
1359 pcre_exec(). However, a pcre_extra block also contains other fields
1360 that can be set by the caller before the block is passed; these are
1361 described below in the section on matching a pattern.
1362
1363 If studying the pattern does not produce any additional information
1364 pcre_study() returns NULL. In that circumstance, if the calling program
1365 wants to pass any of the other fields to pcre_exec(), it must set up
1366 its own pcre_extra block.
1367
1368 The second argument of pcre_study() contains option bits. At present,
1369 no options are defined, and this argument should always be zero.
1370
1371 The third argument for pcre_study() is a pointer for an error message.
1372 If studying succeeds (even if no data is returned), the variable it
1373 points to is set to NULL. Otherwise it is set to point to a textual
1374 error message. This is a static string that is part of the library. You
1375 must not try to free it. You should test the error pointer for NULL
1376 after calling pcre_study(), to be sure that it has run successfully.
1377
1378 This is a typical call to pcre_study():
1379
1380 pcre_extra *pe;
1381 pe = pcre_study(
1382 re, /* result of pcre_compile() */
1383 0, /* no options exist */
1384 &error); /* set to NULL or points to a message */
1385
1386 At present, studying a pattern is useful only for non-anchored patterns
1387 that do not have a single fixed starting character. A bitmap of possi-
1388 ble starting bytes is created.
1389
1390
1391 LOCALE SUPPORT
1392
1393 PCRE handles caseless matching, and determines whether characters are
1394 letters, digits, or whatever, by reference to a set of tables, indexed
1395 by character value. When running in UTF-8 mode, this applies only to
1396 characters with codes less than 128. Higher-valued codes never match
1397 escapes such as \w or \d, but can be tested with \p if PCRE is built
1398 with Unicode character property support. The use of locales with Uni-
1399 code is discouraged. If you are handling characters with codes greater
1400 than 128, you should either use UTF-8 and Unicode, or use locales, but
1401 not try to mix the two.
1402
1403 PCRE contains an internal set of tables that are used when the final
1404 argument of pcre_compile() is NULL. These are sufficient for many
1405 applications. Normally, the internal tables recognize only ASCII char-
1406 acters. However, when PCRE is built, it is possible to cause the inter-
1407 nal tables to be rebuilt in the default "C" locale of the local system,
1408 which may cause them to be different.
1409
1410 The internal tables can always be overridden by tables supplied by the
1411 application that calls PCRE. These may be created in a different locale
1412 from the default. As more and more applications change to using Uni-
1413 code, the need for this locale support is expected to die away.
1414
1415 External tables are built by calling the pcre_maketables() function,
1416 which has no arguments, in the relevant locale. The result can then be
1417 passed to pcre_compile() or pcre_exec() as often as necessary. For
1418 example, to build and use tables that are appropriate for the French
1419 locale (where accented characters with values greater than 128 are
1420 treated as letters), the following code could be used:
1421
1422 setlocale(LC_CTYPE, "fr_FR");
1423 tables = pcre_maketables();
1424 re = pcre_compile(..., tables);
1425
1426 The locale name "fr_FR" is used on Linux and other Unix-like systems;
1427 if you are using Windows, the name for the French locale is "french".
1428
1429 When pcre_maketables() runs, the tables are built in memory that is
1430 obtained via pcre_malloc. It is the caller's responsibility to ensure
1431 that the memory containing the tables remains available for as long as
1432 it is needed.
1433
1434 The pointer that is passed to pcre_compile() is saved with the compiled
1435 pattern, and the same tables are used via this pointer by pcre_study()
1436 and normally also by pcre_exec(). Thus, by default, for any single pat-
1437 tern, compilation, studying and matching all happen in the same locale,
1438 but different patterns can be compiled in different locales.
1439
1440 It is possible to pass a table pointer or NULL (indicating the use of
1441 the internal tables) to pcre_exec(). Although not intended for this
1442 purpose, this facility could be used to match a pattern in a different
1443 locale from the one in which it was compiled. Passing table pointers at
1444 run time is discussed below in the section on matching a pattern.
1445
1446
1447 INFORMATION ABOUT A PATTERN
1448
1449 int pcre_fullinfo(const pcre *code, const pcre_extra *extra,
1450 int what, void *where);
1451
1452 The pcre_fullinfo() function returns information about a compiled pat-
1453 tern. It replaces the obsolete pcre_info() function, which is neverthe-
1454 less retained for backwards compability (and is documented below).
1455
1456 The first argument for pcre_fullinfo() is a pointer to the compiled
1457 pattern. The second argument is the result of pcre_study(), or NULL if
1458 the pattern was not studied. The third argument specifies which piece
1459 of information is required, and the fourth argument is a pointer to a
1460 variable to receive the data. The yield of the function is zero for
1461 success, or one of the following negative numbers:
1462
1463 PCRE_ERROR_NULL the argument code was NULL
1464 the argument where was NULL
1465 PCRE_ERROR_BADMAGIC the "magic number" was not found
1466 PCRE_ERROR_BADOPTION the value of what was invalid
1467
1468 The "magic number" is placed at the start of each compiled pattern as
1469 an simple check against passing an arbitrary memory pointer. Here is a
1470 typical call of pcre_fullinfo(), to obtain the length of the compiled
1471 pattern:
1472
1473 int rc;
1474 size_t length;
1475 rc = pcre_fullinfo(
1476 re, /* result of pcre_compile() */
1477 pe, /* result of pcre_study(), or NULL */
1478 PCRE_INFO_SIZE, /* what is required */
1479 &length); /* where to put the data */
1480
1481 The possible values for the third argument are defined in pcre.h, and
1482 are as follows:
1483
1484 PCRE_INFO_BACKREFMAX
1485
1486 Return the number of the highest back reference in the pattern. The
1487 fourth argument should point to an int variable. Zero is returned if
1488 there are no back references.
1489
1490 PCRE_INFO_CAPTURECOUNT
1491
1492 Return the number of capturing subpatterns in the pattern. The fourth
1493 argument should point to an int variable.
1494
1495 PCRE_INFO_DEFAULT_TABLES
1496
1497 Return a pointer to the internal default character tables within PCRE.
1498 The fourth argument should point to an unsigned char * variable. This
1499 information call is provided for internal use by the pcre_study() func-
1500 tion. External callers can cause PCRE to use its internal tables by
1501 passing a NULL table pointer.
1502
1503 PCRE_INFO_FIRSTBYTE
1504
1505 Return information about the first byte of any matched string, for a
1506 non-anchored pattern. The fourth argument should point to an int vari-
1507 able. (This option used to be called PCRE_INFO_FIRSTCHAR; the old name
1508 is still recognized for backwards compatibility.)
1509
1510 If there is a fixed first byte, for example, from a pattern such as
1511 (cat|cow|coyote), its value is returned. Otherwise, if either
1512
1513 (a) the pattern was compiled with the PCRE_MULTILINE option, and every
1514 branch starts with "^", or
1515
1516 (b) every branch of the pattern starts with ".*" and PCRE_DOTALL is not
1517 set (if it were set, the pattern would be anchored),
1518
1519 -1 is returned, indicating that the pattern matches only at the start
1520 of a subject string or after any newline within the string. Otherwise
1521 -2 is returned. For anchored patterns, -2 is returned.
1522
1523 PCRE_INFO_FIRSTTABLE
1524
1525 If the pattern was studied, and this resulted in the construction of a
1526 256-bit table indicating a fixed set of bytes for the first byte in any
1527 matching string, a pointer to the table is returned. Otherwise NULL is
1528 returned. The fourth argument should point to an unsigned char * vari-
1529 able.
1530
1531 PCRE_INFO_HASCRORLF
1532
1533 Return 1 if the pattern contains any explicit matches for CR or LF
1534 characters, otherwise 0. The fourth argument should point to an int
1535 variable.
1536
1537 PCRE_INFO_JCHANGED
1538
1539 Return 1 if the (?J) option setting is used in the pattern, otherwise
1540 0. The fourth argument should point to an int variable. The (?J) inter-
1541 nal option setting changes the local PCRE_DUPNAMES option.
1542
1543 PCRE_INFO_LASTLITERAL
1544
1545 Return the value of the rightmost literal byte that must exist in any
1546 matched string, other than at its start, if such a byte has been
1547 recorded. The fourth argument should point to an int variable. If there
1548 is no such byte, -1 is returned. For anchored patterns, a last literal
1549 byte is recorded only if it follows something of variable length. For
1550 example, for the pattern /^a\d+z\d+/ the returned value is "z", but for
1551 /^a\dz\d/ the returned value is -1.
1552
1553 PCRE_INFO_NAMECOUNT
1554 PCRE_INFO_NAMEENTRYSIZE
1555 PCRE_INFO_NAMETABLE
1556
1557 PCRE supports the use of named as well as numbered capturing parenthe-
1558 ses. The names are just an additional way of identifying the parenthe-
1559 ses, which still acquire numbers. Several convenience functions such as
1560 pcre_get_named_substring() are provided for extracting captured sub-
1561 strings by name. It is also possible to extract the data directly, by
1562 first converting the name to a number in order to access the correct
1563 pointers in the output vector (described with pcre_exec() below). To do
1564 the conversion, you need to use the name-to-number map, which is
1565 described by these three values.
1566
1567 The map consists of a number of fixed-size entries. PCRE_INFO_NAMECOUNT
1568 gives the number of entries, and PCRE_INFO_NAMEENTRYSIZE gives the size
1569 of each entry; both of these return an int value. The entry size
1570 depends on the length of the longest name. PCRE_INFO_NAMETABLE returns
1571 a pointer to the first entry of the table (a pointer to char). The
1572 first two bytes of each entry are the number of the capturing parenthe-
1573 sis, most significant byte first. The rest of the entry is the corre-
1574 sponding name, zero terminated. The names are in alphabetical order.
1575 When PCRE_DUPNAMES is set, duplicate names are in order of their paren-
1576 theses numbers. For example, consider the following pattern (assume
1577 PCRE_EXTENDED is set, so white space - including newlines - is
1578 ignored):
1579
1580 (?<date> (?<year>(\d\d)?\d\d) -
1581 (?<month>\d\d) - (?<day>\d\d) )
1582
1583 There are four named subpatterns, so the table has four entries, and
1584 each entry in the table is eight bytes long. The table is as follows,
1585 with non-printing bytes shows in hexadecimal, and undefined bytes shown
1586 as ??:
1587
1588 00 01 d a t e 00 ??
1589 00 05 d a y 00 ?? ??
1590 00 04 m o n t h 00
1591 00 02 y e a r 00 ??
1592
1593 When writing code to extract data from named subpatterns using the
1594 name-to-number map, remember that the length of the entries is likely
1595 to be different for each compiled pattern.
1596
1597 PCRE_INFO_OKPARTIAL
1598
1599 Return 1 if the pattern can be used for partial matching, otherwise 0.
1600 The fourth argument should point to an int variable. The pcrepartial
1601 documentation lists the restrictions that apply to patterns when par-
1602 tial matching is used.
1603
1604 PCRE_INFO_OPTIONS
1605
1606 Return a copy of the options with which the pattern was compiled. The
1607 fourth argument should point to an unsigned long int variable. These
1608 option bits are those specified in the call to pcre_compile(), modified
1609 by any top-level option settings at the start of the pattern itself. In
1610 other words, they are the options that will be in force when matching
1611 starts. For example, if the pattern /(?im)abc(?-i)d/ is compiled with
1612 the PCRE_EXTENDED option, the result is PCRE_CASELESS, PCRE_MULTILINE,
1613 and PCRE_EXTENDED.
1614
1615 A pattern is automatically anchored by PCRE if all of its top-level
1616 alternatives begin with one of the following:
1617
1618 ^ unless PCRE_MULTILINE is set
1619 \A always
1620 \G always
1621 .* if PCRE_DOTALL is set and there are no back
1622 references to the subpattern in which .* appears
1623
1624 For such patterns, the PCRE_ANCHORED bit is set in the options returned
1625 by pcre_fullinfo().
1626
1627 PCRE_INFO_SIZE
1628
1629 Return the size of the compiled pattern, that is, the value that was
1630 passed as the argument to pcre_malloc() when PCRE was getting memory in
1631 which to place the compiled data. The fourth argument should point to a
1632 size_t variable.
1633
1634 PCRE_INFO_STUDYSIZE
1635
1636 Return the size of the data block pointed to by the study_data field in
1637 a pcre_extra block. That is, it is the value that was passed to
1638 pcre_malloc() when PCRE was getting memory into which to place the data
1639 created by pcre_study(). The fourth argument should point to a size_t
1640 variable.
1641
1642
1643 OBSOLETE INFO FUNCTION
1644
1645 int pcre_info(const pcre *code, int *optptr, int *firstcharptr);
1646
1647 The pcre_info() function is now obsolete because its interface is too
1648 restrictive to return all the available data about a compiled pattern.
1649 New programs should use pcre_fullinfo() instead. The yield of
1650 pcre_info() is the number of capturing subpatterns, or one of the fol-
1651 lowing negative numbers:
1652
1653 PCRE_ERROR_NULL the argument code was NULL
1654 PCRE_ERROR_BADMAGIC the "magic number" was not found
1655
1656 If the optptr argument is not NULL, a copy of the options with which
1657 the pattern was compiled is placed in the integer it points to (see
1658 PCRE_INFO_OPTIONS above).
1659
1660 If the pattern is not anchored and the firstcharptr argument is not
1661 NULL, it is used to pass back information about the first character of
1662 any matched string (see PCRE_INFO_FIRSTBYTE above).
1663
1664
1665 REFERENCE COUNTS
1666
1667 int pcre_refcount(pcre *code, int adjust);
1668
1669 The pcre_refcount() function is used to maintain a reference count in
1670 the data block that contains a compiled pattern. It is provided for the
1671 benefit of applications that operate in an object-oriented manner,
1672 where different parts of the application may be using the same compiled
1673 pattern, but you want to free the block when they are all done.
1674
1675 When a pattern is compiled, the reference count field is initialized to
1676 zero. It is changed only by calling this function, whose action is to
1677 add the adjust value (which may be positive or negative) to it. The
1678 yield of the function is the new value. However, the value of the count
1679 is constrained to lie between 0 and 65535, inclusive. If the new value
1680 is outside these limits, it is forced to the appropriate limit value.
1681
1682 Except when it is zero, the reference count is not correctly preserved
1683 if a pattern is compiled on one host and then transferred to a host
1684 whose byte-order is different. (This seems a highly unlikely scenario.)
1685
1686
1687 MATCHING A PATTERN: THE TRADITIONAL FUNCTION
1688
1689 int pcre_exec(const pcre *code, const pcre_extra *extra,
1690 const char *subject, int length, int startoffset,
1691 int options, int *ovector, int ovecsize);
1692
1693 The function pcre_exec() is called to match a subject string against a
1694 compiled pattern, which is passed in the code argument. If the pattern
1695 has been studied, the result of the study should be passed in the extra
1696 argument. This function is the main matching facility of the library,
1697 and it operates in a Perl-like manner. For specialist use there is also
1698 an alternative matching function, which is described below in the sec-
1699 tion about the pcre_dfa_exec() function.
1700
1701 In most applications, the pattern will have been compiled (and option-
1702 ally studied) in the same process that calls pcre_exec(). However, it
1703 is possible to save compiled patterns and study data, and then use them
1704 later in different processes, possibly even on different hosts. For a
1705 discussion about this, see the pcreprecompile documentation.
1706
1707 Here is an example of a simple call to pcre_exec():
1708
1709 int rc;
1710 int ovector[30];
1711 rc = pcre_exec(
1712 re, /* result of pcre_compile() */
1713 NULL, /* we didn't study the pattern */
1714 "some string", /* the subject string */
1715 11, /* the length of the subject string */
1716 0, /* start at offset 0 in the subject */
1717 0, /* default options */
1718 ovector, /* vector of integers for substring information */
1719 30); /* number of elements (NOT size in bytes) */
1720
1721 Extra data for pcre_exec()
1722
1723 If the extra argument is not NULL, it must point to a pcre_extra data
1724 block. The pcre_study() function returns such a block (when it doesn't
1725 return NULL), but you can also create one for yourself, and pass addi-
1726 tional information in it. The pcre_extra block contains the following
1727 fields (not necessarily in this order):
1728
1729 unsigned long int flags;
1730 void *study_data;
1731 unsigned long int match_limit;
1732 unsigned long int match_limit_recursion;
1733 void *callout_data;
1734 const unsigned char *tables;
1735
1736 The flags field is a bitmap that specifies which of the other fields
1737 are set. The flag bits are:
1738
1739 PCRE_EXTRA_STUDY_DATA
1740 PCRE_EXTRA_MATCH_LIMIT
1741 PCRE_EXTRA_MATCH_LIMIT_RECURSION
1742 PCRE_EXTRA_CALLOUT_DATA
1743 PCRE_EXTRA_TABLES
1744
1745 Other flag bits should be set to zero. The study_data field is set in
1746 the pcre_extra block that is returned by pcre_study(), together with
1747 the appropriate flag bit. You should not set this yourself, but you may
1748 add to the block by setting the other fields and their corresponding
1749 flag bits.
1750
1751 The match_limit field provides a means of preventing PCRE from using up
1752 a vast amount of resources when running patterns that are not going to
1753 match, but which have a very large number of possibilities in their
1754 search trees. The classic example is the use of nested unlimited
1755 repeats.
1756
1757 Internally, PCRE uses a function called match() which it calls repeat-
1758 edly (sometimes recursively). The limit set by match_limit is imposed
1759 on the number of times this function is called during a match, which
1760 has the effect of limiting the amount of backtracking that can take
1761 place. For patterns that are not anchored, the count restarts from zero
1762 for each position in the subject string.
1763
1764 The default value for the limit can be set when PCRE is built; the
1765 default default is 10 million, which handles all but the most extreme
1766 cases. You can override the default by suppling pcre_exec() with a
1767 pcre_extra block in which match_limit is set, and
1768 PCRE_EXTRA_MATCH_LIMIT is set in the flags field. If the limit is
1769 exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
1770
1771 The match_limit_recursion field is similar to match_limit, but instead
1772 of limiting the total number of times that match() is called, it limits
1773 the depth of recursion. The recursion depth is a smaller number than
1774 the total number of calls, because not all calls to match() are recur-
1775 sive. This limit is of use only if it is set smaller than match_limit.
1776
1777 Limiting the recursion depth limits the amount of stack that can be
1778 used, or, when PCRE has been compiled to use memory on the heap instead
1779 of the stack, the amount of heap memory that can be used.
1780
1781 The default value for match_limit_recursion can be set when PCRE is
1782 built; the default default is the same value as the default for
1783 match_limit. You can override the default by suppling pcre_exec() with
1784 a pcre_extra block in which match_limit_recursion is set, and
1785 PCRE_EXTRA_MATCH_LIMIT_RECURSION is set in the flags field. If the
1786 limit is exceeded, pcre_exec() returns PCRE_ERROR_RECURSIONLIMIT.
1787
1788 The pcre_callout field is used in conjunction with the "callout" fea-
1789 ture, which is described in the pcrecallout documentation.
1790
1791 The tables field is used to pass a character tables pointer to
1792 pcre_exec(); this overrides the value that is stored with the compiled
1793 pattern. A non-NULL value is stored with the compiled pattern only if
1794 custom tables were supplied to pcre_compile() via its tableptr argu-
1795 ment. If NULL is passed to pcre_exec() using this mechanism, it forces
1796 PCRE's internal tables to be used. This facility is helpful when re-
1797 using patterns that have been saved after compiling with an external
1798 set of tables, because the external tables might be at a different
1799 address when pcre_exec() is called. See the pcreprecompile documenta-
1800 tion for a discussion of saving compiled patterns for later use.
1801
1802 Option bits for pcre_exec()
1803
1804 The unused bits of the options argument for pcre_exec() must be zero.
1805 The only bits that may be set are PCRE_ANCHORED, PCRE_NEWLINE_xxx,
1806 PCRE_NOTBOL, PCRE_NOTEOL, PCRE_NOTEMPTY, PCRE_NO_UTF8_CHECK and
1807 PCRE_PARTIAL.
1808
1809 PCRE_ANCHORED
1810
1811 The PCRE_ANCHORED option limits pcre_exec() to matching at the first
1812 matching position. If a pattern was compiled with PCRE_ANCHORED, or
1813 turned out to be anchored by virtue of its contents, it cannot be made
1814 unachored at matching time.
1815
1816 PCRE_NEWLINE_CR
1817 PCRE_NEWLINE_LF
1818 PCRE_NEWLINE_CRLF
1819 PCRE_NEWLINE_ANYCRLF
1820 PCRE_NEWLINE_ANY
1821
1822 These options override the newline definition that was chosen or
1823 defaulted when the pattern was compiled. For details, see the descrip-
1824 tion of pcre_compile() above. During matching, the newline choice
1825 affects the behaviour of the dot, circumflex, and dollar metacharac-
1826 ters. It may also alter the way the match position is advanced after a
1827 match failure for an unanchored pattern.
1828
1829 When PCRE_NEWLINE_CRLF, PCRE_NEWLINE_ANYCRLF, or PCRE_NEWLINE_ANY is
1830 set, and a match attempt for an unanchored pattern fails when the cur-
1831 rent position is at a CRLF sequence, and the pattern contains no
1832 explicit matches for CR or NL characters, the match position is
1833 advanced by two characters instead of one, in other words, to after the
1834 CRLF.
1835
1836 The above rule is a compromise that makes the most common cases work as
1837 expected. For example, if the pattern is .+A (and the PCRE_DOTALL
1838 option is not set), it does not match the string "\r\nA" because, after
1839 failing at the start, it skips both the CR and the LF before retrying.
1840 However, the pattern [\r\n]A does match that string, because it con-
1841 tains an explicit CR or LF reference, and so advances only by one char-
1842 acter after the first failure. Note than an explicit CR or LF refer-
1843 ence occurs for negated character classes such as [^X] because they can
1844 match CR or LF characters.
1845
1846 Notwithstanding the above, anomalous effects may still occur when CRLF
1847 is a valid newline sequence and explicit \r or \n escapes appear in the
1848 pattern.
1849
1850 PCRE_NOTBOL
1851
1852 This option specifies that first character of the subject string is not
1853 the beginning of a line, so the circumflex metacharacter should not
1854 match before it. Setting this without PCRE_MULTILINE (at compile time)
1855 causes circumflex never to match. This option affects only the behav-
1856 iour of the circumflex metacharacter. It does not affect \A.
1857
1858 PCRE_NOTEOL
1859
1860 This option specifies that the end of the subject string is not the end
1861 of a line, so the dollar metacharacter should not match it nor (except
1862 in multiline mode) a newline immediately before it. Setting this with-
1863 out PCRE_MULTILINE (at compile time) causes dollar never to match. This
1864 option affects only the behaviour of the dollar metacharacter. It does
1865 not affect \Z or \z.
1866
1867 PCRE_NOTEMPTY
1868
1869 An empty string is not considered to be a valid match if this option is
1870 set. If there are alternatives in the pattern, they are tried. If all
1871 the alternatives match the empty string, the entire match fails. For
1872 example, if the pattern
1873
1874 a?b?
1875
1876 is applied to a string not beginning with "a" or "b", it matches the
1877 empty string at the start of the subject. With PCRE_NOTEMPTY set, this
1878 match is not valid, so PCRE searches further into the string for occur-
1879 rences of "a" or "b".
1880
1881 Perl has no direct equivalent of PCRE_NOTEMPTY, but it does make a spe-
1882 cial case of a pattern match of the empty string within its split()
1883 function, and when using the /g modifier. It is possible to emulate
1884 Perl's behaviour after matching a null string by first trying the match
1885 again at the same offset with PCRE_NOTEMPTY and PCRE_ANCHORED, and then
1886 if that fails by advancing the starting offset (see below) and trying
1887 an ordinary match again. There is some code that demonstrates how to do
1888 this in the pcredemo.c sample program.
1889
1890 PCRE_NO_UTF8_CHECK
1891
1892 When PCRE_UTF8 is set at compile time, the validity of the subject as a
1893 UTF-8 string is automatically checked when pcre_exec() is subsequently
1894 called. The value of startoffset is also checked to ensure that it
1895 points to the start of a UTF-8 character. There is a discussion about
1896 the validity of UTF-8 strings in the section on UTF-8 support in the
1897 main pcre page. If an invalid UTF-8 sequence of bytes is found,
1898 pcre_exec() returns the error PCRE_ERROR_BADUTF8. If startoffset con-
1899 tains an invalid value, PCRE_ERROR_BADUTF8_OFFSET is returned.
