ViewVC logotype

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

Parent Directory Parent Directory | Revision Log Revision Log

Revision 738 - (show annotations)
Fri Oct 21 09:04:01 2011 UTC (9 years, 6 months ago) by ph10
File MIME type: text/plain
File size: 403072 byte(s)
Source tidies for 8.20 release.
1 -----------------------------------------------------------------------------
2 This file contains a concatenation of the PCRE man pages, converted to plain
3 text format for ease of searching with a text editor, or for use on systems
4 that do not have a man page processor. The small individual files that give
5 synopses of each function in the library have not been included. Neither has
6 the pcredemo program. There are separate text files for the pcregrep and
7 pcretest commands.
8 -----------------------------------------------------------------------------
11 PCRE(3) PCRE(3)
15 PCRE - Perl-compatible regular expressions
20 The PCRE library is a set of functions that implement regular expres-
21 sion pattern matching using the same syntax and semantics as Perl, with
22 just a few differences. Some features that appeared in Python and PCRE
23 before they appeared in Perl are also available using the Python syn-
24 tax, there is some support for one or two .NET and Oniguruma syntax
25 items, and there is an option for requesting some minor changes that
26 give better JavaScript compatibility.
28 The current implementation of PCRE corresponds approximately with Perl
29 5.12, including support for UTF-8 encoded strings and Unicode general
30 category properties. However, UTF-8 and Unicode support has to be
31 explicitly enabled; it is not the default. The Unicode tables corre-
32 spond to Unicode release 6.0.0.
34 In addition to the Perl-compatible matching function, PCRE contains an
35 alternative function that matches the same compiled patterns in a dif-
36 ferent way. In certain circumstances, the alternative function has some
37 advantages. For a discussion of the two matching algorithms, see the
38 pcrematching page.
40 PCRE is written in C and released as a C library. A number of people
41 have written wrappers and interfaces of various kinds. In particular,
42 Google Inc. have provided a comprehensive C++ wrapper. This is now
43 included as part of the PCRE distribution. The pcrecpp page has details
44 of this interface. Other people's contributions can be found in the
45 Contrib directory at the primary FTP site, which is:
47 ftp://ftp.csx.cam.ac.uk/pub/software/programming/pcre
49 Details of exactly which Perl regular expression features are and are
50 not supported by PCRE are given in separate documents. See the pcrepat-
51 tern and pcrecompat pages. There is a syntax summary in the pcresyntax
52 page.
54 Some features of PCRE can be included, excluded, or changed when the
55 library is built. The pcre_config() function makes it possible for a
56 client to discover which features are available. The features them-
57 selves are described in the pcrebuild page. Documentation about build-
58 ing PCRE for various operating systems can be found in the README and
59 NON-UNIX-USE files in the source distribution.
61 The library contains a number of undocumented internal functions and
62 data tables that are used by more than one of the exported external
63 functions, but which are not intended for use by external callers.
64 Their names all begin with "_pcre_", which hopefully will not provoke
65 any name clashes. In some environments, it is possible to control which
66 external symbols are exported when a shared library is built, and in
67 these cases the undocumented symbols are not exported.
72 The user documentation for PCRE comprises a number of different sec-
73 tions. In the "man" format, each of these is a separate "man page". In
74 the HTML format, each is a separate page, linked from the index page.
75 In the plain text format, all the sections, except the pcredemo sec-
76 tion, are concatenated, for ease of searching. The sections are as fol-
77 lows:
79 pcre this document
80 pcre-config show PCRE installation configuration information
81 pcreapi details of PCRE's native C API
82 pcrebuild options for building PCRE
83 pcrecallout details of the callout feature
84 pcrecompat discussion of Perl compatibility
85 pcrecpp details of the C++ wrapper
86 pcredemo a demonstration C program that uses PCRE
87 pcregrep description of the pcregrep command
88 pcrejit discussion of the just-in-time optimization support
89 pcrelimits details of size and other limits
90 pcrematching discussion of the two matching algorithms
91 pcrepartial details of the partial matching facility
92 pcrepattern syntax and semantics of supported
93 regular expressions
94 pcreperform discussion of performance issues
95 pcreposix the POSIX-compatible C API
96 pcreprecompile details of saving and re-using precompiled patterns
97 pcresample discussion of the pcredemo program
98 pcrestack discussion of stack usage
99 pcresyntax quick syntax reference
100 pcretest description of the pcretest testing command
101 pcreunicode discussion of Unicode and UTF-8 support
103 In addition, in the "man" and HTML formats, there is a short page for
104 each C library function, listing its arguments and results.
109 Philip Hazel
110 University Computing Service
111 Cambridge CB2 3QH, England.
113 Putting an actual email address here seems to have been a spam magnet,
114 so I've taken it away. If you want to email me, use my two initials,
115 followed by the two digits 10, at the domain cam.ac.uk.
120 Last updated: 24 August 2011
121 Copyright (c) 1997-2011 University of Cambridge.
122 ------------------------------------------------------------------------------
128 NAME
129 PCRE - Perl-compatible regular expressions
134 This document describes the optional features of PCRE that can be
135 selected when the library is compiled. It assumes use of the configure
136 script, where the optional features are selected or deselected by pro-
137 viding options to configure before running the make command. However,
138 the same options can be selected in both Unix-like and non-Unix-like
139 environments using the GUI facility of cmake-gui if you are using CMake
140 instead of configure to build PCRE.
142 There is a lot more information about building PCRE in non-Unix-like
143 environments in the file called NON_UNIX_USE, which is part of the PCRE
144 distribution. You should consult this file as well as the README file
145 if you are building in a non-Unix-like environment.
147 The complete list of options for configure (which includes the standard
148 ones such as the selection of the installation directory) can be
149 obtained by running
151 ./configure --help
153 The following sections include descriptions of options whose names
154 begin with --enable or --disable. These settings specify changes to the
155 defaults for the configure command. Because of the way that configure
156 works, --enable and --disable always come in pairs, so the complemen-
157 tary option always exists as well, but as it specifies the default, it
158 is not described.
163 The PCRE building process uses libtool to build both shared and static
164 Unix libraries by default. You can suppress one of these by adding one
165 of
167 --disable-shared
168 --disable-static
170 to the configure command, as required.
175 By default, the configure script will search for a C++ compiler and C++
176 header files. If it finds them, it automatically builds the C++ wrapper
177 library for PCRE. You can disable this by adding
179 --disable-cpp
181 to the configure command.
186 To build PCRE with support for UTF-8 Unicode character strings, add
188 --enable-utf8
190 to the configure command. Of itself, this does not make PCRE treat
191 strings as UTF-8. As well as compiling PCRE with this option, you also
192 have have to set the PCRE_UTF8 option when you call the pcre_compile()
193 or pcre_compile2() functions.
195 If you set --enable-utf8 when compiling in an EBCDIC environment, PCRE
196 expects its input to be either ASCII or UTF-8 (depending on the runtime
197 option). It is not possible to support both EBCDIC and UTF-8 codes in
198 the same version of the library. Consequently, --enable-utf8 and
199 --enable-ebcdic are mutually exclusive.
204 UTF-8 support allows PCRE to process character values greater than 255
205 in the strings that it handles. On its own, however, it does not pro-
206 vide any facilities for accessing the properties of such characters. If
207 you want to be able to use the pattern escapes \P, \p, and \X, which
208 refer to Unicode character properties, you must add
210 --enable-unicode-properties
212 to the configure command. This implies UTF-8 support, even if you have
213 not explicitly requested it.
215 Including Unicode property support adds around 30K of tables to the
216 PCRE library. Only the general category properties such as Lu and Nd
217 are supported. Details are given in the pcrepattern documentation.
222 Just-in-time compiler support is included in the build by specifying
224 --enable-jit
226 This support is available only for certain hardware architectures. If
227 this option is set for an unsupported architecture, a compile time
228 error occurs. See the pcrejit documentation for a discussion of JIT
229 usage. When JIT support is enabled, pcregrep automatically makes use of
230 it, unless you add
232 --disable-pcregrep-jit
234 to the "configure" command.
239 By default, PCRE interprets the linefeed (LF) character as indicating
240 the end of a line. This is the normal newline character on Unix-like
241 systems. You can compile PCRE to use carriage return (CR) instead, by
242 adding
244 --enable-newline-is-cr
246 to the configure command. There is also a --enable-newline-is-lf
247 option, which explicitly specifies linefeed as the newline character.
249 Alternatively, you can specify that line endings are to be indicated by
250 the two character sequence CRLF. If you want this, add
252 --enable-newline-is-crlf
254 to the configure command. There is a fourth option, specified by
256 --enable-newline-is-anycrlf
258 which causes PCRE to recognize any of the three sequences CR, LF, or
259 CRLF as indicating a line ending. Finally, a fifth option, specified by
261 --enable-newline-is-any
263 causes PCRE to recognize any Unicode newline sequence.
265 Whatever line ending convention is selected when PCRE is built can be
266 overridden when the library functions are called. At build time it is
267 conventional to use the standard for your operating system.
272 By default, the sequence \R in a pattern matches any Unicode newline
273 sequence, whatever has been selected as the line ending sequence. If
274 you specify
276 --enable-bsr-anycrlf
278 the default is changed so that \R matches only CR, LF, or CRLF. What-
279 ever is selected when PCRE is built can be overridden when the library
280 functions are called.
285 When PCRE is called through the POSIX interface (see the pcreposix doc-
286 umentation), additional working storage is required for holding the
287 pointers to capturing substrings, because PCRE requires three integers
288 per substring, whereas the POSIX interface provides only two. If the
289 number of expected substrings is small, the wrapper function uses space
290 on the stack, because this is faster than using malloc() for each call.
291 The default threshold above which the stack is no longer used is 10; it
292 can be changed by adding a setting such as
294 --with-posix-malloc-threshold=20
296 to the configure command.
301 Within a compiled pattern, offset values are used to point from one
302 part to another (for example, from an opening parenthesis to an alter-
303 nation metacharacter). By default, two-byte values are used for these
304 offsets, leading to a maximum size for a compiled pattern of around
305 64K. This is sufficient to handle all but the most gigantic patterns.
306 Nevertheless, some people do want to process truyl enormous patterns,
307 so it is possible to compile PCRE to use three-byte or four-byte off-
308 sets by adding a setting such as
310 --with-link-size=3
312 to the configure command. The value given must be 2, 3, or 4. Using
313 longer offsets slows down the operation of PCRE because it has to load
314 additional bytes when handling them.
319 When matching with the pcre_exec() function, PCRE implements backtrack-
320 ing by making recursive calls to an internal function called match().
321 In environments where the size of the stack is limited, this can se-
322 verely limit PCRE's operation. (The Unix environment does not usually
323 suffer from this problem, but it may sometimes be necessary to increase
324 the maximum stack size. There is a discussion in the pcrestack docu-
325 mentation.) An alternative approach to recursion that uses memory from
326 the heap to remember data, instead of using recursive function calls,
327 has been implemented to work round the problem of limited stack size.
328 If you want to build a version of PCRE that works this way, add
330 --disable-stack-for-recursion
332 to the configure command. With this configuration, PCRE will use the
333 pcre_stack_malloc and pcre_stack_free variables to call memory manage-
334 ment functions. By default these point to malloc() and free(), but you
335 can replace the pointers so that your own functions are used instead.
337 Separate functions are provided rather than using pcre_malloc and
338 pcre_free because the usage is very predictable: the block sizes
339 requested are always the same, and the blocks are always freed in
340 reverse order. A calling program might be able to implement optimized
341 functions that perform better than malloc() and free(). PCRE runs
342 noticeably more slowly when built in this way. This option affects only
343 the pcre_exec() function; it is not relevant for pcre_dfa_exec().
348 Internally, PCRE has a function called match(), which it calls repeat-
349 edly (sometimes recursively) when matching a pattern with the
350 pcre_exec() function. By controlling the maximum number of times this
351 function may be called during a single matching operation, a limit can
352 be placed on the resources used by a single call to pcre_exec(). The
353 limit can be changed at run time, as described in the pcreapi documen-
354 tation. The default is 10 million, but this can be changed by adding a
355 setting such as
357 --with-match-limit=500000
359 to the configure command. This setting has no effect on the
360 pcre_dfa_exec() matching function.
362 In some environments it is desirable to limit the depth of recursive
363 calls of match() more strictly than the total number of calls, in order
364 to restrict the maximum amount of stack (or heap, if --disable-stack-
365 for-recursion is specified) that is used. A second limit controls this;
366 it defaults to the value that is set for --with-match-limit, which
367 imposes no additional constraints. However, you can set a lower limit
368 by adding, for example,
370 --with-match-limit-recursion=10000
372 to the configure command. This value can also be overridden at run
373 time.
378 PCRE uses fixed tables for processing characters whose code values are
379 less than 256. By default, PCRE is built with a set of tables that are
380 distributed in the file pcre_chartables.c.dist. These tables are for
381 ASCII codes only. If you add
383 --enable-rebuild-chartables
385 to the configure command, the distributed tables are no longer used.
386 Instead, a program called dftables is compiled and run. This outputs
387 the source for new set of tables, created in the default locale of your
388 C runtime system. (This method of replacing the tables does not work if
389 you are cross compiling, because dftables is run on the local host. If
390 you need to create alternative tables when cross compiling, you will
391 have to do so "by hand".)
396 PCRE assumes by default that it will run in an environment where the
397 character code is ASCII (or Unicode, which is a superset of ASCII).
398 This is the case for most computer operating systems. PCRE can, how-
399 ever, be compiled to run in an EBCDIC environment by adding
401 --enable-ebcdic
403 to the configure command. This setting implies --enable-rebuild-charta-
404 bles. You should only use it if you know that you are in an EBCDIC
405 environment (for example, an IBM mainframe operating system). The
406 --enable-ebcdic option is incompatible with --enable-utf8.
411 By default, pcregrep reads all files as plain text. You can build it so
412 that it recognizes files whose names end in .gz or .bz2, and reads them
413 with libz or libbz2, respectively, by adding one or both of
415 --enable-pcregrep-libz
416 --enable-pcregrep-libbz2
418 to the configure command. These options naturally require that the rel-
419 evant libraries are installed on your system. Configuration will fail
420 if they are not.
425 pcregrep uses an internal buffer to hold a "window" on the file it is
426 scanning, in order to be able to output "before" and "after" lines when
427 it finds a match. The size of the buffer is controlled by a parameter
428 whose default value is 20K. The buffer itself is three times this size,
429 but because of the way it is used for holding "before" lines, the long-
430 est line that is guaranteed to be processable is the parameter size.
431 You can change the default parameter value by adding, for example,
433 --with-pcregrep-bufsize=50K
435 to the configure command. The caller of pcregrep can, however, override
436 this value by specifying a run-time option.
441 If you add
443 --enable-pcretest-libreadline
445 to the configure command, pcretest is linked with the libreadline
446 library, and when its input is from a terminal, it reads it using the
447 readline() function. This provides line-editing and history facilities.
448 Note that libreadline is GPL-licensed, so if you distribute a binary of
449 pcretest linked in this way, there may be licensing issues.
451 Setting this option causes the -lreadline option to be added to the
452 pcretest build. In many operating environments with a sytem-installed
453 libreadline this is sufficient. However, in some environments (e.g. if
454 an unmodified distribution version of readline is in use), some extra
455 configuration may be necessary. The INSTALL file for libreadline says
456 this:
458 "Readline uses the termcap functions, but does not link with the
459 termcap or curses library itself, allowing applications which link
460 with readline the to choose an appropriate library."
462 If your environment has not been set up so that an appropriate library
463 is automatically included, you may need to add something like
465 LIBS="-ncurses"
467 immediately before the configure command.
472 pcreapi(3), pcre_config(3).
477 Philip Hazel
478 University Computing Service
479 Cambridge CB2 3QH, England.
484 Last updated: 06 September 2011
485 Copyright (c) 1997-2011 University of Cambridge.
486 ------------------------------------------------------------------------------
492 NAME
493 PCRE - Perl-compatible regular expressions
498 This document describes the two different algorithms that are available
499 in PCRE for matching a compiled regular expression against a given sub-
500 ject string. The "standard" algorithm is the one provided by the
501 pcre_exec() function. This works in the same was as Perl's matching
502 function, and provides a Perl-compatible matching operation.
504 An alternative algorithm is provided by the pcre_dfa_exec() function;
505 this operates in a different way, and is not Perl-compatible. It has
506 advantages and disadvantages compared with the standard algorithm, and
507 these are described below.
509 When there is only one possible way in which a given subject string can
510 match a pattern, the two algorithms give the same answer. A difference
511 arises, however, when there are multiple possibilities. For example, if
512 the pattern
514 ^<.*>
516 is matched against the string
518 <something> <something else> <something further>
520 there are three possible answers. The standard algorithm finds only one
521 of them, whereas the alternative algorithm finds all three.
526 The set of strings that are matched by a regular expression can be rep-
527 resented as a tree structure. An unlimited repetition in the pattern
528 makes the tree of infinite size, but it is still a tree. Matching the
529 pattern to a given subject string (from a given starting point) can be
530 thought of as a search of the tree. There are two ways to search a
531 tree: depth-first and breadth-first, and these correspond to the two
532 matching algorithms provided by PCRE.
537 In the terminology of Jeffrey Friedl's book "Mastering Regular Expres-
538 sions", the standard algorithm is an "NFA algorithm". It conducts a
539 depth-first search of the pattern tree. That is, it proceeds along a
540 single path through the tree, checking that the subject matches what is
541 required. When there is a mismatch, the algorithm tries any alterna-
542 tives at the current point, and if they all fail, it backs up to the
543 previous branch point in the tree, and tries the next alternative
544 branch at that level. This often involves backing up (moving to the
545 left) in the subject string as well. The order in which repetition
546 branches are tried is controlled by the greedy or ungreedy nature of
547 the quantifier.
549 If a leaf node is reached, a matching string has been found, and at
550 that point the algorithm stops. Thus, if there is more than one possi-
551 ble match, this algorithm returns the first one that it finds. Whether
552 this is the shortest, the longest, or some intermediate length depends
553 on the way the greedy and ungreedy repetition quantifiers are specified
554 in the pattern.
556 Because it ends up with a single path through the tree, it is rela-
557 tively straightforward for this algorithm to keep track of the sub-
558 strings that are matched by portions of the pattern in parentheses.
559 This provides support for capturing parentheses and back references.
564 This algorithm conducts a breadth-first search of the tree. Starting
565 from the first matching point in the subject, it scans the subject
566 string from left to right, once, character by character, and as it does
567 this, it remembers all the paths through the tree that represent valid
568 matches. In Friedl's terminology, this is a kind of "DFA algorithm",
569 though it is not implemented as a traditional finite state machine (it
570 keeps multiple states active simultaneously).
572 Although the general principle of this matching algorithm is that it
573 scans the subject string only once, without backtracking, there is one
574 exception: when a lookaround assertion is encountered, the characters
575 following or preceding the current point have to be independently
576 inspected.
578 The scan continues until either the end of the subject is reached, or
579 there are no more unterminated paths. At this point, terminated paths
580 represent the different matching possibilities (if there are none, the
581 match has failed). Thus, if there is more than one possible match,
582 this algorithm finds all of them, and in particular, it finds the long-
583 est. The matches are returned in decreasing order of length. There is
584 an option to stop the algorithm after the first match (which is neces-
585 sarily the shortest) is found.
587 Note that all the matches that are found start at the same point in the
588 subject. If the pattern
590 cat(er(pillar)?)?
592 is matched against the string "the caterpillar catchment", the result
593 will be the three strings "caterpillar", "cater", and "cat" that start
594 at the fifth character of the subject. The algorithm does not automati-
595 cally move on to find matches that start at later positions.
597 There are a number of features of PCRE regular expressions that are not
598 supported by the alternative matching algorithm. They are as follows:
600 1. Because the algorithm finds all possible matches, the greedy or
601 ungreedy nature of repetition quantifiers is not relevant. Greedy and
602 ungreedy quantifiers are treated in exactly the same way. However, pos-
603 sessive quantifiers can make a difference when what follows could also
604 match what is quantified, for example in a pattern like this:
606 ^a++\w!
608 This pattern matches "aaab!" but not "aaa!", which would be matched by
609 a non-possessive quantifier. Similarly, if an atomic group is present,
610 it is matched as if it were a standalone pattern at the current point,
611 and the longest match is then "locked in" for the rest of the overall
612 pattern.
614 2. When dealing with multiple paths through the tree simultaneously, it
615 is not straightforward to keep track of captured substrings for the
616 different matching possibilities, and PCRE's implementation of this
617 algorithm does not attempt to do this. This means that no captured sub-
618 strings are available.
620 3. Because no substrings are captured, back references within the pat-
621 tern are not supported, and cause errors if encountered.
623 4. For the same reason, conditional expressions that use a backrefer-
624 ence as the condition or test for a specific group recursion are not
625 supported.
627 5. Because many paths through the tree may be active, the \K escape
628 sequence, which resets the start of the match when encountered (but may
629 be on some paths and not on others), is not supported. It causes an
630 error if encountered.
632 6. Callouts are supported, but the value of the capture_top field is
633 always 1, and the value of the capture_last field is always -1.
635 7. The \C escape sequence, which (in the standard algorithm) matches a
636 single byte, even in UTF-8 mode, is not supported because the alterna-
637 tive algorithm moves through the subject string one character at a
638 time, for all active paths through the tree.
640 8. Except for (*FAIL), the backtracking control verbs such as (*PRUNE)
641 are not supported. (*FAIL) is supported, and behaves like a failing
642 negative assertion.
647 Using the alternative matching algorithm provides the following advan-
648 tages:
650 1. All possible matches (at a single point in the subject) are automat-
651 ically found, and in particular, the longest match is found. To find
652 more than one match using the standard algorithm, you have to do kludgy
653 things with callouts.
655 2. Because the alternative algorithm scans the subject string just
656 once, and never needs to backtrack, it is possible to pass very long
657 subject strings to the matching function in several pieces, checking
658 for partial matching each time. Although it is possible to do multi-
659 segment matching using the standard algorithm (pcre_exec()), by retain-
660 ing partially matched substrings, it is more complicated. The pcrepar-
661 tial documentation gives details of partial matching and discusses
662 multi-segment matching.
667 The alternative algorithm suffers from a number of disadvantages:
669 1. It is substantially slower than the standard algorithm. This is
670 partly because it has to search for all possible matches, but is also
671 because it is less susceptible to optimization.
673 2. Capturing parentheses and back references are not supported.
675 3. Although atomic groups are supported, their use does not provide the
676 performance advantage that it does for the standard algorithm.
681 Philip Hazel
682 University Computing Service
683 Cambridge CB2 3QH, England.
688 Last updated: 17 November 2010
689 Copyright (c) 1997-2010 University of Cambridge.
690 ------------------------------------------------------------------------------
696 NAME
697 PCRE - Perl-compatible regular expressions
702 #include <pcre.h>
704 pcre *pcre_compile(const char *pattern, int options,
705 const char **errptr, int *erroffset,
706 const unsigned char *tableptr);
708 pcre *pcre_compile2(const char *pattern, int options,
709 int *errorcodeptr,
710 const char **errptr, int *erroffset,
711 const unsigned char *tableptr);
713 pcre_extra *pcre_study(const pcre *code, int options,
714 const char **errptr);
716 void pcre_free_study(pcre_extra *extra);
718 int pcre_exec(const pcre *code, const pcre_extra *extra,
719 const char *subject, int length, int startoffset,
720 int options, int *ovector, int ovecsize);
725 pcre_jit_stack *pcre_jit_stack_alloc(int startsize, int maxsize);
727 void pcre_jit_stack_free(pcre_jit_stack *stack);
729 void pcre_assign_jit_stack(pcre_extra *extra,
730 pcre_jit_callback callback, void *data);
732 int pcre_dfa_exec(const pcre *code, const pcre_extra *extra,
733 const char *subject, int length, int startoffset,
734 int options, int *ovector, int ovecsize,
735 int *workspace, int wscount);
737 int pcre_copy_named_substring(const pcre *code,
738 const char *subject, int *ovector,
739 int stringcount, const char *stringname,
740 char *buffer, int buffersize);
742 int pcre_copy_substring(const char *subject, int *ovector,
743 int stringcount, int stringnumber, char *buffer,
744 int buffersize);
746 int pcre_get_named_substring(const pcre *code,
747 const char *subject, int *ovector,
748 int stringcount, const char *stringname,
749 const char **stringptr);
751 int pcre_get_stringnumber(const pcre *code,
752 const char *name);
754 int pcre_get_stringtable_entries(const pcre *code,
755 const char *name, char **first, char **last);
757 int pcre_get_substring(const char *subject, int *ovector,
758 int stringcount, int stringnumber,
759 const char **stringptr);
761 int pcre_get_substring_list(const char *subject,
762 int *ovector, int stringcount, const char ***listptr);
764 void pcre_free_substring(const char *stringptr);
766 void pcre_free_substring_list(const char **stringptr);
768 const unsigned char *pcre_maketables(void);
770 int pcre_fullinfo(const pcre *code, const pcre_extra *extra,
771 int what, void *where);
773 int pcre_info(const pcre *code, int *optptr, int *firstcharptr);
775 int pcre_refcount(pcre *code, int adjust);
777 int pcre_config(int what, void *where);
779 char *pcre_version(void);
784 void *(*pcre_malloc)(size_t);
786 void (*pcre_free)(void *);
788 void *(*pcre_stack_malloc)(size_t);
790 void (*pcre_stack_free)(void *);
792 int (*pcre_callout)(pcre_callout_block *);
797 PCRE has its own native API, which is described in this document. There
798 are also some wrapper functions that correspond to the POSIX regular
799 expression API, but they do not give access to all the functionality.
800 They are described in the pcreposix documentation. Both of these APIs
801 define a set of C function calls. A C++ wrapper is also distributed
802 with PCRE. It is documented in the pcrecpp page.
804 The native API C function prototypes are defined in the header file
805 pcre.h, and on Unix systems the library itself is called libpcre. It
806 can normally be accessed by adding -lpcre to the command for linking an
807 application that uses PCRE. The header file defines the macros
808 PCRE_MAJOR and PCRE_MINOR to contain the major and minor release num-
809 bers for the library. Applications can use these to include support
810 for different releases of PCRE.
812 In a Windows environment, if you want to statically link an application
813 program against a non-dll pcre.a file, you must define PCRE_STATIC
814 before including pcre.h or pcrecpp.h, because otherwise the pcre_mal-
815 loc() and pcre_free() exported functions will be declared
816 __declspec(dllimport), with unwanted results.
818 The functions pcre_compile(), pcre_compile2(), pcre_study(), and
819 pcre_exec() are used for compiling and matching regular expressions in
820 a Perl-compatible manner. A sample program that demonstrates the sim-
821 plest way of using them is provided in the file called pcredemo.c in
822 the PCRE source distribution. A listing of this program is given in the
823 pcredemo documentation, and the pcresample documentation describes how
824 to compile and run it.
826 Just-in-time compiler support is an optional feature of PCRE that can
827 be built in appropriate hardware environments. It greatly speeds up the
828 matching performance of many patterns. Simple programs can easily
829 request that it be used if available, by setting an option that is
830 ignored when it is not relevant. More complicated programs might need
831 to make use of the functions pcre_jit_stack_alloc(),
832 pcre_jit_stack_free(), and pcre_assign_jit_stack() in order to control
833 the JIT code's memory usage. These functions are discussed in the
834 pcrejit documentation.
836 A second matching function, pcre_dfa_exec(), which is not Perl-compati-
837 ble, is also provided. This uses a different algorithm for the match-
838 ing. The alternative algorithm finds all possible matches (at a given
839 point in the subject), and scans the subject just once (unless there
840 are lookbehind assertions). However, this algorithm does not return
841 captured substrings. A description of the two matching algorithms and
842 their advantages and disadvantages is given in the pcrematching docu-
843 mentation.
845 In addition to the main compiling and matching functions, there are
846 convenience functions for extracting captured substrings from a subject
847 string that is matched by pcre_exec(). They are:
849 pcre_copy_substring()
850 pcre_copy_named_substring()
851 pcre_get_substring()
852 pcre_get_named_substring()
853 pcre_get_substring_list()
854 pcre_get_stringnumber()
855 pcre_get_stringtable_entries()
857 pcre_free_substring() and pcre_free_substring_list() are also provided,
858 to free the memory used for extracted strings.
860 The function pcre_maketables() is used to build a set of character
861 tables in the current locale for passing to pcre_compile(),
862 pcre_exec(), or pcre_dfa_exec(). This is an optional facility that is
863 provided for specialist use. Most commonly, no special tables are
864 passed, in which case internal tables that are generated when PCRE is
865 built are used.
867 The function pcre_fullinfo() is used to find out information about a
868 compiled pattern; pcre_info() is an obsolete version that returns only
869 some of the available information, but is retained for backwards com-
870 patibility. The function pcre_version() returns a pointer to a string
871 containing the version of PCRE and its date of release.
873 The function pcre_refcount() maintains a reference count in a data
874 block containing a compiled pattern. This is provided for the benefit
875 of object-oriented applications.
877 The global variables pcre_malloc and pcre_free initially contain the
878 entry points of the standard malloc() and free() functions, respec-
879 tively. PCRE calls the memory management functions via these variables,
880 so a calling program can replace them if it wishes to intercept the
881 calls. This should be done before calling any PCRE functions.
883 The global variables pcre_stack_malloc and pcre_stack_free are also
884 indirections to memory management functions. These special functions
885 are used only when PCRE is compiled to use the heap for remembering
886 data, instead of recursive function calls, when running the pcre_exec()
887 function. See the pcrebuild documentation for details of how to do
888 this. It is a non-standard way of building PCRE, for use in environ-
889 ments that have limited stacks. Because of the greater use of memory
890 management, it runs more slowly. Separate functions are provided so
891 that special-purpose external code can be used for this case. When
892 used, these functions are always called in a stack-like manner (last
893 obtained, first freed), and always for memory blocks of the same size.
