ViewVC logotype

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

Parent Directory Parent Directory | Revision Log Revision Log

Revision 87 - (hide annotations) (download)
Sat Feb 24 21:41:21 2007 UTC (8 years, 2 months ago) by nigel
File MIME type: text/plain
File size: 231262 byte(s)
Load pcre-6.5 into code/trunk.

1 nigel 75 -----------------------------------------------------------------------------
2 nigel 63 This file contains a concatenation of the PCRE man pages, converted to plain
3     text format for ease of searching with a text editor, or for use on systems
4     that do not have a man page processor. The small individual files that give
5     synopses of each function in the library have not been included. There are
6     separate text files for the pcregrep and pcretest commands.
7     -----------------------------------------------------------------------------
9 nigel 41
10 nigel 79 PCRE(3) PCRE(3)
11 nigel 41
12 nigel 79
13 nigel 73 NAME
14     PCRE - Perl-compatible regular expressions
16 nigel 77
17 nigel 75 INTRODUCTION
18 nigel 41
19 nigel 73 The PCRE library is a set of functions that implement regular expres-
20     sion pattern matching using the same syntax and semantics as Perl, with
21     just a few differences. The current implementation of PCRE (release
22 nigel 77 6.x) corresponds approximately with Perl 5.8, including support for
23 nigel 75 UTF-8 encoded strings and Unicode general category properties. However,
24     this support has to be explicitly enabled; it is not the default.
25 nigel 63
26 nigel 77 In addition to the Perl-compatible matching function, PCRE also con-
27     tains an alternative matching function that matches the same compiled
28     patterns in a different way. In certain circumstances, the alternative
29     function has some advantages. For a discussion of the two matching
30     algorithms, see the pcrematching page.
32 nigel 75 PCRE is written in C and released as a C library. A number of people
33 nigel 77 have written wrappers and interfaces of various kinds. In particular,
34     Google Inc. have provided a comprehensive C++ wrapper. This is now
35     included as part of the PCRE distribution. The pcrecpp page has details
36     of this interface. Other people's contributions can be found in the
37     Contrib directory at the primary FTP site, which is:
38 nigel 63
39 nigel 73 ftp://ftp.csx.cam.ac.uk/pub/software/programming/pcre
40 nigel 63
41 nigel 73 Details of exactly which Perl regular expression features are and are
42     not supported by PCRE are given in separate documents. See the pcrepat-
43     tern and pcrecompat pages.
44 nigel 63
45 nigel 73 Some features of PCRE can be included, excluded, or changed when the
46     library is built. The pcre_config() function makes it possible for a
47 nigel 75 client to discover which features are available. The features them-
48     selves are described in the pcrebuild page. Documentation about build-
49     ing PCRE for various operating systems can be found in the README file
50     in the source distribution.
51 nigel 63
52 nigel 77 The library contains a number of undocumented internal functions and
53     data tables that are used by more than one of the exported external
54     functions, but which are not intended for use by external callers.
55     Their names all begin with "_pcre_", which hopefully will not provoke
56 nigel 83 any name clashes. In some environments, it is possible to control which
57     external symbols are exported when a shared library is built, and in
58     these cases the undocumented symbols are not exported.
59 nigel 63
60 nigel 77
63 nigel 75 The user documentation for PCRE comprises a number of different sec-
64     tions. In the "man" format, each of these is a separate "man page". In
65     the HTML format, each is a separate page, linked from the index page.
66     In the plain text format, all the sections are concatenated, for ease
67     of searching. The sections are as follows:
68 nigel 63
69 nigel 73 pcre this document
70 nigel 77 pcreapi details of PCRE's native C API
71 nigel 73 pcrebuild options for building PCRE
72     pcrecallout details of the callout feature
73     pcrecompat discussion of Perl compatibility
74 nigel 77 pcrecpp details of the C++ wrapper
75 nigel 73 pcregrep description of the pcregrep command
76 nigel 77 pcrematching discussion of the two matching algorithms
77 nigel 75 pcrepartial details of the partial matching facility
78 nigel 73 pcrepattern syntax and semantics of supported
79     regular expressions
80     pcreperform discussion of performance issues
81 nigel 77 pcreposix the POSIX-compatible C API
82 nigel 75 pcreprecompile details of saving and re-using precompiled patterns
83 nigel 73 pcresample discussion of the sample program
84 nigel 75 pcretest description of the pcretest testing command
85 nigel 63
86 nigel 73 In addition, in the "man" and HTML formats, there is a short page for
87 nigel 77 each C library function, listing its arguments and results.
88 nigel 63
92 nigel 73 There are some size limitations in PCRE but it is hoped that they will
93     never in practice be relevant.
94 nigel 63
95 nigel 73 The maximum length of a compiled pattern is 65539 (sic) bytes if PCRE
96     is compiled with the default internal linkage size of 2. If you want to
97     process regular expressions that are truly enormous, you can compile
98     PCRE with an internal linkage size of 3 or 4 (see the README file in
99     the source distribution and the pcrebuild documentation for details).
100 nigel 75 In these cases the limit is substantially larger. However, the speed
101 nigel 73 of execution will be slower.
102 nigel 63
103 nigel 73 All values in repeating quantifiers must be less than 65536. The maxi-
104     mum number of capturing subpatterns is 65535.
105 nigel 63
106 nigel 73 There is no limit to the number of non-capturing subpatterns, but the
107     maximum depth of nesting of all kinds of parenthesized subpattern,
108     including capturing subpatterns, assertions, and other types of subpat-
109     tern, is 200.
110 nigel 63
111 nigel 73 The maximum length of a subject string is the largest positive number
112 nigel 77 that an integer variable can hold. However, when using the traditional
113     matching function, PCRE uses recursion to handle subpatterns and indef-
114     inite repetition. This means that the available stack space may limit
115     the size of a subject string that can be processed by certain patterns.
116 nigel 63
119 nigel 63
120 nigel 77 From release 3.3, PCRE has had some support for character strings
121     encoded in the UTF-8 format. For release 4.0 this was greatly extended
122     to cover most common requirements, and in release 5.0 additional sup-
123 nigel 75 port for Unicode general category properties was added.
124 nigel 63
125 nigel 77 In order process UTF-8 strings, you must build PCRE to include UTF-8
126     support in the code, and, in addition, you must call pcre_compile()
127     with the PCRE_UTF8 option flag. When you do this, both the pattern and
128     any subject strings that are matched against it are treated as UTF-8
129 nigel 73 strings instead of just strings of bytes.
130 nigel 63
131 nigel 77 If you compile PCRE with UTF-8 support, but do not use it at run time,
132     the library will be a bit bigger, but the additional run time overhead
133     is limited to testing the PCRE_UTF8 flag in several places, so should
134 nigel 73 not be very large.
135 nigel 63
136 nigel 75 If PCRE is built with Unicode character property support (which implies
137 nigel 77 UTF-8 support), the escape sequences \p{..}, \P{..}, and \X are sup-
138 nigel 75 ported. The available properties that can be tested are limited to the
139 nigel 77 general category properties such as Lu for an upper case letter or Nd
140 nigel 87 for a decimal number, the Unicode script names such as Arabic or Han,
141     and the derived properties Any and L&. A full list is given in the
142     pcrepattern documentation. Only the short names for properties are sup-
143     ported. For example, \p{L} matches a letter. Its Perl synonym, \p{Let-
144     ter}, is not supported. Furthermore, in Perl, many properties may
145     optionally be prefixed by "Is", for compatibility with Perl 5.6. PCRE
146     does not support this.
147 nigel 75
148 nigel 73 The following comments apply when PCRE is running in UTF-8 mode:
149 nigel 63
150 nigel 77 1. When you set the PCRE_UTF8 flag, the strings passed as patterns and
151     subjects are checked for validity on entry to the relevant functions.
152 nigel 73 If an invalid UTF-8 string is passed, an error return is given. In some
153 nigel 77 situations, you may already know that your strings are valid, and
154 nigel 73 therefore want to skip these checks in order to improve performance. If
155 nigel 77 you set the PCRE_NO_UTF8_CHECK flag at compile time or at run time,
156     PCRE assumes that the pattern or subject it is given (respectively)
157     contains only valid UTF-8 codes. In this case, it does not diagnose an
158     invalid UTF-8 string. If you pass an invalid UTF-8 string to PCRE when
159     PCRE_NO_UTF8_CHECK is set, the results are undefined. Your program may
160 nigel 73 crash.
161 nigel 63
162 nigel 87 2. An unbraced hexadecimal escape sequence (such as \xb3) matches a
163     two-byte UTF-8 character if the value is greater than 127.
164 nigel 63
165 nigel 87 3. Repeat quantifiers apply to complete UTF-8 characters, not to indi-
166 nigel 73 vidual bytes, for example: \x{100}{3}.
167 nigel 63
168 nigel 87 4. The dot metacharacter matches one UTF-8 character instead of a sin-
169 nigel 75 gle byte.
170 nigel 63
171 nigel 87 5. The escape sequence \C can be used to match a single byte in UTF-8
172     mode, but its use can lead to some strange effects. This facility is
173 nigel 77 not available in the alternative matching function, pcre_dfa_exec().
174 nigel 63
175 nigel 87 6. The character escapes \b, \B, \d, \D, \s, \S, \w, and \W correctly
176     test characters of any code value, but the characters that PCRE recog-
177     nizes as digits, spaces, or word characters remain the same set as
178 nigel 75 before, all with values less than 256. This remains true even when PCRE
179 nigel 87 includes Unicode property support, because to do otherwise would slow
180     down PCRE in many common cases. If you really want to test for a wider
181     sense of, say, "digit", you must use Unicode property tests such as
182 nigel 75 \p{Nd}.
183 nigel 63
184 nigel 87 7. Similarly, characters that match the POSIX named character classes
185 nigel 75 are all low-valued characters.
186 nigel 63
187 nigel 87 8. Case-insensitive matching applies only to characters whose values
188     are less than 128, unless PCRE is built with Unicode property support.
189     Even when Unicode property support is available, PCRE still uses its
190     own character tables when checking the case of low-valued characters,
191     so as not to degrade performance. The Unicode property information is
192     used only for characters with higher values. Even when Unicode property
193     support is available, PCRE supports case-insensitive matching only when
194     there is a one-to-one mapping between a letter's cases. There are a
195     small number of many-to-one mappings in Unicode; these are not sup-
196     ported by PCRE.
197 nigel 63
199     AUTHOR
201 nigel 77 Philip Hazel
202 nigel 73 University Computing Service,
203     Cambridge CB2 3QG, England.
204 nigel 63
205 nigel 87 Putting an actual email address here seems to have been a spam magnet,
206 nigel 77 so I've taken it away. If you want to email me, use my initial and sur-
207     name, separated by a dot, at the domain ucs.cam.ac.uk.
209 nigel 87 Last updated: 24 January 2006
210     Copyright (c) 1997-2006 University of Cambridge.
211 nigel 79 ------------------------------------------------------------------------------
212 nigel 63
214 nigel 79 PCREBUILD(3) PCREBUILD(3)
215 nigel 63
216 nigel 79
217 nigel 73 NAME
218     PCRE - Perl-compatible regular expressions
220 nigel 77
223 nigel 73 This document describes the optional features of PCRE that can be
224     selected when the library is compiled. They are all selected, or dese-
225 nigel 75 lected, by providing options to the configure script that is run before
226     the make command. The complete list of options for configure (which
227     includes the standard ones such as the selection of the installation
228     directory) can be obtained by running
229 nigel 63
230 nigel 73 ./configure --help
231 nigel 63
232 nigel 73 The following sections describe certain options whose names begin with
233     --enable or --disable. These settings specify changes to the defaults
234     for the configure command. Because of the way that configure works,
235     --enable and --disable always come in pairs, so the complementary
236     option always exists as well, but as it specifies the default, it is
237     not described.
238 nigel 63
240 nigel 83 C++ SUPPORT
242     By default, the configure script will search for a C++ compiler and C++
243     header files. If it finds them, it automatically builds the C++ wrapper
244     library for PCRE. You can disable this by adding
246     --disable-cpp
248     to the configure command.
251 nigel 63 UTF-8 SUPPORT
253 nigel 73 To build PCRE with support for UTF-8 character strings, add
254 nigel 63
255 nigel 73 --enable-utf8
256 nigel 63
257 nigel 73 to the configure command. Of itself, this does not make PCRE treat
258     strings as UTF-8. As well as compiling PCRE with this option, you also
259     have have to set the PCRE_UTF8 option when you call the pcre_compile()
260     function.
261 nigel 63
265     UTF-8 support allows PCRE to process character values greater than 255
266     in the strings that it handles. On its own, however, it does not pro-
267     vide any facilities for accessing the properties of such characters. If
268     you want to be able to use the pattern escapes \P, \p, and \X, which
269     refer to Unicode character properties, you must add
271     --enable-unicode-properties
273     to the configure command. This implies UTF-8 support, even if you have
274     not explicitly requested it.
276     Including Unicode property support adds around 90K of tables to the
277     PCRE library, approximately doubling its size. Only the general cate-
278     gory properties such as Lu and Nd are supported. Details are given in
279     the pcrepattern documentation.
284 nigel 73 By default, PCRE treats character 10 (linefeed) as the newline charac-
285     ter. This is the normal newline character on Unix-like systems. You can
286     compile PCRE to use character 13 (carriage return) instead by adding
287 nigel 63
288 nigel 73 --enable-newline-is-cr
289 nigel 63
290 nigel 73 to the configure command. For completeness there is also a --enable-
291     newline-is-lf option, which explicitly specifies linefeed as the new-
292     line character.
293 nigel 63
297 nigel 73 The PCRE building process uses libtool to build both shared and static
298     Unix libraries by default. You can suppress one of these by adding one
299     of
300 nigel 63
301 nigel 73 --disable-shared
302     --disable-static
303 nigel 63
304 nigel 73 to the configure command, as required.
305 nigel 63
309 nigel 75 When PCRE is called through the POSIX interface (see the pcreposix doc-
310     umentation), additional working storage is required for holding the
311     pointers to capturing substrings, because PCRE requires three integers
312 nigel 73 per substring, whereas the POSIX interface provides only two. If the
313     number of expected substrings is small, the wrapper function uses space
314     on the stack, because this is faster than using malloc() for each call.
315     The default threshold above which the stack is no longer used is 10; it
316     can be changed by adding a setting such as
317 nigel 63
318 nigel 73 --with-posix-malloc-threshold=20
319 nigel 63
320 nigel 73 to the configure command.
321 nigel 63
325 nigel 75 Internally, PCRE has a function called match(), which it calls repeat-
326 nigel 77 edly (possibly recursively) when matching a pattern with the
327     pcre_exec() function. By controlling the maximum number of times this
328     function may be called during a single matching operation, a limit can
329     be placed on the resources used by a single call to pcre_exec(). The
330     limit can be changed at run time, as described in the pcreapi documen-
331     tation. The default is 10 million, but this can be changed by adding a
332     setting such as
333 nigel 63
334 nigel 73 --with-match-limit=500000
335 nigel 63
336 nigel 77 to the configure command. This setting has no effect on the
337     pcre_dfa_exec() matching function.
338 nigel 63
342 nigel 73 Within a compiled pattern, offset values are used to point from one
343     part to another (for example, from an opening parenthesis to an alter-
344 nigel 75 nation metacharacter). By default, two-byte values are used for these
345 nigel 73 offsets, leading to a maximum size for a compiled pattern of around
346     64K. This is sufficient to handle all but the most gigantic patterns.
347     Nevertheless, some people do want to process enormous patterns, so it
348     is possible to compile PCRE to use three-byte or four-byte offsets by
349     adding a setting such as
350 nigel 63
351 nigel 73 --with-link-size=3
352 nigel 63
353 nigel 73 to the configure command. The value given must be 2, 3, or 4. Using
354     longer offsets slows down the operation of PCRE because it has to load
355     additional bytes when handling them.
356 nigel 63
357 nigel 73 If you build PCRE with an increased link size, test 2 (and test 5 if
358     you are using UTF-8) will fail. Part of the output of these tests is a
359     representation of the compiled pattern, and this changes with the link
360     size.
361 nigel 63
362 nigel 73
365 nigel 77 When matching with the pcre_exec() function, PCRE implements backtrack-
366     ing by making recursive calls to an internal function called match().
367     In environments where the size of the stack is limited, this can se-
368     verely limit PCRE's operation. (The Unix environment does not usually
369     suffer from this problem.) An alternative approach that uses memory
370     from the heap to remember data, instead of using recursive function
371     calls, has been implemented to work round this problem. If you want to
372     build a version of PCRE that works this way, add
373 nigel 73
374     --disable-stack-for-recursion
376     to the configure command. With this configuration, PCRE will use the
377 nigel 75 pcre_stack_malloc and pcre_stack_free variables to call memory manage-
378     ment functions. Separate functions are provided because the usage is
379     very predictable: the block sizes requested are always the same, and
380 nigel 73 the blocks are always freed in reverse order. A calling program might
381     be able to implement optimized functions that perform better than the
382     standard malloc() and free() functions. PCRE runs noticeably more
383 nigel 77 slowly when built in this way. This option affects only the pcre_exec()
384     function; it is not relevant for the the pcre_dfa_exec() function.
385 nigel 73
389 nigel 77 PCRE assumes by default that it will run in an environment where the
390     character code is ASCII (or Unicode, which is a superset of ASCII).
391     PCRE can, however, be compiled to run in an EBCDIC environment by
392 nigel 75 adding
393 nigel 73
394     --enable-ebcdic
396     to the configure command.
398 nigel 83 Last updated: 15 August 2005
399 nigel 77 Copyright (c) 1997-2005 University of Cambridge.
400 nigel 79 ------------------------------------------------------------------------------
401 nigel 63
404 nigel 63
405 nigel 79
406 nigel 77 NAME
407     PCRE - Perl-compatible regular expressions
408 nigel 73
409 nigel 77
412     This document describes the two different algorithms that are available
413     in PCRE for matching a compiled regular expression against a given sub-
414     ject string. The "standard" algorithm is the one provided by the
415     pcre_exec() function. This works in the same was as Perl's matching
416     function, and provides a Perl-compatible matching operation.
418     An alternative algorithm is provided by the pcre_dfa_exec() function;
419     this operates in a different way, and is not Perl-compatible. It has
420     advantages and disadvantages compared with the standard algorithm, and
421     these are described below.
423     When there is only one possible way in which a given subject string can
424     match a pattern, the two algorithms give the same answer. A difference
425     arises, however, when there are multiple possibilities. For example, if
426     the pattern
428     ^<.*>
430     is matched against the string
432     <something> <something else> <something further>
434     there are three possible answers. The standard algorithm finds only one
435     of them, whereas the DFA algorithm finds all three.
440     The set of strings that are matched by a regular expression can be rep-
441     resented as a tree structure. An unlimited repetition in the pattern
442     makes the tree of infinite size, but it is still a tree. Matching the
443     pattern to a given subject string (from a given starting point) can be
444     thought of as a search of the tree. There are two standard ways to
445     search a tree: depth-first and breadth-first, and these correspond to
446     the two matching algorithms provided by PCRE.
451     In the terminology of Jeffrey Friedl's book Mastering Regular Expres-
452     sions, the standard algorithm is an "NFA algorithm". It conducts a
453     depth-first search of the pattern tree. That is, it proceeds along a
454     single path through the tree, checking that the subject matches what is
455     required. When there is a mismatch, the algorithm tries any alterna-
456     tives at the current point, and if they all fail, it backs up to the
457     previous branch point in the tree, and tries the next alternative
458     branch at that level. This often involves backing up (moving to the
459     left) in the subject string as well. The order in which repetition
460     branches are tried is controlled by the greedy or ungreedy nature of
461     the quantifier.
463     If a leaf node is reached, a matching string has been found, and at
464     that point the algorithm stops. Thus, if there is more than one possi-
465     ble match, this algorithm returns the first one that it finds. Whether
466     this is the shortest, the longest, or some intermediate length depends
467     on the way the greedy and ungreedy repetition quantifiers are specified
468     in the pattern.
470     Because it ends up with a single path through the tree, it is rela-
471     tively straightforward for this algorithm to keep track of the sub-
472     strings that are matched by portions of the pattern in parentheses.