1900
1901 If you already know that your subject is valid, and you want to skip
1902 these checks for performance reasons, you can set the
1903 PCRE_NO_UTF8_CHECK option when calling pcre_exec(). You might want to
1904 do this for the second and subsequent calls to pcre_exec() if you are
1905 making repeated calls to find all the matches in a single subject
1906 string. However, you should be sure that the value of startoffset
1907 points to the start of a UTF-8 character. When PCRE_NO_UTF8_CHECK is
1908 set, the effect of passing an invalid UTF-8 string as a subject, or a
1909 value of startoffset that does not point to the start of a UTF-8 char-
1910 acter, is undefined. Your program may crash.
1911
1912 PCRE_PARTIAL
1913
1914 This option turns on the partial matching feature. If the subject
1915 string fails to match the pattern, but at some point during the match-
1916 ing process the end of the subject was reached (that is, the subject
1917 partially matches the pattern and the failure to match occurred only
1918 because there were not enough subject characters), pcre_exec() returns
1919 PCRE_ERROR_PARTIAL instead of PCRE_ERROR_NOMATCH. When PCRE_PARTIAL is
1920 used, there are restrictions on what may appear in the pattern. These
1921 are discussed in the pcrepartial documentation.
1922
1923 The string to be matched by pcre_exec()
1924
1925 The subject string is passed to pcre_exec() as a pointer in subject, a
1926 length in length, and a starting byte offset in startoffset. In UTF-8
1927 mode, the byte offset must point to the start of a UTF-8 character.
1928 Unlike the pattern string, the subject may contain binary zero bytes.
1929 When the starting offset is zero, the search for a match starts at the
1930 beginning of the subject, and this is by far the most common case.
1931
1932 A non-zero starting offset is useful when searching for another match
1933 in the same subject by calling pcre_exec() again after a previous suc-
1934 cess. Setting startoffset differs from just passing over a shortened
1935 string and setting PCRE_NOTBOL in the case of a pattern that begins
1936 with any kind of lookbehind. For example, consider the pattern
1937
1938 \Biss\B
1939
1940 which finds occurrences of "iss" in the middle of words. (\B matches
1941 only if the current position in the subject is not a word boundary.)
1942 When applied to the string "Mississipi" the first call to pcre_exec()
1943 finds the first occurrence. If pcre_exec() is called again with just
1944 the remainder of the subject, namely "issipi", it does not match,
1945 because \B is always false at the start of the subject, which is deemed
1946 to be a word boundary. However, if pcre_exec() is passed the entire
1947 string again, but with startoffset set to 4, it finds the second occur-
1948 rence of "iss" because it is able to look behind the starting point to
1949 discover that it is preceded by a letter.
1950
1951 If a non-zero starting offset is passed when the pattern is anchored,
1952 one attempt to match at the given offset is made. This can only succeed
1953 if the pattern does not require the match to be at the start of the
1954 subject.
1955
1956 How pcre_exec() returns captured substrings
1957
1958 In general, a pattern matches a certain portion of the subject, and in
1959 addition, further substrings from the subject may be picked out by
1960 parts of the pattern. Following the usage in Jeffrey Friedl's book,
1961 this is called "capturing" in what follows, and the phrase "capturing
1962 subpattern" is used for a fragment of a pattern that picks out a sub-
1963 string. PCRE supports several other kinds of parenthesized subpattern
1964 that do not cause substrings to be captured.
1965
1966 Captured substrings are returned to the caller via a vector of integer
1967 offsets whose address is passed in ovector. The number of elements in
1968 the vector is passed in ovecsize, which must be a non-negative number.
1969 Note: this argument is NOT the size of ovector in bytes.
1970
1971 The first two-thirds of the vector is used to pass back captured sub-
1972 strings, each substring using a pair of integers. The remaining third
1973 of the vector is used as workspace by pcre_exec() while matching cap-
1974 turing subpatterns, and is not available for passing back information.
1975 The length passed in ovecsize should always be a multiple of three. If
1976 it is not, it is rounded down.
1977
1978 When a match is successful, information about captured substrings is
1979 returned in pairs of integers, starting at the beginning of ovector,
1980 and continuing up to two-thirds of its length at the most. The first
1981 element of a pair is set to the offset of the first character in a sub-
1982 string, and the second is set to the offset of the first character
1983 after the end of a substring. The first pair, ovector[0] and ovec-
1984 tor[1], identify the portion of the subject string matched by the
1985 entire pattern. The next pair is used for the first capturing subpat-
1986 tern, and so on. The value returned by pcre_exec() is one more than the
1987 highest numbered pair that has been set. For example, if two substrings
1988 have been captured, the returned value is 3. If there are no capturing
1989 subpatterns, the return value from a successful match is 1, indicating
1990 that just the first pair of offsets has been set.
1991
1992 If a capturing subpattern is matched repeatedly, it is the last portion
1993 of the string that it matched that is returned.
1994
1995 If the vector is too small to hold all the captured substring offsets,
1996 it is used as far as possible (up to two-thirds of its length), and the
1997 function returns a value of zero. In particular, if the substring off-
1998 sets are not of interest, pcre_exec() may be called with ovector passed
1999 as NULL and ovecsize as zero. However, if the pattern contains back
2000 references and the ovector is not big enough to remember the related
2001 substrings, PCRE has to get additional memory for use during matching.
2002 Thus it is usually advisable to supply an ovector.
2003
2004 The pcre_info() function can be used to find out how many capturing
2005 subpatterns there are in a compiled pattern. The smallest size for
2006 ovector that will allow for n captured substrings, in addition to the
2007 offsets of the substring matched by the whole pattern, is (n+1)*3.
2008
2009 It is possible for capturing subpattern number n+1 to match some part
2010 of the subject when subpattern n has not been used at all. For example,
2011 if the string "abc" is matched against the pattern (a|(z))(bc) the
2012 return from the function is 4, and subpatterns 1 and 3 are matched, but
2013 2 is not. When this happens, both values in the offset pairs corre-
2014 sponding to unused subpatterns are set to -1.
2015
2016 Offset values that correspond to unused subpatterns at the end of the
2017 expression are also set to -1. For example, if the string "abc" is
2018 matched against the pattern (abc)(x(yz)?)? subpatterns 2 and 3 are not
2019 matched. The return from the function is 2, because the highest used
2020 capturing subpattern number is 1. However, you can refer to the offsets
2021 for the second and third capturing subpatterns if you wish (assuming
2022 the vector is large enough, of course).
2023
2024 Some convenience functions are provided for extracting the captured
2025 substrings as separate strings. These are described below.
2026
2027 Error return values from pcre_exec()
2028
2029 If pcre_exec() fails, it returns a negative number. The following are
2030 defined in the header file:
2031
2032 PCRE_ERROR_NOMATCH (-1)
2033
2034 The subject string did not match the pattern.
2035
2036 PCRE_ERROR_NULL (-2)
2037
2038 Either code or subject was passed as NULL, or ovector was NULL and
2039 ovecsize was not zero.
2040
2041 PCRE_ERROR_BADOPTION (-3)
2042
2043 An unrecognized bit was set in the options argument.
2044
2045 PCRE_ERROR_BADMAGIC (-4)
2046
2047 PCRE stores a 4-byte "magic number" at the start of the compiled code,
2048 to catch the case when it is passed a junk pointer and to detect when a
2049 pattern that was compiled in an environment of one endianness is run in
2050 an environment with the other endianness. This is the error that PCRE
2051 gives when the magic number is not present.
2052
2053 PCRE_ERROR_UNKNOWN_OPCODE (-5)
2054
2055 While running the pattern match, an unknown item was encountered in the
2056 compiled pattern. This error could be caused by a bug in PCRE or by
2057 overwriting of the compiled pattern.
2058
2059 PCRE_ERROR_NOMEMORY (-6)
2060
2061 If a pattern contains back references, but the ovector that is passed
2062 to pcre_exec() is not big enough to remember the referenced substrings,
2063 PCRE gets a block of memory at the start of matching to use for this
2064 purpose. If the call via pcre_malloc() fails, this error is given. The
2065 memory is automatically freed at the end of matching.
2066
2067 PCRE_ERROR_NOSUBSTRING (-7)
2068
2069 This error is used by the pcre_copy_substring(), pcre_get_substring(),
2070 and pcre_get_substring_list() functions (see below). It is never
2071 returned by pcre_exec().
2072
2073 PCRE_ERROR_MATCHLIMIT (-8)
2074
2075 The backtracking limit, as specified by the match_limit field in a
2076 pcre_extra structure (or defaulted) was reached. See the description
2077 above.
2078
2079 PCRE_ERROR_CALLOUT (-9)
2080
2081 This error is never generated by pcre_exec() itself. It is provided for
2082 use by callout functions that want to yield a distinctive error code.
2083 See the pcrecallout documentation for details.
2084
2085 PCRE_ERROR_BADUTF8 (-10)
2086
2087 A string that contains an invalid UTF-8 byte sequence was passed as a
2088 subject.
2089
2090 PCRE_ERROR_BADUTF8_OFFSET (-11)
2091
2092 The UTF-8 byte sequence that was passed as a subject was valid, but the
2093 value of startoffset did not point to the beginning of a UTF-8 charac-
2094 ter.
2095
2096 PCRE_ERROR_PARTIAL (-12)
2097
2098 The subject string did not match, but it did match partially. See the
2099 pcrepartial documentation for details of partial matching.
2100
2101 PCRE_ERROR_BADPARTIAL (-13)
2102
2103 The PCRE_PARTIAL option was used with a compiled pattern containing
2104 items that are not supported for partial matching. See the pcrepartial
2105 documentation for details of partial matching.
2106
2107 PCRE_ERROR_INTERNAL (-14)
2108
2109 An unexpected internal error has occurred. This error could be caused
2110 by a bug in PCRE or by overwriting of the compiled pattern.
2111
2112 PCRE_ERROR_BADCOUNT (-15)
2113
2114 This error is given if the value of the ovecsize argument is negative.
2115
2116 PCRE_ERROR_RECURSIONLIMIT (-21)
2117
2118 The internal recursion limit, as specified by the match_limit_recursion
2119 field in a pcre_extra structure (or defaulted) was reached. See the
2120 description above.
2121
2122 PCRE_ERROR_BADNEWLINE (-23)
2123
2124 An invalid combination of PCRE_NEWLINE_xxx options was given.
2125
2126 Error numbers -16 to -20 and -22 are not used by pcre_exec().
2127
2128
2129 EXTRACTING CAPTURED SUBSTRINGS BY NUMBER
2130
2131 int pcre_copy_substring(const char *subject, int *ovector,
2132 int stringcount, int stringnumber, char *buffer,
2133 int buffersize);
2134
2135 int pcre_get_substring(const char *subject, int *ovector,
2136 int stringcount, int stringnumber,
2137 const char **stringptr);
2138
2139 int pcre_get_substring_list(const char *subject,
2140 int *ovector, int stringcount, const char ***listptr);
2141
2142 Captured substrings can be accessed directly by using the offsets
2143 returned by pcre_exec() in ovector. For convenience, the functions
2144 pcre_copy_substring(), pcre_get_substring(), and pcre_get_sub-
2145 string_list() are provided for extracting captured substrings as new,
2146 separate, zero-terminated strings. These functions identify substrings
2147 by number. The next section describes functions for extracting named
2148 substrings.
2149
2150 A substring that contains a binary zero is correctly extracted and has
2151 a further zero added on the end, but the result is not, of course, a C
2152 string. However, you can process such a string by referring to the
2153 length that is returned by pcre_copy_substring() and pcre_get_sub-
2154 string(). Unfortunately, the interface to pcre_get_substring_list() is
2155 not adequate for handling strings containing binary zeros, because the
2156 end of the final string is not independently indicated.
2157
2158 The first three arguments are the same for all three of these func-
2159 tions: subject is the subject string that has just been successfully
2160 matched, ovector is a pointer to the vector of integer offsets that was
2161 passed to pcre_exec(), and stringcount is the number of substrings that
2162 were captured by the match, including the substring that matched the
2163 entire regular expression. This is the value returned by pcre_exec() if
2164 it is greater than zero. If pcre_exec() returned zero, indicating that
2165 it ran out of space in ovector, the value passed as stringcount should
2166 be the number of elements in the vector divided by three.
2167
2168 The functions pcre_copy_substring() and pcre_get_substring() extract a
2169 single substring, whose number is given as stringnumber. A value of
2170 zero extracts the substring that matched the entire pattern, whereas
2171 higher values extract the captured substrings. For pcre_copy_sub-
2172 string(), the string is placed in buffer, whose length is given by
2173 buffersize, while for pcre_get_substring() a new block of memory is
2174 obtained via pcre_malloc, and its address is returned via stringptr.
2175 The yield of the function is the length of the string, not including
2176 the terminating zero, or one of these error codes:
2177
2178 PCRE_ERROR_NOMEMORY (-6)
2179
2180 The buffer was too small for pcre_copy_substring(), or the attempt to
2181 get memory failed for pcre_get_substring().
2182
2183 PCRE_ERROR_NOSUBSTRING (-7)
2184
2185 There is no substring whose number is stringnumber.
2186
2187 The pcre_get_substring_list() function extracts all available sub-
2188 strings and builds a list of pointers to them. All this is done in a
2189 single block of memory that is obtained via pcre_malloc. The address of
2190 the memory block is returned via listptr, which is also the start of
2191 the list of string pointers. The end of the list is marked by a NULL
2192 pointer. The yield of the function is zero if all went well, or the
2193 error code
2194
2195 PCRE_ERROR_NOMEMORY (-6)
2196
2197 if the attempt to get the memory block failed.
2198
2199 When any of these functions encounter a substring that is unset, which
2200 can happen when capturing subpattern number n+1 matches some part of
2201 the subject, but subpattern n has not been used at all, they return an
2202 empty string. This can be distinguished from a genuine zero-length sub-
2203 string by inspecting the appropriate offset in ovector, which is nega-
2204 tive for unset substrings.
2205
2206 The two convenience functions pcre_free_substring() and pcre_free_sub-
2207 string_list() can be used to free the memory returned by a previous
2208 call of pcre_get_substring() or pcre_get_substring_list(), respec-
2209 tively. They do nothing more than call the function pointed to by
2210 pcre_free, which of course could be called directly from a C program.
2211 However, PCRE is used in some situations where it is linked via a spe-
2212 cial interface to another programming language that cannot use
2213 pcre_free directly; it is for these cases that the functions are pro-
2214 vided.
2215
2216
2217 EXTRACTING CAPTURED SUBSTRINGS BY NAME
2218
2219 int pcre_get_stringnumber(const pcre *code,
2220 const char *name);
2221
2222 int pcre_copy_named_substring(const pcre *code,
2223 const char *subject, int *ovector,
2224 int stringcount, const char *stringname,
2225 char *buffer, int buffersize);
2226
2227 int pcre_get_named_substring(const pcre *code,
2228 const char *subject, int *ovector,
2229 int stringcount, const char *stringname,
2230 const char **stringptr);
2231
2232 To extract a substring by name, you first have to find associated num-
2233 ber. For example, for this pattern
2234
2235 (a+)b(?<xxx>\d+)...
2236
2237 the number of the subpattern called "xxx" is 2. If the name is known to
2238 be unique (PCRE_DUPNAMES was not set), you can find the number from the
2239 name by calling pcre_get_stringnumber(). The first argument is the com-
2240 piled pattern, and the second is the name. The yield of the function is
2241 the subpattern number, or PCRE_ERROR_NOSUBSTRING (-7) if there is no
2242 subpattern of that name.
2243
2244 Given the number, you can extract the substring directly, or use one of
2245 the functions described in the previous section. For convenience, there
2246 are also two functions that do the whole job.
2247
2248 Most of the arguments of pcre_copy_named_substring() and
2249 pcre_get_named_substring() are the same as those for the similarly
2250 named functions that extract by number. As these are described in the
2251 previous section, they are not re-described here. There are just two
2252 differences:
2253
2254 First, instead of a substring number, a substring name is given. Sec-
2255 ond, there is an extra argument, given at the start, which is a pointer
2256 to the compiled pattern. This is needed in order to gain access to the
2257 name-to-number translation table.
2258
2259 These functions call pcre_get_stringnumber(), and if it succeeds, they
2260 then call pcre_copy_substring() or pcre_get_substring(), as appropri-
2261 ate. NOTE: If PCRE_DUPNAMES is set and there are duplicate names, the
2262 behaviour may not be what you want (see the next section).
2263
2264
2265 DUPLICATE SUBPATTERN NAMES
2266
2267 int pcre_get_stringtable_entries(const pcre *code,
2268 const char *name, char **first, char **last);
2269
2270 When a pattern is compiled with the PCRE_DUPNAMES option, names for
2271 subpatterns are not required to be unique. Normally, patterns with
2272 duplicate names are such that in any one match, only one of the named
2273 subpatterns participates. An example is shown in the pcrepattern docu-
2274 mentation.
2275
2276 When duplicates are present, pcre_copy_named_substring() and
2277 pcre_get_named_substring() return the first substring corresponding to
2278 the given name that is set. If none are set, PCRE_ERROR_NOSUBSTRING
2279 (-7) is returned; no data is returned. The pcre_get_stringnumber()
2280 function returns one of the numbers that are associated with the name,
2281 but it is not defined which it is.
2282
2283 If you want to get full details of all captured substrings for a given
2284 name, you must use the pcre_get_stringtable_entries() function. The
2285 first argument is the compiled pattern, and the second is the name. The
2286 third and fourth are pointers to variables which are updated by the
2287 function. After it has run, they point to the first and last entries in
2288 the name-to-number table for the given name. The function itself
2289 returns the length of each entry, or PCRE_ERROR_NOSUBSTRING (-7) if
2290 there are none. The format of the table is described above in the sec-
2291 tion entitled Information about a pattern. Given all the relevant
2292 entries for the name, you can extract each of their numbers, and hence
2293 the captured data, if any.
2294
2295
2296 FINDING ALL POSSIBLE MATCHES
2297
2298 The traditional matching function uses a similar algorithm to Perl,
2299 which stops when it finds the first match, starting at a given point in
2300 the subject. If you want to find all possible matches, or the longest
2301 possible match, consider using the alternative matching function (see
2302 below) instead. If you cannot use the alternative function, but still
2303 need to find all possible matches, you can kludge it up by making use
2304 of the callout facility, which is described in the pcrecallout documen-
2305 tation.
2306
2307 What you have to do is to insert a callout right at the end of the pat-
2308 tern. When your callout function is called, extract and save the cur-
2309 rent matched substring. Then return 1, which forces pcre_exec() to
2310 backtrack and try other alternatives. Ultimately, when it runs out of
2311 matches, pcre_exec() will yield PCRE_ERROR_NOMATCH.
2312
2313
2314 MATCHING A PATTERN: THE ALTERNATIVE FUNCTION
2315
2316 int pcre_dfa_exec(const pcre *code, const pcre_extra *extra,
2317 const char *subject, int length, int startoffset,
2318 int options, int *ovector, int ovecsize,
2319 int *workspace, int wscount);
2320
2321 The function pcre_dfa_exec() is called to match a subject string
2322 against a compiled pattern, using a matching algorithm that scans the
2323 subject string just once, and does not backtrack. This has different
2324 characteristics to the normal algorithm, and is not compatible with
2325 Perl. Some of the features of PCRE patterns are not supported. Never-
2326 theless, there are times when this kind of matching can be useful. For
2327 a discussion of the two matching algorithms, see the pcrematching docu-
2328 mentation.
2329
2330 The arguments for the pcre_dfa_exec() function are the same as for
2331 pcre_exec(), plus two extras. The ovector argument is used in a differ-
2332 ent way, and this is described below. The other common arguments are
2333 used in the same way as for pcre_exec(), so their description is not
2334 repeated here.
2335
2336 The two additional arguments provide workspace for the function. The
2337 workspace vector should contain at least 20 elements. It is used for
2338 keeping track of multiple paths through the pattern tree. More
2339 workspace will be needed for patterns and subjects where there are a
2340 lot of potential matches.
2341
2342 Here is an example of a simple call to pcre_dfa_exec():
2343
2344 int rc;
2345 int ovector[10];
2346 int wspace[20];
2347 rc = pcre_dfa_exec(
2348 re, /* result of pcre_compile() */
2349 NULL, /* we didn't study the pattern */
2350 "some string", /* the subject string */
2351 11, /* the length of the subject string */
2352 0, /* start at offset 0 in the subject */
2353 0, /* default options */
2354 ovector, /* vector of integers for substring information */
2355 10, /* number of elements (NOT size in bytes) */
2356 wspace, /* working space vector */
2357 20); /* number of elements (NOT size in bytes) */
2358
2359 Option bits for pcre_dfa_exec()
2360
2361 The unused bits of the options argument for pcre_dfa_exec() must be
2362 zero. The only bits that may be set are PCRE_ANCHORED, PCRE_NEW-
2363 LINE_xxx, PCRE_NOTBOL, PCRE_NOTEOL, PCRE_NOTEMPTY, PCRE_NO_UTF8_CHECK,
2364 PCRE_PARTIAL, PCRE_DFA_SHORTEST, and PCRE_DFA_RESTART. All but the last
2365 three of these are the same as for pcre_exec(), so their description is
2366 not repeated here.
2367
2368 PCRE_PARTIAL
2369
2370 This has the same general effect as it does for pcre_exec(), but the
2371 details are slightly different. When PCRE_PARTIAL is set for
2372 pcre_dfa_exec(), the return code PCRE_ERROR_NOMATCH is converted into
2373 PCRE_ERROR_PARTIAL if the end of the subject is reached, there have
2374 been no complete matches, but there is still at least one matching pos-
2375 sibility. The portion of the string that provided the partial match is
2376 set as the first matching string.
2377
2378 PCRE_DFA_SHORTEST
2379
2380 Setting the PCRE_DFA_SHORTEST option causes the matching algorithm to
2381 stop as soon as it has found one match. Because of the way the alterna-
2382 tive algorithm works, this is necessarily the shortest possible match
2383 at the first possible matching point in the subject string.
2384
2385 PCRE_DFA_RESTART
2386
2387 When pcre_dfa_exec() is called with the PCRE_PARTIAL option, and
2388 returns a partial match, it is possible to call it again, with addi-
2389 tional subject characters, and have it continue with the same match.
2390 The PCRE_DFA_RESTART option requests this action; when it is set, the
2391 workspace and wscount options must reference the same vector as before
2392 because data about the match so far is left in them after a partial
2393 match. There is more discussion of this facility in the pcrepartial
2394 documentation.
2395
2396 Successful returns from pcre_dfa_exec()
2397
2398 When pcre_dfa_exec() succeeds, it may have matched more than one sub-
2399 string in the subject. Note, however, that all the matches from one run
2400 of the function start at the same point in the subject. The shorter
2401 matches are all initial substrings of the longer matches. For example,
2402 if the pattern
2403
2404 <.*>
2405
2406 is matched against the string
2407
2408 This is <something> <something else> <something further> no more
2409
2410 the three matched strings are
2411
2412 <something>
2413 <something> <something else>
2414 <something> <something else> <something further>
2415
2416 On success, the yield of the function is a number greater than zero,
2417 which is the number of matched substrings. The substrings themselves
2418 are returned in ovector. Each string uses two elements; the first is
2419 the offset to the start, and the second is the offset to the end. In
2420 fact, all the strings have the same start offset. (Space could have
2421 been saved by giving this only once, but it was decided to retain some
2422 compatibility with the way pcre_exec() returns data, even though the
2423 meaning of the strings is different.)
2424
2425 The strings are returned in reverse order of length; that is, the long-
2426 est matching string is given first. If there were too many matches to
2427 fit into ovector, the yield of the function is zero, and the vector is
2428 filled with the longest matches.
2429
2430 Error returns from pcre_dfa_exec()
2431
2432 The pcre_dfa_exec() function returns a negative number when it fails.
2433 Many of the errors are the same as for pcre_exec(), and these are
2434 described above. There are in addition the following errors that are
2435 specific to pcre_dfa_exec():
2436
2437 PCRE_ERROR_DFA_UITEM (-16)
2438
2439 This return is given if pcre_dfa_exec() encounters an item in the pat-
2440 tern that it does not support, for instance, the use of \C or a back
2441 reference.
2442
2443 PCRE_ERROR_DFA_UCOND (-17)
2444
2445 This return is given if pcre_dfa_exec() encounters a condition item
2446 that uses a back reference for the condition, or a test for recursion
2447 in a specific group. These are not supported.
2448
2449 PCRE_ERROR_DFA_UMLIMIT (-18)
2450
2451 This return is given if pcre_dfa_exec() is called with an extra block
2452 that contains a setting of the match_limit field. This is not supported
2453 (it is meaningless).
2454
2455 PCRE_ERROR_DFA_WSSIZE (-19)
2456
2457 This return is given if pcre_dfa_exec() runs out of space in the
2458 workspace vector.
