Annotation of embedaddon/pcre/doc/pcre.txt, revision 1.1.1.4
1.1 misho 1: -----------------------------------------------------------------------------
2: This file contains a concatenation of the PCRE man pages, converted to plain
3: text format for ease of searching with a text editor, or for use on systems
4: that do not have a man page processor. The small individual files that give
5: synopses of each function in the library have not been included. Neither has
6: the pcredemo program. There are separate text files for the pcregrep and
7: pcretest commands.
8: -----------------------------------------------------------------------------
9:
10:
1.1.1.4 ! misho 11: PCRE(3) Library Functions Manual PCRE(3)
! 12:
1.1 misho 13:
14:
15: NAME
16: PCRE - Perl-compatible regular expressions
17:
18: INTRODUCTION
19:
20: The PCRE library is a set of functions that implement regular expres-
21: sion pattern matching using the same syntax and semantics as Perl, with
22: just a few differences. Some features that appeared in Python and PCRE
23: before they appeared in Perl are also available using the Python syn-
24: tax, there is some support for one or two .NET and Oniguruma syntax
25: items, and there is an option for requesting some minor changes that
26: give better JavaScript compatibility.
27:
1.1.1.2 misho 28: Starting with release 8.30, it is possible to compile two separate PCRE
29: libraries: the original, which supports 8-bit character strings
30: (including UTF-8 strings), and a second library that supports 16-bit
31: character strings (including UTF-16 strings). The build process allows
32: either one or both to be built. The majority of the work to make this
33: possible was done by Zoltan Herczeg.
34:
1.1.1.4 ! misho 35: Starting with release 8.32 it is possible to compile a third separate
! 36: PCRE library that supports 32-bit character strings (including UTF-32
! 37: strings). The build process allows any combination of the 8-, 16- and
! 38: 32-bit libraries. The work to make this possible was done by Christian
! 39: Persch.
! 40:
! 41: The three libraries contain identical sets of functions, except that
! 42: the names in the 16-bit library start with pcre16_ instead of pcre_,
! 43: and the names in the 32-bit library start with pcre32_ instead of
! 44: pcre_. To avoid over-complication and reduce the documentation mainte-
! 45: nance load, most of the documentation describes the 8-bit library, with
! 46: the differences for the 16-bit and 32-bit libraries described sepa-
! 47: rately in the pcre16 and pcre32 pages. References to functions or
! 48: structures of the form pcre[16|32]_xxx should be read as meaning
! 49: "pcre_xxx when using the 8-bit library, pcre16_xxx when using the
! 50: 16-bit library, or pcre32_xxx when using the 32-bit library".
1.1.1.2 misho 51:
1.1 misho 52: The current implementation of PCRE corresponds approximately with Perl
1.1.1.4 ! misho 53: 5.12, including support for UTF-8/16/32 encoded strings and Unicode
! 54: general category properties. However, UTF-8/16/32 and Unicode support
! 55: has to be explicitly enabled; it is not the default. The Unicode tables
! 56: correspond to Unicode release 6.2.0.
1.1 misho 57:
58: In addition to the Perl-compatible matching function, PCRE contains an
59: alternative function that matches the same compiled patterns in a dif-
60: ferent way. In certain circumstances, the alternative function has some
61: advantages. For a discussion of the two matching algorithms, see the
62: pcrematching page.
63:
64: PCRE is written in C and released as a C library. A number of people
65: have written wrappers and interfaces of various kinds. In particular,
1.1.1.2 misho 66: Google Inc. have provided a comprehensive C++ wrapper for the 8-bit
67: library. This is now included as part of the PCRE distribution. The
68: pcrecpp page has details of this interface. Other people's contribu-
69: tions can be found in the Contrib directory at the primary FTP site,
70: which is:
1.1 misho 71:
72: ftp://ftp.csx.cam.ac.uk/pub/software/programming/pcre
73:
1.1.1.2 misho 74: Details of exactly which Perl regular expression features are and are
1.1 misho 75: not supported by PCRE are given in separate documents. See the pcrepat-
1.1.1.2 misho 76: tern and pcrecompat pages. There is a syntax summary in the pcresyntax
1.1 misho 77: page.
78:
1.1.1.2 misho 79: Some features of PCRE can be included, excluded, or changed when the
80: library is built. The pcre_config() function makes it possible for a
81: client to discover which features are available. The features them-
82: selves are described in the pcrebuild page. Documentation about build-
83: ing PCRE for various operating systems can be found in the README and
1.1.1.4 ! misho 84: NON-AUTOTOOLS_BUILD files in the source distribution.
1.1 misho 85:
1.1.1.2 misho 86: The libraries contains a number of undocumented internal functions and
87: data tables that are used by more than one of the exported external
88: functions, but which are not intended for use by external callers.
1.1.1.4 ! misho 89: Their names all begin with "_pcre_" or "_pcre16_" or "_pcre32_", which
! 90: hopefully will not provoke any name clashes. In some environments, it
! 91: is possible to control which external symbols are exported when a
! 92: shared library is built, and in these cases the undocumented symbols
! 93: are not exported.
! 94:
! 95:
! 96: SECURITY CONSIDERATIONS
! 97:
! 98: If you are using PCRE in a non-UTF application that permits users to
! 99: supply arbitrary patterns for compilation, you should be aware of a
! 100: feature that allows users to turn on UTF support from within a pattern,
! 101: provided that PCRE was built with UTF support. For example, an 8-bit
! 102: pattern that begins with "(*UTF8)" or "(*UTF)" turns on UTF-8 mode,
! 103: which interprets patterns and subjects as strings of UTF-8 characters
! 104: instead of individual 8-bit characters. This causes both the pattern
! 105: and any data against which it is matched to be checked for UTF-8 valid-
! 106: ity. If the data string is very long, such a check might use suffi-
! 107: ciently many resources as to cause your application to lose perfor-
! 108: mance.
! 109:
! 110: One way of guarding against this possibility is to use the
! 111: pcre_fullinfo() function to check the compiled pattern's options for
! 112: UTF. Alternatively, from release 8.33, you can set the PCRE_NEVER_UTF
! 113: option at compile time. This causes an compile time error if a pattern
! 114: contains a UTF-setting sequence.
! 115:
! 116: If your application is one that supports UTF, be aware that validity
! 117: checking can take time. If the same data string is to be matched many
! 118: times, you can use the PCRE_NO_UTF[8|16|32]_CHECK option for the second
! 119: and subsequent matches to save redundant checks.
! 120:
! 121: Another way that performance can be hit is by running a pattern that
! 122: has a very large search tree against a string that will never match.
! 123: Nested unlimited repeats in a pattern are a common example. PCRE pro-
! 124: vides some protection against this: see the PCRE_EXTRA_MATCH_LIMIT fea-
! 125: ture in the pcreapi page.
1.1 misho 126:
127:
128: USER DOCUMENTATION
129:
1.1.1.2 misho 130: The user documentation for PCRE comprises a number of different sec-
131: tions. In the "man" format, each of these is a separate "man page". In
132: the HTML format, each is a separate page, linked from the index page.
133: In the plain text format, all the sections, except the pcredemo sec-
1.1 misho 134: tion, are concatenated, for ease of searching. The sections are as fol-
135: lows:
136:
137: pcre this document
138: pcre-config show PCRE installation configuration information
1.1.1.4 ! misho 139: pcre16 details of the 16-bit library
! 140: pcre32 details of the 32-bit library
1.1 misho 141: pcreapi details of PCRE's native C API
1.1.1.4 ! misho 142: pcrebuild building PCRE
1.1 misho 143: pcrecallout details of the callout feature
144: pcrecompat discussion of Perl compatibility
1.1.1.2 misho 145: pcrecpp details of the C++ wrapper for the 8-bit library
1.1 misho 146: pcredemo a demonstration C program that uses PCRE
1.1.1.2 misho 147: pcregrep description of the pcregrep command (8-bit only)
1.1 misho 148: pcrejit discussion of the just-in-time optimization support
149: pcrelimits details of size and other limits
150: pcrematching discussion of the two matching algorithms
151: pcrepartial details of the partial matching facility
152: pcrepattern syntax and semantics of supported
153: regular expressions
154: pcreperform discussion of performance issues
1.1.1.2 misho 155: pcreposix the POSIX-compatible C API for the 8-bit library
1.1 misho 156: pcreprecompile details of saving and re-using precompiled patterns
157: pcresample discussion of the pcredemo program
158: pcrestack discussion of stack usage
159: pcresyntax quick syntax reference
160: pcretest description of the pcretest testing command
1.1.1.4 ! misho 161: pcreunicode discussion of Unicode and UTF-8/16/32 support
1.1 misho 162:
1.1.1.2 misho 163: In addition, in the "man" and HTML formats, there is a short page for
1.1.1.4 ! misho 164: each C library function, listing its arguments and results.
1.1 misho 165:
166:
167: AUTHOR
168:
169: Philip Hazel
170: University Computing Service
171: Cambridge CB2 3QH, England.
172:
1.1.1.2 misho 173: Putting an actual email address here seems to have been a spam magnet,
174: so I've taken it away. If you want to email me, use my two initials,
1.1 misho 175: followed by the two digits 10, at the domain cam.ac.uk.
176:
177:
178: REVISION
179:
1.1.1.4 ! misho 180: Last updated: 13 May 2013
! 181: Copyright (c) 1997-2013 University of Cambridge.
1.1.1.2 misho 182: ------------------------------------------------------------------------------
183:
184:
1.1.1.4 ! misho 185: PCRE(3) Library Functions Manual PCRE(3)
! 186:
1.1.1.2 misho 187:
188:
189: NAME
190: PCRE - Perl-compatible regular expressions
191:
192: #include <pcre.h>
193:
194:
195: PCRE 16-BIT API BASIC FUNCTIONS
196:
197: pcre16 *pcre16_compile(PCRE_SPTR16 pattern, int options,
198: const char **errptr, int *erroffset,
199: const unsigned char *tableptr);
200:
201: pcre16 *pcre16_compile2(PCRE_SPTR16 pattern, int options,
202: int *errorcodeptr,
203: const char **errptr, int *erroffset,
204: const unsigned char *tableptr);
205:
206: pcre16_extra *pcre16_study(const pcre16 *code, int options,
207: const char **errptr);
208:
209: void pcre16_free_study(pcre16_extra *extra);
210:
211: int pcre16_exec(const pcre16 *code, const pcre16_extra *extra,
212: PCRE_SPTR16 subject, int length, int startoffset,
213: int options, int *ovector, int ovecsize);
214:
215: int pcre16_dfa_exec(const pcre16 *code, const pcre16_extra *extra,
216: PCRE_SPTR16 subject, int length, int startoffset,
217: int options, int *ovector, int ovecsize,
218: int *workspace, int wscount);
219:
220:
221: PCRE 16-BIT API STRING EXTRACTION FUNCTIONS
222:
223: int pcre16_copy_named_substring(const pcre16 *code,
224: PCRE_SPTR16 subject, int *ovector,
225: int stringcount, PCRE_SPTR16 stringname,
226: PCRE_UCHAR16 *buffer, int buffersize);
227:
228: int pcre16_copy_substring(PCRE_SPTR16 subject, int *ovector,
229: int stringcount, int stringnumber, PCRE_UCHAR16 *buffer,
230: int buffersize);
231:
232: int pcre16_get_named_substring(const pcre16 *code,
233: PCRE_SPTR16 subject, int *ovector,
234: int stringcount, PCRE_SPTR16 stringname,
235: PCRE_SPTR16 *stringptr);
236:
237: int pcre16_get_stringnumber(const pcre16 *code,
238: PCRE_SPTR16 name);
239:
240: int pcre16_get_stringtable_entries(const pcre16 *code,
241: PCRE_SPTR16 name, PCRE_UCHAR16 **first, PCRE_UCHAR16 **last);
242:
243: int pcre16_get_substring(PCRE_SPTR16 subject, int *ovector,
244: int stringcount, int stringnumber,
245: PCRE_SPTR16 *stringptr);
246:
247: int pcre16_get_substring_list(PCRE_SPTR16 subject,
248: int *ovector, int stringcount, PCRE_SPTR16 **listptr);
249:
250: void pcre16_free_substring(PCRE_SPTR16 stringptr);
251:
252: void pcre16_free_substring_list(PCRE_SPTR16 *stringptr);
253:
254:
255: PCRE 16-BIT API AUXILIARY FUNCTIONS
256:
257: pcre16_jit_stack *pcre16_jit_stack_alloc(int startsize, int maxsize);
258:
259: void pcre16_jit_stack_free(pcre16_jit_stack *stack);
260:
261: void pcre16_assign_jit_stack(pcre16_extra *extra,
262: pcre16_jit_callback callback, void *data);
263:
264: const unsigned char *pcre16_maketables(void);
265:
266: int pcre16_fullinfo(const pcre16 *code, const pcre16_extra *extra,
267: int what, void *where);
268:
269: int pcre16_refcount(pcre16 *code, int adjust);
270:
271: int pcre16_config(int what, void *where);
272:
273: const char *pcre16_version(void);
274:
275: int pcre16_pattern_to_host_byte_order(pcre16 *code,
276: pcre16_extra *extra, const unsigned char *tables);
277:
278:
279: PCRE 16-BIT API INDIRECTED FUNCTIONS
280:
281: void *(*pcre16_malloc)(size_t);
282:
283: void (*pcre16_free)(void *);
284:
285: void *(*pcre16_stack_malloc)(size_t);
286:
287: void (*pcre16_stack_free)(void *);
288:
289: int (*pcre16_callout)(pcre16_callout_block *);
290:
291:
292: PCRE 16-BIT API 16-BIT-ONLY FUNCTION
293:
294: int pcre16_utf16_to_host_byte_order(PCRE_UCHAR16 *output,
295: PCRE_SPTR16 input, int length, int *byte_order,
296: int keep_boms);
297:
298:
299: THE PCRE 16-BIT LIBRARY
300:
301: Starting with release 8.30, it is possible to compile a PCRE library
302: that supports 16-bit character strings, including UTF-16 strings, as
303: well as or instead of the original 8-bit library. The majority of the
304: work to make this possible was done by Zoltan Herczeg. The two
305: libraries contain identical sets of functions, used in exactly the same
306: way. Only the names of the functions and the data types of their argu-
307: ments and results are different. To avoid over-complication and reduce
308: the documentation maintenance load, most of the PCRE documentation
309: describes the 8-bit library, with only occasional references to the
310: 16-bit library. This page describes what is different when you use the
311: 16-bit library.
312:
313: WARNING: A single application can be linked with both libraries, but
314: you must take care when processing any particular pattern to use func-
315: tions from just one library. For example, if you want to study a pat-
316: tern that was compiled with pcre16_compile(), you must do so with
317: pcre16_study(), not pcre_study(), and you must free the study data with
318: pcre16_free_study().
319:
320:
321: THE HEADER FILE
322:
323: There is only one header file, pcre.h. It contains prototypes for all
1.1.1.4 ! misho 324: the functions in all libraries, as well as definitions of flags, struc-
! 325: tures, error codes, etc.
1.1.1.2 misho 326:
327:
328: THE LIBRARY NAME
329:
330: In Unix-like systems, the 16-bit library is called libpcre16, and can
331: normally be accesss by adding -lpcre16 to the command for linking an
332: application that uses PCRE.
333:
334:
335: STRING TYPES
336:
337: In the 8-bit library, strings are passed to PCRE library functions as
338: vectors of bytes with the C type "char *". In the 16-bit library,
339: strings are passed as vectors of unsigned 16-bit quantities. The macro
340: PCRE_UCHAR16 specifies an appropriate data type, and PCRE_SPTR16 is
341: defined as "const PCRE_UCHAR16 *". In very many environments, "short
342: int" is a 16-bit data type. When PCRE is built, it defines PCRE_UCHAR16
1.1.1.4 ! misho 343: as "unsigned short int", but checks that it really is a 16-bit data
! 344: type. If it is not, the build fails with an error message telling the
! 345: maintainer to modify the definition appropriately.
1.1.1.2 misho 346:
347:
348: STRUCTURE TYPES
349:
350: The types of the opaque structures that are used for compiled 16-bit
351: patterns and JIT stacks are pcre16 and pcre16_jit_stack respectively.
352: The type of the user-accessible structure that is returned by
353: pcre16_study() is pcre16_extra, and the type of the structure that is
354: used for passing data to a callout function is pcre16_callout_block.
355: These structures contain the same fields, with the same names, as their
356: 8-bit counterparts. The only difference is that pointers to character
357: strings are 16-bit instead of 8-bit types.
358:
359:
360: 16-BIT FUNCTIONS
361:
362: For every function in the 8-bit library there is a corresponding func-
363: tion in the 16-bit library with a name that starts with pcre16_ instead
364: of pcre_. The prototypes are listed above. In addition, there is one
365: extra function, pcre16_utf16_to_host_byte_order(). This is a utility
366: function that converts a UTF-16 character string to host byte order if
367: necessary. The other 16-bit functions expect the strings they are
368: passed to be in host byte order.
369:
370: The input and output arguments of pcre16_utf16_to_host_byte_order() may
371: point to the same address, that is, conversion in place is supported.
372: The output buffer must be at least as long as the input.
373:
374: The length argument specifies the number of 16-bit data units in the
375: input string; a negative value specifies a zero-terminated string.
376:
377: If byte_order is NULL, it is assumed that the string starts off in host
378: byte order. This may be changed by byte-order marks (BOMs) anywhere in
379: the string (commonly as the first character).
380:
381: If byte_order is not NULL, a non-zero value of the integer to which it
382: points means that the input starts off in host byte order, otherwise
383: the opposite order is assumed. Again, BOMs in the string can change
384: this. The final byte order is passed back at the end of processing.
385:
386: If keep_boms is not zero, byte-order mark characters (0xfeff) are
387: copied into the output string. Otherwise they are discarded.
388:
389: The result of the function is the number of 16-bit units placed into
390: the output buffer, including the zero terminator if the string was
391: zero-terminated.
392:
393:
394: SUBJECT STRING OFFSETS
395:
1.1.1.4 ! misho 396: The lengths and starting offsets of subject strings must be specified
! 397: in 16-bit data units, and the offsets within subject strings that are
! 398: returned by the matching functions are in also 16-bit units rather than
! 399: bytes.
1.1.1.2 misho 400:
401:
402: NAMED SUBPATTERNS
403:
404: The name-to-number translation table that is maintained for named sub-
405: patterns uses 16-bit characters. The pcre16_get_stringtable_entries()
406: function returns the length of each entry in the table as the number of
407: 16-bit data units.
408:
409:
410: OPTION NAMES
411:
412: There are two new general option names, PCRE_UTF16 and
413: PCRE_NO_UTF16_CHECK, which correspond to PCRE_UTF8 and
414: PCRE_NO_UTF8_CHECK in the 8-bit library. In fact, these new options
1.1.1.3 misho 415: define the same bits in the options word. There is a discussion about
416: the validity of UTF-16 strings in the pcreunicode page.
1.1.1.2 misho 417:
1.1.1.3 misho 418: For the pcre16_config() function there is an option PCRE_CONFIG_UTF16
419: that returns 1 if UTF-16 support is configured, otherwise 0. If this
1.1.1.4 ! misho 420: option is given to pcre_config() or pcre32_config(), or if the
! 421: PCRE_CONFIG_UTF8 or PCRE_CONFIG_UTF32 option is given to pcre16_con-
! 422: fig(), the result is the PCRE_ERROR_BADOPTION error.
1.1.1.2 misho 423:
424:
425: CHARACTER CODES
426:
1.1.1.4 ! misho 427: In 16-bit mode, when PCRE_UTF16 is not set, character values are
1.1.1.2 misho 428: treated in the same way as in 8-bit, non UTF-8 mode, except, of course,
1.1.1.4 ! misho 429: that they can range from 0 to 0xffff instead of 0 to 0xff. Character
! 430: types for characters less than 0xff can therefore be influenced by the
! 431: locale in the same way as before. Characters greater than 0xff have
1.1.1.2 misho 432: only one case, and no "type" (such as letter or digit).
433:
1.1.1.4 ! misho 434: In UTF-16 mode, the character code is Unicode, in the range 0 to
! 435: 0x10ffff, with the exception of values in the range 0xd800 to 0xdfff
! 436: because those are "surrogate" values that are used in pairs to encode
1.1.1.2 misho 437: values greater than 0xffff.
438:
1.1.1.4 ! misho 439: A UTF-16 string can indicate its endianness by special code knows as a
1.1.1.2 misho 440: byte-order mark (BOM). The PCRE functions do not handle this, expecting
1.1.1.4 ! misho 441: strings to be in host byte order. A utility function called
! 442: pcre16_utf16_to_host_byte_order() is provided to help with this (see
1.1.1.2 misho 443: above).
444:
445:
446: ERROR NAMES
447:
1.1.1.4 ! misho 448: The errors PCRE_ERROR_BADUTF16_OFFSET and PCRE_ERROR_SHORTUTF16 corre-
! 449: spond to their 8-bit counterparts. The error PCRE_ERROR_BADMODE is
! 450: given when a compiled pattern is passed to a function that processes
! 451: patterns in the other mode, for example, if a pattern compiled with
1.1.1.2 misho 452: pcre_compile() is passed to pcre16_exec().
453:
1.1.1.4 ! misho 454: There are new error codes whose names begin with PCRE_UTF16_ERR for
! 455: invalid UTF-16 strings, corresponding to the PCRE_UTF8_ERR codes for
! 456: UTF-8 strings that are described in the section entitled "Reason codes
! 457: for invalid UTF-8 strings" in the main pcreapi page. The UTF-16 errors
1.1.1.2 misho 458: are:
459:
460: PCRE_UTF16_ERR1 Missing low surrogate at end of string
461: PCRE_UTF16_ERR2 Invalid low surrogate follows high surrogate
462: PCRE_UTF16_ERR3 Isolated low surrogate
1.1.1.4 ! misho 463: PCRE_UTF16_ERR4 Non-character
1.1.1.2 misho 464:
465:
466: ERROR TEXTS
467:
1.1.1.4 ! misho 468: If there is an error while compiling a pattern, the error text that is
! 469: passed back by pcre16_compile() or pcre16_compile2() is still an 8-bit
1.1.1.2 misho 470: character string, zero-terminated.
471:
472:
473: CALLOUTS
474:
1.1.1.4 ! misho 475: The subject and mark fields in the callout block that is passed to a
1.1.1.2 misho 476: callout function point to 16-bit vectors.
477:
478:
479: TESTING
480:
1.1.1.4 ! misho 481: The pcretest program continues to operate with 8-bit input and output
! 482: files, but it can be used for testing the 16-bit library. If it is run
1.1.1.2 misho 483: with the command line option -16, patterns and subject strings are con-
484: verted from 8-bit to 16-bit before being passed to PCRE, and the 16-bit
1.1.1.4 ! misho 485: library functions are used instead of the 8-bit ones. Returned 16-bit
! 486: strings are converted to 8-bit for output. If both the 8-bit and the
! 487: 32-bit libraries were not compiled, pcretest defaults to 16-bit and the
! 488: -16 option is ignored.
1.1.1.2 misho 489:
1.1.1.3 misho 490: When PCRE is being built, the RunTest script that is called by "make
1.1.1.4 ! misho 491: check" uses the pcretest -C option to discover which of the 8-bit,
! 492: 16-bit and 32-bit libraries has been built, and runs the tests appro-
! 493: priately.
1.1.1.2 misho 494:
495:
496: NOT SUPPORTED IN 16-BIT MODE
497:
498: Not all the features of the 8-bit library are available with the 16-bit
1.1.1.4 ! misho 499: library. The C++ and POSIX wrapper functions support only the 8-bit
1.1.1.2 misho 500: library, and the pcregrep program is at present 8-bit only.
501:
502:
503: AUTHOR
504:
505: Philip Hazel
506: University Computing Service
507: Cambridge CB2 3QH, England.
508:
509:
510: REVISION
511:
1.1.1.4 ! misho 512: Last updated: 12 May 2013
! 513: Copyright (c) 1997-2013 University of Cambridge.
1.1 misho 514: ------------------------------------------------------------------------------
515:
516:
1.1.1.4 ! misho 517: PCRE(3) Library Functions Manual PCRE(3)
! 518:
1.1 misho 519:
520:
521: NAME
522: PCRE - Perl-compatible regular expressions
523:
1.1.1.4 ! misho 524: #include <pcre.h>
! 525:
! 526:
! 527: PCRE 32-BIT API BASIC FUNCTIONS
! 528:
! 529: pcre32 *pcre32_compile(PCRE_SPTR32 pattern, int options,
! 530: const char **errptr, int *erroffset,
! 531: const unsigned char *tableptr);
! 532:
! 533: pcre32 *pcre32_compile2(PCRE_SPTR32 pattern, int options,
! 534: int *errorcodeptr,
! 535: const char **errptr, int *erroffset,
! 536: const unsigned char *tableptr);
! 537:
! 538: pcre32_extra *pcre32_study(const pcre32 *code, int options,
! 539: const char **errptr);
! 540:
! 541: void pcre32_free_study(pcre32_extra *extra);
! 542:
! 543: int pcre32_exec(const pcre32 *code, const pcre32_extra *extra,
! 544: PCRE_SPTR32 subject, int length, int startoffset,
! 545: int options, int *ovector, int ovecsize);
! 546:
! 547: int pcre32_dfa_exec(const pcre32 *code, const pcre32_extra *extra,
! 548: PCRE_SPTR32 subject, int length, int startoffset,
! 549: int options, int *ovector, int ovecsize,
! 550: int *workspace, int wscount);
! 551:
! 552:
! 553: PCRE 32-BIT API STRING EXTRACTION FUNCTIONS
! 554:
! 555: int pcre32_copy_named_substring(const pcre32 *code,
! 556: PCRE_SPTR32 subject, int *ovector,
! 557: int stringcount, PCRE_SPTR32 stringname,
! 558: PCRE_UCHAR32 *buffer, int buffersize);
! 559:
! 560: int pcre32_copy_substring(PCRE_SPTR32 subject, int *ovector,
! 561: int stringcount, int stringnumber, PCRE_UCHAR32 *buffer,
! 562: int buffersize);
! 563:
! 564: int pcre32_get_named_substring(const pcre32 *code,
! 565: PCRE_SPTR32 subject, int *ovector,
! 566: int stringcount, PCRE_SPTR32 stringname,
! 567: PCRE_SPTR32 *stringptr);
! 568:
! 569: int pcre32_get_stringnumber(const pcre32 *code,
! 570: PCRE_SPTR32 name);
! 571:
! 572: int pcre32_get_stringtable_entries(const pcre32 *code,
! 573: PCRE_SPTR32 name, PCRE_UCHAR32 **first, PCRE_UCHAR32 **last);
! 574:
! 575: int pcre32_get_substring(PCRE_SPTR32 subject, int *ovector,
! 576: int stringcount, int stringnumber,
! 577: PCRE_SPTR32 *stringptr);
! 578:
! 579: int pcre32_get_substring_list(PCRE_SPTR32 subject,
! 580: int *ovector, int stringcount, PCRE_SPTR32 **listptr);
! 581:
! 582: void pcre32_free_substring(PCRE_SPTR32 stringptr);
! 583:
! 584: void pcre32_free_substring_list(PCRE_SPTR32 *stringptr);
! 585:
! 586:
! 587: PCRE 32-BIT API AUXILIARY FUNCTIONS
! 588:
! 589: pcre32_jit_stack *pcre32_jit_stack_alloc(int startsize, int maxsize);
! 590:
! 591: void pcre32_jit_stack_free(pcre32_jit_stack *stack);
! 592:
! 593: void pcre32_assign_jit_stack(pcre32_extra *extra,
! 594: pcre32_jit_callback callback, void *data);
! 595:
! 596: const unsigned char *pcre32_maketables(void);
! 597:
! 598: int pcre32_fullinfo(const pcre32 *code, const pcre32_extra *extra,
! 599: int what, void *where);
! 600:
! 601: int pcre32_refcount(pcre32 *code, int adjust);
! 602:
! 603: int pcre32_config(int what, void *where);
! 604:
! 605: const char *pcre32_version(void);
! 606:
! 607: int pcre32_pattern_to_host_byte_order(pcre32 *code,
! 608: pcre32_extra *extra, const unsigned char *tables);
! 609:
! 610:
! 611: PCRE 32-BIT API INDIRECTED FUNCTIONS
! 612:
! 613: void *(*pcre32_malloc)(size_t);
! 614:
! 615: void (*pcre32_free)(void *);
! 616:
! 617: void *(*pcre32_stack_malloc)(size_t);
! 618:
! 619: void (*pcre32_stack_free)(void *);
! 620:
! 621: int (*pcre32_callout)(pcre32_callout_block *);
! 622:
! 623:
! 624: PCRE 32-BIT API 32-BIT-ONLY FUNCTION
! 625:
! 626: int pcre32_utf32_to_host_byte_order(PCRE_UCHAR32 *output,
! 627: PCRE_SPTR32 input, int length, int *byte_order,
! 628: int keep_boms);
! 629:
! 630:
! 631: THE PCRE 32-BIT LIBRARY
! 632:
! 633: Starting with release 8.32, it is possible to compile a PCRE library
! 634: that supports 32-bit character strings, including UTF-32 strings, as
! 635: well as or instead of the original 8-bit library. This work was done by
! 636: Christian Persch, based on the work done by Zoltan Herczeg for the
! 637: 16-bit library. All three libraries contain identical sets of func-
! 638: tions, used in exactly the same way. Only the names of the functions
! 639: and the data types of their arguments and results are different. To
! 640: avoid over-complication and reduce the documentation maintenance load,
! 641: most of the PCRE documentation describes the 8-bit library, with only
! 642: occasional references to the 16-bit and 32-bit libraries. This page
! 643: describes what is different when you use the 32-bit library.
! 644:
! 645: WARNING: A single application can be linked with all or any of the
! 646: three libraries, but you must take care when processing any particular
! 647: pattern to use functions from just one library. For example, if you
! 648: want to study a pattern that was compiled with pcre32_compile(), you
! 649: must do so with pcre32_study(), not pcre_study(), and you must free the
! 650: study data with pcre32_free_study().
! 651:
! 652:
! 653: THE HEADER FILE
! 654:
! 655: There is only one header file, pcre.h. It contains prototypes for all
! 656: the functions in all libraries, as well as definitions of flags, struc-
! 657: tures, error codes, etc.
! 658:
! 659:
! 660: THE LIBRARY NAME
! 661:
! 662: In Unix-like systems, the 32-bit library is called libpcre32, and can
! 663: normally be accesss by adding -lpcre32 to the command for linking an
! 664: application that uses PCRE.
! 665:
! 666:
! 667: STRING TYPES
! 668:
! 669: In the 8-bit library, strings are passed to PCRE library functions as
! 670: vectors of bytes with the C type "char *". In the 32-bit library,
! 671: strings are passed as vectors of unsigned 32-bit quantities. The macro
! 672: PCRE_UCHAR32 specifies an appropriate data type, and PCRE_SPTR32 is
! 673: defined as "const PCRE_UCHAR32 *". In very many environments, "unsigned
! 674: int" is a 32-bit data type. When PCRE is built, it defines PCRE_UCHAR32
! 675: as "unsigned int", but checks that it really is a 32-bit data type. If
! 676: it is not, the build fails with an error message telling the maintainer
! 677: to modify the definition appropriately.
! 678:
! 679:
! 680: STRUCTURE TYPES
! 681:
! 682: The types of the opaque structures that are used for compiled 32-bit
! 683: patterns and JIT stacks are pcre32 and pcre32_jit_stack respectively.
! 684: The type of the user-accessible structure that is returned by
! 685: pcre32_study() is pcre32_extra, and the type of the structure that is
! 686: used for passing data to a callout function is pcre32_callout_block.
! 687: These structures contain the same fields, with the same names, as their
! 688: 8-bit counterparts. The only difference is that pointers to character
! 689: strings are 32-bit instead of 8-bit types.