894 There is a discussion about PCRE's stack usage in the pcrestack docu-
895 mentation.
897 The global variable pcre_callout initially contains NULL. It can be set
898 by the caller to a "callout" function, which PCRE will then call at
899 specified points during a matching operation. Details are given in the
900 pcrecallout documentation.
905 PCRE supports five different conventions for indicating line breaks in
906 strings: a single CR (carriage return) character, a single LF (line-
907 feed) character, the two-character sequence CRLF, any of the three pre-
908 ceding, or any Unicode newline sequence. The Unicode newline sequences
909 are the three just mentioned, plus the single characters VT (vertical
910 tab, U+000B), FF (formfeed, U+000C), NEL (next line, U+0085), LS (line
911 separator, U+2028), and PS (paragraph separator, U+2029).
913 Each of the first three conventions is used by at least one operating
914 system as its standard newline sequence. When PCRE is built, a default
915 can be specified. The default default is LF, which is the Unix stan-
916 dard. When PCRE is run, the default can be overridden, either when a
917 pattern is compiled, or when it is matched.
919 At compile time, the newline convention can be specified by the options
920 argument of pcre_compile(), or it can be specified by special text at
921 the start of the pattern itself; this overrides any other settings. See
922 the pcrepattern page for details of the special character sequences.
924 In the PCRE documentation the word "newline" is used to mean "the char-
925 acter or pair of characters that indicate a line break". The choice of
926 newline convention affects the handling of the dot, circumflex, and
927 dollar metacharacters, the handling of #-comments in /x mode, and, when
928 CRLF is a recognized line ending sequence, the match position advance-
929 ment for a non-anchored pattern. There is more detail about this in the
930 section on pcre_exec() options below.
932 The choice of newline convention does not affect the interpretation of
933 the \n or \r escape sequences, nor does it affect what \R matches,
934 which is controlled in a similar way, but by separate options.
939 The PCRE functions can be used in multi-threading applications, with
940 the proviso that the memory management functions pointed to by
941 pcre_malloc, pcre_free, pcre_stack_malloc, and pcre_stack_free, and the
942 callout function pointed to by pcre_callout, are shared by all threads.
944 The compiled form of a regular expression is not altered during match-
945 ing, so the same compiled pattern can safely be used by several threads
946 at once.
948 If the just-in-time optimization feature is being used, it needs sepa-
949 rate memory stack areas for each thread. See the pcrejit documentation
950 for more details.
955 The compiled form of a regular expression can be saved and re-used at a
956 later time, possibly by a different program, and even on a host other
957 than the one on which it was compiled. Details are given in the
958 pcreprecompile documentation. However, compiling a regular expression
959 with one version of PCRE for use with a different version is not guar-
960 anteed to work and may cause crashes.
965 int pcre_config(int what, void *where);
967 The function pcre_config() makes it possible for a PCRE client to dis-
968 cover which optional features have been compiled into the PCRE library.
969 The pcrebuild documentation has more details about these optional fea-
970 tures.
972 The first argument for pcre_config() is an integer, specifying which
973 information is required; the second argument is a pointer to a variable
974 into which the information is placed. The following information is
975 available:
979 The output is an integer that is set to one if UTF-8 support is avail-
980 able; otherwise it is set to zero.
984 The output is an integer that is set to one if support for Unicode
985 character properties is available; otherwise it is set to zero.
989 The output is an integer that is set to one if support for just-in-time
990 compiling is available; otherwise it is set to zero.
994 The output is an integer whose value specifies the default character
995 sequence that is recognized as meaning "newline". The four values that
996 are supported are: 10 for LF, 13 for CR, 3338 for CRLF, -2 for ANYCRLF,
997 and -1 for ANY. Though they are derived from ASCII, the same values
998 are returned in EBCDIC environments. The default should normally corre-
999 spond to the standard sequence for your operating system.
1003 The output is an integer whose value indicates what character sequences
1004 the \R escape sequence matches by default. A value of 0 means that \R
1005 matches any Unicode line ending sequence; a value of 1 means that \R
1006 matches only CR, LF, or CRLF. The default can be overridden when a pat-
1007 tern is compiled or matched.
1011 The output is an integer that contains the number of bytes used for
1012 internal linkage in compiled regular expressions. The value is 2, 3, or
1013 4. Larger values allow larger regular expressions to be compiled, at
1014 the expense of slower matching. The default value of 2 is sufficient
1015 for all but the most massive patterns, since it allows the compiled
1016 pattern to be up to 64K in size.
1020 The output is an integer that contains the threshold above which the
1021 POSIX interface uses malloc() for output vectors. Further details are
1022 given in the pcreposix documentation.
1026 The output is a long integer that gives the default limit for the num-
1027 ber of internal matching function calls in a pcre_exec() execution.
1028 Further details are given with pcre_exec() below.
1032 The output is a long integer that gives the default limit for the depth
1033 of recursion when calling the internal matching function in a
1034 pcre_exec() execution. Further details are given with pcre_exec()
1035 below.
1039 The output is an integer that is set to one if internal recursion when
1040 running pcre_exec() is implemented by recursive function calls that use
1041 the stack to remember their state. This is the usual way that PCRE is
1042 compiled. The output is zero if PCRE was compiled to use blocks of data
1043 on the heap instead of recursive function calls. In this case,
1044 pcre_stack_malloc and pcre_stack_free are called to manage memory
1045 blocks on the heap, thus avoiding the use of the stack.
1050 pcre *pcre_compile(const char *pattern, int options,
1051 const char **errptr, int *erroffset,
1052 const unsigned char *tableptr);
1054 pcre *pcre_compile2(const char *pattern, int options,
1055 int *errorcodeptr,
1056 const char **errptr, int *erroffset,
1057 const unsigned char *tableptr);
1059 Either of the functions pcre_compile() or pcre_compile2() can be called
1060 to compile a pattern into an internal form. The only difference between
1061 the two interfaces is that pcre_compile2() has an additional argument,
1062 errorcodeptr, via which a numerical error code can be returned. To
1063 avoid too much repetition, we refer just to pcre_compile() below, but
1064 the information applies equally to pcre_compile2().
1066 The pattern is a C string terminated by a binary zero, and is passed in
1067 the pattern argument. A pointer to a single block of memory that is
1068 obtained via pcre_malloc is returned. This contains the compiled code
1069 and related data. The pcre type is defined for the returned block; this
1070 is a typedef for a structure whose contents are not externally defined.
1071 It is up to the caller to free the memory (via pcre_free) when it is no
1072 longer required.
1074 Although the compiled code of a PCRE regex is relocatable, that is, it
1075 does not depend on memory location, the complete pcre data block is not
1076 fully relocatable, because it may contain a copy of the tableptr argu-
1077 ment, which is an address (see below).
1079 The options argument contains various bit settings that affect the com-
1080 pilation. It should be zero if no options are required. The available
1081 options are described below. Some of them (in particular, those that
1082 are compatible with Perl, but some others as well) can also be set and
1083 unset from within the pattern (see the detailed description in the
1084 pcrepattern documentation). For those options that can be different in
1085 different parts of the pattern, the contents of the options argument
1086 specifies their settings at the start of compilation and execution. The
1088 PCRE_NO_START_OPT options can be set at the time of matching as well as
1089 at compile time.
1091 If errptr is NULL, pcre_compile() returns NULL immediately. Otherwise,
1092 if compilation of a pattern fails, pcre_compile() returns NULL, and
1093 sets the variable pointed to by errptr to point to a textual error mes-
1094 sage. This is a static string that is part of the library. You must not
1095 try to free it. Normally, the offset from the start of the pattern to
1096 the byte that was being processed when the error was discovered is
1097 placed in the variable pointed to by erroffset, which must not be NULL
1098 (if it is, an immediate error is given). However, for an invalid UTF-8
1099 string, the offset is that of the first byte of the failing character.
1100 Also, some errors are not detected until checks are carried out when
1101 the whole pattern has been scanned; in these cases the offset passed
1102 back is the length of the pattern.
1104 Note that the offset is in bytes, not characters, even in UTF-8 mode.
1105 It may sometimes point into the middle of a UTF-8 character.
1107 If pcre_compile2() is used instead of pcre_compile(), and the error-
1108 codeptr argument is not NULL, a non-zero error code number is returned
1109 via this argument in the event of an error. This is in addition to the
1110 textual error message. Error codes and messages are listed below.
1112 If the final argument, tableptr, is NULL, PCRE uses a default set of
1113 character tables that are built when PCRE is compiled, using the
1114 default C locale. Otherwise, tableptr must be an address that is the
1115 result of a call to pcre_maketables(). This value is stored with the
1116 compiled pattern, and used again by pcre_exec(), unless another table
1117 pointer is passed to it. For more discussion, see the section on locale
1118 support below.
1120 This code fragment shows a typical straightforward call to pcre_com-
1121 pile():
1123 pcre *re;
1124 const char *error;
1125 int erroffset;
1126 re = pcre_compile(
1127 "^A.*Z", /* the pattern */
1128 0, /* default options */
1129 &error, /* for error message */
1130 &erroffset, /* for error offset */
1131 NULL); /* use default character tables */
1133 The following names for option bits are defined in the pcre.h header
1134 file:
1138 If this bit is set, the pattern is forced to be "anchored", that is, it
1139 is constrained to match only at the first matching point in the string
1140 that is being searched (the "subject string"). This effect can also be
1141 achieved by appropriate constructs in the pattern itself, which is the
1142 only way to do it in Perl.
1146 If this bit is set, pcre_compile() automatically inserts callout items,
1147 all with number 255, before each pattern item. For discussion of the
1148 callout facility, see the pcrecallout documentation.
1153 These options (which are mutually exclusive) control what the \R escape
1154 sequence matches. The choice is either to match only CR, LF, or CRLF,
1155 or to match any Unicode newline sequence. The default is specified when
1156 PCRE is built. It can be overridden from within the pattern, or by set-
1157 ting an option when a compiled pattern is matched.
1161 If this bit is set, letters in the pattern match both upper and lower
1162 case letters. It is equivalent to Perl's /i option, and it can be
1163 changed within a pattern by a (?i) option setting. In UTF-8 mode, PCRE
1164 always understands the concept of case for characters whose values are
1165 less than 128, so caseless matching is always possible. For characters
1166 with higher values, the concept of case is supported if PCRE is com-
1167 piled with Unicode property support, but not otherwise. If you want to
1168 use caseless matching for characters 128 and above, you must ensure
1169 that PCRE is compiled with Unicode property support as well as with
1170 UTF-8 support.
1174 If this bit is set, a dollar metacharacter in the pattern matches only
1175 at the end of the subject string. Without this option, a dollar also
1176 matches immediately before a newline at the end of the string (but not
1177 before any other newlines). The PCRE_DOLLAR_ENDONLY option is ignored
1178 if PCRE_MULTILINE is set. There is no equivalent to this option in
1179 Perl, and no way to set it within a pattern.
1183 If this bit is set, a dot metacharacter in the pattern matches a char-
1184 acter of any value, including one that indicates a newline. However, it
1185 only ever matches one character, even if newlines are coded as CRLF.
1186 Without this option, a dot does not match when the current position is
1187 at a newline. This option is equivalent to Perl's /s option, and it can
1188 be changed within a pattern by a (?s) option setting. A negative class
1189 such as [^a] always matches newline characters, independent of the set-
1190 ting of this option.
1194 If this bit is set, names used to identify capturing subpatterns need
1195 not be unique. This can be helpful for certain types of pattern when it
1196 is known that only one instance of the named subpattern can ever be
1197 matched. There are more details of named subpatterns below; see also
1198 the pcrepattern documentation.
1202 If this bit is set, whitespace data characters in the pattern are
1203 totally ignored except when escaped or inside a character class. White-
1204 space does not include the VT character (code 11). In addition, charac-
1205 ters between an unescaped # outside a character class and the next new-
1206 line, inclusive, are also ignored. This is equivalent to Perl's /x
1207 option, and it can be changed within a pattern by a (?x) option set-
1208 ting.
1210 Which characters are interpreted as newlines is controlled by the
1211 options passed to pcre_compile() or by a special sequence at the start
1212 of the pattern, as described in the section entitled "Newline conven-
1213 tions" in the pcrepattern documentation. Note that the end of this type
1214 of comment is a literal newline sequence in the pattern; escape
1215 sequences that happen to represent a newline do not count.
1217 This option makes it possible to include comments inside complicated
1218 patterns. Note, however, that this applies only to data characters.
1219 Whitespace characters may never appear within special character
1220 sequences in a pattern, for example within the sequence (?( that intro-
1221 duces a conditional subpattern.
1225 This option was invented in order to turn on additional functionality
1226 of PCRE that is incompatible with Perl, but it is currently of very
1227 little use. When set, any backslash in a pattern that is followed by a
1228 letter that has no special meaning causes an error, thus reserving
1229 these combinations for future expansion. By default, as in Perl, a
1230 backslash followed by a letter with no special meaning is treated as a
1231 literal. (Perl can, however, be persuaded to give an error for this, by
1232 running it with the -w option.) There are at present no other features
1233 controlled by this option. It can also be set by a (?X) option setting
1234 within a pattern.
1238 If this option is set, an unanchored pattern is required to match
1239 before or at the first newline in the subject string, though the
1240 matched text may continue over the newline.
1244 If this option is set, PCRE's behaviour is changed in some ways so that
1245 it is compatible with JavaScript rather than Perl. The changes are as
1246 follows:
1248 (1) A lone closing square bracket in a pattern causes a compile-time
1249 error, because this is illegal in JavaScript (by default it is treated
1250 as a data character). Thus, the pattern AB]CD becomes illegal when this
1251 option is set.
1253 (2) At run time, a back reference to an unset subpattern group matches
1254 an empty string (by default this causes the current matching alterna-
1255 tive to fail). A pattern such as (\1)(a) succeeds when this option is
1256 set (assuming it can find an "a" in the subject), whereas it fails by
1257 default, for Perl compatibility.
1261 By default, PCRE treats the subject string as consisting of a single
1262 line of characters (even if it actually contains newlines). The "start
1263 of line" metacharacter (^) matches only at the start of the string,
1264 while the "end of line" metacharacter ($) matches only at the end of
1265 the string, or before a terminating newline (unless PCRE_DOLLAR_ENDONLY
1266 is set). This is the same as Perl.
1268 When PCRE_MULTILINE it is set, the "start of line" and "end of line"
1269 constructs match immediately following or immediately before internal
1270 newlines in the subject string, respectively, as well as at the very
1271 start and end. This is equivalent to Perl's /m option, and it can be
1272 changed within a pattern by a (?m) option setting. If there are no new-
1273 lines in a subject string, or no occurrences of ^ or $ in a pattern,
1274 setting PCRE_MULTILINE has no effect.
1282 These options override the default newline definition that was chosen
1283 when PCRE was built. Setting the first or the second specifies that a
1284 newline is indicated by a single character (CR or LF, respectively).
1285 Setting PCRE_NEWLINE_CRLF specifies that a newline is indicated by the
1286 two-character CRLF sequence. Setting PCRE_NEWLINE_ANYCRLF specifies
1287 that any of the three preceding sequences should be recognized. Setting
1288 PCRE_NEWLINE_ANY specifies that any Unicode newline sequence should be
1289 recognized. The Unicode newline sequences are the three just mentioned,
1290 plus the single characters VT (vertical tab, U+000B), FF (formfeed,
1291 U+000C), NEL (next line, U+0085), LS (line separator, U+2028), and PS
1292 (paragraph separator, U+2029). The last two are recognized only in
1293 UTF-8 mode.
1295 The newline setting in the options word uses three bits that are
1296 treated as a number, giving eight possibilities. Currently only six are
1297 used (default plus the five values above). This means that if you set
1298 more than one newline option, the combination may or may not be sensi-
1299 ble. For example, PCRE_NEWLINE_CR with PCRE_NEWLINE_LF is equivalent to
1300 PCRE_NEWLINE_CRLF, but other combinations may yield unused numbers and
1301 cause an error.
1303 The only time that a line break in a pattern is specially recognized
1304 when compiling is when PCRE_EXTENDED is set. CR and LF are whitespace
1305 characters, and so are ignored in this mode. Also, an unescaped # out-
1306 side a character class indicates a comment that lasts until after the
1307 next line break sequence. In other circumstances, line break sequences
1308 in patterns are treated as literal data.
1310 The newline option that is set at compile time becomes the default that
1311 is used for pcre_exec() and pcre_dfa_exec(), but it can be overridden.
1315 If this option is set, it disables the use of numbered capturing paren-
1316 theses in the pattern. Any opening parenthesis that is not followed by
1317 ? behaves as if it were followed by ?: but named parentheses can still
1318 be used for capturing (and they acquire numbers in the usual way).
1319 There is no equivalent of this option in Perl.
1323 This is an option that acts at matching time; that is, it is really an
1324 option for pcre_exec() or pcre_dfa_exec(). If it is set at compile
1325 time, it is remembered with the compiled pattern and assumed at match-
1326 ing time. For details see the discussion of PCRE_NO_START_OPTIMIZE
1327 below.
1331 This option changes the way PCRE processes \B, \b, \D, \d, \S, \s, \W,
1332 \w, and some of the POSIX character classes. By default, only ASCII
1333 characters are recognized, but if PCRE_UCP is set, Unicode properties
1334 are used instead to classify characters. More details are given in the
1335 section on generic character types in the pcrepattern page. If you set
1336 PCRE_UCP, matching one of the items it affects takes much longer. The
1337 option is available only if PCRE has been compiled with Unicode prop-
1338 erty support.
1342 This option inverts the "greediness" of the quantifiers so that they
1343 are not greedy by default, but become greedy if followed by "?". It is
1344 not compatible with Perl. It can also be set by a (?U) option setting
1345 within the pattern.
1347 PCRE_UTF8
1349 This option causes PCRE to regard both the pattern and the subject as
1350 strings of UTF-8 characters instead of single-byte character strings.
1351 However, it is available only when PCRE is built to include UTF-8 sup-
1352 port. If not, the use of this option provokes an error. Details of how
1353 this option changes the behaviour of PCRE are given in the pcreunicode
1354 page.
1358 When PCRE_UTF8 is set, the validity of the pattern as a UTF-8 string is
1359 automatically checked. There is a discussion about the validity of
1360 UTF-8 strings in the main pcre page. If an invalid UTF-8 sequence of
1361 bytes is found, pcre_compile() returns an error. If you already know
1362 that your pattern is valid, and you want to skip this check for perfor-
1363 mance reasons, you can set the PCRE_NO_UTF8_CHECK option. When it is
1364 set, the effect of passing an invalid UTF-8 string as a pattern is
1365 undefined. It may cause your program to crash. Note that this option
1366 can also be passed to pcre_exec() and pcre_dfa_exec(), to suppress the
1367 UTF-8 validity checking of subject strings.
1372 The following table lists the error codes than may be returned by
1373 pcre_compile2(), along with the error messages that may be returned by
1374 both compiling functions. As PCRE has developed, some error codes have
1375 fallen out of use. To avoid confusion, they have not been re-used.
1377 0 no error
1378 1 \ at end of pattern
1379 2 \c at end of pattern
1380 3 unrecognized character follows \
1381 4 numbers out of order in {} quantifier
1382 5 number too big in {} quantifier
1383 6 missing terminating ] for character class
1384 7 invalid escape sequence in character class
1385 8 range out of order in character class
1386 9 nothing to repeat
1387 10 [this code is not in use]
1388 11 internal error: unexpected repeat
1389 12 unrecognized character after (? or (?-
1390 13 POSIX named classes are supported only within a class
1391 14 missing )
1392 15 reference to non-existent subpattern
1393 16 erroffset passed as NULL
1394 17 unknown option bit(s) set
1395 18 missing ) after comment
1396 19 [this code is not in use]
1397 20 regular expression is too large
1398 21 failed to get memory
1399 22 unmatched parentheses
1400 23 internal error: code overflow
1401 24 unrecognized character after (?<
1402 25 lookbehind assertion is not fixed length
1403 26 malformed number or name after (?(
1404 27 conditional group contains more than two branches
1405 28 assertion expected after (?(
1406 29 (?R or (?[+-]digits must be followed by )
1407 30 unknown POSIX class name
1408 31 POSIX collating elements are not supported
1409 32 this version of PCRE is not compiled with PCRE_UTF8 support
1410 33 [this code is not in use]
1411 34 character value in \x{...} sequence is too large
1412 35 invalid condition (?(0)
1413 36 \C not allowed in lookbehind assertion
1414 37 PCRE does not support \L, \l, \N{name}, \U, or \u
1415 38 number after (?C is > 255
1416 39 closing ) for (?C expected
1417 40 recursive call could loop indefinitely
1418 41 unrecognized character after (?P
1419 42 syntax error in subpattern name (missing terminator)
1420 43 two named subpatterns have the same name
1421 44 invalid UTF-8 string
1422 45 support for \P, \p, and \X has not been compiled
1423 46 malformed \P or \p sequence
1424 47 unknown property name after \P or \p
1425 48 subpattern name is too long (maximum 32 characters)
1426 49 too many named subpatterns (maximum 10000)
1427 50 [this code is not in use]
1428 51 octal value is greater than \377 (not in UTF-8 mode)
1429 52 internal error: overran compiling workspace
1430 53 internal error: previously-checked referenced subpattern
1431 not found
1432 54 DEFINE group contains more than one branch
1433 55 repeating a DEFINE group is not allowed
1434 56 inconsistent NEWLINE options
1435 57 \g is not followed by a braced, angle-bracketed, or quoted
1436 name/number or by a plain number
1437 58 a numbered reference must not be zero
1438 59 an argument is not allowed for (*ACCEPT), (*FAIL), or (*COMMIT)
1439 60 (*VERB) not recognized
1440 61 number is too big
1441 62 subpattern name expected
1442 63 digit expected after (?+
1443 64 ] is an invalid data character in JavaScript compatibility mode
1444 65 different names for subpatterns of the same number are
1445 not allowed
1446 66 (*MARK) must have an argument
1447 67 this version of PCRE is not compiled with PCRE_UCP support
1448 68 \c must be followed by an ASCII character
1449 69 \k is not followed by a braced, angle-bracketed, or quoted name
1451 The numbers 32 and 10000 in errors 48 and 49 are defaults; different
1452 values may be used if the limits were changed when PCRE was built.
1457 pcre_extra *pcre_study(const pcre *code, int options
1458 const char **errptr);
1460 If a compiled pattern is going to be used several times, it is worth
1461 spending more time analyzing it in order to speed up the time taken for
1462 matching. The function pcre_study() takes a pointer to a compiled pat-
1463 tern as its first argument. If studying the pattern produces additional
1464 information that will help speed up matching, pcre_study() returns a
1465 pointer to a pcre_extra block, in which the study_data field points to
1466 the results of the study.
1468 The returned value from pcre_study() can be passed directly to
1469 pcre_exec() or pcre_dfa_exec(). However, a pcre_extra block also con-
1470 tains other fields that can be set by the caller before the block is
1471 passed; these are described below in the section on matching a pattern.
1473 If studying the pattern does not produce any useful information,
1474 pcre_study() returns NULL. In that circumstance, if the calling program
1475 wants to pass any of the other fields to pcre_exec() or
1476 pcre_dfa_exec(), it must set up its own pcre_extra block.
1478 The second argument of pcre_study() contains option bits. There is only
1479 one option: PCRE_STUDY_JIT_COMPILE. If this is set, and the just-in-
1480 time compiler is available, the pattern is further compiled into
1481 machine code that executes much faster than the pcre_exec() matching
1482 function. If the just-in-time compiler is not available, this option is
1483 ignored. All other bits in the options argument must be zero.
1485 JIT compilation is a heavyweight optimization. It can take some time
1486 for patterns to be analyzed, and for one-off matches and simple pat-
1487 terns the benefit of faster execution might be offset by a much slower
1488 study time. Not all patterns can be optimized by the JIT compiler. For
1489 those that cannot be handled, matching automatically falls back to the
1490 pcre_exec() interpreter. For more details, see the pcrejit documenta-
1491 tion.
1493 The third argument for pcre_study() is a pointer for an error message.
1494 If studying succeeds (even if no data is returned), the variable it
1495 points to is set to NULL. Otherwise it is set to point to a textual
1496 error message. This is a static string that is part of the library. You
1497 must not try to free it. You should test the error pointer for NULL
1498 after calling pcre_study(), to be sure that it has run successfully.
1500 When you are finished with a pattern, you can free the memory used for
1501 the study data by calling pcre_free_study(). This function was added to
1502 the API for release 8.20. For earlier versions, the memory could be
1503 freed with pcre_free(), just like the pattern itself. This will still
1504 work in cases where PCRE_STUDY_JIT_COMPILE is not used, but it is
1505 advisable to change to the new function when convenient.
1507 This is a typical way in which pcre_study() is used (except that in a
1508 real application there should be tests for errors):
1510 int rc;
1511 pcre *re;
1512 pcre_extra *sd;
1513 re = pcre_compile("pattern", 0, &error, &erroroffset, NULL);
1514 sd = pcre_study(
1515 re, /* result of pcre_compile() */
1516 0, /* no options */
1517 &error); /* set to NULL or points to a message */
1518 rc = pcre_exec( /* see below for details of pcre_exec() options */
1519 re, sd, "subject", 7, 0, 0, ovector, 30);
1520 ...
1521 pcre_free_study(sd);
1522 pcre_free(re);
1524 Studying a pattern does two things: first, a lower bound for the length
1525 of subject string that is needed to match the pattern is computed. This
1526 does not mean that there are any strings of that length that match, but
1527 it does guarantee that no shorter strings match. The value is used by
1528 pcre_exec() and pcre_dfa_exec() to avoid wasting time by trying to
1529 match strings that are shorter than the lower bound. You can find out
1530 the value in a calling program via the pcre_fullinfo() function.
1532 Studying a pattern is also useful for non-anchored patterns that do not
1533 have a single fixed starting character. A bitmap of possible starting
1534 bytes is created. This speeds up finding a position in the subject at
1535 which to start matching.
1537 These two optimizations apply to both pcre_exec() and pcre_dfa_exec().
1538 However, they are not used by pcre_exec() if pcre_study() is called
1539 with the PCRE_STUDY_JIT_COMPILE option, and just-in-time compiling is
1540 successful. The optimizations can be disabled by setting the
1541 PCRE_NO_START_OPTIMIZE option when calling pcre_exec() or
1542 pcre_dfa_exec(). You might want to do this if your pattern contains
1543 callouts or (*MARK) (which cannot be handled by the JIT compiler), and
1544 you want to make use of these facilities in cases where matching fails.
1545 See the discussion of PCRE_NO_START_OPTIMIZE below.
1550 PCRE handles caseless matching, and determines whether characters are
1551 letters, digits, or whatever, by reference to a set of tables, indexed
1552 by character value. When running in UTF-8 mode, this applies only to
1553 characters with codes less than 128. By default, higher-valued codes
1554 never match escapes such as \w or \d, but they can be tested with \p if
1555 PCRE is built with Unicode character property support. Alternatively,
1556 the PCRE_UCP option can be set at compile time; this causes \w and
1557 friends to use Unicode property support instead of built-in tables. The
1558 use of locales with Unicode is discouraged. If you are handling charac-
1559 ters with codes greater than 128, you should either use UTF-8 and Uni-
1560 code, or use locales, but not try to mix the two.
1562 PCRE contains an internal set of tables that are used when the final
1563 argument of pcre_compile() is NULL. These are sufficient for many
1564 applications. Normally, the internal tables recognize only ASCII char-
1565 acters. However, when PCRE is built, it is possible to cause the inter-
1566 nal tables to be rebuilt in the default "C" locale of the local system,
1567 which may cause them to be different.
1569 The internal tables can always be overridden by tables supplied by the
1570 application that calls PCRE. These may be created in a different locale
1571 from the default. As more and more applications change to using Uni-
1572 code, the need for this locale support is expected to die away.
1574 External tables are built by calling the pcre_maketables() function,
1575 which has no arguments, in the relevant locale. The result can then be
1576 passed to pcre_compile() or pcre_exec() as often as necessary. For
1577 example, to build and use tables that are appropriate for the French
1578 locale (where accented characters with values greater than 128 are
1579 treated as letters), the following code could be used:
1581 setlocale(LC_CTYPE, "fr_FR");
1582 tables = pcre_maketables();
1583 re = pcre_compile(..., tables);
1585 The locale name "fr_FR" is used on Linux and other Unix-like systems;
1586 if you are using Windows, the name for the French locale is "french".
1588 When pcre_maketables() runs, the tables are built in memory that is
1589 obtained via pcre_malloc. It is the caller's responsibility to ensure
1590 that the memory containing the tables remains available for as long as
1591 it is needed.
1593 The pointer that is passed to pcre_compile() is saved with the compiled
1594 pattern, and the same tables are used via this pointer by pcre_study()
1595 and normally also by pcre_exec(). Thus, by default, for any single pat-
1596 tern, compilation, studying and matching all happen in the same locale,
1597 but different patterns can be compiled in different locales.
1599 It is possible to pass a table pointer or NULL (indicating the use of
1600 the internal tables) to pcre_exec(). Although not intended for this
1601 purpose, this facility could be used to match a pattern in a different
1602 locale from the one in which it was compiled. Passing table pointers at
1603 run time is discussed below in the section on matching a pattern.
1608 int pcre_fullinfo(const pcre *code, const pcre_extra *extra,
1609 int what, void *where);
1611 The pcre_fullinfo() function returns information about a compiled pat-
1612 tern. It replaces the obsolete pcre_info() function, which is neverthe-
1613 less retained for backwards compability (and is documented below).
