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

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

Parent Directory Parent Directory | Revision Log Revision Log


Revision 96 - (hide annotations) (download)
Fri Mar 2 13:10:43 2007 UTC (7 years, 1 month ago) by nigel
File MIME type: text/plain
File size: 263470 byte(s)
 r6896@hex:  nm | 2007-03-02 13:09:14 +0000
 Added EOL and keywork properties throughout

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     -----------------------------------------------------------------------------
8    
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
15    
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 nigel 93 just a few differences. (Certain features that appeared in Python and
22     PCRE before they appeared in Perl are also available using the Python
23     syntax.)
24 nigel 63
25 nigel 93 The current implementation of PCRE (release 7.x) corresponds approxi-
26     mately with Perl 5.10, including support for UTF-8 encoded strings and
27     Unicode general category properties. However, UTF-8 and Unicode support
28     has to be explicitly enabled; it is not the default. The Unicode tables
29     correspond to Unicode release 5.0.0.
30 nigel 77
31 nigel 93 In addition to the Perl-compatible matching function, PCRE contains an
32     alternative matching function that matches the same compiled patterns
33     in a different way. In certain circumstances, the alternative function
34     has some advantages. For a discussion of the two matching algorithms,
35     see the pcrematching page.
36    
37     PCRE is written in C and released as a C library. A number of people
38     have written wrappers and interfaces of various kinds. In particular,
39     Google Inc. have provided a comprehensive C++ wrapper. This is now
40 nigel 77 included as part of the PCRE distribution. The pcrecpp page has details
41 nigel 93 of this interface. Other people's contributions can be found in the
42 nigel 77 Contrib directory at the primary FTP site, which is:
43 nigel 63
44 nigel 73 ftp://ftp.csx.cam.ac.uk/pub/software/programming/pcre
45 nigel 63
46 nigel 93 Details of exactly which Perl regular expression features are and are
47 nigel 73 not supported by PCRE are given in separate documents. See the pcrepat-
48     tern and pcrecompat pages.
49 nigel 63
50 nigel 93 Some features of PCRE can be included, excluded, or changed when the
51     library is built. The pcre_config() function makes it possible for a
52     client to discover which features are available. The features them-
53     selves are described in the pcrebuild page. Documentation about build-
54     ing PCRE for various operating systems can be found in the README file
55 nigel 75 in the source distribution.
56 nigel 63
57 nigel 93 The library contains a number of undocumented internal functions and
58     data tables that are used by more than one of the exported external
59     functions, but which are not intended for use by external callers.
60     Their names all begin with "_pcre_", which hopefully will not provoke
61 nigel 83 any name clashes. In some environments, it is possible to control which
62 nigel 93 external symbols are exported when a shared library is built, and in
63 nigel 83 these cases the undocumented symbols are not exported.
64 nigel 63
65 nigel 77
66 nigel 63 USER DOCUMENTATION
67    
68 nigel 93 The user documentation for PCRE comprises a number of different sec-
69     tions. In the "man" format, each of these is a separate "man page". In
70     the HTML format, each is a separate page, linked from the index page.
71     In the plain text format, all the sections are concatenated, for ease
72 nigel 75 of searching. The sections are as follows:
73 nigel 63
74 nigel 73 pcre this document
75 nigel 77 pcreapi details of PCRE's native C API
76 nigel 73 pcrebuild options for building PCRE
77     pcrecallout details of the callout feature
78     pcrecompat discussion of Perl compatibility
79 nigel 77 pcrecpp details of the C++ wrapper
80 nigel 73 pcregrep description of the pcregrep command
81 nigel 77 pcrematching discussion of the two matching algorithms
82 nigel 75 pcrepartial details of the partial matching facility
83 nigel 73 pcrepattern syntax and semantics of supported
84     regular expressions
85     pcreperform discussion of performance issues
86 nigel 77 pcreposix the POSIX-compatible C API
87 nigel 75 pcreprecompile details of saving and re-using precompiled patterns
88 nigel 73 pcresample discussion of the sample program
89 nigel 91 pcrestack discussion of stack usage
90 nigel 75 pcretest description of the pcretest testing command
91 nigel 63
92 nigel 93 In addition, in the "man" and HTML formats, there is a short page for
93 nigel 77 each C library function, listing its arguments and results.
94 nigel 63
95    
96     LIMITATIONS
97    
98 nigel 93 There are some size limitations in PCRE but it is hoped that they will
99 nigel 73 never in practice be relevant.
100 nigel 63
101 nigel 93 The maximum length of a compiled pattern is 65539 (sic) bytes if PCRE
102 nigel 73 is compiled with the default internal linkage size of 2. If you want to
103 nigel 93 process regular expressions that are truly enormous, you can compile
104     PCRE with an internal linkage size of 3 or 4 (see the README file in
105     the source distribution and the pcrebuild documentation for details).
106     In these cases the limit is substantially larger. However, the speed
107     of execution is slower.
108 nigel 63
109 nigel 93 All values in repeating quantifiers must be less than 65536. The maxi-
110     mum compiled length of subpattern with an explicit repeat count is
111 nigel 91 30000 bytes. The maximum number of capturing subpatterns is 65535.
112 nigel 63
113 nigel 93 There is no limit to the number of parenthesized subpatterns, but there
114     can be no more than 65535 capturing subpatterns.
115 nigel 63
116 nigel 93 The maximum length of name for a named subpattern is 32 characters, and
117     the maximum number of named subpatterns is 10000.
118 nigel 91
119 nigel 93 The maximum length of a subject string is the largest positive number
120     that an integer variable can hold. However, when using the traditional
121 nigel 77 matching function, PCRE uses recursion to handle subpatterns and indef-
122 nigel 93 inite repetition. This means that the available stack space may limit
123 nigel 77 the size of a subject string that can be processed by certain patterns.
124 nigel 91 For a discussion of stack issues, see the pcrestack documentation.
125 nigel 63
126    
127 nigel 75 UTF-8 AND UNICODE PROPERTY SUPPORT
128 nigel 63
129 nigel 93 From release 3.3, PCRE has had some support for character strings
130     encoded in the UTF-8 format. For release 4.0 this was greatly extended
131     to cover most common requirements, and in release 5.0 additional sup-
132 nigel 75 port for Unicode general category properties was added.
133 nigel 63
134 nigel 93 In order process UTF-8 strings, you must build PCRE to include UTF-8
135     support in the code, and, in addition, you must call pcre_compile()
136     with the PCRE_UTF8 option flag. When you do this, both the pattern and
137     any subject strings that are matched against it are treated as UTF-8
138 nigel 73 strings instead of just strings of bytes.
139 nigel 63
140 nigel 93 If you compile PCRE with UTF-8 support, but do not use it at run time,
141     the library will be a bit bigger, but the additional run time overhead
142     is limited to testing the PCRE_UTF8 flag occasionally, so should not be
143     very big.
144 nigel 63
145 nigel 75 If PCRE is built with Unicode character property support (which implies
146 nigel 93 UTF-8 support), the escape sequences \p{..}, \P{..}, and \X are sup-
147 nigel 75 ported. The available properties that can be tested are limited to the
148 nigel 93 general category properties such as Lu for an upper case letter or Nd
149     for a decimal number, the Unicode script names such as Arabic or Han,
150     and the derived properties Any and L&. A full list is given in the
151 nigel 87 pcrepattern documentation. Only the short names for properties are sup-
152 nigel 93 ported. For example, \p{L} matches a letter. Its Perl synonym, \p{Let-
153     ter}, is not supported. Furthermore, in Perl, many properties may
154     optionally be prefixed by "Is", for compatibility with Perl 5.6. PCRE
155 nigel 87 does not support this.
156 nigel 75
157 nigel 73 The following comments apply when PCRE is running in UTF-8 mode:
158 nigel 63
159 nigel 93 1. When you set the PCRE_UTF8 flag, the strings passed as patterns and
160     subjects are checked for validity on entry to the relevant functions.
161 nigel 73 If an invalid UTF-8 string is passed, an error return is given. In some
162 nigel 93 situations, you may already know that your strings are valid, and
163 nigel 73 therefore want to skip these checks in order to improve performance. If
164 nigel 93 you set the PCRE_NO_UTF8_CHECK flag at compile time or at run time,
165     PCRE assumes that the pattern or subject it is given (respectively)
166     contains only valid UTF-8 codes. In this case, it does not diagnose an
167     invalid UTF-8 string. If you pass an invalid UTF-8 string to PCRE when
168     PCRE_NO_UTF8_CHECK is set, the results are undefined. Your program may
169 nigel 73 crash.
170 nigel 63
171 nigel 93 2. An unbraced hexadecimal escape sequence (such as \xb3) matches a
172 nigel 87 two-byte UTF-8 character if the value is greater than 127.
173 nigel 63
174 nigel 93 3. Octal numbers up to \777 are recognized, and match two-byte UTF-8
175 nigel 91 characters for values greater than \177.
176    
177 nigel 93 4. Repeat quantifiers apply to complete UTF-8 characters, not to indi-
178 nigel 73 vidual bytes, for example: \x{100}{3}.
179 nigel 63
180 nigel 93 5. The dot metacharacter matches one UTF-8 character instead of a sin-
181 nigel 75 gle byte.
182 nigel 63
183 nigel 93 6. The escape sequence \C can be used to match a single byte in UTF-8
184     mode, but its use can lead to some strange effects. This facility is
185 nigel 77 not available in the alternative matching function, pcre_dfa_exec().
186 nigel 63
187 nigel 93 7. The character escapes \b, \B, \d, \D, \s, \S, \w, and \W correctly
188     test characters of any code value, but the characters that PCRE recog-
189     nizes as digits, spaces, or word characters remain the same set as
190 nigel 75 before, all with values less than 256. This remains true even when PCRE
191 nigel 93 includes Unicode property support, because to do otherwise would slow
192     down PCRE in many common cases. If you really want to test for a wider
193     sense of, say, "digit", you must use Unicode property tests such as
194 nigel 75 \p{Nd}.
195 nigel 63
196 nigel 93 8. Similarly, characters that match the POSIX named character classes
197 nigel 75 are all low-valued characters.
198 nigel 63
199 nigel 93 9. Case-insensitive matching applies only to characters whose values
200     are less than 128, unless PCRE is built with Unicode property support.
201     Even when Unicode property support is available, PCRE still uses its
202     own character tables when checking the case of low-valued characters,
203     so as not to degrade performance. The Unicode property information is
204 nigel 87 used only for characters with higher values. Even when Unicode property
205     support is available, PCRE supports case-insensitive matching only when
206 nigel 93 there is a one-to-one mapping between a letter's cases. There are a
207     small number of many-to-one mappings in Unicode; these are not sup-
208 nigel 87 ported by PCRE.
209 nigel 63
210    
211     AUTHOR
212    
213 nigel 77 Philip Hazel
214 nigel 73 University Computing Service,
215 nigel 93 Cambridge CB2 3QH, England.
216 nigel 63
217 nigel 93 Putting an actual email address here seems to have been a spam magnet,
218 nigel 77 so I've taken it away. If you want to email me, use my initial and sur-
219     name, separated by a dot, at the domain ucs.cam.ac.uk.
220    
221 nigel 93 Last updated: 23 November 2006
222 nigel 87 Copyright (c) 1997-2006 University of Cambridge.
223 nigel 79 ------------------------------------------------------------------------------
224 nigel 63
225    
226 nigel 79 PCREBUILD(3) PCREBUILD(3)
227 nigel 63
228 nigel 79
229 nigel 73 NAME
230     PCRE - Perl-compatible regular expressions
231    
232 nigel 77
233 nigel 63 PCRE BUILD-TIME OPTIONS
234    
235 nigel 73 This document describes the optional features of PCRE that can be
236     selected when the library is compiled. They are all selected, or dese-
237 nigel 75 lected, by providing options to the configure script that is run before
238     the make command. The complete list of options for configure (which
239     includes the standard ones such as the selection of the installation
240     directory) can be obtained by running
241 nigel 63
242 nigel 73 ./configure --help
243 nigel 63
244 nigel 73 The following sections describe certain options whose names begin with
245     --enable or --disable. These settings specify changes to the defaults
246     for the configure command. Because of the way that configure works,
247     --enable and --disable always come in pairs, so the complementary
248     option always exists as well, but as it specifies the default, it is
249     not described.
250 nigel 63
251    
252 nigel 83 C++ SUPPORT
253    
254     By default, the configure script will search for a C++ compiler and C++
255     header files. If it finds them, it automatically builds the C++ wrapper
256     library for PCRE. You can disable this by adding
257    
258     --disable-cpp
259    
260     to the configure command.
261    
262    
263 nigel 63 UTF-8 SUPPORT
264    
265 nigel 73 To build PCRE with support for UTF-8 character strings, add
266 nigel 63
267 nigel 73 --enable-utf8
268 nigel 63
269 nigel 73 to the configure command. Of itself, this does not make PCRE treat
270     strings as UTF-8. As well as compiling PCRE with this option, you also
271     have have to set the PCRE_UTF8 option when you call the pcre_compile()
272     function.
273 nigel 63
274    
275 nigel 75 UNICODE CHARACTER PROPERTY SUPPORT
276    
277     UTF-8 support allows PCRE to process character values greater than 255
278     in the strings that it handles. On its own, however, it does not pro-
279     vide any facilities for accessing the properties of such characters. If
280     you want to be able to use the pattern escapes \P, \p, and \X, which
281     refer to Unicode character properties, you must add
282    
283     --enable-unicode-properties
284    
285     to the configure command. This implies UTF-8 support, even if you have
286     not explicitly requested it.
287    
288     Including Unicode property support adds around 90K of tables to the
289     PCRE library, approximately doubling its size. Only the general cate-
290     gory properties such as Lu and Nd are supported. Details are given in
291     the pcrepattern documentation.
292    
293    
294 nigel 63 CODE VALUE OF NEWLINE
295    
296 nigel 91 By default, PCRE interprets character 10 (linefeed, LF) as indicating
297     the end of a line. This is the normal newline character on Unix-like
298     systems. You can compile PCRE to use character 13 (carriage return, CR)
299     instead, by adding
300 nigel 63
301 nigel 73 --enable-newline-is-cr
302 nigel 63
303 nigel 91 to the configure command. There is also a --enable-newline-is-lf
304     option, which explicitly specifies linefeed as the newline character.
305 nigel 63
306 nigel 91 Alternatively, you can specify that line endings are to be indicated by
307     the two character sequence CRLF. If you want this, add
308 nigel 63
309 nigel 91 --enable-newline-is-crlf
310    
311 nigel 93 to the configure command. There is a fourth option, specified by
312 nigel 91
313 nigel 93 --enable-newline-is-any
314 nigel 91
315 nigel 93 which causes PCRE to recognize any Unicode newline sequence.
316    
317     Whatever line ending convention is selected when PCRE is built can be
318     overridden when the library functions are called. At build time it is
319     conventional to use the standard for your operating system.
320    
321    
322 nigel 63 BUILDING SHARED AND STATIC LIBRARIES
323    
324 nigel 93 The PCRE building process uses libtool to build both shared and static
325     Unix libraries by default. You can suppress one of these by adding one
326 nigel 73 of
327 nigel 63
328 nigel 73 --disable-shared
329     --disable-static
330 nigel 63
331 nigel 73 to the configure command, as required.
332 nigel 63
333    
334     POSIX MALLOC USAGE
335    
336 nigel 75 When PCRE is called through the POSIX interface (see the pcreposix doc-
337 nigel 93 umentation), additional working storage is required for holding the
338     pointers to capturing substrings, because PCRE requires three integers
339     per substring, whereas the POSIX interface provides only two. If the
340 nigel 73 number of expected substrings is small, the wrapper function uses space
341     on the stack, because this is faster than using malloc() for each call.
342     The default threshold above which the stack is no longer used is 10; it
343     can be changed by adding a setting such as
344 nigel 63
345 nigel 73 --with-posix-malloc-threshold=20
346 nigel 63
347 nigel 73 to the configure command.
348 nigel 63
349    
350     HANDLING VERY LARGE PATTERNS
351    
352 nigel 93 Within a compiled pattern, offset values are used to point from one
353     part to another (for example, from an opening parenthesis to an alter-
354     nation metacharacter). By default, two-byte values are used for these
355     offsets, leading to a maximum size for a compiled pattern of around
356     64K. This is sufficient to handle all but the most gigantic patterns.
357     Nevertheless, some people do want to process enormous patterns, so it
358     is possible to compile PCRE to use three-byte or four-byte offsets by
359 nigel 73 adding a setting such as
360 nigel 63
361 nigel 73 --with-link-size=3
362 nigel 63
363 nigel 93 to the configure command. The value given must be 2, 3, or 4. Using
364     longer offsets slows down the operation of PCRE because it has to load
365 nigel 73 additional bytes when handling them.
366 nigel 63
367 nigel 93 If you build PCRE with an increased link size, test 2 (and test 5 if
368     you are using UTF-8) will fail. Part of the output of these tests is a
369     representation of the compiled pattern, and this changes with the link
370 nigel 73 size.
371 nigel 63
372 nigel 73
373     AVOIDING EXCESSIVE STACK USAGE
374    
375 nigel 77 When matching with the pcre_exec() function, PCRE implements backtrack-
376 nigel 93 ing by making recursive calls to an internal function called match().
377     In environments where the size of the stack is limited, this can se-
378     verely limit PCRE's operation. (The Unix environment does not usually
379 nigel 91 suffer from this problem, but it may sometimes be necessary to increase
380 nigel 93 the maximum stack size. There is a discussion in the pcrestack docu-
381     mentation.) An alternative approach to recursion that uses memory from
382     the heap to remember data, instead of using recursive function calls,
383     has been implemented to work round the problem of limited stack size.
384 nigel 91 If you want to build a version of PCRE that works this way, add
385 nigel 73
386     --disable-stack-for-recursion
387    
388 nigel 93 to the configure command. With this configuration, PCRE will use the
389     pcre_stack_malloc and pcre_stack_free variables to call memory manage-
390     ment functions. Separate functions are provided because the usage is
391     very predictable: the block sizes requested are always the same, and
392     the blocks are always freed in reverse order. A calling program might
393     be able to implement optimized functions that perform better than the
394     standard malloc() and free() functions. PCRE runs noticeably more
395 nigel 77 slowly when built in this way. This option affects only the pcre_exec()
396     function; it is not relevant for the the pcre_dfa_exec() function.
397 nigel 73
398    
399 nigel 91 LIMITING PCRE RESOURCE USAGE
400    
401 nigel 93 Internally, PCRE has a function called match(), which it calls repeat-
402     edly (sometimes recursively) when matching a pattern with the
403     pcre_exec() function. By controlling the maximum number of times this
404     function may be called during a single matching operation, a limit can
405     be placed on the resources used by a single call to pcre_exec(). The
406     limit can be changed at run time, as described in the pcreapi documen-
407     tation. The default is 10 million, but this can be changed by adding a
408 nigel 91 setting such as
409    
410     --with-match-limit=500000
411    
412 nigel 93 to the configure command. This setting has no effect on the
413 nigel 91 pcre_dfa_exec() matching function.
414    
415 nigel 93 In some environments it is desirable to limit the depth of recursive
416 nigel 91 calls of match() more strictly than the total number of calls, in order
417 nigel 93 to restrict the maximum amount of stack (or heap, if --disable-stack-
418 nigel 91 for-recursion is specified) that is used. A second limit controls this;
419 nigel 93 it defaults to the value that is set for --with-match-limit, which
420     imposes no additional constraints. However, you can set a lower limit
421 nigel 91 by adding, for example,
422    
423     --with-match-limit-recursion=10000
424    
425 nigel 93 to the configure command. This value can also be overridden at run
426 nigel 91 time.
427    
428    
429 nigel 73 USING EBCDIC CODE
430    
431 nigel 93 PCRE assumes by default that it will run in an environment where the
432     character code is ASCII (or Unicode, which is a superset of ASCII).
433     PCRE can, however, be compiled to run in an EBCDIC environment by
434 nigel 75 adding
435 nigel 73
436     --enable-ebcdic
437    
438     to the configure command.
439    
440 nigel 93
441     SEE ALSO
442    
443     pcreapi(3), pcre_config(3).
444    
445     Last updated: 30 November 2006
446 nigel 91 Copyright (c) 1997-2006 University of Cambridge.
447 nigel 79 ------------------------------------------------------------------------------
448 nigel 63
449    
450 nigel 79 PCREMATCHING(3) PCREMATCHING(3)
451 nigel 63
452 nigel 79
453 nigel 77 NAME
454     PCRE - Perl-compatible regular expressions
455 nigel 73
456 nigel 77
457     PCRE MATCHING ALGORITHMS
458    
459     This document describes the two different algorithms that are available
460     in PCRE for matching a compiled regular expression against a given sub-
461     ject string. The "standard" algorithm is the one provided by the
462     pcre_exec() function. This works in the same was as Perl's matching
463     function, and provides a Perl-compatible matching operation.
464    
465     An alternative algorithm is provided by the pcre_dfa_exec() function;
466     this operates in a different way, and is not Perl-compatible. It has
467     advantages and disadvantages compared with the standard algorithm, and
468     these are described below.
469    
470     When there is only one possible way in which a given subject string can
471     match a pattern, the two algorithms give the same answer. A difference
472     arises, however, when there are multiple possibilities. For example, if
473     the pattern
474    
475     ^<.*>
476    
477     is matched against the string
478    
479     <something> <something else> <something further>
480    
481     there are three possible answers. The standard algorithm finds only one
482 nigel 93 of them, whereas the alternative algorithm finds all three.