! 690:
! 691:
! 692: 32-BIT FUNCTIONS
! 693:
! 694: For every function in the 8-bit library there is a corresponding func-
! 695: tion in the 32-bit library with a name that starts with pcre32_ instead
! 696: of pcre_. The prototypes are listed above. In addition, there is one
! 697: extra function, pcre32_utf32_to_host_byte_order(). This is a utility
! 698: function that converts a UTF-32 character string to host byte order if
! 699: necessary. The other 32-bit functions expect the strings they are
! 700: passed to be in host byte order.
! 701:
! 702: The input and output arguments of pcre32_utf32_to_host_byte_order() may
! 703: point to the same address, that is, conversion in place is supported.
! 704: The output buffer must be at least as long as the input.
! 705:
! 706: The length argument specifies the number of 32-bit data units in the
! 707: input string; a negative value specifies a zero-terminated string.
! 708:
! 709: If byte_order is NULL, it is assumed that the string starts off in host
! 710: byte order. This may be changed by byte-order marks (BOMs) anywhere in
! 711: the string (commonly as the first character).
! 712:
! 713: If byte_order is not NULL, a non-zero value of the integer to which it
! 714: points means that the input starts off in host byte order, otherwise
! 715: the opposite order is assumed. Again, BOMs in the string can change
! 716: this. The final byte order is passed back at the end of processing.
! 717:
! 718: If keep_boms is not zero, byte-order mark characters (0xfeff) are
! 719: copied into the output string. Otherwise they are discarded.
! 720:
! 721: The result of the function is the number of 32-bit units placed into
! 722: the output buffer, including the zero terminator if the string was
! 723: zero-terminated.
! 724:
! 725:
! 726: SUBJECT STRING OFFSETS
! 727:
! 728: The lengths and starting offsets of subject strings must be specified
! 729: in 32-bit data units, and the offsets within subject strings that are
! 730: returned by the matching functions are in also 32-bit units rather than
! 731: bytes.
! 732:
! 733:
! 734: NAMED SUBPATTERNS
! 735:
! 736: The name-to-number translation table that is maintained for named sub-
! 737: patterns uses 32-bit characters. The pcre32_get_stringtable_entries()
! 738: function returns the length of each entry in the table as the number of
! 739: 32-bit data units.
! 740:
! 741:
! 742: OPTION NAMES
! 743:
! 744: There are two new general option names, PCRE_UTF32 and
! 745: PCRE_NO_UTF32_CHECK, which correspond to PCRE_UTF8 and
! 746: PCRE_NO_UTF8_CHECK in the 8-bit library. In fact, these new options
! 747: define the same bits in the options word. There is a discussion about
! 748: the validity of UTF-32 strings in the pcreunicode page.
! 749:
! 750: For the pcre32_config() function there is an option PCRE_CONFIG_UTF32
! 751: that returns 1 if UTF-32 support is configured, otherwise 0. If this
! 752: option is given to pcre_config() or pcre16_config(), or if the
! 753: PCRE_CONFIG_UTF8 or PCRE_CONFIG_UTF16 option is given to pcre32_con-
! 754: fig(), the result is the PCRE_ERROR_BADOPTION error.
! 755:
! 756:
! 757: CHARACTER CODES
! 758:
! 759: In 32-bit mode, when PCRE_UTF32 is not set, character values are
! 760: treated in the same way as in 8-bit, non UTF-8 mode, except, of course,
! 761: that they can range from 0 to 0x7fffffff instead of 0 to 0xff. Charac-
! 762: ter types for characters less than 0xff can therefore be influenced by
! 763: the locale in the same way as before. Characters greater than 0xff
! 764: have only one case, and no "type" (such as letter or digit).
! 765:
! 766: In UTF-32 mode, the character code is Unicode, in the range 0 to
! 767: 0x10ffff, with the exception of values in the range 0xd800 to 0xdfff
! 768: because those are "surrogate" values that are ill-formed in UTF-32.
! 769:
! 770: A UTF-32 string can indicate its endianness by special code knows as a
! 771: byte-order mark (BOM). The PCRE functions do not handle this, expecting
! 772: strings to be in host byte order. A utility function called
! 773: pcre32_utf32_to_host_byte_order() is provided to help with this (see
! 774: above).
! 775:
! 776:
! 777: ERROR NAMES
! 778:
! 779: The error PCRE_ERROR_BADUTF32 corresponds to its 8-bit counterpart.
! 780: The error PCRE_ERROR_BADMODE is given when a compiled pattern is passed
! 781: to a function that processes patterns in the other mode, for example,
! 782: if a pattern compiled with pcre_compile() is passed to pcre32_exec().
! 783:
! 784: There are new error codes whose names begin with PCRE_UTF32_ERR for
! 785: invalid UTF-32 strings, corresponding to the PCRE_UTF8_ERR codes for
! 786: UTF-8 strings that are described in the section entitled "Reason codes
! 787: for invalid UTF-8 strings" in the main pcreapi page. The UTF-32 errors
! 788: are:
! 789:
! 790: PCRE_UTF32_ERR1 Surrogate character (range from 0xd800 to 0xdfff)
! 791: PCRE_UTF32_ERR2 Non-character
! 792: PCRE_UTF32_ERR3 Character > 0x10ffff
! 793:
! 794:
! 795: ERROR TEXTS
! 796:
! 797: If there is an error while compiling a pattern, the error text that is
! 798: passed back by pcre32_compile() or pcre32_compile2() is still an 8-bit
! 799: character string, zero-terminated.
! 800:
! 801:
! 802: CALLOUTS
! 803:
! 804: The subject and mark fields in the callout block that is passed to a
! 805: callout function point to 32-bit vectors.
! 806:
! 807:
! 808: TESTING
! 809:
! 810: The pcretest program continues to operate with 8-bit input and output
! 811: files, but it can be used for testing the 32-bit library. If it is run
! 812: with the command line option -32, patterns and subject strings are con-
! 813: verted from 8-bit to 32-bit before being passed to PCRE, and the 32-bit
! 814: library functions are used instead of the 8-bit ones. Returned 32-bit
! 815: strings are converted to 8-bit for output. If both the 8-bit and the
! 816: 16-bit libraries were not compiled, pcretest defaults to 32-bit and the
! 817: -32 option is ignored.
! 818:
! 819: When PCRE is being built, the RunTest script that is called by "make
! 820: check" uses the pcretest -C option to discover which of the 8-bit,
! 821: 16-bit and 32-bit libraries has been built, and runs the tests appro-
! 822: priately.
! 823:
! 824:
! 825: NOT SUPPORTED IN 32-BIT MODE
! 826:
! 827: Not all the features of the 8-bit library are available with the 32-bit
! 828: library. The C++ and POSIX wrapper functions support only the 8-bit
! 829: library, and the pcregrep program is at present 8-bit only.
! 830:
! 831:
! 832: AUTHOR
! 833:
! 834: Philip Hazel
! 835: University Computing Service
! 836: Cambridge CB2 3QH, England.
! 837:
! 838:
! 839: REVISION
! 840:
! 841: Last updated: 12 May 2013
! 842: Copyright (c) 1997-2013 University of Cambridge.
! 843: ------------------------------------------------------------------------------
! 844:
! 845:
! 846: PCREBUILD(3) Library Functions Manual PCREBUILD(3)
! 847:
! 848:
! 849:
! 850: NAME
! 851: PCRE - Perl-compatible regular expressions
! 852:
! 853: BUILDING PCRE
! 854:
! 855: PCRE is distributed with a configure script that can be used to build
! 856: the library in Unix-like environments using the applications known as
! 857: Autotools. Also in the distribution are files to support building
! 858: using CMake instead of configure. The text file README contains general
! 859: information about building with Autotools (some of which is repeated
! 860: below), and also has some comments about building on various operating
! 861: systems. There is a lot more information about building PCRE without
! 862: using Autotools (including information about using CMake and building
! 863: "by hand") in the text file called NON-AUTOTOOLS-BUILD. You should
! 864: consult this file as well as the README file if you are building in a
! 865: non-Unix-like environment.
! 866:
1.1 misho 867:
868: PCRE BUILD-TIME OPTIONS
869:
1.1.1.4 ! misho 870: The rest of this document describes the optional features of PCRE that
! 871: can be selected when the library is compiled. It assumes use of the
! 872: configure script, where the optional features are selected or dese-
! 873: lected by providing options to configure before running the make com-
! 874: mand. However, the same options can be selected in both Unix-like and
! 875: non-Unix-like environments using the GUI facility of cmake-gui if you
! 876: are using CMake instead of configure to build PCRE.
! 877:
! 878: If you are not using Autotools or CMake, option selection can be done
! 879: by editing the config.h file, or by passing parameter settings to the
! 880: compiler, as described in NON-AUTOTOOLS-BUILD.
1.1 misho 881:
882: The complete list of options for configure (which includes the standard
1.1.1.4 ! misho 883: ones such as the selection of the installation directory) can be
1.1 misho 884: obtained by running
885:
886: ./configure --help
887:
1.1.1.4 ! misho 888: The following sections include descriptions of options whose names
1.1 misho 889: begin with --enable or --disable. These settings specify changes to the
1.1.1.4 ! misho 890: defaults for the configure command. Because of the way that configure
! 891: works, --enable and --disable always come in pairs, so the complemen-
! 892: tary option always exists as well, but as it specifies the default, it
1.1 misho 893: is not described.
894:
895:
1.1.1.4 ! misho 896: BUILDING 8-BIT, 16-BIT AND 32-BIT LIBRARIES
1.1.1.2 misho 897:
1.1.1.4 ! misho 898: By default, a library called libpcre is built, containing functions
! 899: that take string arguments contained in vectors of bytes, either as
! 900: single-byte characters, or interpreted as UTF-8 strings. You can also
! 901: build a separate library, called libpcre16, in which strings are con-
! 902: tained in vectors of 16-bit data units and interpreted either as sin-
1.1.1.2 misho 903: gle-unit characters or UTF-16 strings, by adding
904:
905: --enable-pcre16
906:
1.1.1.4 ! misho 907: to the configure command. You can also build yet another separate
! 908: library, called libpcre32, in which strings are contained in vectors of
! 909: 32-bit data units and interpreted either as single-unit characters or
! 910: UTF-32 strings, by adding
! 911:
! 912: --enable-pcre32
! 913:
1.1.1.2 misho 914: to the configure command. If you do not want the 8-bit library, add
915:
916: --disable-pcre8
917:
1.1.1.4 ! misho 918: as well. At least one of the three libraries must be built. Note that
! 919: the C++ and POSIX wrappers are for the 8-bit library only, and that
! 920: pcregrep is an 8-bit program. None of these are built if you select
! 921: only the 16-bit or 32-bit libraries.
1.1.1.2 misho 922:
923:
1.1 misho 924: BUILDING SHARED AND STATIC LIBRARIES
925:
1.1.1.4 ! misho 926: The Autotools PCRE building process uses libtool to build both shared
! 927: and static libraries by default. You can suppress one of these by
! 928: adding one of
1.1 misho 929:
930: --disable-shared
931: --disable-static
932:
933: to the configure command, as required.
934:
935:
936: C++ SUPPORT
937:
1.1.1.2 misho 938: By default, if the 8-bit library is being built, the configure script
939: will search for a C++ compiler and C++ header files. If it finds them,
940: it automatically builds the C++ wrapper library (which supports only
941: 8-bit strings). You can disable this by adding
1.1 misho 942:
943: --disable-cpp
944:
945: to the configure command.
946:
947:
1.1.1.4 ! misho 948: UTF-8, UTF-16 AND UTF-32 SUPPORT
1.1 misho 949:
1.1.1.2 misho 950: To build PCRE with support for UTF Unicode character strings, add
1.1 misho 951:
1.1.1.2 misho 952: --enable-utf
1.1 misho 953:
1.1.1.4 ! misho 954: to the configure command. This setting applies to all three libraries,
! 955: adding support for UTF-8 to the 8-bit library, support for UTF-16 to
! 956: the 16-bit library, and support for UTF-32 to the to the 32-bit
! 957: library. There are no separate options for enabling UTF-8, UTF-16 and
! 958: UTF-32 independently because that would allow ridiculous settings such
! 959: as requesting UTF-16 support while building only the 8-bit library. It
! 960: is not possible to build one library with UTF support and another with-
! 961: out in the same configuration. (For backwards compatibility, --enable-
! 962: utf8 is a synonym of --enable-utf.)
! 963:
! 964: Of itself, this setting does not make PCRE treat strings as UTF-8,
! 965: UTF-16 or UTF-32. As well as compiling PCRE with this option, you also
! 966: have have to set the PCRE_UTF8, PCRE_UTF16 or PCRE_UTF32 option (as
! 967: appropriate) when you call one of the pattern compiling functions.
1.1 misho 968:
1.1.1.4 ! misho 969: If you set --enable-utf when compiling in an EBCDIC environment, PCRE
! 970: expects its input to be either ASCII or UTF-8 (depending on the run-
1.1.1.3 misho 971: time option). It is not possible to support both EBCDIC and UTF-8 codes
1.1.1.4 ! misho 972: in the same version of the library. Consequently, --enable-utf and
1.1 misho 973: --enable-ebcdic are mutually exclusive.
974:
975:
976: UNICODE CHARACTER PROPERTY SUPPORT
977:
1.1.1.4 ! misho 978: UTF support allows the libraries to process character codepoints up to
! 979: 0x10ffff in the strings that they handle. On its own, however, it does
1.1.1.2 misho 980: not provide any facilities for accessing the properties of such charac-
981: ters. If you want to be able to use the pattern escapes \P, \p, and \X,
982: which refer to Unicode character properties, you must add
1.1 misho 983:
984: --enable-unicode-properties
985:
1.1.1.4 ! misho 986: to the configure command. This implies UTF support, even if you have
1.1 misho 987: not explicitly requested it.
988:
1.1.1.4 ! misho 989: Including Unicode property support adds around 30K of tables to the
! 990: PCRE library. Only the general category properties such as Lu and Nd
1.1 misho 991: are supported. Details are given in the pcrepattern documentation.
992:
993:
994: JUST-IN-TIME COMPILER SUPPORT
995:
996: Just-in-time compiler support is included in the build by specifying
997:
998: --enable-jit
999:
1.1.1.4 ! misho 1000: This support is available only for certain hardware architectures. If
! 1001: this option is set for an unsupported architecture, a compile time
! 1002: error occurs. See the pcrejit documentation for a discussion of JIT
1.1 misho 1003: usage. When JIT support is enabled, pcregrep automatically makes use of
1004: it, unless you add
1005:
1006: --disable-pcregrep-jit
1007:
1008: to the "configure" command.
1009:
1010:
1011: CODE VALUE OF NEWLINE
1012:
1.1.1.4 ! misho 1013: By default, PCRE interprets the linefeed (LF) character as indicating
! 1014: the end of a line. This is the normal newline character on Unix-like
! 1015: systems. You can compile PCRE to use carriage return (CR) instead, by
1.1 misho 1016: adding
1017:
1018: --enable-newline-is-cr
1019:
1.1.1.4 ! misho 1020: to the configure command. There is also a --enable-newline-is-lf
1.1 misho 1021: option, which explicitly specifies linefeed as the newline character.
1022:
1023: Alternatively, you can specify that line endings are to be indicated by
1024: the two character sequence CRLF. If you want this, add
1025:
1026: --enable-newline-is-crlf
1027:
1028: to the configure command. There is a fourth option, specified by
1029:
1030: --enable-newline-is-anycrlf
1031:
1.1.1.4 ! misho 1032: which causes PCRE to recognize any of the three sequences CR, LF, or
1.1 misho 1033: CRLF as indicating a line ending. Finally, a fifth option, specified by
1034:
1035: --enable-newline-is-any
1036:
1037: causes PCRE to recognize any Unicode newline sequence.
1038:
1.1.1.4 ! misho 1039: Whatever line ending convention is selected when PCRE is built can be
! 1040: overridden when the library functions are called. At build time it is
1.1 misho 1041: conventional to use the standard for your operating system.
1042:
1043:
1044: WHAT \R MATCHES
1045:
1.1.1.4 ! misho 1046: By default, the sequence \R in a pattern matches any Unicode newline
! 1047: sequence, whatever has been selected as the line ending sequence. If
1.1 misho 1048: you specify
1049:
1050: --enable-bsr-anycrlf
1051:
1.1.1.4 ! misho 1052: the default is changed so that \R matches only CR, LF, or CRLF. What-
! 1053: ever is selected when PCRE is built can be overridden when the library
1.1 misho 1054: functions are called.
1055:
1056:
1057: POSIX MALLOC USAGE
1058:
1.1.1.4 ! misho 1059: When the 8-bit library is called through the POSIX interface (see the
! 1060: pcreposix documentation), additional working storage is required for
! 1061: holding the pointers to capturing substrings, because PCRE requires
1.1.1.2 misho 1062: three integers per substring, whereas the POSIX interface provides only
1.1.1.4 ! misho 1063: two. If the number of expected substrings is small, the wrapper func-
! 1064: tion uses space on the stack, because this is faster than using mal-
! 1065: loc() for each call. The default threshold above which the stack is no
1.1.1.2 misho 1066: longer used is 10; it can be changed by adding a setting such as
1.1 misho 1067:
1068: --with-posix-malloc-threshold=20
1069:
1070: to the configure command.
1071:
1072:
1073: HANDLING VERY LARGE PATTERNS
1074:
1.1.1.4 ! misho 1075: Within a compiled pattern, offset values are used to point from one
! 1076: part to another (for example, from an opening parenthesis to an alter-
! 1077: nation metacharacter). By default, in the 8-bit and 16-bit libraries,
! 1078: two-byte values are used for these offsets, leading to a maximum size
! 1079: for a compiled pattern of around 64K. This is sufficient to handle all
! 1080: but the most gigantic patterns. Nevertheless, some people do want to
! 1081: process truly enormous patterns, so it is possible to compile PCRE to
! 1082: use three-byte or four-byte offsets by adding a setting such as
1.1 misho 1083:
1084: --with-link-size=3
1085:
1.1.1.4 ! misho 1086: to the configure command. The value given must be 2, 3, or 4. For the
! 1087: 16-bit library, a value of 3 is rounded up to 4. In these libraries,
! 1088: using longer offsets slows down the operation of PCRE because it has to
! 1089: load additional data when handling them. For the 32-bit library the
! 1090: value is always 4 and cannot be overridden; the value of --with-link-
! 1091: size is ignored.
1.1 misho 1092:
1093:
1094: AVOIDING EXCESSIVE STACK USAGE
1095:
1096: When matching with the pcre_exec() function, PCRE implements backtrack-
1.1.1.4 ! misho 1097: ing by making recursive calls to an internal function called match().
! 1098: In environments where the size of the stack is limited, this can se-
! 1099: verely limit PCRE's operation. (The Unix environment does not usually
1.1 misho 1100: suffer from this problem, but it may sometimes be necessary to increase
1.1.1.4 ! misho 1101: the maximum stack size. There is a discussion in the pcrestack docu-
! 1102: mentation.) An alternative approach to recursion that uses memory from
! 1103: the heap to remember data, instead of using recursive function calls,
! 1104: has been implemented to work round the problem of limited stack size.
1.1 misho 1105: If you want to build a version of PCRE that works this way, add
1106:
1107: --disable-stack-for-recursion
1108:
1.1.1.4 ! misho 1109: to the configure command. With this configuration, PCRE will use the
! 1110: pcre_stack_malloc and pcre_stack_free variables to call memory manage-
! 1111: ment functions. By default these point to malloc() and free(), but you
1.1 misho 1112: can replace the pointers so that your own functions are used instead.
1113:
1.1.1.4 ! misho 1114: Separate functions are provided rather than using pcre_malloc and
! 1115: pcre_free because the usage is very predictable: the block sizes
! 1116: requested are always the same, and the blocks are always freed in
! 1117: reverse order. A calling program might be able to implement optimized
! 1118: functions that perform better than malloc() and free(). PCRE runs
1.1 misho 1119: noticeably more slowly when built in this way. This option affects only
1120: the pcre_exec() function; it is not relevant for pcre_dfa_exec().
1121:
1122:
1123: LIMITING PCRE RESOURCE USAGE
1124:
1.1.1.4 ! misho 1125: Internally, PCRE has a function called match(), which it calls repeat-
! 1126: edly (sometimes recursively) when matching a pattern with the
! 1127: pcre_exec() function. By controlling the maximum number of times this
! 1128: function may be called during a single matching operation, a limit can
! 1129: be placed on the resources used by a single call to pcre_exec(). The
! 1130: limit can be changed at run time, as described in the pcreapi documen-
! 1131: tation. The default is 10 million, but this can be changed by adding a
1.1 misho 1132: setting such as
1133:
1134: --with-match-limit=500000
1135:
1.1.1.4 ! misho 1136: to the configure command. This setting has no effect on the
1.1 misho 1137: pcre_dfa_exec() matching function.
1138:
1.1.1.4 ! misho 1139: In some environments it is desirable to limit the depth of recursive
1.1 misho 1140: calls of match() more strictly than the total number of calls, in order
1.1.1.4 ! misho 1141: to restrict the maximum amount of stack (or heap, if --disable-stack-
1.1 misho 1142: for-recursion is specified) that is used. A second limit controls this;
1.1.1.4 ! misho 1143: it defaults to the value that is set for --with-match-limit, which
! 1144: imposes no additional constraints. However, you can set a lower limit
1.1 misho 1145: by adding, for example,
1146:
1147: --with-match-limit-recursion=10000
1148:
1.1.1.4 ! misho 1149: to the configure command. This value can also be overridden at run
1.1 misho 1150: time.
1151:
1152:
1153: CREATING CHARACTER TABLES AT BUILD TIME
1154:
1.1.1.4 ! misho 1155: PCRE uses fixed tables for processing characters whose code values are
! 1156: less than 256. By default, PCRE is built with a set of tables that are
! 1157: distributed in the file pcre_chartables.c.dist. These tables are for
1.1 misho 1158: ASCII codes only. If you add
1159:
1160: --enable-rebuild-chartables
1161:
1.1.1.4 ! misho 1162: to the configure command, the distributed tables are no longer used.
! 1163: Instead, a program called dftables is compiled and run. This outputs
1.1 misho 1164: the source for new set of tables, created in the default locale of your
1.1.1.4 ! misho 1165: C run-time system. (This method of replacing the tables does not work
! 1166: if you are cross compiling, because dftables is run on the local host.
1.1.1.3 misho 1167: If you need to create alternative tables when cross compiling, you will
1.1 misho 1168: have to do so "by hand".)
1169:
1170:
1171: USING EBCDIC CODE
1172:
1.1.1.4 ! misho 1173: PCRE assumes by default that it will run in an environment where the
! 1174: character code is ASCII (or Unicode, which is a superset of ASCII).
! 1175: This is the case for most computer operating systems. PCRE can, how-
1.1 misho 1176: ever, be compiled to run in an EBCDIC environment by adding
1177:
1178: --enable-ebcdic
1179:
1180: to the configure command. This setting implies --enable-rebuild-charta-
1.1.1.4 ! misho 1181: bles. You should only use it if you know that you are in an EBCDIC
! 1182: environment (for example, an IBM mainframe operating system). The
1.1.1.2 misho 1183: --enable-ebcdic option is incompatible with --enable-utf.
1.1 misho 1184:
1.1.1.4 ! misho 1185: The EBCDIC character that corresponds to an ASCII LF is assumed to have
! 1186: the value 0x15 by default. However, in some EBCDIC environments, 0x25
! 1187: is used. In such an environment you should use
! 1188:
! 1189: --enable-ebcdic-nl25
! 1190:
! 1191: as well as, or instead of, --enable-ebcdic. The EBCDIC character for CR
! 1192: has the same value as in ASCII, namely, 0x0d. Whichever of 0x15 and
! 1193: 0x25 is not chosen as LF is made to correspond to the Unicode NEL char-
! 1194: acter (which, in Unicode, is 0x85).
! 1195:
! 1196: The options that select newline behaviour, such as --enable-newline-is-
! 1197: cr, and equivalent run-time options, refer to these character values in
! 1198: an EBCDIC environment.
! 1199:
1.1 misho 1200:
1201: PCREGREP OPTIONS FOR COMPRESSED FILE SUPPORT
1202:
1203: By default, pcregrep reads all files as plain text. You can build it so
1204: that it recognizes files whose names end in .gz or .bz2, and reads them
1205: with libz or libbz2, respectively, by adding one or both of
1206:
1207: --enable-pcregrep-libz
1208: --enable-pcregrep-libbz2
1209:
1210: to the configure command. These options naturally require that the rel-
1.1.1.2 misho 1211: evant libraries are installed on your system. Configuration will fail
1.1 misho 1212: if they are not.
1213:
1214:
1215: PCREGREP BUFFER SIZE
1216:
1.1.1.2 misho 1217: pcregrep uses an internal buffer to hold a "window" on the file it is
1.1 misho 1218: scanning, in order to be able to output "before" and "after" lines when
1.1.1.2 misho 1219: it finds a match. The size of the buffer is controlled by a parameter
1.1 misho 1220: whose default value is 20K. The buffer itself is three times this size,
1221: but because of the way it is used for holding "before" lines, the long-
1.1.1.2 misho 1222: est line that is guaranteed to be processable is the parameter size.
1.1 misho 1223: You can change the default parameter value by adding, for example,
1224:
1225: --with-pcregrep-bufsize=50K
1226:
1227: to the configure command. The caller of pcregrep can, however, override
1228: this value by specifying a run-time option.
1229:
1230:
1231: PCRETEST OPTION FOR LIBREADLINE SUPPORT
1232:
1233: If you add
1234:
1235: --enable-pcretest-libreadline
1236:
1.1.1.2 misho 1237: to the configure command, pcretest is linked with the libreadline
1238: library, and when its input is from a terminal, it reads it using the
1.1 misho 1239: readline() function. This provides line-editing and history facilities.
1240: Note that libreadline is GPL-licensed, so if you distribute a binary of
1241: pcretest linked in this way, there may be licensing issues.
1242:
1.1.1.2 misho 1243: Setting this option causes the -lreadline option to be added to the
1244: pcretest build. In many operating environments with a sytem-installed
1.1 misho 1245: libreadline this is sufficient. However, in some environments (e.g. if
1.1.1.2 misho 1246: an unmodified distribution version of readline is in use), some extra
1247: configuration may be necessary. The INSTALL file for libreadline says
1.1 misho 1248: this:
1249:
1250: "Readline uses the termcap functions, but does not link with the
1251: termcap or curses library itself, allowing applications which link
1252: with readline the to choose an appropriate library."
1253:
1.1.1.2 misho 1254: If your environment has not been set up so that an appropriate library
1.1 misho 1255: is automatically included, you may need to add something like
1256:
1257: LIBS="-ncurses"
1258:
1259: immediately before the configure command.
1260:
1261:
1.1.1.4 ! misho 1262: DEBUGGING WITH VALGRIND SUPPORT
! 1263:
! 1264: By adding the
! 1265:
! 1266: --enable-valgrind
! 1267:
! 1268: option to to the configure command, PCRE will use valgrind annotations
! 1269: to mark certain memory regions as unaddressable. This allows it to
! 1270: detect invalid memory accesses, and is mostly useful for debugging PCRE
! 1271: itself.
! 1272:
! 1273:
! 1274: CODE COVERAGE REPORTING
! 1275:
! 1276: If your C compiler is gcc, you can build a version of PCRE that can
! 1277: generate a code coverage report for its test suite. To enable this, you
! 1278: must install lcov version 1.6 or above. Then specify
! 1279:
! 1280: --enable-coverage
! 1281:
! 1282: to the configure command and build PCRE in the usual way.
! 1283:
! 1284: Note that using ccache (a caching C compiler) is incompatible with code
! 1285: coverage reporting. If you have configured ccache to run automatically
! 1286: on your system, you must set the environment variable
! 1287:
! 1288: CCACHE_DISABLE=1
! 1289:
! 1290: before running make to build PCRE, so that ccache is not used.
! 1291:
! 1292: When --enable-coverage is used, the following addition targets are
! 1293: added to the Makefile:
! 1294:
! 1295: make coverage
! 1296:
! 1297: This creates a fresh coverage report for the PCRE test suite. It is
! 1298: equivalent to running "make coverage-reset", "make coverage-baseline",
! 1299: "make check", and then "make coverage-report".
! 1300:
! 1301: make coverage-reset
! 1302:
! 1303: This zeroes the coverage counters, but does nothing else.
! 1304:
! 1305: make coverage-baseline
! 1306:
! 1307: This captures baseline coverage information.
! 1308:
! 1309: make coverage-report
! 1310:
! 1311: This creates the coverage report.
! 1312:
! 1313: make coverage-clean-report
! 1314:
! 1315: This removes the generated coverage report without cleaning the cover-
! 1316: age data itself.
! 1317:
! 1318: make coverage-clean-data
! 1319:
! 1320: This removes the captured coverage data without removing the coverage
! 1321: files created at compile time (*.gcno).
! 1322:
! 1323: make coverage-clean
! 1324:
! 1325: This cleans all coverage data including the generated coverage report.
! 1326: For more information about code coverage, see the gcov and lcov docu-
! 1327: mentation.
! 1328:
! 1329:
1.1 misho 1330: SEE ALSO
1331:
1.1.1.4 ! misho 1332: pcreapi(3), pcre16, pcre32, pcre_config(3).
1.1 misho 1333:
1334:
1335: AUTHOR
1336:
1337: Philip Hazel
1338: University Computing Service
1339: Cambridge CB2 3QH, England.
1340:
1341:
1342: REVISION
1343:
1.1.1.4 ! misho 1344: Last updated: 12 May 2013
! 1345: Copyright (c) 1997-2013 University of Cambridge.
1.1 misho 1346: ------------------------------------------------------------------------------
1347:
1348:
1.1.1.4 ! misho 1349: PCREMATCHING(3) Library Functions Manual PCREMATCHING(3)
! 1350:
1.1 misho 1351:
1352:
1353: NAME
1354: PCRE - Perl-compatible regular expressions
1355:
1356: PCRE MATCHING ALGORITHMS
1357:
1358: This document describes the two different algorithms that are available
1359: in PCRE for matching a compiled regular expression against a given sub-
1360: ject string. The "standard" algorithm is the one provided by the
1.1.1.4 ! misho 1361: pcre_exec(), pcre16_exec() and pcre32_exec() functions. These work in
! 1362: the same as as Perl's matching function, and provide a Perl-compatible
! 1363: matching operation. The just-in-time (JIT) optimization that is
! 1364: described in the pcrejit documentation is compatible with these func-
! 1365: tions.
! 1366:
! 1367: An alternative algorithm is provided by the pcre_dfa_exec(),
! 1368: pcre16_dfa_exec() and pcre32_dfa_exec() functions; they operate in a
! 1369: different way, and are not Perl-compatible. This alternative has advan-
! 1370: tages and disadvantages compared with the standard algorithm, and these
! 1371: are described below.
1.1 misho 1372:
1373: When there is only one possible way in which a given subject string can
1374: match a pattern, the two algorithms give the same answer. A difference
1375: arises, however, when there are multiple possibilities. For example, if
1376: the pattern
1377:
1378: ^<.*>
1379:
1380: is matched against the string
1381:
1382: <something> <something else> <something further>
1383:
1384: there are three possible answers. The standard algorithm finds only one
1385: of them, whereas the alternative algorithm finds all three.
1386:
1387:
1388: REGULAR EXPRESSIONS AS TREES
1389:
1390: The set of strings that are matched by a regular expression can be rep-
1391: resented as a tree structure. An unlimited repetition in the pattern
1392: makes the tree of infinite size, but it is still a tree. Matching the
1393: pattern to a given subject string (from a given starting point) can be
1394: thought of as a search of the tree. There are two ways to search a
1395: tree: depth-first and breadth-first, and these correspond to the two
1396: matching algorithms provided by PCRE.
1397:
1398:
1399: THE STANDARD MATCHING ALGORITHM
1400:
1401: In the terminology of Jeffrey Friedl's book "Mastering Regular Expres-
1402: sions", the standard algorithm is an "NFA algorithm". It conducts a
1403: depth-first search of the pattern tree. That is, it proceeds along a
1404: single path through the tree, checking that the subject matches what is
1405: required. When there is a mismatch, the algorithm tries any alterna-
1406: tives at the current point, and if they all fail, it backs up to the
1407: previous branch point in the tree, and tries the next alternative
1408: branch at that level. This often involves backing up (moving to the
1409: left) in the subject string as well. The order in which repetition
1410: branches are tried is controlled by the greedy or ungreedy nature of
1411: the quantifier.