1615 The first argument for pcre_fullinfo() is a pointer to the compiled
1616 pattern. The second argument is the result of pcre_study(), or NULL if
1617 the pattern was not studied. The third argument specifies which piece
1618 of information is required, and the fourth argument is a pointer to a
1619 variable to receive the data. The yield of the function is zero for
1620 success, or one of the following negative numbers:
1622 PCRE_ERROR_NULL the argument code was NULL
1623 the argument where was NULL
1624 PCRE_ERROR_BADMAGIC the "magic number" was not found
1625 PCRE_ERROR_BADOPTION the value of what was invalid
1627 The "magic number" is placed at the start of each compiled pattern as
1628 an simple check against passing an arbitrary memory pointer. Here is a
1629 typical call of pcre_fullinfo(), to obtain the length of the compiled
1630 pattern:
1632 int rc;
1633 size_t length;
1634 rc = pcre_fullinfo(
1635 re, /* result of pcre_compile() */
1636 sd, /* result of pcre_study(), or NULL */
1637 PCRE_INFO_SIZE, /* what is required */
1638 &length); /* where to put the data */
1640 The possible values for the third argument are defined in pcre.h, and
1641 are as follows:
1645 Return the number of the highest back reference in the pattern. The
1646 fourth argument should point to an int variable. Zero is returned if
1647 there are no back references.
1651 Return the number of capturing subpatterns in the pattern. The fourth
1652 argument should point to an int variable.
1656 Return a pointer to the internal default character tables within PCRE.
1657 The fourth argument should point to an unsigned char * variable. This
1658 information call is provided for internal use by the pcre_study() func-
1659 tion. External callers can cause PCRE to use its internal tables by
1660 passing a NULL table pointer.
1664 Return information about the first byte of any matched string, for a
1665 non-anchored pattern. The fourth argument should point to an int vari-
1666 able. (This option used to be called PCRE_INFO_FIRSTCHAR; the old name
1667 is still recognized for backwards compatibility.)
1669 If there is a fixed first byte, for example, from a pattern such as
1670 (cat|cow|coyote), its value is returned. Otherwise, if either
1672 (a) the pattern was compiled with the PCRE_MULTILINE option, and every
1673 branch starts with "^", or
1675 (b) every branch of the pattern starts with ".*" and PCRE_DOTALL is not
1676 set (if it were set, the pattern would be anchored),
1678 -1 is returned, indicating that the pattern matches only at the start
1679 of a subject string or after any newline within the string. Otherwise
1680 -2 is returned. For anchored patterns, -2 is returned.
1684 If the pattern was studied, and this resulted in the construction of a
1685 256-bit table indicating a fixed set of bytes for the first byte in any
1686 matching string, a pointer to the table is returned. Otherwise NULL is
1687 returned. The fourth argument should point to an unsigned char * vari-
1688 able.
1692 Return 1 if the pattern contains any explicit matches for CR or LF
1693 characters, otherwise 0. The fourth argument should point to an int
1694 variable. An explicit match is either a literal CR or LF character, or
1695 \r or \n.
1699 Return 1 if the (?J) or (?-J) option setting is used in the pattern,
1700 otherwise 0. The fourth argument should point to an int variable. (?J)
1701 and (?-J) set and unset the local PCRE_DUPNAMES option, respectively.
1705 Return 1 if the pattern was studied with the PCRE_STUDY_JIT_COMPILE
1706 option, and just-in-time compiling was successful. The fourth argument
1707 should point to an int variable. A return value of 0 means that JIT
1708 support is not available in this version of PCRE, or that the pattern
1709 was not studied with the PCRE_STUDY_JIT_COMPILE option, or that the JIT
1710 compiler could not handle this particular pattern. See the pcrejit doc-
1711 umentation for details of what can and cannot be handled.
1715 Return the value of the rightmost literal byte that must exist in any
1716 matched string, other than at its start, if such a byte has been
1717 recorded. The fourth argument should point to an int variable. If there
1718 is no such byte, -1 is returned. For anchored patterns, a last literal
1719 byte is recorded only if it follows something of variable length. For
1720 example, for the pattern /^a\d+z\d+/ the returned value is "z", but for
1721 /^a\dz\d/ the returned value is -1.
1725 If the pattern was studied and a minimum length for matching subject
1726 strings was computed, its value is returned. Otherwise the returned
1727 value is -1. The value is a number of characters, not bytes (this may
1728 be relevant in UTF-8 mode). The fourth argument should point to an int
1729 variable. A non-negative value is a lower bound to the length of any
1730 matching string. There may not be any strings of that length that do
1731 actually match, but every string that does match is at least that long.
1737 PCRE supports the use of named as well as numbered capturing parenthe-
1738 ses. The names are just an additional way of identifying the parenthe-
1739 ses, which still acquire numbers. Several convenience functions such as
1740 pcre_get_named_substring() are provided for extracting captured sub-
1741 strings by name. It is also possible to extract the data directly, by
1742 first converting the name to a number in order to access the correct
1743 pointers in the output vector (described with pcre_exec() below). To do
1744 the conversion, you need to use the name-to-number map, which is
1745 described by these three values.
1747 The map consists of a number of fixed-size entries. PCRE_INFO_NAMECOUNT
1748 gives the number of entries, and PCRE_INFO_NAMEENTRYSIZE gives the size
1749 of each entry; both of these return an int value. The entry size
1750 depends on the length of the longest name. PCRE_INFO_NAMETABLE returns
1751 a pointer to the first entry of the table (a pointer to char). The
1752 first two bytes of each entry are the number of the capturing parenthe-
1753 sis, most significant byte first. The rest of the entry is the corre-
1754 sponding name, zero terminated.
1756 The names are in alphabetical order. Duplicate names may appear if (?|
1757 is used to create multiple groups with the same number, as described in
1758 the section on duplicate subpattern numbers in the pcrepattern page.
1759 Duplicate names for subpatterns with different numbers are permitted
1760 only if PCRE_DUPNAMES is set. In all cases of duplicate names, they
1761 appear in the table in the order in which they were found in the pat-
1762 tern. In the absence of (?| this is the order of increasing number;
1763 when (?| is used this is not necessarily the case because later subpat-
1764 terns may have lower numbers.
1766 As a simple example of the name/number table, consider the following
1767 pattern (assume PCRE_EXTENDED is set, so white space - including new-
1768 lines - is ignored):
1770 (?<date> (?<year>(\d\d)?\d\d) -
1771 (?<month>\d\d) - (?<day>\d\d) )
1773 There are four named subpatterns, so the table has four entries, and
1774 each entry in the table is eight bytes long. The table is as follows,
1775 with non-printing bytes shows in hexadecimal, and undefined bytes shown
1776 as ??:
1778 00 01 d a t e 00 ??
1779 00 05 d a y 00 ?? ??
1780 00 04 m o n t h 00
1781 00 02 y e a r 00 ??
1783 When writing code to extract data from named subpatterns using the
1784 name-to-number map, remember that the length of the entries is likely
1785 to be different for each compiled pattern.
1789 Return 1 if the pattern can be used for partial matching with
1790 pcre_exec(), otherwise 0. The fourth argument should point to an int
1791 variable. From release 8.00, this always returns 1, because the
1792 restrictions that previously applied to partial matching have been
1793 lifted. The pcrepartial documentation gives details of partial match-
1794 ing.
1798 Return a copy of the options with which the pattern was compiled. The
1799 fourth argument should point to an unsigned long int variable. These
1800 option bits are those specified in the call to pcre_compile(), modified
1801 by any top-level option settings at the start of the pattern itself. In
1802 other words, they are the options that will be in force when matching
1803 starts. For example, if the pattern /(?im)abc(?-i)d/ is compiled with
1804 the PCRE_EXTENDED option, the result is PCRE_CASELESS, PCRE_MULTILINE,
1807 A pattern is automatically anchored by PCRE if all of its top-level
1808 alternatives begin with one of the following:
1810 ^ unless PCRE_MULTILINE is set
1811 \A always
1812 \G always
1813 .* if PCRE_DOTALL is set and there are no back
1814 references to the subpattern in which .* appears
1816 For such patterns, the PCRE_ANCHORED bit is set in the options returned
1817 by pcre_fullinfo().
1821 Return the size of the compiled pattern, that is, the value that was
1822 passed as the argument to pcre_malloc() when PCRE was getting memory in
1823 which to place the compiled data. The fourth argument should point to a
1824 size_t variable.
1828 Return the size of the data block pointed to by the study_data field in
1829 a pcre_extra block. If pcre_extra is NULL, or there is no study data,
1830 zero is returned. The fourth argument should point to a size_t vari-
1831 able. The study_data field is set by pcre_study() to record informa-
1832 tion that will speed up matching (see the section entitled "Studying a
1833 pattern" above). The format of the study_data block is private, but its
1834 length is made available via this option so that it can be saved and
1835 restored (see the pcreprecompile documentation for details).
1840 int pcre_info(const pcre *code, int *optptr, int *firstcharptr);
1842 The pcre_info() function is now obsolete because its interface is too
1843 restrictive to return all the available data about a compiled pattern.
1844 New programs should use pcre_fullinfo() instead. The yield of
1845 pcre_info() is the number of capturing subpatterns, or one of the fol-
1846 lowing negative numbers:
1848 PCRE_ERROR_NULL the argument code was NULL
1849 PCRE_ERROR_BADMAGIC the "magic number" was not found
1851 If the optptr argument is not NULL, a copy of the options with which
1852 the pattern was compiled is placed in the integer it points to (see
1853 PCRE_INFO_OPTIONS above).
1855 If the pattern is not anchored and the firstcharptr argument is not
1856 NULL, it is used to pass back information about the first character of
1857 any matched string (see PCRE_INFO_FIRSTBYTE above).
1862 int pcre_refcount(pcre *code, int adjust);
1864 The pcre_refcount() function is used to maintain a reference count in
1865 the data block that contains a compiled pattern. It is provided for the
1866 benefit of applications that operate in an object-oriented manner,
1867 where different parts of the application may be using the same compiled
1868 pattern, but you want to free the block when they are all done.
1870 When a pattern is compiled, the reference count field is initialized to
1871 zero. It is changed only by calling this function, whose action is to
1872 add the adjust value (which may be positive or negative) to it. The
1873 yield of the function is the new value. However, the value of the count
1874 is constrained to lie between 0 and 65535, inclusive. If the new value
1875 is outside these limits, it is forced to the appropriate limit value.
1877 Except when it is zero, the reference count is not correctly preserved
1878 if a pattern is compiled on one host and then transferred to a host
1879 whose byte-order is different. (This seems a highly unlikely scenario.)
1884 int pcre_exec(const pcre *code, const pcre_extra *extra,
1885 const char *subject, int length, int startoffset,
1886 int options, int *ovector, int ovecsize);
1888 The function pcre_exec() is called to match a subject string against a
1889 compiled pattern, which is passed in the code argument. If the pattern
1890 was studied, the result of the study should be passed in the extra
1891 argument. You can call pcre_exec() with the same code and extra argu-
1892 ments as many times as you like, in order to match different subject
1893 strings with the same pattern.
1895 This function is the main matching facility of the library, and it
1896 operates in a Perl-like manner. For specialist use there is also an
1897 alternative matching function, which is described below in the section
1898 about the pcre_dfa_exec() function.
1900 In most applications, the pattern will have been compiled (and option-
1901 ally studied) in the same process that calls pcre_exec(). However, it
1902 is possible to save compiled patterns and study data, and then use them
1903 later in different processes, possibly even on different hosts. For a
1904 discussion about this, see the pcreprecompile documentation.
1906 Here is an example of a simple call to pcre_exec():
1908 int rc;
1909 int ovector[30];
1910 rc = pcre_exec(
1911 re, /* result of pcre_compile() */
1912 NULL, /* we didn't study the pattern */
1913 "some string", /* the subject string */
1914 11, /* the length of the subject string */
1915 0, /* start at offset 0 in the subject */
1916 0, /* default options */
1917 ovector, /* vector of integers for substring information */
1918 30); /* number of elements (NOT size in bytes) */
1920 Extra data for pcre_exec()
1922 If the extra argument is not NULL, it must point to a pcre_extra data
1923 block. The pcre_study() function returns such a block (when it doesn't
1924 return NULL), but you can also create one for yourself, and pass addi-
1925 tional information in it. The pcre_extra block contains the following
1926 fields (not necessarily in this order):
1928 unsigned long int flags;
1929 void *study_data;
1930 void *executable_jit;
1931 unsigned long int match_limit;
1932 unsigned long int match_limit_recursion;
1933 void *callout_data;
1934 const unsigned char *tables;
1935 unsigned char **mark;
1937 The flags field is a bitmap that specifies which of the other fields
1938 are set. The flag bits are:
1948 Other flag bits should be set to zero. The study_data field and some-
1949 times the executable_jit field are set in the pcre_extra block that is
1950 returned by pcre_study(), together with the appropriate flag bits. You
1951 should not set these yourself, but you may add to the block by setting
1952 the other fields and their corresponding flag bits.
1954 The match_limit field provides a means of preventing PCRE from using up
1955 a vast amount of resources when running patterns that are not going to
1956 match, but which have a very large number of possibilities in their
1957 search trees. The classic example is a pattern that uses nested unlim-
1958 ited repeats.
1960 Internally, pcre_exec() uses a function called match(), which it calls
1961 repeatedly (sometimes recursively). The limit set by match_limit is
1962 imposed on the number of times this function is called during a match,
1963 which has the effect of limiting the amount of backtracking that can
1964 take place. For patterns that are not anchored, the count restarts from
1965 zero for each position in the subject string.
1967 When pcre_exec() is called with a pattern that was successfully studied
1968 with the PCRE_STUDY_JIT_COMPILE option, the way that the matching is
1969 executed is entirely different. However, there is still the possibility
1970 of runaway matching that goes on for a very long time, and so the
1971 match_limit value is also used in this case (but in a different way) to
1972 limit how long the matching can continue.
1974 The default value for the limit can be set when PCRE is built; the
1975 default default is 10 million, which handles all but the most extreme
1976 cases. You can override the default by suppling pcre_exec() with a
1977 pcre_extra block in which match_limit is set, and
1978 PCRE_EXTRA_MATCH_LIMIT is set in the flags field. If the limit is
1979 exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
1981 The match_limit_recursion field is similar to match_limit, but instead
1982 of limiting the total number of times that match() is called, it limits
1983 the depth of recursion. The recursion depth is a smaller number than
1984 the total number of calls, because not all calls to match() are recur-
1985 sive. This limit is of use only if it is set smaller than match_limit.
1987 Limiting the recursion depth limits the amount of machine stack that
1988 can be used, or, when PCRE has been compiled to use memory on the heap
1989 instead of the stack, the amount of heap memory that can be used. This
1990 limit is not relevant, and is ignored, if the pattern was successfully
1991 studied with PCRE_STUDY_JIT_COMPILE.
1993 The default value for match_limit_recursion can be set when PCRE is
1994 built; the default default is the same value as the default for
1995 match_limit. You can override the default by suppling pcre_exec() with
1996 a pcre_extra block in which match_limit_recursion is set, and
1997 PCRE_EXTRA_MATCH_LIMIT_RECURSION is set in the flags field. If the
1998 limit is exceeded, pcre_exec() returns PCRE_ERROR_RECURSIONLIMIT.
2000 The callout_data field is used in conjunction with the "callout" fea-
2001 ture, and is described in the pcrecallout documentation.
2003 The tables field is used to pass a character tables pointer to
2004 pcre_exec(); this overrides the value that is stored with the compiled
2005 pattern. A non-NULL value is stored with the compiled pattern only if
2006 custom tables were supplied to pcre_compile() via its tableptr argu-
2007 ment. If NULL is passed to pcre_exec() using this mechanism, it forces
2008 PCRE's internal tables to be used. This facility is helpful when re-
2009 using patterns that have been saved after compiling with an external
2010 set of tables, because the external tables might be at a different
2011 address when pcre_exec() is called. See the pcreprecompile documenta-
2012 tion for a discussion of saving compiled patterns for later use.
2014 If PCRE_EXTRA_MARK is set in the flags field, the mark field must be
2015 set to point to a char * variable. If the pattern contains any back-
2016 tracking control verbs such as (*MARK:NAME), and the execution ends up
2017 with a name to pass back, a pointer to the name string (zero termi-
2018 nated) is placed in the variable pointed to by the mark field. The
2019 names are within the compiled pattern; if you wish to retain such a
2020 name you must copy it before freeing the memory of a compiled pattern.
2021 If there is no name to pass back, the variable pointed to by the mark
2022 field set to NULL. For details of the backtracking control verbs, see
2023 the section entitled "Backtracking control" in the pcrepattern documen-
2024 tation.
2026 Option bits for pcre_exec()
2028 The unused bits of the options argument for pcre_exec() must be zero.
2029 The only bits that may be set are PCRE_ANCHORED, PCRE_NEWLINE_xxx,
2034 If the pattern was successfully studied with the PCRE_STUDY_JIT_COMPILE
2035 option, the only supported options for JIT execution are
2037 PCRE_NOTEMPTY_ATSTART. Note in particular that partial matching is not
2038 supported. If an unsupported option is used, JIT execution is disabled
2039 and the normal interpretive code in pcre_exec() is run.
2043 The PCRE_ANCHORED option limits pcre_exec() to matching at the first
2044 matching position. If a pattern was compiled with PCRE_ANCHORED, or
2045 turned out to be anchored by virtue of its contents, it cannot be made
2046 unachored at matching time.
2051 These options (which are mutually exclusive) control what the \R escape
2052 sequence matches. The choice is either to match only CR, LF, or CRLF,
2053 or to match any Unicode newline sequence. These options override the
2054 choice that was made or defaulted when the pattern was compiled.
2062 These options override the newline definition that was chosen or
2063 defaulted when the pattern was compiled. For details, see the descrip-
2064 tion of pcre_compile() above. During matching, the newline choice
2065 affects the behaviour of the dot, circumflex, and dollar metacharac-
2066 ters. It may also alter the way the match position is advanced after a
2067 match failure for an unanchored pattern.
2070 set, and a match attempt for an unanchored pattern fails when the cur-
2071 rent position is at a CRLF sequence, and the pattern contains no
2072 explicit matches for CR or LF characters, the match position is
2073 advanced by two characters instead of one, in other words, to after the
2074 CRLF.
2076 The above rule is a compromise that makes the most common cases work as
2077 expected. For example, if the pattern is .+A (and the PCRE_DOTALL
2078 option is not set), it does not match the string "\r\nA" because, after
2079 failing at the start, it skips both the CR and the LF before retrying.
2080 However, the pattern [\r\n]A does match that string, because it con-
2081 tains an explicit CR or LF reference, and so advances only by one char-
2082 acter after the first failure.
2084 An explicit match for CR of LF is either a literal appearance of one of
2085 those characters, or one of the \r or \n escape sequences. Implicit
2086 matches such as [^X] do not count, nor does \s (which includes CR and
2087 LF in the characters that it matches).
2089 Notwithstanding the above, anomalous effects may still occur when CRLF
2090 is a valid newline sequence and explicit \r or \n escapes appear in the
2091 pattern.
2095 This option specifies that first character of the subject string is not
2096 the beginning of a line, so the circumflex metacharacter should not
2097 match before it. Setting this without PCRE_MULTILINE (at compile time)
2098 causes circumflex never to match. This option affects only the behav-
2099 iour of the circumflex metacharacter. It does not affect \A.
2103 This option specifies that the end of the subject string is not the end
2104 of a line, so the dollar metacharacter should not match it nor (except
2105 in multiline mode) a newline immediately before it. Setting this with-
2106 out PCRE_MULTILINE (at compile time) causes dollar never to match. This
2107 option affects only the behaviour of the dollar metacharacter. It does
2108 not affect \Z or \z.
2112 An empty string is not considered to be a valid match if this option is
2113 set. If there are alternatives in the pattern, they are tried. If all
2114 the alternatives match the empty string, the entire match fails. For
2115 example, if the pattern
2117 a?b?
2119 is applied to a string not beginning with "a" or "b", it matches an
2120 empty string at the start of the subject. With PCRE_NOTEMPTY set, this
2121 match is not valid, so PCRE searches further into the string for occur-
2122 rences of "a" or "b".
2126 This is like PCRE_NOTEMPTY, except that an empty string match that is
2127 not at the start of the subject is permitted. If the pattern is
2128 anchored, such a match can occur only if the pattern contains \K.
2130 Perl has no direct equivalent of PCRE_NOTEMPTY or
2131 PCRE_NOTEMPTY_ATSTART, but it does make a special case of a pattern
2132 match of the empty string within its split() function, and when using
2133 the /g modifier. It is possible to emulate Perl's behaviour after
2134 matching a null string by first trying the match again at the same off-
2135 set with PCRE_NOTEMPTY_ATSTART and PCRE_ANCHORED, and then if that
2136 fails, by advancing the starting offset (see below) and trying an ordi-
2137 nary match again. There is some code that demonstrates how to do this
2138 in the pcredemo sample program. In the most general case, you have to
2139 check to see if the newline convention recognizes CRLF as a newline,
2140 and if so, and the current character is CR followed by LF, advance the
2141 starting offset by two characters instead of one.
2145 There are a number of optimizations that pcre_exec() uses at the start
2146 of a match, in order to speed up the process. For example, if it is
2147 known that an unanchored match must start with a specific character, it
2148 searches the subject for that character, and fails immediately if it
2149 cannot find it, without actually running the main matching function.
2150 This means that a special item such as (*COMMIT) at the start of a pat-
2151 tern is not considered until after a suitable starting point for the
2152 match has been found. When callouts or (*MARK) items are in use, these
2153 "start-up" optimizations can cause them to be skipped if the pattern is
2154 never actually used. The start-up optimizations are in effect a pre-
2155 scan of the subject that takes place before the pattern is run.
2157 The PCRE_NO_START_OPTIMIZE option disables the start-up optimizations,
2158 possibly causing performance to suffer, but ensuring that in cases
2159 where the result is "no match", the callouts do occur, and that items
2160 such as (*COMMIT) and (*MARK) are considered at every possible starting
2161 position in the subject string. If PCRE_NO_START_OPTIMIZE is set at
2162 compile time, it cannot be unset at matching time.
2164 Setting PCRE_NO_START_OPTIMIZE can change the outcome of a matching
2165 operation. Consider the pattern
2169 When this is compiled, PCRE records the fact that a match must start
2170 with the character "A". Suppose the subject string is "DEFABC". The
2171 start-up optimization scans along the subject, finds "A" and runs the
2172 first match attempt from there. The (*COMMIT) item means that the pat-
2173 tern must match the current starting position, which in this case, it
2174 does. However, if the same match is run with PCRE_NO_START_OPTIMIZE
2175 set, the initial scan along the subject string does not happen. The
2176 first match attempt is run starting from "D" and when this fails,
2177 (*COMMIT) prevents any further matches being tried, so the overall
2178 result is "no match". If the pattern is studied, more start-up opti-
2179 mizations may be used. For example, a minimum length for the subject
2180 may be recorded. Consider the pattern
2182 (*MARK:A)(X|Y)
2184 The minimum length for a match is one character. If the subject is
2185 "ABC", there will be attempts to match "ABC", "BC", "C", and then
2186 finally an empty string. If the pattern is studied, the final attempt
2187 does not take place, because PCRE knows that the subject is too short,
2188 and so the (*MARK) is never encountered. In this case, studying the
2189 pattern does not affect the overall match result, which is still "no
2190 match", but it does affect the auxiliary information that is returned.
2194 When PCRE_UTF8 is set at compile time, the validity of the subject as a
2195 UTF-8 string is automatically checked when pcre_exec() is subsequently
2196 called. The value of startoffset is also checked to ensure that it
2197 points to the start of a UTF-8 character. There is a discussion about
2198 the validity of UTF-8 strings in the section on UTF-8 support in the
2199 main pcre page. If an invalid UTF-8 sequence of bytes is found,
2200 pcre_exec() returns the error PCRE_ERROR_BADUTF8 or, if PCRE_PAR-
2201 TIAL_HARD is set and the problem is a truncated UTF-8 character at the
2202 end of the subject, PCRE_ERROR_SHORTUTF8. In both cases, information
2203 about the precise nature of the error may also be returned (see the
2204 descriptions of these errors in the section entitled Error return val-
2205 ues from pcre_exec() below). If startoffset contains a value that does
2206 not point to the start of a UTF-8 character (or to the end of the sub-
2207 ject), PCRE_ERROR_BADUTF8_OFFSET is returned.
2209 If you already know that your subject is valid, and you want to skip
2210 these checks for performance reasons, you can set the
2211 PCRE_NO_UTF8_CHECK option when calling pcre_exec(). You might want to
2212 do this for the second and subsequent calls to pcre_exec() if you are
2213 making repeated calls to find all the matches in a single subject
2214 string. However, you should be sure that the value of startoffset
2215 points to the start of a UTF-8 character (or the end of the subject).
2216 When PCRE_NO_UTF8_CHECK is set, the effect of passing an invalid UTF-8
2217 string as a subject or an invalid value of startoffset is undefined.
2218 Your program may crash.
2223 These options turn on the partial matching feature. For backwards com-
2224 patibility, PCRE_PARTIAL is a synonym for PCRE_PARTIAL_SOFT. A partial
2225 match occurs if the end of the subject string is reached successfully,
2226 but there are not enough subject characters to complete the match. If
2227 this happens when PCRE_PARTIAL_SOFT (but not PCRE_PARTIAL_HARD) is set,
2228 matching continues by testing any remaining alternatives. Only if no
2229 complete match can be found is PCRE_ERROR_PARTIAL returned instead of
2230 PCRE_ERROR_NOMATCH. In other words, PCRE_PARTIAL_SOFT says that the
2231 caller is prepared to handle a partial match, but only if no complete
2232 match can be found.
2234 If PCRE_PARTIAL_HARD is set, it overrides PCRE_PARTIAL_SOFT. In this
2235 case, if a partial match is found, pcre_exec() immediately returns
2236 PCRE_ERROR_PARTIAL, without considering any other alternatives. In
2237 other words, when PCRE_PARTIAL_HARD is set, a partial match is consid-
2238 ered to be more important that an alternative complete match.
2240 In both cases, the portion of the string that was inspected when the
2241 partial match was found is set as the first matching string. There is a
2242 more detailed discussion of partial and multi-segment matching, with
2243 examples, in the pcrepartial documentation.
2245 The string to be matched by pcre_exec()
2247 The subject string is passed to pcre_exec() as a pointer in subject, a
2248 length (in bytes) in length, and a starting byte offset in startoffset.
2249 If this is negative or greater than the length of the subject,
2250 pcre_exec() returns PCRE_ERROR_BADOFFSET. When the starting offset is
2251 zero, the search for a match starts at the beginning of the subject,
2252 and this is by far the most common case. In UTF-8 mode, the byte offset
2253 must point to the start of a UTF-8 character (or the end of the sub-
2254 ject). Unlike the pattern string, the subject may contain binary zero
2255 bytes.
2257 A non-zero starting offset is useful when searching for another match
2258 in the same subject by calling pcre_exec() again after a previous suc-
2259 cess. Setting startoffset differs from just passing over a shortened
2260 string and setting PCRE_NOTBOL in the case of a pattern that begins
2261 with any kind of lookbehind. For example, consider the pattern
2263 \Biss\B
2265 which finds occurrences of "iss" in the middle of words. (\B matches
2266 only if the current position in the subject is not a word boundary.)
2267 When applied to the string "Mississipi" the first call to pcre_exec()
2268 finds the first occurrence. If pcre_exec() is called again with just
2269 the remainder of the subject, namely "issipi", it does not match,
2270 because \B is always false at the start of the subject, which is deemed
2271 to be a word boundary. However, if pcre_exec() is passed the entire
2272 string again, but with startoffset set to 4, it finds the second occur-
2273 rence of "iss" because it is able to look behind the starting point to
2274 discover that it is preceded by a letter.
2276 Finding all the matches in a subject is tricky when the pattern can
2277 match an empty string. It is possible to emulate Perl's /g behaviour by
2278 first trying the match again at the same offset, with the
2279 PCRE_NOTEMPTY_ATSTART and PCRE_ANCHORED options, and then if that
2280 fails, advancing the starting offset and trying an ordinary match
2281 again. There is some code that demonstrates how to do this in the pcre-
2282 demo sample program. In the most general case, you have to check to see
2283 if the newline convention recognizes CRLF as a newline, and if so, and
2284 the current character is CR followed by LF, advance the starting offset
2285 by two characters instead of one.
2287 If a non-zero starting offset is passed when the pattern is anchored,
2288 one attempt to match at the given offset is made. This can only succeed
2289 if the pattern does not require the match to be at the start of the
2290 subject.
2292 How pcre_exec() returns captured substrings
2294 In general, a pattern matches a certain portion of the subject, and in
2295 addition, further substrings from the subject may be picked out by
2296 parts of the pattern. Following the usage in Jeffrey Friedl's book,
2297 this is called "capturing" in what follows, and the phrase "capturing
2298 subpattern" is used for a fragment of a pattern that picks out a sub-
2299 string. PCRE supports several other kinds of parenthesized subpattern
2300 that do not cause substrings to be captured.
2302 Captured substrings are returned to the caller via a vector of integers
2303 whose address is passed in ovector. The number of elements in the vec-
2304 tor is passed in ovecsize, which must be a non-negative number. Note:
2305 this argument is NOT the size of ovector in bytes.
2307 The first two-thirds of the vector is used to pass back captured sub-
2308 strings, each substring using a pair of integers. The remaining third
2309 of the vector is used as workspace by pcre_exec() while matching cap-
2310 turing subpatterns, and is not available for passing back information.
2311 The number passed in ovecsize should always be a multiple of three. If
2312 it is not, it is rounded down.
2314 When a match is successful, information about captured substrings is
2315 returned in pairs of integers, starting at the beginning of ovector,
2316 and continuing up to two-thirds of its length at the most. The first
2317 element of each pair is set to the byte offset of the first character
2318 in a substring, and the second is set to the byte offset of the first
2319 character after the end of a substring. Note: these values are always
2320 byte offsets, even in UTF-8 mode. They are not character counts.