483 nigel 77
484    
485     REGULAR EXPRESSIONS AS TREES
486    
487     The set of strings that are matched by a regular expression can be rep-
488     resented as a tree structure. An unlimited repetition in the pattern
489     makes the tree of infinite size, but it is still a tree. Matching the
490     pattern to a given subject string (from a given starting point) can be
491 nigel 91 thought of as a search of the tree. There are two ways to search a
492     tree: depth-first and breadth-first, and these correspond to the two
493     matching algorithms provided by PCRE.
494 nigel 77
495    
496     THE STANDARD MATCHING ALGORITHM
497    
498     In the terminology of Jeffrey Friedl's book Mastering Regular Expres-
499     sions, the standard algorithm is an "NFA algorithm". It conducts a
500     depth-first search of the pattern tree. That is, it proceeds along a
501     single path through the tree, checking that the subject matches what is
502     required. When there is a mismatch, the algorithm tries any alterna-
503     tives at the current point, and if they all fail, it backs up to the
504     previous branch point in the tree, and tries the next alternative
505     branch at that level. This often involves backing up (moving to the
506     left) in the subject string as well. The order in which repetition
507     branches are tried is controlled by the greedy or ungreedy nature of
508     the quantifier.
509    
510     If a leaf node is reached, a matching string has been found, and at
511     that point the algorithm stops. Thus, if there is more than one possi-
512     ble match, this algorithm returns the first one that it finds. Whether
513     this is the shortest, the longest, or some intermediate length depends
514     on the way the greedy and ungreedy repetition quantifiers are specified
515     in the pattern.
516    
517     Because it ends up with a single path through the tree, it is rela-
518     tively straightforward for this algorithm to keep track of the sub-
519     strings that are matched by portions of the pattern in parentheses.
520     This provides support for capturing parentheses and back references.
521    
522    
523 nigel 93 THE ALTERNATIVE MATCHING ALGORITHM
524 nigel 77
525 nigel 93 This algorithm conducts a breadth-first search of the tree. Starting
526     from the first matching point in the subject, it scans the subject
527     string from left to right, once, character by character, and as it does
528     this, it remembers all the paths through the tree that represent valid
529     matches. In Friedl's terminology, this is a kind of "DFA algorithm",
530     though it is not implemented as a traditional finite state machine (it
531     keeps multiple states active simultaneously).
532 nigel 77
533 nigel 93 The scan continues until either the end of the subject is reached, or
534     there are no more unterminated paths. At this point, terminated paths
535     represent the different matching possibilities (if there are none, the
536     match has failed). Thus, if there is more than one possible match,
537 nigel 77 this algorithm finds all of them, and in particular, it finds the long-
538 nigel 93 est. In PCRE, there is an option to stop the algorithm after the first
539 nigel 77 match (which is necessarily the shortest) has been found.
540    
541     Note that all the matches that are found start at the same point in the
542     subject. If the pattern
543    
544     cat(er(pillar)?)
545    
546 nigel 93 is matched against the string "the caterpillar catchment", the result
547     will be the three strings "cat", "cater", and "caterpillar" that start
548 nigel 77 at the fourth character of the subject. The algorithm does not automat-
549     ically move on to find matches that start at later positions.
550    
551     There are a number of features of PCRE regular expressions that are not
552 nigel 93 supported by the alternative matching algorithm. They are as follows:
553 nigel 77
554 nigel 93 1. Because the algorithm finds all possible matches, the greedy or
555     ungreedy nature of repetition quantifiers is not relevant. Greedy and
556     ungreedy quantifiers are treated in exactly the same way. However, pos-
557     sessive quantifiers can make a difference when what follows could also
558     match what is quantified, for example in a pattern like this:
559 nigel 77
560 nigel 93 ^a++\w!
561    
562     This pattern matches "aaab!" but not "aaa!", which would be matched by
563     a non-possessive quantifier. Similarly, if an atomic group is present,
564     it is matched as if it were a standalone pattern at the current point,
565     and the longest match is then "locked in" for the rest of the overall
566     pattern.
567    
568 nigel 77 2. When dealing with multiple paths through the tree simultaneously, it
569 nigel 93 is not straightforward to keep track of captured substrings for the
570     different matching possibilities, and PCRE's implementation of this
571 nigel 77 algorithm does not attempt to do this. This means that no captured sub-
572     strings are available.
573    
574 nigel 93 3. Because no substrings are captured, back references within the pat-
575 nigel 77 tern are not supported, and cause errors if encountered.
576    
577 nigel 93 4. For the same reason, conditional expressions that use a backrefer-
578     ence as the condition or test for a specific group recursion are not
579     supported.
580 nigel 77
581     5. Callouts are supported, but the value of the capture_top field is
582     always 1, and the value of the capture_last field is always -1.
583    
584     6. The \C escape sequence, which (in the standard algorithm) matches a
585 nigel 93 single byte, even in UTF-8 mode, is not supported because the alterna-
586     tive algorithm moves through the subject string one character at a
587     time, for all active paths through the tree.
588 nigel 77
589    
590 nigel 93 ADVANTAGES OF THE ALTERNATIVE ALGORITHM
591 nigel 77
592 nigel 93 Using the alternative matching algorithm provides the following advan-
593     tages:
594 nigel 77
595     1. All possible matches (at a single point in the subject) are automat-
596 nigel 93 ically found, and in particular, the longest match is found. To find
597 nigel 77 more than one match using the standard algorithm, you have to do kludgy
598     things with callouts.
599    
600 nigel 93 2. There is much better support for partial matching. The restrictions
601     on the content of the pattern that apply when using the standard algo-
602     rithm for partial matching do not apply to the alternative algorithm.
603     For non-anchored patterns, the starting position of a partial match is
604     available.
605 nigel 77
606 nigel 93 3. Because the alternative algorithm scans the subject string just
607     once, and never needs to backtrack, it is possible to pass very long
608     subject strings to the matching function in several pieces, checking
609     for partial matching each time.
610 nigel 77
611    
612 nigel 93 DISADVANTAGES OF THE ALTERNATIVE ALGORITHM
613 nigel 77
614 nigel 93 The alternative algorithm suffers from a number of disadvantages:
615 nigel 77
616 nigel 93 1. It is substantially slower than the standard algorithm. This is
617     partly because it has to search for all possible matches, but is also
618 nigel 77 because it is less susceptible to optimization.
619    
620     2. Capturing parentheses and back references are not supported.
621    
622 nigel 93 3. Although atomic groups are supported, their use does not provide the
623     performance advantage that it does for the standard algorithm.
624 nigel 77
625 nigel 93 Last updated: 24 November 2006
626 nigel 91 Copyright (c) 1997-2006 University of Cambridge.
627 nigel 79 ------------------------------------------------------------------------------
628 nigel 77
629    
630 nigel 79 PCREAPI(3) PCREAPI(3)
631 nigel 77
632 nigel 79
633 nigel 73 NAME
634     PCRE - Perl-compatible regular expressions
635    
636 nigel 77
637 nigel 75 PCRE NATIVE API
638 nigel 63
639 nigel 73 #include <pcre.h>
640 nigel 41
641 nigel 73 pcre *pcre_compile(const char *pattern, int options,
642     const char **errptr, int *erroffset,
643     const unsigned char *tableptr);
644 nigel 41
645 nigel 77 pcre *pcre_compile2(const char *pattern, int options,
646     int *errorcodeptr,
647     const char **errptr, int *erroffset,
648     const unsigned char *tableptr);
649    
650 nigel 73 pcre_extra *pcre_study(const pcre *code, int options,
651     const char **errptr);
652 nigel 41
653 nigel 73 int pcre_exec(const pcre *code, const pcre_extra *extra,
654     const char *subject, int length, int startoffset,
655     int options, int *ovector, int ovecsize);
656 nigel 41
657 nigel 77 int pcre_dfa_exec(const pcre *code, const pcre_extra *extra,
658     const char *subject, int length, int startoffset,
659     int options, int *ovector, int ovecsize,
660     int *workspace, int wscount);
661    
662 nigel 73 int pcre_copy_named_substring(const pcre *code,
663     const char *subject, int *ovector,
664     int stringcount, const char *stringname,
665     char *buffer, int buffersize);
666 nigel 63
667 nigel 73 int pcre_copy_substring(const char *subject, int *ovector,
668     int stringcount, int stringnumber, char *buffer,
669     int buffersize);
670 nigel 41
671 nigel 73 int pcre_get_named_substring(const pcre *code,
672     const char *subject, int *ovector,
673     int stringcount, const char *stringname,
674     const char **stringptr);
675 nigel 63
676 nigel 73 int pcre_get_stringnumber(const pcre *code,
677     const char *name);
678 nigel 63
679 nigel 91 int pcre_get_stringtable_entries(const pcre *code,
680     const char *name, char **first, char **last);
681    
682 nigel 73 int pcre_get_substring(const char *subject, int *ovector,
683     int stringcount, int stringnumber,
684     const char **stringptr);
685 nigel 41
686 nigel 73 int pcre_get_substring_list(const char *subject,
687     int *ovector, int stringcount, const char ***listptr);
688 nigel 41
689 nigel 73 void pcre_free_substring(const char *stringptr);
690 nigel 49
691 nigel 73 void pcre_free_substring_list(const char **stringptr);
692 nigel 49
693 nigel 73 const unsigned char *pcre_maketables(void);
694 nigel 41
695 nigel 73 int pcre_fullinfo(const pcre *code, const pcre_extra *extra,
696     int what, void *where);
697 nigel 43
698 nigel 73 int pcre_info(const pcre *code, int *optptr, int *firstcharptr);
699 nigel 63
700 nigel 77 int pcre_refcount(pcre *code, int adjust);
701    
702 nigel 73 int pcre_config(int what, void *where);
703 nigel 41
704 nigel 73 char *pcre_version(void);
705 nigel 63
706 nigel 73 void *(*pcre_malloc)(size_t);
707 nigel 41
708 nigel 73 void (*pcre_free)(void *);
709 nigel 41
710 nigel 73 void *(*pcre_stack_malloc)(size_t);
711 nigel 41
712 nigel 73 void (*pcre_stack_free)(void *);
713 nigel 41
714 nigel 73 int (*pcre_callout)(pcre_callout_block *);
715 nigel 41
716 nigel 73
717 nigel 75 PCRE API OVERVIEW
718 nigel 41
719 nigel 73 PCRE has its own native API, which is described in this document. There
720 nigel 93 are also some wrapper functions that correspond to the POSIX regular
721 nigel 77 expression API. These are described in the pcreposix documentation.
722     Both of these APIs define a set of C function calls. A C++ wrapper is
723     distributed with PCRE. It is documented in the pcrecpp page.
724 nigel 43
725 nigel 77 The native API C function prototypes are defined in the header file
726     pcre.h, and on Unix systems the library itself is called libpcre. It
727 nigel 75 can normally be accessed by adding -lpcre to the command for linking an
728     application that uses PCRE. The header file defines the macros
729     PCRE_MAJOR and PCRE_MINOR to contain the major and minor release num-
730     bers for the library. Applications can use these to include support
731     for different releases of PCRE.
732 nigel 41
733 nigel 77 The functions pcre_compile(), pcre_compile2(), pcre_study(), and
734     pcre_exec() are used for compiling and matching regular expressions in
735     a Perl-compatible manner. A sample program that demonstrates the sim-
736     plest way of using them is provided in the file called pcredemo.c in
737     the source distribution. The pcresample documentation describes how to
738     run it.
739 nigel 49
740 nigel 77 A second matching function, pcre_dfa_exec(), which is not Perl-compati-
741     ble, is also provided. This uses a different algorithm for the match-
742 nigel 91 ing. The alternative algorithm finds all possible matches (at a given
743 nigel 93 point in the subject), and scans the subject just once. However, this
744     algorithm does not return captured substrings. A description of the two
745     matching algorithms and their advantages and disadvantages is given in
746     the pcrematching documentation.
747 nigel 63
748 nigel 77 In addition to the main compiling and matching functions, there are
749     convenience functions for extracting captured substrings from a subject
750     string that is matched by pcre_exec(). They are:
751    
752 nigel 73 pcre_copy_substring()
753     pcre_copy_named_substring()
754     pcre_get_substring()
755     pcre_get_named_substring()
756     pcre_get_substring_list()
757 nigel 75 pcre_get_stringnumber()
758 nigel 91 pcre_get_stringtable_entries()
759 nigel 63
760 nigel 73 pcre_free_substring() and pcre_free_substring_list() are also provided,
761     to free the memory used for extracted strings.
762 nigel 41
763 nigel 77 The function pcre_maketables() is used to build a set of character
764     tables in the current locale for passing to pcre_compile(),
765     pcre_exec(), or pcre_dfa_exec(). This is an optional facility that is
766     provided for specialist use. Most commonly, no special tables are
767     passed, in which case internal tables that are generated when PCRE is
768     built are used.
769 nigel 49
770 nigel 75 The function pcre_fullinfo() is used to find out information about a
771     compiled pattern; pcre_info() is an obsolete version that returns only
772     some of the available information, but is retained for backwards com-
773     patibility. The function pcre_version() returns a pointer to a string
774 nigel 73 containing the version of PCRE and its date of release.
775 nigel 41
776 nigel 77 The function pcre_refcount() maintains a reference count in a data
777     block containing a compiled pattern. This is provided for the benefit
778     of object-oriented applications.
779    
780 nigel 75 The global variables pcre_malloc and pcre_free initially contain the
781     entry points of the standard malloc() and free() functions, respec-
782 nigel 73 tively. PCRE calls the memory management functions via these variables,
783 nigel 75 so a calling program can replace them if it wishes to intercept the
784 nigel 73 calls. This should be done before calling any PCRE functions.
785 nigel 41
786 nigel 75 The global variables pcre_stack_malloc and pcre_stack_free are also
787     indirections to memory management functions. These special functions
788     are used only when PCRE is compiled to use the heap for remembering
789 nigel 77 data, instead of recursive function calls, when running the pcre_exec()
790 nigel 91 function. See the pcrebuild documentation for details of how to do
791     this. It is a non-standard way of building PCRE, for use in environ-
792     ments that have limited stacks. Because of the greater use of memory
793     management, it runs more slowly. Separate functions are provided so
794     that special-purpose external code can be used for this case. When
795     used, these functions are always called in a stack-like manner (last
796     obtained, first freed), and always for memory blocks of the same size.
797     There is a discussion about PCRE's stack usage in the pcrestack docu-
798     mentation.
799 nigel 41
800 nigel 73 The global variable pcre_callout initially contains NULL. It can be set
801 nigel 77 by the caller to a "callout" function, which PCRE will then call at
802     specified points during a matching operation. Details are given in the
803 nigel 73 pcrecallout documentation.
804 nigel 41
805 nigel 73
806 nigel 91 NEWLINES
807    
808 nigel 93 PCRE supports four different conventions for indicating line breaks in
809     strings: a single CR (carriage return) character, a single LF (line-
810     feed) character, the two-character sequence CRLF, or any Unicode new-
811     line sequence. The Unicode newline sequences are the three just men-
812     tioned, plus the single characters VT (vertical tab, U+000B), FF (form-
813     feed, U+000C), NEL (next line, U+0085), LS (line separator, U+2028),
814     and PS (paragraph separator, U+2029).
815    
816     Each of the first three conventions is used by at least one operating
817     system as its standard newline sequence. When PCRE is built, a default
818     can be specified. The default default is LF, which is the Unix stan-
819     dard. When PCRE is run, the default can be overridden, either when a
820     pattern is compiled, or when it is matched.
821    
822 nigel 91 In the PCRE documentation the word "newline" is used to mean "the char-
823 nigel 93 acter or pair of characters that indicate a line break". The choice of
824     newline convention affects the handling of the dot, circumflex, and
825     dollar metacharacters, the handling of #-comments in /x mode, and, when
826     CRLF is a recognized line ending sequence, the match position advance-
827     ment for a non-anchored pattern. The choice of newline convention does
828     not affect the interpretation of the \n or \r escape sequences.
829 nigel 91
830    
831 nigel 63 MULTITHREADING
832    
833 nigel 93 The PCRE functions can be used in multi-threading applications, with
834 nigel 73 the proviso that the memory management functions pointed to by
835     pcre_malloc, pcre_free, pcre_stack_malloc, and pcre_stack_free, and the
836     callout function pointed to by pcre_callout, are shared by all threads.
837 nigel 41
838 nigel 93 The compiled form of a regular expression is not altered during match-
839 nigel 73 ing, so the same compiled pattern can safely be used by several threads
840     at once.
841 nigel 41
842    
843 nigel 75 SAVING PRECOMPILED PATTERNS FOR LATER USE
844    
845     The compiled form of a regular expression can be saved and re-used at a
846 nigel 93 later time, possibly by a different program, and even on a host other
847     than the one on which it was compiled. Details are given in the
848 nigel 75 pcreprecompile documentation.
849    
850    
851 nigel 63 CHECKING BUILD-TIME OPTIONS
852 nigel 41
853 nigel 73 int pcre_config(int what, void *where);
854 nigel 63
855 nigel 93 The function pcre_config() makes it possible for a PCRE client to dis-
856 nigel 73 cover which optional features have been compiled into the PCRE library.
857 nigel 93 The pcrebuild documentation has more details about these optional fea-
858 nigel 73 tures.
859 nigel 63
860 nigel 93 The first argument for pcre_config() is an integer, specifying which
861 nigel 73 information is required; the second argument is a pointer to a variable
862 nigel 93 into which the information is placed. The following information is
863 nigel 73 available:
864 nigel 63
865 nigel 73 PCRE_CONFIG_UTF8
866 nigel 63
867 nigel 93 The output is an integer that is set to one if UTF-8 support is avail-
868 nigel 73 able; otherwise it is set to zero.
869 nigel 63
870 nigel 75 PCRE_CONFIG_UNICODE_PROPERTIES
871    
872 nigel 93 The output is an integer that is set to one if support for Unicode
873 nigel 75 character properties is available; otherwise it is set to zero.
874    
875 nigel 73 PCRE_CONFIG_NEWLINE
876 nigel 63
877 nigel 93 The output is an integer whose value specifies the default character
878     sequence that is recognized as meaning "newline". The four values that
879     are supported are: 10 for LF, 13 for CR, 3338 for CRLF, and -1 for ANY.
880     The default should normally be the standard sequence for your operating
881     system.
882 nigel 63
883 nigel 73 PCRE_CONFIG_LINK_SIZE
884 nigel 63
885 nigel 91 The output is an integer that contains the number of bytes used for
886 nigel 73 internal linkage in compiled regular expressions. The value is 2, 3, or
887 nigel 91 4. Larger values allow larger regular expressions to be compiled, at
888     the expense of slower matching. The default value of 2 is sufficient
889     for all but the most massive patterns, since it allows the compiled
890 nigel 73 pattern to be up to 64K in size.
891 nigel 63
892 nigel 73 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
893 nigel 63
894 nigel 91 The output is an integer that contains the threshold above which the
895     POSIX interface uses malloc() for output vectors. Further details are
896 nigel 73 given in the pcreposix documentation.
897 nigel 63
898 nigel 73 PCRE_CONFIG_MATCH_LIMIT
899 nigel 63
900 nigel 73 The output is an integer that gives the default limit for the number of
901 nigel 91 internal matching function calls in a pcre_exec() execution. Further
902 nigel 73 details are given with pcre_exec() below.
903 nigel 63
904 nigel 87 PCRE_CONFIG_MATCH_LIMIT_RECURSION
905    
906 nigel 91 The output is an integer that gives the default limit for the depth of
907     recursion when calling the internal matching function in a pcre_exec()
908 nigel 87 execution. Further details are given with pcre_exec() below.
909    
910 nigel 73 PCRE_CONFIG_STACKRECURSE
911 nigel 63
912 nigel 91 The output is an integer that is set to one if internal recursion when
913 nigel 77 running pcre_exec() is implemented by recursive function calls that use
914 nigel 91 the stack to remember their state. This is the usual way that PCRE is
915 nigel 77 compiled. The output is zero if PCRE was compiled to use blocks of data
916 nigel 91 on the heap instead of recursive function calls. In this case,
917     pcre_stack_malloc and pcre_stack_free are called to manage memory
918 nigel 77 blocks on the heap, thus avoiding the use of the stack.
919 nigel 73
920    
921 nigel 41 COMPILING A PATTERN
922 nigel 63
923 nigel 73 pcre *pcre_compile(const char *pattern, int options,
924     const char **errptr, int *erroffset,
925     const unsigned char *tableptr);
926 nigel 63
927 nigel 77 pcre *pcre_compile2(const char *pattern, int options,
928     int *errorcodeptr,
929     const char **errptr, int *erroffset,
930     const unsigned char *tableptr);
931 nigel 41
932 nigel 77 Either of the functions pcre_compile() or pcre_compile2() can be called
933     to compile a pattern into an internal form. The only difference between
934 nigel 91 the two interfaces is that pcre_compile2() has an additional argument,
935 nigel 77 errorcodeptr, via which a numerical error code can be returned.
936    
937     The pattern is a C string terminated by a binary zero, and is passed in
938 nigel 91 the pattern argument. A pointer to a single block of memory that is
939     obtained via pcre_malloc is returned. This contains the compiled code
940 nigel 77 and related data. The pcre type is defined for the returned block; this
941     is a typedef for a structure whose contents are not externally defined.