473     This provides support for capturing parentheses and back references.
478     DFA stands for "deterministic finite automaton", but you do not need to
479     understand the origins of that name. This algorithm conducts a breadth-
480     first search of the tree. Starting from the first matching point in the
481     subject, it scans the subject string from left to right, once, charac-
482     ter by character, and as it does this, it remembers all the paths
483     through the tree that represent valid matches.
485     The scan continues until either the end of the subject is reached, or
486     there are no more unterminated paths. At this point, terminated paths
487     represent the different matching possibilities (if there are none, the
488     match has failed). Thus, if there is more than one possible match,
489     this algorithm finds all of them, and in particular, it finds the long-
490     est. In PCRE, there is an option to stop the algorithm after the first
491     match (which is necessarily the shortest) has been found.
493     Note that all the matches that are found start at the same point in the
494     subject. If the pattern
496     cat(er(pillar)?)
498     is matched against the string "the caterpillar catchment", the result
499     will be the three strings "cat", "cater", and "caterpillar" that start
500     at the fourth character of the subject. The algorithm does not automat-
501     ically move on to find matches that start at later positions.
503     There are a number of features of PCRE regular expressions that are not
504     supported by the DFA matching algorithm. They are as follows:
506     1. Because the algorithm finds all possible matches, the greedy or
507     ungreedy nature of repetition quantifiers is not relevant. Greedy and
508     ungreedy quantifiers are treated in exactly the same way.
510     2. When dealing with multiple paths through the tree simultaneously, it
511     is not straightforward to keep track of captured substrings for the
512     different matching possibilities, and PCRE's implementation of this
513     algorithm does not attempt to do this. This means that no captured sub-
514     strings are available.
516     3. Because no substrings are captured, back references within the pat-
517     tern are not supported, and cause errors if encountered.
519     4. For the same reason, conditional expressions that use a backrefer-
520     ence as the condition are not supported.
522     5. Callouts are supported, but the value of the capture_top field is
523     always 1, and the value of the capture_last field is always -1.
525     6. The \C escape sequence, which (in the standard algorithm) matches a
526     single byte, even in UTF-8 mode, is not supported because the DFA algo-
527     rithm moves through the subject string one character at a time, for all
528     active paths through the tree.
533     Using the DFA matching algorithm provides the following advantages:
535     1. All possible matches (at a single point in the subject) are automat-
536     ically found, and in particular, the longest match is found. To find
537     more than one match using the standard algorithm, you have to do kludgy
538     things with callouts.
540     2. There is much better support for partial matching. The restrictions
541     on the content of the pattern that apply when using the standard algo-
542     rithm for partial matching do not apply to the DFA algorithm. For non-
543     anchored patterns, the starting position of a partial match is avail-
544     able.
546     3. Because the DFA algorithm scans the subject string just once, and
547     never needs to backtrack, it is possible to pass very long subject
548     strings to the matching function in several pieces, checking for par-
549     tial matching each time.
554     The DFA algorithm suffers from a number of disadvantages:
556     1. It is substantially slower than the standard algorithm. This is
557     partly because it has to search for all possible matches, but is also
558     because it is less susceptible to optimization.
560     2. Capturing parentheses and back references are not supported.
562     3. The "atomic group" feature of PCRE regular expressions is supported,
563     but does not provide the advantage that it does for the standard algo-
564     rithm.
566     Last updated: 28 February 2005
567     Copyright (c) 1997-2005 University of Cambridge.
568 nigel 79 ------------------------------------------------------------------------------
569 nigel 77
571 nigel 79 PCREAPI(3) PCREAPI(3)
572 nigel 77
573 nigel 79
574 nigel 73 NAME
575     PCRE - Perl-compatible regular expressions
577 nigel 77
578 nigel 75 PCRE NATIVE API
579 nigel 63
580 nigel 73 #include <pcre.h>
581 nigel 41
582 nigel 73 pcre *pcre_compile(const char *pattern, int options,
583     const char **errptr, int *erroffset,
584     const unsigned char *tableptr);
585 nigel 41
586 nigel 77 pcre *pcre_compile2(const char *pattern, int options,
587     int *errorcodeptr,
588     const char **errptr, int *erroffset,
589     const unsigned char *tableptr);
591 nigel 73 pcre_extra *pcre_study(const pcre *code, int options,
592     const char **errptr);
593 nigel 41
594 nigel 73 int pcre_exec(const pcre *code, const pcre_extra *extra,
595     const char *subject, int length, int startoffset,
596     int options, int *ovector, int ovecsize);
597 nigel 41
598 nigel 77 int pcre_dfa_exec(const pcre *code, const pcre_extra *extra,
599     const char *subject, int length, int startoffset,
600     int options, int *ovector, int ovecsize,
601     int *workspace, int wscount);
603 nigel 73 int pcre_copy_named_substring(const pcre *code,
604     const char *subject, int *ovector,
605     int stringcount, const char *stringname,
606     char *buffer, int buffersize);
607 nigel 63
608 nigel 73 int pcre_copy_substring(const char *subject, int *ovector,
609     int stringcount, int stringnumber, char *buffer,
610     int buffersize);
611 nigel 41
612 nigel 73 int pcre_get_named_substring(const pcre *code,
613     const char *subject, int *ovector,
614     int stringcount, const char *stringname,
615     const char **stringptr);
616 nigel 63
617 nigel 73 int pcre_get_stringnumber(const pcre *code,
618     const char *name);
619 nigel 63
620 nigel 73 int pcre_get_substring(const char *subject, int *ovector,
621     int stringcount, int stringnumber,
622     const char **stringptr);
623 nigel 41
624 nigel 73 int pcre_get_substring_list(const char *subject,
625     int *ovector, int stringcount, const char ***listptr);
626 nigel 41
627 nigel 73 void pcre_free_substring(const char *stringptr);
628 nigel 49
629 nigel 73 void pcre_free_substring_list(const char **stringptr);
630 nigel 49
631 nigel 73 const unsigned char *pcre_maketables(void);
632 nigel 41
633 nigel 73 int pcre_fullinfo(const pcre *code, const pcre_extra *extra,
634     int what, void *where);
635 nigel 43
636 nigel 73 int pcre_info(const pcre *code, int *optptr, int *firstcharptr);
637 nigel 63
638 nigel 77 int pcre_refcount(pcre *code, int adjust);
640 nigel 73 int pcre_config(int what, void *where);
641 nigel 41
642 nigel 73 char *pcre_version(void);
643 nigel 63
644 nigel 73 void *(*pcre_malloc)(size_t);
645 nigel 41
646 nigel 73 void (*pcre_free)(void *);
647 nigel 41
648 nigel 73 void *(*pcre_stack_malloc)(size_t);
649 nigel 41
650 nigel 73 void (*pcre_stack_free)(void *);
651 nigel 41
652 nigel 73 int (*pcre_callout)(pcre_callout_block *);
653 nigel 41
654 nigel 73
655 nigel 75 PCRE API OVERVIEW
656 nigel 41
657 nigel 73 PCRE has its own native API, which is described in this document. There
658     is also a set of wrapper functions that correspond to the POSIX regular
659 nigel 77 expression API. These are described in the pcreposix documentation.
660     Both of these APIs define a set of C function calls. A C++ wrapper is
661     distributed with PCRE. It is documented in the pcrecpp page.
662 nigel 43
663 nigel 77 The native API C function prototypes are defined in the header file
664     pcre.h, and on Unix systems the library itself is called libpcre. It
665 nigel 75 can normally be accessed by adding -lpcre to the command for linking an
666     application that uses PCRE. The header file defines the macros
667     PCRE_MAJOR and PCRE_MINOR to contain the major and minor release num-
668     bers for the library. Applications can use these to include support
669     for different releases of PCRE.
670 nigel 41
671 nigel 77 The functions pcre_compile(), pcre_compile2(), pcre_study(), and
672     pcre_exec() are used for compiling and matching regular expressions in
673     a Perl-compatible manner. A sample program that demonstrates the sim-
674     plest way of using them is provided in the file called pcredemo.c in
675     the source distribution. The pcresample documentation describes how to
676     run it.
677 nigel 49
678 nigel 77 A second matching function, pcre_dfa_exec(), which is not Perl-compati-
679     ble, is also provided. This uses a different algorithm for the match-
680     ing. This allows it to find all possible matches (at a given point in
681     the subject), not just one. However, this algorithm does not return
682     captured substrings. A description of the two matching algorithms and
683     their advantages and disadvantages is given in the pcrematching docu-
684     mentation.
685 nigel 63
686 nigel 77 In addition to the main compiling and matching functions, there are
687     convenience functions for extracting captured substrings from a subject
688     string that is matched by pcre_exec(). They are:
690 nigel 73 pcre_copy_substring()
691     pcre_copy_named_substring()
692     pcre_get_substring()
693     pcre_get_named_substring()
694     pcre_get_substring_list()
695 nigel 75 pcre_get_stringnumber()
696 nigel 63
697 nigel 73 pcre_free_substring() and pcre_free_substring_list() are also provided,
698     to free the memory used for extracted strings.
699 nigel 41
700 nigel 77 The function pcre_maketables() is used to build a set of character
701     tables in the current locale for passing to pcre_compile(),
702     pcre_exec(), or pcre_dfa_exec(). This is an optional facility that is
703     provided for specialist use. Most commonly, no special tables are
704     passed, in which case internal tables that are generated when PCRE is
705     built are used.
706 nigel 49
707 nigel 75 The function pcre_fullinfo() is used to find out information about a
708     compiled pattern; pcre_info() is an obsolete version that returns only
709     some of the available information, but is retained for backwards com-
710     patibility. The function pcre_version() returns a pointer to a string
711 nigel 73 containing the version of PCRE and its date of release.
712 nigel 41
713 nigel 77 The function pcre_refcount() maintains a reference count in a data
714     block containing a compiled pattern. This is provided for the benefit
715     of object-oriented applications.
717 nigel 75 The global variables pcre_malloc and pcre_free initially contain the
718     entry points of the standard malloc() and free() functions, respec-
719 nigel 73 tively. PCRE calls the memory management functions via these variables,
720 nigel 75 so a calling program can replace them if it wishes to intercept the
721 nigel 73 calls. This should be done before calling any PCRE functions.
722 nigel 41
723 nigel 75 The global variables pcre_stack_malloc and pcre_stack_free are also
724     indirections to memory management functions. These special functions
725     are used only when PCRE is compiled to use the heap for remembering
726 nigel 77 data, instead of recursive function calls, when running the pcre_exec()
727     function. This is a non-standard way of building PCRE, for use in envi-
728     ronments that have limited stacks. Because of the greater use of memory
729     management, it runs more slowly. Separate functions are provided so
730     that special-purpose external code can be used for this case. When
731     used, these functions are always called in a stack-like manner (last
732     obtained, first freed), and always for memory blocks of the same size.
733 nigel 41
734 nigel 73 The global variable pcre_callout initially contains NULL. It can be set
735 nigel 77 by the caller to a "callout" function, which PCRE will then call at
736     specified points during a matching operation. Details are given in the
737 nigel 73 pcrecallout documentation.
738 nigel 41
739 nigel 73
742 nigel 77 The PCRE functions can be used in multi-threading applications, with
743 nigel 73 the proviso that the memory management functions pointed to by
744     pcre_malloc, pcre_free, pcre_stack_malloc, and pcre_stack_free, and the
745     callout function pointed to by pcre_callout, are shared by all threads.
746 nigel 41
747 nigel 77 The compiled form of a regular expression is not altered during match-
748 nigel 73 ing, so the same compiled pattern can safely be used by several threads
749     at once.
750 nigel 41
754     The compiled form of a regular expression can be saved and re-used at a
755 nigel 77 later time, possibly by a different program, and even on a host other
756     than the one on which it was compiled. Details are given in the
757 nigel 75 pcreprecompile documentation.
761 nigel 41
762 nigel 73 int pcre_config(int what, void *where);
763 nigel 63
764 nigel 77 The function pcre_config() makes it possible for a PCRE client to dis-
765 nigel 73 cover which optional features have been compiled into the PCRE library.
766 nigel 77 The pcrebuild documentation has more details about these optional fea-
767 nigel 73 tures.
768 nigel 63
769 nigel 77 The first argument for pcre_config() is an integer, specifying which
770 nigel 73 information is required; the second argument is a pointer to a variable
771 nigel 77 into which the information is placed. The following information is
772 nigel 73 available:
773 nigel 63
774 nigel 73 PCRE_CONFIG_UTF8
775 nigel 63
776 nigel 77 The output is an integer that is set to one if UTF-8 support is avail-
777 nigel 73 able; otherwise it is set to zero.
778 nigel 63
781 nigel 77 The output is an integer that is set to one if support for Unicode
782 nigel 75 character properties is available; otherwise it is set to zero.
785 nigel 63
786 nigel 77 The output is an integer that is set to the value of the code that is
787     used for the newline character. It is either linefeed (10) or carriage
788     return (13), and should normally be the standard character for your
789 nigel 73 operating system.
790 nigel 63
792 nigel 63
793 nigel 77 The output is an integer that contains the number of bytes used for
794 nigel 73 internal linkage in compiled regular expressions. The value is 2, 3, or
795 nigel 77 4. Larger values allow larger regular expressions to be compiled, at
796     the expense of slower matching. The default value of 2 is sufficient
797     for all but the most massive patterns, since it allows the compiled
798 nigel 73 pattern to be up to 64K in size.
799 nigel 63
801 nigel 63
802 nigel 77 The output is an integer that contains the threshold above which the
803     POSIX interface uses malloc() for output vectors. Further details are
804 nigel 73 given in the pcreposix documentation.
805 nigel 63
807 nigel 63
808 nigel 73 The output is an integer that gives the default limit for the number of
809 nigel 77 internal matching function calls in a pcre_exec() execution. Further
810 nigel 73 details are given with pcre_exec() below.
811 nigel 63
814     The output is an integer that gives the default limit for the depth of
815     recursion when calling the internal matching function in a pcre_exec()
816     execution. Further details are given with pcre_exec() below.
819 nigel 63
820 nigel 77 The output is an integer that is set to one if internal recursion when
821     running pcre_exec() is implemented by recursive function calls that use
822     the stack to remember their state. This is the usual way that PCRE is
823     compiled. The output is zero if PCRE was compiled to use blocks of data
824     on the heap instead of recursive function calls. In this case,
825     pcre_stack_malloc and pcre_stack_free are called to manage memory
826     blocks on the heap, thus avoiding the use of the stack.
827 nigel 73
830 nigel 63
831 nigel 73 pcre *pcre_compile(const char *pattern, int options,
832     const char **errptr, int *erroffset,
833     const unsigned char *tableptr);
834 nigel 63
835 nigel 77 pcre *pcre_compile2(const char *pattern, int options,
836     int *errorcodeptr,
837     const char **errptr, int *erroffset,
838     const unsigned char *tableptr);
839 nigel 41
840 nigel 77 Either of the functions pcre_compile() or pcre_compile2() can be called
841     to compile a pattern into an internal form. The only difference between
842     the two interfaces is that pcre_compile2() has an additional argument,
843     errorcodeptr, via which a numerical error code can be returned.
845     The pattern is a C string terminated by a binary zero, and is passed in
846     the pattern argument. A pointer to a single block of memory that is
847     obtained via pcre_malloc is returned. This contains the compiled code
848     and related data. The pcre type is defined for the returned block; this
849     is a typedef for a structure whose contents are not externally defined.
850     It is up to the caller to free the memory when it is no longer
851     required.
853     Although the compiled code of a PCRE regex is relocatable, that is, it
854 nigel 73 does not depend on memory location, the complete pcre data block is not
855 nigel 77 fully relocatable, because it may contain a copy of the tableptr argu-
856 nigel 75 ment, which is an address (see below).
857 nigel 41
858 nigel 73 The options argument contains independent bits that affect the compila-
859 nigel 77 tion. It should be zero if no options are required. The available
860     options are described below. Some of them, in particular, those that
861     are compatible with Perl, can also be set and unset from within the
862     pattern (see the detailed description in the pcrepattern documenta-
863     tion). For these options, the contents of the options argument speci-
864     fies their initial settings at the start of compilation and execution.
865     The PCRE_ANCHORED option can be set at the time of matching as well as
866 nigel 75 at compile time.
867 nigel 41
868 nigel 73 If errptr is NULL, pcre_compile() returns NULL immediately. Otherwise,
869 nigel 77 if compilation of a pattern fails, pcre_compile() returns NULL, and
870 nigel 73 sets the variable pointed to by errptr to point to a textual error mes-
871 nigel 87 sage. This is a static string that is part of the library. You must not
872     try to free it. The offset from the start of the pattern to the charac-
873     ter where the error was discovered is placed in the variable pointed to
874     by erroffset, which must not be NULL. If it is, an immediate error is
875 nigel 73 given.
876 nigel 53
877 nigel 87 If pcre_compile2() is used instead of pcre_compile(), and the error-
878     codeptr argument is not NULL, a non-zero error code number is returned
879     via this argument in the event of an error. This is in addition to the
880 nigel 77 textual error message. Error codes and messages are listed below.
882 nigel 87 If the final argument, tableptr, is NULL, PCRE uses a default set of
883     character tables that are built when PCRE is compiled, using the
884     default C locale. Otherwise, tableptr must be an address that is the
885     result of a call to pcre_maketables(). This value is stored with the
886     compiled pattern, and used again by pcre_exec(), unless another table
887 nigel 75 pointer is passed to it. For more discussion, see the section on locale
888     support below.
889 nigel 53
890 nigel 87 This code fragment shows a typical straightforward call to pcre_com-
891 nigel 73 pile():
892 nigel 41
893 nigel 73 pcre *re;
894     const char *error;
895     int erroffset;
896     re = pcre_compile(
897     "^A.*Z", /* the pattern */
898     0, /* default options */
899     &error, /* for error message */
900     &erroffset, /* for error offset */
901     NULL); /* use default character tables */
902 nigel 41
903 nigel 87 The following names for option bits are defined in the pcre.h header
904 nigel 75 file:
905 nigel 41
906 nigel 73 PCRE_ANCHORED
907 nigel 41
908 nigel 73 If this bit is set, the pattern is forced to be "anchored", that is, it
909 nigel 87 is constrained to match only at the first matching point in the string
910     that is being searched (the "subject string"). This effect can also be
911     achieved by appropriate constructs in the pattern itself, which is the
912 nigel 73 only way to do it in Perl.
913 nigel 41
914 nigel 75 PCRE_AUTO_CALLOUT
916     If this bit is set, pcre_compile() automatically inserts callout items,
917 nigel 87 all with number 255, before each pattern item. For discussion of the
918 nigel 75 callout facility, see the pcrecallout documentation.
920 nigel 73 PCRE_CASELESS
921 nigel 41
922 nigel 87 If this bit is set, letters in the pattern match both upper and lower
923     case letters. It is equivalent to Perl's /i option, and it can be
924     changed within a pattern by a (?i) option setting. In UTF-8 mode, PCRE
925     always understands the concept of case for characters whose values are
926     less than 128, so caseless matching is always possible. For characters
927     with higher values, the concept of case is supported if PCRE is com-
928     piled with Unicode property support, but not otherwise. If you want to
929     use caseless matching for characters 128 and above, you must ensure
930     that PCRE is compiled with Unicode property support as well as with
931 nigel 77 UTF-8 support.
932 nigel 41
934 nigel 41
935 nigel 87 If this bit is set, a dollar metacharacter in the pattern matches only
936     at the end of the subject string. Without this option, a dollar also
937     matches immediately before the final character if it is a newline (but
938     not before any other newlines). The PCRE_DOLLAR_ENDONLY option is
939 nigel 73 ignored if PCRE_MULTILINE is set. There is no equivalent to this option
940     in Perl, and no way to set it within a pattern.
941 nigel 41
942 nigel 73 PCRE_DOTALL
943 nigel 41
944 nigel 73 If this bit is set, a dot metacharater in the pattern matches all char-
945 nigel 87 acters, including newlines. Without it, newlines are excluded. This
946     option is equivalent to Perl's /s option, and it can be changed within
947     a pattern by a (?s) option setting. A negative class such as [^a]
948     always matches a newline character, independent of the setting of this
949 nigel 73 option.
