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
8 -----------------------------------------------------------------------------
11 PCRE(3) Library Functions Manual PCRE(3)
16 PCRE - Perl-compatible regular expressions (original API)
20 This document relates to PCRE releases that use the original API, with
21 library names libpcre, libpcre16, and libpcre32. January 2015 saw the
22 first release of a new API, known as PCRE2, with release numbers start-
23 ing at 10.00 and library names libpcre2-8, libpcre2-16, and
24 libpcre2-32. The old libraries (now called PCRE1) are still being main-
25 tained for bug fixes, but there will be no new development. New
26 projects are advised to use the new PCRE2 libraries.
31 The PCRE library is a set of functions that implement regular expres-
32 sion pattern matching using the same syntax and semantics as Perl, with
33 just a few differences. Some features that appeared in Python and PCRE
34 before they appeared in Perl are also available using the Python syn-
35 tax, there is some support for one or two .NET and Oniguruma syntax
36 items, and there is an option for requesting some minor changes that
37 give better JavaScript compatibility.
39 Starting with release 8.30, it is possible to compile two separate PCRE
40 libraries: the original, which supports 8-bit character strings
41 (including UTF-8 strings), and a second library that supports 16-bit
42 character strings (including UTF-16 strings). The build process allows
43 either one or both to be built. The majority of the work to make this
44 possible was done by Zoltan Herczeg.
46 Starting with release 8.32 it is possible to compile a third separate
47 PCRE library that supports 32-bit character strings (including UTF-32
48 strings). The build process allows any combination of the 8-, 16- and
49 32-bit libraries. The work to make this possible was done by Christian
52 The three libraries contain identical sets of functions, except that
53 the names in the 16-bit library start with pcre16_ instead of pcre_,
54 and the names in the 32-bit library start with pcre32_ instead of
55 pcre_. To avoid over-complication and reduce the documentation mainte-
56 nance load, most of the documentation describes the 8-bit library, with
57 the differences for the 16-bit and 32-bit libraries described sepa-
58 rately in the pcre16 and pcre32 pages. References to functions or
59 structures of the form pcre[16|32]_xxx should be read as meaning
60 "pcre_xxx when using the 8-bit library, pcre16_xxx when using the
61 16-bit library, or pcre32_xxx when using the 32-bit library".
63 The current implementation of PCRE corresponds approximately with Perl
64 5.12, including support for UTF-8/16/32 encoded strings and Unicode
65 general category properties. However, UTF-8/16/32 and Unicode support
66 has to be explicitly enabled; it is not the default. The Unicode tables
67 correspond to Unicode release 6.3.0.
69 In addition to the Perl-compatible matching function, PCRE contains an
70 alternative function that matches the same compiled patterns in a dif-
71 ferent way. In certain circumstances, the alternative function has some
72 advantages. For a discussion of the two matching algorithms, see the
75 PCRE is written in C and released as a C library. A number of people
76 have written wrappers and interfaces of various kinds. In particular,
77 Google Inc. have provided a comprehensive C++ wrapper for the 8-bit
78 library. This is now included as part of the PCRE distribution. The
79 pcrecpp page has details of this interface. Other people's contribu-
80 tions can be found in the Contrib directory at the primary FTP site,
83 ftp://ftp.csx.cam.ac.uk/pub/software/programming/pcre
85 Details of exactly which Perl regular expression features are and are
86 not supported by PCRE are given in separate documents. See the pcrepat-
87 tern and pcrecompat pages. There is a syntax summary in the pcresyntax
90 Some features of PCRE can be included, excluded, or changed when the
91 library is built. The pcre_config() function makes it possible for a
92 client to discover which features are available. The features them-
93 selves are described in the pcrebuild page. Documentation about build-
94 ing PCRE for various operating systems can be found in the README and
95 NON-AUTOTOOLS_BUILD files in the source distribution.
97 The libraries contains a number of undocumented internal functions and
98 data tables that are used by more than one of the exported external
99 functions, but which are not intended for use by external callers.
100 Their names all begin with "_pcre_" or "_pcre16_" or "_pcre32_", which
101 hopefully will not provoke any name clashes. In some environments, it
102 is possible to control which external symbols are exported when a
103 shared library is built, and in these cases the undocumented symbols
107 SECURITY CONSIDERATIONS
109 If you are using PCRE in a non-UTF application that permits users to
110 supply arbitrary patterns for compilation, you should be aware of a
111 feature that allows users to turn on UTF support from within a pattern,
112 provided that PCRE was built with UTF support. For example, an 8-bit
113 pattern that begins with "(*UTF8)" or "(*UTF)" turns on UTF-8 mode,
114 which interprets patterns and subjects as strings of UTF-8 characters
115 instead of individual 8-bit characters. This causes both the pattern
116 and any data against which it is matched to be checked for UTF-8 valid-
117 ity. If the data string is very long, such a check might use suffi-
118 ciently many resources as to cause your application to lose perfor-
121 One way of guarding against this possibility is to use the
122 pcre_fullinfo() function to check the compiled pattern's options for
123 UTF. Alternatively, from release 8.33, you can set the PCRE_NEVER_UTF
124 option at compile time. This causes an compile time error if a pattern
125 contains a UTF-setting sequence.
127 If your application is one that supports UTF, be aware that validity
128 checking can take time. If the same data string is to be matched many
129 times, you can use the PCRE_NO_UTF[8|16|32]_CHECK option for the second
130 and subsequent matches to save redundant checks.
132 Another way that performance can be hit is by running a pattern that
133 has a very large search tree against a string that will never match.
134 Nested unlimited repeats in a pattern are a common example. PCRE pro-
135 vides some protection against this: see the PCRE_EXTRA_MATCH_LIMIT fea-
136 ture in the pcreapi page.
141 The user documentation for PCRE comprises a number of different sec-
142 tions. In the "man" format, each of these is a separate "man page". In
143 the HTML format, each is a separate page, linked from the index page.
144 In the plain text format, the descriptions of the pcregrep and pcretest
145 programs are in files called pcregrep.txt and pcretest.txt, respec-
146 tively. The remaining sections, except for the pcredemo section (which
147 is a program listing), are concatenated in pcre.txt, for ease of
148 searching. The sections are as follows:
151 pcre-config show PCRE installation configuration information
152 pcre16 details of the 16-bit library
153 pcre32 details of the 32-bit library
154 pcreapi details of PCRE's native C API
155 pcrebuild building PCRE
156 pcrecallout details of the callout feature
157 pcrecompat discussion of Perl compatibility
158 pcrecpp details of the C++ wrapper for the 8-bit library
159 pcredemo a demonstration C program that uses PCRE
160 pcregrep description of the pcregrep command (8-bit only)
161 pcrejit discussion of the just-in-time optimization support
162 pcrelimits details of size and other limits
163 pcrematching discussion of the two matching algorithms
164 pcrepartial details of the partial matching facility
165 pcrepattern syntax and semantics of supported
167 pcreperform discussion of performance issues
168 pcreposix the POSIX-compatible C API for the 8-bit library
169 pcreprecompile details of saving and re-using precompiled patterns
170 pcresample discussion of the pcredemo program
171 pcrestack discussion of stack usage
172 pcresyntax quick syntax reference
173 pcretest description of the pcretest testing command
174 pcreunicode discussion of Unicode and UTF-8/16/32 support
176 In the "man" and HTML formats, there is also a short page for each C
177 library function, listing its arguments and results.
183 University Computing Service
184 Cambridge CB2 3QH, England.
186 Putting an actual email address here seems to have been a spam magnet,
187 so I've taken it away. If you want to email me, use my two initials,
188 followed by the two digits 10, at the domain cam.ac.uk.
193 Last updated: 10 February 2015
194 Copyright (c) 1997-2015 University of Cambridge.
195 ------------------------------------------------------------------------------
198 PCRE(3) Library Functions Manual PCRE(3)
203 PCRE - Perl-compatible regular expressions
208 PCRE 16-BIT API BASIC FUNCTIONS
210 pcre16 *pcre16_compile(PCRE_SPTR16 pattern, int options,
211 const char **errptr, int *erroffset,
212 const unsigned char *tableptr);
214 pcre16 *pcre16_compile2(PCRE_SPTR16 pattern, int options,
216 const char **errptr, int *erroffset,
217 const unsigned char *tableptr);
219 pcre16_extra *pcre16_study(const pcre16 *code, int options,
220 const char **errptr);
222 void pcre16_free_study(pcre16_extra *extra);
224 int pcre16_exec(const pcre16 *code, const pcre16_extra *extra,
225 PCRE_SPTR16 subject, int length, int startoffset,
226 int options, int *ovector, int ovecsize);
228 int pcre16_dfa_exec(const pcre16 *code, const pcre16_extra *extra,
229 PCRE_SPTR16 subject, int length, int startoffset,
230 int options, int *ovector, int ovecsize,
231 int *workspace, int wscount);
234 PCRE 16-BIT API STRING EXTRACTION FUNCTIONS
236 int pcre16_copy_named_substring(const pcre16 *code,
237 PCRE_SPTR16 subject, int *ovector,
238 int stringcount, PCRE_SPTR16 stringname,
239 PCRE_UCHAR16 *buffer, int buffersize);
241 int pcre16_copy_substring(PCRE_SPTR16 subject, int *ovector,
242 int stringcount, int stringnumber, PCRE_UCHAR16 *buffer,
245 int pcre16_get_named_substring(const pcre16 *code,
246 PCRE_SPTR16 subject, int *ovector,
247 int stringcount, PCRE_SPTR16 stringname,
248 PCRE_SPTR16 *stringptr);
250 int pcre16_get_stringnumber(const pcre16 *code,
253 int pcre16_get_stringtable_entries(const pcre16 *code,
254 PCRE_SPTR16 name, PCRE_UCHAR16 **first, PCRE_UCHAR16 **last);
256 int pcre16_get_substring(PCRE_SPTR16 subject, int *ovector,
257 int stringcount, int stringnumber,
258 PCRE_SPTR16 *stringptr);
260 int pcre16_get_substring_list(PCRE_SPTR16 subject,
261 int *ovector, int stringcount, PCRE_SPTR16 **listptr);
263 void pcre16_free_substring(PCRE_SPTR16 stringptr);
265 void pcre16_free_substring_list(PCRE_SPTR16 *stringptr);
268 PCRE 16-BIT API AUXILIARY FUNCTIONS
270 pcre16_jit_stack *pcre16_jit_stack_alloc(int startsize, int maxsize);
272 void pcre16_jit_stack_free(pcre16_jit_stack *stack);
274 void pcre16_assign_jit_stack(pcre16_extra *extra,
275 pcre16_jit_callback callback, void *data);
277 const unsigned char *pcre16_maketables(void);
279 int pcre16_fullinfo(const pcre16 *code, const pcre16_extra *extra,
280 int what, void *where);
282 int pcre16_refcount(pcre16 *code, int adjust);
284 int pcre16_config(int what, void *where);
286 const char *pcre16_version(void);
288 int pcre16_pattern_to_host_byte_order(pcre16 *code,
289 pcre16_extra *extra, const unsigned char *tables);
292 PCRE 16-BIT API INDIRECTED FUNCTIONS
294 void *(*pcre16_malloc)(size_t);
296 void (*pcre16_free)(void *);
298 void *(*pcre16_stack_malloc)(size_t);
300 void (*pcre16_stack_free)(void *);
302 int (*pcre16_callout)(pcre16_callout_block *);
305 PCRE 16-BIT API 16-BIT-ONLY FUNCTION
307 int pcre16_utf16_to_host_byte_order(PCRE_UCHAR16 *output,
308 PCRE_SPTR16 input, int length, int *byte_order,
312 THE PCRE 16-BIT LIBRARY
314 Starting with release 8.30, it is possible to compile a PCRE library
315 that supports 16-bit character strings, including UTF-16 strings, as
316 well as or instead of the original 8-bit library. The majority of the
317 work to make this possible was done by Zoltan Herczeg. The two
318 libraries contain identical sets of functions, used in exactly the same
319 way. Only the names of the functions and the data types of their argu-
320 ments and results are different. To avoid over-complication and reduce
321 the documentation maintenance load, most of the PCRE documentation
322 describes the 8-bit library, with only occasional references to the
323 16-bit library. This page describes what is different when you use the
326 WARNING: A single application can be linked with both libraries, but
327 you must take care when processing any particular pattern to use func-
328 tions from just one library. For example, if you want to study a pat-
329 tern that was compiled with pcre16_compile(), you must do so with
330 pcre16_study(), not pcre_study(), and you must free the study data with
336 There is only one header file, pcre.h. It contains prototypes for all
337 the functions in all libraries, as well as definitions of flags, struc-
338 tures, error codes, etc.
343 In Unix-like systems, the 16-bit library is called libpcre16, and can
344 normally be accesss by adding -lpcre16 to the command for linking an
345 application that uses PCRE.
350 In the 8-bit library, strings are passed to PCRE library functions as
351 vectors of bytes with the C type "char *". In the 16-bit library,
352 strings are passed as vectors of unsigned 16-bit quantities. The macro
353 PCRE_UCHAR16 specifies an appropriate data type, and PCRE_SPTR16 is
354 defined as "const PCRE_UCHAR16 *". In very many environments, "short
355 int" is a 16-bit data type. When PCRE is built, it defines PCRE_UCHAR16
356 as "unsigned short int", but checks that it really is a 16-bit data
357 type. If it is not, the build fails with an error message telling the
358 maintainer to modify the definition appropriately.
363 The types of the opaque structures that are used for compiled 16-bit
364 patterns and JIT stacks are pcre16 and pcre16_jit_stack respectively.
365 The type of the user-accessible structure that is returned by
366 pcre16_study() is pcre16_extra, and the type of the structure that is
367 used for passing data to a callout function is pcre16_callout_block.
368 These structures contain the same fields, with the same names, as their
369 8-bit counterparts. The only difference is that pointers to character
370 strings are 16-bit instead of 8-bit types.
375 For every function in the 8-bit library there is a corresponding func-
376 tion in the 16-bit library with a name that starts with pcre16_ instead
377 of pcre_. The prototypes are listed above. In addition, there is one
378 extra function, pcre16_utf16_to_host_byte_order(). This is a utility
379 function that converts a UTF-16 character string to host byte order if
380 necessary. The other 16-bit functions expect the strings they are
381 passed to be in host byte order.
383 The input and output arguments of pcre16_utf16_to_host_byte_order() may
384 point to the same address, that is, conversion in place is supported.
385 The output buffer must be at least as long as the input.
387 The length argument specifies the number of 16-bit data units in the
388 input string; a negative value specifies a zero-terminated string.
390 If byte_order is NULL, it is assumed that the string starts off in host
391 byte order. This may be changed by byte-order marks (BOMs) anywhere in
392 the string (commonly as the first character).
394 If byte_order is not NULL, a non-zero value of the integer to which it
395 points means that the input starts off in host byte order, otherwise
396 the opposite order is assumed. Again, BOMs in the string can change
397 this. The final byte order is passed back at the end of processing.
399 If keep_boms is not zero, byte-order mark characters (0xfeff) are
400 copied into the output string. Otherwise they are discarded.
402 The result of the function is the number of 16-bit units placed into
403 the output buffer, including the zero terminator if the string was
407 SUBJECT STRING OFFSETS
409 The lengths and starting offsets of subject strings must be specified
410 in 16-bit data units, and the offsets within subject strings that are
411 returned by the matching functions are in also 16-bit units rather than
417 The name-to-number translation table that is maintained for named sub-
418 patterns uses 16-bit characters. The pcre16_get_stringtable_entries()
419 function returns the length of each entry in the table as the number of
425 There are two new general option names, PCRE_UTF16 and
426 PCRE_NO_UTF16_CHECK, which correspond to PCRE_UTF8 and
427 PCRE_NO_UTF8_CHECK in the 8-bit library. In fact, these new options
428 define the same bits in the options word. There is a discussion about
429 the validity of UTF-16 strings in the pcreunicode page.
431 For the pcre16_config() function there is an option PCRE_CONFIG_UTF16
432 that returns 1 if UTF-16 support is configured, otherwise 0. If this
433 option is given to pcre_config() or pcre32_config(), or if the
434 PCRE_CONFIG_UTF8 or PCRE_CONFIG_UTF32 option is given to pcre16_con-
435 fig(), the result is the PCRE_ERROR_BADOPTION error.
440 In 16-bit mode, when PCRE_UTF16 is not set, character values are
441 treated in the same way as in 8-bit, non UTF-8 mode, except, of course,
442 that they can range from 0 to 0xffff instead of 0 to 0xff. Character
443 types for characters less than 0xff can therefore be influenced by the
444 locale in the same way as before. Characters greater than 0xff have
445 only one case, and no "type" (such as letter or digit).
447 In UTF-16 mode, the character code is Unicode, in the range 0 to
448 0x10ffff, with the exception of values in the range 0xd800 to 0xdfff
449 because those are "surrogate" values that are used in pairs to encode
450 values greater than 0xffff.
452 A UTF-16 string can indicate its endianness by special code knows as a
453 byte-order mark (BOM). The PCRE functions do not handle this, expecting
454 strings to be in host byte order. A utility function called
455 pcre16_utf16_to_host_byte_order() is provided to help with this (see
461 The errors PCRE_ERROR_BADUTF16_OFFSET and PCRE_ERROR_SHORTUTF16 corre-
462 spond to their 8-bit counterparts. The error PCRE_ERROR_BADMODE is
463 given when a compiled pattern is passed to a function that processes
464 patterns in the other mode, for example, if a pattern compiled with
465 pcre_compile() is passed to pcre16_exec().
467 There are new error codes whose names begin with PCRE_UTF16_ERR for
468 invalid UTF-16 strings, corresponding to the PCRE_UTF8_ERR codes for
469 UTF-8 strings that are described in the section entitled "Reason codes
470 for invalid UTF-8 strings" in the main pcreapi page. The UTF-16 errors
473 PCRE_UTF16_ERR1 Missing low surrogate at end of string
474 PCRE_UTF16_ERR2 Invalid low surrogate follows high surrogate
475 PCRE_UTF16_ERR3 Isolated low surrogate
476 PCRE_UTF16_ERR4 Non-character
481 If there is an error while compiling a pattern, the error text that is
482 passed back by pcre16_compile() or pcre16_compile2() is still an 8-bit
483 character string, zero-terminated.
488 The subject and mark fields in the callout block that is passed to a
489 callout function point to 16-bit vectors.
494 The pcretest program continues to operate with 8-bit input and output
495 files, but it can be used for testing the 16-bit library. If it is run
496 with the command line option -16, patterns and subject strings are con-
497 verted from 8-bit to 16-bit before being passed to PCRE, and the 16-bit
498 library functions are used instead of the 8-bit ones. Returned 16-bit
499 strings are converted to 8-bit for output. If both the 8-bit and the
500 32-bit libraries were not compiled, pcretest defaults to 16-bit and the
501 -16 option is ignored.
503 When PCRE is being built, the RunTest script that is called by "make
504 check" uses the pcretest -C option to discover which of the 8-bit,
505 16-bit and 32-bit libraries has been built, and runs the tests appro-
509 NOT SUPPORTED IN 16-BIT MODE
511 Not all the features of the 8-bit library are available with the 16-bit
512 library. The C++ and POSIX wrapper functions support only the 8-bit
513 library, and the pcregrep program is at present 8-bit only.
519 University Computing Service
520 Cambridge CB2 3QH, England.
525 Last updated: 12 May 2013
526 Copyright (c) 1997-2013 University of Cambridge.
527 ------------------------------------------------------------------------------
530 PCRE(3) Library Functions Manual PCRE(3)
535 PCRE - Perl-compatible regular expressions
540 PCRE 32-BIT API BASIC FUNCTIONS
542 pcre32 *pcre32_compile(PCRE_SPTR32 pattern, int options,
543 const char **errptr, int *erroffset,
544 const unsigned char *tableptr);
546 pcre32 *pcre32_compile2(PCRE_SPTR32 pattern, int options,
548 const unsigned char *tableptr);
550 pcre32_extra *pcre32_study(const pcre32 *code, int options,
551 const char **errptr);
553 void pcre32_free_study(pcre32_extra *extra);
555 int pcre32_exec(const pcre32 *code, const pcre32_extra *extra,
556 PCRE_SPTR32 subject, int length, int startoffset,
557 int options, int *ovector, int ovecsize);
559 int pcre32_dfa_exec(const pcre32 *code, const pcre32_extra *extra,
560 PCRE_SPTR32 subject, int length, int startoffset,
561 int options, int *ovector, int ovecsize,
562 int *workspace, int wscount);
565 PCRE 32-BIT API STRING EXTRACTION FUNCTIONS
567 int pcre32_copy_named_substring(const pcre32 *code,
568 PCRE_SPTR32 subject, int *ovector,
569 int stringcount, PCRE_SPTR32 stringname,
570 PCRE_UCHAR32 *buffer, int buffersize);
572 int pcre32_copy_substring(PCRE_SPTR32 subject, int *ovector,
573 int stringcount, int stringnumber, PCRE_UCHAR32 *buffer,
576 int pcre32_get_named_substring(const pcre32 *code,
577 PCRE_SPTR32 subject, int *ovector,
578 int stringcount, PCRE_SPTR32 stringname,
579 PCRE_SPTR32 *stringptr);
581 int pcre32_get_stringnumber(const pcre32 *code,
584 int pcre32_get_stringtable_entries(const pcre32 *code,
585 PCRE_SPTR32 name, PCRE_UCHAR32 **first, PCRE_UCHAR32 **last);
587 int pcre32_get_substring(PCRE_SPTR32 subject, int *ovector,
588 int stringcount, int stringnumber,
589 PCRE_SPTR32 *stringptr);
591 int pcre32_get_substring_list(PCRE_SPTR32 subject,
592 int *ovector, int stringcount, PCRE_SPTR32 **listptr);
594 void pcre32_free_substring(PCRE_SPTR32 stringptr);
596 void pcre32_free_substring_list(PCRE_SPTR32 *stringptr);
599 PCRE 32-BIT API AUXILIARY FUNCTIONS
601 pcre32_jit_stack *pcre32_jit_stack_alloc(int startsize, int maxsize);
603 void pcre32_jit_stack_free(pcre32_jit_stack *stack);
605 void pcre32_assign_jit_stack(pcre32_extra *extra,
606 pcre32_jit_callback callback, void *data);
608 const unsigned char *pcre32_maketables(void);
610 int pcre32_fullinfo(const pcre32 *code, const pcre32_extra *extra,
611 int what, void *where);
613 int pcre32_refcount(pcre32 *code, int adjust);
615 int pcre32_config(int what, void *where);
617 const char *pcre32_version(void);
619 int pcre32_pattern_to_host_byte_order(pcre32 *code,
620 pcre32_extra *extra, const unsigned char *tables);
623 PCRE 32-BIT API INDIRECTED FUNCTIONS
625 void *(*pcre32_malloc)(size_t);
627 void (*pcre32_free)(void *);
629 void *(*pcre32_stack_malloc)(size_t);
631 void (*pcre32_stack_free)(void *);
633 int (*pcre32_callout)(pcre32_callout_block *);
636 PCRE 32-BIT API 32-BIT-ONLY FUNCTION
638 int pcre32_utf32_to_host_byte_order(PCRE_UCHAR32 *output,
639 PCRE_SPTR32 input, int length, int *byte_order,
643 THE PCRE 32-BIT LIBRARY
645 Starting with release 8.32, it is possible to compile a PCRE library
646 that supports 32-bit character strings, including UTF-32 strings, as
647 well as or instead of the original 8-bit library. This work was done by
648 Christian Persch, based on the work done by Zoltan Herczeg for the
649 16-bit library. All three libraries contain identical sets of func-
650 tions, used in exactly the same way. Only the names of the functions
651 and the data types of their arguments and results are different. To
652 avoid over-complication and reduce the documentation maintenance load,
653 most of the PCRE documentation describes the 8-bit library, with only
654 occasional references to the 16-bit and 32-bit libraries. This page
655 describes what is different when you use the 32-bit library.
657 WARNING: A single application can be linked with all or any of the
658 three libraries, but you must take care when processing any particular
659 pattern to use functions from just one library. For example, if you
660 want to study a pattern that was compiled with pcre32_compile(), you
661 must do so with pcre32_study(), not pcre_study(), and you must free the
662 study data with pcre32_free_study().
667 There is only one header file, pcre.h. It contains prototypes for all
668 the functions in all libraries, as well as definitions of flags, struc-
669 tures, error codes, etc.
674 In Unix-like systems, the 32-bit library is called libpcre32, and can
675 normally be accesss by adding -lpcre32 to the command for linking an
676 application that uses PCRE.
681 In the 8-bit library, strings are passed to PCRE library functions as
682 vectors of bytes with the C type "char *". In the 32-bit library,
683 strings are passed as vectors of unsigned 32-bit quantities. The macro
684 PCRE_UCHAR32 specifies an appropriate data type, and PCRE_SPTR32 is
685 defined as "const PCRE_UCHAR32 *". In very many environments, "unsigned
686 int" is a 32-bit data type. When PCRE is built, it defines PCRE_UCHAR32
687 as "unsigned int", but checks that it really is a 32-bit data type. If
688 it is not, the build fails with an error message telling the maintainer
689 to modify the definition appropriately.
694 The types of the opaque structures that are used for compiled 32-bit
695 patterns and JIT stacks are pcre32 and pcre32_jit_stack respectively.
696 The type of the user-accessible structure that is returned by
697 pcre32_study() is pcre32_extra, and the type of the structure that is
698 used for passing data to a callout function is pcre32_callout_block.
699 These structures contain the same fields, with the same names, as their
700 8-bit counterparts. The only difference is that pointers to character
701 strings are 32-bit instead of 8-bit types.
706 For every function in the 8-bit library there is a corresponding func-
707 tion in the 32-bit library with a name that starts with pcre32_ instead
708 of pcre_. The prototypes are listed above. In addition, there is one
709 extra function, pcre32_utf32_to_host_byte_order(). This is a utility
710 function that converts a UTF-32 character string to host byte order if
711 necessary. The other 32-bit functions expect the strings they are
712 passed to be in host byte order.
714 The input and output arguments of pcre32_utf32_to_host_byte_order() may
715 point to the same address, that is, conversion in place is supported.
716 The output buffer must be at least as long as the input.
718 The length argument specifies the number of 32-bit data units in the
719 input string; a negative value specifies a zero-terminated string.
721 If byte_order is NULL, it is assumed that the string starts off in host
722 byte order. This may be changed by byte-order marks (BOMs) anywhere in
723 the string (commonly as the first character).
725 If byte_order is not NULL, a non-zero value of the integer to which it
726 points means that the input starts off in host byte order, otherwise
727 the opposite order is assumed. Again, BOMs in the string can change
728 this. The final byte order is passed back at the end of processing.
730 If keep_boms is not zero, byte-order mark characters (0xfeff) are
731 copied into the output string. Otherwise they are discarded.
733 The result of the function is the number of 32-bit units placed into
734 the output buffer, including the zero terminator if the string was
738 SUBJECT STRING OFFSETS
740 The lengths and starting offsets of subject strings must be specified
741 in 32-bit data units, and the offsets within subject strings that are
742 returned by the matching functions are in also 32-bit units rather than
748 The name-to-number translation table that is maintained for named sub-
749 patterns uses 32-bit characters. The pcre32_get_stringtable_entries()
750 function returns the length of each entry in the table as the number of
756 There are two new general option names, PCRE_UTF32 and
757 PCRE_NO_UTF32_CHECK, which correspond to PCRE_UTF8 and
758 PCRE_NO_UTF8_CHECK in the 8-bit library. In fact, these new options
759 define the same bits in the options word. There is a discussion about
760 the validity of UTF-32 strings in the pcreunicode page.
762 For the pcre32_config() function there is an option PCRE_CONFIG_UTF32
763 that returns 1 if UTF-32 support is configured, otherwise 0. If this
764 option is given to pcre_config() or pcre16_config(), or if the
765 PCRE_CONFIG_UTF8 or PCRE_CONFIG_UTF16 option is given to pcre32_con-
766 fig(), the result is the PCRE_ERROR_BADOPTION error.
771 In 32-bit mode, when PCRE_UTF32 is not set, character values are
772 treated in the same way as in 8-bit, non UTF-8 mode, except, of course,
773 that they can range from 0 to 0x7fffffff instead of 0 to 0xff. Charac-
774 ter types for characters less than 0xff can therefore be influenced by
775 the locale in the same way as before. Characters greater than 0xff
776 have only one case, and no "type" (such as letter or digit).
778 In UTF-32 mode, the character code is Unicode, in the range 0 to
779 0x10ffff, with the exception of values in the range 0xd800 to 0xdfff
780 because those are "surrogate" values that are ill-formed in UTF-32.
782 A UTF-32 string can indicate its endianness by special code knows as a
783 byte-order mark (BOM). The PCRE functions do not handle this, expecting
784 strings to be in host byte order. A utility function called
785 pcre32_utf32_to_host_byte_order() is provided to help with this (see
791 The error PCRE_ERROR_BADUTF32 corresponds to its 8-bit counterpart.
792 The error PCRE_ERROR_BADMODE is given when a compiled pattern is passed
793 to a function that processes patterns in the other mode, for example,
794 if a pattern compiled with pcre_compile() is passed to pcre32_exec().
796 There are new error codes whose names begin with PCRE_UTF32_ERR for
797 invalid UTF-32 strings, corresponding to the PCRE_UTF8_ERR codes for
798 UTF-8 strings that are described in the section entitled "Reason codes
799 for invalid UTF-8 strings" in the main pcreapi page. The UTF-32 errors
802 PCRE_UTF32_ERR1 Surrogate character (range from 0xd800 to 0xdfff)
803 PCRE_UTF32_ERR2 Non-character
804 PCRE_UTF32_ERR3 Character > 0x10ffff
809 If there is an error while compiling a pattern, the error text that is
810 passed back by pcre32_compile() or pcre32_compile2() is still an 8-bit
811 character string, zero-terminated.
816 The subject and mark fields in the callout block that is passed to a
817 callout function point to 32-bit vectors.
822 The pcretest program continues to operate with 8-bit input and output
823 files, but it can be used for testing the 32-bit library. If it is run
824 with the command line option -32, patterns and subject strings are con-
825 verted from 8-bit to 32-bit before being passed to PCRE, and the 32-bit
826 library functions are used instead of the 8-bit ones. Returned 32-bit
827 strings are converted to 8-bit for output. If both the 8-bit and the
828 16-bit libraries were not compiled, pcretest defaults to 32-bit and the
829 -32 option is ignored.
831 When PCRE is being built, the RunTest script that is called by "make
832 check" uses the pcretest -C option to discover which of the 8-bit,
833 16-bit and 32-bit libraries has been built, and runs the tests appro-
837 NOT SUPPORTED IN 32-BIT MODE
839 Not all the features of the 8-bit library are available with the 32-bit
840 library. The C++ and POSIX wrapper functions support only the 8-bit
841 library, and the pcregrep program is at present 8-bit only.
847 University Computing Service
848 Cambridge CB2 3QH, England.
853 Last updated: 12 May 2013
854 Copyright (c) 1997-2013 University of Cambridge.
855 ------------------------------------------------------------------------------
858 PCREBUILD(3) Library Functions Manual PCREBUILD(3)
863 PCRE - Perl-compatible regular expressions
867 PCRE is distributed with a configure script that can be used to build
868 the library in Unix-like environments using the applications known as
869 Autotools. Also in the distribution are files to support building
870 using CMake instead of configure. The text file README contains general
871 information about building with Autotools (some of which is repeated
872 below), and also has some comments about building on various operating
873 systems. There is a lot more information about building PCRE without
874 using Autotools (including information about using CMake and building
875 "by hand") in the text file called NON-AUTOTOOLS-BUILD. You should
876 consult this file as well as the README file if you are building in a
877 non-Unix-like environment.
880 PCRE BUILD-TIME OPTIONS
882 The rest of this document describes the optional features of PCRE that
883 can be selected when the library is compiled. It assumes use of the
884 configure script, where the optional features are selected or dese-
885 lected by providing options to configure before running the make com-
886 mand. However, the same options can be selected in both Unix-like and
887 non-Unix-like environments using the GUI facility of cmake-gui if you
888 are using CMake instead of configure to build PCRE.
890 If you are not using Autotools or CMake, option selection can be done
891 by editing the config.h file, or by passing parameter settings to the
892 compiler, as described in NON-AUTOTOOLS-BUILD.
894 The complete list of options for configure (which includes the standard
895 ones such as the selection of the installation directory) can be
900 The following sections include descriptions of options whose names
901 begin with --enable or --disable. These settings specify changes to the
902 defaults for the configure command. Because of the way that configure
903 works, --enable and --disable always come in pairs, so the complemen-
904 tary option always exists as well, but as it specifies the default, it
908 BUILDING 8-BIT, 16-BIT AND 32-BIT LIBRARIES
910 By default, a library called libpcre is built, containing functions
911 that take string arguments contained in vectors of bytes, either as
912 single-byte characters, or interpreted as UTF-8 strings. You can also
913 build a separate library, called libpcre16, in which strings are con-
914 tained in vectors of 16-bit data units and interpreted either as sin-
915 gle-unit characters or UTF-16 strings, by adding
919 to the configure command. You can also build yet another separate
920 library, called libpcre32, in which strings are contained in vectors of
921 32-bit data units and interpreted either as single-unit characters or
922 UTF-32 strings, by adding
926 to the configure command. If you do not want the 8-bit library, add
930 as well. At least one of the three libraries must be built. Note that
931 the C++ and POSIX wrappers are for the 8-bit library only, and that
932 pcregrep is an 8-bit program. None of these are built if you select
933 only the 16-bit or 32-bit libraries.
936 BUILDING SHARED AND STATIC LIBRARIES
938 The Autotools PCRE building process uses libtool to build both shared
939 and static libraries by default. You can suppress one of these by
945 to the configure command, as required.
950 By default, if the 8-bit library is being built, the configure script
951 will search for a C++ compiler and C++ header files. If it finds them,
952 it automatically builds the C++ wrapper library (which supports only
953 8-bit strings). You can disable this by adding
957 to the configure command.
960 UTF-8, UTF-16 AND UTF-32 SUPPORT
962 To build PCRE with support for UTF Unicode character strings, add
966 to the configure command. This setting applies to all three libraries,
967 adding support for UTF-8 to the 8-bit library, support for UTF-16 to
968 the 16-bit library, and support for UTF-32 to the to the 32-bit
969 library. There are no separate options for enabling UTF-8, UTF-16 and
970 UTF-32 independently because that would allow ridiculous settings such
971 as requesting UTF-16 support while building only the 8-bit library. It
972 is not possible to build one library with UTF support and another with-
973 out in the same configuration. (For backwards compatibility, --enable-
974 utf8 is a synonym of --enable-utf.)
976 Of itself, this setting does not make PCRE treat strings as UTF-8,
977 UTF-16 or UTF-32. As well as compiling PCRE with this option, you also
978 have have to set the PCRE_UTF8, PCRE_UTF16 or PCRE_UTF32 option (as
979 appropriate) when you call one of the pattern compiling functions.
981 If you set --enable-utf when compiling in an EBCDIC environment, PCRE
982 expects its input to be either ASCII or UTF-8 (depending on the run-
983 time option). It is not possible to support both EBCDIC and UTF-8 codes
984 in the same version of the library. Consequently, --enable-utf and
985 --enable-ebcdic are mutually exclusive.
988 UNICODE CHARACTER PROPERTY SUPPORT
990 UTF support allows the libraries to process character codepoints up to
991 0x10ffff in the strings that they handle. On its own, however, it does
992 not provide any facilities for accessing the properties of such charac-
993 ters. If you want to be able to use the pattern escapes \P, \p, and \X,
994 which refer to Unicode character properties, you must add
996 --enable-unicode-properties
998 to the configure command. This implies UTF support, even if you have
999 not explicitly requested it.
1001 Including Unicode property support adds around 30K of tables to the
1002 PCRE library. Only the general category properties such as Lu and Nd
1003 are supported. Details are given in the pcrepattern documentation.
1006 JUST-IN-TIME COMPILER SUPPORT
1008 Just-in-time compiler support is included in the build by specifying
1012 This support is available only for certain hardware architectures. If
1013 this option is set for an unsupported architecture, a compile time
1014 error occurs. See the pcrejit documentation for a discussion of JIT
1015 usage. When JIT support is enabled, pcregrep automatically makes use of
1018 --disable-pcregrep-jit
1020 to the "configure" command.
1023 CODE VALUE OF NEWLINE
1025 By default, PCRE interprets the linefeed (LF) character as indicating
1026 the end of a line. This is the normal newline character on Unix-like
1027 systems. You can compile PCRE to use carriage return (CR) instead, by
1030 --enable-newline-is-cr
1032 to the configure command. There is also a --enable-newline-is-lf
1033 option, which explicitly specifies linefeed as the newline character.
1035 Alternatively, you can specify that line endings are to be indicated by
1036 the two character sequence CRLF. If you want this, add
1038 --enable-newline-is-crlf
1040 to the configure command. There is a fourth option, specified by
1042 --enable-newline-is-anycrlf
1044 which causes PCRE to recognize any of the three sequences CR, LF, or
1045 CRLF as indicating a line ending. Finally, a fifth option, specified by
1047 --enable-newline-is-any
1049 causes PCRE to recognize any Unicode newline sequence.
1051 Whatever line ending convention is selected when PCRE is built can be
1052 overridden when the library functions are called. At build time it is
1053 conventional to use the standard for your operating system.
1058 By default, the sequence \R in a pattern matches any Unicode newline
1059 sequence, whatever has been selected as the line ending sequence. If
1062 --enable-bsr-anycrlf
1064 the default is changed so that \R matches only CR, LF, or CRLF. What-
1065 ever is selected when PCRE is built can be overridden when the library
1066 functions are called.
1071 When the 8-bit library is called through the POSIX interface (see the
1072 pcreposix documentation), additional working storage is required for
1073 holding the pointers to capturing substrings, because PCRE requires
1074 three integers per substring, whereas the POSIX interface provides only
1075 two. If the number of expected substrings is small, the wrapper func-
1076 tion uses space on the stack, because this is faster than using mal-
1077 loc() for each call. The default threshold above which the stack is no
1078 longer used is 10; it can be changed by adding a setting such as
1080 --with-posix-malloc-threshold=20
1082 to the configure command.
1085 HANDLING VERY LARGE PATTERNS
1087 Within a compiled pattern, offset values are used to point from one
1088 part to another (for example, from an opening parenthesis to an alter-
1089 nation metacharacter). By default, in the 8-bit and 16-bit libraries,
1090 two-byte values are used for these offsets, leading to a maximum size
1091 for a compiled pattern of around 64K. This is sufficient to handle all
1092 but the most gigantic patterns. Nevertheless, some people do want to
1093 process truly enormous patterns, so it is possible to compile PCRE to
1094 use three-byte or four-byte offsets by adding a setting such as
1098 to the configure command. The value given must be 2, 3, or 4. For the
1099 16-bit library, a value of 3 is rounded up to 4. In these libraries,
1100 using longer offsets slows down the operation of PCRE because it has to
1101 load additional data when handling them. For the 32-bit library the
1102 value is always 4 and cannot be overridden; the value of --with-link-
1106 AVOIDING EXCESSIVE STACK USAGE
1108 When matching with the pcre_exec() function, PCRE implements backtrack-
1109 ing by making recursive calls to an internal function called match().
1110 In environments where the size of the stack is limited, this can se-
1111 verely limit PCRE's operation. (The Unix environment does not usually
1112 suffer from this problem, but it may sometimes be necessary to increase
1113 the maximum stack size. There is a discussion in the pcrestack docu-
1114 mentation.) An alternative approach to recursion that uses memory from
1115 the heap to remember data, instead of using recursive function calls,
1116 has been implemented to work round the problem of limited stack size.
1117 If you want to build a version of PCRE that works this way, add
1119 --disable-stack-for-recursion
1121 to the configure command. With this configuration, PCRE will use the
1122 pcre_stack_malloc and pcre_stack_free variables to call memory manage-
1123 ment functions. By default these point to malloc() and free(), but you
1124 can replace the pointers so that your own functions are used instead.
1126 Separate functions are provided rather than using pcre_malloc and
1127 pcre_free because the usage is very predictable: the block sizes
1128 requested are always the same, and the blocks are always freed in
1129 reverse order. A calling program might be able to implement optimized
1130 functions that perform better than malloc() and free(). PCRE runs
1131 noticeably more slowly when built in this way. This option affects only
1132 the pcre_exec() function; it is not relevant for pcre_dfa_exec().
1135 LIMITING PCRE RESOURCE USAGE
1137 Internally, PCRE has a function called match(), which it calls repeat-
1138 edly (sometimes recursively) when matching a pattern with the
1139 pcre_exec() function. By controlling the maximum number of times this
1140 function may be called during a single matching operation, a limit can
1141 be placed on the resources used by a single call to pcre_exec(). The
1142 limit can be changed at run time, as described in the pcreapi documen-
1143 tation. The default is 10 million, but this can be changed by adding a
1146 --with-match-limit=500000
1148 to the configure command. This setting has no effect on the
1149 pcre_dfa_exec() matching function.
1151 In some environments it is desirable to limit the depth of recursive
1152 calls of match() more strictly than the total number of calls, in order
1153 to restrict the maximum amount of stack (or heap, if --disable-stack-
1154 for-recursion is specified) that is used. A second limit controls this;
1155 it defaults to the value that is set for --with-match-limit, which
1156 imposes no additional constraints. However, you can set a lower limit
1157 by adding, for example,
1159 --with-match-limit-recursion=10000
1161 to the configure command. This value can also be overridden at run
1165 CREATING CHARACTER TABLES AT BUILD TIME
1167 PCRE uses fixed tables for processing characters whose code values are
1168 less than 256. By default, PCRE is built with a set of tables that are
1169 distributed in the file pcre_chartables.c.dist. These tables are for
1170 ASCII codes only. If you add
1172 --enable-rebuild-chartables
1174 to the configure command, the distributed tables are no longer used.
1175 Instead, a program called dftables is compiled and run. This outputs
1176 the source for new set of tables, created in the default locale of your
1177 C run-time system. (This method of replacing the tables does not work
1178 if you are cross compiling, because dftables is run on the local host.
1179 If you need to create alternative tables when cross compiling, you will
1180 have to do so "by hand".)
1185 PCRE assumes by default that it will run in an environment where the
1186 character code is ASCII (or Unicode, which is a superset of ASCII).
1187 This is the case for most computer operating systems. PCRE can, how-
1188 ever, be compiled to run in an EBCDIC environment by adding
1192 to the configure command. This setting implies --enable-rebuild-charta-
1193 bles. You should only use it if you know that you are in an EBCDIC
1194 environment (for example, an IBM mainframe operating system). The
1195 --enable-ebcdic option is incompatible with --enable-utf.
1197 The EBCDIC character that corresponds to an ASCII LF is assumed to have
1198 the value 0x15 by default. However, in some EBCDIC environments, 0x25
1199 is used. In such an environment you should use
1201 --enable-ebcdic-nl25
1203 as well as, or instead of, --enable-ebcdic. The EBCDIC character for CR
1204 has the same value as in ASCII, namely, 0x0d. Whichever of 0x15 and
1205 0x25 is not chosen as LF is made to correspond to the Unicode NEL char-
1206 acter (which, in Unicode, is 0x85).
1208 The options that select newline behaviour, such as --enable-newline-is-
1209 cr, and equivalent run-time options, refer to these character values in
1210 an EBCDIC environment.
1213 PCREGREP OPTIONS FOR COMPRESSED FILE SUPPORT
1215 By default, pcregrep reads all files as plain text. You can build it so
1216 that it recognizes files whose names end in .gz or .bz2, and reads them
1217 with libz or libbz2, respectively, by adding one or both of
1219 --enable-pcregrep-libz
1220 --enable-pcregrep-libbz2
1222 to the configure command. These options naturally require that the rel-
1223 evant libraries are installed on your system. Configuration will fail
1227 PCREGREP BUFFER SIZE
1229 pcregrep uses an internal buffer to hold a "window" on the file it is
1230 scanning, in order to be able to output "before" and "after" lines when
1231 it finds a match. The size of the buffer is controlled by a parameter
1232 whose default value is 20K. The buffer itself is three times this size,
1233 but because of the way it is used for holding "before" lines, the long-
1234 est line that is guaranteed to be processable is the parameter size.
1235 You can change the default parameter value by adding, for example,
1237 --with-pcregrep-bufsize=50K
1239 to the configure command. The caller of pcregrep can, however, override
1240 this value by specifying a run-time option.
1243 PCRETEST OPTION FOR LIBREADLINE SUPPORT
1247 --enable-pcretest-libreadline
1249 to the configure command, pcretest is linked with the libreadline
1250 library, and when its input is from a terminal, it reads it using the
1251 readline() function. This provides line-editing and history facilities.
1252 Note that libreadline is GPL-licensed, so if you distribute a binary of
1253 pcretest linked in this way, there may be licensing issues.
1255 Setting this option causes the -lreadline option to be added to the
1256 pcretest build. In many operating environments with a sytem-installed
1257 libreadline this is sufficient. However, in some environments (e.g. if
1258 an unmodified distribution version of readline is in use), some extra
1259 configuration may be necessary. The INSTALL file for libreadline says
1262 "Readline uses the termcap functions, but does not link with the
1263 termcap or curses library itself, allowing applications which link
1264 with readline the to choose an appropriate library."
1266 If your environment has not been set up so that an appropriate library
1267 is automatically included, you may need to add something like
1271 immediately before the configure command.
1274 DEBUGGING WITH VALGRIND SUPPORT
1280 option to to the configure command, PCRE will use valgrind annotations
1281 to mark certain memory regions as unaddressable. This allows it to
1282 detect invalid memory accesses, and is mostly useful for debugging PCRE
1286 CODE COVERAGE REPORTING
1288 If your C compiler is gcc, you can build a version of PCRE that can
1289 generate a code coverage report for its test suite. To enable this, you
1290 must install lcov version 1.6 or above. Then specify
1294 to the configure command and build PCRE in the usual way.
1296 Note that using ccache (a caching C compiler) is incompatible with code
1297 coverage reporting. If you have configured ccache to run automatically
1298 on your system, you must set the environment variable
1302 before running make to build PCRE, so that ccache is not used.
1304 When --enable-coverage is used, the following addition targets are
1305 added to the Makefile:
1309 This creates a fresh coverage report for the PCRE test suite. It is
1310 equivalent to running "make coverage-reset", "make coverage-baseline",
1311 "make check", and then "make coverage-report".
1315 This zeroes the coverage counters, but does nothing else.
1317 make coverage-baseline
1319 This captures baseline coverage information.
1321 make coverage-report
1323 This creates the coverage report.
1325 make coverage-clean-report
1327 This removes the generated coverage report without cleaning the cover-
1330 make coverage-clean-data
1332 This removes the captured coverage data without removing the coverage
1333 files created at compile time (*.gcno).
1337 This cleans all coverage data including the generated coverage report.
1338 For more information about code coverage, see the gcov and lcov docu-
1344 pcreapi(3), pcre16, pcre32, pcre_config(3).
1350 University Computing Service
1351 Cambridge CB2 3QH, England.
1356 Last updated: 12 May 2013
1357 Copyright (c) 1997-2013 University of Cambridge.
1358 ------------------------------------------------------------------------------
1361 PCREMATCHING(3) Library Functions Manual PCREMATCHING(3)
1366 PCRE - Perl-compatible regular expressions
1368 PCRE MATCHING ALGORITHMS
1370 This document describes the two different algorithms that are available
1371 in PCRE for matching a compiled regular expression against a given sub-
1372 ject string. The "standard" algorithm is the one provided by the
1373 pcre_exec(), pcre16_exec() and pcre32_exec() functions. These work in
1374 the same as as Perl's matching function, and provide a Perl-compatible
1375 matching operation. The just-in-time (JIT) optimization that is
1376 described in the pcrejit documentation is compatible with these func-
1379 An alternative algorithm is provided by the pcre_dfa_exec(),
1380 pcre16_dfa_exec() and pcre32_dfa_exec() functions; they operate in a
1381 different way, and are not Perl-compatible. This alternative has advan-
1382 tages and disadvantages compared with the standard algorithm, and these
1383 are described below.
1385 When there is only one possible way in which a given subject string can
1386 match a pattern, the two algorithms give the same answer. A difference
1387 arises, however, when there are multiple possibilities. For example, if
1392 is matched against the string
1394 <something> <something else> <something further>
1396 there are three possible answers. The standard algorithm finds only one
1397 of them, whereas the alternative algorithm finds all three.
1400 REGULAR EXPRESSIONS AS TREES
1402 The set of strings that are matched by a regular expression can be rep-
1403 resented as a tree structure. An unlimited repetition in the pattern
1404 makes the tree of infinite size, but it is still a tree. Matching the
1405 pattern to a given subject string (from a given starting point) can be
1406 thought of as a search of the tree. There are two ways to search a
1407 tree: depth-first and breadth-first, and these correspond to the two
1408 matching algorithms provided by PCRE.
1411 THE STANDARD MATCHING ALGORITHM
1413 In the terminology of Jeffrey Friedl's book "Mastering Regular Expres-
1414 sions", the standard algorithm is an "NFA algorithm". It conducts a
1415 depth-first search of the pattern tree. That is, it proceeds along a
1416 single path through the tree, checking that the subject matches what is
1417 required. When there is a mismatch, the algorithm tries any alterna-
1418 tives at the current point, and if they all fail, it backs up to the
1419 previous branch point in the tree, and tries the next alternative
1420 branch at that level. This often involves backing up (moving to the
1421 left) in the subject string as well. The order in which repetition
1422 branches are tried is controlled by the greedy or ungreedy nature of
1425 If a leaf node is reached, a matching string has been found, and at
1426 that point the algorithm stops. Thus, if there is more than one possi-
1427 ble match, this algorithm returns the first one that it finds. Whether
1428 this is the shortest, the longest, or some intermediate length depends
1429 on the way the greedy and ungreedy repetition quantifiers are specified
1432 Because it ends up with a single path through the tree, it is rela-
1433 tively straightforward for this algorithm to keep track of the sub-
1434 strings that are matched by portions of the pattern in parentheses.
1435 This provides support for capturing parentheses and back references.
1438 THE ALTERNATIVE MATCHING ALGORITHM
1440 This algorithm conducts a breadth-first search of the tree. Starting
1441 from the first matching point in the subject, it scans the subject
1442 string from left to right, once, character by character, and as it does
1443 this, it remembers all the paths through the tree that represent valid
1444 matches. In Friedl's terminology, this is a kind of "DFA algorithm",
1445 though it is not implemented as a traditional finite state machine (it
1446 keeps multiple states active simultaneously).
1448 Although the general principle of this matching algorithm is that it
1449 scans the subject string only once, without backtracking, there is one
1450 exception: when a lookaround assertion is encountered, the characters
1451 following or preceding the current point have to be independently
1454 The scan continues until either the end of the subject is reached, or
1455 there are no more unterminated paths. At this point, terminated paths
1456 represent the different matching possibilities (if there are none, the
1457 match has failed). Thus, if there is more than one possible match,
1458 this algorithm finds all of them, and in particular, it finds the long-
1459 est. The matches are returned in decreasing order of length. There is
1460 an option to stop the algorithm after the first match (which is neces-
1461 sarily the shortest) is found.