942 nigel 91 It is up to the caller to free the memory (via pcre_free) when it is no
943     longer required.
944 nigel 77
945 nigel 91 Although the compiled code of a PCRE regex is relocatable, that is, it
946 nigel 73 does not depend on memory location, the complete pcre data block is not
947 nigel 91 fully relocatable, because it may contain a copy of the tableptr argu-
948 nigel 75 ment, which is an address (see below).
949 nigel 41
950 nigel 93 The options argument contains various bit settings that affect the com-
951     pilation. It should be zero if no options are required. The available
952 nigel 91 options are described below. Some of them, in particular, those that
953     are compatible with Perl, can also be set and unset from within the
954     pattern (see the detailed description in the pcrepattern documenta-
955     tion). For these options, the contents of the options argument speci-
956     fies their initial settings at the start of compilation and execution.
957     The PCRE_ANCHORED and PCRE_NEWLINE_xxx options can be set at the time
958     of matching as well as at compile time.
959 nigel 41
960 nigel 73 If errptr is NULL, pcre_compile() returns NULL immediately. Otherwise,
961 nigel 91 if compilation of a pattern fails, pcre_compile() returns NULL, and
962 nigel 73 sets the variable pointed to by errptr to point to a textual error mes-
963 nigel 87 sage. This is a static string that is part of the library. You must not
964     try to free it. The offset from the start of the pattern to the charac-
965     ter where the error was discovered is placed in the variable pointed to
966 nigel 91 by erroffset, which must not be NULL. If it is, an immediate error is
967 nigel 73 given.
968 nigel 53
969 nigel 91 If pcre_compile2() is used instead of pcre_compile(), and the error-
970     codeptr argument is not NULL, a non-zero error code number is returned
971     via this argument in the event of an error. This is in addition to the
972 nigel 77 textual error message. Error codes and messages are listed below.
973    
974 nigel 91 If the final argument, tableptr, is NULL, PCRE uses a default set of
975     character tables that are built when PCRE is compiled, using the
976     default C locale. Otherwise, tableptr must be an address that is the
977     result of a call to pcre_maketables(). This value is stored with the
978     compiled pattern, and used again by pcre_exec(), unless another table
979 nigel 75 pointer is passed to it. For more discussion, see the section on locale
980     support below.
981 nigel 53
982 nigel 91 This code fragment shows a typical straightforward call to pcre_com-
983 nigel 73 pile():
984 nigel 41
985 nigel 73 pcre *re;
986     const char *error;
987     int erroffset;
988     re = pcre_compile(
989     "^A.*Z", /* the pattern */
990     0, /* default options */
991     &error, /* for error message */
992     &erroffset, /* for error offset */
993     NULL); /* use default character tables */
994 nigel 41
995 nigel 91 The following names for option bits are defined in the pcre.h header
996 nigel 75 file:
997 nigel 41
998 nigel 73 PCRE_ANCHORED
999 nigel 41
1000 nigel 73 If this bit is set, the pattern is forced to be "anchored", that is, it
1001 nigel 91 is constrained to match only at the first matching point in the string
1002     that is being searched (the "subject string"). This effect can also be
1003     achieved by appropriate constructs in the pattern itself, which is the
1004 nigel 73 only way to do it in Perl.
1005 nigel 41
1006 nigel 75 PCRE_AUTO_CALLOUT
1007    
1008     If this bit is set, pcre_compile() automatically inserts callout items,
1009 nigel 91 all with number 255, before each pattern item. For discussion of the
1010 nigel 75 callout facility, see the pcrecallout documentation.
1011    
1012 nigel 73 PCRE_CASELESS
1013 nigel 41
1014 nigel 91 If this bit is set, letters in the pattern match both upper and lower
1015     case letters. It is equivalent to Perl's /i option, and it can be
1016     changed within a pattern by a (?i) option setting. In UTF-8 mode, PCRE
1017     always understands the concept of case for characters whose values are
1018     less than 128, so caseless matching is always possible. For characters
1019     with higher values, the concept of case is supported if PCRE is com-
1020     piled with Unicode property support, but not otherwise. If you want to
1021     use caseless matching for characters 128 and above, you must ensure
1022     that PCRE is compiled with Unicode property support as well as with
1023 nigel 77 UTF-8 support.
1024 nigel 41
1025 nigel 73 PCRE_DOLLAR_ENDONLY
1026 nigel 41
1027 nigel 91 If this bit is set, a dollar metacharacter in the pattern matches only
1028     at the end of the subject string. Without this option, a dollar also
1029     matches immediately before a newline at the end of the string (but not
1030     before any other newlines). The PCRE_DOLLAR_ENDONLY option is ignored
1031     if PCRE_MULTILINE is set. There is no equivalent to this option in
1032     Perl, and no way to set it within a pattern.
1033 nigel 41
1034 nigel 73 PCRE_DOTALL
1035 nigel 41
1036 nigel 73 If this bit is set, a dot metacharater in the pattern matches all char-
1037 nigel 91 acters, including those that indicate newline. Without it, a dot does
1038     not match when the current position is at a newline. This option is
1039     equivalent to Perl's /s option, and it can be changed within a pattern
1040     by a (?s) option setting. A negative class such as [^a] always matches
1041 nigel 93 newline characters, independent of the setting of this option.
1042 nigel 63
1043 nigel 91 PCRE_DUPNAMES
1044    
1045     If this bit is set, names used to identify capturing subpatterns need
1046     not be unique. This can be helpful for certain types of pattern when it
1047     is known that only one instance of the named subpattern can ever be
1048     matched. There are more details of named subpatterns below; see also
1049     the pcrepattern documentation.
1050    
1051 nigel 73 PCRE_EXTENDED
1052 nigel 41
1053 nigel 91 If this bit is set, whitespace data characters in the pattern are
1054 nigel 77 totally ignored except when escaped or inside a character class. White-
1055     space does not include the VT character (code 11). In addition, charac-
1056     ters between an unescaped # outside a character class and the next new-
1057 nigel 91 line, inclusive, are also ignored. This is equivalent to Perl's /x
1058     option, and it can be changed within a pattern by a (?x) option set-
1059     ting.
1060 nigel 41
1061 nigel 91 This option makes it possible to include comments inside complicated
1062     patterns. Note, however, that this applies only to data characters.
1063     Whitespace characters may never appear within special character
1064     sequences in a pattern, for example within the sequence (?( which
1065 nigel 73 introduces a conditional subpattern.
1066 nigel 41
1067 nigel 73 PCRE_EXTRA
1068 nigel 41
1069 nigel 91 This option was invented in order to turn on additional functionality
1070     of PCRE that is incompatible with Perl, but it is currently of very
1071     little use. When set, any backslash in a pattern that is followed by a
1072     letter that has no special meaning causes an error, thus reserving
1073     these combinations for future expansion. By default, as in Perl, a
1074     backslash followed by a letter with no special meaning is treated as a
1075     literal. (Perl can, however, be persuaded to give a warning for this.)
1076     There are at present no other features controlled by this option. It
1077     can also be set by a (?X) option setting within a pattern.
1078 nigel 41
1079 nigel 77 PCRE_FIRSTLINE
1080    
1081 nigel 87 If this option is set, an unanchored pattern is required to match
1082 nigel 91 before or at the first newline in the subject string, though the
1083     matched text may continue over the newline.
1084 nigel 77
1085 nigel 73 PCRE_MULTILINE
1086 nigel 41
1087 nigel 87 By default, PCRE treats the subject string as consisting of a single
1088     line of characters (even if it actually contains newlines). The "start
1089     of line" metacharacter (^) matches only at the start of the string,
1090     while the "end of line" metacharacter ($) matches only at the end of
1091 nigel 75 the string, or before a terminating newline (unless PCRE_DOLLAR_ENDONLY
1092     is set). This is the same as Perl.
1093 nigel 63
1094 nigel 87 When PCRE_MULTILINE it is set, the "start of line" and "end of line"
1095 nigel 91 constructs match immediately following or immediately before internal
1096     newlines in the subject string, respectively, as well as at the very
1097     start and end. This is equivalent to Perl's /m option, and it can be
1098     changed within a pattern by a (?m) option setting. If there are no new-
1099     lines in a subject string, or no occurrences of ^ or $ in a pattern,
1100 nigel 73 setting PCRE_MULTILINE has no effect.
1101 nigel 63
1102 nigel 91 PCRE_NEWLINE_CR
1103     PCRE_NEWLINE_LF
1104     PCRE_NEWLINE_CRLF
1105 nigel 93 PCRE_NEWLINE_ANY
1106 nigel 91
1107     These options override the default newline definition that was chosen
1108     when PCRE was built. Setting the first or the second specifies that a
1109     newline is indicated by a single character (CR or LF, respectively).
1110 nigel 93 Setting PCRE_NEWLINE_CRLF specifies that a newline is indicated by the
1111     two-character CRLF sequence. Setting PCRE_NEWLINE_ANY specifies that
1112     any Unicode newline sequence should be recognized. The Unicode newline
1113     sequences are the three just mentioned, plus the single characters VT
1114     (vertical tab, U+000B), FF (formfeed, U+000C), NEL (next line, U+0085),
1115     LS (line separator, U+2028), and PS (paragraph separator, U+2029). The
1116     last two are recognized only in UTF-8 mode.
1117 nigel 91
1118 nigel 93 The newline setting in the options word uses three bits that are
1119     treated as a number, giving eight possibilities. Currently only five
1120     are used (default plus the four values above). This means that if you
1121     set more than one newline option, the combination may or may not be
1122     sensible. For example, PCRE_NEWLINE_CR with PCRE_NEWLINE_LF is equiva-
1123     lent to PCRE_NEWLINE_CRLF, but other combinations yield unused numbers
1124     and cause an error.
1125 nigel 91
1126 nigel 93 The only time that a line break is specially recognized when compiling
1127     a pattern is if PCRE_EXTENDED is set, and an unescaped # outside a
1128     character class is encountered. This indicates a comment that lasts
1129     until after the next line break sequence. In other circumstances, line
1130     break sequences are treated as literal data, except that in
1131     PCRE_EXTENDED mode, both CR and LF are treated as whitespace characters
1132     and are therefore ignored.
1133    
1134     The newline option that is set at compile time becomes the default that
1135     is used for pcre_exec() and pcre_dfa_exec(), but it can be overridden.
1136    
1137 nigel 73 PCRE_NO_AUTO_CAPTURE
1138 nigel 41
1139 nigel 73 If this option is set, it disables the use of numbered capturing paren-
1140 nigel 91 theses in the pattern. Any opening parenthesis that is not followed by
1141     ? behaves as if it were followed by ?: but named parentheses can still
1142     be used for capturing (and they acquire numbers in the usual way).
1143 nigel 73 There is no equivalent of this option in Perl.
1144 nigel 41
1145 nigel 73 PCRE_UNGREEDY
1146 nigel 41
1147 nigel 91 This option inverts the "greediness" of the quantifiers so that they
1148     are not greedy by default, but become greedy if followed by "?". It is
1149     not compatible with Perl. It can also be set by a (?U) option setting
1150 nigel 73 within the pattern.
1151 nigel 41
1152 nigel 73 PCRE_UTF8
1153 nigel 49
1154 nigel 91 This option causes PCRE to regard both the pattern and the subject as
1155     strings of UTF-8 characters instead of single-byte character strings.
1156     However, it is available only when PCRE is built to include UTF-8 sup-
1157     port. If not, the use of this option provokes an error. Details of how
1158     this option changes the behaviour of PCRE are given in the section on
1159 nigel 75 UTF-8 support in the main pcre page.
1160 nigel 71
1161 nigel 73 PCRE_NO_UTF8_CHECK
1162 nigel 71
1163 nigel 73 When PCRE_UTF8 is set, the validity of the pattern as a UTF-8 string is
1164 nigel 91 automatically checked. If an invalid UTF-8 sequence of bytes is found,
1165     pcre_compile() returns an error. If you already know that your pattern
1166     is valid, and you want to skip this check for performance reasons, you
1167     can set the PCRE_NO_UTF8_CHECK option. When it is set, the effect of
1168 nigel 73 passing an invalid UTF-8 string as a pattern is undefined. It may cause
1169 nigel 91 your program to crash. Note that this option can also be passed to
1170     pcre_exec() and pcre_dfa_exec(), to suppress the UTF-8 validity check-
1171 nigel 77 ing of subject strings.
1172 nigel 71
1173 nigel 73
1174 nigel 77 COMPILATION ERROR CODES
1175    
1176 nigel 91 The following table lists the error codes than may be returned by
1177     pcre_compile2(), along with the error messages that may be returned by
1178 nigel 93 both compiling functions. As PCRE has developed, some error codes have
1179     fallen out of use. To avoid confusion, they have not been re-used.
1180 nigel 77
1181     0 no error
1182     1 \ at end of pattern
1183     2 \c at end of pattern
1184     3 unrecognized character follows \
1185     4 numbers out of order in {} quantifier
1186     5 number too big in {} quantifier
1187     6 missing terminating ] for character class
1188     7 invalid escape sequence in character class
1189     8 range out of order in character class
1190     9 nothing to repeat
1191 nigel 93 10 [this code is not in use]
1192 nigel 77 11 internal error: unexpected repeat
1193     12 unrecognized character after (?
1194     13 POSIX named classes are supported only within a class
1195     14 missing )
1196     15 reference to non-existent subpattern
1197     16 erroffset passed as NULL
1198     17 unknown option bit(s) set
1199     18 missing ) after comment
1200 nigel 93 19 [this code is not in use]
1201 nigel 77 20 regular expression too large
1202     21 failed to get memory
1203     22 unmatched parentheses
1204     23 internal error: code overflow
1205     24 unrecognized character after (?<
1206     25 lookbehind assertion is not fixed length
1207 nigel 91 26 malformed number or name after (?(
1208 nigel 77 27 conditional group contains more than two branches
1209     28 assertion expected after (?(
1210     29 (?R or (?digits must be followed by )
1211     30 unknown POSIX class name
1212     31 POSIX collating elements are not supported
1213     32 this version of PCRE is not compiled with PCRE_UTF8 support
1214 nigel 93 33 [this code is not in use]
1215 nigel 77 34 character value in \x{...} sequence is too large
1216     35 invalid condition (?(0)
1217     36 \C not allowed in lookbehind assertion
1218     37 PCRE does not support \L, \l, \N, \U, or \u
1219     38 number after (?C is > 255
1220     39 closing ) for (?C expected
1221     40 recursive call could loop indefinitely
1222     41 unrecognized character after (?P
1223 nigel 93 42 syntax error in subpattern name (missing terminator)
1224 nigel 91 43 two named subpatterns have the same name
1225 nigel 77 44 invalid UTF-8 string
1226     45 support for \P, \p, and \X has not been compiled
1227     46 malformed \P or \p sequence
1228     47 unknown property name after \P or \p
1229 nigel 91 48 subpattern name is too long (maximum 32 characters)
1230     49 too many named subpatterns (maximum 10,000)
1231     50 repeated subpattern is too long
1232     51 octal value is greater than \377 (not in UTF-8 mode)
1233 nigel 93 52 internal error: overran compiling workspace
1234     53 internal error: previously-checked referenced subpattern not
1235     found
1236     54 DEFINE group contains more than one branch
1237     55 repeating a DEFINE group is not allowed
1238     56 inconsistent NEWLINE options"
1239 nigel 77
1240    
1241 nigel 63 STUDYING A PATTERN
1242 nigel 49
1243 nigel 77 pcre_extra *pcre_study(const pcre *code, int options
1244 nigel 73 const char **errptr);
1245 nigel 63
1246 nigel 91 If a compiled pattern is going to be used several times, it is worth
1247 nigel 75 spending more time analyzing it in order to speed up the time taken for
1248 nigel 91 matching. The function pcre_study() takes a pointer to a compiled pat-
1249 nigel 75 tern as its first argument. If studying the pattern produces additional
1250 nigel 91 information that will help speed up matching, pcre_study() returns a
1251     pointer to a pcre_extra block, in which the study_data field points to
1252 nigel 75 the results of the study.
1253 nigel 41
1254 nigel 75 The returned value from pcre_study() can be passed directly to
1255 nigel 91 pcre_exec(). However, a pcre_extra block also contains other fields
1256     that can be set by the caller before the block is passed; these are
1257 nigel 75 described below in the section on matching a pattern.
1258 nigel 63
1259 nigel 91 If studying the pattern does not produce any additional information
1260 nigel 75 pcre_study() returns NULL. In that circumstance, if the calling program
1261 nigel 91 wants to pass any of the other fields to pcre_exec(), it must set up
1262 nigel 75 its own pcre_extra block.
1263 nigel 41
1264 nigel 91 The second argument of pcre_study() contains option bits. At present,
1265 nigel 75 no options are defined, and this argument should always be zero.
1266    
1267 nigel 91 The third argument for pcre_study() is a pointer for an error message.
1268     If studying succeeds (even if no data is returned), the variable it
1269     points to is set to NULL. Otherwise it is set to point to a textual
1270 nigel 87 error message. This is a static string that is part of the library. You
1271 nigel 91 must not try to free it. You should test the error pointer for NULL
1272 nigel 87 after calling pcre_study(), to be sure that it has run successfully.
1273 nigel 41
1274 nigel 73 This is a typical call to pcre_study():
1275 nigel 53
1276 nigel 73 pcre_extra *pe;
1277     pe = pcre_study(
1278     re, /* result of pcre_compile() */
1279     0, /* no options exist */
1280     &error); /* set to NULL or points to a message */
1281 nigel 53
1282 nigel 73 At present, studying a pattern is useful only for non-anchored patterns
1283 nigel 91 that do not have a single fixed starting character. A bitmap of possi-
1284 nigel 75 ble starting bytes is created.
1285 nigel 41
1286    
1287 nigel 63 LOCALE SUPPORT
1288 nigel 41
1289 nigel 91 PCRE handles caseless matching, and determines whether characters are
1290     letters digits, or whatever, by reference to a set of tables, indexed
1291     by character value. When running in UTF-8 mode, this applies only to
1292     characters with codes less than 128. Higher-valued codes never match
1293     escapes such as \w or \d, but can be tested with \p if PCRE is built
1294     with Unicode character property support. The use of locales with Uni-
1295 nigel 87 code is discouraged.
1296 nigel 41
1297 nigel 91 An internal set of tables is created in the default C locale when PCRE
1298     is built. This is used when the final argument of pcre_compile() is
1299     NULL, and is sufficient for many applications. An alternative set of
1300     tables can, however, be supplied. These may be created in a different
1301     locale from the default. As more and more applications change to using
1302 nigel 75 Unicode, the need for this locale support is expected to die away.
1303 nigel 41
1304 nigel 91 External tables are built by calling the pcre_maketables() function,
1305     which has no arguments, in the relevant locale. The result can then be
1306     passed to pcre_compile() or pcre_exec() as often as necessary. For
1307     example, to build and use tables that are appropriate for the French
1308     locale (where accented characters with values greater than 128 are
1309 nigel 75 treated as letters), the following code could be used:
1310    
1311     setlocale(LC_CTYPE, "fr_FR");
1312 nigel 73 tables = pcre_maketables();
1313     re = pcre_compile(..., tables);
1314 nigel 41
1315 nigel 91 When pcre_maketables() runs, the tables are built in memory that is
1316     obtained via pcre_malloc. It is the caller's responsibility to ensure
1317     that the memory containing the tables remains available for as long as
1318 nigel 75 it is needed.
1319 nigel 41
1320 nigel 75 The pointer that is passed to pcre_compile() is saved with the compiled
1321 nigel 91 pattern, and the same tables are used via this pointer by pcre_study()
1322 nigel 75 and normally also by pcre_exec(). Thus, by default, for any single pat-
1323     tern, compilation, studying and matching all happen in the same locale,
1324     but different patterns can be compiled in different locales.
1325 nigel 41
1326 nigel 91 It is possible to pass a table pointer or NULL (indicating the use of
1327     the internal tables) to pcre_exec(). Although not intended for this
1328     purpose, this facility could be used to match a pattern in a different
1329 nigel 75 locale from the one in which it was compiled. Passing table pointers at
1330     run time is discussed below in the section on matching a pattern.
1331    
1332    
1333 nigel 63 INFORMATION ABOUT A PATTERN
1334 nigel 41
1335 nigel 73 int pcre_fullinfo(const pcre *code, const pcre_extra *extra,
1336     int what, void *where);
1337 nigel 63
1338 nigel 91 The pcre_fullinfo() function returns information about a compiled pat-
1339 nigel 73 tern. It replaces the obsolete pcre_info() function, which is neverthe-
1340     less retained for backwards compability (and is documented below).