2459
2460 PCRE_ERROR_DFA_RECURSE (-20)
2461
2462 When a recursive subpattern is processed, the matching function calls
2463 itself recursively, using private vectors for ovector and workspace.
2464 This error is given if the output vector is not large enough. This
2465 should be extremely rare, as a vector of size 1000 is used.
2466
2467
2468 SEE ALSO
2469
2470 pcrebuild(3), pcrecallout(3), pcrecpp(3)(3), pcrematching(3), pcrepar-
2471 tial(3), pcreposix(3), pcreprecompile(3), pcresample(3), pcrestack(3).
2472
2473
2474 AUTHOR
2475
2476 Philip Hazel
2477 University Computing Service
2478 Cambridge CB2 3QH, England.
2479
2480
2481 REVISION
2482
2483 Last updated: 21 August 2007
2484 Copyright (c) 1997-2007 University of Cambridge.
2485 ------------------------------------------------------------------------------
2486
2487
2488 PCRECALLOUT(3) PCRECALLOUT(3)
2489
2490
2491 NAME
2492 PCRE - Perl-compatible regular expressions
2493
2494
2495 PCRE CALLOUTS
2496
2497 int (*pcre_callout)(pcre_callout_block *);
2498
2499 PCRE provides a feature called "callout", which is a means of temporar-
2500 ily passing control to the caller of PCRE in the middle of pattern
2501 matching. The caller of PCRE provides an external function by putting
2502 its entry point in the global variable pcre_callout. By default, this
2503 variable contains NULL, which disables all calling out.
2504
2505 Within a regular expression, (?C) indicates the points at which the
2506 external function is to be called. Different callout points can be
2507 identified by putting a number less than 256 after the letter C. The
2508 default value is zero. For example, this pattern has two callout
2509 points:
2510
2511 (?C1)abc(?C2)def
2512
2513 If the PCRE_AUTO_CALLOUT option bit is set when pcre_compile() is
2514 called, PCRE automatically inserts callouts, all with number 255,
2515 before each item in the pattern. For example, if PCRE_AUTO_CALLOUT is
2516 used with the pattern
2517
2518 A(\d{2}|--)
2519
2520 it is processed as if it were
2521
2522 (?C255)A(?C255)((?C255)\d{2}(?C255)|(?C255)-(?C255)-(?C255))(?C255)
2523
2524 Notice that there is a callout before and after each parenthesis and
2525 alternation bar. Automatic callouts can be used for tracking the
2526 progress of pattern matching. The pcretest command has an option that
2527 sets automatic callouts; when it is used, the output indicates how the
2528 pattern is matched. This is useful information when you are trying to
2529 optimize the performance of a particular pattern.
2530
2531
2532 MISSING CALLOUTS
2533
2534 You should be aware that, because of optimizations in the way PCRE
2535 matches patterns, callouts sometimes do not happen. For example, if the
2536 pattern is
2537
2538 ab(?C4)cd
2539
2540 PCRE knows that any matching string must contain the letter "d". If the
2541 subject string is "abyz", the lack of "d" means that matching doesn't
2542 ever start, and the callout is never reached. However, with "abyd",
2543 though the result is still no match, the callout is obeyed.
2544
2545
2546 THE CALLOUT INTERFACE
2547
2548 During matching, when PCRE reaches a callout point, the external func-
2549 tion defined by pcre_callout is called (if it is set). This applies to
2550 both the pcre_exec() and the pcre_dfa_exec() matching functions. The
2551 only argument to the callout function is a pointer to a pcre_callout
2552 block. This structure contains the following fields:
2553
2554 int version;
2555 int callout_number;
2556 int *offset_vector;
2557 const char *subject;
2558 int subject_length;
2559 int start_match;
2560 int current_position;
2561 int capture_top;
2562 int capture_last;
2563 void *callout_data;
2564 int pattern_position;
2565 int next_item_length;
2566
2567 The version field is an integer containing the version number of the
2568 block format. The initial version was 0; the current version is 1. The
2569 version number will change again in future if additional fields are
2570 added, but the intention is never to remove any of the existing fields.
2571
2572 The callout_number field contains the number of the callout, as com-
2573 piled into the pattern (that is, the number after ?C for manual call-
2574 outs, and 255 for automatically generated callouts).
2575
2576 The offset_vector field is a pointer to the vector of offsets that was
2577 passed by the caller to pcre_exec() or pcre_dfa_exec(). When
2578 pcre_exec() is used, the contents can be inspected in order to extract
2579 substrings that have been matched so far, in the same way as for
2580 extracting substrings after a match has completed. For pcre_dfa_exec()
2581 this field is not useful.
2582
2583 The subject and subject_length fields contain copies of the values that
2584 were passed to pcre_exec().
2585
2586 The start_match field normally contains the offset within the subject
2587 at which the current match attempt started. However, if the escape
2588 sequence \K has been encountered, this value is changed to reflect the
2589 modified starting point. If the pattern is not anchored, the callout
2590 function may be called several times from the same point in the pattern
2591 for different starting points in the subject.
2592
2593 The current_position field contains the offset within the subject of
2594 the current match pointer.
2595
2596 When the pcre_exec() function is used, the capture_top field contains
2597 one more than the number of the highest numbered captured substring so
2598 far. If no substrings have been captured, the value of capture_top is
2599 one. This is always the case when pcre_dfa_exec() is used, because it
2600 does not support captured substrings.
2601
2602 The capture_last field contains the number of the most recently cap-
2603 tured substring. If no substrings have been captured, its value is -1.
2604 This is always the case when pcre_dfa_exec() is used.
2605
2606 The callout_data field contains a value that is passed to pcre_exec()
2607 or pcre_dfa_exec() specifically so that it can be passed back in call-
2608 outs. It is passed in the pcre_callout field of the pcre_extra data
2609 structure. If no such data was passed, the value of callout_data in a
2610 pcre_callout block is NULL. There is a description of the pcre_extra
2611 structure in the pcreapi documentation.
2612
2613 The pattern_position field is present from version 1 of the pcre_call-
2614 out structure. It contains the offset to the next item to be matched in
2615 the pattern string.
2616
2617 The next_item_length field is present from version 1 of the pcre_call-
2618 out structure. It contains the length of the next item to be matched in
2619 the pattern string. When the callout immediately precedes an alterna-
2620 tion bar, a closing parenthesis, or the end of the pattern, the length
2621 is zero. When the callout precedes an opening parenthesis, the length
2622 is that of the entire subpattern.
2623
2624 The pattern_position and next_item_length fields are intended to help
2625 in distinguishing between different automatic callouts, which all have
2626 the same callout number. However, they are set for all callouts.
2627
2628
2629 RETURN VALUES
2630
2631 The external callout function returns an integer to PCRE. If the value
2632 is zero, matching proceeds as normal. If the value is greater than
2633 zero, matching fails at the current point, but the testing of other
2634 matching possibilities goes ahead, just as if a lookahead assertion had
2635 failed. If the value is less than zero, the match is abandoned, and
2636 pcre_exec() (or pcre_dfa_exec()) returns the negative value.
2637
2638 Negative values should normally be chosen from the set of
2639 PCRE_ERROR_xxx values. In particular, PCRE_ERROR_NOMATCH forces a stan-
2640 dard "no match" failure. The error number PCRE_ERROR_CALLOUT is
2641 reserved for use by callout functions; it will never be used by PCRE
2642 itself.
2643
2644
2645 AUTHOR
2646
2647 Philip Hazel
2648 University Computing Service
2649 Cambridge CB2 3QH, England.
2650
2651
2652 REVISION
2653
2654 Last updated: 29 May 2007
2655 Copyright (c) 1997-2007 University of Cambridge.
2656 ------------------------------------------------------------------------------
2657
2658
2659 PCRECOMPAT(3) PCRECOMPAT(3)
2660
2661
2662 NAME
2663 PCRE - Perl-compatible regular expressions
2664
2665
2666 DIFFERENCES BETWEEN PCRE AND PERL
2667
2668 This document describes the differences in the ways that PCRE and Perl
2669 handle regular expressions. The differences described here are mainly
2670 with respect to Perl 5.8, though PCRE versions 7.0 and later contain
2671 some features that are expected to be in the forthcoming Perl 5.10.
2672
2673 1. PCRE has only a subset of Perl's UTF-8 and Unicode support. Details
2674 of what it does have are given in the section on UTF-8 support in the
2675 main pcre page.
2676
2677 2. PCRE does not allow repeat quantifiers on lookahead assertions. Perl
2678 permits them, but they do not mean what you might think. For example,
2679 (?!a){3} does not assert that the next three characters are not "a". It
2680 just asserts that the next character is not "a" three times.
2681
2682 3. Capturing subpatterns that occur inside negative lookahead asser-
2683 tions are counted, but their entries in the offsets vector are never
2684 set. Perl sets its numerical variables from any such patterns that are
2685 matched before the assertion fails to match something (thereby succeed-
2686 ing), but only if the negative lookahead assertion contains just one
2687 branch.
2688
2689 4. Though binary zero characters are supported in the subject string,
2690 they are not allowed in a pattern string because it is passed as a nor-
2691 mal C string, terminated by zero. The escape sequence \0 can be used in
2692 the pattern to represent a binary zero.
2693
2694 5. The following Perl escape sequences are not supported: \l, \u, \L,
2695 \U, and \N. In fact these are implemented by Perl's general string-han-
2696 dling and are not part of its pattern matching engine. If any of these
2697 are encountered by PCRE, an error is generated.
2698
2699 6. The Perl escape sequences \p, \P, and \X are supported only if PCRE
2700 is built with Unicode character property support. The properties that
2701 can be tested with \p and \P are limited to the general category prop-
2702 erties such as Lu and Nd, script names such as Greek or Han, and the
2703 derived properties Any and L&.
2704
2705 7. PCRE does support the \Q...\E escape for quoting substrings. Charac-
2706 ters in between are treated as literals. This is slightly different
2707 from Perl in that $ and @ are also handled as literals inside the
2708 quotes. In Perl, they cause variable interpolation (but of course PCRE
2709 does not have variables). Note the following examples:
2710
2711 Pattern PCRE matches Perl matches
2712
2713 \Qabc$xyz\E abc$xyz abc followed by the
2714 contents of $xyz
2715 \Qabc\$xyz\E abc\$xyz abc\$xyz
2716 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
2717
2718 The \Q...\E sequence is recognized both inside and outside character
2719 classes.
2720
2721 8. Fairly obviously, PCRE does not support the (?{code}) and (??{code})
2722 constructions. However, there is support for recursive patterns. This
2723 is not available in Perl 5.8, but will be in Perl 5.10. Also, the PCRE
2724 "callout" feature allows an external function to be called during pat-
2725 tern matching. See the pcrecallout documentation for details.
2726
2727 9. Subpatterns that are called recursively or as "subroutines" are
2728 always treated as atomic groups in PCRE. This is like Python, but
2729 unlike Perl.
2730
2731 10. There are some differences that are concerned with the settings of
2732 captured strings when part of a pattern is repeated. For example,
2733 matching "aba" against the pattern /^(a(b)?)+$/ in Perl leaves $2
2734 unset, but in PCRE it is set to "b".
2735
2736 11. PCRE does support Perl 5.10's backtracking verbs (*ACCEPT),
2737 (*FAIL), (*F), (*COMMIT), (*PRUNE), (*SKIP), and (*THEN), but only in
2738 the forms without an argument. PCRE does not support (*MARK). If
2739 (*ACCEPT) is within capturing parentheses, PCRE does not set that cap-
2740 ture group; this is different to Perl.
2741
2742 12. PCRE provides some extensions to the Perl regular expression facil-
2743 ities. Perl 5.10 will include new features that are not in earlier
2744 versions, some of which (such as named parentheses) have been in PCRE
2745 for some time. This list is with respect to Perl 5.10:
2746
2747 (a) Although lookbehind assertions must match fixed length strings,
2748 each alternative branch of a lookbehind assertion can match a different
2749 length of string. Perl requires them all to have the same length.
2750
2751 (b) If PCRE_DOLLAR_ENDONLY is set and PCRE_MULTILINE is not set, the $
2752 meta-character matches only at the very end of the string.
2753
2754 (c) If PCRE_EXTRA is set, a backslash followed by a letter with no spe-
2755 cial meaning is faulted. Otherwise, like Perl, the backslash is quietly
2756 ignored. (Perl can be made to issue a warning.)
2757
2758 (d) If PCRE_UNGREEDY is set, the greediness of the repetition quanti-
2759 fiers is inverted, that is, by default they are not greedy, but if fol-
2760 lowed by a question mark they are.
2761
2762 (e) PCRE_ANCHORED can be used at matching time to force a pattern to be
2763 tried only at the first matching position in the subject string.
2764
2765 (f) The PCRE_NOTBOL, PCRE_NOTEOL, PCRE_NOTEMPTY, and PCRE_NO_AUTO_CAP-
2766 TURE options for pcre_exec() have no Perl equivalents.
2767
2768 (g) The callout facility is PCRE-specific.
2769
2770 (h) The partial matching facility is PCRE-specific.
2771
2772 (i) Patterns compiled by PCRE can be saved and re-used at a later time,
2773 even on different hosts that have the other endianness.
2774
2775 (j) The alternative matching function (pcre_dfa_exec()) matches in a
2776 different way and is not Perl-compatible.
2777
2778
2779 AUTHOR
2780
2781 Philip Hazel
2782 University Computing Service
2783 Cambridge CB2 3QH, England.
2784
2785
2786 REVISION
2787
2788 Last updated: 08 August 2007
2789 Copyright (c) 1997-2007 University of Cambridge.
2790 ------------------------------------------------------------------------------
2791
2792
2793 PCREPATTERN(3) PCREPATTERN(3)
2794
2795
2796 NAME
2797 PCRE - Perl-compatible regular expressions
2798
2799
2800 PCRE REGULAR EXPRESSION DETAILS
2801
2802 The syntax and semantics of the regular expressions that are supported
2803 by PCRE are described in detail below. There is a quick-reference syn-
2804 tax summary in the pcresyntax page. Perl's regular expressions are
2805 described in its own documentation, and regular expressions in general
2806 are covered in a number of books, some of which have copious examples.
2807 Jeffrey Friedl's "Mastering Regular Expressions", published by
2808 O'Reilly, covers regular expressions in great detail. This description
2809 of PCRE's regular expressions is intended as reference material.
2810
2811 The original operation of PCRE was on strings of one-byte characters.
2812 However, there is now also support for UTF-8 character strings. To use
2813 this, you must build PCRE to include UTF-8 support, and then call
2814 pcre_compile() with the PCRE_UTF8 option. How this affects pattern
2815 matching is mentioned in several places below. There is also a summary
2816 of UTF-8 features in the section on UTF-8 support in the main pcre
2817 page.
2818
2819 The remainder of this document discusses the patterns that are sup-
2820 ported by PCRE when its main matching function, pcre_exec(), is used.
2821 From release 6.0, PCRE offers a second matching function,
2822 pcre_dfa_exec(), which matches using a different algorithm that is not
2823 Perl-compatible. Some of the features discussed below are not available
2824 when pcre_dfa_exec() is used. The advantages and disadvantages of the
2825 alternative function, and how it differs from the normal function, are
2826 discussed in the pcrematching page.
2827
2828
2829 NEWLINE CONVENTIONS
2830
2831 PCRE supports five different conventions for indicating line breaks in
2832 strings: a single CR (carriage return) character, a single LF (line-
2833 feed) character, the two-character sequence CRLF, any of the three pre-
2834 ceding, or any Unicode newline sequence. The pcreapi page has further
2835 discussion about newlines, and shows how to set the newline convention
2836 in the options arguments for the compiling and matching functions.
2837
2838 It is also possible to specify a newline convention by starting a pat-
2839 tern string with one of the following five sequences:
2840
2841 (*CR) carriage return
2842 (*LF) linefeed
2843 (*CRLF) carriage return, followed by linefeed
2844 (*ANYCRLF) any of the three above
2845 (*ANY) all Unicode newline sequences
2846
2847 These override the default and the options given to pcre_compile(). For
2848 example, on a Unix system where LF is the default newline sequence, the
2849 pattern
2850
2851 (*CR)a.b
2852
2853 changes the convention to CR. That pattern matches "a\nb" because LF is
2854 no longer a newline. Note that these special settings, which are not
2855 Perl-compatible, are recognized only at the very start of a pattern,
2856 and that they must be in upper case.
2857
2858
2859 CHARACTERS AND METACHARACTERS
2860
2861 A regular expression is a pattern that is matched against a subject
2862 string from left to right. Most characters stand for themselves in a
2863 pattern, and match the corresponding characters in the subject. As a
2864 trivial example, the pattern
2865
2866 The quick brown fox
2867
2868 matches a portion of a subject string that is identical to itself. When
2869 caseless matching is specified (the PCRE_CASELESS option), letters are
2870 matched independently of case. In UTF-8 mode, PCRE always understands
2871 the concept of case for characters whose values are less than 128, so
2872 caseless matching is always possible. For characters with higher val-
2873 ues, the concept of case is supported if PCRE is compiled with Unicode
2874 property support, but not otherwise. If you want to use caseless
2875 matching for characters 128 and above, you must ensure that PCRE is
2876 compiled with Unicode property support as well as with UTF-8 support.
2877
2878 The power of regular expressions comes from the ability to include
2879 alternatives and repetitions in the pattern. These are encoded in the
2880 pattern by the use of metacharacters, which do not stand for themselves
2881 but instead are interpreted in some special way.
2882
2883 There are two different sets of metacharacters: those that are recog-
2884 nized anywhere in the pattern except within square brackets, and those
2885 that are recognized within square brackets. Outside square brackets,
2886 the metacharacters are as follows:
2887
2888 \ general escape character with several uses
2889 ^ assert start of string (or line, in multiline mode)
2890 $ assert end of string (or line, in multiline mode)
2891 . match any character except newline (by default)
2892 [ start character class definition
2893 | start of alternative branch
2894 ( start subpattern
2895 ) end subpattern
2896 ? extends the meaning of (
2897 also 0 or 1 quantifier
2898 also quantifier minimizer
2899 * 0 or more quantifier
2900 + 1 or more quantifier
2901 also "possessive quantifier"
2902 { start min/max quantifier
2903
2904 Part of a pattern that is in square brackets is called a "character
2905 class". In a character class the only metacharacters are:
2906
2907 \ general escape character
2908 ^ negate the class, but only if the first character
2909 - indicates character range
2910 [ POSIX character class (only if followed by POSIX
2911 syntax)
2912 ] terminates the character class
2913
2914 The following sections describe the use of each of the metacharacters.
2915
2916
2917 BACKSLASH
2918
2919 The backslash character has several uses. Firstly, if it is followed by
2920 a non-alphanumeric character, it takes away any special meaning that
2921 character may have. This use of backslash as an escape character
2922 applies both inside and outside character classes.
2923
2924 For example, if you want to match a * character, you write \* in the
2925 pattern. This escaping action applies whether or not the following
2926 character would otherwise be interpreted as a metacharacter, so it is
2927 always safe to precede a non-alphanumeric with backslash to specify
2928 that it stands for itself. In particular, if you want to match a back-
2929 slash, you write \\.
2930
2931 If a pattern is compiled with the PCRE_EXTENDED option, whitespace in
2932 the pattern (other than in a character class) and characters between a
2933 # outside a character class and the next newline are ignored. An escap-
2934 ing backslash can be used to include a whitespace or # character as
2935 part of the pattern.
2936
2937 If you want to remove the special meaning from a sequence of charac-
2938 ters, you can do so by putting them between \Q and \E. This is differ-
2939 ent from Perl in that $ and @ are handled as literals in \Q...\E
2940 sequences in PCRE, whereas in Perl, $ and @ cause variable interpola-
2941 tion. Note the following examples:
2942
2943 Pattern PCRE matches Perl matches
2944
2945 \Qabc$xyz\E abc$xyz abc followed by the
2946 contents of $xyz
2947 \Qabc\$xyz\E abc\$xyz abc\$xyz
2948 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
2949
2950 The \Q...\E sequence is recognized both inside and outside character
2951 classes.
2952
2953 Non-printing characters
2954
2955 A second use of backslash provides a way of encoding non-printing char-
2956 acters in patterns in a visible manner. There is no restriction on the
2957 appearance of non-printing characters, apart from the binary zero that
2958 terminates a pattern, but when a pattern is being prepared by text
2959 editing, it is usually easier to use one of the following escape
2960 sequences than the binary character it represents:
2961
2962 \a alarm, that is, the BEL character (hex 07)
2963 \cx "control-x", where x is any character
2964 \e escape (hex 1B)
2965 \f formfeed (hex 0C)
2966 \n linefeed (hex 0A)
2967 \r carriage return (hex 0D)
2968 \t tab (hex 09)
2969 \ddd character with octal code ddd, or backreference
2970 \xhh character with hex code hh
2971 \x{hhh..} character with hex code hhh..
2972
2973 The precise effect of \cx is as follows: if x is a lower case letter,
2974 it is converted to upper case. Then bit 6 of the character (hex 40) is
2975 inverted. Thus \cz becomes hex 1A, but \c{ becomes hex 3B, while \c;
2976 becomes hex 7B.
2977
2978 After \x, from zero to two hexadecimal digits are read (letters can be
2979 in upper or lower case). Any number of hexadecimal digits may appear
2980 between \x{ and }, but the value of the character code must be less
2981 than 256 in non-UTF-8 mode, and less than 2**31 in UTF-8 mode. That is,
2982 the maximum value in hexadecimal is 7FFFFFFF. Note that this is bigger
2983 than the largest Unicode code point, which is 10FFFF.
2984
2985 If characters other than hexadecimal digits appear between \x{ and },
2986 or if there is no terminating }, this form of escape is not recognized.
2987 Instead, the initial \x will be interpreted as a basic hexadecimal
2988 escape, with no following digits, giving a character whose value is
2989 zero.
2990
2991 Characters whose value is less than 256 can be defined by either of the
2992 two syntaxes for \x. There is no difference in the way they are han-
2993 dled. For example, \xdc is exactly the same as \x{dc}.
2994
2995 After \0 up to two further octal digits are read. If there are fewer
2996 than two digits, just those that are present are used. Thus the
2997 sequence \0\x\07 specifies two binary zeros followed by a BEL character
2998 (code value 7). Make sure you supply two digits after the initial zero
2999 if the pattern character that follows is itself an octal digit.
3000
3001 The handling of a backslash followed by a digit other than 0 is compli-
3002 cated. Outside a character class, PCRE reads it and any following dig-
3003 its as a decimal number. If the number is less than 10, or if there
3004 have been at least that many previous capturing left parentheses in the
3005 expression, the entire sequence is taken as a back reference. A
3006 description of how this works is given later, following the discussion
3007 of parenthesized subpatterns.
3008
3009 Inside a character class, or if the decimal number is greater than 9
3010 and there have not been that many capturing subpatterns, PCRE re-reads
3011 up to three octal digits following the backslash, and uses them to gen-
3012 erate a data character. Any subsequent digits stand for themselves. In
3013 non-UTF-8 mode, the value of a character specified in octal must be
3014 less than \400. In UTF-8 mode, values up to \777 are permitted. For
3015 example:
3016
3017 \040 is another way of writing a space
3018 \40 is the same, provided there are fewer than 40
3019 previous capturing subpatterns
3020 \7 is always a back reference
3021 \11 might be a back reference, or another way of
3022 writing a tab
3023 \011 is always a tab
3024 \0113 is a tab followed by the character "3"
3025 \113 might be a back reference, otherwise the
3026 character with octal code 113
3027 \377 might be a back reference, otherwise
3028 the byte consisting entirely of 1 bits
3029 \81 is either a back reference, or a binary zero
3030 followed by the two characters "8" and "1"
3031
3032 Note that octal values of 100 or greater must not be introduced by a
3033 leading zero, because no more than three octal digits are ever read.
3034
3035 All the sequences that define a single character value can be used both
3036 inside and outside character classes. In addition, inside a character
3037 class, the sequence \b is interpreted as the backspace character (hex
3038 08), and the sequences \R and \X are interpreted as the characters "R"
3039 and "X", respectively. Outside a character class, these sequences have
3040 different meanings (see below).
3041
3042 Absolute and relative back references
3043
3044 The sequence \g followed by an unsigned or a negative number, option-
3045 ally enclosed in braces, is an absolute or relative back reference. A
3046 named back reference can be coded as \g{name}. Back references are dis-
3047 cussed later, following the discussion of parenthesized subpatterns.
3048
3049 Generic character types
3050
3051 Another use of backslash is for specifying generic character types. The
3052 following are always recognized:
3053
3054 \d any decimal digit
3055 \D any character that is not a decimal digit
3056 \h any horizontal whitespace character
3057 \H any character that is not a horizontal whitespace character
3058 \s any whitespace character
3059 \S any character that is not a whitespace character
3060 \v any vertical whitespace character
3061 \V any character that is not a vertical whitespace character
3062 \w any "word" character
3063 \W any "non-word" character
3064
3065 Each pair of escape sequences partitions the complete set of characters
3066 into two disjoint sets. Any given character matches one, and only one,
3067 of each pair.