1463 Note that all the matches that are found start at the same point in the
1464 subject. If the pattern
1468 is matched against the string "the caterpillar catchment", the result
1469 will be the three strings "caterpillar", "cater", and "cat" that start
1470 at the fifth character of the subject. The algorithm does not automati-
1471 cally move on to find matches that start at later positions.
1473 PCRE's "auto-possessification" optimization usually applies to charac-
1474 ter repeats at the end of a pattern (as well as internally). For exam-
1475 ple, the pattern "a\d+" is compiled as if it were "a\d++" because there
1476 is no point even considering the possibility of backtracking into the
1477 repeated digits. For DFA matching, this means that only one possible
1478 match is found. If you really do want multiple matches in such cases,
1479 either use an ungreedy repeat ("a\d+?") or set the PCRE_NO_AUTO_POSSESS
1480 option when compiling.
1482 There are a number of features of PCRE regular expressions that are not
1483 supported by the alternative matching algorithm. They are as follows:
1485 1. Because the algorithm finds all possible matches, the greedy or
1486 ungreedy nature of repetition quantifiers is not relevant. Greedy and
1487 ungreedy quantifiers are treated in exactly the same way. However, pos-
1488 sessive quantifiers can make a difference when what follows could also
1489 match what is quantified, for example in a pattern like this:
1493 This pattern matches "aaab!" but not "aaa!", which would be matched by
1494 a non-possessive quantifier. Similarly, if an atomic group is present,
1495 it is matched as if it were a standalone pattern at the current point,
1496 and the longest match is then "locked in" for the rest of the overall
1499 2. When dealing with multiple paths through the tree simultaneously, it
1500 is not straightforward to keep track of captured substrings for the
1501 different matching possibilities, and PCRE's implementation of this
1502 algorithm does not attempt to do this. This means that no captured sub-
1503 strings are available.
1505 3. Because no substrings are captured, back references within the pat-
1506 tern are not supported, and cause errors if encountered.
1508 4. For the same reason, conditional expressions that use a backrefer-
1509 ence as the condition or test for a specific group recursion are not
1512 5. Because many paths through the tree may be active, the \K escape
1513 sequence, which resets the start of the match when encountered (but may
1514 be on some paths and not on others), is not supported. It causes an
1515 error if encountered.
1517 6. Callouts are supported, but the value of the capture_top field is
1518 always 1, and the value of the capture_last field is always -1.
1520 7. The \C escape sequence, which (in the standard algorithm) always
1521 matches a single data unit, even in UTF-8, UTF-16 or UTF-32 modes, is
1522 not supported in these modes, because the alternative algorithm moves
1523 through the subject string one character (not data unit) at a time, for
1524 all active paths through the tree.
1526 8. Except for (*FAIL), the backtracking control verbs such as (*PRUNE)
1527 are not supported. (*FAIL) is supported, and behaves like a failing
1531 ADVANTAGES OF THE ALTERNATIVE ALGORITHM
1533 Using the alternative matching algorithm provides the following advan-
1536 1. All possible matches (at a single point in the subject) are automat-
1537 ically found, and in particular, the longest match is found. To find
1538 more than one match using the standard algorithm, you have to do kludgy
1539 things with callouts.
1541 2. Because the alternative algorithm scans the subject string just
1542 once, and never needs to backtrack (except for lookbehinds), it is pos-
1543 sible to pass very long subject strings to the matching function in
1544 several pieces, checking for partial matching each time. Although it is
1545 possible to do multi-segment matching using the standard algorithm by
1546 retaining partially matched substrings, it is more complicated. The
1547 pcrepartial documentation gives details of partial matching and dis-
1548 cusses multi-segment matching.
1551 DISADVANTAGES OF THE ALTERNATIVE ALGORITHM
1553 The alternative algorithm suffers from a number of disadvantages:
1555 1. It is substantially slower than the standard algorithm. This is
1556 partly because it has to search for all possible matches, but is also
1557 because it is less susceptible to optimization.
1559 2. Capturing parentheses and back references are not supported.
1561 3. Although atomic groups are supported, their use does not provide the
1562 performance advantage that it does for the standard algorithm.
1568 University Computing Service
1569 Cambridge CB2 3QH, England.
1574 Last updated: 12 November 2013
1575 Copyright (c) 1997-2012 University of Cambridge.
1576 ------------------------------------------------------------------------------
1579 PCREAPI(3) Library Functions Manual PCREAPI(3)
1584 PCRE - Perl-compatible regular expressions
1589 PCRE NATIVE API BASIC FUNCTIONS
1591 pcre *pcre_compile(const char *pattern, int options,
1592 const char **errptr, int *erroffset,
1593 const unsigned char *tableptr);
1595 pcre *pcre_compile2(const char *pattern, int options,
1597 const char **errptr, int *erroffset,
1598 const unsigned char *tableptr);
1600 pcre_extra *pcre_study(const pcre *code, int options,
1601 const char **errptr);
1603 void pcre_free_study(pcre_extra *extra);
1605 int pcre_exec(const pcre *code, const pcre_extra *extra,
1606 const char *subject, int length, int startoffset,
1607 int options, int *ovector, int ovecsize);
1609 int pcre_dfa_exec(const pcre *code, const pcre_extra *extra,
1610 const char *subject, int length, int startoffset,
1611 int options, int *ovector, int ovecsize,
1612 int *workspace, int wscount);
1615 PCRE NATIVE API STRING EXTRACTION FUNCTIONS
1617 int pcre_copy_named_substring(const pcre *code,
1618 const char *subject, int *ovector,
1619 int stringcount, const char *stringname,
1620 char *buffer, int buffersize);
1622 int pcre_copy_substring(const char *subject, int *ovector,
1623 int stringcount, int stringnumber, char *buffer,
1626 int pcre_get_named_substring(const pcre *code,
1627 const char *subject, int *ovector,
1628 int stringcount, const char *stringname,
1629 const char **stringptr);
1631 int pcre_get_stringnumber(const pcre *code,
1634 int pcre_get_stringtable_entries(const pcre *code,
1635 const char *name, char **first, char **last);
1637 int pcre_get_substring(const char *subject, int *ovector,
1638 int stringcount, int stringnumber,
1639 const char **stringptr);
1641 int pcre_get_substring_list(const char *subject,
1642 int *ovector, int stringcount, const char ***listptr);
1644 void pcre_free_substring(const char *stringptr);
1646 void pcre_free_substring_list(const char **stringptr);
1649 PCRE NATIVE API AUXILIARY FUNCTIONS
1651 int pcre_jit_exec(const pcre *code, const pcre_extra *extra,
1652 const char *subject, int length, int startoffset,
1653 int options, int *ovector, int ovecsize,
1654 pcre_jit_stack *jstack);
1656 pcre_jit_stack *pcre_jit_stack_alloc(int startsize, int maxsize);
1658 void pcre_jit_stack_free(pcre_jit_stack *stack);
1660 void pcre_assign_jit_stack(pcre_extra *extra,
1661 pcre_jit_callback callback, void *data);
1663 const unsigned char *pcre_maketables(void);
1665 int pcre_fullinfo(const pcre *code, const pcre_extra *extra,
1666 int what, void *where);
1668 int pcre_refcount(pcre *code, int adjust);
1670 int pcre_config(int what, void *where);
1672 const char *pcre_version(void);
1674 int pcre_pattern_to_host_byte_order(pcre *code,
1675 pcre_extra *extra, const unsigned char *tables);
1678 PCRE NATIVE API INDIRECTED FUNCTIONS
1680 void *(*pcre_malloc)(size_t);
1682 void (*pcre_free)(void *);
1684 void *(*pcre_stack_malloc)(size_t);
1686 void (*pcre_stack_free)(void *);
1688 int (*pcre_callout)(pcre_callout_block *);
1690 int (*pcre_stack_guard)(void);
1693 PCRE 8-BIT, 16-BIT, AND 32-BIT LIBRARIES
1695 As well as support for 8-bit character strings, PCRE also supports
1696 16-bit strings (from release 8.30) and 32-bit strings (from release
1697 8.32), by means of two additional libraries. They can be built as well
1698 as, or instead of, the 8-bit library. To avoid too much complication,
1699 this document describes the 8-bit versions of the functions, with only
1700 occasional references to the 16-bit and 32-bit libraries.
1702 The 16-bit and 32-bit functions operate in the same way as their 8-bit
1703 counterparts; they just use different data types for their arguments
1704 and results, and their names start with pcre16_ or pcre32_ instead of
1705 pcre_. For every option that has UTF8 in its name (for example,
1706 PCRE_UTF8), there are corresponding 16-bit and 32-bit names with UTF8
1707 replaced by UTF16 or UTF32, respectively. This facility is in fact just
1708 cosmetic; the 16-bit and 32-bit option names define the same bit val-
1711 References to bytes and UTF-8 in this document should be read as refer-
1712 ences to 16-bit data units and UTF-16 when using the 16-bit library, or
1713 32-bit data units and UTF-32 when using the 32-bit library, unless
1714 specified otherwise. More details of the specific differences for the
1715 16-bit and 32-bit libraries are given in the pcre16 and pcre32 pages.
1720 PCRE has its own native API, which is described in this document. There
1721 are also some wrapper functions (for the 8-bit library only) that cor-
1722 respond to the POSIX regular expression API, but they do not give
1723 access to all the functionality. They are described in the pcreposix
1724 documentation. Both of these APIs define a set of C function calls. A
1725 C++ wrapper (again for the 8-bit library only) is also distributed with
1726 PCRE. It is documented in the pcrecpp page.
1728 The native API C function prototypes are defined in the header file
1729 pcre.h, and on Unix-like systems the (8-bit) library itself is called
1730 libpcre. It can normally be accessed by adding -lpcre to the command
1731 for linking an application that uses PCRE. The header file defines the
1732 macros PCRE_MAJOR and PCRE_MINOR to contain the major and minor release
1733 numbers for the library. Applications can use these to include support
1734 for different releases of PCRE.
1736 In a Windows environment, if you want to statically link an application
1737 program against a non-dll pcre.a file, you must define PCRE_STATIC
1738 before including pcre.h or pcrecpp.h, because otherwise the pcre_mal-
1739 loc() and pcre_free() exported functions will be declared
1740 __declspec(dllimport), with unwanted results.
1742 The functions pcre_compile(), pcre_compile2(), pcre_study(), and
1743 pcre_exec() are used for compiling and matching regular expressions in
1744 a Perl-compatible manner. A sample program that demonstrates the sim-
1745 plest way of using them is provided in the file called pcredemo.c in
1746 the PCRE source distribution. A listing of this program is given in the
1747 pcredemo documentation, and the pcresample documentation describes how
1748 to compile and run it.
1750 Just-in-time compiler support is an optional feature of PCRE that can
1751 be built in appropriate hardware environments. It greatly speeds up the
1752 matching performance of many patterns. Simple programs can easily
1753 request that it be used if available, by setting an option that is
1754 ignored when it is not relevant. More complicated programs might need
1755 to make use of the functions pcre_jit_stack_alloc(),
1756 pcre_jit_stack_free(), and pcre_assign_jit_stack() in order to control
1757 the JIT code's memory usage.
1759 From release 8.32 there is also a direct interface for JIT execution,
1760 which gives improved performance. The JIT-specific functions are dis-
1761 cussed in the pcrejit documentation.
1763 A second matching function, pcre_dfa_exec(), which is not Perl-compati-
1764 ble, is also provided. This uses a different algorithm for the match-
1765 ing. The alternative algorithm finds all possible matches (at a given
1766 point in the subject), and scans the subject just once (unless there
1767 are lookbehind assertions). However, this algorithm does not return
1768 captured substrings. A description of the two matching algorithms and
1769 their advantages and disadvantages is given in the pcrematching docu-
1772 In addition to the main compiling and matching functions, there are
1773 convenience functions for extracting captured substrings from a subject
1774 string that is matched by pcre_exec(). They are:
1776 pcre_copy_substring()
1777 pcre_copy_named_substring()
1778 pcre_get_substring()
1779 pcre_get_named_substring()
1780 pcre_get_substring_list()
1781 pcre_get_stringnumber()
1782 pcre_get_stringtable_entries()
1784 pcre_free_substring() and pcre_free_substring_list() are also provided,
1785 to free the memory used for extracted strings.
1787 The function pcre_maketables() is used to build a set of character
1788 tables in the current locale for passing to pcre_compile(),
1789 pcre_exec(), or pcre_dfa_exec(). This is an optional facility that is
1790 provided for specialist use. Most commonly, no special tables are
1791 passed, in which case internal tables that are generated when PCRE is
1794 The function pcre_fullinfo() is used to find out information about a
1795 compiled pattern. The function pcre_version() returns a pointer to a
1796 string containing the version of PCRE and its date of release.
1798 The function pcre_refcount() maintains a reference count in a data
1799 block containing a compiled pattern. This is provided for the benefit
1800 of object-oriented applications.
1802 The global variables pcre_malloc and pcre_free initially contain the
1803 entry points of the standard malloc() and free() functions, respec-
1804 tively. PCRE calls the memory management functions via these variables,
1805 so a calling program can replace them if it wishes to intercept the
1806 calls. This should be done before calling any PCRE functions.
1808 The global variables pcre_stack_malloc and pcre_stack_free are also
1809 indirections to memory management functions. These special functions
1810 are used only when PCRE is compiled to use the heap for remembering
1811 data, instead of recursive function calls, when running the pcre_exec()
1812 function. See the pcrebuild documentation for details of how to do
1813 this. It is a non-standard way of building PCRE, for use in environ-
1814 ments that have limited stacks. Because of the greater use of memory
1815 management, it runs more slowly. Separate functions are provided so
1816 that special-purpose external code can be used for this case. When
1817 used, these functions are always called in a stack-like manner (last
1818 obtained, first freed), and always for memory blocks of the same size.
1819 There is a discussion about PCRE's stack usage in the pcrestack docu-
1822 The global variable pcre_callout initially contains NULL. It can be set
1823 by the caller to a "callout" function, which PCRE will then call at
1824 specified points during a matching operation. Details are given in the
1825 pcrecallout documentation.
1827 The global variable pcre_stack_guard initially contains NULL. It can be
1828 set by the caller to a function that is called by PCRE whenever it
1829 starts to compile a parenthesized part of a pattern. When parentheses
1830 are nested, PCRE uses recursive function calls, which use up the system
1831 stack. This function is provided so that applications with restricted
1832 stacks can force a compilation error if the stack runs out. The func-
1833 tion should return zero if all is well, or non-zero to force an error.
1838 PCRE supports five different conventions for indicating line breaks in
1839 strings: a single CR (carriage return) character, a single LF (line-
1840 feed) character, the two-character sequence CRLF, any of the three pre-
1841 ceding, or any Unicode newline sequence. The Unicode newline sequences
1842 are the three just mentioned, plus the single characters VT (vertical
1843 tab, U+000B), FF (form feed, U+000C), NEL (next line, U+0085), LS (line
1844 separator, U+2028), and PS (paragraph separator, U+2029).
1846 Each of the first three conventions is used by at least one operating
1847 system as its standard newline sequence. When PCRE is built, a default
1848 can be specified. The default default is LF, which is the Unix stan-
1849 dard. When PCRE is run, the default can be overridden, either when a
1850 pattern is compiled, or when it is matched.
1852 At compile time, the newline convention can be specified by the options
1853 argument of pcre_compile(), or it can be specified by special text at
1854 the start of the pattern itself; this overrides any other settings. See
1855 the pcrepattern page for details of the special character sequences.
1857 In the PCRE documentation the word "newline" is used to mean "the char-
1858 acter or pair of characters that indicate a line break". The choice of
1859 newline convention affects the handling of the dot, circumflex, and
1860 dollar metacharacters, the handling of #-comments in /x mode, and, when
1861 CRLF is a recognized line ending sequence, the match position advance-
1862 ment for a non-anchored pattern. There is more detail about this in the
1863 section on pcre_exec() options below.
1865 The choice of newline convention does not affect the interpretation of
1866 the \n or \r escape sequences, nor does it affect what \R matches,
1867 which is controlled in a similar way, but by separate options.
1872 The PCRE functions can be used in multi-threading applications, with
1873 the proviso that the memory management functions pointed to by
1874 pcre_malloc, pcre_free, pcre_stack_malloc, and pcre_stack_free, and the
1875 callout and stack-checking functions pointed to by pcre_callout and
1876 pcre_stack_guard, are shared by all threads.
1878 The compiled form of a regular expression is not altered during match-
1879 ing, so the same compiled pattern can safely be used by several threads
1882 If the just-in-time optimization feature is being used, it needs sepa-
1883 rate memory stack areas for each thread. See the pcrejit documentation
1887 SAVING PRECOMPILED PATTERNS FOR LATER USE
1889 The compiled form of a regular expression can be saved and re-used at a
1890 later time, possibly by a different program, and even on a host other
1891 than the one on which it was compiled. Details are given in the
1892 pcreprecompile documentation, which includes a description of the
1893 pcre_pattern_to_host_byte_order() function. However, compiling a regu-
1894 lar expression with one version of PCRE for use with a different ver-
1895 sion is not guaranteed to work and may cause crashes.
1898 CHECKING BUILD-TIME OPTIONS
1900 int pcre_config(int what, void *where);
1902 The function pcre_config() makes it possible for a PCRE client to dis-
1903 cover which optional features have been compiled into the PCRE library.
1904 The pcrebuild documentation has more details about these optional fea-
1907 The first argument for pcre_config() is an integer, specifying which
1908 information is required; the second argument is a pointer to a variable
1909 into which the information is placed. The returned value is zero on
1910 success, or the negative error code PCRE_ERROR_BADOPTION if the value
1911 in the first argument is not recognized. The following information is
1916 The output is an integer that is set to one if UTF-8 support is avail-
1917 able; otherwise it is set to zero. This value should normally be given
1918 to the 8-bit version of this function, pcre_config(). If it is given to
1919 the 16-bit or 32-bit version of this function, the result is
1920 PCRE_ERROR_BADOPTION.
1924 The output is an integer that is set to one if UTF-16 support is avail-
1925 able; otherwise it is set to zero. This value should normally be given
1926 to the 16-bit version of this function, pcre16_config(). If it is given
1927 to the 8-bit or 32-bit version of this function, the result is
1928 PCRE_ERROR_BADOPTION.
1932 The output is an integer that is set to one if UTF-32 support is avail-
1933 able; otherwise it is set to zero. This value should normally be given
1934 to the 32-bit version of this function, pcre32_config(). If it is given
1935 to the 8-bit or 16-bit version of this function, the result is
1936 PCRE_ERROR_BADOPTION.
1938 PCRE_CONFIG_UNICODE_PROPERTIES
1940 The output is an integer that is set to one if support for Unicode
1941 character properties is available; otherwise it is set to zero.
1945 The output is an integer that is set to one if support for just-in-time
1946 compiling is available; otherwise it is set to zero.
1948 PCRE_CONFIG_JITTARGET
1950 The output is a pointer to a zero-terminated "const char *" string. If
1951 JIT support is available, the string contains the name of the architec-
1952 ture for which the JIT compiler is configured, for example "x86 32bit
1953 (little endian + unaligned)". If JIT support is not available, the
1958 The output is an integer whose value specifies the default character
1959 sequence that is recognized as meaning "newline". The values that are
1960 supported in ASCII/Unicode environments are: 10 for LF, 13 for CR, 3338
1961 for CRLF, -2 for ANYCRLF, and -1 for ANY. In EBCDIC environments, CR,
1962 ANYCRLF, and ANY yield the same values. However, the value for LF is
1963 normally 21, though some EBCDIC environments use 37. The corresponding
1964 values for CRLF are 3349 and 3365. The default should normally corre-
1965 spond to the standard sequence for your operating system.
1969 The output is an integer whose value indicates what character sequences
1970 the \R escape sequence matches by default. A value of 0 means that \R
1971 matches any Unicode line ending sequence; a value of 1 means that \R
1972 matches only CR, LF, or CRLF. The default can be overridden when a pat-
1973 tern is compiled or matched.
1975 PCRE_CONFIG_LINK_SIZE
1977 The output is an integer that contains the number of bytes used for
1978 internal linkage in compiled regular expressions. For the 8-bit
1979 library, the value can be 2, 3, or 4. For the 16-bit library, the value
1980 is either 2 or 4 and is still a number of bytes. For the 32-bit
1981 library, the value is either 2 or 4 and is still a number of bytes. The
1982 default value of 2 is sufficient for all but the most massive patterns,
1983 since it allows the compiled pattern to be up to 64K in size. Larger
1984 values allow larger regular expressions to be compiled, at the expense
1987 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
1989 The output is an integer that contains the threshold above which the
1990 POSIX interface uses malloc() for output vectors. Further details are
1991 given in the pcreposix documentation.
1993 PCRE_CONFIG_PARENS_LIMIT
1995 The output is a long integer that gives the maximum depth of nesting of
1996 parentheses (of any kind) in a pattern. This limit is imposed to cap
1997 the amount of system stack used when a pattern is compiled. It is spec-
1998 ified when PCRE is built; the default is 250. This limit does not take
1999 into account the stack that may already be used by the calling applica-
2000 tion. For finer control over compilation stack usage, you can set a
2001 pointer to an external checking function in pcre_stack_guard.
2003 PCRE_CONFIG_MATCH_LIMIT
2005 The output is a long integer that gives the default limit for the num-
2006 ber of internal matching function calls in a pcre_exec() execution.
2007 Further details are given with pcre_exec() below.
2009 PCRE_CONFIG_MATCH_LIMIT_RECURSION
2011 The output is a long integer that gives the default limit for the depth
2012 of recursion when calling the internal matching function in a
2013 pcre_exec() execution. Further details are given with pcre_exec()
2016 PCRE_CONFIG_STACKRECURSE
2018 The output is an integer that is set to one if internal recursion when
2019 running pcre_exec() is implemented by recursive function calls that use
2020 the stack to remember their state. This is the usual way that PCRE is
2021 compiled. The output is zero if PCRE was compiled to use blocks of data
2022 on the heap instead of recursive function calls. In this case,
2023 pcre_stack_malloc and pcre_stack_free are called to manage memory
2024 blocks on the heap, thus avoiding the use of the stack.
2029 pcre *pcre_compile(const char *pattern, int options,
2030 const char **errptr, int *erroffset,
2031 const unsigned char *tableptr);
2033 pcre *pcre_compile2(const char *pattern, int options,
2035 const char **errptr, int *erroffset,
2036 const unsigned char *tableptr);
2038 Either of the functions pcre_compile() or pcre_compile2() can be called
2039 to compile a pattern into an internal form. The only difference between
2040 the two interfaces is that pcre_compile2() has an additional argument,
2041 errorcodeptr, via which a numerical error code can be returned. To
2042 avoid too much repetition, we refer just to pcre_compile() below, but
2043 the information applies equally to pcre_compile2().
2045 The pattern is a C string terminated by a binary zero, and is passed in
2046 the pattern argument. A pointer to a single block of memory that is
2047 obtained via pcre_malloc is returned. This contains the compiled code
2048 and related data. The pcre type is defined for the returned block; this
2049 is a typedef for a structure whose contents are not externally defined.
2050 It is up to the caller to free the memory (via pcre_free) when it is no
2053 Although the compiled code of a PCRE regex is relocatable, that is, it
2054 does not depend on memory location, the complete pcre data block is not
2055 fully relocatable, because it may contain a copy of the tableptr argu-
2056 ment, which is an address (see below).
2058 The options argument contains various bit settings that affect the com-
2059 pilation. It should be zero if no options are required. The available
2060 options are described below. Some of them (in particular, those that
2061 are compatible with Perl, but some others as well) can also be set and
2062 unset from within the pattern (see the detailed description in the
2063 pcrepattern documentation). For those options that can be different in
2064 different parts of the pattern, the contents of the options argument
2065 specifies their settings at the start of compilation and execution. The
2066 PCRE_ANCHORED, PCRE_BSR_xxx, PCRE_NEWLINE_xxx, PCRE_NO_UTF8_CHECK, and
2067 PCRE_NO_START_OPTIMIZE options can be set at the time of matching as
2068 well as at compile time.
2070 If errptr is NULL, pcre_compile() returns NULL immediately. Otherwise,
2071 if compilation of a pattern fails, pcre_compile() returns NULL, and
2072 sets the variable pointed to by errptr to point to a textual error mes-
2073 sage. This is a static string that is part of the library. You must not
2074 try to free it. Normally, the offset from the start of the pattern to
2075 the data unit that was being processed when the error was discovered is
2076 placed in the variable pointed to by erroffset, which must not be NULL
2077 (if it is, an immediate error is given). However, for an invalid UTF-8
2078 or UTF-16 string, the offset is that of the first data unit of the
2081 Some errors are not detected until the whole pattern has been scanned;
2082 in these cases, the offset passed back is the length of the pattern.
2083 Note that the offset is in data units, not characters, even in a UTF
2084 mode. It may sometimes point into the middle of a UTF-8 or UTF-16 char-
2087 If pcre_compile2() is used instead of pcre_compile(), and the error-
2088 codeptr argument is not NULL, a non-zero error code number is returned
2089 via this argument in the event of an error. This is in addition to the
2090 textual error message. Error codes and messages are listed below.
2092 If the final argument, tableptr, is NULL, PCRE uses a default set of
2093 character tables that are built when PCRE is compiled, using the
2094 default C locale. Otherwise, tableptr must be an address that is the
2095 result of a call to pcre_maketables(). This value is stored with the
2096 compiled pattern, and used again by pcre_exec() and pcre_dfa_exec()
2097 when the pattern is matched. For more discussion, see the section on
2098 locale support below.
2100 This code fragment shows a typical straightforward call to pcre_com-
2107 "^A.*Z", /* the pattern */
2108 0, /* default options */
2109 &error, /* for error message */
2110 &erroffset, /* for error offset */
2111 NULL); /* use default character tables */
2113 The following names for option bits are defined in the pcre.h header
2118 If this bit is set, the pattern is forced to be "anchored", that is, it
2119 is constrained to match only at the first matching point in the string
2120 that is being searched (the "subject string"). This effect can also be
2121 achieved by appropriate constructs in the pattern itself, which is the
2122 only way to do it in Perl.
2126 If this bit is set, pcre_compile() automatically inserts callout items,
2127 all with number 255, before each pattern item. For discussion of the
2128 callout facility, see the pcrecallout documentation.
2133 These options (which are mutually exclusive) control what the \R escape
2134 sequence matches. The choice is either to match only CR, LF, or CRLF,
2135 or to match any Unicode newline sequence. The default is specified when
2136 PCRE is built. It can be overridden from within the pattern, or by set-
2137 ting an option when a compiled pattern is matched.
2141 If this bit is set, letters in the pattern match both upper and lower
2142 case letters. It is equivalent to Perl's /i option, and it can be
2143 changed within a pattern by a (?i) option setting. In UTF-8 mode, PCRE
2144 always understands the concept of case for characters whose values are
2145 less than 128, so caseless matching is always possible. For characters
2146 with higher values, the concept of case is supported if PCRE is com-
2147 piled with Unicode property support, but not otherwise. If you want to
2148 use caseless matching for characters 128 and above, you must ensure
2149 that PCRE is compiled with Unicode property support as well as with
2154 If this bit is set, a dollar metacharacter in the pattern matches only
2155 at the end of the subject string. Without this option, a dollar also
2156 matches immediately before a newline at the end of the string (but not
2157 before any other newlines). The PCRE_DOLLAR_ENDONLY option is ignored
2158 if PCRE_MULTILINE is set. There is no equivalent to this option in
2159 Perl, and no way to set it within a pattern.
2163 If this bit is set, a dot metacharacter in the pattern matches a char-
2164 acter of any value, including one that indicates a newline. However, it
2165 only ever matches one character, even if newlines are coded as CRLF.
2166 Without this option, a dot does not match when the current position is
2167 at a newline. This option is equivalent to Perl's /s option, and it can
2168 be changed within a pattern by a (?s) option setting. A negative class
2169 such as [^a] always matches newline characters, independent of the set-
2170 ting of this option.
2174 If this bit is set, names used to identify capturing subpatterns need
2175 not be unique. This can be helpful for certain types of pattern when it
2176 is known that only one instance of the named subpattern can ever be
2177 matched. There are more details of named subpatterns below; see also
2178 the pcrepattern documentation.
2182 If this bit is set, most white space characters in the pattern are
2183 totally ignored except when escaped or inside a character class. How-
2184 ever, white space is not allowed within sequences such as (?> that
2185 introduce various parenthesized subpatterns, nor within a numerical
2186 quantifier such as {1,3}. However, ignorable white space is permitted
2187 between an item and a following quantifier and between a quantifier and
2188 a following + that indicates possessiveness.
2190 White space did not used to include the VT character (code 11), because
2191 Perl did not treat this character as white space. However, Perl changed
2192 at release 5.18, so PCRE followed at release 8.34, and VT is now
2193 treated as white space.
2195 PCRE_EXTENDED also causes characters between an unescaped # outside a
2196 character class and the next newline, inclusive, to be ignored.
2197 PCRE_EXTENDED is equivalent to Perl's /x option, and it can be changed
2198 within a pattern by a (?x) option setting.
2200 Which characters are interpreted as newlines is controlled by the
2201 options passed to pcre_compile() or by a special sequence at the start
2202 of the pattern, as described in the section entitled "Newline conven-
2203 tions" in the pcrepattern documentation. Note that the end of this type
2204 of comment is a literal newline sequence in the pattern; escape
2205 sequences that happen to represent a newline do not count.
2207 This option makes it possible to include comments inside complicated
2208 patterns. Note, however, that this applies only to data characters.
2209 White space characters may never appear within special character
2210 sequences in a pattern, for example within the sequence (?( that intro-
2211 duces a conditional subpattern.
2215 This option was invented in order to turn on additional functionality
2216 of PCRE that is incompatible with Perl, but it is currently of very
2217 little use. When set, any backslash in a pattern that is followed by a
2218 letter that has no special meaning causes an error, thus reserving
2219 these combinations for future expansion. By default, as in Perl, a
2220 backslash followed by a letter with no special meaning is treated as a
2221 literal. (Perl can, however, be persuaded to give an error for this, by
2222 running it with the -w option.) There are at present no other features
2223 controlled by this option. It can also be set by a (?X) option setting
2228 If this option is set, an unanchored pattern is required to match
2229 before or at the first newline in the subject string, though the
2230 matched text may continue over the newline.
2232 PCRE_JAVASCRIPT_COMPAT
2234 If this option is set, PCRE's behaviour is changed in some ways so that
2235 it is compatible with JavaScript rather than Perl. The changes are as
2238 (1) A lone closing square bracket in a pattern causes a compile-time
2239 error, because this is illegal in JavaScript (by default it is treated
2240 as a data character). Thus, the pattern AB]CD becomes illegal when this
2243 (2) At run time, a back reference to an unset subpattern group matches
2244 an empty string (by default this causes the current matching alterna-
2245 tive to fail). A pattern such as (\1)(a) succeeds when this option is
2246 set (assuming it can find an "a" in the subject), whereas it fails by
2247 default, for Perl compatibility.
2249 (3) \U matches an upper case "U" character; by default \U causes a com-
2250 pile time error (Perl uses \U to upper case subsequent characters).
2252 (4) \u matches a lower case "u" character unless it is followed by four
2253 hexadecimal digits, in which case the hexadecimal number defines the
2254 code point to match. By default, \u causes a compile time error (Perl
2255 uses it to upper case the following character).
2257 (5) \x matches a lower case "x" character unless it is followed by two
2258 hexadecimal digits, in which case the hexadecimal number defines the
2259 code point to match. By default, as in Perl, a hexadecimal number is
2260 always expected after \x, but it may have zero, one, or two digits (so,
2261 for example, \xz matches a binary zero character followed by z).
2265 By default, for the purposes of matching "start of line" and "end of
2266 line", PCRE treats the subject string as consisting of a single line of
2267 characters, even if it actually contains newlines. The "start of line"
2268 metacharacter (^) matches only at the start of the string, and the "end
2269 of line" metacharacter ($) matches only at the end of the string, or
2270 before a terminating newline (except when PCRE_DOLLAR_ENDONLY is set).
2271 Note, however, that unless PCRE_DOTALL is set, the "any character"
2272 metacharacter (.) does not match at a newline. This behaviour (for ^,
2273 $, and dot) is the same as Perl.
2275 When PCRE_MULTILINE it is set, the "start of line" and "end of line"
2276 constructs match immediately following or immediately before internal
2277 newlines in the subject string, respectively, as well as at the very
2278 start and end. This is equivalent to Perl's /m option, and it can be
2279 changed within a pattern by a (?m) option setting. If there are no new-
2280 lines in a subject string, or no occurrences of ^ or $ in a pattern,
2281 setting PCRE_MULTILINE has no effect.
2285 This option locks out interpretation of the pattern as UTF-8 (or UTF-16
2286 or UTF-32 in the 16-bit and 32-bit libraries). In particular, it pre-
2287 vents the creator of the pattern from switching to UTF interpretation
2288 by starting the pattern with (*UTF). This may be useful in applications
2289 that process patterns from external sources. The combination of
2290 PCRE_UTF8 and PCRE_NEVER_UTF also causes an error.
2295 PCRE_NEWLINE_ANYCRLF
2298 These options override the default newline definition that was chosen
2299 when PCRE was built. Setting the first or the second specifies that a
2300 newline is indicated by a single character (CR or LF, respectively).
2301 Setting PCRE_NEWLINE_CRLF specifies that a newline is indicated by the
2302 two-character CRLF sequence. Setting PCRE_NEWLINE_ANYCRLF specifies
2303 that any of the three preceding sequences should be recognized. Setting
2304 PCRE_NEWLINE_ANY specifies that any Unicode newline sequence should be
2307 In an ASCII/Unicode environment, the Unicode newline sequences are the
2308 three just mentioned, plus the single characters VT (vertical tab,
2309 U+000B), FF (form feed, U+000C), NEL (next line, U+0085), LS (line sep-
2310 arator, U+2028), and PS (paragraph separator, U+2029). For the 8-bit
2311 library, the last two are recognized only in UTF-8 mode.
2313 When PCRE is compiled to run in an EBCDIC (mainframe) environment, the
2314 code for CR is 0x0d, the same as ASCII. However, the character code for
2315 LF is normally 0x15, though in some EBCDIC environments 0x25 is used.
2316 Whichever of these is not LF is made to correspond to Unicode's NEL
2317 character. EBCDIC codes are all less than 256. For more details, see
2318 the pcrebuild documentation.
2320 The newline setting in the options word uses three bits that are
2321 treated as a number, giving eight possibilities. Currently only six are
2322 used (default plus the five values above). This means that if you set
2323 more than one newline option, the combination may or may not be sensi-
2324 ble. For example, PCRE_NEWLINE_CR with PCRE_NEWLINE_LF is equivalent to
2325 PCRE_NEWLINE_CRLF, but other combinations may yield unused numbers and
2328 The only time that a line break in a pattern is specially recognized
2329 when compiling is when PCRE_EXTENDED is set. CR and LF are white space
2330 characters, and so are ignored in this mode. Also, an unescaped # out-
2331 side a character class indicates a comment that lasts until after the
2332 next line break sequence. In other circumstances, line break sequences
2333 in patterns are treated as literal data.
2335 The newline option that is set at compile time becomes the default that
2336 is used for pcre_exec() and pcre_dfa_exec(), but it can be overridden.
2338 PCRE_NO_AUTO_CAPTURE
2340 If this option is set, it disables the use of numbered capturing paren-
2341 theses in the pattern. Any opening parenthesis that is not followed by
2342 ? behaves as if it were followed by ?: but named parentheses can still
2343 be used for capturing (and they acquire numbers in the usual way).
2344 There is no equivalent of this option in Perl.
2346 PCRE_NO_AUTO_POSSESS
2348 If this option is set, it disables "auto-possessification". This is an
2349 optimization that, for example, turns a+b into a++b in order to avoid
2350 backtracks into a+ that can never be successful. However, if callouts
2351 are in use, auto-possessification means that some of them are never
2352 taken. You can set this option if you want the matching functions to do
2353 a full unoptimized search and run all the callouts, but it is mainly
2354 provided for testing purposes.
2356 PCRE_NO_START_OPTIMIZE
2358 This is an option that acts at matching time; that is, it is really an
2359 option for pcre_exec() or pcre_dfa_exec(). If it is set at compile
2360 time, it is remembered with the compiled pattern and assumed at match-
2361 ing time. This is necessary if you want to use JIT execution, because
2362 the JIT compiler needs to know whether or not this option is set. For
2363 details see the discussion of PCRE_NO_START_OPTIMIZE below.
2367 This option changes the way PCRE processes \B, \b, \D, \d, \S, \s, \W,
2368 \w, and some of the POSIX character classes. By default, only ASCII
2369 characters are recognized, but if PCRE_UCP is set, Unicode properties
2370 are used instead to classify characters. More details are given in the
2371 section on generic character types in the pcrepattern page. If you set
2372 PCRE_UCP, matching one of the items it affects takes much longer. The
2373 option is available only if PCRE has been compiled with Unicode prop-
2378 This option inverts the "greediness" of the quantifiers so that they
2379 are not greedy by default, but become greedy if followed by "?". It is
2380 not compatible with Perl. It can also be set by a (?U) option setting
2385 This option causes PCRE to regard both the pattern and the subject as
2386 strings of UTF-8 characters instead of single-byte strings. However, it
2387 is available only when PCRE is built to include UTF support. If not,
2388 the use of this option provokes an error. Details of how this option
2389 changes the behaviour of PCRE are given in the pcreunicode page.
2393 When PCRE_UTF8 is set, the validity of the pattern as a UTF-8 string is
2394 automatically checked. There is a discussion about the validity of
2395 UTF-8 strings in the pcreunicode page. If an invalid UTF-8 sequence is
2396 found, pcre_compile() returns an error. If you already know that your
2397 pattern is valid, and you want to skip this check for performance rea-
2398 sons, you can set the PCRE_NO_UTF8_CHECK option. When it is set, the
2399 effect of passing an invalid UTF-8 string as a pattern is undefined. It
2400 may cause your program to crash or loop. Note that this option can also
2401 be passed to pcre_exec() and pcre_dfa_exec(), to suppress the validity
2402 checking of subject strings only. If the same string is being matched
2403 many times, the option can be safely set for the second and subsequent
2404 matchings to improve performance.
2407 COMPILATION ERROR CODES
2409 The following table lists the error codes than may be returned by
2410 pcre_compile2(), along with the error messages that may be returned by
2411 both compiling functions. Note that error messages are always 8-bit
2412 ASCII strings, even in 16-bit or 32-bit mode. As PCRE has developed,
2413 some error codes have fallen out of use. To avoid confusion, they have
2417 1 \ at end of pattern
2418 2 \c at end of pattern
2419 3 unrecognized character follows \
2420 4 numbers out of order in {} quantifier
2421 5 number too big in {} quantifier
2422 6 missing terminating ] for character class
2423 7 invalid escape sequence in character class
2424 8 range out of order in character class
2426 10 [this code is not in use]
2427 11 internal error: unexpected repeat
2428 12 unrecognized character after (? or (?-
2429 13 POSIX named classes are supported only within a class
2431 15 reference to non-existent subpattern
2432 16 erroffset passed as NULL
2433 17 unknown option bit(s) set
2434 18 missing ) after comment
2435 19 [this code is not in use]
2436 20 regular expression is too large
2437 21 failed to get memory
2438 22 unmatched parentheses
2439 23 internal error: code overflow
2440 24 unrecognized character after (?<
2441 25 lookbehind assertion is not fixed length
2442 26 malformed number or name after (?(
2443 27 conditional group contains more than two branches
2444 28 assertion expected after (?(
2445 29 (?R or (?[+-]digits must be followed by )
2446 30 unknown POSIX class name
2447 31 POSIX collating elements are not supported
2448 32 this version of PCRE is compiled without UTF support
2449 33 [this code is not in use]
2450 34 character value in \x{} or \o{} is too large
2451 35 invalid condition (?(0)
2452 36 \C not allowed in lookbehind assertion
2453 37 PCRE does not support \L, \l, \N{name}, \U, or \u
2454 38 number after (?C is > 255
2455 39 closing ) for (?C expected
2456 40 recursive call could loop indefinitely
2457 41 unrecognized character after (?P
2458 42 syntax error in subpattern name (missing terminator)
2459 43 two named subpatterns have the same name
2460 44 invalid UTF-8 string (specifically UTF-8)
2461 45 support for \P, \p, and \X has not been compiled
2462 46 malformed \P or \p sequence
2463 47 unknown property name after \P or \p
2464 48 subpattern name is too long (maximum 32 characters)
2465 49 too many named subpatterns (maximum 10000)
2466 50 [this code is not in use]
2467 51 octal value is greater than \377 in 8-bit non-UTF-8 mode
2468 52 internal error: overran compiling workspace
2469 53 internal error: previously-checked referenced subpattern
2471 54 DEFINE group contains more than one branch
2472 55 repeating a DEFINE group is not allowed
2473 56 inconsistent NEWLINE options
2474 57 \g is not followed by a braced, angle-bracketed, or quoted
2475 name/number or by a plain number
2476 58 a numbered reference must not be zero
2477 59 an argument is not allowed for (*ACCEPT), (*FAIL), or (*COMMIT)
2478 60 (*VERB) not recognized or malformed
2479 61 number is too big
2480 62 subpattern name expected
2481 63 digit expected after (?+
2482 64 ] is an invalid data character in JavaScript compatibility mode
2483 65 different names for subpatterns of the same number are
2485 66 (*MARK) must have an argument
2486 67 this version of PCRE is not compiled with Unicode property
2488 68 \c must be followed by an ASCII character
2489 69 \k is not followed by a braced, angle-bracketed, or quoted name
2490 70 internal error: unknown opcode in find_fixedlength()
2491 71 \N is not supported in a class
2492 72 too many forward references
2493 73 disallowed Unicode code point (>= 0xd800 && <= 0xdfff)
2494 74 invalid UTF-16 string (specifically UTF-16)
2495 75 name is too long in (*MARK), (*PRUNE), (*SKIP), or (*THEN)
2496 76 character value in \u.... sequence is too large
2497 77 invalid UTF-32 string (specifically UTF-32)
2498 78 setting UTF is disabled by the application
2499 79 non-hex character in \x{} (closing brace missing?)
2500 80 non-octal character in \o{} (closing brace missing?)
2501 81 missing opening brace after \o
2502 82 parentheses are too deeply nested
2503 83 invalid range in character class
2504 84 group name must start with a non-digit
2505 85 parentheses are too deeply nested (stack check)
2507 The numbers 32 and 10000 in errors 48 and 49 are defaults; different
2508 values may be used if the limits were changed when PCRE was built.
2513 pcre_extra *pcre_study(const pcre *code, int options,
2514 const char **errptr);
2516 If a compiled pattern is going to be used several times, it is worth
2517 spending more time analyzing it in order to speed up the time taken for
2518 matching. The function pcre_study() takes a pointer to a compiled pat-
2519 tern as its first argument. If studying the pattern produces additional
2520 information that will help speed up matching, pcre_study() returns a
2521 pointer to a pcre_extra block, in which the study_data field points to
2522 the results of the study.
2524 The returned value from pcre_study() can be passed directly to
2525 pcre_exec() or pcre_dfa_exec(). However, a pcre_extra block also con-
2526 tains other fields that can be set by the caller before the block is
2527 passed; these are described below in the section on matching a pattern.
2529 If studying the pattern does not produce any useful information,
2530 pcre_study() returns NULL by default. In that circumstance, if the
2531 calling program wants to pass any of the other fields to pcre_exec() or
2532 pcre_dfa_exec(), it must set up its own pcre_extra block. However, if
2533 pcre_study() is called with the PCRE_STUDY_EXTRA_NEEDED option, it
2534 returns a pcre_extra block even if studying did not find any additional
2535 information. It may still return NULL, however, if an error occurs in
2538 The second argument of pcre_study() contains option bits. There are
2539 three further options in addition to PCRE_STUDY_EXTRA_NEEDED:
2541 PCRE_STUDY_JIT_COMPILE
2542 PCRE_STUDY_JIT_PARTIAL_HARD_COMPILE
2543 PCRE_STUDY_JIT_PARTIAL_SOFT_COMPILE
2545 If any of these are set, and the just-in-time compiler is available,
2546 the pattern is further compiled into machine code that executes much
2547 faster than the pcre_exec() interpretive matching function. If the
2548 just-in-time compiler is not available, these options are ignored. All
2549 undefined bits in the options argument must be zero.
2551 JIT compilation is a heavyweight optimization. It can take some time
2552 for patterns to be analyzed, and for one-off matches and simple pat-
2553 terns the benefit of faster execution might be offset by a much slower
2554 study time. Not all patterns can be optimized by the JIT compiler. For
2555 those that cannot be handled, matching automatically falls back to the
2556 pcre_exec() interpreter. For more details, see the pcrejit documenta-
2559 The third argument for pcre_study() is a pointer for an error message.
2560 If studying succeeds (even if no data is returned), the variable it
2561 points to is set to NULL. Otherwise it is set to point to a textual
2562 error message. This is a static string that is part of the library. You
2563 must not try to free it. You should test the error pointer for NULL
2564 after calling pcre_study(), to be sure that it has run successfully.
2566 When you are finished with a pattern, you can free the memory used for
2567 the study data by calling pcre_free_study(). This function was added to
2568 the API for release 8.20. For earlier versions, the memory could be
2569 freed with pcre_free(), just like the pattern itself. This will still
2570 work in cases where JIT optimization is not used, but it is advisable
2571 to change to the new function when convenient.
2573 This is a typical way in which pcre_study() is used (except that in a
2574 real application there should be tests for errors):
2579 re = pcre_compile("pattern", 0, &error, &erroroffset, NULL);
2581 re, /* result of pcre_compile() */
2583 &error); /* set to NULL or points to a message */
2584 rc = pcre_exec( /* see below for details of pcre_exec() options */
2585 re, sd, "subject", 7, 0, 0, ovector, 30);
2587 pcre_free_study(sd);
2590 Studying a pattern does two things: first, a lower bound for the length
2591 of subject string that is needed to match the pattern is computed. This
2592 does not mean that there are any strings of that length that match, but
2593 it does guarantee that no shorter strings match. The value is used to
2594 avoid wasting time by trying to match strings that are shorter than the
2595 lower bound. You can find out the value in a calling program via the
2596 pcre_fullinfo() function.
2598 Studying a pattern is also useful for non-anchored patterns that do not
2599 have a single fixed starting character. A bitmap of possible starting
2600 bytes is created. This speeds up finding a position in the subject at
2601 which to start matching. (In 16-bit mode, the bitmap is used for 16-bit
2602 values less than 256. In 32-bit mode, the bitmap is used for 32-bit
2603 values less than 256.)
2605 These two optimizations apply to both pcre_exec() and pcre_dfa_exec(),
2606 and the information is also used by the JIT compiler. The optimiza-
2607 tions can be disabled by setting the PCRE_NO_START_OPTIMIZE option.
2608 You might want to do this if your pattern contains callouts or (*MARK)
2609 and you want to make use of these facilities in cases where matching
2612 PCRE_NO_START_OPTIMIZE can be specified at either compile time or exe-
2613 cution time. However, if PCRE_NO_START_OPTIMIZE is passed to
2614 pcre_exec(), (that is, after any JIT compilation has happened) JIT exe-
2615 cution is disabled. For JIT execution to work with PCRE_NO_START_OPTI-
2616 MIZE, the option must be set at compile time.
2618 There is a longer discussion of PCRE_NO_START_OPTIMIZE below.
2623 PCRE handles caseless matching, and determines whether characters are
2624 letters, digits, or whatever, by reference to a set of tables, indexed
2625 by character code point. When running in UTF-8 mode, or in the 16- or
2626 32-bit libraries, this applies only to characters with code points less
2627 than 256. By default, higher-valued code points never match escapes
2628 such as \w or \d. However, if PCRE is built with Unicode property sup-
2629 port, all characters can be tested with \p and \P, or, alternatively,
2630 the PCRE_UCP option can be set when a pattern is compiled; this causes
2631 \w and friends to use Unicode property support instead of the built-in
2634 The use of locales with Unicode is discouraged. If you are handling
2635 characters with code points greater than 128, you should either use
2636 Unicode support, or use locales, but not try to mix the two.
2638 PCRE contains an internal set of tables that are used when the final
2639 argument of pcre_compile() is NULL. These are sufficient for many
2640 applications. Normally, the internal tables recognize only ASCII char-
2641 acters. However, when PCRE is built, it is possible to cause the inter-
2642 nal tables to be rebuilt in the default "C" locale of the local system,
2643 which may cause them to be different.
2645 The internal tables can always be overridden by tables supplied by the
2646 application that calls PCRE. These may be created in a different locale
2647 from the default. As more and more applications change to using Uni-
2648 code, the need for this locale support is expected to die away.
2650 External tables are built by calling the pcre_maketables() function,
2651 which has no arguments, in the relevant locale. The result can then be
2652 passed to pcre_compile() as often as necessary. For example, to build
2653 and use tables that are appropriate for the French locale (where
2654 accented characters with values greater than 128 are treated as let-
2655 ters), the following code could be used:
2657 setlocale(LC_CTYPE, "fr_FR");
2658 tables = pcre_maketables();
2659 re = pcre_compile(..., tables);
2661 The locale name "fr_FR" is used on Linux and other Unix-like systems;
2662 if you are using Windows, the name for the French locale is "french".
2664 When pcre_maketables() runs, the tables are built in memory that is
2665 obtained via pcre_malloc. It is the caller's responsibility to ensure
2666 that the memory containing the tables remains available for as long as
2669 The pointer that is passed to pcre_compile() is saved with the compiled
2670 pattern, and the same tables are used via this pointer by pcre_study()
2671 and also by pcre_exec() and pcre_dfa_exec(). Thus, for any single pat-
2672 tern, compilation, studying and matching all happen in the same locale,
2673 but different patterns can be processed in different locales.
2675 It is possible to pass a table pointer or NULL (indicating the use of
2676 the internal tables) to pcre_exec() or pcre_dfa_exec() (see the discus-
2677 sion below in the section on matching a pattern). This facility is pro-
2678 vided for use with pre-compiled patterns that have been saved and
2679 reloaded. Character tables are not saved with patterns, so if a non-
2680 standard table was used at compile time, it must be provided again when
2681 the reloaded pattern is matched. Attempting to use this facility to
2682 match a pattern in a different locale from the one in which it was com-
2683 piled is likely to lead to anomalous (usually incorrect) results.
2686 INFORMATION ABOUT A PATTERN
2688 int pcre_fullinfo(const pcre *code, const pcre_extra *extra,
2689 int what, void *where);
2691 The pcre_fullinfo() function returns information about a compiled pat-
2692 tern. It replaces the pcre_info() function, which was removed from the
2693 library at version 8.30, after more than 10 years of obsolescence.
