ViewVC logotype

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

Parent Directory Parent Directory | Revision Log Revision Log

Revision 869 - (show annotations)
Sat Jan 14 11:16:23 2012 UTC (5 years ago) by ph10
File MIME type: text/plain
File size: 433351 byte(s)
Bring HTML docs up to date.
1 -----------------------------------------------------------------------------
2 This file contains a concatenation of the PCRE man pages, converted to plain
3 text format for ease of searching with a text editor, or for use on systems
4 that do not have a man page processor. The small individual files that give
5 synopses of each function in the library have not been included. Neither has
6 the pcredemo program. There are separate text files for the pcregrep and
7 pcretest commands.
8 -----------------------------------------------------------------------------
11 PCRE(3) PCRE(3)
15 PCRE - Perl-compatible regular expressions
20 The PCRE library is a set of functions that implement regular expres-
21 sion pattern matching using the same syntax and semantics as Perl, with
22 just a few differences. Some features that appeared in Python and PCRE
23 before they appeared in Perl are also available using the Python syn-
24 tax, there is some support for one or two .NET and Oniguruma syntax
25 items, and there is an option for requesting some minor changes that
26 give better JavaScript compatibility.
28 Starting with release 8.30, it is possible to compile two separate PCRE
29 libraries: the original, which supports 8-bit character strings
30 (including UTF-8 strings), and a second library that supports 16-bit
31 character strings (including UTF-16 strings). The build process allows
32 either one or both to be built. The majority of the work to make this
33 possible was done by Zoltan Herczeg.
35 The two libraries contain identical sets of functions, except that the
36 names in the 16-bit library start with pcre16_ instead of pcre_. To
37 avoid over-complication and reduce the documentation maintenance load,
38 most of the documentation describes the 8-bit library, with the differ-
39 ences for the 16-bit library described separately in the pcre16 page.
40 References to functions or structures of the form pcre[16]_xxx should
41 be read as meaning "pcre_xxx when using the 8-bit library and
42 pcre16_xxx when using the 16-bit library".
44 The current implementation of PCRE corresponds approximately with Perl
45 5.12, including support for UTF-8/16 encoded strings and Unicode gen-
46 eral category properties. However, UTF-8/16 and Unicode support has to
47 be explicitly enabled; it is not the default. The Unicode tables corre-
48 spond to Unicode release 6.0.0.
50 In addition to the Perl-compatible matching function, PCRE contains an
51 alternative function that matches the same compiled patterns in a dif-
52 ferent way. In certain circumstances, the alternative function has some
53 advantages. For a discussion of the two matching algorithms, see the
54 pcrematching page.
56 PCRE is written in C and released as a C library. A number of people
57 have written wrappers and interfaces of various kinds. In particular,
58 Google Inc. have provided a comprehensive C++ wrapper for the 8-bit
59 library. This is now included as part of the PCRE distribution. The
60 pcrecpp page has details of this interface. Other people's contribu-
61 tions can be found in the Contrib directory at the primary FTP site,
62 which is:
64 ftp://ftp.csx.cam.ac.uk/pub/software/programming/pcre
66 Details of exactly which Perl regular expression features are and are
67 not supported by PCRE are given in separate documents. See the pcrepat-
68 tern and pcrecompat pages. There is a syntax summary in the pcresyntax
69 page.
71 Some features of PCRE can be included, excluded, or changed when the
72 library is built. The pcre_config() function makes it possible for a
73 client to discover which features are available. The features them-
74 selves are described in the pcrebuild page. Documentation about build-
75 ing PCRE for various operating systems can be found in the README and
76 NON-UNIX-USE files in the source distribution.
78 The libraries contains a number of undocumented internal functions and
79 data tables that are used by more than one of the exported external
80 functions, but which are not intended for use by external callers.
81 Their names all begin with "_pcre_" or "_pcre16_", which hopefully will
82 not provoke any name clashes. In some environments, it is possible to
83 control which external symbols are exported when a shared library is
84 built, and in these cases the undocumented symbols are not exported.
89 The user documentation for PCRE comprises a number of different sec-
90 tions. In the "man" format, each of these is a separate "man page". In
91 the HTML format, each is a separate page, linked from the index page.
92 In the plain text format, all the sections, except the pcredemo sec-
93 tion, are concatenated, for ease of searching. The sections are as fol-
94 lows:
96 pcre this document
97 pcre16 details of the 16-bit library
98 pcre-config show PCRE installation configuration information
99 pcreapi details of PCRE's native C API
100 pcrebuild options for building PCRE
101 pcrecallout details of the callout feature
102 pcrecompat discussion of Perl compatibility
103 pcrecpp details of the C++ wrapper for the 8-bit library
104 pcredemo a demonstration C program that uses PCRE
105 pcregrep description of the pcregrep command (8-bit only)
106 pcrejit discussion of the just-in-time optimization support
107 pcrelimits details of size and other limits
108 pcrematching discussion of the two matching algorithms
109 pcrepartial details of the partial matching facility
110 pcrepattern syntax and semantics of supported
111 regular expressions
112 pcreperform discussion of performance issues
113 pcreposix the POSIX-compatible C API for the 8-bit library
114 pcreprecompile details of saving and re-using precompiled patterns
115 pcresample discussion of the pcredemo program
116 pcrestack discussion of stack usage
117 pcresyntax quick syntax reference
118 pcretest description of the pcretest testing command
119 pcreunicode discussion of Unicode and UTF-8/16 support
121 In addition, in the "man" and HTML formats, there is a short page for
122 each 8-bit C library function, listing its arguments and results.
127 Philip Hazel
128 University Computing Service
129 Cambridge CB2 3QH, England.
131 Putting an actual email address here seems to have been a spam magnet,
132 so I've taken it away. If you want to email me, use my two initials,
133 followed by the two digits 10, at the domain cam.ac.uk.
138 Last updated: 10 January 2012
139 Copyright (c) 1997-2012 University of Cambridge.
140 ------------------------------------------------------------------------------
143 PCRE(3) PCRE(3)
146 NAME
147 PCRE - Perl-compatible regular expressions
149 #include <pcre.h>
154 pcre16 *pcre16_compile(PCRE_SPTR16 pattern, int options,
155 const char **errptr, int *erroffset,
156 const unsigned char *tableptr);
158 pcre16 *pcre16_compile2(PCRE_SPTR16 pattern, int options,
159 int *errorcodeptr,
160 const char **errptr, int *erroffset,
161 const unsigned char *tableptr);
163 pcre16_extra *pcre16_study(const pcre16 *code, int options,
164 const char **errptr);
166 void pcre16_free_study(pcre16_extra *extra);
168 int pcre16_exec(const pcre16 *code, const pcre16_extra *extra,
169 PCRE_SPTR16 subject, int length, int startoffset,
170 int options, int *ovector, int ovecsize);
172 int pcre16_dfa_exec(const pcre16 *code, const pcre16_extra *extra,
173 PCRE_SPTR16 subject, int length, int startoffset,
174 int options, int *ovector, int ovecsize,
175 int *workspace, int wscount);
180 int pcre16_copy_named_substring(const pcre16 *code,
181 PCRE_SPTR16 subject, int *ovector,
182 int stringcount, PCRE_SPTR16 stringname,
183 PCRE_UCHAR16 *buffer, int buffersize);
185 int pcre16_copy_substring(PCRE_SPTR16 subject, int *ovector,
186 int stringcount, int stringnumber, PCRE_UCHAR16 *buffer,
187 int buffersize);
189 int pcre16_get_named_substring(const pcre16 *code,
190 PCRE_SPTR16 subject, int *ovector,
191 int stringcount, PCRE_SPTR16 stringname,
192 PCRE_SPTR16 *stringptr);
194 int pcre16_get_stringnumber(const pcre16 *code,
195 PCRE_SPTR16 name);
197 int pcre16_get_stringtable_entries(const pcre16 *code,
198 PCRE_SPTR16 name, PCRE_UCHAR16 **first, PCRE_UCHAR16 **last);
200 int pcre16_get_substring(PCRE_SPTR16 subject, int *ovector,
201 int stringcount, int stringnumber,
202 PCRE_SPTR16 *stringptr);
204 int pcre16_get_substring_list(PCRE_SPTR16 subject,
205 int *ovector, int stringcount, PCRE_SPTR16 **listptr);
207 void pcre16_free_substring(PCRE_SPTR16 stringptr);
209 void pcre16_free_substring_list(PCRE_SPTR16 *stringptr);
214 pcre16_jit_stack *pcre16_jit_stack_alloc(int startsize, int maxsize);
216 void pcre16_jit_stack_free(pcre16_jit_stack *stack);
218 void pcre16_assign_jit_stack(pcre16_extra *extra,
219 pcre16_jit_callback callback, void *data);
221 const unsigned char *pcre16_maketables(void);
223 int pcre16_fullinfo(const pcre16 *code, const pcre16_extra *extra,
224 int what, void *where);
226 int pcre16_refcount(pcre16 *code, int adjust);
228 int pcre16_config(int what, void *where);
230 const char *pcre16_version(void);
232 int pcre16_pattern_to_host_byte_order(pcre16 *code,
233 pcre16_extra *extra, const unsigned char *tables);
238 void *(*pcre16_malloc)(size_t);
240 void (*pcre16_free)(void *);
242 void *(*pcre16_stack_malloc)(size_t);
244 void (*pcre16_stack_free)(void *);
246 int (*pcre16_callout)(pcre16_callout_block *);
251 int pcre16_utf16_to_host_byte_order(PCRE_UCHAR16 *output,
252 PCRE_SPTR16 input, int length, int *byte_order,
253 int keep_boms);
258 Starting with release 8.30, it is possible to compile a PCRE library
259 that supports 16-bit character strings, including UTF-16 strings, as
260 well as or instead of the original 8-bit library. The majority of the
261 work to make this possible was done by Zoltan Herczeg. The two
262 libraries contain identical sets of functions, used in exactly the same
263 way. Only the names of the functions and the data types of their argu-
264 ments and results are different. To avoid over-complication and reduce
265 the documentation maintenance load, most of the PCRE documentation
266 describes the 8-bit library, with only occasional references to the
267 16-bit library. This page describes what is different when you use the
268 16-bit library.
270 WARNING: A single application can be linked with both libraries, but
271 you must take care when processing any particular pattern to use func-
272 tions from just one library. For example, if you want to study a pat-
273 tern that was compiled with pcre16_compile(), you must do so with
274 pcre16_study(), not pcre_study(), and you must free the study data with
275 pcre16_free_study().
280 There is only one header file, pcre.h. It contains prototypes for all
281 the functions in both libraries, as well as definitions of flags,
282 structures, error codes, etc.
287 In Unix-like systems, the 16-bit library is called libpcre16, and can
288 normally be accesss by adding -lpcre16 to the command for linking an
289 application that uses PCRE.
294 In the 8-bit library, strings are passed to PCRE library functions as
295 vectors of bytes with the C type "char *". In the 16-bit library,
296 strings are passed as vectors of unsigned 16-bit quantities. The macro
297 PCRE_UCHAR16 specifies an appropriate data type, and PCRE_SPTR16 is
298 defined as "const PCRE_UCHAR16 *". In very many environments, "short
299 int" is a 16-bit data type. When PCRE is built, it defines PCRE_UCHAR16
300 as "short int", but checks that it really is a 16-bit data type. If it
301 is not, the build fails with an error message telling the maintainer to
302 modify the definition appropriately.
307 The types of the opaque structures that are used for compiled 16-bit
308 patterns and JIT stacks are pcre16 and pcre16_jit_stack respectively.
309 The type of the user-accessible structure that is returned by
310 pcre16_study() is pcre16_extra, and the type of the structure that is
311 used for passing data to a callout function is pcre16_callout_block.
312 These structures contain the same fields, with the same names, as their
313 8-bit counterparts. The only difference is that pointers to character
314 strings are 16-bit instead of 8-bit types.
319 For every function in the 8-bit library there is a corresponding func-
320 tion in the 16-bit library with a name that starts with pcre16_ instead
321 of pcre_. The prototypes are listed above. In addition, there is one
322 extra function, pcre16_utf16_to_host_byte_order(). This is a utility
323 function that converts a UTF-16 character string to host byte order if
324 necessary. The other 16-bit functions expect the strings they are
325 passed to be in host byte order.
327 The input and output arguments of pcre16_utf16_to_host_byte_order() may
328 point to the same address, that is, conversion in place is supported.
329 The output buffer must be at least as long as the input.
331 The length argument specifies the number of 16-bit data units in the
332 input string; a negative value specifies a zero-terminated string.
334 If byte_order is NULL, it is assumed that the string starts off in host
335 byte order. This may be changed by byte-order marks (BOMs) anywhere in
336 the string (commonly as the first character).
338 If byte_order is not NULL, a non-zero value of the integer to which it
339 points means that the input starts off in host byte order, otherwise
340 the opposite order is assumed. Again, BOMs in the string can change
341 this. The final byte order is passed back at the end of processing.
343 If keep_boms is not zero, byte-order mark characters (0xfeff) are
344 copied into the output string. Otherwise they are discarded.
346 The result of the function is the number of 16-bit units placed into
347 the output buffer, including the zero terminator if the string was
348 zero-terminated.
353 The offsets within subject strings that are returned by the matching
354 functions are in 16-bit units rather than bytes.
359 The name-to-number translation table that is maintained for named sub-
360 patterns uses 16-bit characters. The pcre16_get_stringtable_entries()
361 function returns the length of each entry in the table as the number of
362 16-bit data units.
367 There are two new general option names, PCRE_UTF16 and
368 PCRE_NO_UTF16_CHECK, which correspond to PCRE_UTF8 and
369 PCRE_NO_UTF8_CHECK in the 8-bit library. In fact, these new options
370 define the same bits in the options word.
372 For the pcre16_config() function there is an option PCRE_CONFIG_UTF16
373 that returns 1 if UTF-16 support is configured, otherwise 0. If this
374 option is given to pcre_config(), or if the PCRE_CONFIG_UTF8 option is
375 given to pcre16_config(), the result is the PCRE_ERROR_BADOPTION error.
380 In 16-bit mode, when PCRE_UTF16 is not set, character values are
381 treated in the same way as in 8-bit, non UTF-8 mode, except, of course,
382 that they can range from 0 to 0xffff instead of 0 to 0xff. Character
383 types for characters less than 0xff can therefore be influenced by the
384 locale in the same way as before. Characters greater than 0xff have
385 only one case, and no "type" (such as letter or digit).
387 In UTF-16 mode, the character code is Unicode, in the range 0 to
388 0x10ffff, with the exception of values in the range 0xd800 to 0xdfff
389 because those are "surrogate" values that are used in pairs to encode
390 values greater than 0xffff.
392 A UTF-16 string can indicate its endianness by special code knows as a
393 byte-order mark (BOM). The PCRE functions do not handle this, expecting
394 strings to be in host byte order. A utility function called
395 pcre16_utf16_to_host_byte_order() is provided to help with this (see
396 above).
402 spond to their 8-bit counterparts. The error PCRE_ERROR_BADMODE is
403 given when a compiled pattern is passed to a function that processes
404 patterns in the other mode, for example, if a pattern compiled with
405 pcre_compile() is passed to pcre16_exec().
407 There are new error codes whose names begin with PCRE_UTF16_ERR for
408 invalid UTF-16 strings, corresponding to the PCRE_UTF8_ERR codes for
409 UTF-8 strings that are described in the section entitled "Reason codes
410 for invalid UTF-8 strings" in the main pcreapi page. The UTF-16 errors
411 are:
413 PCRE_UTF16_ERR1 Missing low surrogate at end of string
414 PCRE_UTF16_ERR2 Invalid low surrogate follows high surrogate
415 PCRE_UTF16_ERR3 Isolated low surrogate
416 PCRE_UTF16_ERR4 Invalid character 0xfffe
421 If there is an error while compiling a pattern, the error text that is
422 passed back by pcre16_compile() or pcre16_compile2() is still an 8-bit
423 character string, zero-terminated.
428 The subject and mark fields in the callout block that is passed to a
429 callout function point to 16-bit vectors.
434 The pcretest program continues to operate with 8-bit input and output
435 files, but it can be used for testing the 16-bit library. If it is run
436 with the command line option -16, patterns and subject strings are con-
437 verted from 8-bit to 16-bit before being passed to PCRE, and the 16-bit
438 library functions are used instead of the 8-bit ones. Returned 16-bit
439 strings are converted to 8-bit for output. If the 8-bit library was not
440 compiled, pcretest defaults to 16-bit and the -16 option is ignored.
442 When PCRE is being built, the RunTest script that is called by "make
443 check" uses the pcretest -C option to discover which of the 8-bit and
444 16-bit libraries has been built, and runs the tests appropriately.
449 Not all the features of the 8-bit library are available with the 16-bit
450 library. The C++ and POSIX wrapper functions support only the 8-bit
451 library, and the pcregrep program is at present 8-bit only.
456 Philip Hazel
457 University Computing Service
458 Cambridge CB2 3QH, England.
463 Last updated: 08 January 2012
464 Copyright (c) 1997-2012 University of Cambridge.
465 ------------------------------------------------------------------------------
471 NAME
472 PCRE - Perl-compatible regular expressions
477 This document describes the optional features of PCRE that can be
478 selected when the library is compiled. It assumes use of the configure
479 script, where the optional features are selected or deselected by pro-
480 viding options to configure before running the make command. However,
481 the same options can be selected in both Unix-like and non-Unix-like
482 environments using the GUI facility of cmake-gui if you are using CMake
483 instead of configure to build PCRE.
485 There is a lot more information about building PCRE in non-Unix-like
486 environments in the file called NON_UNIX_USE, which is part of the PCRE
487 distribution. You should consult this file as well as the README file
488 if you are building in a non-Unix-like environment.
490 The complete list of options for configure (which includes the standard
491 ones such as the selection of the installation directory) can be
492 obtained by running
494 ./configure --help
496 The following sections include descriptions of options whose names
497 begin with --enable or --disable. These settings specify changes to the
498 defaults for the configure command. Because of the way that configure
499 works, --enable and --disable always come in pairs, so the complemen-
500 tary option always exists as well, but as it specifies the default, it
501 is not described.
506 By default, a library called libpcre is built, containing functions
507 that take string arguments contained in vectors of bytes, either as
508 single-byte characters, or interpreted as UTF-8 strings. You can also
509 build a separate library, called libpcre16, in which strings are con-
510 tained in vectors of 16-bit data units and interpreted either as sin-
511 gle-unit characters or UTF-16 strings, by adding
513 --enable-pcre16
515 to the configure command. If you do not want the 8-bit library, add
517 --disable-pcre8
519 as well. At least one of the two libraries must be built. Note that the
520 C++ and POSIX wrappers are for the 8-bit library only, and that pcre-
521 grep is an 8-bit program. None of these are built if you select only
522 the 16-bit library.
527 The PCRE building process uses libtool to build both shared and static
528 Unix libraries by default. You can suppress one of these by adding one
529 of
531 --disable-shared
532 --disable-static
534 to the configure command, as required.
539 By default, if the 8-bit library is being built, the configure script
540 will search for a C++ compiler and C++ header files. If it finds them,
541 it automatically builds the C++ wrapper library (which supports only
542 8-bit strings). You can disable this by adding
544 --disable-cpp
546 to the configure command.
549 UTF-8 and UTF-16 SUPPORT
551 To build PCRE with support for UTF Unicode character strings, add
553 --enable-utf
555 to the configure command. This setting applies to both libraries,
556 adding support for UTF-8 to the 8-bit library and support for UTF-16 to
557 the 16-bit library. It is not possible to build one library with UTF
558 support and the other without in the same configuration. (For backwards
559 compatibility, --enable-utf8 is a synonym of --enable-utf.)
561 Of itself, this setting does not make PCRE treat strings as UTF-8 or
562 UTF-16. As well as compiling PCRE with this option, you also have have
563 to set the PCRE_UTF8 or PCRE_UTF16 option when you call one of the pat-
564 tern compiling functions.
566 If you set --enable-utf when compiling in an EBCDIC environment, PCRE
567 expects its input to be either ASCII or UTF-8 (depending on the runtime
568 option). It is not possible to support both EBCDIC and UTF-8 codes in
569 the same version of the library. Consequently, --enable-utf and
570 --enable-ebcdic are mutually exclusive.
575 UTF support allows the libraries to process character codepoints up to
576 0x10ffff in the strings that they handle. On its own, however, it does
577 not provide any facilities for accessing the properties of such charac-
578 ters. If you want to be able to use the pattern escapes \P, \p, and \X,
579 which refer to Unicode character properties, you must add
581 --enable-unicode-properties
583 to the configure command. This implies UTF support, even if you have
584 not explicitly requested it.
586 Including Unicode property support adds around 30K of tables to the
587 PCRE library. Only the general category properties such as Lu and Nd
588 are supported. Details are given in the pcrepattern documentation.
593 Just-in-time compiler support is included in the build by specifying
595 --enable-jit
597 This support is available only for certain hardware architectures. If
598 this option is set for an unsupported architecture, a compile time
599 error occurs. See the pcrejit documentation for a discussion of JIT
600 usage. When JIT support is enabled, pcregrep automatically makes use of
601 it, unless you add
603 --disable-pcregrep-jit
605 to the "configure" command.
610 By default, PCRE interprets the linefeed (LF) character as indicating
611 the end of a line. This is the normal newline character on Unix-like
612 systems. You can compile PCRE to use carriage return (CR) instead, by
613 adding
615 --enable-newline-is-cr
617 to the configure command. There is also a --enable-newline-is-lf
618 option, which explicitly specifies linefeed as the newline character.
620 Alternatively, you can specify that line endings are to be indicated by
621 the two character sequence CRLF. If you want this, add
623 --enable-newline-is-crlf
625 to the configure command. There is a fourth option, specified by
627 --enable-newline-is-anycrlf
629 which causes PCRE to recognize any of the three sequences CR, LF, or
630 CRLF as indicating a line ending. Finally, a fifth option, specified by
632 --enable-newline-is-any
634 causes PCRE to recognize any Unicode newline sequence.
636 Whatever line ending convention is selected when PCRE is built can be
637 overridden when the library functions are called. At build time it is
638 conventional to use the standard for your operating system.
643 By default, the sequence \R in a pattern matches any Unicode newline
644 sequence, whatever has been selected as the line ending sequence. If
645 you specify
647 --enable-bsr-anycrlf
649 the default is changed so that \R matches only CR, LF, or CRLF. What-
650 ever is selected when PCRE is built can be overridden when the library
651 functions are called.
656 When the 8-bit library is called through the POSIX interface (see the
657 pcreposix documentation), additional working storage is required for
658 holding the pointers to capturing substrings, because PCRE requires
659 three integers per substring, whereas the POSIX interface provides only
660 two. If the number of expected substrings is small, the wrapper func-
661 tion uses space on the stack, because this is faster than using mal-
662 loc() for each call. The default threshold above which the stack is no
663 longer used is 10; it can be changed by adding a setting such as
665 --with-posix-malloc-threshold=20
667 to the configure command.
672 Within a compiled pattern, offset values are used to point from one
673 part to another (for example, from an opening parenthesis to an alter-
674 nation metacharacter). By default, two-byte values are used for these
675 offsets, leading to a maximum size for a compiled pattern of around
676 64K. This is sufficient to handle all but the most gigantic patterns.
677 Nevertheless, some people do want to process truly enormous patterns,
678 so it is possible to compile PCRE to use three-byte or four-byte off-
679 sets by adding a setting such as
681 --with-link-size=3
683 to the configure command. The value given must be 2, 3, or 4. For the
684 16-bit library, a value of 3 is rounded up to 4. Using longer offsets
685 slows down the operation of PCRE because it has to load additional data
686 when handling them.
691 When matching with the pcre_exec() function, PCRE implements backtrack-
692 ing by making recursive calls to an internal function called match().
693 In environments where the size of the stack is limited, this can se-
694 verely limit PCRE's operation. (The Unix environment does not usually
695 suffer from this problem, but it may sometimes be necessary to increase
696 the maximum stack size. There is a discussion in the pcrestack docu-
697 mentation.) An alternative approach to recursion that uses memory from
698 the heap to remember data, instead of using recursive function calls,
699 has been implemented to work round the problem of limited stack size.
700 If you want to build a version of PCRE that works this way, add
702 --disable-stack-for-recursion
704 to the configure command. With this configuration, PCRE will use the
705 pcre_stack_malloc and pcre_stack_free variables to call memory manage-
706 ment functions. By default these point to malloc() and free(), but you
707 can replace the pointers so that your own functions are used instead.
709 Separate functions are provided rather than using pcre_malloc and
710 pcre_free because the usage is very predictable: the block sizes
711 requested are always the same, and the blocks are always freed in
712 reverse order. A calling program might be able to implement optimized
713 functions that perform better than malloc() and free(). PCRE runs
714 noticeably more slowly when built in this way. This option affects only
715 the pcre_exec() function; it is not relevant for pcre_dfa_exec().
720 Internally, PCRE has a function called match(), which it calls repeat-
721 edly (sometimes recursively) when matching a pattern with the
722 pcre_exec() function. By controlling the maximum number of times this
723 function may be called during a single matching operation, a limit can
724 be placed on the resources used by a single call to pcre_exec(). The
725 limit can be changed at run time, as described in the pcreapi documen-
726 tation. The default is 10 million, but this can be changed by adding a
727 setting such as
729 --with-match-limit=500000
731 to the configure command. This setting has no effect on the
732 pcre_dfa_exec() matching function.
734 In some environments it is desirable to limit the depth of recursive
735 calls of match() more strictly than the total number of calls, in order
736 to restrict the maximum amount of stack (or heap, if --disable-stack-
737 for-recursion is specified) that is used. A second limit controls this;
738 it defaults to the value that is set for --with-match-limit, which
739 imposes no additional constraints. However, you can set a lower limit
740 by adding, for example,
742 --with-match-limit-recursion=10000
744 to the configure command. This value can also be overridden at run
745 time.
750 PCRE uses fixed tables for processing characters whose code values are
751 less than 256. By default, PCRE is built with a set of tables that are
752 distributed in the file pcre_chartables.c.dist. These tables are for
753 ASCII codes only. If you add
755 --enable-rebuild-chartables
757 to the configure command, the distributed tables are no longer used.
758 Instead, a program called dftables is compiled and run. This outputs
759 the source for new set of tables, created in the default locale of your
760 C runtime system. (This method of replacing the tables does not work if
761 you are cross compiling, because dftables is run on the local host. If
762 you need to create alternative tables when cross compiling, you will
763 have to do so "by hand".)
768 PCRE assumes by default that it will run in an environment where the
769 character code is ASCII (or Unicode, which is a superset of ASCII).
770 This is the case for most computer operating systems. PCRE can, how-
771 ever, be compiled to run in an EBCDIC environment by adding
773 --enable-ebcdic
775 to the configure command. This setting implies --enable-rebuild-charta-
776 bles. You should only use it if you know that you are in an EBCDIC
777 environment (for example, an IBM mainframe operating system). The
778 --enable-ebcdic option is incompatible with --enable-utf.
783 By default, pcregrep reads all files as plain text. You can build it so
784 that it recognizes files whose names end in .gz or .bz2, and reads them
785 with libz or libbz2, respectively, by adding one or both of
787 --enable-pcregrep-libz
788 --enable-pcregrep-libbz2
790 to the configure command. These options naturally require that the rel-
791 evant libraries are installed on your system. Configuration will fail
792 if they are not.
797 pcregrep uses an internal buffer to hold a "window" on the file it is
798 scanning, in order to be able to output "before" and "after" lines when
799 it finds a match. The size of the buffer is controlled by a parameter
800 whose default value is 20K. The buffer itself is three times this size,
801 but because of the way it is used for holding "before" lines, the long-
802 est line that is guaranteed to be processable is the parameter size.
803 You can change the default parameter value by adding, for example,
805 --with-pcregrep-bufsize=50K
807 to the configure command. The caller of pcregrep can, however, override
808 this value by specifying a run-time option.
813 If you add
815 --enable-pcretest-libreadline
817 to the configure command, pcretest is linked with the libreadline
818 library, and when its input is from a terminal, it reads it using the
819 readline() function. This provides line-editing and history facilities.
820 Note that libreadline is GPL-licensed, so if you distribute a binary of
821 pcretest linked in this way, there may be licensing issues.
823 Setting this option causes the -lreadline option to be added to the
824 pcretest build. In many operating environments with a sytem-installed
825 libreadline this is sufficient. However, in some environments (e.g. if
826 an unmodified distribution version of readline is in use), some extra
827 configuration may be necessary. The INSTALL file for libreadline says
828 this:
830 "Readline uses the termcap functions, but does not link with the
831 termcap or curses library itself, allowing applications which link
832 with readline the to choose an appropriate library."
834 If your environment has not been set up so that an appropriate library
835 is automatically included, you may need to add something like
837 LIBS="-ncurses"
839 immediately before the configure command.
844 pcreapi(3), pcre16, pcre_config(3).
849 Philip Hazel
850 University Computing Service
851 Cambridge CB2 3QH, England.
856 Last updated: 07 January 2012
857 Copyright (c) 1997-2012 University of Cambridge.
858 ------------------------------------------------------------------------------
864 NAME
865 PCRE - Perl-compatible regular expressions
870 This document describes the two different algorithms that are available
871 in PCRE for matching a compiled regular expression against a given sub-
872 ject string. The "standard" algorithm is the one provided by the
873 pcre_exec() and pcre16_exec() functions. These work in the same was as
874 Perl's matching function, and provide a Perl-compatible matching opera-
875 tion. The just-in-time (JIT) optimization that is described in the
876 pcrejit documentation is compatible with these functions.
878 An alternative algorithm is provided by the pcre_dfa_exec() and
879 pcre16_dfa_exec() functions; they operate in a different way, and are
880 not Perl-compatible. This alternative has advantages and disadvantages
881 compared with the standard algorithm, and these are described below.