3068
3069 These character type sequences can appear both inside and outside char-
3070 acter classes. They each match one character of the appropriate type.
3071 If the current matching point is at the end of the subject string, all
3072 of them fail, since there is no character to match.
3073
3074 For compatibility with Perl, \s does not match the VT character (code
3075 11). This makes it different from the the POSIX "space" class. The \s
3076 characters are HT (9), LF (10), FF (12), CR (13), and space (32). If
3077 "use locale;" is included in a Perl script, \s may match the VT charac-
3078 ter. In PCRE, it never does.
3079
3080 In UTF-8 mode, characters with values greater than 128 never match \d,
3081 \s, or \w, and always match \D, \S, and \W. This is true even when Uni-
3082 code character property support is available. These sequences retain
3083 their original meanings from before UTF-8 support was available, mainly
3084 for efficiency reasons.
3085
3086 The sequences \h, \H, \v, and \V are Perl 5.10 features. In contrast to
3087 the other sequences, these do match certain high-valued codepoints in
3088 UTF-8 mode. The horizontal space characters are:
3089
3090 U+0009 Horizontal tab
3091 U+0020 Space
3092 U+00A0 Non-break space
3093 U+1680 Ogham space mark
3094 U+180E Mongolian vowel separator
3095 U+2000 En quad
3096 U+2001 Em quad
3097 U+2002 En space
3098 U+2003 Em space
3099 U+2004 Three-per-em space
3100 U+2005 Four-per-em space
3101 U+2006 Six-per-em space
3102 U+2007 Figure space
3103 U+2008 Punctuation space
3104 U+2009 Thin space
3105 U+200A Hair space
3106 U+202F Narrow no-break space
3107 U+205F Medium mathematical space
3108 U+3000 Ideographic space
3109
3110 The vertical space characters are:
3111
3112 U+000A Linefeed
3113 U+000B Vertical tab
3114 U+000C Formfeed
3115 U+000D Carriage return
3116 U+0085 Next line
3117 U+2028 Line separator
3118 U+2029 Paragraph separator
3119
3120 A "word" character is an underscore or any character less than 256 that
3121 is a letter or digit. The definition of letters and digits is con-
3122 trolled by PCRE's low-valued character tables, and may vary if locale-
3123 specific matching is taking place (see "Locale support" in the pcreapi
3124 page). For example, in a French locale such as "fr_FR" in Unix-like
3125 systems, or "french" in Windows, some character codes greater than 128
3126 are used for accented letters, and these are matched by \w. The use of
3127 locales with Unicode is discouraged.
3128
3129 Newline sequences
3130
3131 Outside a character class, the escape sequence \R matches any Unicode
3132 newline sequence. This is a Perl 5.10 feature. In non-UTF-8 mode \R is
3133 equivalent to the following:
3134
3135 (?>\r\n|\n|\x0b|\f|\r|\x85)
3136
3137 This is an example of an "atomic group", details of which are given
3138 below. This particular group matches either the two-character sequence
3139 CR followed by LF, or one of the single characters LF (linefeed,
3140 U+000A), VT (vertical tab, U+000B), FF (formfeed, U+000C), CR (carriage
3141 return, U+000D), or NEL (next line, U+0085). The two-character sequence
3142 is treated as a single unit that cannot be split.
3143
3144 In UTF-8 mode, two additional characters whose codepoints are greater
3145 than 255 are added: LS (line separator, U+2028) and PS (paragraph sepa-
3146 rator, U+2029). Unicode character property support is not needed for
3147 these characters to be recognized.
3148
3149 Inside a character class, \R matches the letter "R".
3150
3151 Unicode character properties
3152
3153 When PCRE is built with Unicode character property support, three addi-
3154 tional escape sequences that match characters with specific properties
3155 are available. When not in UTF-8 mode, these sequences are of course
3156 limited to testing characters whose codepoints are less than 256, but
3157 they do work in this mode. The extra escape sequences are:
3158
3159 \p{xx} a character with the xx property
3160 \P{xx} a character without the xx property
3161 \X an extended Unicode sequence
3162
3163 The property names represented by xx above are limited to the Unicode
3164 script names, the general category properties, and "Any", which matches
3165 any character (including newline). Other properties such as "InMusical-
3166 Symbols" are not currently supported by PCRE. Note that \P{Any} does
3167 not match any characters, so always causes a match failure.
3168
3169 Sets of Unicode characters are defined as belonging to certain scripts.
3170 A character from one of these sets can be matched using a script name.
3171 For example:
3172
3173 \p{Greek}
3174 \P{Han}
3175
3176 Those that are not part of an identified script are lumped together as
3177 "Common". The current list of scripts is:
3178
3179 Arabic, Armenian, Balinese, Bengali, Bopomofo, Braille, Buginese,
3180 Buhid, Canadian_Aboriginal, Cherokee, Common, Coptic, Cuneiform,
3181 Cypriot, Cyrillic, Deseret, Devanagari, Ethiopic, Georgian, Glagolitic,
3182 Gothic, Greek, Gujarati, Gurmukhi, Han, Hangul, Hanunoo, Hebrew, Hira-
3183 gana, Inherited, Kannada, Katakana, Kharoshthi, Khmer, Lao, Latin,
3184 Limbu, Linear_B, Malayalam, Mongolian, Myanmar, New_Tai_Lue, Nko,
3185 Ogham, Old_Italic, Old_Persian, Oriya, Osmanya, Phags_Pa, Phoenician,
3186 Runic, Shavian, Sinhala, Syloti_Nagri, Syriac, Tagalog, Tagbanwa,
3187 Tai_Le, Tamil, Telugu, Thaana, Thai, Tibetan, Tifinagh, Ugaritic, Yi.
3188
3189 Each character has exactly one general category property, specified by
3190 a two-letter abbreviation. For compatibility with Perl, negation can be
3191 specified by including a circumflex between the opening brace and the
3192 property name. For example, \p{^Lu} is the same as \P{Lu}.
3193
3194 If only one letter is specified with \p or \P, it includes all the gen-
3195 eral category properties that start with that letter. In this case, in
3196 the absence of negation, the curly brackets in the escape sequence are
3197 optional; these two examples have the same effect:
3198
3199 \p{L}
3200 \pL
3201
3202 The following general category property codes are supported:
3203
3204 C Other
3205 Cc Control
3206 Cf Format
3207 Cn Unassigned
3208 Co Private use
3209 Cs Surrogate
3210
3211 L Letter
3212 Ll Lower case letter
3213 Lm Modifier letter
3214 Lo Other letter
3215 Lt Title case letter
3216 Lu Upper case letter
3217
3218 M Mark
3219 Mc Spacing mark
3220 Me Enclosing mark
3221 Mn Non-spacing mark
3222
3223 N Number
3224 Nd Decimal number
3225 Nl Letter number
3226 No Other number
3227
3228 P Punctuation
3229 Pc Connector punctuation
3230 Pd Dash punctuation
3231 Pe Close punctuation
3232 Pf Final punctuation
3233 Pi Initial punctuation
3234 Po Other punctuation
3235 Ps Open punctuation
3236
3237 S Symbol
3238 Sc Currency symbol
3239 Sk Modifier symbol
3240 Sm Mathematical symbol
3241 So Other symbol
3242
3243 Z Separator
3244 Zl Line separator
3245 Zp Paragraph separator
3246 Zs Space separator
3247
3248 The special property L& is also supported: it matches a character that
3249 has the Lu, Ll, or Lt property, in other words, a letter that is not
3250 classified as a modifier or "other".
3251
3252 The Cs (Surrogate) property applies only to characters in the range
3253 U+D800 to U+DFFF. Such characters are not valid in UTF-8 strings (see
3254 RFC 3629) and so cannot be tested by PCRE, unless UTF-8 validity check-
3255 ing has been turned off (see the discussion of PCRE_NO_UTF8_CHECK in
3256 the pcreapi page).
3257
3258 The long synonyms for these properties that Perl supports (such as
3259 \p{Letter}) are not supported by PCRE, nor is it permitted to prefix
3260 any of these properties with "Is".
3261
3262 No character that is in the Unicode table has the Cn (unassigned) prop-
3263 erty. Instead, this property is assumed for any code point that is not
3264 in the Unicode table.
3265
3266 Specifying caseless matching does not affect these escape sequences.
3267 For example, \p{Lu} always matches only upper case letters.
3268
3269 The \X escape matches any number of Unicode characters that form an
3270 extended Unicode sequence. \X is equivalent to
3271
3272 (?>\PM\pM*)
3273
3274 That is, it matches a character without the "mark" property, followed
3275 by zero or more characters with the "mark" property, and treats the
3276 sequence as an atomic group (see below). Characters with the "mark"
3277 property are typically accents that affect the preceding character.
3278 None of them have codepoints less than 256, so in non-UTF-8 mode \X
3279 matches any one character.
3280
3281 Matching characters by Unicode property is not fast, because PCRE has
3282 to search a structure that contains data for over fifteen thousand
3283 characters. That is why the traditional escape sequences such as \d and
3284 \w do not use Unicode properties in PCRE.
3285
3286 Resetting the match start
3287
3288 The escape sequence \K, which is a Perl 5.10 feature, causes any previ-
3289 ously matched characters not to be included in the final matched
3290 sequence. For example, the pattern:
3291
3292 foo\Kbar
3293
3294 matches "foobar", but reports that it has matched "bar". This feature
3295 is similar to a lookbehind assertion (described below). However, in
3296 this case, the part of the subject before the real match does not have
3297 to be of fixed length, as lookbehind assertions do. The use of \K does
3298 not interfere with the setting of captured substrings. For example,
3299 when the pattern
3300
3301 (foo)\Kbar
3302
3303 matches "foobar", the first substring is still set to "foo".
3304
3305 Simple assertions
3306
3307 The final use of backslash is for certain simple assertions. An asser-
3308 tion specifies a condition that has to be met at a particular point in
3309 a match, without consuming any characters from the subject string. The
3310 use of subpatterns for more complicated assertions is described below.
3311 The backslashed assertions are:
3312
3313 \b matches at a word boundary
3314 \B matches when not at a word boundary
3315 \A matches at the start of the subject
3316 \Z matches at the end of the subject
3317 also matches before a newline at the end of the subject
3318 \z matches only at the end of the subject
3319 \G matches at the first matching position in the subject
3320
3321 These assertions may not appear in character classes (but note that \b
3322 has a different meaning, namely the backspace character, inside a char-
3323 acter class).
3324
3325 A word boundary is a position in the subject string where the current
3326 character and the previous character do not both match \w or \W (i.e.
3327 one matches \w and the other matches \W), or the start or end of the
3328 string if the first or last character matches \w, respectively.
3329
3330 The \A, \Z, and \z assertions differ from the traditional circumflex
3331 and dollar (described in the next section) in that they only ever match
3332 at the very start and end of the subject string, whatever options are
3333 set. Thus, they are independent of multiline mode. These three asser-
3334 tions are not affected by the PCRE_NOTBOL or PCRE_NOTEOL options, which
3335 affect only the behaviour of the circumflex and dollar metacharacters.
3336 However, if the startoffset argument of pcre_exec() is non-zero, indi-
3337 cating that matching is to start at a point other than the beginning of
3338 the subject, \A can never match. The difference between \Z and \z is
3339 that \Z matches before a newline at the end of the string as well as at
3340 the very end, whereas \z matches only at the end.
3341
3342 The \G assertion is true only when the current matching position is at
3343 the start point of the match, as specified by the startoffset argument
3344 of pcre_exec(). It differs from \A when the value of startoffset is
3345 non-zero. By calling pcre_exec() multiple times with appropriate argu-
3346 ments, you can mimic Perl's /g option, and it is in this kind of imple-
3347 mentation where \G can be useful.
3348
3349 Note, however, that PCRE's interpretation of \G, as the start of the
3350 current match, is subtly different from Perl's, which defines it as the
3351 end of the previous match. In Perl, these can be different when the
3352 previously matched string was empty. Because PCRE does just one match
3353 at a time, it cannot reproduce this behaviour.
3354
3355 If all the alternatives of a pattern begin with \G, the expression is
3356 anchored to the starting match position, and the "anchored" flag is set
3357 in the compiled regular expression.
3358
3359
3360 CIRCUMFLEX AND DOLLAR
3361
3362 Outside a character class, in the default matching mode, the circumflex
3363 character is an assertion that is true only if the current matching
3364 point is at the start of the subject string. If the startoffset argu-
3365 ment of pcre_exec() is non-zero, circumflex can never match if the
3366 PCRE_MULTILINE option is unset. Inside a character class, circumflex
3367 has an entirely different meaning (see below).
3368
3369 Circumflex need not be the first character of the pattern if a number
3370 of alternatives are involved, but it should be the first thing in each
3371 alternative in which it appears if the pattern is ever to match that
3372 branch. If all possible alternatives start with a circumflex, that is,
3373 if the pattern is constrained to match only at the start of the sub-
3374 ject, it is said to be an "anchored" pattern. (There are also other
3375 constructs that can cause a pattern to be anchored.)
3376
3377 A dollar character is an assertion that is true only if the current
3378 matching point is at the end of the subject string, or immediately
3379 before a newline at the end of the string (by default). Dollar need not
3380 be the last character of the pattern if a number of alternatives are
3381 involved, but it should be the last item in any branch in which it
3382 appears. Dollar has no special meaning in a character class.
3383
3384 The meaning of dollar can be changed so that it matches only at the
3385 very end of the string, by setting the PCRE_DOLLAR_ENDONLY option at
3386 compile time. This does not affect the \Z assertion.
3387
3388 The meanings of the circumflex and dollar characters are changed if the
3389 PCRE_MULTILINE option is set. When this is the case, a circumflex
3390 matches immediately after internal newlines as well as at the start of
3391 the subject string. It does not match after a newline that ends the
3392 string. A dollar matches before any newlines in the string, as well as
3393 at the very end, when PCRE_MULTILINE is set. When newline is specified
3394 as the two-character sequence CRLF, isolated CR and LF characters do
3395 not indicate newlines.
3396
3397 For example, the pattern /^abc$/ matches the subject string "def\nabc"
3398 (where \n represents a newline) in multiline mode, but not otherwise.
3399 Consequently, patterns that are anchored in single line mode because
3400 all branches start with ^ are not anchored in multiline mode, and a
3401 match for circumflex is possible when the startoffset argument of
3402 pcre_exec() is non-zero. The PCRE_DOLLAR_ENDONLY option is ignored if
3403 PCRE_MULTILINE is set.
3404
3405 Note that the sequences \A, \Z, and \z can be used to match the start
3406 and end of the subject in both modes, and if all branches of a pattern
3407 start with \A it is always anchored, whether or not PCRE_MULTILINE is
3408 set.
3409
3410
3411 FULL STOP (PERIOD, DOT)
3412
3413 Outside a character class, a dot in the pattern matches any one charac-
3414 ter in the subject string except (by default) a character that signi-
3415 fies the end of a line. In UTF-8 mode, the matched character may be
3416 more than one byte long.
3417
3418 When a line ending is defined as a single character, dot never matches
3419 that character; when the two-character sequence CRLF is used, dot does
3420 not match CR if it is immediately followed by LF, but otherwise it
3421 matches all characters (including isolated CRs and LFs). When any Uni-
3422 code line endings are being recognized, dot does not match CR or LF or
3423 any of the other line ending characters.
3424
3425 The behaviour of dot with regard to newlines can be changed. If the
3426 PCRE_DOTALL option is set, a dot matches any one character, without
3427 exception. If the two-character sequence CRLF is present in the subject
3428 string, it takes two dots to match it.
3429
3430 The handling of dot is entirely independent of the handling of circum-
3431 flex and dollar, the only relationship being that they both involve
3432 newlines. Dot has no special meaning in a character class.
3433
3434
3435 MATCHING A SINGLE BYTE
3436
3437 Outside a character class, the escape sequence \C matches any one byte,
3438 both in and out of UTF-8 mode. Unlike a dot, it always matches any
3439 line-ending characters. The feature is provided in Perl in order to
3440 match individual bytes in UTF-8 mode. Because it breaks up UTF-8 char-
3441 acters into individual bytes, what remains in the string may be a mal-
3442 formed UTF-8 string. For this reason, the \C escape sequence is best
3443 avoided.
3444
3445 PCRE does not allow \C to appear in lookbehind assertions (described
3446 below), because in UTF-8 mode this would make it impossible to calcu-
3447 late the length of the lookbehind.
3448
3449
3450 SQUARE BRACKETS AND CHARACTER CLASSES
3451
3452 An opening square bracket introduces a character class, terminated by a
3453 closing square bracket. A closing square bracket on its own is not spe-
3454 cial. If a closing square bracket is required as a member of the class,
3455 it should be the first data character in the class (after an initial
3456 circumflex, if present) or escaped with a backslash.
3457
3458 A character class matches a single character in the subject. In UTF-8
3459 mode, the character may occupy more than one byte. A matched character
3460 must be in the set of characters defined by the class, unless the first
3461 character in the class definition is a circumflex, in which case the
3462 subject character must not be in the set defined by the class. If a
3463 circumflex is actually required as a member of the class, ensure it is
3464 not the first character, or escape it with a backslash.
3465
3466 For example, the character class [aeiou] matches any lower case vowel,
3467 while [^aeiou] matches any character that is not a lower case vowel.
3468 Note that a circumflex is just a convenient notation for specifying the
3469 characters that are in the class by enumerating those that are not. A
3470 class that starts with a circumflex is not an assertion: it still con-
3471 sumes a character from the subject string, and therefore it fails if
3472 the current pointer is at the end of the string.
3473
3474 In UTF-8 mode, characters with values greater than 255 can be included
3475 in a class as a literal string of bytes, or by using the \x{ escaping
3476 mechanism.
3477
3478 When caseless matching is set, any letters in a class represent both
3479 their upper case and lower case versions, so for example, a caseless
3480 [aeiou] matches "A" as well as "a", and a caseless [^aeiou] does not
3481 match "A", whereas a caseful version would. In UTF-8 mode, PCRE always
3482 understands the concept of case for characters whose values are less
3483 than 128, so caseless matching is always possible. For characters with
3484 higher values, the concept of case is supported if PCRE is compiled
3485 with Unicode property support, but not otherwise. If you want to use
3486 caseless matching for characters 128 and above, you must ensure that
3487 PCRE is compiled with Unicode property support as well as with UTF-8
3488 support.
3489
3490 Characters that might indicate line breaks are never treated in any
3491 special way when matching character classes, whatever line-ending
3492 sequence is in use, and whatever setting of the PCRE_DOTALL and
3493 PCRE_MULTILINE options is used. A class such as [^a] always matches one
3494 of these characters.
3495
3496 The minus (hyphen) character can be used to specify a range of charac-
3497 ters in a character class. For example, [d-m] matches any letter
3498 between d and m, inclusive. If a minus character is required in a
3499 class, it must be escaped with a backslash or appear in a position
3500 where it cannot be interpreted as indicating a range, typically as the
3501 first or last character in the class.
3502
3503 It is not possible to have the literal character "]" as the end charac-
3504 ter of a range. A pattern such as [W-]46] is interpreted as a class of
3505 two characters ("W" and "-") followed by a literal string "46]", so it
3506 would match "W46]" or "-46]". However, if the "]" is escaped with a
3507 backslash it is interpreted as the end of range, so [W-\]46] is inter-
3508 preted as a class containing a range followed by two other characters.
3509 The octal or hexadecimal representation of "]" can also be used to end
3510 a range.
3511
3512 Ranges operate in the collating sequence of character values. They can
3513 also be used for characters specified numerically, for example
3514 [\000-\037]. In UTF-8 mode, ranges can include characters whose values
3515 are greater than 255, for example [\x{100}-\x{2ff}].
3516
3517 If a range that includes letters is used when caseless matching is set,
3518 it matches the letters in either case. For example, [W-c] is equivalent
3519 to [][\\^_`wxyzabc], matched caselessly, and in non-UTF-8 mode, if
3520 character tables for a French locale are in use, [\xc8-\xcb] matches
3521 accented E characters in both cases. In UTF-8 mode, PCRE supports the
3522 concept of case for characters with values greater than 128 only when
3523 it is compiled with Unicode property support.
3524
3525 The character types \d, \D, \p, \P, \s, \S, \w, and \W may also appear
3526 in a character class, and add the characters that they match to the
3527 class. For example, [\dABCDEF] matches any hexadecimal digit. A circum-
3528 flex can conveniently be used with the upper case character types to
3529 specify a more restricted set of characters than the matching lower
3530 case type. For example, the class [^\W_] matches any letter or digit,
3531 but not underscore.
3532
3533 The only metacharacters that are recognized in character classes are
3534 backslash, hyphen (only where it can be interpreted as specifying a
3535 range), circumflex (only at the start), opening square bracket (only
3536 when it can be interpreted as introducing a POSIX class name - see the
3537 next section), and the terminating closing square bracket. However,
3538 escaping other non-alphanumeric characters does no harm.
3539
3540
3541 POSIX CHARACTER CLASSES
3542
3543 Perl supports the POSIX notation for character classes. This uses names
3544 enclosed by [: and :] within the enclosing square brackets. PCRE also
3545 supports this notation. For example,
3546
3547 [01[:alpha:]%]
3548
3549 matches "0", "1", any alphabetic character, or "%". The supported class
3550 names are
3551
3552 alnum letters and digits
3553 alpha letters
3554 ascii character codes 0 - 127
3555 blank space or tab only
3556 cntrl control characters
3557 digit decimal digits (same as \d)
3558 graph printing characters, excluding space
3559 lower lower case letters
3560 print printing characters, including space
3561 punct printing characters, excluding letters and digits
3562 space white space (not quite the same as \s)
3563 upper upper case letters
3564 word "word" characters (same as \w)
3565 xdigit hexadecimal digits
3566
3567 The "space" characters are HT (9), LF (10), VT (11), FF (12), CR (13),
3568 and space (32). Notice that this list includes the VT character (code
3569 11). This makes "space" different to \s, which does not include VT (for
3570 Perl compatibility).
3571
3572 The name "word" is a Perl extension, and "blank" is a GNU extension
3573 from Perl 5.8. Another Perl extension is negation, which is indicated
3574 by a ^ character after the colon. For example,
3575
3576 [12[:^digit:]]
3577
3578 matches "1", "2", or any non-digit. PCRE (and Perl) also recognize the
3579 POSIX syntax [.ch.] and [=ch=] where "ch" is a "collating element", but
3580 these are not supported, and an error is given if they are encountered.
3581
3582 In UTF-8 mode, characters with values greater than 128 do not match any
3583 of the POSIX character classes.
3584
3585
3586 VERTICAL BAR
3587
3588 Vertical bar characters are used to separate alternative patterns. For
3589 example, the pattern
3590
3591 gilbert|sullivan
3592
3593 matches either "gilbert" or "sullivan". Any number of alternatives may
3594 appear, and an empty alternative is permitted (matching the empty
3595 string). The matching process tries each alternative in turn, from left
3596 to right, and the first one that succeeds is used. If the alternatives
3597 are within a subpattern (defined below), "succeeds" means matching the
3598 rest of the main pattern as well as the alternative in the subpattern.
3599
3600
3601 INTERNAL OPTION SETTING
3602
3603 The settings of the PCRE_CASELESS, PCRE_MULTILINE, PCRE_DOTALL, and
3604 PCRE_EXTENDED options can be changed from within the pattern by a
3605 sequence of Perl option letters enclosed between "(?" and ")". The
3606 option letters are
3607
3608 i for PCRE_CASELESS
3609 m for PCRE_MULTILINE
3610 s for PCRE_DOTALL
3611 x for PCRE_EXTENDED
3612
3613 For example, (?im) sets caseless, multiline matching. It is also possi-
3614 ble to unset these options by preceding the letter with a hyphen, and a
3615 combined setting and unsetting such as (?im-sx), which sets PCRE_CASE-
3616 LESS and PCRE_MULTILINE while unsetting PCRE_DOTALL and PCRE_EXTENDED,
3617 is also permitted. If a letter appears both before and after the
3618 hyphen, the option is unset.
3619
3620 When an option change occurs at top level (that is, not inside subpat-
3621 tern parentheses), the change applies to the remainder of the pattern
3622 that follows. If the change is placed right at the start of a pattern,
3623 PCRE extracts it into the global options (and it will therefore show up
3624 in data extracted by the pcre_fullinfo() function).