2695 The first argument for pcre_fullinfo() is a pointer to the compiled
2696 pattern. The second argument is the result of pcre_study(), or NULL if
2697 the pattern was not studied. The third argument specifies which piece
2698 of information is required, and the fourth argument is a pointer to a
2699 variable to receive the data. The yield of the function is zero for
2700 success, or one of the following negative numbers:
2702 PCRE_ERROR_NULL the argument code was NULL
2703 the argument where was NULL
2704 PCRE_ERROR_BADMAGIC the "magic number" was not found
2705 PCRE_ERROR_BADENDIANNESS the pattern was compiled with different
2707 PCRE_ERROR_BADOPTION the value of what was invalid
2708 PCRE_ERROR_UNSET the requested field is not set
2710 The "magic number" is placed at the start of each compiled pattern as
2711 an simple check against passing an arbitrary memory pointer. The endi-
2712 anness error can occur if a compiled pattern is saved and reloaded on a
2713 different host. Here is a typical call of pcre_fullinfo(), to obtain
2714 the length of the compiled pattern:
2719 re, /* result of pcre_compile() */
2720 sd, /* result of pcre_study(), or NULL */
2721 PCRE_INFO_SIZE, /* what is required */
2722 &length); /* where to put the data */
2724 The possible values for the third argument are defined in pcre.h, and
2727 PCRE_INFO_BACKREFMAX
2729 Return the number of the highest back reference in the pattern. The
2730 fourth argument should point to an int variable. Zero is returned if
2731 there are no back references.
2733 PCRE_INFO_CAPTURECOUNT
2735 Return the number of capturing subpatterns in the pattern. The fourth
2736 argument should point to an int variable.
2738 PCRE_INFO_DEFAULT_TABLES
2740 Return a pointer to the internal default character tables within PCRE.
2741 The fourth argument should point to an unsigned char * variable. This
2742 information call is provided for internal use by the pcre_study() func-
2743 tion. External callers can cause PCRE to use its internal tables by
2744 passing a NULL table pointer.
2746 PCRE_INFO_FIRSTBYTE (deprecated)
2748 Return information about the first data unit of any matched string, for
2749 a non-anchored pattern. The name of this option refers to the 8-bit
2750 library, where data units are bytes. The fourth argument should point
2751 to an int variable. Negative values are used for special cases. How-
2752 ever, this means that when the 32-bit library is in non-UTF-32 mode,
2753 the full 32-bit range of characters cannot be returned. For this rea-
2754 son, this value is deprecated; use PCRE_INFO_FIRSTCHARACTERFLAGS and
2755 PCRE_INFO_FIRSTCHARACTER instead.
2757 If there is a fixed first value, for example, the letter "c" from a
2758 pattern such as (cat|cow|coyote), its value is returned. In the 8-bit
2759 library, the value is always less than 256. In the 16-bit library the
2760 value can be up to 0xffff. In the 32-bit library the value can be up to
2763 If there is no fixed first value, and if either
2765 (a) the pattern was compiled with the PCRE_MULTILINE option, and every
2766 branch starts with "^", or
2768 (b) every branch of the pattern starts with ".*" and PCRE_DOTALL is not
2769 set (if it were set, the pattern would be anchored),
2771 -1 is returned, indicating that the pattern matches only at the start
2772 of a subject string or after any newline within the string. Otherwise
2773 -2 is returned. For anchored patterns, -2 is returned.
2775 PCRE_INFO_FIRSTCHARACTER
2777 Return the value of the first data unit (non-UTF character) of any
2778 matched string in the situation where PCRE_INFO_FIRSTCHARACTERFLAGS
2779 returns 1; otherwise return 0. The fourth argument should point to an
2782 In the 8-bit library, the value is always less than 256. In the 16-bit
2783 library the value can be up to 0xffff. In the 32-bit library in UTF-32
2784 mode the value can be up to 0x10ffff, and up to 0xffffffff when not
2787 PCRE_INFO_FIRSTCHARACTERFLAGS
2789 Return information about the first data unit of any matched string, for
2790 a non-anchored pattern. The fourth argument should point to an int
2793 If there is a fixed first value, for example, the letter "c" from a
2794 pattern such as (cat|cow|coyote), 1 is returned, and the character
2795 value can be retrieved using PCRE_INFO_FIRSTCHARACTER. If there is no
2796 fixed first value, and if either
2798 (a) the pattern was compiled with the PCRE_MULTILINE option, and every
2799 branch starts with "^", or
2801 (b) every branch of the pattern starts with ".*" and PCRE_DOTALL is not
2802 set (if it were set, the pattern would be anchored),
2804 2 is returned, indicating that the pattern matches only at the start of
2805 a subject string or after any newline within the string. Otherwise 0 is
2806 returned. For anchored patterns, 0 is returned.
2808 PCRE_INFO_FIRSTTABLE
2810 If the pattern was studied, and this resulted in the construction of a
2811 256-bit table indicating a fixed set of values for the first data unit
2812 in any matching string, a pointer to the table is returned. Otherwise
2813 NULL is returned. The fourth argument should point to an unsigned char
2818 Return 1 if the pattern contains any explicit matches for CR or LF
2819 characters, otherwise 0. The fourth argument should point to an int
2820 variable. An explicit match is either a literal CR or LF character, or
2825 Return 1 if the (?J) or (?-J) option setting is used in the pattern,
2826 otherwise 0. The fourth argument should point to an int variable. (?J)
2827 and (?-J) set and unset the local PCRE_DUPNAMES option, respectively.
2831 Return 1 if the pattern was studied with one of the JIT options, and
2832 just-in-time compiling was successful. The fourth argument should point
2833 to an int variable. A return value of 0 means that JIT support is not
2834 available in this version of PCRE, or that the pattern was not studied
2835 with a JIT option, or that the JIT compiler could not handle this par-
2836 ticular pattern. See the pcrejit documentation for details of what can
2837 and cannot be handled.
2841 If the pattern was successfully studied with a JIT option, return the
2842 size of the JIT compiled code, otherwise return zero. The fourth argu-
2843 ment should point to a size_t variable.
2845 PCRE_INFO_LASTLITERAL
2847 Return the value of the rightmost literal data unit that must exist in
2848 any matched string, other than at its start, if such a value has been
2849 recorded. The fourth argument should point to an int variable. If there
2850 is no such value, -1 is returned. For anchored patterns, a last literal
2851 value is recorded only if it follows something of variable length. For
2852 example, for the pattern /^a\d+z\d+/ the returned value is "z", but for
2853 /^a\dz\d/ the returned value is -1.
2855 Since for the 32-bit library using the non-UTF-32 mode, this function
2856 is unable to return the full 32-bit range of characters, this value is
2857 deprecated; instead the PCRE_INFO_REQUIREDCHARFLAGS and
2858 PCRE_INFO_REQUIREDCHAR values should be used.
2860 PCRE_INFO_MATCH_EMPTY
2862 Return 1 if the pattern can match an empty string, otherwise 0. The
2863 fourth argument should point to an int variable.
2865 PCRE_INFO_MATCHLIMIT
2867 If the pattern set a match limit by including an item of the form
2868 (*LIMIT_MATCH=nnnn) at the start, the value is returned. The fourth
2869 argument should point to an unsigned 32-bit integer. If no such value
2870 has been set, the call to pcre_fullinfo() returns the error
2873 PCRE_INFO_MAXLOOKBEHIND
2875 Return the number of characters (NB not data units) in the longest
2876 lookbehind assertion in the pattern. This information is useful when
2877 doing multi-segment matching using the partial matching facilities.
2878 Note that the simple assertions \b and \B require a one-character look-
2879 behind. \A also registers a one-character lookbehind, though it does
2880 not actually inspect the previous character. This is to ensure that at
2881 least one character from the old segment is retained when a new segment
2882 is processed. Otherwise, if there are no lookbehinds in the pattern, \A
2883 might match incorrectly at the start of a new segment.
2887 If the pattern was studied and a minimum length for matching subject
2888 strings was computed, its value is returned. Otherwise the returned
2889 value is -1. The value is a number of characters, which in UTF mode may
2890 be different from the number of data units. The fourth argument should
2891 point to an int variable. A non-negative value is a lower bound to the
2892 length of any matching string. There may not be any strings of that
2893 length that do actually match, but every string that does match is at
2897 PCRE_INFO_NAMEENTRYSIZE
2900 PCRE supports the use of named as well as numbered capturing parenthe-
2901 ses. The names are just an additional way of identifying the parenthe-
2902 ses, which still acquire numbers. Several convenience functions such as
2903 pcre_get_named_substring() are provided for extracting captured sub-
2904 strings by name. It is also possible to extract the data directly, by
2905 first converting the name to a number in order to access the correct
2906 pointers in the output vector (described with pcre_exec() below). To do
2907 the conversion, you need to use the name-to-number map, which is
2908 described by these three values.
2910 The map consists of a number of fixed-size entries. PCRE_INFO_NAMECOUNT
2911 gives the number of entries, and PCRE_INFO_NAMEENTRYSIZE gives the size
2912 of each entry; both of these return an int value. The entry size
2913 depends on the length of the longest name. PCRE_INFO_NAMETABLE returns
2914 a pointer to the first entry of the table. This is a pointer to char in
2915 the 8-bit library, where the first two bytes of each entry are the num-
2916 ber of the capturing parenthesis, most significant byte first. In the
2917 16-bit library, the pointer points to 16-bit data units, the first of
2918 which contains the parenthesis number. In the 32-bit library, the
2919 pointer points to 32-bit data units, the first of which contains the
2920 parenthesis number. The rest of the entry is the corresponding name,
2923 The names are in alphabetical order. If (?| is used to create multiple
2924 groups with the same number, as described in the section on duplicate
2925 subpattern numbers in the pcrepattern page, the groups may be given the
2926 same name, but there is only one entry in the table. Different names
2927 for groups of the same number are not permitted. Duplicate names for
2928 subpatterns with different numbers are permitted, but only if PCRE_DUP-
2929 NAMES is set. They appear in the table in the order in which they were
2930 found in the pattern. In the absence of (?| this is the order of
2931 increasing number; when (?| is used this is not necessarily the case
2932 because later subpatterns may have lower numbers.
2934 As a simple example of the name/number table, consider the following
2935 pattern after compilation by the 8-bit library (assume PCRE_EXTENDED is
2936 set, so white space - including newlines - is ignored):
2938 (?<date> (?<year>(\d\d)?\d\d) -
2939 (?<month>\d\d) - (?<day>\d\d) )
2941 There are four named subpatterns, so the table has four entries, and
2942 each entry in the table is eight bytes long. The table is as follows,
2943 with non-printing bytes shows in hexadecimal, and undefined bytes shown
2947 00 05 d a y 00 ?? ??
2951 When writing code to extract data from named subpatterns using the
2952 name-to-number map, remember that the length of the entries is likely
2953 to be different for each compiled pattern.
2957 Return 1 if the pattern can be used for partial matching with
2958 pcre_exec(), otherwise 0. The fourth argument should point to an int
2959 variable. From release 8.00, this always returns 1, because the
2960 restrictions that previously applied to partial matching have been
2961 lifted. The pcrepartial documentation gives details of partial match-
2966 Return a copy of the options with which the pattern was compiled. The
2967 fourth argument should point to an unsigned long int variable. These
2968 option bits are those specified in the call to pcre_compile(), modified
2969 by any top-level option settings at the start of the pattern itself. In
2970 other words, they are the options that will be in force when matching
2971 starts. For example, if the pattern /(?im)abc(?-i)d/ is compiled with
2972 the PCRE_EXTENDED option, the result is PCRE_CASELESS, PCRE_MULTILINE,
2975 A pattern is automatically anchored by PCRE if all of its top-level
2976 alternatives begin with one of the following:
2978 ^ unless PCRE_MULTILINE is set
2981 .* if PCRE_DOTALL is set and there are no back
2982 references to the subpattern in which .* appears
2984 For such patterns, the PCRE_ANCHORED bit is set in the options returned
2987 PCRE_INFO_RECURSIONLIMIT
2989 If the pattern set a recursion limit by including an item of the form
2990 (*LIMIT_RECURSION=nnnn) at the start, the value is returned. The fourth
2991 argument should point to an unsigned 32-bit integer. If no such value
2992 has been set, the call to pcre_fullinfo() returns the error
2997 Return the size of the compiled pattern in bytes (for all three
2998 libraries). The fourth argument should point to a size_t variable. This
2999 value does not include the size of the pcre structure that is returned
3000 by pcre_compile(). The value that is passed as the argument to
3001 pcre_malloc() when pcre_compile() is getting memory in which to place
3002 the compiled data is the value returned by this option plus the size of
3003 the pcre structure. Studying a compiled pattern, with or without JIT,
3004 does not alter the value returned by this option.
3008 Return the size in bytes (for all three libraries) of the data block
3009 pointed to by the study_data field in a pcre_extra block. If pcre_extra
3010 is NULL, or there is no study data, zero is returned. The fourth argu-
3011 ment should point to a size_t variable. The study_data field is set by
3012 pcre_study() to record information that will speed up matching (see the
3013 section entitled "Studying a pattern" above). The format of the
3014 study_data block is private, but its length is made available via this
3015 option so that it can be saved and restored (see the pcreprecompile
3016 documentation for details).
3018 PCRE_INFO_REQUIREDCHARFLAGS
3020 Returns 1 if there is a rightmost literal data unit that must exist in
3021 any matched string, other than at its start. The fourth argument should
3022 point to an int variable. If there is no such value, 0 is returned. If
3023 returning 1, the character value itself can be retrieved using
3024 PCRE_INFO_REQUIREDCHAR.
3026 For anchored patterns, a last literal value is recorded only if it fol-
3027 lows something of variable length. For example, for the pattern
3028 /^a\d+z\d+/ the returned value 1 (with "z" returned from
3029 PCRE_INFO_REQUIREDCHAR), but for /^a\dz\d/ the returned value is 0.
3031 PCRE_INFO_REQUIREDCHAR
3033 Return the value of the rightmost literal data unit that must exist in
3034 any matched string, other than at its start, if such a value has been
3035 recorded. The fourth argument should point to an uint32_t variable. If
3036 there is no such value, 0 is returned.
3041 int pcre_refcount(pcre *code, int adjust);
3043 The pcre_refcount() function is used to maintain a reference count in
3044 the data block that contains a compiled pattern. It is provided for the
3045 benefit of applications that operate in an object-oriented manner,
3046 where different parts of the application may be using the same compiled
3047 pattern, but you want to free the block when they are all done.
3049 When a pattern is compiled, the reference count field is initialized to
3050 zero. It is changed only by calling this function, whose action is to
3051 add the adjust value (which may be positive or negative) to it. The
3052 yield of the function is the new value. However, the value of the count
3053 is constrained to lie between 0 and 65535, inclusive. If the new value
3054 is outside these limits, it is forced to the appropriate limit value.
3056 Except when it is zero, the reference count is not correctly preserved
3057 if a pattern is compiled on one host and then transferred to a host
3058 whose byte-order is different. (This seems a highly unlikely scenario.)
3061 MATCHING A PATTERN: THE TRADITIONAL FUNCTION
3063 int pcre_exec(const pcre *code, const pcre_extra *extra,
3064 const char *subject, int length, int startoffset,
3065 int options, int *ovector, int ovecsize);
3067 The function pcre_exec() is called to match a subject string against a
3068 compiled pattern, which is passed in the code argument. If the pattern
3069 was studied, the result of the study should be passed in the extra
3070 argument. You can call pcre_exec() with the same code and extra argu-
3071 ments as many times as you like, in order to match different subject
3072 strings with the same pattern.
3074 This function is the main matching facility of the library, and it
3075 operates in a Perl-like manner. For specialist use there is also an
3076 alternative matching function, which is described below in the section
3077 about the pcre_dfa_exec() function.
3079 In most applications, the pattern will have been compiled (and option-
3080 ally studied) in the same process that calls pcre_exec(). However, it
3081 is possible to save compiled patterns and study data, and then use them
3082 later in different processes, possibly even on different hosts. For a
3083 discussion about this, see the pcreprecompile documentation.
3085 Here is an example of a simple call to pcre_exec():
3090 re, /* result of pcre_compile() */
3091 NULL, /* we didn't study the pattern */
3092 "some string", /* the subject string */
3093 11, /* the length of the subject string */
3094 0, /* start at offset 0 in the subject */
3095 0, /* default options */
3096 ovector, /* vector of integers for substring information */
3097 30); /* number of elements (NOT size in bytes) */
3099 Extra data for pcre_exec()
3101 If the extra argument is not NULL, it must point to a pcre_extra data
3102 block. The pcre_study() function returns such a block (when it doesn't
3103 return NULL), but you can also create one for yourself, and pass addi-
3104 tional information in it. The pcre_extra block contains the following
3105 fields (not necessarily in this order):
3107 unsigned long int flags;
3109 void *executable_jit;
3110 unsigned long int match_limit;
3111 unsigned long int match_limit_recursion;
3113 const unsigned char *tables;
3114 unsigned char **mark;
3116 In the 16-bit version of this structure, the mark field has type
3119 In the 32-bit version of this structure, the mark field has type
3122 The flags field is used to specify which of the other fields are set.
3125 PCRE_EXTRA_CALLOUT_DATA
3126 PCRE_EXTRA_EXECUTABLE_JIT
3128 PCRE_EXTRA_MATCH_LIMIT
3129 PCRE_EXTRA_MATCH_LIMIT_RECURSION
3130 PCRE_EXTRA_STUDY_DATA
3133 Other flag bits should be set to zero. The study_data field and some-
3134 times the executable_jit field are set in the pcre_extra block that is
3135 returned by pcre_study(), together with the appropriate flag bits. You
3136 should not set these yourself, but you may add to the block by setting
3137 other fields and their corresponding flag bits.
3139 The match_limit field provides a means of preventing PCRE from using up
3140 a vast amount of resources when running patterns that are not going to
3141 match, but which have a very large number of possibilities in their
3142 search trees. The classic example is a pattern that uses nested unlim-
3145 Internally, pcre_exec() uses a function called match(), which it calls
3146 repeatedly (sometimes recursively). The limit set by match_limit is
3147 imposed on the number of times this function is called during a match,
3148 which has the effect of limiting the amount of backtracking that can
3149 take place. For patterns that are not anchored, the count restarts from
3150 zero for each position in the subject string.
3152 When pcre_exec() is called with a pattern that was successfully studied
3153 with a JIT option, the way that the matching is executed is entirely
3154 different. However, there is still the possibility of runaway matching
3155 that goes on for a very long time, and so the match_limit value is also
3156 used in this case (but in a different way) to limit how long the match-
3159 The default value for the limit can be set when PCRE is built; the
3160 default default is 10 million, which handles all but the most extreme
3161 cases. You can override the default by suppling pcre_exec() with a
3162 pcre_extra block in which match_limit is set, and
3163 PCRE_EXTRA_MATCH_LIMIT is set in the flags field. If the limit is
3164 exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3166 A value for the match limit may also be supplied by an item at the
3167 start of a pattern of the form
3171 where d is a decimal number. However, such a setting is ignored unless
3172 d is less than the limit set by the caller of pcre_exec() or, if no
3173 such limit is set, less than the default.
3175 The match_limit_recursion field is similar to match_limit, but instead
3176 of limiting the total number of times that match() is called, it limits
3177 the depth of recursion. The recursion depth is a smaller number than
3178 the total number of calls, because not all calls to match() are recur-
3179 sive. This limit is of use only if it is set smaller than match_limit.
3181 Limiting the recursion depth limits the amount of machine stack that
3182 can be used, or, when PCRE has been compiled to use memory on the heap
3183 instead of the stack, the amount of heap memory that can be used. This
3184 limit is not relevant, and is ignored, when matching is done using JIT
3187 The default value for match_limit_recursion can be set when PCRE is
3188 built; the default default is the same value as the default for
3189 match_limit. You can override the default by suppling pcre_exec() with
3190 a pcre_extra block in which match_limit_recursion is set, and
3191 PCRE_EXTRA_MATCH_LIMIT_RECURSION is set in the flags field. If the
3192 limit is exceeded, pcre_exec() returns PCRE_ERROR_RECURSIONLIMIT.
3194 A value for the recursion limit may also be supplied by an item at the
3195 start of a pattern of the form
3197 (*LIMIT_RECURSION=d)
3199 where d is a decimal number. However, such a setting is ignored unless
3200 d is less than the limit set by the caller of pcre_exec() or, if no
3201 such limit is set, less than the default.
3203 The callout_data field is used in conjunction with the "callout" fea-
3204 ture, and is described in the pcrecallout documentation.
3206 The tables field is provided for use with patterns that have been pre-
3207 compiled using custom character tables, saved to disc or elsewhere, and
3208 then reloaded, because the tables that were used to compile a pattern
3209 are not saved with it. See the pcreprecompile documentation for a dis-
3210 cussion of saving compiled patterns for later use. If NULL is passed
3211 using this mechanism, it forces PCRE's internal tables to be used.
3213 Warning: The tables that pcre_exec() uses must be the same as those
3214 that were used when the pattern was compiled. If this is not the case,
3215 the behaviour of pcre_exec() is undefined. Therefore, when a pattern is
3216 compiled and matched in the same process, this field should never be
3217 set. In this (the most common) case, the correct table pointer is auto-
3218 matically passed with the compiled pattern from pcre_compile() to
3221 If PCRE_EXTRA_MARK is set in the flags field, the mark field must be
3222 set to point to a suitable variable. If the pattern contains any back-
3223 tracking control verbs such as (*MARK:NAME), and the execution ends up
3224 with a name to pass back, a pointer to the name string (zero termi-
3225 nated) is placed in the variable pointed to by the mark field. The
3226 names are within the compiled pattern; if you wish to retain such a
3227 name you must copy it before freeing the memory of a compiled pattern.
3228 If there is no name to pass back, the variable pointed to by the mark
3229 field is set to NULL. For details of the backtracking control verbs,
3230 see the section entitled "Backtracking control" in the pcrepattern doc-
3233 Option bits for pcre_exec()
3235 The unused bits of the options argument for pcre_exec() must be zero.
3236 The only bits that may be set are PCRE_ANCHORED, PCRE_NEWLINE_xxx,
3237 PCRE_NOTBOL, PCRE_NOTEOL, PCRE_NOTEMPTY, PCRE_NOTEMPTY_ATSTART,
3238 PCRE_NO_START_OPTIMIZE, PCRE_NO_UTF8_CHECK, PCRE_PARTIAL_HARD, and
3241 If the pattern was successfully studied with one of the just-in-time
3242 (JIT) compile options, the only supported options for JIT execution are
3243 PCRE_NO_UTF8_CHECK, PCRE_NOTBOL, PCRE_NOTEOL, PCRE_NOTEMPTY,
3244 PCRE_NOTEMPTY_ATSTART, PCRE_PARTIAL_HARD, and PCRE_PARTIAL_SOFT. If an
3245 unsupported option is used, JIT execution is disabled and the normal
3246 interpretive code in pcre_exec() is run.
3250 The PCRE_ANCHORED option limits pcre_exec() to matching at the first
3251 matching position. If a pattern was compiled with PCRE_ANCHORED, or
3252 turned out to be anchored by virtue of its contents, it cannot be made
3253 unachored at matching time.
3258 These options (which are mutually exclusive) control what the \R escape
3259 sequence matches. The choice is either to match only CR, LF, or CRLF,
3260 or to match any Unicode newline sequence. These options override the
3261 choice that was made or defaulted when the pattern was compiled.
3266 PCRE_NEWLINE_ANYCRLF
3269 These options override the newline definition that was chosen or
3270 defaulted when the pattern was compiled. For details, see the descrip-
3271 tion of pcre_compile() above. During matching, the newline choice
3272 affects the behaviour of the dot, circumflex, and dollar metacharac-
3273 ters. It may also alter the way the match position is advanced after a
3274 match failure for an unanchored pattern.
3276 When PCRE_NEWLINE_CRLF, PCRE_NEWLINE_ANYCRLF, or PCRE_NEWLINE_ANY is
3277 set, and a match attempt for an unanchored pattern fails when the cur-
3278 rent position is at a CRLF sequence, and the pattern contains no
3279 explicit matches for CR or LF characters, the match position is
3280 advanced by two characters instead of one, in other words, to after the
3283 The above rule is a compromise that makes the most common cases work as
3284 expected. For example, if the pattern is .+A (and the PCRE_DOTALL
3285 option is not set), it does not match the string "\r\nA" because, after
3286 failing at the start, it skips both the CR and the LF before retrying.
3287 However, the pattern [\r\n]A does match that string, because it con-
3288 tains an explicit CR or LF reference, and so advances only by one char-
3289 acter after the first failure.
3291 An explicit match for CR of LF is either a literal appearance of one of
3292 those characters, or one of the \r or \n escape sequences. Implicit
3293 matches such as [^X] do not count, nor does \s (which includes CR and
3294 LF in the characters that it matches).
3296 Notwithstanding the above, anomalous effects may still occur when CRLF
3297 is a valid newline sequence and explicit \r or \n escapes appear in the
3302 This option specifies that first character of the subject string is not
3303 the beginning of a line, so the circumflex metacharacter should not
3304 match before it. Setting this without PCRE_MULTILINE (at compile time)
3305 causes circumflex never to match. This option affects only the behav-
3306 iour of the circumflex metacharacter. It does not affect \A.
3310 This option specifies that the end of the subject string is not the end
3311 of a line, so the dollar metacharacter should not match it nor (except
3312 in multiline mode) a newline immediately before it. Setting this with-
3313 out PCRE_MULTILINE (at compile time) causes dollar never to match. This
3314 option affects only the behaviour of the dollar metacharacter. It does
3315 not affect \Z or \z.
3319 An empty string is not considered to be a valid match if this option is
3320 set. If there are alternatives in the pattern, they are tried. If all
3321 the alternatives match the empty string, the entire match fails. For
3322 example, if the pattern
3326 is applied to a string not beginning with "a" or "b", it matches an
3327 empty string at the start of the subject. With PCRE_NOTEMPTY set, this
3328 match is not valid, so PCRE searches further into the string for occur-
3329 rences of "a" or "b".
3331 PCRE_NOTEMPTY_ATSTART
3333 This is like PCRE_NOTEMPTY, except that an empty string match that is
3334 not at the start of the subject is permitted. If the pattern is
3335 anchored, such a match can occur only if the pattern contains \K.
3337 Perl has no direct equivalent of PCRE_NOTEMPTY or
3338 PCRE_NOTEMPTY_ATSTART, but it does make a special case of a pattern
3339 match of the empty string within its split() function, and when using
3340 the /g modifier. It is possible to emulate Perl's behaviour after
3341 matching a null string by first trying the match again at the same off-
3342 set with PCRE_NOTEMPTY_ATSTART and PCRE_ANCHORED, and then if that
3343 fails, by advancing the starting offset (see below) and trying an ordi-
3344 nary match again. There is some code that demonstrates how to do this
3345 in the pcredemo sample program. In the most general case, you have to
3346 check to see if the newline convention recognizes CRLF as a newline,
3347 and if so, and the current character is CR followed by LF, advance the
3348 starting offset by two characters instead of one.
3350 PCRE_NO_START_OPTIMIZE
3352 There are a number of optimizations that pcre_exec() uses at the start
3353 of a match, in order to speed up the process. For example, if it is
3354 known that an unanchored match must start with a specific character, it
3355 searches the subject for that character, and fails immediately if it
3356 cannot find it, without actually running the main matching function.
3357 This means that a special item such as (*COMMIT) at the start of a pat-
3358 tern is not considered until after a suitable starting point for the
3359 match has been found. Also, when callouts or (*MARK) items are in use,
3360 these "start-up" optimizations can cause them to be skipped if the pat-
3361 tern is never actually used. The start-up optimizations are in effect a
3362 pre-scan of the subject that takes place before the pattern is run.
3364 The PCRE_NO_START_OPTIMIZE option disables the start-up optimizations,
3365 possibly causing performance to suffer, but ensuring that in cases
3366 where the result is "no match", the callouts do occur, and that items
3367 such as (*COMMIT) and (*MARK) are considered at every possible starting
3368 position in the subject string. If PCRE_NO_START_OPTIMIZE is set at
3369 compile time, it cannot be unset at matching time. The use of
3370 PCRE_NO_START_OPTIMIZE at matching time (that is, passing it to
3371 pcre_exec()) disables JIT execution; in this situation, matching is
3372 always done using interpretively.
3374 Setting PCRE_NO_START_OPTIMIZE can change the outcome of a matching
3375 operation. Consider the pattern
3379 When this is compiled, PCRE records the fact that a match must start
3380 with the character "A". Suppose the subject string is "DEFABC". The
3381 start-up optimization scans along the subject, finds "A" and runs the
3382 first match attempt from there. The (*COMMIT) item means that the pat-
3383 tern must match the current starting position, which in this case, it
3384 does. However, if the same match is run with PCRE_NO_START_OPTIMIZE
3385 set, the initial scan along the subject string does not happen. The
3386 first match attempt is run starting from "D" and when this fails,
3387 (*COMMIT) prevents any further matches being tried, so the overall
3388 result is "no match". If the pattern is studied, more start-up opti-
3389 mizations may be used. For example, a minimum length for the subject
3390 may be recorded. Consider the pattern
3394 The minimum length for a match is one character. If the subject is
3395 "ABC", there will be attempts to match "ABC", "BC", "C", and then
3396 finally an empty string. If the pattern is studied, the final attempt
3397 does not take place, because PCRE knows that the subject is too short,
3398 and so the (*MARK) is never encountered. In this case, studying the
3399 pattern does not affect the overall match result, which is still "no
3400 match", but it does affect the auxiliary information that is returned.
3404 When PCRE_UTF8 is set at compile time, the validity of the subject as a
3405 UTF-8 string is automatically checked when pcre_exec() is subsequently
3406 called. The entire string is checked before any other processing takes
3407 place. The value of startoffset is also checked to ensure that it
3408 points to the start of a UTF-8 character. There is a discussion about
3409 the validity of UTF-8 strings in the pcreunicode page. If an invalid
3410 sequence of bytes is found, pcre_exec() returns the error
3411 PCRE_ERROR_BADUTF8 or, if PCRE_PARTIAL_HARD is set and the problem is a
3412 truncated character at the end of the subject, PCRE_ERROR_SHORTUTF8. In
3413 both cases, information about the precise nature of the error may also
3414 be returned (see the descriptions of these errors in the section enti-
3415 tled Error return values from pcre_exec() below). If startoffset con-
3416 tains a value that does not point to the start of a UTF-8 character (or
3417 to the end of the subject), PCRE_ERROR_BADUTF8_OFFSET is returned.
3419 If you already know that your subject is valid, and you want to skip
3420 these checks for performance reasons, you can set the
3421 PCRE_NO_UTF8_CHECK option when calling pcre_exec(). You might want to
3422 do this for the second and subsequent calls to pcre_exec() if you are
3423 making repeated calls to find all the matches in a single subject
3424 string. However, you should be sure that the value of startoffset
3425 points to the start of a character (or the end of the subject). When
3426 PCRE_NO_UTF8_CHECK is set, the effect of passing an invalid string as a
3427 subject or an invalid value of startoffset is undefined. Your program
3433 These options turn on the partial matching feature. For backwards com-
3434 patibility, PCRE_PARTIAL is a synonym for PCRE_PARTIAL_SOFT. A partial
3435 match occurs if the end of the subject string is reached successfully,
3436 but there are not enough subject characters to complete the match. If
3437 this happens when PCRE_PARTIAL_SOFT (but not PCRE_PARTIAL_HARD) is set,
3438 matching continues by testing any remaining alternatives. Only if no
3439 complete match can be found is PCRE_ERROR_PARTIAL returned instead of
3440 PCRE_ERROR_NOMATCH. In other words, PCRE_PARTIAL_SOFT says that the
3441 caller is prepared to handle a partial match, but only if no complete
3444 If PCRE_PARTIAL_HARD is set, it overrides PCRE_PARTIAL_SOFT. In this
3445 case, if a partial match is found, pcre_exec() immediately returns
3446 PCRE_ERROR_PARTIAL, without considering any other alternatives. In
3447 other words, when PCRE_PARTIAL_HARD is set, a partial match is consid-
3448 ered to be more important that an alternative complete match.
3450 In both cases, the portion of the string that was inspected when the
3451 partial match was found is set as the first matching string. There is a
3452 more detailed discussion of partial and multi-segment matching, with
3453 examples, in the pcrepartial documentation.
3455 The string to be matched by pcre_exec()
3457 The subject string is passed to pcre_exec() as a pointer in subject, a
3458 length in length, and a starting offset in startoffset. The units for
3459 length and startoffset are bytes for the 8-bit library, 16-bit data
3460 items for the 16-bit library, and 32-bit data items for the 32-bit
3463 If startoffset is negative or greater than the length of the subject,
3464 pcre_exec() returns PCRE_ERROR_BADOFFSET. When the starting offset is
3465 zero, the search for a match starts at the beginning of the subject,
3466 and this is by far the most common case. In UTF-8 or UTF-16 mode, the
3467 offset must point to the start of a character, or the end of the sub-
3468 ject (in UTF-32 mode, one data unit equals one character, so all off-
3469 sets are valid). Unlike the pattern string, the subject may contain
3472 A non-zero starting offset is useful when searching for another match
3473 in the same subject by calling pcre_exec() again after a previous suc-
3474 cess. Setting startoffset differs from just passing over a shortened
3475 string and setting PCRE_NOTBOL in the case of a pattern that begins
3476 with any kind of lookbehind. For example, consider the pattern
3480 which finds occurrences of "iss" in the middle of words. (\B matches
3481 only if the current position in the subject is not a word boundary.)
3482 When applied to the string "Mississipi" the first call to pcre_exec()
3483 finds the first occurrence. If pcre_exec() is called again with just
3484 the remainder of the subject, namely "issipi", it does not match,
3485 because \B is always false at the start of the subject, which is deemed
3486 to be a word boundary. However, if pcre_exec() is passed the entire
3487 string again, but with startoffset set to 4, it finds the second occur-
3488 rence of "iss" because it is able to look behind the starting point to
3489 discover that it is preceded by a letter.
3491 Finding all the matches in a subject is tricky when the pattern can
3492 match an empty string. It is possible to emulate Perl's /g behaviour by
3493 first trying the match again at the same offset, with the
3494 PCRE_NOTEMPTY_ATSTART and PCRE_ANCHORED options, and then if that
3495 fails, advancing the starting offset and trying an ordinary match
3496 again. There is some code that demonstrates how to do this in the pcre-
3497 demo sample program. In the most general case, you have to check to see
3498 if the newline convention recognizes CRLF as a newline, and if so, and
3499 the current character is CR followed by LF, advance the starting offset
3500 by two characters instead of one.
3502 If a non-zero starting offset is passed when the pattern is anchored,
3503 one attempt to match at the given offset is made. This can only succeed
3504 if the pattern does not require the match to be at the start of the
3507 How pcre_exec() returns captured substrings
3509 In general, a pattern matches a certain portion of the subject, and in
3510 addition, further substrings from the subject may be picked out by
3511 parts of the pattern. Following the usage in Jeffrey Friedl's book,
3512 this is called "capturing" in what follows, and the phrase "capturing
3513 subpattern" is used for a fragment of a pattern that picks out a sub-
3514 string. PCRE supports several other kinds of parenthesized subpattern
3515 that do not cause substrings to be captured.
3517 Captured substrings are returned to the caller via a vector of integers
3518 whose address is passed in ovector. The number of elements in the vec-
3519 tor is passed in ovecsize, which must be a non-negative number. Note:
3520 this argument is NOT the size of ovector in bytes.
3522 The first two-thirds of the vector is used to pass back captured sub-
3523 strings, each substring using a pair of integers. The remaining third
3524 of the vector is used as workspace by pcre_exec() while matching cap-
3525 turing subpatterns, and is not available for passing back information.
3526 The number passed in ovecsize should always be a multiple of three. If
3527 it is not, it is rounded down.
3529 When a match is successful, information about captured substrings is
3530 returned in pairs of integers, starting at the beginning of ovector,
3531 and continuing up to two-thirds of its length at the most. The first
3532 element of each pair is set to the offset of the first character in a
3533 substring, and the second is set to the offset of the first character
3534 after the end of a substring. These values are always data unit off-
3535 sets, even in UTF mode. They are byte offsets in the 8-bit library,
3536 16-bit data item offsets in the 16-bit library, and 32-bit data item
3537 offsets in the 32-bit library. Note: they are not character counts.
3539 The first pair of integers, ovector[0] and ovector[1], identify the
3540 portion of the subject string matched by the entire pattern. The next
3541 pair is used for the first capturing subpattern, and so on. The value
3542 returned by pcre_exec() is one more than the highest numbered pair that
3543 has been set. For example, if two substrings have been captured, the
3544 returned value is 3. If there are no capturing subpatterns, the return
3545 value from a successful match is 1, indicating that just the first pair
3546 of offsets has been set.
3548 If a capturing subpattern is matched repeatedly, it is the last portion
3549 of the string that it matched that is returned.
3551 If the vector is too small to hold all the captured substring offsets,
3552 it is used as far as possible (up to two-thirds of its length), and the
3553 function returns a value of zero. If neither the actual string matched
3554 nor any captured substrings are of interest, pcre_exec() may be called
3555 with ovector passed as NULL and ovecsize as zero. However, if the pat-
3556 tern contains back references and the ovector is not big enough to
3557 remember the related substrings, PCRE has to get additional memory for
3558 use during matching. Thus it is usually advisable to supply an ovector
3561 There are some cases where zero is returned (indicating vector over-
3562 flow) when in fact the vector is exactly the right size for the final
3563 match. For example, consider the pattern
3567 If a vector of 6 elements (allowing for only 1 captured substring) is
3568 given with subject string "abd", pcre_exec() will try to set the second
3569 captured string, thereby recording a vector overflow, before failing to
3570 match "c" and backing up to try the second alternative. The zero
3571 return, however, does correctly indicate that the maximum number of
3572 slots (namely 2) have been filled. In similar cases where there is tem-
3573 porary overflow, but the final number of used slots is actually less
3574 than the maximum, a non-zero value is returned.
3576 The pcre_fullinfo() function can be used to find out how many capturing
3577 subpatterns there are in a compiled pattern. The smallest size for
3578 ovector that will allow for n captured substrings, in addition to the
3579 offsets of the substring matched by the whole pattern, is (n+1)*3.
3581 It is possible for capturing subpattern number n+1 to match some part
3582 of the subject when subpattern n has not been used at all. For example,
3583 if the string "abc" is matched against the pattern (a|(z))(bc) the
3584 return from the function is 4, and subpatterns 1 and 3 are matched, but
3585 2 is not. When this happens, both values in the offset pairs corre-
3586 sponding to unused subpatterns are set to -1.
3588 Offset values that correspond to unused subpatterns at the end of the
3589 expression are also set to -1. For example, if the string "abc" is
3590 matched against the pattern (abc)(x(yz)?)? subpatterns 2 and 3 are not
3591 matched. The return from the function is 2, because the highest used
3592 capturing subpattern number is 1, and the offsets for for the second
3593 and third capturing subpatterns (assuming the vector is large enough,
3594 of course) are set to -1.
3596 Note: Elements in the first two-thirds of ovector that do not corre-
3597 spond to capturing parentheses in the pattern are never changed. That
3598 is, if a pattern contains n capturing parentheses, no more than ovec-
3599 tor[0] to ovector[2n+1] are set by pcre_exec(). The other elements (in
3600 the first two-thirds) retain whatever values they previously had.
3602 Some convenience functions are provided for extracting the captured
3603 substrings as separate strings. These are described below.
3605 Error return values from pcre_exec()
3607 If pcre_exec() fails, it returns a negative number. The following are
3608 defined in the header file:
3610 PCRE_ERROR_NOMATCH (-1)
3612 The subject string did not match the pattern.
3614 PCRE_ERROR_NULL (-2)
3616 Either code or subject was passed as NULL, or ovector was NULL and
3617 ovecsize was not zero.
3619 PCRE_ERROR_BADOPTION (-3)
3621 An unrecognized bit was set in the options argument.
3623 PCRE_ERROR_BADMAGIC (-4)
3625 PCRE stores a 4-byte "magic number" at the start of the compiled code,
3626 to catch the case when it is passed a junk pointer and to detect when a
3627 pattern that was compiled in an environment of one endianness is run in
3628 an environment with the other endianness. This is the error that PCRE
3629 gives when the magic number is not present.
3631 PCRE_ERROR_UNKNOWN_OPCODE (-5)
3633 While running the pattern match, an unknown item was encountered in the
3634 compiled pattern. This error could be caused by a bug in PCRE or by
3635 overwriting of the compiled pattern.
3637 PCRE_ERROR_NOMEMORY (-6)
3639 If a pattern contains back references, but the ovector that is passed
3640 to pcre_exec() is not big enough to remember the referenced substrings,
3641 PCRE gets a block of memory at the start of matching to use for this
3642 purpose. If the call via pcre_malloc() fails, this error is given. The
3643 memory is automatically freed at the end of matching.
3645 This error is also given if pcre_stack_malloc() fails in pcre_exec().
3646 This can happen only when PCRE has been compiled with --disable-stack-
3649 PCRE_ERROR_NOSUBSTRING (-7)
3651 This error is used by the pcre_copy_substring(), pcre_get_substring(),
3652 and pcre_get_substring_list() functions (see below). It is never
3653 returned by pcre_exec().
3655 PCRE_ERROR_MATCHLIMIT (-8)
3657 The backtracking limit, as specified by the match_limit field in a
3658 pcre_extra structure (or defaulted) was reached. See the description
3661 PCRE_ERROR_CALLOUT (-9)
3663 This error is never generated by pcre_exec() itself. It is provided for
3664 use by callout functions that want to yield a distinctive error code.
3665 See the pcrecallout documentation for details.
3667 PCRE_ERROR_BADUTF8 (-10)
3669 A string that contains an invalid UTF-8 byte sequence was passed as a
3670 subject, and the PCRE_NO_UTF8_CHECK option was not set. If the size of
3671 the output vector (ovecsize) is at least 2, the byte offset to the
3672 start of the the invalid UTF-8 character is placed in the first ele-
3673 ment, and a reason code is placed in the second element. The reason
3674 codes are listed in the following section. For backward compatibility,
3675 if PCRE_PARTIAL_HARD is set and the problem is a truncated UTF-8 char-
3676 acter at the end of the subject (reason codes 1 to 5),
3677 PCRE_ERROR_SHORTUTF8 is returned instead of PCRE_ERROR_BADUTF8.
3679 PCRE_ERROR_BADUTF8_OFFSET (-11)
3681 The UTF-8 byte sequence that was passed as a subject was checked and
3682 found to be valid (the PCRE_NO_UTF8_CHECK option was not set), but the
3683 value of startoffset did not point to the beginning of a UTF-8 charac-
3684 ter or the end of the subject.
3686 PCRE_ERROR_PARTIAL (-12)
3688 The subject string did not match, but it did match partially. See the
3689 pcrepartial documentation for details of partial matching.
3691 PCRE_ERROR_BADPARTIAL (-13)
3693 This code is no longer in use. It was formerly returned when the
3694 PCRE_PARTIAL option was used with a compiled pattern containing items
3695 that were not supported for partial matching. From release 8.00
3696 onwards, there are no restrictions on partial matching.
3698 PCRE_ERROR_INTERNAL (-14)
3700 An unexpected internal error has occurred. This error could be caused
3701 by a bug in PCRE or by overwriting of the compiled pattern.
3703 PCRE_ERROR_BADCOUNT (-15)
3705 This error is given if the value of the ovecsize argument is negative.
3707 PCRE_ERROR_RECURSIONLIMIT (-21)
3709 The internal recursion limit, as specified by the match_limit_recursion
3710 field in a pcre_extra structure (or defaulted) was reached. See the
3713 PCRE_ERROR_BADNEWLINE (-23)
3715 An invalid combination of PCRE_NEWLINE_xxx options was given.
3717 PCRE_ERROR_BADOFFSET (-24)
3719 The value of startoffset was negative or greater than the length of the
3720 subject, that is, the value in length.
3722 PCRE_ERROR_SHORTUTF8 (-25)
3724 This error is returned instead of PCRE_ERROR_BADUTF8 when the subject
3725 string ends with a truncated UTF-8 character and the PCRE_PARTIAL_HARD
3726 option is set. Information about the failure is returned as for
3727 PCRE_ERROR_BADUTF8. It is in fact sufficient to detect this case, but
3728 this special error code for PCRE_PARTIAL_HARD precedes the implementa-
3729 tion of returned information; it is retained for backwards compatibil-
3732 PCRE_ERROR_RECURSELOOP (-26)
3734 This error is returned when pcre_exec() detects a recursion loop within
3735 the pattern. Specifically, it means that either the whole pattern or a
3736 subpattern has been called recursively for the second time at the same
3737 position in the subject string. Some simple patterns that might do this
3738 are detected and faulted at compile time, but more complicated cases,
3739 in particular mutual recursions between two different subpatterns, can-
3740 not be detected until run time.
3742 PCRE_ERROR_JIT_STACKLIMIT (-27)
3744 This error is returned when a pattern that was successfully studied
3745 using a JIT compile option is being matched, but the memory available
3746 for the just-in-time processing stack is not large enough. See the
3747 pcrejit documentation for more details.
3749 PCRE_ERROR_BADMODE (-28)
3751 This error is given if a pattern that was compiled by the 8-bit library
3752 is passed to a 16-bit or 32-bit library function, or vice versa.
3754 PCRE_ERROR_BADENDIANNESS (-29)
3756 This error is given if a pattern that was compiled and saved is
3757 reloaded on a host with different endianness. The utility function
3758 pcre_pattern_to_host_byte_order() can be used to convert such a pattern
3759 so that it runs on the new host.
3761 PCRE_ERROR_JIT_BADOPTION
3763 This error is returned when a pattern that was successfully studied
3764 using a JIT compile option is being matched, but the matching mode
3765 (partial or complete match) does not correspond to any JIT compilation
3766 mode. When the JIT fast path function is used, this error may be also
3767 given for invalid options. See the pcrejit documentation for more
3770 PCRE_ERROR_BADLENGTH (-32)
3772 This error is given if pcre_exec() is called with a negative value for
3773 the length argument.
3775 Error numbers -16 to -20, -22, and 30 are not used by pcre_exec().
3777 Reason codes for invalid UTF-8 strings
3779 This section applies only to the 8-bit library. The corresponding
3780 information for the 16-bit and 32-bit libraries is given in the pcre16
3783 When pcre_exec() returns either PCRE_ERROR_BADUTF8 or PCRE_ERROR_SHORT-
3784 UTF8, and the size of the output vector (ovecsize) is at least 2, the
3785 offset of the start of the invalid UTF-8 character is placed in the
3786 first output vector element (ovector[0]) and a reason code is placed in
3787 the second element (ovector[1]). The reason codes are given names in
3788 the pcre.h header file:
3796 The string ends with a truncated UTF-8 character; the code specifies
3797 how many bytes are missing (1 to 5). Although RFC 3629 restricts UTF-8
3798 characters to be no longer than 4 bytes, the encoding scheme (origi-
3799 nally defined by RFC 2279) allows for up to 6 bytes, and this is
3800 checked first; hence the possibility of 4 or 5 missing bytes.
3808 The two most significant bits of the 2nd, 3rd, 4th, 5th, or 6th byte of
3809 the character do not have the binary value 0b10 (that is, either the
3810 most significant bit is 0, or the next bit is 1).
3815 A character that is valid by the RFC 2279 rules is either 5 or 6 bytes
3816 long; these code points are excluded by RFC 3629.
3820 A 4-byte character has a value greater than 0x10fff; these code points
3821 are excluded by RFC 3629.
3825 A 3-byte character has a value in the range 0xd800 to 0xdfff; this
3826 range of code points are reserved by RFC 3629 for use with UTF-16, and
3827 so are excluded from UTF-8.
3835 A 2-, 3-, 4-, 5-, or 6-byte character is "overlong", that is, it codes
3836 for a value that can be represented by fewer bytes, which is invalid.
3837 For example, the two bytes 0xc0, 0xae give the value 0x2e, whose cor-
3838 rect coding uses just one byte.
3842 The two most significant bits of the first byte of a character have the
3843 binary value 0b10 (that is, the most significant bit is 1 and the sec-
3844 ond is 0). Such a byte can only validly occur as the second or subse-
3845 quent byte of a multi-byte character.
3849 The first byte of a character has the value 0xfe or 0xff. These values
3850 can never occur in a valid UTF-8 string.
3854 This error code was formerly used when the presence of a so-called
3855 "non-character" caused an error. Unicode corrigendum #9 makes it clear
3856 that such characters should not cause a string to be rejected, and so
3857 this code is no longer in use and is never returned.
3860 EXTRACTING CAPTURED SUBSTRINGS BY NUMBER
3862 int pcre_copy_substring(const char *subject, int *ovector,
3863 int stringcount, int stringnumber, char *buffer,
3866 int pcre_get_substring(const char *subject, int *ovector,
3867 int stringcount, int stringnumber,
3868 const char **stringptr);
3870 int pcre_get_substring_list(const char *subject,
3871 int *ovector, int stringcount, const char ***listptr);
3873 Captured substrings can be accessed directly by using the offsets
3874 returned by pcre_exec() in ovector. For convenience, the functions
3875 pcre_copy_substring(), pcre_get_substring(), and pcre_get_sub-
3876 string_list() are provided for extracting captured substrings as new,
3877 separate, zero-terminated strings. These functions identify substrings
3878 by number. The next section describes functions for extracting named
3881 A substring that contains a binary zero is correctly extracted and has
3882 a further zero added on the end, but the result is not, of course, a C
3883 string. However, you can process such a string by referring to the
3884 length that is returned by pcre_copy_substring() and pcre_get_sub-
3885 string(). Unfortunately, the interface to pcre_get_substring_list() is
3886 not adequate for handling strings containing binary zeros, because the
3887 end of the final string is not independently indicated.
3889 The first three arguments are the same for all three of these func-
3890 tions: subject is the subject string that has just been successfully
3891 matched, ovector is a pointer to the vector of integer offsets that was
3892 passed to pcre_exec(), and stringcount is the number of substrings that
3893 were captured by the match, including the substring that matched the
3894 entire regular expression. This is the value returned by pcre_exec() if
3895 it is greater than zero. If pcre_exec() returned zero, indicating that
3896 it ran out of space in ovector, the value passed as stringcount should
3897 be the number of elements in the vector divided by three.
3899 The functions pcre_copy_substring() and pcre_get_substring() extract a
3900 single substring, whose number is given as stringnumber. A value of
3901 zero extracts the substring that matched the entire pattern, whereas
3902 higher values extract the captured substrings. For pcre_copy_sub-
3903 string(), the string is placed in buffer, whose length is given by
3904 buffersize, while for pcre_get_substring() a new block of memory is
3905 obtained via pcre_malloc, and its address is returned via stringptr.
3906 The yield of the function is the length of the string, not including
3907 the terminating zero, or one of these error codes:
3909 PCRE_ERROR_NOMEMORY (-6)
3911 The buffer was too small for pcre_copy_substring(), or the attempt to
3912 get memory failed for pcre_get_substring().