883 When there is only one possible way in which a given subject string can
884 match a pattern, the two algorithms give the same answer. A difference
885 arises, however, when there are multiple possibilities. For example, if
886 the pattern
888 ^<.*>
890 is matched against the string
892 <something> <something else> <something further>
894 there are three possible answers. The standard algorithm finds only one
895 of them, whereas the alternative algorithm finds all three.
900 The set of strings that are matched by a regular expression can be rep-
901 resented as a tree structure. An unlimited repetition in the pattern
902 makes the tree of infinite size, but it is still a tree. Matching the
903 pattern to a given subject string (from a given starting point) can be
904 thought of as a search of the tree. There are two ways to search a
905 tree: depth-first and breadth-first, and these correspond to the two
906 matching algorithms provided by PCRE.
911 In the terminology of Jeffrey Friedl's book "Mastering Regular Expres-
912 sions", the standard algorithm is an "NFA algorithm". It conducts a
913 depth-first search of the pattern tree. That is, it proceeds along a
914 single path through the tree, checking that the subject matches what is
915 required. When there is a mismatch, the algorithm tries any alterna-
916 tives at the current point, and if they all fail, it backs up to the
917 previous branch point in the tree, and tries the next alternative
918 branch at that level. This often involves backing up (moving to the
919 left) in the subject string as well. The order in which repetition
920 branches are tried is controlled by the greedy or ungreedy nature of
921 the quantifier.
923 If a leaf node is reached, a matching string has been found, and at
924 that point the algorithm stops. Thus, if there is more than one possi-
925 ble match, this algorithm returns the first one that it finds. Whether
926 this is the shortest, the longest, or some intermediate length depends
927 on the way the greedy and ungreedy repetition quantifiers are specified
928 in the pattern.
930 Because it ends up with a single path through the tree, it is rela-
931 tively straightforward for this algorithm to keep track of the sub-
932 strings that are matched by portions of the pattern in parentheses.
933 This provides support for capturing parentheses and back references.
938 This algorithm conducts a breadth-first search of the tree. Starting
939 from the first matching point in the subject, it scans the subject
940 string from left to right, once, character by character, and as it does
941 this, it remembers all the paths through the tree that represent valid
942 matches. In Friedl's terminology, this is a kind of "DFA algorithm",
943 though it is not implemented as a traditional finite state machine (it
944 keeps multiple states active simultaneously).
946 Although the general principle of this matching algorithm is that it
947 scans the subject string only once, without backtracking, there is one
948 exception: when a lookaround assertion is encountered, the characters
949 following or preceding the current point have to be independently
950 inspected.
952 The scan continues until either the end of the subject is reached, or
953 there are no more unterminated paths. At this point, terminated paths
954 represent the different matching possibilities (if there are none, the
955 match has failed). Thus, if there is more than one possible match,
956 this algorithm finds all of them, and in particular, it finds the long-
957 est. The matches are returned in decreasing order of length. There is
958 an option to stop the algorithm after the first match (which is neces-
959 sarily the shortest) is found.
961 Note that all the matches that are found start at the same point in the
962 subject. If the pattern
964 cat(er(pillar)?)?
966 is matched against the string "the caterpillar catchment", the result
967 will be the three strings "caterpillar", "cater", and "cat" that start
968 at the fifth character of the subject. The algorithm does not automati-
969 cally move on to find matches that start at later positions.
971 There are a number of features of PCRE regular expressions that are not
972 supported by the alternative matching algorithm. They are as follows:
974 1. Because the algorithm finds all possible matches, the greedy or
975 ungreedy nature of repetition quantifiers is not relevant. Greedy and
976 ungreedy quantifiers are treated in exactly the same way. However, pos-
977 sessive quantifiers can make a difference when what follows could also
978 match what is quantified, for example in a pattern like this:
980 ^a++\w!
982 This pattern matches "aaab!" but not "aaa!", which would be matched by
983 a non-possessive quantifier. Similarly, if an atomic group is present,
984 it is matched as if it were a standalone pattern at the current point,
985 and the longest match is then "locked in" for the rest of the overall
986 pattern.
988 2. When dealing with multiple paths through the tree simultaneously, it
989 is not straightforward to keep track of captured substrings for the
990 different matching possibilities, and PCRE's implementation of this
991 algorithm does not attempt to do this. This means that no captured sub-
992 strings are available.
994 3. Because no substrings are captured, back references within the pat-
995 tern are not supported, and cause errors if encountered.
997 4. For the same reason, conditional expressions that use a backrefer-
998 ence as the condition or test for a specific group recursion are not
999 supported.
1001 5. Because many paths through the tree may be active, the \K escape
1002 sequence, which resets the start of the match when encountered (but may
1003 be on some paths and not on others), is not supported. It causes an
1004 error if encountered.
1006 6. Callouts are supported, but the value of the capture_top field is
1007 always 1, and the value of the capture_last field is always -1.
1009 7. The \C escape sequence, which (in the standard algorithm) always
1010 matches a single data unit, even in UTF-8 or UTF-16 modes, is not sup-
1011 ported in these modes, because the alternative algorithm moves through
1012 the subject string one character (not data unit) at a time, for all
1013 active paths through the tree.
1015 8. Except for (*FAIL), the backtracking control verbs such as (*PRUNE)
1016 are not supported. (*FAIL) is supported, and behaves like a failing
1017 negative assertion.
1022 Using the alternative matching algorithm provides the following advan-
1023 tages:
1025 1. All possible matches (at a single point in the subject) are automat-
1026 ically found, and in particular, the longest match is found. To find
1027 more than one match using the standard algorithm, you have to do kludgy
1028 things with callouts.
1030 2. Because the alternative algorithm scans the subject string just
1031 once, and never needs to backtrack (except for lookbehinds), it is pos-
1032 sible to pass very long subject strings to the matching function in
1033 several pieces, checking for partial matching each time. Although it is
1034 possible to do multi-segment matching using the standard algorithm by
1035 retaining partially matched substrings, it is more complicated. The
1036 pcrepartial documentation gives details of partial matching and dis-
1037 cusses multi-segment matching.
1042 The alternative algorithm suffers from a number of disadvantages:
1044 1. It is substantially slower than the standard algorithm. This is
1045 partly because it has to search for all possible matches, but is also
1046 because it is less susceptible to optimization.
1048 2. Capturing parentheses and back references are not supported.
1050 3. Although atomic groups are supported, their use does not provide the
1051 performance advantage that it does for the standard algorithm.
1056 Philip Hazel
1057 University Computing Service
1058 Cambridge CB2 3QH, England.
1063 Last updated: 08 January 2012
1064 Copyright (c) 1997-2012 University of Cambridge.
1065 ------------------------------------------------------------------------------
1071 NAME
1072 PCRE - Perl-compatible regular expressions
1074 #include <pcre.h>
1079 pcre *pcre_compile(const char *pattern, int options,
1080 const char **errptr, int *erroffset,
1081 const unsigned char *tableptr);
1083 pcre *pcre_compile2(const char *pattern, int options,
1084 int *errorcodeptr,
1085 const char **errptr, int *erroffset,
1086 const unsigned char *tableptr);
1088 pcre_extra *pcre_study(const pcre *code, int options,
1089 const char **errptr);
1091 void pcre_free_study(pcre_extra *extra);
1093 int pcre_exec(const pcre *code, const pcre_extra *extra,
1094 const char *subject, int length, int startoffset,
1095 int options, int *ovector, int ovecsize);
1097 int pcre_dfa_exec(const pcre *code, const pcre_extra *extra,
1098 const char *subject, int length, int startoffset,
1099 int options, int *ovector, int ovecsize,
1100 int *workspace, int wscount);
1105 int pcre_copy_named_substring(const pcre *code,
1106 const char *subject, int *ovector,
1107 int stringcount, const char *stringname,
1108 char *buffer, int buffersize);
1110 int pcre_copy_substring(const char *subject, int *ovector,
1111 int stringcount, int stringnumber, char *buffer,
1112 int buffersize);
1114 int pcre_get_named_substring(const pcre *code,
1115 const char *subject, int *ovector,
1116 int stringcount, const char *stringname,
1117 const char **stringptr);
1119 int pcre_get_stringnumber(const pcre *code,
1120 const char *name);
1122 int pcre_get_stringtable_entries(const pcre *code,
1123 const char *name, char **first, char **last);
1125 int pcre_get_substring(const char *subject, int *ovector,
1126 int stringcount, int stringnumber,
1127 const char **stringptr);
1129 int pcre_get_substring_list(const char *subject,
1130 int *ovector, int stringcount, const char ***listptr);
1132 void pcre_free_substring(const char *stringptr);
1134 void pcre_free_substring_list(const char **stringptr);
1139 pcre_jit_stack *pcre_jit_stack_alloc(int startsize, int maxsize);
1141 void pcre_jit_stack_free(pcre_jit_stack *stack);
1143 void pcre_assign_jit_stack(pcre_extra *extra,
1144 pcre_jit_callback callback, void *data);
1146 const unsigned char *pcre_maketables(void);
1148 int pcre_fullinfo(const pcre *code, const pcre_extra *extra,
1149 int what, void *where);
1151 int pcre_refcount(pcre *code, int adjust);
1153 int pcre_config(int what, void *where);
1155 const char *pcre_version(void);
1157 int pcre_pattern_to_host_byte_order(pcre *code,
1158 pcre_extra *extra, const unsigned char *tables);
1163 void *(*pcre_malloc)(size_t);
1165 void (*pcre_free)(void *);
1167 void *(*pcre_stack_malloc)(size_t);
1169 void (*pcre_stack_free)(void *);
1171 int (*pcre_callout)(pcre_callout_block *);
1176 From release 8.30, PCRE can be compiled as a library for handling
1177 16-bit character strings as well as, or instead of, the original
1178 library that handles 8-bit character strings. To avoid too much compli-
1179 cation, this document describes the 8-bit versions of the functions,
1180 with only occasional references to the 16-bit library.
1182 The 16-bit functions operate in the same way as their 8-bit counter-
1183 parts; they just use different data types for their arguments and
1184 results, and their names start with pcre16_ instead of pcre_. For every
1185 option that has UTF8 in its name (for example, PCRE_UTF8), there is a
1186 corresponding 16-bit name with UTF8 replaced by UTF16. This facility is
1187 in fact just cosmetic; the 16-bit option names define the same bit val-
1188 ues.
1190 References to bytes and UTF-8 in this document should be read as refer-
1191 ences to 16-bit data quantities and UTF-16 when using the 16-bit
1192 library, unless specified otherwise. More details of the specific dif-
1193 ferences for the 16-bit library are given in the pcre16 page.
1198 PCRE has its own native API, which is described in this document. There
1199 are also some wrapper functions (for the 8-bit library only) that cor-
1200 respond to the POSIX regular expression API, but they do not give
1201 access to all the functionality. They are described in the pcreposix
1202 documentation. Both of these APIs define a set of C function calls. A
1203 C++ wrapper (again for the 8-bit library only) is also distributed with
1204 PCRE. It is documented in the pcrecpp page.
1206 The native API C function prototypes are defined in the header file
1207 pcre.h, and on Unix-like systems the (8-bit) library itself is called
1208 libpcre. It can normally be accessed by adding -lpcre to the command
1209 for linking an application that uses PCRE. The header file defines the
1210 macros PCRE_MAJOR and PCRE_MINOR to contain the major and minor release
1211 numbers for the library. Applications can use these to include support
1212 for different releases of PCRE.
1214 In a Windows environment, if you want to statically link an application
1215 program against a non-dll pcre.a file, you must define PCRE_STATIC
1216 before including pcre.h or pcrecpp.h, because otherwise the pcre_mal-
1217 loc() and pcre_free() exported functions will be declared
1218 __declspec(dllimport), with unwanted results.
1220 The functions pcre_compile(), pcre_compile2(), pcre_study(), and
1221 pcre_exec() are used for compiling and matching regular expressions in
1222 a Perl-compatible manner. A sample program that demonstrates the sim-
1223 plest way of using them is provided in the file called pcredemo.c in
1224 the PCRE source distribution. A listing of this program is given in the
1225 pcredemo documentation, and the pcresample documentation describes how
1226 to compile and run it.
1228 Just-in-time compiler support is an optional feature of PCRE that can
1229 be built in appropriate hardware environments. It greatly speeds up the
1230 matching performance of many patterns. Simple programs can easily
1231 request that it be used if available, by setting an option that is
1232 ignored when it is not relevant. More complicated programs might need
1233 to make use of the functions pcre_jit_stack_alloc(),
1234 pcre_jit_stack_free(), and pcre_assign_jit_stack() in order to control
1235 the JIT code's memory usage. These functions are discussed in the
1236 pcrejit documentation.
1238 A second matching function, pcre_dfa_exec(), which is not Perl-compati-
1239 ble, is also provided. This uses a different algorithm for the match-
1240 ing. The alternative algorithm finds all possible matches (at a given
1241 point in the subject), and scans the subject just once (unless there
1242 are lookbehind assertions). However, this algorithm does not return
1243 captured substrings. A description of the two matching algorithms and
1244 their advantages and disadvantages is given in the pcrematching docu-
1245 mentation.
1247 In addition to the main compiling and matching functions, there are
1248 convenience functions for extracting captured substrings from a subject
1249 string that is matched by pcre_exec(). They are:
1251 pcre_copy_substring()
1252 pcre_copy_named_substring()
1253 pcre_get_substring()
1254 pcre_get_named_substring()
1255 pcre_get_substring_list()
1256 pcre_get_stringnumber()
1257 pcre_get_stringtable_entries()
1259 pcre_free_substring() and pcre_free_substring_list() are also provided,
1260 to free the memory used for extracted strings.
1262 The function pcre_maketables() is used to build a set of character
1263 tables in the current locale for passing to pcre_compile(),
1264 pcre_exec(), or pcre_dfa_exec(). This is an optional facility that is
1265 provided for specialist use. Most commonly, no special tables are
1266 passed, in which case internal tables that are generated when PCRE is
1267 built are used.
1269 The function pcre_fullinfo() is used to find out information about a
1270 compiled pattern. The function pcre_version() returns a pointer to a
1271 string containing the version of PCRE and its date of release.
1273 The function pcre_refcount() maintains a reference count in a data
1274 block containing a compiled pattern. This is provided for the benefit
1275 of object-oriented applications.
1277 The global variables pcre_malloc and pcre_free initially contain the
1278 entry points of the standard malloc() and free() functions, respec-
1279 tively. PCRE calls the memory management functions via these variables,
1280 so a calling program can replace them if it wishes to intercept the
1281 calls. This should be done before calling any PCRE functions.
1283 The global variables pcre_stack_malloc and pcre_stack_free are also
1284 indirections to memory management functions. These special functions
1285 are used only when PCRE is compiled to use the heap for remembering
1286 data, instead of recursive function calls, when running the pcre_exec()
1287 function. See the pcrebuild documentation for details of how to do
1288 this. It is a non-standard way of building PCRE, for use in environ-
1289 ments that have limited stacks. Because of the greater use of memory
1290 management, it runs more slowly. Separate functions are provided so
1291 that special-purpose external code can be used for this case. When
1292 used, these functions are always called in a stack-like manner (last
1293 obtained, first freed), and always for memory blocks of the same size.
1294 There is a discussion about PCRE's stack usage in the pcrestack docu-
1295 mentation.
1297 The global variable pcre_callout initially contains NULL. It can be set
1298 by the caller to a "callout" function, which PCRE will then call at
1299 specified points during a matching operation. Details are given in the
1300 pcrecallout documentation.
1305 PCRE supports five different conventions for indicating line breaks in
1306 strings: a single CR (carriage return) character, a single LF (line-
1307 feed) character, the two-character sequence CRLF, any of the three pre-
1308 ceding, or any Unicode newline sequence. The Unicode newline sequences
1309 are the three just mentioned, plus the single characters VT (vertical
1310 tab, U+000B), FF (formfeed, U+000C), NEL (next line, U+0085), LS (line
1311 separator, U+2028), and PS (paragraph separator, U+2029).
1313 Each of the first three conventions is used by at least one operating
1314 system as its standard newline sequence. When PCRE is built, a default
1315 can be specified. The default default is LF, which is the Unix stan-
1316 dard. When PCRE is run, the default can be overridden, either when a
1317 pattern is compiled, or when it is matched.
1319 At compile time, the newline convention can be specified by the options
1320 argument of pcre_compile(), or it can be specified by special text at
1321 the start of the pattern itself; this overrides any other settings. See
1322 the pcrepattern page for details of the special character sequences.
1324 In the PCRE documentation the word "newline" is used to mean "the char-
1325 acter or pair of characters that indicate a line break". The choice of
1326 newline convention affects the handling of the dot, circumflex, and
1327 dollar metacharacters, the handling of #-comments in /x mode, and, when
1328 CRLF is a recognized line ending sequence, the match position advance-
1329 ment for a non-anchored pattern. There is more detail about this in the
1330 section on pcre_exec() options below.
1332 The choice of newline convention does not affect the interpretation of
1333 the \n or \r escape sequences, nor does it affect what \R matches,
1334 which is controlled in a similar way, but by separate options.
1339 The PCRE functions can be used in multi-threading applications, with
1340 the proviso that the memory management functions pointed to by
1341 pcre_malloc, pcre_free, pcre_stack_malloc, and pcre_stack_free, and the
1342 callout function pointed to by pcre_callout, are shared by all threads.
1344 The compiled form of a regular expression is not altered during match-
1345 ing, so the same compiled pattern can safely be used by several threads
1346 at once.
1348 If the just-in-time optimization feature is being used, it needs sepa-
1349 rate memory stack areas for each thread. See the pcrejit documentation
1350 for more details.
1355 The compiled form of a regular expression can be saved and re-used at a
1356 later time, possibly by a different program, and even on a host other
1357 than the one on which it was compiled. Details are given in the
1358 pcreprecompile documentation, which includes a description of the
1359 pcre_pattern_to_host_byte_order() function. However, compiling a regu-
1360 lar expression with one version of PCRE for use with a different ver-
1361 sion is not guaranteed to work and may cause crashes.
1366 int pcre_config(int what, void *where);
1368 The function pcre_config() makes it possible for a PCRE client to dis-
1369 cover which optional features have been compiled into the PCRE library.
1370 The pcrebuild documentation has more details about these optional fea-
1371 tures.
1373 The first argument for pcre_config() is an integer, specifying which
1374 information is required; the second argument is a pointer to a variable
1375 into which the information is placed. The returned value is zero on
1376 success, or the negative error code PCRE_ERROR_BADOPTION if the value
1377 in the first argument is not recognized. The following information is
1378 available:
1382 The output is an integer that is set to one if UTF-8 support is avail-
1383 able; otherwise it is set to zero. If this option is given to the
1384 16-bit version of this function, pcre16_config(), the result is
1389 The output is an integer that is set to one if UTF-16 support is avail-
1390 able; otherwise it is set to zero. This value should normally be given
1391 to the 16-bit version of this function, pcre16_config(). If it is given
1392 to the 8-bit version of this function, the result is PCRE_ERROR_BADOP-
1393 TION.
1397 The output is an integer that is set to one if support for Unicode
1398 character properties is available; otherwise it is set to zero.
1402 The output is an integer that is set to one if support for just-in-time
1403 compiling is available; otherwise it is set to zero.
1407 The output is an integer whose value specifies the default character
1408 sequence that is recognized as meaning "newline". The four values that
1409 are supported are: 10 for LF, 13 for CR, 3338 for CRLF, -2 for ANYCRLF,
1410 and -1 for ANY. Though they are derived from ASCII, the same values
1411 are returned in EBCDIC environments. The default should normally corre-
1412 spond to the standard sequence for your operating system.
1416 The output is an integer whose value indicates what character sequences
1417 the \R escape sequence matches by default. A value of 0 means that \R
1418 matches any Unicode line ending sequence; a value of 1 means that \R
1419 matches only CR, LF, or CRLF. The default can be overridden when a pat-
1420 tern is compiled or matched.
1424 The output is an integer that contains the number of bytes used for
1425 internal linkage in compiled regular expressions. For the 8-bit
1426 library, the value can be 2, 3, or 4. For the 16-bit library, the value
1427 is either 2 or 4 and is still a number of bytes. The default value of 2
1428 is sufficient for all but the most massive patterns, since it allows
1429 the compiled pattern to be up to 64K in size. Larger values allow
1430 larger regular expressions to be compiled, at the expense of slower
1431 matching.
1435 The output is an integer that contains the threshold above which the
1436 POSIX interface uses malloc() for output vectors. Further details are
1437 given in the pcreposix documentation.
1441 The output is a long integer that gives the default limit for the num-
1442 ber of internal matching function calls in a pcre_exec() execution.
1443 Further details are given with pcre_exec() below.
1447 The output is a long integer that gives the default limit for the depth
1448 of recursion when calling the internal matching function in a
1449 pcre_exec() execution. Further details are given with pcre_exec()
1450 below.
1454 The output is an integer that is set to one if internal recursion when
1455 running pcre_exec() is implemented by recursive function calls that use
1456 the stack to remember their state. This is the usual way that PCRE is
1457 compiled. The output is zero if PCRE was compiled to use blocks of data
1458 on the heap instead of recursive function calls. In this case,
1459 pcre_stack_malloc and pcre_stack_free are called to manage memory
1460 blocks on the heap, thus avoiding the use of the stack.
1465 pcre *pcre_compile(const char *pattern, int options,
1466 const char **errptr, int *erroffset,
1467 const unsigned char *tableptr);
1469 pcre *pcre_compile2(const char *pattern, int options,
1470 int *errorcodeptr,
1471 const char **errptr, int *erroffset,
1472 const unsigned char *tableptr);
1474 Either of the functions pcre_compile() or pcre_compile2() can be called
1475 to compile a pattern into an internal form. The only difference between
1476 the two interfaces is that pcre_compile2() has an additional argument,
1477 errorcodeptr, via which a numerical error code can be returned. To
1478 avoid too much repetition, we refer just to pcre_compile() below, but
1479 the information applies equally to pcre_compile2().
1481 The pattern is a C string terminated by a binary zero, and is passed in
1482 the pattern argument. A pointer to a single block of memory that is
1483 obtained via pcre_malloc is returned. This contains the compiled code
1484 and related data. The pcre type is defined for the returned block; this
1485 is a typedef for a structure whose contents are not externally defined.
1486 It is up to the caller to free the memory (via pcre_free) when it is no
1487 longer required.
1489 Although the compiled code of a PCRE regex is relocatable, that is, it
1490 does not depend on memory location, the complete pcre data block is not
1491 fully relocatable, because it may contain a copy of the tableptr argu-
1492 ment, which is an address (see below).
1494 The options argument contains various bit settings that affect the com-
1495 pilation. It should be zero if no options are required. The available
1496 options are described below. Some of them (in particular, those that
1497 are compatible with Perl, but some others as well) can also be set and
1498 unset from within the pattern (see the detailed description in the
1499 pcrepattern documentation). For those options that can be different in
1500 different parts of the pattern, the contents of the options argument
1501 specifies their settings at the start of compilation and execution. The
1503 PCRE_NO_START_OPT options can be set at the time of matching as well as
1504 at compile time.
1506 If errptr is NULL, pcre_compile() returns NULL immediately. Otherwise,
1507 if compilation of a pattern fails, pcre_compile() returns NULL, and
1508 sets the variable pointed to by errptr to point to a textual error mes-
1509 sage. This is a static string that is part of the library. You must not
1510 try to free it. Normally, the offset from the start of the pattern to
1511 the byte that was being processed when the error was discovered is
1512 placed in the variable pointed to by erroffset, which must not be NULL
1513 (if it is, an immediate error is given). However, for an invalid UTF-8
1514 string, the offset is that of the first byte of the failing character.
1516 Some errors are not detected until the whole pattern has been scanned;
1517 in these cases, the offset passed back is the length of the pattern.
1518 Note that the offset is in bytes, not characters, even in UTF-8 mode.
1519 It may sometimes point into the middle of a UTF-8 character.
1521 If pcre_compile2() is used instead of pcre_compile(), and the error-
1522 codeptr argument is not NULL, a non-zero error code number is returned
1523 via this argument in the event of an error. This is in addition to the
1524 textual error message. Error codes and messages are listed below.
1526 If the final argument, tableptr, is NULL, PCRE uses a default set of
1527 character tables that are built when PCRE is compiled, using the
1528 default C locale. Otherwise, tableptr must be an address that is the
1529 result of a call to pcre_maketables(). This value is stored with the
1530 compiled pattern, and used again by pcre_exec(), unless another table
1531 pointer is passed to it. For more discussion, see the section on locale
1532 support below.
1534 This code fragment shows a typical straightforward call to pcre_com-
1535 pile():
1537 pcre *re;
1538 const char *error;
1539 int erroffset;
1540 re = pcre_compile(
1541 "^A.*Z", /* the pattern */
1542 0, /* default options */
1543 &error, /* for error message */
1544 &erroffset, /* for error offset */
1545 NULL); /* use default character tables */
1547 The following names for option bits are defined in the pcre.h header
1548 file:
1552 If this bit is set, the pattern is forced to be "anchored", that is, it
1553 is constrained to match only at the first matching point in the string
1554 that is being searched (the "subject string"). This effect can also be
1555 achieved by appropriate constructs in the pattern itself, which is the
1556 only way to do it in Perl.
1560 If this bit is set, pcre_compile() automatically inserts callout items,
1561 all with number 255, before each pattern item. For discussion of the
1562 callout facility, see the pcrecallout documentation.
1567 These options (which are mutually exclusive) control what the \R escape
1568 sequence matches. The choice is either to match only CR, LF, or CRLF,
1569 or to match any Unicode newline sequence. The default is specified when
1570 PCRE is built. It can be overridden from within the pattern, or by set-
1571 ting an option when a compiled pattern is matched.
1575 If this bit is set, letters in the pattern match both upper and lower
1576 case letters. It is equivalent to Perl's /i option, and it can be
1577 changed within a pattern by a (?i) option setting. In UTF-8 mode, PCRE
1578 always understands the concept of case for characters whose values are
1579 less than 128, so caseless matching is always possible. For characters
1580 with higher values, the concept of case is supported if PCRE is com-
1581 piled with Unicode property support, but not otherwise. If you want to
1582 use caseless matching for characters 128 and above, you must ensure
1583 that PCRE is compiled with Unicode property support as well as with
1584 UTF-8 support.
1588 If this bit is set, a dollar metacharacter in the pattern matches only
1589 at the end of the subject string. Without this option, a dollar also
1590 matches immediately before a newline at the end of the string (but not
1591 before any other newlines). The PCRE_DOLLAR_ENDONLY option is ignored
1592 if PCRE_MULTILINE is set. There is no equivalent to this option in
1593 Perl, and no way to set it within a pattern.
1597 If this bit is set, a dot metacharacter in the pattern matches a char-
1598 acter of any value, including one that indicates a newline. However, it
1599 only ever matches one character, even if newlines are coded as CRLF.
1600 Without this option, a dot does not match when the current position is
1601 at a newline. This option is equivalent to Perl's /s option, and it can
1602 be changed within a pattern by a (?s) option setting. A negative class
1603 such as [^a] always matches newline characters, independent of the set-
1604 ting of this option.
1608 If this bit is set, names used to identify capturing subpatterns need
1609 not be unique. This can be helpful for certain types of pattern when it
1610 is known that only one instance of the named subpattern can ever be
1611 matched. There are more details of named subpatterns below; see also
1612 the pcrepattern documentation.
1616 If this bit is set, whitespace data characters in the pattern are
1617 totally ignored except when escaped or inside a character class. White-
1618 space does not include the VT character (code 11). In addition, charac-
1619 ters between an unescaped # outside a character class and the next new-
1620 line, inclusive, are also ignored. This is equivalent to Perl's /x
1621 option, and it can be changed within a pattern by a (?x) option set-
1622 ting.
1624 Which characters are interpreted as newlines is controlled by the
1625 options passed to pcre_compile() or by a special sequence at the start
1626 of the pattern, as described in the section entitled "Newline conven-
1627 tions" in the pcrepattern documentation. Note that the end of this type
1628 of comment is a literal newline sequence in the pattern; escape
1629 sequences that happen to represent a newline do not count.
1631 This option makes it possible to include comments inside complicated
1632 patterns. Note, however, that this applies only to data characters.
1633 Whitespace characters may never appear within special character
1634 sequences in a pattern, for example within the sequence (?( that intro-
1635 duces a conditional subpattern.
1639 This option was invented in order to turn on additional functionality
1640 of PCRE that is incompatible with Perl, but it is currently of very
1641 little use. When set, any backslash in a pattern that is followed by a
1642 letter that has no special meaning causes an error, thus reserving
1643 these combinations for future expansion. By default, as in Perl, a
1644 backslash followed by a letter with no special meaning is treated as a
1645 literal. (Perl can, however, be persuaded to give an error for this, by
1646 running it with the -w option.) There are at present no other features
1647 controlled by this option. It can also be set by a (?X) option setting
1648 within a pattern.
1652 If this option is set, an unanchored pattern is required to match
1653 before or at the first newline in the subject string, though the
1654 matched text may continue over the newline.
1658 If this option is set, PCRE's behaviour is changed in some ways so that
1659 it is compatible with JavaScript rather than Perl. The changes are as
1660 follows:
1662 (1) A lone closing square bracket in a pattern causes a compile-time
1663 error, because this is illegal in JavaScript (by default it is treated
1664 as a data character). Thus, the pattern AB]CD becomes illegal when this
1665 option is set.
1667 (2) At run time, a back reference to an unset subpattern group matches
1668 an empty string (by default this causes the current matching alterna-
1669 tive to fail). A pattern such as (\1)(a) succeeds when this option is
1670 set (assuming it can find an "a" in the subject), whereas it fails by
1671 default, for Perl compatibility.