1412:
1413: If a leaf node is reached, a matching string has been found, and at
1414: that point the algorithm stops. Thus, if there is more than one possi-
1415: ble match, this algorithm returns the first one that it finds. Whether
1416: this is the shortest, the longest, or some intermediate length depends
1417: on the way the greedy and ungreedy repetition quantifiers are specified
1418: in the pattern.
1419:
1420: Because it ends up with a single path through the tree, it is rela-
1421: tively straightforward for this algorithm to keep track of the sub-
1422: strings that are matched by portions of the pattern in parentheses.
1423: This provides support for capturing parentheses and back references.
1424:
1425:
1426: THE ALTERNATIVE MATCHING ALGORITHM
1427:
1428: This algorithm conducts a breadth-first search of the tree. Starting
1429: from the first matching point in the subject, it scans the subject
1430: string from left to right, once, character by character, and as it does
1431: this, it remembers all the paths through the tree that represent valid
1432: matches. In Friedl's terminology, this is a kind of "DFA algorithm",
1433: though it is not implemented as a traditional finite state machine (it
1434: keeps multiple states active simultaneously).
1435:
1436: Although the general principle of this matching algorithm is that it
1437: scans the subject string only once, without backtracking, there is one
1438: exception: when a lookaround assertion is encountered, the characters
1439: following or preceding the current point have to be independently
1440: inspected.
1441:
1442: The scan continues until either the end of the subject is reached, or
1443: there are no more unterminated paths. At this point, terminated paths
1444: represent the different matching possibilities (if there are none, the
1445: match has failed). Thus, if there is more than one possible match,
1446: this algorithm finds all of them, and in particular, it finds the long-
1447: est. The matches are returned in decreasing order of length. There is
1448: an option to stop the algorithm after the first match (which is neces-
1449: sarily the shortest) is found.
1450:
1451: Note that all the matches that are found start at the same point in the
1452: subject. If the pattern
1453:
1454: cat(er(pillar)?)?
1455:
1456: is matched against the string "the caterpillar catchment", the result
1457: will be the three strings "caterpillar", "cater", and "cat" that start
1458: at the fifth character of the subject. The algorithm does not automati-
1459: cally move on to find matches that start at later positions.
1460:
1461: There are a number of features of PCRE regular expressions that are not
1462: supported by the alternative matching algorithm. They are as follows:
1463:
1464: 1. Because the algorithm finds all possible matches, the greedy or
1465: ungreedy nature of repetition quantifiers is not relevant. Greedy and
1466: ungreedy quantifiers are treated in exactly the same way. However, pos-
1467: sessive quantifiers can make a difference when what follows could also
1468: match what is quantified, for example in a pattern like this:
1469:
1470: ^a++\w!
1471:
1472: This pattern matches "aaab!" but not "aaa!", which would be matched by
1473: a non-possessive quantifier. Similarly, if an atomic group is present,
1474: it is matched as if it were a standalone pattern at the current point,
1475: and the longest match is then "locked in" for the rest of the overall
1476: pattern.
1477:
1478: 2. When dealing with multiple paths through the tree simultaneously, it
1479: is not straightforward to keep track of captured substrings for the
1480: different matching possibilities, and PCRE's implementation of this
1481: algorithm does not attempt to do this. This means that no captured sub-
1482: strings are available.
1483:
1484: 3. Because no substrings are captured, back references within the pat-
1485: tern are not supported, and cause errors if encountered.
1486:
1487: 4. For the same reason, conditional expressions that use a backrefer-
1488: ence as the condition or test for a specific group recursion are not
1489: supported.
1490:
1491: 5. Because many paths through the tree may be active, the \K escape
1492: sequence, which resets the start of the match when encountered (but may
1493: be on some paths and not on others), is not supported. It causes an
1494: error if encountered.
1495:
1496: 6. Callouts are supported, but the value of the capture_top field is
1497: always 1, and the value of the capture_last field is always -1.
1498:
1.1.1.2 misho 1499: 7. The \C escape sequence, which (in the standard algorithm) always
1.1.1.4 ! misho 1500: matches a single data unit, even in UTF-8, UTF-16 or UTF-32 modes, is
! 1501: not supported in these modes, because the alternative algorithm moves
! 1502: through the subject string one character (not data unit) at a time, for
! 1503: all active paths through the tree.
1.1 misho 1504:
1.1.1.2 misho 1505: 8. Except for (*FAIL), the backtracking control verbs such as (*PRUNE)
1506: are not supported. (*FAIL) is supported, and behaves like a failing
1.1 misho 1507: negative assertion.
1508:
1509:
1510: ADVANTAGES OF THE ALTERNATIVE ALGORITHM
1511:
1.1.1.2 misho 1512: Using the alternative matching algorithm provides the following advan-
1.1 misho 1513: tages:
1514:
1515: 1. All possible matches (at a single point in the subject) are automat-
1.1.1.2 misho 1516: ically found, and in particular, the longest match is found. To find
1.1 misho 1517: more than one match using the standard algorithm, you have to do kludgy
1518: things with callouts.
1519:
1.1.1.2 misho 1520: 2. Because the alternative algorithm scans the subject string just
1521: once, and never needs to backtrack (except for lookbehinds), it is pos-
1522: sible to pass very long subject strings to the matching function in
1523: several pieces, checking for partial matching each time. Although it is
1524: possible to do multi-segment matching using the standard algorithm by
1525: retaining partially matched substrings, it is more complicated. The
1526: pcrepartial documentation gives details of partial matching and dis-
1527: cusses multi-segment matching.
1.1 misho 1528:
1529:
1530: DISADVANTAGES OF THE ALTERNATIVE ALGORITHM
1531:
1532: The alternative algorithm suffers from a number of disadvantages:
1533:
1.1.1.2 misho 1534: 1. It is substantially slower than the standard algorithm. This is
1535: partly because it has to search for all possible matches, but is also
1.1 misho 1536: because it is less susceptible to optimization.
1537:
1538: 2. Capturing parentheses and back references are not supported.
1539:
1540: 3. Although atomic groups are supported, their use does not provide the
1541: performance advantage that it does for the standard algorithm.
1542:
1543:
1544: AUTHOR
1545:
1546: Philip Hazel
1547: University Computing Service
1548: Cambridge CB2 3QH, England.
1549:
1550:
1551: REVISION
1552:
1.1.1.2 misho 1553: Last updated: 08 January 2012
1554: Copyright (c) 1997-2012 University of Cambridge.
1.1 misho 1555: ------------------------------------------------------------------------------
1556:
1557:
1.1.1.4 ! misho 1558: PCREAPI(3) Library Functions Manual PCREAPI(3)
! 1559:
1.1 misho 1560:
1561:
1562: NAME
1563: PCRE - Perl-compatible regular expressions
1564:
1.1.1.2 misho 1565: #include <pcre.h>
1.1 misho 1566:
1567:
1.1.1.2 misho 1568: PCRE NATIVE API BASIC FUNCTIONS
1.1 misho 1569:
1570: pcre *pcre_compile(const char *pattern, int options,
1571: const char **errptr, int *erroffset,
1572: const unsigned char *tableptr);
1573:
1574: pcre *pcre_compile2(const char *pattern, int options,
1575: int *errorcodeptr,
1576: const char **errptr, int *erroffset,
1577: const unsigned char *tableptr);
1578:
1579: pcre_extra *pcre_study(const pcre *code, int options,
1580: const char **errptr);
1581:
1582: void pcre_free_study(pcre_extra *extra);
1583:
1584: int pcre_exec(const pcre *code, const pcre_extra *extra,
1585: const char *subject, int length, int startoffset,
1586: int options, int *ovector, int ovecsize);
1587:
1588: int pcre_dfa_exec(const pcre *code, const pcre_extra *extra,
1589: const char *subject, int length, int startoffset,
1590: int options, int *ovector, int ovecsize,
1591: int *workspace, int wscount);
1592:
1.1.1.2 misho 1593:
1594: PCRE NATIVE API STRING EXTRACTION FUNCTIONS
1595:
1.1 misho 1596: int pcre_copy_named_substring(const pcre *code,
1597: const char *subject, int *ovector,
1598: int stringcount, const char *stringname,
1599: char *buffer, int buffersize);
1600:
1601: int pcre_copy_substring(const char *subject, int *ovector,
1602: int stringcount, int stringnumber, char *buffer,
1603: int buffersize);
1604:
1605: int pcre_get_named_substring(const pcre *code,
1606: const char *subject, int *ovector,
1607: int stringcount, const char *stringname,
1608: const char **stringptr);
1609:
1610: int pcre_get_stringnumber(const pcre *code,
1611: const char *name);
1612:
1613: int pcre_get_stringtable_entries(const pcre *code,
1614: const char *name, char **first, char **last);
1615:
1616: int pcre_get_substring(const char *subject, int *ovector,
1617: int stringcount, int stringnumber,
1618: const char **stringptr);
1619:
1620: int pcre_get_substring_list(const char *subject,
1621: int *ovector, int stringcount, const char ***listptr);
1622:
1623: void pcre_free_substring(const char *stringptr);
1624:
1625: void pcre_free_substring_list(const char **stringptr);
1626:
1.1.1.2 misho 1627:
1628: PCRE NATIVE API AUXILIARY FUNCTIONS
1629:
1.1.1.4 ! misho 1630: int pcre_jit_exec(const pcre *code, const pcre_extra *extra,
! 1631: const char *subject, int length, int startoffset,
! 1632: int options, int *ovector, int ovecsize,
! 1633: pcre_jit_stack *jstack);
! 1634:
1.1.1.2 misho 1635: pcre_jit_stack *pcre_jit_stack_alloc(int startsize, int maxsize);
1636:
1637: void pcre_jit_stack_free(pcre_jit_stack *stack);
1638:
1639: void pcre_assign_jit_stack(pcre_extra *extra,
1640: pcre_jit_callback callback, void *data);
1641:
1.1 misho 1642: const unsigned char *pcre_maketables(void);
1643:
1644: int pcre_fullinfo(const pcre *code, const pcre_extra *extra,
1645: int what, void *where);
1646:
1647: int pcre_refcount(pcre *code, int adjust);
1648:
1649: int pcre_config(int what, void *where);
1650:
1.1.1.2 misho 1651: const char *pcre_version(void);
1652:
1653: int pcre_pattern_to_host_byte_order(pcre *code,
1654: pcre_extra *extra, const unsigned char *tables);
1.1 misho 1655:
1656:
1657: PCRE NATIVE API INDIRECTED FUNCTIONS
1658:
1659: void *(*pcre_malloc)(size_t);
1660:
1661: void (*pcre_free)(void *);
1662:
1663: void *(*pcre_stack_malloc)(size_t);
1664:
1665: void (*pcre_stack_free)(void *);
1666:
1667: int (*pcre_callout)(pcre_callout_block *);
1668:
1669:
1.1.1.4 ! misho 1670: PCRE 8-BIT, 16-BIT, AND 32-BIT LIBRARIES
1.1.1.2 misho 1671:
1.1.1.4 ! misho 1672: As well as support for 8-bit character strings, PCRE also supports
! 1673: 16-bit strings (from release 8.30) and 32-bit strings (from release
! 1674: 8.32), by means of two additional libraries. They can be built as well
! 1675: as, or instead of, the 8-bit library. To avoid too much complication,
! 1676: this document describes the 8-bit versions of the functions, with only
! 1677: occasional references to the 16-bit and 32-bit libraries.
! 1678:
! 1679: The 16-bit and 32-bit functions operate in the same way as their 8-bit
! 1680: counterparts; they just use different data types for their arguments
! 1681: and results, and their names start with pcre16_ or pcre32_ instead of
! 1682: pcre_. For every option that has UTF8 in its name (for example,
! 1683: PCRE_UTF8), there are corresponding 16-bit and 32-bit names with UTF8
! 1684: replaced by UTF16 or UTF32, respectively. This facility is in fact just
! 1685: cosmetic; the 16-bit and 32-bit option names define the same bit val-
1.1.1.2 misho 1686: ues.
1687:
1688: References to bytes and UTF-8 in this document should be read as refer-
1.1.1.4 ! misho 1689: ences to 16-bit data units and UTF-16 when using the 16-bit library, or
! 1690: 32-bit data units and UTF-32 when using the 32-bit library, unless
! 1691: specified otherwise. More details of the specific differences for the
! 1692: 16-bit and 32-bit libraries are given in the pcre16 and pcre32 pages.
1.1.1.2 misho 1693:
1694:
1.1 misho 1695: PCRE API OVERVIEW
1696:
1697: PCRE has its own native API, which is described in this document. There
1.1.1.4 ! misho 1698: are also some wrapper functions (for the 8-bit library only) that cor-
! 1699: respond to the POSIX regular expression API, but they do not give
! 1700: access to all the functionality. They are described in the pcreposix
! 1701: documentation. Both of these APIs define a set of C function calls. A
1.1.1.2 misho 1702: C++ wrapper (again for the 8-bit library only) is also distributed with
1703: PCRE. It is documented in the pcrecpp page.
1.1 misho 1704:
1.1.1.4 ! misho 1705: The native API C function prototypes are defined in the header file
! 1706: pcre.h, and on Unix-like systems the (8-bit) library itself is called
! 1707: libpcre. It can normally be accessed by adding -lpcre to the command
! 1708: for linking an application that uses PCRE. The header file defines the
1.1.1.2 misho 1709: macros PCRE_MAJOR and PCRE_MINOR to contain the major and minor release
1.1.1.4 ! misho 1710: numbers for the library. Applications can use these to include support
1.1 misho 1711: for different releases of PCRE.
1712:
1713: In a Windows environment, if you want to statically link an application
1.1.1.4 ! misho 1714: program against a non-dll pcre.a file, you must define PCRE_STATIC
! 1715: before including pcre.h or pcrecpp.h, because otherwise the pcre_mal-
1.1 misho 1716: loc() and pcre_free() exported functions will be declared
1717: __declspec(dllimport), with unwanted results.
1718:
1.1.1.4 ! misho 1719: The functions pcre_compile(), pcre_compile2(), pcre_study(), and
! 1720: pcre_exec() are used for compiling and matching regular expressions in
! 1721: a Perl-compatible manner. A sample program that demonstrates the sim-
! 1722: plest way of using them is provided in the file called pcredemo.c in
1.1 misho 1723: the PCRE source distribution. A listing of this program is given in the
1.1.1.4 ! misho 1724: pcredemo documentation, and the pcresample documentation describes how
1.1 misho 1725: to compile and run it.
1726:
1.1.1.4 ! misho 1727: Just-in-time compiler support is an optional feature of PCRE that can
1.1 misho 1728: be built in appropriate hardware environments. It greatly speeds up the
1.1.1.4 ! misho 1729: matching performance of many patterns. Simple programs can easily
! 1730: request that it be used if available, by setting an option that is
! 1731: ignored when it is not relevant. More complicated programs might need
! 1732: to make use of the functions pcre_jit_stack_alloc(),
! 1733: pcre_jit_stack_free(), and pcre_assign_jit_stack() in order to control
! 1734: the JIT code's memory usage.
! 1735:
! 1736: From release 8.32 there is also a direct interface for JIT execution,
! 1737: which gives improved performance. The JIT-specific functions are dis-
! 1738: cussed in the pcrejit documentation.
1.1 misho 1739:
1740: A second matching function, pcre_dfa_exec(), which is not Perl-compati-
1741: ble, is also provided. This uses a different algorithm for the match-
1742: ing. The alternative algorithm finds all possible matches (at a given
1743: point in the subject), and scans the subject just once (unless there
1744: are lookbehind assertions). However, this algorithm does not return
1745: captured substrings. A description of the two matching algorithms and
1746: their advantages and disadvantages is given in the pcrematching docu-
1747: mentation.
1748:
1749: In addition to the main compiling and matching functions, there are
1750: convenience functions for extracting captured substrings from a subject
1751: string that is matched by pcre_exec(). They are:
1752:
1753: pcre_copy_substring()
1754: pcre_copy_named_substring()
1755: pcre_get_substring()
1756: pcre_get_named_substring()
1757: pcre_get_substring_list()
1758: pcre_get_stringnumber()
1759: pcre_get_stringtable_entries()
1760:
1761: pcre_free_substring() and pcre_free_substring_list() are also provided,
1762: to free the memory used for extracted strings.
1763:
1764: The function pcre_maketables() is used to build a set of character
1765: tables in the current locale for passing to pcre_compile(),
1766: pcre_exec(), or pcre_dfa_exec(). This is an optional facility that is
1767: provided for specialist use. Most commonly, no special tables are
1768: passed, in which case internal tables that are generated when PCRE is
1769: built are used.
1770:
1771: The function pcre_fullinfo() is used to find out information about a
1.1.1.2 misho 1772: compiled pattern. The function pcre_version() returns a pointer to a
1773: string containing the version of PCRE and its date of release.
1.1 misho 1774:
1775: The function pcre_refcount() maintains a reference count in a data
1776: block containing a compiled pattern. This is provided for the benefit
1777: of object-oriented applications.
1778:
1779: The global variables pcre_malloc and pcre_free initially contain the
1780: entry points of the standard malloc() and free() functions, respec-
1781: tively. PCRE calls the memory management functions via these variables,
1782: so a calling program can replace them if it wishes to intercept the
1783: calls. This should be done before calling any PCRE functions.
1784:
1785: The global variables pcre_stack_malloc and pcre_stack_free are also
1786: indirections to memory management functions. These special functions
1787: are used only when PCRE is compiled to use the heap for remembering
1788: data, instead of recursive function calls, when running the pcre_exec()
1789: function. See the pcrebuild documentation for details of how to do
1790: this. It is a non-standard way of building PCRE, for use in environ-
1791: ments that have limited stacks. Because of the greater use of memory
1792: management, it runs more slowly. Separate functions are provided so
1793: that special-purpose external code can be used for this case. When
1794: used, these functions are always called in a stack-like manner (last
1795: obtained, first freed), and always for memory blocks of the same size.
1796: There is a discussion about PCRE's stack usage in the pcrestack docu-
1797: mentation.
1798:
1799: The global variable pcre_callout initially contains NULL. It can be set
1800: by the caller to a "callout" function, which PCRE will then call at
1801: specified points during a matching operation. Details are given in the
1802: pcrecallout documentation.
1803:
1804:
1805: NEWLINES
1806:
1807: PCRE supports five different conventions for indicating line breaks in
1808: strings: a single CR (carriage return) character, a single LF (line-
1809: feed) character, the two-character sequence CRLF, any of the three pre-
1810: ceding, or any Unicode newline sequence. The Unicode newline sequences
1811: are the three just mentioned, plus the single characters VT (vertical
1.1.1.3 misho 1812: tab, U+000B), FF (form feed, U+000C), NEL (next line, U+0085), LS (line
1.1 misho 1813: separator, U+2028), and PS (paragraph separator, U+2029).
1814:
1815: Each of the first three conventions is used by at least one operating
1816: system as its standard newline sequence. When PCRE is built, a default
1817: can be specified. The default default is LF, which is the Unix stan-
1818: dard. When PCRE is run, the default can be overridden, either when a
1819: pattern is compiled, or when it is matched.
1820:
1821: At compile time, the newline convention can be specified by the options
1822: argument of pcre_compile(), or it can be specified by special text at
1823: the start of the pattern itself; this overrides any other settings. See
1824: the pcrepattern page for details of the special character sequences.
1825:
1826: In the PCRE documentation the word "newline" is used to mean "the char-
1827: acter or pair of characters that indicate a line break". The choice of
1828: newline convention affects the handling of the dot, circumflex, and
1829: dollar metacharacters, the handling of #-comments in /x mode, and, when
1830: CRLF is a recognized line ending sequence, the match position advance-
1831: ment for a non-anchored pattern. There is more detail about this in the
1832: section on pcre_exec() options below.
1833:
1834: The choice of newline convention does not affect the interpretation of
1835: the \n or \r escape sequences, nor does it affect what \R matches,
1836: which is controlled in a similar way, but by separate options.
1837:
1838:
1839: MULTITHREADING
1840:
1841: The PCRE functions can be used in multi-threading applications, with
1842: the proviso that the memory management functions pointed to by
1843: pcre_malloc, pcre_free, pcre_stack_malloc, and pcre_stack_free, and the
1844: callout function pointed to by pcre_callout, are shared by all threads.
1845:
1846: The compiled form of a regular expression is not altered during match-
1847: ing, so the same compiled pattern can safely be used by several threads
1848: at once.
1849:
1850: If the just-in-time optimization feature is being used, it needs sepa-
1851: rate memory stack areas for each thread. See the pcrejit documentation
1852: for more details.
1853:
1854:
1855: SAVING PRECOMPILED PATTERNS FOR LATER USE
1856:
1857: The compiled form of a regular expression can be saved and re-used at a
1858: later time, possibly by a different program, and even on a host other
1859: than the one on which it was compiled. Details are given in the
1.1.1.2 misho 1860: pcreprecompile documentation, which includes a description of the
1861: pcre_pattern_to_host_byte_order() function. However, compiling a regu-
1862: lar expression with one version of PCRE for use with a different ver-
1863: sion is not guaranteed to work and may cause crashes.
1.1 misho 1864:
1865:
1866: CHECKING BUILD-TIME OPTIONS
1867:
1868: int pcre_config(int what, void *where);
1869:
1.1.1.2 misho 1870: The function pcre_config() makes it possible for a PCRE client to dis-
1.1 misho 1871: cover which optional features have been compiled into the PCRE library.
1.1.1.2 misho 1872: The pcrebuild documentation has more details about these optional fea-
1.1 misho 1873: tures.
1874:
1.1.1.2 misho 1875: The first argument for pcre_config() is an integer, specifying which
1.1 misho 1876: information is required; the second argument is a pointer to a variable
1.1.1.2 misho 1877: into which the information is placed. The returned value is zero on
1878: success, or the negative error code PCRE_ERROR_BADOPTION if the value
1879: in the first argument is not recognized. The following information is
1.1 misho 1880: available:
1881:
1882: PCRE_CONFIG_UTF8
1883:
1.1.1.2 misho 1884: The output is an integer that is set to one if UTF-8 support is avail-
1.1.1.4 ! misho 1885: able; otherwise it is set to zero. This value should normally be given
! 1886: to the 8-bit version of this function, pcre_config(). If it is given to
! 1887: the 16-bit or 32-bit version of this function, the result is
1.1.1.2 misho 1888: PCRE_ERROR_BADOPTION.
1889:
1890: PCRE_CONFIG_UTF16
1891:
1892: The output is an integer that is set to one if UTF-16 support is avail-
1.1.1.4 ! misho 1893: able; otherwise it is set to zero. This value should normally be given
1.1.1.2 misho 1894: to the 16-bit version of this function, pcre16_config(). If it is given
1.1.1.4 ! misho 1895: to the 8-bit or 32-bit version of this function, the result is
! 1896: PCRE_ERROR_BADOPTION.
! 1897:
! 1898: PCRE_CONFIG_UTF32
! 1899:
! 1900: The output is an integer that is set to one if UTF-32 support is avail-
! 1901: able; otherwise it is set to zero. This value should normally be given
! 1902: to the 32-bit version of this function, pcre32_config(). If it is given
! 1903: to the 8-bit or 16-bit version of this function, the result is
! 1904: PCRE_ERROR_BADOPTION.
1.1 misho 1905:
1906: PCRE_CONFIG_UNICODE_PROPERTIES
1907:
1.1.1.4 ! misho 1908: The output is an integer that is set to one if support for Unicode
1.1 misho 1909: character properties is available; otherwise it is set to zero.
1910:
1911: PCRE_CONFIG_JIT
1912:
1913: The output is an integer that is set to one if support for just-in-time
1914: compiling is available; otherwise it is set to zero.
1915:
1.1.1.2 misho 1916: PCRE_CONFIG_JITTARGET
1917:
1.1.1.4 ! misho 1918: The output is a pointer to a zero-terminated "const char *" string. If
1.1.1.2 misho 1919: JIT support is available, the string contains the name of the architec-
1.1.1.4 ! misho 1920: ture for which the JIT compiler is configured, for example "x86 32bit
! 1921: (little endian + unaligned)". If JIT support is not available, the
1.1.1.2 misho 1922: result is NULL.
1923:
1.1 misho 1924: PCRE_CONFIG_NEWLINE
1925:
1.1.1.4 ! misho 1926: The output is an integer whose value specifies the default character
! 1927: sequence that is recognized as meaning "newline". The values that are
! 1928: supported in ASCII/Unicode environments are: 10 for LF, 13 for CR, 3338
! 1929: for CRLF, -2 for ANYCRLF, and -1 for ANY. In EBCDIC environments, CR,
! 1930: ANYCRLF, and ANY yield the same values. However, the value for LF is
! 1931: normally 21, though some EBCDIC environments use 37. The corresponding
! 1932: values for CRLF are 3349 and 3365. The default should normally corre-
1.1 misho 1933: spond to the standard sequence for your operating system.
1934:
1935: PCRE_CONFIG_BSR
1936:
1937: The output is an integer whose value indicates what character sequences
1.1.1.4 ! misho 1938: the \R escape sequence matches by default. A value of 0 means that \R
! 1939: matches any Unicode line ending sequence; a value of 1 means that \R
1.1 misho 1940: matches only CR, LF, or CRLF. The default can be overridden when a pat-
1941: tern is compiled or matched.
1942:
1943: PCRE_CONFIG_LINK_SIZE
1944:
1.1.1.4 ! misho 1945: The output is an integer that contains the number of bytes used for
1.1.1.2 misho 1946: internal linkage in compiled regular expressions. For the 8-bit
1947: library, the value can be 2, 3, or 4. For the 16-bit library, the value
1.1.1.4 ! misho 1948: is either 2 or 4 and is still a number of bytes. For the 32-bit
! 1949: library, the value is either 2 or 4 and is still a number of bytes. The
! 1950: default value of 2 is sufficient for all but the most massive patterns,
! 1951: since it allows the compiled pattern to be up to 64K in size. Larger
! 1952: values allow larger regular expressions to be compiled, at the expense
! 1953: of slower matching.
1.1 misho 1954:
1955: PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
1956:
1.1.1.2 misho 1957: The output is an integer that contains the threshold above which the
1958: POSIX interface uses malloc() for output vectors. Further details are
1.1 misho 1959: given in the pcreposix documentation.
1960:
1961: PCRE_CONFIG_MATCH_LIMIT
1962:
1.1.1.2 misho 1963: The output is a long integer that gives the default limit for the num-
1964: ber of internal matching function calls in a pcre_exec() execution.
1.1 misho 1965: Further details are given with pcre_exec() below.
1966:
1967: PCRE_CONFIG_MATCH_LIMIT_RECURSION
1968:
1969: The output is a long integer that gives the default limit for the depth
1.1.1.2 misho 1970: of recursion when calling the internal matching function in a
1971: pcre_exec() execution. Further details are given with pcre_exec()
1.1 misho 1972: below.
1973:
1974: PCRE_CONFIG_STACKRECURSE
1975:
1.1.1.2 misho 1976: The output is an integer that is set to one if internal recursion when
1.1 misho 1977: running pcre_exec() is implemented by recursive function calls that use
1.1.1.2 misho 1978: the stack to remember their state. This is the usual way that PCRE is
1.1 misho 1979: compiled. The output is zero if PCRE was compiled to use blocks of data
1.1.1.2 misho 1980: on the heap instead of recursive function calls. In this case,
1981: pcre_stack_malloc and pcre_stack_free are called to manage memory
1.1 misho 1982: blocks on the heap, thus avoiding the use of the stack.
1983:
1984:
1985: COMPILING A PATTERN
1986:
1987: pcre *pcre_compile(const char *pattern, int options,
1988: const char **errptr, int *erroffset,
1989: const unsigned char *tableptr);
1990:
1991: pcre *pcre_compile2(const char *pattern, int options,
1992: int *errorcodeptr,
1993: const char **errptr, int *erroffset,
1994: const unsigned char *tableptr);
1995:
1996: Either of the functions pcre_compile() or pcre_compile2() can be called
1997: to compile a pattern into an internal form. The only difference between
1.1.1.2 misho 1998: the two interfaces is that pcre_compile2() has an additional argument,
1999: errorcodeptr, via which a numerical error code can be returned. To
2000: avoid too much repetition, we refer just to pcre_compile() below, but
1.1 misho 2001: the information applies equally to pcre_compile2().
2002:
2003: The pattern is a C string terminated by a binary zero, and is passed in
1.1.1.2 misho 2004: the pattern argument. A pointer to a single block of memory that is
2005: obtained via pcre_malloc is returned. This contains the compiled code
1.1 misho 2006: and related data. The pcre type is defined for the returned block; this
2007: is a typedef for a structure whose contents are not externally defined.
2008: It is up to the caller to free the memory (via pcre_free) when it is no
2009: longer required.
2010:
1.1.1.2 misho 2011: Although the compiled code of a PCRE regex is relocatable, that is, it
1.1 misho 2012: does not depend on memory location, the complete pcre data block is not
1.1.1.2 misho 2013: fully relocatable, because it may contain a copy of the tableptr argu-
1.1 misho 2014: ment, which is an address (see below).
2015:
2016: The options argument contains various bit settings that affect the com-
1.1.1.2 misho 2017: pilation. It should be zero if no options are required. The available
2018: options are described below. Some of them (in particular, those that
2019: are compatible with Perl, but some others as well) can also be set and
2020: unset from within the pattern (see the detailed description in the
2021: pcrepattern documentation). For those options that can be different in
2022: different parts of the pattern, the contents of the options argument
1.1 misho 2023: specifies their settings at the start of compilation and execution. The
1.1.1.2 misho 2024: PCRE_ANCHORED, PCRE_BSR_xxx, PCRE_NEWLINE_xxx, PCRE_NO_UTF8_CHECK, and
1.1.1.3 misho 2025: PCRE_NO_START_OPTIMIZE options can be set at the time of matching as
2026: well as at compile time.
1.1 misho 2027:
2028: If errptr is NULL, pcre_compile() returns NULL immediately. Otherwise,
1.1.1.2 misho 2029: if compilation of a pattern fails, pcre_compile() returns NULL, and
1.1 misho 2030: sets the variable pointed to by errptr to point to a textual error mes-
2031: sage. This is a static string that is part of the library. You must not
1.1.1.2 misho 2032: try to free it. Normally, the offset from the start of the pattern to
1.1.1.4 ! misho 2033: the data unit that was being processed when the error was discovered is
1.1.1.2 misho 2034: placed in the variable pointed to by erroffset, which must not be NULL
2035: (if it is, an immediate error is given). However, for an invalid UTF-8
1.1.1.4 ! misho 2036: or UTF-16 string, the offset is that of the first data unit of the
! 2037: failing character.
1.1 misho 2038:
1.1.1.4 ! misho 2039: Some errors are not detected until the whole pattern has been scanned;
! 2040: in these cases, the offset passed back is the length of the pattern.
! 2041: Note that the offset is in data units, not characters, even in a UTF
! 2042: mode. It may sometimes point into the middle of a UTF-8 or UTF-16 char-
! 2043: acter.
1.1 misho 2044:
2045: If pcre_compile2() is used instead of pcre_compile(), and the error-
2046: codeptr argument is not NULL, a non-zero error code number is returned
2047: via this argument in the event of an error. This is in addition to the
2048: textual error message. Error codes and messages are listed below.
2049:
2050: If the final argument, tableptr, is NULL, PCRE uses a default set of
2051: character tables that are built when PCRE is compiled, using the
2052: default C locale. Otherwise, tableptr must be an address that is the
2053: result of a call to pcre_maketables(). This value is stored with the
2054: compiled pattern, and used again by pcre_exec(), unless another table
2055: pointer is passed to it. For more discussion, see the section on locale
2056: support below.