3914 PCRE_ERROR_NOSUBSTRING (-7)
3916 There is no substring whose number is stringnumber.
3918 The pcre_get_substring_list() function extracts all available sub-
3919 strings and builds a list of pointers to them. All this is done in a
3920 single block of memory that is obtained via pcre_malloc. The address of
3921 the memory block is returned via listptr, which is also the start of
3922 the list of string pointers. The end of the list is marked by a NULL
3923 pointer. The yield of the function is zero if all went well, or the
3926 PCRE_ERROR_NOMEMORY (-6)
3928 if the attempt to get the memory block failed.
3930 When any of these functions encounter a substring that is unset, which
3931 can happen when capturing subpattern number n+1 matches some part of
3932 the subject, but subpattern n has not been used at all, they return an
3933 empty string. This can be distinguished from a genuine zero-length sub-
3934 string by inspecting the appropriate offset in ovector, which is nega-
3935 tive for unset substrings.
3937 The two convenience functions pcre_free_substring() and pcre_free_sub-
3938 string_list() can be used to free the memory returned by a previous
3939 call of pcre_get_substring() or pcre_get_substring_list(), respec-
3940 tively. They do nothing more than call the function pointed to by
3941 pcre_free, which of course could be called directly from a C program.
3942 However, PCRE is used in some situations where it is linked via a spe-
3943 cial interface to another programming language that cannot use
3944 pcre_free directly; it is for these cases that the functions are pro-
3948 EXTRACTING CAPTURED SUBSTRINGS BY NAME
3950 int pcre_get_stringnumber(const pcre *code,
3953 int pcre_copy_named_substring(const pcre *code,
3954 const char *subject, int *ovector,
3955 int stringcount, const char *stringname,
3956 char *buffer, int buffersize);
3958 int pcre_get_named_substring(const pcre *code,
3959 const char *subject, int *ovector,
3960 int stringcount, const char *stringname,
3961 const char **stringptr);
3963 To extract a substring by name, you first have to find associated num-
3964 ber. For example, for this pattern
3968 the number of the subpattern called "xxx" is 2. If the name is known to
3969 be unique (PCRE_DUPNAMES was not set), you can find the number from the
3970 name by calling pcre_get_stringnumber(). The first argument is the com-
3971 piled pattern, and the second is the name. The yield of the function is
3972 the subpattern number, or PCRE_ERROR_NOSUBSTRING (-7) if there is no
3973 subpattern of that name.
3975 Given the number, you can extract the substring directly, or use one of
3976 the functions described in the previous section. For convenience, there
3977 are also two functions that do the whole job.
3979 Most of the arguments of pcre_copy_named_substring() and
3980 pcre_get_named_substring() are the same as those for the similarly
3981 named functions that extract by number. As these are described in the
3982 previous section, they are not re-described here. There are just two
3985 First, instead of a substring number, a substring name is given. Sec-
3986 ond, there is an extra argument, given at the start, which is a pointer
3987 to the compiled pattern. This is needed in order to gain access to the
3988 name-to-number translation table.
3990 These functions call pcre_get_stringnumber(), and if it succeeds, they
3991 then call pcre_copy_substring() or pcre_get_substring(), as appropri-
3992 ate. NOTE: If PCRE_DUPNAMES is set and there are duplicate names, the
3993 behaviour may not be what you want (see the next section).
3995 Warning: If the pattern uses the (?| feature to set up multiple subpat-
3996 terns with the same number, as described in the section on duplicate
3997 subpattern numbers in the pcrepattern page, you cannot use names to
3998 distinguish the different subpatterns, because names are not included
3999 in the compiled code. The matching process uses only numbers. For this
4000 reason, the use of different names for subpatterns of the same number
4001 causes an error at compile time.
4004 DUPLICATE SUBPATTERN NAMES
4006 int pcre_get_stringtable_entries(const pcre *code,
4007 const char *name, char **first, char **last);
4009 When a pattern is compiled with the PCRE_DUPNAMES option, names for
4010 subpatterns are not required to be unique. (Duplicate names are always
4011 allowed for subpatterns with the same number, created by using the (?|
4012 feature. Indeed, if such subpatterns are named, they are required to
4013 use the same names.)
4015 Normally, patterns with duplicate names are such that in any one match,
4016 only one of the named subpatterns participates. An example is shown in
4017 the pcrepattern documentation.
4019 When duplicates are present, pcre_copy_named_substring() and
4020 pcre_get_named_substring() return the first substring corresponding to
4021 the given name that is set. If none are set, PCRE_ERROR_NOSUBSTRING
4022 (-7) is returned; no data is returned. The pcre_get_stringnumber()
4023 function returns one of the numbers that are associated with the name,
4024 but it is not defined which it is.
4026 If you want to get full details of all captured substrings for a given
4027 name, you must use the pcre_get_stringtable_entries() function. The
4028 first argument is the compiled pattern, and the second is the name. The
4029 third and fourth are pointers to variables which are updated by the
4030 function. After it has run, they point to the first and last entries in
4031 the name-to-number table for the given name. The function itself
4032 returns the length of each entry, or PCRE_ERROR_NOSUBSTRING (-7) if
4033 there are none. The format of the table is described above in the sec-
4034 tion entitled Information about a pattern above. Given all the rele-
4035 vant entries for the name, you can extract each of their numbers, and
4036 hence the captured data, if any.
4039 FINDING ALL POSSIBLE MATCHES
4041 The traditional matching function uses a similar algorithm to Perl,
4042 which stops when it finds the first match, starting at a given point in
4043 the subject. If you want to find all possible matches, or the longest
4044 possible match, consider using the alternative matching function (see
4045 below) instead. If you cannot use the alternative function, but still
4046 need to find all possible matches, you can kludge it up by making use
4047 of the callout facility, which is described in the pcrecallout documen-
4050 What you have to do is to insert a callout right at the end of the pat-
4051 tern. When your callout function is called, extract and save the cur-
4052 rent matched substring. Then return 1, which forces pcre_exec() to
4053 backtrack and try other alternatives. Ultimately, when it runs out of
4054 matches, pcre_exec() will yield PCRE_ERROR_NOMATCH.
4057 OBTAINING AN ESTIMATE OF STACK USAGE
4059 Matching certain patterns using pcre_exec() can use a lot of process
4060 stack, which in certain environments can be rather limited in size.
4061 Some users find it helpful to have an estimate of the amount of stack
4062 that is used by pcre_exec(), to help them set recursion limits, as
4063 described in the pcrestack documentation. The estimate that is output
4064 by pcretest when called with the -m and -C options is obtained by call-
4065 ing pcre_exec with the values NULL, NULL, NULL, -999, and -999 for its
4066 first five arguments.
4068 Normally, if its first argument is NULL, pcre_exec() immediately
4069 returns the negative error code PCRE_ERROR_NULL, but with this special
4070 combination of arguments, it returns instead a negative number whose
4071 absolute value is the approximate stack frame size in bytes. (A nega-
4072 tive number is used so that it is clear that no match has happened.)
4073 The value is approximate because in some cases, recursive calls to
4074 pcre_exec() occur when there are one or two additional variables on the
4077 If PCRE has been compiled to use the heap instead of the stack for
4078 recursion, the value returned is the size of each block that is
4079 obtained from the heap.
4082 MATCHING A PATTERN: THE ALTERNATIVE FUNCTION
4084 int pcre_dfa_exec(const pcre *code, const pcre_extra *extra,
4085 const char *subject, int length, int startoffset,
4086 int options, int *ovector, int ovecsize,
4087 int *workspace, int wscount);
4089 The function pcre_dfa_exec() is called to match a subject string
4090 against a compiled pattern, using a matching algorithm that scans the
4091 subject string just once, and does not backtrack. This has different
4092 characteristics to the normal algorithm, and is not compatible with
4093 Perl. Some of the features of PCRE patterns are not supported. Never-
4094 theless, there are times when this kind of matching can be useful. For
4095 a discussion of the two matching algorithms, and a list of features
4096 that pcre_dfa_exec() does not support, see the pcrematching documenta-
4099 The arguments for the pcre_dfa_exec() function are the same as for
4100 pcre_exec(), plus two extras. The ovector argument is used in a differ-
4101 ent way, and this is described below. The other common arguments are
4102 used in the same way as for pcre_exec(), so their description is not
4105 The two additional arguments provide workspace for the function. The
4106 workspace vector should contain at least 20 elements. It is used for
4107 keeping track of multiple paths through the pattern tree. More
4108 workspace will be needed for patterns and subjects where there are a
4109 lot of potential matches.
4111 Here is an example of a simple call to pcre_dfa_exec():
4117 re, /* result of pcre_compile() */
4118 NULL, /* we didn't study the pattern */
4119 "some string", /* the subject string */
4120 11, /* the length of the subject string */
4121 0, /* start at offset 0 in the subject */
4122 0, /* default options */
4123 ovector, /* vector of integers for substring information */
4124 10, /* number of elements (NOT size in bytes) */
4125 wspace, /* working space vector */
4126 20); /* number of elements (NOT size in bytes) */
4128 Option bits for pcre_dfa_exec()
4130 The unused bits of the options argument for pcre_dfa_exec() must be
4131 zero. The only bits that may be set are PCRE_ANCHORED, PCRE_NEW-
4132 LINE_xxx, PCRE_NOTBOL, PCRE_NOTEOL, PCRE_NOTEMPTY,
4133 PCRE_NOTEMPTY_ATSTART, PCRE_NO_UTF8_CHECK, PCRE_BSR_ANYCRLF,
4134 PCRE_BSR_UNICODE, PCRE_NO_START_OPTIMIZE, PCRE_PARTIAL_HARD, PCRE_PAR-
4135 TIAL_SOFT, PCRE_DFA_SHORTEST, and PCRE_DFA_RESTART. All but the last
4136 four of these are exactly the same as for pcre_exec(), so their
4137 description is not repeated here.
4142 These have the same general effect as they do for pcre_exec(), but the
4143 details are slightly different. When PCRE_PARTIAL_HARD is set for
4144 pcre_dfa_exec(), it returns PCRE_ERROR_PARTIAL if the end of the sub-
4145 ject is reached and there is still at least one matching possibility
4146 that requires additional characters. This happens even if some complete
4147 matches have also been found. When PCRE_PARTIAL_SOFT is set, the return
4148 code PCRE_ERROR_NOMATCH is converted into PCRE_ERROR_PARTIAL if the end
4149 of the subject is reached, there have been no complete matches, but
4150 there is still at least one matching possibility. The portion of the
4151 string that was inspected when the longest partial match was found is
4152 set as the first matching string in both cases. There is a more
4153 detailed discussion of partial and multi-segment matching, with exam-
4154 ples, in the pcrepartial documentation.
4158 Setting the PCRE_DFA_SHORTEST option causes the matching algorithm to
4159 stop as soon as it has found one match. Because of the way the alterna-
4160 tive algorithm works, this is necessarily the shortest possible match
4161 at the first possible matching point in the subject string.
4165 When pcre_dfa_exec() returns a partial match, it is possible to call it
4166 again, with additional subject characters, and have it continue with
4167 the same match. The PCRE_DFA_RESTART option requests this action; when
4168 it is set, the workspace and wscount options must reference the same
4169 vector as before because data about the match so far is left in them
4170 after a partial match. There is more discussion of this facility in the
4171 pcrepartial documentation.
4173 Successful returns from pcre_dfa_exec()
4175 When pcre_dfa_exec() succeeds, it may have matched more than one sub-
4176 string in the subject. Note, however, that all the matches from one run
4177 of the function start at the same point in the subject. The shorter
4178 matches are all initial substrings of the longer matches. For example,
4183 is matched against the string
4185 This is <something> <something else> <something further> no more
4187 the three matched strings are
4190 <something> <something else>
4191 <something> <something else> <something further>
4193 On success, the yield of the function is a number greater than zero,
4194 which is the number of matched substrings. The substrings themselves
4195 are returned in ovector. Each string uses two elements; the first is
4196 the offset to the start, and the second is the offset to the end. In
4197 fact, all the strings have the same start offset. (Space could have
4198 been saved by giving this only once, but it was decided to retain some
4199 compatibility with the way pcre_exec() returns data, even though the
4200 meaning of the strings is different.)
4202 The strings are returned in reverse order of length; that is, the long-
4203 est matching string is given first. If there were too many matches to
4204 fit into ovector, the yield of the function is zero, and the vector is
4205 filled with the longest matches. Unlike pcre_exec(), pcre_dfa_exec()
4206 can use the entire ovector for returning matched strings.
4208 NOTE: PCRE's "auto-possessification" optimization usually applies to
4209 character repeats at the end of a pattern (as well as internally). For
4210 example, the pattern "a\d+" is compiled as if it were "a\d++" because
4211 there is no point even considering the possibility of backtracking into
4212 the repeated digits. For DFA matching, this means that only one possi-
4213 ble match is found. If you really do want multiple matches in such
4214 cases, either use an ungreedy repeat ("a\d+?") or set the
4215 PCRE_NO_AUTO_POSSESS option when compiling.
4217 Error returns from pcre_dfa_exec()
4219 The pcre_dfa_exec() function returns a negative number when it fails.
4220 Many of the errors are the same as for pcre_exec(), and these are
4221 described above. There are in addition the following errors that are
4222 specific to pcre_dfa_exec():
4224 PCRE_ERROR_DFA_UITEM (-16)
4226 This return is given if pcre_dfa_exec() encounters an item in the pat-
4227 tern that it does not support, for instance, the use of \C or a back
4230 PCRE_ERROR_DFA_UCOND (-17)
4232 This return is given if pcre_dfa_exec() encounters a condition item
4233 that uses a back reference for the condition, or a test for recursion
4234 in a specific group. These are not supported.
4236 PCRE_ERROR_DFA_UMLIMIT (-18)
4238 This return is given if pcre_dfa_exec() is called with an extra block
4239 that contains a setting of the match_limit or match_limit_recursion
4240 fields. This is not supported (these fields are meaningless for DFA
4243 PCRE_ERROR_DFA_WSSIZE (-19)
4245 This return is given if pcre_dfa_exec() runs out of space in the
4248 PCRE_ERROR_DFA_RECURSE (-20)
4250 When a recursive subpattern is processed, the matching function calls
4251 itself recursively, using private vectors for ovector and workspace.
4252 This error is given if the output vector is not large enough. This
4253 should be extremely rare, as a vector of size 1000 is used.
4255 PCRE_ERROR_DFA_BADRESTART (-30)
4257 When pcre_dfa_exec() is called with the PCRE_DFA_RESTART option, some
4258 plausibility checks are made on the contents of the workspace, which
4259 should contain data about the previous partial match. If any of these
4260 checks fail, this error is given.
4265 pcre16(3), pcre32(3), pcrebuild(3), pcrecallout(3), pcrecpp(3)(3),
4266 pcrematching(3), pcrepartial(3), pcreposix(3), pcreprecompile(3), pcre-
4267 sample(3), pcrestack(3).
4273 University Computing Service
4274 Cambridge CB2 3QH, England.
4279 Last updated: 09 February 2014
4280 Copyright (c) 1997-2014 University of Cambridge.
4281 ------------------------------------------------------------------------------
4284 PCRECALLOUT(3) Library Functions Manual PCRECALLOUT(3)
4289 PCRE - Perl-compatible regular expressions
4295 int (*pcre_callout)(pcre_callout_block *);
4297 int (*pcre16_callout)(pcre16_callout_block *);
4299 int (*pcre32_callout)(pcre32_callout_block *);
4304 PCRE provides a feature called "callout", which is a means of temporar-
4305 ily passing control to the caller of PCRE in the middle of pattern
4306 matching. The caller of PCRE provides an external function by putting
4307 its entry point in the global variable pcre_callout (pcre16_callout for
4308 the 16-bit library, pcre32_callout for the 32-bit library). By default,
4309 this variable contains NULL, which disables all calling out.
4311 Within a regular expression, (?C) indicates the points at which the
4312 external function is to be called. Different callout points can be
4313 identified by putting a number less than 256 after the letter C. The
4314 default value is zero. For example, this pattern has two callout
4319 If the PCRE_AUTO_CALLOUT option bit is set when a pattern is compiled,
4320 PCRE automatically inserts callouts, all with number 255, before each
4321 item in the pattern. For example, if PCRE_AUTO_CALLOUT is used with the
4326 it is processed as if it were
4328 (?C255)A(?C255)((?C255)\d{2}(?C255)|(?C255)-(?C255)-(?C255))(?C255)
4330 Notice that there is a callout before and after each parenthesis and
4331 alternation bar. If the pattern contains a conditional group whose con-
4332 dition is an assertion, an automatic callout is inserted immediately
4333 before the condition. Such a callout may also be inserted explicitly,
4338 This applies only to assertion conditions (because they are themselves
4339 independent groups).
4341 Automatic callouts can be used for tracking the progress of pattern
4342 matching. The pcretest program has a pattern qualifier (/C) that sets
4343 automatic callouts; when it is used, the output indicates how the pat-
4344 tern is being matched. This is useful information when you are trying
4345 to optimize the performance of a particular pattern.
4350 You should be aware that, because of optimizations in the way PCRE com-
4351 piles and matches patterns, callouts sometimes do not happen exactly as
4354 At compile time, PCRE "auto-possessifies" repeated items when it knows
4355 that what follows cannot be part of the repeat. For example, a+[bc] is
4356 compiled as if it were a++[bc]. The pcretest output when this pattern
4357 is anchored and then applied with automatic callouts to the string
4366 This indicates that when matching [bc] fails, there is no backtracking
4367 into a+ and therefore the callouts that would be taken for the back-
4368 tracks do not occur. You can disable the auto-possessify feature by
4369 passing PCRE_NO_AUTO_POSSESS to pcre_compile(), or starting the pattern
4370 with (*NO_AUTO_POSSESS). If this is done in pcretest (using the /O
4371 qualifier), the output changes to this:
4382 This time, when matching [bc] fails, the matcher backtracks into a+ and
4383 tries again, repeatedly, until a+ itself fails.
4385 Other optimizations that provide fast "no match" results also affect
4386 callouts. For example, if the pattern is
4390 PCRE knows that any matching string must contain the letter "d". If the
4391 subject string is "abyz", the lack of "d" means that matching doesn't
4392 ever start, and the callout is never reached. However, with "abyd",
4393 though the result is still no match, the callout is obeyed.
4395 If the pattern is studied, PCRE knows the minimum length of a matching
4396 string, and will immediately give a "no match" return without actually
4397 running a match if the subject is not long enough, or, for unanchored
4398 patterns, if it has been scanned far enough.
4400 You can disable these optimizations by passing the PCRE_NO_START_OPTI-
4401 MIZE option to the matching function, or by starting the pattern with
4402 (*NO_START_OPT). This slows down the matching process, but does ensure
4403 that callouts such as the example above are obeyed.
4406 THE CALLOUT INTERFACE
4408 During matching, when PCRE reaches a callout point, the external func-
4409 tion defined by pcre_callout or pcre[16|32]_callout is called (if it is
4410 set). This applies to both normal and DFA matching. The only argument
4411 to the callout function is a pointer to a pcre_callout or
4412 pcre[16|32]_callout block. These structures contains the following
4418 const char *subject; (8-bit version)
4419 PCRE_SPTR16 subject; (16-bit version)
4420 PCRE_SPTR32 subject; (32-bit version)
4423 int current_position;
4427 int pattern_position;
4428 int next_item_length;
4429 const unsigned char *mark; (8-bit version)
4430 const PCRE_UCHAR16 *mark; (16-bit version)
4431 const PCRE_UCHAR32 *mark; (32-bit version)
4433 The version field is an integer containing the version number of the
4434 block format. The initial version was 0; the current version is 2. The
4435 version number will change again in future if additional fields are
4436 added, but the intention is never to remove any of the existing fields.
4438 The callout_number field contains the number of the callout, as com-
4439 piled into the pattern (that is, the number after ?C for manual call-
4440 outs, and 255 for automatically generated callouts).
4442 The offset_vector field is a pointer to the vector of offsets that was
4443 passed by the caller to the matching function. When pcre_exec() or
4444 pcre[16|32]_exec() is used, the contents can be inspected, in order to
4445 extract substrings that have been matched so far, in the same way as
4446 for extracting substrings after a match has completed. For the DFA
4447 matching functions, this field is not useful.
4449 The subject and subject_length fields contain copies of the values that
4450 were passed to the matching function.
4452 The start_match field normally contains the offset within the subject
4453 at which the current match attempt started. However, if the escape
4454 sequence \K has been encountered, this value is changed to reflect the
4455 modified starting point. If the pattern is not anchored, the callout
4456 function may be called several times from the same point in the pattern
4457 for different starting points in the subject.
4459 The current_position field contains the offset within the subject of
4460 the current match pointer.
4462 When the pcre_exec() or pcre[16|32]_exec() is used, the capture_top
4463 field contains one more than the number of the highest numbered cap-
4464 tured substring so far. If no substrings have been captured, the value
4465 of capture_top is one. This is always the case when the DFA functions
4466 are used, because they do not support captured substrings.
4468 The capture_last field contains the number of the most recently cap-
4469 tured substring. However, when a recursion exits, the value reverts to
4470 what it was outside the recursion, as do the values of all captured
4471 substrings. If no substrings have been captured, the value of cap-
4472 ture_last is -1. This is always the case for the DFA matching func-
4475 The callout_data field contains a value that is passed to a matching
4476 function specifically so that it can be passed back in callouts. It is
4477 passed in the callout_data field of a pcre_extra or pcre[16|32]_extra
4478 data structure. If no such data was passed, the value of callout_data
4479 in a callout block is NULL. There is a description of the pcre_extra
4480 structure in the pcreapi documentation.
4482 The pattern_position field is present from version 1 of the callout
4483 structure. It contains the offset to the next item to be matched in the
4486 The next_item_length field is present from version 1 of the callout
4487 structure. It contains the length of the next item to be matched in the
4488 pattern string. When the callout immediately precedes an alternation
4489 bar, a closing parenthesis, or the end of the pattern, the length is
4490 zero. When the callout precedes an opening parenthesis, the length is
4491 that of the entire subpattern.
4493 The pattern_position and next_item_length fields are intended to help
4494 in distinguishing between different automatic callouts, which all have
4495 the same callout number. However, they are set for all callouts.
4497 The mark field is present from version 2 of the callout structure. In
4498 callouts from pcre_exec() or pcre[16|32]_exec() it contains a pointer
4499 to the zero-terminated name of the most recently passed (*MARK),
4500 (*PRUNE), or (*THEN) item in the match, or NULL if no such items have
4501 been passed. Instances of (*PRUNE) or (*THEN) without a name do not
4502 obliterate a previous (*MARK). In callouts from the DFA matching func-
4503 tions this field always contains NULL.
4508 The external callout function returns an integer to PCRE. If the value
4509 is zero, matching proceeds as normal. If the value is greater than
4510 zero, matching fails at the current point, but the testing of other
4511 matching possibilities goes ahead, just as if a lookahead assertion had
4512 failed. If the value is less than zero, the match is abandoned, the
4513 matching function returns the negative value.
4515 Negative values should normally be chosen from the set of
4516 PCRE_ERROR_xxx values. In particular, PCRE_ERROR_NOMATCH forces a stan-
4517 dard "no match" failure. The error number PCRE_ERROR_CALLOUT is
4518 reserved for use by callout functions; it will never be used by PCRE
4525 University Computing Service
4526 Cambridge CB2 3QH, England.
4531 Last updated: 12 November 2013
4532 Copyright (c) 1997-2013 University of Cambridge.
4533 ------------------------------------------------------------------------------
4536 PCRECOMPAT(3) Library Functions Manual PCRECOMPAT(3)
4541 PCRE - Perl-compatible regular expressions
4543 DIFFERENCES BETWEEN PCRE AND PERL
4545 This document describes the differences in the ways that PCRE and Perl
4546 handle regular expressions. The differences described here are with
4547 respect to Perl versions 5.10 and above.
4549 1. PCRE has only a subset of Perl's Unicode support. Details of what it
4550 does have are given in the pcreunicode page.
4552 2. PCRE allows repeat quantifiers only on parenthesized assertions, but
4553 they do not mean what you might think. For example, (?!a){3} does not
4554 assert that the next three characters are not "a". It just asserts that
4555 the next character is not "a" three times (in principle: PCRE optimizes
4556 this to run the assertion just once). Perl allows repeat quantifiers on
4557 other assertions such as \b, but these do not seem to have any use.
4559 3. Capturing subpatterns that occur inside negative lookahead asser-
4560 tions are counted, but their entries in the offsets vector are never
4561 set. Perl sometimes (but not always) sets its numerical variables from
4562 inside negative assertions.
4564 4. Though binary zero characters are supported in the subject string,
4565 they are not allowed in a pattern string because it is passed as a nor-
4566 mal C string, terminated by zero. The escape sequence \0 can be used in
4567 the pattern to represent a binary zero.
4569 5. The following Perl escape sequences are not supported: \l, \u, \L,
4570 \U, and \N when followed by a character name or Unicode value. (\N on
4571 its own, matching a non-newline character, is supported.) In fact these
4572 are implemented by Perl's general string-handling and are not part of
4573 its pattern matching engine. If any of these are encountered by PCRE,
4574 an error is generated by default. However, if the PCRE_JAVASCRIPT_COM-
4575 PAT option is set, \U and \u are interpreted as JavaScript interprets
4578 6. The Perl escape sequences \p, \P, and \X are supported only if PCRE
4579 is built with Unicode character property support. The properties that
4580 can be tested with \p and \P are limited to the general category prop-
4581 erties such as Lu and Nd, script names such as Greek or Han, and the
4582 derived properties Any and L&. PCRE does support the Cs (surrogate)
4583 property, which Perl does not; the Perl documentation says "Because
4584 Perl hides the need for the user to understand the internal representa-
4585 tion of Unicode characters, there is no need to implement the somewhat
4586 messy concept of surrogates."
4588 7. PCRE does support the \Q...\E escape for quoting substrings. Charac-
4589 ters in between are treated as literals. This is slightly different
4590 from Perl in that $ and @ are also handled as literals inside the
4591 quotes. In Perl, they cause variable interpolation (but of course PCRE
4592 does not have variables). Note the following examples:
4594 Pattern PCRE matches Perl matches
4596 \Qabc$xyz\E abc$xyz abc followed by the
4598 \Qabc\$xyz\E abc\$xyz abc\$xyz
4599 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4601 The \Q...\E sequence is recognized both inside and outside character
4604 8. Fairly obviously, PCRE does not support the (?{code}) and (??{code})
4605 constructions. However, there is support for recursive patterns. This
4606 is not available in Perl 5.8, but it is in Perl 5.10. Also, the PCRE
4607 "callout" feature allows an external function to be called during pat-
4608 tern matching. See the pcrecallout documentation for details.
4610 9. Subpatterns that are called as subroutines (whether or not recur-
4611 sively) are always treated as atomic groups in PCRE. This is like
4612 Python, but unlike Perl. Captured values that are set outside a sub-
4613 routine call can be reference from inside in PCRE, but not in Perl.
4614 There is a discussion that explains these differences in more detail in
4615 the section on recursion differences from Perl in the pcrepattern page.
4617 10. If any of the backtracking control verbs are used in a subpattern
4618 that is called as a subroutine (whether or not recursively), their
4619 effect is confined to that subpattern; it does not extend to the sur-
4620 rounding pattern. This is not always the case in Perl. In particular,
4621 if (*THEN) is present in a group that is called as a subroutine, its
4622 action is limited to that group, even if the group does not contain any
4623 | characters. Note that such subpatterns are processed as anchored at
4624 the point where they are tested.
4626 11. If a pattern contains more than one backtracking control verb, the
4627 first one that is backtracked onto acts. For example, in the pattern
4628 A(*COMMIT)B(*PRUNE)C a failure in B triggers (*COMMIT), but a failure
4629 in C triggers (*PRUNE). Perl's behaviour is more complex; in many cases
4630 it is the same as PCRE, but there are examples where it differs.
4632 12. Most backtracking verbs in assertions have their normal actions.
4633 They are not confined to the assertion.
4635 13. There are some differences that are concerned with the settings of
4636 captured strings when part of a pattern is repeated. For example,
4637 matching "aba" against the pattern /^(a(b)?)+$/ in Perl leaves $2
4638 unset, but in PCRE it is set to "b".
4640 14. PCRE's handling of duplicate subpattern numbers and duplicate sub-
4641 pattern names is not as general as Perl's. This is a consequence of the
4642 fact the PCRE works internally just with numbers, using an external ta-
4643 ble to translate between numbers and names. In particular, a pattern
4644 such as (?|(?<a>A)|(?<b)B), where the two capturing parentheses have
4645 the same number but different names, is not supported, and causes an
4646 error at compile time. If it were allowed, it would not be possible to
4647 distinguish which parentheses matched, because both names map to cap-
4648 turing subpattern number 1. To avoid this confusing situation, an error
4649 is given at compile time.
4651 15. Perl recognizes comments in some places that PCRE does not, for
4652 example, between the ( and ? at the start of a subpattern. If the /x
4653 modifier is set, Perl allows white space between ( and ? (though cur-
4654 rent Perls warn that this is deprecated) but PCRE never does, even if
4655 the PCRE_EXTENDED option is set.
4657 16. Perl, when in warning mode, gives warnings for character classes
4658 such as [A-\d] or [a-[:digit:]]. It then treats the hyphens as liter-
4659 als. PCRE has no warning features, so it gives an error in these cases
4660 because they are almost certainly user mistakes.
4662 17. In PCRE, the upper/lower case character properties Lu and Ll are
4663 not affected when case-independent matching is specified. For example,
4664 \p{Lu} always matches an upper case letter. I think Perl has changed in
4665 this respect; in the release at the time of writing (5.16), \p{Lu} and
4666 \p{Ll} match all letters, regardless of case, when case independence is
4669 18. PCRE provides some extensions to the Perl regular expression facil-
4670 ities. Perl 5.10 includes new features that are not in earlier ver-
4671 sions of Perl, some of which (such as named parentheses) have been in
4672 PCRE for some time. This list is with respect to Perl 5.10:
4674 (a) Although lookbehind assertions in PCRE must match fixed length
4675 strings, each alternative branch of a lookbehind assertion can match a
4676 different length of string. Perl requires them all to have the same
4679 (b) If PCRE_DOLLAR_ENDONLY is set and PCRE_MULTILINE is not set, the $
4680 meta-character matches only at the very end of the string.
4682 (c) If PCRE_EXTRA is set, a backslash followed by a letter with no spe-
4683 cial meaning is faulted. Otherwise, like Perl, the backslash is quietly
4684 ignored. (Perl can be made to issue a warning.)
4686 (d) If PCRE_UNGREEDY is set, the greediness of the repetition quanti-
4687 fiers is inverted, that is, by default they are not greedy, but if fol-
4688 lowed by a question mark they are.
4690 (e) PCRE_ANCHORED can be used at matching time to force a pattern to be
4691 tried only at the first matching position in the subject string.
4693 (f) The PCRE_NOTBOL, PCRE_NOTEOL, PCRE_NOTEMPTY, PCRE_NOTEMPTY_ATSTART,
4694 and PCRE_NO_AUTO_CAPTURE options for pcre_exec() have no Perl equiva-
4697 (g) The \R escape sequence can be restricted to match only CR, LF, or
4698 CRLF by the PCRE_BSR_ANYCRLF option.
4700 (h) The callout facility is PCRE-specific.
4702 (i) The partial matching facility is PCRE-specific.
4704 (j) Patterns compiled by PCRE can be saved and re-used at a later time,
4705 even on different hosts that have the other endianness. However, this
4706 does not apply to optimized data created by the just-in-time compiler.
4708 (k) The alternative matching functions (pcre_dfa_exec(),
4709 pcre16_dfa_exec() and pcre32_dfa_exec(),) match in a different way and
4710 are not Perl-compatible.
4712 (l) PCRE recognizes some special sequences such as (*CR) at the start
4713 of a pattern that set overall options that cannot be changed within the
4720 University Computing Service
4721 Cambridge CB2 3QH, England.
4726 Last updated: 10 November 2013
4727 Copyright (c) 1997-2013 University of Cambridge.
4728 ------------------------------------------------------------------------------
4731 PCREPATTERN(3) Library Functions Manual PCREPATTERN(3)
4736 PCRE - Perl-compatible regular expressions
4738 PCRE REGULAR EXPRESSION DETAILS
4740 The syntax and semantics of the regular expressions that are supported
4741 by PCRE are described in detail below. There is a quick-reference syn-
4742 tax summary in the pcresyntax page. PCRE tries to match Perl syntax and
4743 semantics as closely as it can. PCRE also supports some alternative
4744 regular expression syntax (which does not conflict with the Perl syn-
4745 tax) in order to provide some compatibility with regular expressions in
4746 Python, .NET, and Oniguruma.
4748 Perl's regular expressions are described in its own documentation, and
4749 regular expressions in general are covered in a number of books, some
4750 of which have copious examples. Jeffrey Friedl's "Mastering Regular
4751 Expressions", published by O'Reilly, covers regular expressions in
4752 great detail. This description of PCRE's regular expressions is
4753 intended as reference material.
4755 This document discusses the patterns that are supported by PCRE when
4756 one its main matching functions, pcre_exec() (8-bit) or
4757 pcre[16|32]_exec() (16- or 32-bit), is used. PCRE also has alternative
4758 matching functions, pcre_dfa_exec() and pcre[16|32_dfa_exec(), which
4759 match using a different algorithm that is not Perl-compatible. Some of
4760 the features discussed below are not available when DFA matching is
4761 used. The advantages and disadvantages of the alternative functions,
4762 and how they differ from the normal functions, are discussed in the
4766 SPECIAL START-OF-PATTERN ITEMS
4768 A number of options that can be passed to pcre_compile() can also be
4769 set by special items at the start of a pattern. These are not Perl-com-
4770 patible, but are provided to make these options accessible to pattern
4771 writers who are not able to change the program that processes the pat-
4772 tern. Any number of these items may appear, but they must all be
4773 together right at the start of the pattern string, and the letters must
4778 The original operation of PCRE was on strings of one-byte characters.
4779 However, there is now also support for UTF-8 strings in the original
4780 library, an extra library that supports 16-bit and UTF-16 character
4781 strings, and a third library that supports 32-bit and UTF-32 character
4782 strings. To use these features, PCRE must be built to include appropri-
4783 ate support. When using UTF strings you must either call the compiling
4784 function with the PCRE_UTF8, PCRE_UTF16, or PCRE_UTF32 option, or the
4785 pattern must start with one of these special sequences:
4792 (*UTF) is a generic sequence that can be used with any of the
4793 libraries. Starting a pattern with such a sequence is equivalent to
4794 setting the relevant option. How setting a UTF mode affects pattern
4795 matching is mentioned in several places below. There is also a summary
4796 of features in the pcreunicode page.
4798 Some applications that allow their users to supply patterns may wish to
4799 restrict them to non-UTF data for security reasons. If the
4800 PCRE_NEVER_UTF option is set at compile time, (*UTF) etc. are not
4801 allowed, and their appearance causes an error.
4803 Unicode property support
4805 Another special sequence that may appear at the start of a pattern is
4806 (*UCP). This has the same effect as setting the PCRE_UCP option: it
4807 causes sequences such as \d and \w to use Unicode properties to deter-
4808 mine character types, instead of recognizing only characters with codes
4809 less than 128 via a lookup table.
4811 Disabling auto-possessification
4813 If a pattern starts with (*NO_AUTO_POSSESS), it has the same effect as
4814 setting the PCRE_NO_AUTO_POSSESS option at compile time. This stops
4815 PCRE from making quantifiers possessive when what follows cannot match
4816 the repeated item. For example, by default a+b is treated as a++b. For
4817 more details, see the pcreapi documentation.
4819 Disabling start-up optimizations
4821 If a pattern starts with (*NO_START_OPT), it has the same effect as
4822 setting the PCRE_NO_START_OPTIMIZE option either at compile or matching
4823 time. This disables several optimizations for quickly reaching "no
4824 match" results. For more details, see the pcreapi documentation.
4828 PCRE supports five different conventions for indicating line breaks in
4829 strings: a single CR (carriage return) character, a single LF (line-
4830 feed) character, the two-character sequence CRLF, any of the three pre-
4831 ceding, or any Unicode newline sequence. The pcreapi page has further
4832 discussion about newlines, and shows how to set the newline convention
4833 in the options arguments for the compiling and matching functions.
4835 It is also possible to specify a newline convention by starting a pat-
4836 tern string with one of the following five sequences:
4838 (*CR) carriage return
4840 (*CRLF) carriage return, followed by linefeed
4841 (*ANYCRLF) any of the three above
4842 (*ANY) all Unicode newline sequences
4844 These override the default and the options given to the compiling func-
4845 tion. For example, on a Unix system where LF is the default newline
4846 sequence, the pattern
4850 changes the convention to CR. That pattern matches "a\nb" because LF is
4851 no longer a newline. If more than one of these settings is present, the
4854 The newline convention affects where the circumflex and dollar asser-
4855 tions are true. It also affects the interpretation of the dot metachar-
4856 acter when PCRE_DOTALL is not set, and the behaviour of \N. However, it
4857 does not affect what the \R escape sequence matches. By default, this
4858 is any Unicode newline sequence, for Perl compatibility. However, this
4859 can be changed; see the description of \R in the section entitled "New-
4860 line sequences" below. A change of \R setting can be combined with a
4861 change of newline convention.
4863 Setting match and recursion limits
4865 The caller of pcre_exec() can set a limit on the number of times the
4866 internal match() function is called and on the maximum depth of recur-
4867 sive calls. These facilities are provided to catch runaway matches that
4868 are provoked by patterns with huge matching trees (a typical example is
4869 a pattern with nested unlimited repeats) and to avoid running out of
4870 system stack by too much recursion. When one of these limits is
4871 reached, pcre_exec() gives an error return. The limits can also be set
4872 by items at the start of the pattern of the form
4875 (*LIMIT_RECURSION=d)
4877 where d is any number of decimal digits. However, the value of the set-
4878 ting must be less than the value set (or defaulted) by the caller of
4879 pcre_exec() for it to have any effect. In other words, the pattern
4880 writer can lower the limits set by the programmer, but not raise them.
4881 If there is more than one setting of one of these limits, the lower
4885 EBCDIC CHARACTER CODES
4887 PCRE can be compiled to run in an environment that uses EBCDIC as its
4888 character code rather than ASCII or Unicode (typically a mainframe sys-
4889 tem). In the sections below, character code values are ASCII or Uni-
4890 code; in an EBCDIC environment these characters may have different code
4891 values, and there are no code points greater than 255.
4894 CHARACTERS AND METACHARACTERS
4896 A regular expression is a pattern that is matched against a subject
4897 string from left to right. Most characters stand for themselves in a
4898 pattern, and match the corresponding characters in the subject. As a
4899 trivial example, the pattern
4903 matches a portion of a subject string that is identical to itself. When
4904 caseless matching is specified (the PCRE_CASELESS option), letters are
4905 matched independently of case. In a UTF mode, PCRE always understands
4906 the concept of case for characters whose values are less than 128, so
4907 caseless matching is always possible. For characters with higher val-
4908 ues, the concept of case is supported if PCRE is compiled with Unicode
4909 property support, but not otherwise. If you want to use caseless
4910 matching for characters 128 and above, you must ensure that PCRE is
4911 compiled with Unicode property support as well as with UTF support.
4913 The power of regular expressions comes from the ability to include
4914 alternatives and repetitions in the pattern. These are encoded in the
4915 pattern by the use of metacharacters, which do not stand for themselves
4916 but instead are interpreted in some special way.
4918 There are two different sets of metacharacters: those that are recog-
4919 nized anywhere in the pattern except within square brackets, and those
4920 that are recognized within square brackets. Outside square brackets,
4921 the metacharacters are as follows:
4923 \ general escape character with several uses
4924 ^ assert start of string (or line, in multiline mode)
4925 $ assert end of string (or line, in multiline mode)
4926 . match any character except newline (by default)
4927 [ start character class definition
4928 | start of alternative branch
4931 ? extends the meaning of (
4932 also 0 or 1 quantifier
4933 also quantifier minimizer
4934 * 0 or more quantifier
4935 + 1 or more quantifier
4936 also "possessive quantifier"
4937 { start min/max quantifier
4939 Part of a pattern that is in square brackets is called a "character
4940 class". In a character class the only metacharacters are:
4942 \ general escape character
4943 ^ negate the class, but only if the first character
4944 - indicates character range
4945 [ POSIX character class (only if followed by POSIX
4947 ] terminates the character class
4949 The following sections describe the use of each of the metacharacters.
4954 The backslash character has several uses. Firstly, if it is followed by
4955 a character that is not a number or a letter, it takes away any special
4956 meaning that character may have. This use of backslash as an escape
4957 character applies both inside and outside character classes.
4959 For example, if you want to match a * character, you write \* in the
4960 pattern. This escaping action applies whether or not the following
4961 character would otherwise be interpreted as a metacharacter, so it is
4962 always safe to precede a non-alphanumeric with backslash to specify
4963 that it stands for itself. In particular, if you want to match a back-
4964 slash, you write \\.
4966 In a UTF mode, only ASCII numbers and letters have any special meaning
4967 after a backslash. All other characters (in particular, those whose
4968 codepoints are greater than 127) are treated as literals.
4970 If a pattern is compiled with the PCRE_EXTENDED option, most white
4971 space in the pattern (other than in a character class), and characters
4972 between a # outside a character class and the next newline, inclusive,
4973 are ignored. An escaping backslash can be used to include a white space
4974 or # character as part of the pattern.
4976 If you want to remove the special meaning from a sequence of charac-
4977 ters, you can do so by putting them between \Q and \E. This is differ-
4978 ent from Perl in that $ and @ are handled as literals in \Q...\E
4979 sequences in PCRE, whereas in Perl, $ and @ cause variable interpola-
4980 tion. Note the following examples:
4982 Pattern PCRE matches Perl matches
4984 \Qabc$xyz\E abc$xyz abc followed by the
4986 \Qabc\$xyz\E abc\$xyz abc\$xyz
4987 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4989 The \Q...\E sequence is recognized both inside and outside character
4990 classes. An isolated \E that is not preceded by \Q is ignored. If \Q
4991 is not followed by \E later in the pattern, the literal interpretation
4992 continues to the end of the pattern (that is, \E is assumed at the
4993 end). If the isolated \Q is inside a character class, this causes an
4994 error, because the character class is not terminated.
4996 Non-printing characters
4998 A second use of backslash provides a way of encoding non-printing char-
4999 acters in patterns in a visible manner. There is no restriction on the
5000 appearance of non-printing characters, apart from the binary zero that
5001 terminates a pattern, but when a pattern is being prepared by text
5002 editing, it is often easier to use one of the following escape
5003 sequences than the binary character it represents. In an ASCII or Uni-
5004 code environment, these escapes are as follows:
5006 \a alarm, that is, the BEL character (hex 07)
5007 \cx "control-x", where x is any ASCII character
5009 \f form feed (hex 0C)
5010 \n linefeed (hex 0A)
5011 \r carriage return (hex 0D)
5013 \0dd character with octal code 0dd
5014 \ddd character with octal code ddd, or back reference
5015 \o{ddd..} character with octal code ddd..
5016 \xhh character with hex code hh
5017 \x{hhh..} character with hex code hhh.. (non-JavaScript mode)
5018 \uhhhh character with hex code hhhh (JavaScript mode only)
5020 The precise effect of \cx on ASCII characters is as follows: if x is a
5021 lower case letter, it is converted to upper case. Then bit 6 of the
5022 character (hex 40) is inverted. Thus \cA to \cZ become hex 01 to hex 1A
5023 (A is 41, Z is 5A), but \c{ becomes hex 3B ({ is 7B), and \c; becomes
5024 hex 7B (; is 3B). If the data item (byte or 16-bit value) following \c
5025 has a value greater than 127, a compile-time error occurs. This locks
5026 out non-ASCII characters in all modes.
5028 When PCRE is compiled in EBCDIC mode, \a, \e, \f, \n, \r, and \t gener-
5029 ate the appropriate EBCDIC code values. The \c escape is processed as
5030 specified for Perl in the perlebcdic document. The only characters that
5031 are allowed after \c are A-Z, a-z, or one of @, [, \, ], ^, _, or ?.
5032 Any other character provokes a compile-time error. The sequence \@
5033 encodes character code 0; the letters (in either case) encode charac-
5034 ters 1-26 (hex 01 to hex 1A); [, \, ], ^, and _ encode characters 27-31
5035 (hex 1B to hex 1F), and \? becomes either 255 (hex FF) or 95 (hex 5F).
5037 Thus, apart from \?, these escapes generate the same character code
5038 values as they do in an ASCII environment, though the meanings of the
5039 values mostly differ. For example, \G always generates code value 7,
5040 which is BEL in ASCII but DEL in EBCDIC.
5042 The sequence \? generates DEL (127, hex 7F) in an ASCII environment,
5043 but because 127 is not a control character in EBCDIC, Perl makes it
5044 generate the APC character. Unfortunately, there are several variants
5045 of EBCDIC. In most of them the APC character has the value 255 (hex
5046 FF), but in the one Perl calls POSIX-BC its value is 95 (hex 5F). If
5047 certain other characters have POSIX-BC values, PCRE makes \? generate
5048 95; otherwise it generates 255.
5050 After \0 up to two further octal digits are read. If there are fewer
5051 than two digits, just those that are present are used. Thus the
5052 sequence \0\x\015 specifies two binary zeros followed by a CR character
5053 (code value 13). Make sure you supply two digits after the initial zero
5054 if the pattern character that follows is itself an octal digit.
5056 The escape \o must be followed by a sequence of octal digits, enclosed
5057 in braces. An error occurs if this is not the case. This escape is a
5058 recent addition to Perl; it provides way of specifying character code
5059 points as octal numbers greater than 0777, and it also allows octal
5060 numbers and back references to be unambiguously specified.
5062 For greater clarity and unambiguity, it is best to avoid following \ by
5063 a digit greater than zero. Instead, use \o{} or \x{} to specify charac-
5064 ter numbers, and \g{} to specify back references. The following para-
5065 graphs describe the old, ambiguous syntax.
5067 The handling of a backslash followed by a digit other than 0 is compli-
5068 cated, and Perl has changed in recent releases, causing PCRE also to
5069 change. Outside a character class, PCRE reads the digit and any follow-
5070 ing digits as a decimal number. If the number is less than 8, or if
5071 there have been at least that many previous capturing left parentheses
5072 in the expression, the entire sequence is taken as a back reference. A
5073 description of how this works is given later, following the discussion
5074 of parenthesized subpatterns.
5076 Inside a character class, or if the decimal number following \ is
5077 greater than 7 and there have not been that many capturing subpatterns,
5078 PCRE handles \8 and \9 as the literal characters "8" and "9", and oth-
5079 erwise re-reads up to three octal digits following the backslash, using
5080 them to generate a data character. Any subsequent digits stand for
5081 themselves. For example:
5083 \040 is another way of writing an ASCII space
5084 \40 is the same, provided there are fewer than 40
5085 previous capturing subpatterns
5086 \7 is always a back reference
5087 \11 might be a back reference, or another way of
5089 \011 is always a tab
5090 \0113 is a tab followed by the character "3"
5091 \113 might be a back reference, otherwise the
5092 character with octal code 113
5093 \377 might be a back reference, otherwise
5094 the value 255 (decimal)
5095 \81 is either a back reference, or the two
5096 characters "8" and "1"
5098 Note that octal values of 100 or greater that are specified using this
5099 syntax must not be introduced by a leading zero, because no more than
5100 three octal digits are ever read.
5102 By default, after \x that is not followed by {, from zero to two hexa-
5103 decimal digits are read (letters can be in upper or lower case). Any
5104 number of hexadecimal digits may appear between \x{ and }. If a charac-
5105 ter other than a hexadecimal digit appears between \x{ and }, or if
5106 there is no terminating }, an error occurs.
5108 If the PCRE_JAVASCRIPT_COMPAT option is set, the interpretation of \x
5109 is as just described only when it is followed by two hexadecimal dig-
5110 its. Otherwise, it matches a literal "x" character. In JavaScript
5111 mode, support for code points greater than 256 is provided by \u, which
5112 must be followed by four hexadecimal digits; otherwise it matches a
5113 literal "u" character.
5115 Characters whose value is less than 256 can be defined by either of the
5116 two syntaxes for \x (or by \u in JavaScript mode). There is no differ-
5117 ence in the way they are handled. For example, \xdc is exactly the same
5118 as \x{dc} (or \u00dc in JavaScript mode).
5120 Constraints on character values
5122 Characters that are specified using octal or hexadecimal numbers are
5123 limited to certain values, as follows:
5125 8-bit non-UTF mode less than 0x100
5126 8-bit UTF-8 mode less than 0x10ffff and a valid codepoint
5127 16-bit non-UTF mode less than 0x10000
5128 16-bit UTF-16 mode less than 0x10ffff and a valid codepoint
5129 32-bit non-UTF mode less than 0x100000000
5130 32-bit UTF-32 mode less than 0x10ffff and a valid codepoint
5132 Invalid Unicode codepoints are the range 0xd800 to 0xdfff (the so-
5133 called "surrogate" codepoints), and 0xffef.
5135 Escape sequences in character classes
5137 All the sequences that define a single character value can be used both
5138 inside and outside character classes. In addition, inside a character
5139 class, \b is interpreted as the backspace character (hex 08).
5141 \N is not allowed in a character class. \B, \R, and \X are not special
5142 inside a character class. Like other unrecognized escape sequences,
5143 they are treated as the literal characters "B", "R", and "X" by
5144 default, but cause an error if the PCRE_EXTRA option is set. Outside a
5145 character class, these sequences have different meanings.
5147 Unsupported escape sequences
5149 In Perl, the sequences \l, \L, \u, and \U are recognized by its string
5150 handler and used to modify the case of following characters. By
5151 default, PCRE does not support these escape sequences. However, if the
5152 PCRE_JAVASCRIPT_COMPAT option is set, \U matches a "U" character, and
5153 \u can be used to define a character by code point, as described in the
5156 Absolute and relative back references
5158 The sequence \g followed by an unsigned or a negative number, option-
5159 ally enclosed in braces, is an absolute or relative back reference. A
5160 named back reference can be coded as \g{name}. Back references are dis-
5161 cussed later, following the discussion of parenthesized subpatterns.
5163 Absolute and relative subroutine calls
5165 For compatibility with Oniguruma, the non-Perl syntax \g followed by a
5166 name or a number enclosed either in angle brackets or single quotes, is
5167 an alternative syntax for referencing a subpattern as a "subroutine".
5168 Details are discussed later. Note that \g{...} (Perl syntax) and
5169 \g<...> (Oniguruma syntax) are not synonymous. The former is a back
5170 reference; the latter is a subroutine call.
5172 Generic character types
5174 Another use of backslash is for specifying generic character types:
5176 \d any decimal digit
5177 \D any character that is not a decimal digit
5178 \h any horizontal white space character
5179 \H any character that is not a horizontal white space character
5180 \s any white space character
5181 \S any character that is not a white space character
5182 \v any vertical white space character
5183 \V any character that is not a vertical white space character
5184 \w any "word" character
5185 \W any "non-word" character
5187 There is also the single sequence \N, which matches a non-newline char-
5188 acter. This is the same as the "." metacharacter when PCRE_DOTALL is
5189 not set. Perl also uses \N to match characters by name; PCRE does not
5192 Each pair of lower and upper case escape sequences partitions the com-
5193 plete set of characters into two disjoint sets. Any given character
5194 matches one, and only one, of each pair. The sequences can appear both
5195 inside and outside character classes. They each match one character of
5196 the appropriate type. If the current matching point is at the end of
5197 the subject string, all of them fail, because there is no character to
5200 For compatibility with Perl, \s did not used to match the VT character
5201 (code 11), which made it different from the the POSIX "space" class.