1673 (3) \U matches an upper case "U" character; by default \U causes a com-
1674 pile time error (Perl uses \U to upper case subsequent characters).
1676 (4) \u matches a lower case "u" character unless it is followed by four
1677 hexadecimal digits, in which case the hexadecimal number defines the
1678 code point to match. By default, \u causes a compile time error (Perl
1679 uses it to upper case the following character).
1681 (5) \x matches a lower case "x" character unless it is followed by two
1682 hexadecimal digits, in which case the hexadecimal number defines the
1683 code point to match. By default, as in Perl, a hexadecimal number is
1684 always expected after \x, but it may have zero, one, or two digits (so,
1685 for example, \xz matches a binary zero character followed by z).
1689 By default, PCRE treats the subject string as consisting of a single
1690 line of characters (even if it actually contains newlines). The "start
1691 of line" metacharacter (^) matches only at the start of the string,
1692 while the "end of line" metacharacter ($) matches only at the end of
1693 the string, or before a terminating newline (unless PCRE_DOLLAR_ENDONLY
1694 is set). This is the same as Perl.
1696 When PCRE_MULTILINE it is set, the "start of line" and "end of line"
1697 constructs match immediately following or immediately before internal
1698 newlines in the subject string, respectively, as well as at the very
1699 start and end. This is equivalent to Perl's /m option, and it can be
1700 changed within a pattern by a (?m) option setting. If there are no new-
1701 lines in a subject string, or no occurrences of ^ or $ in a pattern,
1702 setting PCRE_MULTILINE has no effect.
1710 These options override the default newline definition that was chosen
1711 when PCRE was built. Setting the first or the second specifies that a
1712 newline is indicated by a single character (CR or LF, respectively).
1713 Setting PCRE_NEWLINE_CRLF specifies that a newline is indicated by the
1714 two-character CRLF sequence. Setting PCRE_NEWLINE_ANYCRLF specifies
1715 that any of the three preceding sequences should be recognized. Setting
1716 PCRE_NEWLINE_ANY specifies that any Unicode newline sequence should be
1717 recognized. The Unicode newline sequences are the three just mentioned,
1718 plus the single characters VT (vertical tab, U+000B), FF (formfeed,
1719 U+000C), NEL (next line, U+0085), LS (line separator, U+2028), and PS
1720 (paragraph separator, U+2029). For the 8-bit library, the last two are
1721 recognized only in UTF-8 mode.
1723 The newline setting in the options word uses three bits that are
1724 treated as a number, giving eight possibilities. Currently only six are
1725 used (default plus the five values above). This means that if you set
1726 more than one newline option, the combination may or may not be sensi-
1727 ble. For example, PCRE_NEWLINE_CR with PCRE_NEWLINE_LF is equivalent to
1728 PCRE_NEWLINE_CRLF, but other combinations may yield unused numbers and
1729 cause an error.
1731 The only time that a line break in a pattern is specially recognized
1732 when compiling is when PCRE_EXTENDED is set. CR and LF are whitespace
1733 characters, and so are ignored in this mode. Also, an unescaped # out-
1734 side a character class indicates a comment that lasts until after the
1735 next line break sequence. In other circumstances, line break sequences
1736 in patterns are treated as literal data.
1738 The newline option that is set at compile time becomes the default that
1739 is used for pcre_exec() and pcre_dfa_exec(), but it can be overridden.
1743 If this option is set, it disables the use of numbered capturing paren-
1744 theses in the pattern. Any opening parenthesis that is not followed by
1745 ? behaves as if it were followed by ?: but named parentheses can still
1746 be used for capturing (and they acquire numbers in the usual way).
1747 There is no equivalent of this option in Perl.
1751 This is an option that acts at matching time; that is, it is really an
1752 option for pcre_exec() or pcre_dfa_exec(). If it is set at compile
1753 time, it is remembered with the compiled pattern and assumed at match-
1754 ing time. For details see the discussion of PCRE_NO_START_OPTIMIZE
1755 below.
1759 This option changes the way PCRE processes \B, \b, \D, \d, \S, \s, \W,
1760 \w, and some of the POSIX character classes. By default, only ASCII
1761 characters are recognized, but if PCRE_UCP is set, Unicode properties
1762 are used instead to classify characters. More details are given in the
1763 section on generic character types in the pcrepattern page. If you set
1764 PCRE_UCP, matching one of the items it affects takes much longer. The
1765 option is available only if PCRE has been compiled with Unicode prop-
1766 erty support.
1770 This option inverts the "greediness" of the quantifiers so that they
1771 are not greedy by default, but become greedy if followed by "?". It is
1772 not compatible with Perl. It can also be set by a (?U) option setting
1773 within the pattern.
1775 PCRE_UTF8
1777 This option causes PCRE to regard both the pattern and the subject as
1778 strings of UTF-8 characters instead of single-byte strings. However, it
1779 is available only when PCRE is built to include UTF support. If not,
1780 the use of this option provokes an error. Details of how this option
1781 changes the behaviour of PCRE are given in the pcreunicode page.
1785 When PCRE_UTF8 is set, the validity of the pattern as a UTF-8 string is
1786 automatically checked. There is a discussion about the validity of
1787 UTF-8 strings in the pcreunicode page. If an invalid UTF-8 sequence is
1788 found, pcre_compile() returns an error. If you already know that your
1789 pattern is valid, and you want to skip this check for performance rea-
1790 sons, you can set the PCRE_NO_UTF8_CHECK option. When it is set, the
1791 effect of passing an invalid UTF-8 string as a pattern is undefined. It
1792 may cause your program to crash. Note that this option can also be
1793 passed to pcre_exec() and pcre_dfa_exec(), to suppress the validity
1794 checking of subject strings.
1799 The following table lists the error codes than may be returned by
1800 pcre_compile2(), along with the error messages that may be returned by
1801 both compiling functions. Note that error messages are always 8-bit
1802 ASCII strings, even in 16-bit mode. As PCRE has developed, some error
1803 codes have fallen out of use. To avoid confusion, they have not been
1804 re-used.
1806 0 no error
1807 1 \ at end of pattern
1808 2 \c at end of pattern
1809 3 unrecognized character follows \
1810 4 numbers out of order in {} quantifier
1811 5 number too big in {} quantifier
1812 6 missing terminating ] for character class
1813 7 invalid escape sequence in character class
1814 8 range out of order in character class
1815 9 nothing to repeat
1816 10 [this code is not in use]
1817 11 internal error: unexpected repeat
1818 12 unrecognized character after (? or (?-
1819 13 POSIX named classes are supported only within a class
1820 14 missing )
1821 15 reference to non-existent subpattern
1822 16 erroffset passed as NULL
1823 17 unknown option bit(s) set
1824 18 missing ) after comment
1825 19 [this code is not in use]
1826 20 regular expression is too large
1827 21 failed to get memory
1828 22 unmatched parentheses
1829 23 internal error: code overflow
1830 24 unrecognized character after (?<
1831 25 lookbehind assertion is not fixed length
1832 26 malformed number or name after (?(
1833 27 conditional group contains more than two branches
1834 28 assertion expected after (?(
1835 29 (?R or (?[+-]digits must be followed by )
1836 30 unknown POSIX class name
1837 31 POSIX collating elements are not supported
1838 32 this version of PCRE is compiled without UTF support
1839 33 [this code is not in use]
1840 34 character value in \x{...} sequence is too large
1841 35 invalid condition (?(0)
1842 36 \C not allowed in lookbehind assertion
1843 37 PCRE does not support \L, \l, \N{name}, \U, or \u
1844 38 number after (?C is > 255
1845 39 closing ) for (?C expected
1846 40 recursive call could loop indefinitely
1847 41 unrecognized character after (?P
1848 42 syntax error in subpattern name (missing terminator)
1849 43 two named subpatterns have the same name
1850 44 invalid UTF-8 string (specifically UTF-8)
1851 45 support for \P, \p, and \X has not been compiled
1852 46 malformed \P or \p sequence
1853 47 unknown property name after \P or \p
1854 48 subpattern name is too long (maximum 32 characters)
1855 49 too many named subpatterns (maximum 10000)
1856 50 [this code is not in use]
1857 51 octal value is greater than \377 in 8-bit non-UTF-8 mode
1858 52 internal error: overran compiling workspace
1859 53 internal error: previously-checked referenced subpattern
1860 not found
1861 54 DEFINE group contains more than one branch
1862 55 repeating a DEFINE group is not allowed
1863 56 inconsistent NEWLINE options
1864 57 \g is not followed by a braced, angle-bracketed, or quoted
1865 name/number or by a plain number
1866 58 a numbered reference must not be zero
1867 59 an argument is not allowed for (*ACCEPT), (*FAIL), or (*COMMIT)
1868 60 (*VERB) not recognized
1869 61 number is too big
1870 62 subpattern name expected
1871 63 digit expected after (?+
1872 64 ] is an invalid data character in JavaScript compatibility mode
1873 65 different names for subpatterns of the same number are
1874 not allowed
1875 66 (*MARK) must have an argument
1876 67 this version of PCRE is not compiled with Unicode property
1877 support
1878 68 \c must be followed by an ASCII character
1879 69 \k is not followed by a braced, angle-bracketed, or quoted name
1880 70 internal error: unknown opcode in find_fixedlength()
1881 71 \N is not supported in a class
1882 72 too many forward references
1883 73 disallowed Unicode code point (>= 0xd800 && <= 0xdfff)
1884 74 invalid UTF-16 string (specifically UTF-16)
1886 The numbers 32 and 10000 in errors 48 and 49 are defaults; different
1887 values may be used if the limits were changed when PCRE was built.
1892 pcre_extra *pcre_study(const pcre *code, int options
1893 const char **errptr);
1895 If a compiled pattern is going to be used several times, it is worth
1896 spending more time analyzing it in order to speed up the time taken for
1897 matching. The function pcre_study() takes a pointer to a compiled pat-
1898 tern as its first argument. If studying the pattern produces additional
1899 information that will help speed up matching, pcre_study() returns a
1900 pointer to a pcre_extra block, in which the study_data field points to
1901 the results of the study.
1903 The returned value from pcre_study() can be passed directly to
1904 pcre_exec() or pcre_dfa_exec(). However, a pcre_extra block also con-
1905 tains other fields that can be set by the caller before the block is
1906 passed; these are described below in the section on matching a pattern.
1908 If studying the pattern does not produce any useful information,
1909 pcre_study() returns NULL. In that circumstance, if the calling program
1910 wants to pass any of the other fields to pcre_exec() or
1911 pcre_dfa_exec(), it must set up its own pcre_extra block.
1913 The second argument of pcre_study() contains option bits. There is only
1914 one option: PCRE_STUDY_JIT_COMPILE. If this is set, and the just-in-
1915 time compiler is available, the pattern is further compiled into
1916 machine code that executes much faster than the pcre_exec() matching
1917 function. If the just-in-time compiler is not available, this option is
1918 ignored. All other bits in the options argument must be zero.
1920 JIT compilation is a heavyweight optimization. It can take some time
1921 for patterns to be analyzed, and for one-off matches and simple pat-
1922 terns the benefit of faster execution might be offset by a much slower
1923 study time. Not all patterns can be optimized by the JIT compiler. For
1924 those that cannot be handled, matching automatically falls back to the
1925 pcre_exec() interpreter. For more details, see the pcrejit documenta-
1926 tion.
1928 The third argument for pcre_study() is a pointer for an error message.
1929 If studying succeeds (even if no data is returned), the variable it
1930 points to is set to NULL. Otherwise it is set to point to a textual
1931 error message. This is a static string that is part of the library. You
1932 must not try to free it. You should test the error pointer for NULL
1933 after calling pcre_study(), to be sure that it has run successfully.
1935 When you are finished with a pattern, you can free the memory used for
1936 the study data by calling pcre_free_study(). This function was added to
1937 the API for release 8.20. For earlier versions, the memory could be
1938 freed with pcre_free(), just like the pattern itself. This will still
1939 work in cases where PCRE_STUDY_JIT_COMPILE is not used, but it is
1940 advisable to change to the new function when convenient.
1942 This is a typical way in which pcre_study() is used (except that in a
1943 real application there should be tests for errors):
1945 int rc;
1946 pcre *re;
1947 pcre_extra *sd;
1948 re = pcre_compile("pattern", 0, &error, &erroroffset, NULL);
1949 sd = pcre_study(
1950 re, /* result of pcre_compile() */
1951 0, /* no options */
1952 &error); /* set to NULL or points to a message */
1953 rc = pcre_exec( /* see below for details of pcre_exec() options */
1954 re, sd, "subject", 7, 0, 0, ovector, 30);
1955 ...
1956 pcre_free_study(sd);
1957 pcre_free(re);
1959 Studying a pattern does two things: first, a lower bound for the length
1960 of subject string that is needed to match the pattern is computed. This
1961 does not mean that there are any strings of that length that match, but
1962 it does guarantee that no shorter strings match. The value is used by
1963 pcre_exec() and pcre_dfa_exec() to avoid wasting time by trying to
1964 match strings that are shorter than the lower bound. You can find out
1965 the value in a calling program via the pcre_fullinfo() function.
1967 Studying a pattern is also useful for non-anchored patterns that do not
1968 have a single fixed starting character. A bitmap of possible starting
1969 bytes is created. This speeds up finding a position in the subject at
1970 which to start matching. (In 16-bit mode, the bitmap is used for 16-bit
1971 values less than 256.)
1973 These two optimizations apply to both pcre_exec() and pcre_dfa_exec().
1974 However, they are not used by pcre_exec() if pcre_study() is called
1975 with the PCRE_STUDY_JIT_COMPILE option, and just-in-time compiling is
1976 successful. The optimizations can be disabled by setting the
1977 PCRE_NO_START_OPTIMIZE option when calling pcre_exec() or
1978 pcre_dfa_exec(). You might want to do this if your pattern contains
1979 callouts or (*MARK) (which cannot be handled by the JIT compiler), and
1980 you want to make use of these facilities in cases where matching fails.
1981 See the discussion of PCRE_NO_START_OPTIMIZE below.
1986 PCRE handles caseless matching, and determines whether characters are
1987 letters, digits, or whatever, by reference to a set of tables, indexed
1988 by character value. When running in UTF-8 mode, this applies only to
1989 characters with codes less than 128. By default, higher-valued codes
1990 never match escapes such as \w or \d, but they can be tested with \p if
1991 PCRE is built with Unicode character property support. Alternatively,
1992 the PCRE_UCP option can be set at compile time; this causes \w and
1993 friends to use Unicode property support instead of built-in tables. The
1994 use of locales with Unicode is discouraged. If you are handling charac-
1995 ters with codes greater than 128, you should either use UTF-8 and Uni-
1996 code, or use locales, but not try to mix the two.
1998 PCRE contains an internal set of tables that are used when the final
1999 argument of pcre_compile() is NULL. These are sufficient for many
2000 applications. Normally, the internal tables recognize only ASCII char-
2001 acters. However, when PCRE is built, it is possible to cause the inter-
2002 nal tables to be rebuilt in the default "C" locale of the local system,
2003 which may cause them to be different.
2005 The internal tables can always be overridden by tables supplied by the
2006 application that calls PCRE. These may be created in a different locale
2007 from the default. As more and more applications change to using Uni-
2008 code, the need for this locale support is expected to die away.
2010 External tables are built by calling the pcre_maketables() function,
2011 which has no arguments, in the relevant locale. The result can then be
2012 passed to pcre_compile() or pcre_exec() as often as necessary. For
2013 example, to build and use tables that are appropriate for the French
2014 locale (where accented characters with values greater than 128 are
2015 treated as letters), the following code could be used:
2017 setlocale(LC_CTYPE, "fr_FR");
2018 tables = pcre_maketables();
2019 re = pcre_compile(..., tables);
2021 The locale name "fr_FR" is used on Linux and other Unix-like systems;
2022 if you are using Windows, the name for the French locale is "french".
2024 When pcre_maketables() runs, the tables are built in memory that is
2025 obtained via pcre_malloc. It is the caller's responsibility to ensure
2026 that the memory containing the tables remains available for as long as
2027 it is needed.
2029 The pointer that is passed to pcre_compile() is saved with the compiled
2030 pattern, and the same tables are used via this pointer by pcre_study()
2031 and normally also by pcre_exec(). Thus, by default, for any single pat-
2032 tern, compilation, studying and matching all happen in the same locale,
2033 but different patterns can be compiled in different locales.
2035 It is possible to pass a table pointer or NULL (indicating the use of
2036 the internal tables) to pcre_exec(). Although not intended for this
2037 purpose, this facility could be used to match a pattern in a different
2038 locale from the one in which it was compiled. Passing table pointers at
2039 run time is discussed below in the section on matching a pattern.
2044 int pcre_fullinfo(const pcre *code, const pcre_extra *extra,
2045 int what, void *where);
2047 The pcre_fullinfo() function returns information about a compiled pat-
2048 tern. It replaces the pcre_info() function, which was removed from the
2049 library at version 8.30, after more than 10 years of obsolescence.
2051 The first argument for pcre_fullinfo() is a pointer to the compiled
2052 pattern. The second argument is the result of pcre_study(), or NULL if
2053 the pattern was not studied. The third argument specifies which piece
2054 of information is required, and the fourth argument is a pointer to a
2055 variable to receive the data. The yield of the function is zero for
2056 success, or one of the following negative numbers:
2058 PCRE_ERROR_NULL the argument code was NULL
2059 the argument where was NULL
2060 PCRE_ERROR_BADMAGIC the "magic number" was not found
2061 PCRE_ERROR_BADENDIANNESS the pattern was compiled with different
2062 endianness
2063 PCRE_ERROR_BADOPTION the value of what was invalid
2065 The "magic number" is placed at the start of each compiled pattern as
2066 an simple check against passing an arbitrary memory pointer. The endi-
2067 anness error can occur if a compiled pattern is saved and reloaded on a
2068 different host. Here is a typical call of pcre_fullinfo(), to obtain
2069 the length of the compiled pattern:
2071 int rc;
2072 size_t length;
2073 rc = pcre_fullinfo(
2074 re, /* result of pcre_compile() */
2075 sd, /* result of pcre_study(), or NULL */
2076 PCRE_INFO_SIZE, /* what is required */
2077 &length); /* where to put the data */
2079 The possible values for the third argument are defined in pcre.h, and
2080 are as follows:
2084 Return the number of the highest back reference in the pattern. The
2085 fourth argument should point to an int variable. Zero is returned if
2086 there are no back references.
2090 Return the number of capturing subpatterns in the pattern. The fourth
2091 argument should point to an int variable.
2095 Return a pointer to the internal default character tables within PCRE.
2096 The fourth argument should point to an unsigned char * variable. This
2097 information call is provided for internal use by the pcre_study() func-
2098 tion. External callers can cause PCRE to use its internal tables by
2099 passing a NULL table pointer.
2103 Return information about the first data unit of any matched string, for
2104 a non-anchored pattern. (The name of this option refers to the 8-bit
2105 library, where data units are bytes.) The fourth argument should point
2106 to an int variable.
2108 If there is a fixed first value, for example, the letter "c" from a
2109 pattern such as (cat|cow|coyote), its value is returned. In the 8-bit
2110 library, the value is always less than 256; in the 16-bit library the
2111 value can be up to 0xffff.
2113 If there is no fixed first value, and if either
2115 (a) the pattern was compiled with the PCRE_MULTILINE option, and every
2116 branch starts with "^", or
2118 (b) every branch of the pattern starts with ".*" and PCRE_DOTALL is not
2119 set (if it were set, the pattern would be anchored),
2121 -1 is returned, indicating that the pattern matches only at the start
2122 of a subject string or after any newline within the string. Otherwise
2123 -2 is returned. For anchored patterns, -2 is returned.
2127 If the pattern was studied, and this resulted in the construction of a
2128 256-bit table indicating a fixed set of values for the first data unit
2129 in any matching string, a pointer to the table is returned. Otherwise
2130 NULL is returned. The fourth argument should point to an unsigned char
2131 * variable.
2135 Return 1 if the pattern contains any explicit matches for CR or LF
2136 characters, otherwise 0. The fourth argument should point to an int
2137 variable. An explicit match is either a literal CR or LF character, or
2138 \r or \n.
2142 Return 1 if the (?J) or (?-J) option setting is used in the pattern,
2143 otherwise 0. The fourth argument should point to an int variable. (?J)
2144 and (?-J) set and unset the local PCRE_DUPNAMES option, respectively.
2148 Return 1 if the pattern was studied with the PCRE_STUDY_JIT_COMPILE
2149 option, and just-in-time compiling was successful. The fourth argument
2150 should point to an int variable. A return value of 0 means that JIT
2151 support is not available in this version of PCRE, or that the pattern
2152 was not studied with the PCRE_STUDY_JIT_COMPILE option, or that the JIT
2153 compiler could not handle this particular pattern. See the pcrejit doc-
2154 umentation for details of what can and cannot be handled.
2158 If the pattern was successfully studied with the PCRE_STUDY_JIT_COMPILE
2159 option, return the size of the JIT compiled code, otherwise return
2160 zero. The fourth argument should point to a size_t variable.
2164 Return the value of the rightmost literal data unit that must exist in
2165 any matched string, other than at its start, if such a value has been
2166 recorded. The fourth argument should point to an int variable. If there
2167 is no such value, -1 is returned. For anchored patterns, a last literal
2168 value is recorded only if it follows something of variable length. For
2169 example, for the pattern /^a\d+z\d+/ the returned value is "z", but for
2170 /^a\dz\d/ the returned value is -1.
2174 If the pattern was studied and a minimum length for matching subject
2175 strings was computed, its value is returned. Otherwise the returned
2176 value is -1. The value is a number of characters, which in UTF-8 mode
2177 may be different from the number of bytes. The fourth argument should
2178 point to an int variable. A non-negative value is a lower bound to the
2179 length of any matching string. There may not be any strings of that
2180 length that do actually match, but every string that does match is at
2181 least that long.
2187 PCRE supports the use of named as well as numbered capturing parenthe-
2188 ses. The names are just an additional way of identifying the parenthe-
2189 ses, which still acquire numbers. Several convenience functions such as
2190 pcre_get_named_substring() are provided for extracting captured sub-
2191 strings by name. It is also possible to extract the data directly, by
2192 first converting the name to a number in order to access the correct
2193 pointers in the output vector (described with pcre_exec() below). To do
2194 the conversion, you need to use the name-to-number map, which is
2195 described by these three values.
2197 The map consists of a number of fixed-size entries. PCRE_INFO_NAMECOUNT
2198 gives the number of entries, and PCRE_INFO_NAMEENTRYSIZE gives the size
2199 of each entry; both of these return an int value. The entry size
2200 depends on the length of the longest name. PCRE_INFO_NAMETABLE returns
2201 a pointer to the first entry of the table. This is a pointer to char in
2202 the 8-bit library, where the first two bytes of each entry are the num-
2203 ber of the capturing parenthesis, most significant byte first. In the
2204 16-bit library, the pointer points to 16-bit data units, the first of
2205 which contains the parenthesis number. The rest of the entry is the
2206 corresponding name, zero terminated.
2208 The names are in alphabetical order. Duplicate names may appear if (?|
2209 is used to create multiple groups with the same number, as described in
2210 the section on duplicate subpattern numbers in the pcrepattern page.
2211 Duplicate names for subpatterns with different numbers are permitted
2212 only if PCRE_DUPNAMES is set. In all cases of duplicate names, they
2213 appear in the table in the order in which they were found in the pat-
2214 tern. In the absence of (?| this is the order of increasing number;
2215 when (?| is used this is not necessarily the case because later subpat-
2216 terns may have lower numbers.
2218 As a simple example of the name/number table, consider the following
2219 pattern after compilation by the 8-bit library (assume PCRE_EXTENDED is
2220 set, so white space - including newlines - is ignored):
2222 (?<date> (?<year>(\d\d)?\d\d) -
2223 (?<month>\d\d) - (?<day>\d\d) )
2225 There are four named subpatterns, so the table has four entries, and
2226 each entry in the table is eight bytes long. The table is as follows,
2227 with non-printing bytes shows in hexadecimal, and undefined bytes shown
2228 as ??:
2230 00 01 d a t e 00 ??
2231 00 05 d a y 00 ?? ??
2232 00 04 m o n t h 00
2233 00 02 y e a r 00 ??
2235 When writing code to extract data from named subpatterns using the
2236 name-to-number map, remember that the length of the entries is likely
2237 to be different for each compiled pattern.
2241 Return 1 if the pattern can be used for partial matching with
2242 pcre_exec(), otherwise 0. The fourth argument should point to an int
2243 variable. From release 8.00, this always returns 1, because the
2244 restrictions that previously applied to partial matching have been
2245 lifted. The pcrepartial documentation gives details of partial match-
2246 ing.
2250 Return a copy of the options with which the pattern was compiled. The
2251 fourth argument should point to an unsigned long int variable. These
2252 option bits are those specified in the call to pcre_compile(), modified
2253 by any top-level option settings at the start of the pattern itself. In
2254 other words, they are the options that will be in force when matching
2255 starts. For example, if the pattern /(?im)abc(?-i)d/ is compiled with
2256 the PCRE_EXTENDED option, the result is PCRE_CASELESS, PCRE_MULTILINE,
2259 A pattern is automatically anchored by PCRE if all of its top-level
2260 alternatives begin with one of the following:
2262 ^ unless PCRE_MULTILINE is set
2263 \A always
2264 \G always
2265 .* if PCRE_DOTALL is set and there are no back
2266 references to the subpattern in which .* appears
2268 For such patterns, the PCRE_ANCHORED bit is set in the options returned
2269 by pcre_fullinfo().
2273 Return the size of the compiled pattern in bytes (for both libraries).
2274 The fourth argument should point to a size_t variable. This value does
2275 not include the size of the pcre structure that is returned by
2276 pcre_compile(). The value that is passed as the argument to pcre_mal-
2277 loc() when pcre_compile() is getting memory in which to place the com-
2278 piled data is the value returned by this option plus the size of the
2279 pcre structure. Studying a compiled pattern, with or without JIT, does
2280 not alter the value returned by this option.
2284 Return the size in bytes of the data block pointed to by the study_data
2285 field in a pcre_extra block. If pcre_extra is NULL, or there is no
2286 study data, zero is returned. The fourth argument should point to a
2287 size_t variable. The study_data field is set by pcre_study() to record
2288 information that will speed up matching (see the section entitled
2289 "Studying a pattern" above). The format of the study_data block is pri-
2290 vate, but its length is made available via this option so that it can
2291 be saved and restored (see the pcreprecompile documentation for
2292 details).
2297 int pcre_refcount(pcre *code, int adjust);
2299 The pcre_refcount() function is used to maintain a reference count in
2300 the data block that contains a compiled pattern. It is provided for the
2301 benefit of applications that operate in an object-oriented manner,
2302 where different parts of the application may be using the same compiled
2303 pattern, but you want to free the block when they are all done.
2305 When a pattern is compiled, the reference count field is initialized to
2306 zero. It is changed only by calling this function, whose action is to
2307 add the adjust value (which may be positive or negative) to it. The
2308 yield of the function is the new value. However, the value of the count
2309 is constrained to lie between 0 and 65535, inclusive. If the new value
2310 is outside these limits, it is forced to the appropriate limit value.
2312 Except when it is zero, the reference count is not correctly preserved
2313 if a pattern is compiled on one host and then transferred to a host
2314 whose byte-order is different. (This seems a highly unlikely scenario.)
2319 int pcre_exec(const pcre *code, const pcre_extra *extra,
2320 const char *subject, int length, int startoffset,
2321 int options, int *ovector, int ovecsize);
2323 The function pcre_exec() is called to match a subject string against a
2324 compiled pattern, which is passed in the code argument. If the pattern
2325 was studied, the result of the study should be passed in the extra
2326 argument. You can call pcre_exec() with the same code and extra argu-
2327 ments as many times as you like, in order to match different subject
2328 strings with the same pattern.
2330 This function is the main matching facility of the library, and it
2331 operates in a Perl-like manner. For specialist use there is also an
2332 alternative matching function, which is described below in the section
2333 about the pcre_dfa_exec() function.
2335 In most applications, the pattern will have been compiled (and option-
2336 ally studied) in the same process that calls pcre_exec(). However, it
2337 is possible to save compiled patterns and study data, and then use them
2338 later in different processes, possibly even on different hosts. For a
2339 discussion about this, see the pcreprecompile documentation.
2341 Here is an example of a simple call to pcre_exec():
2343 int rc;
2344 int ovector[30];
2345 rc = pcre_exec(
2346 re, /* result of pcre_compile() */
2347 NULL, /* we didn't study the pattern */
2348 "some string", /* the subject string */
2349 11, /* the length of the subject string */
2350 0, /* start at offset 0 in the subject */
2351 0, /* default options */
2352 ovector, /* vector of integers for substring information */
2353 30); /* number of elements (NOT size in bytes) */
2355 Extra data for pcre_exec()
2357 If the extra argument is not NULL, it must point to a pcre_extra data
2358 block. The pcre_study() function returns such a block (when it doesn't
2359 return NULL), but you can also create one for yourself, and pass addi-
2360 tional information in it. The pcre_extra block contains the following
2361 fields (not necessarily in this order):
2363 unsigned long int flags;
2364 void *study_data;
2365 void *executable_jit;
2366 unsigned long int match_limit;
2367 unsigned long int match_limit_recursion;
2368 void *callout_data;
2369 const unsigned char *tables;
2370 unsigned char **mark;
2372 In the 16-bit version of this structure, the mark field has type
2373 "PCRE_UCHAR16 **".
2375 The flags field is a bitmap that specifies which of the other fields
2376 are set. The flag bits are:
2386 Other flag bits should be set to zero. The study_data field and some-
2387 times the executable_jit field are set in the pcre_extra block that is
2388 returned by pcre_study(), together with the appropriate flag bits. You
2389 should not set these yourself, but you may add to the block by setting
2390 the other fields and their corresponding flag bits.