2057:
2058: This code fragment shows a typical straightforward call to pcre_com-
2059: pile():
2060:
2061: pcre *re;
2062: const char *error;
2063: int erroffset;
2064: re = pcre_compile(
2065: "^A.*Z", /* the pattern */
2066: 0, /* default options */
2067: &error, /* for error message */
2068: &erroffset, /* for error offset */
2069: NULL); /* use default character tables */
2070:
2071: The following names for option bits are defined in the pcre.h header
2072: file:
2073:
2074: PCRE_ANCHORED
2075:
2076: If this bit is set, the pattern is forced to be "anchored", that is, it
2077: is constrained to match only at the first matching point in the string
2078: that is being searched (the "subject string"). This effect can also be
2079: achieved by appropriate constructs in the pattern itself, which is the
2080: only way to do it in Perl.
2081:
2082: PCRE_AUTO_CALLOUT
2083:
2084: If this bit is set, pcre_compile() automatically inserts callout items,
2085: all with number 255, before each pattern item. For discussion of the
2086: callout facility, see the pcrecallout documentation.
2087:
2088: PCRE_BSR_ANYCRLF
2089: PCRE_BSR_UNICODE
2090:
2091: These options (which are mutually exclusive) control what the \R escape
2092: sequence matches. The choice is either to match only CR, LF, or CRLF,
2093: or to match any Unicode newline sequence. The default is specified when
2094: PCRE is built. It can be overridden from within the pattern, or by set-
2095: ting an option when a compiled pattern is matched.
2096:
2097: PCRE_CASELESS
2098:
2099: If this bit is set, letters in the pattern match both upper and lower
2100: case letters. It is equivalent to Perl's /i option, and it can be
2101: changed within a pattern by a (?i) option setting. In UTF-8 mode, PCRE
2102: always understands the concept of case for characters whose values are
2103: less than 128, so caseless matching is always possible. For characters
2104: with higher values, the concept of case is supported if PCRE is com-
2105: piled with Unicode property support, but not otherwise. If you want to
2106: use caseless matching for characters 128 and above, you must ensure
2107: that PCRE is compiled with Unicode property support as well as with
2108: UTF-8 support.
2109:
2110: PCRE_DOLLAR_ENDONLY
2111:
2112: If this bit is set, a dollar metacharacter in the pattern matches only
2113: at the end of the subject string. Without this option, a dollar also
2114: matches immediately before a newline at the end of the string (but not
2115: before any other newlines). The PCRE_DOLLAR_ENDONLY option is ignored
2116: if PCRE_MULTILINE is set. There is no equivalent to this option in
2117: Perl, and no way to set it within a pattern.
2118:
2119: PCRE_DOTALL
2120:
2121: If this bit is set, a dot metacharacter in the pattern matches a char-
2122: acter of any value, including one that indicates a newline. However, it
2123: only ever matches one character, even if newlines are coded as CRLF.
2124: Without this option, a dot does not match when the current position is
2125: at a newline. This option is equivalent to Perl's /s option, and it can
2126: be changed within a pattern by a (?s) option setting. A negative class
2127: such as [^a] always matches newline characters, independent of the set-
2128: ting of this option.
2129:
2130: PCRE_DUPNAMES
2131:
2132: If this bit is set, names used to identify capturing subpatterns need
2133: not be unique. This can be helpful for certain types of pattern when it
2134: is known that only one instance of the named subpattern can ever be
2135: matched. There are more details of named subpatterns below; see also
2136: the pcrepattern documentation.
2137:
2138: PCRE_EXTENDED
2139:
1.1.1.3 misho 2140: If this bit is set, white space data characters in the pattern are
2141: totally ignored except when escaped or inside a character class. White
1.1 misho 2142: space does not include the VT character (code 11). In addition, charac-
2143: ters between an unescaped # outside a character class and the next new-
2144: line, inclusive, are also ignored. This is equivalent to Perl's /x
2145: option, and it can be changed within a pattern by a (?x) option set-
2146: ting.
2147:
2148: Which characters are interpreted as newlines is controlled by the
2149: options passed to pcre_compile() or by a special sequence at the start
2150: of the pattern, as described in the section entitled "Newline conven-
2151: tions" in the pcrepattern documentation. Note that the end of this type
2152: of comment is a literal newline sequence in the pattern; escape
2153: sequences that happen to represent a newline do not count.
2154:
2155: This option makes it possible to include comments inside complicated
2156: patterns. Note, however, that this applies only to data characters.
1.1.1.3 misho 2157: White space characters may never appear within special character
1.1 misho 2158: sequences in a pattern, for example within the sequence (?( that intro-
2159: duces a conditional subpattern.
2160:
2161: PCRE_EXTRA
2162:
2163: This option was invented in order to turn on additional functionality
2164: of PCRE that is incompatible with Perl, but it is currently of very
2165: little use. When set, any backslash in a pattern that is followed by a
2166: letter that has no special meaning causes an error, thus reserving
2167: these combinations for future expansion. By default, as in Perl, a
2168: backslash followed by a letter with no special meaning is treated as a
2169: literal. (Perl can, however, be persuaded to give an error for this, by
2170: running it with the -w option.) There are at present no other features
2171: controlled by this option. It can also be set by a (?X) option setting
2172: within a pattern.
2173:
2174: PCRE_FIRSTLINE
2175:
2176: If this option is set, an unanchored pattern is required to match
2177: before or at the first newline in the subject string, though the
2178: matched text may continue over the newline.
2179:
2180: PCRE_JAVASCRIPT_COMPAT
2181:
2182: If this option is set, PCRE's behaviour is changed in some ways so that
2183: it is compatible with JavaScript rather than Perl. The changes are as
2184: follows:
2185:
2186: (1) A lone closing square bracket in a pattern causes a compile-time
2187: error, because this is illegal in JavaScript (by default it is treated
2188: as a data character). Thus, the pattern AB]CD becomes illegal when this
2189: option is set.
2190:
2191: (2) At run time, a back reference to an unset subpattern group matches
2192: an empty string (by default this causes the current matching alterna-
2193: tive to fail). A pattern such as (\1)(a) succeeds when this option is
2194: set (assuming it can find an "a" in the subject), whereas it fails by
2195: default, for Perl compatibility.
2196:
2197: (3) \U matches an upper case "U" character; by default \U causes a com-
2198: pile time error (Perl uses \U to upper case subsequent characters).
2199:
2200: (4) \u matches a lower case "u" character unless it is followed by four
2201: hexadecimal digits, in which case the hexadecimal number defines the
2202: code point to match. By default, \u causes a compile time error (Perl
2203: uses it to upper case the following character).
2204:
2205: (5) \x matches a lower case "x" character unless it is followed by two
2206: hexadecimal digits, in which case the hexadecimal number defines the
2207: code point to match. By default, as in Perl, a hexadecimal number is
2208: always expected after \x, but it may have zero, one, or two digits (so,
2209: for example, \xz matches a binary zero character followed by z).
2210:
2211: PCRE_MULTILINE
2212:
1.1.1.4 ! misho 2213: By default, for the purposes of matching "start of line" and "end of
! 2214: line", PCRE treats the subject string as consisting of a single line of
! 2215: characters, even if it actually contains newlines. The "start of line"
! 2216: metacharacter (^) matches only at the start of the string, and the "end
! 2217: of line" metacharacter ($) matches only at the end of the string, or
! 2218: before a terminating newline (except when PCRE_DOLLAR_ENDONLY is set).
! 2219: Note, however, that unless PCRE_DOTALL is set, the "any character"
! 2220: metacharacter (.) does not match at a newline. This behaviour (for ^,
! 2221: $, and dot) is the same as Perl.
! 2222:
! 2223: When PCRE_MULTILINE it is set, the "start of line" and "end of line"
! 2224: constructs match immediately following or immediately before internal
! 2225: newlines in the subject string, respectively, as well as at the very
! 2226: start and end. This is equivalent to Perl's /m option, and it can be
1.1 misho 2227: changed within a pattern by a (?m) option setting. If there are no new-
1.1.1.4 ! misho 2228: lines in a subject string, or no occurrences of ^ or $ in a pattern,
1.1 misho 2229: setting PCRE_MULTILINE has no effect.
2230:
1.1.1.4 ! misho 2231: PCRE_NEVER_UTF
! 2232:
! 2233: This option locks out interpretation of the pattern as UTF-8 (or UTF-16
! 2234: or UTF-32 in the 16-bit and 32-bit libraries). In particular, it pre-
! 2235: vents the creator of the pattern from switching to UTF interpretation
! 2236: by starting the pattern with (*UTF). This may be useful in applications
! 2237: that process patterns from external sources. The combination of
! 2238: PCRE_UTF8 and PCRE_NEVER_UTF also causes an error.
! 2239:
1.1 misho 2240: PCRE_NEWLINE_CR
2241: PCRE_NEWLINE_LF
2242: PCRE_NEWLINE_CRLF
2243: PCRE_NEWLINE_ANYCRLF
2244: PCRE_NEWLINE_ANY
2245:
2246: These options override the default newline definition that was chosen
2247: when PCRE was built. Setting the first or the second specifies that a
2248: newline is indicated by a single character (CR or LF, respectively).
2249: Setting PCRE_NEWLINE_CRLF specifies that a newline is indicated by the
2250: two-character CRLF sequence. Setting PCRE_NEWLINE_ANYCRLF specifies
2251: that any of the three preceding sequences should be recognized. Setting
2252: PCRE_NEWLINE_ANY specifies that any Unicode newline sequence should be
1.1.1.4 ! misho 2253: recognized.
1.1 misho 2254:
1.1.1.4 ! misho 2255: In an ASCII/Unicode environment, the Unicode newline sequences are the
! 2256: three just mentioned, plus the single characters VT (vertical tab,
! 2257: U+000B), FF (form feed, U+000C), NEL (next line, U+0085), LS (line sep-
! 2258: arator, U+2028), and PS (paragraph separator, U+2029). For the 8-bit
! 2259: library, the last two are recognized only in UTF-8 mode.
! 2260:
! 2261: When PCRE is compiled to run in an EBCDIC (mainframe) environment, the
! 2262: code for CR is 0x0d, the same as ASCII. However, the character code for
! 2263: LF is normally 0x15, though in some EBCDIC environments 0x25 is used.
! 2264: Whichever of these is not LF is made to correspond to Unicode's NEL
! 2265: character. EBCDIC codes are all less than 256. For more details, see
! 2266: the pcrebuild documentation.
! 2267:
! 2268: The newline setting in the options word uses three bits that are
1.1 misho 2269: treated as a number, giving eight possibilities. Currently only six are
1.1.1.4 ! misho 2270: used (default plus the five values above). This means that if you set
! 2271: more than one newline option, the combination may or may not be sensi-
1.1 misho 2272: ble. For example, PCRE_NEWLINE_CR with PCRE_NEWLINE_LF is equivalent to
1.1.1.4 ! misho 2273: PCRE_NEWLINE_CRLF, but other combinations may yield unused numbers and
1.1 misho 2274: cause an error.
2275:
1.1.1.4 ! misho 2276: The only time that a line break in a pattern is specially recognized
! 2277: when compiling is when PCRE_EXTENDED is set. CR and LF are white space
! 2278: characters, and so are ignored in this mode. Also, an unescaped # out-
! 2279: side a character class indicates a comment that lasts until after the
! 2280: next line break sequence. In other circumstances, line break sequences
1.1 misho 2281: in patterns are treated as literal data.
2282:
2283: The newline option that is set at compile time becomes the default that
2284: is used for pcre_exec() and pcre_dfa_exec(), but it can be overridden.
2285:
2286: PCRE_NO_AUTO_CAPTURE
2287:
2288: If this option is set, it disables the use of numbered capturing paren-
1.1.1.4 ! misho 2289: theses in the pattern. Any opening parenthesis that is not followed by
! 2290: ? behaves as if it were followed by ?: but named parentheses can still
! 2291: be used for capturing (and they acquire numbers in the usual way).
1.1 misho 2292: There is no equivalent of this option in Perl.
2293:
1.1.1.4 ! misho 2294: PCRE_NO_START_OPTIMIZE
1.1 misho 2295:
1.1.1.4 ! misho 2296: This is an option that acts at matching time; that is, it is really an
! 2297: option for pcre_exec() or pcre_dfa_exec(). If it is set at compile
! 2298: time, it is remembered with the compiled pattern and assumed at match-
! 2299: ing time. This is necessary if you want to use JIT execution, because
! 2300: the JIT compiler needs to know whether or not this option is set. For
! 2301: details see the discussion of PCRE_NO_START_OPTIMIZE below.
1.1 misho 2302:
2303: PCRE_UCP
2304:
2305: This option changes the way PCRE processes \B, \b, \D, \d, \S, \s, \W,
2306: \w, and some of the POSIX character classes. By default, only ASCII
2307: characters are recognized, but if PCRE_UCP is set, Unicode properties
2308: are used instead to classify characters. More details are given in the
2309: section on generic character types in the pcrepattern page. If you set
2310: PCRE_UCP, matching one of the items it affects takes much longer. The
2311: option is available only if PCRE has been compiled with Unicode prop-
2312: erty support.
2313:
2314: PCRE_UNGREEDY
2315:
2316: This option inverts the "greediness" of the quantifiers so that they
2317: are not greedy by default, but become greedy if followed by "?". It is
2318: not compatible with Perl. It can also be set by a (?U) option setting
2319: within the pattern.
2320:
2321: PCRE_UTF8
2322:
2323: This option causes PCRE to regard both the pattern and the subject as
1.1.1.2 misho 2324: strings of UTF-8 characters instead of single-byte strings. However, it
2325: is available only when PCRE is built to include UTF support. If not,
2326: the use of this option provokes an error. Details of how this option
2327: changes the behaviour of PCRE are given in the pcreunicode page.
1.1 misho 2328:
2329: PCRE_NO_UTF8_CHECK
2330:
2331: When PCRE_UTF8 is set, the validity of the pattern as a UTF-8 string is
1.1.1.2 misho 2332: automatically checked. There is a discussion about the validity of
2333: UTF-8 strings in the pcreunicode page. If an invalid UTF-8 sequence is
2334: found, pcre_compile() returns an error. If you already know that your
2335: pattern is valid, and you want to skip this check for performance rea-
2336: sons, you can set the PCRE_NO_UTF8_CHECK option. When it is set, the
2337: effect of passing an invalid UTF-8 string as a pattern is undefined. It
2338: may cause your program to crash. Note that this option can also be
2339: passed to pcre_exec() and pcre_dfa_exec(), to suppress the validity
1.1.1.4 ! misho 2340: checking of subject strings only. If the same string is being matched
! 2341: many times, the option can be safely set for the second and subsequent
! 2342: matchings to improve performance.
1.1 misho 2343:
2344:
2345: COMPILATION ERROR CODES
2346:
1.1.1.2 misho 2347: The following table lists the error codes than may be returned by
2348: pcre_compile2(), along with the error messages that may be returned by
2349: both compiling functions. Note that error messages are always 8-bit
1.1.1.4 ! misho 2350: ASCII strings, even in 16-bit or 32-bit mode. As PCRE has developed,
! 2351: some error codes have fallen out of use. To avoid confusion, they have
! 2352: not been re-used.
1.1 misho 2353:
2354: 0 no error
2355: 1 \ at end of pattern
2356: 2 \c at end of pattern
2357: 3 unrecognized character follows \
2358: 4 numbers out of order in {} quantifier
2359: 5 number too big in {} quantifier
2360: 6 missing terminating ] for character class
2361: 7 invalid escape sequence in character class
2362: 8 range out of order in character class
2363: 9 nothing to repeat
2364: 10 [this code is not in use]
2365: 11 internal error: unexpected repeat
2366: 12 unrecognized character after (? or (?-
2367: 13 POSIX named classes are supported only within a class
2368: 14 missing )
2369: 15 reference to non-existent subpattern
2370: 16 erroffset passed as NULL
2371: 17 unknown option bit(s) set
2372: 18 missing ) after comment
2373: 19 [this code is not in use]
2374: 20 regular expression is too large
2375: 21 failed to get memory
2376: 22 unmatched parentheses
2377: 23 internal error: code overflow
2378: 24 unrecognized character after (?<
2379: 25 lookbehind assertion is not fixed length
2380: 26 malformed number or name after (?(
2381: 27 conditional group contains more than two branches
2382: 28 assertion expected after (?(
2383: 29 (?R or (?[+-]digits must be followed by )
2384: 30 unknown POSIX class name
2385: 31 POSIX collating elements are not supported
1.1.1.2 misho 2386: 32 this version of PCRE is compiled without UTF support
1.1 misho 2387: 33 [this code is not in use]
2388: 34 character value in \x{...} sequence is too large
2389: 35 invalid condition (?(0)
2390: 36 \C not allowed in lookbehind assertion
2391: 37 PCRE does not support \L, \l, \N{name}, \U, or \u
2392: 38 number after (?C is > 255
2393: 39 closing ) for (?C expected
2394: 40 recursive call could loop indefinitely
2395: 41 unrecognized character after (?P
2396: 42 syntax error in subpattern name (missing terminator)
2397: 43 two named subpatterns have the same name
1.1.1.2 misho 2398: 44 invalid UTF-8 string (specifically UTF-8)
1.1 misho 2399: 45 support for \P, \p, and \X has not been compiled
2400: 46 malformed \P or \p sequence
2401: 47 unknown property name after \P or \p
2402: 48 subpattern name is too long (maximum 32 characters)
2403: 49 too many named subpatterns (maximum 10000)
2404: 50 [this code is not in use]
1.1.1.2 misho 2405: 51 octal value is greater than \377 in 8-bit non-UTF-8 mode
1.1 misho 2406: 52 internal error: overran compiling workspace
2407: 53 internal error: previously-checked referenced subpattern
2408: not found
2409: 54 DEFINE group contains more than one branch
2410: 55 repeating a DEFINE group is not allowed
2411: 56 inconsistent NEWLINE options
2412: 57 \g is not followed by a braced, angle-bracketed, or quoted
2413: name/number or by a plain number
2414: 58 a numbered reference must not be zero
2415: 59 an argument is not allowed for (*ACCEPT), (*FAIL), or (*COMMIT)
1.1.1.4 ! misho 2416: 60 (*VERB) not recognized or malformed
1.1 misho 2417: 61 number is too big
2418: 62 subpattern name expected
2419: 63 digit expected after (?+
2420: 64 ] is an invalid data character in JavaScript compatibility mode
2421: 65 different names for subpatterns of the same number are
2422: not allowed
2423: 66 (*MARK) must have an argument
1.1.1.2 misho 2424: 67 this version of PCRE is not compiled with Unicode property
2425: support
1.1 misho 2426: 68 \c must be followed by an ASCII character
2427: 69 \k is not followed by a braced, angle-bracketed, or quoted name
1.1.1.2 misho 2428: 70 internal error: unknown opcode in find_fixedlength()
2429: 71 \N is not supported in a class
2430: 72 too many forward references
2431: 73 disallowed Unicode code point (>= 0xd800 && <= 0xdfff)
2432: 74 invalid UTF-16 string (specifically UTF-16)
1.1.1.3 misho 2433: 75 name is too long in (*MARK), (*PRUNE), (*SKIP), or (*THEN)
2434: 76 character value in \u.... sequence is too large
1.1.1.4 ! misho 2435: 77 invalid UTF-32 string (specifically UTF-32)
1.1 misho 2436:
1.1.1.2 misho 2437: The numbers 32 and 10000 in errors 48 and 49 are defaults; different
1.1 misho 2438: values may be used if the limits were changed when PCRE was built.
2439:
2440:
2441: STUDYING A PATTERN
2442:
2443: pcre_extra *pcre_study(const pcre *code, int options
2444: const char **errptr);
2445:
1.1.1.2 misho 2446: If a compiled pattern is going to be used several times, it is worth
1.1 misho 2447: spending more time analyzing it in order to speed up the time taken for
1.1.1.2 misho 2448: matching. The function pcre_study() takes a pointer to a compiled pat-
1.1 misho 2449: tern as its first argument. If studying the pattern produces additional
1.1.1.2 misho 2450: information that will help speed up matching, pcre_study() returns a
2451: pointer to a pcre_extra block, in which the study_data field points to
1.1 misho 2452: the results of the study.
2453:
2454: The returned value from pcre_study() can be passed directly to
1.1.1.2 misho 2455: pcre_exec() or pcre_dfa_exec(). However, a pcre_extra block also con-
2456: tains other fields that can be set by the caller before the block is
1.1 misho 2457: passed; these are described below in the section on matching a pattern.
2458:
1.1.1.2 misho 2459: If studying the pattern does not produce any useful information,
1.1.1.4 ! misho 2460: pcre_study() returns NULL by default. In that circumstance, if the
! 2461: calling program wants to pass any of the other fields to pcre_exec() or
! 2462: pcre_dfa_exec(), it must set up its own pcre_extra block. However, if
! 2463: pcre_study() is called with the PCRE_STUDY_EXTRA_NEEDED option, it
! 2464: returns a pcre_extra block even if studying did not find any additional
! 2465: information. It may still return NULL, however, if an error occurs in
! 2466: pcre_study().
1.1 misho 2467:
1.1.1.3 misho 2468: The second argument of pcre_study() contains option bits. There are
1.1.1.4 ! misho 2469: three further options in addition to PCRE_STUDY_EXTRA_NEEDED:
1.1.1.3 misho 2470:
2471: PCRE_STUDY_JIT_COMPILE
2472: PCRE_STUDY_JIT_PARTIAL_HARD_COMPILE
2473: PCRE_STUDY_JIT_PARTIAL_SOFT_COMPILE
2474:
2475: If any of these are set, and the just-in-time compiler is available,
2476: the pattern is further compiled into machine code that executes much
2477: faster than the pcre_exec() interpretive matching function. If the
2478: just-in-time compiler is not available, these options are ignored. All
1.1.1.4 ! misho 2479: undefined bits in the options argument must be zero.
1.1 misho 2480:
1.1.1.2 misho 2481: JIT compilation is a heavyweight optimization. It can take some time
2482: for patterns to be analyzed, and for one-off matches and simple pat-
2483: terns the benefit of faster execution might be offset by a much slower
1.1 misho 2484: study time. Not all patterns can be optimized by the JIT compiler. For
1.1.1.2 misho 2485: those that cannot be handled, matching automatically falls back to the
2486: pcre_exec() interpreter. For more details, see the pcrejit documenta-
1.1 misho 2487: tion.
2488:
1.1.1.2 misho 2489: The third argument for pcre_study() is a pointer for an error message.
2490: If studying succeeds (even if no data is returned), the variable it
2491: points to is set to NULL. Otherwise it is set to point to a textual
1.1 misho 2492: error message. This is a static string that is part of the library. You
1.1.1.2 misho 2493: must not try to free it. You should test the error pointer for NULL
1.1 misho 2494: after calling pcre_study(), to be sure that it has run successfully.
2495:
1.1.1.2 misho 2496: When you are finished with a pattern, you can free the memory used for
1.1 misho 2497: the study data by calling pcre_free_study(). This function was added to
1.1.1.2 misho 2498: the API for release 8.20. For earlier versions, the memory could be
2499: freed with pcre_free(), just like the pattern itself. This will still
1.1.1.3 misho 2500: work in cases where JIT optimization is not used, but it is advisable
2501: to change to the new function when convenient.
1.1 misho 2502:
1.1.1.2 misho 2503: This is a typical way in which pcre_study() is used (except that in a
1.1 misho 2504: real application there should be tests for errors):
2505:
2506: int rc;
2507: pcre *re;
2508: pcre_extra *sd;
2509: re = pcre_compile("pattern", 0, &error, &erroroffset, NULL);
2510: sd = pcre_study(
2511: re, /* result of pcre_compile() */
2512: 0, /* no options */
2513: &error); /* set to NULL or points to a message */
2514: rc = pcre_exec( /* see below for details of pcre_exec() options */
2515: re, sd, "subject", 7, 0, 0, ovector, 30);
2516: ...
2517: pcre_free_study(sd);
2518: pcre_free(re);
2519:
2520: Studying a pattern does two things: first, a lower bound for the length
2521: of subject string that is needed to match the pattern is computed. This
2522: does not mean that there are any strings of that length that match, but
1.1.1.4 ! misho 2523: it does guarantee that no shorter strings match. The value is used to
! 2524: avoid wasting time by trying to match strings that are shorter than the
! 2525: lower bound. You can find out the value in a calling program via the
! 2526: pcre_fullinfo() function.
1.1 misho 2527:
2528: Studying a pattern is also useful for non-anchored patterns that do not
1.1.1.2 misho 2529: have a single fixed starting character. A bitmap of possible starting
2530: bytes is created. This speeds up finding a position in the subject at
2531: which to start matching. (In 16-bit mode, the bitmap is used for 16-bit
1.1.1.4 ! misho 2532: values less than 256. In 32-bit mode, the bitmap is used for 32-bit
1.1.1.2 misho 2533: values less than 256.)
1.1 misho 2534:
1.1.1.4 ! misho 2535: These two optimizations apply to both pcre_exec() and pcre_dfa_exec(),
! 2536: and the information is also used by the JIT compiler. The optimiza-
! 2537: tions can be disabled by setting the PCRE_NO_START_OPTIMIZE option.
! 2538: You might want to do this if your pattern contains callouts or (*MARK)
! 2539: and you want to make use of these facilities in cases where matching
! 2540: fails.
! 2541:
! 2542: PCRE_NO_START_OPTIMIZE can be specified at either compile time or exe-
! 2543: cution time. However, if PCRE_NO_START_OPTIMIZE is passed to
! 2544: pcre_exec(), (that is, after any JIT compilation has happened) JIT exe-
! 2545: cution is disabled. For JIT execution to work with PCRE_NO_START_OPTI-
! 2546: MIZE, the option must be set at compile time.
! 2547:
! 2548: There is a longer discussion of PCRE_NO_START_OPTIMIZE below.
1.1 misho 2549:
2550:
2551: LOCALE SUPPORT
2552:
1.1.1.4 ! misho 2553: PCRE handles caseless matching, and determines whether characters are
! 2554: letters, digits, or whatever, by reference to a set of tables, indexed
! 2555: by character value. When running in UTF-8 mode, this applies only to
! 2556: characters with codes less than 128. By default, higher-valued codes
1.1 misho 2557: never match escapes such as \w or \d, but they can be tested with \p if
1.1.1.4 ! misho 2558: PCRE is built with Unicode character property support. Alternatively,
! 2559: the PCRE_UCP option can be set at compile time; this causes \w and
1.1 misho 2560: friends to use Unicode property support instead of built-in tables. The
2561: use of locales with Unicode is discouraged. If you are handling charac-
1.1.1.4 ! misho 2562: ters with codes greater than 128, you should either use UTF-8 and Uni-
1.1 misho 2563: code, or use locales, but not try to mix the two.
2564:
1.1.1.4 ! misho 2565: PCRE contains an internal set of tables that are used when the final
! 2566: argument of pcre_compile() is NULL. These are sufficient for many
1.1 misho 2567: applications. Normally, the internal tables recognize only ASCII char-
2568: acters. However, when PCRE is built, it is possible to cause the inter-
2569: nal tables to be rebuilt in the default "C" locale of the local system,
2570: which may cause them to be different.
2571:
1.1.1.4 ! misho 2572: The internal tables can always be overridden by tables supplied by the
1.1 misho 2573: application that calls PCRE. These may be created in a different locale
1.1.1.4 ! misho 2574: from the default. As more and more applications change to using Uni-
1.1 misho 2575: code, the need for this locale support is expected to die away.
2576:
1.1.1.4 ! misho 2577: External tables are built by calling the pcre_maketables() function,
! 2578: which has no arguments, in the relevant locale. The result can then be
! 2579: passed to pcre_compile() or pcre_exec() as often as necessary. For
! 2580: example, to build and use tables that are appropriate for the French
! 2581: locale (where accented characters with values greater than 128 are
1.1 misho 2582: treated as letters), the following code could be used:
2583:
2584: setlocale(LC_CTYPE, "fr_FR");
2585: tables = pcre_maketables();
2586: re = pcre_compile(..., tables);
2587:
1.1.1.4 ! misho 2588: The locale name "fr_FR" is used on Linux and other Unix-like systems;
1.1 misho 2589: if you are using Windows, the name for the French locale is "french".
2590:
1.1.1.4 ! misho 2591: When pcre_maketables() runs, the tables are built in memory that is
! 2592: obtained via pcre_malloc. It is the caller's responsibility to ensure
! 2593: that the memory containing the tables remains available for as long as
1.1 misho 2594: it is needed.
2595:
2596: The pointer that is passed to pcre_compile() is saved with the compiled
1.1.1.4 ! misho 2597: pattern, and the same tables are used via this pointer by pcre_study()
1.1 misho 2598: and normally also by pcre_exec(). Thus, by default, for any single pat-
2599: tern, compilation, studying and matching all happen in the same locale,
2600: but different patterns can be compiled in different locales.
2601:
1.1.1.4 ! misho 2602: It is possible to pass a table pointer or NULL (indicating the use of
! 2603: the internal tables) to pcre_exec(). Although not intended for this
! 2604: purpose, this facility could be used to match a pattern in a different
1.1 misho 2605: locale from the one in which it was compiled. Passing table pointers at
2606: run time is discussed below in the section on matching a pattern.
2607:
2608:
2609: INFORMATION ABOUT A PATTERN
2610:
2611: int pcre_fullinfo(const pcre *code, const pcre_extra *extra,
2612: int what, void *where);
2613:
1.1.1.4 ! misho 2614: The pcre_fullinfo() function returns information about a compiled pat-
! 2615: tern. It replaces the pcre_info() function, which was removed from the
1.1.1.2 misho 2616: library at version 8.30, after more than 10 years of obsolescence.
1.1 misho 2617:
1.1.1.4 ! misho 2618: The first argument for pcre_fullinfo() is a pointer to the compiled
! 2619: pattern. The second argument is the result of pcre_study(), or NULL if
! 2620: the pattern was not studied. The third argument specifies which piece
! 2621: of information is required, and the fourth argument is a pointer to a
! 2622: variable to receive the data. The yield of the function is zero for
1.1 misho 2623: success, or one of the following negative numbers:
2624:
1.1.1.2 misho 2625: PCRE_ERROR_NULL the argument code was NULL
2626: the argument where was NULL
2627: PCRE_ERROR_BADMAGIC the "magic number" was not found
2628: PCRE_ERROR_BADENDIANNESS the pattern was compiled with different
2629: endianness
2630: PCRE_ERROR_BADOPTION the value of what was invalid
1.1.1.4 ! misho 2631: PCRE_ERROR_UNSET the requested field is not set
1.1 misho 2632:
1.1.1.4 ! misho 2633: The "magic number" is placed at the start of each compiled pattern as
! 2634: an simple check against passing an arbitrary memory pointer. The endi-
1.1.1.2 misho 2635: anness error can occur if a compiled pattern is saved and reloaded on a
1.1.1.4 ! misho 2636: different host. Here is a typical call of pcre_fullinfo(), to obtain
1.1.1.2 misho 2637: the length of the compiled pattern:
1.1 misho 2638:
2639: int rc;
2640: size_t length;
2641: rc = pcre_fullinfo(
2642: re, /* result of pcre_compile() */
2643: sd, /* result of pcre_study(), or NULL */
2644: PCRE_INFO_SIZE, /* what is required */
2645: &length); /* where to put the data */
2646:
1.1.1.4 ! misho 2647: The possible values for the third argument are defined in pcre.h, and
1.1 misho 2648: are as follows:
2649:
2650: PCRE_INFO_BACKREFMAX
2651:
1.1.1.4 ! misho 2652: Return the number of the highest back reference in the pattern. The
! 2653: fourth argument should point to an int variable. Zero is returned if
1.1 misho 2654: there are no back references.
2655:
2656: PCRE_INFO_CAPTURECOUNT
2657:
1.1.1.4 ! misho 2658: Return the number of capturing subpatterns in the pattern. The fourth
1.1 misho 2659: argument should point to an int variable.
2660:
2661: PCRE_INFO_DEFAULT_TABLES
2662:
1.1.1.4 ! misho 2663: Return a pointer to the internal default character tables within PCRE.