950 nigel 63
951 nigel 73 PCRE_EXTENDED
952 nigel 41
953 nigel 87 If this bit is set, whitespace data characters in the pattern are
954 nigel 77 totally ignored except when escaped or inside a character class. White-
955     space does not include the VT character (code 11). In addition, charac-
956     ters between an unescaped # outside a character class and the next new-
957 nigel 87 line character, inclusive, are also ignored. This is equivalent to
958     Perl's /x option, and it can be changed within a pattern by a (?x)
959 nigel 73 option setting.
960 nigel 41
961 nigel 87 This option makes it possible to include comments inside complicated
962     patterns. Note, however, that this applies only to data characters.
963     Whitespace characters may never appear within special character
964     sequences in a pattern, for example within the sequence (?( which
965 nigel 73 introduces a conditional subpattern.
966 nigel 41
967 nigel 73 PCRE_EXTRA
968 nigel 41
969 nigel 87 This option was invented in order to turn on additional functionality
970     of PCRE that is incompatible with Perl, but it is currently of very
971     little use. When set, any backslash in a pattern that is followed by a
972     letter that has no special meaning causes an error, thus reserving
973     these combinations for future expansion. By default, as in Perl, a
974     backslash followed by a letter with no special meaning is treated as a
975     literal. There are at present no other features controlled by this
976 nigel 73 option. It can also be set by a (?X) option setting within a pattern.
977 nigel 41
978 nigel 77 PCRE_FIRSTLINE
980 nigel 87 If this option is set, an unanchored pattern is required to match
981     before or at the first newline character in the subject string, though
982 nigel 77 the matched text may continue over the newline.
984 nigel 73 PCRE_MULTILINE
985 nigel 41
986 nigel 87 By default, PCRE treats the subject string as consisting of a single
987     line of characters (even if it actually contains newlines). The "start
988     of line" metacharacter (^) matches only at the start of the string,
989     while the "end of line" metacharacter ($) matches only at the end of
990 nigel 75 the string, or before a terminating newline (unless PCRE_DOLLAR_ENDONLY
991     is set). This is the same as Perl.
992 nigel 63
993 nigel 87 When PCRE_MULTILINE it is set, the "start of line" and "end of line"
994     constructs match immediately following or immediately before any new-
995     line in the subject string, respectively, as well as at the very start
996     and end. This is equivalent to Perl's /m option, and it can be changed
997 nigel 73 within a pattern by a (?m) option setting. If there are no "\n" charac-
998 nigel 87 ters in a subject string, or no occurrences of ^ or $ in a pattern,
999 nigel 73 setting PCRE_MULTILINE has no effect.
1000 nigel 63
1001 nigel 73 PCRE_NO_AUTO_CAPTURE
1002 nigel 41
1003 nigel 73 If this option is set, it disables the use of numbered capturing paren-
1004 nigel 87 theses in the pattern. Any opening parenthesis that is not followed by
1005     ? behaves as if it were followed by ?: but named parentheses can still
1006     be used for capturing (and they acquire numbers in the usual way).
1007 nigel 73 There is no equivalent of this option in Perl.
1008 nigel 41
1009 nigel 73 PCRE_UNGREEDY
1010 nigel 41
1011 nigel 87 This option inverts the "greediness" of the quantifiers so that they
1012     are not greedy by default, but become greedy if followed by "?". It is
1013     not compatible with Perl. It can also be set by a (?U) option setting
1014 nigel 73 within the pattern.
1015 nigel 41
1016 nigel 73 PCRE_UTF8
1017 nigel 49
1018 nigel 87 This option causes PCRE to regard both the pattern and the subject as
1019     strings of UTF-8 characters instead of single-byte character strings.
1020     However, it is available only when PCRE is built to include UTF-8 sup-
1021     port. If not, the use of this option provokes an error. Details of how
1022     this option changes the behaviour of PCRE are given in the section on
1023 nigel 75 UTF-8 support in the main pcre page.
1024 nigel 71
1025 nigel 73 PCRE_NO_UTF8_CHECK
1026 nigel 71
1027 nigel 73 When PCRE_UTF8 is set, the validity of the pattern as a UTF-8 string is
1028 nigel 87 automatically checked. If an invalid UTF-8 sequence of bytes is found,
1029     pcre_compile() returns an error. If you already know that your pattern
1030     is valid, and you want to skip this check for performance reasons, you
1031     can set the PCRE_NO_UTF8_CHECK option. When it is set, the effect of
1032 nigel 73 passing an invalid UTF-8 string as a pattern is undefined. It may cause
1033 nigel 87 your program to crash. Note that this option can also be passed to
1034     pcre_exec() and pcre_dfa_exec(), to suppress the UTF-8 validity check-
1035 nigel 77 ing of subject strings.
1036 nigel 71
1037 nigel 73
1040 nigel 87 The following table lists the error codes than may be returned by
1041     pcre_compile2(), along with the error messages that may be returned by
1042 nigel 77 both compiling functions.
1044     0 no error
1045     1 \ at end of pattern
1046     2 \c at end of pattern
1047     3 unrecognized character follows \
1048     4 numbers out of order in {} quantifier
1049     5 number too big in {} quantifier
1050     6 missing terminating ] for character class
1051     7 invalid escape sequence in character class
1052     8 range out of order in character class
1053     9 nothing to repeat
1054     10 operand of unlimited repeat could match the empty string
1055     11 internal error: unexpected repeat
1056     12 unrecognized character after (?
1057     13 POSIX named classes are supported only within a class
1058     14 missing )
1059     15 reference to non-existent subpattern
1060     16 erroffset passed as NULL
1061     17 unknown option bit(s) set
1062     18 missing ) after comment
1063     19 parentheses nested too deeply
1064     20 regular expression too large
1065     21 failed to get memory
1066     22 unmatched parentheses
1067     23 internal error: code overflow
1068     24 unrecognized character after (?<
1069     25 lookbehind assertion is not fixed length
1070     26 malformed number after (?(
1071     27 conditional group contains more than two branches
1072     28 assertion expected after (?(
1073     29 (?R or (?digits must be followed by )
1074     30 unknown POSIX class name
1075     31 POSIX collating elements are not supported
1076     32 this version of PCRE is not compiled with PCRE_UTF8 support
1077     33 spare error
1078     34 character value in \x{...} sequence is too large
1079     35 invalid condition (?(0)
1080     36 \C not allowed in lookbehind assertion
1081     37 PCRE does not support \L, \l, \N, \U, or \u
1082     38 number after (?C is > 255
1083     39 closing ) for (?C expected
1084     40 recursive call could loop indefinitely
1085     41 unrecognized character after (?P
1086     42 syntax error after (?P
1087     43 two named groups have the same name
1088     44 invalid UTF-8 string
1089     45 support for \P, \p, and \X has not been compiled
1090     46 malformed \P or \p sequence
1091     47 unknown property name after \P or \p
1094 nigel 63 STUDYING A PATTERN
1095 nigel 49
1096 nigel 77 pcre_extra *pcre_study(const pcre *code, int options
1097 nigel 73 const char **errptr);
1098 nigel 63
1099 nigel 87 If a compiled pattern is going to be used several times, it is worth
1100 nigel 75 spending more time analyzing it in order to speed up the time taken for
1101 nigel 87 matching. The function pcre_study() takes a pointer to a compiled pat-
1102 nigel 75 tern as its first argument. If studying the pattern produces additional
1103 nigel 87 information that will help speed up matching, pcre_study() returns a
1104     pointer to a pcre_extra block, in which the study_data field points to
1105 nigel 75 the results of the study.
1106 nigel 41
1107 nigel 75 The returned value from pcre_study() can be passed directly to
1108 nigel 87 pcre_exec(). However, a pcre_extra block also contains other fields
1109     that can be set by the caller before the block is passed; these are
1110 nigel 75 described below in the section on matching a pattern.
1111 nigel 63
1112 nigel 87 If studying the pattern does not produce any additional information
1113 nigel 75 pcre_study() returns NULL. In that circumstance, if the calling program
1114 nigel 87 wants to pass any of the other fields to pcre_exec(), it must set up
1115 nigel 75 its own pcre_extra block.
1116 nigel 41
1117 nigel 87 The second argument of pcre_study() contains option bits. At present,
1118 nigel 75 no options are defined, and this argument should always be zero.
1120 nigel 87 The third argument for pcre_study() is a pointer for an error message.
1121     If studying succeeds (even if no data is returned), the variable it
1122     points to is set to NULL. Otherwise it is set to point to a textual
1123     error message. This is a static string that is part of the library. You
1124     must not try to free it. You should test the error pointer for NULL
1125     after calling pcre_study(), to be sure that it has run successfully.
1126 nigel 41
1127 nigel 73 This is a typical call to pcre_study():
1128 nigel 53
1129 nigel 73 pcre_extra *pe;
1130     pe = pcre_study(
1131     re, /* result of pcre_compile() */
1132     0, /* no options exist */
1133     &error); /* set to NULL or points to a message */
1134 nigel 53
1135 nigel 73 At present, studying a pattern is useful only for non-anchored patterns
1136 nigel 77 that do not have a single fixed starting character. A bitmap of possi-
1137 nigel 75 ble starting bytes is created.
1138 nigel 41
1140 nigel 63 LOCALE SUPPORT
1141 nigel 41
1142 nigel 77 PCRE handles caseless matching, and determines whether characters are
1143     letters digits, or whatever, by reference to a set of tables, indexed
1144     by character value. When running in UTF-8 mode, this applies only to
1145     characters with codes less than 128. Higher-valued codes never match
1146     escapes such as \w or \d, but can be tested with \p if PCRE is built
1147 nigel 87 with Unicode character property support. The use of locales with Uni-
1148     code is discouraged.
1149 nigel 41
1150 nigel 87 An internal set of tables is created in the default C locale when PCRE
1151     is built. This is used when the final argument of pcre_compile() is
1152     NULL, and is sufficient for many applications. An alternative set of
1153     tables can, however, be supplied. These may be created in a different
1154     locale from the default. As more and more applications change to using
1155 nigel 75 Unicode, the need for this locale support is expected to die away.
1156 nigel 41
1157 nigel 87 External tables are built by calling the pcre_maketables() function,
1158     which has no arguments, in the relevant locale. The result can then be
1159     passed to pcre_compile() or pcre_exec() as often as necessary. For
1160     example, to build and use tables that are appropriate for the French
1161     locale (where accented characters with values greater than 128 are
1162 nigel 75 treated as letters), the following code could be used:
1164     setlocale(LC_CTYPE, "fr_FR");
1165 nigel 73 tables = pcre_maketables();
1166     re = pcre_compile(..., tables);
1167 nigel 41
1168 nigel 87 When pcre_maketables() runs, the tables are built in memory that is
1169     obtained via pcre_malloc. It is the caller's responsibility to ensure
1170     that the memory containing the tables remains available for as long as
1171 nigel 75 it is needed.
1172 nigel 41
1173 nigel 75 The pointer that is passed to pcre_compile() is saved with the compiled
1174 nigel 87 pattern, and the same tables are used via this pointer by pcre_study()
1175 nigel 75 and normally also by pcre_exec(). Thus, by default, for any single pat-
1176     tern, compilation, studying and matching all happen in the same locale,
1177     but different patterns can be compiled in different locales.
1178 nigel 41
1179 nigel 87 It is possible to pass a table pointer or NULL (indicating the use of
1180     the internal tables) to pcre_exec(). Although not intended for this
1181     purpose, this facility could be used to match a pattern in a different
1182 nigel 75 locale from the one in which it was compiled. Passing table pointers at
1183     run time is discussed below in the section on matching a pattern.
1187 nigel 41
1188 nigel 73 int pcre_fullinfo(const pcre *code, const pcre_extra *extra,
1189     int what, void *where);
1190 nigel 63
1191 nigel 87 The pcre_fullinfo() function returns information about a compiled pat-
1192 nigel 73 tern. It replaces the obsolete pcre_info() function, which is neverthe-
1193     less retained for backwards compability (and is documented below).
1194 nigel 43
1195 nigel 87 The first argument for pcre_fullinfo() is a pointer to the compiled
1196     pattern. The second argument is the result of pcre_study(), or NULL if
1197     the pattern was not studied. The third argument specifies which piece
1198     of information is required, and the fourth argument is a pointer to a
1199     variable to receive the data. The yield of the function is zero for
1200 nigel 73 success, or one of the following negative numbers:
1201 nigel 41
1202 nigel 73 PCRE_ERROR_NULL the argument code was NULL
1203     the argument where was NULL
1204     PCRE_ERROR_BADMAGIC the "magic number" was not found
1205     PCRE_ERROR_BADOPTION the value of what was invalid
1206 nigel 53
1207 nigel 87 The "magic number" is placed at the start of each compiled pattern as
1208     an simple check against passing an arbitrary memory pointer. Here is a
1209     typical call of pcre_fullinfo(), to obtain the length of the compiled
1210 nigel 75 pattern:
1211 nigel 53
1212 nigel 73 int rc;
1213     unsigned long int length;
1214     rc = pcre_fullinfo(
1215     re, /* result of pcre_compile() */
1216     pe, /* result of pcre_study(), or NULL */
1217     PCRE_INFO_SIZE, /* what is required */
1218     &length); /* where to put the data */
1219 nigel 43
1220 nigel 87 The possible values for the third argument are defined in pcre.h, and
1221 nigel 73 are as follows:
1222 nigel 43
1224 nigel 41
1225 nigel 87 Return the number of the highest back reference in the pattern. The
1226     fourth argument should point to an int variable. Zero is returned if
1227 nigel 73 there are no back references.
1228 nigel 43
1230 nigel 43
1231 nigel 87 Return the number of capturing subpatterns in the pattern. The fourth
1232 nigel 73 argument should point to an int variable.
1233 nigel 43
1235 nigel 75
1236 nigel 87 Return a pointer to the internal default character tables within PCRE.
1237     The fourth argument should point to an unsigned char * variable. This
1238 nigel 75 information call is provided for internal use by the pcre_study() func-
1239 nigel 87 tion. External callers can cause PCRE to use its internal tables by
1240 nigel 75 passing a NULL table pointer.
1243 nigel 43
1244 nigel 87 Return information about the first byte of any matched string, for a
1245     non-anchored pattern. (This option used to be called
1246     PCRE_INFO_FIRSTCHAR; the old name is still recognized for backwards
1247 nigel 73 compatibility.)
1248 nigel 41
1249 nigel 87 If there is a fixed first byte, for example, from a pattern such as
1250     (cat|cow|coyote), it is returned in the integer pointed to by where.
1251 nigel 73 Otherwise, if either
1252 nigel 41
1253 nigel 87 (a) the pattern was compiled with the PCRE_MULTILINE option, and every
1254 nigel 73 branch starts with "^", or
1255 nigel 43
1256 nigel 73 (b) every branch of the pattern starts with ".*" and PCRE_DOTALL is not
1257     set (if it were set, the pattern would be anchored),
1258 nigel 41
1259 nigel 87 -1 is returned, indicating that the pattern matches only at the start
1260     of a subject string or after any newline within the string. Otherwise
1261 nigel 73 -2 is returned. For anchored patterns, -2 is returned.
1262 nigel 41
1264 nigel 41
1265 nigel 87 If the pattern was studied, and this resulted in the construction of a
1266 nigel 73 256-bit table indicating a fixed set of bytes for the first byte in any
1267 nigel 87 matching string, a pointer to the table is returned. Otherwise NULL is
1268     returned. The fourth argument should point to an unsigned char * vari-
1269 nigel 73 able.
1270 nigel 43
1272 nigel 43
1273 nigel 87 Return the value of the rightmost literal byte that must exist in any
1274     matched string, other than at its start, if such a byte has been
1275 nigel 73 recorded. The fourth argument should point to an int variable. If there
1276 nigel 87 is no such byte, -1 is returned. For anchored patterns, a last literal
1277     byte is recorded only if it follows something of variable length. For
1278 nigel 73 example, for the pattern /^a\d+z\d+/ the returned value is "z", but for
1279     /^a\dz\d/ the returned value is -1.
1280 nigel 63
1284 nigel 63
1285 nigel 87 PCRE supports the use of named as well as numbered capturing parenthe-
1286     ses. The names are just an additional way of identifying the parenthe-
1287 nigel 75 ses, which still acquire numbers. A convenience function called
1288 nigel 87 pcre_get_named_substring() is provided for extracting an individual
1289     captured substring by name. It is also possible to extract the data
1290     directly, by first converting the name to a number in order to access
1291     the correct pointers in the output vector (described with pcre_exec()
1292     below). To do the conversion, you need to use the name-to-number map,
1293 nigel 75 which is described by these three values.
1294 nigel 63
1295 nigel 73 The map consists of a number of fixed-size entries. PCRE_INFO_NAMECOUNT
1296     gives the number of entries, and PCRE_INFO_NAMEENTRYSIZE gives the size
1297 nigel 87 of each entry; both of these return an int value. The entry size
1298     depends on the length of the longest name. PCRE_INFO_NAMETABLE returns
1299     a pointer to the first entry of the table (a pointer to char). The
1300 nigel 73 first two bytes of each entry are the number of the capturing parenthe-
1301 nigel 87 sis, most significant byte first. The rest of the entry is the corre-
1302     sponding name, zero terminated. The names are in alphabetical order.
1303     For example, consider the following pattern (assume PCRE_EXTENDED is
1304 nigel 73 set, so white space - including newlines - is ignored):
1305 nigel 63
1306 nigel 73 (?P<date> (?P<year>(\d\d)?\d\d) -
1307     (?P<month>\d\d) - (?P<day>\d\d) )
1308 nigel 63
1309 nigel 87 There are four named subpatterns, so the table has four entries, and
1310     each entry in the table is eight bytes long. The table is as follows,
1311 nigel 75 with non-printing bytes shows in hexadecimal, and undefined bytes shown
1312     as ??:
1313 nigel 63
1314 nigel 73 00 01 d a t e 00 ??
1315     00 05 d a y 00 ?? ??
1316     00 04 m o n t h 00
1317     00 02 y e a r 00 ??
1318 nigel 63
1319 nigel 87 When writing code to extract data from named subpatterns using the
1320 nigel 75 name-to-number map, remember that the length of each entry is likely to
1321     be different for each compiled pattern.
1322 nigel 63
1323 nigel 73 PCRE_INFO_OPTIONS
1324 nigel 63
1325 nigel 87 Return a copy of the options with which the pattern was compiled. The
1326     fourth argument should point to an unsigned long int variable. These
1327 nigel 73 option bits are those specified in the call to pcre_compile(), modified
1328     by any top-level option settings within the pattern itself.
1329 nigel 63
1330 nigel 87 A pattern is automatically anchored by PCRE if all of its top-level
1331 nigel 73 alternatives begin with one of the following:
1332 nigel 63
1333 nigel 73 ^ unless PCRE_MULTILINE is set
1334     \A always
1335     \G always
1336     .* if PCRE_DOTALL is set and there are no back
1337     references to the subpattern in which .* appears
1338 nigel 63
1339 nigel 73 For such patterns, the PCRE_ANCHORED bit is set in the options returned
1340     by pcre_fullinfo().
1341 nigel 63
1342 nigel 73 PCRE_INFO_SIZE
1343 nigel 63
1344 nigel 87 Return the size of the compiled pattern, that is, the value that was
1345 nigel 73 passed as the argument to pcre_malloc() when PCRE was getting memory in
1346     which to place the compiled data. The fourth argument should point to a
1347     size_t variable.
1348 nigel 63
1350 nigel 63
1351 nigel 75 Return the size of the data block pointed to by the study_data field in
1352 nigel 87 a pcre_extra block. That is, it is the value that was passed to
1353 nigel 73 pcre_malloc() when PCRE was getting memory into which to place the data
1354 nigel 87 created by pcre_study(). The fourth argument should point to a size_t
1355 nigel 73 variable.
1356 nigel 63
1357 nigel 73
1360 nigel 73 int pcre_info(const pcre *code, int *optptr, int *firstcharptr);
1361 nigel 63
1362 nigel 87 The pcre_info() function is now obsolete because its interface is too
1363     restrictive to return all the available data about a compiled pattern.