2392 The match_limit field provides a means of preventing PCRE from using up
2393 a vast amount of resources when running patterns that are not going to
2394 match, but which have a very large number of possibilities in their
2395 search trees. The classic example is a pattern that uses nested unlim-
2396 ited repeats.
2398 Internally, pcre_exec() uses a function called match(), which it calls
2399 repeatedly (sometimes recursively). The limit set by match_limit is
2400 imposed on the number of times this function is called during a match,
2401 which has the effect of limiting the amount of backtracking that can
2402 take place. For patterns that are not anchored, the count restarts from
2403 zero for each position in the subject string.
2405 When pcre_exec() is called with a pattern that was successfully studied
2406 with the PCRE_STUDY_JIT_COMPILE option, the way that the matching is
2407 executed is entirely different. However, there is still the possibility
2408 of runaway matching that goes on for a very long time, and so the
2409 match_limit value is also used in this case (but in a different way) to
2410 limit how long the matching can continue.
2412 The default value for the limit can be set when PCRE is built; the
2413 default default is 10 million, which handles all but the most extreme
2414 cases. You can override the default by suppling pcre_exec() with a
2415 pcre_extra block in which match_limit is set, and
2416 PCRE_EXTRA_MATCH_LIMIT is set in the flags field. If the limit is
2417 exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
2419 The match_limit_recursion field is similar to match_limit, but instead
2420 of limiting the total number of times that match() is called, it limits
2421 the depth of recursion. The recursion depth is a smaller number than
2422 the total number of calls, because not all calls to match() are recur-
2423 sive. This limit is of use only if it is set smaller than match_limit.
2425 Limiting the recursion depth limits the amount of machine stack that
2426 can be used, or, when PCRE has been compiled to use memory on the heap
2427 instead of the stack, the amount of heap memory that can be used. This
2428 limit is not relevant, and is ignored, if the pattern was successfully
2429 studied with PCRE_STUDY_JIT_COMPILE.
2431 The default value for match_limit_recursion can be set when PCRE is
2432 built; the default default is the same value as the default for
2433 match_limit. You can override the default by suppling pcre_exec() with
2434 a pcre_extra block in which match_limit_recursion is set, and
2435 PCRE_EXTRA_MATCH_LIMIT_RECURSION is set in the flags field. If the
2436 limit is exceeded, pcre_exec() returns PCRE_ERROR_RECURSIONLIMIT.
2438 The callout_data field is used in conjunction with the "callout" fea-
2439 ture, and is described in the pcrecallout documentation.
2441 The tables field is used to pass a character tables pointer to
2442 pcre_exec(); this overrides the value that is stored with the compiled
2443 pattern. A non-NULL value is stored with the compiled pattern only if
2444 custom tables were supplied to pcre_compile() via its tableptr argu-
2445 ment. If NULL is passed to pcre_exec() using this mechanism, it forces
2446 PCRE's internal tables to be used. This facility is helpful when re-
2447 using patterns that have been saved after compiling with an external
2448 set of tables, because the external tables might be at a different
2449 address when pcre_exec() is called. See the pcreprecompile documenta-
2450 tion for a discussion of saving compiled patterns for later use.
2452 If PCRE_EXTRA_MARK is set in the flags field, the mark field must be
2453 set to point to a suitable variable. If the pattern contains any back-
2454 tracking control verbs such as (*MARK:NAME), and the execution ends up
2455 with a name to pass back, a pointer to the name string (zero termi-
2456 nated) is placed in the variable pointed to by the mark field. The
2457 names are within the compiled pattern; if you wish to retain such a
2458 name you must copy it before freeing the memory of a compiled pattern.
2459 If there is no name to pass back, the variable pointed to by the mark
2460 field is set to NULL. For details of the backtracking control verbs,
2461 see the section entitled "Backtracking control" in the pcrepattern doc-
2462 umentation.
2464 Option bits for pcre_exec()
2466 The unused bits of the options argument for pcre_exec() must be zero.
2467 The only bits that may be set are PCRE_ANCHORED, PCRE_NEWLINE_xxx,
2472 If the pattern was successfully studied with the PCRE_STUDY_JIT_COMPILE
2473 option, the only supported options for JIT execution are
2475 PCRE_NOTEMPTY_ATSTART. Note in particular that partial matching is not
2476 supported. If an unsupported option is used, JIT execution is disabled
2477 and the normal interpretive code in pcre_exec() is run.
2481 The PCRE_ANCHORED option limits pcre_exec() to matching at the first
2482 matching position. If a pattern was compiled with PCRE_ANCHORED, or
2483 turned out to be anchored by virtue of its contents, it cannot be made
2484 unachored at matching time.
2489 These options (which are mutually exclusive) control what the \R escape
2490 sequence matches. The choice is either to match only CR, LF, or CRLF,
2491 or to match any Unicode newline sequence. These options override the
2492 choice that was made or defaulted when the pattern was compiled.
2500 These options override the newline definition that was chosen or
2501 defaulted when the pattern was compiled. For details, see the descrip-
2502 tion of pcre_compile() above. During matching, the newline choice
2503 affects the behaviour of the dot, circumflex, and dollar metacharac-
2504 ters. It may also alter the way the match position is advanced after a
2505 match failure for an unanchored pattern.
2508 set, and a match attempt for an unanchored pattern fails when the cur-
2509 rent position is at a CRLF sequence, and the pattern contains no
2510 explicit matches for CR or LF characters, the match position is
2511 advanced by two characters instead of one, in other words, to after the
2512 CRLF.
2514 The above rule is a compromise that makes the most common cases work as
2515 expected. For example, if the pattern is .+A (and the PCRE_DOTALL
2516 option is not set), it does not match the string "\r\nA" because, after
2517 failing at the start, it skips both the CR and the LF before retrying.
2518 However, the pattern [\r\n]A does match that string, because it con-
2519 tains an explicit CR or LF reference, and so advances only by one char-
2520 acter after the first failure.
2522 An explicit match for CR of LF is either a literal appearance of one of
2523 those characters, or one of the \r or \n escape sequences. Implicit
2524 matches such as [^X] do not count, nor does \s (which includes CR and
2525 LF in the characters that it matches).
2527 Notwithstanding the above, anomalous effects may still occur when CRLF
2528 is a valid newline sequence and explicit \r or \n escapes appear in the
2529 pattern.
2533 This option specifies that first character of the subject string is not
2534 the beginning of a line, so the circumflex metacharacter should not
2535 match before it. Setting this without PCRE_MULTILINE (at compile time)
2536 causes circumflex never to match. This option affects only the behav-
2537 iour of the circumflex metacharacter. It does not affect \A.
2541 This option specifies that the end of the subject string is not the end
2542 of a line, so the dollar metacharacter should not match it nor (except
2543 in multiline mode) a newline immediately before it. Setting this with-
2544 out PCRE_MULTILINE (at compile time) causes dollar never to match. This
2545 option affects only the behaviour of the dollar metacharacter. It does
2546 not affect \Z or \z.
2550 An empty string is not considered to be a valid match if this option is
2551 set. If there are alternatives in the pattern, they are tried. If all
2552 the alternatives match the empty string, the entire match fails. For
2553 example, if the pattern
2555 a?b?
2557 is applied to a string not beginning with "a" or "b", it matches an
2558 empty string at the start of the subject. With PCRE_NOTEMPTY set, this
2559 match is not valid, so PCRE searches further into the string for occur-
2560 rences of "a" or "b".
2564 This is like PCRE_NOTEMPTY, except that an empty string match that is
2565 not at the start of the subject is permitted. If the pattern is
2566 anchored, such a match can occur only if the pattern contains \K.
2568 Perl has no direct equivalent of PCRE_NOTEMPTY or
2569 PCRE_NOTEMPTY_ATSTART, but it does make a special case of a pattern
2570 match of the empty string within its split() function, and when using
2571 the /g modifier. It is possible to emulate Perl's behaviour after
2572 matching a null string by first trying the match again at the same off-
2573 set with PCRE_NOTEMPTY_ATSTART and PCRE_ANCHORED, and then if that
2574 fails, by advancing the starting offset (see below) and trying an ordi-
2575 nary match again. There is some code that demonstrates how to do this
2576 in the pcredemo sample program. In the most general case, you have to
2577 check to see if the newline convention recognizes CRLF as a newline,
2578 and if so, and the current character is CR followed by LF, advance the
2579 starting offset by two characters instead of one.
2583 There are a number of optimizations that pcre_exec() uses at the start
2584 of a match, in order to speed up the process. For example, if it is
2585 known that an unanchored match must start with a specific character, it
2586 searches the subject for that character, and fails immediately if it
2587 cannot find it, without actually running the main matching function.
2588 This means that a special item such as (*COMMIT) at the start of a pat-
2589 tern is not considered until after a suitable starting point for the
2590 match has been found. When callouts or (*MARK) items are in use, these
2591 "start-up" optimizations can cause them to be skipped if the pattern is
2592 never actually used. The start-up optimizations are in effect a pre-
2593 scan of the subject that takes place before the pattern is run.
2595 The PCRE_NO_START_OPTIMIZE option disables the start-up optimizations,
2596 possibly causing performance to suffer, but ensuring that in cases
2597 where the result is "no match", the callouts do occur, and that items
2598 such as (*COMMIT) and (*MARK) are considered at every possible starting
2599 position in the subject string. If PCRE_NO_START_OPTIMIZE is set at
2600 compile time, it cannot be unset at matching time.
2602 Setting PCRE_NO_START_OPTIMIZE can change the outcome of a matching
2603 operation. Consider the pattern
2607 When this is compiled, PCRE records the fact that a match must start
2608 with the character "A". Suppose the subject string is "DEFABC". The
2609 start-up optimization scans along the subject, finds "A" and runs the
2610 first match attempt from there. The (*COMMIT) item means that the pat-
2611 tern must match the current starting position, which in this case, it
2612 does. However, if the same match is run with PCRE_NO_START_OPTIMIZE
2613 set, the initial scan along the subject string does not happen. The
2614 first match attempt is run starting from "D" and when this fails,
2615 (*COMMIT) prevents any further matches being tried, so the overall
2616 result is "no match". If the pattern is studied, more start-up opti-
2617 mizations may be used. For example, a minimum length for the subject
2618 may be recorded. Consider the pattern
2620 (*MARK:A)(X|Y)
2622 The minimum length for a match is one character. If the subject is
2623 "ABC", there will be attempts to match "ABC", "BC", "C", and then
2624 finally an empty string. If the pattern is studied, the final attempt
2625 does not take place, because PCRE knows that the subject is too short,
2626 and so the (*MARK) is never encountered. In this case, studying the
2627 pattern does not affect the overall match result, which is still "no
2628 match", but it does affect the auxiliary information that is returned.
2632 When PCRE_UTF8 is set at compile time, the validity of the subject as a
2633 UTF-8 string is automatically checked when pcre_exec() is subsequently
2634 called. The value of startoffset is also checked to ensure that it
2635 points to the start of a UTF-8 character. There is a discussion about
2636 the validity of UTF-8 strings in the pcreunicode page. If an invalid
2637 sequence of bytes is found, pcre_exec() returns the error
2638 PCRE_ERROR_BADUTF8 or, if PCRE_PARTIAL_HARD is set and the problem is a
2639 truncated character at the end of the subject, PCRE_ERROR_SHORTUTF8. In
2640 both cases, information about the precise nature of the error may also
2641 be returned (see the descriptions of these errors in the section enti-
2642 tled Error return values from pcre_exec() below). If startoffset con-
2643 tains a value that does not point to the start of a UTF-8 character (or
2644 to the end of the subject), PCRE_ERROR_BADUTF8_OFFSET is returned.
2646 If you already know that your subject is valid, and you want to skip
2647 these checks for performance reasons, you can set the
2648 PCRE_NO_UTF8_CHECK option when calling pcre_exec(). You might want to
2649 do this for the second and subsequent calls to pcre_exec() if you are
2650 making repeated calls to find all the matches in a single subject
2651 string. However, you should be sure that the value of startoffset
2652 points to the start of a character (or the end of the subject). When
2653 PCRE_NO_UTF8_CHECK is set, the effect of passing an invalid string as a
2654 subject or an invalid value of startoffset is undefined. Your program
2655 may crash.
2660 These options turn on the partial matching feature. For backwards com-
2661 patibility, PCRE_PARTIAL is a synonym for PCRE_PARTIAL_SOFT. A partial
2662 match occurs if the end of the subject string is reached successfully,
2663 but there are not enough subject characters to complete the match. If
2664 this happens when PCRE_PARTIAL_SOFT (but not PCRE_PARTIAL_HARD) is set,
2665 matching continues by testing any remaining alternatives. Only if no
2666 complete match can be found is PCRE_ERROR_PARTIAL returned instead of
2667 PCRE_ERROR_NOMATCH. In other words, PCRE_PARTIAL_SOFT says that the
2668 caller is prepared to handle a partial match, but only if no complete
2669 match can be found.
2671 If PCRE_PARTIAL_HARD is set, it overrides PCRE_PARTIAL_SOFT. In this
2672 case, if a partial match is found, pcre_exec() immediately returns
2673 PCRE_ERROR_PARTIAL, without considering any other alternatives. In
2674 other words, when PCRE_PARTIAL_HARD is set, a partial match is consid-
2675 ered to be more important that an alternative complete match.
2677 In both cases, the portion of the string that was inspected when the
2678 partial match was found is set as the first matching string. There is a
2679 more detailed discussion of partial and multi-segment matching, with
2680 examples, in the pcrepartial documentation.
2682 The string to be matched by pcre_exec()
2684 The subject string is passed to pcre_exec() as a pointer in subject, a
2685 length in bytes in length, and a starting byte offset in startoffset.
2686 If this is negative or greater than the length of the subject,
2687 pcre_exec() returns PCRE_ERROR_BADOFFSET. When the starting offset is
2688 zero, the search for a match starts at the beginning of the subject,
2689 and this is by far the most common case. In UTF-8 mode, the byte offset
2690 must point to the start of a UTF-8 character (or the end of the sub-
2691 ject). Unlike the pattern string, the subject may contain binary zero
2692 bytes.
2694 A non-zero starting offset is useful when searching for another match
2695 in the same subject by calling pcre_exec() again after a previous suc-
2696 cess. Setting startoffset differs from just passing over a shortened
2697 string and setting PCRE_NOTBOL in the case of a pattern that begins
2698 with any kind of lookbehind. For example, consider the pattern
2700 \Biss\B
2702 which finds occurrences of "iss" in the middle of words. (\B matches
2703 only if the current position in the subject is not a word boundary.)
2704 When applied to the string "Mississipi" the first call to pcre_exec()
2705 finds the first occurrence. If pcre_exec() is called again with just
2706 the remainder of the subject, namely "issipi", it does not match,
2707 because \B is always false at the start of the subject, which is deemed
2708 to be a word boundary. However, if pcre_exec() is passed the entire
2709 string again, but with startoffset set to 4, it finds the second occur-
2710 rence of "iss" because it is able to look behind the starting point to
2711 discover that it is preceded by a letter.
2713 Finding all the matches in a subject is tricky when the pattern can
2714 match an empty string. It is possible to emulate Perl's /g behaviour by
2715 first trying the match again at the same offset, with the
2716 PCRE_NOTEMPTY_ATSTART and PCRE_ANCHORED options, and then if that
2717 fails, advancing the starting offset and trying an ordinary match
2718 again. There is some code that demonstrates how to do this in the pcre-
2719 demo sample program. In the most general case, you have to check to see
2720 if the newline convention recognizes CRLF as a newline, and if so, and
2721 the current character is CR followed by LF, advance the starting offset
2722 by two characters instead of one.
2724 If a non-zero starting offset is passed when the pattern is anchored,
2725 one attempt to match at the given offset is made. This can only succeed
2726 if the pattern does not require the match to be at the start of the
2727 subject.
2729 How pcre_exec() returns captured substrings
2731 In general, a pattern matches a certain portion of the subject, and in
2732 addition, further substrings from the subject may be picked out by
2733 parts of the pattern. Following the usage in Jeffrey Friedl's book,
2734 this is called "capturing" in what follows, and the phrase "capturing
2735 subpattern" is used for a fragment of a pattern that picks out a sub-
2736 string. PCRE supports several other kinds of parenthesized subpattern
2737 that do not cause substrings to be captured.
2739 Captured substrings are returned to the caller via a vector of integers
2740 whose address is passed in ovector. The number of elements in the vec-
2741 tor is passed in ovecsize, which must be a non-negative number. Note:
2742 this argument is NOT the size of ovector in bytes.
2744 The first two-thirds of the vector is used to pass back captured sub-
2745 strings, each substring using a pair of integers. The remaining third
2746 of the vector is used as workspace by pcre_exec() while matching cap-
2747 turing subpatterns, and is not available for passing back information.
2748 The number passed in ovecsize should always be a multiple of three. If
2749 it is not, it is rounded down.
2751 When a match is successful, information about captured substrings is
2752 returned in pairs of integers, starting at the beginning of ovector,
2753 and continuing up to two-thirds of its length at the most. The first
2754 element of each pair is set to the byte offset of the first character
2755 in a substring, and the second is set to the byte offset of the first
2756 character after the end of a substring. Note: these values are always
2757 byte offsets, even in UTF-8 mode. They are not character counts.
2759 The first pair of integers, ovector[0] and ovector[1], identify the
2760 portion of the subject string matched by the entire pattern. The next
2761 pair is used for the first capturing subpattern, and so on. The value
2762 returned by pcre_exec() is one more than the highest numbered pair that
2763 has been set. For example, if two substrings have been captured, the
2764 returned value is 3. If there are no capturing subpatterns, the return
2765 value from a successful match is 1, indicating that just the first pair
2766 of offsets has been set.
2768 If a capturing subpattern is matched repeatedly, it is the last portion
2769 of the string that it matched that is returned.
2771 If the vector is too small to hold all the captured substring offsets,
2772 it is used as far as possible (up to two-thirds of its length), and the
2773 function returns a value of zero. If neither the actual string matched
2774 not any captured substrings are of interest, pcre_exec() may be called
2775 with ovector passed as NULL and ovecsize as zero. However, if the pat-
2776 tern contains back references and the ovector is not big enough to
2777 remember the related substrings, PCRE has to get additional memory for
2778 use during matching. Thus it is usually advisable to supply an ovector
2779 of reasonable size.
2781 There are some cases where zero is returned (indicating vector over-
2782 flow) when in fact the vector is exactly the right size for the final
2783 match. For example, consider the pattern
2785 (a)(?:(b)c|bd)
2787 If a vector of 6 elements (allowing for only 1 captured substring) is
2788 given with subject string "abd", pcre_exec() will try to set the second
2789 captured string, thereby recording a vector overflow, before failing to
2790 match "c" and backing up to try the second alternative. The zero
2791 return, however, does correctly indicate that the maximum number of
2792 slots (namely 2) have been filled. In similar cases where there is tem-
2793 porary overflow, but the final number of used slots is actually less
2794 than the maximum, a non-zero value is returned.
2796 The pcre_fullinfo() function can be used to find out how many capturing
2797 subpatterns there are in a compiled pattern. The smallest size for
2798 ovector that will allow for n captured substrings, in addition to the
2799 offsets of the substring matched by the whole pattern, is (n+1)*3.
2801 It is possible for capturing subpattern number n+1 to match some part
2802 of the subject when subpattern n has not been used at all. For example,
2803 if the string "abc" is matched against the pattern (a|(z))(bc) the
2804 return from the function is 4, and subpatterns 1 and 3 are matched, but
2805 2 is not. When this happens, both values in the offset pairs corre-
2806 sponding to unused subpatterns are set to -1.
2808 Offset values that correspond to unused subpatterns at the end of the
2809 expression are also set to -1. For example, if the string "abc" is
2810 matched against the pattern (abc)(x(yz)?)? subpatterns 2 and 3 are not
2811 matched. The return from the function is 2, because the highest used
2812 capturing subpattern number is 1, and the offsets for for the second
2813 and third capturing subpatterns (assuming the vector is large enough,
2814 of course) are set to -1.
2816 Note: Elements in the first two-thirds of ovector that do not corre-
2817 spond to capturing parentheses in the pattern are never changed. That
2818 is, if a pattern contains n capturing parentheses, no more than ovec-
2819 tor[0] to ovector[2n+1] are set by pcre_exec(). The other elements (in
2820 the first two-thirds) retain whatever values they previously had.
2822 Some convenience functions are provided for extracting the captured
2823 substrings as separate strings. These are described below.
2825 Error return values from pcre_exec()
2827 If pcre_exec() fails, it returns a negative number. The following are
2828 defined in the header file:
2832 The subject string did not match the pattern.
2836 Either code or subject was passed as NULL, or ovector was NULL and
2837 ovecsize was not zero.
2841 An unrecognized bit was set in the options argument.
2845 PCRE stores a 4-byte "magic number" at the start of the compiled code,
2846 to catch the case when it is passed a junk pointer and to detect when a
2847 pattern that was compiled in an environment of one endianness is run in
2848 an environment with the other endianness. This is the error that PCRE
2849 gives when the magic number is not present.
2853 While running the pattern match, an unknown item was encountered in the
2854 compiled pattern. This error could be caused by a bug in PCRE or by
2855 overwriting of the compiled pattern.
2859 If a pattern contains back references, but the ovector that is passed
2860 to pcre_exec() is not big enough to remember the referenced substrings,
2861 PCRE gets a block of memory at the start of matching to use for this
2862 purpose. If the call via pcre_malloc() fails, this error is given. The
2863 memory is automatically freed at the end of matching.
2865 This error is also given if pcre_stack_malloc() fails in pcre_exec().
2866 This can happen only when PCRE has been compiled with --disable-stack-
2867 for-recursion.
2871 This error is used by the pcre_copy_substring(), pcre_get_substring(),
2872 and pcre_get_substring_list() functions (see below). It is never
2873 returned by pcre_exec().
2877 The backtracking limit, as specified by the match_limit field in a
2878 pcre_extra structure (or defaulted) was reached. See the description
2879 above.
2883 This error is never generated by pcre_exec() itself. It is provided for
2884 use by callout functions that want to yield a distinctive error code.
2885 See the pcrecallout documentation for details.
2889 A string that contains an invalid UTF-8 byte sequence was passed as a
2890 subject, and the PCRE_NO_UTF8_CHECK option was not set. If the size of
2891 the output vector (ovecsize) is at least 2, the byte offset to the
2892 start of the the invalid UTF-8 character is placed in the first ele-
2893 ment, and a reason code is placed in the second element. The reason
2894 codes are listed in the following section. For backward compatibility,
2895 if PCRE_PARTIAL_HARD is set and the problem is a truncated UTF-8 char-
2896 acter at the end of the subject (reason codes 1 to 5),
2897 PCRE_ERROR_SHORTUTF8 is returned instead of PCRE_ERROR_BADUTF8.
2901 The UTF-8 byte sequence that was passed as a subject was checked and
2902 found to be valid (the PCRE_NO_UTF8_CHECK option was not set), but the
2903 value of startoffset did not point to the beginning of a UTF-8 charac-
2904 ter or the end of the subject.
2908 The subject string did not match, but it did match partially. See the
2909 pcrepartial documentation for details of partial matching.
2913 This code is no longer in use. It was formerly returned when the
2914 PCRE_PARTIAL option was used with a compiled pattern containing items
2915 that were not supported for partial matching. From release 8.00
2916 onwards, there are no restrictions on partial matching.
2920 An unexpected internal error has occurred. This error could be caused
2921 by a bug in PCRE or by overwriting of the compiled pattern.
2925 This error is given if the value of the ovecsize argument is negative.
2929 The internal recursion limit, as specified by the match_limit_recursion
2930 field in a pcre_extra structure (or defaulted) was reached. See the
2931 description above.
2935 An invalid combination of PCRE_NEWLINE_xxx options was given.
2939 The value of startoffset was negative or greater than the length of the
2940 subject, that is, the value in length.
2944 This error is returned instead of PCRE_ERROR_BADUTF8 when the subject
2945 string ends with a truncated UTF-8 character and the PCRE_PARTIAL_HARD
2946 option is set. Information about the failure is returned as for
2947 PCRE_ERROR_BADUTF8. It is in fact sufficient to detect this case, but
2948 this special error code for PCRE_PARTIAL_HARD precedes the implementa-
2949 tion of returned information; it is retained for backwards compatibil-
2950 ity.
2954 This error is returned when pcre_exec() detects a recursion loop within
2955 the pattern. Specifically, it means that either the whole pattern or a
2956 subpattern has been called recursively for the second time at the same
2957 position in the subject string. Some simple patterns that might do this
2958 are detected and faulted at compile time, but more complicated cases,
2959 in particular mutual recursions between two different subpatterns, can-
2960 not be detected until run time.
2964 This error is returned when a pattern that was successfully studied
2965 using the PCRE_STUDY_JIT_COMPILE option is being matched, but the mem-
2966 ory available for the just-in-time processing stack is not large
2967 enough. See the pcrejit documentation for more details.
2971 This error is given if a pattern that was compiled by the 8-bit library
2972 is passed to a 16-bit library function, or vice versa.
2976 This error is given if a pattern that was compiled and saved is
2977 reloaded on a host with different endianness. The utility function
2978 pcre_pattern_to_host_byte_order() can be used to convert such a pattern
2979 so that it runs on the new host.
2981 Error numbers -16 to -20 and -22 are not used by pcre_exec().
2983 Reason codes for invalid UTF-8 strings
2985 This section applies only to the 8-bit library. The corresponding
2986 information for the 16-bit library is given in the pcre16 page.
2988 When pcre_exec() returns either PCRE_ERROR_BADUTF8 or PCRE_ERROR_SHORT-
2989 UTF8, and the size of the output vector (ovecsize) is at least 2, the
2990 offset of the start of the invalid UTF-8 character is placed in the
2991 first output vector element (ovector[0]) and a reason code is placed in
2992 the second element (ovector[1]). The reason codes are given names in
2993 the pcre.h header file:
3001 The string ends with a truncated UTF-8 character; the code specifies
3002 how many bytes are missing (1 to 5). Although RFC 3629 restricts UTF-8
3003 characters to be no longer than 4 bytes, the encoding scheme (origi-
3004 nally defined by RFC 2279) allows for up to 6 bytes, and this is
3005 checked first; hence the possibility of 4 or 5 missing bytes.
3011 PCRE_UTF8_ERR10
3013 The two most significant bits of the 2nd, 3rd, 4th, 5th, or 6th byte of
3014 the character do not have the binary value 0b10 (that is, either the
3015 most significant bit is 0, or the next bit is 1).
3017 PCRE_UTF8_ERR11
3018 PCRE_UTF8_ERR12
3020 A character that is valid by the RFC 2279 rules is either 5 or 6 bytes
3021 long; these code points are excluded by RFC 3629.
3023 PCRE_UTF8_ERR13
3025 A 4-byte character has a value greater than 0x10fff; these code points
3026 are excluded by RFC 3629.
3028 PCRE_UTF8_ERR14
3030 A 3-byte character has a value in the range 0xd800 to 0xdfff; this
3031 range of code points are reserved by RFC 3629 for use with UTF-16, and
3032 so are excluded from UTF-8.
3034 PCRE_UTF8_ERR15
3035 PCRE_UTF8_ERR16
3036 PCRE_UTF8_ERR17
3037 PCRE_UTF8_ERR18
3038 PCRE_UTF8_ERR19
3040 A 2-, 3-, 4-, 5-, or 6-byte character is "overlong", that is, it codes
3041 for a value that can be represented by fewer bytes, which is invalid.
3042 For example, the two bytes 0xc0, 0xae give the value 0x2e, whose cor-
3043 rect coding uses just one byte.
3045 PCRE_UTF8_ERR20
3047 The two most significant bits of the first byte of a character have the
3048 binary value 0b10 (that is, the most significant bit is 1 and the sec-
3049 ond is 0). Such a byte can only validly occur as the second or subse-
3050 quent byte of a multi-byte character.
3052 PCRE_UTF8_ERR21
3054 The first byte of a character has the value 0xfe or 0xff. These values
3055 can never occur in a valid UTF-8 string.
3060 int pcre_copy_substring(const char *subject, int *ovector,
3061 int stringcount, int stringnumber, char *buffer,
3062 int buffersize);
3064 int pcre_get_substring(const char *subject, int *ovector,
3065 int stringcount, int stringnumber,
3066 const char **stringptr);
3068 int pcre_get_substring_list(const char *subject,
3069 int *ovector, int stringcount, const char ***listptr);
3071 Captured substrings can be accessed directly by using the offsets
3072 returned by pcre_exec() in ovector. For convenience, the functions
3073 pcre_copy_substring(), pcre_get_substring(), and pcre_get_sub-
3074 string_list() are provided for extracting captured substrings as new,
3075 separate, zero-terminated strings. These functions identify substrings
3076 by number. The next section describes functions for extracting named
3077 substrings.
3079 A substring that contains a binary zero is correctly extracted and has
3080 a further zero added on the end, but the result is not, of course, a C
3081 string. However, you can process such a string by referring to the
3082 length that is returned by pcre_copy_substring() and pcre_get_sub-
3083 string(). Unfortunately, the interface to pcre_get_substring_list() is
3084 not adequate for handling strings containing binary zeros, because the
3085 end of the final string is not independently indicated.
3087 The first three arguments are the same for all three of these func-
3088 tions: subject is the subject string that has just been successfully
3089 matched, ovector is a pointer to the vector of integer offsets that was
3090 passed to pcre_exec(), and stringcount is the number of substrings that
3091 were captured by the match, including the substring that matched the
3092 entire regular expression. This is the value returned by pcre_exec() if
3093 it is greater than zero. If pcre_exec() returned zero, indicating that
3094 it ran out of space in ovector, the value passed as stringcount should
3095 be the number of elements in the vector divided by three.