2322 The first pair of integers, ovector[0] and ovector[1], identify the
2323 portion of the subject string matched by the entire pattern. The next
2324 pair is used for the first capturing subpattern, and so on. The value
2325 returned by pcre_exec() is one more than the highest numbered pair that
2326 has been set. For example, if two substrings have been captured, the
2327 returned value is 3. If there are no capturing subpatterns, the return
2328 value from a successful match is 1, indicating that just the first pair
2329 of offsets has been set.
2331 If a capturing subpattern is matched repeatedly, it is the last portion
2332 of the string that it matched that is returned.
2334 If the vector is too small to hold all the captured substring offsets,
2335 it is used as far as possible (up to two-thirds of its length), and the
2336 function returns a value of zero. If neither the actual string matched
2337 not any captured substrings are of interest, pcre_exec() may be called
2338 with ovector passed as NULL and ovecsize as zero. However, if the pat-
2339 tern contains back references and the ovector is not big enough to
2340 remember the related substrings, PCRE has to get additional memory for
2341 use during matching. Thus it is usually advisable to supply an ovector
2342 of reasonable size.
2344 There are some cases where zero is returned (indicating vector over-
2345 flow) when in fact the vector is exactly the right size for the final
2346 match. For example, consider the pattern
2348 (a)(?:(b)c|bd)
2350 If a vector of 6 elements (allowing for only 1 captured substring) is
2351 given with subject string "abd", pcre_exec() will try to set the second
2352 captured string, thereby recording a vector overflow, before failing to
2353 match "c" and backing up to try the second alternative. The zero
2354 return, however, does correctly indicate that the maximum number of
2355 slots (namely 2) have been filled. In similar cases where there is tem-
2356 porary overflow, but the final number of used slots is actually less
2357 than the maximum, a non-zero value is returned.
2359 The pcre_fullinfo() function can be used to find out how many capturing
2360 subpatterns there are in a compiled pattern. The smallest size for
2361 ovector that will allow for n captured substrings, in addition to the
2362 offsets of the substring matched by the whole pattern, is (n+1)*3.
2364 It is possible for capturing subpattern number n+1 to match some part
2365 of the subject when subpattern n has not been used at all. For example,
2366 if the string "abc" is matched against the pattern (a|(z))(bc) the
2367 return from the function is 4, and subpatterns 1 and 3 are matched, but
2368 2 is not. When this happens, both values in the offset pairs corre-
2369 sponding to unused subpatterns are set to -1.
2371 Offset values that correspond to unused subpatterns at the end of the
2372 expression are also set to -1. For example, if the string "abc" is
2373 matched against the pattern (abc)(x(yz)?)? subpatterns 2 and 3 are not
2374 matched. The return from the function is 2, because the highest used
2375 capturing subpattern number is 1, and the offsets for for the second
2376 and third capturing subpatterns (assuming the vector is large enough,
2377 of course) are set to -1.
2379 Note: Elements in the first two-thirds of ovector that do not corre-
2380 spond to capturing parentheses in the pattern are never changed. That
2381 is, if a pattern contains n capturing parentheses, no more than ovec-
2382 tor[0] to ovector[2n+1] are set by pcre_exec(). The other elements (in
2383 the first two-thirds) retain whatever values they previously had.
2385 Some convenience functions are provided for extracting the captured
2386 substrings as separate strings. These are described below.
2388 Error return values from pcre_exec()
2390 If pcre_exec() fails, it returns a negative number. The following are
2391 defined in the header file:
2395 The subject string did not match the pattern.
2399 Either code or subject was passed as NULL, or ovector was NULL and
2400 ovecsize was not zero.
2404 An unrecognized bit was set in the options argument.
2408 PCRE stores a 4-byte "magic number" at the start of the compiled code,
2409 to catch the case when it is passed a junk pointer and to detect when a
2410 pattern that was compiled in an environment of one endianness is run in
2411 an environment with the other endianness. This is the error that PCRE
2412 gives when the magic number is not present.
2416 While running the pattern match, an unknown item was encountered in the
2417 compiled pattern. This error could be caused by a bug in PCRE or by
2418 overwriting of the compiled pattern.
2422 If a pattern contains back references, but the ovector that is passed
2423 to pcre_exec() is not big enough to remember the referenced substrings,
2424 PCRE gets a block of memory at the start of matching to use for this
2425 purpose. If the call via pcre_malloc() fails, this error is given. The
2426 memory is automatically freed at the end of matching.
2428 This error is also given if pcre_stack_malloc() fails in pcre_exec().
2429 This can happen only when PCRE has been compiled with --disable-stack-
2430 for-recursion.
2434 This error is used by the pcre_copy_substring(), pcre_get_substring(),
2435 and pcre_get_substring_list() functions (see below). It is never
2436 returned by pcre_exec().
2440 The backtracking limit, as specified by the match_limit field in a
2441 pcre_extra structure (or defaulted) was reached. See the description
2442 above.
2446 This error is never generated by pcre_exec() itself. It is provided for
2447 use by callout functions that want to yield a distinctive error code.
2448 See the pcrecallout documentation for details.
2452 A string that contains an invalid UTF-8 byte sequence was passed as a
2453 subject, and the PCRE_NO_UTF8_CHECK option was not set. If the size of
2454 the output vector (ovecsize) is at least 2, the byte offset to the
2455 start of the the invalid UTF-8 character is placed in the first ele-
2456 ment, and a reason code is placed in the second element. The reason
2457 codes are listed in the following section. For backward compatibility,
2458 if PCRE_PARTIAL_HARD is set and the problem is a truncated UTF-8 char-
2459 acter at the end of the subject (reason codes 1 to 5),
2460 PCRE_ERROR_SHORTUTF8 is returned instead of PCRE_ERROR_BADUTF8.
2464 The UTF-8 byte sequence that was passed as a subject was checked and
2465 found to be valid (the PCRE_NO_UTF8_CHECK option was not set), but the
2466 value of startoffset did not point to the beginning of a UTF-8 charac-
2467 ter or the end of the subject.
2471 The subject string did not match, but it did match partially. See the
2472 pcrepartial documentation for details of partial matching.
2476 This code is no longer in use. It was formerly returned when the
2477 PCRE_PARTIAL option was used with a compiled pattern containing items
2478 that were not supported for partial matching. From release 8.00
2479 onwards, there are no restrictions on partial matching.
2483 An unexpected internal error has occurred. This error could be caused
2484 by a bug in PCRE or by overwriting of the compiled pattern.
2488 This error is given if the value of the ovecsize argument is negative.
2492 The internal recursion limit, as specified by the match_limit_recursion
2493 field in a pcre_extra structure (or defaulted) was reached. See the
2494 description above.
2498 An invalid combination of PCRE_NEWLINE_xxx options was given.
2502 The value of startoffset was negative or greater than the length of the
2503 subject, that is, the value in length.
2507 This error is returned instead of PCRE_ERROR_BADUTF8 when the subject
2508 string ends with a truncated UTF-8 character and the PCRE_PARTIAL_HARD
2509 option is set. Information about the failure is returned as for
2510 PCRE_ERROR_BADUTF8. It is in fact sufficient to detect this case, but
2511 this special error code for PCRE_PARTIAL_HARD precedes the implementa-
2512 tion of returned information; it is retained for backwards compatibil-
2513 ity.
2517 This error is returned when pcre_exec() detects a recursion loop within
2518 the pattern. Specifically, it means that either the whole pattern or a
2519 subpattern has been called recursively for the second time at the same
2520 position in the subject string. Some simple patterns that might do this
2521 are detected and faulted at compile time, but more complicated cases,
2522 in particular mutual recursions between two different subpatterns, can-
2523 not be detected until run time.
2527 This error is returned when a pattern that was successfully studied
2528 using the PCRE_STUDY_JIT_COMPILE option is being matched, but the mem-
2529 ory available for the just-in-time processing stack is not large
2530 enough. See the pcrejit documentation for more details.
2532 Error numbers -16 to -20 and -22 are not used by pcre_exec().
2534 Reason codes for invalid UTF-8 strings
2536 When pcre_exec() returns either PCRE_ERROR_BADUTF8 or PCRE_ERROR_SHORT-
2537 UTF8, and the size of the output vector (ovecsize) is at least 2, the
2538 offset of the start of the invalid UTF-8 character is placed in the
2539 first output vector element (ovector[0]) and a reason code is placed in
2540 the second element (ovector[1]). The reason codes are given names in
2541 the pcre.h header file:
2549 The string ends with a truncated UTF-8 character; the code specifies
2550 how many bytes are missing (1 to 5). Although RFC 3629 restricts UTF-8
2551 characters to be no longer than 4 bytes, the encoding scheme (origi-
2552 nally defined by RFC 2279) allows for up to 6 bytes, and this is
2553 checked first; hence the possibility of 4 or 5 missing bytes.
2559 PCRE_UTF8_ERR10
2561 The two most significant bits of the 2nd, 3rd, 4th, 5th, or 6th byte of
2562 the character do not have the binary value 0b10 (that is, either the
2563 most significant bit is 0, or the next bit is 1).
2565 PCRE_UTF8_ERR11
2566 PCRE_UTF8_ERR12
2568 A character that is valid by the RFC 2279 rules is either 5 or 6 bytes
2569 long; these code points are excluded by RFC 3629.
2571 PCRE_UTF8_ERR13
2573 A 4-byte character has a value greater than 0x10fff; these code points
2574 are excluded by RFC 3629.
2576 PCRE_UTF8_ERR14
2578 A 3-byte character has a value in the range 0xd800 to 0xdfff; this
2579 range of code points are reserved by RFC 3629 for use with UTF-16, and
2580 so are excluded from UTF-8.
2582 PCRE_UTF8_ERR15
2583 PCRE_UTF8_ERR16
2584 PCRE_UTF8_ERR17
2585 PCRE_UTF8_ERR18
2586 PCRE_UTF8_ERR19
2588 A 2-, 3-, 4-, 5-, or 6-byte character is "overlong", that is, it codes
2589 for a value that can be represented by fewer bytes, which is invalid.
2590 For example, the two bytes 0xc0, 0xae give the value 0x2e, whose cor-
2591 rect coding uses just one byte.
2593 PCRE_UTF8_ERR20
2595 The two most significant bits of the first byte of a character have the
2596 binary value 0b10 (that is, the most significant bit is 1 and the sec-
2597 ond is 0). Such a byte can only validly occur as the second or subse-
2598 quent byte of a multi-byte character.
2600 PCRE_UTF8_ERR21
2602 The first byte of a character has the value 0xfe or 0xff. These values
2603 can never occur in a valid UTF-8 string.
2608 int pcre_copy_substring(const char *subject, int *ovector,
2609 int stringcount, int stringnumber, char *buffer,
2610 int buffersize);
2612 int pcre_get_substring(const char *subject, int *ovector,
2613 int stringcount, int stringnumber,
2614 const char **stringptr);
2616 int pcre_get_substring_list(const char *subject,
2617 int *ovector, int stringcount, const char ***listptr);
2619 Captured substrings can be accessed directly by using the offsets
2620 returned by pcre_exec() in ovector. For convenience, the functions
2621 pcre_copy_substring(), pcre_get_substring(), and pcre_get_sub-
2622 string_list() are provided for extracting captured substrings as new,
2623 separate, zero-terminated strings. These functions identify substrings
2624 by number. The next section describes functions for extracting named
2625 substrings.
2627 A substring that contains a binary zero is correctly extracted and has
2628 a further zero added on the end, but the result is not, of course, a C
2629 string. However, you can process such a string by referring to the
2630 length that is returned by pcre_copy_substring() and pcre_get_sub-
2631 string(). Unfortunately, the interface to pcre_get_substring_list() is
2632 not adequate for handling strings containing binary zeros, because the
2633 end of the final string is not independently indicated.
2635 The first three arguments are the same for all three of these func-
2636 tions: subject is the subject string that has just been successfully
2637 matched, ovector is a pointer to the vector of integer offsets that was
2638 passed to pcre_exec(), and stringcount is the number of substrings that
2639 were captured by the match, including the substring that matched the
2640 entire regular expression. This is the value returned by pcre_exec() if
2641 it is greater than zero. If pcre_exec() returned zero, indicating that
2642 it ran out of space in ovector, the value passed as stringcount should
2643 be the number of elements in the vector divided by three.
2645 The functions pcre_copy_substring() and pcre_get_substring() extract a
2646 single substring, whose number is given as stringnumber. A value of
2647 zero extracts the substring that matched the entire pattern, whereas
2648 higher values extract the captured substrings. For pcre_copy_sub-
2649 string(), the string is placed in buffer, whose length is given by
2650 buffersize, while for pcre_get_substring() a new block of memory is
2651 obtained via pcre_malloc, and its address is returned via stringptr.
2652 The yield of the function is the length of the string, not including
2653 the terminating zero, or one of these error codes:
2657 The buffer was too small for pcre_copy_substring(), or the attempt to
2658 get memory failed for pcre_get_substring().
2662 There is no substring whose number is stringnumber.
2664 The pcre_get_substring_list() function extracts all available sub-
2665 strings and builds a list of pointers to them. All this is done in a
2666 single block of memory that is obtained via pcre_malloc. The address of
2667 the memory block is returned via listptr, which is also the start of
2668 the list of string pointers. The end of the list is marked by a NULL
2669 pointer. The yield of the function is zero if all went well, or the
2670 error code
2674 if the attempt to get the memory block failed.
2676 When any of these functions encounter a substring that is unset, which
2677 can happen when capturing subpattern number n+1 matches some part of
2678 the subject, but subpattern n has not been used at all, they return an
2679 empty string. This can be distinguished from a genuine zero-length sub-
2680 string by inspecting the appropriate offset in ovector, which is nega-
2681 tive for unset substrings.
2683 The two convenience functions pcre_free_substring() and pcre_free_sub-
2684 string_list() can be used to free the memory returned by a previous
2685 call of pcre_get_substring() or pcre_get_substring_list(), respec-
2686 tively. They do nothing more than call the function pointed to by
2687 pcre_free, which of course could be called directly from a C program.
2688 However, PCRE is used in some situations where it is linked via a spe-
2689 cial interface to another programming language that cannot use
2690 pcre_free directly; it is for these cases that the functions are pro-
2691 vided.
2696 int pcre_get_stringnumber(const pcre *code,
2697 const char *name);
2699 int pcre_copy_named_substring(const pcre *code,
2700 const char *subject, int *ovector,
2701 int stringcount, const char *stringname,
2702 char *buffer, int buffersize);
2704 int pcre_get_named_substring(const pcre *code,
2705 const char *subject, int *ovector,
2706 int stringcount, const char *stringname,
2707 const char **stringptr);
2709 To extract a substring by name, you first have to find associated num-
2710 ber. For example, for this pattern
2712 (a+)b(?<xxx>\d+)...
2714 the number of the subpattern called "xxx" is 2. If the name is known to
2715 be unique (PCRE_DUPNAMES was not set), you can find the number from the
2716 name by calling pcre_get_stringnumber(). The first argument is the com-
2717 piled pattern, and the second is the name. The yield of the function is
2718 the subpattern number, or PCRE_ERROR_NOSUBSTRING (-7) if there is no
2719 subpattern of that name.
2721 Given the number, you can extract the substring directly, or use one of
2722 the functions described in the previous section. For convenience, there
2723 are also two functions that do the whole job.
2725 Most of the arguments of pcre_copy_named_substring() and
2726 pcre_get_named_substring() are the same as those for the similarly
2727 named functions that extract by number. As these are described in the
2728 previous section, they are not re-described here. There are just two
2729 differences:
2731 First, instead of a substring number, a substring name is given. Sec-
2732 ond, there is an extra argument, given at the start, which is a pointer
2733 to the compiled pattern. This is needed in order to gain access to the
2734 name-to-number translation table.
2736 These functions call pcre_get_stringnumber(), and if it succeeds, they
2737 then call pcre_copy_substring() or pcre_get_substring(), as appropri-
2738 ate. NOTE: If PCRE_DUPNAMES is set and there are duplicate names, the
2739 behaviour may not be what you want (see the next section).
2741 Warning: If the pattern uses the (?| feature to set up multiple subpat-
2742 terns with the same number, as described in the section on duplicate
2743 subpattern numbers in the pcrepattern page, you cannot use names to
2744 distinguish the different subpatterns, because names are not included
2745 in the compiled code. The matching process uses only numbers. For this
2746 reason, the use of different names for subpatterns of the same number
2747 causes an error at compile time.
2752 int pcre_get_stringtable_entries(const pcre *code,
2753 const char *name, char **first, char **last);
2755 When a pattern is compiled with the PCRE_DUPNAMES option, names for
2756 subpatterns are not required to be unique. (Duplicate names are always
2757 allowed for subpatterns with the same number, created by using the (?|
2758 feature. Indeed, if such subpatterns are named, they are required to
2759 use the same names.)
2761 Normally, patterns with duplicate names are such that in any one match,
2762 only one of the named subpatterns participates. An example is shown in
2763 the pcrepattern documentation.
2765 When duplicates are present, pcre_copy_named_substring() and
2766 pcre_get_named_substring() return the first substring corresponding to
2767 the given name that is set. If none are set, PCRE_ERROR_NOSUBSTRING
2768 (-7) is returned; no data is returned. The pcre_get_stringnumber()
2769 function returns one of the numbers that are associated with the name,
2770 but it is not defined which it is.
2772 If you want to get full details of all captured substrings for a given
2773 name, you must use the pcre_get_stringtable_entries() function. The
2774 first argument is the compiled pattern, and the second is the name. The
2775 third and fourth are pointers to variables which are updated by the
2776 function. After it has run, they point to the first and last entries in
2777 the name-to-number table for the given name. The function itself
2778 returns the length of each entry, or PCRE_ERROR_NOSUBSTRING (-7) if
2779 there are none. The format of the table is described above in the sec-
2780 tion entitled Information about a pattern above. Given all the rele-
2781 vant entries for the name, you can extract each of their numbers, and
2782 hence the captured data, if any.
2787 The traditional matching function uses a similar algorithm to Perl,
2788 which stops when it finds the first match, starting at a given point in
2789 the subject. If you want to find all possible matches, or the longest
2790 possible match, consider using the alternative matching function (see
2791 below) instead. If you cannot use the alternative function, but still
2792 need to find all possible matches, you can kludge it up by making use
2793 of the callout facility, which is described in the pcrecallout documen-
2794 tation.
2796 What you have to do is to insert a callout right at the end of the pat-
2797 tern. When your callout function is called, extract and save the cur-
2798 rent matched substring. Then return 1, which forces pcre_exec() to
2799 backtrack and try other alternatives. Ultimately, when it runs out of
2800 matches, pcre_exec() will yield PCRE_ERROR_NOMATCH.
2805 int pcre_dfa_exec(const pcre *code, const pcre_extra *extra,
2806 const char *subject, int length, int startoffset,
2807 int options, int *ovector, int ovecsize,
2808 int *workspace, int wscount);
2810 The function pcre_dfa_exec() is called to match a subject string
2811 against a compiled pattern, using a matching algorithm that scans the
2812 subject string just once, and does not backtrack. This has different
2813 characteristics to the normal algorithm, and is not compatible with
2814 Perl. Some of the features of PCRE patterns are not supported. Never-
2815 theless, there are times when this kind of matching can be useful. For
2816 a discussion of the two matching algorithms, and a list of features
2817 that pcre_dfa_exec() does not support, see the pcrematching documenta-
2818 tion.
2820 The arguments for the pcre_dfa_exec() function are the same as for
2821 pcre_exec(), plus two extras. The ovector argument is used in a differ-
2822 ent way, and this is described below. The other common arguments are
2823 used in the same way as for pcre_exec(), so their description is not
2824 repeated here.
2826 The two additional arguments provide workspace for the function. The
2827 workspace vector should contain at least 20 elements. It is used for
2828 keeping track of multiple paths through the pattern tree. More
2829 workspace will be needed for patterns and subjects where there are a
2830 lot of potential matches.
2832 Here is an example of a simple call to pcre_dfa_exec():
2834 int rc;
2835 int ovector[10];
2836 int wspace[20];
2837 rc = pcre_dfa_exec(
2838 re, /* result of pcre_compile() */
2839 NULL, /* we didn't study the pattern */
2840 "some string", /* the subject string */
2841 11, /* the length of the subject string */
2842 0, /* start at offset 0 in the subject */
2843 0, /* default options */
2844 ovector, /* vector of integers for substring information */
2845 10, /* number of elements (NOT size in bytes) */
2846 wspace, /* working space vector */
2847 20); /* number of elements (NOT size in bytes) */
2849 Option bits for pcre_dfa_exec()
2851 The unused bits of the options argument for pcre_dfa_exec() must be
2852 zero. The only bits that may be set are PCRE_ANCHORED, PCRE_NEW-
2857 four of these are exactly the same as for pcre_exec(), so their
2858 description is not repeated here.
2863 These have the same general effect as they do for pcre_exec(), but the
2864 details are slightly different. When PCRE_PARTIAL_HARD is set for
2865 pcre_dfa_exec(), it returns PCRE_ERROR_PARTIAL if the end of the sub-
2866 ject is reached and there is still at least one matching possibility
2867 that requires additional characters. This happens even if some complete
2868 matches have also been found. When PCRE_PARTIAL_SOFT is set, the return
2869 code PCRE_ERROR_NOMATCH is converted into PCRE_ERROR_PARTIAL if the end
2870 of the subject is reached, there have been no complete matches, but
2871 there is still at least one matching possibility. The portion of the
2872 string that was inspected when the longest partial match was found is
2873 set as the first matching string in both cases. There is a more
2874 detailed discussion of partial and multi-segment matching, with exam-
2875 ples, in the pcrepartial documentation.
2879 Setting the PCRE_DFA_SHORTEST option causes the matching algorithm to
2880 stop as soon as it has found one match. Because of the way the alterna-
2881 tive algorithm works, this is necessarily the shortest possible match
2882 at the first possible matching point in the subject string.
2886 When pcre_dfa_exec() returns a partial match, it is possible to call it
2887 again, with additional subject characters, and have it continue with
2888 the same match. The PCRE_DFA_RESTART option requests this action; when
2889 it is set, the workspace and wscount options must reference the same
2890 vector as before because data about the match so far is left in them
2891 after a partial match. There is more discussion of this facility in the
2892 pcrepartial documentation.
2894 Successful returns from pcre_dfa_exec()
2896 When pcre_dfa_exec() succeeds, it may have matched more than one sub-
2897 string in the subject. Note, however, that all the matches from one run
2898 of the function start at the same point in the subject. The shorter
2899 matches are all initial substrings of the longer matches. For example,
2900 if the pattern
2902 <.*>
2904 is matched against the string
2906 This is <something> <something else> <something further> no more
2908 the three matched strings are
2910 <something>
2911 <something> <something else>
2912 <something> <something else> <something further>
2914 On success, the yield of the function is a number greater than zero,
2915 which is the number of matched substrings. The substrings themselves
2916 are returned in ovector. Each string uses two elements; the first is
2917 the offset to the start, and the second is the offset to the end. In
2918 fact, all the strings have the same start offset. (Space could have
2919 been saved by giving this only once, but it was decided to retain some
2920 compatibility with the way pcre_exec() returns data, even though the
2921 meaning of the strings is different.)
2923 The strings are returned in reverse order of length; that is, the long-
2924 est matching string is given first. If there were too many matches to
2925 fit into ovector, the yield of the function is zero, and the vector is
2926 filled with the longest matches. Unlike pcre_exec(), pcre_dfa_exec()
2927 can use the entire ovector for returning matched strings.
2929 Error returns from pcre_dfa_exec()
2931 The pcre_dfa_exec() function returns a negative number when it fails.
2932 Many of the errors are the same as for pcre_exec(), and these are
2933 described above. There are in addition the following errors that are
2934 specific to pcre_dfa_exec():
2938 This return is given if pcre_dfa_exec() encounters an item in the pat-
2939 tern that it does not support, for instance, the use of \C or a back
2940 reference.
2944 This return is given if pcre_dfa_exec() encounters a condition item
2945 that uses a back reference for the condition, or a test for recursion
2946 in a specific group. These are not supported.
2950 This return is given if pcre_dfa_exec() is called with an extra block
2951 that contains a setting of the match_limit or match_limit_recursion
2952 fields. This is not supported (these fields are meaningless for DFA
2953 matching).
2957 This return is given if pcre_dfa_exec() runs out of space in the
2958 workspace vector.
2962 When a recursive subpattern is processed, the matching function calls
2963 itself recursively, using private vectors for ovector and workspace.
2964 This error is given if the output vector is not large enough. This
2965 should be extremely rare, as a vector of size 1000 is used.
2970 pcrebuild(3), pcrecallout(3), pcrecpp(3)(3), pcrematching(3), pcrepar-
2971 tial(3), pcreposix(3), pcreprecompile(3), pcresample(3), pcrestack(3).
2976 Philip Hazel
2977 University Computing Service
2978 Cambridge CB2 3QH, England.
2983 Last updated: 23 September 2011
2984 Copyright (c) 1997-2011 University of Cambridge.
2985 ------------------------------------------------------------------------------
2991 NAME
2992 PCRE - Perl-compatible regular expressions
2997 int (*pcre_callout)(pcre_callout_block *);
2999 PCRE provides a feature called "callout", which is a means of temporar-
3000 ily passing control to the caller of PCRE in the middle of pattern
3001 matching. The caller of PCRE provides an external function by putting
3002 its entry point in the global variable pcre_callout. By default, this
3003 variable contains NULL, which disables all calling out.
3005 Within a regular expression, (?C) indicates the points at which the
3006 external function is to be called. Different callout points can be
3007 identified by putting a number less than 256 after the letter C. The
3008 default value is zero. For example, this pattern has two callout
3009 points:
3011 (?C1)abc(?C2)def
3013 If the PCRE_AUTO_CALLOUT option bit is set when pcre_compile() or
3014 pcre_compile2() is called, PCRE automatically inserts callouts, all
3015 with number 255, before each item in the pattern. For example, if
3016 PCRE_AUTO_CALLOUT is used with the pattern
3018 A(\d{2}|--)
3020 it is processed as if it were
3022 (?C255)A(?C255)((?C255)\d{2}(?C255)|(?C255)-(?C255)-(?C255))(?C255)
3024 Notice that there is a callout before and after each parenthesis and
3025 alternation bar. Automatic callouts can be used for tracking the
3026 progress of pattern matching. The pcretest command has an option that
3027 sets automatic callouts; when it is used, the output indicates how the
3028 pattern is matched. This is useful information when you are trying to
3029 optimize the performance of a particular pattern.
3031 The use of callouts in a pattern makes it ineligible for optimization
3032 by the just-in-time compiler. Studying such a pattern with the
3033 PCRE_STUDY_JIT_COMPILE option always fails.
3038 You should be aware that, because of optimizations in the way PCRE
3039 matches patterns by default, callouts sometimes do not happen. For
3040 example, if the pattern is
3042 ab(?C4)cd
3044 PCRE knows that any matching string must contain the letter "d". If the
3045 subject string is "abyz", the lack of "d" means that matching doesn't
3046 ever start, and the callout is never reached. However, with "abyd",
3047 though the result is still no match, the callout is obeyed.
3049 If the pattern is studied, PCRE knows the minimum length of a matching
3050 string, and will immediately give a "no match" return without actually
3051 running a match if the subject is not long enough, or, for unanchored
3052 patterns, if it has been scanned far enough.
3054 You can disable these optimizations by passing the PCRE_NO_START_OPTI-
3055 MIZE option to pcre_compile(), pcre_exec(), or pcre_dfa_exec(), or by
3056 starting the pattern with (*NO_START_OPT). This slows down the matching
3057 process, but does ensure that callouts such as the example above are
3058 obeyed.
3063 During matching, when PCRE reaches a callout point, the external func-
3064 tion defined by pcre_callout is called (if it is set). This applies to
3065 both the pcre_exec() and the pcre_dfa_exec() matching functions. The
3066 only argument to the callout function is a pointer to a pcre_callout
3067 block. This structure contains the following fields:
3069 int version;
3070 int callout_number;
3071 int *offset_vector;
3072 const char *subject;
3073 int subject_length;
3074 int start_match;
3075 int current_position;
3076 int capture_top;
3077 int capture_last;
3078 void *callout_data;
3079 int pattern_position;
3080 int next_item_length;
3081 const unsigned char *mark;
3083 The version field is an integer containing the version number of the
3084 block format. The initial version was 0; the current version is 2. The
3085 version number will change again in future if additional fields are
3086 added, but the intention is never to remove any of the existing fields.
3088 The callout_number field contains the number of the callout, as com-
3089 piled into the pattern (that is, the number after ?C for manual call-
3090 outs, and 255 for automatically generated callouts).
3092 The offset_vector field is a pointer to the vector of offsets that was
3093 passed by the caller to pcre_exec() or pcre_dfa_exec(). When
3094 pcre_exec() is used, the contents can be inspected in order to extract
3095 substrings that have been matched so far, in the same way as for
3096 extracting substrings after a match has completed. For pcre_dfa_exec()
3097 this field is not useful.
3099 The subject and subject_length fields contain copies of the values that
3100 were passed to pcre_exec().
3102 The start_match field normally contains the offset within the subject
3103 at which the current match attempt started. However, if the escape
3104 sequence \K has been encountered, this value is changed to reflect the
3105 modified starting point. If the pattern is not anchored, the callout
3106 function may be called several times from the same point in the pattern
3107 for different starting points in the subject.
3109 The current_position field contains the offset within the subject of
3110 the current match pointer.
3112 When the pcre_exec() function is used, the capture_top field contains
3113 one more than the number of the highest numbered captured substring so
3114 far. If no substrings have been captured, the value of capture_top is
3115 one. This is always the case when pcre_dfa_exec() is used, because it
3116 does not support captured substrings.
3118 The capture_last field contains the number of the most recently cap-
3119 tured substring. If no substrings have been captured, its value is -1.