1364     New programs should use pcre_fullinfo() instead. The yield of
1365     pcre_info() is the number of capturing subpatterns, or one of the fol-
1366 nigel 73 lowing negative numbers:
1367 nigel 43
1368 nigel 73 PCRE_ERROR_NULL the argument code was NULL
1369     PCRE_ERROR_BADMAGIC the "magic number" was not found
1370 nigel 43
1371 nigel 87 If the optptr argument is not NULL, a copy of the options with which
1372     the pattern was compiled is placed in the integer it points to (see
1373 nigel 73 PCRE_INFO_OPTIONS above).
1374 nigel 43
1375 nigel 87 If the pattern is not anchored and the firstcharptr argument is not
1376     NULL, it is used to pass back information about the first character of
1377 nigel 73 any matched string (see PCRE_INFO_FIRSTBYTE above).
1378 nigel 43
1380 nigel 77 REFERENCE COUNTS
1381 nigel 53
1382 nigel 77 int pcre_refcount(pcre *code, int adjust);
1384 nigel 87 The pcre_refcount() function is used to maintain a reference count in
1385 nigel 77 the data block that contains a compiled pattern. It is provided for the
1386 nigel 87 benefit of applications that operate in an object-oriented manner,
1387 nigel 77 where different parts of the application may be using the same compiled
1388     pattern, but you want to free the block when they are all done.
1390     When a pattern is compiled, the reference count field is initialized to
1391 nigel 87 zero. It is changed only by calling this function, whose action is to
1392     add the adjust value (which may be positive or negative) to it. The
1393 nigel 77 yield of the function is the new value. However, the value of the count
1394 nigel 87 is constrained to lie between 0 and 65535, inclusive. If the new value
1395 nigel 77 is outside these limits, it is forced to the appropriate limit value.
1397 nigel 87 Except when it is zero, the reference count is not correctly preserved
1398     if a pattern is compiled on one host and then transferred to a host
1399 nigel 77 whose byte-order is different. (This seems a highly unlikely scenario.)
1404 nigel 73 int pcre_exec(const pcre *code, const pcre_extra *extra,
1405     const char *subject, int length, int startoffset,
1406     int options, int *ovector, int ovecsize);
1407 nigel 53
1408 nigel 87 The function pcre_exec() is called to match a subject string against a
1409     compiled pattern, which is passed in the code argument. If the pattern
1410 nigel 75 has been studied, the result of the study should be passed in the extra
1411 nigel 87 argument. This function is the main matching facility of the library,
1412 nigel 77 and it operates in a Perl-like manner. For specialist use there is also
1413 nigel 87 an alternative matching function, which is described below in the sec-
1414 nigel 77 tion about the pcre_dfa_exec() function.
1415 nigel 41
1416 nigel 87 In most applications, the pattern will have been compiled (and option-
1417     ally studied) in the same process that calls pcre_exec(). However, it
1418 nigel 75 is possible to save compiled patterns and study data, and then use them
1419 nigel 87 later in different processes, possibly even on different hosts. For a
1420 nigel 75 discussion about this, see the pcreprecompile documentation.
1422 nigel 73 Here is an example of a simple call to pcre_exec():
1423 nigel 53
1424 nigel 73 int rc;
1425     int ovector[30];
1426     rc = pcre_exec(
1427     re, /* result of pcre_compile() */
1428     NULL, /* we didn't study the pattern */
1429     "some string", /* the subject string */
1430     11, /* the length of the subject string */
1431     0, /* start at offset 0 in the subject */
1432     0, /* default options */
1433 nigel 75 ovector, /* vector of integers for substring information */
1434 nigel 77 30); /* number of elements (NOT size in bytes) */
1435 nigel 53
1436 nigel 75 Extra data for pcre_exec()
1437 nigel 63
1438 nigel 87 If the extra argument is not NULL, it must point to a pcre_extra data
1439     block. The pcre_study() function returns such a block (when it doesn't
1440     return NULL), but you can also create one for yourself, and pass addi-
1441     tional information in it. The pcre_extra block contains the following
1442     fields (not necessarily in this order):
1443 nigel 75
1444 nigel 73 unsigned long int flags;
1445     void *study_data;
1446     unsigned long int match_limit;
1447 nigel 87 unsigned long int match_limit_recursion;
1448 nigel 73 void *callout_data;
1449 nigel 75 const unsigned char *tables;
1450 nigel 63
1451 nigel 87 The flags field is a bitmap that specifies which of the other fields
1452 nigel 73 are set. The flag bits are:
1453 nigel 63
1458 nigel 75 PCRE_EXTRA_TABLES
1459 nigel 63
1460 nigel 87 Other flag bits should be set to zero. The study_data field is set in
1461     the pcre_extra block that is returned by pcre_study(), together with
1462 nigel 75 the appropriate flag bit. You should not set this yourself, but you may
1463 nigel 87 add to the block by setting the other fields and their corresponding
1464 nigel 75 flag bits.
1465 nigel 63
1466 nigel 73 The match_limit field provides a means of preventing PCRE from using up
1467 nigel 87 a vast amount of resources when running patterns that are not going to
1468     match, but which have a very large number of possibilities in their
1469     search trees. The classic example is the use of nested unlimited
1470 nigel 75 repeats.
1471 nigel 63
1472 nigel 87 Internally, PCRE uses a function called match() which it calls repeat-
1473     edly (sometimes recursively). The limit set by match_limit is imposed
1474     on the number of times this function is called during a match, which
1475     has the effect of limiting the amount of backtracking that can take
1476     place. For patterns that are not anchored, the count restarts from zero
1477     for each position in the subject string.
1478 nigel 75
1479 nigel 87 The default value for the limit can be set when PCRE is built; the
1480     default default is 10 million, which handles all but the most extreme
1481     cases. You can override the default by suppling pcre_exec() with a
1482     pcre_extra block in which match_limit is set, and
1483     PCRE_EXTRA_MATCH_LIMIT is set in the flags field. If the limit is
1484 nigel 73 exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
1485 nigel 63
1486 nigel 87 The match_limit_recursion field is similar to match_limit, but instead
1487     of limiting the total number of times that match() is called, it limits
1488     the depth of recursion. The recursion depth is a smaller number than
1489     the total number of calls, because not all calls to match() are recur-
1490     sive. This limit is of use only if it is set smaller than match_limit.
1492     Limiting the recursion depth limits the amount of stack that can be
1493     used, or, when PCRE has been compiled to use memory on the heap instead
1494     of the stack, the amount of heap memory that can be used.
1496     The default value for match_limit_recursion can be set when PCRE is
1497     built; the default default is the same value as the default for
1498     match_limit. You can override the default by suppling pcre_exec() with
1499     a pcre_extra block in which match_limit_recursion is set, and
1500     PCRE_EXTRA_MATCH_LIMIT_RECURSION is set in the flags field. If the
1501     limit is exceeded, pcre_exec() returns PCRE_ERROR_RECURSIONLIMIT.
1503     The pcre_callout field is used in conjunction with the "callout" fea-
1504 nigel 73 ture, which is described in the pcrecallout documentation.
1505 nigel 63
1506 nigel 87 The tables field is used to pass a character tables pointer to
1507     pcre_exec(); this overrides the value that is stored with the compiled
1508     pattern. A non-NULL value is stored with the compiled pattern only if
1509     custom tables were supplied to pcre_compile() via its tableptr argu-
1510 nigel 75 ment. If NULL is passed to pcre_exec() using this mechanism, it forces
1511 nigel 87 PCRE's internal tables to be used. This facility is helpful when re-
1512     using patterns that have been saved after compiling with an external
1513     set of tables, because the external tables might be at a different
1514     address when pcre_exec() is called. See the pcreprecompile documenta-
1515 nigel 75 tion for a discussion of saving compiled patterns for later use.
1516 nigel 41
1517 nigel 75 Option bits for pcre_exec()
1518 nigel 71
1519 nigel 87 The unused bits of the options argument for pcre_exec() must be zero.
1520     The only bits that may be set are PCRE_ANCHORED, PCRE_NOTBOL,
1522 nigel 41
1523 nigel 75 PCRE_ANCHORED
1524 nigel 41
1525 nigel 87 The PCRE_ANCHORED option limits pcre_exec() to matching at the first
1526     matching position. If a pattern was compiled with PCRE_ANCHORED, or
1527     turned out to be anchored by virtue of its contents, it cannot be made
1528 nigel 75 unachored at matching time.
1530 nigel 73 PCRE_NOTBOL
1531 nigel 41
1532 nigel 75 This option specifies that first character of the subject string is not
1533 nigel 87 the beginning of a line, so the circumflex metacharacter should not
1534     match before it. Setting this without PCRE_MULTILINE (at compile time)
1535     causes circumflex never to match. This option affects only the behav-
1536 nigel 77 iour of the circumflex metacharacter. It does not affect \A.
1537 nigel 41
1538 nigel 73 PCRE_NOTEOL
1539 nigel 41
1540 nigel 75 This option specifies that the end of the subject string is not the end
1541 nigel 87 of a line, so the dollar metacharacter should not match it nor (except
1542     in multiline mode) a newline immediately before it. Setting this with-
1543 nigel 75 out PCRE_MULTILINE (at compile time) causes dollar never to match. This
1544 nigel 87 option affects only the behaviour of the dollar metacharacter. It does
1545 nigel 75 not affect \Z or \z.
1546 nigel 41
1547 nigel 73 PCRE_NOTEMPTY
1548 nigel 41
1549 nigel 73 An empty string is not considered to be a valid match if this option is
1550 nigel 87 set. If there are alternatives in the pattern, they are tried. If all
1551     the alternatives match the empty string, the entire match fails. For
1552 nigel 73 example, if the pattern
1553 nigel 41
1554 nigel 73 a?b?
1555 nigel 41
1556 nigel 87 is applied to a string not beginning with "a" or "b", it matches the
1557     empty string at the start of the subject. With PCRE_NOTEMPTY set, this
1558 nigel 73 match is not valid, so PCRE searches further into the string for occur-
1559     rences of "a" or "b".
1560 nigel 41
1561 nigel 73 Perl has no direct equivalent of PCRE_NOTEMPTY, but it does make a spe-
1562 nigel 87 cial case of a pattern match of the empty string within its split()
1563     function, and when using the /g modifier. It is possible to emulate
1564 nigel 73 Perl's behaviour after matching a null string by first trying the match
1565 nigel 75 again at the same offset with PCRE_NOTEMPTY and PCRE_ANCHORED, and then
1566 nigel 87 if that fails by advancing the starting offset (see below) and trying
1567 nigel 75 an ordinary match again. There is some code that demonstrates how to do
1568     this in the pcredemo.c sample program.
1569 nigel 41
1570 nigel 75 PCRE_NO_UTF8_CHECK
1572     When PCRE_UTF8 is set at compile time, the validity of the subject as a
1573 nigel 87 UTF-8 string is automatically checked when pcre_exec() is subsequently
1574     called. The value of startoffset is also checked to ensure that it
1575     points to the start of a UTF-8 character. If an invalid UTF-8 sequence
1576 nigel 75 of bytes is found, pcre_exec() returns the error PCRE_ERROR_BADUTF8. If
1577 nigel 87 startoffset contains an invalid value, PCRE_ERROR_BADUTF8_OFFSET is
1578 nigel 75 returned.
1580 nigel 87 If you already know that your subject is valid, and you want to skip
1581     these checks for performance reasons, you can set the
1582     PCRE_NO_UTF8_CHECK option when calling pcre_exec(). You might want to
1583     do this for the second and subsequent calls to pcre_exec() if you are
1584     making repeated calls to find all the matches in a single subject
1585     string. However, you should be sure that the value of startoffset
1586     points to the start of a UTF-8 character. When PCRE_NO_UTF8_CHECK is
1587     set, the effect of passing an invalid UTF-8 string as a subject, or a
1588     value of startoffset that does not point to the start of a UTF-8 char-
1589 nigel 75 acter, is undefined. Your program may crash.
1593 nigel 87 This option turns on the partial matching feature. If the subject
1594     string fails to match the pattern, but at some point during the match-
1595     ing process the end of the subject was reached (that is, the subject
1596     partially matches the pattern and the failure to match occurred only
1597     because there were not enough subject characters), pcre_exec() returns
1599     used, there are restrictions on what may appear in the pattern. These
1600 nigel 75 are discussed in the pcrepartial documentation.
1602     The string to be matched by pcre_exec()
1604 nigel 87 The subject string is passed to pcre_exec() as a pointer in subject, a
1605     length in length, and a starting byte offset in startoffset. In UTF-8
1606     mode, the byte offset must point to the start of a UTF-8 character.
1607     Unlike the pattern string, the subject may contain binary zero bytes.
1608     When the starting offset is zero, the search for a match starts at the
1609 nigel 75 beginning of the subject, and this is by far the most common case.
1610 nigel 63
1611 nigel 87 A non-zero starting offset is useful when searching for another match
1612     in the same subject by calling pcre_exec() again after a previous suc-
1613     cess. Setting startoffset differs from just passing over a shortened
1614     string and setting PCRE_NOTBOL in the case of a pattern that begins
1615 nigel 73 with any kind of lookbehind. For example, consider the pattern
1616 nigel 41
1617 nigel 73 \Biss\B
1618 nigel 41
1619 nigel 87 which finds occurrences of "iss" in the middle of words. (\B matches
1620     only if the current position in the subject is not a word boundary.)
1621     When applied to the string "Mississipi" the first call to pcre_exec()
1622     finds the first occurrence. If pcre_exec() is called again with just
1623     the remainder of the subject, namely "issipi", it does not match,
1624 nigel 73 because \B is always false at the start of the subject, which is deemed
1625 nigel 87 to be a word boundary. However, if pcre_exec() is passed the entire
1626 nigel 75 string again, but with startoffset set to 4, it finds the second occur-
1627 nigel 87 rence of "iss" because it is able to look behind the starting point to
1628 nigel 75 discover that it is preceded by a letter.
1629 nigel 41
1630 nigel 87 If a non-zero starting offset is passed when the pattern is anchored,
1631 nigel 75 one attempt to match at the given offset is made. This can only succeed
1632 nigel 87 if the pattern does not require the match to be at the start of the
1633 nigel 75 subject.
1634 nigel 41
1635 nigel 75 How pcre_exec() returns captured substrings
1637 nigel 87 In general, a pattern matches a certain portion of the subject, and in
1638     addition, further substrings from the subject may be picked out by
1639     parts of the pattern. Following the usage in Jeffrey Friedl's book,
1640     this is called "capturing" in what follows, and the phrase "capturing
1641     subpattern" is used for a fragment of a pattern that picks out a sub-
1642     string. PCRE supports several other kinds of parenthesized subpattern
1643 nigel 73 that do not cause substrings to be captured.
1644 nigel 65
1645 nigel 87 Captured substrings are returned to the caller via a vector of integer
1646     offsets whose address is passed in ovector. The number of elements in
1647     the vector is passed in ovecsize, which must be a non-negative number.
1648 nigel 75 Note: this argument is NOT the size of ovector in bytes.
1649 nigel 41
1650 nigel 87 The first two-thirds of the vector is used to pass back captured sub-
1651     strings, each substring using a pair of integers. The remaining third
1652     of the vector is used as workspace by pcre_exec() while matching cap-
1653     turing subpatterns, and is not available for passing back information.
1654     The length passed in ovecsize should always be a multiple of three. If
1655 nigel 75 it is not, it is rounded down.
1657 nigel 87 When a match is successful, information about captured substrings is
1658     returned in pairs of integers, starting at the beginning of ovector,
1659     and continuing up to two-thirds of its length at the most. The first
1660 nigel 73 element of a pair is set to the offset of the first character in a sub-
1661 nigel 87 string, and the second is set to the offset of the first character
1662     after the end of a substring. The first pair, ovector[0] and ovec-
1663     tor[1], identify the portion of the subject string matched by the
1664     entire pattern. The next pair is used for the first capturing subpat-
1665     tern, and so on. The value returned by pcre_exec() is the number of
1666     pairs that have been set. If there are no capturing subpatterns, the
1667     return value from a successful match is 1, indicating that just the
1668 nigel 73 first pair of offsets has been set.
1669 nigel 41
1670 nigel 87 Some convenience functions are provided for extracting the captured
1671     substrings as separate strings. These are described in the following
1672 nigel 73 section.
1673 nigel 41
1674 nigel 87 It is possible for an capturing subpattern number n+1 to match some
1675     part of the subject when subpattern n has not been used at all. For
1676 nigel 73 example, if the string "abc" is matched against the pattern (a|(z))(bc)
1677 nigel 87 subpatterns 1 and 3 are matched, but 2 is not. When this happens, both
1678 nigel 73 offset values corresponding to the unused subpattern are set to -1.
1679 nigel 41
1680 nigel 73 If a capturing subpattern is matched repeatedly, it is the last portion
1681 nigel 75 of the string that it matched that is returned.
1682 nigel 41
1683 nigel 87 If the vector is too small to hold all the captured substring offsets,
1684 nigel 75 it is used as far as possible (up to two-thirds of its length), and the
1685 nigel 87 function returns a value of zero. In particular, if the substring off-
1686 nigel 75 sets are not of interest, pcre_exec() may be called with ovector passed
1687 nigel 87 as NULL and ovecsize as zero. However, if the pattern contains back
1688     references and the ovector is not big enough to remember the related
1689     substrings, PCRE has to get additional memory for use during matching.
1690 nigel 73 Thus it is usually advisable to supply an ovector.
1691 nigel 41
1692 nigel 87 Note that pcre_info() can be used to find out how many capturing sub-
1693 nigel 73 patterns there are in a compiled pattern. The smallest size for ovector
1694 nigel 87 that will allow for n captured substrings, in addition to the offsets
1695 nigel 73 of the substring matched by the whole pattern, is (n+1)*3.
1696 nigel 41
1697 nigel 75 Return values from pcre_exec()
1699 nigel 87 If pcre_exec() fails, it returns a negative number. The following are
1700 nigel 73 defined in the header file:
1701 nigel 41
1702 nigel 73 PCRE_ERROR_NOMATCH (-1)
1703 nigel 41
1704 nigel 73 The subject string did not match the pattern.
1705 nigel 41
1706 nigel 73 PCRE_ERROR_NULL (-2)
1707 nigel 41
1708 nigel 87 Either code or subject was passed as NULL, or ovector was NULL and
1709 nigel 73 ovecsize was not zero.
1710 nigel 41
1711 nigel 73 PCRE_ERROR_BADOPTION (-3)
1712 nigel 41
1713 nigel 73 An unrecognized bit was set in the options argument.
1714 nigel 41
1715 nigel 73 PCRE_ERROR_BADMAGIC (-4)
1716 nigel 41
1717 nigel 87 PCRE stores a 4-byte "magic number" at the start of the compiled code,
1718 nigel 75 to catch the case when it is passed a junk pointer and to detect when a
1719     pattern that was compiled in an environment of one endianness is run in
1720 nigel 87 an environment with the other endianness. This is the error that PCRE
1721 nigel 75 gives when the magic number is not present.
1722 nigel 41
1723 nigel 73 PCRE_ERROR_UNKNOWN_NODE (-5)
1724 nigel 41
1725 nigel 73 While running the pattern match, an unknown item was encountered in the
1726 nigel 87 compiled pattern. This error could be caused by a bug in PCRE or by
1727 nigel 73 overwriting of the compiled pattern.
1728 nigel 41
1729 nigel 73 PCRE_ERROR_NOMEMORY (-6)
1730 nigel 41
1731 nigel 87 If a pattern contains back references, but the ovector that is passed
1732 nigel 73 to pcre_exec() is not big enough to remember the referenced substrings,
1733 nigel 87 PCRE gets a block of memory at the start of matching to use for this
1734     purpose. If the call via pcre_malloc() fails, this error is given. The
1735 nigel 75 memory is automatically freed at the end of matching.
1736 nigel 41
1737 nigel 73 PCRE_ERROR_NOSUBSTRING (-7)
1738 nigel 53
1739 nigel 87 This error is used by the pcre_copy_substring(), pcre_get_substring(),
1740 nigel 73 and pcre_get_substring_list() functions (see below). It is never
1741     returned by pcre_exec().