1341 nigel 43
1342 nigel 91 The first argument for pcre_fullinfo() is a pointer to the compiled
1343     pattern. The second argument is the result of pcre_study(), or NULL if
1344     the pattern was not studied. The third argument specifies which piece
1345     of information is required, and the fourth argument is a pointer to a
1346     variable to receive the data. The yield of the function is zero for
1347 nigel 73 success, or one of the following negative numbers:
1348 nigel 41
1349 nigel 73 PCRE_ERROR_NULL the argument code was NULL
1350     the argument where was NULL
1351     PCRE_ERROR_BADMAGIC the "magic number" was not found
1352     PCRE_ERROR_BADOPTION the value of what was invalid
1353 nigel 53
1354 nigel 91 The "magic number" is placed at the start of each compiled pattern as
1355     an simple check against passing an arbitrary memory pointer. Here is a
1356     typical call of pcre_fullinfo(), to obtain the length of the compiled
1357 nigel 75 pattern:
1358 nigel 53
1359 nigel 73 int rc;
1360 nigel 91 size_t length;
1361 nigel 73 rc = pcre_fullinfo(
1362     re, /* result of pcre_compile() */
1363     pe, /* result of pcre_study(), or NULL */
1364     PCRE_INFO_SIZE, /* what is required */
1365     &length); /* where to put the data */
1366 nigel 43
1367 nigel 91 The possible values for the third argument are defined in pcre.h, and
1368 nigel 73 are as follows:
1369 nigel 43
1370 nigel 73 PCRE_INFO_BACKREFMAX
1371 nigel 41
1372 nigel 91 Return the number of the highest back reference in the pattern. The
1373     fourth argument should point to an int variable. Zero is returned if
1374 nigel 73 there are no back references.
1375 nigel 43
1376 nigel 73 PCRE_INFO_CAPTURECOUNT
1377 nigel 43
1378 nigel 91 Return the number of capturing subpatterns in the pattern. The fourth
1379 nigel 73 argument should point to an int variable.
1380 nigel 43
1381 nigel 77 PCRE_INFO_DEFAULT_TABLES
1382 nigel 75
1383 nigel 91 Return a pointer to the internal default character tables within PCRE.
1384     The fourth argument should point to an unsigned char * variable. This
1385 nigel 75 information call is provided for internal use by the pcre_study() func-
1386 nigel 91 tion. External callers can cause PCRE to use its internal tables by
1387 nigel 75 passing a NULL table pointer.
1388    
1389 nigel 73 PCRE_INFO_FIRSTBYTE
1390 nigel 43
1391 nigel 91 Return information about the first byte of any matched string, for a
1392     non-anchored pattern. The fourth argument should point to an int vari-
1393     able. (This option used to be called PCRE_INFO_FIRSTCHAR; the old name
1394     is still recognized for backwards compatibility.)
1395 nigel 41
1396 nigel 91 If there is a fixed first byte, for example, from a pattern such as
1397 nigel 93 (cat|cow|coyote), its value is returned. Otherwise, if either
1398 nigel 41
1399 nigel 87 (a) the pattern was compiled with the PCRE_MULTILINE option, and every
1400 nigel 73 branch starts with "^", or
1401 nigel 43
1402 nigel 73 (b) every branch of the pattern starts with ".*" and PCRE_DOTALL is not
1403     set (if it were set, the pattern would be anchored),
1404 nigel 41
1405 nigel 87 -1 is returned, indicating that the pattern matches only at the start
1406     of a subject string or after any newline within the string. Otherwise
1407 nigel 73 -2 is returned. For anchored patterns, -2 is returned.
1408 nigel 41
1409 nigel 73 PCRE_INFO_FIRSTTABLE
1410 nigel 41
1411 nigel 87 If the pattern was studied, and this resulted in the construction of a
1412 nigel 73 256-bit table indicating a fixed set of bytes for the first byte in any
1413 nigel 87 matching string, a pointer to the table is returned. Otherwise NULL is
1414     returned. The fourth argument should point to an unsigned char * vari-
1415 nigel 73 able.
1416 nigel 43
1417 nigel 73 PCRE_INFO_LASTLITERAL
1418 nigel 43
1419 nigel 87 Return the value of the rightmost literal byte that must exist in any
1420     matched string, other than at its start, if such a byte has been
1421 nigel 73 recorded. The fourth argument should point to an int variable. If there
1422 nigel 87 is no such byte, -1 is returned. For anchored patterns, a last literal
1423     byte is recorded only if it follows something of variable length. For
1424 nigel 73 example, for the pattern /^a\d+z\d+/ the returned value is "z", but for
1425     /^a\dz\d/ the returned value is -1.
1426 nigel 63
1427 nigel 73 PCRE_INFO_NAMECOUNT
1428     PCRE_INFO_NAMEENTRYSIZE
1429     PCRE_INFO_NAMETABLE
1430 nigel 63
1431 nigel 87 PCRE supports the use of named as well as numbered capturing parenthe-
1432     ses. The names are just an additional way of identifying the parenthe-
1433 nigel 91 ses, which still acquire numbers. Several convenience functions such as
1434     pcre_get_named_substring() are provided for extracting captured sub-
1435     strings by name. It is also possible to extract the data directly, by
1436     first converting the name to a number in order to access the correct
1437     pointers in the output vector (described with pcre_exec() below). To do
1438     the conversion, you need to use the name-to-number map, which is
1439     described by these three values.
1440 nigel 63
1441 nigel 73 The map consists of a number of fixed-size entries. PCRE_INFO_NAMECOUNT
1442     gives the number of entries, and PCRE_INFO_NAMEENTRYSIZE gives the size
1443 nigel 87 of each entry; both of these return an int value. The entry size
1444     depends on the length of the longest name. PCRE_INFO_NAMETABLE returns
1445     a pointer to the first entry of the table (a pointer to char). The
1446 nigel 73 first two bytes of each entry are the number of the capturing parenthe-
1447 nigel 87 sis, most significant byte first. The rest of the entry is the corre-
1448     sponding name, zero terminated. The names are in alphabetical order.
1449 nigel 91 When PCRE_DUPNAMES is set, duplicate names are in order of their paren-
1450     theses numbers. For example, consider the following pattern (assume
1451     PCRE_EXTENDED is set, so white space - including newlines - is
1452     ignored):
1453 nigel 63
1454 nigel 93 (?<date> (?<year>(\d\d)?\d\d) -
1455     (?<month>\d\d) - (?<day>\d\d) )
1456 nigel 63
1457 nigel 87 There are four named subpatterns, so the table has four entries, and
1458     each entry in the table is eight bytes long. The table is as follows,
1459 nigel 75 with non-printing bytes shows in hexadecimal, and undefined bytes shown
1460     as ??:
1461 nigel 63
1462 nigel 73 00 01 d a t e 00 ??
1463     00 05 d a y 00 ?? ??
1464     00 04 m o n t h 00
1465     00 02 y e a r 00 ??
1466 nigel 63
1467 nigel 87 When writing code to extract data from named subpatterns using the
1468 nigel 91 name-to-number map, remember that the length of the entries is likely
1469     to be different for each compiled pattern.
1470 nigel 63
1471 nigel 73 PCRE_INFO_OPTIONS
1472 nigel 63
1473 nigel 87 Return a copy of the options with which the pattern was compiled. The
1474     fourth argument should point to an unsigned long int variable. These
1475 nigel 73 option bits are those specified in the call to pcre_compile(), modified
1476     by any top-level option settings within the pattern itself.
1477 nigel 63
1478 nigel 87 A pattern is automatically anchored by PCRE if all of its top-level
1479 nigel 73 alternatives begin with one of the following:
1480 nigel 63
1481 nigel 73 ^ unless PCRE_MULTILINE is set
1482     \A always
1483     \G always
1484     .* if PCRE_DOTALL is set and there are no back
1485     references to the subpattern in which .* appears
1486 nigel 63
1487 nigel 73 For such patterns, the PCRE_ANCHORED bit is set in the options returned
1488     by pcre_fullinfo().
1489 nigel 63
1490 nigel 73 PCRE_INFO_SIZE
1491 nigel 63
1492 nigel 87 Return the size of the compiled pattern, that is, the value that was
1493 nigel 73 passed as the argument to pcre_malloc() when PCRE was getting memory in
1494     which to place the compiled data. The fourth argument should point to a
1495     size_t variable.
1496 nigel 63
1497 nigel 73 PCRE_INFO_STUDYSIZE
1498 nigel 63
1499 nigel 75 Return the size of the data block pointed to by the study_data field in
1500 nigel 87 a pcre_extra block. That is, it is the value that was passed to
1501 nigel 73 pcre_malloc() when PCRE was getting memory into which to place the data
1502 nigel 87 created by pcre_study(). The fourth argument should point to a size_t
1503 nigel 73 variable.
1504 nigel 63
1505 nigel 73
1506 nigel 63 OBSOLETE INFO FUNCTION
1507    
1508 nigel 73 int pcre_info(const pcre *code, int *optptr, int *firstcharptr);
1509 nigel 63
1510 nigel 87 The pcre_info() function is now obsolete because its interface is too
1511     restrictive to return all the available data about a compiled pattern.
1512     New programs should use pcre_fullinfo() instead. The yield of
1513     pcre_info() is the number of capturing subpatterns, or one of the fol-
1514 nigel 73 lowing negative numbers:
1515 nigel 43
1516 nigel 73 PCRE_ERROR_NULL the argument code was NULL
1517     PCRE_ERROR_BADMAGIC the "magic number" was not found
1518 nigel 43
1519 nigel 87 If the optptr argument is not NULL, a copy of the options with which
1520     the pattern was compiled is placed in the integer it points to (see
1521 nigel 73 PCRE_INFO_OPTIONS above).
1522 nigel 43
1523 nigel 87 If the pattern is not anchored and the firstcharptr argument is not
1524     NULL, it is used to pass back information about the first character of
1525 nigel 73 any matched string (see PCRE_INFO_FIRSTBYTE above).
1526 nigel 43
1527    
1528 nigel 77 REFERENCE COUNTS
1529 nigel 53
1530 nigel 77 int pcre_refcount(pcre *code, int adjust);
1531    
1532 nigel 87 The pcre_refcount() function is used to maintain a reference count in
1533 nigel 77 the data block that contains a compiled pattern. It is provided for the
1534 nigel 87 benefit of applications that operate in an object-oriented manner,
1535 nigel 77 where different parts of the application may be using the same compiled
1536     pattern, but you want to free the block when they are all done.
1537    
1538     When a pattern is compiled, the reference count field is initialized to
1539 nigel 87 zero. It is changed only by calling this function, whose action is to
1540     add the adjust value (which may be positive or negative) to it. The
1541 nigel 77 yield of the function is the new value. However, the value of the count
1542 nigel 87 is constrained to lie between 0 and 65535, inclusive. If the new value
1543 nigel 77 is outside these limits, it is forced to the appropriate limit value.
1544    
1545 nigel 87 Except when it is zero, the reference count is not correctly preserved
1546     if a pattern is compiled on one host and then transferred to a host
1547 nigel 77 whose byte-order is different. (This seems a highly unlikely scenario.)
1548    
1549    
1550     MATCHING A PATTERN: THE TRADITIONAL FUNCTION
1551    
1552 nigel 73 int pcre_exec(const pcre *code, const pcre_extra *extra,
1553     const char *subject, int length, int startoffset,
1554     int options, int *ovector, int ovecsize);
1555 nigel 53
1556 nigel 87 The function pcre_exec() is called to match a subject string against a
1557     compiled pattern, which is passed in the code argument. If the pattern
1558 nigel 75 has been studied, the result of the study should be passed in the extra
1559 nigel 87 argument. This function is the main matching facility of the library,
1560 nigel 77 and it operates in a Perl-like manner. For specialist use there is also
1561 nigel 87 an alternative matching function, which is described below in the sec-
1562 nigel 77 tion about the pcre_dfa_exec() function.
1563 nigel 41
1564 nigel 87 In most applications, the pattern will have been compiled (and option-
1565     ally studied) in the same process that calls pcre_exec(). However, it
1566 nigel 75 is possible to save compiled patterns and study data, and then use them
1567 nigel 87 later in different processes, possibly even on different hosts. For a
1568 nigel 75 discussion about this, see the pcreprecompile documentation.
1569    
1570 nigel 73 Here is an example of a simple call to pcre_exec():
1571 nigel 53
1572 nigel 73 int rc;
1573     int ovector[30];
1574     rc = pcre_exec(
1575     re, /* result of pcre_compile() */
1576     NULL, /* we didn't study the pattern */
1577     "some string", /* the subject string */
1578     11, /* the length of the subject string */
1579     0, /* start at offset 0 in the subject */
1580     0, /* default options */
1581 nigel 75 ovector, /* vector of integers for substring information */
1582 nigel 77 30); /* number of elements (NOT size in bytes) */
1583 nigel 53
1584 nigel 75 Extra data for pcre_exec()
1585 nigel 63
1586 nigel 87 If the extra argument is not NULL, it must point to a pcre_extra data
1587     block. The pcre_study() function returns such a block (when it doesn't
1588     return NULL), but you can also create one for yourself, and pass addi-
1589     tional information in it. The pcre_extra block contains the following
1590     fields (not necessarily in this order):
1591 nigel 75
1592 nigel 73 unsigned long int flags;
1593     void *study_data;
1594     unsigned long int match_limit;
1595 nigel 87 unsigned long int match_limit_recursion;
1596 nigel 73 void *callout_data;
1597 nigel 75 const unsigned char *tables;
1598 nigel 63
1599 nigel 87 The flags field is a bitmap that specifies which of the other fields
1600 nigel 73 are set. The flag bits are:
1601 nigel 63
1602 nigel 73 PCRE_EXTRA_STUDY_DATA
1603     PCRE_EXTRA_MATCH_LIMIT
1604 nigel 87 PCRE_EXTRA_MATCH_LIMIT_RECURSION
1605 nigel 73 PCRE_EXTRA_CALLOUT_DATA
1606 nigel 75 PCRE_EXTRA_TABLES
1607 nigel 63
1608 nigel 87 Other flag bits should be set to zero. The study_data field is set in
1609     the pcre_extra block that is returned by pcre_study(), together with
1610 nigel 75 the appropriate flag bit. You should not set this yourself, but you may
1611 nigel 87 add to the block by setting the other fields and their corresponding
1612 nigel 75 flag bits.
1613 nigel 63
1614 nigel 73 The match_limit field provides a means of preventing PCRE from using up
1615 nigel 87 a vast amount of resources when running patterns that are not going to
1616     match, but which have a very large number of possibilities in their
1617     search trees. The classic example is the use of nested unlimited
1618 nigel 75 repeats.
1619 nigel 63
1620 nigel 87 Internally, PCRE uses a function called match() which it calls repeat-
1621     edly (sometimes recursively). The limit set by match_limit is imposed
1622     on the number of times this function is called during a match, which
1623     has the effect of limiting the amount of backtracking that can take
1624     place. For patterns that are not anchored, the count restarts from zero
1625     for each position in the subject string.
1626 nigel 75
1627 nigel 87 The default value for the limit can be set when PCRE is built; the
1628     default default is 10 million, which handles all but the most extreme
1629     cases. You can override the default by suppling pcre_exec() with a
1630     pcre_extra block in which match_limit is set, and
1631     PCRE_EXTRA_MATCH_LIMIT is set in the flags field. If the limit is
1632 nigel 73 exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
1633 nigel 63
1634 nigel 87 The match_limit_recursion field is similar to match_limit, but instead
1635     of limiting the total number of times that match() is called, it limits
1636     the depth of recursion. The recursion depth is a smaller number than
1637     the total number of calls, because not all calls to match() are recur-
1638     sive. This limit is of use only if it is set smaller than match_limit.
1639    
1640     Limiting the recursion depth limits the amount of stack that can be
1641     used, or, when PCRE has been compiled to use memory on the heap instead
1642     of the stack, the amount of heap memory that can be used.
1643    
1644     The default value for match_limit_recursion can be set when PCRE is
1645     built; the default default is the same value as the default for
1646     match_limit. You can override the default by suppling pcre_exec() with
1647     a pcre_extra block in which match_limit_recursion is set, and
1648     PCRE_EXTRA_MATCH_LIMIT_RECURSION is set in the flags field. If the
1649     limit is exceeded, pcre_exec() returns PCRE_ERROR_RECURSIONLIMIT.
1650    
1651     The pcre_callout field is used in conjunction with the "callout" fea-
1652 nigel 73 ture, which is described in the pcrecallout documentation.
1653 nigel 63
1654 nigel 87 The tables field is used to pass a character tables pointer to
1655     pcre_exec(); this overrides the value that is stored with the compiled
1656     pattern. A non-NULL value is stored with the compiled pattern only if
1657     custom tables were supplied to pcre_compile() via its tableptr argu-
1658 nigel 75 ment. If NULL is passed to pcre_exec() using this mechanism, it forces
1659 nigel 87 PCRE's internal tables to be used. This facility is helpful when re-
1660     using patterns that have been saved after compiling with an external
1661     set of tables, because the external tables might be at a different
1662     address when pcre_exec() is called. See the pcreprecompile documenta-
1663 nigel 75 tion for a discussion of saving compiled patterns for later use.
1664 nigel 41
1665 nigel 75 Option bits for pcre_exec()
1666 nigel 71
1667 nigel 87 The unused bits of the options argument for pcre_exec() must be zero.
1668 nigel 91 The only bits that may be set are PCRE_ANCHORED, PCRE_NEWLINE_xxx,
1669     PCRE_NOTBOL, PCRE_NOTEOL, PCRE_NOTEMPTY, PCRE_NO_UTF8_CHECK and
1670     PCRE_PARTIAL.
1671 nigel 41
1672 nigel 75 PCRE_ANCHORED
1673 nigel 41
1674 nigel 91 The PCRE_ANCHORED option limits pcre_exec() to matching at the first
1675     matching position. If a pattern was compiled with PCRE_ANCHORED, or
1676     turned out to be anchored by virtue of its contents, it cannot be made
1677 nigel 75 unachored at matching time.
1678    
1679 nigel 91 PCRE_NEWLINE_CR
1680     PCRE_NEWLINE_LF
1681     PCRE_NEWLINE_CRLF
1682 nigel 93 PCRE_NEWLINE_ANY
1683 nigel 91
1684     These options override the newline definition that was chosen or
1685     defaulted when the pattern was compiled. For details, see the descrip-
1686 nigel 93 tion of pcre_compile() above. During matching, the newline choice
1687     affects the behaviour of the dot, circumflex, and dollar metacharac-
1688     ters. It may also alter the way the match position is advanced after a
1689     match failure for an unanchored pattern. When PCRE_NEWLINE_CRLF or
1690     PCRE_NEWLINE_ANY is set, and a match attempt fails when the current
1691     position is at a CRLF sequence, the match position is advanced by two
1692     characters instead of one, in other words, to after the CRLF.
1693 nigel 91
1694 nigel 73 PCRE_NOTBOL
1695 nigel 41
1696 nigel 75 This option specifies that first character of the subject string is not
1697 nigel 93 the beginning of a line, so the circumflex metacharacter should not
1698     match before it. Setting this without PCRE_MULTILINE (at compile time)
1699     causes circumflex never to match. This option affects only the behav-
1700 nigel 77 iour of the circumflex metacharacter. It does not affect \A.
1701 nigel 41
1702 nigel 73 PCRE_NOTEOL
1703 nigel 41
1704 nigel 75 This option specifies that the end of the subject string is not the end
1705 nigel 93 of a line, so the dollar metacharacter should not match it nor (except
1706     in multiline mode) a newline immediately before it. Setting this with-
1707 nigel 75 out PCRE_MULTILINE (at compile time) causes dollar never to match. This
1708 nigel 93 option affects only the behaviour of the dollar metacharacter. It does
1709 nigel 75 not affect \Z or \z.
1710 nigel 41
1711 nigel 73 PCRE_NOTEMPTY
1712 nigel 41
1713 nigel 73 An empty string is not considered to be a valid match if this option is
1714 nigel 93 set. If there are alternatives in the pattern, they are tried. If all
1715     the alternatives match the empty string, the entire match fails. For
1716 nigel 73 example, if the pattern
1717 nigel 41
1718 nigel 73 a?b?
1719 nigel 41
1720 nigel 93 is applied to a string not beginning with "a" or "b", it matches the
1721     empty string at the start of the subject. With PCRE_NOTEMPTY set, this
1722 nigel 73 match is not valid, so PCRE searches further into the string for occur-
1723     rences of "a" or "b".
1724 nigel 41
1725 nigel 73 Perl has no direct equivalent of PCRE_NOTEMPTY, but it does make a spe-
1726 nigel 93 cial case of a pattern match of the empty string within its split()
1727     function, and when using the /g modifier. It is possible to emulate
1728 nigel 73 Perl's behaviour after matching a null string by first trying the match
1729 nigel 75 again at the same offset with PCRE_NOTEMPTY and PCRE_ANCHORED, and then
1730 nigel 93 if that fails by advancing the starting offset (see below) and trying
1731 nigel 75 an ordinary match again. There is some code that demonstrates how to do
1732     this in the pcredemo.c sample program.
1733 nigel 41
1734 nigel 75 PCRE_NO_UTF8_CHECK
1735    
1736     When PCRE_UTF8 is set at compile time, the validity of the subject as a
1737 nigel 93 UTF-8 string is automatically checked when pcre_exec() is subsequently
1738     called. The value of startoffset is also checked to ensure that it
1739     points to the start of a UTF-8 character. If an invalid UTF-8 sequence
1740 nigel 75 of bytes is found, pcre_exec() returns the error PCRE_ERROR_BADUTF8. If
1741 nigel 93 startoffset contains an invalid value, PCRE_ERROR_BADUTF8_OFFSET is
1742 nigel 75 returned.