3120 This is always the case when pcre_dfa_exec() is used.
3122 The callout_data field contains a value that is passed to pcre_exec()
3123 or pcre_dfa_exec() specifically so that it can be passed back in call-
3124 outs. It is passed in the pcre_callout field of the pcre_extra data
3125 structure. If no such data was passed, the value of callout_data in a
3126 pcre_callout block is NULL. There is a description of the pcre_extra
3127 structure in the pcreapi documentation.
3129 The pattern_position field is present from version 1 of the pcre_call-
3130 out structure. It contains the offset to the next item to be matched in
3131 the pattern string.
3133 The next_item_length field is present from version 1 of the pcre_call-
3134 out structure. It contains the length of the next item to be matched in
3135 the pattern string. When the callout immediately precedes an alterna-
3136 tion bar, a closing parenthesis, or the end of the pattern, the length
3137 is zero. When the callout precedes an opening parenthesis, the length
3138 is that of the entire subpattern.
3140 The pattern_position and next_item_length fields are intended to help
3141 in distinguishing between different automatic callouts, which all have
3142 the same callout number. However, they are set for all callouts.
3144 The mark field is present from version 2 of the pcre_callout structure.
3145 In callouts from pcre_exec() it contains a pointer to the zero-termi-
3146 nated name of the most recently passed (*MARK) item in the match, or
3147 NULL if there are no (*MARK)s in the current matching path. In callouts
3148 from pcre_dfa_exec() this field always contains NULL.
3153 The external callout function returns an integer to PCRE. If the value
3154 is zero, matching proceeds as normal. If the value is greater than
3155 zero, matching fails at the current point, but the testing of other
3156 matching possibilities goes ahead, just as if a lookahead assertion had
3157 failed. If the value is less than zero, the match is abandoned, and
3158 pcre_exec() or pcre_dfa_exec() returns the negative value.
3160 Negative values should normally be chosen from the set of
3161 PCRE_ERROR_xxx values. In particular, PCRE_ERROR_NOMATCH forces a stan-
3162 dard "no match" failure. The error number PCRE_ERROR_CALLOUT is
3163 reserved for use by callout functions; it will never be used by PCRE
3164 itself.
3169 Philip Hazel
3170 University Computing Service
3171 Cambridge CB2 3QH, England.
3176 Last updated: 26 August 2011
3177 Copyright (c) 1997-2011 University of Cambridge.
3178 ------------------------------------------------------------------------------
3184 NAME
3185 PCRE - Perl-compatible regular expressions
3190 This document describes the differences in the ways that PCRE and Perl
3191 handle regular expressions. The differences described here are with
3192 respect to Perl versions 5.10 and above.
3194 1. PCRE has only a subset of Perl's UTF-8 and Unicode support. Details
3195 of what it does have are given in the pcreunicode page.
3197 2. PCRE allows repeat quantifiers only on parenthesized assertions, but
3198 they do not mean what you might think. For example, (?!a){3} does not
3199 assert that the next three characters are not "a". It just asserts that
3200 the next character is not "a" three times (in principle: PCRE optimizes
3201 this to run the assertion just once). Perl allows repeat quantifiers on
3202 other assertions such as \b, but these do not seem to have any use.
3204 3. Capturing subpatterns that occur inside negative lookahead asser-
3205 tions are counted, but their entries in the offsets vector are never
3206 set. Perl sets its numerical variables from any such patterns that are
3207 matched before the assertion fails to match something (thereby succeed-
3208 ing), but only if the negative lookahead assertion contains just one
3209 branch.
3211 4. Though binary zero characters are supported in the subject string,
3212 they are not allowed in a pattern string because it is passed as a nor-
3213 mal C string, terminated by zero. The escape sequence \0 can be used in
3214 the pattern to represent a binary zero.
3216 5. The following Perl escape sequences are not supported: \l, \u, \L,
3217 \U, and \N when followed by a character name or Unicode value. (\N on
3218 its own, matching a non-newline character, is supported.) In fact these
3219 are implemented by Perl's general string-handling and are not part of
3220 its pattern matching engine. If any of these are encountered by PCRE,
3221 an error is generated.
3223 6. The Perl escape sequences \p, \P, and \X are supported only if PCRE
3224 is built with Unicode character property support. The properties that
3225 can be tested with \p and \P are limited to the general category prop-
3226 erties such as Lu and Nd, script names such as Greek or Han, and the
3227 derived properties Any and L&. PCRE does support the Cs (surrogate)
3228 property, which Perl does not; the Perl documentation says "Because
3229 Perl hides the need for the user to understand the internal representa-
3230 tion of Unicode characters, there is no need to implement the somewhat
3231 messy concept of surrogates."
3233 7. PCRE implements a simpler version of \X than Perl, which changed to
3234 make \X match what Unicode calls an "extended grapheme cluster". This
3235 is more complicated than an extended Unicode sequence, which is what
3236 PCRE matches.
3238 8. PCRE does support the \Q...\E escape for quoting substrings. Charac-
3239 ters in between are treated as literals. This is slightly different
3240 from Perl in that $ and @ are also handled as literals inside the
3241 quotes. In Perl, they cause variable interpolation (but of course PCRE
3242 does not have variables). Note the following examples:
3244 Pattern PCRE matches Perl matches
3246 \Qabc$xyz\E abc$xyz abc followed by the
3247 contents of $xyz
3248 \Qabc\$xyz\E abc\$xyz abc\$xyz
3249 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3251 The \Q...\E sequence is recognized both inside and outside character
3252 classes.
3254 9. Fairly obviously, PCRE does not support the (?{code}) and (??{code})
3255 constructions. However, there is support for recursive patterns. This
3256 is not available in Perl 5.8, but it is in Perl 5.10. Also, the PCRE
3257 "callout" feature allows an external function to be called during pat-
3258 tern matching. See the pcrecallout documentation for details.
3260 10. Subpatterns that are called as subroutines (whether or not recur-
3261 sively) are always treated as atomic groups in PCRE. This is like
3262 Python, but unlike Perl. Captured values that are set outside a sub-
3263 routine call can be reference from inside in PCRE, but not in Perl.
3264 There is a discussion that explains these differences in more detail in
3265 the section on recursion differences from Perl in the pcrepattern page.
3267 11. If (*THEN) is present in a group that is called as a subroutine,
3268 its action is limited to that group, even if the group does not contain
3269 any | characters.
3271 12. There are some differences that are concerned with the settings of
3272 captured strings when part of a pattern is repeated. For example,
3273 matching "aba" against the pattern /^(a(b)?)+$/ in Perl leaves $2
3274 unset, but in PCRE it is set to "b".
3276 13. PCRE's handling of duplicate subpattern numbers and duplicate sub-
3277 pattern names is not as general as Perl's. This is a consequence of the
3278 fact the PCRE works internally just with numbers, using an external ta-
3279 ble to translate between numbers and names. In particular, a pattern
3280 such as (?|(?<a>A)|(?<b)B), where the two capturing parentheses have
3281 the same number but different names, is not supported, and causes an
3282 error at compile time. If it were allowed, it would not be possible to
3283 distinguish which parentheses matched, because both names map to cap-
3284 turing subpattern number 1. To avoid this confusing situation, an error
3285 is given at compile time.
3287 14. Perl recognizes comments in some places that PCRE does not, for
3288 example, between the ( and ? at the start of a subpattern. If the /x
3289 modifier is set, Perl allows whitespace between ( and ? but PCRE never
3290 does, even if the PCRE_EXTENDED option is set.
3292 15. PCRE provides some extensions to the Perl regular expression facil-
3293 ities. Perl 5.10 includes new features that are not in earlier ver-
3294 sions of Perl, some of which (such as named parentheses) have been in
3295 PCRE for some time. This list is with respect to Perl 5.10:
3297 (a) Although lookbehind assertions in PCRE must match fixed length
3298 strings, each alternative branch of a lookbehind assertion can match a
3299 different length of string. Perl requires them all to have the same
3300 length.
3302 (b) If PCRE_DOLLAR_ENDONLY is set and PCRE_MULTILINE is not set, the $
3303 meta-character matches only at the very end of the string.
3305 (c) If PCRE_EXTRA is set, a backslash followed by a letter with no spe-
3306 cial meaning is faulted. Otherwise, like Perl, the backslash is quietly
3307 ignored. (Perl can be made to issue a warning.)
3309 (d) If PCRE_UNGREEDY is set, the greediness of the repetition quanti-
3310 fiers is inverted, that is, by default they are not greedy, but if fol-
3311 lowed by a question mark they are.
3313 (e) PCRE_ANCHORED can be used at matching time to force a pattern to be
3314 tried only at the first matching position in the subject string.
3317 and PCRE_NO_AUTO_CAPTURE options for pcre_exec() have no Perl equiva-
3318 lents.
3320 (g) The \R escape sequence can be restricted to match only CR, LF, or
3321 CRLF by the PCRE_BSR_ANYCRLF option.
3323 (h) The callout facility is PCRE-specific.
3325 (i) The partial matching facility is PCRE-specific.
3327 (j) Patterns compiled by PCRE can be saved and re-used at a later time,
3328 even on different hosts that have the other endianness. However, this
3329 does not apply to optimized data created by the just-in-time compiler.
3331 (k) The alternative matching function (pcre_dfa_exec()) matches in a
3332 different way and is not Perl-compatible.
3334 (l) PCRE recognizes some special sequences such as (*CR) at the start
3335 of a pattern that set overall options that cannot be changed within the
3336 pattern.
3341 Philip Hazel
3342 University Computing Service
3343 Cambridge CB2 3QH, England.
3348 Last updated: 09 October 2011
3349 Copyright (c) 1997-2011 University of Cambridge.
3350 ------------------------------------------------------------------------------
3356 NAME
3357 PCRE - Perl-compatible regular expressions
3362 The syntax and semantics of the regular expressions that are supported
3363 by PCRE are described in detail below. There is a quick-reference syn-
3364 tax summary in the pcresyntax page. PCRE tries to match Perl syntax and
3365 semantics as closely as it can. PCRE also supports some alternative
3366 regular expression syntax (which does not conflict with the Perl syn-
3367 tax) in order to provide some compatibility with regular expressions in
3368 Python, .NET, and Oniguruma.
3370 Perl's regular expressions are described in its own documentation, and
3371 regular expressions in general are covered in a number of books, some
3372 of which have copious examples. Jeffrey Friedl's "Mastering Regular
3373 Expressions", published by O'Reilly, covers regular expressions in
3374 great detail. This description of PCRE's regular expressions is
3375 intended as reference material.
3377 The original operation of PCRE was on strings of one-byte characters.
3378 However, there is now also support for UTF-8 character strings. To use
3379 this, PCRE must be built to include UTF-8 support, and you must call
3380 pcre_compile() or pcre_compile2() with the PCRE_UTF8 option. There is
3381 also a special sequence that can be given at the start of a pattern:
3383 (*UTF8)
3385 Starting a pattern with this sequence is equivalent to setting the
3386 PCRE_UTF8 option. This feature is not Perl-compatible. How setting
3387 UTF-8 mode affects pattern matching is mentioned in several places
3388 below. There is also a summary of UTF-8 features in the pcreunicode
3389 page.
3391 Another special sequence that may appear at the start of a pattern or
3392 in combination with (*UTF8) is:
3394 (*UCP)
3396 This has the same effect as setting the PCRE_UCP option: it causes
3397 sequences such as \d and \w to use Unicode properties to determine
3398 character types, instead of recognizing only characters with codes less
3399 than 128 via a lookup table.
3401 If a pattern starts with (*NO_START_OPT), it has the same effect as
3402 setting the PCRE_NO_START_OPTIMIZE option either at compile or matching
3403 time. There are also some more of these special sequences that are con-
3404 cerned with the handling of newlines; they are described below.
3406 The remainder of this document discusses the patterns that are sup-
3407 ported by PCRE when its main matching function, pcre_exec(), is used.
3408 From release 6.0, PCRE offers a second matching function,
3409 pcre_dfa_exec(), which matches using a different algorithm that is not
3410 Perl-compatible. Some of the features discussed below are not available
3411 when pcre_dfa_exec() is used. The advantages and disadvantages of the
3412 alternative function, and how it differs from the normal function, are
3413 discussed in the pcrematching page.
3418 PCRE supports five different conventions for indicating line breaks in
3419 strings: a single CR (carriage return) character, a single LF (line-
3420 feed) character, the two-character sequence CRLF, any of the three pre-
3421 ceding, or any Unicode newline sequence. The pcreapi page has further
3422 discussion about newlines, and shows how to set the newline convention
3423 in the options arguments for the compiling and matching functions.
3425 It is also possible to specify a newline convention by starting a pat-
3426 tern string with one of the following five sequences:
3428 (*CR) carriage return
3429 (*LF) linefeed
3430 (*CRLF) carriage return, followed by linefeed
3431 (*ANYCRLF) any of the three above
3432 (*ANY) all Unicode newline sequences
3434 These override the default and the options given to pcre_compile() or
3435 pcre_compile2(). For example, on a Unix system where LF is the default
3436 newline sequence, the pattern
3438 (*CR)a.b
3440 changes the convention to CR. That pattern matches "a\nb" because LF is
3441 no longer a newline. Note that these special settings, which are not
3442 Perl-compatible, are recognized only at the very start of a pattern,
3443 and that they must be in upper case. If more than one of them is
3444 present, the last one is used.
3446 The newline convention affects the interpretation of the dot metachar-
3447 acter when PCRE_DOTALL is not set, and also the behaviour of \N. How-
3448 ever, it does not affect what the \R escape sequence matches. By
3449 default, this is any Unicode newline sequence, for Perl compatibility.
3450 However, this can be changed; see the description of \R in the section
3451 entitled "Newline sequences" below. A change of \R setting can be com-
3452 bined with a change of newline convention.
3457 A regular expression is a pattern that is matched against a subject
3458 string from left to right. Most characters stand for themselves in a
3459 pattern, and match the corresponding characters in the subject. As a
3460 trivial example, the pattern
3462 The quick brown fox
3464 matches a portion of a subject string that is identical to itself. When
3465 caseless matching is specified (the PCRE_CASELESS option), letters are
3466 matched independently of case. In UTF-8 mode, PCRE always understands
3467 the concept of case for characters whose values are less than 128, so
3468 caseless matching is always possible. For characters with higher val-
3469 ues, the concept of case is supported if PCRE is compiled with Unicode
3470 property support, but not otherwise. If you want to use caseless
3471 matching for characters 128 and above, you must ensure that PCRE is
3472 compiled with Unicode property support as well as with UTF-8 support.
3474 The power of regular expressions comes from the ability to include
3475 alternatives and repetitions in the pattern. These are encoded in the
3476 pattern by the use of metacharacters, which do not stand for themselves
3477 but instead are interpreted in some special way.
3479 There are two different sets of metacharacters: those that are recog-
3480 nized anywhere in the pattern except within square brackets, and those
3481 that are recognized within square brackets. Outside square brackets,
3482 the metacharacters are as follows:
3484 \ general escape character with several uses
3485 ^ assert start of string (or line, in multiline mode)
3486 $ assert end of string (or line, in multiline mode)
3487 . match any character except newline (by default)
3488 [ start character class definition
3489 | start of alternative branch
3490 ( start subpattern
3491 ) end subpattern
3492 ? extends the meaning of (
3493 also 0 or 1 quantifier
3494 also quantifier minimizer
3495 * 0 or more quantifier
3496 + 1 or more quantifier
3497 also "possessive quantifier"
3498 { start min/max quantifier
3500 Part of a pattern that is in square brackets is called a "character
3501 class". In a character class the only metacharacters are:
3503 \ general escape character
3504 ^ negate the class, but only if the first character
3505 - indicates character range
3506 [ POSIX character class (only if followed by POSIX
3507 syntax)
3508 ] terminates the character class
3510 The following sections describe the use of each of the metacharacters.
3515 The backslash character has several uses. Firstly, if it is followed by
3516 a character that is not a number or a letter, it takes away any special
3517 meaning that character may have. This use of backslash as an escape
3518 character applies both inside and outside character classes.
3520 For example, if you want to match a * character, you write \* in the
3521 pattern. This escaping action applies whether or not the following
3522 character would otherwise be interpreted as a metacharacter, so it is
3523 always safe to precede a non-alphanumeric with backslash to specify
3524 that it stands for itself. In particular, if you want to match a back-
3525 slash, you write \\.
3527 In UTF-8 mode, only ASCII numbers and letters have any special meaning
3528 after a backslash. All other characters (in particular, those whose
3529 codepoints are greater than 127) are treated as literals.
3531 If a pattern is compiled with the PCRE_EXTENDED option, whitespace in
3532 the pattern (other than in a character class) and characters between a
3533 # outside a character class and the next newline are ignored. An escap-
3534 ing backslash can be used to include a whitespace or # character as
3535 part of the pattern.
3537 If you want to remove the special meaning from a sequence of charac-
3538 ters, you can do so by putting them between \Q and \E. This is differ-
3539 ent from Perl in that $ and @ are handled as literals in \Q...\E
3540 sequences in PCRE, whereas in Perl, $ and @ cause variable interpola-
3541 tion. Note the following examples:
3543 Pattern PCRE matches Perl matches
3545 \Qabc$xyz\E abc$xyz abc followed by the
3546 contents of $xyz
3547 \Qabc\$xyz\E abc\$xyz abc\$xyz
3548 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3550 The \Q...\E sequence is recognized both inside and outside character
3551 classes. An isolated \E that is not preceded by \Q is ignored. If \Q
3552 is not followed by \E later in the pattern, the literal interpretation
3553 continues to the end of the pattern (that is, \E is assumed at the
3554 end). If the isolated \Q is inside a character class, this causes an
3555 error, because the character class is not terminated.
3557 Non-printing characters
3559 A second use of backslash provides a way of encoding non-printing char-
3560 acters in patterns in a visible manner. There is no restriction on the
3561 appearance of non-printing characters, apart from the binary zero that
3562 terminates a pattern, but when a pattern is being prepared by text
3563 editing, it is often easier to use one of the following escape
3564 sequences than the binary character it represents:
3566 \a alarm, that is, the BEL character (hex 07)
3567 \cx "control-x", where x is any ASCII character
3568 \e escape (hex 1B)
3569 \f formfeed (hex 0C)
3570 \n linefeed (hex 0A)
3571 \r carriage return (hex 0D)
3572 \t tab (hex 09)
3573 \ddd character with octal code ddd, or back reference
3574 \xhh character with hex code hh
3575 \x{hhh..} character with hex code hhh..
3577 The precise effect of \cx is as follows: if x is a lower case letter,
3578 it is converted to upper case. Then bit 6 of the character (hex 40) is
3579 inverted. Thus \cz becomes hex 1A (z is 7A), but \c{ becomes hex 3B ({
3580 is 7B), while \c; becomes hex 7B (; is 3B). If the byte following \c
3581 has a value greater than 127, a compile-time error occurs. This locks
3582 out non-ASCII characters in both byte mode and UTF-8 mode. (When PCRE
3583 is compiled in EBCDIC mode, all byte values are valid. A lower case
3584 letter is converted to upper case, and then the 0xc0 bits are flipped.)
3586 After \x, from zero to two hexadecimal digits are read (letters can be
3587 in upper or lower case). Any number of hexadecimal digits may appear
3588 between \x{ and }, but the value of the character code must be less
3589 than 256 in non-UTF-8 mode, and less than 2**31 in UTF-8 mode. That is,
3590 the maximum value in hexadecimal is 7FFFFFFF. Note that this is bigger
3591 than the largest Unicode code point, which is 10FFFF.
3593 If characters other than hexadecimal digits appear between \x{ and },
3594 or if there is no terminating }, this form of escape is not recognized.
3595 Instead, the initial \x will be interpreted as a basic hexadecimal
3596 escape, with no following digits, giving a character whose value is
3597 zero.
3599 Characters whose value is less than 256 can be defined by either of the
3600 two syntaxes for \x. There is no difference in the way they are han-
3601 dled. For example, \xdc is exactly the same as \x{dc}.
3603 After \0 up to two further octal digits are read. If there are fewer
3604 than two digits, just those that are present are used. Thus the
3605 sequence \0\x\07 specifies two binary zeros followed by a BEL character
3606 (code value 7). Make sure you supply two digits after the initial zero
3607 if the pattern character that follows is itself an octal digit.
3609 The handling of a backslash followed by a digit other than 0 is compli-
3610 cated. Outside a character class, PCRE reads it and any following dig-
3611 its as a decimal number. If the number is less than 10, or if there
3612 have been at least that many previous capturing left parentheses in the
3613 expression, the entire sequence is taken as a back reference. A
3614 description of how this works is given later, following the discussion
3615 of parenthesized subpatterns.
3617 Inside a character class, or if the decimal number is greater than 9
3618 and there have not been that many capturing subpatterns, PCRE re-reads
3619 up to three octal digits following the backslash, and uses them to gen-
3620 erate a data character. Any subsequent digits stand for themselves. In
3621 non-UTF-8 mode, the value of a character specified in octal must be
3622 less than \400. In UTF-8 mode, values up to \777 are permitted. For
3623 example:
3625 \040 is another way of writing a space
3626 \40 is the same, provided there are fewer than 40
3627 previous capturing subpatterns
3628 \7 is always a back reference
3629 \11 might be a back reference, or another way of
3630 writing a tab
3631 \011 is always a tab
3632 \0113 is a tab followed by the character "3"
3633 \113 might be a back reference, otherwise the
3634 character with octal code 113
3635 \377 might be a back reference, otherwise
3636 the byte consisting entirely of 1 bits
3637 \81 is either a back reference, or a binary zero
3638 followed by the two characters "8" and "1"
3640 Note that octal values of 100 or greater must not be introduced by a
3641 leading zero, because no more than three octal digits are ever read.
3643 All the sequences that define a single character value can be used both
3644 inside and outside character classes. In addition, inside a character
3645 class, the sequence \b is interpreted as the backspace character (hex
3646 08). The sequences \B, \N, \R, and \X are not special inside a charac-
3647 ter class. Like any other unrecognized escape sequences, they are
3648 treated as the literal characters "B", "N", "R", and "X" by default,
3649 but cause an error if the PCRE_EXTRA option is set. Outside a character
3650 class, these sequences have different meanings.
3652 Absolute and relative back references
3654 The sequence \g followed by an unsigned or a negative number, option-
3655 ally enclosed in braces, is an absolute or relative back reference. A
3656 named back reference can be coded as \g{name}. Back references are dis-
3657 cussed later, following the discussion of parenthesized subpatterns.
3659 Absolute and relative subroutine calls
3661 For compatibility with Oniguruma, the non-Perl syntax \g followed by a
3662 name or a number enclosed either in angle brackets or single quotes, is
3663 an alternative syntax for referencing a subpattern as a "subroutine".
3664 Details are discussed later. Note that \g{...} (Perl syntax) and
3665 \g<...> (Oniguruma syntax) are not synonymous. The former is a back
3666 reference; the latter is a subroutine call.
3668 Generic character types
3670 Another use of backslash is for specifying generic character types:
3672 \d any decimal digit
3673 \D any character that is not a decimal digit
3674 \h any horizontal whitespace character
3675 \H any character that is not a horizontal whitespace character
3676 \s any whitespace character
3677 \S any character that is not a whitespace character
3678 \v any vertical whitespace character
3679 \V any character that is not a vertical whitespace character
3680 \w any "word" character
3681 \W any "non-word" character
3683 There is also the single sequence \N, which matches a non-newline char-
3684 acter. This is the same as the "." metacharacter when PCRE_DOTALL is
3685 not set.
3687 Each pair of lower and upper case escape sequences partitions the com-
3688 plete set of characters into two disjoint sets. Any given character
3689 matches one, and only one, of each pair. The sequences can appear both
3690 inside and outside character classes. They each match one character of
3691 the appropriate type. If the current matching point is at the end of
3692 the subject string, all of them fail, because there is no character to
3693 match.
3695 For compatibility with Perl, \s does not match the VT character (code
3696 11). This makes it different from the the POSIX "space" class. The \s
3697 characters are HT (9), LF (10), FF (12), CR (13), and space (32). If
3698 "use locale;" is included in a Perl script, \s may match the VT charac-
3699 ter. In PCRE, it never does.
3701 A "word" character is an underscore or any character that is a letter
3702 or digit. By default, the definition of letters and digits is con-
3703 trolled by PCRE's low-valued character tables, and may vary if locale-
3704 specific matching is taking place (see "Locale support" in the pcreapi
3705 page). For example, in a French locale such as "fr_FR" in Unix-like
3706 systems, or "french" in Windows, some character codes greater than 128
3707 are used for accented letters, and these are then matched by \w. The
3708 use of locales with Unicode is discouraged.
3710 By default, in UTF-8 mode, characters with values greater than 128
3711 never match \d, \s, or \w, and always match \D, \S, and \W. These
3712 sequences retain their original meanings from before UTF-8 support was
3713 available, mainly for efficiency reasons. However, if PCRE is compiled
3714 with Unicode property support, and the PCRE_UCP option is set, the be-
3715 haviour is changed so that Unicode properties are used to determine
3716 character types, as follows:
3718 \d any character that \p{Nd} matches (decimal digit)
3719 \s any character that \p{Z} matches, plus HT, LF, FF, CR
3720 \w any character that \p{L} or \p{N} matches, plus underscore
3722 The upper case escapes match the inverse sets of characters. Note that
3723 \d matches only decimal digits, whereas \w matches any Unicode digit,
3724 as well as any Unicode letter, and underscore. Note also that PCRE_UCP
3725 affects \b, and \B because they are defined in terms of \w and \W.
3726 Matching these sequences is noticeably slower when PCRE_UCP is set.
3728 The sequences \h, \H, \v, and \V are features that were added to Perl
3729 at release 5.10. In contrast to the other sequences, which match only
3730 ASCII characters by default, these always match certain high-valued
3731 codepoints in UTF-8 mode, whether or not PCRE_UCP is set. The horizon-
3732 tal space characters are:
3734 U+0009 Horizontal tab
3735 U+0020 Space
3736 U+00A0 Non-break space
3737 U+1680 Ogham space mark
3738 U+180E Mongolian vowel separator
3739 U+2000 En quad
3740 U+2001 Em quad
3741 U+2002 En space
3742 U+2003 Em space
3743 U+2004 Three-per-em space
3744 U+2005 Four-per-em space
3745 U+2006 Six-per-em space
3746 U+2007 Figure space
3747 U+2008 Punctuation space
3748 U+2009 Thin space
3749 U+200A Hair space
3750 U+202F Narrow no-break space
3751 U+205F Medium mathematical space
3752 U+3000 Ideographic space
3754 The vertical space characters are:
3756 U+000A Linefeed
3757 U+000B Vertical tab
3758 U+000C Formfeed
3759 U+000D Carriage return
3760 U+0085 Next line
3761 U+2028 Line separator
3762 U+2029 Paragraph separator
3764 Newline sequences
3766 Outside a character class, by default, the escape sequence \R matches
3767 any Unicode newline sequence. In non-UTF-8 mode \R is equivalent to the
3768 following:
3770 (?>\r\n|\n|\x0b|\f|\r|\x85)
3772 This is an example of an "atomic group", details of which are given
3773 below. This particular group matches either the two-character sequence
3774 CR followed by LF, or one of the single characters LF (linefeed,
3775 U+000A), VT (vertical tab, U+000B), FF (formfeed, U+000C), CR (carriage
3776 return, U+000D), or NEL (next line, U+0085). The two-character sequence
3777 is treated as a single unit that cannot be split.
3779 In UTF-8 mode, two additional characters whose codepoints are greater
3780 than 255 are added: LS (line separator, U+2028) and PS (paragraph sepa-
3781 rator, U+2029). Unicode character property support is not needed for
3782 these characters to be recognized.
3784 It is possible to restrict \R to match only CR, LF, or CRLF (instead of
3785 the complete set of Unicode line endings) by setting the option
3786 PCRE_BSR_ANYCRLF either at compile time or when the pattern is matched.
3787 (BSR is an abbrevation for "backslash R".) This can be made the default
3788 when PCRE is built; if this is the case, the other behaviour can be
3789 requested via the PCRE_BSR_UNICODE option. It is also possible to
3790 specify these settings by starting a pattern string with one of the
3791 following sequences:
3793 (*BSR_ANYCRLF) CR, LF, or CRLF only
3794 (*BSR_UNICODE) any Unicode newline sequence
3796 These override the default and the options given to pcre_compile() or
3797 pcre_compile2(), but they can be overridden by options given to
3798 pcre_exec() or pcre_dfa_exec(). Note that these special settings, which
3799 are not Perl-compatible, are recognized only at the very start of a
3800 pattern, and that they must be in upper case. If more than one of them
3801 is present, the last one is used. They can be combined with a change of
3802 newline convention; for example, a pattern can start with:
3806 They can also be combined with the (*UTF8) or (*UCP) special sequences.
3807 Inside a character class, \R is treated as an unrecognized escape
3808 sequence, and so matches the letter "R" by default, but causes an error
3809 if PCRE_EXTRA is set.
3811 Unicode character properties
3813 When PCRE is built with Unicode character property support, three addi-
3814 tional escape sequences that match characters with specific properties
3815 are available. When not in UTF-8 mode, these sequences are of course
3816 limited to testing characters whose codepoints are less than 256, but
3817 they do work in this mode. The extra escape sequences are:
3819 \p{xx} a character with the xx property
3820 \P{xx} a character without the xx property
3821 \X an extended Unicode sequence
3823 The property names represented by xx above are limited to the Unicode
3824 script names, the general category properties, "Any", which matches any
3825 character (including newline), and some special PCRE properties
3826 (described in the next section). Other Perl properties such as "InMu-
3827 sicalSymbols" are not currently supported by PCRE. Note that \P{Any}
3828 does not match any characters, so always causes a match failure.