1742 nigel 63
1743 nigel 73 PCRE_ERROR_MATCHLIMIT (-8)
1744 nigel 63
1745 nigel 87 The backtracking limit, as specified by the match_limit field in a
1746     pcre_extra structure (or defaulted) was reached. See the description
1747     above.
1751     The internal recursion limit, as specified by the match_limit_recursion
1752     field in a pcre_extra structure (or defaulted) was reached. See the
1753 nigel 73 description above.
1754 nigel 63
1755 nigel 73 PCRE_ERROR_CALLOUT (-9)
1756 nigel 63
1757 nigel 73 This error is never generated by pcre_exec() itself. It is provided for
1758 nigel 87 use by callout functions that want to yield a distinctive error code.
1759 nigel 73 See the pcrecallout documentation for details.
1760 nigel 71
1761 nigel 73 PCRE_ERROR_BADUTF8 (-10)
1762 nigel 71
1763 nigel 87 A string that contains an invalid UTF-8 byte sequence was passed as a
1764 nigel 73 subject.
1768     The UTF-8 byte sequence that was passed as a subject was valid, but the
1769 nigel 87 value of startoffset did not point to the beginning of a UTF-8 charac-
1770 nigel 73 ter.
1772 nigel 77 PCRE_ERROR_PARTIAL (-12)
1773 nigel 73
1774 nigel 87 The subject string did not match, but it did match partially. See the
1775 nigel 75 pcrepartial documentation for details of partial matching.
1777 nigel 77 PCRE_ERROR_BADPARTIAL (-13)
1778 nigel 75
1779 nigel 87 The PCRE_PARTIAL option was used with a compiled pattern containing
1780     items that are not supported for partial matching. See the pcrepartial
1781 nigel 75 documentation for details of partial matching.
1783 nigel 77 PCRE_ERROR_INTERNAL (-14)
1784 nigel 75
1785 nigel 87 An unexpected internal error has occurred. This error could be caused
1786 nigel 75 by a bug in PCRE or by overwriting of the compiled pattern.
1788 nigel 77 PCRE_ERROR_BADCOUNT (-15)
1789 nigel 75
1790 nigel 87 This error is given if the value of the ovecsize argument is negative.
1791 nigel 75
1795 nigel 73 int pcre_copy_substring(const char *subject, int *ovector,
1796     int stringcount, int stringnumber, char *buffer,
1797     int buffersize);
1798 nigel 63
1799 nigel 73 int pcre_get_substring(const char *subject, int *ovector,
1800     int stringcount, int stringnumber,
1801     const char **stringptr);
1802 nigel 63
1803 nigel 73 int pcre_get_substring_list(const char *subject,
1804     int *ovector, int stringcount, const char ***listptr);
1805 nigel 63
1806 nigel 87 Captured substrings can be accessed directly by using the offsets
1807     returned by pcre_exec() in ovector. For convenience, the functions
1808 nigel 73 pcre_copy_substring(), pcre_get_substring(), and pcre_get_sub-
1809 nigel 87 string_list() are provided for extracting captured substrings as new,
1810     separate, zero-terminated strings. These functions identify substrings
1811     by number. The next section describes functions for extracting named
1812     substrings. A substring that contains a binary zero is correctly
1813     extracted and has a further zero added on the end, but the result is
1814 nigel 73 not, of course, a C string.
1815 nigel 41
1816 nigel 87 The first three arguments are the same for all three of these func-
1817     tions: subject is the subject string that has just been successfully
1818 nigel 73 matched, ovector is a pointer to the vector of integer offsets that was
1819     passed to pcre_exec(), and stringcount is the number of substrings that
1820 nigel 87 were captured by the match, including the substring that matched the
1821 nigel 75 entire regular expression. This is the value returned by pcre_exec() if
1822 nigel 87 it is greater than zero. If pcre_exec() returned zero, indicating that
1823     it ran out of space in ovector, the value passed as stringcount should
1824 nigel 75 be the number of elements in the vector divided by three.
1825 nigel 41
1826 nigel 87 The functions pcre_copy_substring() and pcre_get_substring() extract a
1827     single substring, whose number is given as stringnumber. A value of
1828     zero extracts the substring that matched the entire pattern, whereas
1829     higher values extract the captured substrings. For pcre_copy_sub-
1830     string(), the string is placed in buffer, whose length is given by
1831     buffersize, while for pcre_get_substring() a new block of memory is
1832     obtained via pcre_malloc, and its address is returned via stringptr.
1833     The yield of the function is the length of the string, not including
1834 nigel 73 the terminating zero, or one of
1835 nigel 41
1836 nigel 73 PCRE_ERROR_NOMEMORY (-6)
1837 nigel 41
1838 nigel 87 The buffer was too small for pcre_copy_substring(), or the attempt to
1839 nigel 73 get memory failed for pcre_get_substring().
1840 nigel 41
1841 nigel 73 PCRE_ERROR_NOSUBSTRING (-7)
1842 nigel 41
1843 nigel 73 There is no substring whose number is stringnumber.
1844 nigel 41
1845 nigel 87 The pcre_get_substring_list() function extracts all available sub-
1846     strings and builds a list of pointers to them. All this is done in a
1847 nigel 75 single block of memory that is obtained via pcre_malloc. The address of
1848 nigel 87 the memory block is returned via listptr, which is also the start of
1849     the list of string pointers. The end of the list is marked by a NULL
1850 nigel 73 pointer. The yield of the function is zero if all went well, or
1851 nigel 41
1852 nigel 73 PCRE_ERROR_NOMEMORY (-6)
1853 nigel 41
1854 nigel 73 if the attempt to get the memory block failed.
1855 nigel 41
1856 nigel 87 When any of these functions encounter a substring that is unset, which
1857     can happen when capturing subpattern number n+1 matches some part of
1858     the subject, but subpattern n has not been used at all, they return an
1859 nigel 73 empty string. This can be distinguished from a genuine zero-length sub-
1860 nigel 87 string by inspecting the appropriate offset in ovector, which is nega-
1861 nigel 73 tive for unset substrings.
1862 nigel 41
1863 nigel 87 The two convenience functions pcre_free_substring() and pcre_free_sub-
1864     string_list() can be used to free the memory returned by a previous
1865 nigel 75 call of pcre_get_substring() or pcre_get_substring_list(), respec-
1866 nigel 87 tively. They do nothing more than call the function pointed to by
1867     pcre_free, which of course could be called directly from a C program.
1868     However, PCRE is used in some situations where it is linked via a spe-
1869 nigel 73 cial interface to another programming language which cannot use
1870 nigel 87 pcre_free directly; it is for these cases that the functions are pro-
1871 nigel 77 vided.
1872 nigel 41
1873 nigel 73
1875 nigel 41
1876 nigel 75 int pcre_get_stringnumber(const pcre *code,
1877     const char *name);
1879 nigel 73 int pcre_copy_named_substring(const pcre *code,
1880     const char *subject, int *ovector,
1881     int stringcount, const char *stringname,
1882     char *buffer, int buffersize);
1883 nigel 41
1884 nigel 73 int pcre_get_named_substring(const pcre *code,
1885     const char *subject, int *ovector,
1886     int stringcount, const char *stringname,
1887     const char **stringptr);
1888 nigel 41
1889 nigel 87 To extract a substring by name, you first have to find associated num-
1890 nigel 75 ber. For example, for this pattern
1891 nigel 41
1892 nigel 79 (a+)b(?P<xxx>\d+)...
1893 nigel 63
1894 nigel 75 the number of the subpattern called "xxx" is 2. You can find the number
1895     from the name by calling pcre_get_stringnumber(). The first argument is
1896 nigel 87 the compiled pattern, and the second is the name. The yield of the
1897     function is the subpattern number, or PCRE_ERROR_NOSUBSTRING (-7) if
1898 nigel 75 there is no subpattern of that name.
1899 nigel 63
1900 nigel 75 Given the number, you can extract the substring directly, or use one of
1901     the functions described in the previous section. For convenience, there
1902     are also two functions that do the whole job.
1904 nigel 87 Most of the arguments of pcre_copy_named_substring() and
1905     pcre_get_named_substring() are the same as those for the similarly
1906     named functions that extract by number. As these are described in the
1907     previous section, they are not re-described here. There are just two
1908 nigel 75 differences:
1909 nigel 63
1910 nigel 87 First, instead of a substring number, a substring name is given. Sec-
1911 nigel 73 ond, there is an extra argument, given at the start, which is a pointer
1912 nigel 87 to the compiled pattern. This is needed in order to gain access to the
1913 nigel 73 name-to-number translation table.
1914 nigel 63
1915 nigel 87 These functions call pcre_get_stringnumber(), and if it succeeds, they
1916     then call pcre_copy_substring() or pcre_get_substring(), as appropri-
1917 nigel 73 ate.
1918 nigel 63
1919 nigel 77
1922 nigel 87 The traditional matching function uses a similar algorithm to Perl,
1923 nigel 77 which stops when it finds the first match, starting at a given point in
1924 nigel 87 the subject. If you want to find all possible matches, or the longest
1925     possible match, consider using the alternative matching function (see
1926     below) instead. If you cannot use the alternative function, but still
1927     need to find all possible matches, you can kludge it up by making use
1928 nigel 77 of the callout facility, which is described in the pcrecallout documen-
1929     tation.
1931     What you have to do is to insert a callout right at the end of the pat-
1932 nigel 87 tern. When your callout function is called, extract and save the cur-
1933     rent matched substring. Then return 1, which forces pcre_exec() to
1934     backtrack and try other alternatives. Ultimately, when it runs out of
1935 nigel 77 matches, pcre_exec() will yield PCRE_ERROR_NOMATCH.
1940     int pcre_dfa_exec(const pcre *code, const pcre_extra *extra,
1941     const char *subject, int length, int startoffset,
1942     int options, int *ovector, int ovecsize,
1943     int *workspace, int wscount);
1945 nigel 87 The function pcre_dfa_exec() is called to match a subject string
1946     against a compiled pattern, using a "DFA" matching algorithm. This has
1947     different characteristics to the normal algorithm, and is not compati-
1948 nigel 77 ble with Perl. Some of the features of PCRE patterns are not supported.
1949     Nevertheless, there are times when this kind of matching can be useful.
1950 nigel 87 For a discussion of the two matching algorithms, see the pcrematching
1951 nigel 77 documentation.
1953 nigel 87 The arguments for the pcre_dfa_exec() function are the same as for
1954 nigel 77 pcre_exec(), plus two extras. The ovector argument is used in a differ-
1955 nigel 87 ent way, and this is described below. The other common arguments are
1956     used in the same way as for pcre_exec(), so their description is not
1957 nigel 77 repeated here.
1959 nigel 87 The two additional arguments provide workspace for the function. The
1960     workspace vector should contain at least 20 elements. It is used for
1961 nigel 77 keeping track of multiple paths through the pattern tree. More
1962 nigel 87 workspace will be needed for patterns and subjects where there are a
1963 nigel 77 lot of possible matches.
1965 nigel 87 Here is an example of a simple call to pcre_dfa_exec():
1966 nigel 77
1967     int rc;
1968     int ovector[10];
1969     int wspace[20];
1970 nigel 87 rc = pcre_dfa_exec(
1971 nigel 77 re, /* result of pcre_compile() */
1972     NULL, /* we didn't study the pattern */
1973     "some string", /* the subject string */
1974     11, /* the length of the subject string */
1975     0, /* start at offset 0 in the subject */
1976     0, /* default options */
1977     ovector, /* vector of integers for substring information */
1978     10, /* number of elements (NOT size in bytes) */
1979     wspace, /* working space vector */
1980     20); /* number of elements (NOT size in bytes) */
1982     Option bits for pcre_dfa_exec()
1984 nigel 87 The unused bits of the options argument for pcre_dfa_exec() must be
1985     zero. The only bits that may be set are PCRE_ANCHORED, PCRE_NOTBOL,
1987     PCRE_DFA_SHORTEST, and PCRE_DFA_RESTART. All but the last three of
1988     these are the same as for pcre_exec(), so their description is not
1989 nigel 77 repeated here.
1993 nigel 87 This has the same general effect as it does for pcre_exec(), but the
1994     details are slightly different. When PCRE_PARTIAL is set for
1995     pcre_dfa_exec(), the return code PCRE_ERROR_NOMATCH is converted into
1996     PCRE_ERROR_PARTIAL if the end of the subject is reached, there have
1997 nigel 77 been no complete matches, but there is still at least one matching pos-
1998 nigel 87 sibility. The portion of the string that provided the partial match is
1999 nigel 77 set as the first matching string.
2003 nigel 87 Setting the PCRE_DFA_SHORTEST option causes the matching algorithm to
2004     stop as soon as it has found one match. Because of the way the DFA
2005 nigel 77 algorithm works, this is necessarily the shortest possible match at the
2006     first possible matching point in the subject string.
2010 nigel 87 When pcre_dfa_exec() is called with the PCRE_PARTIAL option, and
2011     returns a partial match, it is possible to call it again, with addi-
2012     tional subject characters, and have it continue with the same match.
2013     The PCRE_DFA_RESTART option requests this action; when it is set, the
2014     workspace and wscount options must reference the same vector as before
2015     because data about the match so far is left in them after a partial
2016     match. There is more discussion of this facility in the pcrepartial
2017 nigel 77 documentation.
2019     Successful returns from pcre_dfa_exec()
2021 nigel 87 When pcre_dfa_exec() succeeds, it may have matched more than one sub-
2022 nigel 77 string in the subject. Note, however, that all the matches from one run
2023 nigel 87 of the function start at the same point in the subject. The shorter
2024     matches are all initial substrings of the longer matches. For example,
2025 nigel 77 if the pattern
2027     <.*>
2029     is matched against the string
2031     This is <something> <something else> <something further> no more
2033     the three matched strings are
2035     <something>
2036     <something> <something else>
2037     <something> <something else> <something further>
2039 nigel 87 On success, the yield of the function is a number greater than zero,
2040     which is the number of matched substrings. The substrings themselves
2041     are returned in ovector. Each string uses two elements; the first is
2042     the offset to the start, and the second is the offset to the end. All
2043 nigel 77 the strings have the same start offset. (Space could have been saved by
2044 nigel 87 giving this only once, but it was decided to retain some compatibility
2045     with the way pcre_exec() returns data, even though the meaning of the
2046 nigel 77 strings is different.)
2048     The strings are returned in reverse order of length; that is, the long-
2049 nigel 87 est matching string is given first. If there were too many matches to
2050     fit into ovector, the yield of the function is zero, and the vector is
2051 nigel 77 filled with the longest matches.
2053     Error returns from pcre_dfa_exec()
2055 nigel 87 The pcre_dfa_exec() function returns a negative number when it fails.
2056     Many of the errors are the same as for pcre_exec(), and these are
2057     described above. There are in addition the following errors that are
2058 nigel 77 specific to pcre_dfa_exec():
2060     PCRE_ERROR_DFA_UITEM (-16)
2062 nigel 87 This return is given if pcre_dfa_exec() encounters an item in the pat-
2063     tern that it does not support, for instance, the use of \C or a back
2064 nigel 77 reference.
2066     PCRE_ERROR_DFA_UCOND (-17)
2068 nigel 87 This return is given if pcre_dfa_exec() encounters a condition item in
2069     a pattern that uses a back reference for the condition. This is not
2070 nigel 77 supported.
2074 nigel 87 This return is given if pcre_dfa_exec() is called with an extra block
2075 nigel 77 that contains a setting of the match_limit field. This is not supported
2076     (it is meaningless).
2078     PCRE_ERROR_DFA_WSSIZE (-19)
2080 nigel 87 This return is given if pcre_dfa_exec() runs out of space in the
2081 nigel 77 workspace vector.
2085 nigel 87 When a recursive subpattern is processed, the matching function calls
2086     itself recursively, using private vectors for ovector and workspace.
2087     This error is given if the output vector is not large enough. This
2088 nigel 77 should be extremely rare, as a vector of size 1000 is used.
2090 nigel 87 Last updated: 18 January 2006
2091     Copyright (c) 1997-2006 University of Cambridge.
2092 nigel 79 ------------------------------------------------------------------------------
2093 nigel 63
2096 nigel 63
2097 nigel 79
2098 nigel 73 NAME
2099     PCRE - Perl-compatible regular expressions
2101 nigel 77
2102 nigel 63 PCRE CALLOUTS
2104 nigel 73 int (*pcre_callout)(pcre_callout_block *);
2105 nigel 63
2106 nigel 73 PCRE provides a feature called "callout", which is a means of temporar-
2107     ily passing control to the caller of PCRE in the middle of pattern
2108     matching. The caller of PCRE provides an external function by putting
2109     its entry point in the global variable pcre_callout. By default, this
2110     variable contains NULL, which disables all calling out.
2111 nigel 63
2112 nigel 73 Within a regular expression, (?C) indicates the points at which the
2113     external function is to be called. Different callout points can be
2114     identified by putting a number less than 256 after the letter C. The
2115     default value is zero. For example, this pattern has two callout
2116     points:
2117 nigel 63
2118 nigel 75 (?C1)eabc(?C2)def
2119 nigel 63
2120 nigel 75 If the PCRE_AUTO_CALLOUT option bit is set when pcre_compile() is
2121     called, PCRE automatically inserts callouts, all with number 255,
2122     before each item in the pattern. For example, if PCRE_AUTO_CALLOUT is
2123     used with the pattern
2124 nigel 63
2125 nigel 75 A(\d{2}|--)
2127     it is processed as if it were
2129     (?C255)A(?C255)((?C255)\d{2}(?C255)|(?C255)-(?C255)-(?C255))(?C255)
2131     Notice that there is a callout before and after each parenthesis and
2132     alternation bar. Automatic callouts can be used for tracking the
2133     progress of pattern matching. The pcretest command has an option that
2134     sets automatic callouts; when it is used, the output indicates how the
2135     pattern is matched. This is useful information when you are trying to
2136     optimize the performance of a particular pattern.
2141     You should be aware that, because of optimizations in the way PCRE
2142     matches patterns, callouts sometimes do not happen. For example, if the
2143     pattern is
2145     ab(?C4)cd
2147     PCRE knows that any matching string must contain the letter "d". If the
2148     subject string is "abyz", the lack of "d" means that matching doesn't
2149     ever start, and the callout is never reached. However, with "abyd",
2150     though the result is still no match, the callout is obeyed.
2155     During matching, when PCRE reaches a callout point, the external func-
2156 nigel 77 tion defined by pcre_callout is called (if it is set). This applies to
2157     both the pcre_exec() and the pcre_dfa_exec() matching functions. The
2158     only argument to the callout function is a pointer to a pcre_callout
2159     block. This structure contains the following fields:
2160 nigel 75
2161 nigel 73 int version;
2162     int callout_number;
2163     int *offset_vector;
2164     const char *subject;
2165     int subject_length;
2166     int start_match;
2167     int current_position;
2168     int capture_top;
2169     int capture_last;
2170     void *callout_data;
2171 nigel 75 int pattern_position;
2172     int next_item_length;
2173 nigel 63
2174 nigel 77 The version field is an integer containing the version number of the
2175     block format. The initial version was 0; the current version is 1. The
2176     version number will change again in future if additional fields are
2177 nigel 75 added, but the intention is never to remove any of the existing fields.
2178 nigel 63
2179 nigel 77 The callout_number field contains the number of the callout, as com-
2180     piled into the pattern (that is, the number after ?C for manual call-
2181 nigel 75 outs, and 255 for automatically generated callouts).
2182 nigel 63
2183 nigel 77 The offset_vector field is a pointer to the vector of offsets that was
2184     passed by the caller to pcre_exec() or pcre_dfa_exec(). When
2185     pcre_exec() is used, the contents can be inspected in order to extract
2186     substrings that have been matched so far, in the same way as for
2187     extracting substrings after a match has completed. For pcre_dfa_exec()
2188     this field is not useful.
2189 nigel 63
2190 nigel 75 The subject and subject_length fields contain copies of the values that
2191 nigel 73 were passed to pcre_exec().
2192 nigel 63
2193 nigel 77 The start_match field contains the offset within the subject at which
2194     the current match attempt started. If the pattern is not anchored, the
2195 nigel 75 callout function may be called several times from the same point in the
2196     pattern for different starting points in the subject.