1743    
1744 nigel 93 If you already know that your subject is valid, and you want to skip
1745     these checks for performance reasons, you can set the
1746     PCRE_NO_UTF8_CHECK option when calling pcre_exec(). You might want to
1747     do this for the second and subsequent calls to pcre_exec() if you are
1748     making repeated calls to find all the matches in a single subject
1749     string. However, you should be sure that the value of startoffset
1750     points to the start of a UTF-8 character. When PCRE_NO_UTF8_CHECK is
1751     set, the effect of passing an invalid UTF-8 string as a subject, or a
1752     value of startoffset that does not point to the start of a UTF-8 char-
1753 nigel 75 acter, is undefined. Your program may crash.
1754    
1755     PCRE_PARTIAL
1756    
1757 nigel 93 This option turns on the partial matching feature. If the subject
1758     string fails to match the pattern, but at some point during the match-
1759     ing process the end of the subject was reached (that is, the subject
1760     partially matches the pattern and the failure to match occurred only
1761     because there were not enough subject characters), pcre_exec() returns
1762     PCRE_ERROR_PARTIAL instead of PCRE_ERROR_NOMATCH. When PCRE_PARTIAL is
1763     used, there are restrictions on what may appear in the pattern. These
1764 nigel 75 are discussed in the pcrepartial documentation.
1765    
1766     The string to be matched by pcre_exec()
1767    
1768 nigel 93 The subject string is passed to pcre_exec() as a pointer in subject, a
1769     length in length, and a starting byte offset in startoffset. In UTF-8
1770     mode, the byte offset must point to the start of a UTF-8 character.
1771     Unlike the pattern string, the subject may contain binary zero bytes.
1772     When the starting offset is zero, the search for a match starts at the
1773 nigel 75 beginning of the subject, and this is by far the most common case.
1774 nigel 63
1775 nigel 93 A non-zero starting offset is useful when searching for another match
1776     in the same subject by calling pcre_exec() again after a previous suc-
1777     cess. Setting startoffset differs from just passing over a shortened
1778     string and setting PCRE_NOTBOL in the case of a pattern that begins
1779 nigel 73 with any kind of lookbehind. For example, consider the pattern
1780 nigel 41
1781 nigel 73 \Biss\B
1782 nigel 41
1783 nigel 93 which finds occurrences of "iss" in the middle of words. (\B matches
1784     only if the current position in the subject is not a word boundary.)
1785     When applied to the string "Mississipi" the first call to pcre_exec()
1786     finds the first occurrence. If pcre_exec() is called again with just
1787     the remainder of the subject, namely "issipi", it does not match,
1788 nigel 73 because \B is always false at the start of the subject, which is deemed
1789 nigel 93 to be a word boundary. However, if pcre_exec() is passed the entire
1790 nigel 75 string again, but with startoffset set to 4, it finds the second occur-
1791 nigel 93 rence of "iss" because it is able to look behind the starting point to
1792 nigel 75 discover that it is preceded by a letter.
1793 nigel 41
1794 nigel 93 If a non-zero starting offset is passed when the pattern is anchored,
1795 nigel 75 one attempt to match at the given offset is made. This can only succeed
1796 nigel 93 if the pattern does not require the match to be at the start of the
1797 nigel 75 subject.
1798 nigel 41
1799 nigel 75 How pcre_exec() returns captured substrings
1800    
1801 nigel 93 In general, a pattern matches a certain portion of the subject, and in
1802     addition, further substrings from the subject may be picked out by
1803     parts of the pattern. Following the usage in Jeffrey Friedl's book,
1804     this is called "capturing" in what follows, and the phrase "capturing
1805     subpattern" is used for a fragment of a pattern that picks out a sub-
1806     string. PCRE supports several other kinds of parenthesized subpattern
1807 nigel 73 that do not cause substrings to be captured.
1808 nigel 65
1809 nigel 93 Captured substrings are returned to the caller via a vector of integer
1810     offsets whose address is passed in ovector. The number of elements in
1811     the vector is passed in ovecsize, which must be a non-negative number.
1812 nigel 75 Note: this argument is NOT the size of ovector in bytes.
1813 nigel 41
1814 nigel 93 The first two-thirds of the vector is used to pass back captured sub-
1815     strings, each substring using a pair of integers. The remaining third
1816     of the vector is used as workspace by pcre_exec() while matching cap-
1817     turing subpatterns, and is not available for passing back information.
1818     The length passed in ovecsize should always be a multiple of three. If
1819 nigel 75 it is not, it is rounded down.
1820    
1821 nigel 93 When a match is successful, information about captured substrings is
1822     returned in pairs of integers, starting at the beginning of ovector,
1823     and continuing up to two-thirds of its length at the most. The first
1824 nigel 73 element of a pair is set to the offset of the first character in a sub-
1825 nigel 93 string, and the second is set to the offset of the first character
1826     after the end of a substring. The first pair, ovector[0] and ovec-
1827     tor[1], identify the portion of the subject string matched by the
1828     entire pattern. The next pair is used for the first capturing subpat-
1829 nigel 91 tern, and so on. The value returned by pcre_exec() is one more than the
1830     highest numbered pair that has been set. For example, if two substrings
1831 nigel 93 have been captured, the returned value is 3. If there are no capturing
1832     subpatterns, the return value from a successful match is 1, indicating
1833 nigel 91 that just the first pair of offsets has been set.
1834 nigel 41
1835 nigel 73 If a capturing subpattern is matched repeatedly, it is the last portion
1836 nigel 75 of the string that it matched that is returned.
1837 nigel 41
1838 nigel 93 If the vector is too small to hold all the captured substring offsets,
1839 nigel 75 it is used as far as possible (up to two-thirds of its length), and the
1840 nigel 93 function returns a value of zero. In particular, if the substring off-
1841 nigel 75 sets are not of interest, pcre_exec() may be called with ovector passed
1842 nigel 93 as NULL and ovecsize as zero. However, if the pattern contains back
1843     references and the ovector is not big enough to remember the related
1844     substrings, PCRE has to get additional memory for use during matching.
1845 nigel 73 Thus it is usually advisable to supply an ovector.
1846 nigel 41
1847 nigel 93 The pcre_info() function can be used to find out how many capturing
1848     subpatterns there are in a compiled pattern. The smallest size for
1849     ovector that will allow for n captured substrings, in addition to the
1850 nigel 91 offsets of the substring matched by the whole pattern, is (n+1)*3.
1851 nigel 41
1852 nigel 93 It is possible for capturing subpattern number n+1 to match some part
1853 nigel 91 of the subject when subpattern n has not been used at all. For example,
1854 nigel 93 if the string "abc" is matched against the pattern (a|(z))(bc) the
1855 nigel 91 return from the function is 4, and subpatterns 1 and 3 are matched, but
1856 nigel 93 2 is not. When this happens, both values in the offset pairs corre-
1857 nigel 91 sponding to unused subpatterns are set to -1.
1858 nigel 75
1859 nigel 93 Offset values that correspond to unused subpatterns at the end of the
1860     expression are also set to -1. For example, if the string "abc" is
1861     matched against the pattern (abc)(x(yz)?)? subpatterns 2 and 3 are not
1862     matched. The return from the function is 2, because the highest used
1863 nigel 91 capturing subpattern number is 1. However, you can refer to the offsets
1864 nigel 93 for the second and third capturing subpatterns if you wish (assuming
1865 nigel 91 the vector is large enough, of course).
1866    
1867 nigel 93 Some convenience functions are provided for extracting the captured
1868 nigel 91 substrings as separate strings. These are described below.
1869    
1870     Error return values from pcre_exec()
1871    
1872 nigel 93 If pcre_exec() fails, it returns a negative number. The following are
1873 nigel 73 defined in the header file:
1874 nigel 41
1875 nigel 73 PCRE_ERROR_NOMATCH (-1)
1876 nigel 41
1877 nigel 73 The subject string did not match the pattern.
1878 nigel 41
1879 nigel 73 PCRE_ERROR_NULL (-2)
1880 nigel 41
1881 nigel 93 Either code or subject was passed as NULL, or ovector was NULL and
1882 nigel 73 ovecsize was not zero.
1883 nigel 41
1884 nigel 73 PCRE_ERROR_BADOPTION (-3)
1885 nigel 41
1886 nigel 73 An unrecognized bit was set in the options argument.
1887 nigel 41
1888 nigel 73 PCRE_ERROR_BADMAGIC (-4)
1889 nigel 41
1890 nigel 93 PCRE stores a 4-byte "magic number" at the start of the compiled code,
1891 nigel 75 to catch the case when it is passed a junk pointer and to detect when a
1892     pattern that was compiled in an environment of one endianness is run in
1893 nigel 93 an environment with the other endianness. This is the error that PCRE
1894 nigel 75 gives when the magic number is not present.
1895 nigel 41
1896 nigel 93 PCRE_ERROR_UNKNOWN_OPCODE (-5)
1897 nigel 41
1898 nigel 73 While running the pattern match, an unknown item was encountered in the
1899 nigel 93 compiled pattern. This error could be caused by a bug in PCRE or by
1900 nigel 73 overwriting of the compiled pattern.
1901 nigel 41
1902 nigel 73 PCRE_ERROR_NOMEMORY (-6)
1903 nigel 41
1904 nigel 93 If a pattern contains back references, but the ovector that is passed
1905 nigel 73 to pcre_exec() is not big enough to remember the referenced substrings,
1906 nigel 93 PCRE gets a block of memory at the start of matching to use for this
1907     purpose. If the call via pcre_malloc() fails, this error is given. The
1908 nigel 75 memory is automatically freed at the end of matching.
1909 nigel 41
1910 nigel 73 PCRE_ERROR_NOSUBSTRING (-7)
1911 nigel 53
1912 nigel 93 This error is used by the pcre_copy_substring(), pcre_get_substring(),
1913 nigel 73 and pcre_get_substring_list() functions (see below). It is never
1914     returned by pcre_exec().
1915 nigel 63
1916 nigel 73 PCRE_ERROR_MATCHLIMIT (-8)
1917 nigel 63
1918 nigel 93 The backtracking limit, as specified by the match_limit field in a
1919     pcre_extra structure (or defaulted) was reached. See the description
1920 nigel 87 above.
1921    
1922 nigel 73 PCRE_ERROR_CALLOUT (-9)
1923 nigel 63
1924 nigel 73 This error is never generated by pcre_exec() itself. It is provided for
1925 nigel 93 use by callout functions that want to yield a distinctive error code.
1926 nigel 73 See the pcrecallout documentation for details.
1927 nigel 71
1928 nigel 73 PCRE_ERROR_BADUTF8 (-10)
1929 nigel 71
1930 nigel 93 A string that contains an invalid UTF-8 byte sequence was passed as a
1931 nigel 73 subject.
1932    
1933     PCRE_ERROR_BADUTF8_OFFSET (-11)
1934    
1935     The UTF-8 byte sequence that was passed as a subject was valid, but the
1936 nigel 93 value of startoffset did not point to the beginning of a UTF-8 charac-
1937 nigel 73 ter.
1938    
1939 nigel 77 PCRE_ERROR_PARTIAL (-12)
1940 nigel 73
1941 nigel 93 The subject string did not match, but it did match partially. See the
1942 nigel 75 pcrepartial documentation for details of partial matching.
1943    
1944 nigel 77 PCRE_ERROR_BADPARTIAL (-13)
1945 nigel 75
1946 nigel 93 The PCRE_PARTIAL option was used with a compiled pattern containing
1947     items that are not supported for partial matching. See the pcrepartial
1948 nigel 75 documentation for details of partial matching.
1949    
1950 nigel 77 PCRE_ERROR_INTERNAL (-14)
1951 nigel 75
1952 nigel 93 An unexpected internal error has occurred. This error could be caused
1953 nigel 75 by a bug in PCRE or by overwriting of the compiled pattern.
1954    
1955 nigel 77 PCRE_ERROR_BADCOUNT (-15)
1956 nigel 75
1957 nigel 93 This error is given if the value of the ovecsize argument is negative.
1958 nigel 75
1959 nigel 93 PCRE_ERROR_RECURSIONLIMIT (-21)
1960 nigel 75
1961 nigel 93 The internal recursion limit, as specified by the match_limit_recursion
1962     field in a pcre_extra structure (or defaulted) was reached. See the
1963     description above.
1964    
1965     PCRE_ERROR_NULLWSLIMIT (-22)
1966    
1967     When a group that can match an empty substring is repeated with an
1968     unbounded upper limit, the subject position at the start of the group
1969     must be remembered, so that a test for an empty string can be made when
1970     the end of the group is reached. Some workspace is required for this;
1971     if it runs out, this error is given.
1972    
1973     PCRE_ERROR_BADNEWLINE (-23)
1974    
1975     An invalid combination of PCRE_NEWLINE_xxx options was given.
1976    
1977     Error numbers -16 to -20 are not used by pcre_exec().
1978    
1979    
1980 nigel 63 EXTRACTING CAPTURED SUBSTRINGS BY NUMBER
1981    
1982 nigel 73 int pcre_copy_substring(const char *subject, int *ovector,
1983     int stringcount, int stringnumber, char *buffer,
1984     int buffersize);
1985 nigel 63
1986 nigel 73 int pcre_get_substring(const char *subject, int *ovector,
1987     int stringcount, int stringnumber,
1988     const char **stringptr);
1989 nigel 63
1990 nigel 73 int pcre_get_substring_list(const char *subject,
1991     int *ovector, int stringcount, const char ***listptr);
1992 nigel 63
1993 nigel 93 Captured substrings can be accessed directly by using the offsets
1994     returned by pcre_exec() in ovector. For convenience, the functions
1995 nigel 73 pcre_copy_substring(), pcre_get_substring(), and pcre_get_sub-
1996 nigel 93 string_list() are provided for extracting captured substrings as new,
1997     separate, zero-terminated strings. These functions identify substrings
1998     by number. The next section describes functions for extracting named
1999 nigel 91 substrings.
2000 nigel 41
2001 nigel 93 A substring that contains a binary zero is correctly extracted and has
2002     a further zero added on the end, but the result is not, of course, a C
2003     string. However, you can process such a string by referring to the
2004     length that is returned by pcre_copy_substring() and pcre_get_sub-
2005 nigel 91 string(). Unfortunately, the interface to pcre_get_substring_list() is
2006 nigel 93 not adequate for handling strings containing binary zeros, because the
2007 nigel 91 end of the final string is not independently indicated.
2008    
2009 nigel 93 The first three arguments are the same for all three of these func-
2010     tions: subject is the subject string that has just been successfully
2011 nigel 73 matched, ovector is a pointer to the vector of integer offsets that was
2012     passed to pcre_exec(), and stringcount is the number of substrings that
2013 nigel 93 were captured by the match, including the substring that matched the
2014 nigel 75 entire regular expression. This is the value returned by pcre_exec() if
2015 nigel 93 it is greater than zero. If pcre_exec() returned zero, indicating that
2016     it ran out of space in ovector, the value passed as stringcount should
2017 nigel 75 be the number of elements in the vector divided by three.
2018 nigel 41
2019 nigel 93 The functions pcre_copy_substring() and pcre_get_substring() extract a
2020     single substring, whose number is given as stringnumber. A value of
2021     zero extracts the substring that matched the entire pattern, whereas
2022     higher values extract the captured substrings. For pcre_copy_sub-
2023     string(), the string is placed in buffer, whose length is given by
2024     buffersize, while for pcre_get_substring() a new block of memory is
2025     obtained via pcre_malloc, and its address is returned via stringptr.
2026     The yield of the function is the length of the string, not including
2027     the terminating zero, or one of these error codes:
2028 nigel 41
2029 nigel 73 PCRE_ERROR_NOMEMORY (-6)
2030 nigel 41
2031 nigel 93 The buffer was too small for pcre_copy_substring(), or the attempt to
2032 nigel 73 get memory failed for pcre_get_substring().
2033 nigel 41
2034 nigel 73 PCRE_ERROR_NOSUBSTRING (-7)
2035 nigel 41
2036 nigel 73 There is no substring whose number is stringnumber.
2037 nigel 41
2038 nigel 93 The pcre_get_substring_list() function extracts all available sub-
2039     strings and builds a list of pointers to them. All this is done in a
2040 nigel 75 single block of memory that is obtained via pcre_malloc. The address of
2041 nigel 93 the memory block is returned via listptr, which is also the start of
2042     the list of string pointers. The end of the list is marked by a NULL
2043     pointer. The yield of the function is zero if all went well, or the
2044     error code
2045 nigel 41
2046 nigel 73 PCRE_ERROR_NOMEMORY (-6)
2047 nigel 41
2048 nigel 73 if the attempt to get the memory block failed.
2049 nigel 41
2050 nigel 91 When any of these functions encounter a substring that is unset, which
2051     can happen when capturing subpattern number n+1 matches some part of
2052     the subject, but subpattern n has not been used at all, they return an
2053 nigel 73 empty string. This can be distinguished from a genuine zero-length sub-
2054 nigel 91 string by inspecting the appropriate offset in ovector, which is nega-
2055 nigel 73 tive for unset substrings.
2056 nigel 41
2057 nigel 91 The two convenience functions pcre_free_substring() and pcre_free_sub-
2058     string_list() can be used to free the memory returned by a previous
2059 nigel 75 call of pcre_get_substring() or pcre_get_substring_list(), respec-
2060 nigel 91 tively. They do nothing more than call the function pointed to by
2061     pcre_free, which of course could be called directly from a C program.
2062     However, PCRE is used in some situations where it is linked via a spe-
2063     cial interface to another programming language that cannot use
2064     pcre_free directly; it is for these cases that the functions are pro-
2065 nigel 77 vided.
2066 nigel 41
2067 nigel 73
2068 nigel 63 EXTRACTING CAPTURED SUBSTRINGS BY NAME
2069 nigel 41
2070 nigel 75 int pcre_get_stringnumber(const pcre *code,
2071     const char *name);
2072    
2073 nigel 73 int pcre_copy_named_substring(const pcre *code,
2074     const char *subject, int *ovector,
2075     int stringcount, const char *stringname,
2076     char *buffer, int buffersize);
2077 nigel 41
2078 nigel 73 int pcre_get_named_substring(const pcre *code,
2079     const char *subject, int *ovector,
2080     int stringcount, const char *stringname,
2081     const char **stringptr);
2082 nigel 41
2083 nigel 91 To extract a substring by name, you first have to find associated num-
2084 nigel 75 ber. For example, for this pattern
2085 nigel 41
2086 nigel 93 (a+)b(?<xxx>\d+)...
2087 nigel 63
2088 nigel 91 the number of the subpattern called "xxx" is 2. If the name is known to
2089     be unique (PCRE_DUPNAMES was not set), you can find the number from the
2090     name by calling pcre_get_stringnumber(). The first argument is the com-
2091     piled pattern, and the second is the name. The yield of the function is
2092     the subpattern number, or PCRE_ERROR_NOSUBSTRING (-7) if there is no
2093     subpattern of that name.
2094 nigel 63
2095 nigel 75 Given the number, you can extract the substring directly, or use one of
2096     the functions described in the previous section. For convenience, there
2097     are also two functions that do the whole job.
2098    
2099 nigel 87 Most of the arguments of pcre_copy_named_substring() and
2100     pcre_get_named_substring() are the same as those for the similarly
2101     named functions that extract by number. As these are described in the
2102     previous section, they are not re-described here. There are just two
2103 nigel 75 differences:
2104 nigel 63
2105 nigel 87 First, instead of a substring number, a substring name is given. Sec-
2106 nigel 73 ond, there is an extra argument, given at the start, which is a pointer
2107 nigel 87 to the compiled pattern. This is needed in order to gain access to the
2108 nigel 73 name-to-number translation table.
2109 nigel 63
2110 nigel 87 These functions call pcre_get_stringnumber(), and if it succeeds, they
2111     then call pcre_copy_substring() or pcre_get_substring(), as appropri-
2112 nigel 73 ate.
2113 nigel 63
2114 nigel 77
2115 nigel 91 DUPLICATE SUBPATTERN NAMES
2116    
2117     int pcre_get_stringtable_entries(const pcre *code,
2118     const char *name, char **first, char **last);
2119    
2120     When a pattern is compiled with the PCRE_DUPNAMES option, names for
2121     subpatterns are not required to be unique. Normally, patterns with
2122     duplicate names are such that in any one match, only one of the named
2123     subpatterns participates. An example is shown in the pcrepattern docu-
2124     mentation. When duplicates are present, pcre_copy_named_substring() and
2125     pcre_get_named_substring() return the first substring corresponding to
2126     the given name that is set. If none are set, an empty string is
2127     returned. The pcre_get_stringnumber() function returns one of the num-
2128     bers that are associated with the name, but it is not defined which it
2129     is.
2130    
2131     If you want to get full details of all captured substrings for a given
2132     name, you must use the pcre_get_stringtable_entries() function. The
2133     first argument is the compiled pattern, and the second is the name. The
2134     third and fourth are pointers to variables which are updated by the
2135     function. After it has run, they point to the first and last entries in
2136     the name-to-number table for the given name. The function itself
2137 nigel 93 returns the length of each entry, or PCRE_ERROR_NOSUBSTRING (-7) if
2138     there are none. The format of the table is described above in the sec-
2139     tion entitled Information about a pattern. Given all the relevant
2140     entries for the name, you can extract each of their numbers, and hence
2141     the captured data, if any.