5202 However, Perl added VT at release 5.18, and PCRE followed suit at
5203 release 8.34. The default \s characters are now HT (9), LF (10), VT
5204 (11), FF (12), CR (13), and space (32), which are defined as white
5205 space in the "C" locale. This list may vary if locale-specific matching
5206 is taking place. For example, in some locales the "non-breaking space"
5207 character (\xA0) is recognized as white space, and in others the VT
5210 A "word" character is an underscore or any character that is a letter
5211 or digit. By default, the definition of letters and digits is con-
5212 trolled by PCRE's low-valued character tables, and may vary if locale-
5213 specific matching is taking place (see "Locale support" in the pcreapi
5214 page). For example, in a French locale such as "fr_FR" in Unix-like
5215 systems, or "french" in Windows, some character codes greater than 127
5216 are used for accented letters, and these are then matched by \w. The
5217 use of locales with Unicode is discouraged.
5219 By default, characters whose code points are greater than 127 never
5220 match \d, \s, or \w, and always match \D, \S, and \W, although this may
5221 vary for characters in the range 128-255 when locale-specific matching
5222 is happening. These escape sequences retain their original meanings
5223 from before Unicode support was available, mainly for efficiency rea-
5224 sons. If PCRE is compiled with Unicode property support, and the
5225 PCRE_UCP option is set, the behaviour is changed so that Unicode prop-
5226 erties are used to determine character types, as follows:
5228 \d any character that matches \p{Nd} (decimal digit)
5229 \s any character that matches \p{Z} or \h or \v
5230 \w any character that matches \p{L} or \p{N}, plus underscore
5232 The upper case escapes match the inverse sets of characters. Note that
5233 \d matches only decimal digits, whereas \w matches any Unicode digit,
5234 as well as any Unicode letter, and underscore. Note also that PCRE_UCP
5235 affects \b, and \B because they are defined in terms of \w and \W.
5236 Matching these sequences is noticeably slower when PCRE_UCP is set.
5238 The sequences \h, \H, \v, and \V are features that were added to Perl
5239 at release 5.10. In contrast to the other sequences, which match only
5240 ASCII characters by default, these always match certain high-valued
5241 code points, whether or not PCRE_UCP is set. The horizontal space char-
5244 U+0009 Horizontal tab (HT)
5246 U+00A0 Non-break space
5247 U+1680 Ogham space mark
5248 U+180E Mongolian vowel separator
5253 U+2004 Three-per-em space
5254 U+2005 Four-per-em space
5255 U+2006 Six-per-em space
5257 U+2008 Punctuation space
5260 U+202F Narrow no-break space
5261 U+205F Medium mathematical space
5262 U+3000 Ideographic space
5264 The vertical space characters are:
5266 U+000A Linefeed (LF)
5267 U+000B Vertical tab (VT)
5268 U+000C Form feed (FF)
5269 U+000D Carriage return (CR)
5270 U+0085 Next line (NEL)
5271 U+2028 Line separator
5272 U+2029 Paragraph separator
5274 In 8-bit, non-UTF-8 mode, only the characters with codepoints less than
5279 Outside a character class, by default, the escape sequence \R matches
5280 any Unicode newline sequence. In 8-bit non-UTF-8 mode \R is equivalent
5283 (?>\r\n|\n|\x0b|\f|\r|\x85)
5285 This is an example of an "atomic group", details of which are given
5286 below. This particular group matches either the two-character sequence
5287 CR followed by LF, or one of the single characters LF (linefeed,
5288 U+000A), VT (vertical tab, U+000B), FF (form feed, U+000C), CR (car-
5289 riage return, U+000D), or NEL (next line, U+0085). The two-character
5290 sequence is treated as a single unit that cannot be split.
5292 In other modes, two additional characters whose codepoints are greater
5293 than 255 are added: LS (line separator, U+2028) and PS (paragraph sepa-
5294 rator, U+2029). Unicode character property support is not needed for
5295 these characters to be recognized.
5297 It is possible to restrict \R to match only CR, LF, or CRLF (instead of
5298 the complete set of Unicode line endings) by setting the option
5299 PCRE_BSR_ANYCRLF either at compile time or when the pattern is matched.
5300 (BSR is an abbrevation for "backslash R".) This can be made the default
5301 when PCRE is built; if this is the case, the other behaviour can be
5302 requested via the PCRE_BSR_UNICODE option. It is also possible to
5303 specify these settings by starting a pattern string with one of the
5304 following sequences:
5306 (*BSR_ANYCRLF) CR, LF, or CRLF only
5307 (*BSR_UNICODE) any Unicode newline sequence
5309 These override the default and the options given to the compiling func-
5310 tion, but they can themselves be overridden by options given to a
5311 matching function. Note that these special settings, which are not
5312 Perl-compatible, are recognized only at the very start of a pattern,
5313 and that they must be in upper case. If more than one of them is
5314 present, the last one is used. They can be combined with a change of
5315 newline convention; for example, a pattern can start with:
5317 (*ANY)(*BSR_ANYCRLF)
5319 They can also be combined with the (*UTF8), (*UTF16), (*UTF32), (*UTF)
5320 or (*UCP) special sequences. Inside a character class, \R is treated as
5321 an unrecognized escape sequence, and so matches the letter "R" by
5322 default, but causes an error if PCRE_EXTRA is set.
5324 Unicode character properties
5326 When PCRE is built with Unicode character property support, three addi-
5327 tional escape sequences that match characters with specific properties
5328 are available. When in 8-bit non-UTF-8 mode, these sequences are of
5329 course limited to testing characters whose codepoints are less than
5330 256, but they do work in this mode. The extra escape sequences are:
5332 \p{xx} a character with the xx property
5333 \P{xx} a character without the xx property
5334 \X a Unicode extended grapheme cluster
5336 The property names represented by xx above are limited to the Unicode
5337 script names, the general category properties, "Any", which matches any
5338 character (including newline), and some special PCRE properties
5339 (described in the next section). Other Perl properties such as "InMu-
5340 sicalSymbols" are not currently supported by PCRE. Note that \P{Any}
5341 does not match any characters, so always causes a match failure.
5343 Sets of Unicode characters are defined as belonging to certain scripts.
5344 A character from one of these sets can be matched using a script name.
5350 Those that are not part of an identified script are lumped together as
5351 "Common". The current list of scripts is:
5353 Arabic, Armenian, Avestan, Balinese, Bamum, Bassa_Vah, Batak, Bengali,
5354 Bopomofo, Brahmi, Braille, Buginese, Buhid, Canadian_Aboriginal, Car-
5355 ian, Caucasian_Albanian, Chakma, Cham, Cherokee, Common, Coptic, Cunei-
5356 form, Cypriot, Cyrillic, Deseret, Devanagari, Duployan, Egyptian_Hiero-
5357 glyphs, Elbasan, Ethiopic, Georgian, Glagolitic, Gothic, Grantha,
5358 Greek, Gujarati, Gurmukhi, Han, Hangul, Hanunoo, Hebrew, Hiragana,
5359 Imperial_Aramaic, Inherited, Inscriptional_Pahlavi, Inscrip-
5360 tional_Parthian, Javanese, Kaithi, Kannada, Katakana, Kayah_Li,
5361 Kharoshthi, Khmer, Khojki, Khudawadi, Lao, Latin, Lepcha, Limbu, Lin-
5362 ear_A, Linear_B, Lisu, Lycian, Lydian, Mahajani, Malayalam, Mandaic,
5363 Manichaean, Meetei_Mayek, Mende_Kikakui, Meroitic_Cursive,
5364 Meroitic_Hieroglyphs, Miao, Modi, Mongolian, Mro, Myanmar, Nabataean,
5365 New_Tai_Lue, Nko, Ogham, Ol_Chiki, Old_Italic, Old_North_Arabian,
5366 Old_Permic, Old_Persian, Old_South_Arabian, Old_Turkic, Oriya, Osmanya,
5367 Pahawh_Hmong, Palmyrene, Pau_Cin_Hau, Phags_Pa, Phoenician,
5368 Psalter_Pahlavi, Rejang, Runic, Samaritan, Saurashtra, Sharada, Sha-
5369 vian, Siddham, Sinhala, Sora_Sompeng, Sundanese, Syloti_Nagri, Syriac,
5370 Tagalog, Tagbanwa, Tai_Le, Tai_Tham, Tai_Viet, Takri, Tamil, Telugu,
5371 Thaana, Thai, Tibetan, Tifinagh, Tirhuta, Ugaritic, Vai, Warang_Citi,
5374 Each character has exactly one Unicode general category property, spec-
5375 ified by a two-letter abbreviation. For compatibility with Perl, nega-
5376 tion can be specified by including a circumflex between the opening
5377 brace and the property name. For example, \p{^Lu} is the same as
5380 If only one letter is specified with \p or \P, it includes all the gen-
5381 eral category properties that start with that letter. In this case, in
5382 the absence of negation, the curly brackets in the escape sequence are
5383 optional; these two examples have the same effect:
5388 The following general category property codes are supported:
5398 Ll Lower case letter
5401 Lt Title case letter
5402 Lu Upper case letter
5415 Pc Connector punctuation
5417 Pe Close punctuation
5418 Pf Final punctuation
5419 Pi Initial punctuation
5420 Po Other punctuation
5426 Sm Mathematical symbol
5431 Zp Paragraph separator
5434 The special property L& is also supported: it matches a character that
5435 has the Lu, Ll, or Lt property, in other words, a letter that is not
5436 classified as a modifier or "other".
5438 The Cs (Surrogate) property applies only to characters in the range
5439 U+D800 to U+DFFF. Such characters are not valid in Unicode strings and
5440 so cannot be tested by PCRE, unless UTF validity checking has been
5441 turned off (see the discussion of PCRE_NO_UTF8_CHECK,
5442 PCRE_NO_UTF16_CHECK and PCRE_NO_UTF32_CHECK in the pcreapi page). Perl
5443 does not support the Cs property.
5445 The long synonyms for property names that Perl supports (such as
5446 \p{Letter}) are not supported by PCRE, nor is it permitted to prefix
5447 any of these properties with "Is".
5449 No character that is in the Unicode table has the Cn (unassigned) prop-
5450 erty. Instead, this property is assumed for any code point that is not
5451 in the Unicode table.
5453 Specifying caseless matching does not affect these escape sequences.
5454 For example, \p{Lu} always matches only upper case letters. This is
5455 different from the behaviour of current versions of Perl.
5457 Matching characters by Unicode property is not fast, because PCRE has
5458 to do a multistage table lookup in order to find a character's prop-
5459 erty. That is why the traditional escape sequences such as \d and \w do
5460 not use Unicode properties in PCRE by default, though you can make them
5461 do so by setting the PCRE_UCP option or by starting the pattern with
5464 Extended grapheme clusters
5466 The \X escape matches any number of Unicode characters that form an
5467 "extended grapheme cluster", and treats the sequence as an atomic group
5468 (see below). Up to and including release 8.31, PCRE matched an ear-
5469 lier, simpler definition that was equivalent to
5473 That is, it matched a character without the "mark" property, followed
5474 by zero or more characters with the "mark" property. Characters with
5475 the "mark" property are typically non-spacing accents that affect the
5476 preceding character.
5478 This simple definition was extended in Unicode to include more compli-
5479 cated kinds of composite character by giving each character a grapheme
5480 breaking property, and creating rules that use these properties to
5481 define the boundaries of extended grapheme clusters. In releases of
5482 PCRE later than 8.31, \X matches one of these clusters.
5484 \X always matches at least one character. Then it decides whether to
5485 add additional characters according to the following rules for ending a
5488 1. End at the end of the subject string.
5490 2. Do not end between CR and LF; otherwise end after any control char-
5493 3. Do not break Hangul (a Korean script) syllable sequences. Hangul
5494 characters are of five types: L, V, T, LV, and LVT. An L character may
5495 be followed by an L, V, LV, or LVT character; an LV or V character may
5496 be followed by a V or T character; an LVT or T character may be follwed
5497 only by a T character.
5499 4. Do not end before extending characters or spacing marks. Characters
5500 with the "mark" property always have the "extend" grapheme breaking
5503 5. Do not end after prepend characters.
5505 6. Otherwise, end the cluster.
5507 PCRE's additional properties
5509 As well as the standard Unicode properties described above, PCRE sup-
5510 ports four more that make it possible to convert traditional escape
5511 sequences such as \w and \s to use Unicode properties. PCRE uses these
5512 non-standard, non-Perl properties internally when PCRE_UCP is set. How-
5513 ever, they may also be used explicitly. These properties are:
5515 Xan Any alphanumeric character
5516 Xps Any POSIX space character
5517 Xsp Any Perl space character
5518 Xwd Any Perl "word" character
5520 Xan matches characters that have either the L (letter) or the N (num-
5521 ber) property. Xps matches the characters tab, linefeed, vertical tab,
5522 form feed, or carriage return, and any other character that has the Z
5523 (separator) property. Xsp is the same as Xps; it used to exclude ver-
5524 tical tab, for Perl compatibility, but Perl changed, and so PCRE fol-
5525 lowed at release 8.34. Xwd matches the same characters as Xan, plus
5528 There is another non-standard property, Xuc, which matches any charac-
5529 ter that can be represented by a Universal Character Name in C++ and
5530 other programming languages. These are the characters $, @, ` (grave
5531 accent), and all characters with Unicode code points greater than or
5532 equal to U+00A0, except for the surrogates U+D800 to U+DFFF. Note that
5533 most base (ASCII) characters are excluded. (Universal Character Names
5534 are of the form \uHHHH or \UHHHHHHHH where H is a hexadecimal digit.
5535 Note that the Xuc property does not match these sequences but the char-
5536 acters that they represent.)
5538 Resetting the match start
5540 The escape sequence \K causes any previously matched characters not to
5541 be included in the final matched sequence. For example, the pattern:
5545 matches "foobar", but reports that it has matched "bar". This feature
5546 is similar to a lookbehind assertion (described below). However, in
5547 this case, the part of the subject before the real match does not have
5548 to be of fixed length, as lookbehind assertions do. The use of \K does
5549 not interfere with the setting of captured substrings. For example,
5554 matches "foobar", the first substring is still set to "foo".
5556 Perl documents that the use of \K within assertions is "not well
5557 defined". In PCRE, \K is acted upon when it occurs inside positive
5558 assertions, but is ignored in negative assertions. Note that when a
5559 pattern such as (?=ab\K) matches, the reported start of the match can
5560 be greater than the end of the match.
5564 The final use of backslash is for certain simple assertions. An asser-
5565 tion specifies a condition that has to be met at a particular point in
5566 a match, without consuming any characters from the subject string. The
5567 use of subpatterns for more complicated assertions is described below.
5568 The backslashed assertions are:
5570 \b matches at a word boundary
5571 \B matches when not at a word boundary
5572 \A matches at the start of the subject
5573 \Z matches at the end of the subject
5574 also matches before a newline at the end of the subject
5575 \z matches only at the end of the subject
5576 \G matches at the first matching position in the subject
5578 Inside a character class, \b has a different meaning; it matches the
5579 backspace character. If any other of these assertions appears in a
5580 character class, by default it matches the corresponding literal char-
5581 acter (for example, \B matches the letter B). However, if the
5582 PCRE_EXTRA option is set, an "invalid escape sequence" error is gener-
5585 A word boundary is a position in the subject string where the current
5586 character and the previous character do not both match \w or \W (i.e.
5587 one matches \w and the other matches \W), or the start or end of the
5588 string if the first or last character matches \w, respectively. In a
5589 UTF mode, the meanings of \w and \W can be changed by setting the
5590 PCRE_UCP option. When this is done, it also affects \b and \B. Neither
5591 PCRE nor Perl has a separate "start of word" or "end of word" metase-
5592 quence. However, whatever follows \b normally determines which it is.
5593 For example, the fragment \ba matches "a" at the start of a word.
5595 The \A, \Z, and \z assertions differ from the traditional circumflex
5596 and dollar (described in the next section) in that they only ever match
5597 at the very start and end of the subject string, whatever options are
5598 set. Thus, they are independent of multiline mode. These three asser-
5599 tions are not affected by the PCRE_NOTBOL or PCRE_NOTEOL options, which
5600 affect only the behaviour of the circumflex and dollar metacharacters.
5601 However, if the startoffset argument of pcre_exec() is non-zero, indi-
5602 cating that matching is to start at a point other than the beginning of
5603 the subject, \A can never match. The difference between \Z and \z is
5604 that \Z matches before a newline at the end of the string as well as at
5605 the very end, whereas \z matches only at the end.
5607 The \G assertion is true only when the current matching position is at
5608 the start point of the match, as specified by the startoffset argument
5609 of pcre_exec(). It differs from \A when the value of startoffset is
5610 non-zero. By calling pcre_exec() multiple times with appropriate argu-
5611 ments, you can mimic Perl's /g option, and it is in this kind of imple-
5612 mentation where \G can be useful.
5614 Note, however, that PCRE's interpretation of \G, as the start of the
5615 current match, is subtly different from Perl's, which defines it as the
5616 end of the previous match. In Perl, these can be different when the
5617 previously matched string was empty. Because PCRE does just one match
5618 at a time, it cannot reproduce this behaviour.
5620 If all the alternatives of a pattern begin with \G, the expression is
5621 anchored to the starting match position, and the "anchored" flag is set
5622 in the compiled regular expression.
5625 CIRCUMFLEX AND DOLLAR
5627 The circumflex and dollar metacharacters are zero-width assertions.
5628 That is, they test for a particular condition being true without con-
5629 suming any characters from the subject string.
5631 Outside a character class, in the default matching mode, the circumflex
5632 character is an assertion that is true only if the current matching
5633 point is at the start of the subject string. If the startoffset argu-
5634 ment of pcre_exec() is non-zero, circumflex can never match if the
5635 PCRE_MULTILINE option is unset. Inside a character class, circumflex
5636 has an entirely different meaning (see below).
5638 Circumflex need not be the first character of the pattern if a number
5639 of alternatives are involved, but it should be the first thing in each
5640 alternative in which it appears if the pattern is ever to match that
5641 branch. If all possible alternatives start with a circumflex, that is,
5642 if the pattern is constrained to match only at the start of the sub-
5643 ject, it is said to be an "anchored" pattern. (There are also other
5644 constructs that can cause a pattern to be anchored.)
5646 The dollar character is an assertion that is true only if the current
5647 matching point is at the end of the subject string, or immediately
5648 before a newline at the end of the string (by default). Note, however,
5649 that it does not actually match the newline. Dollar need not be the
5650 last character of the pattern if a number of alternatives are involved,
5651 but it should be the last item in any branch in which it appears. Dol-
5652 lar has no special meaning in a character class.
5654 The meaning of dollar can be changed so that it matches only at the
5655 very end of the string, by setting the PCRE_DOLLAR_ENDONLY option at
5656 compile time. This does not affect the \Z assertion.
5658 The meanings of the circumflex and dollar characters are changed if the
5659 PCRE_MULTILINE option is set. When this is the case, a circumflex
5660 matches immediately after internal newlines as well as at the start of
5661 the subject string. It does not match after a newline that ends the
5662 string. A dollar matches before any newlines in the string, as well as
5663 at the very end, when PCRE_MULTILINE is set. When newline is specified
5664 as the two-character sequence CRLF, isolated CR and LF characters do
5665 not indicate newlines.
5667 For example, the pattern /^abc$/ matches the subject string "def\nabc"
5668 (where \n represents a newline) in multiline mode, but not otherwise.
5669 Consequently, patterns that are anchored in single line mode because
5670 all branches start with ^ are not anchored in multiline mode, and a
5671 match for circumflex is possible when the startoffset argument of
5672 pcre_exec() is non-zero. The PCRE_DOLLAR_ENDONLY option is ignored if
5673 PCRE_MULTILINE is set.
5675 Note that the sequences \A, \Z, and \z can be used to match the start
5676 and end of the subject in both modes, and if all branches of a pattern
5677 start with \A it is always anchored, whether or not PCRE_MULTILINE is
5681 FULL STOP (PERIOD, DOT) AND \N
5683 Outside a character class, a dot in the pattern matches any one charac-
5684 ter in the subject string except (by default) a character that signi-
5685 fies the end of a line.
5687 When a line ending is defined as a single character, dot never matches
5688 that character; when the two-character sequence CRLF is used, dot does
5689 not match CR if it is immediately followed by LF, but otherwise it
5690 matches all characters (including isolated CRs and LFs). When any Uni-
5691 code line endings are being recognized, dot does not match CR or LF or
5692 any of the other line ending characters.
5694 The behaviour of dot with regard to newlines can be changed. If the
5695 PCRE_DOTALL option is set, a dot matches any one character, without
5696 exception. If the two-character sequence CRLF is present in the subject
5697 string, it takes two dots to match it.
5699 The handling of dot is entirely independent of the handling of circum-
5700 flex and dollar, the only relationship being that they both involve
5701 newlines. Dot has no special meaning in a character class.
5703 The escape sequence \N behaves like a dot, except that it is not
5704 affected by the PCRE_DOTALL option. In other words, it matches any
5705 character except one that signifies the end of a line. Perl also uses
5706 \N to match characters by name; PCRE does not support this.
5709 MATCHING A SINGLE DATA UNIT
5711 Outside a character class, the escape sequence \C matches any one data
5712 unit, whether or not a UTF mode is set. In the 8-bit library, one data
5713 unit is one byte; in the 16-bit library it is a 16-bit unit; in the
5714 32-bit library it is a 32-bit unit. Unlike a dot, \C always matches
5715 line-ending characters. The feature is provided in Perl in order to
5716 match individual bytes in UTF-8 mode, but it is unclear how it can use-
5717 fully be used. Because \C breaks up characters into individual data
5718 units, matching one unit with \C in a UTF mode means that the rest of
5719 the string may start with a malformed UTF character. This has undefined
5720 results, because PCRE assumes that it is dealing with valid UTF strings
5721 (and by default it checks this at the start of processing unless the
5722 PCRE_NO_UTF8_CHECK, PCRE_NO_UTF16_CHECK or PCRE_NO_UTF32_CHECK option
5725 PCRE does not allow \C to appear in lookbehind assertions (described
5726 below) in a UTF mode, because this would make it impossible to calcu-
5727 late the length of the lookbehind.
5729 In general, the \C escape sequence is best avoided. However, one way of
5730 using it that avoids the problem of malformed UTF characters is to use
5731 a lookahead to check the length of the next character, as in this pat-
5732 tern, which could be used with a UTF-8 string (ignore white space and
5735 (?| (?=[\x00-\x7f])(\C) |
5736 (?=[\x80-\x{7ff}])(\C)(\C) |
5737 (?=[\x{800}-\x{ffff}])(\C)(\C)(\C) |
5738 (?=[\x{10000}-\x{1fffff}])(\C)(\C)(\C)(\C))
5740 A group that starts with (?| resets the capturing parentheses numbers
5741 in each alternative (see "Duplicate Subpattern Numbers" below). The
5742 assertions at the start of each branch check the next UTF-8 character
5743 for values whose encoding uses 1, 2, 3, or 4 bytes, respectively. The
5744 character's individual bytes are then captured by the appropriate num-
5748 SQUARE BRACKETS AND CHARACTER CLASSES
5750 An opening square bracket introduces a character class, terminated by a
5751 closing square bracket. A closing square bracket on its own is not spe-
5752 cial by default. However, if the PCRE_JAVASCRIPT_COMPAT option is set,
5753 a lone closing square bracket causes a compile-time error. If a closing
5754 square bracket is required as a member of the class, it should be the
5755 first data character in the class (after an initial circumflex, if
5756 present) or escaped with a backslash.
5758 A character class matches a single character in the subject. In a UTF
5759 mode, the character may be more than one data unit long. A matched
5760 character must be in the set of characters defined by the class, unless
5761 the first character in the class definition is a circumflex, in which
5762 case the subject character must not be in the set defined by the class.
5763 If a circumflex is actually required as a member of the class, ensure
5764 it is not the first character, or escape it with a backslash.
5766 For example, the character class [aeiou] matches any lower case vowel,
5767 while [^aeiou] matches any character that is not a lower case vowel.
5768 Note that a circumflex is just a convenient notation for specifying the
5769 characters that are in the class by enumerating those that are not. A
5770 class that starts with a circumflex is not an assertion; it still con-
5771 sumes a character from the subject string, and therefore it fails if
5772 the current pointer is at the end of the string.
5774 In UTF-8 (UTF-16, UTF-32) mode, characters with values greater than 255
5775 (0xffff) can be included in a class as a literal string of data units,
5776 or by using the \x{ escaping mechanism.
5778 When caseless matching is set, any letters in a class represent both
5779 their upper case and lower case versions, so for example, a caseless
5780 [aeiou] matches "A" as well as "a", and a caseless [^aeiou] does not
5781 match "A", whereas a caseful version would. In a UTF mode, PCRE always
5782 understands the concept of case for characters whose values are less
5783 than 128, so caseless matching is always possible. For characters with
5784 higher values, the concept of case is supported if PCRE is compiled
5785 with Unicode property support, but not otherwise. If you want to use
5786 caseless matching in a UTF mode for characters 128 and above, you must
5787 ensure that PCRE is compiled with Unicode property support as well as
5790 Characters that might indicate line breaks are never treated in any
5791 special way when matching character classes, whatever line-ending
5792 sequence is in use, and whatever setting of the PCRE_DOTALL and
5793 PCRE_MULTILINE options is used. A class such as [^a] always matches one
5794 of these characters.
5796 The minus (hyphen) character can be used to specify a range of charac-
5797 ters in a character class. For example, [d-m] matches any letter
5798 between d and m, inclusive. If a minus character is required in a
5799 class, it must be escaped with a backslash or appear in a position
5800 where it cannot be interpreted as indicating a range, typically as the
5801 first or last character in the class, or immediately after a range. For
5802 example, [b-d-z] matches letters in the range b to d, a hyphen charac-
5805 It is not possible to have the literal character "]" as the end charac-
5806 ter of a range. A pattern such as [W-]46] is interpreted as a class of
5807 two characters ("W" and "-") followed by a literal string "46]", so it
5808 would match "W46]" or "-46]". However, if the "]" is escaped with a
5809 backslash it is interpreted as the end of range, so [W-\]46] is inter-
5810 preted as a class containing a range followed by two other characters.
5811 The octal or hexadecimal representation of "]" can also be used to end
5814 An error is generated if a POSIX character class (see below) or an
5815 escape sequence other than one that defines a single character appears
5816 at a point where a range ending character is expected. For example,
5817 [z-\xff] is valid, but [A-\d] and [A-[:digit:]] are not.
5819 Ranges operate in the collating sequence of character values. They can
5820 also be used for characters specified numerically, for example
5821 [\000-\037]. Ranges can include any characters that are valid for the
5824 If a range that includes letters is used when caseless matching is set,
5825 it matches the letters in either case. For example, [W-c] is equivalent
5826 to [][\\^_`wxyzabc], matched caselessly, and in a non-UTF mode, if
5827 character tables for a French locale are in use, [\xc8-\xcb] matches
5828 accented E characters in both cases. In UTF modes, PCRE supports the
5829 concept of case for characters with values greater than 128 only when
5830 it is compiled with Unicode property support.
5832 The character escape sequences \d, \D, \h, \H, \p, \P, \s, \S, \v, \V,
5833 \w, and \W may appear in a character class, and add the characters that
5834 they match to the class. For example, [\dABCDEF] matches any hexadeci-
5835 mal digit. In UTF modes, the PCRE_UCP option affects the meanings of
5836 \d, \s, \w and their upper case partners, just as it does when they
5837 appear outside a character class, as described in the section entitled
5838 "Generic character types" above. The escape sequence \b has a different
5839 meaning inside a character class; it matches the backspace character.
5840 The sequences \B, \N, \R, and \X are not special inside a character
5841 class. Like any other unrecognized escape sequences, they are treated
5842 as the literal characters "B", "N", "R", and "X" by default, but cause
5843 an error if the PCRE_EXTRA option is set.
5845 A circumflex can conveniently be used with the upper case character
5846 types to specify a more restricted set of characters than the matching
5847 lower case type. For example, the class [^\W_] matches any letter or
5848 digit, but not underscore, whereas [\w] includes underscore. A positive
5849 character class should be read as "something OR something OR ..." and a
5850 negative class as "NOT something AND NOT something AND NOT ...".
5852 The only metacharacters that are recognized in character classes are
5853 backslash, hyphen (only where it can be interpreted as specifying a
5854 range), circumflex (only at the start), opening square bracket (only
5855 when it can be interpreted as introducing a POSIX class name, or for a
5856 special compatibility feature - see the next two sections), and the
5857 terminating closing square bracket. However, escaping other non-
5858 alphanumeric characters does no harm.
5861 POSIX CHARACTER CLASSES
5863 Perl supports the POSIX notation for character classes. This uses names
5864 enclosed by [: and :] within the enclosing square brackets. PCRE also
5865 supports this notation. For example,
5869 matches "0", "1", any alphabetic character, or "%". The supported class
5872 alnum letters and digits
5874 ascii character codes 0 - 127
5875 blank space or tab only
5876 cntrl control characters
5877 digit decimal digits (same as \d)
5878 graph printing characters, excluding space
5879 lower lower case letters
5880 print printing characters, including space
5881 punct printing characters, excluding letters and digits and space
5882 space white space (the same as \s from PCRE 8.34)
5883 upper upper case letters
5884 word "word" characters (same as \w)
5885 xdigit hexadecimal digits
5887 The default "space" characters are HT (9), LF (10), VT (11), FF (12),
5888 CR (13), and space (32). If locale-specific matching is taking place,
5889 the list of space characters may be different; there may be fewer or
5890 more of them. "Space" used to be different to \s, which did not include
5891 VT, for Perl compatibility. However, Perl changed at release 5.18, and
5892 PCRE followed at release 8.34. "Space" and \s now match the same set
5895 The name "word" is a Perl extension, and "blank" is a GNU extension
5896 from Perl 5.8. Another Perl extension is negation, which is indicated
5897 by a ^ character after the colon. For example,
5901 matches "1", "2", or any non-digit. PCRE (and Perl) also recognize the
5902 POSIX syntax [.ch.] and [=ch=] where "ch" is a "collating element", but
5903 these are not supported, and an error is given if they are encountered.
5905 By default, characters with values greater than 128 do not match any of
5906 the POSIX character classes. However, if the PCRE_UCP option is passed
5907 to pcre_compile(), some of the classes are changed so that Unicode
5908 character properties are used. This is achieved by replacing certain
5909 POSIX classes by other sequences, as follows:
5911 [:alnum:] becomes \p{Xan}
5912 [:alpha:] becomes \p{L}
5913 [:blank:] becomes \h
5914 [:digit:] becomes \p{Nd}
5915 [:lower:] becomes \p{Ll}
5916 [:space:] becomes \p{Xps}
5917 [:upper:] becomes \p{Lu}
5918 [:word:] becomes \p{Xwd}
5920 Negated versions, such as [:^alpha:] use \P instead of \p. Three other
5921 POSIX classes are handled specially in UCP mode:
5923 [:graph:] This matches characters that have glyphs that mark the page
5924 when printed. In Unicode property terms, it matches all char-
5925 acters with the L, M, N, P, S, or Cf properties, except for:
5927 U+061C Arabic Letter Mark
5928 U+180E Mongolian Vowel Separator
5929 U+2066 - U+2069 Various "isolate"s
5932 [:print:] This matches the same characters as [:graph:] plus space
5933 characters that are not controls, that is, characters with
5936 [:punct:] This matches all characters that have the Unicode P (punctua-
5937 tion) property, plus those characters whose code points are
5938 less than 128 that have the S (Symbol) property.
5940 The other POSIX classes are unchanged, and match only characters with
5941 code points less than 128.
5944 COMPATIBILITY FEATURE FOR WORD BOUNDARIES
5946 In the POSIX.2 compliant library that was included in 4.4BSD Unix, the
5947 ugly syntax [[:<:]] and [[:>:]] is used for matching "start of word"
5948 and "end of word". PCRE treats these items as follows:
5950 [[:<:]] is converted to \b(?=\w)
5951 [[:>:]] is converted to \b(?<=\w)
5953 Only these exact character sequences are recognized. A sequence such as
5954 [a[:<:]b] provokes error for an unrecognized POSIX class name. This
5955 support is not compatible with Perl. It is provided to help migrations
5956 from other environments, and is best not used in any new patterns. Note
5957 that \b matches at the start and the end of a word (see "Simple asser-
5958 tions" above), and in a Perl-style pattern the preceding or following
5959 character normally shows which is wanted, without the need for the
5960 assertions that are used above in order to give exactly the POSIX be-
5966 Vertical bar characters are used to separate alternative patterns. For
5967 example, the pattern
5971 matches either "gilbert" or "sullivan". Any number of alternatives may
5972 appear, and an empty alternative is permitted (matching the empty
5973 string). The matching process tries each alternative in turn, from left
5974 to right, and the first one that succeeds is used. If the alternatives
5975 are within a subpattern (defined below), "succeeds" means matching the
5976 rest of the main pattern as well as the alternative in the subpattern.
5979 INTERNAL OPTION SETTING
5981 The settings of the PCRE_CASELESS, PCRE_MULTILINE, PCRE_DOTALL, and
5982 PCRE_EXTENDED options (which are Perl-compatible) can be changed from
5983 within the pattern by a sequence of Perl option letters enclosed
5984 between "(?" and ")". The option letters are
5987 m for PCRE_MULTILINE
5991 For example, (?im) sets caseless, multiline matching. It is also possi-
5992 ble to unset these options by preceding the letter with a hyphen, and a
5993 combined setting and unsetting such as (?im-sx), which sets PCRE_CASE-
5994 LESS and PCRE_MULTILINE while unsetting PCRE_DOTALL and PCRE_EXTENDED,
5995 is also permitted. If a letter appears both before and after the
5996 hyphen, the option is unset.
5998 The PCRE-specific options PCRE_DUPNAMES, PCRE_UNGREEDY, and PCRE_EXTRA
5999 can be changed in the same way as the Perl-compatible options by using
6000 the characters J, U and X respectively.
6002 When one of these option changes occurs at top level (that is, not
6003 inside subpattern parentheses), the change applies to the remainder of
6004 the pattern that follows. If the change is placed right at the start of
6005 a pattern, PCRE extracts it into the global options (and it will there-
6006 fore show up in data extracted by the pcre_fullinfo() function).
6008 An option change within a subpattern (see below for a description of
6009 subpatterns) affects only that part of the subpattern that follows it,
6014 matches abc and aBc and no other strings (assuming PCRE_CASELESS is not
6015 used). By this means, options can be made to have different settings
6016 in different parts of the pattern. Any changes made in one alternative
6017 do carry on into subsequent branches within the same subpattern. For
6022 matches "ab", "aB", "c", and "C", even though when matching "C" the
6023 first branch is abandoned before the option setting. This is because
6024 the effects of option settings happen at compile time. There would be
6025 some very weird behaviour otherwise.
6027 Note: There are other PCRE-specific options that can be set by the
6028 application when the compiling or matching functions are called. In
6029 some cases the pattern can contain special leading sequences such as
6030 (*CRLF) to override what the application has set or what has been
6031 defaulted. Details are given in the section entitled "Newline
6032 sequences" above. There are also the (*UTF8), (*UTF16),(*UTF32), and
6033 (*UCP) leading sequences that can be used to set UTF and Unicode prop-
6034 erty modes; they are equivalent to setting the PCRE_UTF8, PCRE_UTF16,
6035 PCRE_UTF32 and the PCRE_UCP options, respectively. The (*UTF) sequence
6036 is a generic version that can be used with any of the libraries. How-
6037 ever, the application can set the PCRE_NEVER_UTF option, which locks
6038 out the use of the (*UTF) sequences.
6043 Subpatterns are delimited by parentheses (round brackets), which can be
6044 nested. Turning part of a pattern into a subpattern does two things:
6046 1. It localizes a set of alternatives. For example, the pattern
6048 cat(aract|erpillar|)
6050 matches "cataract", "caterpillar", or "cat". Without the parentheses,
6051 it would match "cataract", "erpillar" or an empty string.
6053 2. It sets up the subpattern as a capturing subpattern. This means
6054 that, when the whole pattern matches, that portion of the subject
6055 string that matched the subpattern is passed back to the caller via the
6056 ovector argument of the matching function. (This applies only to the
6057 traditional matching functions; the DFA matching functions do not sup-
6060 Opening parentheses are counted from left to right (starting from 1) to
6061 obtain numbers for the capturing subpatterns. For example, if the
6062 string "the red king" is matched against the pattern
6064 the ((red|white) (king|queen))
6066 the captured substrings are "red king", "red", and "king", and are num-
6067 bered 1, 2, and 3, respectively.
6069 The fact that plain parentheses fulfil two functions is not always
6070 helpful. There are often times when a grouping subpattern is required
6071 without a capturing requirement. If an opening parenthesis is followed
6072 by a question mark and a colon, the subpattern does not do any captur-
6073 ing, and is not counted when computing the number of any subsequent
6074 capturing subpatterns. For example, if the string "the white queen" is
6075 matched against the pattern
6077 the ((?:red|white) (king|queen))
6079 the captured substrings are "white queen" and "queen", and are numbered
6080 1 and 2. The maximum number of capturing subpatterns is 65535.
6082 As a convenient shorthand, if any option settings are required at the
6083 start of a non-capturing subpattern, the option letters may appear
6084 between the "?" and the ":". Thus the two patterns
6086 (?i:saturday|sunday)
6087 (?:(?i)saturday|sunday)
6089 match exactly the same set of strings. Because alternative branches are
6090 tried from left to right, and options are not reset until the end of
6091 the subpattern is reached, an option setting in one branch does affect
6092 subsequent branches, so the above patterns match "SUNDAY" as well as
6096 DUPLICATE SUBPATTERN NUMBERS
6098 Perl 5.10 introduced a feature whereby each alternative in a subpattern
6099 uses the same numbers for its capturing parentheses. Such a subpattern
6100 starts with (?| and is itself a non-capturing subpattern. For example,
6101 consider this pattern:
6103 (?|(Sat)ur|(Sun))day
6105 Because the two alternatives are inside a (?| group, both sets of cap-
6106 turing parentheses are numbered one. Thus, when the pattern matches,
6107 you can look at captured substring number one, whichever alternative
6108 matched. This construct is useful when you want to capture part, but
6109 not all, of one of a number of alternatives. Inside a (?| group, paren-
6110 theses are numbered as usual, but the number is reset at the start of
6111 each branch. The numbers of any capturing parentheses that follow the
6112 subpattern start after the highest number used in any branch. The fol-
6113 lowing example is taken from the Perl documentation. The numbers under-
6114 neath show in which buffer the captured content will be stored.
6116 # before ---------------branch-reset----------- after
6117 / ( a ) (?| x ( y ) z | (p (q) r) | (t) u (v) ) ( z ) /x
6120 A back reference to a numbered subpattern uses the most recent value
6121 that is set for that number by any subpattern. The following pattern
6122 matches "abcabc" or "defdef":
6126 In contrast, a subroutine call to a numbered subpattern always refers
6127 to the first one in the pattern with the given number. The following
6128 pattern matches "abcabc" or "defabc":
6130 /(?|(abc)|(def))(?1)/
6132 If a condition test for a subpattern's having matched refers to a non-
6133 unique number, the test is true if any of the subpatterns of that num-
6136 An alternative approach to using this "branch reset" feature is to use
6137 duplicate named subpatterns, as described in the next section.
6142 Identifying capturing parentheses by number is simple, but it can be
6143 very hard to keep track of the numbers in complicated regular expres-
6144 sions. Furthermore, if an expression is modified, the numbers may
6145 change. To help with this difficulty, PCRE supports the naming of sub-
6146 patterns. This feature was not added to Perl until release 5.10. Python
6147 had the feature earlier, and PCRE introduced it at release 4.0, using
6148 the Python syntax. PCRE now supports both the Perl and the Python syn-
6149 tax. Perl allows identically numbered subpatterns to have different
6150 names, but PCRE does not.
6152 In PCRE, a subpattern can be named in one of three ways: (?<name>...)
6153 or (?'name'...) as in Perl, or (?P<name>...) as in Python. References
6154 to capturing parentheses from other parts of the pattern, such as back
6155 references, recursion, and conditions, can be made by name as well as
6158 Names consist of up to 32 alphanumeric characters and underscores, but
6159 must start with a non-digit. Named capturing parentheses are still
6160 allocated numbers as well as names, exactly as if the names were not
6161 present. The PCRE API provides function calls for extracting the name-
6162 to-number translation table from a compiled pattern. There is also a
6163 convenience function for extracting a captured substring by name.
6165 By default, a name must be unique within a pattern, but it is possible
6166 to relax this constraint by setting the PCRE_DUPNAMES option at compile
6167 time. (Duplicate names are also always permitted for subpatterns with
6168 the same number, set up as described in the previous section.) Dupli-
6169 cate names can be useful for patterns where only one instance of the
6170 named parentheses can match. Suppose you want to match the name of a
6171 weekday, either as a 3-letter abbreviation or as the full name, and in
6172 both cases you want to extract the abbreviation. This pattern (ignoring
6173 the line breaks) does the job:
6175 (?<DN>Mon|Fri|Sun)(?:day)?|
6176 (?<DN>Tue)(?:sday)?|
6177 (?<DN>Wed)(?:nesday)?|
6178 (?<DN>Thu)(?:rsday)?|
6179 (?<DN>Sat)(?:urday)?
6181 There are five capturing substrings, but only one is ever set after a
6182 match. (An alternative way of solving this problem is to use a "branch
6183 reset" subpattern, as described in the previous section.)
6185 The convenience function for extracting the data by name returns the
6186 substring for the first (and in this example, the only) subpattern of
6187 that name that matched. This saves searching to find which numbered
6190 If you make a back reference to a non-unique named subpattern from
6191 elsewhere in the pattern, the subpatterns to which the name refers are
6192 checked in the order in which they appear in the overall pattern. The
6193 first one that is set is used for the reference. For example, this pat-
6194 tern matches both "foofoo" and "barbar" but not "foobar" or "barfoo":
6196 (?:(?<n>foo)|(?<n>bar))\k<n>
6199 If you make a subroutine call to a non-unique named subpattern, the one
6200 that corresponds to the first occurrence of the name is used. In the
6201 absence of duplicate numbers (see the previous section) this is the one
6202 with the lowest number.
6204 If you use a named reference in a condition test (see the section about
6205 conditions below), either to check whether a subpattern has matched, or
6206 to check for recursion, all subpatterns with the same name are tested.
6207 If the condition is true for any one of them, the overall condition is
6208 true. This is the same behaviour as testing by number. For further
6209 details of the interfaces for handling named subpatterns, see the
6210 pcreapi documentation.
6212 Warning: You cannot use different names to distinguish between two sub-
6213 patterns with the same number because PCRE uses only the numbers when
6214 matching. For this reason, an error is given at compile time if differ-
6215 ent names are given to subpatterns with the same number. However, you
6216 can always give the same name to subpatterns with the same number, even
6217 when PCRE_DUPNAMES is not set.
6222 Repetition is specified by quantifiers, which can follow any of the
6225 a literal data character
6226 the dot metacharacter
6227 the \C escape sequence
6228 the \X escape sequence
6229 the \R escape sequence
6230 an escape such as \d or \pL that matches a single character
6232 a back reference (see next section)
6233 a parenthesized subpattern (including assertions)
6234 a subroutine call to a subpattern (recursive or otherwise)
6236 The general repetition quantifier specifies a minimum and maximum num-
6237 ber of permitted matches, by giving the two numbers in curly brackets
6238 (braces), separated by a comma. The numbers must be less than 65536,
6239 and the first must be less than or equal to the second. For example:
6243 matches "zz", "zzz", or "zzzz". A closing brace on its own is not a
6244 special character. If the second number is omitted, but the comma is
6245 present, there is no upper limit; if the second number and the comma
6246 are both omitted, the quantifier specifies an exact number of required
6251 matches at least 3 successive vowels, but may match many more, while
6255 matches exactly 8 digits. An opening curly bracket that appears in a
6256 position where a quantifier is not allowed, or one that does not match
6257 the syntax of a quantifier, is taken as a literal character. For exam-
6258 ple, {,6} is not a quantifier, but a literal string of four characters.
6260 In UTF modes, quantifiers apply to characters rather than to individual
6261 data units. Thus, for example, \x{100}{2} matches two characters, each
6262 of which is represented by a two-byte sequence in a UTF-8 string. Simi-
6263 larly, \X{3} matches three Unicode extended grapheme clusters, each of
6264 which may be several data units long (and they may be of different
6267 The quantifier {0} is permitted, causing the expression to behave as if
6268 the previous item and the quantifier were not present. This may be use-
6269 ful for subpatterns that are referenced as subroutines from elsewhere
6270 in the pattern (but see also the section entitled "Defining subpatterns
6271 for use by reference only" below). Items other than subpatterns that
6272 have a {0} quantifier are omitted from the compiled pattern.
6274 For convenience, the three most common quantifiers have single-charac-
6277 * is equivalent to {0,}
6278 + is equivalent to {1,}
6279 ? is equivalent to {0,1}
6281 It is possible to construct infinite loops by following a subpattern
6282 that can match no characters with a quantifier that has no upper limit,
6287 Earlier versions of Perl and PCRE used to give an error at compile time
6288 for such patterns. However, because there are cases where this can be
6289 useful, such patterns are now accepted, but if any repetition of the
6290 subpattern does in fact match no characters, the loop is forcibly bro-
6293 By default, the quantifiers are "greedy", that is, they match as much
6294 as possible (up to the maximum number of permitted times), without
6295 causing the rest of the pattern to fail. The classic example of where
6296 this gives problems is in trying to match comments in C programs. These
6297 appear between /* and */ and within the comment, individual * and /
6298 characters may appear. An attempt to match C comments by applying the
6305 /* first comment */ not comment /* second comment */
6307 fails, because it matches the entire string owing to the greediness of
6310 However, if a quantifier is followed by a question mark, it ceases to
6311 be greedy, and instead matches the minimum number of times possible, so
6316 does the right thing with the C comments. The meaning of the various
6317 quantifiers is not otherwise changed, just the preferred number of
6318 matches. Do not confuse this use of question mark with its use as a
6319 quantifier in its own right. Because it has two uses, it can sometimes
6320 appear doubled, as in
6324 which matches one digit by preference, but can match two if that is the
6325 only way the rest of the pattern matches.
6327 If the PCRE_UNGREEDY option is set (an option that is not available in
6328 Perl), the quantifiers are not greedy by default, but individual ones
6329 can be made greedy by following them with a question mark. In other
6330 words, it inverts the default behaviour.
6332 When a parenthesized subpattern is quantified with a minimum repeat
6333 count that is greater than 1 or with a limited maximum, more memory is
6334 required for the compiled pattern, in proportion to the size of the
6337 If a pattern starts with .* or .{0,} and the PCRE_DOTALL option (equiv-
6338 alent to Perl's /s) is set, thus allowing the dot to match newlines,
6339 the pattern is implicitly anchored, because whatever follows will be
6340 tried against every character position in the subject string, so there
6341 is no point in retrying the overall match at any position after the
6342 first. PCRE normally treats such a pattern as though it were preceded
6345 In cases where it is known that the subject string contains no new-
6346 lines, it is worth setting PCRE_DOTALL in order to obtain this opti-
6347 mization, or alternatively using ^ to indicate anchoring explicitly.
6349 However, there are some cases where the optimization cannot be used.
6350 When .* is inside capturing parentheses that are the subject of a back
6351 reference elsewhere in the pattern, a match at the start may fail where
6352 a later one succeeds. Consider, for example:
6356 If the subject is "xyz123abc123" the match point is the fourth charac-
6357 ter. For this reason, such a pattern is not implicitly anchored.
6359 Another case where implicit anchoring is not applied is when the lead-
6360 ing .* is inside an atomic group. Once again, a match at the start may
6361 fail where a later one succeeds. Consider this pattern:
6365 It matches "ab" in the subject "aab". The use of the backtracking con-
6366 trol verbs (*PRUNE) and (*SKIP) also disable this optimization.
6368 When a capturing subpattern is repeated, the value captured is the sub-
6369 string that matched the final iteration. For example, after
6371 (tweedle[dume]{3}\s*)+
6373 has matched "tweedledum tweedledee" the value of the captured substring
6374 is "tweedledee". However, if there are nested capturing subpatterns,
6375 the corresponding captured values may have been set in previous itera-
6376 tions. For example, after
6380 matches "aba" the value of the second captured substring is "b".
6383 ATOMIC GROUPING AND POSSESSIVE QUANTIFIERS
6385 With both maximizing ("greedy") and minimizing ("ungreedy" or "lazy")
6386 repetition, failure of what follows normally causes the repeated item
6387 to be re-evaluated to see if a different number of repeats allows the
6388 rest of the pattern to match. Sometimes it is useful to prevent this,
6389 either to change the nature of the match, or to cause it fail earlier
6390 than it otherwise might, when the author of the pattern knows there is
6391 no point in carrying on.
6393 Consider, for example, the pattern \d+foo when applied to the subject
6398 After matching all 6 digits and then failing to match "foo", the normal
6399 action of the matcher is to try again with only 5 digits matching the
6400 \d+ item, and then with 4, and so on, before ultimately failing.
6401 "Atomic grouping" (a term taken from Jeffrey Friedl's book) provides
6402 the means for specifying that once a subpattern has matched, it is not
6403 to be re-evaluated in this way.
6405 If we use atomic grouping for the previous example, the matcher gives
6406 up immediately on failing to match "foo" the first time. The notation
6407 is a kind of special parenthesis, starting with (?> as in this example:
6411 This kind of parenthesis "locks up" the part of the pattern it con-
6412 tains once it has matched, and a failure further into the pattern is
6413 prevented from backtracking into it. Backtracking past it to previous
6414 items, however, works as normal.
6416 An alternative description is that a subpattern of this type matches
6417 the string of characters that an identical standalone pattern would
6418 match, if anchored at the current point in the subject string.
6420 Atomic grouping subpatterns are not capturing subpatterns. Simple cases
6421 such as the above example can be thought of as a maximizing repeat that
6422 must swallow everything it can. So, while both \d+ and \d+? are pre-
6423 pared to adjust the number of digits they match in order to make the
6424 rest of the pattern match, (?>\d+) can only match an entire sequence of
6427 Atomic groups in general can of course contain arbitrarily complicated
6428 subpatterns, and can be nested. However, when the subpattern for an
6429 atomic group is just a single repeated item, as in the example above, a
6430 simpler notation, called a "possessive quantifier" can be used. This
6431 consists of an additional + character following a quantifier. Using
6432 this notation, the previous example can be rewritten as
6436 Note that a possessive quantifier can be used with an entire group, for
6441 Possessive quantifiers are always greedy; the setting of the
6442 PCRE_UNGREEDY option is ignored. They are a convenient notation for the
6443 simpler forms of atomic group. However, there is no difference in the
6444 meaning of a possessive quantifier and the equivalent atomic group,
6445 though there may be a performance difference; possessive quantifiers
6446 should be slightly faster.