! 2664: The fourth argument should point to an unsigned char * variable. This
1.1 misho 2665: information call is provided for internal use by the pcre_study() func-
1.1.1.4 ! misho 2666: tion. External callers can cause PCRE to use its internal tables by
1.1 misho 2667: passing a NULL table pointer.
2668:
2669: PCRE_INFO_FIRSTBYTE
2670:
1.1.1.2 misho 2671: Return information about the first data unit of any matched string, for
1.1.1.4 ! misho 2672: a non-anchored pattern. (The name of this option refers to the 8-bit
! 2673: library, where data units are bytes.) The fourth argument should point
1.1.1.2 misho 2674: to an int variable.
2675:
1.1.1.4 ! misho 2676: If there is a fixed first value, for example, the letter "c" from a
! 2677: pattern such as (cat|cow|coyote), its value is returned. In the 8-bit
! 2678: library, the value is always less than 256. In the 16-bit library the
! 2679: value can be up to 0xffff. In the 32-bit library the value can be up to
! 2680: 0x10ffff.
1.1 misho 2681:
1.1.1.2 misho 2682: If there is no fixed first value, and if either
1.1 misho 2683:
1.1.1.3 misho 2684: (a) the pattern was compiled with the PCRE_MULTILINE option, and every
1.1 misho 2685: branch starts with "^", or
2686:
2687: (b) every branch of the pattern starts with ".*" and PCRE_DOTALL is not
2688: set (if it were set, the pattern would be anchored),
2689:
1.1.1.3 misho 2690: -1 is returned, indicating that the pattern matches only at the start
2691: of a subject string or after any newline within the string. Otherwise
1.1 misho 2692: -2 is returned. For anchored patterns, -2 is returned.
2693:
1.1.1.4 ! misho 2694: Since for the 32-bit library using the non-UTF-32 mode, this function
! 2695: is unable to return the full 32-bit range of the character, this value
! 2696: is deprecated; instead the PCRE_INFO_FIRSTCHARACTERFLAGS and
! 2697: PCRE_INFO_FIRSTCHARACTER values should be used.
! 2698:
1.1 misho 2699: PCRE_INFO_FIRSTTABLE
2700:
1.1.1.4 ! misho 2701: If the pattern was studied, and this resulted in the construction of a
! 2702: 256-bit table indicating a fixed set of values for the first data unit
! 2703: in any matching string, a pointer to the table is returned. Otherwise
! 2704: NULL is returned. The fourth argument should point to an unsigned char
1.1.1.2 misho 2705: * variable.
1.1 misho 2706:
2707: PCRE_INFO_HASCRORLF
2708:
1.1.1.4 ! misho 2709: Return 1 if the pattern contains any explicit matches for CR or LF
! 2710: characters, otherwise 0. The fourth argument should point to an int
! 2711: variable. An explicit match is either a literal CR or LF character, or
1.1 misho 2712: \r or \n.
2713:
2714: PCRE_INFO_JCHANGED
2715:
1.1.1.4 ! misho 2716: Return 1 if the (?J) or (?-J) option setting is used in the pattern,
! 2717: otherwise 0. The fourth argument should point to an int variable. (?J)
1.1 misho 2718: and (?-J) set and unset the local PCRE_DUPNAMES option, respectively.
2719:
2720: PCRE_INFO_JIT
2721:
1.1.1.4 ! misho 2722: Return 1 if the pattern was studied with one of the JIT options, and
1.1.1.3 misho 2723: just-in-time compiling was successful. The fourth argument should point
1.1.1.4 ! misho 2724: to an int variable. A return value of 0 means that JIT support is not
! 2725: available in this version of PCRE, or that the pattern was not studied
! 2726: with a JIT option, or that the JIT compiler could not handle this par-
! 2727: ticular pattern. See the pcrejit documentation for details of what can
1.1.1.3 misho 2728: and cannot be handled.
1.1 misho 2729:
2730: PCRE_INFO_JITSIZE
2731:
1.1.1.4 ! misho 2732: If the pattern was successfully studied with a JIT option, return the
! 2733: size of the JIT compiled code, otherwise return zero. The fourth argu-
1.1.1.3 misho 2734: ment should point to a size_t variable.
1.1 misho 2735:
2736: PCRE_INFO_LASTLITERAL
2737:
1.1.1.4 ! misho 2738: Return the value of the rightmost literal data unit that must exist in
! 2739: any matched string, other than at its start, if such a value has been
1.1 misho 2740: recorded. The fourth argument should point to an int variable. If there
1.1.1.2 misho 2741: is no such value, -1 is returned. For anchored patterns, a last literal
1.1.1.4 ! misho 2742: value is recorded only if it follows something of variable length. For
1.1 misho 2743: example, for the pattern /^a\d+z\d+/ the returned value is "z", but for
2744: /^a\dz\d/ the returned value is -1.
2745:
1.1.1.4 ! misho 2746: Since for the 32-bit library using the non-UTF-32 mode, this function
! 2747: is unable to return the full 32-bit range of the character, this value
! 2748: is deprecated; instead the PCRE_INFO_REQUIREDCHARFLAGS and
! 2749: PCRE_INFO_REQUIREDCHAR values should be used.
! 2750:
! 2751: PCRE_INFO_MATCHLIMIT
! 2752:
! 2753: If the pattern set a match limit by including an item of the form
! 2754: (*LIMIT_MATCH=nnnn) at the start, the value is returned. The fourth
! 2755: argument should point to an unsigned 32-bit integer. If no such value
! 2756: has been set, the call to pcre_fullinfo() returns the error
! 2757: PCRE_ERROR_UNSET.
! 2758:
1.1.1.3 misho 2759: PCRE_INFO_MAXLOOKBEHIND
2760:
1.1.1.4 ! misho 2761: Return the number of characters (NB not data units) in the longest
! 2762: lookbehind assertion in the pattern. This information is useful when
! 2763: doing multi-segment matching using the partial matching facilities.
! 2764: Note that the simple assertions \b and \B require a one-character look-
! 2765: behind. \A also registers a one-character lookbehind, though it does
! 2766: not actually inspect the previous character. This is to ensure that at
! 2767: least one character from the old segment is retained when a new segment
! 2768: is processed. Otherwise, if there are no lookbehinds in the pattern, \A
! 2769: might match incorrectly at the start of a new segment.
1.1.1.3 misho 2770:
1.1 misho 2771: PCRE_INFO_MINLENGTH
2772:
1.1.1.4 ! misho 2773: If the pattern was studied and a minimum length for matching subject
! 2774: strings was computed, its value is returned. Otherwise the returned
! 2775: value is -1. The value is a number of characters, which in UTF mode may
! 2776: be different from the number of data units. The fourth argument should
! 2777: point to an int variable. A non-negative value is a lower bound to the
! 2778: length of any matching string. There may not be any strings of that
! 2779: length that do actually match, but every string that does match is at
1.1.1.2 misho 2780: least that long.
1.1 misho 2781:
2782: PCRE_INFO_NAMECOUNT
2783: PCRE_INFO_NAMEENTRYSIZE
2784: PCRE_INFO_NAMETABLE
2785:
1.1.1.4 ! misho 2786: PCRE supports the use of named as well as numbered capturing parenthe-
! 2787: ses. The names are just an additional way of identifying the parenthe-
1.1 misho 2788: ses, which still acquire numbers. Several convenience functions such as
1.1.1.4 ! misho 2789: pcre_get_named_substring() are provided for extracting captured sub-
! 2790: strings by name. It is also possible to extract the data directly, by
! 2791: first converting the name to a number in order to access the correct
1.1 misho 2792: pointers in the output vector (described with pcre_exec() below). To do
1.1.1.4 ! misho 2793: the conversion, you need to use the name-to-number map, which is
1.1 misho 2794: described by these three values.
2795:
2796: The map consists of a number of fixed-size entries. PCRE_INFO_NAMECOUNT
2797: gives the number of entries, and PCRE_INFO_NAMEENTRYSIZE gives the size
1.1.1.4 ! misho 2798: of each entry; both of these return an int value. The entry size
! 2799: depends on the length of the longest name. PCRE_INFO_NAMETABLE returns
1.1.1.2 misho 2800: a pointer to the first entry of the table. This is a pointer to char in
2801: the 8-bit library, where the first two bytes of each entry are the num-
1.1.1.4 ! misho 2802: ber of the capturing parenthesis, most significant byte first. In the
! 2803: 16-bit library, the pointer points to 16-bit data units, the first of
! 2804: which contains the parenthesis number. In the 32-bit library, the
! 2805: pointer points to 32-bit data units, the first of which contains the
! 2806: parenthesis number. The rest of the entry is the corresponding name,
! 2807: zero terminated.
1.1 misho 2808:
1.1.1.4 ! misho 2809: The names are in alphabetical order. Duplicate names may appear if (?|
1.1 misho 2810: is used to create multiple groups with the same number, as described in
1.1.1.4 ! misho 2811: the section on duplicate subpattern numbers in the pcrepattern page.
! 2812: Duplicate names for subpatterns with different numbers are permitted
! 2813: only if PCRE_DUPNAMES is set. In all cases of duplicate names, they
! 2814: appear in the table in the order in which they were found in the pat-
! 2815: tern. In the absence of (?| this is the order of increasing number;
1.1 misho 2816: when (?| is used this is not necessarily the case because later subpat-
2817: terns may have lower numbers.
2818:
1.1.1.4 ! misho 2819: As a simple example of the name/number table, consider the following
1.1.1.2 misho 2820: pattern after compilation by the 8-bit library (assume PCRE_EXTENDED is
2821: set, so white space - including newlines - is ignored):
1.1 misho 2822:
2823: (?<date> (?<year>(\d\d)?\d\d) -
2824: (?<month>\d\d) - (?<day>\d\d) )
2825:
1.1.1.4 ! misho 2826: There are four named subpatterns, so the table has four entries, and
! 2827: each entry in the table is eight bytes long. The table is as follows,
1.1 misho 2828: with non-printing bytes shows in hexadecimal, and undefined bytes shown
2829: as ??:
2830:
2831: 00 01 d a t e 00 ??
2832: 00 05 d a y 00 ?? ??
2833: 00 04 m o n t h 00
2834: 00 02 y e a r 00 ??
2835:
1.1.1.4 ! misho 2836: When writing code to extract data from named subpatterns using the
! 2837: name-to-number map, remember that the length of the entries is likely
1.1 misho 2838: to be different for each compiled pattern.
2839:
2840: PCRE_INFO_OKPARTIAL
2841:
1.1.1.4 ! misho 2842: Return 1 if the pattern can be used for partial matching with
! 2843: pcre_exec(), otherwise 0. The fourth argument should point to an int
! 2844: variable. From release 8.00, this always returns 1, because the
! 2845: restrictions that previously applied to partial matching have been
! 2846: lifted. The pcrepartial documentation gives details of partial match-
1.1 misho 2847: ing.
2848:
2849: PCRE_INFO_OPTIONS
2850:
1.1.1.4 ! misho 2851: Return a copy of the options with which the pattern was compiled. The
! 2852: fourth argument should point to an unsigned long int variable. These
1.1 misho 2853: option bits are those specified in the call to pcre_compile(), modified
2854: by any top-level option settings at the start of the pattern itself. In
1.1.1.4 ! misho 2855: other words, they are the options that will be in force when matching
! 2856: starts. For example, if the pattern /(?im)abc(?-i)d/ is compiled with
! 2857: the PCRE_EXTENDED option, the result is PCRE_CASELESS, PCRE_MULTILINE,
1.1 misho 2858: and PCRE_EXTENDED.
2859:
1.1.1.4 ! misho 2860: A pattern is automatically anchored by PCRE if all of its top-level
1.1 misho 2861: alternatives begin with one of the following:
2862:
2863: ^ unless PCRE_MULTILINE is set
2864: \A always
2865: \G always
2866: .* if PCRE_DOTALL is set and there are no back
2867: references to the subpattern in which .* appears
2868:
2869: For such patterns, the PCRE_ANCHORED bit is set in the options returned
2870: by pcre_fullinfo().
2871:
1.1.1.4 ! misho 2872: PCRE_INFO_RECURSIONLIMIT
! 2873:
! 2874: If the pattern set a recursion limit by including an item of the form
! 2875: (*LIMIT_RECURSION=nnnn) at the start, the value is returned. The fourth
! 2876: argument should point to an unsigned 32-bit integer. If no such value
! 2877: has been set, the call to pcre_fullinfo() returns the error
! 2878: PCRE_ERROR_UNSET.
! 2879:
1.1 misho 2880: PCRE_INFO_SIZE
2881:
1.1.1.4 ! misho 2882: Return the size of the compiled pattern in bytes (for all three
! 2883: libraries). The fourth argument should point to a size_t variable. This
! 2884: value does not include the size of the pcre structure that is returned
! 2885: by pcre_compile(). The value that is passed as the argument to
! 2886: pcre_malloc() when pcre_compile() is getting memory in which to place
! 2887: the compiled data is the value returned by this option plus the size of
! 2888: the pcre structure. Studying a compiled pattern, with or without JIT,
! 2889: does not alter the value returned by this option.
1.1 misho 2890:
2891: PCRE_INFO_STUDYSIZE
2892:
1.1.1.4 ! misho 2893: Return the size in bytes (for all three libraries) of the data block
! 2894: pointed to by the study_data field in a pcre_extra block. If pcre_extra
! 2895: is NULL, or there is no study data, zero is returned. The fourth argu-
! 2896: ment should point to a size_t variable. The study_data field is set by
! 2897: pcre_study() to record information that will speed up matching (see the
! 2898: section entitled "Studying a pattern" above). The format of the
! 2899: study_data block is private, but its length is made available via this
! 2900: option so that it can be saved and restored (see the pcreprecompile
! 2901: documentation for details).
! 2902:
! 2903: PCRE_INFO_FIRSTCHARACTERFLAGS
! 2904:
! 2905: Return information about the first data unit of any matched string, for
! 2906: a non-anchored pattern. The fourth argument should point to an int
! 2907: variable.
! 2908:
! 2909: If there is a fixed first value, for example, the letter "c" from a
! 2910: pattern such as (cat|cow|coyote), 1 is returned, and the character
! 2911: value can be retrieved using PCRE_INFO_FIRSTCHARACTER.
! 2912:
! 2913: If there is no fixed first value, and if either
! 2914:
! 2915: (a) the pattern was compiled with the PCRE_MULTILINE option, and every
! 2916: branch starts with "^", or
! 2917:
! 2918: (b) every branch of the pattern starts with ".*" and PCRE_DOTALL is not
! 2919: set (if it were set, the pattern would be anchored),
! 2920:
! 2921: 2 is returned, indicating that the pattern matches only at the start of
! 2922: a subject string or after any newline within the string. Otherwise 0 is
! 2923: returned. For anchored patterns, 0 is returned.
! 2924:
! 2925: PCRE_INFO_FIRSTCHARACTER
! 2926:
! 2927: Return the fixed first character value, if PCRE_INFO_FIRSTCHARACTER-
! 2928: FLAGS returned 1; otherwise returns 0. The fourth argument should point
! 2929: to an uint_t variable.
! 2930:
! 2931: In the 8-bit library, the value is always less than 256. In the 16-bit
! 2932: library the value can be up to 0xffff. In the 32-bit library in UTF-32
! 2933: mode the value can be up to 0x10ffff, and up to 0xffffffff when not
! 2934: using UTF-32 mode.
! 2935:
! 2936: If there is no fixed first value, and if either
! 2937:
! 2938: (a) the pattern was compiled with the PCRE_MULTILINE option, and every
! 2939: branch starts with "^", or
! 2940:
! 2941: (b) every branch of the pattern starts with ".*" and PCRE_DOTALL is not
! 2942: set (if it were set, the pattern would be anchored),
! 2943:
! 2944: -1 is returned, indicating that the pattern matches only at the start
! 2945: of a subject string or after any newline within the string. Otherwise
! 2946: -2 is returned. For anchored patterns, -2 is returned.
! 2947:
! 2948: PCRE_INFO_REQUIREDCHARFLAGS
! 2949:
! 2950: Returns 1 if there is a rightmost literal data unit that must exist in
! 2951: any matched string, other than at its start. The fourth argument should
! 2952: point to an int variable. If there is no such value, 0 is returned. If
! 2953: returning 1, the character value itself can be retrieved using
! 2954: PCRE_INFO_REQUIREDCHAR.
! 2955:
! 2956: For anchored patterns, a last literal value is recorded only if it fol-
! 2957: lows something of variable length. For example, for the pattern
! 2958: /^a\d+z\d+/ the returned value 1 (with "z" returned from
! 2959: PCRE_INFO_REQUIREDCHAR), but for /^a\dz\d/ the returned value is 0.
! 2960:
! 2961: PCRE_INFO_REQUIREDCHAR
! 2962:
! 2963: Return the value of the rightmost literal data unit that must exist in
! 2964: any matched string, other than at its start, if such a value has been
! 2965: recorded. The fourth argument should point to an uint32_t variable. If
! 2966: there is no such value, 0 is returned.
1.1 misho 2967:
2968:
2969: REFERENCE COUNTS
2970:
2971: int pcre_refcount(pcre *code, int adjust);
2972:
1.1.1.2 misho 2973: The pcre_refcount() function is used to maintain a reference count in
1.1 misho 2974: the data block that contains a compiled pattern. It is provided for the
1.1.1.2 misho 2975: benefit of applications that operate in an object-oriented manner,
1.1 misho 2976: where different parts of the application may be using the same compiled
2977: pattern, but you want to free the block when they are all done.
2978:
2979: When a pattern is compiled, the reference count field is initialized to
1.1.1.2 misho 2980: zero. It is changed only by calling this function, whose action is to
2981: add the adjust value (which may be positive or negative) to it. The
1.1 misho 2982: yield of the function is the new value. However, the value of the count
1.1.1.2 misho 2983: is constrained to lie between 0 and 65535, inclusive. If the new value
1.1 misho 2984: is outside these limits, it is forced to the appropriate limit value.
2985:
1.1.1.2 misho 2986: Except when it is zero, the reference count is not correctly preserved
2987: if a pattern is compiled on one host and then transferred to a host
1.1 misho 2988: whose byte-order is different. (This seems a highly unlikely scenario.)
2989:
2990:
2991: MATCHING A PATTERN: THE TRADITIONAL FUNCTION
2992:
2993: int pcre_exec(const pcre *code, const pcre_extra *extra,
2994: const char *subject, int length, int startoffset,
2995: int options, int *ovector, int ovecsize);
2996:
1.1.1.2 misho 2997: The function pcre_exec() is called to match a subject string against a
2998: compiled pattern, which is passed in the code argument. If the pattern
2999: was studied, the result of the study should be passed in the extra
3000: argument. You can call pcre_exec() with the same code and extra argu-
3001: ments as many times as you like, in order to match different subject
1.1 misho 3002: strings with the same pattern.
3003:
1.1.1.2 misho 3004: This function is the main matching facility of the library, and it
3005: operates in a Perl-like manner. For specialist use there is also an
3006: alternative matching function, which is described below in the section
1.1 misho 3007: about the pcre_dfa_exec() function.
3008:
1.1.1.2 misho 3009: In most applications, the pattern will have been compiled (and option-
3010: ally studied) in the same process that calls pcre_exec(). However, it
1.1 misho 3011: is possible to save compiled patterns and study data, and then use them
1.1.1.2 misho 3012: later in different processes, possibly even on different hosts. For a
1.1 misho 3013: discussion about this, see the pcreprecompile documentation.
3014:
3015: Here is an example of a simple call to pcre_exec():
3016:
3017: int rc;
3018: int ovector[30];
3019: rc = pcre_exec(
3020: re, /* result of pcre_compile() */
3021: NULL, /* we didn't study the pattern */
3022: "some string", /* the subject string */
3023: 11, /* the length of the subject string */
3024: 0, /* start at offset 0 in the subject */
3025: 0, /* default options */
3026: ovector, /* vector of integers for substring information */
3027: 30); /* number of elements (NOT size in bytes) */
3028:
3029: Extra data for pcre_exec()
3030:
1.1.1.2 misho 3031: If the extra argument is not NULL, it must point to a pcre_extra data
3032: block. The pcre_study() function returns such a block (when it doesn't
3033: return NULL), but you can also create one for yourself, and pass addi-
3034: tional information in it. The pcre_extra block contains the following
1.1 misho 3035: fields (not necessarily in this order):
3036:
3037: unsigned long int flags;
3038: void *study_data;
3039: void *executable_jit;
3040: unsigned long int match_limit;
3041: unsigned long int match_limit_recursion;
3042: void *callout_data;
3043: const unsigned char *tables;
3044: unsigned char **mark;
3045:
1.1.1.2 misho 3046: In the 16-bit version of this structure, the mark field has type
3047: "PCRE_UCHAR16 **".
3048:
1.1.1.4 ! misho 3049: In the 32-bit version of this structure, the mark field has type
! 3050: "PCRE_UCHAR32 **".
! 3051:
! 3052: The flags field is used to specify which of the other fields are set.
1.1.1.3 misho 3053: The flag bits are:
1.1 misho 3054:
1.1.1.3 misho 3055: PCRE_EXTRA_CALLOUT_DATA
1.1 misho 3056: PCRE_EXTRA_EXECUTABLE_JIT
1.1.1.3 misho 3057: PCRE_EXTRA_MARK
1.1 misho 3058: PCRE_EXTRA_MATCH_LIMIT
3059: PCRE_EXTRA_MATCH_LIMIT_RECURSION
1.1.1.3 misho 3060: PCRE_EXTRA_STUDY_DATA
1.1 misho 3061: PCRE_EXTRA_TABLES
3062:
1.1.1.4 ! misho 3063: Other flag bits should be set to zero. The study_data field and some-
! 3064: times the executable_jit field are set in the pcre_extra block that is
! 3065: returned by pcre_study(), together with the appropriate flag bits. You
! 3066: should not set these yourself, but you may add to the block by setting
1.1.1.3 misho 3067: other fields and their corresponding flag bits.
1.1 misho 3068:
3069: The match_limit field provides a means of preventing PCRE from using up
1.1.1.4 ! misho 3070: a vast amount of resources when running patterns that are not going to
! 3071: match, but which have a very large number of possibilities in their
! 3072: search trees. The classic example is a pattern that uses nested unlim-
1.1 misho 3073: ited repeats.
3074:
1.1.1.4 ! misho 3075: Internally, pcre_exec() uses a function called match(), which it calls
! 3076: repeatedly (sometimes recursively). The limit set by match_limit is
! 3077: imposed on the number of times this function is called during a match,
! 3078: which has the effect of limiting the amount of backtracking that can
1.1 misho 3079: take place. For patterns that are not anchored, the count restarts from
3080: zero for each position in the subject string.
3081:
3082: When pcre_exec() is called with a pattern that was successfully studied
1.1.1.4 ! misho 3083: with a JIT option, the way that the matching is executed is entirely
1.1.1.3 misho 3084: different. However, there is still the possibility of runaway matching
3085: that goes on for a very long time, and so the match_limit value is also
3086: used in this case (but in a different way) to limit how long the match-
3087: ing can continue.
1.1 misho 3088:
1.1.1.4 ! misho 3089: The default value for the limit can be set when PCRE is built; the
! 3090: default default is 10 million, which handles all but the most extreme
! 3091: cases. You can override the default by suppling pcre_exec() with a
! 3092: pcre_extra block in which match_limit is set, and
! 3093: PCRE_EXTRA_MATCH_LIMIT is set in the flags field. If the limit is
1.1 misho 3094: exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3095:
1.1.1.4 ! misho 3096: A value for the match limit may also be supplied by an item at the
! 3097: start of a pattern of the form
! 3098:
! 3099: (*LIMIT_MATCH=d)
! 3100:
! 3101: where d is a decimal number. However, such a setting is ignored unless
! 3102: d is less than the limit set by the caller of pcre_exec() or, if no
! 3103: such limit is set, less than the default.
! 3104:
1.1 misho 3105: The match_limit_recursion field is similar to match_limit, but instead
3106: of limiting the total number of times that match() is called, it limits
3107: the depth of recursion. The recursion depth is a smaller number than
3108: the total number of calls, because not all calls to match() are recur-
3109: sive. This limit is of use only if it is set smaller than match_limit.
3110:
3111: Limiting the recursion depth limits the amount of machine stack that
3112: can be used, or, when PCRE has been compiled to use memory on the heap
3113: instead of the stack, the amount of heap memory that can be used. This
1.1.1.3 misho 3114: limit is not relevant, and is ignored, when matching is done using JIT
3115: compiled code.
1.1 misho 3116:
3117: The default value for match_limit_recursion can be set when PCRE is
3118: built; the default default is the same value as the default for
3119: match_limit. You can override the default by suppling pcre_exec() with
3120: a pcre_extra block in which match_limit_recursion is set, and
3121: PCRE_EXTRA_MATCH_LIMIT_RECURSION is set in the flags field. If the
3122: limit is exceeded, pcre_exec() returns PCRE_ERROR_RECURSIONLIMIT.
3123:
1.1.1.4 ! misho 3124: A value for the recursion limit may also be supplied by an item at the
! 3125: start of a pattern of the form
! 3126:
! 3127: (*LIMIT_RECURSION=d)
! 3128:
! 3129: where d is a decimal number. However, such a setting is ignored unless
! 3130: d is less than the limit set by the caller of pcre_exec() or, if no
! 3131: such limit is set, less than the default.
! 3132:
! 3133: The callout_data field is used in conjunction with the "callout" fea-
1.1 misho 3134: ture, and is described in the pcrecallout documentation.
3135:
1.1.1.4 ! misho 3136: The tables field is used to pass a character tables pointer to
! 3137: pcre_exec(); this overrides the value that is stored with the compiled
! 3138: pattern. A non-NULL value is stored with the compiled pattern only if
! 3139: custom tables were supplied to pcre_compile() via its tableptr argu-
1.1 misho 3140: ment. If NULL is passed to pcre_exec() using this mechanism, it forces
1.1.1.4 ! misho 3141: PCRE's internal tables to be used. This facility is helpful when re-
! 3142: using patterns that have been saved after compiling with an external
! 3143: set of tables, because the external tables might be at a different
! 3144: address when pcre_exec() is called. See the pcreprecompile documenta-
1.1 misho 3145: tion for a discussion of saving compiled patterns for later use.
3146:
1.1.1.4 ! misho 3147: If PCRE_EXTRA_MARK is set in the flags field, the mark field must be
! 3148: set to point to a suitable variable. If the pattern contains any back-
! 3149: tracking control verbs such as (*MARK:NAME), and the execution ends up
! 3150: with a name to pass back, a pointer to the name string (zero termi-
! 3151: nated) is placed in the variable pointed to by the mark field. The
! 3152: names are within the compiled pattern; if you wish to retain such a
! 3153: name you must copy it before freeing the memory of a compiled pattern.
! 3154: If there is no name to pass back, the variable pointed to by the mark
! 3155: field is set to NULL. For details of the backtracking control verbs,
1.1.1.2 misho 3156: see the section entitled "Backtracking control" in the pcrepattern doc-
3157: umentation.
1.1 misho 3158:
3159: Option bits for pcre_exec()
3160:
1.1.1.4 ! misho 3161: The unused bits of the options argument for pcre_exec() must be zero.
! 3162: The only bits that may be set are PCRE_ANCHORED, PCRE_NEWLINE_xxx,
! 3163: PCRE_NOTBOL, PCRE_NOTEOL, PCRE_NOTEMPTY, PCRE_NOTEMPTY_ATSTART,
! 3164: PCRE_NO_START_OPTIMIZE, PCRE_NO_UTF8_CHECK, PCRE_PARTIAL_HARD, and
1.1.1.3 misho 3165: PCRE_PARTIAL_SOFT.
1.1 misho 3166:
1.1.1.4 ! misho 3167: If the pattern was successfully studied with one of the just-in-time
1.1.1.3 misho 3168: (JIT) compile options, the only supported options for JIT execution are
1.1.1.4 ! misho 3169: PCRE_NO_UTF8_CHECK, PCRE_NOTBOL, PCRE_NOTEOL, PCRE_NOTEMPTY,
! 3170: PCRE_NOTEMPTY_ATSTART, PCRE_PARTIAL_HARD, and PCRE_PARTIAL_SOFT. If an
! 3171: unsupported option is used, JIT execution is disabled and the normal
1.1.1.3 misho 3172: interpretive code in pcre_exec() is run.
1.1 misho 3173:
3174: PCRE_ANCHORED
3175:
1.1.1.4 ! misho 3176: The PCRE_ANCHORED option limits pcre_exec() to matching at the first
! 3177: matching position. If a pattern was compiled with PCRE_ANCHORED, or
! 3178: turned out to be anchored by virtue of its contents, it cannot be made
1.1 misho 3179: unachored at matching time.
3180:
3181: PCRE_BSR_ANYCRLF
3182: PCRE_BSR_UNICODE
3183:
3184: These options (which are mutually exclusive) control what the \R escape
1.1.1.4 ! misho 3185: sequence matches. The choice is either to match only CR, LF, or CRLF,
! 3186: or to match any Unicode newline sequence. These options override the
1.1 misho 3187: choice that was made or defaulted when the pattern was compiled.
3188:
3189: PCRE_NEWLINE_CR
3190: PCRE_NEWLINE_LF
3191: PCRE_NEWLINE_CRLF
3192: PCRE_NEWLINE_ANYCRLF
3193: PCRE_NEWLINE_ANY
3194:
1.1.1.4 ! misho 3195: These options override the newline definition that was chosen or
! 3196: defaulted when the pattern was compiled. For details, see the descrip-
! 3197: tion of pcre_compile() above. During matching, the newline choice
! 3198: affects the behaviour of the dot, circumflex, and dollar metacharac-
! 3199: ters. It may also alter the way the match position is advanced after a
1.1 misho 3200: match failure for an unanchored pattern.
3201:
1.1.1.4 ! misho 3202: When PCRE_NEWLINE_CRLF, PCRE_NEWLINE_ANYCRLF, or PCRE_NEWLINE_ANY is
! 3203: set, and a match attempt for an unanchored pattern fails when the cur-
! 3204: rent position is at a CRLF sequence, and the pattern contains no
! 3205: explicit matches for CR or LF characters, the match position is
1.1 misho 3206: advanced by two characters instead of one, in other words, to after the
3207: CRLF.
3208:
3209: The above rule is a compromise that makes the most common cases work as
1.1.1.4 ! misho 3210: expected. For example, if the pattern is .+A (and the PCRE_DOTALL
1.1 misho 3211: option is not set), it does not match the string "\r\nA" because, after
1.1.1.4 ! misho 3212: failing at the start, it skips both the CR and the LF before retrying.
! 3213: However, the pattern [\r\n]A does match that string, because it con-
1.1 misho 3214: tains an explicit CR or LF reference, and so advances only by one char-
3215: acter after the first failure.
3216:
3217: An explicit match for CR of LF is either a literal appearance of one of
1.1.1.4 ! misho 3218: those characters, or one of the \r or \n escape sequences. Implicit
! 3219: matches such as [^X] do not count, nor does \s (which includes CR and
1.1 misho 3220: LF in the characters that it matches).
3221:
1.1.1.4 ! misho 3222: Notwithstanding the above, anomalous effects may still occur when CRLF
1.1 misho 3223: is a valid newline sequence and explicit \r or \n escapes appear in the
3224: pattern.
3225:
3226: PCRE_NOTBOL
3227:
3228: This option specifies that first character of the subject string is not
1.1.1.4 ! misho 3229: the beginning of a line, so the circumflex metacharacter should not
! 3230: match before it. Setting this without PCRE_MULTILINE (at compile time)
! 3231: causes circumflex never to match. This option affects only the behav-
1.1 misho 3232: iour of the circumflex metacharacter. It does not affect \A.
3233:
3234: PCRE_NOTEOL
3235:
3236: This option specifies that the end of the subject string is not the end
1.1.1.4 ! misho 3237: of a line, so the dollar metacharacter should not match it nor (except
! 3238: in multiline mode) a newline immediately before it. Setting this with-
1.1 misho 3239: out PCRE_MULTILINE (at compile time) causes dollar never to match. This
1.1.1.4 ! misho 3240: option affects only the behaviour of the dollar metacharacter. It does
1.1 misho 3241: not affect \Z or \z.