2142 nigel 91
2143    
2144 nigel 77 FINDING ALL POSSIBLE MATCHES
2145    
2146 nigel 91 The traditional matching function uses a similar algorithm to Perl,
2147 nigel 77 which stops when it finds the first match, starting at a given point in
2148 nigel 91 the subject. If you want to find all possible matches, or the longest
2149     possible match, consider using the alternative matching function (see
2150     below) instead. If you cannot use the alternative function, but still
2151     need to find all possible matches, you can kludge it up by making use
2152 nigel 77 of the callout facility, which is described in the pcrecallout documen-
2153     tation.
2154    
2155     What you have to do is to insert a callout right at the end of the pat-
2156 nigel 91 tern. When your callout function is called, extract and save the cur-
2157     rent matched substring. Then return 1, which forces pcre_exec() to
2158     backtrack and try other alternatives. Ultimately, when it runs out of
2159 nigel 77 matches, pcre_exec() will yield PCRE_ERROR_NOMATCH.
2160    
2161    
2162     MATCHING A PATTERN: THE ALTERNATIVE FUNCTION
2163    
2164     int pcre_dfa_exec(const pcre *code, const pcre_extra *extra,
2165     const char *subject, int length, int startoffset,
2166     int options, int *ovector, int ovecsize,
2167     int *workspace, int wscount);
2168    
2169 nigel 91 The function pcre_dfa_exec() is called to match a subject string
2170 nigel 93 against a compiled pattern, using a matching algorithm that scans the
2171     subject string just once, and does not backtrack. This has different
2172     characteristics to the normal algorithm, and is not compatible with
2173     Perl. Some of the features of PCRE patterns are not supported. Never-
2174     theless, there are times when this kind of matching can be useful. For
2175     a discussion of the two matching algorithms, see the pcrematching docu-
2176     mentation.
2177 nigel 77
2178 nigel 93 The arguments for the pcre_dfa_exec() function are the same as for
2179 nigel 77 pcre_exec(), plus two extras. The ovector argument is used in a differ-
2180 nigel 93 ent way, and this is described below. The other common arguments are
2181     used in the same way as for pcre_exec(), so their description is not
2182 nigel 77 repeated here.
2183    
2184 nigel 93 The two additional arguments provide workspace for the function. The
2185     workspace vector should contain at least 20 elements. It is used for
2186 nigel 77 keeping track of multiple paths through the pattern tree. More
2187 nigel 93 workspace will be needed for patterns and subjects where there are a
2188 nigel 91 lot of potential matches.
2189 nigel 77
2190 nigel 87 Here is an example of a simple call to pcre_dfa_exec():
2191 nigel 77
2192     int rc;
2193     int ovector[10];
2194     int wspace[20];
2195 nigel 87 rc = pcre_dfa_exec(
2196 nigel 77 re, /* result of pcre_compile() */
2197     NULL, /* we didn't study the pattern */
2198     "some string", /* the subject string */
2199     11, /* the length of the subject string */
2200     0, /* start at offset 0 in the subject */
2201     0, /* default options */
2202     ovector, /* vector of integers for substring information */
2203     10, /* number of elements (NOT size in bytes) */
2204     wspace, /* working space vector */
2205     20); /* number of elements (NOT size in bytes) */
2206    
2207     Option bits for pcre_dfa_exec()
2208    
2209 nigel 93 The unused bits of the options argument for pcre_dfa_exec() must be
2210     zero. The only bits that may be set are PCRE_ANCHORED, PCRE_NEW-
2211     LINE_xxx, PCRE_NOTBOL, PCRE_NOTEOL, PCRE_NOTEMPTY, PCRE_NO_UTF8_CHECK,
2212 nigel 91 PCRE_PARTIAL, PCRE_DFA_SHORTEST, and PCRE_DFA_RESTART. All but the last
2213     three of these are the same as for pcre_exec(), so their description is
2214     not repeated here.
2215 nigel 77
2216     PCRE_PARTIAL
2217    
2218 nigel 93 This has the same general effect as it does for pcre_exec(), but the
2219     details are slightly different. When PCRE_PARTIAL is set for
2220     pcre_dfa_exec(), the return code PCRE_ERROR_NOMATCH is converted into
2221     PCRE_ERROR_PARTIAL if the end of the subject is reached, there have
2222 nigel 77 been no complete matches, but there is still at least one matching pos-
2223 nigel 93 sibility. The portion of the string that provided the partial match is
2224 nigel 77 set as the first matching string.
2225    
2226     PCRE_DFA_SHORTEST
2227    
2228 nigel 93 Setting the PCRE_DFA_SHORTEST option causes the matching algorithm to
2229     stop as soon as it has found one match. Because of the way the alterna-
2230     tive algorithm works, this is necessarily the shortest possible match
2231     at the first possible matching point in the subject string.
2232 nigel 77
2233     PCRE_DFA_RESTART
2234    
2235 nigel 93 When pcre_dfa_exec() is called with the PCRE_PARTIAL option, and
2236     returns a partial match, it is possible to call it again, with addi-
2237     tional subject characters, and have it continue with the same match.
2238     The PCRE_DFA_RESTART option requests this action; when it is set, the
2239     workspace and wscount options must reference the same vector as before
2240     because data about the match so far is left in them after a partial
2241     match. There is more discussion of this facility in the pcrepartial
2242 nigel 77 documentation.
2243    
2244     Successful returns from pcre_dfa_exec()
2245    
2246 nigel 93 When pcre_dfa_exec() succeeds, it may have matched more than one sub-
2247 nigel 77 string in the subject. Note, however, that all the matches from one run
2248 nigel 93 of the function start at the same point in the subject. The shorter
2249     matches are all initial substrings of the longer matches. For example,
2250 nigel 77 if the pattern
2251    
2252     <.*>
2253    
2254     is matched against the string
2255    
2256     This is <something> <something else> <something further> no more
2257    
2258     the three matched strings are
2259    
2260     <something>
2261     <something> <something else>
2262     <something> <something else> <something further>
2263    
2264 nigel 93 On success, the yield of the function is a number greater than zero,
2265     which is the number of matched substrings. The substrings themselves
2266     are returned in ovector. Each string uses two elements; the first is
2267     the offset to the start, and the second is the offset to the end. In
2268     fact, all the strings have the same start offset. (Space could have
2269     been saved by giving this only once, but it was decided to retain some
2270     compatibility with the way pcre_exec() returns data, even though the
2271     meaning of the strings is different.)
2272 nigel 77
2273     The strings are returned in reverse order of length; that is, the long-
2274 nigel 93 est matching string is given first. If there were too many matches to
2275     fit into ovector, the yield of the function is zero, and the vector is
2276 nigel 77 filled with the longest matches.
2277    
2278     Error returns from pcre_dfa_exec()
2279    
2280 nigel 93 The pcre_dfa_exec() function returns a negative number when it fails.
2281     Many of the errors are the same as for pcre_exec(), and these are
2282     described above. There are in addition the following errors that are
2283 nigel 77 specific to pcre_dfa_exec():
2284    
2285     PCRE_ERROR_DFA_UITEM (-16)
2286    
2287 nigel 93 This return is given if pcre_dfa_exec() encounters an item in the pat-
2288     tern that it does not support, for instance, the use of \C or a back
2289 nigel 77 reference.
2290    
2291     PCRE_ERROR_DFA_UCOND (-17)
2292    
2293 nigel 93 This return is given if pcre_dfa_exec() encounters a condition item
2294     that uses a back reference for the condition, or a test for recursion
2295     in a specific group. These are not supported.
2296 nigel 77
2297     PCRE_ERROR_DFA_UMLIMIT (-18)
2298    
2299 nigel 93 This return is given if pcre_dfa_exec() is called with an extra block
2300 nigel 77 that contains a setting of the match_limit field. This is not supported
2301     (it is meaningless).
2302    
2303     PCRE_ERROR_DFA_WSSIZE (-19)
2304    
2305 nigel 93 This return is given if pcre_dfa_exec() runs out of space in the
2306 nigel 77 workspace vector.
2307    
2308     PCRE_ERROR_DFA_RECURSE (-20)
2309    
2310 nigel 93 When a recursive subpattern is processed, the matching function calls
2311     itself recursively, using private vectors for ovector and workspace.
2312     This error is given if the output vector is not large enough. This
2313 nigel 77 should be extremely rare, as a vector of size 1000 is used.
2314    
2315 nigel 93
2316     SEE ALSO
2317    
2318     pcrebuild(3), pcrecallout(3), pcrecpp(3)(3), pcrematching(3), pcrepar-
2319     tial(3), pcreposix(3), pcreprecompile(3), pcresample(3), pcrestack(3).
2320    
2321     Last updated: 30 November 2006
2322 nigel 87 Copyright (c) 1997-2006 University of Cambridge.
2323 nigel 79 ------------------------------------------------------------------------------
2324 nigel 63
2325    
2326 nigel 79 PCRECALLOUT(3) PCRECALLOUT(3)
2327 nigel 63
2328 nigel 79
2329 nigel 73 NAME
2330     PCRE - Perl-compatible regular expressions
2331    
2332 nigel 77
2333 nigel 63 PCRE CALLOUTS
2334    
2335 nigel 73 int (*pcre_callout)(pcre_callout_block *);
2336 nigel 63
2337 nigel 73 PCRE provides a feature called "callout", which is a means of temporar-
2338     ily passing control to the caller of PCRE in the middle of pattern
2339     matching. The caller of PCRE provides an external function by putting
2340     its entry point in the global variable pcre_callout. By default, this
2341     variable contains NULL, which disables all calling out.
2342 nigel 63
2343 nigel 73 Within a regular expression, (?C) indicates the points at which the
2344     external function is to be called. Different callout points can be
2345     identified by putting a number less than 256 after the letter C. The
2346     default value is zero. For example, this pattern has two callout
2347     points:
2348 nigel 63
2349 nigel 75 (?C1)eabc(?C2)def
2350 nigel 63
2351 nigel 75 If the PCRE_AUTO_CALLOUT option bit is set when pcre_compile() is
2352     called, PCRE automatically inserts callouts, all with number 255,
2353     before each item in the pattern. For example, if PCRE_AUTO_CALLOUT is
2354     used with the pattern
2355 nigel 63
2356 nigel 75 A(\d{2}|--)
2357    
2358     it is processed as if it were
2359    
2360     (?C255)A(?C255)((?C255)\d{2}(?C255)|(?C255)-(?C255)-(?C255))(?C255)
2361    
2362     Notice that there is a callout before and after each parenthesis and
2363     alternation bar. Automatic callouts can be used for tracking the
2364     progress of pattern matching. The pcretest command has an option that
2365     sets automatic callouts; when it is used, the output indicates how the
2366     pattern is matched. This is useful information when you are trying to
2367     optimize the performance of a particular pattern.
2368    
2369    
2370     MISSING CALLOUTS
2371    
2372     You should be aware that, because of optimizations in the way PCRE
2373     matches patterns, callouts sometimes do not happen. For example, if the
2374     pattern is
2375    
2376     ab(?C4)cd
2377    
2378     PCRE knows that any matching string must contain the letter "d". If the
2379     subject string is "abyz", the lack of "d" means that matching doesn't
2380     ever start, and the callout is never reached. However, with "abyd",
2381     though the result is still no match, the callout is obeyed.
2382    
2383    
2384     THE CALLOUT INTERFACE
2385    
2386     During matching, when PCRE reaches a callout point, the external func-
2387 nigel 77 tion defined by pcre_callout is called (if it is set). This applies to
2388     both the pcre_exec() and the pcre_dfa_exec() matching functions. The
2389     only argument to the callout function is a pointer to a pcre_callout
2390     block. This structure contains the following fields:
2391 nigel 75
2392 nigel 73 int version;
2393     int callout_number;
2394     int *offset_vector;
2395     const char *subject;
2396     int subject_length;
2397     int start_match;
2398     int current_position;
2399     int capture_top;
2400     int capture_last;
2401     void *callout_data;
2402 nigel 75 int pattern_position;
2403     int next_item_length;
2404 nigel 63
2405 nigel 77 The version field is an integer containing the version number of the
2406     block format. The initial version was 0; the current version is 1. The
2407     version number will change again in future if additional fields are
2408 nigel 75 added, but the intention is never to remove any of the existing fields.
2409 nigel 63
2410 nigel 77 The callout_number field contains the number of the callout, as com-
2411     piled into the pattern (that is, the number after ?C for manual call-
2412 nigel 75 outs, and 255 for automatically generated callouts).
2413 nigel 63
2414 nigel 77 The offset_vector field is a pointer to the vector of offsets that was
2415     passed by the caller to pcre_exec() or pcre_dfa_exec(). When
2416     pcre_exec() is used, the contents can be inspected in order to extract
2417     substrings that have been matched so far, in the same way as for
2418     extracting substrings after a match has completed. For pcre_dfa_exec()
2419     this field is not useful.
2420 nigel 63
2421 nigel 75 The subject and subject_length fields contain copies of the values that
2422 nigel 73 were passed to pcre_exec().
2423 nigel 63
2424 nigel 77 The start_match field contains the offset within the subject at which
2425     the current match attempt started. If the pattern is not anchored, the
2426 nigel 75 callout function may be called several times from the same point in the
2427     pattern for different starting points in the subject.
2428 nigel 63
2429 nigel 77 The current_position field contains the offset within the subject of
2430 nigel 73 the current match pointer.
2431 nigel 63
2432 nigel 77 When the pcre_exec() function is used, the capture_top field contains
2433     one more than the number of the highest numbered captured substring so
2434     far. If no substrings have been captured, the value of capture_top is
2435     one. This is always the case when pcre_dfa_exec() is used, because it
2436     does not support captured substrings.
2437 nigel 63
2438 nigel 77 The capture_last field contains the number of the most recently cap-
2439     tured substring. If no substrings have been captured, its value is -1.
2440     This is always the case when pcre_dfa_exec() is used.
2441 nigel 63
2442 nigel 77 The callout_data field contains a value that is passed to pcre_exec()
2443     or pcre_dfa_exec() specifically so that it can be passed back in call-
2444     outs. It is passed in the pcre_callout field of the pcre_extra data
2445     structure. If no such data was passed, the value of callout_data in a
2446     pcre_callout block is NULL. There is a description of the pcre_extra
2447 nigel 73 structure in the pcreapi documentation.
2448 nigel 63
2449 nigel 77 The pattern_position field is present from version 1 of the pcre_call-
2450 nigel 75 out structure. It contains the offset to the next item to be matched in
2451     the pattern string.
2452 nigel 63
2453 nigel 77 The next_item_length field is present from version 1 of the pcre_call-
2454 nigel 75 out structure. It contains the length of the next item to be matched in
2455 nigel 77 the pattern string. When the callout immediately precedes an alterna-
2456     tion bar, a closing parenthesis, or the end of the pattern, the length
2457     is zero. When the callout precedes an opening parenthesis, the length
2458 nigel 75 is that of the entire subpattern.
2459 nigel 73
2460 nigel 77 The pattern_position and next_item_length fields are intended to help
2461     in distinguishing between different automatic callouts, which all have
2462 nigel 75 the same callout number. However, they are set for all callouts.
2463    
2464    
2465 nigel 63 RETURN VALUES
2466    
2467 nigel 77 The external callout function returns an integer to PCRE. If the value
2468     is zero, matching proceeds as normal. If the value is greater than
2469     zero, matching fails at the current point, but the testing of other
2470     matching possibilities goes ahead, just as if a lookahead assertion had
2471     failed. If the value is less than zero, the match is abandoned, and
2472     pcre_exec() (or pcre_dfa_exec()) returns the negative value.
2473 nigel 63
2474 nigel 77 Negative values should normally be chosen from the set of
2475 nigel 73 PCRE_ERROR_xxx values. In particular, PCRE_ERROR_NOMATCH forces a stan-
2476 nigel 77 dard "no match" failure. The error number PCRE_ERROR_CALLOUT is
2477     reserved for use by callout functions; it will never be used by PCRE
2478 nigel 73 itself.
2479 nigel 63
2480 nigel 77 Last updated: 28 February 2005
2481     Copyright (c) 1997-2005 University of Cambridge.
2482 nigel 79 ------------------------------------------------------------------------------
2483 nigel 63
2484    
2485 nigel 79 PCRECOMPAT(3) PCRECOMPAT(3)
2486 nigel 63
2487 nigel 79
2488 nigel 73 NAME
2489     PCRE - Perl-compatible regular expressions
2490    
2491 nigel 77
2492 nigel 75 DIFFERENCES BETWEEN PCRE AND PERL
2493 nigel 41
2494 nigel 73 This document describes the differences in the ways that PCRE and Perl
2495 nigel 93 handle regular expressions. The differences described here are mainly
2496     with respect to Perl 5.8, though PCRE version 7.0 contains some fea-
2497     tures that are expected to be in the forthcoming Perl 5.10.
2498 nigel 41
2499 nigel 93 1. PCRE has only a subset of Perl's UTF-8 and Unicode support. Details
2500     of what it does have are given in the section on UTF-8 support in the
2501 nigel 87 main pcre page.
2502 nigel 41
2503 nigel 73 2. PCRE does not allow repeat quantifiers on lookahead assertions. Perl
2504 nigel 93 permits them, but they do not mean what you might think. For example,
2505 nigel 73 (?!a){3} does not assert that the next three characters are not "a". It
2506     just asserts that the next character is not "a" three times.
2507 nigel 41
2508 nigel 93 3. Capturing subpatterns that occur inside negative lookahead asser-
2509     tions are counted, but their entries in the offsets vector are never
2510     set. Perl sets its numerical variables from any such patterns that are
2511 nigel 73 matched before the assertion fails to match something (thereby succeed-
2512 nigel 93 ing), but only if the negative lookahead assertion contains just one
2513 nigel 73 branch.
2514 nigel 41
2515 nigel 93 4. Though binary zero characters are supported in the subject string,
2516 nigel 73 they are not allowed in a pattern string because it is passed as a nor-
2517 nigel 75 mal C string, terminated by zero. The escape sequence \0 can be used in
2518     the pattern to represent a binary zero.
2519 nigel 41
2520 nigel 93 5. The following Perl escape sequences are not supported: \l, \u, \L,
2521 nigel 75 \U, and \N. In fact these are implemented by Perl's general string-han-
2522 nigel 93 dling and are not part of its pattern matching engine. If any of these
2523 nigel 75 are encountered by PCRE, an error is generated.
2524 nigel 41
2525 nigel 93 6. The Perl escape sequences \p, \P, and \X are supported only if PCRE
2526     is built with Unicode character property support. The properties that
2527     can be tested with \p and \P are limited to the general category prop-
2528     erties such as Lu and Nd, script names such as Greek or Han, and the
2529 nigel 87 derived properties Any and L&.
2530 nigel 75
2531     7. PCRE does support the \Q...\E escape for quoting substrings. Charac-
2532 nigel 93 ters in between are treated as literals. This is slightly different
2533     from Perl in that $ and @ are also handled as literals inside the
2534     quotes. In Perl, they cause variable interpolation (but of course PCRE
2535 nigel 73 does not have variables). Note the following examples:
2536 nigel 49
2537 nigel 73 Pattern PCRE matches Perl matches
2538 nigel 41
2539 nigel 73 \Qabc$xyz\E abc$xyz abc followed by the
2540     contents of $xyz
2541     \Qabc\$xyz\E abc\$xyz abc\$xyz
2542     \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
2543 nigel 41
2544 nigel 93 The \Q...\E sequence is recognized both inside and outside character
2545 nigel 73 classes.
2546 nigel 41
2547 nigel 93 8. Fairly obviously, PCRE does not support the (?{code}) and (??{code})
2548     constructions. However, there is support for recursive patterns. This
2549     is not available in Perl 5.8, but will be in Perl 5.10. Also, the PCRE
2550     "callout" feature allows an external function to be called during pat-
2551 nigel 75 tern matching. See the pcrecallout documentation for details.
2552 nigel 63
2553 nigel 93 9. Subpatterns that are called recursively or as "subroutines" are
2554     always treated as atomic groups in PCRE. This is like Python, but
2555     unlike Perl.
2556    
2557     10. There are some differences that are concerned with the settings of
2558     captured strings when part of a pattern is repeated. For example,
2559     matching "aba" against the pattern /^(a(b)?)+$/ in Perl leaves $2
2560 nigel 73 unset, but in PCRE it is set to "b".
2561 nigel 41
2562 nigel 93 11. PCRE provides some extensions to the Perl regular expression facil-
2563     ities. Perl 5.10 will include new features that are not in earlier
2564     versions, some of which (such as named parentheses) have been in PCRE
2565     for some time. This list is with respect to Perl 5.10:
2566 nigel 41
2567 nigel 93 (a) Although lookbehind assertions must match fixed length strings,
2568 nigel 73 each alternative branch of a lookbehind assertion can match a different
2569     length of string. Perl requires them all to have the same length.
2570 nigel 41
2571 nigel 93 (b) If PCRE_DOLLAR_ENDONLY is set and PCRE_MULTILINE is not set, the $
2572 nigel 73 meta-character matches only at the very end of the string.
2573 nigel 41
2574 nigel 73 (c) If PCRE_EXTRA is set, a backslash followed by a letter with no spe-
2575 nigel 93 cial meaning is faulted. Otherwise, like Perl, the backslash is
2576 nigel 91 ignored. (Perl can be made to issue a warning.)
2577 nigel 41
2578 nigel 93 (d) If PCRE_UNGREEDY is set, the greediness of the repetition quanti-
2579 nigel 73 fiers is inverted, that is, by default they are not greedy, but if fol-
2580     lowed by a question mark they are.