3097 The functions pcre_copy_substring() and pcre_get_substring() extract a
3098 single substring, whose number is given as stringnumber. A value of
3099 zero extracts the substring that matched the entire pattern, whereas
3100 higher values extract the captured substrings. For pcre_copy_sub-
3101 string(), the string is placed in buffer, whose length is given by
3102 buffersize, while for pcre_get_substring() a new block of memory is
3103 obtained via pcre_malloc, and its address is returned via stringptr.
3104 The yield of the function is the length of the string, not including
3105 the terminating zero, or one of these error codes:
3109 The buffer was too small for pcre_copy_substring(), or the attempt to
3110 get memory failed for pcre_get_substring().
3114 There is no substring whose number is stringnumber.
3116 The pcre_get_substring_list() function extracts all available sub-
3117 strings and builds a list of pointers to them. All this is done in a
3118 single block of memory that is obtained via pcre_malloc. The address of
3119 the memory block is returned via listptr, which is also the start of
3120 the list of string pointers. The end of the list is marked by a NULL
3121 pointer. The yield of the function is zero if all went well, or the
3122 error code
3126 if the attempt to get the memory block failed.
3128 When any of these functions encounter a substring that is unset, which
3129 can happen when capturing subpattern number n+1 matches some part of
3130 the subject, but subpattern n has not been used at all, they return an
3131 empty string. This can be distinguished from a genuine zero-length sub-
3132 string by inspecting the appropriate offset in ovector, which is nega-
3133 tive for unset substrings.
3135 The two convenience functions pcre_free_substring() and pcre_free_sub-
3136 string_list() can be used to free the memory returned by a previous
3137 call of pcre_get_substring() or pcre_get_substring_list(), respec-
3138 tively. They do nothing more than call the function pointed to by
3139 pcre_free, which of course could be called directly from a C program.
3140 However, PCRE is used in some situations where it is linked via a spe-
3141 cial interface to another programming language that cannot use
3142 pcre_free directly; it is for these cases that the functions are pro-
3143 vided.
3148 int pcre_get_stringnumber(const pcre *code,
3149 const char *name);
3151 int pcre_copy_named_substring(const pcre *code,
3152 const char *subject, int *ovector,
3153 int stringcount, const char *stringname,
3154 char *buffer, int buffersize);
3156 int pcre_get_named_substring(const pcre *code,
3157 const char *subject, int *ovector,
3158 int stringcount, const char *stringname,
3159 const char **stringptr);
3161 To extract a substring by name, you first have to find associated num-
3162 ber. For example, for this pattern
3164 (a+)b(?<xxx>\d+)...
3166 the number of the subpattern called "xxx" is 2. If the name is known to
3167 be unique (PCRE_DUPNAMES was not set), you can find the number from the
3168 name by calling pcre_get_stringnumber(). The first argument is the com-
3169 piled pattern, and the second is the name. The yield of the function is
3170 the subpattern number, or PCRE_ERROR_NOSUBSTRING (-7) if there is no
3171 subpattern of that name.
3173 Given the number, you can extract the substring directly, or use one of
3174 the functions described in the previous section. For convenience, there
3175 are also two functions that do the whole job.
3177 Most of the arguments of pcre_copy_named_substring() and
3178 pcre_get_named_substring() are the same as those for the similarly
3179 named functions that extract by number. As these are described in the
3180 previous section, they are not re-described here. There are just two
3181 differences:
3183 First, instead of a substring number, a substring name is given. Sec-
3184 ond, there is an extra argument, given at the start, which is a pointer
3185 to the compiled pattern. This is needed in order to gain access to the
3186 name-to-number translation table.
3188 These functions call pcre_get_stringnumber(), and if it succeeds, they
3189 then call pcre_copy_substring() or pcre_get_substring(), as appropri-
3190 ate. NOTE: If PCRE_DUPNAMES is set and there are duplicate names, the
3191 behaviour may not be what you want (see the next section).
3193 Warning: If the pattern uses the (?| feature to set up multiple subpat-
3194 terns with the same number, as described in the section on duplicate
3195 subpattern numbers in the pcrepattern page, you cannot use names to
3196 distinguish the different subpatterns, because names are not included
3197 in the compiled code. The matching process uses only numbers. For this
3198 reason, the use of different names for subpatterns of the same number
3199 causes an error at compile time.
3204 int pcre_get_stringtable_entries(const pcre *code,
3205 const char *name, char **first, char **last);
3207 When a pattern is compiled with the PCRE_DUPNAMES option, names for
3208 subpatterns are not required to be unique. (Duplicate names are always
3209 allowed for subpatterns with the same number, created by using the (?|
3210 feature. Indeed, if such subpatterns are named, they are required to
3211 use the same names.)
3213 Normally, patterns with duplicate names are such that in any one match,
3214 only one of the named subpatterns participates. An example is shown in
3215 the pcrepattern documentation.
3217 When duplicates are present, pcre_copy_named_substring() and
3218 pcre_get_named_substring() return the first substring corresponding to
3219 the given name that is set. If none are set, PCRE_ERROR_NOSUBSTRING
3220 (-7) is returned; no data is returned. The pcre_get_stringnumber()
3221 function returns one of the numbers that are associated with the name,
3222 but it is not defined which it is.
3224 If you want to get full details of all captured substrings for a given
3225 name, you must use the pcre_get_stringtable_entries() function. The
3226 first argument is the compiled pattern, and the second is the name. The
3227 third and fourth are pointers to variables which are updated by the
3228 function. After it has run, they point to the first and last entries in
3229 the name-to-number table for the given name. The function itself
3230 returns the length of each entry, or PCRE_ERROR_NOSUBSTRING (-7) if
3231 there are none. The format of the table is described above in the sec-
3232 tion entitled Information about a pattern above. Given all the rele-
3233 vant entries for the name, you can extract each of their numbers, and
3234 hence the captured data, if any.
3239 The traditional matching function uses a similar algorithm to Perl,
3240 which stops when it finds the first match, starting at a given point in
3241 the subject. If you want to find all possible matches, or the longest
3242 possible match, consider using the alternative matching function (see
3243 below) instead. If you cannot use the alternative function, but still
3244 need to find all possible matches, you can kludge it up by making use
3245 of the callout facility, which is described in the pcrecallout documen-
3246 tation.
3248 What you have to do is to insert a callout right at the end of the pat-
3249 tern. When your callout function is called, extract and save the cur-
3250 rent matched substring. Then return 1, which forces pcre_exec() to
3251 backtrack and try other alternatives. Ultimately, when it runs out of
3252 matches, pcre_exec() will yield PCRE_ERROR_NOMATCH.
3257 int pcre_dfa_exec(const pcre *code, const pcre_extra *extra,
3258 const char *subject, int length, int startoffset,
3259 int options, int *ovector, int ovecsize,
3260 int *workspace, int wscount);
3262 The function pcre_dfa_exec() is called to match a subject string
3263 against a compiled pattern, using a matching algorithm that scans the
3264 subject string just once, and does not backtrack. This has different
3265 characteristics to the normal algorithm, and is not compatible with
3266 Perl. Some of the features of PCRE patterns are not supported. Never-
3267 theless, there are times when this kind of matching can be useful. For
3268 a discussion of the two matching algorithms, and a list of features
3269 that pcre_dfa_exec() does not support, see the pcrematching documenta-
3270 tion.
3272 The arguments for the pcre_dfa_exec() function are the same as for
3273 pcre_exec(), plus two extras. The ovector argument is used in a differ-
3274 ent way, and this is described below. The other common arguments are
3275 used in the same way as for pcre_exec(), so their description is not
3276 repeated here.
3278 The two additional arguments provide workspace for the function. The
3279 workspace vector should contain at least 20 elements. It is used for
3280 keeping track of multiple paths through the pattern tree. More
3281 workspace will be needed for patterns and subjects where there are a
3282 lot of potential matches.
3284 Here is an example of a simple call to pcre_dfa_exec():
3286 int rc;
3287 int ovector[10];
3288 int wspace[20];
3289 rc = pcre_dfa_exec(
3290 re, /* result of pcre_compile() */
3291 NULL, /* we didn't study the pattern */
3292 "some string", /* the subject string */
3293 11, /* the length of the subject string */
3294 0, /* start at offset 0 in the subject */
3295 0, /* default options */
3296 ovector, /* vector of integers for substring information */
3297 10, /* number of elements (NOT size in bytes) */
3298 wspace, /* working space vector */
3299 20); /* number of elements (NOT size in bytes) */
3301 Option bits for pcre_dfa_exec()
3303 The unused bits of the options argument for pcre_dfa_exec() must be
3304 zero. The only bits that may be set are PCRE_ANCHORED, PCRE_NEW-
3309 four of these are exactly the same as for pcre_exec(), so their
3310 description is not repeated here.
3315 These have the same general effect as they do for pcre_exec(), but the
3316 details are slightly different. When PCRE_PARTIAL_HARD is set for
3317 pcre_dfa_exec(), it returns PCRE_ERROR_PARTIAL if the end of the sub-
3318 ject is reached and there is still at least one matching possibility
3319 that requires additional characters. This happens even if some complete
3320 matches have also been found. When PCRE_PARTIAL_SOFT is set, the return
3321 code PCRE_ERROR_NOMATCH is converted into PCRE_ERROR_PARTIAL if the end
3322 of the subject is reached, there have been no complete matches, but
3323 there is still at least one matching possibility. The portion of the
3324 string that was inspected when the longest partial match was found is
3325 set as the first matching string in both cases. There is a more
3326 detailed discussion of partial and multi-segment matching, with exam-
3327 ples, in the pcrepartial documentation.
3331 Setting the PCRE_DFA_SHORTEST option causes the matching algorithm to
3332 stop as soon as it has found one match. Because of the way the alterna-
3333 tive algorithm works, this is necessarily the shortest possible match
3334 at the first possible matching point in the subject string.
3338 When pcre_dfa_exec() returns a partial match, it is possible to call it
3339 again, with additional subject characters, and have it continue with
3340 the same match. The PCRE_DFA_RESTART option requests this action; when
3341 it is set, the workspace and wscount options must reference the same
3342 vector as before because data about the match so far is left in them
3343 after a partial match. There is more discussion of this facility in the
3344 pcrepartial documentation.
3346 Successful returns from pcre_dfa_exec()
3348 When pcre_dfa_exec() succeeds, it may have matched more than one sub-
3349 string in the subject. Note, however, that all the matches from one run
3350 of the function start at the same point in the subject. The shorter
3351 matches are all initial substrings of the longer matches. For example,
3352 if the pattern
3354 <.*>
3356 is matched against the string
3358 This is <something> <something else> <something further> no more
3360 the three matched strings are
3362 <something>
3363 <something> <something else>
3364 <something> <something else> <something further>
3366 On success, the yield of the function is a number greater than zero,
3367 which is the number of matched substrings. The substrings themselves
3368 are returned in ovector. Each string uses two elements; the first is
3369 the offset to the start, and the second is the offset to the end. In
3370 fact, all the strings have the same start offset. (Space could have
3371 been saved by giving this only once, but it was decided to retain some
3372 compatibility with the way pcre_exec() returns data, even though the
3373 meaning of the strings is different.)
3375 The strings are returned in reverse order of length; that is, the long-
3376 est matching string is given first. If there were too many matches to
3377 fit into ovector, the yield of the function is zero, and the vector is
3378 filled with the longest matches. Unlike pcre_exec(), pcre_dfa_exec()
3379 can use the entire ovector for returning matched strings.
3381 Error returns from pcre_dfa_exec()
3383 The pcre_dfa_exec() function returns a negative number when it fails.
3384 Many of the errors are the same as for pcre_exec(), and these are
3385 described above. There are in addition the following errors that are
3386 specific to pcre_dfa_exec():
3390 This return is given if pcre_dfa_exec() encounters an item in the pat-
3391 tern that it does not support, for instance, the use of \C or a back
3392 reference.
3396 This return is given if pcre_dfa_exec() encounters a condition item
3397 that uses a back reference for the condition, or a test for recursion
3398 in a specific group. These are not supported.
3402 This return is given if pcre_dfa_exec() is called with an extra block
3403 that contains a setting of the match_limit or match_limit_recursion
3404 fields. This is not supported (these fields are meaningless for DFA
3405 matching).
3409 This return is given if pcre_dfa_exec() runs out of space in the
3410 workspace vector.
3414 When a recursive subpattern is processed, the matching function calls
3415 itself recursively, using private vectors for ovector and workspace.
3416 This error is given if the output vector is not large enough. This
3417 should be extremely rare, as a vector of size 1000 is used.
3422 pcre16(3), pcrebuild(3), pcrecallout(3), pcrecpp(3)(3), pcrematch-
3423 ing(3), pcrepartial(3), pcreposix(3), pcreprecompile(3), pcresample(3),
3424 pcrestack(3).
3429 Philip Hazel
3430 University Computing Service
3431 Cambridge CB2 3QH, England.
3436 Last updated: 07 January 2012
3437 Copyright (c) 1997-2012 University of Cambridge.
3438 ------------------------------------------------------------------------------
3444 NAME
3445 PCRE - Perl-compatible regular expressions
3450 int (*pcre_callout)(pcre_callout_block *);
3452 int (*pcre16_callout)(pcre16_callout_block *);
3454 PCRE provides a feature called "callout", which is a means of temporar-
3455 ily passing control to the caller of PCRE in the middle of pattern
3456 matching. The caller of PCRE provides an external function by putting
3457 its entry point in the global variable pcre_callout (pcre16_callout for
3458 the 16-bit library). By default, this variable contains NULL, which
3459 disables all calling out.
3461 Within a regular expression, (?C) indicates the points at which the
3462 external function is to be called. Different callout points can be
3463 identified by putting a number less than 256 after the letter C. The
3464 default value is zero. For example, this pattern has two callout
3465 points:
3467 (?C1)abc(?C2)def
3469 If the PCRE_AUTO_CALLOUT option bit is set when a pattern is compiled,
3470 PCRE automatically inserts callouts, all with number 255, before each
3471 item in the pattern. For example, if PCRE_AUTO_CALLOUT is used with the
3472 pattern
3474 A(\d{2}|--)
3476 it is processed as if it were
3478 (?C255)A(?C255)((?C255)\d{2}(?C255)|(?C255)-(?C255)-(?C255))(?C255)
3480 Notice that there is a callout before and after each parenthesis and
3481 alternation bar. Automatic callouts can be used for tracking the
3482 progress of pattern matching. The pcretest command has an option that
3483 sets automatic callouts; when it is used, the output indicates how the
3484 pattern is matched. This is useful information when you are trying to
3485 optimize the performance of a particular pattern.
3487 The use of callouts in a pattern makes it ineligible for optimization
3488 by the just-in-time compiler. Studying such a pattern with the
3489 PCRE_STUDY_JIT_COMPILE option always fails.
3494 You should be aware that, because of optimizations in the way PCRE
3495 matches patterns by default, callouts sometimes do not happen. For
3496 example, if the pattern is
3498 ab(?C4)cd
3500 PCRE knows that any matching string must contain the letter "d". If the
3501 subject string is "abyz", the lack of "d" means that matching doesn't
3502 ever start, and the callout is never reached. However, with "abyd",
3503 though the result is still no match, the callout is obeyed.
3505 If the pattern is studied, PCRE knows the minimum length of a matching
3506 string, and will immediately give a "no match" return without actually
3507 running a match if the subject is not long enough, or, for unanchored
3508 patterns, if it has been scanned far enough.
3510 You can disable these optimizations by passing the PCRE_NO_START_OPTI-
3511 MIZE option to the matching function, or by starting the pattern with
3512 (*NO_START_OPT). This slows down the matching process, but does ensure
3513 that callouts such as the example above are obeyed.
3518 During matching, when PCRE reaches a callout point, the external func-
3519 tion defined by pcre_callout or pcre16_callout is called (if it is
3520 set). This applies to both normal and DFA matching. The only argument
3521 to the callout function is a pointer to a pcre_callout or pcre16_call-
3522 out block. These structures contains the following fields:
3524 int version;
3525 int callout_number;
3526 int *offset_vector;
3527 const char *subject; (8-bit version)
3528 PCRE_SPTR16 subject; (16-bit version)
3529 int subject_length;
3530 int start_match;
3531 int current_position;
3532 int capture_top;
3533 int capture_last;
3534 void *callout_data;
3535 int pattern_position;
3536 int next_item_length;
3537 const unsigned char *mark; (8-bit version)
3538 const PCRE_UCHAR16 *mark; (16-bit version)
3540 The version field is an integer containing the version number of the
3541 block format. The initial version was 0; the current version is 2. The
3542 version number will change again in future if additional fields are
3543 added, but the intention is never to remove any of the existing fields.
3545 The callout_number field contains the number of the callout, as com-
3546 piled into the pattern (that is, the number after ?C for manual call-
3547 outs, and 255 for automatically generated callouts).
3549 The offset_vector field is a pointer to the vector of offsets that was
3550 passed by the caller to the matching function. When pcre_exec() or
3551 pcre16_exec() is used, the contents can be inspected, in order to
3552 extract substrings that have been matched so far, in the same way as
3553 for extracting substrings after a match has completed. For the DFA
3554 matching functions, this field is not useful.
3556 The subject and subject_length fields contain copies of the values that
3557 were passed to the matching function.
3559 The start_match field normally contains the offset within the subject
3560 at which the current match attempt started. However, if the escape
3561 sequence \K has been encountered, this value is changed to reflect the
3562 modified starting point. If the pattern is not anchored, the callout
3563 function may be called several times from the same point in the pattern
3564 for different starting points in the subject.
3566 The current_position field contains the offset within the subject of
3567 the current match pointer.
3569 When the pcre_exec() or pcre16_exec() is used, the capture_top field
3570 contains one more than the number of the highest numbered captured sub-
3571 string so far. If no substrings have been captured, the value of cap-
3572 ture_top is one. This is always the case when the DFA functions are
3573 used, because they do not support captured substrings.
3575 The capture_last field contains the number of the most recently cap-
3576 tured substring. If no substrings have been captured, its value is -1.
3577 This is always the case for the DFA matching functions.
3579 The callout_data field contains a value that is passed to a matching
3580 function specifically so that it can be passed back in callouts. It is
3581 passed in the callout_data field of a pcre_extra or pcre16_extra data
3582 structure. If no such data was passed, the value of callout_data in a
3583 callout block is NULL. There is a description of the pcre_extra struc-
3584 ture in the pcreapi documentation.
3586 The pattern_position field is present from version 1 of the callout
3587 structure. It contains the offset to the next item to be matched in the
3588 pattern string.
3590 The next_item_length field is present from version 1 of the callout
3591 structure. It contains the length of the next item to be matched in the
3592 pattern string. When the callout immediately precedes an alternation
3593 bar, a closing parenthesis, or the end of the pattern, the length is
3594 zero. When the callout precedes an opening parenthesis, the length is
3595 that of the entire subpattern.
3597 The pattern_position and next_item_length fields are intended to help
3598 in distinguishing between different automatic callouts, which all have
3599 the same callout number. However, they are set for all callouts.
3601 The mark field is present from version 2 of the callout structure. In
3602 callouts from pcre_exec() or pcre16_exec() it contains a pointer to the
3603 zero-terminated name of the most recently passed (*MARK), (*PRUNE), or
3604 (*THEN) item in the match, or NULL if no such items have been passed.
3605 Instances of (*PRUNE) or (*THEN) without a name do not obliterate a
3606 previous (*MARK). In callouts from the DFA matching functions this
3607 field always contains NULL.
3612 The external callout function returns an integer to PCRE. If the value
3613 is zero, matching proceeds as normal. If the value is greater than
3614 zero, matching fails at the current point, but the testing of other
3615 matching possibilities goes ahead, just as if a lookahead assertion had
3616 failed. If the value is less than zero, the match is abandoned, the
3617 matching function returns the negative value.
3619 Negative values should normally be chosen from the set of
3620 PCRE_ERROR_xxx values. In particular, PCRE_ERROR_NOMATCH forces a stan-
3621 dard "no match" failure. The error number PCRE_ERROR_CALLOUT is
3622 reserved for use by callout functions; it will never be used by PCRE
3623 itself.
3628 Philip Hazel
3629 University Computing Service
3630 Cambridge CB2 3QH, England.
3635 Last updated: 08 Janurary 2012
3636 Copyright (c) 1997-2012 University of Cambridge.
3637 ------------------------------------------------------------------------------
3643 NAME
3644 PCRE - Perl-compatible regular expressions
3649 This document describes the differences in the ways that PCRE and Perl
3650 handle regular expressions. The differences described here are with
3651 respect to Perl versions 5.10 and above.
3653 1. PCRE has only a subset of Perl's Unicode support. Details of what it
3654 does have are given in the pcreunicode page.
3656 2. PCRE allows repeat quantifiers only on parenthesized assertions, but
3657 they do not mean what you might think. For example, (?!a){3} does not
3658 assert that the next three characters are not "a". It just asserts that
3659 the next character is not "a" three times (in principle: PCRE optimizes
3660 this to run the assertion just once). Perl allows repeat quantifiers on
3661 other assertions such as \b, but these do not seem to have any use.
3663 3. Capturing subpatterns that occur inside negative lookahead asser-
3664 tions are counted, but their entries in the offsets vector are never
3665 set. Perl sets its numerical variables from any such patterns that are
3666 matched before the assertion fails to match something (thereby succeed-
3667 ing), but only if the negative lookahead assertion contains just one
3668 branch.
3670 4. Though binary zero characters are supported in the subject string,
3671 they are not allowed in a pattern string because it is passed as a nor-
3672 mal C string, terminated by zero. The escape sequence \0 can be used in
3673 the pattern to represent a binary zero.
3675 5. The following Perl escape sequences are not supported: \l, \u, \L,
3676 \U, and \N when followed by a character name or Unicode value. (\N on
3677 its own, matching a non-newline character, is supported.) In fact these
3678 are implemented by Perl's general string-handling and are not part of
3679 its pattern matching engine. If any of these are encountered by PCRE,
3680 an error is generated by default. However, if the PCRE_JAVASCRIPT_COM-
3681 PAT option is set, \U and \u are interpreted as JavaScript interprets
3682 them.
3684 6. The Perl escape sequences \p, \P, and \X are supported only if PCRE
3685 is built with Unicode character property support. The properties that
3686 can be tested with \p and \P are limited to the general category prop-
3687 erties such as Lu and Nd, script names such as Greek or Han, and the
3688 derived properties Any and L&. PCRE does support the Cs (surrogate)
3689 property, which Perl does not; the Perl documentation says "Because
3690 Perl hides the need for the user to understand the internal representa-
3691 tion of Unicode characters, there is no need to implement the somewhat
3692 messy concept of surrogates."
3694 7. PCRE implements a simpler version of \X than Perl, which changed to
3695 make \X match what Unicode calls an "extended grapheme cluster". This
3696 is more complicated than an extended Unicode sequence, which is what
3697 PCRE matches.
3699 8. PCRE does support the \Q...\E escape for quoting substrings. Charac-
3700 ters in between are treated as literals. This is slightly different
3701 from Perl in that $ and @ are also handled as literals inside the
3702 quotes. In Perl, they cause variable interpolation (but of course PCRE
3703 does not have variables). Note the following examples:
3705 Pattern PCRE matches Perl matches
3707 \Qabc$xyz\E abc$xyz abc followed by the
3708 contents of $xyz
3709 \Qabc\$xyz\E abc\$xyz abc\$xyz
3710 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3712 The \Q...\E sequence is recognized both inside and outside character
3713 classes.
3715 9. Fairly obviously, PCRE does not support the (?{code}) and (??{code})
3716 constructions. However, there is support for recursive patterns. This
3717 is not available in Perl 5.8, but it is in Perl 5.10. Also, the PCRE
3718 "callout" feature allows an external function to be called during pat-
3719 tern matching. See the pcrecallout documentation for details.
3721 10. Subpatterns that are called as subroutines (whether or not recur-
3722 sively) are always treated as atomic groups in PCRE. This is like
3723 Python, but unlike Perl. Captured values that are set outside a sub-
3724 routine call can be reference from inside in PCRE, but not in Perl.
3725 There is a discussion that explains these differences in more detail in
3726 the section on recursion differences from Perl in the pcrepattern page.
3728 11. If (*THEN) is present in a group that is called as a subroutine,
3729 its action is limited to that group, even if the group does not contain
3730 any | characters.
3732 12. There are some differences that are concerned with the settings of
3733 captured strings when part of a pattern is repeated. For example,
3734 matching "aba" against the pattern /^(a(b)?)+$/ in Perl leaves $2
3735 unset, but in PCRE it is set to "b".
3737 13. PCRE's handling of duplicate subpattern numbers and duplicate sub-
3738 pattern names is not as general as Perl's. This is a consequence of the
3739 fact the PCRE works internally just with numbers, using an external ta-
3740 ble to translate between numbers and names. In particular, a pattern
3741 such as (?|(?<a>A)|(?<b)B), where the two capturing parentheses have
3742 the same number but different names, is not supported, and causes an
3743 error at compile time. If it were allowed, it would not be possible to
3744 distinguish which parentheses matched, because both names map to cap-
3745 turing subpattern number 1. To avoid this confusing situation, an error
3746 is given at compile time.
3748 14. Perl recognizes comments in some places that PCRE does not, for
3749 example, between the ( and ? at the start of a subpattern. If the /x
3750 modifier is set, Perl allows whitespace between ( and ? but PCRE never
3751 does, even if the PCRE_EXTENDED option is set.
3753 15. PCRE provides some extensions to the Perl regular expression facil-
3754 ities. Perl 5.10 includes new features that are not in earlier ver-
3755 sions of Perl, some of which (such as named parentheses) have been in
3756 PCRE for some time. This list is with respect to Perl 5.10:
3758 (a) Although lookbehind assertions in PCRE must match fixed length
3759 strings, each alternative branch of a lookbehind assertion can match a
3760 different length of string. Perl requires them all to have the same
3761 length.
3763 (b) If PCRE_DOLLAR_ENDONLY is set and PCRE_MULTILINE is not set, the $
3764 meta-character matches only at the very end of the string.
3766 (c) If PCRE_EXTRA is set, a backslash followed by a letter with no spe-
3767 cial meaning is faulted. Otherwise, like Perl, the backslash is quietly
3768 ignored. (Perl can be made to issue a warning.)
3770 (d) If PCRE_UNGREEDY is set, the greediness of the repetition quanti-
3771 fiers is inverted, that is, by default they are not greedy, but if fol-
3772 lowed by a question mark they are.
3774 (e) PCRE_ANCHORED can be used at matching time to force a pattern to be
3775 tried only at the first matching position in the subject string.
3778 and PCRE_NO_AUTO_CAPTURE options for pcre_exec() have no Perl equiva-
3779 lents.
3781 (g) The \R escape sequence can be restricted to match only CR, LF, or
3782 CRLF by the PCRE_BSR_ANYCRLF option.
3784 (h) The callout facility is PCRE-specific.
3786 (i) The partial matching facility is PCRE-specific.
3788 (j) Patterns compiled by PCRE can be saved and re-used at a later time,
3789 even on different hosts that have the other endianness. However, this
3790 does not apply to optimized data created by the just-in-time compiler.
3792 (k) The alternative matching functions (pcre_dfa_exec() and
3793 pcre16_dfa_exec()) match in a different way and are not Perl-compati-
3794 ble.
3796 (l) PCRE recognizes some special sequences such as (*CR) at the start
3797 of a pattern that set overall options that cannot be changed within the
3798 pattern.
3803 Philip Hazel
3804 University Computing Service
3805 Cambridge CB2 3QH, England.
3810 Last updated: 08 Januray 2012
3811 Copyright (c) 1997-2012 University of Cambridge.
3812 ------------------------------------------------------------------------------
3818 NAME
3819 PCRE - Perl-compatible regular expressions
3824 The syntax and semantics of the regular expressions that are supported
3825 by PCRE are described in detail below. There is a quick-reference syn-
3826 tax summary in the pcresyntax page. PCRE tries to match Perl syntax and
3827 semantics as closely as it can. PCRE also supports some alternative
3828 regular expression syntax (which does not conflict with the Perl syn-
3829 tax) in order to provide some compatibility with regular expressions in
3830 Python, .NET, and Oniguruma.
3832 Perl's regular expressions are described in its own documentation, and
3833 regular expressions in general are covered in a number of books, some
3834 of which have copious examples. Jeffrey Friedl's "Mastering Regular
3835 Expressions", published by O'Reilly, covers regular expressions in
3836 great detail. This description of PCRE's regular expressions is
3837 intended as reference material.
3839 The original operation of PCRE was on strings of one-byte characters.
3840 However, there is now also support for UTF-8 strings in the original
3841 library, and a second library that supports 16-bit and UTF-16 character
3842 strings. To use these features, PCRE must be built to include appropri-
3843 ate support. When using UTF strings you must either call the compiling
3844 function with the PCRE_UTF8 or PCRE_UTF16 option, or the pattern must
3845 start with one of these special sequences:
3847 (*UTF8)
3848 (*UTF16)
3850 Starting a pattern with such a sequence is equivalent to setting the
3851 relevant option. This feature is not Perl-compatible. How setting a UTF
3852 mode affects pattern matching is mentioned in several places below.
3853 There is also a summary of features in the pcreunicode page.
3855 Another special sequence that may appear at the start of a pattern or
3856 in combination with (*UTF8) or (*UTF16) is:
3858 (*UCP)
3860 This has the same effect as setting the PCRE_UCP option: it causes
3861 sequences such as \d and \w to use Unicode properties to determine
3862 character types, instead of recognizing only characters with codes less
3863 than 128 via a lookup table.
3865 If a pattern starts with (*NO_START_OPT), it has the same effect as
3866 setting the PCRE_NO_START_OPTIMIZE option either at compile or matching
3867 time. There are also some more of these special sequences that are con-
3868 cerned with the handling of newlines; they are described below.