3242:
3243: PCRE_NOTEMPTY
3244:
3245: An empty string is not considered to be a valid match if this option is
1.1.1.4 ! misho 3246: set. If there are alternatives in the pattern, they are tried. If all
! 3247: the alternatives match the empty string, the entire match fails. For
1.1 misho 3248: example, if the pattern
3249:
3250: a?b?
3251:
1.1.1.4 ! misho 3252: is applied to a string not beginning with "a" or "b", it matches an
! 3253: empty string at the start of the subject. With PCRE_NOTEMPTY set, this
1.1 misho 3254: match is not valid, so PCRE searches further into the string for occur-
3255: rences of "a" or "b".
3256:
3257: PCRE_NOTEMPTY_ATSTART
3258:
1.1.1.4 ! misho 3259: This is like PCRE_NOTEMPTY, except that an empty string match that is
! 3260: not at the start of the subject is permitted. If the pattern is
1.1 misho 3261: anchored, such a match can occur only if the pattern contains \K.
3262:
1.1.1.4 ! misho 3263: Perl has no direct equivalent of PCRE_NOTEMPTY or
! 3264: PCRE_NOTEMPTY_ATSTART, but it does make a special case of a pattern
! 3265: match of the empty string within its split() function, and when using
! 3266: the /g modifier. It is possible to emulate Perl's behaviour after
1.1 misho 3267: matching a null string by first trying the match again at the same off-
1.1.1.4 ! misho 3268: set with PCRE_NOTEMPTY_ATSTART and PCRE_ANCHORED, and then if that
1.1 misho 3269: fails, by advancing the starting offset (see below) and trying an ordi-
1.1.1.4 ! misho 3270: nary match again. There is some code that demonstrates how to do this
! 3271: in the pcredemo sample program. In the most general case, you have to
! 3272: check to see if the newline convention recognizes CRLF as a newline,
! 3273: and if so, and the current character is CR followed by LF, advance the
1.1 misho 3274: starting offset by two characters instead of one.
3275:
3276: PCRE_NO_START_OPTIMIZE
3277:
1.1.1.4 ! misho 3278: There are a number of optimizations that pcre_exec() uses at the start
! 3279: of a match, in order to speed up the process. For example, if it is
1.1 misho 3280: known that an unanchored match must start with a specific character, it
1.1.1.4 ! misho 3281: searches the subject for that character, and fails immediately if it
! 3282: cannot find it, without actually running the main matching function.
1.1 misho 3283: This means that a special item such as (*COMMIT) at the start of a pat-
1.1.1.4 ! misho 3284: tern is not considered until after a suitable starting point for the
! 3285: match has been found. Also, when callouts or (*MARK) items are in use,
! 3286: these "start-up" optimizations can cause them to be skipped if the pat-
! 3287: tern is never actually used. The start-up optimizations are in effect a
! 3288: pre-scan of the subject that takes place before the pattern is run.
! 3289:
! 3290: The PCRE_NO_START_OPTIMIZE option disables the start-up optimizations,
! 3291: possibly causing performance to suffer, but ensuring that in cases
! 3292: where the result is "no match", the callouts do occur, and that items
1.1 misho 3293: such as (*COMMIT) and (*MARK) are considered at every possible starting
1.1.1.4 ! misho 3294: position in the subject string. If PCRE_NO_START_OPTIMIZE is set at
! 3295: compile time, it cannot be unset at matching time. The use of
! 3296: PCRE_NO_START_OPTIMIZE at matching time (that is, passing it to
! 3297: pcre_exec()) disables JIT execution; in this situation, matching is
! 3298: always done using interpretively.
1.1 misho 3299:
3300: Setting PCRE_NO_START_OPTIMIZE can change the outcome of a matching
3301: operation. Consider the pattern
3302:
3303: (*COMMIT)ABC
3304:
3305: When this is compiled, PCRE records the fact that a match must start
3306: with the character "A". Suppose the subject string is "DEFABC". The
3307: start-up optimization scans along the subject, finds "A" and runs the
3308: first match attempt from there. The (*COMMIT) item means that the pat-
3309: tern must match the current starting position, which in this case, it
3310: does. However, if the same match is run with PCRE_NO_START_OPTIMIZE
3311: set, the initial scan along the subject string does not happen. The
3312: first match attempt is run starting from "D" and when this fails,
3313: (*COMMIT) prevents any further matches being tried, so the overall
3314: result is "no match". If the pattern is studied, more start-up opti-
3315: mizations may be used. For example, a minimum length for the subject
3316: may be recorded. Consider the pattern
3317:
3318: (*MARK:A)(X|Y)
3319:
3320: The minimum length for a match is one character. If the subject is
3321: "ABC", there will be attempts to match "ABC", "BC", "C", and then
3322: finally an empty string. If the pattern is studied, the final attempt
3323: does not take place, because PCRE knows that the subject is too short,
3324: and so the (*MARK) is never encountered. In this case, studying the
3325: pattern does not affect the overall match result, which is still "no
3326: match", but it does affect the auxiliary information that is returned.
3327:
3328: PCRE_NO_UTF8_CHECK
3329:
3330: When PCRE_UTF8 is set at compile time, the validity of the subject as a
3331: UTF-8 string is automatically checked when pcre_exec() is subsequently
1.1.1.3 misho 3332: called. The entire string is checked before any other processing takes
3333: place. The value of startoffset is also checked to ensure that it
3334: points to the start of a UTF-8 character. There is a discussion about
3335: the validity of UTF-8 strings in the pcreunicode page. If an invalid
3336: sequence of bytes is found, pcre_exec() returns the error
1.1.1.2 misho 3337: PCRE_ERROR_BADUTF8 or, if PCRE_PARTIAL_HARD is set and the problem is a
3338: truncated character at the end of the subject, PCRE_ERROR_SHORTUTF8. In
1.1.1.3 misho 3339: both cases, information about the precise nature of the error may also
3340: be returned (see the descriptions of these errors in the section enti-
3341: tled Error return values from pcre_exec() below). If startoffset con-
1.1.1.2 misho 3342: tains a value that does not point to the start of a UTF-8 character (or
3343: to the end of the subject), PCRE_ERROR_BADUTF8_OFFSET is returned.
3344:
1.1.1.3 misho 3345: If you already know that your subject is valid, and you want to skip
3346: these checks for performance reasons, you can set the
3347: PCRE_NO_UTF8_CHECK option when calling pcre_exec(). You might want to
3348: do this for the second and subsequent calls to pcre_exec() if you are
3349: making repeated calls to find all the matches in a single subject
3350: string. However, you should be sure that the value of startoffset
3351: points to the start of a character (or the end of the subject). When
1.1.1.2 misho 3352: PCRE_NO_UTF8_CHECK is set, the effect of passing an invalid string as a
1.1.1.3 misho 3353: subject or an invalid value of startoffset is undefined. Your program
1.1.1.2 misho 3354: may crash.
1.1 misho 3355:
3356: PCRE_PARTIAL_HARD
3357: PCRE_PARTIAL_SOFT
3358:
1.1.1.3 misho 3359: These options turn on the partial matching feature. For backwards com-
3360: patibility, PCRE_PARTIAL is a synonym for PCRE_PARTIAL_SOFT. A partial
3361: match occurs if the end of the subject string is reached successfully,
3362: but there are not enough subject characters to complete the match. If
1.1 misho 3363: this happens when PCRE_PARTIAL_SOFT (but not PCRE_PARTIAL_HARD) is set,
1.1.1.3 misho 3364: matching continues by testing any remaining alternatives. Only if no
3365: complete match can be found is PCRE_ERROR_PARTIAL returned instead of
3366: PCRE_ERROR_NOMATCH. In other words, PCRE_PARTIAL_SOFT says that the
3367: caller is prepared to handle a partial match, but only if no complete
1.1 misho 3368: match can be found.
3369:
1.1.1.3 misho 3370: If PCRE_PARTIAL_HARD is set, it overrides PCRE_PARTIAL_SOFT. In this
3371: case, if a partial match is found, pcre_exec() immediately returns
3372: PCRE_ERROR_PARTIAL, without considering any other alternatives. In
3373: other words, when PCRE_PARTIAL_HARD is set, a partial match is consid-
1.1 misho 3374: ered to be more important that an alternative complete match.
3375:
1.1.1.3 misho 3376: In both cases, the portion of the string that was inspected when the
1.1 misho 3377: partial match was found is set as the first matching string. There is a
1.1.1.3 misho 3378: more detailed discussion of partial and multi-segment matching, with
1.1 misho 3379: examples, in the pcrepartial documentation.
3380:
3381: The string to be matched by pcre_exec()
3382:
1.1.1.3 misho 3383: The subject string is passed to pcre_exec() as a pointer in subject, a
1.1.1.4 ! misho 3384: length in length, and a starting offset in startoffset. The units for
! 3385: length and startoffset are bytes for the 8-bit library, 16-bit data
! 3386: items for the 16-bit library, and 32-bit data items for the 32-bit
! 3387: library.
! 3388:
! 3389: If startoffset is negative or greater than the length of the subject,
1.1.1.3 misho 3390: pcre_exec() returns PCRE_ERROR_BADOFFSET. When the starting offset is
3391: zero, the search for a match starts at the beginning of the subject,
1.1.1.4 ! misho 3392: and this is by far the most common case. In UTF-8 or UTF-16 mode, the
! 3393: offset must point to the start of a character, or the end of the sub-
! 3394: ject (in UTF-32 mode, one data unit equals one character, so all off-
! 3395: sets are valid). Unlike the pattern string, the subject may contain
! 3396: binary zeroes.
! 3397:
! 3398: A non-zero starting offset is useful when searching for another match
! 3399: in the same subject by calling pcre_exec() again after a previous suc-
! 3400: cess. Setting startoffset differs from just passing over a shortened
! 3401: string and setting PCRE_NOTBOL in the case of a pattern that begins
1.1 misho 3402: with any kind of lookbehind. For example, consider the pattern
3403:
3404: \Biss\B
3405:
1.1.1.4 ! misho 3406: which finds occurrences of "iss" in the middle of words. (\B matches
! 3407: only if the current position in the subject is not a word boundary.)
! 3408: When applied to the string "Mississipi" the first call to pcre_exec()
! 3409: finds the first occurrence. If pcre_exec() is called again with just
! 3410: the remainder of the subject, namely "issipi", it does not match,
1.1 misho 3411: because \B is always false at the start of the subject, which is deemed
1.1.1.4 ! misho 3412: to be a word boundary. However, if pcre_exec() is passed the entire
1.1 misho 3413: string again, but with startoffset set to 4, it finds the second occur-
1.1.1.4 ! misho 3414: rence of "iss" because it is able to look behind the starting point to
1.1 misho 3415: discover that it is preceded by a letter.
3416:
1.1.1.4 ! misho 3417: Finding all the matches in a subject is tricky when the pattern can
1.1 misho 3418: match an empty string. It is possible to emulate Perl's /g behaviour by
1.1.1.4 ! misho 3419: first trying the match again at the same offset, with the
! 3420: PCRE_NOTEMPTY_ATSTART and PCRE_ANCHORED options, and then if that
! 3421: fails, advancing the starting offset and trying an ordinary match
1.1 misho 3422: again. There is some code that demonstrates how to do this in the pcre-
3423: demo sample program. In the most general case, you have to check to see
1.1.1.4 ! misho 3424: if the newline convention recognizes CRLF as a newline, and if so, and
1.1 misho 3425: the current character is CR followed by LF, advance the starting offset
3426: by two characters instead of one.
3427:
1.1.1.4 ! misho 3428: If a non-zero starting offset is passed when the pattern is anchored,
1.1 misho 3429: one attempt to match at the given offset is made. This can only succeed
1.1.1.4 ! misho 3430: if the pattern does not require the match to be at the start of the
1.1 misho 3431: subject.
3432:
3433: How pcre_exec() returns captured substrings
3434:
1.1.1.4 ! misho 3435: In general, a pattern matches a certain portion of the subject, and in
! 3436: addition, further substrings from the subject may be picked out by
! 3437: parts of the pattern. Following the usage in Jeffrey Friedl's book,
! 3438: this is called "capturing" in what follows, and the phrase "capturing
! 3439: subpattern" is used for a fragment of a pattern that picks out a sub-
! 3440: string. PCRE supports several other kinds of parenthesized subpattern
1.1 misho 3441: that do not cause substrings to be captured.
3442:
3443: Captured substrings are returned to the caller via a vector of integers
1.1.1.4 ! misho 3444: whose address is passed in ovector. The number of elements in the vec-
! 3445: tor is passed in ovecsize, which must be a non-negative number. Note:
1.1 misho 3446: this argument is NOT the size of ovector in bytes.
3447:
1.1.1.4 ! misho 3448: The first two-thirds of the vector is used to pass back captured sub-
! 3449: strings, each substring using a pair of integers. The remaining third
! 3450: of the vector is used as workspace by pcre_exec() while matching cap-
! 3451: turing subpatterns, and is not available for passing back information.
! 3452: The number passed in ovecsize should always be a multiple of three. If
1.1 misho 3453: it is not, it is rounded down.
3454:
1.1.1.4 ! misho 3455: When a match is successful, information about captured substrings is
! 3456: returned in pairs of integers, starting at the beginning of ovector,
! 3457: and continuing up to two-thirds of its length at the most. The first
! 3458: element of each pair is set to the offset of the first character in a
! 3459: substring, and the second is set to the offset of the first character
! 3460: after the end of a substring. These values are always data unit off-
! 3461: sets, even in UTF mode. They are byte offsets in the 8-bit library,
! 3462: 16-bit data item offsets in the 16-bit library, and 32-bit data item
! 3463: offsets in the 32-bit library. Note: they are not character counts.
! 3464:
! 3465: The first pair of integers, ovector[0] and ovector[1], identify the
! 3466: portion of the subject string matched by the entire pattern. The next
! 3467: pair is used for the first capturing subpattern, and so on. The value
1.1 misho 3468: returned by pcre_exec() is one more than the highest numbered pair that
1.1.1.4 ! misho 3469: has been set. For example, if two substrings have been captured, the
! 3470: returned value is 3. If there are no capturing subpatterns, the return
1.1 misho 3471: value from a successful match is 1, indicating that just the first pair
3472: of offsets has been set.
3473:
3474: If a capturing subpattern is matched repeatedly, it is the last portion
3475: of the string that it matched that is returned.
3476:
1.1.1.4 ! misho 3477: If the vector is too small to hold all the captured substring offsets,
1.1 misho 3478: it is used as far as possible (up to two-thirds of its length), and the
1.1.1.4 ! misho 3479: function returns a value of zero. If neither the actual string matched
! 3480: nor any captured substrings are of interest, pcre_exec() may be called
! 3481: with ovector passed as NULL and ovecsize as zero. However, if the pat-
! 3482: tern contains back references and the ovector is not big enough to
! 3483: remember the related substrings, PCRE has to get additional memory for
! 3484: use during matching. Thus it is usually advisable to supply an ovector
1.1 misho 3485: of reasonable size.
3486:
1.1.1.4 ! misho 3487: There are some cases where zero is returned (indicating vector over-
! 3488: flow) when in fact the vector is exactly the right size for the final
1.1 misho 3489: match. For example, consider the pattern
3490:
3491: (a)(?:(b)c|bd)
3492:
1.1.1.4 ! misho 3493: If a vector of 6 elements (allowing for only 1 captured substring) is
1.1 misho 3494: given with subject string "abd", pcre_exec() will try to set the second
3495: captured string, thereby recording a vector overflow, before failing to
1.1.1.4 ! misho 3496: match "c" and backing up to try the second alternative. The zero
! 3497: return, however, does correctly indicate that the maximum number of
1.1 misho 3498: slots (namely 2) have been filled. In similar cases where there is tem-
1.1.1.4 ! misho 3499: porary overflow, but the final number of used slots is actually less
1.1 misho 3500: than the maximum, a non-zero value is returned.
3501:
3502: The pcre_fullinfo() function can be used to find out how many capturing
1.1.1.4 ! misho 3503: subpatterns there are in a compiled pattern. The smallest size for
! 3504: ovector that will allow for n captured substrings, in addition to the
1.1 misho 3505: offsets of the substring matched by the whole pattern, is (n+1)*3.
3506:
1.1.1.4 ! misho 3507: It is possible for capturing subpattern number n+1 to match some part
1.1 misho 3508: of the subject when subpattern n has not been used at all. For example,
1.1.1.4 ! misho 3509: if the string "abc" is matched against the pattern (a|(z))(bc) the
1.1 misho 3510: return from the function is 4, and subpatterns 1 and 3 are matched, but
1.1.1.4 ! misho 3511: 2 is not. When this happens, both values in the offset pairs corre-
1.1 misho 3512: sponding to unused subpatterns are set to -1.
3513:
1.1.1.4 ! misho 3514: Offset values that correspond to unused subpatterns at the end of the
! 3515: expression are also set to -1. For example, if the string "abc" is
! 3516: matched against the pattern (abc)(x(yz)?)? subpatterns 2 and 3 are not
! 3517: matched. The return from the function is 2, because the highest used
! 3518: capturing subpattern number is 1, and the offsets for for the second
! 3519: and third capturing subpatterns (assuming the vector is large enough,
1.1 misho 3520: of course) are set to -1.
3521:
1.1.1.4 ! misho 3522: Note: Elements in the first two-thirds of ovector that do not corre-
! 3523: spond to capturing parentheses in the pattern are never changed. That
! 3524: is, if a pattern contains n capturing parentheses, no more than ovec-
! 3525: tor[0] to ovector[2n+1] are set by pcre_exec(). The other elements (in
1.1 misho 3526: the first two-thirds) retain whatever values they previously had.
3527:
1.1.1.4 ! misho 3528: Some convenience functions are provided for extracting the captured
1.1 misho 3529: substrings as separate strings. These are described below.
3530:
3531: Error return values from pcre_exec()
3532:
1.1.1.4 ! misho 3533: If pcre_exec() fails, it returns a negative number. The following are
1.1 misho 3534: defined in the header file:
3535:
3536: PCRE_ERROR_NOMATCH (-1)
3537:
3538: The subject string did not match the pattern.
3539:
3540: PCRE_ERROR_NULL (-2)
3541:
1.1.1.4 ! misho 3542: Either code or subject was passed as NULL, or ovector was NULL and
1.1 misho 3543: ovecsize was not zero.
3544:
3545: PCRE_ERROR_BADOPTION (-3)
3546:
3547: An unrecognized bit was set in the options argument.
3548:
3549: PCRE_ERROR_BADMAGIC (-4)
3550:
1.1.1.4 ! misho 3551: PCRE stores a 4-byte "magic number" at the start of the compiled code,
1.1 misho 3552: to catch the case when it is passed a junk pointer and to detect when a
3553: pattern that was compiled in an environment of one endianness is run in
1.1.1.4 ! misho 3554: an environment with the other endianness. This is the error that PCRE
1.1 misho 3555: gives when the magic number is not present.
3556:
3557: PCRE_ERROR_UNKNOWN_OPCODE (-5)
3558:
3559: While running the pattern match, an unknown item was encountered in the
1.1.1.4 ! misho 3560: compiled pattern. This error could be caused by a bug in PCRE or by
1.1 misho 3561: overwriting of the compiled pattern.
3562:
3563: PCRE_ERROR_NOMEMORY (-6)
3564:
1.1.1.4 ! misho 3565: If a pattern contains back references, but the ovector that is passed
1.1 misho 3566: to pcre_exec() is not big enough to remember the referenced substrings,
1.1.1.4 ! misho 3567: PCRE gets a block of memory at the start of matching to use for this
! 3568: purpose. If the call via pcre_malloc() fails, this error is given. The
1.1 misho 3569: memory is automatically freed at the end of matching.
3570:
1.1.1.4 ! misho 3571: This error is also given if pcre_stack_malloc() fails in pcre_exec().
! 3572: This can happen only when PCRE has been compiled with --disable-stack-
1.1 misho 3573: for-recursion.
3574:
3575: PCRE_ERROR_NOSUBSTRING (-7)
3576:
1.1.1.4 ! misho 3577: This error is used by the pcre_copy_substring(), pcre_get_substring(),
1.1 misho 3578: and pcre_get_substring_list() functions (see below). It is never
3579: returned by pcre_exec().
3580:
3581: PCRE_ERROR_MATCHLIMIT (-8)
3582:
1.1.1.4 ! misho 3583: The backtracking limit, as specified by the match_limit field in a
! 3584: pcre_extra structure (or defaulted) was reached. See the description
1.1 misho 3585: above.
3586:
3587: PCRE_ERROR_CALLOUT (-9)
3588:
3589: This error is never generated by pcre_exec() itself. It is provided for
1.1.1.4 ! misho 3590: use by callout functions that want to yield a distinctive error code.
1.1 misho 3591: See the pcrecallout documentation for details.
3592:
3593: PCRE_ERROR_BADUTF8 (-10)
3594:
1.1.1.4 ! misho 3595: A string that contains an invalid UTF-8 byte sequence was passed as a
! 3596: subject, and the PCRE_NO_UTF8_CHECK option was not set. If the size of
! 3597: the output vector (ovecsize) is at least 2, the byte offset to the
! 3598: start of the the invalid UTF-8 character is placed in the first ele-
! 3599: ment, and a reason code is placed in the second element. The reason
1.1 misho 3600: codes are listed in the following section. For backward compatibility,
1.1.1.4 ! misho 3601: if PCRE_PARTIAL_HARD is set and the problem is a truncated UTF-8 char-
! 3602: acter at the end of the subject (reason codes 1 to 5),
1.1 misho 3603: PCRE_ERROR_SHORTUTF8 is returned instead of PCRE_ERROR_BADUTF8.
3604:
3605: PCRE_ERROR_BADUTF8_OFFSET (-11)
3606:
1.1.1.4 ! misho 3607: The UTF-8 byte sequence that was passed as a subject was checked and
! 3608: found to be valid (the PCRE_NO_UTF8_CHECK option was not set), but the
! 3609: value of startoffset did not point to the beginning of a UTF-8 charac-
1.1 misho 3610: ter or the end of the subject.
3611:
3612: PCRE_ERROR_PARTIAL (-12)
3613:
1.1.1.4 ! misho 3614: The subject string did not match, but it did match partially. See the
1.1 misho 3615: pcrepartial documentation for details of partial matching.
3616:
3617: PCRE_ERROR_BADPARTIAL (-13)
3618:
1.1.1.4 ! misho 3619: This code is no longer in use. It was formerly returned when the
! 3620: PCRE_PARTIAL option was used with a compiled pattern containing items
! 3621: that were not supported for partial matching. From release 8.00
1.1 misho 3622: onwards, there are no restrictions on partial matching.
3623:
3624: PCRE_ERROR_INTERNAL (-14)
3625:
1.1.1.4 ! misho 3626: An unexpected internal error has occurred. This error could be caused
1.1 misho 3627: by a bug in PCRE or by overwriting of the compiled pattern.
3628:
3629: PCRE_ERROR_BADCOUNT (-15)
3630:
3631: This error is given if the value of the ovecsize argument is negative.
3632:
3633: PCRE_ERROR_RECURSIONLIMIT (-21)
3634:
3635: The internal recursion limit, as specified by the match_limit_recursion
1.1.1.4 ! misho 3636: field in a pcre_extra structure (or defaulted) was reached. See the
1.1 misho 3637: description above.
3638:
3639: PCRE_ERROR_BADNEWLINE (-23)
3640:
3641: An invalid combination of PCRE_NEWLINE_xxx options was given.
3642:
3643: PCRE_ERROR_BADOFFSET (-24)
3644:
3645: The value of startoffset was negative or greater than the length of the
3646: subject, that is, the value in length.
3647:
3648: PCRE_ERROR_SHORTUTF8 (-25)
3649:
1.1.1.4 ! misho 3650: This error is returned instead of PCRE_ERROR_BADUTF8 when the subject
! 3651: string ends with a truncated UTF-8 character and the PCRE_PARTIAL_HARD
! 3652: option is set. Information about the failure is returned as for
! 3653: PCRE_ERROR_BADUTF8. It is in fact sufficient to detect this case, but
! 3654: this special error code for PCRE_PARTIAL_HARD precedes the implementa-
! 3655: tion of returned information; it is retained for backwards compatibil-
1.1 misho 3656: ity.
3657:
3658: PCRE_ERROR_RECURSELOOP (-26)
3659:
3660: This error is returned when pcre_exec() detects a recursion loop within
1.1.1.4 ! misho 3661: the pattern. Specifically, it means that either the whole pattern or a
! 3662: subpattern has been called recursively for the second time at the same
1.1 misho 3663: position in the subject string. Some simple patterns that might do this
1.1.1.4 ! misho 3664: are detected and faulted at compile time, but more complicated cases,
1.1 misho 3665: in particular mutual recursions between two different subpatterns, can-
3666: not be detected until run time.
3667:
3668: PCRE_ERROR_JIT_STACKLIMIT (-27)
3669:
1.1.1.4 ! misho 3670: This error is returned when a pattern that was successfully studied
! 3671: using a JIT compile option is being matched, but the memory available
! 3672: for the just-in-time processing stack is not large enough. See the
1.1.1.3 misho 3673: pcrejit documentation for more details.
1.1 misho 3674:
1.1.1.3 misho 3675: PCRE_ERROR_BADMODE (-28)
1.1.1.2 misho 3676:
3677: This error is given if a pattern that was compiled by the 8-bit library
1.1.1.4 ! misho 3678: is passed to a 16-bit or 32-bit library function, or vice versa.
1.1.1.2 misho 3679:
1.1.1.3 misho 3680: PCRE_ERROR_BADENDIANNESS (-29)
1.1.1.2 misho 3681:
1.1.1.4 ! misho 3682: This error is given if a pattern that was compiled and saved is
! 3683: reloaded on a host with different endianness. The utility function
1.1.1.2 misho 3684: pcre_pattern_to_host_byte_order() can be used to convert such a pattern
3685: so that it runs on the new host.
3686:
1.1.1.4 ! misho 3687: PCRE_ERROR_JIT_BADOPTION
! 3688:
! 3689: This error is returned when a pattern that was successfully studied
! 3690: using a JIT compile option is being matched, but the matching mode
! 3691: (partial or complete match) does not correspond to any JIT compilation
! 3692: mode. When the JIT fast path function is used, this error may be also
! 3693: given for invalid options. See the pcrejit documentation for more
! 3694: details.
! 3695:
! 3696: PCRE_ERROR_BADLENGTH (-32)
! 3697:
! 3698: This error is given if pcre_exec() is called with a negative value for
! 3699: the length argument.
! 3700:
! 3701: Error numbers -16 to -20, -22, and 30 are not used by pcre_exec().
1.1 misho 3702:
3703: Reason codes for invalid UTF-8 strings
3704:
1.1.1.4 ! misho 3705: This section applies only to the 8-bit library. The corresponding
! 3706: information for the 16-bit and 32-bit libraries is given in the pcre16
! 3707: and pcre32 pages.
1.1.1.2 misho 3708:
1.1 misho 3709: When pcre_exec() returns either PCRE_ERROR_BADUTF8 or PCRE_ERROR_SHORT-
1.1.1.3 misho 3710: UTF8, and the size of the output vector (ovecsize) is at least 2, the
3711: offset of the start of the invalid UTF-8 character is placed in the
1.1 misho 3712: first output vector element (ovector[0]) and a reason code is placed in
1.1.1.3 misho 3713: the second element (ovector[1]). The reason codes are given names in
1.1 misho 3714: the pcre.h header file:
3715:
3716: PCRE_UTF8_ERR1
3717: PCRE_UTF8_ERR2
3718: PCRE_UTF8_ERR3
3719: PCRE_UTF8_ERR4
3720: PCRE_UTF8_ERR5
3721:
1.1.1.3 misho 3722: The string ends with a truncated UTF-8 character; the code specifies
3723: how many bytes are missing (1 to 5). Although RFC 3629 restricts UTF-8
3724: characters to be no longer than 4 bytes, the encoding scheme (origi-
3725: nally defined by RFC 2279) allows for up to 6 bytes, and this is
1.1 misho 3726: checked first; hence the possibility of 4 or 5 missing bytes.
3727:
3728: PCRE_UTF8_ERR6
3729: PCRE_UTF8_ERR7
3730: PCRE_UTF8_ERR8
3731: PCRE_UTF8_ERR9
3732: PCRE_UTF8_ERR10
3733:
3734: The two most significant bits of the 2nd, 3rd, 4th, 5th, or 6th byte of
1.1.1.3 misho 3735: the character do not have the binary value 0b10 (that is, either the
1.1 misho 3736: most significant bit is 0, or the next bit is 1).
3737:
3738: PCRE_UTF8_ERR11
3739: PCRE_UTF8_ERR12
3740:
1.1.1.3 misho 3741: A character that is valid by the RFC 2279 rules is either 5 or 6 bytes
1.1 misho 3742: long; these code points are excluded by RFC 3629.
3743:
3744: PCRE_UTF8_ERR13
3745:
1.1.1.3 misho 3746: A 4-byte character has a value greater than 0x10fff; these code points
1.1 misho 3747: are excluded by RFC 3629.
3748:
3749: PCRE_UTF8_ERR14
3750:
1.1.1.3 misho 3751: A 3-byte character has a value in the range 0xd800 to 0xdfff; this
3752: range of code points are reserved by RFC 3629 for use with UTF-16, and
1.1 misho 3753: so are excluded from UTF-8.
3754:
3755: PCRE_UTF8_ERR15
3756: PCRE_UTF8_ERR16
3757: PCRE_UTF8_ERR17
3758: PCRE_UTF8_ERR18
3759: PCRE_UTF8_ERR19
3760:
1.1.1.3 misho 3761: A 2-, 3-, 4-, 5-, or 6-byte character is "overlong", that is, it codes
3762: for a value that can be represented by fewer bytes, which is invalid.
3763: For example, the two bytes 0xc0, 0xae give the value 0x2e, whose cor-
1.1 misho 3764: rect coding uses just one byte.
3765:
3766: PCRE_UTF8_ERR20
3767:
3768: The two most significant bits of the first byte of a character have the
1.1.1.3 misho 3769: binary value 0b10 (that is, the most significant bit is 1 and the sec-
3770: ond is 0). Such a byte can only validly occur as the second or subse-
1.1 misho 3771: quent byte of a multi-byte character.
3772:
3773: PCRE_UTF8_ERR21
3774:
1.1.1.3 misho 3775: The first byte of a character has the value 0xfe or 0xff. These values
1.1 misho 3776: can never occur in a valid UTF-8 string.
3777:
1.1.1.4 ! misho 3778: PCRE_UTF8_ERR22
! 3779:
! 3780: This error code was formerly used when the presence of a so-called
! 3781: "non-character" caused an error. Unicode corrigendum #9 makes it clear
! 3782: that such characters should not cause a string to be rejected, and so
! 3783: this code is no longer in use and is never returned.
! 3784:
1.1 misho 3785:
3786: EXTRACTING CAPTURED SUBSTRINGS BY NUMBER
3787:
3788: int pcre_copy_substring(const char *subject, int *ovector,
3789: int stringcount, int stringnumber, char *buffer,
3790: int buffersize);
3791:
3792: int pcre_get_substring(const char *subject, int *ovector,
3793: int stringcount, int stringnumber,
3794: const char **stringptr);
3795:
3796: int pcre_get_substring_list(const char *subject,
3797: int *ovector, int stringcount, const char ***listptr);
3798:
1.1.1.4 ! misho 3799: Captured substrings can be accessed directly by using the offsets
! 3800: returned by pcre_exec() in ovector. For convenience, the functions
1.1 misho 3801: pcre_copy_substring(), pcre_get_substring(), and pcre_get_sub-
1.1.1.4 ! misho 3802: string_list() are provided for extracting captured substrings as new,
! 3803: separate, zero-terminated strings. These functions identify substrings
! 3804: by number. The next section describes functions for extracting named
1.1 misho 3805: substrings.