2197 nigel 63
2198 nigel 77 The current_position field contains the offset within the subject of
2199 nigel 73 the current match pointer.
2200 nigel 63
2201 nigel 77 When the pcre_exec() function is used, the capture_top field contains
2202     one more than the number of the highest numbered captured substring so
2203     far. If no substrings have been captured, the value of capture_top is
2204     one. This is always the case when pcre_dfa_exec() is used, because it
2205     does not support captured substrings.
2206 nigel 63
2207 nigel 77 The capture_last field contains the number of the most recently cap-
2208     tured substring. If no substrings have been captured, its value is -1.
2209     This is always the case when pcre_dfa_exec() is used.
2210 nigel 63
2211 nigel 77 The callout_data field contains a value that is passed to pcre_exec()
2212     or pcre_dfa_exec() specifically so that it can be passed back in call-
2213     outs. It is passed in the pcre_callout field of the pcre_extra data
2214     structure. If no such data was passed, the value of callout_data in a
2215     pcre_callout block is NULL. There is a description of the pcre_extra
2216 nigel 73 structure in the pcreapi documentation.
2217 nigel 63
2218 nigel 77 The pattern_position field is present from version 1 of the pcre_call-
2219 nigel 75 out structure. It contains the offset to the next item to be matched in
2220     the pattern string.
2221 nigel 63
2222 nigel 77 The next_item_length field is present from version 1 of the pcre_call-
2223 nigel 75 out structure. It contains the length of the next item to be matched in
2224 nigel 77 the pattern string. When the callout immediately precedes an alterna-
2225     tion bar, a closing parenthesis, or the end of the pattern, the length
2226     is zero. When the callout precedes an opening parenthesis, the length
2227 nigel 75 is that of the entire subpattern.
2228 nigel 73
2229 nigel 77 The pattern_position and next_item_length fields are intended to help
2230     in distinguishing between different automatic callouts, which all have
2231 nigel 75 the same callout number. However, they are set for all callouts.
2234 nigel 63 RETURN VALUES
2236 nigel 77 The external callout function returns an integer to PCRE. If the value
2237     is zero, matching proceeds as normal. If the value is greater than
2238     zero, matching fails at the current point, but the testing of other
2239     matching possibilities goes ahead, just as if a lookahead assertion had
2240     failed. If the value is less than zero, the match is abandoned, and
2241     pcre_exec() (or pcre_dfa_exec()) returns the negative value.
2242 nigel 63
2243 nigel 77 Negative values should normally be chosen from the set of
2244 nigel 73 PCRE_ERROR_xxx values. In particular, PCRE_ERROR_NOMATCH forces a stan-
2245 nigel 77 dard "no match" failure. The error number PCRE_ERROR_CALLOUT is
2246     reserved for use by callout functions; it will never be used by PCRE
2247 nigel 73 itself.
2248 nigel 63
2249 nigel 77 Last updated: 28 February 2005
2250     Copyright (c) 1997-2005 University of Cambridge.
2251 nigel 79 ------------------------------------------------------------------------------
2252 nigel 63
2254 nigel 79 PCRECOMPAT(3) PCRECOMPAT(3)
2255 nigel 63
2256 nigel 79
2257 nigel 73 NAME
2258     PCRE - Perl-compatible regular expressions
2260 nigel 77
2262 nigel 41
2263 nigel 73 This document describes the differences in the ways that PCRE and Perl
2264     handle regular expressions. The differences described here are with
2265     respect to Perl 5.8.
2266 nigel 41
2267 nigel 87 1. PCRE has only a subset of Perl's UTF-8 and Unicode support. Details
2268     of what it does have are given in the section on UTF-8 support in the
2269     main pcre page.
2270 nigel 41
2271 nigel 73 2. PCRE does not allow repeat quantifiers on lookahead assertions. Perl
2272 nigel 87 permits them, but they do not mean what you might think. For example,
2273 nigel 73 (?!a){3} does not assert that the next three characters are not "a". It
2274     just asserts that the next character is not "a" three times.
2275 nigel 41
2276 nigel 87 3. Capturing subpatterns that occur inside negative lookahead asser-
2277     tions are counted, but their entries in the offsets vector are never
2278     set. Perl sets its numerical variables from any such patterns that are
2279 nigel 73 matched before the assertion fails to match something (thereby succeed-
2280 nigel 87 ing), but only if the negative lookahead assertion contains just one
2281 nigel 73 branch.
2282 nigel 41
2283 nigel 87 4. Though binary zero characters are supported in the subject string,
2284 nigel 73 they are not allowed in a pattern string because it is passed as a nor-
2285 nigel 75 mal C string, terminated by zero. The escape sequence \0 can be used in
2286     the pattern to represent a binary zero.
2287 nigel 41
2288 nigel 87 5. The following Perl escape sequences are not supported: \l, \u, \L,
2289 nigel 75 \U, and \N. In fact these are implemented by Perl's general string-han-
2290 nigel 87 dling and are not part of its pattern matching engine. If any of these
2291 nigel 75 are encountered by PCRE, an error is generated.
2292 nigel 41
2293 nigel 87 6. The Perl escape sequences \p, \P, and \X are supported only if PCRE
2294     is built with Unicode character property support. The properties that
2295     can be tested with \p and \P are limited to the general category prop-
2296     erties such as Lu and Nd, script names such as Greek or Han, and the
2297     derived properties Any and L&.
2298 nigel 75
2299     7. PCRE does support the \Q...\E escape for quoting substrings. Charac-
2300     ters in between are treated as literals. This is slightly different
2301     from Perl in that $ and @ are also handled as literals inside the
2302     quotes. In Perl, they cause variable interpolation (but of course PCRE
2303 nigel 73 does not have variables). Note the following examples:
2304 nigel 49
2305 nigel 73 Pattern PCRE matches Perl matches
2306 nigel 41
2307 nigel 73 \Qabc$xyz\E abc$xyz abc followed by the
2308     contents of $xyz
2309     \Qabc\$xyz\E abc\$xyz abc\$xyz
2310     \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
2311 nigel 41
2312 nigel 75 The \Q...\E sequence is recognized both inside and outside character
2313 nigel 73 classes.
2314 nigel 41
2315 nigel 75 8. Fairly obviously, PCRE does not support the (?{code}) and (?p{code})
2316     constructions. However, there is support for recursive patterns using
2317     the non-Perl items (?R), (?number), and (?P>name). Also, the PCRE
2318     "callout" feature allows an external function to be called during pat-
2319     tern matching. See the pcrecallout documentation for details.
2320 nigel 63
2321 nigel 75 9. There are some differences that are concerned with the settings of
2322     captured strings when part of a pattern is repeated. For example,
2323     matching "aba" against the pattern /^(a(b)?)+$/ in Perl leaves $2
2324 nigel 73 unset, but in PCRE it is set to "b".
2325 nigel 41
2326 nigel 75 10. PCRE provides some extensions to the Perl regular expression facil-
2327     ities:
2328 nigel 41
2329 nigel 75 (a) Although lookbehind assertions must match fixed length strings,
2330 nigel 73 each alternative branch of a lookbehind assertion can match a different
2331     length of string. Perl requires them all to have the same length.
2332 nigel 41
2333 nigel 75 (b) If PCRE_DOLLAR_ENDONLY is set and PCRE_MULTILINE is not set, the $
2334 nigel 73 meta-character matches only at the very end of the string.
2335 nigel 41
2336 nigel 73 (c) If PCRE_EXTRA is set, a backslash followed by a letter with no spe-
2337     cial meaning is faulted.
2338 nigel 41
2339 nigel 75 (d) If PCRE_UNGREEDY is set, the greediness of the repetition quanti-
2340 nigel 73 fiers is inverted, that is, by default they are not greedy, but if fol-
2341     lowed by a question mark they are.
2342 nigel 41
2343 nigel 75 (e) PCRE_ANCHORED can be used at matching time to force a pattern to be
2344     tried only at the first matching position in the subject string.
2345 nigel 41
2347 nigel 73 TURE options for pcre_exec() have no Perl equivalents.
2348 nigel 41
2349 nigel 75 (g) The (?R), (?number), and (?P>name) constructs allows for recursive
2350     pattern matching (Perl can do this using the (?p{code}) construct,
2351 nigel 73 which PCRE cannot support.)
2352 nigel 41
2353 nigel 75 (h) PCRE supports named capturing substrings, using the Python syntax.
2354 nigel 43
2355 nigel 75 (i) PCRE supports the possessive quantifier "++" syntax, taken from
2356 nigel 73 Sun's Java package.
2357 nigel 63
2358 nigel 73 (j) The (R) condition, for testing recursion, is a PCRE extension.
2359 nigel 63
2360 nigel 73 (k) The callout facility is PCRE-specific.
2362 nigel 75 (l) The partial matching facility is PCRE-specific.
2364     (m) Patterns compiled by PCRE can be saved and re-used at a later time,
2365     even on different hosts that have the other endianness.
2367 nigel 77 (n) The alternative matching function (pcre_dfa_exec()) matches in a
2368     different way and is not Perl-compatible.
2370 nigel 87 Last updated: 24 January 2006
2371     Copyright (c) 1997-2006 University of Cambridge.
2372 nigel 79 ------------------------------------------------------------------------------
2373 nigel 63
2376 nigel 63
2377 nigel 79
2378 nigel 73 NAME
2379     PCRE - Perl-compatible regular expressions
2381 nigel 77
2384 nigel 73 The syntax and semantics of the regular expressions supported by PCRE
2385     are described below. Regular expressions are also described in the Perl
2386 nigel 75 documentation and in a number of books, some of which have copious
2387     examples. Jeffrey Friedl's "Mastering Regular Expressions", published
2388     by O'Reilly, covers regular expressions in great detail. This descrip-
2389     tion of PCRE's regular expressions is intended as reference material.
2390 nigel 49
2391 nigel 75 The original operation of PCRE was on strings of one-byte characters.
2392     However, there is now also support for UTF-8 character strings. To use
2393     this, you must build PCRE to include UTF-8 support, and then call
2394     pcre_compile() with the PCRE_UTF8 option. How this affects pattern
2395     matching is mentioned in several places below. There is also a summary
2396     of UTF-8 features in the section on UTF-8 support in the main pcre
2397     page.
2398 nigel 41
2399 nigel 77 The remainder of this document discusses the patterns that are sup-
2400     ported by PCRE when its main matching function, pcre_exec(), is used.
2401     From release 6.0, PCRE offers a second matching function,
2402     pcre_dfa_exec(), which matches using a different algorithm that is not
2403     Perl-compatible. The advantages and disadvantages of the alternative
2404     function, and how it differs from the normal function, are discussed in
2405     the pcrematching page.
2407 nigel 75 A regular expression is a pattern that is matched against a subject
2408     string from left to right. Most characters stand for themselves in a
2409     pattern, and match the corresponding characters in the subject. As a
2410 nigel 73 trivial example, the pattern
2411 nigel 41
2412 nigel 73 The quick brown fox
2413 nigel 41
2414 nigel 77 matches a portion of a subject string that is identical to itself. When
2415     caseless matching is specified (the PCRE_CASELESS option), letters are
2416     matched independently of case. In UTF-8 mode, PCRE always understands
2417     the concept of case for characters whose values are less than 128, so
2418     caseless matching is always possible. For characters with higher val-
2419     ues, the concept of case is supported if PCRE is compiled with Unicode
2420     property support, but not otherwise. If you want to use caseless
2421     matching for characters 128 and above, you must ensure that PCRE is
2422     compiled with Unicode property support as well as with UTF-8 support.
2423 nigel 41
2424 nigel 77 The power of regular expressions comes from the ability to include
2425     alternatives and repetitions in the pattern. These are encoded in the
2426     pattern by the use of metacharacters, which do not stand for themselves
2427     but instead are interpreted in some special way.
2429     There are two different sets of metacharacters: those that are recog-
2430     nized anywhere in the pattern except within square brackets, and those
2431     that are recognized in square brackets. Outside square brackets, the
2432 nigel 75 metacharacters are as follows:
2433 nigel 41
2434 nigel 73 \ general escape character with several uses
2435     ^ assert start of string (or line, in multiline mode)
2436     $ assert end of string (or line, in multiline mode)
2437     . match any character except newline (by default)
2438     [ start character class definition
2439     | start of alternative branch
2440     ( start subpattern
2441     ) end subpattern
2442     ? extends the meaning of (
2443     also 0 or 1 quantifier
2444     also quantifier minimizer
2445     * 0 or more quantifier
2446     + 1 or more quantifier
2447     also "possessive quantifier"
2448     { start min/max quantifier
2449 nigel 41
2450 nigel 77 Part of a pattern that is in square brackets is called a "character
2451 nigel 75 class". In a character class the only metacharacters are:
2452 nigel 41
2453 nigel 73 \ general escape character
2454     ^ negate the class, but only if the first character
2455     - indicates character range
2456     [ POSIX character class (only if followed by POSIX
2457     syntax)
2458     ] terminates the character class
2459 nigel 41
2460 nigel 77 The following sections describe the use of each of the metacharacters.
2461 nigel 41
2463 nigel 63 BACKSLASH
2464 nigel 41
2465 nigel 73 The backslash character has several uses. Firstly, if it is followed by
2466 nigel 77 a non-alphanumeric character, it takes away any special meaning that
2467     character may have. This use of backslash as an escape character
2468 nigel 73 applies both inside and outside character classes.
2469 nigel 41
2470 nigel 77 For example, if you want to match a * character, you write \* in the
2471     pattern. This escaping action applies whether or not the following
2472     character would otherwise be interpreted as a metacharacter, so it is
2473     always safe to precede a non-alphanumeric with backslash to specify
2474     that it stands for itself. In particular, if you want to match a back-
2475 nigel 75 slash, you write \\.
2476 nigel 41
2477 nigel 77 If a pattern is compiled with the PCRE_EXTENDED option, whitespace in
2478     the pattern (other than in a character class) and characters between a
2479 nigel 73 # outside a character class and the next newline character are ignored.
2480 nigel 77 An escaping backslash can be used to include a whitespace or # charac-
2481 nigel 73 ter as part of the pattern.
2482 nigel 41
2483 nigel 77 If you want to remove the special meaning from a sequence of charac-
2484     ters, you can do so by putting them between \Q and \E. This is differ-
2485     ent from Perl in that $ and @ are handled as literals in \Q...\E
2486     sequences in PCRE, whereas in Perl, $ and @ cause variable interpola-
2487 nigel 73 tion. Note the following examples:
2488 nigel 63
2489 nigel 73 Pattern PCRE matches Perl matches
2490 nigel 63
2491 nigel 73 \Qabc$xyz\E abc$xyz abc followed by the
2492     contents of $xyz
2493     \Qabc\$xyz\E abc\$xyz abc\$xyz
2494     \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
2495 nigel 63
2496 nigel 77 The \Q...\E sequence is recognized both inside and outside character
2497 nigel 73 classes.
2498 nigel 63
2499 nigel 75 Non-printing characters
2501 nigel 73 A second use of backslash provides a way of encoding non-printing char-
2502 nigel 77 acters in patterns in a visible manner. There is no restriction on the
2503     appearance of non-printing characters, apart from the binary zero that
2504     terminates a pattern, but when a pattern is being prepared by text
2505     editing, it is usually easier to use one of the following escape
2506 nigel 73 sequences than the binary character it represents:
2507 nigel 63
2508 nigel 73 \a alarm, that is, the BEL character (hex 07)
2509     \cx "control-x", where x is any character
2510     \e escape (hex 1B)
2511     \f formfeed (hex 0C)
2512     \n newline (hex 0A)
2513     \r carriage return (hex 0D)
2514     \t tab (hex 09)
2515     \ddd character with octal code ddd, or backreference
2516     \xhh character with hex code hh
2517 nigel 87 \x{hhh..} character with hex code hhh..
2518 nigel 41
2519 nigel 77 The precise effect of \cx is as follows: if x is a lower case letter,
2520     it is converted to upper case. Then bit 6 of the character (hex 40) is
2521     inverted. Thus \cz becomes hex 1A, but \c{ becomes hex 3B, while \c;
2522 nigel 73 becomes hex 7B.
2523 nigel 41
2524 nigel 77 After \x, from zero to two hexadecimal digits are read (letters can be
2525 nigel 87 in upper or lower case). Any number of hexadecimal digits may appear
2526     between \x{ and }, but the value of the character code must be less
2527     than 256 in non-UTF-8 mode, and less than 2**31 in UTF-8 mode (that is,
2528     the maximum hexadecimal value is 7FFFFFFF). If characters other than
2529     hexadecimal digits appear between \x{ and }, or if there is no termi-
2530     nating }, this form of escape is not recognized. Instead, the initial
2531     \x will be interpreted as a basic hexadecimal escape, with no following
2532     digits, giving a character whose value is zero.
2533 nigel 41
2534 nigel 73 Characters whose value is less than 256 can be defined by either of the
2535 nigel 87 two syntaxes for \x. There is no difference in the way they are han-
2536     dled. For example, \xdc is exactly the same as \x{dc}.
2537 nigel 41
2538 nigel 87 After \0 up to two further octal digits are read. In both cases, if
2539     there are fewer than two digits, just those that are present are used.
2540     Thus the sequence \0\x\07 specifies two binary zeros followed by a BEL
2541     character (code value 7). Make sure you supply two digits after the
2542     initial zero if the pattern character that follows is itself an octal
2543 nigel 75 digit.
2544 nigel 63
2545 nigel 73 The handling of a backslash followed by a digit other than 0 is compli-
2546     cated. Outside a character class, PCRE reads it and any following dig-
2547 nigel 87 its as a decimal number. If the number is less than 10, or if there
2548 nigel 73 have been at least that many previous capturing left parentheses in the
2549 nigel 87 expression, the entire sequence is taken as a back reference. A
2550     description of how this works is given later, following the discussion
2551 nigel 73 of parenthesized subpatterns.
2552 nigel 41
2553 nigel 87 Inside a character class, or if the decimal number is greater than 9
2554     and there have not been that many capturing subpatterns, PCRE re-reads
2555     up to three octal digits following the backslash, and generates a sin-
2556 nigel 73 gle byte from the least significant 8 bits of the value. Any subsequent
2557     digits stand for themselves. For example:
2558 nigel 41
2559 nigel 73 \040 is another way of writing a space
2560     \40 is the same, provided there are fewer than 40
2561     previous capturing subpatterns
2562     \7 is always a back reference
2563     \11 might be a back reference, or another way of
2564     writing a tab
2565     \011 is always a tab
2566     \0113 is a tab followed by the character "3"
2567     \113 might be a back reference, otherwise the
2568     character with octal code 113
2569     \377 might be a back reference, otherwise
2570     the byte consisting entirely of 1 bits
2571     \81 is either a back reference, or a binary zero
2572     followed by the two characters "8" and "1"
2573 nigel 41
2574 nigel 87 Note that octal values of 100 or greater must not be introduced by a
2575 nigel 73 leading zero, because no more than three octal digits are ever read.
2576 nigel 41
2577 nigel 87 All the sequences that define a single byte value or a single UTF-8
2578 nigel 73 character (in UTF-8 mode) can be used both inside and outside character
2579 nigel 87 classes. In addition, inside a character class, the sequence \b is
2580 nigel 75 interpreted as the backspace character (hex 08), and the sequence \X is
2581 nigel 87 interpreted as the character "X". Outside a character class, these
2582 nigel 75 sequences have different meanings (see below).
2583 nigel 43
2584 nigel 75 Generic character types
2585 nigel 41
2586 nigel 87 The third use of backslash is for specifying generic character types.
2587 nigel 75 The following are always recognized:
2589 nigel 73 \d any decimal digit
2590     \D any character that is not a decimal digit
2591     \s any whitespace character
2592     \S any character that is not a whitespace character
2593     \w any "word" character
2594     \W any "non-word" character
2595 nigel 41
2596 nigel 73 Each pair of escape sequences partitions the complete set of characters
2597 nigel 87 into two disjoint sets. Any given character matches one, and only one,
2598 nigel 73 of each pair.
2599 nigel 41
2600 nigel 75 These character type sequences can appear both inside and outside char-
2601 nigel 87 acter classes. They each match one character of the appropriate type.