3830 Sets of Unicode characters are defined as belonging to certain scripts.
3831 A character from one of these sets can be matched using a script name.
3832 For example:
3834 \p{Greek}
3835 \P{Han}
3837 Those that are not part of an identified script are lumped together as
3838 "Common". The current list of scripts is:
3840 Arabic, Armenian, Avestan, Balinese, Bamum, Bengali, Bopomofo, Braille,
3841 Buginese, Buhid, Canadian_Aboriginal, Carian, Cham, Cherokee, Common,
3842 Coptic, Cuneiform, Cypriot, Cyrillic, Deseret, Devanagari, Egyp-
3843 tian_Hieroglyphs, Ethiopic, Georgian, Glagolitic, Gothic, Greek,
3844 Gujarati, Gurmukhi, Han, Hangul, Hanunoo, Hebrew, Hiragana, Impe-
3845 rial_Aramaic, Inherited, Inscriptional_Pahlavi, Inscriptional_Parthian,
3846 Javanese, Kaithi, Kannada, Katakana, Kayah_Li, Kharoshthi, Khmer, Lao,
3847 Latin, Lepcha, Limbu, Linear_B, Lisu, Lycian, Lydian, Malayalam,
3848 Meetei_Mayek, Mongolian, Myanmar, New_Tai_Lue, Nko, Ogham, Old_Italic,
3849 Old_Persian, Old_South_Arabian, Old_Turkic, Ol_Chiki, Oriya, Osmanya,
3850 Phags_Pa, Phoenician, Rejang, Runic, Samaritan, Saurashtra, Shavian,
3851 Sinhala, Sundanese, Syloti_Nagri, Syriac, Tagalog, Tagbanwa, Tai_Le,
3852 Tai_Tham, Tai_Viet, Tamil, Telugu, Thaana, Thai, Tibetan, Tifinagh,
3853 Ugaritic, Vai, Yi.
3855 Each character has exactly one Unicode general category property, spec-
3856 ified by a two-letter abbreviation. For compatibility with Perl, nega-
3857 tion can be specified by including a circumflex between the opening
3858 brace and the property name. For example, \p{^Lu} is the same as
3859 \P{Lu}.
3861 If only one letter is specified with \p or \P, it includes all the gen-
3862 eral category properties that start with that letter. In this case, in
3863 the absence of negation, the curly brackets in the escape sequence are
3864 optional; these two examples have the same effect:
3866 \p{L}
3867 \pL
3869 The following general category property codes are supported:
3871 C Other
3872 Cc Control
3873 Cf Format
3874 Cn Unassigned
3875 Co Private use
3876 Cs Surrogate
3878 L Letter
3879 Ll Lower case letter
3880 Lm Modifier letter
3881 Lo Other letter
3882 Lt Title case letter
3883 Lu Upper case letter
3885 M Mark
3886 Mc Spacing mark
3887 Me Enclosing mark
3888 Mn Non-spacing mark
3890 N Number
3891 Nd Decimal number
3892 Nl Letter number
3893 No Other number
3895 P Punctuation
3896 Pc Connector punctuation
3897 Pd Dash punctuation
3898 Pe Close punctuation
3899 Pf Final punctuation
3900 Pi Initial punctuation
3901 Po Other punctuation
3902 Ps Open punctuation
3904 S Symbol
3905 Sc Currency symbol
3906 Sk Modifier symbol
3907 Sm Mathematical symbol
3908 So Other symbol
3910 Z Separator
3911 Zl Line separator
3912 Zp Paragraph separator
3913 Zs Space separator
3915 The special property L& is also supported: it matches a character that
3916 has the Lu, Ll, or Lt property, in other words, a letter that is not
3917 classified as a modifier or "other".
3919 The Cs (Surrogate) property applies only to characters in the range
3920 U+D800 to U+DFFF. Such characters are not valid in UTF-8 strings (see
3921 RFC 3629) and so cannot be tested by PCRE, unless UTF-8 validity check-
3922 ing has been turned off (see the discussion of PCRE_NO_UTF8_CHECK in
3923 the pcreapi page). Perl does not support the Cs property.
3925 The long synonyms for property names that Perl supports (such as
3926 \p{Letter}) are not supported by PCRE, nor is it permitted to prefix
3927 any of these properties with "Is".
3929 No character that is in the Unicode table has the Cn (unassigned) prop-
3930 erty. Instead, this property is assumed for any code point that is not
3931 in the Unicode table.
3933 Specifying caseless matching does not affect these escape sequences.
3934 For example, \p{Lu} always matches only upper case letters.
3936 The \X escape matches any number of Unicode characters that form an
3937 extended Unicode sequence. \X is equivalent to
3939 (?>\PM\pM*)
3941 That is, it matches a character without the "mark" property, followed
3942 by zero or more characters with the "mark" property, and treats the
3943 sequence as an atomic group (see below). Characters with the "mark"
3944 property are typically accents that affect the preceding character.
3945 None of them have codepoints less than 256, so in non-UTF-8 mode \X
3946 matches any one character.
3948 Note that recent versions of Perl have changed \X to match what Unicode
3949 calls an "extended grapheme cluster", which has a more complicated def-
3950 inition.
3952 Matching characters by Unicode property is not fast, because PCRE has
3953 to search a structure that contains data for over fifteen thousand
3954 characters. That is why the traditional escape sequences such as \d and
3955 \w do not use Unicode properties in PCRE by default, though you can
3956 make them do so by setting the PCRE_UCP option for pcre_compile() or by
3957 starting the pattern with (*UCP).
3959 PCRE's additional properties
3961 As well as the standard Unicode properties described in the previous
3962 section, PCRE supports four more that make it possible to convert tra-
3963 ditional escape sequences such as \w and \s and POSIX character classes
3964 to use Unicode properties. PCRE uses these non-standard, non-Perl prop-
3965 erties internally when PCRE_UCP is set. They are:
3967 Xan Any alphanumeric character
3968 Xps Any POSIX space character
3969 Xsp Any Perl space character
3970 Xwd Any Perl "word" character
3972 Xan matches characters that have either the L (letter) or the N (num-
3973 ber) property. Xps matches the characters tab, linefeed, vertical tab,
3974 formfeed, or carriage return, and any other character that has the Z
3975 (separator) property. Xsp is the same as Xps, except that vertical tab
3976 is excluded. Xwd matches the same characters as Xan, plus underscore.
3978 Resetting the match start
3980 The escape sequence \K causes any previously matched characters not to
3981 be included in the final matched sequence. For example, the pattern:
3983 foo\Kbar
3985 matches "foobar", but reports that it has matched "bar". This feature
3986 is similar to a lookbehind assertion (described below). However, in
3987 this case, the part of the subject before the real match does not have
3988 to be of fixed length, as lookbehind assertions do. The use of \K does
3989 not interfere with the setting of captured substrings. For example,
3990 when the pattern
3992 (foo)\Kbar
3994 matches "foobar", the first substring is still set to "foo".
3996 Perl documents that the use of \K within assertions is "not well
3997 defined". In PCRE, \K is acted upon when it occurs inside positive
3998 assertions, but is ignored in negative assertions.
4000 Simple assertions
4002 The final use of backslash is for certain simple assertions. An asser-
4003 tion specifies a condition that has to be met at a particular point in
4004 a match, without consuming any characters from the subject string. The
4005 use of subpatterns for more complicated assertions is described below.
4006 The backslashed assertions are:
4008 \b matches at a word boundary
4009 \B matches when not at a word boundary
4010 \A matches at the start of the subject
4011 \Z matches at the end of the subject
4012 also matches before a newline at the end of the subject
4013 \z matches only at the end of the subject
4014 \G matches at the first matching position in the subject
4016 Inside a character class, \b has a different meaning; it matches the
4017 backspace character. If any other of these assertions appears in a
4018 character class, by default it matches the corresponding literal char-
4019 acter (for example, \B matches the letter B). However, if the
4020 PCRE_EXTRA option is set, an "invalid escape sequence" error is gener-
4021 ated instead.
4023 A word boundary is a position in the subject string where the current
4024 character and the previous character do not both match \w or \W (i.e.
4025 one matches \w and the other matches \W), or the start or end of the
4026 string if the first or last character matches \w, respectively. In
4027 UTF-8 mode, the meanings of \w and \W can be changed by setting the
4028 PCRE_UCP option. When this is done, it also affects \b and \B. Neither
4029 PCRE nor Perl has a separate "start of word" or "end of word" metase-
4030 quence. However, whatever follows \b normally determines which it is.
4031 For example, the fragment \ba matches "a" at the start of a word.
4033 The \A, \Z, and \z assertions differ from the traditional circumflex
4034 and dollar (described in the next section) in that they only ever match
4035 at the very start and end of the subject string, whatever options are
4036 set. Thus, they are independent of multiline mode. These three asser-
4037 tions are not affected by the PCRE_NOTBOL or PCRE_NOTEOL options, which
4038 affect only the behaviour of the circumflex and dollar metacharacters.
4039 However, if the startoffset argument of pcre_exec() is non-zero, indi-
4040 cating that matching is to start at a point other than the beginning of
4041 the subject, \A can never match. The difference between \Z and \z is
4042 that \Z matches before a newline at the end of the string as well as at
4043 the very end, whereas \z matches only at the end.
4045 The \G assertion is true only when the current matching position is at
4046 the start point of the match, as specified by the startoffset argument
4047 of pcre_exec(). It differs from \A when the value of startoffset is
4048 non-zero. By calling pcre_exec() multiple times with appropriate argu-
4049 ments, you can mimic Perl's /g option, and it is in this kind of imple-
4050 mentation where \G can be useful.
4052 Note, however, that PCRE's interpretation of \G, as the start of the
4053 current match, is subtly different from Perl's, which defines it as the
4054 end of the previous match. In Perl, these can be different when the
4055 previously matched string was empty. Because PCRE does just one match
4056 at a time, it cannot reproduce this behaviour.
4058 If all the alternatives of a pattern begin with \G, the expression is
4059 anchored to the starting match position, and the "anchored" flag is set
4060 in the compiled regular expression.
4065 Outside a character class, in the default matching mode, the circumflex
4066 character is an assertion that is true only if the current matching
4067 point is at the start of the subject string. If the startoffset argu-
4068 ment of pcre_exec() is non-zero, circumflex can never match if the
4069 PCRE_MULTILINE option is unset. Inside a character class, circumflex
4070 has an entirely different meaning (see below).
4072 Circumflex need not be the first character of the pattern if a number
4073 of alternatives are involved, but it should be the first thing in each
4074 alternative in which it appears if the pattern is ever to match that
4075 branch. If all possible alternatives start with a circumflex, that is,
4076 if the pattern is constrained to match only at the start of the sub-
4077 ject, it is said to be an "anchored" pattern. (There are also other
4078 constructs that can cause a pattern to be anchored.)
4080 A dollar character is an assertion that is true only if the current
4081 matching point is at the end of the subject string, or immediately
4082 before a newline at the end of the string (by default). Dollar need not
4083 be the last character of the pattern if a number of alternatives are
4084 involved, but it should be the last item in any branch in which it
4085 appears. Dollar has no special meaning in a character class.
4087 The meaning of dollar can be changed so that it matches only at the
4088 very end of the string, by setting the PCRE_DOLLAR_ENDONLY option at
4089 compile time. This does not affect the \Z assertion.
4091 The meanings of the circumflex and dollar characters are changed if the
4092 PCRE_MULTILINE option is set. When this is the case, a circumflex
4093 matches immediately after internal newlines as well as at the start of
4094 the subject string. It does not match after a newline that ends the
4095 string. A dollar matches before any newlines in the string, as well as
4096 at the very end, when PCRE_MULTILINE is set. When newline is specified
4097 as the two-character sequence CRLF, isolated CR and LF characters do
4098 not indicate newlines.
4100 For example, the pattern /^abc$/ matches the subject string "def\nabc"
4101 (where \n represents a newline) in multiline mode, but not otherwise.
4102 Consequently, patterns that are anchored in single line mode because
4103 all branches start with ^ are not anchored in multiline mode, and a
4104 match for circumflex is possible when the startoffset argument of
4105 pcre_exec() is non-zero. The PCRE_DOLLAR_ENDONLY option is ignored if
4106 PCRE_MULTILINE is set.
4108 Note that the sequences \A, \Z, and \z can be used to match the start
4109 and end of the subject in both modes, and if all branches of a pattern
4110 start with \A it is always anchored, whether or not PCRE_MULTILINE is
4111 set.
4116 Outside a character class, a dot in the pattern matches any one charac-
4117 ter in the subject string except (by default) a character that signi-
4118 fies the end of a line. In UTF-8 mode, the matched character may be
4119 more than one byte long.
4121 When a line ending is defined as a single character, dot never matches
4122 that character; when the two-character sequence CRLF is used, dot does
4123 not match CR if it is immediately followed by LF, but otherwise it
4124 matches all characters (including isolated CRs and LFs). When any Uni-
4125 code line endings are being recognized, dot does not match CR or LF or
4126 any of the other line ending characters.
4128 The behaviour of dot with regard to newlines can be changed. If the
4129 PCRE_DOTALL option is set, a dot matches any one character, without
4130 exception. If the two-character sequence CRLF is present in the subject
4131 string, it takes two dots to match it.
4133 The handling of dot is entirely independent of the handling of circum-
4134 flex and dollar, the only relationship being that they both involve
4135 newlines. Dot has no special meaning in a character class.
4137 The escape sequence \N behaves like a dot, except that it is not
4138 affected by the PCRE_DOTALL option. In other words, it matches any
4139 character except one that signifies the end of a line.
4144 Outside a character class, the escape sequence \C matches any one byte,
4145 both in and out of UTF-8 mode. Unlike a dot, it always matches line-
4146 ending characters. The feature is provided in Perl in order to match
4147 individual bytes in UTF-8 mode, but it is unclear how it can usefully
4148 be used. Because \C breaks up characters into individual bytes, match-
4149 ing one byte with \C in UTF-8 mode means that the rest of the string
4150 may start with a malformed UTF-8 character. This has undefined results,
4151 because PCRE assumes that it is dealing with valid UTF-8 strings (and
4152 by default it checks this at the start of processing unless the
4153 PCRE_NO_UTF8_CHECK option is used).
4155 PCRE does not allow \C to appear in lookbehind assertions (described
4156 below), because in UTF-8 mode this would make it impossible to calcu-
4157 late the length of the lookbehind.
4159 In general, the \C escape sequence is best avoided in UTF-8 mode. How-
4160 ever, one way of using it that avoids the problem of malformed UTF-8
4161 characters is to use a lookahead to check the length of the next char-
4162 acter, as in this pattern (ignore white space and line breaks):
4164 (?| (?=[\x00-\x7f])(\C) |
4165 (?=[\x80-\x{7ff}])(\C)(\C) |
4166 (?=[\x{800}-\x{ffff}])(\C)(\C)(\C) |
4167 (?=[\x{10000}-\x{1fffff}])(\C)(\C)(\C)(\C))
4169 A group that starts with (?| resets the capturing parentheses numbers
4170 in each alternative (see "Duplicate Subpattern Numbers" below). The
4171 assertions at the start of each branch check the next UTF-8 character
4172 for values whose encoding uses 1, 2, 3, or 4 bytes, respectively. The
4173 character's individual bytes are then captured by the appropriate num-
4174 ber of groups.
4179 An opening square bracket introduces a character class, terminated by a
4180 closing square bracket. A closing square bracket on its own is not spe-
4181 cial by default. However, if the PCRE_JAVASCRIPT_COMPAT option is set,
4182 a lone closing square bracket causes a compile-time error. If a closing
4183 square bracket is required as a member of the class, it should be the
4184 first data character in the class (after an initial circumflex, if
4185 present) or escaped with a backslash.
4187 A character class matches a single character in the subject. In UTF-8
4188 mode, the character may be more than one byte long. A matched character
4189 must be in the set of characters defined by the class, unless the first
4190 character in the class definition is a circumflex, in which case the
4191 subject character must not be in the set defined by the class. If a
4192 circumflex is actually required as a member of the class, ensure it is
4193 not the first character, or escape it with a backslash.
4195 For example, the character class [aeiou] matches any lower case vowel,
4196 while [^aeiou] matches any character that is not a lower case vowel.
4197 Note that a circumflex is just a convenient notation for specifying the
4198 characters that are in the class by enumerating those that are not. A
4199 class that starts with a circumflex is not an assertion; it still con-
4200 sumes a character from the subject string, and therefore it fails if
4201 the current pointer is at the end of the string.
4203 In UTF-8 mode, characters with values greater than 255 can be included
4204 in a class as a literal string of bytes, or by using the \x{ escaping
4205 mechanism.
4207 When caseless matching is set, any letters in a class represent both
4208 their upper case and lower case versions, so for example, a caseless
4209 [aeiou] matches "A" as well as "a", and a caseless [^aeiou] does not
4210 match "A", whereas a caseful version would. In UTF-8 mode, PCRE always
4211 understands the concept of case for characters whose values are less
4212 than 128, so caseless matching is always possible. For characters with
4213 higher values, the concept of case is supported if PCRE is compiled
4214 with Unicode property support, but not otherwise. If you want to use
4215 caseless matching in UTF8-mode for characters 128 and above, you must
4216 ensure that PCRE is compiled with Unicode property support as well as
4217 with UTF-8 support.
4219 Characters that might indicate line breaks are never treated in any
4220 special way when matching character classes, whatever line-ending
4221 sequence is in use, and whatever setting of the PCRE_DOTALL and
4222 PCRE_MULTILINE options is used. A class such as [^a] always matches one
4223 of these characters.
4225 The minus (hyphen) character can be used to specify a range of charac-
4226 ters in a character class. For example, [d-m] matches any letter
4227 between d and m, inclusive. If a minus character is required in a
4228 class, it must be escaped with a backslash or appear in a position
4229 where it cannot be interpreted as indicating a range, typically as the
4230 first or last character in the class.
4232 It is not possible to have the literal character "]" as the end charac-
4233 ter of a range. A pattern such as [W-]46] is interpreted as a class of
4234 two characters ("W" and "-") followed by a literal string "46]", so it
4235 would match "W46]" or "-46]". However, if the "]" is escaped with a
4236 backslash it is interpreted as the end of range, so [W-\]46] is inter-
4237 preted as a class containing a range followed by two other characters.
4238 The octal or hexadecimal representation of "]" can also be used to end
4239 a range.
4241 Ranges operate in the collating sequence of character values. They can
4242 also be used for characters specified numerically, for example
4243 [\000-\037]. In UTF-8 mode, ranges can include characters whose values
4244 are greater than 255, for example [\x{100}-\x{2ff}].
4246 If a range that includes letters is used when caseless matching is set,
4247 it matches the letters in either case. For example, [W-c] is equivalent
4248 to [][\\^_`wxyzabc], matched caselessly, and in non-UTF-8 mode, if
4249 character tables for a French locale are in use, [\xc8-\xcb] matches
4250 accented E characters in both cases. In UTF-8 mode, PCRE supports the
4251 concept of case for characters with values greater than 128 only when
4252 it is compiled with Unicode property support.
4254 The character escape sequences \d, \D, \h, \H, \p, \P, \s, \S, \v, \V,
4255 \w, and \W may appear in a character class, and add the characters that
4256 they match to the class. For example, [\dABCDEF] matches any hexadeci-
4257 mal digit. In UTF-8 mode, the PCRE_UCP option affects the meanings of
4258 \d, \s, \w and their upper case partners, just as it does when they
4259 appear outside a character class, as described in the section entitled
4260 "Generic character types" above. The escape sequence \b has a different
4261 meaning inside a character class; it matches the backspace character.
4262 The sequences \B, \N, \R, and \X are not special inside a character
4263 class. Like any other unrecognized escape sequences, they are treated
4264 as the literal characters "B", "N", "R", and "X" by default, but cause
4265 an error if the PCRE_EXTRA option is set.
4267 A circumflex can conveniently be used with the upper case character
4268 types to specify a more restricted set of characters than the matching
4269 lower case type. For example, the class [^\W_] matches any letter or
4270 digit, but not underscore, whereas [\w] includes underscore. A positive
4271 character class should be read as "something OR something OR ..." and a
4272 negative class as "NOT something AND NOT something AND NOT ...".
4274 The only metacharacters that are recognized in character classes are
4275 backslash, hyphen (only where it can be interpreted as specifying a
4276 range), circumflex (only at the start), opening square bracket (only
4277 when it can be interpreted as introducing a POSIX class name - see the
4278 next section), and the terminating closing square bracket. However,
4279 escaping other non-alphanumeric characters does no harm.
4284 Perl supports the POSIX notation for character classes. This uses names
4285 enclosed by [: and :] within the enclosing square brackets. PCRE also
4286 supports this notation. For example,
4288 [01[:alpha:]%]
4290 matches "0", "1", any alphabetic character, or "%". The supported class
4291 names are:
4293 alnum letters and digits
4294 alpha letters
4295 ascii character codes 0 - 127
4296 blank space or tab only
4297 cntrl control characters
4298 digit decimal digits (same as \d)
4299 graph printing characters, excluding space
4300 lower lower case letters
4301 print printing characters, including space
4302 punct printing characters, excluding letters and digits and space
4303 space white space (not quite the same as \s)
4304 upper upper case letters
4305 word "word" characters (same as \w)
4306 xdigit hexadecimal digits
4308 The "space" characters are HT (9), LF (10), VT (11), FF (12), CR (13),
4309 and space (32). Notice that this list includes the VT character (code
4310 11). This makes "space" different to \s, which does not include VT (for
4311 Perl compatibility).
4313 The name "word" is a Perl extension, and "blank" is a GNU extension
4314 from Perl 5.8. Another Perl extension is negation, which is indicated
4315 by a ^ character after the colon. For example,
4317 [12[:^digit:]]
4319 matches "1", "2", or any non-digit. PCRE (and Perl) also recognize the
4320 POSIX syntax [.ch.] and [=ch=] where "ch" is a "collating element", but
4321 these are not supported, and an error is given if they are encountered.
4323 By default, in UTF-8 mode, characters with values greater than 128 do
4324 not match any of the POSIX character classes. However, if the PCRE_UCP
4325 option is passed to pcre_compile(), some of the classes are changed so
4326 that Unicode character properties are used. This is achieved by replac-
4327 ing the POSIX classes by other sequences, as follows:
4329 [:alnum:] becomes \p{Xan}
4330 [:alpha:] becomes \p{L}
4331 [:blank:] becomes \h
4332 [:digit:] becomes \p{Nd}
4333 [:lower:] becomes \p{Ll}
4334 [:space:] becomes \p{Xps}
4335 [:upper:] becomes \p{Lu}
4336 [:word:] becomes \p{Xwd}
4338 Negated versions, such as [:^alpha:] use \P instead of \p. The other
4339 POSIX classes are unchanged, and match only characters with code points
4340 less than 128.
4345 Vertical bar characters are used to separate alternative patterns. For
4346 example, the pattern
4348 gilbert|sullivan
4350 matches either "gilbert" or "sullivan". Any number of alternatives may
4351 appear, and an empty alternative is permitted (matching the empty
4352 string). The matching process tries each alternative in turn, from left
4353 to right, and the first one that succeeds is used. If the alternatives
4354 are within a subpattern (defined below), "succeeds" means matching the
4355 rest of the main pattern as well as the alternative in the subpattern.
4360 The settings of the PCRE_CASELESS, PCRE_MULTILINE, PCRE_DOTALL, and
4361 PCRE_EXTENDED options (which are Perl-compatible) can be changed from
4362 within the pattern by a sequence of Perl option letters enclosed
4363 between "(?" and ")". The option letters are
4365 i for PCRE_CASELESS
4367 s for PCRE_DOTALL
4368 x for PCRE_EXTENDED
4370 For example, (?im) sets caseless, multiline matching. It is also possi-
4371 ble to unset these options by preceding the letter with a hyphen, and a
4372 combined setting and unsetting such as (?im-sx), which sets PCRE_CASE-
4374 is also permitted. If a letter appears both before and after the
4375 hyphen, the option is unset.
4378 can be changed in the same way as the Perl-compatible options by using
4379 the characters J, U and X respectively.
4381 When one of these option changes occurs at top level (that is, not
4382 inside subpattern parentheses), the change applies to the remainder of
4383 the pattern that follows. If the change is placed right at the start of
4384 a pattern, PCRE extracts it into the global options (and it will there-
4385 fore show up in data extracted by the pcre_fullinfo() function).
4387 An option change within a subpattern (see below for a description of
4388 subpatterns) affects only that part of the subpattern that follows it,
4389 so
4391 (a(?i)b)c
4393 matches abc and aBc and no other strings (assuming PCRE_CASELESS is not
4394 used). By this means, options can be made to have different settings
4395 in different parts of the pattern. Any changes made in one alternative
4396 do carry on into subsequent branches within the same subpattern. For
4397 example,
4399 (a(?i)b|c)
4401 matches "ab", "aB", "c", and "C", even though when matching "C" the
4402 first branch is abandoned before the option setting. This is because
4403 the effects of option settings happen at compile time. There would be
4404 some very weird behaviour otherwise.
4406 Note: There are other PCRE-specific options that can be set by the
4407 application when the compile or match functions are called. In some
4408 cases the pattern can contain special leading sequences such as (*CRLF)
4409 to override what the application has set or what has been defaulted.
4410 Details are given in the section entitled "Newline sequences" above.
4411 There are also the (*UTF8) and (*UCP) leading sequences that can be
4412 used to set UTF-8 and Unicode property modes; they are equivalent to
4413 setting the PCRE_UTF8 and the PCRE_UCP options, respectively.
4418 Subpatterns are delimited by parentheses (round brackets), which can be
4419 nested. Turning part of a pattern into a subpattern does two things:
4421 1. It localizes a set of alternatives. For example, the pattern
4423 cat(aract|erpillar|)
4425 matches "cataract", "caterpillar", or "cat". Without the parentheses,
4426 it would match "cataract", "erpillar" or an empty string.
4428 2. It sets up the subpattern as a capturing subpattern. This means
4429 that, when the whole pattern matches, that portion of the subject
4430 string that matched the subpattern is passed back to the caller via the
4431 ovector argument of pcre_exec(). Opening parentheses are counted from
4432 left to right (starting from 1) to obtain numbers for the capturing
4433 subpatterns. For example, if the string "the red king" is matched
4434 against the pattern
4436 the ((red|white) (king|queen))
4438 the captured substrings are "red king", "red", and "king", and are num-
4439 bered 1, 2, and 3, respectively.
4441 The fact that plain parentheses fulfil two functions is not always
4442 helpful. There are often times when a grouping subpattern is required
4443 without a capturing requirement. If an opening parenthesis is followed
4444 by a question mark and a colon, the subpattern does not do any captur-
4445 ing, and is not counted when computing the number of any subsequent
4446 capturing subpatterns. For example, if the string "the white queen" is
4447 matched against the pattern
4449 the ((?:red|white) (king|queen))
4451 the captured substrings are "white queen" and "queen", and are numbered
4452 1 and 2. The maximum number of capturing subpatterns is 65535.
4454 As a convenient shorthand, if any option settings are required at the
4455 start of a non-capturing subpattern, the option letters may appear
4456 between the "?" and the ":". Thus the two patterns
4458 (?i:saturday|sunday)
4459 (?:(?i)saturday|sunday)
4461 match exactly the same set of strings. Because alternative branches are
4462 tried from left to right, and options are not reset until the end of
4463 the subpattern is reached, an option setting in one branch does affect
4464 subsequent branches, so the above patterns match "SUNDAY" as well as
4465 "Saturday".
4470 Perl 5.10 introduced a feature whereby each alternative in a subpattern
4471 uses the same numbers for its capturing parentheses. Such a subpattern
4472 starts with (?| and is itself a non-capturing subpattern. For example,
4473 consider this pattern:
4475 (?|(Sat)ur|(Sun))day
4477 Because the two alternatives are inside a (?| group, both sets of cap-
4478 turing parentheses are numbered one. Thus, when the pattern matches,
4479 you can look at captured substring number one, whichever alternative
4480 matched. This construct is useful when you want to capture part, but
4481 not all, of one of a number of alternatives. Inside a (?| group, paren-
4482 theses are numbered as usual, but the number is reset at the start of
4483 each branch. The numbers of any capturing parentheses that follow the
4484 subpattern start after the highest number used in any branch. The fol-
4485 lowing example is taken from the Perl documentation. The numbers under-
4486 neath show in which buffer the captured content will be stored.
4488 # before ---------------branch-reset----------- after
4489 / ( a ) (?| x ( y ) z | (p (q) r) | (t) u (v) ) ( z ) /x
4490 # 1 2 2 3 2 3 4
4492 A back reference to a numbered subpattern uses the most recent value
4493 that is set for that number by any subpattern. The following pattern
4494 matches "abcabc" or "defdef":
4496 /(?|(abc)|(def))\1/
4498 In contrast, a subroutine call to a numbered subpattern always refers
4499 to the first one in the pattern with the given number. The following
4500 pattern matches "abcabc" or "defabc":
4502 /(?|(abc)|(def))(?1)/
4504 If a condition test for a subpattern's having matched refers to a non-
4505 unique number, the test is true if any of the subpatterns of that num-
4506 ber have matched.
4508 An alternative approach to using this "branch reset" feature is to use
4509 duplicate named subpatterns, as described in the next section.
4514 Identifying capturing parentheses by number is simple, but it can be
4515 very hard to keep track of the numbers in complicated regular expres-
4516 sions. Furthermore, if an expression is modified, the numbers may
4517 change. To help with this difficulty, PCRE supports the naming of sub-
4518 patterns. This feature was not added to Perl until release 5.10. Python
4519 had the feature earlier, and PCRE introduced it at release 4.0, using
4520 the Python syntax. PCRE now supports both the Perl and the Python syn-
4521 tax. Perl allows identically numbered subpatterns to have different
4522 names, but PCRE does not.
4524 In PCRE, a subpattern can be named in one of three ways: (?<name>...)
4525 or (?'name'...) as in Perl, or (?P<name>...) as in Python. References
4526 to capturing parentheses from other parts of the pattern, such as back
4527 references, recursion, and conditions, can be made by name as well as
4528 by number.