3870 The remainder of this document discusses the patterns that are sup-
3871 ported by PCRE when one its main matching functions, pcre_exec()
3872 (8-bit) or pcre16_exec() (16-bit), is used. PCRE also has alternative
3873 matching functions, pcre_dfa_exec() and pcre16_dfa_exec(), which match
3874 using a different algorithm that is not Perl-compatible. Some of the
3875 features discussed below are not available when DFA matching is used.
3876 The advantages and disadvantages of the alternative functions, and how
3877 they differ from the normal functions, are discussed in the pcrematch-
3878 ing page.
3883 PCRE supports five different conventions for indicating line breaks in
3884 strings: a single CR (carriage return) character, a single LF (line-
3885 feed) character, the two-character sequence CRLF, any of the three pre-
3886 ceding, or any Unicode newline sequence. The pcreapi page has further
3887 discussion about newlines, and shows how to set the newline convention
3888 in the options arguments for the compiling and matching functions.
3890 It is also possible to specify a newline convention by starting a pat-
3891 tern string with one of the following five sequences:
3893 (*CR) carriage return
3894 (*LF) linefeed
3895 (*CRLF) carriage return, followed by linefeed
3896 (*ANYCRLF) any of the three above
3897 (*ANY) all Unicode newline sequences
3899 These override the default and the options given to the compiling func-
3900 tion. For example, on a Unix system where LF is the default newline
3901 sequence, the pattern
3903 (*CR)a.b
3905 changes the convention to CR. That pattern matches "a\nb" because LF is
3906 no longer a newline. Note that these special settings, which are not
3907 Perl-compatible, are recognized only at the very start of a pattern,
3908 and that they must be in upper case. If more than one of them is
3909 present, the last one is used.
3911 The newline convention affects the interpretation of the dot metachar-
3912 acter when PCRE_DOTALL is not set, and also the behaviour of \N. How-
3913 ever, it does not affect what the \R escape sequence matches. By
3914 default, this is any Unicode newline sequence, for Perl compatibility.
3915 However, this can be changed; see the description of \R in the section
3916 entitled "Newline sequences" below. A change of \R setting can be com-
3917 bined with a change of newline convention.
3922 A regular expression is a pattern that is matched against a subject
3923 string from left to right. Most characters stand for themselves in a
3924 pattern, and match the corresponding characters in the subject. As a
3925 trivial example, the pattern
3927 The quick brown fox
3929 matches a portion of a subject string that is identical to itself. When
3930 caseless matching is specified (the PCRE_CASELESS option), letters are
3931 matched independently of case. In a UTF mode, PCRE always understands
3932 the concept of case for characters whose values are less than 128, so
3933 caseless matching is always possible. For characters with higher val-
3934 ues, the concept of case is supported if PCRE is compiled with Unicode
3935 property support, but not otherwise. If you want to use caseless
3936 matching for characters 128 and above, you must ensure that PCRE is
3937 compiled with Unicode property support as well as with UTF support.
3939 The power of regular expressions comes from the ability to include
3940 alternatives and repetitions in the pattern. These are encoded in the
3941 pattern by the use of metacharacters, which do not stand for themselves
3942 but instead are interpreted in some special way.
3944 There are two different sets of metacharacters: those that are recog-
3945 nized anywhere in the pattern except within square brackets, and those
3946 that are recognized within square brackets. Outside square brackets,
3947 the metacharacters are as follows:
3949 \ general escape character with several uses
3950 ^ assert start of string (or line, in multiline mode)
3951 $ assert end of string (or line, in multiline mode)
3952 . match any character except newline (by default)
3953 [ start character class definition
3954 | start of alternative branch
3955 ( start subpattern
3956 ) end subpattern
3957 ? extends the meaning of (
3958 also 0 or 1 quantifier
3959 also quantifier minimizer
3960 * 0 or more quantifier
3961 + 1 or more quantifier
3962 also "possessive quantifier"
3963 { start min/max quantifier
3965 Part of a pattern that is in square brackets is called a "character
3966 class". In a character class the only metacharacters are:
3968 \ general escape character
3969 ^ negate the class, but only if the first character
3970 - indicates character range
3971 [ POSIX character class (only if followed by POSIX
3972 syntax)
3973 ] terminates the character class
3975 The following sections describe the use of each of the metacharacters.
3980 The backslash character has several uses. Firstly, if it is followed by
3981 a character that is not a number or a letter, it takes away any special
3982 meaning that character may have. This use of backslash as an escape
3983 character applies both inside and outside character classes.
3985 For example, if you want to match a * character, you write \* in the
3986 pattern. This escaping action applies whether or not the following
3987 character would otherwise be interpreted as a metacharacter, so it is
3988 always safe to precede a non-alphanumeric with backslash to specify
3989 that it stands for itself. In particular, if you want to match a back-
3990 slash, you write \\.
3992 In a UTF mode, only ASCII numbers and letters have any special meaning
3993 after a backslash. All other characters (in particular, those whose
3994 codepoints are greater than 127) are treated as literals.
3996 If a pattern is compiled with the PCRE_EXTENDED option, whitespace in
3997 the pattern (other than in a character class) and characters between a
3998 # outside a character class and the next newline are ignored. An escap-
3999 ing backslash can be used to include a whitespace or # character as
4000 part of the pattern.
4002 If you want to remove the special meaning from a sequence of charac-
4003 ters, you can do so by putting them between \Q and \E. This is differ-
4004 ent from Perl in that $ and @ are handled as literals in \Q...\E
4005 sequences in PCRE, whereas in Perl, $ and @ cause variable interpola-
4006 tion. Note the following examples:
4008 Pattern PCRE matches Perl matches
4010 \Qabc$xyz\E abc$xyz abc followed by the
4011 contents of $xyz
4012 \Qabc\$xyz\E abc\$xyz abc\$xyz
4013 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4015 The \Q...\E sequence is recognized both inside and outside character
4016 classes. An isolated \E that is not preceded by \Q is ignored. If \Q
4017 is not followed by \E later in the pattern, the literal interpretation
4018 continues to the end of the pattern (that is, \E is assumed at the
4019 end). If the isolated \Q is inside a character class, this causes an
4020 error, because the character class is not terminated.
4022 Non-printing characters
4024 A second use of backslash provides a way of encoding non-printing char-
4025 acters in patterns in a visible manner. There is no restriction on the
4026 appearance of non-printing characters, apart from the binary zero that
4027 terminates a pattern, but when a pattern is being prepared by text
4028 editing, it is often easier to use one of the following escape
4029 sequences than the binary character it represents:
4031 \a alarm, that is, the BEL character (hex 07)
4032 \cx "control-x", where x is any ASCII character
4033 \e escape (hex 1B)
4034 \f formfeed (hex 0C)
4035 \n linefeed (hex 0A)
4036 \r carriage return (hex 0D)
4037 \t tab (hex 09)
4038 \ddd character with octal code ddd, or back reference
4039 \xhh character with hex code hh
4040 \x{hhh..} character with hex code hhh.. (non-JavaScript mode)
4041 \uhhhh character with hex code hhhh (JavaScript mode only)
4043 The precise effect of \cx is as follows: if x is a lower case letter,
4044 it is converted to upper case. Then bit 6 of the character (hex 40) is
4045 inverted. Thus \cz becomes hex 1A (z is 7A), but \c{ becomes hex 3B ({
4046 is 7B), while \c; becomes hex 7B (; is 3B). If the byte following \c
4047 has a value greater than 127, a compile-time error occurs. This locks
4048 out non-ASCII characters in all modes. (When PCRE is compiled in EBCDIC
4049 mode, all byte values are valid. A lower case letter is converted to
4050 upper case, and then the 0xc0 bits are flipped.)
4052 By default, after \x, from zero to two hexadecimal digits are read
4053 (letters can be in upper or lower case). Any number of hexadecimal dig-
4054 its may appear between \x{ and }, but the character code is constrained
4055 as follows:
4057 8-bit non-UTF mode less than 0x100
4058 8-bit UTF-8 mode less than 0x10ffff and a valid codepoint
4059 16-bit non-UTF mode less than 0x10000
4060 16-bit UTF-16 mode less than 0x10ffff and a valid codepoint
4062 Invalid Unicode codepoints are the range 0xd800 to 0xdfff (the so-
4063 called "surrogate" codepoints).
4065 If characters other than hexadecimal digits appear between \x{ and },
4066 or if there is no terminating }, this form of escape is not recognized.
4067 Instead, the initial \x will be interpreted as a basic hexadecimal
4068 escape, with no following digits, giving a character whose value is
4069 zero.
4071 If the PCRE_JAVASCRIPT_COMPAT option is set, the interpretation of \x
4072 is as just described only when it is followed by two hexadecimal dig-
4073 its. Otherwise, it matches a literal "x" character. In JavaScript
4074 mode, support for code points greater than 256 is provided by \u, which
4075 must be followed by four hexadecimal digits; otherwise it matches a
4076 literal "u" character.
4078 Characters whose value is less than 256 can be defined by either of the
4079 two syntaxes for \x (or by \u in JavaScript mode). There is no differ-
4080 ence in the way they are handled. For example, \xdc is exactly the same
4081 as \x{dc} (or \u00dc in JavaScript mode).
4083 After \0 up to two further octal digits are read. If there are fewer
4084 than two digits, just those that are present are used. Thus the
4085 sequence \0\x\07 specifies two binary zeros followed by a BEL character
4086 (code value 7). Make sure you supply two digits after the initial zero
4087 if the pattern character that follows is itself an octal digit.
4089 The handling of a backslash followed by a digit other than 0 is compli-
4090 cated. Outside a character class, PCRE reads it and any following dig-
4091 its as a decimal number. If the number is less than 10, or if there
4092 have been at least that many previous capturing left parentheses in the
4093 expression, the entire sequence is taken as a back reference. A
4094 description of how this works is given later, following the discussion
4095 of parenthesized subpatterns.
4097 Inside a character class, or if the decimal number is greater than 9
4098 and there have not been that many capturing subpatterns, PCRE re-reads
4099 up to three octal digits following the backslash, and uses them to gen-
4100 erate a data character. Any subsequent digits stand for themselves. The
4101 value of the character is constrained in the same way as characters
4102 specified in hexadecimal. For example:
4104 \040 is another way of writing a space
4105 \40 is the same, provided there are fewer than 40
4106 previous capturing subpatterns
4107 \7 is always a back reference
4108 \11 might be a back reference, or another way of
4109 writing a tab
4110 \011 is always a tab
4111 \0113 is a tab followed by the character "3"
4112 \113 might be a back reference, otherwise the
4113 character with octal code 113
4114 \377 might be a back reference, otherwise
4115 the value 255 (decimal)
4116 \81 is either a back reference, or a binary zero
4117 followed by the two characters "8" and "1"
4119 Note that octal values of 100 or greater must not be introduced by a
4120 leading zero, because no more than three octal digits are ever read.
4122 All the sequences that define a single character value can be used both
4123 inside and outside character classes. In addition, inside a character
4124 class, \b is interpreted as the backspace character (hex 08).
4126 \N is not allowed in a character class. \B, \R, and \X are not special
4127 inside a character class. Like other unrecognized escape sequences,
4128 they are treated as the literal characters "B", "R", and "X" by
4129 default, but cause an error if the PCRE_EXTRA option is set. Outside a
4130 character class, these sequences have different meanings.
4132 Unsupported escape sequences
4134 In Perl, the sequences \l, \L, \u, and \U are recognized by its string
4135 handler and used to modify the case of following characters. By
4136 default, PCRE does not support these escape sequences. However, if the
4137 PCRE_JAVASCRIPT_COMPAT option is set, \U matches a "U" character, and
4138 \u can be used to define a character by code point, as described in the
4139 previous section.
4141 Absolute and relative back references
4143 The sequence \g followed by an unsigned or a negative number, option-
4144 ally enclosed in braces, is an absolute or relative back reference. A
4145 named back reference can be coded as \g{name}. Back references are dis-
4146 cussed later, following the discussion of parenthesized subpatterns.
4148 Absolute and relative subroutine calls
4150 For compatibility with Oniguruma, the non-Perl syntax \g followed by a
4151 name or a number enclosed either in angle brackets or single quotes, is
4152 an alternative syntax for referencing a subpattern as a "subroutine".
4153 Details are discussed later. Note that \g{...} (Perl syntax) and
4154 \g<...> (Oniguruma syntax) are not synonymous. The former is a back
4155 reference; the latter is a subroutine call.
4157 Generic character types
4159 Another use of backslash is for specifying generic character types:
4161 \d any decimal digit
4162 \D any character that is not a decimal digit
4163 \h any horizontal whitespace character
4164 \H any character that is not a horizontal whitespace character
4165 \s any whitespace character
4166 \S any character that is not a whitespace character
4167 \v any vertical whitespace character
4168 \V any character that is not a vertical whitespace character
4169 \w any "word" character
4170 \W any "non-word" character
4172 There is also the single sequence \N, which matches a non-newline char-
4173 acter. This is the same as the "." metacharacter when PCRE_DOTALL is
4174 not set. Perl also uses \N to match characters by name; PCRE does not
4175 support this.
4177 Each pair of lower and upper case escape sequences partitions the com-
4178 plete set of characters into two disjoint sets. Any given character
4179 matches one, and only one, of each pair. The sequences can appear both
4180 inside and outside character classes. They each match one character of
4181 the appropriate type. If the current matching point is at the end of
4182 the subject string, all of them fail, because there is no character to
4183 match.
4185 For compatibility with Perl, \s does not match the VT character (code
4186 11). This makes it different from the the POSIX "space" class. The \s
4187 characters are HT (9), LF (10), FF (12), CR (13), and space (32). If
4188 "use locale;" is included in a Perl script, \s may match the VT charac-
4189 ter. In PCRE, it never does.
4191 A "word" character is an underscore or any character that is a letter
4192 or digit. By default, the definition of letters and digits is con-
4193 trolled by PCRE's low-valued character tables, and may vary if locale-
4194 specific matching is taking place (see "Locale support" in the pcreapi
4195 page). For example, in a French locale such as "fr_FR" in Unix-like
4196 systems, or "french" in Windows, some character codes greater than 128
4197 are used for accented letters, and these are then matched by \w. The
4198 use of locales with Unicode is discouraged.
4200 By default, in a UTF mode, characters with values greater than 128
4201 never match \d, \s, or \w, and always match \D, \S, and \W. These
4202 sequences retain their original meanings from before UTF support was
4203 available, mainly for efficiency reasons. However, if PCRE is compiled
4204 with Unicode property support, and the PCRE_UCP option is set, the be-
4205 haviour is changed so that Unicode properties are used to determine
4206 character types, as follows:
4208 \d any character that \p{Nd} matches (decimal digit)
4209 \s any character that \p{Z} matches, plus HT, LF, FF, CR
4210 \w any character that \p{L} or \p{N} matches, plus underscore
4212 The upper case escapes match the inverse sets of characters. Note that
4213 \d matches only decimal digits, whereas \w matches any Unicode digit,
4214 as well as any Unicode letter, and underscore. Note also that PCRE_UCP
4215 affects \b, and \B because they are defined in terms of \w and \W.
4216 Matching these sequences is noticeably slower when PCRE_UCP is set.
4218 The sequences \h, \H, \v, and \V are features that were added to Perl
4219 at release 5.10. In contrast to the other sequences, which match only
4220 ASCII characters by default, these always match certain high-valued
4221 codepoints, whether or not PCRE_UCP is set. The horizontal space char-
4222 acters are:
4224 U+0009 Horizontal tab
4225 U+0020 Space
4226 U+00A0 Non-break space
4227 U+1680 Ogham space mark
4228 U+180E Mongolian vowel separator
4229 U+2000 En quad
4230 U+2001 Em quad
4231 U+2002 En space
4232 U+2003 Em space
4233 U+2004 Three-per-em space
4234 U+2005 Four-per-em space
4235 U+2006 Six-per-em space
4236 U+2007 Figure space
4237 U+2008 Punctuation space
4238 U+2009 Thin space
4239 U+200A Hair space
4240 U+202F Narrow no-break space
4241 U+205F Medium mathematical space
4242 U+3000 Ideographic space
4244 The vertical space characters are:
4246 U+000A Linefeed
4247 U+000B Vertical tab
4248 U+000C Formfeed
4249 U+000D Carriage return
4250 U+0085 Next line
4251 U+2028 Line separator
4252 U+2029 Paragraph separator
4254 In 8-bit, non-UTF-8 mode, only the characters with codepoints less than
4255 256 are relevant.
4257 Newline sequences
4259 Outside a character class, by default, the escape sequence \R matches
4260 any Unicode newline sequence. In 8-bit non-UTF-8 mode \R is equivalent
4261 to the following:
4263 (?>\r\n|\n|\x0b|\f|\r|\x85)
4265 This is an example of an "atomic group", details of which are given
4266 below. This particular group matches either the two-character sequence
4267 CR followed by LF, or one of the single characters LF (linefeed,
4268 U+000A), VT (vertical tab, U+000B), FF (formfeed, U+000C), CR (carriage
4269 return, U+000D), or NEL (next line, U+0085). The two-character sequence
4270 is treated as a single unit that cannot be split.
4272 In other modes, two additional characters whose codepoints are greater
4273 than 255 are added: LS (line separator, U+2028) and PS (paragraph sepa-
4274 rator, U+2029). Unicode character property support is not needed for
4275 these characters to be recognized.
4277 It is possible to restrict \R to match only CR, LF, or CRLF (instead of
4278 the complete set of Unicode line endings) by setting the option
4279 PCRE_BSR_ANYCRLF either at compile time or when the pattern is matched.
4280 (BSR is an abbrevation for "backslash R".) This can be made the default
4281 when PCRE is built; if this is the case, the other behaviour can be
4282 requested via the PCRE_BSR_UNICODE option. It is also possible to
4283 specify these settings by starting a pattern string with one of the
4284 following sequences:
4286 (*BSR_ANYCRLF) CR, LF, or CRLF only
4287 (*BSR_UNICODE) any Unicode newline sequence
4289 These override the default and the options given to the compiling func-
4290 tion, but they can themselves be overridden by options given to a
4291 matching function. Note that these special settings, which are not
4292 Perl-compatible, are recognized only at the very start of a pattern,
4293 and that they must be in upper case. If more than one of them is
4294 present, the last one is used. They can be combined with a change of
4295 newline convention; for example, a pattern can start with:
4299 They can also be combined with the (*UTF8), (*UTF16), or (*UCP) special
4300 sequences. Inside a character class, \R is treated as an unrecognized
4301 escape sequence, and so matches the letter "R" by default, but causes
4302 an error if PCRE_EXTRA is set.
4304 Unicode character properties
4306 When PCRE is built with Unicode character property support, three addi-
4307 tional escape sequences that match characters with specific properties
4308 are available. When in 8-bit non-UTF-8 mode, these sequences are of
4309 course limited to testing characters whose codepoints are less than
4310 256, but they do work in this mode. The extra escape sequences are:
4312 \p{xx} a character with the xx property
4313 \P{xx} a character without the xx property
4314 \X an extended Unicode sequence
4316 The property names represented by xx above are limited to the Unicode
4317 script names, the general category properties, "Any", which matches any
4318 character (including newline), and some special PCRE properties
4319 (described in the next section). Other Perl properties such as "InMu-
4320 sicalSymbols" are not currently supported by PCRE. Note that \P{Any}
4321 does not match any characters, so always causes a match failure.
4323 Sets of Unicode characters are defined as belonging to certain scripts.
4324 A character from one of these sets can be matched using a script name.
4325 For example:
4327 \p{Greek}
4328 \P{Han}
4330 Those that are not part of an identified script are lumped together as
4331 "Common". The current list of scripts is:
4333 Arabic, Armenian, Avestan, Balinese, Bamum, Bengali, Bopomofo, Braille,
4334 Buginese, Buhid, Canadian_Aboriginal, Carian, Cham, Cherokee, Common,
4335 Coptic, Cuneiform, Cypriot, Cyrillic, Deseret, Devanagari, Egyp-
4336 tian_Hieroglyphs, Ethiopic, Georgian, Glagolitic, Gothic, Greek,
4337 Gujarati, Gurmukhi, Han, Hangul, Hanunoo, Hebrew, Hiragana, Impe-
4338 rial_Aramaic, Inherited, Inscriptional_Pahlavi, Inscriptional_Parthian,
4339 Javanese, Kaithi, Kannada, Katakana, Kayah_Li, Kharoshthi, Khmer, Lao,
4340 Latin, Lepcha, Limbu, Linear_B, Lisu, Lycian, Lydian, Malayalam,
4341 Meetei_Mayek, Mongolian, Myanmar, New_Tai_Lue, Nko, Ogham, Old_Italic,
4342 Old_Persian, Old_South_Arabian, Old_Turkic, Ol_Chiki, Oriya, Osmanya,
4343 Phags_Pa, Phoenician, Rejang, Runic, Samaritan, Saurashtra, Shavian,
4344 Sinhala, Sundanese, Syloti_Nagri, Syriac, Tagalog, Tagbanwa, Tai_Le,
4345 Tai_Tham, Tai_Viet, Tamil, Telugu, Thaana, Thai, Tibetan, Tifinagh,
4346 Ugaritic, Vai, Yi.
4348 Each character has exactly one Unicode general category property, spec-
4349 ified by a two-letter abbreviation. For compatibility with Perl, nega-
4350 tion can be specified by including a circumflex between the opening
4351 brace and the property name. For example, \p{^Lu} is the same as
4352 \P{Lu}.
4354 If only one letter is specified with \p or \P, it includes all the gen-
4355 eral category properties that start with that letter. In this case, in
4356 the absence of negation, the curly brackets in the escape sequence are
4357 optional; these two examples have the same effect:
4359 \p{L}
4360 \pL
4362 The following general category property codes are supported:
4364 C Other
4365 Cc Control
4366 Cf Format
4367 Cn Unassigned
4368 Co Private use
4369 Cs Surrogate
4371 L Letter
4372 Ll Lower case letter
4373 Lm Modifier letter
4374 Lo Other letter
4375 Lt Title case letter
4376 Lu Upper case letter
4378 M Mark
4379 Mc Spacing mark
4380 Me Enclosing mark
4381 Mn Non-spacing mark
4383 N Number
4384 Nd Decimal number
4385 Nl Letter number
4386 No Other number
4388 P Punctuation
4389 Pc Connector punctuation
4390 Pd Dash punctuation
4391 Pe Close punctuation
4392 Pf Final punctuation
4393 Pi Initial punctuation
4394 Po Other punctuation
4395 Ps Open punctuation
4397 S Symbol
4398 Sc Currency symbol
4399 Sk Modifier symbol
4400 Sm Mathematical symbol
4401 So Other symbol
4403 Z Separator
4404 Zl Line separator
4405 Zp Paragraph separator
4406 Zs Space separator
4408 The special property L& is also supported: it matches a character that
4409 has the Lu, Ll, or Lt property, in other words, a letter that is not
4410 classified as a modifier or "other".
4412 The Cs (Surrogate) property applies only to characters in the range
4413 U+D800 to U+DFFF. Such characters are not valid in Unicode strings and
4414 so cannot be tested by PCRE, unless UTF validity checking has been
4415 turned off (see the discussion of PCRE_NO_UTF8_CHECK and
4416 PCRE_NO_UTF16_CHECK in the pcreapi page). Perl does not support the Cs
4417 property.
4419 The long synonyms for property names that Perl supports (such as
4420 \p{Letter}) are not supported by PCRE, nor is it permitted to prefix
4421 any of these properties with "Is".
4423 No character that is in the Unicode table has the Cn (unassigned) prop-
4424 erty. Instead, this property is assumed for any code point that is not
4425 in the Unicode table.
4427 Specifying caseless matching does not affect these escape sequences.
4428 For example, \p{Lu} always matches only upper case letters.
4430 The \X escape matches any number of Unicode characters that form an
4431 extended Unicode sequence. \X is equivalent to
4433 (?>\PM\pM*)
4435 That is, it matches a character without the "mark" property, followed
4436 by zero or more characters with the "mark" property, and treats the
4437 sequence as an atomic group (see below). Characters with the "mark"
4438 property are typically accents that affect the preceding character.
4439 None of them have codepoints less than 256, so in 8-bit non-UTF-8 mode
4440 \X matches any one character.
4442 Note that recent versions of Perl have changed \X to match what Unicode
4443 calls an "extended grapheme cluster", which has a more complicated def-
4444 inition.
4446 Matching characters by Unicode property is not fast, because PCRE has
4447 to search a structure that contains data for over fifteen thousand
4448 characters. That is why the traditional escape sequences such as \d and
4449 \w do not use Unicode properties in PCRE by default, though you can
4450 make them do so by setting the PCRE_UCP option or by starting the pat-
4451 tern with (*UCP).
4453 PCRE's additional properties
4455 As well as the standard Unicode properties described in the previous
4456 section, PCRE supports four more that make it possible to convert tra-
4457 ditional escape sequences such as \w and \s and POSIX character classes
4458 to use Unicode properties. PCRE uses these non-standard, non-Perl prop-
4459 erties internally when PCRE_UCP is set. They are:
4461 Xan Any alphanumeric character
4462 Xps Any POSIX space character
4463 Xsp Any Perl space character
4464 Xwd Any Perl "word" character
4466 Xan matches characters that have either the L (letter) or the N (num-
4467 ber) property. Xps matches the characters tab, linefeed, vertical tab,
4468 formfeed, or carriage return, and any other character that has the Z
4469 (separator) property. Xsp is the same as Xps, except that vertical tab
4470 is excluded. Xwd matches the same characters as Xan, plus underscore.
4472 Resetting the match start
4474 The escape sequence \K causes any previously matched characters not to
4475 be included in the final matched sequence. For example, the pattern:
4477 foo\Kbar
4479 matches "foobar", but reports that it has matched "bar". This feature
4480 is similar to a lookbehind assertion (described below). However, in
4481 this case, the part of the subject before the real match does not have
4482 to be of fixed length, as lookbehind assertions do. The use of \K does
4483 not interfere with the setting of captured substrings. For example,
4484 when the pattern
4486 (foo)\Kbar
4488 matches "foobar", the first substring is still set to "foo".
4490 Perl documents that the use of \K within assertions is "not well
4491 defined". In PCRE, \K is acted upon when it occurs inside positive
4492 assertions, but is ignored in negative assertions.
4494 Simple assertions
4496 The final use of backslash is for certain simple assertions. An asser-
4497 tion specifies a condition that has to be met at a particular point in
4498 a match, without consuming any characters from the subject string. The
4499 use of subpatterns for more complicated assertions is described below.
4500 The backslashed assertions are:
4502 \b matches at a word boundary
4503 \B matches when not at a word boundary
4504 \A matches at the start of the subject
4505 \Z matches at the end of the subject
4506 also matches before a newline at the end of the subject
4507 \z matches only at the end of the subject
4508 \G matches at the first matching position in the subject
4510 Inside a character class, \b has a different meaning; it matches the
4511 backspace character. If any other of these assertions appears in a
4512 character class, by default it matches the corresponding literal char-
4513 acter (for example, \B matches the letter B). However, if the
4514 PCRE_EXTRA option is set, an "invalid escape sequence" error is gener-
4515 ated instead.
4517 A word boundary is a position in the subject string where the current
4518 character and the previous character do not both match \w or \W (i.e.
4519 one matches \w and the other matches \W), or the start or end of the
4520 string if the first or last character matches \w, respectively. In a
4521 UTF mode, the meanings of \w and \W can be changed by setting the
4522 PCRE_UCP option. When this is done, it also affects \b and \B. Neither
4523 PCRE nor Perl has a separate "start of word" or "end of word" metase-
4524 quence. However, whatever follows \b normally determines which it is.
4525 For example, the fragment \ba matches "a" at the start of a word.
4527 The \A, \Z, and \z assertions differ from the traditional circumflex
4528 and dollar (described in the next section) in that they only ever match
4529 at the very start and end of the subject string, whatever options are
4530 set. Thus, they are independent of multiline mode. These three asser-
4531 tions are not affected by the PCRE_NOTBOL or PCRE_NOTEOL options, which
4532 affect only the behaviour of the circumflex and dollar metacharacters.
4533 However, if the startoffset argument of pcre_exec() is non-zero, indi-
4534 cating that matching is to start at a point other than the beginning of
4535 the subject, \A can never match. The difference between \Z and \z is
4536 that \Z matches before a newline at the end of the string as well as at
4537 the very end, whereas \z matches only at the end.
4539 The \G assertion is true only when the current matching position is at
4540 the start point of the match, as specified by the startoffset argument
4541 of pcre_exec(). It differs from \A when the value of startoffset is
4542 non-zero. By calling pcre_exec() multiple times with appropriate argu-
4543 ments, you can mimic Perl's /g option, and it is in this kind of imple-
4544 mentation where \G can be useful.
4546 Note, however, that PCRE's interpretation of \G, as the start of the
4547 current match, is subtly different from Perl's, which defines it as the
4548 end of the previous match. In Perl, these can be different when the
4549 previously matched string was empty. Because PCRE does just one match
4550 at a time, it cannot reproduce this behaviour.
4552 If all the alternatives of a pattern begin with \G, the expression is
4553 anchored to the starting match position, and the "anchored" flag is set
4554 in the compiled regular expression.
4559 Outside a character class, in the default matching mode, the circumflex
4560 character is an assertion that is true only if the current matching
4561 point is at the start of the subject string. If the startoffset argu-
4562 ment of pcre_exec() is non-zero, circumflex can never match if the
4563 PCRE_MULTILINE option is unset. Inside a character class, circumflex
4564 has an entirely different meaning (see below).
4566 Circumflex need not be the first character of the pattern if a number
4567 of alternatives are involved, but it should be the first thing in each
4568 alternative in which it appears if the pattern is ever to match that
4569 branch. If all possible alternatives start with a circumflex, that is,
4570 if the pattern is constrained to match only at the start of the sub-
4571 ject, it is said to be an "anchored" pattern. (There are also other
4572 constructs that can cause a pattern to be anchored.)