2581 nigel 41
2582 nigel 75 (e) PCRE_ANCHORED can be used at matching time to force a pattern to be
2583     tried only at the first matching position in the subject string.
2584 nigel 41
2585 nigel 93 (f) The PCRE_NOTBOL, PCRE_NOTEOL, PCRE_NOTEMPTY, and PCRE_NO_AUTO_CAP-
2586 nigel 73 TURE options for pcre_exec() have no Perl equivalents.
2587 nigel 41
2588 nigel 93 (g) The callout facility is PCRE-specific.
2589 nigel 41
2590 nigel 93 (h) The partial matching facility is PCRE-specific.
2591 nigel 43
2592 nigel 93 (i) Patterns compiled by PCRE can be saved and re-used at a later time,
2593 nigel 75 even on different hosts that have the other endianness.
2594    
2595 nigel 93 (j) The alternative matching function (pcre_dfa_exec()) matches in a
2596 nigel 77 different way and is not Perl-compatible.
2597    
2598 nigel 93 Last updated: 28 November 2006
2599 nigel 87 Copyright (c) 1997-2006 University of Cambridge.
2600 nigel 79 ------------------------------------------------------------------------------
2601 nigel 63
2602    
2603 nigel 79 PCREPATTERN(3) PCREPATTERN(3)
2604 nigel 63
2605 nigel 79
2606 nigel 73 NAME
2607     PCRE - Perl-compatible regular expressions
2608    
2609 nigel 77
2610 nigel 63 PCRE REGULAR EXPRESSION DETAILS
2611    
2612 nigel 73 The syntax and semantics of the regular expressions supported by PCRE
2613     are described below. Regular expressions are also described in the Perl
2614 nigel 75 documentation and in a number of books, some of which have copious
2615     examples. Jeffrey Friedl's "Mastering Regular Expressions", published
2616     by O'Reilly, covers regular expressions in great detail. This descrip-
2617     tion of PCRE's regular expressions is intended as reference material.
2618 nigel 49
2619 nigel 75 The original operation of PCRE was on strings of one-byte characters.
2620     However, there is now also support for UTF-8 character strings. To use
2621     this, you must build PCRE to include UTF-8 support, and then call
2622     pcre_compile() with the PCRE_UTF8 option. How this affects pattern
2623     matching is mentioned in several places below. There is also a summary
2624     of UTF-8 features in the section on UTF-8 support in the main pcre
2625     page.
2626 nigel 41
2627 nigel 77 The remainder of this document discusses the patterns that are sup-
2628     ported by PCRE when its main matching function, pcre_exec(), is used.
2629     From release 6.0, PCRE offers a second matching function,
2630     pcre_dfa_exec(), which matches using a different algorithm that is not
2631     Perl-compatible. The advantages and disadvantages of the alternative
2632     function, and how it differs from the normal function, are discussed in
2633     the pcrematching page.
2634    
2635 nigel 93
2636     CHARACTERS AND METACHARACTERS
2637    
2638 nigel 75 A regular expression is a pattern that is matched against a subject
2639     string from left to right. Most characters stand for themselves in a
2640     pattern, and match the corresponding characters in the subject. As a
2641 nigel 73 trivial example, the pattern
2642 nigel 41
2643 nigel 73 The quick brown fox
2644 nigel 41
2645 nigel 77 matches a portion of a subject string that is identical to itself. When
2646     caseless matching is specified (the PCRE_CASELESS option), letters are
2647     matched independently of case. In UTF-8 mode, PCRE always understands
2648     the concept of case for characters whose values are less than 128, so
2649     caseless matching is always possible. For characters with higher val-
2650     ues, the concept of case is supported if PCRE is compiled with Unicode
2651     property support, but not otherwise. If you want to use caseless
2652     matching for characters 128 and above, you must ensure that PCRE is
2653     compiled with Unicode property support as well as with UTF-8 support.
2654 nigel 41
2655 nigel 77 The power of regular expressions comes from the ability to include
2656     alternatives and repetitions in the pattern. These are encoded in the
2657     pattern by the use of metacharacters, which do not stand for themselves
2658     but instead are interpreted in some special way.
2659    
2660     There are two different sets of metacharacters: those that are recog-
2661     nized anywhere in the pattern except within square brackets, and those
2662 nigel 93 that are recognized within square brackets. Outside square brackets,
2663     the metacharacters are as follows:
2664 nigel 41
2665 nigel 73 \ general escape character with several uses
2666     ^ assert start of string (or line, in multiline mode)
2667     $ assert end of string (or line, in multiline mode)
2668     . match any character except newline (by default)
2669     [ start character class definition
2670     | start of alternative branch
2671     ( start subpattern
2672     ) end subpattern
2673     ? extends the meaning of (
2674     also 0 or 1 quantifier
2675     also quantifier minimizer
2676     * 0 or more quantifier
2677     + 1 or more quantifier
2678     also "possessive quantifier"
2679     { start min/max quantifier
2680 nigel 41
2681 nigel 77 Part of a pattern that is in square brackets is called a "character
2682 nigel 75 class". In a character class the only metacharacters are:
2683 nigel 41
2684 nigel 73 \ general escape character
2685     ^ negate the class, but only if the first character
2686     - indicates character range
2687     [ POSIX character class (only if followed by POSIX
2688     syntax)
2689     ] terminates the character class
2690 nigel 41
2691 nigel 77 The following sections describe the use of each of the metacharacters.
2692 nigel 41
2693    
2694 nigel 63 BACKSLASH
2695 nigel 41
2696 nigel 73 The backslash character has several uses. Firstly, if it is followed by
2697 nigel 77 a non-alphanumeric character, it takes away any special meaning that
2698     character may have. This use of backslash as an escape character
2699 nigel 73 applies both inside and outside character classes.
2700 nigel 41
2701 nigel 77 For example, if you want to match a * character, you write \* in the
2702     pattern. This escaping action applies whether or not the following
2703     character would otherwise be interpreted as a metacharacter, so it is
2704     always safe to precede a non-alphanumeric with backslash to specify
2705     that it stands for itself. In particular, if you want to match a back-
2706 nigel 75 slash, you write \\.
2707 nigel 41
2708 nigel 77 If a pattern is compiled with the PCRE_EXTENDED option, whitespace in
2709     the pattern (other than in a character class) and characters between a
2710 nigel 91 # outside a character class and the next newline are ignored. An escap-
2711     ing backslash can be used to include a whitespace or # character as
2712     part of the pattern.
2713 nigel 41
2714 nigel 77 If you want to remove the special meaning from a sequence of charac-
2715     ters, you can do so by putting them between \Q and \E. This is differ-
2716     ent from Perl in that $ and @ are handled as literals in \Q...\E
2717     sequences in PCRE, whereas in Perl, $ and @ cause variable interpola-
2718 nigel 73 tion. Note the following examples:
2719 nigel 63
2720 nigel 73 Pattern PCRE matches Perl matches
2721 nigel 63
2722 nigel 73 \Qabc$xyz\E abc$xyz abc followed by the
2723     contents of $xyz
2724     \Qabc\$xyz\E abc\$xyz abc\$xyz
2725     \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
2726 nigel 63
2727 nigel 77 The \Q...\E sequence is recognized both inside and outside character
2728 nigel 73 classes.
2729 nigel 63
2730 nigel 75 Non-printing characters
2731    
2732 nigel 73 A second use of backslash provides a way of encoding non-printing char-
2733 nigel 77 acters in patterns in a visible manner. There is no restriction on the
2734     appearance of non-printing characters, apart from the binary zero that
2735     terminates a pattern, but when a pattern is being prepared by text
2736     editing, it is usually easier to use one of the following escape
2737 nigel 73 sequences than the binary character it represents:
2738 nigel 63
2739 nigel 73 \a alarm, that is, the BEL character (hex 07)
2740     \cx "control-x", where x is any character
2741     \e escape (hex 1B)
2742     \f formfeed (hex 0C)
2743     \n newline (hex 0A)
2744     \r carriage return (hex 0D)
2745     \t tab (hex 09)
2746     \ddd character with octal code ddd, or backreference
2747     \xhh character with hex code hh
2748 nigel 87 \x{hhh..} character with hex code hhh..
2749 nigel 41
2750 nigel 77 The precise effect of \cx is as follows: if x is a lower case letter,
2751     it is converted to upper case. Then bit 6 of the character (hex 40) is
2752     inverted. Thus \cz becomes hex 1A, but \c{ becomes hex 3B, while \c;
2753 nigel 73 becomes hex 7B.
2754 nigel 41
2755 nigel 77 After \x, from zero to two hexadecimal digits are read (letters can be
2756 nigel 87 in upper or lower case). Any number of hexadecimal digits may appear
2757     between \x{ and }, but the value of the character code must be less
2758     than 256 in non-UTF-8 mode, and less than 2**31 in UTF-8 mode (that is,
2759     the maximum hexadecimal value is 7FFFFFFF). If characters other than
2760     hexadecimal digits appear between \x{ and }, or if there is no termi-
2761     nating }, this form of escape is not recognized. Instead, the initial
2762     \x will be interpreted as a basic hexadecimal escape, with no following
2763     digits, giving a character whose value is zero.
2764 nigel 41
2765 nigel 73 Characters whose value is less than 256 can be defined by either of the
2766 nigel 87 two syntaxes for \x. There is no difference in the way they are han-
2767     dled. For example, \xdc is exactly the same as \x{dc}.
2768 nigel 41
2769 nigel 91 After \0 up to two further octal digits are read. If there are fewer
2770     than two digits, just those that are present are used. Thus the
2771     sequence \0\x\07 specifies two binary zeros followed by a BEL character
2772     (code value 7). Make sure you supply two digits after the initial zero
2773     if the pattern character that follows is itself an octal digit.
2774 nigel 63
2775 nigel 73 The handling of a backslash followed by a digit other than 0 is compli-
2776     cated. Outside a character class, PCRE reads it and any following dig-
2777 nigel 91 its as a decimal number. If the number is less than 10, or if there
2778 nigel 73 have been at least that many previous capturing left parentheses in the
2779 nigel 91 expression, the entire sequence is taken as a back reference. A
2780     description of how this works is given later, following the discussion
2781 nigel 73 of parenthesized subpatterns.
2782 nigel 41
2783 nigel 91 Inside a character class, or if the decimal number is greater than 9
2784     and there have not been that many capturing subpatterns, PCRE re-reads
2785 nigel 93 up to three octal digits following the backslash, and uses them to gen-
2786 nigel 91 erate a data character. Any subsequent digits stand for themselves. In
2787     non-UTF-8 mode, the value of a character specified in octal must be
2788     less than \400. In UTF-8 mode, values up to \777 are permitted. For
2789     example:
2790 nigel 41
2791 nigel 73 \040 is another way of writing a space
2792     \40 is the same, provided there are fewer than 40
2793     previous capturing subpatterns
2794     \7 is always a back reference
2795     \11 might be a back reference, or another way of
2796     writing a tab
2797     \011 is always a tab
2798     \0113 is a tab followed by the character "3"
2799     \113 might be a back reference, otherwise the
2800     character with octal code 113
2801     \377 might be a back reference, otherwise
2802     the byte consisting entirely of 1 bits
2803     \81 is either a back reference, or a binary zero
2804     followed by the two characters "8" and "1"
2805 nigel 41
2806 nigel 91 Note that octal values of 100 or greater must not be introduced by a
2807 nigel 73 leading zero, because no more than three octal digits are ever read.
2808 nigel 41
2809 nigel 91 All the sequences that define a single character value can be used both
2810     inside and outside character classes. In addition, inside a character
2811     class, the sequence \b is interpreted as the backspace character (hex
2812 nigel 93 08), and the sequences \R and \X are interpreted as the characters "R"
2813     and "X", respectively. Outside a character class, these sequences have
2814     different meanings (see below).
2815 nigel 43
2816 nigel 93 Absolute and relative back references
2817    
2818     The sequence \g followed by a positive or negative number, optionally
2819     enclosed in braces, is an absolute or relative back reference. Back
2820     references are discussed later, following the discussion of parenthe-
2821     sized subpatterns.
2822    
2823 nigel 75 Generic character types
2824 nigel 41
2825 nigel 93 Another use of backslash is for specifying generic character types. The
2826     following are always recognized:
2827 nigel 75
2828 nigel 73 \d any decimal digit
2829     \D any character that is not a decimal digit
2830     \s any whitespace character
2831     \S any character that is not a whitespace character
2832     \w any "word" character
2833     \W any "non-word" character
2834 nigel 41
2835 nigel 73 Each pair of escape sequences partitions the complete set of characters
2836 nigel 87 into two disjoint sets. Any given character matches one, and only one,
2837 nigel 73 of each pair.
2838 nigel 41
2839 nigel 75 These character type sequences can appear both inside and outside char-
2840 nigel 87 acter classes. They each match one character of the appropriate type.
2841     If the current matching point is at the end of the subject string, all
2842 nigel 75 of them fail, since there is no character to match.
2843 nigel 41
2844 nigel 87 For compatibility with Perl, \s does not match the VT character (code
2845     11). This makes it different from the the POSIX "space" class. The \s
2846 nigel 91 characters are HT (9), LF (10), FF (12), CR (13), and space (32). (If
2847     "use locale;" is included in a Perl script, \s may match the VT charac-
2848     ter. In PCRE, it never does.)
2849 nigel 63
2850 nigel 75 A "word" character is an underscore or any character less than 256 that
2851 nigel 87 is a letter or digit. The definition of letters and digits is con-
2852     trolled by PCRE's low-valued character tables, and may vary if locale-
2853     specific matching is taking place (see "Locale support" in the pcreapi
2854     page). For example, in the "fr_FR" (French) locale, some character
2855     codes greater than 128 are used for accented letters, and these are
2856 nigel 75 matched by \w.
2857 nigel 63
2858 nigel 87 In UTF-8 mode, characters with values greater than 128 never match \d,
2859 nigel 75 \s, or \w, and always match \D, \S, and \W. This is true even when Uni-
2860 nigel 87 code character property support is available. The use of locales with
2861     Unicode is discouraged.
2862 nigel 41
2863 nigel 93 Newline sequences
2864    
2865     Outside a character class, the escape sequence \R matches any Unicode
2866     newline sequence. This is an extension to Perl. In non-UTF-8 mode \R is
2867     equivalent to the following:
2868    
2869     (?>\r\n|\n|\x0b|\f|\r|\x85)
2870    
2871     This is an example of an "atomic group", details of which are given
2872     below. This particular group matches either the two-character sequence
2873     CR followed by LF, or one of the single characters LF (linefeed,
2874     U+000A), VT (vertical tab, U+000B), FF (formfeed, U+000C), CR (carriage
2875     return, U+000D), or NEL (next line, U+0085). The two-character sequence
2876     is treated as a single unit that cannot be split.
2877    
2878     In UTF-8 mode, two additional characters whose codepoints are greater
2879     than 255 are added: LS (line separator, U+2028) and PS (paragraph sepa-
2880     rator, U+2029). Unicode character property support is not needed for
2881     these characters to be recognized.
2882    
2883     Inside a character class, \R matches the letter "R".
2884    
2885 nigel 75 Unicode character properties
2886    
2887     When PCRE is built with Unicode character property support, three addi-
2888 nigel 87 tional escape sequences to match character properties are available
2889 nigel 75 when UTF-8 mode is selected. They are:
2890    
2891 nigel 87 \p{xx} a character with the xx property
2892     \P{xx} a character without the xx property
2893     \X an extended Unicode sequence
2894 nigel 75
2895 nigel 77 The property names represented by xx above are limited to the Unicode
2896 nigel 87 script names, the general category properties, and "Any", which matches
2897     any character (including newline). Other properties such as "InMusical-
2898     Symbols" are not currently supported by PCRE. Note that \P{Any} does
2899     not match any characters, so always causes a match failure.
2900 nigel 75
2901 nigel 87 Sets of Unicode characters are defined as belonging to certain scripts.
2902     A character from one of these sets can be matched using a script name.
2903     For example:
2904 nigel 75
2905 nigel 87 \p{Greek}
2906     \P{Han}
2907    
2908     Those that are not part of an identified script are lumped together as
2909     "Common". The current list of scripts is:
2910    
2911 nigel 93 Arabic, Armenian, Balinese, Bengali, Bopomofo, Braille, Buginese,
2912     Buhid, Canadian_Aboriginal, Cherokee, Common, Coptic, Cuneiform,
2913     Cypriot, Cyrillic, Deseret, Devanagari, Ethiopic, Georgian, Glagolitic,
2914     Gothic, Greek, Gujarati, Gurmukhi, Han, Hangul, Hanunoo, Hebrew, Hira-
2915     gana, Inherited, Kannada, Katakana, Kharoshthi, Khmer, Lao, Latin,
2916     Limbu, Linear_B, Malayalam, Mongolian, Myanmar, New_Tai_Lue, Nko,
2917     Ogham, Old_Italic, Old_Persian, Oriya, Osmanya, Phags_Pa, Phoenician,
2918     Runic, Shavian, Sinhala, Syloti_Nagri, Syriac, Tagalog, Tagbanwa,
2919     Tai_Le, Tamil, Telugu, Thaana, Thai, Tibetan, Tifinagh, Ugaritic, Yi.
2920 nigel 87
2921     Each character has exactly one general category property, specified by
2922     a two-letter abbreviation. For compatibility with Perl, negation can be
2923     specified by including a circumflex between the opening brace and the
2924     property name. For example, \p{^Lu} is the same as \P{Lu}.
2925    
2926     If only one letter is specified with \p or \P, it includes all the gen-
2927     eral category properties that start with that letter. In this case, in
2928     the absence of negation, the curly brackets in the escape sequence are
2929     optional; these two examples have the same effect:
2930    
2931 nigel 75 \p{L}
2932     \pL
2933    
2934 nigel 87 The following general category property codes are supported:
2935 nigel 75
2936     C Other
2937     Cc Control
2938     Cf Format
2939     Cn Unassigned
2940     Co Private use
2941     Cs Surrogate
2942    
2943     L Letter
2944     Ll Lower case letter
2945     Lm Modifier letter
2946     Lo Other letter
2947     Lt Title case letter
2948     Lu Upper case letter
2949    
2950     M Mark
2951     Mc Spacing mark
2952     Me Enclosing mark
2953     Mn Non-spacing mark
2954    
2955     N Number
2956     Nd Decimal number
2957     Nl Letter number
2958     No Other number
2959    
2960     P Punctuation
2961     Pc Connector punctuation
2962     Pd Dash punctuation
2963     Pe Close punctuation
2964     Pf Final punctuation
2965     Pi Initial punctuation
2966     Po Other punctuation
2967     Ps Open punctuation
2968    
2969     S Symbol
2970     Sc Currency symbol
2971     Sk Modifier symbol
2972     Sm Mathematical symbol
2973     So Other symbol
2974    
2975     Z Separator
2976     Zl Line separator
2977     Zp Paragraph separator
2978     Zs Space separator
2979    
2980 nigel 87 The special property L& is also supported: it matches a character that
2981     has the Lu, Ll, or Lt property, in other words, a letter that is not
2982     classified as a modifier or "other".
2983 nigel 75
2984 nigel 87 The long synonyms for these properties that Perl supports (such as
2985 nigel 91 \p{Letter}) are not supported by PCRE, nor is it permitted to prefix
2986 nigel 87 any of these properties with "Is".
2987    
2988     No character that is in the Unicode table has the Cn (unassigned) prop-
2989     erty. Instead, this property is assumed for any code point that is not
2990     in the Unicode table.
2991    
2992 nigel 77 Specifying caseless matching does not affect these escape sequences.
2993 nigel 75 For example, \p{Lu} always matches only upper case letters.
2994    
2995 nigel 77 The \X escape matches any number of Unicode characters that form an
2996 nigel 75 extended Unicode sequence. \X is equivalent to
2997    
2998     (?>\PM\pM*)
2999    
3000 nigel 77 That is, it matches a character without the "mark" property, followed
3001     by zero or more characters with the "mark" property, and treats the
3002     sequence as an atomic group (see below). Characters with the "mark"
3003 nigel 75 property are typically accents that affect the preceding character.
3004    
3005 nigel 77 Matching characters by Unicode property is not fast, because PCRE has
3006     to search a structure that contains data for over fifteen thousand
3007 nigel 75 characters. That is why the traditional escape sequences such as \d and
3008     \w do not use Unicode properties in PCRE.
3009    
3010     Simple assertions
3011    
3012 nigel 93 The final use of backslash is for certain simple assertions. An asser-
3013 nigel 77 tion specifies a condition that has to be met at a particular point in
3014     a match, without consuming any characters from the subject string. The
3015     use of subpatterns for more complicated assertions is described below.
3016 nigel 75 The backslashed assertions are:
3017 nigel 41
3018 nigel 73 \b matches at a word boundary
3019     \B matches when not at a word boundary
3020 nigel 93 \A matches at the start of the subject
3021     \Z matches at the end of the subject
3022     also matches before a newline at the end of the subject
3023     \z matches only at the end of the subject
3024     \G matches at the first matching position in the subject
3025 nigel 41
3026 nigel 77 These assertions may not appear in character classes (but note that \b
3027 nigel 73 has a different meaning, namely the backspace character, inside a char-
3028     acter class).
3029 nigel 41
3030 nigel 77 A word boundary is a position in the subject string where the current
3031     character and the previous character do not both match \w or \W (i.e.
3032     one matches \w and the other matches \W), or the start or end of the
3033 nigel 73 string if the first or last character matches \w, respectively.