4530 Names consist of up to 32 alphanumeric characters and underscores.
4531 Named capturing parentheses are still allocated numbers as well as
4532 names, exactly as if the names were not present. The PCRE API provides
4533 function calls for extracting the name-to-number translation table from
4534 a compiled pattern. There is also a convenience function for extracting
4535 a captured substring by name.
4537 By default, a name must be unique within a pattern, but it is possible
4538 to relax this constraint by setting the PCRE_DUPNAMES option at compile
4539 time. (Duplicate names are also always permitted for subpatterns with
4540 the same number, set up as described in the previous section.) Dupli-
4541 cate names can be useful for patterns where only one instance of the
4542 named parentheses can match. Suppose you want to match the name of a
4543 weekday, either as a 3-letter abbreviation or as the full name, and in
4544 both cases you want to extract the abbreviation. This pattern (ignoring
4545 the line breaks) does the job:
4547 (?<DN>Mon|Fri|Sun)(?:day)?|
4548 (?<DN>Tue)(?:sday)?|
4549 (?<DN>Wed)(?:nesday)?|
4550 (?<DN>Thu)(?:rsday)?|
4551 (?<DN>Sat)(?:urday)?
4553 There are five capturing substrings, but only one is ever set after a
4554 match. (An alternative way of solving this problem is to use a "branch
4555 reset" subpattern, as described in the previous section.)
4557 The convenience function for extracting the data by name returns the
4558 substring for the first (and in this example, the only) subpattern of
4559 that name that matched. This saves searching to find which numbered
4560 subpattern it was.
4562 If you make a back reference to a non-unique named subpattern from
4563 elsewhere in the pattern, the one that corresponds to the first occur-
4564 rence of the name is used. In the absence of duplicate numbers (see the
4565 previous section) this is the one with the lowest number. If you use a
4566 named reference in a condition test (see the section about conditions
4567 below), either to check whether a subpattern has matched, or to check
4568 for recursion, all subpatterns with the same name are tested. If the
4569 condition is true for any one of them, the overall condition is true.
4570 This is the same behaviour as testing by number. For further details of
4571 the interfaces for handling named subpatterns, see the pcreapi documen-
4572 tation.
4574 Warning: You cannot use different names to distinguish between two sub-
4575 patterns with the same number because PCRE uses only the numbers when
4576 matching. For this reason, an error is given at compile time if differ-
4577 ent names are given to subpatterns with the same number. However, you
4578 can give the same name to subpatterns with the same number, even when
4579 PCRE_DUPNAMES is not set.
4584 Repetition is specified by quantifiers, which can follow any of the
4585 following items:
4587 a literal data character
4588 the dot metacharacter
4589 the \C escape sequence
4590 the \X escape sequence (in UTF-8 mode with Unicode properties)
4591 the \R escape sequence
4592 an escape such as \d or \pL that matches a single character
4593 a character class
4594 a back reference (see next section)
4595 a parenthesized subpattern (including assertions)
4596 a subroutine call to a subpattern (recursive or otherwise)
4598 The general repetition quantifier specifies a minimum and maximum num-
4599 ber of permitted matches, by giving the two numbers in curly brackets
4600 (braces), separated by a comma. The numbers must be less than 65536,
4601 and the first must be less than or equal to the second. For example:
4603 z{2,4}
4605 matches "zz", "zzz", or "zzzz". A closing brace on its own is not a
4606 special character. If the second number is omitted, but the comma is
4607 present, there is no upper limit; if the second number and the comma
4608 are both omitted, the quantifier specifies an exact number of required
4609 matches. Thus
4611 [aeiou]{3,}
4613 matches at least 3 successive vowels, but may match many more, while
4615 \d{8}
4617 matches exactly 8 digits. An opening curly bracket that appears in a
4618 position where a quantifier is not allowed, or one that does not match
4619 the syntax of a quantifier, is taken as a literal character. For exam-
4620 ple, {,6} is not a quantifier, but a literal string of four characters.
4622 In UTF-8 mode, quantifiers apply to UTF-8 characters rather than to
4623 individual bytes. Thus, for example, \x{100}{2} matches two UTF-8 char-
4624 acters, each of which is represented by a two-byte sequence. Similarly,
4625 when Unicode property support is available, \X{3} matches three Unicode
4626 extended sequences, each of which may be several bytes long (and they
4627 may be of different lengths).
4629 The quantifier {0} is permitted, causing the expression to behave as if
4630 the previous item and the quantifier were not present. This may be use-
4631 ful for subpatterns that are referenced as subroutines from elsewhere
4632 in the pattern (but see also the section entitled "Defining subpatterns
4633 for use by reference only" below). Items other than subpatterns that
4634 have a {0} quantifier are omitted from the compiled pattern.
4636 For convenience, the three most common quantifiers have single-charac-
4637 ter abbreviations:
4639 * is equivalent to {0,}
4640 + is equivalent to {1,}
4641 ? is equivalent to {0,1}
4643 It is possible to construct infinite loops by following a subpattern
4644 that can match no characters with a quantifier that has no upper limit,
4645 for example:
4647 (a?)*
4649 Earlier versions of Perl and PCRE used to give an error at compile time
4650 for such patterns. However, because there are cases where this can be
4651 useful, such patterns are now accepted, but if any repetition of the
4652 subpattern does in fact match no characters, the loop is forcibly bro-
4653 ken.
4655 By default, the quantifiers are "greedy", that is, they match as much
4656 as possible (up to the maximum number of permitted times), without
4657 causing the rest of the pattern to fail. The classic example of where
4658 this gives problems is in trying to match comments in C programs. These
4659 appear between /* and */ and within the comment, individual * and /
4660 characters may appear. An attempt to match C comments by applying the
4661 pattern
4663 /\*.*\*/
4665 to the string
4667 /* first comment */ not comment /* second comment */
4669 fails, because it matches the entire string owing to the greediness of
4670 the .* item.
4672 However, if a quantifier is followed by a question mark, it ceases to
4673 be greedy, and instead matches the minimum number of times possible, so
4674 the pattern
4676 /\*.*?\*/
4678 does the right thing with the C comments. The meaning of the various
4679 quantifiers is not otherwise changed, just the preferred number of
4680 matches. Do not confuse this use of question mark with its use as a
4681 quantifier in its own right. Because it has two uses, it can sometimes
4682 appear doubled, as in
4684 \d??\d
4686 which matches one digit by preference, but can match two if that is the
4687 only way the rest of the pattern matches.
4689 If the PCRE_UNGREEDY option is set (an option that is not available in
4690 Perl), the quantifiers are not greedy by default, but individual ones
4691 can be made greedy by following them with a question mark. In other
4692 words, it inverts the default behaviour.
4694 When a parenthesized subpattern is quantified with a minimum repeat
4695 count that is greater than 1 or with a limited maximum, more memory is
4696 required for the compiled pattern, in proportion to the size of the
4697 minimum or maximum.
4699 If a pattern starts with .* or .{0,} and the PCRE_DOTALL option (equiv-
4700 alent to Perl's /s) is set, thus allowing the dot to match newlines,
4701 the pattern is implicitly anchored, because whatever follows will be
4702 tried against every character position in the subject string, so there
4703 is no point in retrying the overall match at any position after the
4704 first. PCRE normally treats such a pattern as though it were preceded
4705 by \A.
4707 In cases where it is known that the subject string contains no new-
4708 lines, it is worth setting PCRE_DOTALL in order to obtain this opti-
4709 mization, or alternatively using ^ to indicate anchoring explicitly.
4711 However, there is one situation where the optimization cannot be used.
4712 When .* is inside capturing parentheses that are the subject of a back
4713 reference elsewhere in the pattern, a match at the start may fail where
4714 a later one succeeds. Consider, for example:
4716 (.*)abc\1
4718 If the subject is "xyz123abc123" the match point is the fourth charac-
4719 ter. For this reason, such a pattern is not implicitly anchored.
4721 When a capturing subpattern is repeated, the value captured is the sub-
4722 string that matched the final iteration. For example, after
4724 (tweedle[dume]{3}\s*)+
4726 has matched "tweedledum tweedledee" the value of the captured substring
4727 is "tweedledee". However, if there are nested capturing subpatterns,
4728 the corresponding captured values may have been set in previous itera-
4729 tions. For example, after
4731 /(a|(b))+/
4733 matches "aba" the value of the second captured substring is "b".
4738 With both maximizing ("greedy") and minimizing ("ungreedy" or "lazy")
4739 repetition, failure of what follows normally causes the repeated item
4740 to be re-evaluated to see if a different number of repeats allows the
4741 rest of the pattern to match. Sometimes it is useful to prevent this,
4742 either to change the nature of the match, or to cause it fail earlier
4743 than it otherwise might, when the author of the pattern knows there is
4744 no point in carrying on.
4746 Consider, for example, the pattern \d+foo when applied to the subject
4747 line
4749 123456bar
4751 After matching all 6 digits and then failing to match "foo", the normal
4752 action of the matcher is to try again with only 5 digits matching the
4753 \d+ item, and then with 4, and so on, before ultimately failing.
4754 "Atomic grouping" (a term taken from Jeffrey Friedl's book) provides
4755 the means for specifying that once a subpattern has matched, it is not
4756 to be re-evaluated in this way.
4758 If we use atomic grouping for the previous example, the matcher gives
4759 up immediately on failing to match "foo" the first time. The notation
4760 is a kind of special parenthesis, starting with (?> as in this example:
4762 (?>\d+)foo
4764 This kind of parenthesis "locks up" the part of the pattern it con-
4765 tains once it has matched, and a failure further into the pattern is
4766 prevented from backtracking into it. Backtracking past it to previous
4767 items, however, works as normal.
4769 An alternative description is that a subpattern of this type matches
4770 the string of characters that an identical standalone pattern would
4771 match, if anchored at the current point in the subject string.
4773 Atomic grouping subpatterns are not capturing subpatterns. Simple cases
4774 such as the above example can be thought of as a maximizing repeat that
4775 must swallow everything it can. So, while both \d+ and \d+? are pre-
4776 pared to adjust the number of digits they match in order to make the
4777 rest of the pattern match, (?>\d+) can only match an entire sequence of
4778 digits.
4780 Atomic groups in general can of course contain arbitrarily complicated
4781 subpatterns, and can be nested. However, when the subpattern for an
4782 atomic group is just a single repeated item, as in the example above, a
4783 simpler notation, called a "possessive quantifier" can be used. This
4784 consists of an additional + character following a quantifier. Using
4785 this notation, the previous example can be rewritten as
4787 \d++foo
4789 Note that a possessive quantifier can be used with an entire group, for
4790 example:
4792 (abc|xyz){2,3}+
4794 Possessive quantifiers are always greedy; the setting of the
4795 PCRE_UNGREEDY option is ignored. They are a convenient notation for the
4796 simpler forms of atomic group. However, there is no difference in the
4797 meaning of a possessive quantifier and the equivalent atomic group,
4798 though there may be a performance difference; possessive quantifiers
4799 should be slightly faster.
4801 The possessive quantifier syntax is an extension to the Perl 5.8 syn-
4802 tax. Jeffrey Friedl originated the idea (and the name) in the first
4803 edition of his book. Mike McCloskey liked it, so implemented it when he
4804 built Sun's Java package, and PCRE copied it from there. It ultimately
4805 found its way into Perl at release 5.10.
4807 PCRE has an optimization that automatically "possessifies" certain sim-
4808 ple pattern constructs. For example, the sequence A+B is treated as
4809 A++B because there is no point in backtracking into a sequence of A's
4810 when B must follow.
4812 When a pattern contains an unlimited repeat inside a subpattern that
4813 can itself be repeated an unlimited number of times, the use of an
4814 atomic group is the only way to avoid some failing matches taking a
4815 very long time indeed. The pattern
4817 (\D+|<\d+>)*[!?]
4819 matches an unlimited number of substrings that either consist of non-
4820 digits, or digits enclosed in <>, followed by either ! or ?. When it
4821 matches, it runs quickly. However, if it is applied to
4823 aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
4825 it takes a long time before reporting failure. This is because the
4826 string can be divided between the internal \D+ repeat and the external
4827 * repeat in a large number of ways, and all have to be tried. (The
4828 example uses [!?] rather than a single character at the end, because
4829 both PCRE and Perl have an optimization that allows for fast failure
4830 when a single character is used. They remember the last single charac-
4831 ter that is required for a match, and fail early if it is not present
4832 in the string.) If the pattern is changed so that it uses an atomic
4833 group, like this:
4835 ((?>\D+)|<\d+>)*[!?]
4837 sequences of non-digits cannot be broken, and failure happens quickly.
4842 Outside a character class, a backslash followed by a digit greater than
4843 0 (and possibly further digits) is a back reference to a capturing sub-
4844 pattern earlier (that is, to its left) in the pattern, provided there
4845 have been that many previous capturing left parentheses.
4847 However, if the decimal number following the backslash is less than 10,
4848 it is always taken as a back reference, and causes an error only if
4849 there are not that many capturing left parentheses in the entire pat-
4850 tern. In other words, the parentheses that are referenced need not be
4851 to the left of the reference for numbers less than 10. A "forward back
4852 reference" of this type can make sense when a repetition is involved
4853 and the subpattern to the right has participated in an earlier itera-
4854 tion.
4856 It is not possible to have a numerical "forward back reference" to a
4857 subpattern whose number is 10 or more using this syntax because a
4858 sequence such as \50 is interpreted as a character defined in octal.
4859 See the subsection entitled "Non-printing characters" above for further
4860 details of the handling of digits following a backslash. There is no
4861 such problem when named parentheses are used. A back reference to any
4862 subpattern is possible using named parentheses (see below).
4864 Another way of avoiding the ambiguity inherent in the use of digits
4865 following a backslash is to use the \g escape sequence. This escape
4866 must be followed by an unsigned number or a negative number, optionally
4867 enclosed in braces. These examples are all identical:
4869 (ring), \1
4870 (ring), \g1
4871 (ring), \g{1}
4873 An unsigned number specifies an absolute reference without the ambigu-
4874 ity that is present in the older syntax. It is also useful when literal
4875 digits follow the reference. A negative number is a relative reference.
4876 Consider this example:
4878 (abc(def)ghi)\g{-1}
4880 The sequence \g{-1} is a reference to the most recently started captur-
4881 ing subpattern before \g, that is, is it equivalent to \2 in this exam-
4882 ple. Similarly, \g{-2} would be equivalent to \1. The use of relative
4883 references can be helpful in long patterns, and also in patterns that
4884 are created by joining together fragments that contain references
4885 within themselves.
4887 A back reference matches whatever actually matched the capturing sub-
4888 pattern in the current subject string, rather than anything matching
4889 the subpattern itself (see "Subpatterns as subroutines" below for a way
4890 of doing that). So the pattern
4892 (sens|respons)e and \1ibility
4894 matches "sense and sensibility" and "response and responsibility", but
4895 not "sense and responsibility". If caseful matching is in force at the
4896 time of the back reference, the case of letters is relevant. For exam-
4897 ple,
4899 ((?i)rah)\s+\1
4901 matches "rah rah" and "RAH RAH", but not "RAH rah", even though the
4902 original capturing subpattern is matched caselessly.
4904 There are several different ways of writing back references to named
4905 subpatterns. The .NET syntax \k{name} and the Perl syntax \k<name> or
4906 \k'name' are supported, as is the Python syntax (?P=name). Perl 5.10's
4907 unified back reference syntax, in which \g can be used for both numeric
4908 and named references, is also supported. We could rewrite the above
4909 example in any of the following ways:
4911 (?<p1>(?i)rah)\s+\k<p1>
4912 (?'p1'(?i)rah)\s+\k{p1}
4913 (?P<p1>(?i)rah)\s+(?P=p1)
4914 (?<p1>(?i)rah)\s+\g{p1}
4916 A subpattern that is referenced by name may appear in the pattern
4917 before or after the reference.
4919 There may be more than one back reference to the same subpattern. If a
4920 subpattern has not actually been used in a particular match, any back
4921 references to it always fail by default. For example, the pattern
4923 (a|(bc))\2
4925 always fails if it starts to match "a" rather than "bc". However, if
4926 the PCRE_JAVASCRIPT_COMPAT option is set at compile time, a back refer-
4927 ence to an unset value matches an empty string.
4929 Because there may be many capturing parentheses in a pattern, all dig-
4930 its following a backslash are taken as part of a potential back refer-
4931 ence number. If the pattern continues with a digit character, some
4932 delimiter must be used to terminate the back reference. If the
4933 PCRE_EXTENDED option is set, this can be whitespace. Otherwise, the \g{
4934 syntax or an empty comment (see "Comments" below) can be used.
4936 Recursive back references
4938 A back reference that occurs inside the parentheses to which it refers
4939 fails when the subpattern is first used, so, for example, (a\1) never
4940 matches. However, such references can be useful inside repeated sub-
4941 patterns. For example, the pattern
4943 (a|b\1)+
4945 matches any number of "a"s and also "aba", "ababbaa" etc. At each iter-
4946 ation of the subpattern, the back reference matches the character
4947 string corresponding to the previous iteration. In order for this to
4948 work, the pattern must be such that the first iteration does not need
4949 to match the back reference. This can be done using alternation, as in
4950 the example above, or by a quantifier with a minimum of zero.
4952 Back references of this type cause the group that they reference to be
4953 treated as an atomic group. Once the whole group has been matched, a
4954 subsequent matching failure cannot cause backtracking into the middle
4955 of the group.
4960 An assertion is a test on the characters following or preceding the
4961 current matching point that does not actually consume any characters.
4962 The simple assertions coded as \b, \B, \A, \G, \Z, \z, ^ and $ are
4963 described above.
4965 More complicated assertions are coded as subpatterns. There are two
4966 kinds: those that look ahead of the current position in the subject
4967 string, and those that look behind it. An assertion subpattern is
4968 matched in the normal way, except that it does not cause the current
4969 matching position to be changed.
4971 Assertion subpatterns are not capturing subpatterns. If such an asser-
4972 tion contains capturing subpatterns within it, these are counted for
4973 the purposes of numbering the capturing subpatterns in the whole pat-
4974 tern. However, substring capturing is carried out only for positive
4975 assertions, because it does not make sense for negative assertions.
4977 For compatibility with Perl, assertion subpatterns may be repeated;
4978 though it makes no sense to assert the same thing several times, the
4979 side effect of capturing parentheses may occasionally be useful. In
4980 practice, there only three cases:
4982 (1) If the quantifier is {0}, the assertion is never obeyed during
4983 matching. However, it may contain internal capturing parenthesized
4984 groups that are called from elsewhere via the subroutine mechanism.
4986 (2) If quantifier is {0,n} where n is greater than zero, it is treated
4987 as if it were {0,1}. At run time, the rest of the pattern match is
4988 tried with and without the assertion, the order depending on the greed-
4989 iness of the quantifier.
4991 (3) If the minimum repetition is greater than zero, the quantifier is
4992 ignored. The assertion is obeyed just once when encountered during
4993 matching.
4995 Lookahead assertions
4997 Lookahead assertions start with (?= for positive assertions and (?! for
4998 negative assertions. For example,
5000 \w+(?=;)
5002 matches a word followed by a semicolon, but does not include the semi-
5003 colon in the match, and
5005 foo(?!bar)
5007 matches any occurrence of "foo" that is not followed by "bar". Note
5008 that the apparently similar pattern
5010 (?!foo)bar
5012 does not find an occurrence of "bar" that is preceded by something
5013 other than "foo"; it finds any occurrence of "bar" whatsoever, because
5014 the assertion (?!foo) is always true when the next three characters are
5015 "bar". A lookbehind assertion is needed to achieve the other effect.
5017 If you want to force a matching failure at some point in a pattern, the
5018 most convenient way to do it is with (?!) because an empty string
5019 always matches, so an assertion that requires there not to be an empty
5020 string must always fail. The backtracking control verb (*FAIL) or (*F)
5021 is a synonym for (?!).
5023 Lookbehind assertions
5025 Lookbehind assertions start with (?<= for positive assertions and (?<!
5026 for negative assertions. For example,
5028 (?<!foo)bar
5030 does find an occurrence of "bar" that is not preceded by "foo". The
5031 contents of a lookbehind assertion are restricted such that all the
5032 strings it matches must have a fixed length. However, if there are sev-
5033 eral top-level alternatives, they do not all have to have the same
5034 fixed length. Thus
5036 (?<=bullock|donkey)
5038 is permitted, but
5040 (?<!dogs?|cats?)
5042 causes an error at compile time. Branches that match different length
5043 strings are permitted only at the top level of a lookbehind assertion.
5044 This is an extension compared with Perl, which requires all branches to
5045 match the same length of string. An assertion such as
5047 (?<=ab(c|de))
5049 is not permitted, because its single top-level branch can match two
5050 different lengths, but it is acceptable to PCRE if rewritten to use two
5051 top-level branches:
5053 (?<=abc|abde)
5055 In some cases, the escape sequence \K (see above) can be used instead
5056 of a lookbehind assertion to get round the fixed-length restriction.
5058 The implementation of lookbehind assertions is, for each alternative,
5059 to temporarily move the current position back by the fixed length and
5060 then try to match. If there are insufficient characters before the cur-
5061 rent position, the assertion fails.
5063 PCRE does not allow the \C escape (which matches a single byte in UTF-8
5064 mode) to appear in lookbehind assertions, because it makes it impossi-
5065 ble to calculate the length of the lookbehind. The \X and \R escapes,
5066 which can match different numbers of bytes, are also not permitted.
5068 "Subroutine" calls (see below) such as (?2) or (?&X) are permitted in
5069 lookbehinds, as long as the subpattern matches a fixed-length string.
5070 Recursion, however, is not supported.
5072 Possessive quantifiers can be used in conjunction with lookbehind
5073 assertions to specify efficient matching of fixed-length strings at the
5074 end of subject strings. Consider a simple pattern such as
5076 abcd$
5078 when applied to a long string that does not match. Because matching
5079 proceeds from left to right, PCRE will look for each "a" in the subject
5080 and then see if what follows matches the rest of the pattern. If the
5081 pattern is specified as
5083 ^.*abcd$
5085 the initial .* matches the entire string at first, but when this fails
5086 (because there is no following "a"), it backtracks to match all but the
5087 last character, then all but the last two characters, and so on. Once
5088 again the search for "a" covers the entire string, from right to left,
5089 so we are no better off. However, if the pattern is written as
5091 ^.*+(?<=abcd)
5093 there can be no backtracking for the .*+ item; it can match only the
5094 entire string. The subsequent lookbehind assertion does a single test
5095 on the last four characters. If it fails, the match fails immediately.
5096 For long strings, this approach makes a significant difference to the
5097 processing time.
5099 Using multiple assertions
5101 Several assertions (of any sort) may occur in succession. For example,
5103 (?<=\d{3})(?<!999)foo
5105 matches "foo" preceded by three digits that are not "999". Notice that
5106 each of the assertions is applied independently at the same point in
5107 the subject string. First there is a check that the previous three
5108 characters are all digits, and then there is a check that the same
5109 three characters are not "999". This pattern does not match "foo" pre-
5110 ceded by six characters, the first of which are digits and the last
5111 three of which are not "999". For example, it doesn't match "123abc-
5112 foo". A pattern to do that is
5114 (?<=\d{3}...)(?<!999)foo
5116 This time the first assertion looks at the preceding six characters,
5117 checking that the first three are digits, and then the second assertion
5118 checks that the preceding three characters are not "999".
5120 Assertions can be nested in any combination. For example,
5122 (?<=(?<!foo)bar)baz
5124 matches an occurrence of "baz" that is preceded by "bar" which in turn
5125 is not preceded by "foo", while
5127 (?<=\d{3}(?!999)...)foo
5129 is another pattern that matches "foo" preceded by three digits and any
5130 three characters that are not "999".
5135 It is possible to cause the matching process to obey a subpattern con-
5136 ditionally or to choose between two alternative subpatterns, depending
5137 on the result of an assertion, or whether a specific capturing subpat-
5138 tern has already been matched. The two possible forms of conditional
5139 subpattern are:
5141 (?(condition)yes-pattern)
5142 (?(condition)yes-pattern|no-pattern)
5144 If the condition is satisfied, the yes-pattern is used; otherwise the
5145 no-pattern (if present) is used. If there are more than two alterna-
5146 tives in the subpattern, a compile-time error occurs. Each of the two
5147 alternatives may itself contain nested subpatterns of any form, includ-
5148 ing conditional subpatterns; the restriction to two alternatives
5149 applies only at the level of the condition. This pattern fragment is an
5150 example where the alternatives are complex:
5152 (?(1) (A|B|C) | (D | (?(2)E|F) | E) )
5155 There are four kinds of condition: references to subpatterns, refer-
5156 ences to recursion, a pseudo-condition called DEFINE, and assertions.
5158 Checking for a used subpattern by number
5160 If the text between the parentheses consists of a sequence of digits,
5161 the condition is true if a capturing subpattern of that number has pre-
5162 viously matched. If there is more than one capturing subpattern with
5163 the same number (see the earlier section about duplicate subpattern
5164 numbers), the condition is true if any of them have matched. An alter-
5165 native notation is to precede the digits with a plus or minus sign. In
5166 this case, the subpattern number is relative rather than absolute. The
5167 most recently opened parentheses can be referenced by (?(-1), the next
5168 most recent by (?(-2), and so on. Inside loops it can also make sense
5169 to refer to subsequent groups. The next parentheses to be opened can be
5170 referenced as (?(+1), and so on. (The value zero in any of these forms
5171 is not used; it provokes a compile-time error.)
5173 Consider the following pattern, which contains non-significant white
5174 space to make it more readable (assume the PCRE_EXTENDED option) and to
5175 divide it into three parts for ease of discussion:
5177 ( \( )? [^()]+ (?(1) \) )
5179 The first part matches an optional opening parenthesis, and if that
5180 character is present, sets it as the first captured substring. The sec-
5181 ond part matches one or more characters that are not parentheses. The
5182 third part is a conditional subpattern that tests whether or not the
5183 first set of parentheses matched. If they did, that is, if subject
5184 started with an opening parenthesis, the condition is true, and so the
5185 yes-pattern is executed and a closing parenthesis is required. Other-
5186 wise, since no-pattern is not present, the subpattern matches nothing.
5187 In other words, this pattern matches a sequence of non-parentheses,
5188 optionally enclosed in parentheses.
5190 If you were embedding this pattern in a larger one, you could use a
5191 relative reference:
5193 ...other stuff... ( \( )? [^()]+ (?(-1) \) ) ...
5195 This makes the fragment independent of the parentheses in the larger
5196 pattern.
5198 Checking for a used subpattern by name
5200 Perl uses the syntax (?(<name>)...) or (?('name')...) to test for a
5201 used subpattern by name. For compatibility with earlier versions of
5202 PCRE, which had this facility before Perl, the syntax (?(name)...) is
5203 also recognized. However, there is a possible ambiguity with this syn-
5204 tax, because subpattern names may consist entirely of digits. PCRE
5205 looks first for a named subpattern; if it cannot find one and the name
5206 consists entirely of digits, PCRE looks for a subpattern of that num-
5207 ber, which must be greater than zero. Using subpattern names that con-
5208 sist entirely of digits is not recommended.
5210 Rewriting the above example to use a named subpattern gives this:
5212 (?<OPEN> \( )? [^()]+ (?(<OPEN>) \) )
5214 If the name used in a condition of this kind is a duplicate, the test
5215 is applied to all subpatterns of the same name, and is true if any one
5216 of them has matched.
5218 Checking for pattern recursion
5220 If the condition is the string (R), and there is no subpattern with the
5221 name R, the condition is true if a recursive call to the whole pattern
5222 or any subpattern has been made. If digits or a name preceded by amper-
5223 sand follow the letter R, for example:
5225 (?(R3)...) or (?(R&name)...)
5227 the condition is true if the most recent recursion is into a subpattern
5228 whose number or name is given. This condition does not check the entire
5229 recursion stack. If the name used in a condition of this kind is a
5230 duplicate, the test is applied to all subpatterns of the same name, and
5231 is true if any one of them is the most recent recursion.
5233 At "top level", all these recursion test conditions are false. The
5234 syntax for recursive patterns is described below.
5236 Defining subpatterns for use by reference only
5238 If the condition is the string (DEFINE), and there is no subpattern
5239 with the name DEFINE, the condition is always false. In this case,
5240 there may be only one alternative in the subpattern. It is always
5241 skipped if control reaches this point in the pattern; the idea of
5242 DEFINE is that it can be used to define subroutines that can be refer-
5243 enced from elsewhere. (The use of subroutines is described below.) For
5244 example, a pattern to match an IPv4 address such as ""
5245 could be written like this (ignore whitespace and line breaks):
5247 (?(DEFINE) (?<byte> 2[0-4]\d | 25[0-5] | 1\d\d | [1-9]?\d) )
5248 \b (?&byte) (\.(?&byte)){3} \b
5250 The first part of the pattern is a DEFINE group inside which a another
5251 group named "byte" is defined. This matches an individual component of
5252 an IPv4 address (a number less than 256). When matching takes place,
5253 this part of the pattern is skipped because DEFINE acts like a false
5254 condition. The rest of the pattern uses references to the named group
5255 to match the four dot-separated components of an IPv4 address, insist-
5256 ing on a word boundary at each end.
5258 Assertion conditions
5260 If the condition is not in any of the above formats, it must be an
5261 assertion. This may be a positive or negative lookahead or lookbehind
5262 assertion. Consider this pattern, again containing non-significant
5263 white space, and with the two alternatives on the second line:
5265 (?(?=[^a-z]*[a-z])
5266 \d{2}-[a-z]{3}-\d{2} | \d{2}-\d{2}-\d{2} )
5268 The condition is a positive lookahead assertion that matches an
5269 optional sequence of non-letters followed by a letter. In other words,
5270 it tests for the presence of at least one letter in the subject. If a
5271 letter is found, the subject is matched against the first alternative;
5272 otherwise it is matched against the second. This pattern matches
5273 strings in one of the two forms dd-aaa-dd or dd-dd-dd, where aaa are
5274 letters and dd are digits.