4574 A dollar character is an assertion that is true only if the current
4575 matching point is at the end of the subject string, or immediately
4576 before a newline at the end of the string (by default). Dollar need not
4577 be the last character of the pattern if a number of alternatives are
4578 involved, but it should be the last item in any branch in which it
4579 appears. Dollar has no special meaning in a character class.
4581 The meaning of dollar can be changed so that it matches only at the
4582 very end of the string, by setting the PCRE_DOLLAR_ENDONLY option at
4583 compile time. This does not affect the \Z assertion.
4585 The meanings of the circumflex and dollar characters are changed if the
4586 PCRE_MULTILINE option is set. When this is the case, a circumflex
4587 matches immediately after internal newlines as well as at the start of
4588 the subject string. It does not match after a newline that ends the
4589 string. A dollar matches before any newlines in the string, as well as
4590 at the very end, when PCRE_MULTILINE is set. When newline is specified
4591 as the two-character sequence CRLF, isolated CR and LF characters do
4592 not indicate newlines.
4594 For example, the pattern /^abc$/ matches the subject string "def\nabc"
4595 (where \n represents a newline) in multiline mode, but not otherwise.
4596 Consequently, patterns that are anchored in single line mode because
4597 all branches start with ^ are not anchored in multiline mode, and a
4598 match for circumflex is possible when the startoffset argument of
4599 pcre_exec() is non-zero. The PCRE_DOLLAR_ENDONLY option is ignored if
4600 PCRE_MULTILINE is set.
4602 Note that the sequences \A, \Z, and \z can be used to match the start
4603 and end of the subject in both modes, and if all branches of a pattern
4604 start with \A it is always anchored, whether or not PCRE_MULTILINE is
4605 set.
4610 Outside a character class, a dot in the pattern matches any one charac-
4611 ter in the subject string except (by default) a character that signi-
4612 fies the end of a line.
4614 When a line ending is defined as a single character, dot never matches
4615 that character; when the two-character sequence CRLF is used, dot does
4616 not match CR if it is immediately followed by LF, but otherwise it
4617 matches all characters (including isolated CRs and LFs). When any Uni-
4618 code line endings are being recognized, dot does not match CR or LF or
4619 any of the other line ending characters.
4621 The behaviour of dot with regard to newlines can be changed. If the
4622 PCRE_DOTALL option is set, a dot matches any one character, without
4623 exception. If the two-character sequence CRLF is present in the subject
4624 string, it takes two dots to match it.
4626 The handling of dot is entirely independent of the handling of circum-
4627 flex and dollar, the only relationship being that they both involve
4628 newlines. Dot has no special meaning in a character class.
4630 The escape sequence \N behaves like a dot, except that it is not
4631 affected by the PCRE_DOTALL option. In other words, it matches any
4632 character except one that signifies the end of a line. Perl also uses
4633 \N to match characters by name; PCRE does not support this.
4638 Outside a character class, the escape sequence \C matches any one data
4639 unit, whether or not a UTF mode is set. In the 8-bit library, one data
4640 unit is one byte; in the 16-bit library it is a 16-bit unit. Unlike a
4641 dot, \C always matches line-ending characters. The feature is provided
4642 in Perl in order to match individual bytes in UTF-8 mode, but it is
4643 unclear how it can usefully be used. Because \C breaks up characters
4644 into individual data units, matching one unit with \C in a UTF mode
4645 means that the rest of the string may start with a malformed UTF char-
4646 acter. This has undefined results, because PCRE assumes that it is
4647 dealing with valid UTF strings (and by default it checks this at the
4648 start of processing unless the PCRE_NO_UTF8_CHECK option is used).
4650 PCRE does not allow \C to appear in lookbehind assertions (described
4651 below) in a UTF mode, because this would make it impossible to calcu-
4652 late the length of the lookbehind.
4654 In general, the \C escape sequence is best avoided. However, one way of
4655 using it that avoids the problem of malformed UTF characters is to use
4656 a lookahead to check the length of the next character, as in this pat-
4657 tern, which could be used with a UTF-8 string (ignore white space and
4658 line breaks):
4660 (?| (?=[\x00-\x7f])(\C) |
4661 (?=[\x80-\x{7ff}])(\C)(\C) |
4662 (?=[\x{800}-\x{ffff}])(\C)(\C)(\C) |
4663 (?=[\x{10000}-\x{1fffff}])(\C)(\C)(\C)(\C))
4665 A group that starts with (?| resets the capturing parentheses numbers
4666 in each alternative (see "Duplicate Subpattern Numbers" below). The
4667 assertions at the start of each branch check the next UTF-8 character
4668 for values whose encoding uses 1, 2, 3, or 4 bytes, respectively. The
4669 character's individual bytes are then captured by the appropriate num-
4670 ber of groups.
4675 An opening square bracket introduces a character class, terminated by a
4676 closing square bracket. A closing square bracket on its own is not spe-
4677 cial by default. However, if the PCRE_JAVASCRIPT_COMPAT option is set,
4678 a lone closing square bracket causes a compile-time error. If a closing
4679 square bracket is required as a member of the class, it should be the
4680 first data character in the class (after an initial circumflex, if
4681 present) or escaped with a backslash.
4683 A character class matches a single character in the subject. In a UTF
4684 mode, the character may be more than one data unit long. A matched
4685 character must be in the set of characters defined by the class, unless
4686 the first character in the class definition is a circumflex, in which
4687 case the subject character must not be in the set defined by the class.
4688 If a circumflex is actually required as a member of the class, ensure
4689 it is not the first character, or escape it with a backslash.
4691 For example, the character class [aeiou] matches any lower case vowel,
4692 while [^aeiou] matches any character that is not a lower case vowel.
4693 Note that a circumflex is just a convenient notation for specifying the
4694 characters that are in the class by enumerating those that are not. A
4695 class that starts with a circumflex is not an assertion; it still con-
4696 sumes a character from the subject string, and therefore it fails if
4697 the current pointer is at the end of the string.
4699 In UTF-8 (UTF-16) mode, characters with values greater than 255
4700 (0xffff) can be included in a class as a literal string of data units,
4701 or by using the \x{ escaping mechanism.
4703 When caseless matching is set, any letters in a class represent both
4704 their upper case and lower case versions, so for example, a caseless
4705 [aeiou] matches "A" as well as "a", and a caseless [^aeiou] does not
4706 match "A", whereas a caseful version would. In a UTF mode, PCRE always
4707 understands the concept of case for characters whose values are less
4708 than 128, so caseless matching is always possible. For characters with
4709 higher values, the concept of case is supported if PCRE is compiled
4710 with Unicode property support, but not otherwise. If you want to use
4711 caseless matching in a UTF mode for characters 128 and above, you must
4712 ensure that PCRE is compiled with Unicode property support as well as
4713 with UTF support.
4715 Characters that might indicate line breaks are never treated in any
4716 special way when matching character classes, whatever line-ending
4717 sequence is in use, and whatever setting of the PCRE_DOTALL and
4718 PCRE_MULTILINE options is used. A class such as [^a] always matches one
4719 of these characters.
4721 The minus (hyphen) character can be used to specify a range of charac-
4722 ters in a character class. For example, [d-m] matches any letter
4723 between d and m, inclusive. If a minus character is required in a
4724 class, it must be escaped with a backslash or appear in a position
4725 where it cannot be interpreted as indicating a range, typically as the
4726 first or last character in the class.
4728 It is not possible to have the literal character "]" as the end charac-
4729 ter of a range. A pattern such as [W-]46] is interpreted as a class of
4730 two characters ("W" and "-") followed by a literal string "46]", so it
4731 would match "W46]" or "-46]". However, if the "]" is escaped with a
4732 backslash it is interpreted as the end of range, so [W-\]46] is inter-
4733 preted as a class containing a range followed by two other characters.
4734 The octal or hexadecimal representation of "]" can also be used to end
4735 a range.
4737 Ranges operate in the collating sequence of character values. They can
4738 also be used for characters specified numerically, for example
4739 [\000-\037]. Ranges can include any characters that are valid for the
4740 current mode.
4742 If a range that includes letters is used when caseless matching is set,
4743 it matches the letters in either case. For example, [W-c] is equivalent
4744 to [][\\^_`wxyzabc], matched caselessly, and in a non-UTF mode, if
4745 character tables for a French locale are in use, [\xc8-\xcb] matches
4746 accented E characters in both cases. In UTF modes, PCRE supports the
4747 concept of case for characters with values greater than 128 only when
4748 it is compiled with Unicode property support.
4750 The character escape sequences \d, \D, \h, \H, \p, \P, \s, \S, \v, \V,
4751 \w, and \W may appear in a character class, and add the characters that
4752 they match to the class. For example, [\dABCDEF] matches any hexadeci-
4753 mal digit. In UTF modes, the PCRE_UCP option affects the meanings of
4754 \d, \s, \w and their upper case partners, just as it does when they
4755 appear outside a character class, as described in the section entitled
4756 "Generic character types" above. The escape sequence \b has a different
4757 meaning inside a character class; it matches the backspace character.
4758 The sequences \B, \N, \R, and \X are not special inside a character
4759 class. Like any other unrecognized escape sequences, they are treated
4760 as the literal characters "B", "N", "R", and "X" by default, but cause
4761 an error if the PCRE_EXTRA option is set.
4763 A circumflex can conveniently be used with the upper case character
4764 types to specify a more restricted set of characters than the matching
4765 lower case type. For example, the class [^\W_] matches any letter or
4766 digit, but not underscore, whereas [\w] includes underscore. A positive
4767 character class should be read as "something OR something OR ..." and a
4768 negative class as "NOT something AND NOT something AND NOT ...".
4770 The only metacharacters that are recognized in character classes are
4771 backslash, hyphen (only where it can be interpreted as specifying a
4772 range), circumflex (only at the start), opening square bracket (only
4773 when it can be interpreted as introducing a POSIX class name - see the
4774 next section), and the terminating closing square bracket. However,
4775 escaping other non-alphanumeric characters does no harm.
4780 Perl supports the POSIX notation for character classes. This uses names
4781 enclosed by [: and :] within the enclosing square brackets. PCRE also
4782 supports this notation. For example,
4784 [01[:alpha:]%]
4786 matches "0", "1", any alphabetic character, or "%". The supported class
4787 names are:
4789 alnum letters and digits
4790 alpha letters
4791 ascii character codes 0 - 127
4792 blank space or tab only
4793 cntrl control characters
4794 digit decimal digits (same as \d)
4795 graph printing characters, excluding space
4796 lower lower case letters
4797 print printing characters, including space
4798 punct printing characters, excluding letters and digits and space
4799 space white space (not quite the same as \s)
4800 upper upper case letters
4801 word "word" characters (same as \w)
4802 xdigit hexadecimal digits
4804 The "space" characters are HT (9), LF (10), VT (11), FF (12), CR (13),
4805 and space (32). Notice that this list includes the VT character (code
4806 11). This makes "space" different to \s, which does not include VT (for
4807 Perl compatibility).
4809 The name "word" is a Perl extension, and "blank" is a GNU extension
4810 from Perl 5.8. Another Perl extension is negation, which is indicated
4811 by a ^ character after the colon. For example,
4813 [12[:^digit:]]
4815 matches "1", "2", or any non-digit. PCRE (and Perl) also recognize the
4816 POSIX syntax [.ch.] and [=ch=] where "ch" is a "collating element", but
4817 these are not supported, and an error is given if they are encountered.
4819 By default, in UTF modes, characters with values greater than 128 do
4820 not match any of the POSIX character classes. However, if the PCRE_UCP
4821 option is passed to pcre_compile(), some of the classes are changed so
4822 that Unicode character properties are used. This is achieved by replac-
4823 ing the POSIX classes by other sequences, as follows:
4825 [:alnum:] becomes \p{Xan}
4826 [:alpha:] becomes \p{L}
4827 [:blank:] becomes \h
4828 [:digit:] becomes \p{Nd}
4829 [:lower:] becomes \p{Ll}
4830 [:space:] becomes \p{Xps}
4831 [:upper:] becomes \p{Lu}
4832 [:word:] becomes \p{Xwd}
4834 Negated versions, such as [:^alpha:] use \P instead of \p. The other
4835 POSIX classes are unchanged, and match only characters with code points
4836 less than 128.
4841 Vertical bar characters are used to separate alternative patterns. For
4842 example, the pattern
4844 gilbert|sullivan
4846 matches either "gilbert" or "sullivan". Any number of alternatives may
4847 appear, and an empty alternative is permitted (matching the empty
4848 string). The matching process tries each alternative in turn, from left
4849 to right, and the first one that succeeds is used. If the alternatives
4850 are within a subpattern (defined below), "succeeds" means matching the
4851 rest of the main pattern as well as the alternative in the subpattern.
4856 The settings of the PCRE_CASELESS, PCRE_MULTILINE, PCRE_DOTALL, and
4857 PCRE_EXTENDED options (which are Perl-compatible) can be changed from
4858 within the pattern by a sequence of Perl option letters enclosed
4859 between "(?" and ")". The option letters are
4861 i for PCRE_CASELESS
4863 s for PCRE_DOTALL
4864 x for PCRE_EXTENDED
4866 For example, (?im) sets caseless, multiline matching. It is also possi-
4867 ble to unset these options by preceding the letter with a hyphen, and a
4868 combined setting and unsetting such as (?im-sx), which sets PCRE_CASE-
4870 is also permitted. If a letter appears both before and after the
4871 hyphen, the option is unset.
4874 can be changed in the same way as the Perl-compatible options by using
4875 the characters J, U and X respectively.
4877 When one of these option changes occurs at top level (that is, not
4878 inside subpattern parentheses), the change applies to the remainder of
4879 the pattern that follows. If the change is placed right at the start of
4880 a pattern, PCRE extracts it into the global options (and it will there-
4881 fore show up in data extracted by the pcre_fullinfo() function).
4883 An option change within a subpattern (see below for a description of
4884 subpatterns) affects only that part of the subpattern that follows it,
4885 so
4887 (a(?i)b)c
4889 matches abc and aBc and no other strings (assuming PCRE_CASELESS is not
4890 used). By this means, options can be made to have different settings
4891 in different parts of the pattern. Any changes made in one alternative
4892 do carry on into subsequent branches within the same subpattern. For
4893 example,
4895 (a(?i)b|c)
4897 matches "ab", "aB", "c", and "C", even though when matching "C" the
4898 first branch is abandoned before the option setting. This is because
4899 the effects of option settings happen at compile time. There would be
4900 some very weird behaviour otherwise.
4902 Note: There are other PCRE-specific options that can be set by the
4903 application when the compiling or matching functions are called. In
4904 some cases the pattern can contain special leading sequences such as
4905 (*CRLF) to override what the application has set or what has been
4906 defaulted. Details are given in the section entitled "Newline
4907 sequences" above. There are also the (*UTF8), (*UTF16), and (*UCP)
4908 leading sequences that can be used to set UTF and Unicode property
4909 modes; they are equivalent to setting the PCRE_UTF8, PCRE_UTF16, and
4910 the PCRE_UCP options, respectively.
4915 Subpatterns are delimited by parentheses (round brackets), which can be
4916 nested. Turning part of a pattern into a subpattern does two things:
4918 1. It localizes a set of alternatives. For example, the pattern
4920 cat(aract|erpillar|)
4922 matches "cataract", "caterpillar", or "cat". Without the parentheses,
4923 it would match "cataract", "erpillar" or an empty string.
4925 2. It sets up the subpattern as a capturing subpattern. This means
4926 that, when the whole pattern matches, that portion of the subject
4927 string that matched the subpattern is passed back to the caller via the
4928 ovector argument of the matching function. (This applies only to the
4929 traditional matching functions; the DFA matching functions do not sup-
4930 port capturing.)
4932 Opening parentheses are counted from left to right (starting from 1) to
4933 obtain numbers for the capturing subpatterns. For example, if the
4934 string "the red king" is matched against the pattern
4936 the ((red|white) (king|queen))
4938 the captured substrings are "red king", "red", and "king", and are num-
4939 bered 1, 2, and 3, respectively.
4941 The fact that plain parentheses fulfil two functions is not always
4942 helpful. There are often times when a grouping subpattern is required
4943 without a capturing requirement. If an opening parenthesis is followed
4944 by a question mark and a colon, the subpattern does not do any captur-
4945 ing, and is not counted when computing the number of any subsequent
4946 capturing subpatterns. For example, if the string "the white queen" is
4947 matched against the pattern
4949 the ((?:red|white) (king|queen))
4951 the captured substrings are "white queen" and "queen", and are numbered
4952 1 and 2. The maximum number of capturing subpatterns is 65535.
4954 As a convenient shorthand, if any option settings are required at the
4955 start of a non-capturing subpattern, the option letters may appear
4956 between the "?" and the ":". Thus the two patterns
4958 (?i:saturday|sunday)
4959 (?:(?i)saturday|sunday)
4961 match exactly the same set of strings. Because alternative branches are
4962 tried from left to right, and options are not reset until the end of
4963 the subpattern is reached, an option setting in one branch does affect
4964 subsequent branches, so the above patterns match "SUNDAY" as well as
4965 "Saturday".
4970 Perl 5.10 introduced a feature whereby each alternative in a subpattern
4971 uses the same numbers for its capturing parentheses. Such a subpattern
4972 starts with (?| and is itself a non-capturing subpattern. For example,
4973 consider this pattern:
4975 (?|(Sat)ur|(Sun))day
4977 Because the two alternatives are inside a (?| group, both sets of cap-
4978 turing parentheses are numbered one. Thus, when the pattern matches,
4979 you can look at captured substring number one, whichever alternative
4980 matched. This construct is useful when you want to capture part, but
4981 not all, of one of a number of alternatives. Inside a (?| group, paren-
4982 theses are numbered as usual, but the number is reset at the start of
4983 each branch. The numbers of any capturing parentheses that follow the
4984 subpattern start after the highest number used in any branch. The fol-
4985 lowing example is taken from the Perl documentation. The numbers under-
4986 neath show in which buffer the captured content will be stored.
4988 # before ---------------branch-reset----------- after
4989 / ( a ) (?| x ( y ) z | (p (q) r) | (t) u (v) ) ( z ) /x
4990 # 1 2 2 3 2 3 4
4992 A back reference to a numbered subpattern uses the most recent value
4993 that is set for that number by any subpattern. The following pattern
4994 matches "abcabc" or "defdef":
4996 /(?|(abc)|(def))\1/
4998 In contrast, a subroutine call to a numbered subpattern always refers
4999 to the first one in the pattern with the given number. The following
5000 pattern matches "abcabc" or "defabc":
5002 /(?|(abc)|(def))(?1)/
5004 If a condition test for a subpattern's having matched refers to a non-
5005 unique number, the test is true if any of the subpatterns of that num-
5006 ber have matched.
5008 An alternative approach to using this "branch reset" feature is to use
5009 duplicate named subpatterns, as described in the next section.
5014 Identifying capturing parentheses by number is simple, but it can be
5015 very hard to keep track of the numbers in complicated regular expres-
5016 sions. Furthermore, if an expression is modified, the numbers may
5017 change. To help with this difficulty, PCRE supports the naming of sub-
5018 patterns. This feature was not added to Perl until release 5.10. Python
5019 had the feature earlier, and PCRE introduced it at release 4.0, using
5020 the Python syntax. PCRE now supports both the Perl and the Python syn-
5021 tax. Perl allows identically numbered subpatterns to have different
5022 names, but PCRE does not.
5024 In PCRE, a subpattern can be named in one of three ways: (?<name>...)
5025 or (?'name'...) as in Perl, or (?P<name>...) as in Python. References
5026 to capturing parentheses from other parts of the pattern, such as back
5027 references, recursion, and conditions, can be made by name as well as
5028 by number.
5030 Names consist of up to 32 alphanumeric characters and underscores.
5031 Named capturing parentheses are still allocated numbers as well as
5032 names, exactly as if the names were not present. The PCRE API provides
5033 function calls for extracting the name-to-number translation table from
5034 a compiled pattern. There is also a convenience function for extracting
5035 a captured substring by name.
5037 By default, a name must be unique within a pattern, but it is possible
5038 to relax this constraint by setting the PCRE_DUPNAMES option at compile
5039 time. (Duplicate names are also always permitted for subpatterns with
5040 the same number, set up as described in the previous section.) Dupli-
5041 cate names can be useful for patterns where only one instance of the
5042 named parentheses can match. Suppose you want to match the name of a
5043 weekday, either as a 3-letter abbreviation or as the full name, and in
5044 both cases you want to extract the abbreviation. This pattern (ignoring
5045 the line breaks) does the job:
5047 (?<DN>Mon|Fri|Sun)(?:day)?|
5048 (?<DN>Tue)(?:sday)?|
5049 (?<DN>Wed)(?:nesday)?|
5050 (?<DN>Thu)(?:rsday)?|
5051 (?<DN>Sat)(?:urday)?
5053 There are five capturing substrings, but only one is ever set after a
5054 match. (An alternative way of solving this problem is to use a "branch
5055 reset" subpattern, as described in the previous section.)
5057 The convenience function for extracting the data by name returns the
5058 substring for the first (and in this example, the only) subpattern of
5059 that name that matched. This saves searching to find which numbered
5060 subpattern it was.
5062 If you make a back reference to a non-unique named subpattern from
5063 elsewhere in the pattern, the one that corresponds to the first occur-
5064 rence of the name is used. In the absence of duplicate numbers (see the
5065 previous section) this is the one with the lowest number. If you use a
5066 named reference in a condition test (see the section about conditions
5067 below), either to check whether a subpattern has matched, or to check
5068 for recursion, all subpatterns with the same name are tested. If the
5069 condition is true for any one of them, the overall condition is true.
5070 This is the same behaviour as testing by number. For further details of
5071 the interfaces for handling named subpatterns, see the pcreapi documen-
5072 tation.
5074 Warning: You cannot use different names to distinguish between two sub-
5075 patterns with the same number because PCRE uses only the numbers when
5076 matching. For this reason, an error is given at compile time if differ-
5077 ent names are given to subpatterns with the same number. However, you
5078 can give the same name to subpatterns with the same number, even when
5079 PCRE_DUPNAMES is not set.
5084 Repetition is specified by quantifiers, which can follow any of the
5085 following items:
5087 a literal data character
5088 the dot metacharacter
5089 the \C escape sequence
5090 the \X escape sequence
5091 the \R escape sequence
5092 an escape such as \d or \pL that matches a single character
5093 a character class
5094 a back reference (see next section)
5095 a parenthesized subpattern (including assertions)
5096 a subroutine call to a subpattern (recursive or otherwise)
5098 The general repetition quantifier specifies a minimum and maximum num-
5099 ber of permitted matches, by giving the two numbers in curly brackets
5100 (braces), separated by a comma. The numbers must be less than 65536,
5101 and the first must be less than or equal to the second. For example:
5103 z{2,4}
5105 matches "zz", "zzz", or "zzzz". A closing brace on its own is not a
5106 special character. If the second number is omitted, but the comma is
5107 present, there is no upper limit; if the second number and the comma
5108 are both omitted, the quantifier specifies an exact number of required
5109 matches. Thus
5111 [aeiou]{3,}
5113 matches at least 3 successive vowels, but may match many more, while
5115 \d{8}
5117 matches exactly 8 digits. An opening curly bracket that appears in a
5118 position where a quantifier is not allowed, or one that does not match
5119 the syntax of a quantifier, is taken as a literal character. For exam-
5120 ple, {,6} is not a quantifier, but a literal string of four characters.
5122 In UTF modes, quantifiers apply to characters rather than to individual
5123 data units. Thus, for example, \x{100}{2} matches two characters, each
5124 of which is represented by a two-byte sequence in a UTF-8 string. Simi-
5125 larly, \X{3} matches three Unicode extended sequences, each of which
5126 may be several data units long (and they may be of different lengths).
5128 The quantifier {0} is permitted, causing the expression to behave as if
5129 the previous item and the quantifier were not present. This may be use-
5130 ful for subpatterns that are referenced as subroutines from elsewhere
5131 in the pattern (but see also the section entitled "Defining subpatterns
5132 for use by reference only" below). Items other than subpatterns that
5133 have a {0} quantifier are omitted from the compiled pattern.
5135 For convenience, the three most common quantifiers have single-charac-
5136 ter abbreviations:
5138 * is equivalent to {0,}
5139 + is equivalent to {1,}
5140 ? is equivalent to {0,1}
5142 It is possible to construct infinite loops by following a subpattern
5143 that can match no characters with a quantifier that has no upper limit,
5144 for example:
5146 (a?)*
5148 Earlier versions of Perl and PCRE used to give an error at compile time
5149 for such patterns. However, because there are cases where this can be
5150 useful, such patterns are now accepted, but if any repetition of the
5151 subpattern does in fact match no characters, the loop is forcibly bro-
5152 ken.
5154 By default, the quantifiers are "greedy", that is, they match as much
5155 as possible (up to the maximum number of permitted times), without
5156 causing the rest of the pattern to fail. The classic example of where
5157 this gives problems is in trying to match comments in C programs. These
5158 appear between /* and */ and within the comment, individual * and /
5159 characters may appear. An attempt to match C comments by applying the
5160 pattern
5162 /\*.*\*/
5164 to the string
5166 /* first comment */ not comment /* second comment */
5168 fails, because it matches the entire string owing to the greediness of
5169 the .* item.
5171 However, if a quantifier is followed by a question mark, it ceases to
5172 be greedy, and instead matches the minimum number of times possible, so
5173 the pattern
5175 /\*.*?\*/
5177 does the right thing with the C comments. The meaning of the various
5178 quantifiers is not otherwise changed, just the preferred number of
5179 matches. Do not confuse this use of question mark with its use as a
5180 quantifier in its own right. Because it has two uses, it can sometimes
5181 appear doubled, as in
5183 \d??\d
5185 which matches one digit by preference, but can match two if that is the
5186 only way the rest of the pattern matches.
5188 If the PCRE_UNGREEDY option is set (an option that is not available in
5189 Perl), the quantifiers are not greedy by default, but individual ones
5190 can be made greedy by following them with a question mark. In other
5191 words, it inverts the default behaviour.
5193 When a parenthesized subpattern is quantified with a minimum repeat
5194 count that is greater than 1 or with a limited maximum, more memory is
5195 required for the compiled pattern, in proportion to the size of the
5196 minimum or maximum.
5198 If a pattern starts with .* or .{0,} and the PCRE_DOTALL option (equiv-
5199 alent to Perl's /s) is set, thus allowing the dot to match newlines,
5200 the pattern is implicitly anchored, because whatever follows will be
5201 tried against every character position in the subject string, so there
5202 is no point in retrying the overall match at any position after the
5203 first. PCRE normally treats such a pattern as though it were preceded
5204 by \A.
5206 In cases where it is known that the subject string contains no new-
5207 lines, it is worth setting PCRE_DOTALL in order to obtain this opti-
5208 mization, or alternatively using ^ to indicate anchoring explicitly.
5210 However, there is one situation where the optimization cannot be used.
5211 When .* is inside capturing parentheses that are the subject of a back
5212 reference elsewhere in the pattern, a match at the start may fail where
5213 a later one succeeds. Consider, for example:
5215 (.*)abc\1
5217 If the subject is "xyz123abc123" the match point is the fourth charac-
5218 ter. For this reason, such a pattern is not implicitly anchored.
5220 When a capturing subpattern is repeated, the value captured is the sub-
5221 string that matched the final iteration. For example, after
5223 (tweedle[dume]{3}\s*)+
5225 has matched "tweedledum tweedledee" the value of the captured substring
5226 is "tweedledee". However, if there are nested capturing subpatterns,
5227 the corresponding captured values may have been set in previous itera-
5228 tions. For example, after
5230 /(a|(b))+/
5232 matches "aba" the value of the second captured substring is "b".
5237 With both maximizing ("greedy") and minimizing ("ungreedy" or "lazy")
5238 repetition, failure of what follows normally causes the repeated item
5239 to be re-evaluated to see if a different number of repeats allows the
5240 rest of the pattern to match. Sometimes it is useful to prevent this,
5241 either to change the nature of the match, or to cause it fail earlier
5242 than it otherwise might, when the author of the pattern knows there is
5243 no point in carrying on.
5245 Consider, for example, the pattern \d+foo when applied to the subject
5246 line
5248 123456bar
5250 After matching all 6 digits and then failing to match "foo", the normal
5251 action of the matcher is to try again with only 5 digits matching the
5252 \d+ item, and then with 4, and so on, before ultimately failing.
5253 "Atomic grouping" (a term taken from Jeffrey Friedl's book) provides
5254 the means for specifying that once a subpattern has matched, it is not
5255 to be re-evaluated in this way.
5257 If we use atomic grouping for the previous example, the matcher gives
5258 up immediately on failing to match "foo" the first time. The notation
5259 is a kind of special parenthesis, starting with (?> as in this example:
5261 (?>\d+)foo
5263 This kind of parenthesis "locks up" the part of the pattern it con-
5264 tains once it has matched, and a failure further into the pattern is
5265 prevented from backtracking into it. Backtracking past it to previous
5266 items, however, works as normal.
5268 An alternative description is that a subpattern of this type matches
5269 the string of characters that an identical standalone pattern would
5270 match, if anchored at the current point in the subject string.
5272 Atomic grouping subpatterns are not capturing subpatterns. Simple cases
5273 such as the above example can be thought of as a maximizing repeat that
5274 must swallow everything it can. So, while both \d+ and \d+? are pre-
5275 pared to adjust the number of digits they match in order to make the
5276 rest of the pattern match, (?>\d+) can only match an entire sequence of
5277 digits.