2602     If the current matching point is at the end of the subject string, all
2603 nigel 75 of them fail, since there is no character to match.
2604 nigel 41
2605 nigel 87 For compatibility with Perl, \s does not match the VT character (code
2606     11). This makes it different from the the POSIX "space" class. The \s
2607 nigel 73 characters are HT (9), LF (10), FF (12), CR (13), and space (32).
2608 nigel 63
2609 nigel 75 A "word" character is an underscore or any character less than 256 that
2610 nigel 87 is a letter or digit. The definition of letters and digits is con-
2611     trolled by PCRE's low-valued character tables, and may vary if locale-
2612     specific matching is taking place (see "Locale support" in the pcreapi
2613     page). For example, in the "fr_FR" (French) locale, some character
2614     codes greater than 128 are used for accented letters, and these are
2615 nigel 75 matched by \w.
2616 nigel 63
2617 nigel 87 In UTF-8 mode, characters with values greater than 128 never match \d,
2618 nigel 75 \s, or \w, and always match \D, \S, and \W. This is true even when Uni-
2619 nigel 87 code character property support is available. The use of locales with
2620     Unicode is discouraged.
2621 nigel 41
2622 nigel 75 Unicode character properties
2624     When PCRE is built with Unicode character property support, three addi-
2625 nigel 87 tional escape sequences to match character properties are available
2626 nigel 75 when UTF-8 mode is selected. They are:
2628 nigel 87 \p{xx} a character with the xx property
2629     \P{xx} a character without the xx property
2630     \X an extended Unicode sequence
2631 nigel 75
2632 nigel 77 The property names represented by xx above are limited to the Unicode
2633 nigel 87 script names, the general category properties, and "Any", which matches
2634     any character (including newline). Other properties such as "InMusical-
2635     Symbols" are not currently supported by PCRE. Note that \P{Any} does
2636     not match any characters, so always causes a match failure.
2637 nigel 75
2638 nigel 87 Sets of Unicode characters are defined as belonging to certain scripts.
2639     A character from one of these sets can be matched using a script name.
2640     For example:
2641 nigel 75
2642 nigel 87 \p{Greek}
2643     \P{Han}
2645     Those that are not part of an identified script are lumped together as
2646     "Common". The current list of scripts is:
2648     Arabic, Armenian, Bengali, Bopomofo, Braille, Buginese, Buhid, Cana-
2649     dian_Aboriginal, Cherokee, Common, Coptic, Cypriot, Cyrillic, Deseret,
2650     Devanagari, Ethiopic, Georgian, Glagolitic, Gothic, Greek, Gujarati,
2651     Gurmukhi, Han, Hangul, Hanunoo, Hebrew, Hiragana, Inherited, Kannada,
2652     Katakana, Kharoshthi, Khmer, Lao, Latin, Limbu, Linear_B, Malayalam,
2653     Mongolian, Myanmar, New_Tai_Lue, Ogham, Old_Italic, Old_Persian, Oriya,
2654     Osmanya, Runic, Shavian, Sinhala, Syloti_Nagri, Syriac, Tagalog, Tag-
2655     banwa, Tai_Le, Tamil, Telugu, Thaana, Thai, Tibetan, Tifinagh,
2656     Ugaritic, Yi.
2658     Each character has exactly one general category property, specified by
2659     a two-letter abbreviation. For compatibility with Perl, negation can be
2660     specified by including a circumflex between the opening brace and the
2661     property name. For example, \p{^Lu} is the same as \P{Lu}.
2663     If only one letter is specified with \p or \P, it includes all the gen-
2664     eral category properties that start with that letter. In this case, in
2665     the absence of negation, the curly brackets in the escape sequence are
2666     optional; these two examples have the same effect:
2668 nigel 75 \p{L}
2669     \pL
2671 nigel 87 The following general category property codes are supported:
2672 nigel 75
2673     C Other
2674     Cc Control
2675     Cf Format
2676     Cn Unassigned
2677     Co Private use
2678     Cs Surrogate
2680     L Letter
2681     Ll Lower case letter
2682     Lm Modifier letter
2683     Lo Other letter
2684     Lt Title case letter
2685     Lu Upper case letter
2687     M Mark
2688     Mc Spacing mark
2689     Me Enclosing mark
2690     Mn Non-spacing mark
2692     N Number
2693     Nd Decimal number
2694     Nl Letter number
2695     No Other number
2697     P Punctuation
2698     Pc Connector punctuation
2699     Pd Dash punctuation
2700     Pe Close punctuation
2701     Pf Final punctuation
2702     Pi Initial punctuation
2703     Po Other punctuation
2704     Ps Open punctuation
2706     S Symbol
2707     Sc Currency symbol
2708     Sk Modifier symbol
2709     Sm Mathematical symbol
2710     So Other symbol
2712     Z Separator
2713     Zl Line separator
2714     Zp Paragraph separator
2715     Zs Space separator
2717 nigel 87 The special property L& is also supported: it matches a character that
2718     has the Lu, Ll, or Lt property, in other words, a letter that is not
2719     classified as a modifier or "other".
2720 nigel 75
2721 nigel 87 The long synonyms for these properties that Perl supports (such as
2722     \p{Letter}) are not supported by PCRE. Nor is is permitted to prefix
2723     any of these properties with "Is".
2725     No character that is in the Unicode table has the Cn (unassigned) prop-
2726     erty. Instead, this property is assumed for any code point that is not
2727     in the Unicode table.
2729 nigel 77 Specifying caseless matching does not affect these escape sequences.
2730 nigel 75 For example, \p{Lu} always matches only upper case letters.
2732 nigel 77 The \X escape matches any number of Unicode characters that form an
2733 nigel 75 extended Unicode sequence. \X is equivalent to
2735     (?>\PM\pM*)
2737 nigel 77 That is, it matches a character without the "mark" property, followed
2738     by zero or more characters with the "mark" property, and treats the
2739     sequence as an atomic group (see below). Characters with the "mark"
2740 nigel 75 property are typically accents that affect the preceding character.
2742 nigel 77 Matching characters by Unicode property is not fast, because PCRE has
2743     to search a structure that contains data for over fifteen thousand
2744 nigel 75 characters. That is why the traditional escape sequences such as \d and
2745     \w do not use Unicode properties in PCRE.
2747     Simple assertions
2749 nigel 73 The fourth use of backslash is for certain simple assertions. An asser-
2750 nigel 77 tion specifies a condition that has to be met at a particular point in
2751     a match, without consuming any characters from the subject string. The
2752     use of subpatterns for more complicated assertions is described below.
2753 nigel 75 The backslashed assertions are:
2754 nigel 41
2755 nigel 73 \b matches at a word boundary
2756     \B matches when not at a word boundary
2757     \A matches at start of subject
2758     \Z matches at end of subject or before newline at end
2759     \z matches at end of subject
2760     \G matches at first matching position in subject
2761 nigel 41
2762 nigel 77 These assertions may not appear in character classes (but note that \b
2763 nigel 73 has a different meaning, namely the backspace character, inside a char-
2764     acter class).
2765 nigel 41
2766 nigel 77 A word boundary is a position in the subject string where the current
2767     character and the previous character do not both match \w or \W (i.e.
2768     one matches \w and the other matches \W), or the start or end of the
2769 nigel 73 string if the first or last character matches \w, respectively.
2770 nigel 43
2771 nigel 77 The \A, \Z, and \z assertions differ from the traditional circumflex
2772 nigel 75 and dollar (described in the next section) in that they only ever match
2773 nigel 77 at the very start and end of the subject string, whatever options are
2774     set. Thus, they are independent of multiline mode. These three asser-
2775 nigel 75 tions are not affected by the PCRE_NOTBOL or PCRE_NOTEOL options, which
2776 nigel 77 affect only the behaviour of the circumflex and dollar metacharacters.
2777     However, if the startoffset argument of pcre_exec() is non-zero, indi-
2778 nigel 75 cating that matching is to start at a point other than the beginning of
2779 nigel 77 the subject, \A can never match. The difference between \Z and \z is
2780     that \Z matches before a newline that is the last character of the
2781     string as well as at the end of the string, whereas \z matches only at
2782 nigel 75 the end.
2783 nigel 63
2784 nigel 77 The \G assertion is true only when the current matching position is at
2785     the start point of the match, as specified by the startoffset argument
2786     of pcre_exec(). It differs from \A when the value of startoffset is
2787     non-zero. By calling pcre_exec() multiple times with appropriate argu-
2788 nigel 73 ments, you can mimic Perl's /g option, and it is in this kind of imple-
2789     mentation where \G can be useful.
2790 nigel 41
2791 nigel 77 Note, however, that PCRE's interpretation of \G, as the start of the
2792 nigel 73 current match, is subtly different from Perl's, which defines it as the
2793 nigel 77 end of the previous match. In Perl, these can be different when the
2794     previously matched string was empty. Because PCRE does just one match
2795 nigel 73 at a time, it cannot reproduce this behaviour.
2796 nigel 41
2797 nigel 77 If all the alternatives of a pattern begin with \G, the expression is
2798 nigel 73 anchored to the starting match position, and the "anchored" flag is set
2799     in the compiled regular expression.
2800 nigel 63
2803 nigel 63
2804 nigel 73 Outside a character class, in the default matching mode, the circumflex
2805 nigel 77 character is an assertion that is true only if the current matching
2806     point is at the start of the subject string. If the startoffset argu-
2807     ment of pcre_exec() is non-zero, circumflex can never match if the
2808     PCRE_MULTILINE option is unset. Inside a character class, circumflex
2809 nigel 73 has an entirely different meaning (see below).
2810 nigel 41
2811 nigel 77 Circumflex need not be the first character of the pattern if a number
2812     of alternatives are involved, but it should be the first thing in each
2813     alternative in which it appears if the pattern is ever to match that
2814     branch. If all possible alternatives start with a circumflex, that is,
2815     if the pattern is constrained to match only at the start of the sub-
2816     ject, it is said to be an "anchored" pattern. (There are also other
2817 nigel 73 constructs that can cause a pattern to be anchored.)
2818 nigel 41
2819 nigel 77 A dollar character is an assertion that is true only if the current
2820     matching point is at the end of the subject string, or immediately
2821 nigel 73 before a newline character that is the last character in the string (by
2822 nigel 77 default). Dollar need not be the last character of the pattern if a
2823     number of alternatives are involved, but it should be the last item in
2824     any branch in which it appears. Dollar has no special meaning in a
2825 nigel 73 character class.
2826 nigel 41
2827 nigel 77 The meaning of dollar can be changed so that it matches only at the
2828     very end of the string, by setting the PCRE_DOLLAR_ENDONLY option at
2829 nigel 73 compile time. This does not affect the \Z assertion.
2830 nigel 41
2831 nigel 73 The meanings of the circumflex and dollar characters are changed if the
2832     PCRE_MULTILINE option is set. When this is the case, they match immedi-
2833 nigel 77 ately after and immediately before an internal newline character,
2834     respectively, in addition to matching at the start and end of the sub-
2835     ject string. For example, the pattern /^abc$/ matches the subject
2836     string "def\nabc" (where \n represents a newline character) in multi-
2837 nigel 75 line mode, but not otherwise. Consequently, patterns that are anchored
2838 nigel 77 in single line mode because all branches start with ^ are not anchored
2839     in multiline mode, and a match for circumflex is possible when the
2840     startoffset argument of pcre_exec() is non-zero. The PCRE_DOL-
2841 nigel 75 LAR_ENDONLY option is ignored if PCRE_MULTILINE is set.
2842 nigel 41
2843 nigel 77 Note that the sequences \A, \Z, and \z can be used to match the start
2844     and end of the subject in both modes, and if all branches of a pattern
2845     start with \A it is always anchored, whether PCRE_MULTILINE is set or
2846 nigel 73 not.
2847 nigel 41
2849 nigel 63 FULL STOP (PERIOD, DOT)
2850 nigel 41
2851 nigel 73 Outside a character class, a dot in the pattern matches any one charac-
2852 nigel 77 ter in the subject, including a non-printing character, but not (by
2853     default) newline. In UTF-8 mode, a dot matches any UTF-8 character,
2854 nigel 75 which might be more than one byte long, except (by default) newline. If
2855 nigel 77 the PCRE_DOTALL option is set, dots match newlines as well. The han-
2856     dling of dot is entirely independent of the handling of circumflex and
2857     dollar, the only relationship being that they both involve newline
2858 nigel 75 characters. Dot has no special meaning in a character class.
2859 nigel 41
2863 nigel 73 Outside a character class, the escape sequence \C matches any one byte,
2864 nigel 77 both in and out of UTF-8 mode. Unlike a dot, it can match a newline.
2865     The feature is provided in Perl in order to match individual bytes in
2866     UTF-8 mode. Because it breaks up UTF-8 characters into individual
2867     bytes, what remains in the string may be a malformed UTF-8 string. For
2868 nigel 75 this reason, the \C escape sequence is best avoided.
2869 nigel 63
2870 nigel 77 PCRE does not allow \C to appear in lookbehind assertions (described
2871     below), because in UTF-8 mode this would make it impossible to calcu-
2872 nigel 75 late the length of the lookbehind.
2873 nigel 63
2876 nigel 63
2877 nigel 73 An opening square bracket introduces a character class, terminated by a
2878     closing square bracket. A closing square bracket on its own is not spe-
2879     cial. If a closing square bracket is required as a member of the class,
2880 nigel 77 it should be the first data character in the class (after an initial
2881 nigel 73 circumflex, if present) or escaped with a backslash.
2882 nigel 41
2883 nigel 77 A character class matches a single character in the subject. In UTF-8
2884     mode, the character may occupy more than one byte. A matched character
2885 nigel 73 must be in the set of characters defined by the class, unless the first
2886 nigel 77 character in the class definition is a circumflex, in which case the
2887     subject character must not be in the set defined by the class. If a
2888     circumflex is actually required as a member of the class, ensure it is
2889 nigel 73 not the first character, or escape it with a backslash.
2890 nigel 41
2891 nigel 77 For example, the character class [aeiou] matches any lower case vowel,
2892     while [^aeiou] matches any character that is not a lower case vowel.
2893 nigel 73 Note that a circumflex is just a convenient notation for specifying the
2894 nigel 77 characters that are in the class by enumerating those that are not. A
2895     class that starts with a circumflex is not an assertion: it still con-
2896     sumes a character from the subject string, and therefore it fails if
2897 nigel 75 the current pointer is at the end of the string.
2898 nigel 41
2899 nigel 77 In UTF-8 mode, characters with values greater than 255 can be included
2900     in a class as a literal string of bytes, or by using the \x{ escaping
2901 nigel 73 mechanism.
2902 nigel 63
2903 nigel 77 When caseless matching is set, any letters in a class represent both
2904     their upper case and lower case versions, so for example, a caseless
2905     [aeiou] matches "A" as well as "a", and a caseless [^aeiou] does not
2906     match "A", whereas a caseful version would. In UTF-8 mode, PCRE always
2907     understands the concept of case for characters whose values are less
2908     than 128, so caseless matching is always possible. For characters with
2909     higher values, the concept of case is supported if PCRE is compiled
2910     with Unicode property support, but not otherwise. If you want to use
2911     caseless matching for characters 128 and above, you must ensure that
2912     PCRE is compiled with Unicode property support as well as with UTF-8
2913     support.
2914 nigel 41
2915 nigel 75 The newline character is never treated in any special way in character
2916     classes, whatever the setting of the PCRE_DOTALL or PCRE_MULTILINE
2917 nigel 73 options is. A class such as [^a] will always match a newline.
2918 nigel 41
2919 nigel 75 The minus (hyphen) character can be used to specify a range of charac-
2920     ters in a character class. For example, [d-m] matches any letter
2921     between d and m, inclusive. If a minus character is required in a
2922     class, it must be escaped with a backslash or appear in a position
2923     where it cannot be interpreted as indicating a range, typically as the
2924 nigel 73 first or last character in the class.
2925 nigel 41
2926 nigel 73 It is not possible to have the literal character "]" as the end charac-
2927 nigel 75 ter of a range. A pattern such as [W-]46] is interpreted as a class of
2928     two characters ("W" and "-") followed by a literal string "46]", so it
2929     would match "W46]" or "-46]". However, if the "]" is escaped with a
2930     backslash it is interpreted as the end of range, so [W-\]46] is inter-
2931     preted as a class containing a range followed by two other characters.
2932     The octal or hexadecimal representation of "]" can also be used to end
2933     a range.
2934 nigel 41
2935 nigel 75 Ranges operate in the collating sequence of character values. They can
2936     also be used for characters specified numerically, for example
2937     [\000-\037]. In UTF-8 mode, ranges can include characters whose values
2938 nigel 73 are greater than 255, for example [\x{100}-\x{2ff}].
2939 nigel 63
2940 nigel 73 If a range that includes letters is used when caseless matching is set,
2941     it matches the letters in either case. For example, [W-c] is equivalent
2942 nigel 75 to [][\\^_`wxyzabc], matched caselessly, and in non-UTF-8 mode, if
2943     character tables for the "fr_FR" locale are in use, [\xc8-\xcb] matches
2944     accented E characters in both cases. In UTF-8 mode, PCRE supports the
2945     concept of case for characters with values greater than 128 only when
2946     it is compiled with Unicode property support.
2947 nigel 41
2948 nigel 75 The character types \d, \D, \p, \P, \s, \S, \w, and \W may also appear
2949     in a character class, and add the characters that they match to the
2950     class. For example, [\dABCDEF] matches any hexadecimal digit. A circum-
2951     flex can conveniently be used with the upper case character types to
2952     specify a more restricted set of characters than the matching lower
2953     case type. For example, the class [^\W_] matches any letter or digit,
2954     but not underscore.
2955 nigel 41
2956 nigel 75 The only metacharacters that are recognized in character classes are
2957     backslash, hyphen (only where it can be interpreted as specifying a
2958     range), circumflex (only at the start), opening square bracket (only
2959     when it can be interpreted as introducing a POSIX class name - see the
2960     next section), and the terminating closing square bracket. However,
2961     escaping other non-alphanumeric characters does no harm.
2962 nigel 41
2963 nigel 73
2966 nigel 75 Perl supports the POSIX notation for character classes. This uses names
2967     enclosed by [: and :] within the enclosing square brackets. PCRE also
2968     supports this notation. For example,
2969 nigel 63
2970 nigel 73 [01[:alpha:]%]
2971 nigel 43
2972 nigel 73 matches "0", "1", any alphabetic character, or "%". The supported class
2973     names are
2974 nigel 43
2975 nigel 73 alnum letters and digits
2976     alpha letters
2977     ascii character codes 0 - 127
2978     blank space or tab only
2979     cntrl control characters
2980     digit decimal digits (same as \d)
2981     graph printing characters, excluding space
2982     lower lower case letters
2983     print printing characters, including space
2984     punct printing characters, excluding letters and digits
2985     space white space (not quite the same as \s)
2986     upper upper case letters
2987     word "word" characters (same as \w)
2988     xdigit hexadecimal digits
2989 nigel 43
2990 nigel 73 The "space" characters are HT (9), LF (10), VT (11), FF (12), CR (13),
2991     and space (32). Notice that this list includes the VT character (code
2992     11). This makes "space" different to \s, which does not include VT (for
2993     Perl compatibility).
2994 nigel 43
2995 nigel 73 The name "word" is a Perl extension, and "blank" is a GNU extension
2996     from Perl 5.8. Another Perl extension is negation, which is indicated
2997     by a ^ character after the colon. For example,
2998 nigel 63
2999 nigel 73 [12[:^digit:]]
3000 nigel 43
3001 nigel 73 matches "1", "2", or any non-digit. PCRE (and Perl) also recognize the
3002     POSIX syntax [.ch.] and [=ch=] where "ch" is a "collating element", but
3003     these are not supported, and an error is given if they are encountered.
3004 nigel 43
3005 nigel 75 In UTF-8 mode, characters with values greater than 128 do not match any
3006 nigel 73 of the POSIX character classes.