5279 There are two ways of including comments in patterns that are processed
5280 by PCRE. In both cases, the start of the comment must not be in a char-
5281 acter class, nor in the middle of any other sequence of related charac-
5282 ters such as (?: or a subpattern name or number. The characters that
5283 make up a comment play no part in the pattern matching.
5285 The sequence (?# marks the start of a comment that continues up to the
5286 next closing parenthesis. Nested parentheses are not permitted. If the
5287 PCRE_EXTENDED option is set, an unescaped # character also introduces a
5288 comment, which in this case continues to immediately after the next
5289 newline character or character sequence in the pattern. Which charac-
5290 ters are interpreted as newlines is controlled by the options passed to
5291 pcre_compile() or by a special sequence at the start of the pattern, as
5292 described in the section entitled "Newline conventions" above. Note
5293 that the end of this type of comment is a literal newline sequence in
5294 the pattern; escape sequences that happen to represent a newline do not
5295 count. For example, consider this pattern when PCRE_EXTENDED is set,
5296 and the default newline convention is in force:
5298 abc #comment \n still comment
5300 On encountering the # character, pcre_compile() skips along, looking
5301 for a newline in the pattern. The sequence \n is still literal at this
5302 stage, so it does not terminate the comment. Only an actual character
5303 with the code value 0x0a (the default newline) does so.
5308 Consider the problem of matching a string in parentheses, allowing for
5309 unlimited nested parentheses. Without the use of recursion, the best
5310 that can be done is to use a pattern that matches up to some fixed
5311 depth of nesting. It is not possible to handle an arbitrary nesting
5312 depth.
5314 For some time, Perl has provided a facility that allows regular expres-
5315 sions to recurse (amongst other things). It does this by interpolating
5316 Perl code in the expression at run time, and the code can refer to the
5317 expression itself. A Perl pattern using code interpolation to solve the
5318 parentheses problem can be created like this:
5320 $re = qr{\( (?: (?>[^()]+) | (?p{$re}) )* \)}x;
5322 The (?p{...}) item interpolates Perl code at run time, and in this case
5323 refers recursively to the pattern in which it appears.
5325 Obviously, PCRE cannot support the interpolation of Perl code. Instead,
5326 it supports special syntax for recursion of the entire pattern, and
5327 also for individual subpattern recursion. After its introduction in
5328 PCRE and Python, this kind of recursion was subsequently introduced
5329 into Perl at release 5.10.
5331 A special item that consists of (? followed by a number greater than
5332 zero and a closing parenthesis is a recursive subroutine call of the
5333 subpattern of the given number, provided that it occurs inside that
5334 subpattern. (If not, it is a non-recursive subroutine call, which is
5335 described in the next section.) The special item (?R) or (?0) is a
5336 recursive call of the entire regular expression.
5338 This PCRE pattern solves the nested parentheses problem (assume the
5339 PCRE_EXTENDED option is set so that white space is ignored):
5341 \( ( [^()]++ | (?R) )* \)
5343 First it matches an opening parenthesis. Then it matches any number of
5344 substrings which can either be a sequence of non-parentheses, or a
5345 recursive match of the pattern itself (that is, a correctly parenthe-
5346 sized substring). Finally there is a closing parenthesis. Note the use
5347 of a possessive quantifier to avoid backtracking into sequences of non-
5348 parentheses.
5350 If this were part of a larger pattern, you would not want to recurse
5351 the entire pattern, so instead you could use this:
5353 ( \( ( [^()]++ | (?1) )* \) )
5355 We have put the pattern into parentheses, and caused the recursion to
5356 refer to them instead of the whole pattern.
5358 In a larger pattern, keeping track of parenthesis numbers can be
5359 tricky. This is made easier by the use of relative references. Instead
5360 of (?1) in the pattern above you can write (?-2) to refer to the second
5361 most recently opened parentheses preceding the recursion. In other
5362 words, a negative number counts capturing parentheses leftwards from
5363 the point at which it is encountered.
5365 It is also possible to refer to subsequently opened parentheses, by
5366 writing references such as (?+2). However, these cannot be recursive
5367 because the reference is not inside the parentheses that are refer-
5368 enced. They are always non-recursive subroutine calls, as described in
5369 the next section.
5371 An alternative approach is to use named parentheses instead. The Perl
5372 syntax for this is (?&name); PCRE's earlier syntax (?P>name) is also
5373 supported. We could rewrite the above example as follows:
5375 (?<pn> \( ( [^()]++ | (?&pn) )* \) )
5377 If there is more than one subpattern with the same name, the earliest
5378 one is used.
5380 This particular example pattern that we have been looking at contains
5381 nested unlimited repeats, and so the use of a possessive quantifier for
5382 matching strings of non-parentheses is important when applying the pat-
5383 tern to strings that do not match. For example, when this pattern is
5384 applied to
5386 (aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa()
5388 it yields "no match" quickly. However, if a possessive quantifier is
5389 not used, the match runs for a very long time indeed because there are
5390 so many different ways the + and * repeats can carve up the subject,
5391 and all have to be tested before failure can be reported.
5393 At the end of a match, the values of capturing parentheses are those
5394 from the outermost level. If you want to obtain intermediate values, a
5395 callout function can be used (see below and the pcrecallout documenta-
5396 tion). If the pattern above is matched against
5398 (ab(cd)ef)
5400 the value for the inner capturing parentheses (numbered 2) is "ef",
5401 which is the last value taken on at the top level. If a capturing sub-
5402 pattern is not matched at the top level, its final captured value is
5403 unset, even if it was (temporarily) set at a deeper level during the
5404 matching process.
5406 If there are more than 15 capturing parentheses in a pattern, PCRE has
5407 to obtain extra memory to store data during a recursion, which it does
5408 by using pcre_malloc, freeing it via pcre_free afterwards. If no memory
5409 can be obtained, the match fails with the PCRE_ERROR_NOMEMORY error.
5411 Do not confuse the (?R) item with the condition (R), which tests for
5412 recursion. Consider this pattern, which matches text in angle brack-
5413 ets, allowing for arbitrary nesting. Only digits are allowed in nested
5414 brackets (that is, when recursing), whereas any characters are permit-
5415 ted at the outer level.
5417 < (?: (?(R) \d++ | [^<>]*+) | (?R)) * >
5419 In this pattern, (?(R) is the start of a conditional subpattern, with
5420 two different alternatives for the recursive and non-recursive cases.
5421 The (?R) item is the actual recursive call.
5423 Differences in recursion processing between PCRE and Perl
5425 Recursion processing in PCRE differs from Perl in two important ways.
5426 In PCRE (like Python, but unlike Perl), a recursive subpattern call is
5427 always treated as an atomic group. That is, once it has matched some of
5428 the subject string, it is never re-entered, even if it contains untried
5429 alternatives and there is a subsequent matching failure. This can be
5430 illustrated by the following pattern, which purports to match a palin-
5431 dromic string that contains an odd number of characters (for example,
5432 "a", "aba", "abcba", "abcdcba"):
5434 ^(.|(.)(?1)\2)$
5436 The idea is that it either matches a single character, or two identical
5437 characters surrounding a sub-palindrome. In Perl, this pattern works;
5438 in PCRE it does not if the pattern is longer than three characters.
5439 Consider the subject string "abcba":
5441 At the top level, the first character is matched, but as it is not at
5442 the end of the string, the first alternative fails; the second alterna-
5443 tive is taken and the recursion kicks in. The recursive call to subpat-
5444 tern 1 successfully matches the next character ("b"). (Note that the
5445 beginning and end of line tests are not part of the recursion).
5447 Back at the top level, the next character ("c") is compared with what
5448 subpattern 2 matched, which was "a". This fails. Because the recursion
5449 is treated as an atomic group, there are now no backtracking points,
5450 and so the entire match fails. (Perl is able, at this point, to re-
5451 enter the recursion and try the second alternative.) However, if the
5452 pattern is written with the alternatives in the other order, things are
5453 different:
5455 ^((.)(?1)\2|.)$
5457 This time, the recursing alternative is tried first, and continues to
5458 recurse until it runs out of characters, at which point the recursion
5459 fails. But this time we do have another alternative to try at the
5460 higher level. That is the big difference: in the previous case the
5461 remaining alternative is at a deeper recursion level, which PCRE cannot
5462 use.
5464 To change the pattern so that it matches all palindromic strings, not
5465 just those with an odd number of characters, it is tempting to change
5466 the pattern to this:
5468 ^((.)(?1)\2|.?)$
5470 Again, this works in Perl, but not in PCRE, and for the same reason.
5471 When a deeper recursion has matched a single character, it cannot be
5472 entered again in order to match an empty string. The solution is to
5473 separate the two cases, and write out the odd and even cases as alter-
5474 natives at the higher level:
5476 ^(?:((.)(?1)\2|)|((.)(?3)\4|.))
5478 If you want to match typical palindromic phrases, the pattern has to
5479 ignore all non-word characters, which can be done like this:
5481 ^\W*+(?:((.)\W*+(?1)\W*+\2|)|((.)\W*+(?3)\W*+\4|\W*+.\W*+))\W*+$
5483 If run with the PCRE_CASELESS option, this pattern matches phrases such
5484 as "A man, a plan, a canal: Panama!" and it works well in both PCRE and
5485 Perl. Note the use of the possessive quantifier *+ to avoid backtrack-
5486 ing into sequences of non-word characters. Without this, PCRE takes a
5487 great deal longer (ten times or more) to match typical phrases, and
5488 Perl takes so long that you think it has gone into a loop.
5490 WARNING: The palindrome-matching patterns above work only if the sub-
5491 ject string does not start with a palindrome that is shorter than the
5492 entire string. For example, although "abcba" is correctly matched, if
5493 the subject is "ababa", PCRE finds the palindrome "aba" at the start,
5494 then fails at top level because the end of the string does not follow.
5495 Once again, it cannot jump back into the recursion to try other alter-
5496 natives, so the entire match fails.
5498 The second way in which PCRE and Perl differ in their recursion pro-
5499 cessing is in the handling of captured values. In Perl, when a subpat-
5500 tern is called recursively or as a subpattern (see the next section),
5501 it has no access to any values that were captured outside the recur-
5502 sion, whereas in PCRE these values can be referenced. Consider this
5503 pattern:
5505 ^(.)(\1|a(?2))
5507 In PCRE, this pattern matches "bab". The first capturing parentheses
5508 match "b", then in the second group, when the back reference \1 fails
5509 to match "b", the second alternative matches "a" and then recurses. In
5510 the recursion, \1 does now match "b" and so the whole match succeeds.
5511 In Perl, the pattern fails to match because inside the recursive call
5512 \1 cannot access the externally set value.
5517 If the syntax for a recursive subpattern call (either by number or by
5518 name) is used outside the parentheses to which it refers, it operates
5519 like a subroutine in a programming language. The called subpattern may
5520 be defined before or after the reference. A numbered reference can be
5521 absolute or relative, as in these examples:
5523 (...(absolute)...)...(?2)...
5524 (...(relative)...)...(?-1)...
5525 (...(?+1)...(relative)...
5527 An earlier example pointed out that the pattern
5529 (sens|respons)e and \1ibility
5531 matches "sense and sensibility" and "response and responsibility", but
5532 not "sense and responsibility". If instead the pattern
5534 (sens|respons)e and (?1)ibility
5536 is used, it does match "sense and responsibility" as well as the other
5537 two strings. Another example is given in the discussion of DEFINE
5538 above.
5540 All subroutine calls, whether recursive or not, are always treated as
5541 atomic groups. That is, once a subroutine has matched some of the sub-
5542 ject string, it is never re-entered, even if it contains untried alter-
5543 natives and there is a subsequent matching failure. Any capturing
5544 parentheses that are set during the subroutine call revert to their
5545 previous values afterwards.
5547 Processing options such as case-independence are fixed when a subpat-
5548 tern is defined, so if it is used as a subroutine, such options cannot
5549 be changed for different calls. For example, consider this pattern:
5551 (abc)(?i:(?-1))
5553 It matches "abcabc". It does not match "abcABC" because the change of
5554 processing option does not affect the called subpattern.
5559 For compatibility with Oniguruma, the non-Perl syntax \g followed by a
5560 name or a number enclosed either in angle brackets or single quotes, is
5561 an alternative syntax for referencing a subpattern as a subroutine,
5562 possibly recursively. Here are two of the examples used above, rewrit-
5563 ten using this syntax:
5565 (?<pn> \( ( (?>[^()]+) | \g<pn> )* \) )
5566 (sens|respons)e and \g'1'ibility
5568 PCRE supports an extension to Oniguruma: if a number is preceded by a
5569 plus or a minus sign it is taken as a relative reference. For example:
5571 (abc)(?i:\g<-1>)
5573 Note that \g{...} (Perl syntax) and \g<...> (Oniguruma syntax) are not
5574 synonymous. The former is a back reference; the latter is a subroutine
5575 call.
5580 Perl has a feature whereby using the sequence (?{...}) causes arbitrary
5581 Perl code to be obeyed in the middle of matching a regular expression.
5582 This makes it possible, amongst other things, to extract different sub-
5583 strings that match the same pair of parentheses when there is a repeti-
5584 tion.
5586 PCRE provides a similar feature, but of course it cannot obey arbitrary
5587 Perl code. The feature is called "callout". The caller of PCRE provides
5588 an external function by putting its entry point in the global variable
5589 pcre_callout. By default, this variable contains NULL, which disables
5590 all calling out.
5592 Within a regular expression, (?C) indicates the points at which the
5593 external function is to be called. If you want to identify different
5594 callout points, you can put a number less than 256 after the letter C.
5595 The default value is zero. For example, this pattern has two callout
5596 points:
5598 (?C1)abc(?C2)def
5600 If the PCRE_AUTO_CALLOUT flag is passed to pcre_compile(), callouts are
5601 automatically installed before each item in the pattern. They are all
5602 numbered 255.
5604 During matching, when PCRE reaches a callout point (and pcre_callout is
5605 set), the external function is called. It is provided with the number
5606 of the callout, the position in the pattern, and, optionally, one item
5607 of data originally supplied by the caller of pcre_exec(). The callout
5608 function may cause matching to proceed, to backtrack, or to fail alto-
5609 gether. A complete description of the interface to the callout function
5610 is given in the pcrecallout documentation.
5615 Perl 5.10 introduced a number of "Special Backtracking Control Verbs",
5616 which are described in the Perl documentation as "experimental and sub-
5617 ject to change or removal in a future version of Perl". It goes on to
5618 say: "Their usage in production code should be noted to avoid problems
5619 during upgrades." The same remarks apply to the PCRE features described
5620 in this section.
5622 Since these verbs are specifically related to backtracking, most of
5623 them can be used only when the pattern is to be matched using
5624 pcre_exec(), which uses a backtracking algorithm. With the exception of
5625 (*FAIL), which behaves like a failing negative assertion, they cause an
5626 error if encountered by pcre_dfa_exec().
5628 If any of these verbs are used in an assertion or in a subpattern that
5629 is called as a subroutine (whether or not recursively), their effect is
5630 confined to that subpattern; it does not extend to the surrounding pat-
5631 tern, with one exception: a *MARK that is encountered in a positive
5632 assertion is passed back (compare capturing parentheses in assertions).
5633 Note that such subpatterns are processed as anchored at the point where
5634 they are tested. Note also that Perl's treatment of subroutines is dif-
5635 ferent in some cases.
5637 The new verbs make use of what was previously invalid syntax: an open-
5638 ing parenthesis followed by an asterisk. They are generally of the form
5639 (*VERB) or (*VERB:NAME). Some may take either form, with differing be-
5640 haviour, depending on whether or not an argument is present. A name is
5641 any sequence of characters that does not include a closing parenthesis.
5642 If the name is empty, that is, if the closing parenthesis immediately
5643 follows the colon, the effect is as if the colon were not there. Any
5644 number of these verbs may occur in a pattern.
5646 PCRE contains some optimizations that are used to speed up matching by
5647 running some checks at the start of each match attempt. For example, it
5648 may know the minimum length of matching subject, or that a particular
5649 character must be present. When one of these optimizations suppresses
5650 the running of a match, any included backtracking verbs will not, of
5651 course, be processed. You can suppress the start-of-match optimizations
5652 by setting the PCRE_NO_START_OPTIMIZE option when calling pcre_com-
5653 pile() or pcre_exec(), or by starting the pattern with (*NO_START_OPT).
5655 Verbs that act immediately
5657 The following verbs act as soon as they are encountered. They may not
5658 be followed by a name.
5660 (*ACCEPT)
5662 This verb causes the match to end successfully, skipping the remainder
5663 of the pattern. However, when it is inside a subpattern that is called
5664 as a subroutine, only that subpattern is ended successfully. Matching
5665 then continues at the outer level. If (*ACCEPT) is inside capturing
5666 parentheses, the data so far is captured. For example:
5668 A((?:A|B(*ACCEPT)|C)D)
5670 This matches "AB", "AAD", or "ACD"; when it matches "AB", "B" is cap-
5671 tured by the outer parentheses.
5673 (*FAIL) or (*F)
5675 This verb causes a matching failure, forcing backtracking to occur. It
5676 is equivalent to (?!) but easier to read. The Perl documentation notes
5677 that it is probably useful only when combined with (?{}) or (??{}).
5678 Those are, of course, Perl features that are not present in PCRE. The
5679 nearest equivalent is the callout feature, as for example in this pat-
5680 tern:
5682 a+(?C)(*FAIL)
5684 A match with the string "aaaa" always fails, but the callout is taken
5685 before each backtrack happens (in this example, 10 times).
5687 Recording which path was taken
5689 There is one verb whose main purpose is to track how a match was
5690 arrived at, though it also has a secondary use in conjunction with
5691 advancing the match starting point (see (*SKIP) below).
5693 (*MARK:NAME) or (*:NAME)
5695 A name is always required with this verb. There may be as many
5696 instances of (*MARK) as you like in a pattern, and their names do not
5697 have to be unique.
5699 When a match succeeds, the name of the last-encountered (*MARK) is
5700 passed back to the caller via the pcre_extra data structure, as
5701 described in the section on pcre_extra in the pcreapi documentation. No
5702 data is returned for a partial match. Here is an example of pcretest
5703 output, where the /K modifier requests the retrieval and outputting of
5704 (*MARK) data:
5706 /X(*MARK:A)Y|X(*MARK:B)Z/K
5707 XY
5708 0: XY
5709 MK: A
5710 XZ
5711 0: XZ
5712 MK: B
5714 The (*MARK) name is tagged with "MK:" in this output, and in this exam-
5715 ple it indicates which of the two alternatives matched. This is a more
5716 efficient way of obtaining this information than putting each alterna-
5717 tive in its own capturing parentheses.
5719 If (*MARK) is encountered in a positive assertion, its name is recorded
5720 and passed back if it is the last-encountered. This does not happen for
5721 negative assertions.
5723 A name may also be returned after a failed match if the final path
5724 through the pattern involves (*MARK). However, unless (*MARK) used in
5725 conjunction with (*COMMIT), this is unlikely to happen for an unan-
5726 chored pattern because, as the starting point for matching is advanced,
5727 the final check is often with an empty string, causing a failure before
5728 (*MARK) is reached. For example:
5730 /X(*MARK:A)Y|X(*MARK:B)Z/K
5731 XP
5732 No match
5734 There are three potential starting points for this match (starting with
5735 X, starting with P, and with an empty string). If the pattern is
5736 anchored, the result is different:
5738 /^X(*MARK:A)Y|^X(*MARK:B)Z/K
5739 XP
5740 No match, mark = B
5742 PCRE's start-of-match optimizations can also interfere with this. For
5743 example, if, as a result of a call to pcre_study(), it knows the mini-
5744 mum subject length for a match, a shorter subject will not be scanned
5745 at all.
5747 Note that similar anomalies (though different in detail) exist in Perl,
5748 no doubt for the same reasons. The use of (*MARK) data after a failed
5749 match of an unanchored pattern is not recommended, unless (*COMMIT) is
5750 involved.
5752 Verbs that act after backtracking
5754 The following verbs do nothing when they are encountered. Matching con-
5755 tinues with what follows, but if there is no subsequent match, causing
5756 a backtrack to the verb, a failure is forced. That is, backtracking
5757 cannot pass to the left of the verb. However, when one of these verbs
5758 appears inside an atomic group, its effect is confined to that group,
5759 because once the group has been matched, there is never any backtrack-
5760 ing into it. In this situation, backtracking can "jump back" to the
5761 left of the entire atomic group. (Remember also, as stated above, that
5762 this localization also applies in subroutine calls and assertions.)
5764 These verbs differ in exactly what kind of failure occurs when back-
5765 tracking reaches them.
5767 (*COMMIT)
5769 This verb, which may not be followed by a name, causes the whole match
5770 to fail outright if the rest of the pattern does not match. Even if the
5771 pattern is unanchored, no further attempts to find a match by advancing
5772 the starting point take place. Once (*COMMIT) has been passed,
5773 pcre_exec() is committed to finding a match at the current starting
5774 point, or not at all. For example:
5776 a+(*COMMIT)b
5778 This matches "xxaab" but not "aacaab". It can be thought of as a kind
5779 of dynamic anchor, or "I've started, so I must finish." The name of the
5780 most recently passed (*MARK) in the path is passed back when (*COMMIT)
5781 forces a match failure.
5783 Note that (*COMMIT) at the start of a pattern is not the same as an
5784 anchor, unless PCRE's start-of-match optimizations are turned off, as
5785 shown in this pcretest example:
5787 /(*COMMIT)abc/
5788 xyzabc
5789 0: abc
5790 xyzabc\Y
5791 No match
5793 PCRE knows that any match must start with "a", so the optimization
5794 skips along the subject to "a" before running the first match attempt,
5795 which succeeds. When the optimization is disabled by the \Y escape in
5796 the second subject, the match starts at "x" and so the (*COMMIT) causes
5797 it to fail without trying any other starting points.
5799 (*PRUNE) or (*PRUNE:NAME)
5801 This verb causes the match to fail at the current starting position in
5802 the subject if the rest of the pattern does not match. If the pattern
5803 is unanchored, the normal "bumpalong" advance to the next starting
5804 character then happens. Backtracking can occur as usual to the left of
5805 (*PRUNE), before it is reached, or when matching to the right of
5806 (*PRUNE), but if there is no match to the right, backtracking cannot
5807 cross (*PRUNE). In simple cases, the use of (*PRUNE) is just an alter-
5808 native to an atomic group or possessive quantifier, but there are some
5809 uses of (*PRUNE) that cannot be expressed in any other way. The behav-
5810 iour of (*PRUNE:NAME) is the same as (*MARK:NAME)(*PRUNE) when the
5811 match fails completely; the name is passed back if this is the final
5812 attempt. (*PRUNE:NAME) does not pass back a name if the match suc-
5813 ceeds. In an anchored pattern (*PRUNE) has the same effect as (*COM-
5814 MIT).
5816 (*SKIP)
5818 This verb, when given without a name, is like (*PRUNE), except that if
5819 the pattern is unanchored, the "bumpalong" advance is not to the next
5820 character, but to the position in the subject where (*SKIP) was encoun-
5821 tered. (*SKIP) signifies that whatever text was matched leading up to
5822 it cannot be part of a successful match. Consider:
5824 a+(*SKIP)b
5826 If the subject is "aaaac...", after the first match attempt fails
5827 (starting at the first character in the string), the starting point
5828 skips on to start the next attempt at "c". Note that a possessive quan-
5829 tifer does not have the same effect as this example; although it would
5830 suppress backtracking during the first match attempt, the second
5831 attempt would start at the second character instead of skipping on to
5832 "c".
5834 (*SKIP:NAME)
5836 When (*SKIP) has an associated name, its behaviour is modified. If the
5837 following pattern fails to match, the previous path through the pattern
5838 is searched for the most recent (*MARK) that has the same name. If one
5839 is found, the "bumpalong" advance is to the subject position that cor-
5840 responds to that (*MARK) instead of to where (*SKIP) was encountered.
5841 If no (*MARK) with a matching name is found, normal "bumpalong" of one
5842 character happens (that is, the (*SKIP) is ignored).
5844 (*THEN) or (*THEN:NAME)
5846 This verb causes a skip to the next innermost alternative if the rest
5847 of the pattern does not match. That is, it cancels pending backtrack-
5848 ing, but only within the current alternative. Its name comes from the
5849 observation that it can be used for a pattern-based if-then-else block:
5851 ( COND1 (*THEN) FOO | COND2 (*THEN) BAR | COND3 (*THEN) BAZ ) ...
5853 If the COND1 pattern matches, FOO is tried (and possibly further items
5854 after the end of the group if FOO succeeds); on failure, the matcher
5855 skips to the second alternative and tries COND2, without backtracking
5856 into COND1. The behaviour of (*THEN:NAME) is exactly the same as
5857 (*MARK:NAME)(*THEN) if the overall match fails. If (*THEN) is not
5858 inside an alternation, it acts like (*PRUNE).
5860 Note that a subpattern that does not contain a | character is just a
5861 part of the enclosing alternative; it is not a nested alternation with
5862 only one alternative. The effect of (*THEN) extends beyond such a sub-
5863 pattern to the enclosing alternative. Consider this pattern, where A,
5864 B, etc. are complex pattern fragments that do not contain any | charac-
5865 ters at this level:
5867 A (B(*THEN)C) | D
5869 If A and B are matched, but there is a failure in C, matching does not
5870 backtrack into A; instead it moves to the next alternative, that is, D.
5871 However, if the subpattern containing (*THEN) is given an alternative,
5872 it behaves differently:
5874 A (B(*THEN)C | (*FAIL)) | D
5876 The effect of (*THEN) is now confined to the inner subpattern. After a
5877 failure in C, matching moves to (*FAIL), which causes the whole subpat-
5878 tern to fail because there are no more alternatives to try. In this
5879 case, matching does now backtrack into A.
5881 Note also that a conditional subpattern is not considered as having two
5882 alternatives, because only one is ever used. In other words, the |
5883 character in a conditional subpattern has a different meaning. Ignoring
5884 white space, consider:
5886 ^.*? (?(?=a) a | b(*THEN)c )
5888 If the subject is "ba", this pattern does not match. Because .*? is
5889 ungreedy, it initially matches zero characters. The condition (?=a)
5890 then fails, the character "b" is matched, but "c" is not. At this
5891 point, matching does not backtrack to .*? as might perhaps be expected
5892 from the presence of the | character. The conditional subpattern is
5893 part of the single alternative that comprises the whole pattern, and so
5894 the match fails. (If there was a backtrack into .*?, allowing it to
5895 match "b", the match would succeed.)
5897 The verbs just described provide four different "strengths" of control
5898 when subsequent matching fails. (*THEN) is the weakest, carrying on the
5899 match at the next alternative. (*PRUNE) comes next, failing the match
5900 at the current starting position, but allowing an advance to the next
5901 character (for an unanchored pattern). (*SKIP) is similar, except that
5902 the advance may be more than one character. (*COMMIT) is the strongest,
5903 causing the entire match to fail.
5905 If more than one such verb is present in a pattern, the "strongest" one
5906 wins. For example, consider this pattern, where A, B, etc. are complex
5907 pattern fragments:
5909 (A(*COMMIT)B(*THEN)C|D)
5911 Once A has matched, PCRE is committed to this match, at the current
5912 starting position. If subsequently B matches, but C does not, the nor-
5913 mal (*THEN) action of trying the next alternative (that is, D) does not
5914 happen because (*COMMIT) overrides.
5919 pcreapi(3), pcrecallout(3), pcrematching(3), pcresyntax(3), pcre(3).
5924 Philip Hazel
5925 University Computing Service
5926 Cambridge CB2 3QH, England.
5931 Last updated: 19 October 2011
5932 Copyright (c) 1997-2011 University of Cambridge.
5933 ------------------------------------------------------------------------------
5939 NAME
5940 PCRE - Perl-compatible regular expressions
5945 The full syntax and semantics of the regular expressions that are sup-
5946 ported by PCRE are described in the pcrepattern documentation. This
5947 document contains just a quick-reference summary of the syntax.
5952 \x where x is non-alphanumeric is a literal x
5953 \Q...\E treat enclosed characters as literal
5958 \a alarm, that is, the BEL character (hex 07)
5959 \cx "control-x", where x is any ASCII character
5960 \e escape (hex 1B)
5961 \f formfeed (hex 0C)
5962 \n newline (hex 0A)
5963 \r carriage return (hex 0D)
5964 \t tab (hex 09)
5965 \ddd character with octal code ddd, or backreference
5966 \xhh character with hex code hh
5967 \x{hhh..} character with hex code hhh..
5972 . any character except newline;
5973 in dotall mode, any character whatsoever
5974 \C one byte, even in UTF-8 mode (best avoided)
5975 \d a decimal digit
5976 \D a character that is not a decimal digit
5977 \h a horizontal whitespace character
5978 \H a character that is not a horizontal whitespace character
5979 \N a character that is not a newline
5980 \p{xx} a character with the xx property
5981 \P{xx} a character without the xx property
5982 \R a newline sequence
5983 \s a whitespace character
5984 \S a character that is not a whitespace character
5985 \v a vertical whitespace character
5986 \V a character that is not a vertical whitespace character
5987 \w a "word" character
5988 \W a "non-word" character
5989 \X an extended Unicode sequence
5991 In PCRE, by default, \d, \D, \s, \S, \w, and \W recognize only ASCII
5992 characters, even in UTF-8 mode. However, this can be changed by setting
5993 the PCRE_UCP option.
5998 C Other
5999 Cc Control
6000 Cf Format
6001 Cn Unassigned
6002 Co Private use
6003 Cs Surrogate
6005 L Letter
6006 Ll Lower case letter
6007 Lm Modifier letter
6008 Lo Other letter
6009 Lt Title case letter
6010 Lu Upper case letter
6011 L& Ll, Lu, or Lt