5279 Atomic groups in general can of course contain arbitrarily complicated
5280 subpatterns, and can be nested. However, when the subpattern for an
5281 atomic group is just a single repeated item, as in the example above, a
5282 simpler notation, called a "possessive quantifier" can be used. This
5283 consists of an additional + character following a quantifier. Using
5284 this notation, the previous example can be rewritten as
5286 \d++foo
5288 Note that a possessive quantifier can be used with an entire group, for
5289 example:
5291 (abc|xyz){2,3}+
5293 Possessive quantifiers are always greedy; the setting of the
5294 PCRE_UNGREEDY option is ignored. They are a convenient notation for the
5295 simpler forms of atomic group. However, there is no difference in the
5296 meaning of a possessive quantifier and the equivalent atomic group,
5297 though there may be a performance difference; possessive quantifiers
5298 should be slightly faster.
5300 The possessive quantifier syntax is an extension to the Perl 5.8 syn-
5301 tax. Jeffrey Friedl originated the idea (and the name) in the first
5302 edition of his book. Mike McCloskey liked it, so implemented it when he
5303 built Sun's Java package, and PCRE copied it from there. It ultimately
5304 found its way into Perl at release 5.10.
5306 PCRE has an optimization that automatically "possessifies" certain sim-
5307 ple pattern constructs. For example, the sequence A+B is treated as
5308 A++B because there is no point in backtracking into a sequence of A's
5309 when B must follow.
5311 When a pattern contains an unlimited repeat inside a subpattern that
5312 can itself be repeated an unlimited number of times, the use of an
5313 atomic group is the only way to avoid some failing matches taking a
5314 very long time indeed. The pattern
5316 (\D+|<\d+>)*[!?]
5318 matches an unlimited number of substrings that either consist of non-
5319 digits, or digits enclosed in <>, followed by either ! or ?. When it
5320 matches, it runs quickly. However, if it is applied to
5322 aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
5324 it takes a long time before reporting failure. This is because the
5325 string can be divided between the internal \D+ repeat and the external
5326 * repeat in a large number of ways, and all have to be tried. (The
5327 example uses [!?] rather than a single character at the end, because
5328 both PCRE and Perl have an optimization that allows for fast failure
5329 when a single character is used. They remember the last single charac-
5330 ter that is required for a match, and fail early if it is not present
5331 in the string.) If the pattern is changed so that it uses an atomic
5332 group, like this:
5334 ((?>\D+)|<\d+>)*[!?]
5336 sequences of non-digits cannot be broken, and failure happens quickly.
5341 Outside a character class, a backslash followed by a digit greater than
5342 0 (and possibly further digits) is a back reference to a capturing sub-
5343 pattern earlier (that is, to its left) in the pattern, provided there
5344 have been that many previous capturing left parentheses.
5346 However, if the decimal number following the backslash is less than 10,
5347 it is always taken as a back reference, and causes an error only if
5348 there are not that many capturing left parentheses in the entire pat-
5349 tern. In other words, the parentheses that are referenced need not be
5350 to the left of the reference for numbers less than 10. A "forward back
5351 reference" of this type can make sense when a repetition is involved
5352 and the subpattern to the right has participated in an earlier itera-
5353 tion.
5355 It is not possible to have a numerical "forward back reference" to a
5356 subpattern whose number is 10 or more using this syntax because a
5357 sequence such as \50 is interpreted as a character defined in octal.
5358 See the subsection entitled "Non-printing characters" above for further
5359 details of the handling of digits following a backslash. There is no
5360 such problem when named parentheses are used. A back reference to any
5361 subpattern is possible using named parentheses (see below).
5363 Another way of avoiding the ambiguity inherent in the use of digits
5364 following a backslash is to use the \g escape sequence. This escape
5365 must be followed by an unsigned number or a negative number, optionally
5366 enclosed in braces. These examples are all identical:
5368 (ring), \1
5369 (ring), \g1
5370 (ring), \g{1}
5372 An unsigned number specifies an absolute reference without the ambigu-
5373 ity that is present in the older syntax. It is also useful when literal
5374 digits follow the reference. A negative number is a relative reference.
5375 Consider this example:
5377 (abc(def)ghi)\g{-1}
5379 The sequence \g{-1} is a reference to the most recently started captur-
5380 ing subpattern before \g, that is, is it equivalent to \2 in this exam-
5381 ple. Similarly, \g{-2} would be equivalent to \1. The use of relative
5382 references can be helpful in long patterns, and also in patterns that
5383 are created by joining together fragments that contain references
5384 within themselves.
5386 A back reference matches whatever actually matched the capturing sub-
5387 pattern in the current subject string, rather than anything matching
5388 the subpattern itself (see "Subpatterns as subroutines" below for a way
5389 of doing that). So the pattern
5391 (sens|respons)e and \1ibility
5393 matches "sense and sensibility" and "response and responsibility", but
5394 not "sense and responsibility". If caseful matching is in force at the
5395 time of the back reference, the case of letters is relevant. For exam-
5396 ple,
5398 ((?i)rah)\s+\1
5400 matches "rah rah" and "RAH RAH", but not "RAH rah", even though the
5401 original capturing subpattern is matched caselessly.
5403 There are several different ways of writing back references to named
5404 subpatterns. The .NET syntax \k{name} and the Perl syntax \k<name> or
5405 \k'name' are supported, as is the Python syntax (?P=name). Perl 5.10's
5406 unified back reference syntax, in which \g can be used for both numeric
5407 and named references, is also supported. We could rewrite the above
5408 example in any of the following ways:
5410 (?<p1>(?i)rah)\s+\k<p1>
5411 (?'p1'(?i)rah)\s+\k{p1}
5412 (?P<p1>(?i)rah)\s+(?P=p1)
5413 (?<p1>(?i)rah)\s+\g{p1}
5415 A subpattern that is referenced by name may appear in the pattern
5416 before or after the reference.
5418 There may be more than one back reference to the same subpattern. If a
5419 subpattern has not actually been used in a particular match, any back
5420 references to it always fail by default. For example, the pattern
5422 (a|(bc))\2
5424 always fails if it starts to match "a" rather than "bc". However, if
5425 the PCRE_JAVASCRIPT_COMPAT option is set at compile time, a back refer-
5426 ence to an unset value matches an empty string.
5428 Because there may be many capturing parentheses in a pattern, all dig-
5429 its following a backslash are taken as part of a potential back refer-
5430 ence number. If the pattern continues with a digit character, some
5431 delimiter must be used to terminate the back reference. If the
5432 PCRE_EXTENDED option is set, this can be whitespace. Otherwise, the \g{
5433 syntax or an empty comment (see "Comments" below) can be used.
5435 Recursive back references
5437 A back reference that occurs inside the parentheses to which it refers
5438 fails when the subpattern is first used, so, for example, (a\1) never
5439 matches. However, such references can be useful inside repeated sub-
5440 patterns. For example, the pattern
5442 (a|b\1)+
5444 matches any number of "a"s and also "aba", "ababbaa" etc. At each iter-
5445 ation of the subpattern, the back reference matches the character
5446 string corresponding to the previous iteration. In order for this to
5447 work, the pattern must be such that the first iteration does not need
5448 to match the back reference. This can be done using alternation, as in
5449 the example above, or by a quantifier with a minimum of zero.
5451 Back references of this type cause the group that they reference to be
5452 treated as an atomic group. Once the whole group has been matched, a
5453 subsequent matching failure cannot cause backtracking into the middle
5454 of the group.
5459 An assertion is a test on the characters following or preceding the
5460 current matching point that does not actually consume any characters.
5461 The simple assertions coded as \b, \B, \A, \G, \Z, \z, ^ and $ are
5462 described above.
5464 More complicated assertions are coded as subpatterns. There are two
5465 kinds: those that look ahead of the current position in the subject
5466 string, and those that look behind it. An assertion subpattern is
5467 matched in the normal way, except that it does not cause the current
5468 matching position to be changed.
5470 Assertion subpatterns are not capturing subpatterns. If such an asser-
5471 tion contains capturing subpatterns within it, these are counted for
5472 the purposes of numbering the capturing subpatterns in the whole pat-
5473 tern. However, substring capturing is carried out only for positive
5474 assertions, because it does not make sense for negative assertions.
5476 For compatibility with Perl, assertion subpatterns may be repeated;
5477 though it makes no sense to assert the same thing several times, the
5478 side effect of capturing parentheses may occasionally be useful. In
5479 practice, there only three cases:
5481 (1) If the quantifier is {0}, the assertion is never obeyed during
5482 matching. However, it may contain internal capturing parenthesized
5483 groups that are called from elsewhere via the subroutine mechanism.
5485 (2) If quantifier is {0,n} where n is greater than zero, it is treated
5486 as if it were {0,1}. At run time, the rest of the pattern match is
5487 tried with and without the assertion, the order depending on the greed-
5488 iness of the quantifier.
5490 (3) If the minimum repetition is greater than zero, the quantifier is
5491 ignored. The assertion is obeyed just once when encountered during
5492 matching.
5494 Lookahead assertions
5496 Lookahead assertions start with (?= for positive assertions and (?! for
5497 negative assertions. For example,
5499 \w+(?=;)
5501 matches a word followed by a semicolon, but does not include the semi-
5502 colon in the match, and
5504 foo(?!bar)
5506 matches any occurrence of "foo" that is not followed by "bar". Note
5507 that the apparently similar pattern
5509 (?!foo)bar
5511 does not find an occurrence of "bar" that is preceded by something
5512 other than "foo"; it finds any occurrence of "bar" whatsoever, because
5513 the assertion (?!foo) is always true when the next three characters are
5514 "bar". A lookbehind assertion is needed to achieve the other effect.
5516 If you want to force a matching failure at some point in a pattern, the
5517 most convenient way to do it is with (?!) because an empty string
5518 always matches, so an assertion that requires there not to be an empty
5519 string must always fail. The backtracking control verb (*FAIL) or (*F)
5520 is a synonym for (?!).
5522 Lookbehind assertions
5524 Lookbehind assertions start with (?<= for positive assertions and (?<!
5525 for negative assertions. For example,
5527 (?<!foo)bar
5529 does find an occurrence of "bar" that is not preceded by "foo". The
5530 contents of a lookbehind assertion are restricted such that all the
5531 strings it matches must have a fixed length. However, if there are sev-
5532 eral top-level alternatives, they do not all have to have the same
5533 fixed length. Thus
5535 (?<=bullock|donkey)
5537 is permitted, but
5539 (?<!dogs?|cats?)
5541 causes an error at compile time. Branches that match different length
5542 strings are permitted only at the top level of a lookbehind assertion.
5543 This is an extension compared with Perl, which requires all branches to
5544 match the same length of string. An assertion such as
5546 (?<=ab(c|de))
5548 is not permitted, because its single top-level branch can match two
5549 different lengths, but it is acceptable to PCRE if rewritten to use two
5550 top-level branches:
5552 (?<=abc|abde)
5554 In some cases, the escape sequence \K (see above) can be used instead
5555 of a lookbehind assertion to get round the fixed-length restriction.
5557 The implementation of lookbehind assertions is, for each alternative,
5558 to temporarily move the current position back by the fixed length and
5559 then try to match. If there are insufficient characters before the cur-
5560 rent position, the assertion fails.
5562 In a UTF mode, PCRE does not allow the \C escape (which matches a sin-
5563 gle data unit even in a UTF mode) to appear in lookbehind assertions,
5564 because it makes it impossible to calculate the length of the lookbe-
5565 hind. The \X and \R escapes, which can match different numbers of data
5566 units, are also not permitted.
5568 "Subroutine" calls (see below) such as (?2) or (?&X) are permitted in
5569 lookbehinds, as long as the subpattern matches a fixed-length string.
5570 Recursion, however, is not supported.
5572 Possessive quantifiers can be used in conjunction with lookbehind
5573 assertions to specify efficient matching of fixed-length strings at the
5574 end of subject strings. Consider a simple pattern such as
5576 abcd$
5578 when applied to a long string that does not match. Because matching
5579 proceeds from left to right, PCRE will look for each "a" in the subject
5580 and then see if what follows matches the rest of the pattern. If the
5581 pattern is specified as
5583 ^.*abcd$
5585 the initial .* matches the entire string at first, but when this fails
5586 (because there is no following "a"), it backtracks to match all but the
5587 last character, then all but the last two characters, and so on. Once
5588 again the search for "a" covers the entire string, from right to left,
5589 so we are no better off. However, if the pattern is written as
5591 ^.*+(?<=abcd)
5593 there can be no backtracking for the .*+ item; it can match only the
5594 entire string. The subsequent lookbehind assertion does a single test
5595 on the last four characters. If it fails, the match fails immediately.
5596 For long strings, this approach makes a significant difference to the
5597 processing time.
5599 Using multiple assertions
5601 Several assertions (of any sort) may occur in succession. For example,
5603 (?<=\d{3})(?<!999)foo
5605 matches "foo" preceded by three digits that are not "999". Notice that
5606 each of the assertions is applied independently at the same point in
5607 the subject string. First there is a check that the previous three
5608 characters are all digits, and then there is a check that the same
5609 three characters are not "999". This pattern does not match "foo" pre-
5610 ceded by six characters, the first of which are digits and the last
5611 three of which are not "999". For example, it doesn't match "123abc-
5612 foo". A pattern to do that is
5614 (?<=\d{3}...)(?<!999)foo
5616 This time the first assertion looks at the preceding six characters,
5617 checking that the first three are digits, and then the second assertion
5618 checks that the preceding three characters are not "999".
5620 Assertions can be nested in any combination. For example,
5622 (?<=(?<!foo)bar)baz
5624 matches an occurrence of "baz" that is preceded by "bar" which in turn
5625 is not preceded by "foo", while
5627 (?<=\d{3}(?!999)...)foo
5629 is another pattern that matches "foo" preceded by three digits and any
5630 three characters that are not "999".
5635 It is possible to cause the matching process to obey a subpattern con-
5636 ditionally or to choose between two alternative subpatterns, depending
5637 on the result of an assertion, or whether a specific capturing subpat-
5638 tern has already been matched. The two possible forms of conditional
5639 subpattern are:
5641 (?(condition)yes-pattern)
5642 (?(condition)yes-pattern|no-pattern)
5644 If the condition is satisfied, the yes-pattern is used; otherwise the
5645 no-pattern (if present) is used. If there are more than two alterna-
5646 tives in the subpattern, a compile-time error occurs. Each of the two
5647 alternatives may itself contain nested subpatterns of any form, includ-
5648 ing conditional subpatterns; the restriction to two alternatives
5649 applies only at the level of the condition. This pattern fragment is an
5650 example where the alternatives are complex:
5652 (?(1) (A|B|C) | (D | (?(2)E|F) | E) )
5655 There are four kinds of condition: references to subpatterns, refer-
5656 ences to recursion, a pseudo-condition called DEFINE, and assertions.
5658 Checking for a used subpattern by number
5660 If the text between the parentheses consists of a sequence of digits,
5661 the condition is true if a capturing subpattern of that number has pre-
5662 viously matched. If there is more than one capturing subpattern with
5663 the same number (see the earlier section about duplicate subpattern
5664 numbers), the condition is true if any of them have matched. An alter-
5665 native notation is to precede the digits with a plus or minus sign. In
5666 this case, the subpattern number is relative rather than absolute. The
5667 most recently opened parentheses can be referenced by (?(-1), the next
5668 most recent by (?(-2), and so on. Inside loops it can also make sense
5669 to refer to subsequent groups. The next parentheses to be opened can be
5670 referenced as (?(+1), and so on. (The value zero in any of these forms
5671 is not used; it provokes a compile-time error.)
5673 Consider the following pattern, which contains non-significant white
5674 space to make it more readable (assume the PCRE_EXTENDED option) and to
5675 divide it into three parts for ease of discussion:
5677 ( \( )? [^()]+ (?(1) \) )
5679 The first part matches an optional opening parenthesis, and if that
5680 character is present, sets it as the first captured substring. The sec-
5681 ond part matches one or more characters that are not parentheses. The
5682 third part is a conditional subpattern that tests whether or not the
5683 first set of parentheses matched. If they did, that is, if subject
5684 started with an opening parenthesis, the condition is true, and so the
5685 yes-pattern is executed and a closing parenthesis is required. Other-
5686 wise, since no-pattern is not present, the subpattern matches nothing.
5687 In other words, this pattern matches a sequence of non-parentheses,
5688 optionally enclosed in parentheses.
5690 If you were embedding this pattern in a larger one, you could use a
5691 relative reference:
5693 ...other stuff... ( \( )? [^()]+ (?(-1) \) ) ...
5695 This makes the fragment independent of the parentheses in the larger
5696 pattern.
5698 Checking for a used subpattern by name
5700 Perl uses the syntax (?(<name>)...) or (?('name')...) to test for a
5701 used subpattern by name. For compatibility with earlier versions of
5702 PCRE, which had this facility before Perl, the syntax (?(name)...) is
5703 also recognized. However, there is a possible ambiguity with this syn-
5704 tax, because subpattern names may consist entirely of digits. PCRE
5705 looks first for a named subpattern; if it cannot find one and the name
5706 consists entirely of digits, PCRE looks for a subpattern of that num-
5707 ber, which must be greater than zero. Using subpattern names that con-
5708 sist entirely of digits is not recommended.
5710 Rewriting the above example to use a named subpattern gives this:
5712 (?<OPEN> \( )? [^()]+ (?(<OPEN>) \) )
5714 If the name used in a condition of this kind is a duplicate, the test
5715 is applied to all subpatterns of the same name, and is true if any one
5716 of them has matched.
5718 Checking for pattern recursion
5720 If the condition is the string (R), and there is no subpattern with the
5721 name R, the condition is true if a recursive call to the whole pattern
5722 or any subpattern has been made. If digits or a name preceded by amper-
5723 sand follow the letter R, for example:
5725 (?(R3)...) or (?(R&name)...)
5727 the condition is true if the most recent recursion is into a subpattern
5728 whose number or name is given. This condition does not check the entire
5729 recursion stack. If the name used in a condition of this kind is a
5730 duplicate, the test is applied to all subpatterns of the same name, and
5731 is true if any one of them is the most recent recursion.
5733 At "top level", all these recursion test conditions are false. The
5734 syntax for recursive patterns is described below.
5736 Defining subpatterns for use by reference only
5738 If the condition is the string (DEFINE), and there is no subpattern
5739 with the name DEFINE, the condition is always false. In this case,
5740 there may be only one alternative in the subpattern. It is always
5741 skipped if control reaches this point in the pattern; the idea of
5742 DEFINE is that it can be used to define subroutines that can be refer-
5743 enced from elsewhere. (The use of subroutines is described below.) For
5744 example, a pattern to match an IPv4 address such as ""
5745 could be written like this (ignore whitespace and line breaks):
5747 (?(DEFINE) (?<byte> 2[0-4]\d | 25[0-5] | 1\d\d | [1-9]?\d) )
5748 \b (?&byte) (\.(?&byte)){3} \b
5750 The first part of the pattern is a DEFINE group inside which a another
5751 group named "byte" is defined. This matches an individual component of
5752 an IPv4 address (a number less than 256). When matching takes place,
5753 this part of the pattern is skipped because DEFINE acts like a false
5754 condition. The rest of the pattern uses references to the named group
5755 to match the four dot-separated components of an IPv4 address, insist-
5756 ing on a word boundary at each end.
5758 Assertion conditions
5760 If the condition is not in any of the above formats, it must be an
5761 assertion. This may be a positive or negative lookahead or lookbehind
5762 assertion. Consider this pattern, again containing non-significant
5763 white space, and with the two alternatives on the second line:
5765 (?(?=[^a-z]*[a-z])
5766 \d{2}-[a-z]{3}-\d{2} | \d{2}-\d{2}-\d{2} )
5768 The condition is a positive lookahead assertion that matches an
5769 optional sequence of non-letters followed by a letter. In other words,
5770 it tests for the presence of at least one letter in the subject. If a
5771 letter is found, the subject is matched against the first alternative;
5772 otherwise it is matched against the second. This pattern matches
5773 strings in one of the two forms dd-aaa-dd or dd-dd-dd, where aaa are
5774 letters and dd are digits.
5779 There are two ways of including comments in patterns that are processed
5780 by PCRE. In both cases, the start of the comment must not be in a char-
5781 acter class, nor in the middle of any other sequence of related charac-
5782 ters such as (?: or a subpattern name or number. The characters that
5783 make up a comment play no part in the pattern matching.
5785 The sequence (?# marks the start of a comment that continues up to the
5786 next closing parenthesis. Nested parentheses are not permitted. If the
5787 PCRE_EXTENDED option is set, an unescaped # character also introduces a
5788 comment, which in this case continues to immediately after the next
5789 newline character or character sequence in the pattern. Which charac-
5790 ters are interpreted as newlines is controlled by the options passed to
5791 a compiling function or by a special sequence at the start of the pat-
5792 tern, as described in the section entitled "Newline conventions" above.
5793 Note that the end of this type of comment is a literal newline sequence
5794 in the pattern; escape sequences that happen to represent a newline do
5795 not count. For example, consider this pattern when PCRE_EXTENDED is
5796 set, and the default newline convention is in force:
5798 abc #comment \n still comment
5800 On encountering the # character, pcre_compile() skips along, looking
5801 for a newline in the pattern. The sequence \n is still literal at this
5802 stage, so it does not terminate the comment. Only an actual character
5803 with the code value 0x0a (the default newline) does so.
5808 Consider the problem of matching a string in parentheses, allowing for
5809 unlimited nested parentheses. Without the use of recursion, the best
5810 that can be done is to use a pattern that matches up to some fixed
5811 depth of nesting. It is not possible to handle an arbitrary nesting
5812 depth.
5814 For some time, Perl has provided a facility that allows regular expres-
5815 sions to recurse (amongst other things). It does this by interpolating
5816 Perl code in the expression at run time, and the code can refer to the
5817 expression itself. A Perl pattern using code interpolation to solve the
5818 parentheses problem can be created like this:
5820 $re = qr{\( (?: (?>[^()]+) | (?p{$re}) )* \)}x;
5822 The (?p{...}) item interpolates Perl code at run time, and in this case
5823 refers recursively to the pattern in which it appears.
5825 Obviously, PCRE cannot support the interpolation of Perl code. Instead,
5826 it supports special syntax for recursion of the entire pattern, and
5827 also for individual subpattern recursion. After its introduction in
5828 PCRE and Python, this kind of recursion was subsequently introduced
5829 into Perl at release 5.10.
5831 A special item that consists of (? followed by a number greater than
5832 zero and a closing parenthesis is a recursive subroutine call of the
5833 subpattern of the given number, provided that it occurs inside that
5834 subpattern. (If not, it is a non-recursive subroutine call, which is
5835 described in the next section.) The special item (?R) or (?0) is a
5836 recursive call of the entire regular expression.
5838 This PCRE pattern solves the nested parentheses problem (assume the
5839 PCRE_EXTENDED option is set so that white space is ignored):
5841 \( ( [^()]++ | (?R) )* \)
5843 First it matches an opening parenthesis. Then it matches any number of
5844 substrings which can either be a sequence of non-parentheses, or a
5845 recursive match of the pattern itself (that is, a correctly parenthe-
5846 sized substring). Finally there is a closing parenthesis. Note the use
5847 of a possessive quantifier to avoid backtracking into sequences of non-
5848 parentheses.
5850 If this were part of a larger pattern, you would not want to recurse
5851 the entire pattern, so instead you could use this:
5853 ( \( ( [^()]++ | (?1) )* \) )
5855 We have put the pattern into parentheses, and caused the recursion to
5856 refer to them instead of the whole pattern.
5858 In a larger pattern, keeping track of parenthesis numbers can be
5859 tricky. This is made easier by the use of relative references. Instead
5860 of (?1) in the pattern above you can write (?-2) to refer to the second
5861 most recently opened parentheses preceding the recursion. In other
5862 words, a negative number counts capturing parentheses leftwards from
5863 the point at which it is encountered.
5865 It is also possible to refer to subsequently opened parentheses, by
5866 writing references such as (?+2). However, these cannot be recursive
5867 because the reference is not inside the parentheses that are refer-
5868 enced. They are always non-recursive subroutine calls, as described in
5869 the next section.
5871 An alternative approach is to use named parentheses instead. The Perl
5872 syntax for this is (?&name); PCRE's earlier syntax (?P>name) is also
5873 supported. We could rewrite the above example as follows:
5875 (?<pn> \( ( [^()]++ | (?&pn) )* \) )
5877 If there is more than one subpattern with the same name, the earliest
5878 one is used.
5880 This particular example pattern that we have been looking at contains
5881 nested unlimited repeats, and so the use of a possessive quantifier for
5882 matching strings of non-parentheses is important when applying the pat-
5883 tern to strings that do not match. For example, when this pattern is
5884 applied to
5886 (aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa()
5888 it yields "no match" quickly. However, if a possessive quantifier is
5889 not used, the match runs for a very long time indeed because there are
5890 so many different ways the + and * repeats can carve up the subject,
5891 and all have to be tested before failure can be reported.
5893 At the end of a match, the values of capturing parentheses are those
5894 from the outermost level. If you want to obtain intermediate values, a
5895 callout function can be used (see below and the pcrecallout documenta-
5896 tion). If the pattern above is matched against
5898 (ab(cd)ef)
5900 the value for the inner capturing parentheses (numbered 2) is "ef",
5901 which is the last value taken on at the top level. If a capturing sub-
5902 pattern is not matched at the top level, its final captured value is
5903 unset, even if it was (temporarily) set at a deeper level during the
5904 matching process.
5906 If there are more than 15 capturing parentheses in a pattern, PCRE has
5907 to obtain extra memory to store data during a recursion, which it does
5908 by using pcre_malloc, freeing it via pcre_free afterwards. If no memory
5909 can be obtained, the match fails with the PCRE_ERROR_NOMEMORY error.
5911 Do not confuse the (?R) item with the condition (R), which tests for
5912 recursion. Consider this pattern, which matches text in angle brack-
5913 ets, allowing for arbitrary nesting. Only digits are allowed in nested
5914 brackets (that is, when recursing), whereas any characters are permit-
5915 ted at the outer level.
5917 < (?: (?(R) \d++ | [^<>]*+) | (?R)) * >
5919 In this pattern, (?(R) is the start of a conditional subpattern, with
5920 two different alternatives for the recursive and non-recursive cases.
5921 The (?R) item is the actual recursive call.
5923 Differences in recursion processing between PCRE and Perl
5925 Recursion processing in PCRE differs from Perl in two important ways.
5926 In PCRE (like Python, but unlike Perl), a recursive subpattern call is
5927 always treated as an atomic group. That is, once it has matched some of
5928 the subject string, it is never re-entered, even if it contains untried
5929 alternatives and there is a subsequent matching failure. This can be
5930 illustrated by the following pattern, which purports to match a palin-
5931 dromic string that contains an odd number of characters (for example,
5932 "a", "aba", "abcba", "abcdcba"):
5934 ^(.|(.)(?1)\2)$
5936 The idea is that it either matches a single character, or two identical
5937 characters surrounding a sub-palindrome. In Perl, this pattern works;
5938 in PCRE it does not if the pattern is longer than three characters.
5939 Consider the subject string "abcba":
5941 At the top level, the first character is matched, but as it is not at
5942 the end of the string, the first alternative fails; the second alterna-
5943 tive is taken and the recursion kicks in. The recursive call to subpat-
5944 tern 1 successfully matches the next character ("b"). (Note that the
5945 beginning and end of line tests are not part of the recursion).
5947 Back at the top level, the next character ("c") is compared with what
5948 subpattern 2 matched, which was "a". This fails. Because the recursion
5949 is treated as an atomic group, there are now no backtracking points,
5950 and so the entire match fails. (Perl is able, at this point, to re-
5951 enter the recursion and try the second alternative.) However, if the
5952 pattern is written with the alternatives in the other order, things are
5953 different:
5955 ^((.)(?1)\2|.)$
5957 This time, the recursing alternative is tried first, and continues to
5958 recurse until it runs out of characters, at which point the recursion
5959 fails. But this time we do have another alternative to try at the
5960 higher level. That is the big difference: in the previous case the
5961 remaining alternative is at a deeper recursion level, which PCRE cannot
5962 use.
5964 To change the pattern so that it matches all palindromic strings, not
5965 just those with an odd number of characters, it is tempting to change
5966 the pattern to this:
5968 ^((.)(?1)\2|.?)$
5970 Again, this works in Perl, but not in PCRE, and for the same reason.
5971 When a deeper recursion has matched a single character, it cannot be
5972 entered again in order to match an empty string. The solution is to
5973 separate the two cases, and write out the odd and even cases as alter-
5974 natives at the higher level:
5976 ^(?:((.)(?1)\2|)|((.)(?3)\4|.))
5978 If you want to match typical palindromic phrases, the pattern has to
5979 ignore all non-word characters, which can be done like this:
5981 ^\W*+(?:((.)\W*+(?1)\W*+\2|)|((.)\W*+(?3)\W*+\4|\W*+.\W*+))\W*+$
5983 If run with the PCRE_CASELESS option, this pattern matches phrases such
5984 as "A man, a plan, a canal: Panama!" and it works well in both PCRE and
5985 Perl. Note the use of the possessive quantifier *+ to avoid backtrack-
5986 ing into sequences of non-word characters. Without this, PCRE takes a
5987 great deal longer (ten times or more) to match typical phrases, and
5988 Perl takes so long that you think it has gone into a loop.
5990 WARNING: The palindrome-matching patterns above work only if the sub-
5991 ject string does not start with a palindrome that is shorter than the
5992 entire string. For example, although "abcba" is correctly matched, if
5993 the subject is "ababa", PCRE finds the palindrome "aba" at the start,
5994 then fails at top level because the end of the string does not follow.
5995 Once again, it cannot jump back into the recursion to try other alter-
5996 natives, so the entire match fails.
5998 The second way in which PCRE and Perl differ in their recursion pro-
5999 cessing is in the handling of captured values. In Perl, when a subpat-
6000 tern is called recursively or as a subpattern (see the next section),
6001 it has no access to any values that were captured outside the recur-
6002 sion, whereas in PCR