6448 The possessive quantifier syntax is an extension to the Perl 5.8 syn-
6449 tax. Jeffrey Friedl originated the idea (and the name) in the first
6450 edition of his book. Mike McCloskey liked it, so implemented it when he
6451 built Sun's Java package, and PCRE copied it from there. It ultimately
6452 found its way into Perl at release 5.10.
6454 PCRE has an optimization that automatically "possessifies" certain sim-
6455 ple pattern constructs. For example, the sequence A+B is treated as
6456 A++B because there is no point in backtracking into a sequence of A's
6459 When a pattern contains an unlimited repeat inside a subpattern that
6460 can itself be repeated an unlimited number of times, the use of an
6461 atomic group is the only way to avoid some failing matches taking a
6462 very long time indeed. The pattern
6466 matches an unlimited number of substrings that either consist of non-
6467 digits, or digits enclosed in <>, followed by either ! or ?. When it
6468 matches, it runs quickly. However, if it is applied to
6470 aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
6472 it takes a long time before reporting failure. This is because the
6473 string can be divided between the internal \D+ repeat and the external
6474 * repeat in a large number of ways, and all have to be tried. (The
6475 example uses [!?] rather than a single character at the end, because
6476 both PCRE and Perl have an optimization that allows for fast failure
6477 when a single character is used. They remember the last single charac-
6478 ter that is required for a match, and fail early if it is not present
6479 in the string.) If the pattern is changed so that it uses an atomic
6482 ((?>\D+)|<\d+>)*[!?]
6484 sequences of non-digits cannot be broken, and failure happens quickly.
6489 Outside a character class, a backslash followed by a digit greater than
6490 0 (and possibly further digits) is a back reference to a capturing sub-
6491 pattern earlier (that is, to its left) in the pattern, provided there
6492 have been that many previous capturing left parentheses.
6494 However, if the decimal number following the backslash is less than 10,
6495 it is always taken as a back reference, and causes an error only if
6496 there are not that many capturing left parentheses in the entire pat-
6497 tern. In other words, the parentheses that are referenced need not be
6498 to the left of the reference for numbers less than 10. A "forward back
6499 reference" of this type can make sense when a repetition is involved
6500 and the subpattern to the right has participated in an earlier itera-
6503 It is not possible to have a numerical "forward back reference" to a
6504 subpattern whose number is 10 or more using this syntax because a
6505 sequence such as \50 is interpreted as a character defined in octal.
6506 See the subsection entitled "Non-printing characters" above for further
6507 details of the handling of digits following a backslash. There is no
6508 such problem when named parentheses are used. A back reference to any
6509 subpattern is possible using named parentheses (see below).
6511 Another way of avoiding the ambiguity inherent in the use of digits
6512 following a backslash is to use the \g escape sequence. This escape
6513 must be followed by an unsigned number or a negative number, optionally
6514 enclosed in braces. These examples are all identical:
6520 An unsigned number specifies an absolute reference without the ambigu-
6521 ity that is present in the older syntax. It is also useful when literal
6522 digits follow the reference. A negative number is a relative reference.
6523 Consider this example:
6527 The sequence \g{-1} is a reference to the most recently started captur-
6528 ing subpattern before \g, that is, is it equivalent to \2 in this exam-
6529 ple. Similarly, \g{-2} would be equivalent to \1. The use of relative
6530 references can be helpful in long patterns, and also in patterns that
6531 are created by joining together fragments that contain references
6534 A back reference matches whatever actually matched the capturing sub-
6535 pattern in the current subject string, rather than anything matching
6536 the subpattern itself (see "Subpatterns as subroutines" below for a way
6537 of doing that). So the pattern
6539 (sens|respons)e and \1ibility
6541 matches "sense and sensibility" and "response and responsibility", but
6542 not "sense and responsibility". If caseful matching is in force at the
6543 time of the back reference, the case of letters is relevant. For exam-
6548 matches "rah rah" and "RAH RAH", but not "RAH rah", even though the
6549 original capturing subpattern is matched caselessly.
6551 There are several different ways of writing back references to named
6552 subpatterns. The .NET syntax \k{name} and the Perl syntax \k<name> or
6553 \k'name' are supported, as is the Python syntax (?P=name). Perl 5.10's
6554 unified back reference syntax, in which \g can be used for both numeric
6555 and named references, is also supported. We could rewrite the above
6556 example in any of the following ways:
6558 (?<p1>(?i)rah)\s+\k<p1>
6559 (?'p1'(?i)rah)\s+\k{p1}
6560 (?P<p1>(?i)rah)\s+(?P=p1)
6561 (?<p1>(?i)rah)\s+\g{p1}
6563 A subpattern that is referenced by name may appear in the pattern
6564 before or after the reference.
6566 There may be more than one back reference to the same subpattern. If a
6567 subpattern has not actually been used in a particular match, any back
6568 references to it always fail by default. For example, the pattern
6572 always fails if it starts to match "a" rather than "bc". However, if
6573 the PCRE_JAVASCRIPT_COMPAT option is set at compile time, a back refer-
6574 ence to an unset value matches an empty string.
6576 Because there may be many capturing parentheses in a pattern, all dig-
6577 its following a backslash are taken as part of a potential back refer-
6578 ence number. If the pattern continues with a digit character, some
6579 delimiter must be used to terminate the back reference. If the
6580 PCRE_EXTENDED option is set, this can be white space. Otherwise, the
6581 \g{ syntax or an empty comment (see "Comments" below) can be used.
6583 Recursive back references
6585 A back reference that occurs inside the parentheses to which it refers
6586 fails when the subpattern is first used, so, for example, (a\1) never
6587 matches. However, such references can be useful inside repeated sub-
6588 patterns. For example, the pattern
6592 matches any number of "a"s and also "aba", "ababbaa" etc. At each iter-
6593 ation of the subpattern, the back reference matches the character
6594 string corresponding to the previous iteration. In order for this to
6595 work, the pattern must be such that the first iteration does not need
6596 to match the back reference. This can be done using alternation, as in
6597 the example above, or by a quantifier with a minimum of zero.
6599 Back references of this type cause the group that they reference to be
6600 treated as an atomic group. Once the whole group has been matched, a
6601 subsequent matching failure cannot cause backtracking into the middle
6607 An assertion is a test on the characters following or preceding the
6608 current matching point that does not actually consume any characters.
6609 The simple assertions coded as \b, \B, \A, \G, \Z, \z, ^ and $ are
6612 More complicated assertions are coded as subpatterns. There are two
6613 kinds: those that look ahead of the current position in the subject
6614 string, and those that look behind it. An assertion subpattern is
6615 matched in the normal way, except that it does not cause the current
6616 matching position to be changed.
6618 Assertion subpatterns are not capturing subpatterns. If such an asser-
6619 tion contains capturing subpatterns within it, these are counted for
6620 the purposes of numbering the capturing subpatterns in the whole pat-
6621 tern. However, substring capturing is carried out only for positive
6622 assertions. (Perl sometimes, but not always, does do capturing in nega-
6625 For compatibility with Perl, assertion subpatterns may be repeated;
6626 though it makes no sense to assert the same thing several times, the
6627 side effect of capturing parentheses may occasionally be useful. In
6628 practice, there only three cases:
6630 (1) If the quantifier is {0}, the assertion is never obeyed during
6631 matching. However, it may contain internal capturing parenthesized
6632 groups that are called from elsewhere via the subroutine mechanism.
6634 (2) If quantifier is {0,n} where n is greater than zero, it is treated
6635 as if it were {0,1}. At run time, the rest of the pattern match is
6636 tried with and without the assertion, the order depending on the greed-
6637 iness of the quantifier.
6639 (3) If the minimum repetition is greater than zero, the quantifier is
6640 ignored. The assertion is obeyed just once when encountered during
6643 Lookahead assertions
6645 Lookahead assertions start with (?= for positive assertions and (?! for
6646 negative assertions. For example,
6650 matches a word followed by a semicolon, but does not include the semi-
6651 colon in the match, and
6655 matches any occurrence of "foo" that is not followed by "bar". Note
6656 that the apparently similar pattern
6660 does not find an occurrence of "bar" that is preceded by something
6661 other than "foo"; it finds any occurrence of "bar" whatsoever, because
6662 the assertion (?!foo) is always true when the next three characters are
6663 "bar". A lookbehind assertion is needed to achieve the other effect.
6665 If you want to force a matching failure at some point in a pattern, the
6666 most convenient way to do it is with (?!) because an empty string
6667 always matches, so an assertion that requires there not to be an empty
6668 string must always fail. The backtracking control verb (*FAIL) or (*F)
6669 is a synonym for (?!).
6671 Lookbehind assertions
6673 Lookbehind assertions start with (?<= for positive assertions and (?<!
6674 for negative assertions. For example,
6678 does find an occurrence of "bar" that is not preceded by "foo". The
6679 contents of a lookbehind assertion are restricted such that all the
6680 strings it matches must have a fixed length. However, if there are sev-
6681 eral top-level alternatives, they do not all have to have the same
6690 causes an error at compile time. Branches that match different length
6691 strings are permitted only at the top level of a lookbehind assertion.
6692 This is an extension compared with Perl, which requires all branches to
6693 match the same length of string. An assertion such as
6697 is not permitted, because its single top-level branch can match two
6698 different lengths, but it is acceptable to PCRE if rewritten to use two
6703 In some cases, the escape sequence \K (see above) can be used instead
6704 of a lookbehind assertion to get round the fixed-length restriction.
6706 The implementation of lookbehind assertions is, for each alternative,
6707 to temporarily move the current position back by the fixed length and
6708 then try to match. If there are insufficient characters before the cur-
6709 rent position, the assertion fails.
6711 In a UTF mode, PCRE does not allow the \C escape (which matches a sin-
6712 gle data unit even in a UTF mode) to appear in lookbehind assertions,
6713 because it makes it impossible to calculate the length of the lookbe-
6714 hind. The \X and \R escapes, which can match different numbers of data
6715 units, are also not permitted.
6717 "Subroutine" calls (see below) such as (?2) or (?&X) are permitted in
6718 lookbehinds, as long as the subpattern matches a fixed-length string.
6719 Recursion, however, is not supported.
6721 Possessive quantifiers can be used in conjunction with lookbehind
6722 assertions to specify efficient matching of fixed-length strings at the
6723 end of subject strings. Consider a simple pattern such as
6727 when applied to a long string that does not match. Because matching
6728 proceeds from left to right, PCRE will look for each "a" in the subject
6729 and then see if what follows matches the rest of the pattern. If the
6730 pattern is specified as
6734 the initial .* matches the entire string at first, but when this fails
6735 (because there is no following "a"), it backtracks to match all but the
6736 last character, then all but the last two characters, and so on. Once
6737 again the search for "a" covers the entire string, from right to left,
6738 so we are no better off. However, if the pattern is written as
6742 there can be no backtracking for the .*+ item; it can match only the
6743 entire string. The subsequent lookbehind assertion does a single test
6744 on the last four characters. If it fails, the match fails immediately.
6745 For long strings, this approach makes a significant difference to the
6748 Using multiple assertions
6750 Several assertions (of any sort) may occur in succession. For example,
6752 (?<=\d{3})(?<!999)foo
6754 matches "foo" preceded by three digits that are not "999". Notice that
6755 each of the assertions is applied independently at the same point in
6756 the subject string. First there is a check that the previous three
6757 characters are all digits, and then there is a check that the same
6758 three characters are not "999". This pattern does not match "foo" pre-
6759 ceded by six characters, the first of which are digits and the last
6760 three of which are not "999". For example, it doesn't match "123abc-
6761 foo". A pattern to do that is
6763 (?<=\d{3}...)(?<!999)foo
6765 This time the first assertion looks at the preceding six characters,
6766 checking that the first three are digits, and then the second assertion
6767 checks that the preceding three characters are not "999".
6769 Assertions can be nested in any combination. For example,
6773 matches an occurrence of "baz" that is preceded by "bar" which in turn
6774 is not preceded by "foo", while
6776 (?<=\d{3}(?!999)...)foo
6778 is another pattern that matches "foo" preceded by three digits and any
6779 three characters that are not "999".
6782 CONDITIONAL SUBPATTERNS
6784 It is possible to cause the matching process to obey a subpattern con-
6785 ditionally or to choose between two alternative subpatterns, depending
6786 on the result of an assertion, or whether a specific capturing subpat-
6787 tern has already been matched. The two possible forms of conditional
6790 (?(condition)yes-pattern)
6791 (?(condition)yes-pattern|no-pattern)
6793 If the condition is satisfied, the yes-pattern is used; otherwise the
6794 no-pattern (if present) is used. If there are more than two alterna-
6795 tives in the subpattern, a compile-time error occurs. Each of the two
6796 alternatives may itself contain nested subpatterns of any form, includ-
6797 ing conditional subpatterns; the restriction to two alternatives
6798 applies only at the level of the condition. This pattern fragment is an
6799 example where the alternatives are complex:
6801 (?(1) (A|B|C) | (D | (?(2)E|F) | E) )
6804 There are four kinds of condition: references to subpatterns, refer-
6805 ences to recursion, a pseudo-condition called DEFINE, and assertions.
6807 Checking for a used subpattern by number
6809 If the text between the parentheses consists of a sequence of digits,
6810 the condition is true if a capturing subpattern of that number has pre-
6811 viously matched. If there is more than one capturing subpattern with
6812 the same number (see the earlier section about duplicate subpattern
6813 numbers), the condition is true if any of them have matched. An alter-
6814 native notation is to precede the digits with a plus or minus sign. In
6815 this case, the subpattern number is relative rather than absolute. The
6816 most recently opened parentheses can be referenced by (?(-1), the next
6817 most recent by (?(-2), and so on. Inside loops it can also make sense
6818 to refer to subsequent groups. The next parentheses to be opened can be
6819 referenced as (?(+1), and so on. (The value zero in any of these forms
6820 is not used; it provokes a compile-time error.)
6822 Consider the following pattern, which contains non-significant white
6823 space to make it more readable (assume the PCRE_EXTENDED option) and to
6824 divide it into three parts for ease of discussion:
6826 ( \( )? [^()]+ (?(1) \) )
6828 The first part matches an optional opening parenthesis, and if that
6829 character is present, sets it as the first captured substring. The sec-
6830 ond part matches one or more characters that are not parentheses. The
6831 third part is a conditional subpattern that tests whether or not the
6832 first set of parentheses matched. If they did, that is, if subject
6833 started with an opening parenthesis, the condition is true, and so the
6834 yes-pattern is executed and a closing parenthesis is required. Other-
6835 wise, since no-pattern is not present, the subpattern matches nothing.
6836 In other words, this pattern matches a sequence of non-parentheses,
6837 optionally enclosed in parentheses.
6839 If you were embedding this pattern in a larger one, you could use a
6842 ...other stuff... ( \( )? [^()]+ (?(-1) \) ) ...
6844 This makes the fragment independent of the parentheses in the larger
6847 Checking for a used subpattern by name
6849 Perl uses the syntax (?(<name>)...) or (?('name')...) to test for a
6850 used subpattern by name. For compatibility with earlier versions of
6851 PCRE, which had this facility before Perl, the syntax (?(name)...) is
6854 Rewriting the above example to use a named subpattern gives this:
6856 (?<OPEN> \( )? [^()]+ (?(<OPEN>) \) )
6858 If the name used in a condition of this kind is a duplicate, the test
6859 is applied to all subpatterns of the same name, and is true if any one
6860 of them has matched.
6862 Checking for pattern recursion
6864 If the condition is the string (R), and there is no subpattern with the
6865 name R, the condition is true if a recursive call to the whole pattern
6866 or any subpattern has been made. If digits or a name preceded by amper-
6867 sand follow the letter R, for example:
6869 (?(R3)...) or (?(R&name)...)
6871 the condition is true if the most recent recursion is into a subpattern
6872 whose number or name is given. This condition does not check the entire
6873 recursion stack. If the name used in a condition of this kind is a
6874 duplicate, the test is applied to all subpatterns of the same name, and
6875 is true if any one of them is the most recent recursion.
6877 At "top level", all these recursion test conditions are false. The
6878 syntax for recursive patterns is described below.
6880 Defining subpatterns for use by reference only
6882 If the condition is the string (DEFINE), and there is no subpattern
6883 with the name DEFINE, the condition is always false. In this case,
6884 there may be only one alternative in the subpattern. It is always
6885 skipped if control reaches this point in the pattern; the idea of
6886 DEFINE is that it can be used to define subroutines that can be refer-
6887 enced from elsewhere. (The use of subroutines is described below.) For
6888 example, a pattern to match an IPv4 address such as "192.168.23.245"
6889 could be written like this (ignore white space and line breaks):
6891 (?(DEFINE) (?<byte> 2[0-4]\d | 25[0-5] | 1\d\d | [1-9]?\d) )
6892 \b (?&byte) (\.(?&byte)){3} \b
6894 The first part of the pattern is a DEFINE group inside which a another
6895 group named "byte" is defined. This matches an individual component of
6896 an IPv4 address (a number less than 256). When matching takes place,
6897 this part of the pattern is skipped because DEFINE acts like a false
6898 condition. The rest of the pattern uses references to the named group
6899 to match the four dot-separated components of an IPv4 address, insist-
6900 ing on a word boundary at each end.
6902 Assertion conditions
6904 If the condition is not in any of the above formats, it must be an
6905 assertion. This may be a positive or negative lookahead or lookbehind
6906 assertion. Consider this pattern, again containing non-significant
6907 white space, and with the two alternatives on the second line:
6910 \d{2}-[a-z]{3}-\d{2} | \d{2}-\d{2}-\d{2} )
6912 The condition is a positive lookahead assertion that matches an
6913 optional sequence of non-letters followed by a letter. In other words,
6914 it tests for the presence of at least one letter in the subject. If a
6915 letter is found, the subject is matched against the first alternative;
6916 otherwise it is matched against the second. This pattern matches
6917 strings in one of the two forms dd-aaa-dd or dd-dd-dd, where aaa are
6918 letters and dd are digits.
6923 There are two ways of including comments in patterns that are processed
6924 by PCRE. In both cases, the start of the comment must not be in a char-
6925 acter class, nor in the middle of any other sequence of related charac-
6926 ters such as (?: or a subpattern name or number. The characters that
6927 make up a comment play no part in the pattern matching.
6929 The sequence (?# marks the start of a comment that continues up to the
6930 next closing parenthesis. Nested parentheses are not permitted. If the
6931 PCRE_EXTENDED option is set, an unescaped # character also introduces a
6932 comment, which in this case continues to immediately after the next
6933 newline character or character sequence in the pattern. Which charac-
6934 ters are interpreted as newlines is controlled by the options passed to
6935 a compiling function or by a special sequence at the start of the pat-
6936 tern, as described in the section entitled "Newline conventions" above.
6937 Note that the end of this type of comment is a literal newline sequence
6938 in the pattern; escape sequences that happen to represent a newline do
6939 not count. For example, consider this pattern when PCRE_EXTENDED is
6940 set, and the default newline convention is in force:
6942 abc #comment \n still comment
6944 On encountering the # character, pcre_compile() skips along, looking
6945 for a newline in the pattern. The sequence \n is still literal at this
6946 stage, so it does not terminate the comment. Only an actual character
6947 with the code value 0x0a (the default newline) does so.
6952 Consider the problem of matching a string in parentheses, allowing for
6953 unlimited nested parentheses. Without the use of recursion, the best
6954 that can be done is to use a pattern that matches up to some fixed
6955 depth of nesting. It is not possible to handle an arbitrary nesting
6958 For some time, Perl has provided a facility that allows regular expres-
6959 sions to recurse (amongst other things). It does this by interpolating
6960 Perl code in the expression at run time, and the code can refer to the
6961 expression itself. A Perl pattern using code interpolation to solve the
6962 parentheses problem can be created like this:
6964 $re = qr{\( (?: (?>[^()]+) | (?p{$re}) )* \)}x;
6966 The (?p{...}) item interpolates Perl code at run time, and in this case
6967 refers recursively to the pattern in which it appears.
6969 Obviously, PCRE cannot support the interpolation of Perl code. Instead,
6970 it supports special syntax for recursion of the entire pattern, and
6971 also for individual subpattern recursion. After its introduction in
6972 PCRE and Python, this kind of recursion was subsequently introduced
6973 into Perl at release 5.10.
6975 A special item that consists of (? followed by a number greater than
6976 zero and a closing parenthesis is a recursive subroutine call of the
6977 subpattern of the given number, provided that it occurs inside that
6978 subpattern. (If not, it is a non-recursive subroutine call, which is
6979 described in the next section.) The special item (?R) or (?0) is a
6980 recursive call of the entire regular expression.
6982 This PCRE pattern solves the nested parentheses problem (assume the
6983 PCRE_EXTENDED option is set so that white space is ignored):
6985 \( ( [^()]++ | (?R) )* \)
6987 First it matches an opening parenthesis. Then it matches any number of
6988 substrings which can either be a sequence of non-parentheses, or a
6989 recursive match of the pattern itself (that is, a correctly parenthe-
6990 sized substring). Finally there is a closing parenthesis. Note the use
6991 of a possessive quantifier to avoid backtracking into sequences of non-
6994 If this were part of a larger pattern, you would not want to recurse
6995 the entire pattern, so instead you could use this:
6997 ( \( ( [^()]++ | (?1) )* \) )
6999 We have put the pattern into parentheses, and caused the recursion to
7000 refer to them instead of the whole pattern.
7002 In a larger pattern, keeping track of parenthesis numbers can be
7003 tricky. This is made easier by the use of relative references. Instead
7004 of (?1) in the pattern above you can write (?-2) to refer to the second
7005 most recently opened parentheses preceding the recursion. In other
7006 words, a negative number counts capturing parentheses leftwards from
7007 the point at which it is encountered.
7009 It is also possible to refer to subsequently opened parentheses, by
7010 writing references such as (?+2). However, these cannot be recursive
7011 because the reference is not inside the parentheses that are refer-
7012 enced. They are always non-recursive subroutine calls, as described in
7015 An alternative approach is to use named parentheses instead. The Perl
7016 syntax for this is (?&name); PCRE's earlier syntax (?P>name) is also
7017 supported. We could rewrite the above example as follows:
7019 (?<pn> \( ( [^()]++ | (?&pn) )* \) )
7021 If there is more than one subpattern with the same name, the earliest
7024 This particular example pattern that we have been looking at contains
7025 nested unlimited repeats, and so the use of a possessive quantifier for
7026 matching strings of non-parentheses is important when applying the pat-
7027 tern to strings that do not match. For example, when this pattern is
7030 (aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa()
7032 it yields "no match" quickly. However, if a possessive quantifier is
7033 not used, the match runs for a very long time indeed because there are
7034 so many different ways the + and * repeats can carve up the subject,
7035 and all have to be tested before failure can be reported.
7037 At the end of a match, the values of capturing parentheses are those
7038 from the outermost level. If you want to obtain intermediate values, a
7039 callout function can be used (see below and the pcrecallout documenta-
7040 tion). If the pattern above is matched against
7044 the value for the inner capturing parentheses (numbered 2) is "ef",
7045 which is the last value taken on at the top level. If a capturing sub-
7046 pattern is not matched at the top level, its final captured value is
7047 unset, even if it was (temporarily) set at a deeper level during the
7050 If there are more than 15 capturing parentheses in a pattern, PCRE has
7051 to obtain extra memory to store data during a recursion, which it does
7052 by using pcre_malloc, freeing it via pcre_free afterwards. If no memory
7053 can be obtained, the match fails with the PCRE_ERROR_NOMEMORY error.
7055 Do not confuse the (?R) item with the condition (R), which tests for
7056 recursion. Consider this pattern, which matches text in angle brack-
7057 ets, allowing for arbitrary nesting. Only digits are allowed in nested
7058 brackets (that is, when recursing), whereas any characters are permit-
7059 ted at the outer level.
7061 < (?: (?(R) \d++ | [^<>]*+) | (?R)) * >
7063 In this pattern, (?(R) is the start of a conditional subpattern, with
7064 two different alternatives for the recursive and non-recursive cases.
7065 The (?R) item is the actual recursive call.
7067 Differences in recursion processing between PCRE and Perl
7069 Recursion processing in PCRE differs from Perl in two important ways.
7070 In PCRE (like Python, but unlike Perl), a recursive subpattern call is
7071 always treated as an atomic group. That is, once it has matched some of
7072 the subject string, it is never re-entered, even if it contains untried
7073 alternatives and there is a subsequent matching failure. This can be
7074 illustrated by the following pattern, which purports to match a palin-
7075 dromic string that contains an odd number of characters (for example,
7076 "a", "aba", "abcba", "abcdcba"):
7080 The idea is that it either matches a single character, or two identical
7081 characters surrounding a sub-palindrome. In Perl, this pattern works;
7082 in PCRE it does not if the pattern is longer than three characters.
7083 Consider the subject string "abcba":
7085 At the top level, the first character is matched, but as it is not at
7086 the end of the string, the first alternative fails; the second alterna-
7087 tive is taken and the recursion kicks in. The recursive call to subpat-
7088 tern 1 successfully matches the next character ("b"). (Note that the
7089 beginning and end of line tests are not part of the recursion).
7091 Back at the top level, the next character ("c") is compared with what
7092 subpattern 2 matched, which was "a". This fails. Because the recursion
7093 is treated as an atomic group, there are now no backtracking points,
7094 and so the entire match fails. (Perl is able, at this point, to re-
7095 enter the recursion and try the second alternative.) However, if the
7096 pattern is written with the alternatives in the other order, things are
7101 This time, the recursing alternative is tried first, and continues to
7102 recurse until it runs out of characters, at which point the recursion
7103 fails. But this time we do have another alternative to try at the
7104 higher level. That is the big difference: in the previous case the
7105 remaining alternative is at a deeper recursion level, which PCRE cannot
7108 To change the pattern so that it matches all palindromic strings, not
7109 just those with an odd number of characters, it is tempting to change
7110 the pattern to this:
7114 Again, this works in Perl, but not in PCRE, and for the same reason.
7115 When a deeper recursion has matched a single character, it cannot be
7116 entered again in order to match an empty string. The solution is to
7117 separate the two cases, and write out the odd and even cases as alter-
7118 natives at the higher level:
7120 ^(?:((.)(?1)\2|)|((.)(?3)\4|.))
7122 If you want to match typical palindromic phrases, the pattern has to
7123 ignore all non-word characters, which can be done like this:
7125 ^\W*+(?:((.)\W*+(?1)\W*+\2|)|((.)\W*+(?3)\W*+\4|\W*+.\W*+))\W*+$
7127 If run with the PCRE_CASELESS option, this pattern matches phrases such
7128 as "A man, a plan, a canal: Panama!" and it works well in both PCRE and
7129 Perl. Note the use of the possessive quantifier *+ to avoid backtrack-
7130 ing into sequences of non-word characters. Without this, PCRE takes a
7131 great deal longer (ten times or more) to match typical phrases, and
7132 Perl takes so long that you think it has gone into a loop.
7134 WARNING: The palindrome-matching patterns above work only if the sub-
7135 ject string does not start with a palindrome that is shorter than the
7136 entire string. For example, although "abcba" is correctly matched, if
7137 the subject is "ababa", PCRE finds the palindrome "aba" at the start,
7138 then fails at top level because the end of the string does not follow.
7139 Once again, it cannot jump back into the recursion to try other alter-
7140 natives, so the entire match fails.
7142 The second way in which PCRE and Perl differ in their recursion pro-
7143 cessing is in the handling of captured values. In Perl, when a subpat-
7144 tern is called recursively or as a subpattern (see the next section),
7145 it has no access to any values that were captured outside the recur-
7146 sion, whereas in PCRE these values can be referenced. Consider this
7151 In PCRE, this pattern matches "bab". The first capturing parentheses
7152 match "b", then in the second group, when the back reference \1 fails
7153 to match "b", the second alternative matches "a" and then recurses. In
7154 the recursion, \1 does now match "b" and so the whole match succeeds.
7155 In Perl, the pattern fails to match because inside the recursive call
7156 \1 cannot access the externally set value.
7159 SUBPATTERNS AS SUBROUTINES
7161 If the syntax for a recursive subpattern call (either by number or by
7162 name) is used outside the parentheses to which it refers, it operates
7163 like a subroutine in a programming language. The called subpattern may
7164 be defined before or after the reference. A numbered reference can be
7165 absolute or relative, as in these examples:
7167 (...(absolute)...)...(?2)...
7168 (...(relative)...)...(?-1)...
7169 (...(?+1)...(relative)...
7171 An earlier example pointed out that the pattern
7173 (sens|respons)e and \1ibility
7175 matches "sense and sensibility" and "response and responsibility", but
7176 not "sense and responsibility". If instead the pattern
7178 (sens|respons)e and (?1)ibility
7180 is used, it does match "sense and responsibility" as well as the other
7181 two strings. Another example is given in the discussion of DEFINE
7184 All subroutine calls, whether recursive or not, are always treated as
7185 atomic groups. That is, once a subroutine has matched some of the sub-
7186 ject string, it is never re-entered, even if it contains untried alter-
7187 natives and there is a subsequent matching failure. Any capturing
7188 parentheses that are set during the subroutine call revert to their
7189 previous values afterwards.
7191 Processing options such as case-independence are fixed when a subpat-
7192 tern is defined, so if it is used as a subroutine, such options cannot
7193 be changed for different calls. For example, consider this pattern:
7197 It matches "abcabc". It does not match "abcABC" because the change of
7198 processing option does not affect the called subpattern.
7201 ONIGURUMA SUBROUTINE SYNTAX
7203 For compatibility with Oniguruma, the non-Perl syntax \g followed by a
7204 name or a number enclosed either in angle brackets or single quotes, is
7205 an alternative syntax for referencing a subpattern as a subroutine,
7206 possibly recursively. Here are two of the examples used above, rewrit-
7207 ten using this syntax:
7209 (?<pn> \( ( (?>[^()]+) | \g<pn> )* \) )
7210 (sens|respons)e and \g'1'ibility
7212 PCRE supports an extension to Oniguruma: if a number is preceded by a
7213 plus or a minus sign it is taken as a relative reference. For example:
7217 Note that \g{...} (Perl syntax) and \g<...> (Oniguruma syntax) are not
7218 synonymous. The former is a back reference; the latter is a subroutine
7224 Perl has a feature whereby using the sequence (?{...}) causes arbitrary
7225 Perl code to be obeyed in the middle of matching a regular expression.
7226 This makes it possible, amongst other things, to extract different sub-
7227 strings that match the same pair of parentheses when there is a repeti-
7230 PCRE provides a similar feature, but of course it cannot obey arbitrary
7231 Perl code. The feature is called "callout". The caller of PCRE provides
7232 an external function by putting its entry point in the global variable
7233 pcre_callout (8-bit library) or pcre[16|32]_callout (16-bit or 32-bit
7234 library). By default, this variable contains NULL, which disables all
7237 Within a regular expression, (?C) indicates the points at which the
7238 external function is to be called. If you want to identify different
7239 callout points, you can put a number less than 256 after the letter C.
7240 The default value is zero. For example, this pattern has two callout
7245 If the PCRE_AUTO_CALLOUT flag is passed to a compiling function, call-
7246 outs are automatically installed before each item in the pattern. They
7247 are all numbered 255. If there is a conditional group in the pattern
7248 whose condition is an assertion, an additional callout is inserted just
7249 before the condition. An explicit callout may also be set at this posi-
7250 tion, as in this example:
7252 (?(?C9)(?=a)abc|def)
7254 Note that this applies only to assertion conditions, not to other types
7257 During matching, when PCRE reaches a callout point, the external func-
7258 tion is called. It is provided with the number of the callout, the
7259 position in the pattern, and, optionally, one item of data originally
7260 supplied by the caller of the matching function. The callout function
7261 may cause matching to proceed, to backtrack, or to fail altogether.
7263 By default, PCRE implements a number of optimizations at compile time
7264 and matching time, and one side-effect is that sometimes callouts are
7265 skipped. If you need all possible callouts to happen, you need to set
7266 options that disable the relevant optimizations. More details, and a
7267 complete description of the interface to the callout function, are
7268 given in the pcrecallout documentation.
7271 BACKTRACKING CONTROL
7273 Perl 5.10 introduced a number of "Special Backtracking Control Verbs",
7274 which are still described in the Perl documentation as "experimental
7275 and subject to change or removal in a future version of Perl". It goes
7276 on to say: "Their usage in production code should be noted to avoid
7277 problems during upgrades." The same remarks apply to the PCRE features
7278 described in this section.
7280 The new verbs make use of what was previously invalid syntax: an open-
7281 ing parenthesis followed by an asterisk. They are generally of the form
7282 (*VERB) or (*VERB:NAME). Some may take either form, possibly behaving
7283 differently depending on whether or not a name is present. A name is
7284 any sequence of characters that does not include a closing parenthesis.
7285 The maximum length of name is 255 in the 8-bit library and 65535 in the
7286 16-bit and 32-bit libraries. If the name is empty, that is, if the
7287 closing parenthesis immediately follows the colon, the effect is as if
7288 the colon were not there. Any number of these verbs may occur in a
7291 Since these verbs are specifically related to backtracking, most of
7292 them can be used only when the pattern is to be matched using one of
7293 the traditional matching functions, because these use a backtracking
7294 algorithm. With the exception of (*FAIL), which behaves like a failing
7295 negative assertion, the backtracking control verbs cause an error if
7296 encountered by a DFA matching function.
7298 The behaviour of these verbs in repeated groups, assertions, and in
7299 subpatterns called as subroutines (whether or not recursively) is docu-
7302 Optimizations that affect backtracking verbs
7304 PCRE contains some optimizations that are used to speed up matching by
7305 running some checks at the start of each match attempt. For example, it
7306 may know the minimum length of matching subject, or that a particular
7307 character must be present. When one of these optimizations bypasses the
7308 running of a match, any included backtracking verbs will not, of
7309 course, be processed. You can suppress the start-of-match optimizations
7310 by setting the PCRE_NO_START_OPTIMIZE option when calling pcre_com-
7311 pile() or pcre_exec(), or by starting the pattern with (*NO_START_OPT).
7312 There is more discussion of this option in the section entitled "Option
7313 bits for pcre_exec()" in the pcreapi documentation.
7315 Experiments with Perl suggest that it too has similar optimizations,
7316 sometimes leading to anomalous results.
7318 Verbs that act immediately
7320 The following verbs act as soon as they are encountered. They may not
7321 be followed by a name.
7325 This verb causes the match to end successfully, skipping the remainder
7326 of the pattern. However, when it is inside a subpattern that is called
7327 as a subroutine, only that subpattern is ended successfully. Matching
7328 then continues at the outer level. If (*ACCEPT) in triggered in a posi-
7329 tive assertion, the assertion succeeds; in a negative assertion, the
7332 If (*ACCEPT) is inside capturing parentheses, the data so far is cap-
7335 A((?:A|B(*ACCEPT)|C)D)
7337 This matches "AB", "AAD", or "ACD"; when it matches "AB", "B" is cap-
7338 tured by the outer parentheses.
7342 This verb causes a matching failure, forcing backtracking to occur. It
7343 is equivalent to (?!) but easier to read. The Perl documentation notes
7344 that it is probably useful only when combined with (?{}) or (??{}).
7345 Those are, of course, Perl features that are not present in PCRE. The
7346 nearest equivalent is the callout feature, as for example in this pat-
7351 A match with the string "aaaa" always fails, but the callout is taken
7352 before each backtrack happens (in this example, 10 times).
7354 Recording which path was taken
7356 There is one verb whose main purpose is to track how a match was
7357 arrived at, though it also has a secondary use in conjunction with
7358 advancing the match starting point (see (*SKIP) below).
7360 (*MARK:NAME) or (*:NAME)
7362 A name is always required with this verb. There may be as many
7363 instances of (*MARK) as you like in a pattern, and their names do not
7366 When a match succeeds, the name of the last-encountered (*MARK:NAME),
7367 (*PRUNE:NAME), or (*THEN:NAME) on the matching path is passed back to
7368 the caller as described in the section entitled "Extra data for
7369 pcre_exec()" in the pcreapi documentation. Here is an example of
7370 pcretest output, where the /K modifier requests the retrieval and out-
7371 putting of (*MARK) data:
7373 re> /X(*MARK:A)Y|X(*MARK:B)Z/K
7381 The (*MARK) name is tagged with "MK:" in this output, and in this exam-
7382 ple it indicates which of the two alternatives matched. This is a more
7383 efficient way of obtaining this information than putting each alterna-
7384 tive in its own capturing parentheses.
7386 If a verb with a name is encountered in a positive assertion that is
7387 true, the name is recorded and passed back if it is the last-encoun-
7388 tered. This does not happen for negative assertions or failing positive
7391 After a partial match or a failed match, the last encountered name in
7392 the entire match process is returned. For example:
7394 re> /X(*MARK:A)Y|X(*MARK:B)Z/K
7398 Note that in this unanchored example the mark is retained from the
7399 match attempt that started at the letter "X" in the subject. Subsequent
7400 match attempts starting at "P" and then with an empty string do not get
7401 as far as the (*MARK) item, but nevertheless do not reset it.
7403 If you are interested in (*MARK) values after failed matches, you
7404 should probably set the PCRE_NO_START_OPTIMIZE option (see above) to
7405 ensure that the match is always attempted.
7407 Verbs that act after backtracking
7409 The following verbs do nothing when they are encountered. Matching con-
7410 tinues with what follows, but if there is no subsequent match, causing
7411 a backtrack to the verb, a failure is forced. That is, backtracking
7412 cannot pass to the left of the verb. However, when one of these verbs
7413 appears inside an atomic group or an assertion that is true, its effect
7414 is confined to that group, because once the group has been matched,
7415 there is never any backtracking into it. In this situation, backtrack-
7416 ing can "jump back" to the left of the entire atomic group or asser-
7417 tion. (Remember also, as stated above, that this localization also
7418 applies in subroutine calls.)
7420 These verbs differ in exactly what kind of failure occurs when back-
7421 tracking reaches them. The behaviour described below is what happens
7422 when the verb is not in a subroutine or an assertion. Subsequent sec-
7423 tions cover these special cases.
7427 This verb, which may not be followed by a name, causes the whole match
7428 to fail outright if there is a later matching failure that causes back-
7429 tracking to reach it. Even if the pattern is unanchored, no further
7430 attempts to find a match by advancing the starting point take place. If
7431 (*COMMIT) is the only backtracking verb that is encountered, once it
7432 has been passed pcre_exec() is committed to finding a match at the cur-
7433 rent starting point, or not at all. For example:
7437 This matches "xxaab" but not "aacaab". It can be thought of as a kind
7438 of dynamic anchor, or "I've started, so I must finish." The name of the
7439 most recently passed (*MARK) in the path is passed back when (*COMMIT)
7440 forces a match failure.
7442 If there is more than one backtracking verb in a pattern, a different
7443 one that follows (*COMMIT) may be triggered first, so merely passing
7444 (*COMMIT) during a match does not always guarantee that a match must be
7445 at this starting point.
7447 Note that (*COMMIT) at the start of a pattern is not the same as an
7448 anchor, unless PCRE's start-of-match optimizations are turned off, as
7449 shown in this output from pcretest:
7457 For this pattern, PCRE knows that any match must start with "a", so the
7458 optimization skips along the subject to "a" before applying the pattern
7459 to the first set of data. The match attempt then succeeds. In the sec-
7460 ond set of data, the escape sequence \Y is interpreted by the pcretest
7461 program. It causes the PCRE_NO_START_OPTIMIZE option to be set when
7462 pcre_exec() is called. This disables the optimization that skips along
7463 to the first character. The pattern is now applied starting at "x", and
7464 so the (*COMMIT) causes the match to fail without trying any other
7467 (*PRUNE) or (*PRUNE:NAME)
7469 This verb causes the match to fail at the current starting position in
7470 the subject if there is a later matching failure that causes backtrack-
7471 ing to reach it. If the pattern is unanchored, the normal "bumpalong"
7472 advance to the next starting character then happens. Backtracking can
7473 occur as usual to the left of (*PRUNE), before it is reached, or when
7474 matching to the right of (*PRUNE), but if there is no match to the
7475 right, backtracking cannot cross (*PRUNE). In simple cases, the use of
7476 (*PRUNE) is just an alternative to an atomic group or possessive quan-
7477 tifier, but there are some uses of (*PRUNE) that cannot be expressed in
7478 any other way. In an anchored pattern (*PRUNE) has the same effect as
7481 The behaviour of (*PRUNE:NAME) is the not the same as
7482 (*MARK:NAME)(*PRUNE). It is like (*MARK:NAME) in that the name is
7483 remembered for passing back to the caller. However, (*SKIP:NAME)
7484 searches only for names set with (*MARK).
7488 This verb, when given without a name, is like (*PRUNE), except that if
7489 the pattern is unanchored, the "bumpalong" advance is not to the next
7490 character, but to the position in the subject where (*SKIP) was encoun-
7491 tered. (*SKIP) signifies that whatever text was matched leading up to
7492 it cannot be part of a successful match. Consider:
7496 If the subject is "aaaac...", after the first match attempt fails
7497 (starting at the first character in the string), the starting point
7498 skips on to start the next attempt at "c". Note that a possessive quan-
7499 tifer does not have the same effect as this example; although it would
7500 suppress backtracking during the first match attempt, the second
7501 attempt would start at the second character instead of skipping on to
7506 When (*SKIP) has an associated name, its behaviour is modified. When it
7507 is triggered, the previous path through the pattern is searched for the
7508 most recent (*MARK) that has the same name. If one is found, the
7509 "bumpalong" advance is to the subject position that corresponds to that
7510 (*MARK) instead of to where (*SKIP) was encountered. If no (*MARK) with
7511 a matching name is found, the (*SKIP) is ignored.
7513 Note that (*SKIP:NAME) searches only for names set by (*MARK:NAME). It
7514 ignores names that are set by (*PRUNE:NAME) or (*THEN:NAME).
7516 (*THEN) or (*THEN:NAME)
7518 This verb causes a skip to the next innermost alternative when back-
7519 tracking reaches it. That is, it cancels any further backtracking
7520 within the current alternative. Its name comes from the observation
7521 that it can be used for a pattern-based if-then-else block:
7523 ( COND1 (*THEN) FOO | COND2 (*THEN) BAR | COND3 (*THEN) BAZ ) ...
7525 If the COND1 pattern matches, FOO is tried (and possibly further items
7526 after the end of the group if FOO succeeds); on failure, the matcher
7527 skips to the second alternative and tries COND2, without backtracking
7528 into COND1. If that succeeds and BAR fails, COND3 is tried. If subse-
7529 quently BAZ fails, there are no more alternatives, so there is a back-
7530 track to whatever came before the entire group. If (*THEN) is not
7531 inside an alternation, it acts like (*PRUNE).
7533 The behaviour of (*THEN:NAME) is the not the same as
7534 (*MARK:NAME)(*THEN). It is like (*MARK:NAME) in that the name is
7535 remembered for passing back to the caller. However, (*SKIP:NAME)
7536 searches only for names set with (*MARK).
7538 A subpattern that does not contain a | character is just a part of the
7539 enclosing alternative; it is not a nested alternation with only one
7540 alternative. The effect of (*THEN) extends beyond such a subpattern to
7541 the enclosing alternative. Consider this pattern, where A, B, etc. are
7542 complex pattern fragments that do not contain any | characters at this
7547 If A and B are matched, but there is a failure in C, matching does not
7548 backtrack into A; instead it moves to the next alternative, that is, D.
7549 However, if the subpattern containing (*THEN) is given an alternative,
7550 it behaves differently:
7552 A (B(*THEN)C | (*FAIL)) | D
7554 The effect of (*THEN) is now confined to the inner subpattern. After a
7555 failure in C, matching moves to (*FAIL), which causes the whole subpat-
7556 tern to fail because there are no more alternatives to try. In this
7557 case, matching does now backtrack into A.
7559 Note that a conditional subpattern is not considered as having two
7560 alternatives, because only one is ever used. In other words, the |
7561 character in a conditional subpattern has a different meaning. Ignoring
7562 white space, consider:
7564 ^.*? (?(?=a) a | b(*THEN)c )
7566 If the subject is "ba", this pattern does not match. Because .*? is
7567 ungreedy, it initially matches zero characters. The condition (?=a)
7568 then fails, the character "b" is matched, but "c" is not. At this
7569 point, matching does not backtrack to .*? as might perhaps be expected
7570 from the presence of the | character. The conditional subpattern is
7571 part of the single alternative that comprises the whole pattern, and so
7572 the match fails. (If there was a backtrack into .*?, allowing it to
7573 match "b", the match would succeed.)
7575 The verbs just described provide four different "strengths" of control
7576 when subsequent matching fails. (*THEN) is the weakest, carrying on the
7577 match at the next alternative. (*PRUNE) comes next, failing the match
7578 at the current starting position, but allowing an advance to the next
7579 character (for an unanchored pattern). (*SKIP) is similar, except that
7580 the advance may be more than one character. (*COMMIT) is the strongest,
7581 causing the entire match to fail.
7583 More than one backtracking verb
7585 If more than one backtracking verb is present in a pattern, the one
7586 that is backtracked onto first acts. For example, consider this pat-
7587 tern, where A, B, etc. are complex pattern fragments:
7589 (A(*COMMIT)B(*THEN)C|ABD)
7591 If A matches but B fails, the backtrack to (*COMMIT) causes the entire
7592 match to fail. However, if A and B match, but C fails, the backtrack to
7593 (*THEN) causes the next alternative (ABD) to be tried. This behaviour
7594 is consistent, but is not always the same as Perl's. It means that if
7595 two or more backtracking verbs appear in succession, all the the last
7596 of them has no effect. Consider this example:
7598 ...(*COMMIT)(*PRUNE)...
7600 If there is a matching failure to the right, backtracking onto (*PRUNE)
7601 causes it to be triggered, and its action is taken. There can never be
7602 a backtrack onto (*COMMIT).
7604 Backtracking verbs in repeated groups
7606 PCRE differs from Perl in its handling of backtracking verbs in
7607 repeated groups. For example, consider:
7611 If the subject is "abac", Perl matches, but PCRE fails because the
7612 (*COMMIT) in the second repeat of the group acts.
7614 Backtracking verbs in assertions
7616 (*FAIL) in an assertion has its normal effect: it forces an immediate
7619 (*ACCEPT) in a positive assertion causes the assertion to succeed with-
7620 out any further processing. In a negative assertion, (*ACCEPT) causes
7621 the assertion to fail without any further processing.
7623 The other backtracking verbs are not treated specially if they appear
7624 in a positive assertion. In particular, (*THEN) skips to the next
7625 alternative in the innermost enclosing group that has alternations,
7626 whether or not this is within the assertion.
7628 Negative assertions are, however, different, in order to ensure that
7629 changing a positive assertion into a negative assertion changes its
7630 result. Backtracking into (*COMMIT), (*SKIP), or (*PRUNE) causes a neg-
7631 ative assertion to be true, without considering any further alternative
7632 branches in the assertion. Backtracking into (*THEN) causes it to skip
7633 to the next enclosing alternative within the assertion (the normal be-
7634 haviour), but if the assertion does not have such an alternative,
7635 (*THEN) behaves like (*PRUNE).
7637 Backtracking verbs in subroutines
7639 These behaviours occur whether or not the subpattern is called recur-
7640 sively. Perl's treatment of subroutines is different in some cases.
7642 (*FAIL) in a subpattern called as a subroutine has its normal effect:
7643 it forces an immediate backtrack.
7645 (*ACCEPT) in a subpattern called as a subroutine causes the subroutine
7646 match to succeed without any further processing. Matching then contin-
7647 ues after the subroutine call.
7649 (*COMMIT), (*SKIP), and (*PRUNE) in a subpattern called as a subroutine
7650 cause the subroutine match to fail.
7652 (*THEN) skips to the next alternative in the innermost enclosing group
7653 within the subpattern that has alternatives. If there is no such group
7654 within the subpattern, (*THEN) causes the subroutine match to fail.
7659 pcreapi(3), pcrecallout(3), pcrematching(3), pcresyntax(3), pcre(3),
7660 pcre16(3), pcre32(3).
7666 University Computing Service
7667 Cambridge CB2 3QH, England.
7672 Last updated: 14 June 2015
7673 Copyright (c) 1997-2015 University of Cambridge.
7674 ------------------------------------------------------------------------------
7677 PCRESYNTAX(3) Library Functions Manual PCRESYNTAX(3)
7682 PCRE - Perl-compatible regular expressions
7684 PCRE REGULAR EXPRESSION SYNTAX SUMMARY
7686 The full syntax and semantics of the regular expressions that are sup-
7687 ported by PCRE are described in the pcrepattern documentation. This
7688 document contains a quick-reference summary of the syntax.
7693 \x where x is non-alphanumeric is a literal x
7694 \Q...\E treat enclosed characters as literal
7699 \a alarm, that is, the BEL character (hex 07)
7700 \cx "control-x", where x is any ASCII character
7702 \f form feed (hex 0C)
7704 \r carriage return (hex 0D)
7706 \0dd character with octal code 0dd
7707 \ddd character with octal code ddd, or backreference
7708 \o{ddd..} character with octal code ddd..
7709 \xhh character with hex code hh
7710 \x{hhh..} character with hex code hhh..
7712 Note that \0dd is always an octal code, and that \8 and \9 are the lit-
7713 eral characters "8" and "9".
7718 . any character except newline;
7719 in dotall mode, any character whatsoever
7720 \C one data unit, even in UTF mode (best avoided)
7722 \D a character that is not a decimal digit
7723 \h a horizontal white space character
7724 \H a character that is not a horizontal white space character
7725 \N a character that is not a newline
7726 \p{xx} a character with the xx property
7727 \P{xx} a character without the xx property
7728 \R a newline sequence
7729 \s a white space character
7730 \S a character that is not a white space character
7731 \v a vertical white space character
7732 \V a character that is not a vertical white space character
7733 \w a "word" character
7734 \W a "non-word" character
7735 \X a Unicode extended grapheme cluster
7737 By default, \d, \s, and \w match only ASCII characters, even in UTF-8
7738 mode or in the 16- bit and 32-bit libraries. However, if locale-spe-
7739 cific matching is happening, \s and \w may also match characters with
7740 code points in the range 128-255. If the PCRE_UCP option is set, the
7741 behaviour of these escape sequences is changed to use Unicode proper-
7742 ties and they match many more characters.