3806:
1.1.1.4 ! misho 3807: A substring that contains a binary zero is correctly extracted and has
! 3808: a further zero added on the end, but the result is not, of course, a C
! 3809: string. However, you can process such a string by referring to the
! 3810: length that is returned by pcre_copy_substring() and pcre_get_sub-
1.1 misho 3811: string(). Unfortunately, the interface to pcre_get_substring_list() is
1.1.1.4 ! misho 3812: not adequate for handling strings containing binary zeros, because the
1.1 misho 3813: end of the final string is not independently indicated.
3814:
1.1.1.4 ! misho 3815: The first three arguments are the same for all three of these func-
! 3816: tions: subject is the subject string that has just been successfully
1.1 misho 3817: matched, ovector is a pointer to the vector of integer offsets that was
3818: passed to pcre_exec(), and stringcount is the number of substrings that
1.1.1.4 ! misho 3819: were captured by the match, including the substring that matched the
1.1 misho 3820: entire regular expression. This is the value returned by pcre_exec() if
1.1.1.4 ! misho 3821: it is greater than zero. If pcre_exec() returned zero, indicating that
! 3822: it ran out of space in ovector, the value passed as stringcount should
1.1 misho 3823: be the number of elements in the vector divided by three.
3824:
1.1.1.4 ! misho 3825: The functions pcre_copy_substring() and pcre_get_substring() extract a
! 3826: single substring, whose number is given as stringnumber. A value of
! 3827: zero extracts the substring that matched the entire pattern, whereas
! 3828: higher values extract the captured substrings. For pcre_copy_sub-
! 3829: string(), the string is placed in buffer, whose length is given by
! 3830: buffersize, while for pcre_get_substring() a new block of memory is
! 3831: obtained via pcre_malloc, and its address is returned via stringptr.
! 3832: The yield of the function is the length of the string, not including
1.1 misho 3833: the terminating zero, or one of these error codes:
3834:
3835: PCRE_ERROR_NOMEMORY (-6)
3836:
1.1.1.4 ! misho 3837: The buffer was too small for pcre_copy_substring(), or the attempt to
1.1 misho 3838: get memory failed for pcre_get_substring().
3839:
3840: PCRE_ERROR_NOSUBSTRING (-7)
3841:
3842: There is no substring whose number is stringnumber.
3843:
1.1.1.4 ! misho 3844: The pcre_get_substring_list() function extracts all available sub-
! 3845: strings and builds a list of pointers to them. All this is done in a
1.1 misho 3846: single block of memory that is obtained via pcre_malloc. The address of
1.1.1.4 ! misho 3847: the memory block is returned via listptr, which is also the start of
! 3848: the list of string pointers. The end of the list is marked by a NULL
! 3849: pointer. The yield of the function is zero if all went well, or the
1.1 misho 3850: error code
3851:
3852: PCRE_ERROR_NOMEMORY (-6)
3853:
3854: if the attempt to get the memory block failed.
3855:
1.1.1.4 ! misho 3856: When any of these functions encounter a substring that is unset, which
! 3857: can happen when capturing subpattern number n+1 matches some part of
! 3858: the subject, but subpattern n has not been used at all, they return an
1.1 misho 3859: empty string. This can be distinguished from a genuine zero-length sub-
1.1.1.4 ! misho 3860: string by inspecting the appropriate offset in ovector, which is nega-
1.1 misho 3861: tive for unset substrings.
3862:
1.1.1.4 ! misho 3863: The two convenience functions pcre_free_substring() and pcre_free_sub-
! 3864: string_list() can be used to free the memory returned by a previous
1.1 misho 3865: call of pcre_get_substring() or pcre_get_substring_list(), respec-
1.1.1.4 ! misho 3866: tively. They do nothing more than call the function pointed to by
! 3867: pcre_free, which of course could be called directly from a C program.
! 3868: However, PCRE is used in some situations where it is linked via a spe-
! 3869: cial interface to another programming language that cannot use
! 3870: pcre_free directly; it is for these cases that the functions are pro-
1.1 misho 3871: vided.
3872:
3873:
3874: EXTRACTING CAPTURED SUBSTRINGS BY NAME
3875:
3876: int pcre_get_stringnumber(const pcre *code,
3877: const char *name);
3878:
3879: int pcre_copy_named_substring(const pcre *code,
3880: const char *subject, int *ovector,
3881: int stringcount, const char *stringname,
3882: char *buffer, int buffersize);
3883:
3884: int pcre_get_named_substring(const pcre *code,
3885: const char *subject, int *ovector,
3886: int stringcount, const char *stringname,
3887: const char **stringptr);
3888:
1.1.1.4 ! misho 3889: To extract a substring by name, you first have to find associated num-
1.1 misho 3890: ber. For example, for this pattern
3891:
3892: (a+)b(?<xxx>\d+)...
3893:
3894: the number of the subpattern called "xxx" is 2. If the name is known to
3895: be unique (PCRE_DUPNAMES was not set), you can find the number from the
3896: name by calling pcre_get_stringnumber(). The first argument is the com-
3897: piled pattern, and the second is the name. The yield of the function is
1.1.1.4 ! misho 3898: the subpattern number, or PCRE_ERROR_NOSUBSTRING (-7) if there is no
1.1 misho 3899: subpattern of that name.
3900:
3901: Given the number, you can extract the substring directly, or use one of
3902: the functions described in the previous section. For convenience, there
3903: are also two functions that do the whole job.
3904:
1.1.1.4 ! misho 3905: Most of the arguments of pcre_copy_named_substring() and
! 3906: pcre_get_named_substring() are the same as those for the similarly
! 3907: named functions that extract by number. As these are described in the
! 3908: previous section, they are not re-described here. There are just two
1.1 misho 3909: differences:
3910:
1.1.1.4 ! misho 3911: First, instead of a substring number, a substring name is given. Sec-
1.1 misho 3912: ond, there is an extra argument, given at the start, which is a pointer
1.1.1.4 ! misho 3913: to the compiled pattern. This is needed in order to gain access to the
1.1 misho 3914: name-to-number translation table.
3915:
1.1.1.4 ! misho 3916: These functions call pcre_get_stringnumber(), and if it succeeds, they
! 3917: then call pcre_copy_substring() or pcre_get_substring(), as appropri-
! 3918: ate. NOTE: If PCRE_DUPNAMES is set and there are duplicate names, the
1.1 misho 3919: behaviour may not be what you want (see the next section).
3920:
3921: Warning: If the pattern uses the (?| feature to set up multiple subpat-
1.1.1.4 ! misho 3922: terns with the same number, as described in the section on duplicate
! 3923: subpattern numbers in the pcrepattern page, you cannot use names to
! 3924: distinguish the different subpatterns, because names are not included
! 3925: in the compiled code. The matching process uses only numbers. For this
! 3926: reason, the use of different names for subpatterns of the same number
1.1 misho 3927: causes an error at compile time.
3928:
3929:
3930: DUPLICATE SUBPATTERN NAMES
3931:
3932: int pcre_get_stringtable_entries(const pcre *code,
3933: const char *name, char **first, char **last);
3934:
1.1.1.4 ! misho 3935: When a pattern is compiled with the PCRE_DUPNAMES option, names for
! 3936: subpatterns are not required to be unique. (Duplicate names are always
! 3937: allowed for subpatterns with the same number, created by using the (?|
! 3938: feature. Indeed, if such subpatterns are named, they are required to
1.1 misho 3939: use the same names.)
3940:
3941: Normally, patterns with duplicate names are such that in any one match,
1.1.1.4 ! misho 3942: only one of the named subpatterns participates. An example is shown in
1.1 misho 3943: the pcrepattern documentation.
3944:
1.1.1.4 ! misho 3945: When duplicates are present, pcre_copy_named_substring() and
! 3946: pcre_get_named_substring() return the first substring corresponding to
! 3947: the given name that is set. If none are set, PCRE_ERROR_NOSUBSTRING
! 3948: (-7) is returned; no data is returned. The pcre_get_stringnumber()
! 3949: function returns one of the numbers that are associated with the name,
1.1 misho 3950: but it is not defined which it is.
3951:
1.1.1.4 ! misho 3952: If you want to get full details of all captured substrings for a given
! 3953: name, you must use the pcre_get_stringtable_entries() function. The
1.1 misho 3954: first argument is the compiled pattern, and the second is the name. The
1.1.1.4 ! misho 3955: third and fourth are pointers to variables which are updated by the
1.1 misho 3956: function. After it has run, they point to the first and last entries in
1.1.1.4 ! misho 3957: the name-to-number table for the given name. The function itself
! 3958: returns the length of each entry, or PCRE_ERROR_NOSUBSTRING (-7) if
! 3959: there are none. The format of the table is described above in the sec-
! 3960: tion entitled Information about a pattern above. Given all the rele-
! 3961: vant entries for the name, you can extract each of their numbers, and
1.1 misho 3962: hence the captured data, if any.
3963:
3964:
3965: FINDING ALL POSSIBLE MATCHES
3966:
1.1.1.4 ! misho 3967: The traditional matching function uses a similar algorithm to Perl,
1.1 misho 3968: which stops when it finds the first match, starting at a given point in
1.1.1.4 ! misho 3969: the subject. If you want to find all possible matches, or the longest
! 3970: possible match, consider using the alternative matching function (see
! 3971: below) instead. If you cannot use the alternative function, but still
! 3972: need to find all possible matches, you can kludge it up by making use
1.1 misho 3973: of the callout facility, which is described in the pcrecallout documen-
3974: tation.
3975:
3976: What you have to do is to insert a callout right at the end of the pat-
1.1.1.4 ! misho 3977: tern. When your callout function is called, extract and save the cur-
! 3978: rent matched substring. Then return 1, which forces pcre_exec() to
! 3979: backtrack and try other alternatives. Ultimately, when it runs out of
1.1 misho 3980: matches, pcre_exec() will yield PCRE_ERROR_NOMATCH.
3981:
3982:
1.1.1.2 misho 3983: OBTAINING AN ESTIMATE OF STACK USAGE
3984:
1.1.1.4 ! misho 3985: Matching certain patterns using pcre_exec() can use a lot of process
! 3986: stack, which in certain environments can be rather limited in size.
! 3987: Some users find it helpful to have an estimate of the amount of stack
! 3988: that is used by pcre_exec(), to help them set recursion limits, as
! 3989: described in the pcrestack documentation. The estimate that is output
1.1.1.2 misho 3990: by pcretest when called with the -m and -C options is obtained by call-
1.1.1.4 ! misho 3991: ing pcre_exec with the values NULL, NULL, NULL, -999, and -999 for its
1.1.1.2 misho 3992: first five arguments.
3993:
1.1.1.4 ! misho 3994: Normally, if its first argument is NULL, pcre_exec() immediately
! 3995: returns the negative error code PCRE_ERROR_NULL, but with this special
! 3996: combination of arguments, it returns instead a negative number whose
! 3997: absolute value is the approximate stack frame size in bytes. (A nega-
! 3998: tive number is used so that it is clear that no match has happened.)
! 3999: The value is approximate because in some cases, recursive calls to
1.1.1.2 misho 4000: pcre_exec() occur when there are one or two additional variables on the
4001: stack.
4002:
1.1.1.4 ! misho 4003: If PCRE has been compiled to use the heap instead of the stack for
! 4004: recursion, the value returned is the size of each block that is
1.1.1.2 misho 4005: obtained from the heap.
4006:
4007:
1.1 misho 4008: MATCHING A PATTERN: THE ALTERNATIVE FUNCTION
4009:
4010: int pcre_dfa_exec(const pcre *code, const pcre_extra *extra,
4011: const char *subject, int length, int startoffset,
4012: int options, int *ovector, int ovecsize,
4013: int *workspace, int wscount);
4014:
1.1.1.4 ! misho 4015: The function pcre_dfa_exec() is called to match a subject string
! 4016: against a compiled pattern, using a matching algorithm that scans the
! 4017: subject string just once, and does not backtrack. This has different
! 4018: characteristics to the normal algorithm, and is not compatible with
! 4019: Perl. Some of the features of PCRE patterns are not supported. Never-
! 4020: theless, there are times when this kind of matching can be useful. For
! 4021: a discussion of the two matching algorithms, and a list of features
! 4022: that pcre_dfa_exec() does not support, see the pcrematching documenta-
1.1 misho 4023: tion.
4024:
1.1.1.4 ! misho 4025: The arguments for the pcre_dfa_exec() function are the same as for
1.1 misho 4026: pcre_exec(), plus two extras. The ovector argument is used in a differ-
1.1.1.4 ! misho 4027: ent way, and this is described below. The other common arguments are
! 4028: used in the same way as for pcre_exec(), so their description is not
1.1 misho 4029: repeated here.
4030:
1.1.1.4 ! misho 4031: The two additional arguments provide workspace for the function. The
! 4032: workspace vector should contain at least 20 elements. It is used for
1.1 misho 4033: keeping track of multiple paths through the pattern tree. More
1.1.1.4 ! misho 4034: workspace will be needed for patterns and subjects where there are a
1.1 misho 4035: lot of potential matches.
4036:
4037: Here is an example of a simple call to pcre_dfa_exec():
4038:
4039: int rc;
4040: int ovector[10];
4041: int wspace[20];
4042: rc = pcre_dfa_exec(
4043: re, /* result of pcre_compile() */
4044: NULL, /* we didn't study the pattern */
4045: "some string", /* the subject string */
4046: 11, /* the length of the subject string */
4047: 0, /* start at offset 0 in the subject */
4048: 0, /* default options */
4049: ovector, /* vector of integers for substring information */
4050: 10, /* number of elements (NOT size in bytes) */
4051: wspace, /* working space vector */
4052: 20); /* number of elements (NOT size in bytes) */
4053:
4054: Option bits for pcre_dfa_exec()
4055:
1.1.1.4 ! misho 4056: The unused bits of the options argument for pcre_dfa_exec() must be
! 4057: zero. The only bits that may be set are PCRE_ANCHORED, PCRE_NEW-
1.1 misho 4058: LINE_xxx, PCRE_NOTBOL, PCRE_NOTEOL, PCRE_NOTEMPTY,
1.1.1.4 ! misho 4059: PCRE_NOTEMPTY_ATSTART, PCRE_NO_UTF8_CHECK, PCRE_BSR_ANYCRLF,
! 4060: PCRE_BSR_UNICODE, PCRE_NO_START_OPTIMIZE, PCRE_PARTIAL_HARD, PCRE_PAR-
! 4061: TIAL_SOFT, PCRE_DFA_SHORTEST, and PCRE_DFA_RESTART. All but the last
! 4062: four of these are exactly the same as for pcre_exec(), so their
1.1 misho 4063: description is not repeated here.
4064:
4065: PCRE_PARTIAL_HARD
4066: PCRE_PARTIAL_SOFT
4067:
1.1.1.4 ! misho 4068: These have the same general effect as they do for pcre_exec(), but the
! 4069: details are slightly different. When PCRE_PARTIAL_HARD is set for
! 4070: pcre_dfa_exec(), it returns PCRE_ERROR_PARTIAL if the end of the sub-
! 4071: ject is reached and there is still at least one matching possibility
1.1 misho 4072: that requires additional characters. This happens even if some complete
4073: matches have also been found. When PCRE_PARTIAL_SOFT is set, the return
4074: code PCRE_ERROR_NOMATCH is converted into PCRE_ERROR_PARTIAL if the end
1.1.1.4 ! misho 4075: of the subject is reached, there have been no complete matches, but
! 4076: there is still at least one matching possibility. The portion of the
! 4077: string that was inspected when the longest partial match was found is
! 4078: set as the first matching string in both cases. There is a more
! 4079: detailed discussion of partial and multi-segment matching, with exam-
1.1 misho 4080: ples, in the pcrepartial documentation.
4081:
4082: PCRE_DFA_SHORTEST
4083:
1.1.1.4 ! misho 4084: Setting the PCRE_DFA_SHORTEST option causes the matching algorithm to
1.1 misho 4085: stop as soon as it has found one match. Because of the way the alterna-
1.1.1.4 ! misho 4086: tive algorithm works, this is necessarily the shortest possible match
1.1 misho 4087: at the first possible matching point in the subject string.
4088:
4089: PCRE_DFA_RESTART
4090:
4091: When pcre_dfa_exec() returns a partial match, it is possible to call it
1.1.1.4 ! misho 4092: again, with additional subject characters, and have it continue with
! 4093: the same match. The PCRE_DFA_RESTART option requests this action; when
! 4094: it is set, the workspace and wscount options must reference the same
! 4095: vector as before because data about the match so far is left in them
1.1 misho 4096: after a partial match. There is more discussion of this facility in the
4097: pcrepartial documentation.
4098:
4099: Successful returns from pcre_dfa_exec()
4100:
1.1.1.4 ! misho 4101: When pcre_dfa_exec() succeeds, it may have matched more than one sub-
1.1 misho 4102: string in the subject. Note, however, that all the matches from one run
1.1.1.4 ! misho 4103: of the function start at the same point in the subject. The shorter
! 4104: matches are all initial substrings of the longer matches. For example,
1.1 misho 4105: if the pattern
4106:
4107: <.*>
4108:
4109: is matched against the string
4110:
4111: This is <something> <something else> <something further> no more
4112:
4113: the three matched strings are
4114:
4115: <something>
4116: <something> <something else>
4117: <something> <something else> <something further>
4118:
1.1.1.4 ! misho 4119: On success, the yield of the function is a number greater than zero,
! 4120: which is the number of matched substrings. The substrings themselves
! 4121: are returned in ovector. Each string uses two elements; the first is
! 4122: the offset to the start, and the second is the offset to the end. In
! 4123: fact, all the strings have the same start offset. (Space could have
! 4124: been saved by giving this only once, but it was decided to retain some
! 4125: compatibility with the way pcre_exec() returns data, even though the
1.1 misho 4126: meaning of the strings is different.)
4127:
4128: The strings are returned in reverse order of length; that is, the long-
1.1.1.4 ! misho 4129: est matching string is given first. If there were too many matches to
! 4130: fit into ovector, the yield of the function is zero, and the vector is
! 4131: filled with the longest matches. Unlike pcre_exec(), pcre_dfa_exec()
1.1 misho 4132: can use the entire ovector for returning matched strings.
4133:
4134: Error returns from pcre_dfa_exec()
4135:
1.1.1.4 ! misho 4136: The pcre_dfa_exec() function returns a negative number when it fails.
! 4137: Many of the errors are the same as for pcre_exec(), and these are
! 4138: described above. There are in addition the following errors that are
1.1 misho 4139: specific to pcre_dfa_exec():
4140:
4141: PCRE_ERROR_DFA_UITEM (-16)
4142:
1.1.1.4 ! misho 4143: This return is given if pcre_dfa_exec() encounters an item in the pat-
! 4144: tern that it does not support, for instance, the use of \C or a back
1.1 misho 4145: reference.
4146:
4147: PCRE_ERROR_DFA_UCOND (-17)
4148:
1.1.1.4 ! misho 4149: This return is given if pcre_dfa_exec() encounters a condition item
! 4150: that uses a back reference for the condition, or a test for recursion
1.1 misho 4151: in a specific group. These are not supported.
4152:
4153: PCRE_ERROR_DFA_UMLIMIT (-18)
4154:
1.1.1.4 ! misho 4155: This return is given if pcre_dfa_exec() is called with an extra block
! 4156: that contains a setting of the match_limit or match_limit_recursion
! 4157: fields. This is not supported (these fields are meaningless for DFA
1.1 misho 4158: matching).
4159:
4160: PCRE_ERROR_DFA_WSSIZE (-19)
4161:
1.1.1.4 ! misho 4162: This return is given if pcre_dfa_exec() runs out of space in the
1.1 misho 4163: workspace vector.
4164:
4165: PCRE_ERROR_DFA_RECURSE (-20)
4166:
1.1.1.4 ! misho 4167: When a recursive subpattern is processed, the matching function calls
! 4168: itself recursively, using private vectors for ovector and workspace.
! 4169: This error is given if the output vector is not large enough. This
1.1 misho 4170: should be extremely rare, as a vector of size 1000 is used.
4171:
1.1.1.3 misho 4172: PCRE_ERROR_DFA_BADRESTART (-30)
4173:
1.1.1.4 ! misho 4174: When pcre_dfa_exec() is called with the PCRE_DFA_RESTART option, some
! 4175: plausibility checks are made on the contents of the workspace, which
! 4176: should contain data about the previous partial match. If any of these
1.1.1.3 misho 4177: checks fail, this error is given.
4178:
1.1 misho 4179:
4180: SEE ALSO
4181:
1.1.1.4 ! misho 4182: pcre16(3), pcre32(3), pcrebuild(3), pcrecallout(3), pcrecpp(3)(3),
! 4183: pcrematching(3), pcrepartial(3), pcreposix(3), pcreprecompile(3), pcre-
! 4184: sample(3), pcrestack(3).
1.1 misho 4185:
4186:
4187: AUTHOR
4188:
4189: Philip Hazel
4190: University Computing Service
4191: Cambridge CB2 3QH, England.
4192:
4193:
4194: REVISION
4195:
1.1.1.4 ! misho 4196: Last updated: 12 May 2013
! 4197: Copyright (c) 1997-2013 University of Cambridge.
1.1 misho 4198: ------------------------------------------------------------------------------
4199:
4200:
1.1.1.4 ! misho 4201: PCRECALLOUT(3) Library Functions Manual PCRECALLOUT(3)
! 4202:
1.1 misho 4203:
4204:
4205: NAME
4206: PCRE - Perl-compatible regular expressions
4207:
1.1.1.4 ! misho 4208: SYNOPSIS
1.1 misho 4209:
1.1.1.4 ! misho 4210: #include <pcre.h>
1.1 misho 4211:
4212: int (*pcre_callout)(pcre_callout_block *);
4213:
1.1.1.2 misho 4214: int (*pcre16_callout)(pcre16_callout_block *);
4215:
1.1.1.4 ! misho 4216: int (*pcre32_callout)(pcre32_callout_block *);
! 4217:
! 4218:
! 4219: DESCRIPTION
! 4220:
1.1 misho 4221: PCRE provides a feature called "callout", which is a means of temporar-
4222: ily passing control to the caller of PCRE in the middle of pattern
4223: matching. The caller of PCRE provides an external function by putting
1.1.1.2 misho 4224: its entry point in the global variable pcre_callout (pcre16_callout for
1.1.1.4 ! misho 4225: the 16-bit library, pcre32_callout for the 32-bit library). By default,
! 4226: this variable contains NULL, which disables all calling out.
1.1 misho 4227:
1.1.1.2 misho 4228: Within a regular expression, (?C) indicates the points at which the
4229: external function is to be called. Different callout points can be
4230: identified by putting a number less than 256 after the letter C. The
4231: default value is zero. For example, this pattern has two callout
1.1 misho 4232: points:
4233:
4234: (?C1)abc(?C2)def
4235:
1.1.1.2 misho 4236: If the PCRE_AUTO_CALLOUT option bit is set when a pattern is compiled,
4237: PCRE automatically inserts callouts, all with number 255, before each
4238: item in the pattern. For example, if PCRE_AUTO_CALLOUT is used with the
4239: pattern
1.1 misho 4240:
4241: A(\d{2}|--)
4242:
4243: it is processed as if it were
4244:
4245: (?C255)A(?C255)((?C255)\d{2}(?C255)|(?C255)-(?C255)-(?C255))(?C255)
4246:
1.1.1.2 misho 4247: Notice that there is a callout before and after each parenthesis and
1.1.1.4 ! misho 4248: alternation bar. If the pattern contains a conditional group whose con-
! 4249: dition is an assertion, an automatic callout is inserted immediately
! 4250: before the condition. Such a callout may also be inserted explicitly,
! 4251: for example:
! 4252:
! 4253: (?(?C9)(?=a)ab|de)
! 4254:
! 4255: This applies only to assertion conditions (because they are themselves
! 4256: independent groups).
! 4257:
! 4258: Automatic callouts can be used for tracking the progress of pattern
! 4259: matching. The pcretest command has an option that sets automatic call-
! 4260: outs; when it is used, the output indicates how the pattern is matched.
! 4261: This is useful information when you are trying to optimize the perfor-
! 4262: mance of a particular pattern.
1.1 misho 4263:
4264:
4265: MISSING CALLOUTS
4266:
1.1.1.2 misho 4267: You should be aware that, because of optimizations in the way PCRE
4268: matches patterns by default, callouts sometimes do not happen. For
1.1 misho 4269: example, if the pattern is
4270:
4271: ab(?C4)cd
4272:
4273: PCRE knows that any matching string must contain the letter "d". If the
1.1.1.2 misho 4274: subject string is "abyz", the lack of "d" means that matching doesn't
4275: ever start, and the callout is never reached. However, with "abyd",
1.1 misho 4276: though the result is still no match, the callout is obeyed.
4277:
1.1.1.2 misho 4278: If the pattern is studied, PCRE knows the minimum length of a matching
4279: string, and will immediately give a "no match" return without actually
4280: running a match if the subject is not long enough, or, for unanchored
1.1 misho 4281: patterns, if it has been scanned far enough.
4282:
1.1.1.2 misho 4283: You can disable these optimizations by passing the PCRE_NO_START_OPTI-
4284: MIZE option to the matching function, or by starting the pattern with
4285: (*NO_START_OPT). This slows down the matching process, but does ensure
4286: that callouts such as the example above are obeyed.
1.1 misho 4287:
4288:
4289: THE CALLOUT INTERFACE
4290:
4291: During matching, when PCRE reaches a callout point, the external func-
1.1.1.4 ! misho 4292: tion defined by pcre_callout or pcre[16|32]_callout is called (if it is
! 4293: set). This applies to both normal and DFA matching. The only argument
! 4294: to the callout function is a pointer to a pcre_callout or
! 4295: pcre[16|32]_callout block. These structures contains the following
! 4296: fields:
1.1.1.2 misho 4297:
4298: int version;
4299: int callout_number;
4300: int *offset_vector;
4301: const char *subject; (8-bit version)
4302: PCRE_SPTR16 subject; (16-bit version)
1.1.1.4 ! misho 4303: PCRE_SPTR32 subject; (32-bit version)
1.1.1.2 misho 4304: int subject_length;
4305: int start_match;
4306: int current_position;
4307: int capture_top;
4308: int capture_last;
4309: void *callout_data;
4310: int pattern_position;
4311: int next_item_length;
4312: const unsigned char *mark; (8-bit version)
4313: const PCRE_UCHAR16 *mark; (16-bit version)
1.1.1.4 ! misho 4314: const PCRE_UCHAR32 *mark; (32-bit version)
1.1 misho 4315:
1.1.1.4 ! misho 4316: The version field is an integer containing the version number of the
! 4317: block format. The initial version was 0; the current version is 2. The
! 4318: version number will change again in future if additional fields are
1.1 misho 4319: added, but the intention is never to remove any of the existing fields.
4320:
1.1.1.4 ! misho 4321: The callout_number field contains the number of the callout, as com-
! 4322: piled into the pattern (that is, the number after ?C for manual call-
1.1 misho 4323: outs, and 255 for automatically generated callouts).
4324:
1.1.1.4 ! misho 4325: The offset_vector field is a pointer to the vector of offsets that was
! 4326: passed by the caller to the matching function. When pcre_exec() or
! 4327: pcre[16|32]_exec() is used, the contents can be inspected, in order to
! 4328: extract substrings that have been matched so far, in the same way as
! 4329: for extracting substrings after a match has completed. For the DFA
1.1.1.2 misho 4330: matching functions, this field is not useful.
1.1 misho 4331:
4332: The subject and subject_length fields contain copies of the values that
1.1.1.2 misho 4333: were passed to the matching function.
1.1 misho 4334:
1.1.1.4 ! misho 4335: The start_match field normally contains the offset within the subject
! 4336: at which the current match attempt started. However, if the escape
! 4337: sequence \K has been encountered, this value is changed to reflect the
! 4338: modified starting point. If the pattern is not anchored, the callout
1.1 misho 4339: function may be called several times from the same point in the pattern
4340: for different starting points in the subject.
4341:
1.1.1.4 ! misho 4342: The current_position field contains the offset within the subject of
1.1 misho 4343: the current match pointer.
4344:
1.1.1.4 ! misho 4345: When the pcre_exec() or pcre[16|32]_exec() is used, the capture_top
! 4346: field contains one more than the number of the highest numbered cap-
! 4347: tured substring so far. If no substrings have been captured, the value
! 4348: of capture_top is one. This is always the case when the DFA functions
! 4349: are used, because they do not support captured substrings.
! 4350:
! 4351: The capture_last field contains the number of the most recently cap-
! 4352: tured substring. However, when a recursion exits, the value reverts to
! 4353: what it was outside the recursion, as do the values of all captured
! 4354: substrings. If no substrings have been captured, the value of cap-
! 4355: ture_last is -1. This is always the case for the DFA matching func-
! 4356: tions.
1.1 misho 4357:
1.1.1.2 misho 4358: The callout_data field contains a value that is passed to a matching
4359: function specifically so that it can be passed back in callouts. It is
1.1.1.4 ! misho 4360: passed in the callout_data field of a pcre_extra or pcre[16|32]_extra
! 4361: data structure. If no such data was passed, the value of callout_data
! 4362: in a callout block is NULL. There is a description of the pcre_extra
! 4363: structure in the pcreapi documentation.
1.1 misho 4364:
1.1.1.2 misho 4365: The pattern_position field is present from version 1 of the callout
4366: structure. It contains the offset to the next item to be matched in the
4367: pattern string.
4368:
4369: The next_item_length field is present from version 1 of the callout
4370: structure. It contains the length of the next item to be matched in the
4371: pattern string. When the callout immediately precedes an alternation
4372: bar, a closing parenthesis, or the end of the pattern, the length is
4373: zero. When the callout precedes an opening parenthesis, the length is
4374: that of the entire subpattern.
1.1 misho 4375:
4376: The pattern_position and next_item_length fields are intended to help
4377: in distinguishing between different automatic callouts, which all have
4378: the same callout number. However, they are set for all callouts.
4379:
1.1.1.2 misho 4380: The mark field is present from version 2 of the callout structure. In
1.1.1.4 ! misho 4381: callouts from pcre_exec() or pcre[16|32]_exec() it contains a pointer
! 4382: to the zero-terminated name of the most recently passed (*MARK),
! 4383: (*PRUNE), or (*THEN) item in the match, or NULL if no such items have
! 4384: been passed. Instances of (*PRUNE) or (*THEN) without a name do not
! 4385: obliterate a previous (*MARK). In callouts from the DFA matching func-
! 4386: tions this field always contains NULL.
1.1 misho 4387:
4388:
4389: RETURN VALUES
4390:
4391: The external callout function returns an integer to PCRE. If the value
4392: is zero, matching proceeds as normal. If the value is greater than
4393: zero, matching fails at the current point, but the testing of other
4394: matching possibilities goes ahead, just as if a lookahead assertion had
1.1.1.2 misho 4395: failed. If the value is less than zero, the match is abandoned, the
4396: matching function returns the negative value.
1.1 misho 4397:
4398: Negative values should normally be chosen from the set of
4399: PCRE_ERROR_xxx values. In particular, PCRE_ERROR_NOMATCH forces a stan-
4400: dard "no match" failure. The error number PCRE_ERROR_CALLOUT is
4401: reserved for use by callout functions; it will never be used by PCRE
4402: itself.
4403:
4404:
4405: AUTHOR
4406:
4407: Philip Hazel
4408: University Computing Service
4409: Cambridge CB2 3QH, England.
4410:
4411:
4412: REVISION
4413:
1.1.1.4 ! misho 4414: Last updated: 03 March 2013
! 4415: Copyright (c) 1997-2013 University of Cambridge.
1.1 misho 4416: ------------------------------------------------------------------------------
4417:
4418:
1.1.1.4 ! misho 4419: PCRECOMPAT(3) Library Functions Manual PCRECOMPAT(3)
! 4420:
1.1 misho 4421:
4422:
4423: NAME
4424: PCRE - Perl-compatible regular expressions
4425:
4426: DIFFERENCES BETWEEN PCRE AND PERL
4427:
4428: This document describes the differences in the ways that PCRE and Perl
4429: handle regular expressions. The differences described here are with
4430: respect to Perl versions 5.10 and above.