3007 nigel 43
3009 nigel 41 VERTICAL BAR
3010 nigel 63
3011 nigel 73 Vertical bar characters are used to separate alternative patterns. For
3012     example, the pattern
3013 nigel 41
3014 nigel 73 gilbert|sullivan
3015 nigel 41
3016 nigel 73 matches either "gilbert" or "sullivan". Any number of alternatives may
3017     appear, and an empty alternative is permitted (matching the empty
3018     string). The matching process tries each alternative in turn, from
3019     left to right, and the first one that succeeds is used. If the alterna-
3020     tives are within a subpattern (defined below), "succeeds" means match-
3021     ing the rest of the main pattern as well as the alternative in the sub-
3022     pattern.
3023 nigel 41
3027 nigel 73 The settings of the PCRE_CASELESS, PCRE_MULTILINE, PCRE_DOTALL, and
3028     PCRE_EXTENDED options can be changed from within the pattern by a
3029     sequence of Perl option letters enclosed between "(?" and ")". The
3030     option letters are
3031 nigel 63
3032 nigel 73 i for PCRE_CASELESS
3033     m for PCRE_MULTILINE
3034     s for PCRE_DOTALL
3035     x for PCRE_EXTENDED
3036 nigel 41
3037 nigel 73 For example, (?im) sets caseless, multiline matching. It is also possi-
3038     ble to unset these options by preceding the letter with a hyphen, and a
3039     combined setting and unsetting such as (?im-sx), which sets PCRE_CASE-
3040     LESS and PCRE_MULTILINE while unsetting PCRE_DOTALL and PCRE_EXTENDED,
3041     is also permitted. If a letter appears both before and after the
3042     hyphen, the option is unset.
3043 nigel 41
3044 nigel 73 When an option change occurs at top level (that is, not inside subpat-
3045     tern parentheses), the change applies to the remainder of the pattern
3046     that follows. If the change is placed right at the start of a pattern,
3047     PCRE extracts it into the global options (and it will therefore show up
3048     in data extracted by the pcre_fullinfo() function).
3049 nigel 41
3050 nigel 73 An option change within a subpattern affects only that part of the cur-
3051     rent pattern that follows it, so
3052 nigel 41
3053 nigel 73 (a(?i)b)c
3054 nigel 41
3055 nigel 73 matches abc and aBc and no other strings (assuming PCRE_CASELESS is not
3056     used). By this means, options can be made to have different settings
3057     in different parts of the pattern. Any changes made in one alternative
3058     do carry on into subsequent branches within the same subpattern. For
3059     example,
3060 nigel 41
3061 nigel 73 (a(?i)b|c)
3062 nigel 41
3063 nigel 73 matches "ab", "aB", "c", and "C", even though when matching "C" the
3064     first branch is abandoned before the option setting. This is because
3065     the effects of option settings happen at compile time. There would be
3066     some very weird behaviour otherwise.
3067 nigel 41
3068 nigel 73 The PCRE-specific options PCRE_UNGREEDY and PCRE_EXTRA can be changed
3069     in the same way as the Perl-compatible options by using the characters
3070     U and X respectively. The (?X) flag setting is special in that it must
3071     always occur earlier in the pattern than any of the additional features
3072 nigel 75 it turns on, even when it is at top level. It is best to put it at the
3073     start.
3074 nigel 41
3076 nigel 63 SUBPATTERNS
3077 nigel 41
3078 nigel 73 Subpatterns are delimited by parentheses (round brackets), which can be
3079 nigel 75 nested. Turning part of a pattern into a subpattern does two things:
3080 nigel 41
3081 nigel 73 1. It localizes a set of alternatives. For example, the pattern
3082 nigel 41
3083 nigel 73 cat(aract|erpillar|)
3084 nigel 41
3085 nigel 73 matches one of the words "cat", "cataract", or "caterpillar". Without
3086     the parentheses, it would match "cataract", "erpillar" or the empty
3087     string.
3088 nigel 41
3089 nigel 75 2. It sets up the subpattern as a capturing subpattern. This means
3090     that, when the whole pattern matches, that portion of the subject
3091 nigel 73 string that matched the subpattern is passed back to the caller via the
3092     ovector argument of pcre_exec(). Opening parentheses are counted from
3093 nigel 75 left to right (starting from 1) to obtain numbers for the capturing
3094 nigel 73 subpatterns.
3095 nigel 41
3096 nigel 73 For example, if the string "the red king" is matched against the pat-
3097     tern
3098 nigel 41
3099 nigel 73 the ((red|white) (king|queen))
3100 nigel 41
3101 nigel 73 the captured substrings are "red king", "red", and "king", and are num-
3102     bered 1, 2, and 3, respectively.
3103 nigel 41
3104 nigel 73 The fact that plain parentheses fulfil two functions is not always
3105     helpful. There are often times when a grouping subpattern is required
3106     without a capturing requirement. If an opening parenthesis is followed
3107     by a question mark and a colon, the subpattern does not do any captur-
3108     ing, and is not counted when computing the number of any subsequent
3109     capturing subpatterns. For example, if the string "the white queen" is
3110     matched against the pattern
3111 nigel 41
3112 nigel 73 the ((?:red|white) (king|queen))
3113 nigel 41
3114 nigel 73 the captured substrings are "white queen" and "queen", and are numbered
3115     1 and 2. The maximum number of capturing subpatterns is 65535, and the
3116     maximum depth of nesting of all subpatterns, both capturing and non-
3117     capturing, is 200.
3118 nigel 41
3119 nigel 73 As a convenient shorthand, if any option settings are required at the
3120     start of a non-capturing subpattern, the option letters may appear
3121     between the "?" and the ":". Thus the two patterns
3122 nigel 41
3123 nigel 73 (?i:saturday|sunday)
3124     (?:(?i)saturday|sunday)
3125 nigel 41
3126 nigel 73 match exactly the same set of strings. Because alternative branches are
3127     tried from left to right, and options are not reset until the end of
3128     the subpattern is reached, an option setting in one branch does affect
3129     subsequent branches, so the above patterns match "SUNDAY" as well as
3130     "Saturday".
3131 nigel 41
3134 nigel 41
3135 nigel 73 Identifying capturing parentheses by number is simple, but it can be
3136     very hard to keep track of the numbers in complicated regular expres-
3137     sions. Furthermore, if an expression is modified, the numbers may
3138 nigel 75 change. To help with this difficulty, PCRE supports the naming of sub-
3139 nigel 73 patterns, something that Perl does not provide. The Python syntax
3140     (?P<name>...) is used. Names consist of alphanumeric characters and
3141     underscores, and must be unique within a pattern.
3142 nigel 63
3143 nigel 73 Named capturing parentheses are still allocated numbers as well as
3144     names. The PCRE API provides function calls for extracting the name-to-
3145 nigel 75 number translation table from a compiled pattern. There is also a con-
3146     venience function for extracting a captured substring by name. For fur-
3147     ther details see the pcreapi documentation.
3148 nigel 63
3150 nigel 41 REPETITION
3151 nigel 63
3152 nigel 75 Repetition is specified by quantifiers, which can follow any of the
3153 nigel 73 following items:
3154 nigel 41
3155 nigel 73 a literal data character
3156     the . metacharacter
3157     the \C escape sequence
3158 nigel 75 the \X escape sequence (in UTF-8 mode with Unicode properties)
3159     an escape such as \d that matches a single character
3160 nigel 73 a character class
3161     a back reference (see next section)
3162     a parenthesized subpattern (unless it is an assertion)
3163 nigel 41
3164 nigel 75 The general repetition quantifier specifies a minimum and maximum num-
3165     ber of permitted matches, by giving the two numbers in curly brackets
3166     (braces), separated by a comma. The numbers must be less than 65536,
3167 nigel 73 and the first must be less than or equal to the second. For example:
3168 nigel 41
3169 nigel 73 z{2,4}
3170 nigel 41
3171 nigel 75 matches "zz", "zzz", or "zzzz". A closing brace on its own is not a
3172     special character. If the second number is omitted, but the comma is
3173     present, there is no upper limit; if the second number and the comma
3174     are both omitted, the quantifier specifies an exact number of required
3175 nigel 73 matches. Thus
3176 nigel 41
3177 nigel 73 [aeiou]{3,}
3178 nigel 41
3179 nigel 73 matches at least 3 successive vowels, but may match many more, while
3180 nigel 41
3181 nigel 73 \d{8}
3182 nigel 41
3183 nigel 75 matches exactly 8 digits. An opening curly bracket that appears in a
3184     position where a quantifier is not allowed, or one that does not match
3185     the syntax of a quantifier, is taken as a literal character. For exam-
3186 nigel 73 ple, {,6} is not a quantifier, but a literal string of four characters.
3187 nigel 63
3188 nigel 75 In UTF-8 mode, quantifiers apply to UTF-8 characters rather than to
3189 nigel 73 individual bytes. Thus, for example, \x{100}{2} matches two UTF-8 char-
3190 nigel 75 acters, each of which is represented by a two-byte sequence. Similarly,
3191     when Unicode property support is available, \X{3} matches three Unicode
3192     extended sequences, each of which may be several bytes long (and they
3193     may be of different lengths).
3194 nigel 63
3195 nigel 73 The quantifier {0} is permitted, causing the expression to behave as if
3196     the previous item and the quantifier were not present.
3197 nigel 41
3198 nigel 73 For convenience (and historical compatibility) the three most common
3199     quantifiers have single-character abbreviations:
3200 nigel 41
3201 nigel 73 * is equivalent to {0,}
3202     + is equivalent to {1,}
3203     ? is equivalent to {0,1}
3204 nigel 41
3205 nigel 73 It is possible to construct infinite loops by following a subpattern
3206     that can match no characters with a quantifier that has no upper limit,
3207     for example:
3208 nigel 41
3209 nigel 73 (a?)*
3210 nigel 41
3211 nigel 73 Earlier versions of Perl and PCRE used to give an error at compile time
3212     for such patterns. However, because there are cases where this can be
3213     useful, such patterns are now accepted, but if any repetition of the
3214     subpattern does in fact match no characters, the loop is forcibly bro-
3215     ken.
3216 nigel 41
3217 nigel 73 By default, the quantifiers are "greedy", that is, they match as much
3218     as possible (up to the maximum number of permitted times), without
3219     causing the rest of the pattern to fail. The classic example of where
3220     this gives problems is in trying to match comments in C programs. These
3221 nigel 75 appear between /* and */ and within the comment, individual * and /
3222     characters may appear. An attempt to match C comments by applying the
3223     pattern
3224 nigel 41
3225 nigel 73 /\*.*\*/
3226 nigel 41
3227 nigel 73 to the string
3228 nigel 41
3229 nigel 75 /* first comment */ not comment /* second comment */
3230 nigel 41
3231 nigel 73 fails, because it matches the entire string owing to the greediness of
3232     the .* item.
3233 nigel 41
3234 nigel 73 However, if a quantifier is followed by a question mark, it ceases to
3235     be greedy, and instead matches the minimum number of times possible, so
3236     the pattern
3237 nigel 41
3238 nigel 73 /\*.*?\*/
3239 nigel 41
3240 nigel 73 does the right thing with the C comments. The meaning of the various
3241     quantifiers is not otherwise changed, just the preferred number of
3242     matches. Do not confuse this use of question mark with its use as a
3243     quantifier in its own right. Because it has two uses, it can sometimes
3244     appear doubled, as in
3245 nigel 41
3246 nigel 73 \d??\d
3247 nigel 41
3248 nigel 73 which matches one digit by preference, but can match two if that is the
3249     only way the rest of the pattern matches.
3250 nigel 41
3251 nigel 73 If the PCRE_UNGREEDY option is set (an option which is not available in
3252     Perl), the quantifiers are not greedy by default, but individual ones
3253     can be made greedy by following them with a question mark. In other
3254     words, it inverts the default behaviour.
3255 nigel 41
3256 nigel 73 When a parenthesized subpattern is quantified with a minimum repeat
3257 nigel 75 count that is greater than 1 or with a limited maximum, more memory is
3258 nigel 73 required for the compiled pattern, in proportion to the size of the
3259     minimum or maximum.
3260 nigel 41
3261 nigel 73 If a pattern starts with .* or .{0,} and the PCRE_DOTALL option (equiv-
3262     alent to Perl's /s) is set, thus allowing the . to match newlines, the
3263     pattern is implicitly anchored, because whatever follows will be tried
3264     against every character position in the subject string, so there is no
3265     point in retrying the overall match at any position after the first.
3266     PCRE normally treats such a pattern as though it were preceded by \A.
3267 nigel 63
3268 nigel 73 In cases where it is known that the subject string contains no new-
3269     lines, it is worth setting PCRE_DOTALL in order to obtain this opti-
3270     mization, or alternatively using ^ to indicate anchoring explicitly.
3271 nigel 63
3272 nigel 73 However, there is one situation where the optimization cannot be used.
3273     When .* is inside capturing parentheses that are the subject of a
3274     backreference elsewhere in the pattern, a match at the start may fail,
3275     and a later one succeed. Consider, for example:
3276 nigel 63
3277 nigel 73 (.*)abc\1
3278 nigel 63
3279 nigel 73 If the subject is "xyz123abc123" the match point is the fourth charac-
3280     ter. For this reason, such a pattern is not implicitly anchored.
3281 nigel 41
3282 nigel 73 When a capturing subpattern is repeated, the value captured is the sub-
3283     string that matched the final iteration. For example, after
3284 nigel 41
3285 nigel 73 (tweedle[dume]{3}\s*)+
3286 nigel 41
3287 nigel 73 has matched "tweedledum tweedledee" the value of the captured substring
3288     is "tweedledee". However, if there are nested capturing subpatterns,
3289     the corresponding captured values may have been set in previous itera-
3290     tions. For example, after
3291 nigel 41
3292 nigel 73 /(a|(b))+/
3293 nigel 41
3294 nigel 73 matches "aba" the value of the second captured substring is "b".
3295 nigel 41
3296 nigel 73
3298 nigel 41
3299 nigel 73 With both maximizing and minimizing repetition, failure of what follows
3300     normally causes the repeated item to be re-evaluated to see if a dif-
3301     ferent number of repeats allows the rest of the pattern to match. Some-
3302     times it is useful to prevent this, either to change the nature of the
3303     match, or to cause it fail earlier than it otherwise might, when the
3304     author of the pattern knows there is no point in carrying on.
3305 nigel 53
3306 nigel 73 Consider, for example, the pattern \d+foo when applied to the subject
3307     line
3308 nigel 53
3309 nigel 73 123456bar
3310 nigel 53
3311 nigel 73 After matching all 6 digits and then failing to match "foo", the normal
3312     action of the matcher is to try again with only 5 digits matching the
3313     \d+ item, and then with 4, and so on, before ultimately failing.
3314     "Atomic grouping" (a term taken from Jeffrey Friedl's book) provides
3315     the means for specifying that once a subpattern has matched, it is not
3316     to be re-evaluated in this way.
3317 nigel 53
3318 nigel 73 If we use atomic grouping for the previous example, the matcher would
3319     give up immediately on failing to match "foo" the first time. The nota-
3320     tion is a kind of special parenthesis, starting with (?> as in this
3321     example:
3322 nigel 53
3323 nigel 73 (?>\d+)foo
3324 nigel 53
3325 nigel 73 This kind of parenthesis "locks up" the part of the pattern it con-
3326     tains once it has matched, and a failure further into the pattern is
3327     prevented from backtracking into it. Backtracking past it to previous
3328     items, however, works as normal.
3329 nigel 53
3330 nigel 73 An alternative description is that a subpattern of this type matches
3331     the string of characters that an identical standalone pattern would
3332     match, if anchored at the current point in the subject string.
3333 nigel 63
3334 nigel 73 Atomic grouping subpatterns are not capturing subpatterns. Simple cases
3335     such as the above example can be thought of as a maximizing repeat that
3336     must swallow everything it can. So, while both \d+ and \d+? are pre-
3337     pared to adjust the number of digits they match in order to make the
3338     rest of the pattern match, (?>\d+) can only match an entire sequence of
3339     digits.
3340 nigel 63
3341 nigel 73 Atomic groups in general can of course contain arbitrarily complicated
3342     subpatterns, and can be nested. However, when the subpattern for an
3343     atomic group is just a single repeated item, as in the example above, a
3344     simpler notation, called a "possessive quantifier" can be used. This
3345     consists of an additional + character following a quantifier. Using
3346     this notation, the previous example can be rewritten as
3347 nigel 63
3348 nigel 75 \d++foo
3349 nigel 63
3350 nigel 73 Possessive quantifiers are always greedy; the setting of the
3351     PCRE_UNGREEDY option is ignored. They are a convenient notation for the
3352     simpler forms of atomic group. However, there is no difference in the
3353     meaning or processing of a possessive quantifier and the equivalent
3354     atomic group.
3355 nigel 63
3356 nigel 73 The possessive quantifier syntax is an extension to the Perl syntax. It
3357     originates in Sun's Java package.
3358 nigel 63
3359 nigel 73 When a pattern contains an unlimited repeat inside a subpattern that
3360     can itself be repeated an unlimited number of times, the use of an
3361     atomic group is the only way to avoid some failing matches taking a
3362     very long time indeed. The pattern
3363 nigel 63
3364 nigel 73 (\D+|<\d+>)*[!?]
3365 nigel 63
3366 nigel 73 matches an unlimited number of substrings that either consist of non-
3367     digits, or digits enclosed in <>, followed by either ! or ?. When it
3368     matches, it runs quickly. However, if it is applied to
3369 nigel 63
3370 nigel 73 aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
3371 nigel 63
3372 nigel 73 it takes a long time before reporting failure. This is because the
3373 nigel 75 string can be divided between the internal \D+ repeat and the external
3374     * repeat in a large number of ways, and all have to be tried. (The
3375     example uses [!?] rather than a single character at the end, because
3376     both PCRE and Perl have an optimization that allows for fast failure
3377     when a single character is used. They remember the last single charac-
3378     ter that is required for a match, and fail early if it is not present
3379     in the string.) If the pattern is changed so that it uses an atomic
3380     group, like this:
3381 nigel 63
3382 nigel 73 ((?>\D+)|<\d+>)*[!?]
3383 nigel 63
3384 nigel 75 sequences of non-digits cannot be broken, and failure happens quickly.
3385 nigel 63
3389 nigel 73 Outside a character class, a backslash followed by a digit greater than
3390     0 (and possibly further digits) is a back reference to a capturing sub-
3391 nigel 75 pattern earlier (that is, to its left) in the pattern, provided there
3392 nigel 73 have been that many previous capturing left parentheses.
3393 nigel 41
3394 nigel 73 However, if the decimal number following the backslash is less than 10,
3395 nigel 75 it is always taken as a back reference, and causes an error only if
3396     there are not that many capturing left parentheses in the entire pat-
3397     tern. In other words, the parentheses that are referenced need not be
3398     to the left of the reference for numbers less than 10. See the subsec-
3399     tion entitled "Non-printing characters" above for further details of
3400     the handling of digits following a backslash.
3401 nigel 41
3402 nigel 75 A back reference matches whatever actually matched the capturing sub-
3403     pattern in the current subject string, rather than anything matching
3404 nigel 73 the subpattern itself (see "Subpatterns as subroutines" below for a way
3405     of doing that). So the pattern
3406 nigel 41
3407 nigel 73 (sens|respons)e and \1ibility
3408 nigel 41
3409 nigel 75 matches "sense and sensibility" and "response and responsibility", but
3410     not "sense and responsibility". If caseful matching is in force at the
3411     time of the back reference, the case of letters is relevant. For exam-
3412 nigel 73 ple,
3413 nigel 41
3414 nigel 73 ((?i)rah)\s+\1
3415 nigel 41
3416 nigel 75 matches "rah rah" and "RAH RAH", but not "RAH rah", even though the
3417 nigel 73 original capturing subpattern is matched caselessly.
3418 nigel 41
3419 nigel 75 Back references to named subpatterns use the Python syntax (?P=name).
3420 nigel 73 We could rewrite the above example as follows:
3421 nigel 63
3422 nigel 73 (?<p1>(?i)rah)\s+(?P=p1)
3423 nigel 63
3424 nigel 75 There may be more than one back reference to the same subpattern. If a
3425     subpattern has not actually been used in a particular match, any back
3426 nigel 73 references to it always fail. For example, the pattern
3427 nigel 41
3428 nigel 73 (a|(bc))\2
3429 nigel 41