7745 GENERAL CATEGORY PROPERTIES FOR \p and \P
7755 Ll Lower case letter
7758 Lt Title case letter
7759 Lu Upper case letter
7773 Pc Connector punctuation
7775 Pe Close punctuation
7776 Pf Final punctuation
7777 Pi Initial punctuation
7778 Po Other punctuation
7784 Sm Mathematical symbol
7789 Zp Paragraph separator
7793 PCRE SPECIAL CATEGORY PROPERTIES FOR \p and \P
7795 Xan Alphanumeric: union of properties L and N
7796 Xps POSIX space: property Z or tab, NL, VT, FF, CR
7797 Xsp Perl space: property Z or tab, NL, VT, FF, CR
7798 Xuc Univerally-named character: one that can be
7799 represented by a Universal Character Name
7800 Xwd Perl word: property Xan or underscore
7802 Perl and POSIX space are now the same. Perl added VT to its space char-
7803 acter set at release 5.18 and PCRE changed at release 8.34.
7806 SCRIPT NAMES FOR \p AND \P
7808 Arabic, Armenian, Avestan, Balinese, Bamum, Bassa_Vah, Batak, Bengali,
7809 Bopomofo, Brahmi, Braille, Buginese, Buhid, Canadian_Aboriginal, Car-
7810 ian, Caucasian_Albanian, Chakma, Cham, Cherokee, Common, Coptic, Cunei-
7811 form, Cypriot, Cyrillic, Deseret, Devanagari, Duployan, Egyptian_Hiero-
7812 glyphs, Elbasan, Ethiopic, Georgian, Glagolitic, Gothic, Grantha,
7813 Greek, Gujarati, Gurmukhi, Han, Hangul, Hanunoo, Hebrew, Hiragana,
7814 Imperial_Aramaic, Inherited, Inscriptional_Pahlavi, Inscrip-
7815 tional_Parthian, Javanese, Kaithi, Kannada, Katakana, Kayah_Li,
7816 Kharoshthi, Khmer, Khojki, Khudawadi, Lao, Latin, Lepcha, Limbu, Lin-
7817 ear_A, Linear_B, Lisu, Lycian, Lydian, Mahajani, Malayalam, Mandaic,
7818 Manichaean, Meetei_Mayek, Mende_Kikakui, Meroitic_Cursive,
7819 Meroitic_Hieroglyphs, Miao, Modi, Mongolian, Mro, Myanmar, Nabataean,
7820 New_Tai_Lue, Nko, Ogham, Ol_Chiki, Old_Italic, Old_North_Arabian,
7821 Old_Permic, Old_Persian, Old_South_Arabian, Old_Turkic, Oriya, Osmanya,
7822 Pahawh_Hmong, Palmyrene, Pau_Cin_Hau, Phags_Pa, Phoenician,
7823 Psalter_Pahlavi, Rejang, Runic, Samaritan, Saurashtra, Sharada, Sha-
7824 vian, Siddham, Sinhala, Sora_Sompeng, Sundanese, Syloti_Nagri, Syriac,
7825 Tagalog, Tagbanwa, Tai_Le, Tai_Tham, Tai_Viet, Takri, Tamil, Telugu,
7826 Thaana, Thai, Tibetan, Tifinagh, Tirhuta, Ugaritic, Vai, Warang_Citi,
7832 [...] positive character class
7833 [^...] negative character class
7834 [x-y] range (can be used for hex characters)
7835 [[:xxx:]] positive POSIX named set
7836 [[:^xxx:]] negative POSIX named set
7842 cntrl control character
7844 graph printing, excluding space
7845 lower lower case letter
7846 print printing, including space
7847 punct printing, excluding alphanumeric
7849 upper upper case letter
7851 xdigit hexadecimal digit
7853 In PCRE, POSIX character set names recognize only ASCII characters by
7854 default, but some of them use Unicode properties if PCRE_UCP is set.
7855 You can use \Q...\E inside a character class.
7861 ?+ 0 or 1, possessive
7864 *+ 0 or more, possessive
7867 ++ 1 or more, possessive
7870 {n,m} at least n, no more than m, greedy
7871 {n,m}+ at least n, no more than m, possessive
7872 {n,m}? at least n, no more than m, lazy
7873 {n,} n or more, greedy
7874 {n,}+ n or more, possessive
7875 {n,}? n or more, lazy
7878 ANCHORS AND SIMPLE ASSERTIONS
7881 \B not a word boundary
7883 also after internal newline in multiline mode
7886 also before newline at end of subject
7887 also before internal newline in multiline mode
7889 also before newline at end of subject
7891 \G first matching position in subject
7896 \K reset start of match
7898 \K is honoured in positive assertions, but ignored in negative ones.
7908 (...) capturing group
7909 (?<name>...) named capturing group (Perl)
7910 (?'name'...) named capturing group (Perl)
7911 (?P<name>...) named capturing group (Python)
7912 (?:...) non-capturing group
7913 (?|...) non-capturing group; reset group numbers for
7914 capturing groups in each alternative
7919 (?>...) atomic, non-capturing group
7924 (?#....) comment (not nestable)
7930 (?J) allow duplicate names
7932 (?s) single line (dotall)
7933 (?U) default ungreedy (lazy)
7934 (?x) extended (ignore white space)
7935 (?-...) unset option(s)
7937 The following are recognized only at the very start of a pattern or
7938 after one of the newline or \R options with similar syntax. More than
7939 one of them may appear.
7941 (*LIMIT_MATCH=d) set the match limit to d (decimal number)
7942 (*LIMIT_RECURSION=d) set the recursion limit to d (decimal number)
7943 (*NO_AUTO_POSSESS) no auto-possessification (PCRE_NO_AUTO_POSSESS)
7944 (*NO_START_OPT) no start-match optimization (PCRE_NO_START_OPTIMIZE)
7945 (*UTF8) set UTF-8 mode: 8-bit library (PCRE_UTF8)
7946 (*UTF16) set UTF-16 mode: 16-bit library (PCRE_UTF16)
7947 (*UTF32) set UTF-32 mode: 32-bit library (PCRE_UTF32)
7948 (*UTF) set appropriate UTF mode for the library in use
7949 (*UCP) set PCRE_UCP (use Unicode properties for \d etc)
7951 Note that LIMIT_MATCH and LIMIT_RECURSION can only reduce the value of
7952 the limits set by the caller of pcre_exec(), not increase them.
7957 These are recognized only at the very start of the pattern or after
7958 option settings with a similar syntax.
7960 (*CR) carriage return only
7962 (*CRLF) carriage return followed by linefeed
7963 (*ANYCRLF) all three of the above
7964 (*ANY) any Unicode newline sequence
7969 These are recognized only at the very start of the pattern or after
7970 option setting with a similar syntax.
7972 (*BSR_ANYCRLF) CR, LF, or CRLF
7973 (*BSR_UNICODE) any Unicode newline sequence
7976 LOOKAHEAD AND LOOKBEHIND ASSERTIONS
7978 (?=...) positive look ahead
7979 (?!...) negative look ahead
7980 (?<=...) positive look behind
7981 (?<!...) negative look behind
7983 Each top-level branch of a look behind must be of a fixed length.
7988 \n reference by number (can be ambiguous)
7989 \gn reference by number
7990 \g{n} reference by number
7991 \g{-n} relative reference by number
7992 \k<name> reference by name (Perl)
7993 \k'name' reference by name (Perl)
7994 \g{name} reference by name (Perl)
7995 \k{name} reference by name (.NET)
7996 (?P=name) reference by name (Python)
7999 SUBROUTINE REFERENCES (POSSIBLY RECURSIVE)
8001 (?R) recurse whole pattern
8002 (?n) call subpattern by absolute number
8003 (?+n) call subpattern by relative number
8004 (?-n) call subpattern by relative number
8005 (?&name) call subpattern by name (Perl)
8006 (?P>name) call subpattern by name (Python)
8007 \g<name> call subpattern by name (Oniguruma)
8008 \g'name' call subpattern by name (Oniguruma)
8009 \g<n> call subpattern by absolute number (Oniguruma)
8010 \g'n' call subpattern by absolute number (Oniguruma)
8011 \g<+n> call subpattern by relative number (PCRE extension)
8012 \g'+n' call subpattern by relative number (PCRE extension)
8013 \g<-n> call subpattern by relative number (PCRE extension)
8014 \g'-n' call subpattern by relative number (PCRE extension)
8017 CONDITIONAL PATTERNS
8019 (?(condition)yes-pattern)
8020 (?(condition)yes-pattern|no-pattern)
8022 (?(n)... absolute reference condition
8023 (?(+n)... relative reference condition
8024 (?(-n)... relative reference condition
8025 (?(<name>)... named reference condition (Perl)
8026 (?('name')... named reference condition (Perl)
8027 (?(name)... named reference condition (PCRE)
8028 (?(R)... overall recursion condition
8029 (?(Rn)... specific group recursion condition
8030 (?(R&name)... specific recursion condition
8031 (?(DEFINE)... define subpattern for reference
8032 (?(assert)... assertion condition
8035 BACKTRACKING CONTROL
8037 The following act immediately they are reached:
8039 (*ACCEPT) force successful match
8040 (*FAIL) force backtrack; synonym (*F)
8041 (*MARK:NAME) set name to be passed back; synonym (*:NAME)
8043 The following act only when a subsequent match failure causes a back-
8044 track to reach them. They all force a match failure, but they differ in
8045 what happens afterwards. Those that advance the start-of-match point do
8046 so only if the pattern is not anchored.
8048 (*COMMIT) overall failure, no advance of starting point
8049 (*PRUNE) advance to next starting character
8050 (*PRUNE:NAME) equivalent to (*MARK:NAME)(*PRUNE)
8051 (*SKIP) advance to current matching position
8052 (*SKIP:NAME) advance to position corresponding to an earlier
8053 (*MARK:NAME); if not found, the (*SKIP) is ignored
8054 (*THEN) local failure, backtrack to next alternation
8055 (*THEN:NAME) equivalent to (*MARK:NAME)(*THEN)
8061 (?Cn) callout with data n
8066 pcrepattern(3), pcreapi(3), pcrecallout(3), pcrematching(3), pcre(3).
8072 University Computing Service
8073 Cambridge CB2 3QH, England.
8078 Last updated: 08 January 2014
8079 Copyright (c) 1997-2014 University of Cambridge.
8080 ------------------------------------------------------------------------------
8083 PCREUNICODE(3) Library Functions Manual PCREUNICODE(3)
8088 PCRE - Perl-compatible regular expressions
8090 UTF-8, UTF-16, UTF-32, AND UNICODE PROPERTY SUPPORT
8092 As well as UTF-8 support, PCRE also supports UTF-16 (from release 8.30)
8093 and UTF-32 (from release 8.32), by means of two additional libraries.
8094 They can be built as well as, or instead of, the 8-bit library.
8099 In order process UTF-8 strings, you must build PCRE's 8-bit library
8100 with UTF support, and, in addition, you must call pcre_compile() with
8101 the PCRE_UTF8 option flag, or the pattern must start with the sequence
8102 (*UTF8) or (*UTF). When either of these is the case, both the pattern
8103 and any subject strings that are matched against it are treated as
8104 UTF-8 strings instead of strings of individual 1-byte characters.
8107 UTF-16 AND UTF-32 SUPPORT
8109 In order process UTF-16 or UTF-32 strings, you must build PCRE's 16-bit
8110 or 32-bit library with UTF support, and, in addition, you must call
8111 pcre16_compile() or pcre32_compile() with the PCRE_UTF16 or PCRE_UTF32
8112 option flag, as appropriate. Alternatively, the pattern must start with
8113 the sequence (*UTF16), (*UTF32), as appropriate, or (*UTF), which can
8114 be used with either library. When UTF mode is set, both the pattern and
8115 any subject strings that are matched against it are treated as UTF-16
8116 or UTF-32 strings instead of strings of individual 16-bit or 32-bit
8120 UTF SUPPORT OVERHEAD
8122 If you compile PCRE with UTF support, but do not use it at run time,
8123 the library will be a bit bigger, but the additional run time overhead
8124 is limited to testing the PCRE_UTF[8|16|32] flag occasionally, so
8125 should not be very big.
8128 UNICODE PROPERTY SUPPORT
8130 If PCRE is built with Unicode character property support (which implies
8131 UTF support), the escape sequences \p{..}, \P{..}, and \X can be used.
8132 The available properties that can be tested are limited to the general
8133 category properties such as Lu for an upper case letter or Nd for a
8134 decimal number, the Unicode script names such as Arabic or Han, and the
8135 derived properties Any and L&. Full lists is given in the pcrepattern
8136 and pcresyntax documentation. Only the short names for properties are
8137 supported. For example, \p{L} matches a letter. Its Perl synonym,
8138 \p{Letter}, is not supported. Furthermore, in Perl, many properties
8139 may optionally be prefixed by "Is", for compatibility with Perl 5.6.
8140 PCRE does not support this.
8142 Validity of UTF-8 strings
8144 When you set the PCRE_UTF8 flag, the byte strings passed as patterns
8145 and subjects are (by default) checked for validity on entry to the rel-
8146 evant functions. The entire string is checked before any other process-
8147 ing takes place. From release 7.3 of PCRE, the check is according the
8148 rules of RFC 3629, which are themselves derived from the Unicode speci-
8149 fication. Earlier releases of PCRE followed the rules of RFC 2279,
8150 which allows the full range of 31-bit values (0 to 0x7FFFFFFF). The
8151 current check allows only values in the range U+0 to U+10FFFF, exclud-
8152 ing the surrogate area. (From release 8.33 the so-called "non-charac-
8153 ter" code points are no longer excluded because Unicode corrigendum #9
8154 makes it clear that they should not be.)
8156 Characters in the "Surrogate Area" of Unicode are reserved for use by
8157 UTF-16, where they are used in pairs to encode codepoints with values
8158 greater than 0xFFFF. The code points that are encoded by UTF-16 pairs
8159 are available independently in the UTF-8 and UTF-32 encodings. (In
8160 other words, the whole surrogate thing is a fudge for UTF-16 which
8161 unfortunately messes up UTF-8 and UTF-32.)
8163 If an invalid UTF-8 string is passed to PCRE, an error return is given.
8164 At compile time, the only additional information is the offset to the
8165 first byte of the failing character. The run-time functions pcre_exec()
8166 and pcre_dfa_exec() also pass back this information, as well as a more
8167 detailed reason code if the caller has provided memory in which to do
8170 In some situations, you may already know that your strings are valid,
8171 and therefore want to skip these checks in order to improve perfor-
8172 mance, for example in the case of a long subject string that is being
8173 scanned repeatedly. If you set the PCRE_NO_UTF8_CHECK flag at compile
8174 time or at run time, PCRE assumes that the pattern or subject it is
8175 given (respectively) contains only valid UTF-8 codes. In this case, it
8176 does not diagnose an invalid UTF-8 string.
8178 Note that passing PCRE_NO_UTF8_CHECK to pcre_compile() just disables
8179 the check for the pattern; it does not also apply to subject strings.
8180 If you want to disable the check for a subject string you must pass
8181 this option to pcre_exec() or pcre_dfa_exec().
8183 If you pass an invalid UTF-8 string when PCRE_NO_UTF8_CHECK is set, the
8184 result is undefined and your program may crash.
8186 Validity of UTF-16 strings
8188 When you set the PCRE_UTF16 flag, the strings of 16-bit data units that
8189 are passed as patterns and subjects are (by default) checked for valid-
8190 ity on entry to the relevant functions. Values other than those in the
8191 surrogate range U+D800 to U+DFFF are independent code points. Values in
8192 the surrogate range must be used in pairs in the correct manner.
8194 If an invalid UTF-16 string is passed to PCRE, an error return is
8195 given. At compile time, the only additional information is the offset
8196 to the first data unit of the failing character. The run-time functions
8197 pcre16_exec() and pcre16_dfa_exec() also pass back this information, as
8198 well as a more detailed reason code if the caller has provided memory
8199 in which to do this.
8201 In some situations, you may already know that your strings are valid,
8202 and therefore want to skip these checks in order to improve perfor-
8203 mance. If you set the PCRE_NO_UTF16_CHECK flag at compile time or at
8204 run time, PCRE assumes that the pattern or subject it is given (respec-
8205 tively) contains only valid UTF-16 sequences. In this case, it does not
8206 diagnose an invalid UTF-16 string. However, if an invalid string is
8207 passed, the result is undefined.
8209 Validity of UTF-32 strings
8211 When you set the PCRE_UTF32 flag, the strings of 32-bit data units that
8212 are passed as patterns and subjects are (by default) checked for valid-
8213 ity on entry to the relevant functions. This check allows only values
8214 in the range U+0 to U+10FFFF, excluding the surrogate area U+D800 to
8217 If an invalid UTF-32 string is passed to PCRE, an error return is
8218 given. At compile time, the only additional information is the offset
8219 to the first data unit of the failing character. The run-time functions
8220 pcre32_exec() and pcre32_dfa_exec() also pass back this information, as
8221 well as a more detailed reason code if the caller has provided memory
8222 in which to do this.
8224 In some situations, you may already know that your strings are valid,
8225 and therefore want to skip these checks in order to improve perfor-
8226 mance. If you set the PCRE_NO_UTF32_CHECK flag at compile time or at
8227 run time, PCRE assumes that the pattern or subject it is given (respec-
8228 tively) contains only valid UTF-32 sequences. In this case, it does not
8229 diagnose an invalid UTF-32 string. However, if an invalid string is
8230 passed, the result is undefined.
8232 General comments about UTF modes
8234 1. Codepoints less than 256 can be specified in patterns by either
8235 braced or unbraced hexadecimal escape sequences (for example, \x{b3} or
8236 \xb3). Larger values have to use braced sequences.
8238 2. Octal numbers up to \777 are recognized, and in UTF-8 mode they
8239 match two-byte characters for values greater than \177.
8241 3. Repeat quantifiers apply to complete UTF characters, not to individ-
8242 ual data units, for example: \x{100}{3}.
8244 4. The dot metacharacter matches one UTF character instead of a single
8247 5. The escape sequence \C can be used to match a single byte in UTF-8
8248 mode, or a single 16-bit data unit in UTF-16 mode, or a single 32-bit
8249 data unit in UTF-32 mode, but its use can lead to some strange effects
8250 because it breaks up multi-unit characters (see the description of \C
8251 in the pcrepattern documentation). The use of \C is not supported in
8252 the alternative matching function pcre[16|32]_dfa_exec(), nor is it
8253 supported in UTF mode by the JIT optimization of pcre[16|32]_exec(). If
8254 JIT optimization is requested for a UTF pattern that contains \C, it
8255 will not succeed, and so the matching will be carried out by the normal
8256 interpretive function.
8258 6. The character escapes \b, \B, \d, \D, \s, \S, \w, and \W correctly
8259 test characters of any code value, but, by default, the characters that
8260 PCRE recognizes as digits, spaces, or word characters remain the same
8261 set as in non-UTF mode, all with values less than 256. This remains
8262 true even when PCRE is built to include Unicode property support,
8263 because to do otherwise would slow down PCRE in many common cases. Note
8264 in particular that this applies to \b and \B, because they are defined
8265 in terms of \w and \W. If you really want to test for a wider sense of,
8266 say, "digit", you can use explicit Unicode property tests such as
8267 \p{Nd}. Alternatively, if you set the PCRE_UCP option, the way that the
8268 character escapes work is changed so that Unicode properties are used
8269 to determine which characters match. There are more details in the sec-
8270 tion on generic character types in the pcrepattern documentation.
8272 7. Similarly, characters that match the POSIX named character classes
8273 are all low-valued characters, unless the PCRE_UCP option is set.
8275 8. However, the horizontal and vertical white space matching escapes
8276 (\h, \H, \v, and \V) do match all the appropriate Unicode characters,
8277 whether or not PCRE_UCP is set.
8279 9. Case-insensitive matching applies only to characters whose values
8280 are less than 128, unless PCRE is built with Unicode property support.
8281 A few Unicode characters such as Greek sigma have more than two code-
8282 points that are case-equivalent. Up to and including PCRE release 8.31,
8283 only one-to-one case mappings were supported, but later releases (with
8284 Unicode property support) do treat as case-equivalent all versions of
8285 characters such as Greek sigma.
8291 University Computing Service
8292 Cambridge CB2 3QH, England.
8297 Last updated: 27 February 2013
8298 Copyright (c) 1997-2013 University of Cambridge.
8299 ------------------------------------------------------------------------------
8302 PCREJIT(3) Library Functions Manual PCREJIT(3)
8307 PCRE - Perl-compatible regular expressions
8309 PCRE JUST-IN-TIME COMPILER SUPPORT
8311 Just-in-time compiling is a heavyweight optimization that can greatly
8312 speed up pattern matching. However, it comes at the cost of extra pro-
8313 cessing before the match is performed. Therefore, it is of most benefit
8314 when the same pattern is going to be matched many times. This does not
8315 necessarily mean many calls of a matching function; if the pattern is
8316 not anchored, matching attempts may take place many times at various
8317 positions in the subject, even for a single call. Therefore, if the
8318 subject string is very long, it may still pay to use JIT for one-off
8321 JIT support applies only to the traditional Perl-compatible matching
8322 function. It does not apply when the DFA matching function is being
8323 used. The code for this support was written by Zoltan Herczeg.
8326 8-BIT, 16-BIT AND 32-BIT SUPPORT
8328 JIT support is available for all of the 8-bit, 16-bit and 32-bit PCRE
8329 libraries. To keep this documentation simple, only the 8-bit interface
8330 is described in what follows. If you are using the 16-bit library, sub-
8331 stitute the 16-bit functions and 16-bit structures (for example,
8332 pcre16_jit_stack instead of pcre_jit_stack). If you are using the
8333 32-bit library, substitute the 32-bit functions and 32-bit structures
8334 (for example, pcre32_jit_stack instead of pcre_jit_stack).
8337 AVAILABILITY OF JIT SUPPORT
8339 JIT support is an optional feature of PCRE. The "configure" option
8340 --enable-jit (or equivalent CMake option) must be set when PCRE is
8341 built if you want to use JIT. The support is limited to the following
8344 ARM v5, v7, and Thumb2
8345 Intel x86 32-bit and 64-bit
8347 Power PC 32-bit and 64-bit
8348 SPARC 32-bit (experimental)
8350 If --enable-jit is set on an unsupported platform, compilation fails.
8352 A program that is linked with PCRE 8.20 or later can tell if JIT sup-
8353 port is available by calling pcre_config() with the PCRE_CONFIG_JIT
8354 option. The result is 1 when JIT is available, and 0 otherwise. How-
8355 ever, a simple program does not need to check this in order to use JIT.
8356 The normal API is implemented in a way that falls back to the interpre-
8357 tive code if JIT is not available. For programs that need the best pos-
8358 sible performance, there is also a "fast path" API that is JIT-spe-
8361 If your program may sometimes be linked with versions of PCRE that are
8362 older than 8.20, but you want to use JIT when it is available, you can
8363 test the values of PCRE_MAJOR and PCRE_MINOR, or the existence of a JIT
8364 macro such as PCRE_CONFIG_JIT, for compile-time control of your code.
8369 You have to do two things to make use of the JIT support in the sim-
8372 (1) Call pcre_study() with the PCRE_STUDY_JIT_COMPILE option for
8373 each compiled pattern, and pass the resulting pcre_extra block to
8376 (2) Use pcre_free_study() to free the pcre_extra block when it is
8377 no longer needed, instead of just freeing it yourself. This
8379 any JIT data is also freed.
8381 For a program that may be linked with pre-8.20 versions of PCRE, you
8384 #ifndef PCRE_STUDY_JIT_COMPILE
8385 #define PCRE_STUDY_JIT_COMPILE 0
8388 so that no option is passed to pcre_study(), and then use something
8389 like this to free the study data:
8391 #ifdef PCRE_CONFIG_JIT
8392 pcre_free_study(study_ptr);
8394 pcre_free(study_ptr);
8397 PCRE_STUDY_JIT_COMPILE requests the JIT compiler to generate code for
8398 complete matches. If you want to run partial matches using the
8399 PCRE_PARTIAL_HARD or PCRE_PARTIAL_SOFT options of pcre_exec(), you
8400 should set one or both of the following options in addition to, or
8401 instead of, PCRE_STUDY_JIT_COMPILE when you call pcre_study():
8403 PCRE_STUDY_JIT_PARTIAL_HARD_COMPILE
8404 PCRE_STUDY_JIT_PARTIAL_SOFT_COMPILE
8406 The JIT compiler generates different optimized code for each of the
8407 three modes (normal, soft partial, hard partial). When pcre_exec() is
8408 called, the appropriate code is run if it is available. Otherwise, the
8409 pattern is matched using interpretive code.
8411 In some circumstances you may need to call additional functions. These
8412 are described in the section entitled "Controlling the JIT stack"
8415 If JIT support is not available, PCRE_STUDY_JIT_COMPILE etc. are
8416 ignored, and no JIT data is created. Otherwise, the compiled pattern is
8417 passed to the JIT compiler, which turns it into machine code that exe-
8418 cutes much faster than the normal interpretive code. When pcre_exec()
8419 is passed a pcre_extra block containing a pointer to JIT code of the
8420 appropriate mode (normal or hard/soft partial), it obeys that code
8421 instead of running the interpreter. The result is identical, but the
8422 compiled JIT code runs much faster.
8424 There are some pcre_exec() options that are not supported for JIT exe-
8425 cution. There are also some pattern items that JIT cannot handle.
8426 Details are given below. In both cases, execution automatically falls
8427 back to the interpretive code. If you want to know whether JIT was
8428 actually used for a particular match, you should arrange for a JIT
8429 callback function to be set up as described in the section entitled
8430 "Controlling the JIT stack" below, even if you do not need to supply a
8431 non-default JIT stack. Such a callback function is called whenever JIT
8432 code is about to be obeyed. If the execution options are not right for
8433 JIT execution, the callback function is not obeyed.
8435 If the JIT compiler finds an unsupported item, no JIT data is gener-
8436 ated. You can find out if JIT execution is available after studying a
8437 pattern by calling pcre_fullinfo() with the PCRE_INFO_JIT option. A
8438 result of 1 means that JIT compilation was successful. A result of 0
8439 means that JIT support is not available, or the pattern was not studied
8440 with PCRE_STUDY_JIT_COMPILE etc., or the JIT compiler was not able to
8443 Once a pattern has been studied, with or without JIT, it can be used as
8444 many times as you like for matching different subject strings.
8447 UNSUPPORTED OPTIONS AND PATTERN ITEMS
8449 The only pcre_exec() options that are supported for JIT execution are
8450 PCRE_NO_UTF8_CHECK, PCRE_NO_UTF16_CHECK, PCRE_NO_UTF32_CHECK, PCRE_NOT-
8451 BOL, PCRE_NOTEOL, PCRE_NOTEMPTY, PCRE_NOTEMPTY_ATSTART, PCRE_PAR-
8452 TIAL_HARD, and PCRE_PARTIAL_SOFT.
8454 The only unsupported pattern items are \C (match a single data unit)
8455 when running in a UTF mode, and a callout immediately before an asser-
8456 tion condition in a conditional group.
8459 RETURN VALUES FROM JIT EXECUTION
8461 When a pattern is matched using JIT execution, the return values are
8462 the same as those given by the interpretive pcre_exec() code, with the
8463 addition of one new error code: PCRE_ERROR_JIT_STACKLIMIT. This means
8464 that the memory used for the JIT stack was insufficient. See "Control-
8465 ling the JIT stack" below for a discussion of JIT stack usage. For com-
8466 patibility with the interpretive pcre_exec() code, no more than two-
8467 thirds of the ovector argument is used for passing back captured sub-
8470 The error code PCRE_ERROR_MATCHLIMIT is returned by the JIT code if
8471 searching a very large pattern tree goes on for too long, as it is in
8472 the same circumstance when JIT is not used, but the details of exactly
8473 what is counted are not the same. The PCRE_ERROR_RECURSIONLIMIT error
8474 code is never returned by JIT execution.
8477 SAVING AND RESTORING COMPILED PATTERNS
8479 The code that is generated by the JIT compiler is architecture-spe-
8480 cific, and is also position dependent. For those reasons it cannot be
8481 saved (in a file or database) and restored later like the bytecode and
8482 other data of a compiled pattern. Saving and restoring compiled pat-
8483 terns is not something many people do. More detail about this facility
8484 is given in the pcreprecompile documentation. It should be possible to
8485 run pcre_study() on a saved and restored pattern, and thereby recreate
8486 the JIT data, but because JIT compilation uses significant resources,
8487 it is probably not worth doing this; you might as well recompile the
8491 CONTROLLING THE JIT STACK
8493 When the compiled JIT code runs, it needs a block of memory to use as a
8494 stack. By default, it uses 32K on the machine stack. However, some
8495 large or complicated patterns need more than this. The error
8496 PCRE_ERROR_JIT_STACKLIMIT is given when there is not enough stack.
8497 Three functions are provided for managing blocks of memory for use as
8498 JIT stacks. There is further discussion about the use of JIT stacks in
8499 the section entitled "JIT stack FAQ" below.
8501 The pcre_jit_stack_alloc() function creates a JIT stack. Its arguments
8502 are a starting size and a maximum size, and it returns a pointer to an
8503 opaque structure of type pcre_jit_stack, or NULL if there is an error.
8504 The pcre_jit_stack_free() function can be used to free a stack that is
8505 no longer needed. (For the technically minded: the address space is
8506 allocated by mmap or VirtualAlloc.)
8508 JIT uses far less memory for recursion than the interpretive code, and
8509 a maximum stack size of 512K to 1M should be more than enough for any
8512 The pcre_assign_jit_stack() function specifies which stack JIT code
8513 should use. Its arguments are as follows:
8516 pcre_jit_callback callback
8519 The extra argument must be the result of studying a pattern with
8520 PCRE_STUDY_JIT_COMPILE etc. There are three cases for the values of the
8523 (1) If callback is NULL and data is NULL, an internal 32K block
8524 on the machine stack is used.
8526 (2) If callback is NULL and data is not NULL, data must be
8527 a valid JIT stack, the result of calling pcre_jit_stack_alloc().
8529 (3) If callback is not NULL, it must point to a function that is
8530 called with data as an argument at the start of matching, in
8531 order to set up a JIT stack. If the return from the callback
8532 function is NULL, the internal 32K stack is used; otherwise the
8533 return value must be a valid JIT stack, the result of calling
8534 pcre_jit_stack_alloc().
8536 A callback function is obeyed whenever JIT code is about to be run; it
8537 is not obeyed when pcre_exec() is called with options that are incom-
8538 patible for JIT execution. A callback function can therefore be used to
8539 determine whether a match operation was executed by JIT or by the
8542 You may safely use the same JIT stack for more than one pattern (either
8543 by assigning directly or by callback), as long as the patterns are all
8544 matched sequentially in the same thread. In a multithread application,
8545 if you do not specify a JIT stack, or if you assign or pass back NULL
8546 from a callback, that is thread-safe, because each thread has its own
8547 machine stack. However, if you assign or pass back a non-NULL JIT
8548 stack, this must be a different stack for each thread so that the
8549 application is thread-safe.
8551 Strictly speaking, even more is allowed. You can assign the same non-
8552 NULL stack to any number of patterns as long as they are not used for
8553 matching by multiple threads at the same time. For example, you can
8554 assign the same stack to all compiled patterns, and use a global mutex
8555 in the callback to wait until the stack is available for use. However,
8556 this is an inefficient solution, and not recommended.
8558 This is a suggestion for how a multithreaded program that needs to set
8559 up non-default JIT stacks might operate:
8561 During thread initalization
8562 thread_local_var = pcre_jit_stack_alloc(...)
8565 pcre_jit_stack_free(thread_local_var)
8567 Use a one-line callback function
8568 return thread_local_var
8570 All the functions described in this section do nothing if JIT is not
8571 available, and pcre_assign_jit_stack() does nothing unless the extra
8572 argument is non-NULL and points to a pcre_extra block that is the
8573 result of a successful study with PCRE_STUDY_JIT_COMPILE etc.
8578 (1) Why do we need JIT stacks?
8580 PCRE (and JIT) is a recursive, depth-first engine, so it needs a stack
8581 where the local data of the current node is pushed before checking its
8582 child nodes. Allocating real machine stack on some platforms is diffi-
8583 cult. For example, the stack chain needs to be updated every time if we
8584 extend the stack on PowerPC. Although it is possible, its updating
8585 time overhead decreases performance. So we do the recursion in memory.
8587 (2) Why don't we simply allocate blocks of memory with malloc()?
8589 Modern operating systems have a nice feature: they can reserve an
8590 address space instead of allocating memory. We can safely allocate mem-
8591 ory pages inside this address space, so the stack could grow without
8592 moving memory data (this is important because of pointers). Thus we can
8593 allocate 1M address space, and use only a single memory page (usually
8594 4K) if that is enough. However, we can still grow up to 1M anytime if
8597 (3) Who "owns" a JIT stack?
8599 The owner of the stack is the user program, not the JIT studied pattern
8600 or anything else. The user program must ensure that if a stack is used
8601 by pcre_exec(), (that is, it is assigned to the pattern currently run-
8602 ning), that stack must not be used by any other threads (to avoid over-
8603 writing the same memory area). The best practice for multithreaded pro-
8604 grams is to allocate a stack for each thread, and return this stack
8605 through the JIT callback function.
8607 (4) When should a JIT stack be freed?
8609 You can free a JIT stack at any time, as long as it will not be used by
8610 pcre_exec() again. When you assign the stack to a pattern, only a
8611 pointer is set. There is no reference counting or any other magic. You
8612 can free the patterns and stacks in any order, anytime. Just do not
8613 call pcre_exec() with a pattern pointing to an already freed stack, as
8614 that will cause SEGFAULT. (Also, do not free a stack currently used by
8615 pcre_exec() in another thread). You can also replace the stack for a
8616 pattern at any time. You can even free the previous stack before
8617 assigning a replacement.
8619 (5) Should I allocate/free a stack every time before/after calling
8622 No, because this is too costly in terms of resources. However, you
8623 could implement some clever idea which release the stack if it is not
8624 used in let's say two minutes. The JIT callback can help to achieve
8625 this without keeping a list of the currently JIT studied patterns.
8627 (6) OK, the stack is for long term memory allocation. But what happens
8628 if a pattern causes stack overflow with a stack of 1M? Is that 1M kept
8629 until the stack is freed?
8631 Especially on embedded sytems, it might be a good idea to release mem-
8632 ory sometimes without freeing the stack. There is no API for this at
8633 the moment. Probably a function call which returns with the currently
8634 allocated memory for any stack and another which allows releasing mem-
8635 ory (shrinking the stack) would be a good idea if someone needs this.
8637 (7) This is too much of a headache. Isn't there any better solution for
8640 No, thanks to Windows. If POSIX threads were used everywhere, we could
8641 throw out this complicated API.
8646 This is a single-threaded example that specifies a JIT stack without
8653 pcre_jit_stack *jit_stack;
8655 re = pcre_compile(pattern, 0, &error, &erroffset, NULL);
8656 /* Check for errors */
8657 extra = pcre_study(re, PCRE_STUDY_JIT_COMPILE, &error);
8658 jit_stack = pcre_jit_stack_alloc(32*1024, 512*1024);
8659 /* Check for error (NULL) */
8660 pcre_assign_jit_stack(extra, NULL, jit_stack);
8661 rc = pcre_exec(re, extra, subject, length, 0, 0, ovector, 30);
8664 pcre_free_study(extra);
8665 pcre_jit_stack_free(jit_stack);
8670 Because the API described above falls back to interpreted execution
8671 when JIT is not available, it is convenient for programs that are writ-
8672 ten for general use in many environments. However, calling JIT via
8673 pcre_exec() does have a performance impact. Programs that are written
8674 for use where JIT is known to be available, and which need the best
8675 possible performance, can instead use a "fast path" API to call JIT
8676 execution directly instead of calling pcre_exec() (obviously only for
8677 patterns that have been successfully studied by JIT).
8679 The fast path function is called pcre_jit_exec(), and it takes exactly
8680 the same arguments as pcre_exec(), plus one additional argument that
8681 must point to a JIT stack. The JIT stack arrangements described above
8682 do not apply. The return values are the same as for pcre_exec().
8684 When you call pcre_exec(), as well as testing for invalid options, a
8685 number of other sanity checks are performed on the arguments. For exam-
8686 ple, if the subject pointer is NULL, or its length is negative, an
8687 immediate error is given. Also, unless PCRE_NO_UTF[8|16|32] is set, a
8688 UTF subject string is tested for validity. In the interests of speed,
8689 these checks do not happen on the JIT fast path, and if invalid data is
8690 passed, the result is undefined.
8692 Bypassing the sanity checks and the pcre_exec() wrapping can give
8693 speedups of more than 10%.
8703 Philip Hazel (FAQ by Zoltan Herczeg)
8704 University Computing Service
8705 Cambridge CB2 3QH, England.
8710 Last updated: 17 March 2013
8711 Copyright (c) 1997-2013 University of Cambridge.
8712 ------------------------------------------------------------------------------
8715 PCREPARTIAL(3) Library Functions Manual PCREPARTIAL(3)
8720 PCRE - Perl-compatible regular expressions
8722 PARTIAL MATCHING IN PCRE
8724 In normal use of PCRE, if the subject string that is passed to a match-
8725 ing function matches as far as it goes, but is too short to match the
8726 entire pattern, PCRE_ERROR_NOMATCH is returned. There are circumstances
8727 where it might be helpful to distinguish this case from other cases in
8728 which there is no match.
8730 Consider, for example, an application where a human is required to type
8731 in data for a field with specific formatting requirements. An example
8732 might be a date in the form ddmmmyy, defined by this pattern:
8734 ^\d?\d(jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)\d\d$
8736 If the application sees the user's keystrokes one by one, and can check
8737 that what has been typed so far is potentially valid, it is able to
8738 raise an error as soon as a mistake is made, by beeping and not
8739 reflecting the character that has been typed, for example. This immedi-
8740 ate feedback is likely to be a better user interface than a check that
8741 is delayed until the entire string has been entered. Partial matching
8742 can also be useful when the subject string is very long and is not all
8745 PCRE supports partial matching by means of the PCRE_PARTIAL_SOFT and
8746 PCRE_PARTIAL_HARD options, which can be set when calling any of the
8747 matching functions. For backwards compatibility, PCRE_PARTIAL is a syn-
8748 onym for PCRE_PARTIAL_SOFT. The essential difference between the two
8749 options is whether or not a partial match is preferred to an alterna-
8750 tive complete match, though the details differ between the two types of
8751 matching function. If both options are set, PCRE_PARTIAL_HARD takes
8754 If you want to use partial matching with just-in-time optimized code,
8755 you must call pcre_study(), pcre16_study() or pcre32_study() with one
8756 or both of these options:
8758 PCRE_STUDY_JIT_PARTIAL_SOFT_COMPILE
8759 PCRE_STUDY_JIT_PARTIAL_HARD_COMPILE
8761 PCRE_STUDY_JIT_COMPILE should also be set if you are going to run non-
8762 partial matches on the same pattern. If the appropriate JIT study mode
8763 has not been set for a match, the interpretive matching code is used.
8765 Setting a partial matching option disables two of PCRE's standard opti-
8766 mizations. PCRE remembers the last literal data unit in a pattern, and
8767 abandons matching immediately if it is not present in the subject
8768 string. This optimization cannot be used for a subject string that
8769 might match only partially. If the pattern was studied, PCRE knows the
8770 minimum length of a matching string, and does not bother to run the
8771 matching function on shorter strings. This optimization is also dis-
8772 abled for partial matching.
8775 PARTIAL MATCHING USING pcre_exec() OR pcre[16|32]_exec()
8777 A partial match occurs during a call to pcre_exec() or
8778 pcre[16|32]_exec() when the end of the subject string is reached suc-
8779 cessfully, but matching cannot continue because more characters are
8780 needed. However, at least one character in the subject must have been
8781 inspected. This character need not form part of the final matched
8782 string; lookbehind assertions and the \K escape sequence provide ways
8783 of inspecting characters before the start of a matched substring. The
8784 requirement for inspecting at least one character exists because an
8785 empty string can always be matched; without such a restriction there
8786 would always be a partial match of an empty string at the end of the
8789 If there are at least two slots in the offsets vector when a partial
8790 match is returned, the first slot is set to the offset of the earliest
8791 character that was inspected. For convenience, the second offset points
8792 to the end of the subject so that a substring can easily be identified.
8793 If there are at least three slots in the offsets vector, the third slot
8794 is set to the offset of the character where matching started.
8796 For the majority of patterns, the contents of the first and third slots
8797 will be the same. However, for patterns that contain lookbehind asser-
8798 tions, or begin with \b or \B, characters before the one where matching
8799 started may have been inspected while carrying out the match. For exam-
8800 ple, consider this pattern:
8804 This pattern matches "123", but only if it is preceded by "abc". If the
8805 subject string is "xyzabc12", the first two offsets after a partial
8806 match are for the substring "abc12", because all these characters were
8807 inspected. However, the third offset is set to 6, because that is the
8808 offset where matching began.
8810 What happens when a partial match is identified depends on which of the
8811 two partial matching options are set.
8813 PCRE_PARTIAL_SOFT WITH pcre_exec() OR pcre[16|32]_exec()
8815 If PCRE_PARTIAL_SOFT is set when pcre_exec() or pcre[16|32]_exec()
8816 identifies a partial match, the partial match is remembered, but match-
8817 ing continues as normal, and other alternatives in the pattern are
8818 tried. If no complete match can be found, PCRE_ERROR_PARTIAL is
8819 returned instead of PCRE_ERROR_NOMATCH.
8821 This option is "soft" because it prefers a complete match over a par-
8822 tial match. All the various matching items in a pattern behave as if
8823 the subject string is potentially complete. For example, \z, \Z, and $
8824 match at the end of the subject, as normal, and for \b and \B the end
8825 of the subject is treated as a non-alphanumeric.
8827 If there is more than one partial match, the first one that was found
8828 provides the data that is returned. Consider this pattern:
8832 If this is matched against the subject string "abc123dog", both alter-
8833 natives fail to match, but the end of the subject is reached during
8834 matching, so PCRE_ERROR_PARTIAL is returned. The offsets are set to 3
8835 and 9, identifying "123dog" as the first partial match that was found.
8836 (In this example, there are two partial matches, because "dog" on its
8837 own partially matches the second alternative.)
8839 PCRE_PARTIAL_HARD WITH pcre_exec() OR pcre[16|32]_exec()
8841 If PCRE_PARTIAL_HARD is set for pcre_exec() or pcre[16|32]_exec(),
8842 PCRE_ERROR_PARTIAL is returned as soon as a partial match is found,
8843 without continuing to search for possible complete matches. This option
8844 is "hard" because it prefers an earlier partial match over a later com-
8845 plete match. For this reason, the assumption is made that the end of
8846 the supplied subject string may not be the true end of the available
8847 data, and so, if \z, \Z, \b, \B, or $ are encountered at the end of the
8848 subject, the result is PCRE_ERROR_PARTIAL, provided that at least one
8849 character in the subject has been inspected.
8851 Setting PCRE_PARTIAL_HARD also affects the way UTF-8 and UTF-16 subject
8852 strings are checked for validity. Normally, an invalid sequence causes
8853 the error PCRE_ERROR_BADUTF8 or PCRE_ERROR_BADUTF16. However, in the
8854 special case of a truncated character at the end of the subject,
8855 PCRE_ERROR_SHORTUTF8 or PCRE_ERROR_SHORTUTF16 is returned when
8856 PCRE_PARTIAL_HARD is set.
8858 Comparing hard and soft partial matching
8860 The difference between the two partial matching options can be illus-
8861 trated by a pattern such as:
8865 This matches either "dog" or "dogsbody", greedily (that is, it prefers
8866 the longer string if possible). If it is matched against the string
8867 "dog" with PCRE_PARTIAL_SOFT, it yields a complete match for "dog".
8868 However, if PCRE_PARTIAL_HARD is set, the result is PCRE_ERROR_PARTIAL.
8869 On the other hand, if the pattern is made ungreedy the result is dif-
8874 In this case the result is always a complete match because that is
8875 found first, and matching never continues after finding a complete
8876 match. It might be easier to follow this explanation by thinking of the
8877 two patterns like this:
8879 /dog(sbody)?/ is the same as /dogsbody|dog/
8880 /dog(sbody)??/ is the same as /dog|dogsbody/
8882 The second pattern will never match "dogsbody", because it will always
8883 find the shorter match first.
8886 PARTIAL MATCHING USING pcre_dfa_exec() OR pcre[16|32]_dfa_exec()
8888 The DFA functions move along the subject string character by character,
8889 without backtracking, searching for all possible matches simultane-
8890 ously. If the end of the subject is reached before the end of the pat-
8891 tern, there is the possibility of a partial match, again provided that
8892 at least one character has been inspected.
8894 When PCRE_PARTIAL_SOFT is set, PCRE_ERROR_PARTIAL is returned only if
8895 there have been no complete matches. Otherwise, the complete matches
8896 are returned. However, if PCRE_PARTIAL_HARD is set, a partial match
8897 takes precedence over any complete matches. The portion of the string
8898 that was inspected when the longest partial match was found is set as
8899 the first matching string, provided there are at least two slots in the
8902 Because the DFA functions always search for all possible matches, and
8903 there is no difference between greedy and ungreedy repetition, their
8904 behaviour is different from the standard functions when PCRE_PAR-
8905 TIAL_HARD is set. Consider the string "dog" matched against the
8906 ungreedy pattern shown above:
8910 Whereas the standard functions stop as soon as they find the complete
8911 match for "dog", the DFA functions also find the partial match for
8912 "dogsbody", and so return that when PCRE_PARTIAL_HARD is set.
8915 PARTIAL MATCHING AND WORD BOUNDARIES
8917 If a pattern ends with one of sequences \b or \B, which test for word
8918 boundaries, partial matching with PCRE_PARTIAL_SOFT can give counter-
8919 intuitive results. Consider this pattern:
8923 This matches "cat", provided there is a word boundary at either end. If
8924 the subject string is "the cat", the comparison of the final "t" with a
8925 following character cannot take place, so a partial match is found.
8926 However, normal matching carries on, and \b matches at the end of the
8927 subject when the last character is a letter, so a complete match is
8928 found. The result, therefore, is not PCRE_ERROR_PARTIAL. Using
8929 PCRE_PARTIAL_HARD in this case does yield PCRE_ERROR_PARTIAL, because
8930 then the partial match takes precedence.
8933 FORMERLY RESTRICTED PATTERNS
8935 For releases of PCRE prior to 8.00, because of the way certain internal
8936 optimizations were implemented in the pcre_exec() function, the
8937 PCRE_PARTIAL option (predecessor of PCRE_PARTIAL_SOFT) could not be
8938 used with all patterns. From release 8.00 onwards, the restrictions no
8939 longer apply, and partial matching with can be requested for any pat-
8942 Items that were formerly restricted were repeated single characters and
8943 repeated metasequences. If PCRE_PARTIAL was set for a pattern that did
8944 not conform to the restrictions, pcre_exec() returned the error code
8945 PCRE_ERROR_BADPARTIAL (-13). This error code is no longer in use. The
8946 PCRE_INFO_OKPARTIAL call to pcre_fullinfo() to find out if a compiled
8947 pattern can be used for partial matching now always returns 1.
8950 EXAMPLE OF PARTIAL MATCHING USING PCRETEST
8952 If the escape sequence \P is present in a pcretest data line, the
8953 PCRE_PARTIAL_SOFT option is used for the match. Here is a run of
8954 pcretest that uses the date example quoted above:
8956 re> /^\d?\d(jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)\d\d$/
8961 Partial match: 23dec3
8969 The first data string is matched completely, so pcretest shows the
8970 matched substrings. The remaining four strings do not match the com-
8971 plete pattern, but the first two are partial matches. Similar output is
8972 obtained if DFA matching is used.
8974 If the escape sequence \P is present more than once in a pcretest data
8975 line, the PCRE_PARTIAL_HARD option is set for the match.
8978 MULTI-SEGMENT MATCHING WITH pcre_dfa_exec() OR pcre[16|32]_dfa_exec()
8980 When a partial match has been found using a DFA matching function, it
8981 is possible to continue the match by providing additional subject data
8982 and calling the function again with the same compiled regular expres-
8983 sion, this time setting the PCRE_DFA_RESTART option. You must pass the
8984 same working space as before, because this is where details of the pre-
8985 vious partial match are stored. Here is an example using pcretest,
8986 using the \R escape sequence to set the PCRE_DFA_RESTART option (\D
8987 specifies the use of the DFA matching function):
8989 re> /^\d?\d(jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)\d\d$/
8995 The first call has "23ja" as the subject, and requests partial match-
8996 ing; the second call has "n05" as the subject for the continued
8997 (restarted) match. Notice that when the match is complete, only the
8998 last part is shown; PCRE does not retain the previously partially-
8999 matched string. It is up to the calling program to do that if it needs
9002 That means that, for an unanchored pattern, if a continued match fails,
9003 it is not possible to try again at a new starting point. All this
9004 facility is capable of doing is continuing with the previous match
9005 attempt. In the previous example, if the second set of data is "ug23"
9006 the result is no match, even though there would be a match for "aug23"
9007 if the entire string were given at once. Depending on the application,
9008 this may or may not be what you want. The only way to allow for start-
9009 ing again at the next character is to retain the matched part of the
9010 subject and try a new complete match.
9012 You can set the PCRE_PARTIAL_SOFT or PCRE_PARTIAL_HARD options with
9013 PCRE_DFA_RESTART to continue partial matching over multiple segments.
9014 This facility can be used to pass very long subject strings to the DFA
9018 MULTI-SEGMENT MATCHING WITH pcre_exec() OR pcre[16|32]_exec()
9020 From release 8.00, the standard matching functions can also be used to
9021 do multi-segment matching. Unlike the DFA functions, it is not possible
9022 to restart the previous match with a new segment of data. Instead, new
9023 data must be added to the previous subject string, and the entire match
9024 re-run, starting from the point where the partial match occurred. Ear-
9025 lier data can be discarded.
9027 It is best to use PCRE_PARTIAL_HARD in this situation, because it does
9028 not treat the end of a segment as the end of the subject when matching
9029 \z, \Z, \b, \B, and $. Consider an unanchored pattern that matches
9032 re> /\d?\d(jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)\d\d/
9033 data> The date is 23ja\P\P
9036 At this stage, an application could discard the text preceding "23ja",
9037 add on text from the next segment, and call the matching function
9038 again. Unlike the DFA matching functions, the entire matching string
9039 must always be available, and the complete matching process occurs for
9040 each call, so more memory and more processing time is needed.
9042 Note: If the pattern contains lookbehind assertions, or \K, or starts
9043 with \b or \B, the string that is returned for a partial match includes
9044 characters that precede the start of what would be returned for a com-
9045 plete match, because it contains all the characters that were inspected
9046 during the partial match.
9049 ISSUES WITH MULTI-SEGMENT MATCHING
9051 Certain types of pattern may give problems with multi-segment matching,
9052 whichever matching function is used.
9054 1. If the pattern contains a test for the beginning of a line, you need
9055 to pass the PCRE_NOTBOL option when the subject string for any call
9056 does start at the beginning of a line. There is also a PCRE_NOTEOL
9057 option, but in practice when doing multi-segment matching you should be
9058 using PCRE_PARTIAL_HARD, which includes the effect of PCRE_NOTEOL.