3625
3626 An option change within a subpattern (see below for a description of
3627 subpatterns) affects only that part of the current pattern that follows
3628 it, so
3629
3630 (a(?i)b)c
3631
3632 matches abc and aBc and no other strings (assuming PCRE_CASELESS is not
3633 used). By this means, options can be made to have different settings
3634 in different parts of the pattern. Any changes made in one alternative
3635 do carry on into subsequent branches within the same subpattern. For
3636 example,
3637
3638 (a(?i)b|c)
3639
3640 matches "ab", "aB", "c", and "C", even though when matching "C" the
3641 first branch is abandoned before the option setting. This is because
3642 the effects of option settings happen at compile time. There would be
3643 some very weird behaviour otherwise.
3644
3645 The PCRE-specific options PCRE_DUPNAMES, PCRE_UNGREEDY, and PCRE_EXTRA
3646 can be changed in the same way as the Perl-compatible options by using
3647 the characters J, U and X respectively.
3648
3649
3650 SUBPATTERNS
3651
3652 Subpatterns are delimited by parentheses (round brackets), which can be
3653 nested. Turning part of a pattern into a subpattern does two things:
3654
3655 1. It localizes a set of alternatives. For example, the pattern
3656
3657 cat(aract|erpillar|)
3658
3659 matches one of the words "cat", "cataract", or "caterpillar". Without
3660 the parentheses, it would match "cataract", "erpillar" or an empty
3661 string.
3662
3663 2. It sets up the subpattern as a capturing subpattern. This means
3664 that, when the whole pattern matches, that portion of the subject
3665 string that matched the subpattern is passed back to the caller via the
3666 ovector argument of pcre_exec(). Opening parentheses are counted from
3667 left to right (starting from 1) to obtain numbers for the capturing
3668 subpatterns.
3669
3670 For example, if the string "the red king" is matched against the pat-
3671 tern
3672
3673 the ((red|white) (king|queen))
3674
3675 the captured substrings are "red king", "red", and "king", and are num-
3676 bered 1, 2, and 3, respectively.
3677
3678 The fact that plain parentheses fulfil two functions is not always
3679 helpful. There are often times when a grouping subpattern is required
3680 without a capturing requirement. If an opening parenthesis is followed
3681 by a question mark and a colon, the subpattern does not do any captur-
3682 ing, and is not counted when computing the number of any subsequent
3683 capturing subpatterns. For example, if the string "the white queen" is
3684 matched against the pattern
3685
3686 the ((?:red|white) (king|queen))
3687
3688 the captured substrings are "white queen" and "queen", and are numbered
3689 1 and 2. The maximum number of capturing subpatterns is 65535.
3690
3691 As a convenient shorthand, if any option settings are required at the
3692 start of a non-capturing subpattern, the option letters may appear
3693 between the "?" and the ":". Thus the two patterns
3694
3695 (?i:saturday|sunday)
3696 (?:(?i)saturday|sunday)
3697
3698 match exactly the same set of strings. Because alternative branches are
3699 tried from left to right, and options are not reset until the end of
3700 the subpattern is reached, an option setting in one branch does affect
3701 subsequent branches, so the above patterns match "SUNDAY" as well as
3702 "Saturday".
3703
3704
3705 DUPLICATE SUBPATTERN NUMBERS
3706
3707 Perl 5.10 introduced a feature whereby each alternative in a subpattern
3708 uses the same numbers for its capturing parentheses. Such a subpattern
3709 starts with (?| and is itself a non-capturing subpattern. For example,
3710 consider this pattern:
3711
3712 (?|(Sat)ur|(Sun))day
3713
3714 Because the two alternatives are inside a (?| group, both sets of cap-
3715 turing parentheses are numbered one. Thus, when the pattern matches,
3716 you can look at captured substring number one, whichever alternative
3717 matched. This construct is useful when you want to capture part, but
3718 not all, of one of a number of alternatives. Inside a (?| group, paren-
3719 theses are numbered as usual, but the number is reset at the start of
3720 each branch. The numbers of any capturing buffers that follow the sub-
3721 pattern start after the highest number used in any branch. The follow-
3722 ing example is taken from the Perl documentation. The numbers under-
3723 neath show in which buffer the captured content will be stored.
3724
3725 # before ---------------branch-reset----------- after
3726 / ( a ) (?| x ( y ) z | (p (q) r) | (t) u (v) ) ( z ) /x
3727 # 1 2 2 3 2 3 4
3728
3729 A backreference or a recursive call to a numbered subpattern always
3730 refers to the first one in the pattern with the given number.
3731
3732 An alternative approach to using this "branch reset" feature is to use
3733 duplicate named subpatterns, as described in the next section.
3734
3735
3736 NAMED SUBPATTERNS
3737
3738 Identifying capturing parentheses by number is simple, but it can be
3739 very hard to keep track of the numbers in complicated regular expres-
3740 sions. Furthermore, if an expression is modified, the numbers may
3741 change. To help with this difficulty, PCRE supports the naming of sub-
3742 patterns. This feature was not added to Perl until release 5.10. Python
3743 had the feature earlier, and PCRE introduced it at release 4.0, using
3744 the Python syntax. PCRE now supports both the Perl and the Python syn-
3745 tax.
3746
3747 In PCRE, a subpattern can be named in one of three ways: (?<name>...)
3748 or (?'name'...) as in Perl, or (?P<name>...) as in Python. References
3749 to capturing parentheses from other parts of the pattern, such as back-
3750 references, recursion, and conditions, can be made by name as well as
3751 by number.
3752
3753 Names consist of up to 32 alphanumeric characters and underscores.
3754 Named capturing parentheses are still allocated numbers as well as
3755 names, exactly as if the names were not present. The PCRE API provides
3756 function calls for extracting the name-to-number translation table from
3757 a compiled pattern. There is also a convenience function for extracting
3758 a captured substring by name.
3759
3760 By default, a name must be unique within a pattern, but it is possible
3761 to relax this constraint by setting the PCRE_DUPNAMES option at compile
3762 time. This can be useful for patterns where only one instance of the
3763 named parentheses can match. Suppose you want to match the name of a
3764 weekday, either as a 3-letter abbreviation or as the full name, and in
3765 both cases you want to extract the abbreviation. This pattern (ignoring
3766 the line breaks) does the job:
3767
3768 (?<DN>Mon|Fri|Sun)(?:day)?|
3769 (?<DN>Tue)(?:sday)?|
3770 (?<DN>Wed)(?:nesday)?|
3771 (?<DN>Thu)(?:rsday)?|
3772 (?<DN>Sat)(?:urday)?
3773
3774 There are five capturing substrings, but only one is ever set after a
3775 match. (An alternative way of solving this problem is to use a "branch
3776 reset" subpattern, as described in the previous section.)
3777
3778 The convenience function for extracting the data by name returns the
3779 substring for the first (and in this example, the only) subpattern of
3780 that name that matched. This saves searching to find which numbered
3781 subpattern it was. If you make a reference to a non-unique named sub-
3782 pattern from elsewhere in the pattern, the one that corresponds to the
3783 lowest number is used. For further details of the interfaces for han-
3784 dling named subpatterns, see the pcreapi documentation.
3785
3786
3787 REPETITION
3788
3789 Repetition is specified by quantifiers, which can follow any of the
3790 following items:
3791
3792 a literal data character
3793 the dot metacharacter
3794 the \C escape sequence
3795 the \X escape sequence (in UTF-8 mode with Unicode properties)
3796 the \R escape sequence
3797 an escape such as \d that matches a single character
3798 a character class
3799 a back reference (see next section)
3800 a parenthesized subpattern (unless it is an assertion)
3801
3802 The general repetition quantifier specifies a minimum and maximum num-
3803 ber of permitted matches, by giving the two numbers in curly brackets
3804 (braces), separated by a comma. The numbers must be less than 65536,
3805 and the first must be less than or equal to the second. For example:
3806
3807 z{2,4}
3808
3809 matches "zz", "zzz", or "zzzz". A closing brace on its own is not a
3810 special character. If the second number is omitted, but the comma is
3811 present, there is no upper limit; if the second number and the comma
3812 are both omitted, the quantifier specifies an exact number of required
3813 matches. Thus
3814
3815 [aeiou]{3,}
3816
3817 matches at least 3 successive vowels, but may match many more, while
3818
3819 \d{8}
3820
3821 matches exactly 8 digits. An opening curly bracket that appears in a
3822 position where a quantifier is not allowed, or one that does not match
3823 the syntax of a quantifier, is taken as a literal character. For exam-
3824 ple, {,6} is not a quantifier, but a literal string of four characters.
3825
3826 In UTF-8 mode, quantifiers apply to UTF-8 characters rather than to
3827 individual bytes. Thus, for example, \x{100}{2} matches two UTF-8 char-
3828 acters, each of which is represented by a two-byte sequence. Similarly,
3829 when Unicode property support is available, \X{3} matches three Unicode
3830 extended sequences, each of which may be several bytes long (and they
3831 may be of different lengths).
3832
3833 The quantifier {0} is permitted, causing the expression to behave as if
3834 the previous item and the quantifier were not present.
3835
3836 For convenience, the three most common quantifiers have single-charac-
3837 ter abbreviations:
3838
3839 * is equivalent to {0,}
3840 + is equivalent to {1,}
3841 ? is equivalent to {0,1}
3842
3843 It is possible to construct infinite loops by following a subpattern
3844 that can match no characters with a quantifier that has no upper limit,
3845 for example:
3846
3847 (a?)*
3848
3849 Earlier versions of Perl and PCRE used to give an error at compile time
3850 for such patterns. However, because there are cases where this can be
3851 useful, such patterns are now accepted, but if any repetition of the
3852 subpattern does in fact match no characters, the loop is forcibly bro-
3853 ken.
3854
3855 By default, the quantifiers are "greedy", that is, they match as much
3856 as possible (up to the maximum number of permitted times), without
3857 causing the rest of the pattern to fail. The classic example of where
3858 this gives problems is in trying to match comments in C programs. These
3859 appear between /* and */ and within the comment, individual * and /
3860 characters may appear. An attempt to match C comments by applying the
3861 pattern
3862
3863 /\*.*\*/
3864
3865 to the string
3866
3867 /* first comment */ not comment /* second comment */
3868
3869 fails, because it matches the entire string owing to the greediness of
3870 the .* item.
3871
3872 However, if a quantifier is followed by a question mark, it ceases to
3873 be greedy, and instead matches the minimum number of times possible, so
3874 the pattern
3875
3876 /\*.*?\*/
3877
3878 does the right thing with the C comments. The meaning of the various
3879 quantifiers is not otherwise changed, just the preferred number of
3880 matches. Do not confuse this use of question mark with its use as a
3881 quantifier in its own right. Because it has two uses, it can sometimes
3882 appear doubled, as in
3883
3884 \d??\d
3885
3886 which matches one digit by preference, but can match two if that is the
3887 only way the rest of the pattern matches.
3888
3889 If the PCRE_UNGREEDY option is set (an option that is not available in
3890 Perl), the quantifiers are not greedy by default, but individual ones
3891 can be made greedy by following them with a question mark. In other
3892 words, it inverts the default behaviour.
3893
3894 When a parenthesized subpattern is quantified with a minimum repeat
3895 count that is greater than 1 or with a limited maximum, more memory is
3896 required for the compiled pattern, in proportion to the size of the
3897 minimum or maximum.
3898
3899 If a pattern starts with .* or .{0,} and the PCRE_DOTALL option (equiv-
3900 alent to Perl's /s) is set, thus allowing the dot to match newlines,
3901 the pattern is implicitly anchored, because whatever follows will be
3902 tried against every character position in the subject string, so there
3903 is no point in retrying the overall match at any position after the
3904 first. PCRE normally treats such a pattern as though it were preceded
3905 by \A.
3906
3907 In cases where it is known that the subject string contains no new-
3908 lines, it is worth setting PCRE_DOTALL in order to obtain this opti-
3909 mization, or alternatively using ^ to indicate anchoring explicitly.
3910
3911 However, there is one situation where the optimization cannot be used.
3912 When .* is inside capturing parentheses that are the subject of a
3913 backreference elsewhere in the pattern, a match at the start may fail
3914 where a later one succeeds. Consider, for example:
3915
3916 (.*)abc\1
3917
3918 If the subject is "xyz123abc123" the match point is the fourth charac-
3919 ter. For this reason, such a pattern is not implicitly anchored.
3920
3921 When a capturing subpattern is repeated, the value captured is the sub-
3922 string that matched the final iteration. For example, after
3923
3924 (tweedle[dume]{3}\s*)+
3925
3926 has matched "tweedledum tweedledee" the value of the captured substring
3927 is "tweedledee". However, if there are nested capturing subpatterns,
3928 the corresponding captured values may have been set in previous itera-
3929 tions. For example, after
3930
3931 /(a|(b))+/
3932
3933 matches "aba" the value of the second captured substring is "b".
3934
3935
3936 ATOMIC GROUPING AND POSSESSIVE QUANTIFIERS
3937
3938 With both maximizing ("greedy") and minimizing ("ungreedy" or "lazy")
3939 repetition, failure of what follows normally causes the repeated item
3940 to be re-evaluated to see if a different number of repeats allows the
3941 rest of the pattern to match. Sometimes it is useful to prevent this,
3942 either to change the nature of the match, or to cause it fail earlier
3943 than it otherwise might, when the author of the pattern knows there is
3944 no point in carrying on.
3945
3946 Consider, for example, the pattern \d+foo when applied to the subject
3947 line
3948
3949 123456bar
3950
3951 After matching all 6 digits and then failing to match "foo", the normal
3952 action of the matcher is to try again with only 5 digits matching the
3953 \d+ item, and then with 4, and so on, before ultimately failing.
3954 "Atomic grouping" (a term taken from Jeffrey Friedl's book) provides
3955 the means for specifying that once a subpattern has matched, it is not
3956 to be re-evaluated in this way.
3957
3958 If we use atomic grouping for the previous example, the matcher gives
3959 up immediately on failing to match "foo" the first time. The notation
3960 is a kind of special parenthesis, starting with (?> as in this example:
3961
3962 (?>\d+)foo
3963
3964 This kind of parenthesis "locks up" the part of the pattern it con-
3965 tains once it has matched, and a failure further into the pattern is
3966 prevented from backtracking into it. Backtracking past it to previous
3967 items, however, works as normal.
3968
3969 An alternative description is that a subpattern of this type matches
3970 the string of characters that an identical standalone pattern would
3971 match, if anchored at the current point in the subject string.
3972
3973 Atomic grouping subpatterns are not capturing subpatterns. Simple cases
3974 such as the above example can be thought of as a maximizing repeat that
3975 must swallow everything it can. So, while both \d+ and \d+? are pre-
3976 pared to adjust the number of digits they match in order to make the
3977 rest of the pattern match, (?>\d+) can only match an entire sequence of
3978 digits.
3979
3980 Atomic groups in general can of course contain arbitrarily complicated
3981 subpatterns, and can be nested. However, when the subpattern for an
3982 atomic group is just a single repeated item, as in the example above, a
3983 simpler notation, called a "possessive quantifier" can be used. This
3984 consists of an additional + character following a quantifier. Using
3985 this notation, the previous example can be rewritten as
3986
3987 \d++foo
3988
3989 Note that a possessive quantifier can be used with an entire group, for
3990 example:
3991
3992 (abc|xyz){2,3}+
3993
3994 Possessive quantifiers are always greedy; the setting of the
3995 PCRE_UNGREEDY option is ignored. They are a convenient notation for the
3996 simpler forms of atomic group. However, there is no difference in the
3997 meaning of a possessive quantifier and the equivalent atomic group,
3998 though there may be a performance difference; possessive quantifiers
3999 should be slightly faster.
4000
4001 The possessive quantifier syntax is an extension to the Perl 5.8 syn-
4002 tax. Jeffrey Friedl originated the idea (and the name) in the first
4003 edition of his book. Mike McCloskey liked it, so implemented it when he
4004 built Sun's Java package, and PCRE copied it from there. It ultimately
4005 found its way into Perl at release 5.10.
4006
4007 PCRE has an optimization that automatically "possessifies" certain sim-
4008 ple pattern constructs. For example, the sequence A+B is treated as
4009 A++B because there is no point in backtracking into a sequence of A's
4010 when B must follow.
4011
4012 When a pattern contains an unlimited repeat inside a subpattern that
4013 can itself be repeated an unlimited number of times, the use of an
4014 atomic group is the only way to avoid some failing matches taking a
4015 very long time indeed. The pattern
4016
4017 (\D+|<\d+>)*[!?]
4018
4019 matches an unlimited number of substrings that either consist of non-
4020 digits, or digits enclosed in <>, followed by either ! or ?. When it
4021 matches, it runs quickly. However, if it is applied to
4022
4023 aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
4024
4025 it takes a long time before reporting failure. This is because the
4026 string can be divided between the internal \D+ repeat and the external
4027 * repeat in a large number of ways, and all have to be tried. (The
4028 example uses [!?] rather than a single character at the end, because
4029 both PCRE and Perl have an optimization that allows for fast failure
4030 when a single character is used. They remember the last single charac-
4031 ter that is required for a match, and fail early if it is not present
4032 in the string.) If the pattern is changed so that it uses an atomic
4033 group, like this:
4034
4035 ((?>\D+)|<\d+>)*[!?]
4036
4037 sequences of non-digits cannot be broken, and failure happens quickly.
4038
4039
4040 BACK REFERENCES
4041
4042 Outside a character class, a backslash followed by a digit greater than
4043 0 (and possibly further digits) is a back reference to a capturing sub-
4044 pattern earlier (that is, to its left) in the pattern, provided there
4045 have been that many previous capturing left parentheses.
4046
4047 However, if the decimal number following the backslash is less than 10,
4048 it is always taken as a back reference, and causes an error only if
4049 there are not that many capturing left parentheses in the entire pat-
4050 tern. In other words, the parentheses that are referenced need not be
4051 to the left of the reference for numbers less than 10. A "forward back
4052 reference" of this type can make sense when a repetition is involved
4053 and the subpattern to the right has participated in an earlier itera-
4054 tion.
4055
4056 It is not possible to have a numerical "forward back reference" to a
4057 subpattern whose number is 10 or more using this syntax because a
4058 sequence such as \50 is interpreted as a character defined in octal.
4059 See the subsection entitled "Non-printing characters" above for further
4060 details of the handling of digits following a backslash. There is no
4061 such problem when named parentheses are used. A back reference to any
4062 subpattern is possible using named parentheses (see below).
4063
4064 Another way of avoiding the ambiguity inherent in the use of digits
4065 following a backslash is to use the \g escape sequence, which is a fea-
4066 ture introduced in Perl 5.10. This escape must be followed by an
4067 unsigned number or a negative number, optionally enclosed in braces.
4068 These examples are all identical:
4069
4070 (ring), \1
4071 (ring), \g1
4072 (ring), \g{1}
4073
4074 An unsigned number specifies an absolute reference without the ambigu-
4075 ity that is present in the older syntax. It is also useful when literal
4076 digits follow the reference. A negative number is a relative reference.
4077 Consider this example:
4078
4079 (abc(def)ghi)\g{-1}
4080
4081 The sequence \g{-1} is a reference to the most recently started captur-
4082 ing subpattern before \g, that is, is it equivalent to \2. Similarly,
4083 \g{-2} would be equivalent to \1. The use of relative references can be
4084 helpful in long patterns, and also in patterns that are created by
4085 joining together fragments that contain references within themselves.
4086
4087 A back reference matches whatever actually matched the capturing sub-
4088 pattern in the current subject string, rather than anything matching
4089 the subpattern itself (see "Subpatterns as subroutines" below for a way
4090 of doing that). So the pattern
4091
4092 (sens|respons)e and \1ibility
4093
4094 matches "sense and sensibility" and "response and responsibility", but
4095 not "sense and responsibility". If caseful matching is in force at the
4096 time of the back reference, the case of letters is relevant. For exam-
4097 ple,
4098
4099 ((?i)rah)\s+\1
4100
4101 matches "rah rah" and "RAH RAH", but not "RAH rah", even though the
4102 original capturing subpattern is matched caselessly.
4103
4104 There are several different ways of writing back references to named
4105 subpatterns. The .NET syntax \k{name} and the Perl syntax \k<name> or
4106 \k'name' are supported, as is the Python syntax (?P=name). Perl 5.10's
4107 unified back reference syntax, in which \g can be used for both numeric
4108 and named references, is also supported. We could rewrite the above
4109 example in any of the following ways:
4110
4111 (?<p1>(?i)rah)\s+\k<p1>
4112 (?'p1'(?i)rah)\s+\k{p1}
4113 (?P<p1>(?i)rah)\s+(?P=p1)
4114 (?<p1>(?i)rah)\s+\g{p1}
4115
4116 A subpattern that is referenced by name may appear in the pattern
4117 before or after the reference.
4118
4119 There may be more than one back reference to the same subpattern. If a
4120 subpattern has not actually been used in a particular match, any back
4121 references to it always fail. For example, the pattern
4122
4123 (a|(bc))\2
4124
4125 always fails if it starts to match "a" rather than "bc". Because there
4126 may be many capturing parentheses in a pattern, all digits following
4127 the backslash are taken as part of a potential back reference number.
4128 If the pattern continues with a digit character, some delimiter must be
4129 used to terminate the back reference. If the PCRE_EXTENDED option is
4130 set, this can be whitespace. Otherwise an empty comment (see "Com-
4131 ments" below) can be used.
4132
4133 A back reference that occurs inside the parentheses to which it refers
4134 fails when the subpattern is first used, so, for example, (a\1) never
4135 matches. However, such references can be useful inside repeated sub-
4136 patterns. For example, the pattern
4137
4138 (a|b\1)+
4139
4140 matches any number of "a"s and also "aba", "ababbaa" etc. At each iter-
4141 ation of the subpattern, the back reference matches the character
4142 string corresponding to the previous iteration. In order for this to
4143 work, the pattern must be such that the first iteration does not need
4144 to match the back reference. This can be done using alternation, as in
4145 the example above, or by a quantifier with a minimum of zero.
4146
4147
4148 ASSERTIONS
4149
4150 An assertion is a test on the characters following or preceding the
4151 current matching point that does not actually consume any characters.
4152 The simple assertions coded as \b, \B, \A, \G, \Z, \z, ^ and $ are
4153 described above.
4154
4155 More complicated assertions are coded as subpatterns. There are two
4156 kinds: those that look ahead of the current position in the subject
4157 string, and those that look behind it. An assertion subpattern is
4158 matched in the normal way, except that it does not cause the current
4159 matching position to be changed.
4160
4161 Assertion subpatterns are not capturing subpatterns, and may not be
4162 repeated, because it makes no sense to assert the same thing several
4163 times. If any kind of assertion contains capturing subpatterns within
4164 it, these are counted for the purposes of numbering the capturing sub-
4165 patterns in the whole pattern. However, substring capturing is carried
4166 out only for positive assertions, because it does not make sense for
4167 negative assertions.
4168
4169 Lookahead assertions
4170
4171 Lookahead assertions start with (?= for positive assertions and (?! for
4172 negative assertions. For example,
4173
4174 \w+(?=;)
4175
4176 matches a word followed by a semicolon, but does not include the semi-
4177 colon in the match, and
4178
4179 foo(?!bar)
4180
4181 matches any occurrence of "foo" that is not followed by "bar". Note
4182 that the apparently similar pattern
4183
4184 (?!foo)bar
4185
4186 does not find an occurrence of "bar" that is preceded by something
4187 other than "foo"; it finds any occurrence of "bar" whatsoever, because
4188 the assertion (?!foo) is always true when the next three characters are
4189 "bar". A lookbehind assertion is needed to achieve the other effect.
4190
4191 If you want to force a matching failure at some point in a pattern, the
4192 most convenient way to do it is with (?!) because an empty string
4193 always matches, so an assertion that requires there not to be an empty
4194 string must always fail.
4195
4196 Lookbehind assertions
4197
4198 Lookbehind assertions start with (?<= for positive assertions and (?<!
4199 for negative assertions. For example,
4200
4201 (?<!foo)bar
4202
4203 does find an occurrence of "bar" that is not preceded by "foo". The
4204 contents of a lookbehind assertion are restricted such that all the
4205 strings it matches must have a fixed length. However, if there are sev-
4206 eral top-level alternatives, they do not all have to have the same
4207 fixed length. Thus
4208
4209 (?<=bullock|donkey)
4210
4211 is permitted, but
4212
4213 (?<!dogs?|cats?)
4214
4215 causes an error at compile time. Branches that match different length
4216 strings are permitted only at the top level of a lookbehind assertion.