3034 nigel 43
3035 nigel 77 The \A, \Z, and \z assertions differ from the traditional circumflex
3036 nigel 75 and dollar (described in the next section) in that they only ever match
3037 nigel 77 at the very start and end of the subject string, whatever options are
3038     set. Thus, they are independent of multiline mode. These three asser-
3039 nigel 75 tions are not affected by the PCRE_NOTBOL or PCRE_NOTEOL options, which
3040 nigel 77 affect only the behaviour of the circumflex and dollar metacharacters.
3041     However, if the startoffset argument of pcre_exec() is non-zero, indi-
3042 nigel 75 cating that matching is to start at a point other than the beginning of
3043 nigel 77 the subject, \A can never match. The difference between \Z and \z is
3044 nigel 91 that \Z matches before a newline at the end of the string as well as at
3045     the very end, whereas \z matches only at the end.
3046 nigel 63
3047 nigel 91 The \G assertion is true only when the current matching position is at
3048     the start point of the match, as specified by the startoffset argument
3049     of pcre_exec(). It differs from \A when the value of startoffset is
3050     non-zero. By calling pcre_exec() multiple times with appropriate argu-
3051 nigel 73 ments, you can mimic Perl's /g option, and it is in this kind of imple-
3052     mentation where \G can be useful.
3053 nigel 41
3054 nigel 91 Note, however, that PCRE's interpretation of \G, as the start of the
3055 nigel 73 current match, is subtly different from Perl's, which defines it as the
3056 nigel 91 end of the previous match. In Perl, these can be different when the
3057     previously matched string was empty. Because PCRE does just one match
3058 nigel 73 at a time, it cannot reproduce this behaviour.
3059 nigel 41
3060 nigel 91 If all the alternatives of a pattern begin with \G, the expression is
3061 nigel 73 anchored to the starting match position, and the "anchored" flag is set
3062     in the compiled regular expression.
3063 nigel 63
3064    
3065 nigel 41 CIRCUMFLEX AND DOLLAR
3066 nigel 63
3067 nigel 73 Outside a character class, in the default matching mode, the circumflex
3068 nigel 91 character is an assertion that is true only if the current matching
3069     point is at the start of the subject string. If the startoffset argu-
3070     ment of pcre_exec() is non-zero, circumflex can never match if the
3071     PCRE_MULTILINE option is unset. Inside a character class, circumflex
3072 nigel 73 has an entirely different meaning (see below).
3073 nigel 41
3074 nigel 91 Circumflex need not be the first character of the pattern if a number
3075     of alternatives are involved, but it should be the first thing in each
3076     alternative in which it appears if the pattern is ever to match that
3077     branch. If all possible alternatives start with a circumflex, that is,
3078     if the pattern is constrained to match only at the start of the sub-
3079     ject, it is said to be an "anchored" pattern. (There are also other
3080 nigel 73 constructs that can cause a pattern to be anchored.)
3081 nigel 41
3082 nigel 91 A dollar character is an assertion that is true only if the current
3083     matching point is at the end of the subject string, or immediately
3084     before a newline at the end of the string (by default). Dollar need not
3085     be the last character of the pattern if a number of alternatives are
3086     involved, but it should be the last item in any branch in which it
3087     appears. Dollar has no special meaning in a character class.
3088 nigel 41
3089 nigel 77 The meaning of dollar can be changed so that it matches only at the
3090     very end of the string, by setting the PCRE_DOLLAR_ENDONLY option at
3091 nigel 73 compile time. This does not affect the \Z assertion.
3092 nigel 41
3093 nigel 73 The meanings of the circumflex and dollar characters are changed if the
3094 nigel 91 PCRE_MULTILINE option is set. When this is the case, a circumflex
3095     matches immediately after internal newlines as well as at the start of
3096     the subject string. It does not match after a newline that ends the
3097     string. A dollar matches before any newlines in the string, as well as
3098     at the very end, when PCRE_MULTILINE is set. When newline is specified
3099     as the two-character sequence CRLF, isolated CR and LF characters do
3100     not indicate newlines.
3101 nigel 41
3102 nigel 91 For example, the pattern /^abc$/ matches the subject string "def\nabc"
3103     (where \n represents a newline) in multiline mode, but not otherwise.
3104     Consequently, patterns that are anchored in single line mode because
3105     all branches start with ^ are not anchored in multiline mode, and a
3106     match for circumflex is possible when the startoffset argument of
3107     pcre_exec() is non-zero. The PCRE_DOLLAR_ENDONLY option is ignored if
3108     PCRE_MULTILINE is set.
3109 nigel 41
3110 nigel 91 Note that the sequences \A, \Z, and \z can be used to match the start
3111     and end of the subject in both modes, and if all branches of a pattern
3112     start with \A it is always anchored, whether or not PCRE_MULTILINE is
3113     set.
3114 nigel 41
3115 nigel 91
3116 nigel 63 FULL STOP (PERIOD, DOT)
3117 nigel 41
3118 nigel 73 Outside a character class, a dot in the pattern matches any one charac-
3119 nigel 91 ter in the subject string except (by default) a character that signi-
3120     fies the end of a line. In UTF-8 mode, the matched character may be
3121 nigel 93 more than one byte long.
3122 nigel 41
3123 nigel 93 When a line ending is defined as a single character, dot never matches
3124     that character; when the two-character sequence CRLF is used, dot does
3125     not match CR if it is immediately followed by LF, but otherwise it
3126     matches all characters (including isolated CRs and LFs). When any Uni-
3127     code line endings are being recognized, dot does not match CR or LF or
3128     any of the other line ending characters.
3129 nigel 41
3130 nigel 93 The behaviour of dot with regard to newlines can be changed. If the
3131     PCRE_DOTALL option is set, a dot matches any one character, without
3132     exception. If the two-character sequence CRLF is present in the subject
3133     string, it takes two dots to match it.
3134    
3135     The handling of dot is entirely independent of the handling of circum-
3136     flex and dollar, the only relationship being that they both involve
3137 nigel 91 newlines. Dot has no special meaning in a character class.
3138    
3139    
3140 nigel 63 MATCHING A SINGLE BYTE
3141    
3142 nigel 73 Outside a character class, the escape sequence \C matches any one byte,
3143 nigel 93 both in and out of UTF-8 mode. Unlike a dot, it always matches any
3144     line-ending characters. The feature is provided in Perl in order to
3145     match individual bytes in UTF-8 mode. Because it breaks up UTF-8 char-
3146     acters into individual bytes, what remains in the string may be a mal-
3147     formed UTF-8 string. For this reason, the \C escape sequence is best
3148     avoided.
3149 nigel 63
3150 nigel 77 PCRE does not allow \C to appear in lookbehind assertions (described
3151     below), because in UTF-8 mode this would make it impossible to calcu-
3152 nigel 75 late the length of the lookbehind.
3153 nigel 63
3154    
3155 nigel 75 SQUARE BRACKETS AND CHARACTER CLASSES
3156 nigel 63
3157 nigel 73 An opening square bracket introduces a character class, terminated by a
3158     closing square bracket. A closing square bracket on its own is not spe-
3159     cial. If a closing square bracket is required as a member of the class,
3160 nigel 77 it should be the first data character in the class (after an initial
3161 nigel 73 circumflex, if present) or escaped with a backslash.
3162 nigel 41
3163 nigel 77 A character class matches a single character in the subject. In UTF-8
3164     mode, the character may occupy more than one byte. A matched character
3165 nigel 73 must be in the set of characters defined by the class, unless the first
3166 nigel 77 character in the class definition is a circumflex, in which case the
3167     subject character must not be in the set defined by the class. If a
3168     circumflex is actually required as a member of the class, ensure it is
3169 nigel 73 not the first character, or escape it with a backslash.
3170 nigel 41
3171 nigel 77 For example, the character class [aeiou] matches any lower case vowel,
3172     while [^aeiou] matches any character that is not a lower case vowel.
3173 nigel 73 Note that a circumflex is just a convenient notation for specifying the
3174 nigel 77 characters that are in the class by enumerating those that are not. A
3175     class that starts with a circumflex is not an assertion: it still con-
3176     sumes a character from the subject string, and therefore it fails if
3177 nigel 75 the current pointer is at the end of the string.
3178 nigel 41
3179 nigel 77 In UTF-8 mode, characters with values greater than 255 can be included
3180     in a class as a literal string of bytes, or by using the \x{ escaping
3181 nigel 73 mechanism.
3182 nigel 63
3183 nigel 77 When caseless matching is set, any letters in a class represent both
3184     their upper case and lower case versions, so for example, a caseless
3185     [aeiou] matches "A" as well as "a", and a caseless [^aeiou] does not
3186     match "A", whereas a caseful version would. In UTF-8 mode, PCRE always
3187     understands the concept of case for characters whose values are less
3188     than 128, so caseless matching is always possible. For characters with
3189     higher values, the concept of case is supported if PCRE is compiled
3190     with Unicode property support, but not otherwise. If you want to use
3191     caseless matching for characters 128 and above, you must ensure that
3192     PCRE is compiled with Unicode property support as well as with UTF-8
3193     support.
3194 nigel 41
3195 nigel 93 Characters that might indicate line breaks are never treated in any
3196     special way when matching character classes, whatever line-ending
3197     sequence is in use, and whatever setting of the PCRE_DOTALL and
3198     PCRE_MULTILINE options is used. A class such as [^a] always matches one
3199     of these characters.
3200 nigel 41
3201 nigel 75 The minus (hyphen) character can be used to specify a range of charac-
3202     ters in a character class. For example, [d-m] matches any letter
3203     between d and m, inclusive. If a minus character is required in a
3204     class, it must be escaped with a backslash or appear in a position
3205     where it cannot be interpreted as indicating a range, typically as the
3206 nigel 73 first or last character in the class.
3207 nigel 41
3208 nigel 73 It is not possible to have the literal character "]" as the end charac-
3209 nigel 75 ter of a range. A pattern such as [W-]46] is interpreted as a class of
3210     two characters ("W" and "-") followed by a literal string "46]", so it
3211     would match "W46]" or "-46]". However, if the "]" is escaped with a
3212     backslash it is interpreted as the end of range, so [W-\]46] is inter-
3213     preted as a class containing a range followed by two other characters.
3214     The octal or hexadecimal representation of "]" can also be used to end
3215     a range.
3216 nigel 41
3217 nigel 75 Ranges operate in the collating sequence of character values. They can
3218     also be used for characters specified numerically, for example
3219     [\000-\037]. In UTF-8 mode, ranges can include characters whose values
3220 nigel 73 are greater than 255, for example [\x{100}-\x{2ff}].
3221 nigel 63
3222 nigel 73 If a range that includes letters is used when caseless matching is set,
3223     it matches the letters in either case. For example, [W-c] is equivalent
3224 nigel 75 to [][\\^_`wxyzabc], matched caselessly, and in non-UTF-8 mode, if
3225     character tables for the "fr_FR" locale are in use, [\xc8-\xcb] matches
3226     accented E characters in both cases. In UTF-8 mode, PCRE supports the
3227     concept of case for characters with values greater than 128 only when
3228     it is compiled with Unicode property support.
3229 nigel 41
3230 nigel 75 The character types \d, \D, \p, \P, \s, \S, \w, and \W may also appear
3231     in a character class, and add the characters that they match to the
3232     class. For example, [\dABCDEF] matches any hexadecimal digit. A circum-
3233     flex can conveniently be used with the upper case character types to
3234     specify a more restricted set of characters than the matching lower
3235     case type. For example, the class [^\W_] matches any letter or digit,
3236     but not underscore.
3237 nigel 41
3238 nigel 75 The only metacharacters that are recognized in character classes are
3239     backslash, hyphen (only where it can be interpreted as specifying a
3240     range), circumflex (only at the start), opening square bracket (only
3241     when it can be interpreted as introducing a POSIX class name - see the
3242     next section), and the terminating closing square bracket. However,
3243     escaping other non-alphanumeric characters does no harm.
3244 nigel 41
3245 nigel 73
3246 nigel 43 POSIX CHARACTER CLASSES
3247    
3248 nigel 75 Perl supports the POSIX notation for character classes. This uses names
3249     enclosed by [: and :] within the enclosing square brackets. PCRE also
3250     supports this notation. For example,
3251 nigel 63
3252 nigel 73 [01[:alpha:]%]
3253 nigel 43
3254 nigel 73 matches "0", "1", any alphabetic character, or "%". The supported class
3255     names are
3256 nigel 43
3257 nigel 73 alnum letters and digits
3258     alpha letters
3259     ascii character codes 0 - 127
3260     blank space or tab only
3261     cntrl control characters
3262     digit decimal digits (same as \d)
3263     graph printing characters, excluding space
3264     lower lower case letters
3265     print printing characters, including space
3266     punct printing characters, excluding letters and digits
3267     space white space (not quite the same as \s)
3268     upper upper case letters
3269     word "word" characters (same as \w)
3270     xdigit hexadecimal digits
3271 nigel 43
3272 nigel 73 The "space" characters are HT (9), LF (10), VT (11), FF (12), CR (13),
3273     and space (32). Notice that this list includes the VT character (code
3274     11). This makes "space" different to \s, which does not include VT (for
3275     Perl compatibility).
3276 nigel 43
3277 nigel 73 The name "word" is a Perl extension, and "blank" is a GNU extension
3278     from Perl 5.8. Another Perl extension is negation, which is indicated
3279     by a ^ character after the colon. For example,
3280 nigel 63
3281 nigel 73 [12[:^digit:]]
3282 nigel 43
3283 nigel 73 matches "1", "2", or any non-digit. PCRE (and Perl) also recognize the
3284     POSIX syntax [.ch.] and [=ch=] where "ch" is a "collating element", but
3285     these are not supported, and an error is given if they are encountered.
3286 nigel 43
3287 nigel 75 In UTF-8 mode, characters with values greater than 128 do not match any
3288 nigel 73 of the POSIX character classes.
3289 nigel 43
3290    
3291 nigel 41 VERTICAL BAR
3292 nigel 63
3293 nigel 73 Vertical bar characters are used to separate alternative patterns. For
3294     example, the pattern
3295 nigel 41
3296 nigel 73 gilbert|sullivan
3297 nigel 41
3298 nigel 73 matches either "gilbert" or "sullivan". Any number of alternatives may
3299     appear, and an empty alternative is permitted (matching the empty
3300 nigel 91 string). The matching process tries each alternative in turn, from left
3301     to right, and the first one that succeeds is used. If the alternatives
3302     are within a subpattern (defined below), "succeeds" means matching the
3303     rest of the main pattern as well as the alternative in the subpattern.
3304 nigel 41
3305    
3306     INTERNAL OPTION SETTING
3307    
3308 nigel 73 The settings of the PCRE_CASELESS, PCRE_MULTILINE, PCRE_DOTALL, and
3309     PCRE_EXTENDED options can be changed from within the pattern by a
3310     sequence of Perl option letters enclosed between "(?" and ")". The
3311     option letters are
3312 nigel 63
3313 nigel 73 i for PCRE_CASELESS
3314     m for PCRE_MULTILINE
3315     s for PCRE_DOTALL
3316     x for PCRE_EXTENDED
3317 nigel 41
3318 nigel 73 For example, (?im) sets caseless, multiline matching. It is also possi-
3319     ble to unset these options by preceding the letter with a hyphen, and a
3320     combined setting and unsetting such as (?im-sx), which sets PCRE_CASE-
3321     LESS and PCRE_MULTILINE while unsetting PCRE_DOTALL and PCRE_EXTENDED,
3322     is also permitted. If a letter appears both before and after the
3323     hyphen, the option is unset.
3324 nigel 41
3325 nigel 73 When an option change occurs at top level (that is, not inside subpat-
3326     tern parentheses), the change applies to the remainder of the pattern
3327     that follows. If the change is placed right at the start of a pattern,
3328     PCRE extracts it into the global options (and it will therefore show up
3329     in data extracted by the pcre_fullinfo() function).
3330 nigel 41
3331 nigel 93 An option change within a subpattern (see below for a description of
3332     subpatterns) affects only that part of the current pattern that follows
3333     it, so
3334 nigel 41
3335 nigel 73 (a(?i)b)c
3336 nigel 41
3337 nigel 73 matches abc and aBc and no other strings (assuming PCRE_CASELESS is not
3338 nigel 93 used). By this means, options can be made to have different settings
3339     in different parts of the pattern. Any changes made in one alternative
3340     do carry on into subsequent branches within the same subpattern. For
3341 nigel 73 example,
3342 nigel 41
3343 nigel 73 (a(?i)b|c)
3344 nigel 41
3345 nigel 93 matches "ab", "aB", "c", and "C", even though when matching "C" the
3346     first branch is abandoned before the option setting. This is because
3347     the effects of option settings happen at compile time. There would be
3348 nigel 73 some very weird behaviour otherwise.
3349 nigel 41
3350 nigel 93 The PCRE-specific options PCRE_DUPNAMES, PCRE_UNGREEDY, and PCRE_EXTRA
3351     can be changed in the same way as the Perl-compatible options by using
3352 nigel 91 the characters J, U and X respectively.
3353 nigel 41
3354    
3355 nigel 63 SUBPATTERNS
3356 nigel 41
3357 nigel 73 Subpatterns are delimited by parentheses (round brackets), which can be
3358 nigel 75 nested. Turning part of a pattern into a subpattern does two things:
3359 nigel 41
3360 nigel 73 1. It localizes a set of alternatives. For example, the pattern
3361 nigel 41
3362 nigel 73 cat(aract|erpillar|)
3363 nigel 41
3364 nigel 93 matches one of the words "cat", "cataract", or "caterpillar". Without
3365     the parentheses, it would match "cataract", "erpillar" or an empty
3366 nigel 73 string.
3367 nigel 41
3368 nigel 93 2. It sets up the subpattern as a capturing subpattern. This means
3369     that, when the whole pattern matches, that portion of the subject
3370 nigel 73 string that matched the subpattern is passed back to the caller via the
3371 nigel 93 ovector argument of pcre_exec(). Opening parentheses are counted from
3372     left to right (starting from 1) to obtain numbers for the capturing
3373 nigel 73 subpatterns.
3374 nigel 41
3375 nigel 93 For example, if the string "the red king" is matched against the pat-
3376 nigel 73 tern
3377 nigel 41
3378 nigel 73 the ((red|white) (king|queen))
3379 nigel 41
3380 nigel 73 the captured substrings are "red king", "red", and "king", and are num-
3381     bered 1, 2, and 3, respectively.
3382 nigel 41
3383 nigel 93 The fact that plain parentheses fulfil two functions is not always
3384     helpful. There are often times when a grouping subpattern is required
3385     without a capturing requirement. If an opening parenthesis is followed
3386     by a question mark and a colon, the subpattern does not do any captur-
3387     ing, and is not counted when computing the number of any subsequent
3388     capturing subpatterns. For example, if the string "the white queen" is
3389 nigel 73 matched against the pattern
3390 nigel 41
3391 nigel 73 the ((?:red|white) (king|queen))
3392 nigel 41
3393 nigel 73 the captured substrings are "white queen" and "queen", and are numbered
3394 nigel 93 1 and 2. The maximum number of capturing subpatterns is 65535.
3395 nigel 41
3396 nigel 93 As a convenient shorthand, if any option settings are required at the
3397     start of a non-capturing subpattern, the option letters may appear
3398 nigel 73 between the "?" and the ":". Thus the two patterns
3399 nigel 41
3400 nigel 73 (?i:saturday|sunday)
3401     (?:(?i)saturday|sunday)
3402 nigel 41
3403 nigel 73 match exactly the same set of strings. Because alternative branches are
3404 nigel 93 tried from left to right, and options are not reset until the end of
3405     the subpattern is reached, an option setting in one branch does affect
3406     subsequent branches, so the above patterns match "SUNDAY" as well as
3407 nigel 73 "Saturday".
3408 nigel 41
3409    
3410 nigel 63 NAMED SUBPATTERNS
3411 nigel 41
3412 nigel 93 Identifying capturing parentheses by number is simple, but it can be
3413     very hard to keep track of the numbers in complicated regular expres-
3414     sions. Furthermore, if an expression is modified, the numbers may
3415     change. To help with this difficulty, PCRE supports the naming of sub-
3416     patterns. This feature was not added to Perl until release 5.10. Python
3417     had the feature earlier, and PCRE introduced it at release 4.0, using
3418     the Python syntax. PCRE now supports both the Perl and the Python syn-
3419     tax.
3420 nigel 63
3421 nigel 93 In PCRE, a subpattern can be named in one of three ways: (?<name>...)
3422     or (?'name'...) as in Perl, or (?P<name>...) as in Python. References
3423     to capturing parentheses from other parts of the pattern, such as back-
3424     references, recursion, and conditions, can be made by name as well as
3425     by number.
3426 nigel 63
3427 nigel 93 Names consist of up to 32 alphanumeric characters and underscores.
3428     Named capturing parentheses are still allocated numbers as well as
3429     names, exactly as if the names were not present. The PCRE API provides
3430     function calls for extracting the name-to-number translation table from
3431     a compiled pattern. There is also a convenience function for extracting
3432     a captured substring by name.
3433    
3434 nigel 91 By default, a name must be unique within a pattern, but it is possible
3435     to relax this constraint by setting the PCRE_DUPNAMES option at compile
3436     time. This can be useful for patterns where only one instance of the
3437     named parentheses can m