9060 2. Lookbehind assertions that have already been obeyed are catered for
9061 in the offsets that are returned for a partial match. However a lookbe-
9062 hind assertion later in the pattern could require even earlier charac-
9063 ters to be inspected. You can handle this case by using the
9064 PCRE_INFO_MAXLOOKBEHIND option of the pcre_fullinfo() or
9065 pcre[16|32]_fullinfo() functions to obtain the length of the longest
9066 lookbehind in the pattern. This length is given in characters, not
9067 bytes. If you always retain at least that many characters before the
9068 partially matched string, all should be well. (Of course, near the
9069 start of the subject, fewer characters may be present; in that case all
9070 characters should be retained.)
9072 From release 8.33, there is a more accurate way of deciding which char-
9073 acters to retain. Instead of subtracting the length of the longest
9074 lookbehind from the earliest inspected character (offsets[0]), the
9075 match start position (offsets[2]) should be used, and the next match
9076 attempt started at the offsets[2] character by setting the startoffset
9077 argument of pcre_exec() or pcre_dfa_exec().
9079 For example, if the pattern "(?<=123)abc" is partially matched against
9080 the string "xx123a", the three offset values returned are 2, 6, and 5.
9081 This indicates that the matching process that gave a partial match
9082 started at offset 5, but the characters "123a" were all inspected. The
9083 maximum lookbehind for that pattern is 3, so taking that away from 5
9084 shows that we need only keep "123a", and the next match attempt can be
9085 started at offset 3 (that is, at "a") when further characters have been
9086 added. When the match start is not the earliest inspected character,
9087 pcretest shows it explicitly:
9091 Partial match at offset 5: 123a
9093 3. Because a partial match must always contain at least one character,
9094 what might be considered a partial match of an empty string actually
9095 gives a "no match" result. For example:
9101 If the next segment begins "cx", a match should be found, but this will
9102 only happen if characters from the previous segment are retained. For
9103 this reason, a "no match" result should be interpreted as "partial
9104 match of an empty string" when the pattern contains lookbehinds.
9106 4. Matching a subject string that is split into multiple segments may
9107 not always produce exactly the same result as matching over one single
9108 long string, especially when PCRE_PARTIAL_SOFT is used. The section
9109 "Partial Matching and Word Boundaries" above describes an issue that
9110 arises if the pattern ends with \b or \B. Another kind of difference
9111 may occur when there are multiple matching possibilities, because (for
9112 PCRE_PARTIAL_SOFT) a partial match result is given only when there are
9113 no completed matches. This means that as soon as the shortest match has
9114 been found, continuation to a new subject segment is no longer possi-
9115 ble. Consider again this pcretest example:
9128 The first data line passes the string "dogsb" to a standard matching
9129 function, setting the PCRE_PARTIAL_SOFT option. Although the string is
9130 a partial match for "dogsbody", the result is not PCRE_ERROR_PARTIAL,
9131 because the shorter string "dog" is a complete match. Similarly, when
9132 the subject is presented to a DFA matching function in several parts
9133 ("do" and "gsb" being the first two) the match stops when "dog" has
9134 been found, and it is not possible to continue. On the other hand, if
9135 "dogsbody" is presented as a single string, a DFA matching function
9138 Because of these problems, it is best to use PCRE_PARTIAL_HARD when
9139 matching multi-segment data. The example above then behaves differ-
9144 Partial match: dogsb
9150 5. Patterns that contain alternatives at the top level which do not all
9151 start with the same pattern item may not work as expected when
9152 PCRE_DFA_RESTART is used. For example, consider this pattern:
9156 If the first part of the subject is "ABC123", a partial match of the
9157 first alternative is found at offset 3. There is no partial match for
9158 the second alternative, because such a match does not start at the same
9159 point in the subject string. Attempting to continue with the string
9160 "7890" does not yield a match because only those alternatives that
9161 match at one point in the subject are remembered. The problem arises
9162 because the start of the second alternative matches within the first
9163 alternative. There is no problem with anchored patterns or patterns
9168 where no string can be a partial match for both alternatives. This is
9169 not a problem if a standard matching function is used, because the
9170 entire match has to be rerun each time:
9178 Of course, instead of using PCRE_DFA_RESTART, the same technique of re-
9179 running the entire match can also be used with the DFA matching func-
9180 tions. Another possibility is to work with two buffers. If a partial
9181 match at offset n in the first buffer is followed by "no match" when
9182 PCRE_DFA_RESTART is used on the second buffer, you can then try a new
9183 match starting at offset n+1 in the first buffer.
9189 University Computing Service
9190 Cambridge CB2 3QH, England.
9195 Last updated: 02 July 2013
9196 Copyright (c) 1997-2013 University of Cambridge.
9197 ------------------------------------------------------------------------------
9200 PCREPRECOMPILE(3) Library Functions Manual PCREPRECOMPILE(3)
9205 PCRE - Perl-compatible regular expressions
9207 SAVING AND RE-USING PRECOMPILED PCRE PATTERNS
9209 If you are running an application that uses a large number of regular
9210 expression patterns, it may be useful to store them in a precompiled
9211 form instead of having to compile them every time the application is
9212 run. If you are not using any private character tables (see the
9213 pcre_maketables() documentation), this is relatively straightforward.
9214 If you are using private tables, it is a little bit more complicated.
9215 However, if you are using the just-in-time optimization feature, it is
9216 not possible to save and reload the JIT data.
9218 If you save compiled patterns to a file, you can copy them to a differ-
9219 ent host and run them there. If the two hosts have different endianness
9220 (byte order), you should run the pcre[16|32]_pat-
9221 tern_to_host_byte_order() function on the new host before trying to
9222 match the pattern. The matching functions return PCRE_ERROR_BADENDIAN-
9223 NESS if they detect a pattern with the wrong endianness.
9225 Compiling regular expressions with one version of PCRE for use with a
9226 different version is not guaranteed to work and may cause crashes, and
9227 saving and restoring a compiled pattern loses any JIT optimization
9231 SAVING A COMPILED PATTERN
9233 The value returned by pcre[16|32]_compile() points to a single block of
9234 memory that holds the compiled pattern and associated data. You can
9235 find the length of this block in bytes by calling
9236 pcre[16|32]_fullinfo() with an argument of PCRE_INFO_SIZE. You can then
9237 save the data in any appropriate manner. Here is sample code for the
9238 8-bit library that compiles a pattern and writes it to a file. It
9239 assumes that the variable fd refers to a file that is open for output:
9241 int erroroffset, rc, size;
9245 re = pcre_compile("my pattern", 0, &error, &erroroffset, NULL);
9246 if (re == NULL) { ... handle errors ... }
9247 rc = pcre_fullinfo(re, NULL, PCRE_INFO_SIZE, &size);
9248 if (rc < 0) { ... handle errors ... }
9249 rc = fwrite(re, 1, size, fd);
9250 if (rc != size) { ... handle errors ... }
9252 In this example, the bytes that comprise the compiled pattern are
9253 copied exactly. Note that this is binary data that may contain any of
9254 the 256 possible byte values. On systems that make a distinction
9255 between binary and non-binary data, be sure that the file is opened for
9258 If you want to write more than one pattern to a file, you will have to
9259 devise a way of separating them. For binary data, preceding each pat-
9260 tern with its length is probably the most straightforward approach.
9261 Another possibility is to write out the data in hexadecimal instead of
9262 binary, one pattern to a line.
9264 Saving compiled patterns in a file is only one possible way of storing
9265 them for later use. They could equally well be saved in a database, or
9266 in the memory of some daemon process that passes them via sockets to
9267 the processes that want them.
9269 If the pattern has been studied, it is also possible to save the normal
9270 study data in a similar way to the compiled pattern itself. However, if
9271 the PCRE_STUDY_JIT_COMPILE was used, the just-in-time data that is cre-
9272 ated cannot be saved because it is too dependent on the current envi-
9273 ronment. When studying generates additional information,
9274 pcre[16|32]_study() returns a pointer to a pcre[16|32]_extra data
9275 block. Its format is defined in the section on matching a pattern in
9276 the pcreapi documentation. The study_data field points to the binary
9277 study data, and this is what you must save (not the pcre[16|32]_extra
9278 block itself). The length of the study data can be obtained by calling
9279 pcre[16|32]_fullinfo() with an argument of PCRE_INFO_STUDYSIZE. Remem-
9280 ber to check that pcre[16|32]_study() did return a non-NULL value
9281 before trying to save the study data.
9284 RE-USING A PRECOMPILED PATTERN
9286 Re-using a precompiled pattern is straightforward. Having reloaded it
9287 into main memory, called pcre[16|32]_pattern_to_host_byte_order() if
9288 necessary, you pass its pointer to pcre[16|32]_exec() or
9289 pcre[16|32]_dfa_exec() in the usual way.
9291 However, if you passed a pointer to custom character tables when the
9292 pattern was compiled (the tableptr argument of pcre[16|32]_compile()),
9293 you must now pass a similar pointer to pcre[16|32]_exec() or
9294 pcre[16|32]_dfa_exec(), because the value saved with the compiled pat-
9295 tern will obviously be nonsense. A field in a pcre[16|32]_extra() block
9296 is used to pass this data, as described in the section on matching a
9297 pattern in the pcreapi documentation.
9299 Warning: The tables that pcre_exec() and pcre_dfa_exec() use must be
9300 the same as those that were used when the pattern was compiled. If this
9301 is not the case, the behaviour is undefined.
9303 If you did not provide custom character tables when the pattern was
9304 compiled, the pointer in the compiled pattern is NULL, which causes the
9305 matching functions to use PCRE's internal tables. Thus, you do not need
9306 to take any special action at run time in this case.
9308 If you saved study data with the compiled pattern, you need to create
9309 your own pcre[16|32]_extra data block and set the study_data field to
9310 point to the reloaded study data. You must also set the
9311 PCRE_EXTRA_STUDY_DATA bit in the flags field to indicate that study
9312 data is present. Then pass the pcre[16|32]_extra block to the matching
9313 function in the usual way. If the pattern was studied for just-in-time
9314 optimization, that data cannot be saved, and so is lost by a
9318 COMPATIBILITY WITH DIFFERENT PCRE RELEASES
9320 In general, it is safest to recompile all saved patterns when you
9321 update to a new PCRE release, though not all updates actually require
9328 University Computing Service
9329 Cambridge CB2 3QH, England.
9334 Last updated: 12 November 2013
9335 Copyright (c) 1997-2013 University of Cambridge.
9336 ------------------------------------------------------------------------------
9339 PCREPERFORM(3) Library Functions Manual PCREPERFORM(3)
9344 PCRE - Perl-compatible regular expressions
9348 Two aspects of performance are discussed below: memory usage and pro-
9349 cessing time. The way you express your pattern as a regular expression
9350 can affect both of them.
9353 COMPILED PATTERN MEMORY USAGE
9355 Patterns are compiled by PCRE into a reasonably efficient interpretive
9356 code, so that most simple patterns do not use much memory. However,
9357 there is one case where the memory usage of a compiled pattern can be
9358 unexpectedly large. If a parenthesized subpattern has a quantifier with
9359 a minimum greater than 1 and/or a limited maximum, the whole subpattern
9360 is repeated in the compiled code. For example, the pattern
9364 is compiled as if it were
9366 (abc|def)(abc|def)((abc|def)(abc|def)?)?
9368 (Technical aside: It is done this way so that backtrack points within
9369 each of the repetitions can be independently maintained.)
9371 For regular expressions whose quantifiers use only small numbers, this
9372 is not usually a problem. However, if the numbers are large, and par-
9373 ticularly if such repetitions are nested, the memory usage can become
9374 an embarrassment. For example, the very simple pattern
9376 ((ab){1,1000}c){1,3}
9378 uses 51K bytes when compiled using the 8-bit library. When PCRE is com-
9379 piled with its default internal pointer size of two bytes, the size
9380 limit on a compiled pattern is 64K data units, and this is reached with
9381 the above pattern if the outer repetition is increased from 3 to 4.
9382 PCRE can be compiled to use larger internal pointers and thus handle
9383 larger compiled patterns, but it is better to try to rewrite your pat-
9384 tern to use less memory if you can.
9386 One way of reducing the memory usage for such patterns is to make use
9387 of PCRE's "subroutine" facility. Re-writing the above pattern as
9389 ((ab)(?2){0,999}c)(?1){0,2}
9391 reduces the memory requirements to 18K, and indeed it remains under 20K
9392 even with the outer repetition increased to 100. However, this pattern
9393 is not exactly equivalent, because the "subroutine" calls are treated
9394 as atomic groups into which there can be no backtracking if there is a
9395 subsequent matching failure. Therefore, PCRE cannot do this kind of
9396 rewriting automatically. Furthermore, there is a noticeable loss of
9397 speed when executing the modified pattern. Nevertheless, if the atomic
9398 grouping is not a problem and the loss of speed is acceptable, this
9399 kind of rewriting will allow you to process patterns that PCRE cannot
9403 STACK USAGE AT RUN TIME
9405 When pcre_exec() or pcre[16|32]_exec() is used for matching, certain
9406 kinds of pattern can cause it to use large amounts of the process
9407 stack. In some environments the default process stack is quite small,
9408 and if it runs out the result is often SIGSEGV. This issue is probably
9409 the most frequently raised problem with PCRE. Rewriting your pattern
9410 can often help. The pcrestack documentation discusses this issue in
9416 Certain items in regular expression patterns are processed more effi-
9417 ciently than others. It is more efficient to use a character class like
9418 [aeiou] than a set of single-character alternatives such as
9419 (a|e|i|o|u). In general, the simplest construction that provides the
9420 required behaviour is usually the most efficient. Jeffrey Friedl's book
9421 contains a lot of useful general discussion about optimizing regular
9422 expressions for efficient performance. This document contains a few
9423 observations about PCRE.
9425 Using Unicode character properties (the \p, \P, and \X escapes) is
9426 slow, because PCRE has to use a multi-stage table lookup whenever it
9427 needs a character's property. If you can find an alternative pattern
9428 that does not use character properties, it will probably be faster.
9430 By default, the escape sequences \b, \d, \s, and \w, and the POSIX
9431 character classes such as [:alpha:] do not use Unicode properties,
9432 partly for backwards compatibility, and partly for performance reasons.
9433 However, you can set PCRE_UCP if you want Unicode character properties
9434 to be used. This can double the matching time for items such as \d,
9435 when matched with a traditional matching function; the performance loss
9436 is less with a DFA matching function, and in both cases there is not
9437 much difference for \b.
9439 When a pattern begins with .* not in parentheses, or in parentheses
9440 that are not the subject of a backreference, and the PCRE_DOTALL option
9441 is set, the pattern is implicitly anchored by PCRE, since it can match
9442 only at the start of a subject string. However, if PCRE_DOTALL is not
9443 set, PCRE cannot make this optimization, because the . metacharacter
9444 does not then match a newline, and if the subject string contains new-
9445 lines, the pattern may match from the character immediately following
9446 one of them instead of from the very start. For example, the pattern
9450 matches the subject "first\nand second" (where \n stands for a newline
9451 character), with the match starting at the seventh character. In order
9452 to do this, PCRE has to retry the match starting after every newline in
9455 If you are using such a pattern with subject strings that do not con-
9456 tain newlines, the best performance is obtained by setting PCRE_DOTALL,
9457 or starting the pattern with ^.* or ^.*? to indicate explicit anchor-
9458 ing. That saves PCRE from having to scan along the subject looking for
9459 a newline to restart at.
9461 Beware of patterns that contain nested indefinite repeats. These can
9462 take a long time to run when applied to a string that does not match.
9463 Consider the pattern fragment
9467 This can match "aaaa" in 16 different ways, and this number increases
9468 very rapidly as the string gets longer. (The * repeat can match 0, 1,
9469 2, 3, or 4 times, and for each of those cases other than 0 or 4, the +
9470 repeats can match different numbers of times.) When the remainder of
9471 the pattern is such that the entire match is going to fail, PCRE has in
9472 principle to try every possible variation, and this can take an
9473 extremely long time, even for relatively short strings.
9475 An optimization catches some of the more simple cases such as
9479 where a literal character follows. Before embarking on the standard
9480 matching procedure, PCRE checks that there is a "b" later in the sub-
9481 ject string, and if there is not, it fails the match immediately. How-
9482 ever, when there is no following literal this optimization cannot be
9483 used. You can see the difference by comparing the behaviour of
9487 with the pattern above. The former gives a failure almost instantly
9488 when applied to a whole line of "a" characters, whereas the latter
9489 takes an appreciable time with strings longer than about 20 characters.
9491 In many cases, the solution to this kind of performance issue is to use
9492 an atomic group or a possessive quantifier.
9498 University Computing Service
9499 Cambridge CB2 3QH, England.
9504 Last updated: 25 August 2012
9505 Copyright (c) 1997-2012 University of Cambridge.
9506 ------------------------------------------------------------------------------
9509 PCREPOSIX(3) Library Functions Manual PCREPOSIX(3)
9514 PCRE - Perl-compatible regular expressions.
9518 #include <pcreposix.h>
9520 int regcomp(regex_t *preg, const char *pattern,
9523 int regexec(regex_t *preg, const char *string,
9524 size_t nmatch, regmatch_t pmatch[], int eflags);
9525 size_t regerror(int errcode, const regex_t *preg,
9526 char *errbuf, size_t errbuf_size);
9528 void regfree(regex_t *preg);
9533 This set of functions provides a POSIX-style API for the PCRE regular
9534 expression 8-bit library. See the pcreapi documentation for a descrip-
9535 tion of PCRE's native API, which contains much additional functional-
9536 ity. There is no POSIX-style wrapper for PCRE's 16-bit and 32-bit
9539 The functions described here are just wrapper functions that ultimately
9540 call the PCRE native API. Their prototypes are defined in the
9541 pcreposix.h header file, and on Unix systems the library itself is
9542 called pcreposix.a, so can be accessed by adding -lpcreposix to the
9543 command for linking an application that uses them. Because the POSIX
9544 functions call the native ones, it is also necessary to add -lpcre.
9546 I have implemented only those POSIX option bits that can be reasonably
9547 mapped to PCRE native options. In addition, the option REG_EXTENDED is
9548 defined with the value zero. This has no effect, but since programs
9549 that are written to the POSIX interface often use it, this makes it
9550 easier to slot in PCRE as a replacement library. Other POSIX options
9551 are not even defined.
9553 There are also some other options that are not defined by POSIX. These
9554 have been added at the request of users who want to make use of certain
9555 PCRE-specific features via the POSIX calling interface.
9557 When PCRE is called via these functions, it is only the API that is
9558 POSIX-like in style. The syntax and semantics of the regular expres-
9559 sions themselves are still those of Perl, subject to the setting of
9560 various PCRE options, as described below. "POSIX-like in style" means
9561 that the API approximates to the POSIX definition; it is not fully
9562 POSIX-compatible, and in multi-byte encoding domains it is probably
9563 even less compatible.
9565 The header for these functions is supplied as pcreposix.h to avoid any
9566 potential clash with other POSIX libraries. It can, of course, be
9567 renamed or aliased as regex.h, which is the "correct" name. It provides
9568 two structure types, regex_t for compiled internal forms, and reg-
9569 match_t for returning captured substrings. It also defines some con-
9570 stants whose names start with "REG_"; these are used for setting
9571 options and identifying error codes.
9576 The function regcomp() is called to compile a pattern into an internal
9577 form. The pattern is a C string terminated by a binary zero, and is
9578 passed in the argument pattern. The preg argument is a pointer to a
9579 regex_t structure that is used as a base for storing information about
9580 the compiled regular expression.
9582 The argument cflags is either zero, or contains one or more of the bits
9583 defined by the following macros:
9587 The PCRE_DOTALL option is set when the regular expression is passed for
9588 compilation to the native function. Note that REG_DOTALL is not part of
9593 The PCRE_CASELESS option is set when the regular expression is passed
9594 for compilation to the native function.
9598 The PCRE_MULTILINE option is set when the regular expression is passed
9599 for compilation to the native function. Note that this does not mimic
9600 the defined POSIX behaviour for REG_NEWLINE (see the following sec-
9605 The PCRE_NO_AUTO_CAPTURE option is set when the regular expression is
9606 passed for compilation to the native function. In addition, when a pat-
9607 tern that is compiled with this flag is passed to regexec() for match-
9608 ing, the nmatch and pmatch arguments are ignored, and no captured
9609 strings are returned.
9613 The PCRE_UCP option is set when the regular expression is passed for
9614 compilation to the native function. This causes PCRE to use Unicode
9615 properties when matchine \d, \w, etc., instead of just recognizing
9616 ASCII values. Note that REG_UTF8 is not part of the POSIX standard.
9620 The PCRE_UNGREEDY option is set when the regular expression is passed
9621 for compilation to the native function. Note that REG_UNGREEDY is not
9622 part of the POSIX standard.
9626 The PCRE_UTF8 option is set when the regular expression is passed for
9627 compilation to the native function. This causes the pattern itself and
9628 all data strings used for matching it to be treated as UTF-8 strings.
9629 Note that REG_UTF8 is not part of the POSIX standard.
9631 In the absence of these flags, no options are passed to the native
9632 function. This means the the regex is compiled with PCRE default
9633 semantics. In particular, the way it handles newline characters in the
9634 subject string is the Perl way, not the POSIX way. Note that setting
9635 PCRE_MULTILINE has only some of the effects specified for REG_NEWLINE.
9636 It does not affect the way newlines are matched by . (they are not) or
9637 by a negative class such as [^a] (they are).
9639 The yield of regcomp() is zero on success, and non-zero otherwise. The
9640 preg structure is filled in on success, and one member of the structure
9641 is public: re_nsub contains the number of capturing subpatterns in the
9642 regular expression. Various error codes are defined in the header file.
9644 NOTE: If the yield of regcomp() is non-zero, you must not attempt to
9645 use the contents of the preg structure. If, for example, you pass it to
9646 regexec(), the result is undefined and your program is likely to crash.
9649 MATCHING NEWLINE CHARACTERS
9651 This area is not simple, because POSIX and Perl take different views of
9652 things. It is not possible to get PCRE to obey POSIX semantics, but
9653 then PCRE was never intended to be a POSIX engine. The following table
9654 lists the different possibilities for matching newline characters in
9659 . matches newline no PCRE_DOTALL
9660 newline matches [^a] yes not changeable
9661 $ matches \n at end yes PCRE_DOLLARENDONLY
9662 $ matches \n in middle no PCRE_MULTILINE
9663 ^ matches \n in middle no PCRE_MULTILINE
9665 This is the equivalent table for POSIX:
9669 . matches newline yes REG_NEWLINE
9670 newline matches [^a] yes REG_NEWLINE
9671 $ matches \n at end no REG_NEWLINE
9672 $ matches \n in middle no REG_NEWLINE
9673 ^ matches \n in middle no REG_NEWLINE
9675 PCRE's behaviour is the same as Perl's, except that there is no equiva-
9676 lent for PCRE_DOLLAR_ENDONLY in Perl. In both PCRE and Perl, there is
9677 no way to stop newline from matching [^a].
9679 The default POSIX newline handling can be obtained by setting
9680 PCRE_DOTALL and PCRE_DOLLAR_ENDONLY, but there is no way to make PCRE
9681 behave exactly as for the REG_NEWLINE action.
9686 The function regexec() is called to match a compiled pattern preg
9687 against a given string, which is by default terminated by a zero byte
9688 (but see REG_STARTEND below), subject to the options in eflags. These
9693 The PCRE_NOTBOL option is set when calling the underlying PCRE matching
9698 The PCRE_NOTEMPTY option is set when calling the underlying PCRE match-
9699 ing function. Note that REG_NOTEMPTY is not part of the POSIX standard.
9700 However, setting this option can give more POSIX-like behaviour in some
9705 The PCRE_NOTEOL option is set when calling the underlying PCRE matching
9710 The string is considered to start at string + pmatch[0].rm_so and to
9711 have a terminating NUL located at string + pmatch[0].rm_eo (there need
9712 not actually be a NUL at that location), regardless of the value of
9713 nmatch. This is a BSD extension, compatible with but not specified by
9714 IEEE Standard 1003.2 (POSIX.2), and should be used with caution in
9715 software intended to be portable to other systems. Note that a non-zero
9716 rm_so does not imply REG_NOTBOL; REG_STARTEND affects only the location
9717 of the string, not how it is matched.
9719 If the pattern was compiled with the REG_NOSUB flag, no data about any
9720 matched strings is returned. The nmatch and pmatch arguments of
9721 regexec() are ignored.
9723 If the value of nmatch is zero, or if the value pmatch is NULL, no data
9724 about any matched strings is returned.
9726 Otherwise,the portion of the string that was matched, and also any cap-
9727 tured substrings, are returned via the pmatch argument, which points to
9728 an array of nmatch structures of type regmatch_t, containing the mem-
9729 bers rm_so and rm_eo. These contain the offset to the first character
9730 of each substring and the offset to the first character after the end
9731 of each substring, respectively. The 0th element of the vector relates
9732 to the entire portion of string that was matched; subsequent elements
9733 relate to the capturing subpatterns of the regular expression. Unused
9734 entries in the array have both structure members set to -1.
9736 A successful match yields a zero return; various error codes are
9737 defined in the header file, of which REG_NOMATCH is the "expected"
9743 The regerror() function maps a non-zero errorcode from either regcomp()
9744 or regexec() to a printable message. If preg is not NULL, the error
9745 should have arisen from the use of that structure. A message terminated
9746 by a binary zero is placed in errbuf. The length of the message,
9747 including the zero, is limited to errbuf_size. The yield of the func-
9748 tion is the size of buffer needed to hold the whole message.
9753 Compiling a regular expression causes memory to be allocated and asso-
9754 ciated with the preg structure. The function regfree() frees all such
9755 memory, after which preg may no longer be used as a compiled expres-
9762 University Computing Service
9763 Cambridge CB2 3QH, England.
9768 Last updated: 09 January 2012
9769 Copyright (c) 1997-2012 University of Cambridge.
9770 ------------------------------------------------------------------------------
9773 PCRECPP(3) Library Functions Manual PCRECPP(3)
9778 PCRE - Perl-compatible regular expressions.
9780 SYNOPSIS OF C++ WRAPPER
9782 #include <pcrecpp.h>
9787 The C++ wrapper for PCRE was provided by Google Inc. Some additional
9788 functionality was added by Giuseppe Maxia. This brief man page was con-
9789 structed from the notes in the pcrecpp.h file, which should be con-
9790 sulted for further details. Note that the C++ wrapper supports only the
9791 original 8-bit PCRE library. There is no 16-bit or 32-bit support at
9797 The "FullMatch" operation checks that supplied text matches a supplied
9798 pattern exactly. If pointer arguments are supplied, it copies matched
9799 sub-strings that match sub-patterns into them.
9801 Example: successful match
9802 pcrecpp::RE re("h.*o");
9803 re.FullMatch("hello");
9805 Example: unsuccessful match (requires full match):
9806 pcrecpp::RE re("e");
9807 !re.FullMatch("hello");
9809 Example: creating a temporary RE object:
9810 pcrecpp::RE("h.*o").FullMatch("hello");
9812 You can pass in a "const char*" or a "string" for "text". The examples
9813 below tend to use a const char*. You can, as in the different examples
9814 above, store the RE object explicitly in a variable or use a temporary
9815 RE object. The examples below use one mode or the other arbitrarily.
9816 Either could correctly be used for any of these examples.
9818 You must supply extra pointer arguments to extract matched subpieces.
9820 Example: extracts "ruby" into "s" and 1234 into "i"
9823 pcrecpp::RE re("(\\w+):(\\d+)");
9824 re.FullMatch("ruby:1234", &s, &i);
9826 Example: does not try to extract any extra sub-patterns
9827 re.FullMatch("ruby:1234", &s);
9829 Example: does not try to extract into NULL
9830 re.FullMatch("ruby:1234", NULL, &i);
9832 Example: integer overflow causes failure
9833 !re.FullMatch("ruby:1234567891234", NULL, &i);
9835 Example: fails because there aren't enough sub-patterns:
9836 !pcrecpp::RE("\\w+:\\d+").FullMatch("ruby:1234", &s);
9838 Example: fails because string cannot be stored in integer
9839 !pcrecpp::RE("(.*)").FullMatch("ruby", &i);
9841 The provided pointer arguments can be pointers to any scalar numeric
9844 string (matched piece is copied to string)
9845 StringPiece (StringPiece is mutated to point to matched piece)
9846 T (where "bool T::ParseFrom(const char*, int)" exists)
9847 NULL (the corresponding matched sub-pattern is not copied)
9849 The function returns true iff all of the following conditions are sat-
9852 a. "text" matches "pattern" exactly;
9854 b. The number of matched sub-patterns is >= number of supplied
9857 c. The "i"th argument has a suitable type for holding the
9858 string captured as the "i"th sub-pattern. If you pass in
9859 void * NULL for the "i"th argument, or a non-void * NULL
9860 of the correct type, or pass fewer arguments than the
9861 number of sub-patterns, "i"th captured sub-pattern is
9864 CAVEAT: An optional sub-pattern that does not exist in the matched
9865 string is assigned the empty string. Therefore, the following will
9866 return false (because the empty string is not a valid number):
9869 pcrecpp::RE::FullMatch("abc", "[a-z]+(\\d+)?", &number);
9871 The matching interface supports at most 16 arguments per call. If you
9872 need more, consider using the more general interface
9873 pcrecpp::RE::DoMatch. See pcrecpp.h for the signature for DoMatch.
9875 NOTE: Do not use no_arg, which is used internally to mark the end of a
9876 list of optional arguments, as a placeholder for missing arguments, as
9877 this can lead to segfaults.
9880 QUOTING METACHARACTERS
9882 You can use the "QuoteMeta" operation to insert backslashes before all
9883 potentially meaningful characters in a string. The returned string,
9884 used as a regular expression, will exactly match the original string.
9887 string quoted = RE::QuoteMeta(unquoted);
9889 Note that it's legal to escape a character even if it has no special
9890 meaning in a regular expression -- so this function does that. (This
9891 also makes it identical to the perl function of the same name; see
9892 "perldoc -f quotemeta".) For example, "1.5-2.0?" becomes
9898 You can use the "PartialMatch" operation when you want the pattern to
9899 match any substring of the text.
9901 Example: simple search for a string:
9902 pcrecpp::RE("ell").PartialMatch("hello");
9904 Example: find first number in a string:
9906 pcrecpp::RE re("(\\d+)");
9907 re.PartialMatch("x*100 + 20", &number);
9908 assert(number == 100);
9911 UTF-8 AND THE MATCHING INTERFACE
9913 By default, pattern and text are plain text, one byte per character.
9914 The UTF8 flag, passed to the constructor, causes both pattern and
9915 string to be treated as UTF-8 text, still a byte stream but potentially
9916 multiple bytes per character. In practice, the text is likelier to be
9917 UTF-8 than the pattern, but the match returned may depend on the UTF8
9918 flag, so always use it when matching UTF8 text. For example, "." will
9919 match one byte normally but with UTF8 set may match up to three bytes
9920 of a multi-byte character.
9923 pcrecpp::RE_Options options;
9925 pcrecpp::RE re(utf8_pattern, options);
9926 re.FullMatch(utf8_string);
9928 Example: using the convenience function UTF8():
9929 pcrecpp::RE re(utf8_pattern, pcrecpp::UTF8());
9930 re.FullMatch(utf8_string);
9932 NOTE: The UTF8 flag is ignored if pcre was not configured with the
9936 PASSING MODIFIERS TO THE REGULAR EXPRESSION ENGINE
9938 PCRE defines some modifiers to change the behavior of the regular
9939 expression engine. The C++ wrapper defines an auxiliary class,
9940 RE_Options, as a vehicle to pass such modifiers to a RE class. Cur-
9941 rently, the following modifiers are supported:
9943 modifier description Perl corresponding
9945 PCRE_CASELESS case insensitive match /i
9946 PCRE_MULTILINE multiple lines match /m
9947 PCRE_DOTALL dot matches newlines /s
9948 PCRE_DOLLAR_ENDONLY $ matches only at end N/A
9949 PCRE_EXTRA strict escape parsing N/A
9950 PCRE_EXTENDED ignore white spaces /x
9951 PCRE_UTF8 handles UTF8 chars built-in
9952 PCRE_UNGREEDY reverses * and *? N/A
9953 PCRE_NO_AUTO_CAPTURE disables capturing parens N/A (*)
9955 (*) Both Perl and PCRE allow non capturing parentheses by means of the
9956 "?:" modifier within the pattern itself. e.g. (?:ab|cd) does not cap-
9957 ture, while (ab|cd) does.
9959 For a full account on how each modifier works, please check the PCRE
9962 For each modifier, there are two member functions whose name is made
9963 out of the modifier in lowercase, without the "PCRE_" prefix. For
9964 instance, PCRE_CASELESS is handled by
9968 which returns true if the modifier is set, and
9970 RE_Options & set_caseless(bool)
9972 which sets or unsets the modifier. Moreover, PCRE_EXTRA_MATCH_LIMIT can
9973 be accessed through the set_match_limit() and match_limit() member
9974 functions. Setting match_limit to a non-zero value will limit the exe-
9975 cution of pcre to keep it from doing bad things like blowing the stack
9976 or taking an eternity to return a result. A value of 5000 is good
9977 enough to stop stack blowup in a 2MB thread stack. Setting match_limit
9978 to zero disables match limiting. Alternatively, you can call
9979 match_limit_recursion() which uses PCRE_EXTRA_MATCH_LIMIT_RECURSION to
9980 limit how much PCRE recurses. match_limit() limits the number of
9981 matches PCRE does; match_limit_recursion() limits the depth of internal
9982 recursion, and therefore the amount of stack that is used.
9984 Normally, to pass one or more modifiers to a RE class, you declare a
9985 RE_Options object, set the appropriate options, and pass this object to
9986 a RE constructor. Example:
9989 opt.set_caseless(true);
9990 if (RE("HELLO", opt).PartialMatch("hello world")) ...
9992 RE_options has two constructors. The default constructor takes no argu-
9993 ments and creates a set of flags that are off by default. The optional
9994 parameter option_flags is to facilitate transfer of legacy code from C
9995 programs. This lets you do
9998 RE_Options(PCRE_CASELESS|PCRE_MULTILINE)).PartialMatch(str);
10000 However, new code is better off doing
10003 RE_Options().set_caseless(true).set_multiline(true))
10004 .PartialMatch(str);
10006 If you are going to pass one of the most used modifiers, there are some
10007 convenience functions that return a RE_Options class with the appropri-
10008 ate modifier already set: CASELESS(), UTF8(), MULTILINE(), DOTALL(),
10011 If you need to set several options at once, and you don't want to go
10012 through the pains of declaring a RE_Options object and setting several
10013 options, there is a parallel method that give you such ability on the
10014 fly. You can concatenate several set_xxxxx() member functions, since
10015 each of them returns a reference to its class object. For example, to
10016 pass PCRE_CASELESS, PCRE_EXTENDED, and PCRE_MULTILINE to a RE with one
10017 statement, you may write:
10019 RE(" ^ xyz \\s+ .* blah$",
10021 .set_caseless(true)
10022 .set_extended(true)
10023 .set_multiline(true)).PartialMatch(sometext);
10026 SCANNING TEXT INCREMENTALLY
10028 The "Consume" operation may be useful if you want to repeatedly match
10029 regular expressions at the front of a string and skip over them as they
10030 match. This requires use of the "StringPiece" type, which represents a
10031 sub-range of a real string. Like RE, StringPiece is defined in the
10034 Example: read lines of the form "var = value" from a string.
10035 string contents = ...; // Fill string somehow
10036 pcrecpp::StringPiece input(contents); // Wrap in a StringPiece
10040 pcrecpp::RE re("(\\w+) = (\\d+)\n");
10041 while (re.Consume(&input, &var, &value)) {
10045 Each successful call to "Consume" will set "var/value", and also
10046 advance "input" so it points past the matched text.
10048 The "FindAndConsume" operation is similar to "Consume" but does not
10049 anchor your match at the beginning of the string. For example, you
10050 could extract all words from a string by repeatedly calling
10052 pcrecpp::RE("(\\w+)").FindAndConsume(&input, &word)
10055 PARSING HEX/OCTAL/C-RADIX NUMBERS
10057 By default, if you pass a pointer to a numeric value, the corresponding
10058 text is interpreted as a base-10 number. You can instead wrap the
10059 pointer with a call to one of the operators Hex(), Octal(), or CRadix()
10060 to interpret the text in another base. The CRadix operator interprets
10061 C-style "0" (base-8) and "0x" (base-16) prefixes, but defaults to
10066 pcrecpp::RE re("(.*) (.*) (.*) (.*)");
10067 re.FullMatch("100 40 0100 0x40",
10068 pcrecpp::Octal(&a), pcrecpp::Hex(&b),
10069 pcrecpp::CRadix(&c), pcrecpp::CRadix(&d));
10071 will leave 64 in a, b, c, and d.
10074 REPLACING PARTS OF STRINGS
10076 You can replace the first match of "pattern" in "str" with "rewrite".
10077 Within "rewrite", backslash-escaped digits (\1 to \9) can be used to
10078 insert text matching corresponding parenthesized group from the pat-
10079 tern. \0 in "rewrite" refers to the entire matching text. For example:
10081 string s = "yabba dabba doo";
10082 pcrecpp::RE("b+").Replace("d", &s);
10084 will leave "s" containing "yada dabba doo". The result is true if the
10085 pattern matches and a replacement occurs, false otherwise.
10087 GlobalReplace is like Replace except that it replaces all occurrences
10088 of the pattern in the string with the rewrite. Replacements are not
10089 subject to re-matching. For example:
10091 string s = "yabba dabba doo";
10092 pcrecpp::RE("b+").GlobalReplace("d", &s);
10094 will leave "s" containing "yada dada doo". It returns the number of
10097 Extract is like Replace, except that if the pattern matches, "rewrite"
10098 is copied into "out" (an additional argument) with substitutions. The
10099 non-matching portions of "text" are ignored. Returns true iff a match
10100 occurred and the extraction happened successfully; if no match occurs,
10101 the string is left unaffected.
10106 The C++ wrapper was contributed by Google Inc.
10107 Copyright (c) 2007 Google Inc.
10112 Last updated: 08 January 2012
10113 ------------------------------------------------------------------------------
10116 PCRESAMPLE(3) Library Functions Manual PCRESAMPLE(3)
10121 PCRE - Perl-compatible regular expressions
10123 PCRE SAMPLE PROGRAM
10125 A simple, complete demonstration program, to get you started with using
10126 PCRE, is supplied in the file pcredemo.c in the PCRE distribution. A
10127 listing of this program is given in the pcredemo documentation. If you
10128 do not have a copy of the PCRE distribution, you can save this listing
10129 to re-create pcredemo.c.
10131 The demonstration program, which uses the original PCRE 8-bit library,
10132 compiles the regular expression that is its first argument, and matches
10133 it against the subject string in its second argument. No PCRE options
10134 are set, and default character tables are used. If matching succeeds,
10135 the program outputs the portion of the subject that matched, together
10136 with the contents of any captured substrings.
10138 If the -g option is given on the command line, the program then goes on
10139 to check for further matches of the same regular expression in the same
10140 subject string. The logic is a little bit tricky because of the possi-
10141 bility of matching an empty string. Comments in the code explain what
10144 If PCRE is installed in the standard include and library directories
10145 for your operating system, you should be able to compile the demonstra-
10146 tion program using this command:
10148 gcc -o pcredemo pcredemo.c -lpcre
10150 If PCRE is installed elsewhere, you may need to add additional options
10151 to the command line. For example, on a Unix-like system that has PCRE
10152 installed in /usr/local, you can compile the demonstration program
10153 using a command like this:
10155 gcc -o pcredemo -I/usr/local/include pcredemo.c \
10156 -L/usr/local/lib -lpcre
10158 In a Windows environment, if you want to statically link the program
10159 against a non-dll pcre.a file, you must uncomment the line that defines
10160 PCRE_STATIC before including pcre.h, because otherwise the pcre_mal-
10161 loc() and pcre_free() exported functions will be declared
10162 __declspec(dllimport), with unwanted results.
10164 Once you have compiled and linked the demonstration program, you can
10165 run simple tests like this:
10167 ./pcredemo 'cat|dog' 'the cat sat on the mat'
10168 ./pcredemo -g 'cat|dog' 'the dog sat on the cat'
10170 Note that there is a much more comprehensive test program, called
10171 pcretest, which supports many more facilities for testing regular
10172 expressions and both PCRE libraries. The pcredemo program is provided
10173 as a simple coding example.
10175 If you try to run pcredemo when PCRE is not installed in the standard
10176 library directory, you may get an error like this on some operating
10177 systems (e.g. Solaris):
10179 ld.so.1: a.out: fatal: libpcre.so.0: open failed: No such file or
10182 This is caused by the way shared library support works on those sys-
10183 tems. You need to add
10187 (for example) to the compile command to get round this problem.
10193 University Computing Service
10194 Cambridge CB2 3QH, England.
10199 Last updated: 10 January 2012
10200 Copyright (c) 1997-2012 University of Cambridge.
10201 ------------------------------------------------------------------------------
10202 PCRELIMITS(3) Library Functions Manual PCRELIMITS(3)
10207 PCRE - Perl-compatible regular expressions
10209 SIZE AND OTHER LIMITATIONS
10211 There are some size limitations in PCRE but it is hoped that they will
10212 never in practice be relevant.
10214 The maximum length of a compiled pattern is approximately 64K data
10215 units (bytes for the 8-bit library, 16-bit units for the 16-bit
10216 library, and 32-bit units for the 32-bit library) if PCRE is compiled
10217 with the default internal linkage size, which is 2 bytes for the 8-bit
10218 and 16-bit libraries, and 4 bytes for the 32-bit library. If you want
10219 to process regular expressions that are truly enormous, you can compile
10220 PCRE with an internal linkage size of 3 or 4 (when building the 16-bit
10221 or 32-bit library, 3 is rounded up to 4). See the README file in the
10222 source distribution and the pcrebuild documentation for details. In
10223 these cases the limit is substantially larger. However, the speed of
10224 execution is slower.
10226 All values in repeating quantifiers must be less than 65536.
10228 There is no limit to the number of parenthesized subpatterns, but there
10229 can be no more than 65535 capturing subpatterns. There is, however, a
10230 limit to the depth of nesting of parenthesized subpatterns of all
10231 kinds. This is imposed in order to limit the amount of system stack
10232 used at compile time. The limit can be specified when PCRE is built;
10233 the default is 250.
10235 There is a limit to the number of forward references to subsequent sub-
10236 patterns of around 200,000. Repeated forward references with fixed
10237 upper limits, for example, (?2){0,100} when subpattern number 2 is to
10238 the right, are included in the count. There is no limit to the number
10239 of backward references.
10241 The maximum length of name for a named subpattern is 32 characters, and
10242 the maximum number of named subpatterns is 10000.
10244 The maximum length of a name in a (*MARK), (*PRUNE), (*SKIP), or
10245 (*THEN) verb is 255 for the 8-bit library and 65535 for the 16-bit and
10248 The maximum length of a subject string is the largest positive number
10249 that an integer variable can hold. However, when using the traditional
10250 matching function, PCRE uses recursion to handle subpatterns and indef-
10251 inite repetition. This means that the available stack space may limit
10252 the size of a subject string that can be processed by certain patterns.
10253 For a discussion of stack issues, see the pcrestack documentation.
10259 University Computing Service
10260 Cambridge CB2 3QH, England.
10265 Last updated: 05 November 2013
10266 Copyright (c) 1997-2013 University of Cambridge.
10267 ------------------------------------------------------------------------------
10270 PCRESTACK(3) Library Functions Manual PCRESTACK(3)
10275 PCRE - Perl-compatible regular expressions
10277 PCRE DISCUSSION OF STACK USAGE
10279 When you call pcre[16|32]_exec(), it makes use of an internal function
10280 called match(). This calls itself recursively at branch points in the
10281 pattern, in order to remember the state of the match so that it can
10282 back up and try a different alternative if the first one fails. As
10283 matching proceeds deeper and deeper into the tree of possibilities, the
10284 recursion depth increases. The match() function is also called in other
10285 circumstances, for example, whenever a parenthesized sub-pattern is
10286 entered, and in certain cases of repetition.
10288 Not all calls of match() increase the recursion depth; for an item such
10289 as a* it may be called several times at the same level, after matching
10290 different numbers of a's. Furthermore, in a number of cases where the
10291 result of the recursive call would immediately be passed back as the
10292 result of the current call (a "tail recursion"), the function is just
10295 The above comments apply when pcre[16|32]_exec() is run in its normal
10296 interpretive manner. If the pattern was studied with the
10297 PCRE_STUDY_JIT_COMPILE option, and just-in-time compiling was success-
10298 ful, and the options passed to pcre[16|32]_exec() were not incompati-
10299 ble, the matching process uses the JIT-compiled code instead of the
10300 match() function. In this case, the memory requirements are handled
10301 entirely differently. See the pcrejit documentation for details.
10303 The pcre[16|32]_dfa_exec() function operates in an entirely different
10304 way, and uses recursion only when there is a regular expression recur-
10305 sion or subroutine call in the pattern. This includes the processing of
10306 assertion and "once-only" subpatterns, which are handled like subrou-
10307 tine calls. Normally, these are never very deep, and the limit on the
10308 complexity of pcre[16|32]_dfa_exec() is controlled by the amount of
10309 workspace it is given. However, it is possible to write patterns with
10310 runaway infinite recursions; such patterns will cause
10311 pcre[16|32]_dfa_exec() to run out of stack. At present, there is no
10312 protection against this.
10314 The comments that follow do NOT apply to pcre[16|32]_dfa_exec(); they
10315 are relevant only for pcre[16|32]_exec() without the JIT optimization.
10317 Reducing pcre[16|32]_exec()'s stack usage
10319 Each time that match() is actually called recursively, it uses memory
10320 from the process stack. For certain kinds of pattern and data, very
10321 large amounts of stack may be needed, despite the recognition of "tail
10322 recursion". You can often reduce the amount of recursion, and there-
10323 fore the amount of stack used, by modifying the pattern that is being
10324 matched. Consider, for example, this pattern:
10328 It matches from wherever it starts until it encounters "<inet" or the
10329 end of the data, and is the kind of pattern that might be used when
10330 processing an XML file. Each iteration of the outer parentheses matches
10331 either one character that is not "<" or a "<" that is not followed by
10332 "inet". However, each time a parenthesis is processed, a recursion
10333 occurs, so this formulation uses a stack frame for each matched charac-
10334 ter. For a long string, a lot of stack is required. Consider now this
10335 rewritten pattern, which matches exactly the same strings:
10337 ([^<]++|<(?!inet))+
10339 This uses very much less stack, because runs of characters that do not
10340 contain "<" are "swallowed" in one item inside the parentheses. Recur-
10341 sion happens only when a "<" character that is not followed by "inet"
10342 is encountered (and we assume this is relatively rare). A possessive
10343 quantifier is used to stop any backtracking into the runs of non-"<"
10344 characters, but that is not related to stack usage.
10346 This example shows that one way of avoiding stack problems when match-
10347 ing long subject strings is to write repeated parenthesized subpatterns
10348 to match more than one character whenever possible.
10350 Compiling PCRE to use heap instead of stack for pcre[16|32]_exec()
10352 In environments where stack memory is constrained, you might want to
10353 compile PCRE to use heap memory instead of stack for remembering back-
10354 up points when pcre[16|32]_exec() is running. This makes it run a lot
10355 more slowly, however. Details of how to do this are given in the pcre-
10356 build documentation. When built in this way, instead of using the
10357 stack, PCRE obtains and frees memory by calling the functions that are
10358 pointed to by the pcre[16|32]_stack_malloc and pcre[16|32]_stack_free
10359 variables. By default, these point to malloc() and free(), but you can
10360 replace the pointers to cause PCRE to use your own functions. Since the
10361 block sizes are always the same, and are always freed in reverse order,
10362 it may be possible to implement customized memory handlers that are
10363 more efficient than the standard functions.
10365 Limiting pcre[16|32]_exec()'s stack usage
10367 You can set limits on the number of times that match() is called, both
10368 in total and recursively. If a limit is exceeded, pcre[16|32]_exec()
10369 returns an error code. Setting suitable limits should prevent it from
10370 running out of stack. The default values of the limits are very large,
10371 and unlikely ever to operate. They can be changed when PCRE is built,
10372 and they can also be set when pcre[16|32]_exec() is called. For details
10373 of these interfaces, see the pcrebuild documentation and the section on
10374 extra data for pcre[16|32]_exec() in the pcreapi documentation.
10376 As a very rough rule of thumb, you should reckon on about 500 bytes per
10377 recursion. Thus, if you want to limit your stack usage to 8Mb, you
10378 should set the limit at 16000 recursions. A 64Mb stack, on the other
10379 hand, can support around 128000 recursions.
10381 In Unix-like environments, the pcretest test program has a command line
10382 option (-S) that can be used to increase the size of its stack. As long
10383 as the stack is large enough, another option (-M) can be used to find
10384 the smallest limits that allow a particular pattern to match a given
10385 subject string. This is done by calling pcre[16|32]_exec() repeatedly
10386 with different limits.
10388 Obtaining an estimate of stack usage
10390 The actual amount of stack used per recursion can vary quite a lot,
10391 depending on the compiler that was used to build PCRE and the optimiza-
10392 tion or debugging options that were set for it. The rule of thumb value
10393 of 500 bytes mentioned above may be larger or smaller than what is
10394 actually needed. A better approximation can be obtained by running this
10399 The -C option causes pcretest to output information about the options
10400 with which PCRE was compiled. When -m is also given (before -C), infor-
10401 mation about stack use is given in a line like this:
10403 Match recursion uses stack: approximate frame size = 640 bytes
10405 The value is approximate because some recursions need a bit more (up to
10406 perhaps 16 more bytes).
10408 If the above command is given when PCRE is compiled to use the heap
10409 instead of the stack for recursion, the value that is output is the
10410 size of each block that is obtained from the heap.
10412 Changing stack size in Unix-like systems
10414 In Unix-like environments, there is not often a problem with the stack
10415 unless very long strings are involved, though the default limit on
10416 stack size varies from system to system. Values from 8Mb to 64Mb are
10417 common. You can find your default limit by running the command:
10421 Unfortunately, the effect of running out of stack is often SIGSEGV,
10422 though sometimes a more explicit error message is given. You can nor-
10423 mally increase the limit on stack size by code such as this:
10425 struct rlimit rlim;
10426 getrlimit(RLIMIT_STACK, &rlim);
10427 rlim.rlim_cur = 100*1024*1024;
10428 setrlimit(RLIMIT_STACK, &rlim);
10430 This reads the current limits (soft and hard) using getrlimit(), then
10431 attempts to increase the soft limit to 100Mb using setrlimit(). You
10432 must do this before calling pcre[16|32]_exec().
10434 Changing stack size in Mac OS X
10436 Using setrlimit(), as described above, should also work on Mac OS X. It
10437 is also possible to set a stack size when linking a program. There is a
10438 discussion about stack sizes in Mac OS X at this web site:
10439 http://developer.apple.com/qa/qa2005/qa1419.html.
10445 University Computing Service
10446 Cambridge CB2 3QH, England.
10451 Last updated: 24 June 2012
10452 Copyright (c) 1997-2012 University of Cambridge.
10453 ------------------------------------------------------------------------------