4217 This is an extension compared with Perl (at least for 5.8), which
4218 requires all branches to match the same length of string. An assertion
4219 such as
4220
4221 (?<=ab(c|de))
4222
4223 is not permitted, because its single top-level branch can match two
4224 different lengths, but it is acceptable if rewritten to use two top-
4225 level branches:
4226
4227 (?<=abc|abde)
4228
4229 In some cases, the Perl 5.10 escape sequence \K (see above) can be used
4230 instead of a lookbehind assertion; this is not restricted to a fixed-
4231 length.
4232
4233 The implementation of lookbehind assertions is, for each alternative,
4234 to temporarily move the current position back by the fixed length and
4235 then try to match. If there are insufficient characters before the cur-
4236 rent position, the assertion fails.
4237
4238 PCRE does not allow the \C escape (which matches a single byte in UTF-8
4239 mode) to appear in lookbehind assertions, because it makes it impossi-
4240 ble to calculate the length of the lookbehind. The \X and \R escapes,
4241 which can match different numbers of bytes, are also not permitted.
4242
4243 Possessive quantifiers can be used in conjunction with lookbehind
4244 assertions to specify efficient matching at the end of the subject
4245 string. Consider a simple pattern such as
4246
4247 abcd$
4248
4249 when applied to a long string that does not match. Because matching
4250 proceeds from left to right, PCRE will look for each "a" in the subject
4251 and then see if what follows matches the rest of the pattern. If the
4252 pattern is specified as
4253
4254 ^.*abcd$
4255
4256 the initial .* matches the entire string at first, but when this fails
4257 (because there is no following "a"), it backtracks to match all but the
4258 last character, then all but the last two characters, and so on. Once
4259 again the search for "a" covers the entire string, from right to left,
4260 so we are no better off. However, if the pattern is written as
4261
4262 ^.*+(?<=abcd)
4263
4264 there can be no backtracking for the .*+ item; it can match only the
4265 entire string. The subsequent lookbehind assertion does a single test
4266 on the last four characters. If it fails, the match fails immediately.
4267 For long strings, this approach makes a significant difference to the
4268 processing time.
4269
4270 Using multiple assertions
4271
4272 Several assertions (of any sort) may occur in succession. For example,
4273
4274 (?<=\d{3})(?<!999)foo
4275
4276 matches "foo" preceded by three digits that are not "999". Notice that
4277 each of the assertions is applied independently at the same point in
4278 the subject string. First there is a check that the previous three
4279 characters are all digits, and then there is a check that the same
4280 three characters are not "999". This pattern does not match "foo" pre-
4281 ceded by six characters, the first of which are digits and the last
4282 three of which are not "999". For example, it doesn't match "123abc-
4283 foo". A pattern to do that is
4284
4285 (?<=\d{3}...)(?<!999)foo
4286
4287 This time the first assertion looks at the preceding six characters,
4288 checking that the first three are digits, and then the second assertion
4289 checks that the preceding three characters are not "999".
4290
4291 Assertions can be nested in any combination. For example,
4292
4293 (?<=(?<!foo)bar)baz
4294
4295 matches an occurrence of "baz" that is preceded by "bar" which in turn
4296 is not preceded by "foo", while
4297
4298 (?<=\d{3}(?!999)...)foo
4299
4300 is another pattern that matches "foo" preceded by three digits and any
4301 three characters that are not "999".
4302
4303
4304 CONDITIONAL SUBPATTERNS
4305
4306 It is possible to cause the matching process to obey a subpattern con-
4307 ditionally or to choose between two alternative subpatterns, depending
4308 on the result of an assertion, or whether a previous capturing subpat-
4309 tern matched or not. The two possible forms of conditional subpattern
4310 are
4311
4312 (?(condition)yes-pattern)
4313 (?(condition)yes-pattern|no-pattern)
4314
4315 If the condition is satisfied, the yes-pattern is used; otherwise the
4316 no-pattern (if present) is used. If there are more than two alterna-
4317 tives in the subpattern, a compile-time error occurs.
4318
4319 There are four kinds of condition: references to subpatterns, refer-
4320 ences to recursion, a pseudo-condition called DEFINE, and assertions.
4321
4322 Checking for a used subpattern by number
4323
4324 If the text between the parentheses consists of a sequence of digits,
4325 the condition is true if the capturing subpattern of that number has
4326 previously matched. An alternative notation is to precede the digits
4327 with a plus or minus sign. In this case, the subpattern number is rela-
4328 tive rather than absolute. The most recently opened parentheses can be
4329 referenced by (?(-1), the next most recent by (?(-2), and so on. In
4330 looping constructs it can also make sense to refer to subsequent groups
4331 with constructs such as (?(+2).
4332
4333 Consider the following pattern, which contains non-significant white
4334 space to make it more readable (assume the PCRE_EXTENDED option) and to
4335 divide it into three parts for ease of discussion:
4336
4337 ( \( )? [^()]+ (?(1) \) )
4338
4339 The first part matches an optional opening parenthesis, and if that
4340 character is present, sets it as the first captured substring. The sec-
4341 ond part matches one or more characters that are not parentheses. The
4342 third part is a conditional subpattern that tests whether the first set
4343 of parentheses matched or not. If they did, that is, if subject started
4344 with an opening parenthesis, the condition is true, and so the yes-pat-
4345 tern is executed and a closing parenthesis is required. Otherwise,
4346 since no-pattern is not present, the subpattern matches nothing. In
4347 other words, this pattern matches a sequence of non-parentheses,
4348 optionally enclosed in parentheses.
4349
4350 If you were embedding this pattern in a larger one, you could use a
4351 relative reference:
4352
4353 ...other stuff... ( \( )? [^()]+ (?(-1) \) ) ...
4354
4355 This makes the fragment independent of the parentheses in the larger
4356 pattern.
4357
4358 Checking for a used subpattern by name
4359
4360 Perl uses the syntax (?(<name>)...) or (?('name')...) to test for a
4361 used subpattern by name. For compatibility with earlier versions of
4362 PCRE, which had this facility before Perl, the syntax (?(name)...) is
4363 also recognized. However, there is a possible ambiguity with this syn-
4364 tax, because subpattern names may consist entirely of digits. PCRE
4365 looks first for a named subpattern; if it cannot find one and the name
4366 consists entirely of digits, PCRE looks for a subpattern of that num-
4367 ber, which must be greater than zero. Using subpattern names that con-
4368 sist entirely of digits is not recommended.
4369
4370 Rewriting the above example to use a named subpattern gives this:
4371
4372 (?<OPEN> \( )? [^()]+ (?(<OPEN>) \) )
4373
4374
4375 Checking for pattern recursion
4376
4377 If the condition is the string (R), and there is no subpattern with the
4378 name R, the condition is true if a recursive call to the whole pattern
4379 or any subpattern has been made. If digits or a name preceded by amper-
4380 sand follow the letter R, for example:
4381
4382 (?(R3)...) or (?(R&name)...)
4383
4384 the condition is true if the most recent recursion is into the subpat-
4385 tern whose number or name is given. This condition does not check the
4386 entire recursion stack.
4387
4388 At "top level", all these recursion test conditions are false. Recur-
4389 sive patterns are described below.
4390
4391 Defining subpatterns for use by reference only
4392
4393 If the condition is the string (DEFINE), and there is no subpattern
4394 with the name DEFINE, the condition is always false. In this case,
4395 there may be only one alternative in the subpattern. It is always
4396 skipped if control reaches this point in the pattern; the idea of
4397 DEFINE is that it can be used to define "subroutines" that can be ref-
4398 erenced from elsewhere. (The use of "subroutines" is described below.)
4399 For example, a pattern to match an IPv4 address could be written like
4400 this (ignore whitespace and line breaks):
4401
4402 (?(DEFINE) (?<byte> 2[0-4]\d | 25[0-5] | 1\d\d | [1-9]?\d) )
4403 \b (?&byte) (\.(?&byte)){3} \b
4404
4405 The first part of the pattern is a DEFINE group inside which a another
4406 group named "byte" is defined. This matches an individual component of
4407 an IPv4 address (a number less than 256). When matching takes place,
4408 this part of the pattern is skipped because DEFINE acts like a false
4409 condition.
4410
4411 The rest of the pattern uses references to the named group to match the
4412 four dot-separated components of an IPv4 address, insisting on a word
4413 boundary at each end.
4414
4415 Assertion conditions
4416
4417 If the condition is not in any of the above formats, it must be an
4418 assertion. This may be a positive or negative lookahead or lookbehind
4419 assertion. Consider this pattern, again containing non-significant
4420 white space, and with the two alternatives on the second line:
4421
4422 (?(?=[^a-z]*[a-z])
4423 \d{2}-[a-z]{3}-\d{2} | \d{2}-\d{2}-\d{2} )
4424
4425 The condition is a positive lookahead assertion that matches an
4426 optional sequence of non-letters followed by a letter. In other words,
4427 it tests for the presence of at least one letter in the subject. If a
4428 letter is found, the subject is matched against the first alternative;
4429 otherwise it is matched against the second. This pattern matches
4430 strings in one of the two forms dd-aaa-dd or dd-dd-dd, where aaa are
4431 letters and dd are digits.
4432
4433
4434 COMMENTS
4435
4436 The sequence (?# marks the start of a comment that continues up to the
4437 next closing parenthesis. Nested parentheses are not permitted. The
4438 characters that make up a comment play no part in the pattern matching
4439 at all.
4440
4441 If the PCRE_EXTENDED option is set, an unescaped # character outside a
4442 character class introduces a comment that continues to immediately
4443 after the next newline in the pattern.
4444
4445
4446 RECURSIVE PATTERNS
4447
4448 Consider the problem of matching a string in parentheses, allowing for
4449 unlimited nested parentheses. Without the use of recursion, the best
4450 that can be done is to use a pattern that matches up to some fixed
4451 depth of nesting. It is not possible to handle an arbitrary nesting
4452 depth.
4453
4454 For some time, Perl has provided a facility that allows regular expres-
4455 sions to recurse (amongst other things). It does this by interpolating
4456 Perl code in the expression at run time, and the code can refer to the
4457 expression itself. A Perl pattern using code interpolation to solve the
4458 parentheses problem can be created like this:
4459
4460 $re = qr{\( (?: (?>[^()]+) | (?p{$re}) )* \)}x;
4461
4462 The (?p{...}) item interpolates Perl code at run time, and in this case
4463 refers recursively to the pattern in which it appears.
4464
4465 Obviously, PCRE cannot support the interpolation of Perl code. Instead,
4466 it supports special syntax for recursion of the entire pattern, and
4467 also for individual subpattern recursion. After its introduction in
4468 PCRE and Python, this kind of recursion was introduced into Perl at
4469 release 5.10.
4470
4471 A special item that consists of (? followed by a number greater than
4472 zero and a closing parenthesis is a recursive call of the subpattern of
4473 the given number, provided that it occurs inside that subpattern. (If
4474 not, it is a "subroutine" call, which is described in the next sec-
4475 tion.) The special item (?R) or (?0) is a recursive call of the entire
4476 regular expression.
4477
4478 In PCRE (like Python, but unlike Perl), a recursive subpattern call is
4479 always treated as an atomic group. That is, once it has matched some of
4480 the subject string, it is never re-entered, even if it contains untried
4481 alternatives and there is a subsequent matching failure.
4482
4483 This PCRE pattern solves the nested parentheses problem (assume the
4484 PCRE_EXTENDED option is set so that white space is ignored):
4485
4486 \( ( (?>[^()]+) | (?R) )* \)
4487
4488 First it matches an opening parenthesis. Then it matches any number of
4489 substrings which can either be a sequence of non-parentheses, or a
4490 recursive match of the pattern itself (that is, a correctly parenthe-
4491 sized substring). Finally there is a closing parenthesis.
4492
4493 If this were part of a larger pattern, you would not want to recurse
4494 the entire pattern, so instead you could use this:
4495
4496 ( \( ( (?>[^()]+) | (?1) )* \) )
4497
4498 We have put the pattern into parentheses, and caused the recursion to
4499 refer to them instead of the whole pattern.
4500
4501 In a larger pattern, keeping track of parenthesis numbers can be
4502 tricky. This is made easier by the use of relative references. (A Perl
4503 5.10 feature.) Instead of (?1) in the pattern above you can write
4504 (?-2) to refer to the second most recently opened parentheses preceding
4505 the recursion. In other words, a negative number counts capturing
4506 parentheses leftwards from the point at which it is encountered.
4507
4508 It is also possible to refer to subsequently opened parentheses, by
4509 writing references such as (?+2). However, these cannot be recursive
4510 because the reference is not inside the parentheses that are refer-
4511 enced. They are always "subroutine" calls, as described in the next
4512 section.
4513
4514 An alternative approach is to use named parentheses instead. The Perl
4515 syntax for this is (?&name); PCRE's earlier syntax (?P>name) is also
4516 supported. We could rewrite the above example as follows:
4517
4518 (?<pn> \( ( (?>[^()]+) | (?&pn) )* \) )
4519
4520 If there is more than one subpattern with the same name, the earliest
4521 one is used.
4522
4523 This particular example pattern that we have been looking at contains
4524 nested unlimited repeats, and so the use of atomic grouping for match-
4525 ing strings of non-parentheses is important when applying the pattern
4526 to strings that do not match. For example, when this pattern is applied
4527 to
4528
4529 (aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa()
4530
4531 it yields "no match" quickly. However, if atomic grouping is not used,
4532 the match runs for a very long time indeed because there are so many
4533 different ways the + and * repeats can carve up the subject, and all
4534 have to be tested before failure can be reported.
4535
4536 At the end of a match, the values set for any capturing subpatterns are
4537 those from the outermost level of the recursion at which the subpattern
4538 value is set. If you want to obtain intermediate values, a callout
4539 function can be used (see below and the pcrecallout documentation). If
4540 the pattern above is matched against
4541
4542 (ab(cd)ef)
4543
4544 the value for the capturing parentheses is "ef", which is the last
4545 value taken on at the top level. If additional parentheses are added,
4546 giving
4547
4548 \( ( ( (?>[^()]+) | (?R) )* ) \)
4549 ^ ^
4550 ^ ^
4551
4552 the string they capture is "ab(cd)ef", the contents of the top level
4553 parentheses. If there are more than 15 capturing parentheses in a pat-
4554 tern, PCRE has to obtain extra memory to store data during a recursion,
4555 which it does by using pcre_malloc, freeing it via pcre_free after-
4556 wards. If no memory can be obtained, the match fails with the
4557 PCRE_ERROR_NOMEMORY error.
4558
4559 Do not confuse the (?R) item with the condition (R), which tests for
4560 recursion. Consider this pattern, which matches text in angle brack-
4561 ets, allowing for arbitrary nesting. Only digits are allowed in nested
4562 brackets (that is, when recursing), whereas any characters are permit-
4563 ted at the outer level.
4564
4565 < (?: (?(R) \d++ | [^<>]*+) | (?R)) * >
4566
4567 In this pattern, (?(R) is the start of a conditional subpattern, with
4568 two different alternatives for the recursive and non-recursive cases.
4569 The (?R) item is the actual recursive call.
4570
4571
4572 SUBPATTERNS AS SUBROUTINES
4573
4574 If the syntax for a recursive subpattern reference (either by number or
4575 by name) is used outside the parentheses to which it refers, it oper-
4576 ates like a subroutine in a programming language. The "called" subpat-
4577 tern may be defined before or after the reference. A numbered reference
4578 can be absolute or relative, as in these examples:
4579
4580 (...(absolute)...)...(?2)...
4581 (...(relative)...)...(?-1)...
4582 (...(?+1)...(relative)...
4583
4584 An earlier example pointed out that the pattern
4585
4586 (sens|respons)e and \1ibility
4587
4588 matches "sense and sensibility" and "response and responsibility", but
4589 not "sense and responsibility". If instead the pattern
4590
4591 (sens|respons)e and (?1)ibility
4592
4593 is used, it does match "sense and responsibility" as well as the other
4594 two strings. Another example is given in the discussion of DEFINE
4595 above.
4596
4597 Like recursive subpatterns, a "subroutine" call is always treated as an
4598 atomic group. That is, once it has matched some of the subject string,
4599 it is never re-entered, even if it contains untried alternatives and
4600 there is a subsequent matching failure.
4601
4602 When a subpattern is used as a subroutine, processing options such as
4603 case-independence are fixed when the subpattern is defined. They cannot
4604 be changed for different calls. For example, consider this pattern:
4605
4606 (abc)(?i:(?-1))
4607
4608 It matches "abcabc". It does not match "abcABC" because the change of
4609 processing option does not affect the called subpattern.
4610
4611
4612 CALLOUTS
4613
4614 Perl has a feature whereby using the sequence (?{...}) causes arbitrary
4615 Perl code to be obeyed in the middle of matching a regular expression.
4616 This makes it possible, amongst other things, to extract different sub-
4617 strings that match the same pair of parentheses when there is a repeti-
4618 tion.
4619
4620 PCRE provides a similar feature, but of course it cannot obey arbitrary
4621 Perl code. The feature is called "callout". The caller of PCRE provides
4622 an external function by putting its entry point in the global variable
4623 pcre_callout. By default, this variable contains NULL, which disables
4624 all calling out.
4625
4626 Within a regular expression, (?C) indicates the points at which the
4627 external function is to be called. If you want to identify different
4628 callout points, you can put a number less than 256 after the letter C.
4629 The default value is zero. For example, this pattern has two callout
4630 points:
4631
4632 (?C1)abc(?C2)def
4633
4634 If the PCRE_AUTO_CALLOUT flag is passed to pcre_compile(), callouts are
4635 automatically installed before each item in the pattern. They are all
4636 numbered 255.
4637
4638 During matching, when PCRE reaches a callout point (and pcre_callout is
4639 set), the external function is called. It is provided with the number
4640 of the callout, the position in the pattern, and, optionally, one item
4641 of data originally supplied by the caller of pcre_exec(). The callout
4642 function may cause matching to proceed, to backtrack, or to fail alto-
4643 gether. A complete description of the interface to the callout function
4644 is given in the pcrecallout documentation.
4645
4646
4647 BACTRACKING CONTROL
4648
4649 Perl 5.10 introduced a number of "Special Backtracking Control Verbs",
4650 which are described in the Perl documentation as "experimental and sub-
4651 ject to change or removal in a future version of Perl". It goes on to
4652 say: "Their usage in production code should be noted to avoid problems
4653 during upgrades." The same remarks apply to the PCRE features described
4654 in this section.
4655
4656 Since these verbs are specifically related to backtracking, they can be
4657 used only when the pattern is to be matched using pcre_exec(), which
4658 uses a backtracking algorithm. They cause an error if encountered by
4659 pcre_dfa_exec().
4660
4661 The new verbs make use of what was previously invalid syntax: an open-
4662 ing parenthesis followed by an asterisk. In Perl, they are generally of
4663 the form (*VERB:ARG) but PCRE does not support the use of arguments, so
4664 its general form is just (*VERB). Any number of these verbs may occur
4665 in a pattern. There are two kinds:
4666
4667 Verbs that act immediately
4668
4669 The following verbs act as soon as they are encountered:
4670
4671 (*ACCEPT)
4672
4673 This verb causes the match to end successfully, skipping the remainder
4674 of the pattern. When inside a recursion, only the innermost pattern is
4675 ended immediately. PCRE differs from Perl in what happens if the
4676 (*ACCEPT) is inside capturing parentheses. In Perl, the data so far is
4677 captured: in PCRE no data is captured. For example:
4678
4679 A(A|B(*ACCEPT)|C)D
4680
4681 This matches "AB", "AAD", or "ACD", but when it matches "AB", no data
4682 is captured.
4683
4684 (*FAIL) or (*F)
4685
4686 This verb causes the match to fail, forcing backtracking to occur. It
4687 is equivalent to (?!) but easier to read. The Perl documentation notes
4688 that it is probably useful only when combined with (?{}) or (??{}).
4689 Those are, of course, Perl features that are not present in PCRE. The
4690 nearest equivalent is the callout feature, as for example in this pat-
4691 tern:
4692
4693 a+(?C)(*FAIL)
4694
4695 A match with the string "aaaa" always fails, but the callout is taken
4696 before each backtrack happens (in this example, 10 times).
4697
4698 Verbs that act after backtracking
4699
4700 The following verbs do nothing when they are encountered. Matching con-
4701 tinues with what follows, but if there is no subsequent match, a fail-
4702 ure is forced. The verbs differ in exactly what kind of failure
4703 occurs.
4704
4705 (*COMMIT)
4706
4707 This verb causes the whole match to fail outright if the rest of the
4708 pattern does not match. Even if the pattern is unanchored, no further
4709 attempts to find a match by advancing the start point take place. Once
4710 (*COMMIT) has been passed, pcre_exec() is committed to finding a match
4711 at the current starting point, or not at all. For example:
4712
4713 a+(*COMMIT)b
4714
4715 This matches "xxaab" but not "aacaab". It can be thought of as a kind
4716 of dynamic anchor, or "I've started, so I must finish."
4717
4718 (*PRUNE)
4719
4720 This verb causes the match to fail at the current position if the rest
4721 of the pattern does not match. If the pattern is unanchored, the normal
4722 "bumpalong" advance to the next starting character then happens. Back-
4723 tracking can occur as usual to the left of (*PRUNE), or when matching
4724 to the right of (*PRUNE), but if there is no match to the right, back-
4725 tracking cannot cross (*PRUNE). In simple cases, the use of (*PRUNE)
4726 is just an alternative to an atomic group or possessive quantifier, but
4727 there are some uses of (*PRUNE) that cannot be expressed in any other
4728 way.
4729
4730 (*SKIP)
4731
4732 This verb is like (*PRUNE), except that if the pattern is unanchored,
4733 the "bumpalong" advance is not to the next character, but to the posi-
4734 tion in the subject where (*SKIP) was encountered. (*SKIP) signifies
4735 that whatever text was matched leading up to it cannot be part of a
4736 successful match. Consider:
4737
4738 a+(*SKIP)b
4739
4740 If the subject is "aaaac...", after the first match attempt fails
4741 (starting at the first character in the string), the starting point
4742 skips on to start the next attempt at "c". Note that a possessive quan-
4743 tifer does not have the same effect in this example; although it would
4744 suppress backtracking during the first match attempt, the second
4745 attempt would start at the second character instead of skipping on to
4746 "c".
4747
4748 (*THEN)
4749
4750 This verb causes a skip to the next alternation if the rest of the pat-
4751 tern does not match. That is, it cancels pending backtracking, but only
4752 within the current alternation. Its name comes from the observation
4753 that it can be used for a pattern-based if-then-else block:
4754
4755 ( COND1 (*THEN) FOO | COND2 (*THEN) BAR | COND3 (*THEN) BAZ ) ...
4756
4757 If the COND1 pattern matches, FOO is tried (and possibly further items
4758 after the end of the group if FOO succeeds); on failure the matcher
4759 skips to the second alternative and tries COND2, without backtracking
4760 into COND1. If (*THEN) is used outside of any alternation, it acts
4761 exactly like (*PRUNE).
4762
4763
4764 SEE ALSO
4765
4766 pcreapi(3), pcrecallout(3), pcrematching(3), pcre(3).
4767
4768
4769 AUTHOR
4770
4771 Philip Hazel
4772 University Computing Service
4773 Cambridge CB2 3QH, England.
4774
4775
4776 REVISION
4777
4778 Last updated: 21 August 2007
4779 Copyright (c) 1997-2007 University of Cambridge.
4780 ------------------------------------------------------------------------------
4781
4782
4783 PCRESYNTAX(3) PCRESYNTAX(3)
4784
4785
4786 NAME
4787 PCRE - Perl-compatible regular expressions
4788
4789
4790 PCRE REGULAR EXPRESSION SYNTAX SUMMARY
4791
4792 The full syntax and semantics of the regular expressions that are sup-
4793 ported by PCRE are described in the pcrepattern documentation. This
4794 document contains just a quick-reference summary of the syntax.
4795
4796
4797 QUOTING
4798
4799 \x where x is non-alphanumeric is a literal x
4800 \Q...\E treat enclosed characters as literal
4801
4802
4803 CHARACTERS
4804
4805 \a alarm, that is, the BEL character (hex 07)
4806 \cx "control-x", where x is any character
4807 \e escape (hex 1B)
4808 \f formfeed (hex 0C)
4809 \n newline (hex 0A)
4810 \r carriage return (hex 0D)
4811 \t tab (hex 09)
4812 \ddd character with octal code ddd, or backreference
4813 \xhh character with hex code hh
4814 \x{hhh..} character with hex code hhh..
4815
4816
4817 CHARACTER TYPES
4818
4819 . any character except newline;
4820 in dotall mode, any character whatsoever
4821 \C one byte, even in UTF-8 mode (best avoided)
4822 \d a decimal digit
4823 \D a character that is not a decimal digit
4824 \h a horizontal whitespace character
4825 \H a character that is not a horizontal whitespace character
4826 \p{xx} a character with the xx property
4827 \P{xx} a character without the xx property
4828 \R a newline sequence
4829 \s a whitespace character
4830 \S a character that is not a whitespace character
4831 \v a vertical whitespace character
4832 \V a character that is not a vertical whitespace character
4833 \w a "word" character
4834 \W a "non-word" character
4835 \X an extended Unicode sequence
4836
4837 In PCRE, \d, \D, \s, \S, \w, and \W recognize only ASCII characters.
4838
4839
4840 GENERAL CATEGORY PROPERTY CODES FOR \p and \P
4841
4842 C Other
4843 Cc Control
4844 Cf Format
4845 Cn Unassigned
4846 Co Private use
4847 Cs Surrogate
4848
4849 L Letter
4850 Ll Lower case letter
4851 Lm Modifier letter
4852 Lo Other letter
4853 Lt Title case letter
4854 Lu Upper case letter
4855 L& Ll, Lu, or Lt
4856
4857 M Mark
4858 Mc Spacing mark
4859 Me Enclosing mark
4860 Mn Non-spacing mark
4861
4862 N Number
4863 Nd Decimal number
4864 Nl Letter number
4865 No Other number
4866
4867 P Punctuation
4868 Pc Connector punctuation
4869 Pd Dash punctuation
4870 Pe Close punctuation
4871 Pf Final punctuation
4872 Pi Initial punctuation
4873 Po Other punctuation
4874 Ps Open punctuation
4875
4876 S Symbol
4877 Sc Currency symbol
4878 Sk Modifier symbol
4879 Sm Mathematical symbol
4880 So Other symbol
4881
4882 Z Separator
4883 Zl Line separator
4884 Zp Paragraph separator
4885 Zs Space separator
4886
4887
4888 SCRIPT NAMES FOR \p AND \P
4889
4890 Arabic, Armenian, Balinese, Bengali, Bopomofo, Braille, Buginese,
4891 Buhid, Canadian_Aboriginal, Cherokee, Common, Coptic, Cuneiform,
4892 Cypriot, Cyrillic, Deseret, Devanagari, Ethiopic, Georgian, Glagolitic,
4893 Gothic, Greek, Gujarati, Gurmukhi, Han, Hangul, Hanunoo, Hebrew, Hira-
4894 gana, Inherited, Kannada, Katakana, Kharoshthi, Khmer, Lao, Latin,
4895 Limbu, Linear_B, Malayalam, Mongolian, Myanmar, New_Tai_Lue, Nko,
4896 Ogham, Old_Italic, Old_Persian, Oriya, Osmanya, Phags_Pa, Phoenician,
4897 Runic, Shavian, Sinhala, Syloti_Nagri, Syriac, Tagalog, Tagbanwa,
4898 Tai_Le, Tamil, Telugu, Thaana, Thai, Tibetan, Tifinagh, Ugaritic, Yi.
4899
4900
4901 CHARACTER CLASSES
4902
4903 [...] positive character class
4904 [^...] negative character class
4905 [x-y] range (can be used for hex characters)
4906 [[:xxx:]] positive POSIX named set
4907 [[^:xxx:]] negative POSIX named set
4908
4909 alnum alphanumeric
4910 alpha alphabetic
4911 ascii 0-127
4912 blank space or tab
4913 cntrl control character
4914 digit decimal digit
4915 graph printing, excluding space
4916 lower lower case letter
4917 print printing, including space
4918 punct printing, excluding alphanumeric
4919 space whitespace
4920 upper upper case letter
4921 word same as \w
4922 xdigit hexadecimal digit
4923
4924 In PCRE, POSIX character set names recognize only ASCII characters. You
4925 can use \Q...\E inside a character class.
4926
4927
4928 QUANTIFIERS
4929
4930 ? 0 or 1, greedy
4931 ?+ 0 or 1, possessive
4932 ?? 0 or 1, lazy
4933 * 0 or more, greedy
4934 *+ 0 or more, possessive
4935 *? 0 or more, lazy
4936 + 1 or more, greedy
4937 ++ 1 or more, possessive
4938 +? 1 or more, lazy
4939 {n} exactly n
4940 {n,m} at least n, no more than m, greedy
4941 {n,m}+ at least n, no more than m, possessive
4942 {n,m}? at least n, no more than m, lazy
4943 {n,} n or more, greedy
4944 {n,}+ n or more, possessive
4945 {n,}? n or more, lazy
4946
4947
4948 ANCHORS AND SIMPLE ASSERTIONS
4949
4950 \b word boundary
4951 \B not a word boundary
4952 ^ start of subject
4953 also after internal newline in multiline mode
4954 \A start of subject
4955 $ end of subject
4956 also before newline at end of subject
4957 also before internal newline in multiline mode
4958 \Z end of subject
4959 also before newline at end of subject
4960 \z end of subject
4961 \G first matching position in subject
4962
4963
4964 MATCH POINT RESET
4965
4966 \K reset start of match
4967
4968
4969 ALTERNATION
4970
4971 expr|expr|expr...
4972
4973
4974 CAPTURING
4975
4976 (...) capturing group
4977 (?<name>...) named capturing group (Perl)
4978 (?'name'...) named capturing group (Perl)
4979 (?P<name>...) named capturing group (Python)
4980 (?:...) non-capturing group
4981 (?|...) non-capturing group; reset group numbers for
4982 capturing groups in each alternative
4983
4984
4985 ATOMIC GROUPS
4986
4987 (?>...) atomic, non-capturing group
4988
4989
4990 COMMENT
4991
4992 (?#....) comment (not nestable)
4993
4994
4995 OPTION SETTING
4996
4997 (?i) caseless
4998 (?J) allow duplicate names
4999 (?m) multiline
5000 (?s) single line (dotall)
5001 (?U) default ungreedy (lazy)
5002 (?x) extended (ignore white space)
5003 (?-...) unset option(s)
5004
5005
5006 LOOKAHEAD AND LOOKBEHIND ASSERTIONS
5007
5008 (?=...) positive look ahead
5009 (?!...) negative look ahead
5010 (?<=...) positive look behind
5011 (?<!...) negative look behind
5012
5013 Each top-level branch of a look behind must be of a fixed length.
5014
5015
5016 BACKREFERENCES
5017
5018 \n reference by number (can be ambiguous)
5019 \gn reference by number
5020 \g{n} reference by number
5021 \g{-n} relative reference by number
5022 \k<name> reference by name (Perl)
5023 \k'name' reference by name (Perl)
5024 \g{name} reference by name (Perl)
5025 \k{name} reference by name (.NET)
5026 (?P=name) reference by name (Python)
5027
5028
5029 SUBROUTINE REFERENCES (POSSIBLY RECURSIVE)
5030
5031 (?R) recurse whole pattern
5032 (?n) call subpattern by absolute number
5033 (?+n) call subpattern by relative number
5034 (?-n) call subpattern by relative number
5035 (?&name) call subpattern by name (Perl)
5036 (?P>name) call subpattern by name (Python)
5037
5038
5039 CONDITIONAL PATTERNS
5040
5041 (?(condition)yes-pattern)
5042 (?(condition)yes-pattern|no-pattern)
5043
5044 (?(n)... absolute reference condition
5045 (?(+n)... relative reference condition
5046 (?(-n)... relative reference condition
5047 (?(<name>)... named reference condition (Perl)
5048 (?('name')... named reference condition (Perl)
5049 (?(name)... named reference condition (PCRE)
5050 (?(R)... overall recursion condition
5051 (?(Rn)... specific group recursion condition
5052 (?(R&name)... specific recursion condition
5053 (?(DEFINE)... define subpattern for reference
5054 (?(assert)... assertion condition
5055
5056
5057 BACKTRACKING CONTROL
5058
5059 The following act immediately they are reached:
5060
5061 (*ACCEPT) force successful match
5062 (*FAIL) force backtrack; synonym (*F)
5063
5064 The following act only when a subsequent match failure causes a back-
5065 track to reach them. They all force a match failure, but they differ in
5066 what happens afterwards. Those that advance the start-of-match point do
5067 so only if the pattern is not anchored.
5068
5069 (*COMMIT) overall failure, no advance of starting point
5070 (*PRUNE) advance to next starting character
5071 (*SKIP) advance start to current matching position
5072 (*THEN) local failure, backtrack to next alternation
5073
5074
5075 NEWLINE CONVENTIONS
5076
5077 These are recognized only at the very start of a pattern.
5078
5079 (*CR)
5080 (*LF)
5081 (*CRLF)
5082 (*ANYCRLF)
5083 (*ANY)
5084
5085
5086 CALLOUTS
5087
5088 (?C) callout
5089 (?Cn) callout with data n
5090
5091
5092 SEE ALSO
5093
5094 pcrepattern(3), pcreapi(3), pcrecallout(3), pcrematching(3), pcre(3).
5095
5096
5097 AUTHOR
5098
5099 Philip Hazel
5100 University Computing Service
5101 Cambridge CB2 3QH, England.
5102
5103
5104 REVISION
5105
5106 Last updated: 21 August 2007
5107 Copyright (c) 1997-2007 University of Cambridge.
5108 ------------------------------------------------------------------------------
5109
5110
5111 PCREPARTIAL(3) PCREPARTIAL(3)
5112
5113
5114 NAME
5115 PCRE - Perl-compatible regular expressions
5116
5117
5118 PARTIAL MATCHING IN PCRE
5119
5120 In normal use of PCRE, if the subject string that is passed to
5121 pcre_exec() or pcre_dfa_exec() matches as far as it goes, but is too
5122 short to match the entire pattern, PCRE_ERROR_NOMATCH is returned.
5123 There are circumstances where it might be helpful to distinguish this
5124 case from other cases in which there is no match.
5125
5126 Consider, for example, an application where a human is required to type
5127 in data for a field with specific formatting requirements. An example
5128 might be a date in the form ddmmmyy, defined by this pattern:
5129
5130 ^\d?\d(jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)\d\d$
5131
5132 If the application sees the user's keystrokes one by one, and can check
5133 that what has been typed so far is potentially valid, it is able to
5134 raise an error as soon as a mistake is made, possibly beeping and not
5135 reflecting the character that has been typed. This immediate feedback
5136 is likely to be a better user interface than a check that is delayed
5137 until the entire string has been entered.
5138
5139 PCRE supports the concept of partial matching by means of the PCRE_PAR-
5140 TIAL option, which can be set when calling pcre_exec() or
5141 pcre_dfa_exec(). When this flag is set for pcre_exec(), the return code
5142 PCRE_ERROR_NOMATCH is converted into PCRE_ERROR_PARTIAL if at any time
5143 during the matching process the last part of the subject string matched
5144 part of the pattern. Unfortunately, for non-anchored matching, it is
5145 not possible to obtain the position of the start of the partial match.
5146 No captured data is set when PCRE_ERROR_PARTIAL is returned.
5147
5148 When PCRE_PARTIAL is set for pcre_dfa_exec(), the return code
5149 PCRE_ERROR_NOMATCH is converted into PCRE_ERROR_PARTIAL if the end of
5150 the subject is reached, there have been no complete matches, but there
5151 is still at least one matching possibility. The portion of the string
5152 that provided the partial match is set as the first matching string.
5153
5154 Using PCRE_PARTIAL disables one of PCRE's optimizations. PCRE remembers
5155 the last literal byte in a pattern, and abandons matching immediately
5156 if such a byte is not present in the subject string. This optimization
5157 cannot be used for a subject string that might match only partially.
5158
5159
5160 RESTRICTED PATTERNS FOR PCRE_PARTIAL
5161
5162 Because of the way certain internal optimizations are implemented in
5163 the pcre_exec() function, the PCRE_PARTIAL option cannot be used with
5164 all patterns. These restrictions do not apply when pcre_dfa_exec() is
5165 used. For pcre_exec(), repeated single characters such as
5166
5167 a{2,4}
5168
5169 and repeated single metasequences such as
5170
5171 \d+
5172
5173 are not permitted if the maximum number of occurrences is greater than
5174 one. Optional items such as \d? (where the maximum is one) are permit-
5175 ted. Quantifiers with any values are permitted after parentheses, so
5176 the invalid examples above can be coded thus:
5177
5178 (a){2,4}
5179 (\d)+
5180
5181 These constructions run more slowly, but for the kinds of application
5182 that are envisaged for this facility, this is not felt to be a major
5183 restriction.
5184
5185 If PCRE_PARTIAL is set for a pattern that does not conform to the
5186 restrictions, pcre_exec() returns the error code PCRE_ERROR_BADPARTIAL
5187 (-13). You can use the PCRE_INFO_OKPARTIAL call to pcre_fullinfo() to
5188 find out if a compiled pattern can be used for partial matching.
5189
5190
5191 EXAMPLE OF PARTIAL MATCHING USING PCRETEST
5192
5193 If the escape sequence \P is present in a pcretest data line, the
5194 PCRE_PARTIAL flag is used for the match. Here is a run of pcretest that
5195 uses the date example quoted above:
5196
5197 re> /^\d?\d(jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)\d\d$/
5198 data> 25jun04\P
5199 0: 25jun04
5200 1: jun
5201 data> 25dec3\P
5202 Partial match
5203 data> 3ju\P
5204 Partial match
5205 data> 3juj\P
5206 No match
5207 data> j\P
5208 No match
5209
5210 The first data string is matched completely, so pcretest shows the
5211 matched substrings. The remaining four strings do not match the com-
5212 plete pattern, but the first two are partial matches. The same test,
5213 using pcre_dfa_exec() matching (by means of the \D escape sequence),
5214 produces the following output:
5215
5216 re> /^\d?\d(jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)\d\d$/
5217 data> 25jun04\P\D
5218 0: 25jun04
5219 data> 23dec3\P\D
5220 Partial match: 23dec3
5221 data> 3ju\P\D
5222 Partial match: 3ju
5223 data> 3juj\P\D
5224 No match
5225 data> j\P\D
5226 No match
5227
5228 Notice that in this case the portion of the string that was matched is
5229 made available.
5230
5231
5232 MULTI-SEGMENT MATCHING WITH pcre_dfa_exec()
5233
5234 When a partial match has been found using pcre_dfa_exec(), it is possi-
5235 ble to continue the match by providing additional subject data and
5236 calling pcre_dfa_exec() again with the same compiled regular expres-
5237 sion, this time setting the PCRE_DFA_RESTART option. You must also pass
5238 the same working space as before, because this is where details of the
5239 previous partial match are stored. Here is an example using pcretest,
5240 using the \R escape sequence to set the PCRE_DFA_RESTART option (\P and
5241 \D are as above):
5242
5243 re> /^\d?\d(jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)\d\d$/
5244 data> 23ja\P\D
5245 Partial match: 23ja
5246 data> n05\R\D
5247 0: n05
5248
5249 The first call has "23ja" as the subject, and requests partial match-
5250 ing; the second call has "n05" as the subject for the continued
5251 (restarted) match. Notice that when the match is complete, only the
5252 last part is shown; PCRE does not retain the previously partially-
5253 matched string. It is up to the calling program to do that if it needs
5254 to.
5255
5256 You can set PCRE_PARTIAL with PCRE_DFA_RESTART to continue partial
5257 matching over multiple segments. This facility can be used to pass very
5258 long subject strings to pcre_dfa_exec(). However, some care is needed
5259 for certain types of pattern.
5260
5261 1. If the pattern contains tests for the beginning or end of a line,
5262 you need to pass the PCRE_NOTBOL or PCRE_NOTEOL options, as appropri-
5263 ate, when the subject string for any call does not contain the begin-
5264 ning or end of a line.
5265
5266 2. If the pattern contains backward assertions (including \b or \B),
5267 you need to arrange for some overlap in the subject strings to allow
5268 for this. For example, you could pass the subject in chunks that are
5269 500 bytes long, but in a buffer of 700 bytes, with the starting offset
5270 set to 200 and the previous 200 bytes at the start of the buffer.
5271
5272 3. Matching a subject string that is split into multiple segments does
5273 not always produce exactly the same result as matching over one single
5274 long string. The difference arises when there are multiple matching
5275 possibilities, because a partial match result is given only when there
5276 are no completed matches in a call to pcre_dfa_exec(). This means that
5277 as soon as the shortest match has been found, continuation to a new
5278 subject segment is no longer possible. Consider this pcretest example:
5279
5280 re> /dog(sbody)?/
5281 data> do\P\D
5282 Partial match: do
5283 data> gsb\R\P\D
5284 0: g
5285 data> dogsbody\D
5286 0: dogsbody
5287 1: dog
5288
5289 The pattern matches the words "dog" or "dogsbody". When the subject is
5290 presented in several parts ("do" and "gsb" being the first two) the
5291 match stops when "dog" has been found, and it is not possible to con-
5292 tinue. On the other hand, if "dogsbody" is presented as a single
5293 string, both matches are found.
5294
5295 Because of this phenomenon, it does not usually make sense to end a
5296 pattern that is going to be matched in this way with a variable repeat.
5297
5298 4. Patterns that contain alternatives at the top level which do not all
5299 start with the same pattern item may not work as expected. For example,
5300 consider this pattern:
5301
5302 1234|3789
5303
5304 If the first part of the subject is "ABC123", a partial match of the
5305 first alternative is found at offset 3. There is no partial match for
5306 the second alternative, because such a match does not start at the same
5307 point in the subject string. Attempting to continue with the string
5308 "789" does not yield a match because only those alternatives that match
5309 at one point in the subject are remembered. The problem arises because
5310 the start of the second alternative matches within the first alterna-
5311 tive. There is no problem with anchored patterns or patterns such as:
5312
5313 1234|ABCD
5314
5315 where no string can be a partial match for both alternatives.
5316
5317
5318 AUTHOR
5319
5320 Philip Hazel
5321 University Computing Service
5322 Cambridge CB2 3QH, England.
5323
5324
5325 REVISION
5326
5327 Last updated: 04 June 2007
5328 Copyright (c) 1997-2007 University of Cambridge.
5329 ------------------------------------------------------------------------------
5330
5331
5332 PCREPRECOMPILE(3) PCREPRECOMPILE(3)
5333
5334
5335 NAME
5336 PCRE - Perl-compatible regular expressions
5337
5338
5339 SAVING AND RE-USING PRECOMPILED PCRE PATTERNS
5340
5341 If you are running an application that uses a large number of regular
5342 expression patterns, it may be useful to store them in a precompiled
5343 form instead of having to compile them every time the application is
5344 run. If you are not using any private character tables (see the
5345 pcre_maketables() documentation), this is relatively straightforward.
5346 If you are using private tables, it is a little bit more complicated.
5347
5348 If you save compiled patterns to a file, you can copy them to a differ-
5349 ent host and run them there. This works even if the new host has the
5350 opposite endianness to the one on which the patterns were compiled.
5351 There may be a small performance penalty, but it should be insignifi-
5352 cant. However, compiling regular expressions with one version of PCRE
5353 for use with a different version is not guaranteed to work and may
5354 cause crashes.
5355
5356
5357 SAVING A COMPILED PATTERN
5358 The value returned by pcre_compile() points to a single block of memory
5359 that holds the compiled pattern and associated data. You can find the
5360 length of this block in bytes by calling pcre_fullinfo() with an argu-
5361 ment of PCRE_INFO_SIZE. You can then save the data in any appropriate
5362 manner. Here is sample code that compiles a pattern and writes it to a
5363 file. It assumes that the variable fd refers to a file that is open for
5364 output:
5365
5366 int erroroffset, rc, size;
5367 char *error;
5368 pcre *re;
5369
5370 re = pcre_compile("my pattern", 0, &error, &erroroffset, NULL);
5371 if (re == NULL) { ... handle errors ... }
5372 rc = pcre_fullinfo(re, NULL, PCRE_INFO_SIZE, &size);
5373 if (rc < 0) { ... handle errors ... }
5374 rc = fwrite(re, 1, size, fd);
5375 if (rc != size) { ... handle errors ... }
5376
5377 In this example, the bytes that comprise the compiled pattern are
5378 copied exactly. Note that this is binary data that may contain any of
5379 the 256 possible byte values. On systems that make a distinction
5380 between binary and non-binary data, be sure that the file is opened for
5381 binary output.
5382
5383 If you want to write more than one pattern to a file, you will have to
5384 devise a way of separating them. For binary data, preceding each pat-
5385 tern with its length is probably the most straightforward approach.
5386 Another possibility is to write out the data in hexadecimal instead of
5387 binary, one pattern to a line.
5388
5389 Saving compiled patterns in a file is only one possible way of storing
5390 them for later use. They could equally well be saved in a database, or
5391 in the memory of some daemon process that passes them via sockets to
5392 the processes that want them.
5393
5394 If the pattern has been studied, it is also possible to save the study
5395 data in a similar way to the compiled pattern itself. When studying
5396 generates additional information, pcre_study() returns a pointer to a
5397 pcre_extra data block. Its format is defined in the section on matching
5398 a pattern in the pcreapi documentation. The study_data field points to
5399 the binary study data, and this is what you must save (not the
5400 pcre_extra block itself). The length of the study data can be obtained
5401 by calling pcre_fullinfo() with an argument of PCRE_INFO_STUDYSIZE.
5402 Remember to check that pcre_study() did return a non-NULL value before
5403 trying to save the study data.
5404
5405
5406 RE-USING A PRECOMPILED PATTERN
5407
5408 Re-using a precompiled pattern is straightforward. Having reloaded it
5409 into main memory, you pass its pointer to pcre_exec() or
5410 pcre_dfa_exec() in the usual way. This should work even on another
5411 host, and even if that host has the opposite endianness to the one
5412 where the pattern was compiled.
5413
5414 However, if you passed a pointer to custom character tables when the
5415 pattern was compiled (the tableptr argument of pcre_compile()), you
5416 must now pass a similar pointer to pcre_exec() or pcre_dfa_exec(),
5417 because the value saved with the compiled pattern will obviously be
5418 nonsense. A field in a pcre_extra() block is used to pass this data, as
5419 described in the section on matching a pattern in the pcreapi documen-
5420 tation.
5421
5422 If you did not provide custom character tables when the pattern was
5423 compiled, the pointer in the compiled pattern is NULL, which causes
5424 pcre_exec() to use PCRE's internal tables. Thus, you do not need to
5425 take any special action at run time in this case.
5426
5427 If you saved study data with the compiled pattern, you need to create
5428 your own pcre_extra data block and set the study_data field to point to
5429 the reloaded study data. You must also set the PCRE_EXTRA_STUDY_DATA
5430 bit in the flags field to indicate that study data is present. Then
5431 pass the pcre_extra block to pcre_exec() or pcre_dfa_exec() in the
5432 usual way.
5433
5434
5435 COMPATIBILITY WITH DIFFERENT PCRE RELEASES
5436
5437 In general, it is safest to recompile all saved patterns when you
5438 update to a new PCRE release, though not all updates actually require
5439 this. Recompiling is definitely needed for release 7.2.
5440
5441
5442 AUTHOR
5443
5444 Philip Hazel
5445 University Computing Service
5446 Cambridge CB2 3QH, England.
5447
5448
5449 REVISION
5450
5451 Last updated: 13 June 2007
5452 Copyright (c) 1997-2007 University of Cambridge.
5453 ------------------------------------------------------------------------------
5454
5455
5456 PCREPERFORM(3) PCREPERFORM(3)
5457
5458
5459 NAME
5460 PCRE - Perl-compatible regular expressions
5461
5462
5463 PCRE PERFORMANCE
5464
5465 Two aspects of performance are discussed below: memory usage and pro-
5466 cessing time. The way you express your pattern as a regular expression
5467 can affect both of them.
5468
5469
5470 MEMORY USAGE
5471
5472 Patterns are compiled by PCRE into a reasonably efficient byte code, so
5473 that most simple patterns do not use much memory. However, there is one
5474 case where memory usage can be unexpectedly large. When a parenthesized
5475 subpattern has a quantifier with a minimum greater than 1 and/or a lim-
5476 ited maximum, the whole subpattern is repeated in the compiled code.
5477 For example, the pattern
5478
5479 (abc|def){2,4}
5480
5481 is compiled as if it were
5482
5483 (abc|def)(abc|def)((abc|def)(abc|def)?)?
5484
5485 (Technical aside: It is done this way so that backtrack points within
5486 each of the repetitions can be independently maintained.)
5487
5488 For regular expressions whose quantifiers use only small numbers, this
5489 is not usually a problem. However, if the numbers are large, and par-
5490 ticularly if such repetitions are nested, the memory usage can become
5491 an embarrassment. For example, the very simple pattern
5492
5493 ((ab){1,1000}c){1,3}
5494
5495 uses 51K bytes when compiled. When PCRE is compiled with its default
5496 internal pointer size of two bytes, the size limit on a compiled pat-
5497 tern is 64K, and this is reached with the above pattern if the outer
5498 repetition is increased from 3 to 4. PCRE can be compiled to use larger
5499 internal pointers and thus handle larger compiled patterns, but it is
5500 better to try to rewrite your pattern to use less memory if you can.
5501
5502 One way of reducing the memory usage for such patterns is to make use
5503 of PCRE's "subroutine" facility. Re-writing the above pattern as
5504
5505 ((ab)(?2){0,999}c)(?1){0,2}
5506
5507 reduces the memory requirements to 18K, and indeed it remains under 20K
5508 even with the outer repetition increased to 100. However, this pattern
5509 is not exactly equivalent, because the "subroutine" calls are treated
5510 as atomic groups into which there can be no backtracking if there is a
5511 subsequent matching failure. Therefore, PCRE cannot do this kind of
5512 rewriting automatically. Furthermore, there is a noticeable loss of
5513 speed when executing the modified pattern. Nevertheless, if the atomic
5514 grouping is not a problem and the loss of speed is acceptable, this
5515 kind of rewriting will allow you to process patterns that PCRE cannot
5516 otherwise handle.
5517
5518
5519 PROCESSING TIME
5520
5521 Certain items in regular expression patterns are processed more effi-
5522 ciently than others. It is more efficient to use a character class like
5523 [aeiou] than a set of single-character alternatives such as
5524 (a|e|i|o|u). In general, the simplest construction that provides the
5525 required behaviour is usually the most efficient. Jeffrey Friedl's book
5526 contains a lot of useful general discussion about optimizing regular
5527 expressions for efficient performance. This document contains a few
5528 observations about PCRE.
5529
5530 Using Unicode character properties (the \p, \P, and \X escapes) is
5531 slow, because PCRE has to scan a structure that contains data for over
5532 fifteen thousand characters whenever it needs a character's property.
5533 If you can find an alternative pattern that does not use character
5534 properties, it will probably be faster.
5535
5536 When a pattern begins with .* not in parentheses, or in parentheses
5537 that are not the subject of a backreference, and the PCRE_DOTALL option
5538 is set, the pattern is implicitly anchored by PCRE, since it can match
5539 only at the start of a subject string. However, if PCRE_DOTALL is not
5540 set, PCRE cannot make this optimization, because the . metacharacter
5541 does not then match a newline, and if the subject string contains new-
5542 lines, the pattern may match from the character immediately following
5543 one of them instead of from the very start. For example, the pattern
5544
5545 .*second
5546
5547 matches the subject "first\nand second" (where \n stands for a newline
5548 character), with the match starting at the seventh character. In order
5549 to do this, PCRE has to retry the match starting after every newline in
5550 the subject.
5551
5552 If you are using such a pattern with subject strings that do not con-
5553 tain newlines, the best performance is obtained by setting PCRE_DOTALL,
5554 or starting the pattern with ^.* or ^.*? to indicate explicit anchor-
5555 ing. That saves PCRE from having to scan along the subject looking for
5556 a newline to restart at.
5557
5558 Beware of patterns that contain nested indefinite repeats. These can
5559 take a long time to run when applied to a string that does not match.
5560 Consider the pattern fragment
5561
5562 ^(a+)*
5563
5564 This can match "aaaa" in 16 different ways, and this number increases
5565 very rapidly as the string gets longer. (The * repeat can match 0, 1,
5566 2, 3, or 4 times, and for each of those cases other than 0 or 4, the +
5567 repeats can match different numbers of times.) When the remainder of
5568 the pattern is such that the entire match is going to fail, PCRE has in
5569 principle to try every possible variation, and this can take an
5570 extremely long time, even for relatively short strings.
5571
5572 An optimization catches some of the more simple cases such as
5573
5574 (a+)*b
5575
5576 where a literal character follows. Before embarking on the standard
5577 matching procedure, PCRE checks that there is a "b" later in the sub-
5578 ject string, and if there is not, it fails the match immediately. How-
5579 ever, when there is no following literal this optimization cannot be
5580 used. You can see the difference by comparing the behaviour of
5581
5582 (a+)*\d
5583
5584 with the pattern above. The former gives a failure almost instantly
5585 when applied to a whole line of "a" characters, whereas the latter
5586 takes an appreciable time with strings longer than about 20 characters.
5587
5588 In many cases, the solution to this kind of performance issue is to use
5589 an atomic group or a possessive quantifier.
5590
5591
5592 AUTHOR
5593
5594 Philip Hazel
5595 University Computing Service
5596 Cambridge CB2 3QH, England.
5597
5598
5599 REVISION
5600
5601 Last updated: 06 March 2007
5602 Copyright (c) 1997-2007 University of Cambridge.
5603 ------------------------------------------------------------------------------
5604
5605
5606 PCREPOSIX(3) PCREPOSIX(3)
5607
5608
5609 NAME
5610 PCRE - Perl-compatible regular expressions.
5611
5612
5613 SYNOPSIS OF POSIX API
5614
5615 #include <pcreposix.h>
5616
5617 int regcomp(regex_t *preg, const char *pattern,
5618 int cflags);
5619
5620 int regexec(regex_t *preg, const char *string,
5621 size_t nmatch, regmatch_t pmatch[], int eflags);
5622
5623 size_t regerror(int errcode, const regex_t *preg,
5624 char *errbuf, size_t errbuf_size);
5625
5626 void regfree(regex_t *preg);
5627
5628
5629 DESCRIPTION
5630
5631 This set of functions provides a POSIX-style API to the PCRE regular
5632 expression package. See the pcreapi documentation for a description of
5633 PCRE's native API, which contains much additional functionality.
5634
5635 The functions described here are just wrapper functions that ultimately
5636 call the PCRE native API. Their prototypes are defined in the
5637 pcreposix.h header file, and on Unix systems the library itself is
5638 called pcreposix.a, so can be accessed by adding -lpcreposix to the
5639 command for linking an application that uses them. Because the POSIX
5640 functions call the native ones, it is also necessary to add -lpcre.
5641
5642 I have implemented only those option bits that can be reasonably mapped
5643 to PCRE native options. In addition, the option REG_EXTENDED is defined
5644 with the value zero. This has no effect, but since programs that are
5645 written to the POSIX interface often use it, this makes it easier to
5646 slot in PCRE as a replacement library. Other POSIX options are not even
5647 defined.
5648
5649 When PCRE is called via these functions, it is only the API that is
5650 POSIX-like in style. The syntax and semantics of the regular expres-
5651 sions themselves are still those of Perl, subject to the setting of
5652 various PCRE options, as described below. "POSIX-like in style" means
5653 that the API approximates to the POSIX definition; it is not fully
5654 POSIX-compatible, and in multi-byte encoding domains it is probably
5655 even less compatible.
5656
5657 The header for these functions is supplied as pcreposix.h to avoid any
5658 potential clash with other POSIX libraries. It can, of course, be
5659 renamed or aliased as regex.h, which is the "correct" name. It provides
5660 two structure types, regex_t for compiled internal forms, and reg-
5661 match_t for returning captured substrings. It also defines some con-
5662 stants whose names start with "REG_"; these are used for setting
5663 options and identifying error codes.
5664
5665
5666 COMPILING A PATTERN
5667
5668 The function regcomp() is called to compile a pattern into an internal
5669 form. The pattern is a C string terminated by a binary zero, and is
5670 passed in the argument pattern. The preg argument is a pointer to a
5671 regex_t structure that is used as a base for storing information about
5672 the compiled regular expression.
5673
5674 The argument cflags is either zero, or contains one or more of the bits
5675 defined by the following macros:
5676
5677 REG_DOTALL
5678
5679 The PCRE_DOTALL option is set when the regular expression is passed for
5680 compilation to the native function. Note that REG_DOTALL is not part of
5681 the POSIX standard.
5682
5683 REG_ICASE
5684
5685 The PCRE_CASELESS option is set when the regular expression is passed
5686 for compilation to the native function.
5687
5688 REG_NEWLINE
5689
5690 The PCRE_MULTILINE option is set when the regular expression is passed
5691 for compilation to the native function. Note that this does not mimic
5692 the defined POSIX behaviour for REG_NEWLINE (see the following sec-
5693 tion).
5694
5695 REG_NOSUB
5696
5697 The PCRE_NO_AUTO_CAPTURE option is set when the regular expression is
5698 passed for compilation to the native function. In addition, when a pat-
5699 tern that is compiled with this flag is passed to regexec() for match-
5700 ing, the nmatch and pmatch arguments are ignored, and no captured
5701 strings are returned.
5702
5703 REG_UTF8
5704
5705 The PCRE_UTF8 option is set when the regular expression is passed for
5706 compilation to the native function. This causes the pattern itself and
5707 all data strings used for matching it to be treated as UTF-8 strings.
5708 Note that REG_UTF8 is not part of the POSIX standard.
5709
5710 In the absence of these flags, no options are passed to the native
5711 function. This means the the regex is compiled with PCRE default
5712 semantics. In particular, the way it handles newline characters in the
5713 subject string is the Perl way, not the POSIX way. Note that setting
5714 PCRE_MULTILINE has only some of the effects specified for REG_NEWLINE.
5715 It does not affect the way newlines are matched by . (they aren't) or
5716 by a negative class such as [^a] (they are).
5717
5718 The yield of regcomp() is zero on success, and non-zero otherwise. The
5719 preg structure is filled in on success, and one member of the structure
5720 is public: re_nsub contains the number of capturing subpatterns in the
5721 regular expression. Various error codes are defined in the header file.
5722
5723
5724 MATCHING NEWLINE CHARACTERS
5725
5726 This area is not simple, because POSIX and Perl take different views of
5727 things. It is not possible to get PCRE to obey POSIX semantics, but
5728 then PCRE was never intended to be a POSIX engine. The following table
5729 lists the different possibilities for matching newline characters in
5730 PCRE:
5731
5732 Default Change with
5733
5734 . matches newline no PCRE_DOTALL
5735 newline matches [^a] yes not changeable
5736 $ matches \n at end yes PCRE_DOLLARENDONLY
5737 $ matches \n in middle no PCRE_MULTILINE
5738 ^ matches \n in middle no PCRE_MULTILINE
5739
5740 This is the equivalent table for POSIX:
5741
5742 Default Change with
5743
5744 . matches newline yes REG_NEWLINE
5745 newline matches [^a] yes REG_NEWLINE
5746 $ matches \n at end no REG_NEWLINE
5747 $ matches \n in middle no REG_NEWLINE
5748 ^ matches \n in middle no REG_NEWLINE
5749
5750 PCRE's behaviour is the same as Perl's, except that there is no equiva-
5751 lent for PCRE_DOLLAR_ENDONLY in Perl. In both PCRE and Perl, there is
5752 no way to stop newline from matching [^a].
5753
5754 The default POSIX newline handling can be obtained by setting
5755 PCRE_DOTALL and PCRE_DOLLAR_ENDONLY, but there is no way to make PCRE
5756 behave exactly as for the REG_NEWLINE action.
5757
5758
5759 MATCHING A PATTERN
5760
5761 The function regexec() is called to match a compiled pattern preg
5762 against a given string, which is terminated by a zero byte, subject to
5763 the options in eflags. These can be:
5764
5765 REG_NOTBOL
5766
5767 The PCRE_NOTBOL option is set when calling the underlying PCRE matching
5768 function.
5769
5770 REG_NOTEOL
5771
5772 The PCRE_NOTEOL option is set when calling the underlying PCRE matching
5773 function.
5774
5775 If the pattern was compiled with the REG_NOSUB flag, no data about any
5776 matched strings is returned. The nmatch and pmatch arguments of
5777 regexec() are ignored.
5778
5779 Otherwise,the portion of the string that was matched, and also any cap-
5780 tured substrings, are returned via the pmatch argument, which points to
5781 an array of nmatch structures of type regmatch_t, containing the mem-
5782 bers rm_so and rm_eo. These contain the offset to the first character
5783 of each substring and the offset to the first character after the end
5784 of each substring, respectively. The 0th element of the vector relates
5785 to the entire portion of string that was matched; subsequent elements
5786 relate to the capturing subpatterns of the regular expression. Unused
5787 entries in the array have both structure members set to -1.
5788
5789 A successful match yields a zero return; various error codes are
5790 defined in the header file, of which REG_NOMATCH is the "expected"
5791 failure code.
5792
5793
5794 ERROR MESSAGES
5795
5796 The regerror() function maps a non-zero errorcode from either regcomp()
5797 or regexec() to a printable message. If preg is not NULL, the error
5798 should have arisen from the use of that structure. A message terminated
5799 by a binary zero is placed in errbuf. The length of the message,
5800 including the zero, is limited to errbuf_size. The yield of the func-
5801 tion is the size of buffer needed to hold the whole message.
5802
5803
5804 MEMORY USAGE
5805
5806 Compiling a regular expression causes memory to be allocated and asso-
5807 ciated with the preg structure. The function regfree() frees all such
5808 memory, after which preg may no longer be used as a compiled expres-
5809 sion.
5810
5811
5812 AUTHOR
5813
5814 Philip Hazel
5815 University Computing Service
5816 Cambridge CB2 3QH, England.
5817
5818
5819 REVISION
5820
5821 Last updated: 06 March 2007
5822 Copyright (c) 1997-2007 University of Cambridge.
5823 ------------------------------------------------------------------------------
5824
5825
5826 PCRECPP(3) PCRECPP(3)
5827
5828
5829 NAME
5830 PCRE - Perl-compatible regular expressions.
5831
5832
5833 SYNOPSIS OF C++ WRAPPER
5834
5835 #include <pcrecpp.h>
5836
5837
5838 DESCRIPTION
5839
5840 The C++ wrapper for PCRE was provided by Google Inc. Some additional
5841 functionality was added by Giuseppe Maxia. This brief man page was con-
5842 structed from the notes in the pcrecpp.h file, which should be con-
5843 sulted for further details.
5844
5845
5846 MATCHING INTERFACE
5847
5848 The "FullMatch" operation checks that supplied text matches a supplied
5849 pattern exactly. If pointer arguments are supplied, it copies matched
5850 sub-strings that match sub-patterns into them.
5851
5852 Example: successful match
5853 pcrecpp::RE re("h.*o");
5854 re.FullMatch("hello");
5855
5856 Example: unsuccessful match (requires full match):
5857 pcrecpp::RE re("e");
5858 !re.FullMatch("hello");
5859
5860 Example: creating a temporary RE object:
5861 pcrecpp::RE("h.*o").FullMatch("hello");
5862
5863 You can pass in a "const char*" or a "string" for "text". The examples
5864 below tend to use a const char*. You can, as in the different examples
5865 above, store the RE object explicitly in a variable or use a temporary
5866 RE object. The examples below use one mode or the other arbitrarily.
5867 Either could correctly be used for any of these examples.
5868
5869 You must supply extra pointer arguments to extract matched subpieces.
5870
5871 Example: extracts "ruby" into "s" and 1234 into "i"
5872 int i;
5873 string s;
5874 pcrecpp::RE re("(\\w+):(\\d+)");
5875 re.FullMatch("ruby:1234", &s, &i);
5876
5877 Example: does not try to extract any extra sub-patterns
5878 re.FullMatch("ruby:1234", &s);
5879
5880 Example: does not try to extract into NULL
5881 re.FullMatch("ruby:1234", NULL, &i);
5882
5883 Example: integer overflow causes failure
5884 !re.FullMatch("ruby:1234567891234", NULL, &i);
5885
5886 Example: fails because there aren't enough sub-patterns:
5887 !pcrecpp::RE("\\w+:\\d+").FullMatch("ruby:1234", &s);
5888
5889 Example: fails because string cannot be stored in integer
5890 !pcrecpp::RE("(.*)").FullMatch("ruby", &i);
5891
5892 The provided pointer arguments can be pointers to any scalar numeric
5893 type, or one of:
5894
5895 string (matched piece is copied to string)
5896 StringPiece (StringPiece is mutated to point to matched piece)
5897 T (where "bool T::ParseFrom(const char*, int)" exists)
5898 NULL (the corresponding matched sub-pattern is not copied)
5899
5900 The function returns true iff all of the following conditions are sat-
5901 isfied:
5902
5903 a. "text" matches "pattern" exactly;
5904
5905 b. The number of matched sub-patterns is >= number of supplied
5906 pointers;
5907
5908 c. The "i"th argument has a suitable type for holding the
5909 string captured as the "i"th sub-pattern. If you pass in
5910 NULL for the "i"th argument, or pass fewer arguments than
5911 number of sub-patterns, "i"th captured sub-pattern is
5912 ignored.
5913
5914 CAVEAT: An optional sub-pattern that does not exist in the matched
5915 string is assigned the empty string. Therefore, the following will
5916 return false (because the empty string is not a valid number):
5917
5918 int number;
5919 pcrecpp::RE::FullMatch("abc", "[a-z]+(\\d+)?", &number);
5920
5921 The matching interface supports at most 16 arguments per call. If you
5922 need more, consider using the more general interface
5923 pcrecpp::RE::DoMatch. See pcrecpp.h for the signature for DoMatch.
5924
5925
5926 QUOTING METACHARACTERS
5927
5928 You can use the "QuoteMeta" operation to insert backslashes before all
5929 potentially meaningful characters in a string. The returned string,
5930 used as a regular expression, will exactly match the original string.
5931
5932 Example:
5933 string quoted = RE::QuoteMeta(unquoted);
5934
5935 Note that it's legal to escape a character even if it has no special
5936 meaning in a regular expression -- so this function does that. (This
5937 also makes it identical to the perl function of the same name; see
5938 "perldoc -f quotemeta".) For example, "1.5-2.0?" becomes
5939 "1\.5\-2\.0\?".
5940
5941
5942 PARTIAL MATCHES
5943
5944 You can use the "PartialMatch" operation when you want the pattern to
5945 match any substring of the text.
5946
5947 Example: simple search for a string:
5948 pcrecpp::RE("ell").PartialMatch("hello");
5949
5950 Example: find first number in a string:
5951 int number;
5952 pcrecpp::RE re("(\\d+)");
5953 re.PartialMatch("x*100 + 20", &number);
5954 assert(number == 100);
5955
5956
5957 UTF-8 AND THE MATCHING INTERFACE
5958
5959 By default, pattern and text are plain text, one byte per character.
5960 The UTF8 flag, passed to the constructor, causes both pattern and
5961 string to be treated as UTF-8 text, still a byte stream but potentially
5962 multiple bytes per character. In practice, the text is likelier to be
5963 UTF-8 than the pattern, but the match returned may depend on the UTF8
5964 flag, so always use it when matching UTF8 text. For example, "." will
5965 match one byte normally but with UTF8 set may match up to three bytes
5966 of a multi-byte character.
5967
5968 Example:
5969 pcrecpp::RE_Options options;
5970 options.set_utf8();
5971 pcrecpp::RE re(utf8_pattern, options);
5972 re.FullMatch(utf8_string);
5973
5974 Example: using the convenience function UTF8():
5975 pcrecpp::RE re(utf8_pattern, pcrecpp::UTF8());
5976 re.FullMatch(utf8_string);
5977
5978 NOTE: The UTF8 flag is ignored if pcre was not configured with the
5979 --enable-utf8 flag.
5980
5981
5982 PASSING MODIFIERS TO THE REGULAR EXPRESSION ENGINE
5983
5984 PCRE defines some modifiers to change the behavior of the regular
5985 expression engine. The C++ wrapper defines an auxiliary class,
5986 RE_Options, as a vehicle to pass such modifiers to a RE class. Cur-
5987 rently, the following modifiers are supported:
5988
5989 modifier description Perl corresponding
5990
5991 PCRE_CASELESS case insensitive match /i
5992 PCRE_MULTILINE multiple lines match /m
5993 PCRE_DOTALL dot matches newlines /s
5994 PCRE_DOLLAR_ENDONLY $ matches only at end N/A
5995 PCRE_EXTRA strict escape parsing N/A
5996 PCRE_EXTENDED ignore whitespaces /x
5997 PCRE_UTF8 handles UTF8 chars built-in
5998 PCRE_UNGREEDY reverses * and *? N/A
5999 PCRE_NO_AUTO_CAPTURE disables capturing parens N/A (*)
6000
6001 (*) Both Perl and PCRE allow non capturing parentheses by means of the
6002 "?:" modifier within the pattern itself. e.g. (?:ab|cd) does not cap-
6003 ture, while (ab|cd) does.
6004
6005 For a full account on how each modifier works, please check the PCRE
6006 API reference page.
6007
6008 For each modifier, there are two member functions whose name is made
6009 out of the modifier in lowercase, without the "PCRE_" prefix. For
6010 instance, PCRE_CASELESS is handled by
6011
6012 bool caseless()
6013
6014 which returns true if the modifier is set, and
6015
6016 RE_Options & set_caseless(bool)
6017
6018 which sets or unsets the modifier. Moreover, PCRE_EXTRA_MATCH_LIMIT can
6019 be accessed through the set_match_limit() and match_limit() member
6020 functions. Setting match_limit to a non-zero value will limit the exe-
6021 cution of pcre to keep it from doing bad things like blowing the stack
6022 or taking an eternity to return a result. A value of 5000 is good
6023 enough to stop stack blowup in a 2MB thread stack. Setting match_limit
6024 to zero disables match limiting. Alternatively, you can call
6025 match_limit_recursion() which uses PCRE_EXTRA_MATCH_LIMIT_RECURSION to
6026 limit how much PCRE recurses. match_limit() limits the number of
6027 matches PCRE does; match_limit_recursion() limits the depth of internal
6028 recursion, and therefore the amount of stack that is used.
6029
6030 Normally, to pass one or more modifiers to a RE class, you declare a
6031 RE_Options object, set the appropriate options, and pass this object to
6032 a RE constructor. Example:
6033
6034 RE_options opt;
6035 opt.set_caseless(true);
6036 if (RE("HELLO", opt).PartialMatch("hello world")) ...
6037
6038 RE_options has two constructors. The default constructor takes no argu-
6039 ments and creates a set of flags that are off by default. The optional
6040 parameter option_flags is to facilitate transfer of legacy code from C
6041 programs. This lets you do
6042
6043 RE(pattern,
6044 RE_Options(PCRE_CASELESS|PCRE_MULTILINE)).PartialMatch(str);
6045
6046 However, new code is better off doing
6047
6048 RE(pattern,
6049 RE_Options().set_caseless(true).set_multiline(true))
6050 .PartialMatch(str);
6051
6052 If you are going to pass one of the most used modifiers, there are some
6053 convenience functions that return a RE_Options class with the appropri-
6054 ate modifier already set: CASELESS(), UTF8(), MULTILINE(), DOTALL(),
6055 and EXTENDED().
6056
6057 If you need to set several options at once, and you don't want to go
6058 through the pains of declaring a RE_Options object and setting several
6059 options, there is a parallel method that give you such ability on the
6060 fly. You can concatenate several set_xxxxx() member functions, since
6061 each of them returns a reference to its class object. For example, to
6062 pass PCRE_CASELESS, PCRE_EXTENDED, and PCRE_MULTILINE to a RE with one
6063 statement, you may write:
6064
6065 RE(" ^ xyz \\s+ .* blah$",
6066 RE_Options()
6067 .set_caseless(true)
6068 .set_extended(true)
6069 .set_multiline(true)).PartialMatch(sometext);
6070
6071
6072 SCANNING TEXT INCREMENTALLY
6073
6074 The "Consume" operation may be useful if you want to repeatedly match
6075 regular expressions at the front of a string and skip over them as they
6076 match. This requires use of the "StringPiece" type, which represents a
6077 sub-range of a real string. Like RE, StringPiece is defined in the
6078 pcrecpp namespace.
6079
6080 Example: read lines of the form "var = value" from a string.
6081 string contents = ...; // Fill string somehow
6082 pcrecpp::StringPiece input(contents); // Wrap in a StringPiece
6083
6084 string var;
6085 int value;
6086 pcrecpp::RE re("(\\w+) = (\\d+)\n");
6087 while (re.Consume(&input, &var, &value)) {
6088 ...;
6089 }
6090
6091 Each successful call to "Consume" will set "var/value", and also
6092 advance "input" so it points past the matched text.
6093
6094 The "FindAndConsume" operation is similar to "Consume" but does not
6095 anchor your match at the beginning of the string. For example, you
6096 could extract all words from a string by repeatedly calling
6097
6098 pcrecpp::RE("(\\w+)").FindAndConsume(&input, &word)
6099
6100
6101 PARSING HEX/OCTAL/C-RADIX NUMBERS
6102
6103 By default, if you pass a pointer to a numeric value, the corresponding
6104 text is interpreted as a base-10 number. You can instead wrap the
6105 pointer with a call to one of the operators Hex(), Octal(), or CRadix()
6106 to interpret the text in another base. The CRadix operator interprets
6107 C-style "0" (base-8) and "0x" (base-16) prefixes, but defaults to
6108 base-10.
6109
6110 Example:
6111 int a, b, c, d;
6112