4431:
1.1.1.2 misho 4432: 1. PCRE has only a subset of Perl's Unicode support. Details of what it
4433: does have are given in the pcreunicode page.
1.1 misho 4434:
4435: 2. PCRE allows repeat quantifiers only on parenthesized assertions, but
4436: they do not mean what you might think. For example, (?!a){3} does not
4437: assert that the next three characters are not "a". It just asserts that
4438: the next character is not "a" three times (in principle: PCRE optimizes
4439: this to run the assertion just once). Perl allows repeat quantifiers on
4440: other assertions such as \b, but these do not seem to have any use.
4441:
4442: 3. Capturing subpatterns that occur inside negative lookahead asser-
4443: tions are counted, but their entries in the offsets vector are never
1.1.1.4 ! misho 4444: set. Perl sometimes (but not always) sets its numerical variables from
! 4445: inside negative assertions.
1.1 misho 4446:
4447: 4. Though binary zero characters are supported in the subject string,
4448: they are not allowed in a pattern string because it is passed as a nor-
4449: mal C string, terminated by zero. The escape sequence \0 can be used in
4450: the pattern to represent a binary zero.
4451:
4452: 5. The following Perl escape sequences are not supported: \l, \u, \L,
4453: \U, and \N when followed by a character name or Unicode value. (\N on
4454: its own, matching a non-newline character, is supported.) In fact these
4455: are implemented by Perl's general string-handling and are not part of
4456: its pattern matching engine. If any of these are encountered by PCRE,
4457: an error is generated by default. However, if the PCRE_JAVASCRIPT_COM-
4458: PAT option is set, \U and \u are interpreted as JavaScript interprets
4459: them.
4460:
4461: 6. The Perl escape sequences \p, \P, and \X are supported only if PCRE
4462: is built with Unicode character property support. The properties that
4463: can be tested with \p and \P are limited to the general category prop-
4464: erties such as Lu and Nd, script names such as Greek or Han, and the
4465: derived properties Any and L&. PCRE does support the Cs (surrogate)
4466: property, which Perl does not; the Perl documentation says "Because
4467: Perl hides the need for the user to understand the internal representa-
4468: tion of Unicode characters, there is no need to implement the somewhat
4469: messy concept of surrogates."
4470:
1.1.1.4 ! misho 4471: 7. PCRE does support the \Q...\E escape for quoting substrings. Charac-
! 4472: ters in between are treated as literals. This is slightly different
! 4473: from Perl in that $ and @ are also handled as literals inside the
! 4474: quotes. In Perl, they cause variable interpolation (but of course PCRE
1.1 misho 4475: does not have variables). Note the following examples:
4476:
4477: Pattern PCRE matches Perl matches
4478:
4479: \Qabc$xyz\E abc$xyz abc followed by the
4480: contents of $xyz
4481: \Qabc\$xyz\E abc\$xyz abc\$xyz
4482: \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4483:
1.1.1.4 ! misho 4484: The \Q...\E sequence is recognized both inside and outside character
1.1 misho 4485: classes.
4486:
1.1.1.4 ! misho 4487: 8. Fairly obviously, PCRE does not support the (?{code}) and (??{code})
! 4488: constructions. However, there is support for recursive patterns. This
! 4489: is not available in Perl 5.8, but it is in Perl 5.10. Also, the PCRE
! 4490: "callout" feature allows an external function to be called during pat-
1.1 misho 4491: tern matching. See the pcrecallout documentation for details.
4492:
1.1.1.4 ! misho 4493: 9. Subpatterns that are called as subroutines (whether or not recur-
! 4494: sively) are always treated as atomic groups in PCRE. This is like
! 4495: Python, but unlike Perl. Captured values that are set outside a sub-
! 4496: routine call can be reference from inside in PCRE, but not in Perl.
1.1 misho 4497: There is a discussion that explains these differences in more detail in
4498: the section on recursion differences from Perl in the pcrepattern page.
4499:
1.1.1.4 ! misho 4500: 10. If any of the backtracking control verbs are used in a subpattern
! 4501: that is called as a subroutine (whether or not recursively), their
! 4502: effect is confined to that subpattern; it does not extend to the sur-
! 4503: rounding pattern. This is not always the case in Perl. In particular,
! 4504: if (*THEN) is present in a group that is called as a subroutine, its
! 4505: action is limited to that group, even if the group does not contain any
! 4506: | characters. Note that such subpatterns are processed as anchored at
! 4507: the point where they are tested.
! 4508:
! 4509: 11. If a pattern contains more than one backtracking control verb, the
! 4510: first one that is backtracked onto acts. For example, in the pattern
! 4511: A(*COMMIT)B(*PRUNE)C a failure in B triggers (*COMMIT), but a failure
! 4512: in C triggers (*PRUNE). Perl's behaviour is more complex; in many cases
! 4513: it is the same as PCRE, but there are examples where it differs.
! 4514:
! 4515: 12. Most backtracking verbs in assertions have their normal actions.
! 4516: They are not confined to the assertion.
! 4517:
! 4518: 13. There are some differences that are concerned with the settings of
! 4519: captured strings when part of a pattern is repeated. For example,
! 4520: matching "aba" against the pattern /^(a(b)?)+$/ in Perl leaves $2
1.1 misho 4521: unset, but in PCRE it is set to "b".
4522:
1.1.1.4 ! misho 4523: 14. PCRE's handling of duplicate subpattern numbers and duplicate sub-
1.1 misho 4524: pattern names is not as general as Perl's. This is a consequence of the
4525: fact the PCRE works internally just with numbers, using an external ta-
1.1.1.4 ! misho 4526: ble to translate between numbers and names. In particular, a pattern
! 4527: such as (?|(?<a>A)|(?<b)B), where the two capturing parentheses have
! 4528: the same number but different names, is not supported, and causes an
! 4529: error at compile time. If it were allowed, it would not be possible to
! 4530: distinguish which parentheses matched, because both names map to cap-
1.1 misho 4531: turing subpattern number 1. To avoid this confusing situation, an error
4532: is given at compile time.
4533:
1.1.1.4 ! misho 4534: 15. Perl recognizes comments in some places that PCRE does not, for
! 4535: example, between the ( and ? at the start of a subpattern. If the /x
1.1.1.3 misho 4536: modifier is set, Perl allows white space between ( and ? but PCRE never
1.1 misho 4537: does, even if the PCRE_EXTENDED option is set.
4538:
1.1.1.4 ! misho 4539: 16. In PCRE, the upper/lower case character properties Lu and Ll are
! 4540: not affected when case-independent matching is specified. For example,
! 4541: \p{Lu} always matches an upper case letter. I think Perl has changed in
! 4542: this respect; in the release at the time of writing (5.16), \p{Lu} and
! 4543: \p{Ll} match all letters, regardless of case, when case independence is
! 4544: specified.
! 4545:
! 4546: 17. PCRE provides some extensions to the Perl regular expression facil-
1.1 misho 4547: ities. Perl 5.10 includes new features that are not in earlier ver-
4548: sions of Perl, some of which (such as named parentheses) have been in
4549: PCRE for some time. This list is with respect to Perl 5.10:
4550:
4551: (a) Although lookbehind assertions in PCRE must match fixed length
4552: strings, each alternative branch of a lookbehind assertion can match a
4553: different length of string. Perl requires them all to have the same
4554: length.
4555:
4556: (b) If PCRE_DOLLAR_ENDONLY is set and PCRE_MULTILINE is not set, the $
4557: meta-character matches only at the very end of the string.
4558:
4559: (c) If PCRE_EXTRA is set, a backslash followed by a letter with no spe-
4560: cial meaning is faulted. Otherwise, like Perl, the backslash is quietly
4561: ignored. (Perl can be made to issue a warning.)
4562:
4563: (d) If PCRE_UNGREEDY is set, the greediness of the repetition quanti-
4564: fiers is inverted, that is, by default they are not greedy, but if fol-
4565: lowed by a question mark they are.
4566:
4567: (e) PCRE_ANCHORED can be used at matching time to force a pattern to be
4568: tried only at the first matching position in the subject string.
4569:
4570: (f) The PCRE_NOTBOL, PCRE_NOTEOL, PCRE_NOTEMPTY, PCRE_NOTEMPTY_ATSTART,
4571: and PCRE_NO_AUTO_CAPTURE options for pcre_exec() have no Perl equiva-
4572: lents.
4573:
4574: (g) The \R escape sequence can be restricted to match only CR, LF, or
4575: CRLF by the PCRE_BSR_ANYCRLF option.
4576:
4577: (h) The callout facility is PCRE-specific.
4578:
4579: (i) The partial matching facility is PCRE-specific.
4580:
4581: (j) Patterns compiled by PCRE can be saved and re-used at a later time,
4582: even on different hosts that have the other endianness. However, this
4583: does not apply to optimized data created by the just-in-time compiler.
4584:
1.1.1.4 ! misho 4585: (k) The alternative matching functions (pcre_dfa_exec(),
! 4586: pcre16_dfa_exec() and pcre32_dfa_exec(),) match in a different way and
! 4587: are not Perl-compatible.
1.1 misho 4588:
1.1.1.2 misho 4589: (l) PCRE recognizes some special sequences such as (*CR) at the start
1.1 misho 4590: of a pattern that set overall options that cannot be changed within the
4591: pattern.
4592:
4593:
4594: AUTHOR
4595:
4596: Philip Hazel
4597: University Computing Service
4598: Cambridge CB2 3QH, England.
4599:
4600:
4601: REVISION
4602:
1.1.1.4 ! misho 4603: Last updated: 19 March 2013
! 4604: Copyright (c) 1997-2013 University of Cambridge.
1.1 misho 4605: ------------------------------------------------------------------------------
4606:
4607:
1.1.1.4 ! misho 4608: PCREPATTERN(3) Library Functions Manual PCREPATTERN(3)
! 4609:
1.1 misho 4610:
4611:
4612: NAME
4613: PCRE - Perl-compatible regular expressions
4614:
4615: PCRE REGULAR EXPRESSION DETAILS
4616:
4617: The syntax and semantics of the regular expressions that are supported
4618: by PCRE are described in detail below. There is a quick-reference syn-
4619: tax summary in the pcresyntax page. PCRE tries to match Perl syntax and
4620: semantics as closely as it can. PCRE also supports some alternative
4621: regular expression syntax (which does not conflict with the Perl syn-
4622: tax) in order to provide some compatibility with regular expressions in
4623: Python, .NET, and Oniguruma.
4624:
4625: Perl's regular expressions are described in its own documentation, and
4626: regular expressions in general are covered in a number of books, some
4627: of which have copious examples. Jeffrey Friedl's "Mastering Regular
4628: Expressions", published by O'Reilly, covers regular expressions in
4629: great detail. This description of PCRE's regular expressions is
4630: intended as reference material.
4631:
1.1.1.4 ! misho 4632: This document discusses the patterns that are supported by PCRE when
! 4633: one its main matching functions, pcre_exec() (8-bit) or
! 4634: pcre[16|32]_exec() (16- or 32-bit), is used. PCRE also has alternative
! 4635: matching functions, pcre_dfa_exec() and pcre[16|32_dfa_exec(), which
! 4636: match using a different algorithm that is not Perl-compatible. Some of
! 4637: the features discussed below are not available when DFA matching is
! 4638: used. The advantages and disadvantages of the alternative functions,
! 4639: and how they differ from the normal functions, are discussed in the
! 4640: pcrematching page.
! 4641:
! 4642:
! 4643: SPECIAL START-OF-PATTERN ITEMS
! 4644:
! 4645: A number of options that can be passed to pcre_compile() can also be
! 4646: set by special items at the start of a pattern. These are not Perl-com-
! 4647: patible, but are provided to make these options accessible to pattern
! 4648: writers who are not able to change the program that processes the pat-
! 4649: tern. Any number of these items may appear, but they must all be
! 4650: together right at the start of the pattern string, and the letters must
! 4651: be in upper case.
! 4652:
! 4653: UTF support
! 4654:
1.1 misho 4655: The original operation of PCRE was on strings of one-byte characters.
1.1.1.2 misho 4656: However, there is now also support for UTF-8 strings in the original
1.1.1.4 ! misho 4657: library, an extra library that supports 16-bit and UTF-16 character
! 4658: strings, and a third library that supports 32-bit and UTF-32 character
1.1.1.2 misho 4659: strings. To use these features, PCRE must be built to include appropri-
1.1.1.4 ! misho 4660: ate support. When using UTF strings you must either call the compiling
! 4661: function with the PCRE_UTF8, PCRE_UTF16, or PCRE_UTF32 option, or the
! 4662: pattern must start with one of these special sequences:
1.1 misho 4663:
4664: (*UTF8)
1.1.1.2 misho 4665: (*UTF16)
1.1.1.4 ! misho 4666: (*UTF32)
! 4667: (*UTF)
! 4668:
! 4669: (*UTF) is a generic sequence that can be used with any of the
! 4670: libraries. Starting a pattern with such a sequence is equivalent to
! 4671: setting the relevant option. How setting a UTF mode affects pattern
! 4672: matching is mentioned in several places below. There is also a summary
! 4673: of features in the pcreunicode page.
! 4674:
! 4675: Some applications that allow their users to supply patterns may wish to
! 4676: restrict them to non-UTF data for security reasons. If the
! 4677: PCRE_NEVER_UTF option is set at compile time, (*UTF) etc. are not
! 4678: allowed, and their appearance causes an error.
1.1 misho 4679:
1.1.1.4 ! misho 4680: Unicode property support
1.1 misho 4681:
1.1.1.4 ! misho 4682: Another special sequence that may appear at the start of a pattern is
1.1 misho 4683:
4684: (*UCP)
4685:
1.1.1.2 misho 4686: This has the same effect as setting the PCRE_UCP option: it causes
4687: sequences such as \d and \w to use Unicode properties to determine
1.1 misho 4688: character types, instead of recognizing only characters with codes less
4689: than 128 via a lookup table.
4690:
1.1.1.4 ! misho 4691: Disabling start-up optimizations
! 4692:
1.1.1.2 misho 4693: If a pattern starts with (*NO_START_OPT), it has the same effect as
1.1 misho 4694: setting the PCRE_NO_START_OPTIMIZE option either at compile or matching
1.1.1.4 ! misho 4695: time.
1.1 misho 4696:
1.1.1.4 ! misho 4697: Newline conventions
1.1 misho 4698:
1.1.1.4 ! misho 4699: PCRE supports five different conventions for indicating line breaks in
! 4700: strings: a single CR (carriage return) character, a single LF (line-
1.1 misho 4701: feed) character, the two-character sequence CRLF, any of the three pre-
1.1.1.4 ! misho 4702: ceding, or any Unicode newline sequence. The pcreapi page has further
! 4703: discussion about newlines, and shows how to set the newline convention
1.1 misho 4704: in the options arguments for the compiling and matching functions.
4705:
1.1.1.4 ! misho 4706: It is also possible to specify a newline convention by starting a pat-
1.1 misho 4707: tern string with one of the following five sequences:
4708:
4709: (*CR) carriage return
4710: (*LF) linefeed
4711: (*CRLF) carriage return, followed by linefeed
4712: (*ANYCRLF) any of the three above
4713: (*ANY) all Unicode newline sequences
4714:
1.1.1.2 misho 4715: These override the default and the options given to the compiling func-
1.1.1.4 ! misho 4716: tion. For example, on a Unix system where LF is the default newline
1.1.1.2 misho 4717: sequence, the pattern
1.1 misho 4718:
4719: (*CR)a.b
4720:
4721: changes the convention to CR. That pattern matches "a\nb" because LF is
1.1.1.4 ! misho 4722: no longer a newline. If more than one of these settings is present, the
! 4723: last one is used.
! 4724:
! 4725: The newline convention affects where the circumflex and dollar asser-
! 4726: tions are true. It also affects the interpretation of the dot metachar-
! 4727: acter when PCRE_DOTALL is not set, and the behaviour of \N. However, it
! 4728: does not affect what the \R escape sequence matches. By default, this
! 4729: is any Unicode newline sequence, for Perl compatibility. However, this
! 4730: can be changed; see the description of \R in the section entitled "New-
! 4731: line sequences" below. A change of \R setting can be combined with a
! 4732: change of newline convention.
! 4733:
! 4734: Setting match and recursion limits
! 4735:
! 4736: The caller of pcre_exec() can set a limit on the number of times the
! 4737: internal match() function is called and on the maximum depth of recur-
! 4738: sive calls. These facilities are provided to catch runaway matches that
! 4739: are provoked by patterns with huge matching trees (a typical example is
! 4740: a pattern with nested unlimited repeats) and to avoid running out of
! 4741: system stack by too much recursion. When one of these limits is
! 4742: reached, pcre_exec() gives an error return. The limits can also be set
! 4743: by items at the start of the pattern of the form
! 4744:
! 4745: (*LIMIT_MATCH=d)
! 4746: (*LIMIT_RECURSION=d)
! 4747:
! 4748: where d is any number of decimal digits. However, the value of the set-
! 4749: ting must be less than the value set by the caller of pcre_exec() for
! 4750: it to have any effect. In other words, the pattern writer can lower the
! 4751: limit set by the programmer, but not raise it. If there is more than
! 4752: one setting of one of these limits, the lower value is used.
! 4753:
! 4754:
! 4755: EBCDIC CHARACTER CODES
! 4756:
! 4757: PCRE can be compiled to run in an environment that uses EBCDIC as its
! 4758: character code rather than ASCII or Unicode (typically a mainframe sys-
! 4759: tem). In the sections below, character code values are ASCII or Uni-
! 4760: code; in an EBCDIC environment these characters may have different code
! 4761: values, and there are no code points greater than 255.
1.1 misho 4762:
4763:
4764: CHARACTERS AND METACHARACTERS
4765:
1.1.1.4 ! misho 4766: A regular expression is a pattern that is matched against a subject
! 4767: string from left to right. Most characters stand for themselves in a
! 4768: pattern, and match the corresponding characters in the subject. As a
1.1 misho 4769: trivial example, the pattern
4770:
4771: The quick brown fox
4772:
4773: matches a portion of a subject string that is identical to itself. When
1.1.1.4 ! misho 4774: caseless matching is specified (the PCRE_CASELESS option), letters are
! 4775: matched independently of case. In a UTF mode, PCRE always understands
! 4776: the concept of case for characters whose values are less than 128, so
! 4777: caseless matching is always possible. For characters with higher val-
! 4778: ues, the concept of case is supported if PCRE is compiled with Unicode
! 4779: property support, but not otherwise. If you want to use caseless
! 4780: matching for characters 128 and above, you must ensure that PCRE is
1.1.1.2 misho 4781: compiled with Unicode property support as well as with UTF support.
1.1 misho 4782:
1.1.1.4 ! misho 4783: The power of regular expressions comes from the ability to include
! 4784: alternatives and repetitions in the pattern. These are encoded in the
1.1 misho 4785: pattern by the use of metacharacters, which do not stand for themselves
4786: but instead are interpreted in some special way.
4787:
1.1.1.4 ! misho 4788: There are two different sets of metacharacters: those that are recog-
! 4789: nized anywhere in the pattern except within square brackets, and those
! 4790: that are recognized within square brackets. Outside square brackets,
1.1 misho 4791: the metacharacters are as follows:
4792:
4793: \ general escape character with several uses
4794: ^ assert start of string (or line, in multiline mode)
4795: $ assert end of string (or line, in multiline mode)
4796: . match any character except newline (by default)
4797: [ start character class definition
4798: | start of alternative branch
4799: ( start subpattern
4800: ) end subpattern
4801: ? extends the meaning of (
4802: also 0 or 1 quantifier
4803: also quantifier minimizer
4804: * 0 or more quantifier
4805: + 1 or more quantifier
4806: also "possessive quantifier"
4807: { start min/max quantifier
4808:
1.1.1.4 ! misho 4809: Part of a pattern that is in square brackets is called a "character
1.1 misho 4810: class". In a character class the only metacharacters are:
4811:
4812: \ general escape character
4813: ^ negate the class, but only if the first character
4814: - indicates character range
4815: [ POSIX character class (only if followed by POSIX
4816: syntax)
4817: ] terminates the character class
4818:
4819: The following sections describe the use of each of the metacharacters.
4820:
4821:
4822: BACKSLASH
4823:
4824: The backslash character has several uses. Firstly, if it is followed by
4825: a character that is not a number or a letter, it takes away any special
1.1.1.4 ! misho 4826: meaning that character may have. This use of backslash as an escape
1.1 misho 4827: character applies both inside and outside character classes.
4828:
1.1.1.4 ! misho 4829: For example, if you want to match a * character, you write \* in the
! 4830: pattern. This escaping action applies whether or not the following
! 4831: character would otherwise be interpreted as a metacharacter, so it is
! 4832: always safe to precede a non-alphanumeric with backslash to specify
! 4833: that it stands for itself. In particular, if you want to match a back-
1.1 misho 4834: slash, you write \\.
4835:
1.1.1.4 ! misho 4836: In a UTF mode, only ASCII numbers and letters have any special meaning
! 4837: after a backslash. All other characters (in particular, those whose
1.1 misho 4838: codepoints are greater than 127) are treated as literals.
4839:
1.1.1.4 ! misho 4840: If a pattern is compiled with the PCRE_EXTENDED option, white space in
! 4841: the pattern (other than in a character class) and characters between a
1.1 misho 4842: # outside a character class and the next newline are ignored. An escap-
1.1.1.4 ! misho 4843: ing backslash can be used to include a white space or # character as
1.1 misho 4844: part of the pattern.
4845:
1.1.1.4 ! misho 4846: If you want to remove the special meaning from a sequence of charac-
! 4847: ters, you can do so by putting them between \Q and \E. This is differ-
! 4848: ent from Perl in that $ and @ are handled as literals in \Q...\E
! 4849: sequences in PCRE, whereas in Perl, $ and @ cause variable interpola-
1.1 misho 4850: tion. Note the following examples:
4851:
4852: Pattern PCRE matches Perl matches
4853:
4854: \Qabc$xyz\E abc$xyz abc followed by the
4855: contents of $xyz
4856: \Qabc\$xyz\E abc\$xyz abc\$xyz
4857: \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4858:
1.1.1.4 ! misho 4859: The \Q...\E sequence is recognized both inside and outside character
! 4860: classes. An isolated \E that is not preceded by \Q is ignored. If \Q
! 4861: is not followed by \E later in the pattern, the literal interpretation
! 4862: continues to the end of the pattern (that is, \E is assumed at the
! 4863: end). If the isolated \Q is inside a character class, this causes an
1.1 misho 4864: error, because the character class is not terminated.
4865:
4866: Non-printing characters
4867:
4868: A second use of backslash provides a way of encoding non-printing char-
1.1.1.4 ! misho 4869: acters in patterns in a visible manner. There is no restriction on the
! 4870: appearance of non-printing characters, apart from the binary zero that
! 4871: terminates a pattern, but when a pattern is being prepared by text
! 4872: editing, it is often easier to use one of the following escape
1.1 misho 4873: sequences than the binary character it represents:
4874:
4875: \a alarm, that is, the BEL character (hex 07)
4876: \cx "control-x", where x is any ASCII character
4877: \e escape (hex 1B)
1.1.1.3 misho 4878: \f form feed (hex 0C)
1.1 misho 4879: \n linefeed (hex 0A)
4880: \r carriage return (hex 0D)
4881: \t tab (hex 09)
4882: \ddd character with octal code ddd, or back reference
4883: \xhh character with hex code hh
4884: \x{hhh..} character with hex code hhh.. (non-JavaScript mode)
4885: \uhhhh character with hex code hhhh (JavaScript mode only)
4886:
1.1.1.4 ! misho 4887: The precise effect of \cx on ASCII characters is as follows: if x is a
! 4888: lower case letter, it is converted to upper case. Then bit 6 of the
! 4889: character (hex 40) is inverted. Thus \cA to \cZ become hex 01 to hex 1A
! 4890: (A is 41, Z is 5A), but \c{ becomes hex 3B ({ is 7B), and \c; becomes
! 4891: hex 7B (; is 3B). If the data item (byte or 16-bit value) following \c
1.1 misho 4892: has a value greater than 127, a compile-time error occurs. This locks
1.1.1.4 ! misho 4893: out non-ASCII characters in all modes.
! 4894:
! 4895: The \c facility was designed for use with ASCII characters, but with
! 4896: the extension to Unicode it is even less useful than it once was. It
! 4897: is, however, recognized when PCRE is compiled in EBCDIC mode, where
! 4898: data items are always bytes. In this mode, all values are valid after
! 4899: \c. If the next character is a lower case letter, it is converted to
! 4900: upper case. Then the 0xc0 bits of the byte are inverted. Thus \cA
! 4901: becomes hex 01, as in ASCII (A is C1), but because the EBCDIC letters
! 4902: are disjoint, \cZ becomes hex 29 (Z is E9), and other characters also
! 4903: generate different values.
1.1 misho 4904:
4905: By default, after \x, from zero to two hexadecimal digits are read
4906: (letters can be in upper or lower case). Any number of hexadecimal dig-
1.1.1.2 misho 4907: its may appear between \x{ and }, but the character code is constrained
4908: as follows:
4909:
4910: 8-bit non-UTF mode less than 0x100
4911: 8-bit UTF-8 mode less than 0x10ffff and a valid codepoint
4912: 16-bit non-UTF mode less than 0x10000
4913: 16-bit UTF-16 mode less than 0x10ffff and a valid codepoint
1.1.1.4 ! misho 4914: 32-bit non-UTF mode less than 0x80000000
! 4915: 32-bit UTF-32 mode less than 0x10ffff and a valid codepoint
1.1 misho 4916:
1.1.1.2 misho 4917: Invalid Unicode codepoints are the range 0xd800 to 0xdfff (the so-
1.1.1.4 ! misho 4918: called "surrogate" codepoints), and 0xffef.
1.1.1.2 misho 4919:
4920: If characters other than hexadecimal digits appear between \x{ and },
1.1 misho 4921: or if there is no terminating }, this form of escape is not recognized.
1.1.1.2 misho 4922: Instead, the initial \x will be interpreted as a basic hexadecimal
4923: escape, with no following digits, giving a character whose value is
1.1 misho 4924: zero.
4925:
1.1.1.2 misho 4926: If the PCRE_JAVASCRIPT_COMPAT option is set, the interpretation of \x
4927: is as just described only when it is followed by two hexadecimal dig-
4928: its. Otherwise, it matches a literal "x" character. In JavaScript
1.1 misho 4929: mode, support for code points greater than 256 is provided by \u, which
1.1.1.2 misho 4930: must be followed by four hexadecimal digits; otherwise it matches a
1.1.1.3 misho 4931: literal "u" character. Character codes specified by \u in JavaScript
4932: mode are constrained in the same was as those specified by \x in non-
4933: JavaScript mode.
1.1 misho 4934:
4935: Characters whose value is less than 256 can be defined by either of the
1.1.1.2 misho 4936: two syntaxes for \x (or by \u in JavaScript mode). There is no differ-
1.1 misho 4937: ence in the way they are handled. For example, \xdc is exactly the same
4938: as \x{dc} (or \u00dc in JavaScript mode).
4939:
1.1.1.2 misho 4940: After \0 up to two further octal digits are read. If there are fewer
4941: than two digits, just those that are present are used. Thus the
1.1 misho 4942: sequence \0\x\07 specifies two binary zeros followed by a BEL character
1.1.1.2 misho 4943: (code value 7). Make sure you supply two digits after the initial zero
1.1 misho 4944: if the pattern character that follows is itself an octal digit.
4945:
4946: The handling of a backslash followed by a digit other than 0 is compli-
4947: cated. Outside a character class, PCRE reads it and any following dig-
1.1.1.2 misho 4948: its as a decimal number. If the number is less than 10, or if there
1.1 misho 4949: have been at least that many previous capturing left parentheses in the
1.1.1.2 misho 4950: expression, the entire sequence is taken as a back reference. A
4951: description of how this works is given later, following the discussion
1.1 misho 4952: of parenthesized subpatterns.
4953:
1.1.1.2 misho 4954: Inside a character class, or if the decimal number is greater than 9
4955: and there have not been that many capturing subpatterns, PCRE re-reads
1.1 misho 4956: up to three octal digits following the backslash, and uses them to gen-
1.1.1.2 misho 4957: erate a data character. Any subsequent digits stand for themselves. The
4958: value of the character is constrained in the same way as characters
4959: specified in hexadecimal. For example:
1.1 misho 4960:
1.1.1.4 ! misho 4961: \040 is another way of writing an ASCII space
1.1 misho 4962: \40 is the same, provided there are fewer than 40
4963: previous capturing subpatterns
4964: \7 is always a back reference
4965: \11 might be a back reference, or another way of
4966: writing a tab
4967: \011 is always a tab
4968: \0113 is a tab followed by the character "3"
4969: \113 might be a back reference, otherwise the
4970: character with octal code 113
4971: \377 might be a back reference, otherwise
1.1.1.2 misho 4972: the value 255 (decimal)
1.1 misho 4973: \81 is either a back reference, or a binary zero
4974: followed by the two characters "8" and "1"
4975:
4976: Note that octal values of 100 or greater must not be introduced by a
4977: leading zero, because no more than three octal digits are ever read.
4978:
4979: All the sequences that define a single character value can be used both
4980: inside and outside character classes. In addition, inside a character
4981: class, \b is interpreted as the backspace character (hex 08).
4982:
4983: \N is not allowed in a character class. \B, \R, and \X are not special
4984: inside a character class. Like other unrecognized escape sequences,
4985: they are treated as the literal characters "B", "R", and "X" by
4986: default, but cause an error if the PCRE_EXTRA option is set. Outside a
4987: character class, these sequences have different meanings.
4988:
4989: Unsupported escape sequences
4990:
4991: In Perl, the sequences \l, \L, \u, and \U are recognized by its string
4992: handler and used to modify the case of following characters. By
4993: default, PCRE does not support these escape sequences. However, if the
4994: PCRE_JAVASCRIPT_COMPAT option is set, \U matches a "U" character, and
4995: \u can be used to define a character by code point, as described in the
4996: previous section.
4997:
4998: Absolute and relative back references
4999:
5000: The sequence \g followed by an unsigned or a negative number, option-
5001: ally enclosed in braces, is an absolute or relative back reference. A
5002: named back reference can be coded as \g{name}. Back references are dis-
5003: cussed later, following the discussion of parenthesized subpatterns.
5004:
5005: Absolute and relative subroutine calls
5006:
5007: For compatibility with Oniguruma, the non-Perl syntax \g followed by a
5008: name or a number enclosed either in angle brackets or single quotes, is
5009: an alternative syntax for referencing a subpattern as a "subroutine".
5010: Details are discussed later. Note that \g{...} (Perl syntax) and
5011: \g<...> (Oniguruma syntax) are not synonymous. The former is a back
5012: reference; the latter is a subroutine call.
5013:
5014: Generic character types
5015:
5016: Another use of backslash is for specifying generic character types:
5017:
5018: \d any decimal digit
5019: \D any character that is not a decimal digit
1.1.1.3 misho 5020: \h any horizontal white space character
5021: \H any character that is not a horizontal white space character
5022: \s any white space character
5023: \S any character that is not a white space character
5024: \v any vertical white space character
5025: \V any character that is not a vertical white space character
1.1 misho 5026: \w any "word" character
5027: \W any "non-word" character
5028:
5029: There is also the single sequence \N, which matches a non-newline char-
5030: acter. This is the same as the "." metacharacter when PCRE_DOTALL is
5031: not set. Perl also uses \N to match characters by name; PCRE does not
5032: support this.
5033:
5034: Each pair of lower and upper case escape sequences partitions the com-
5035: plete set of characters into two disjoint sets. Any given character
5036: matches one, and only one, of each pair. The sequences can appear both
5037: inside and outside character classes. They each match one character of
5038: the appropriate type. If the current matching point is at the end of
5039: the subject string, all of them fail, because there is no character to
5040: match.
5041:
5042: For compatibility with Perl, \s does not match the VT character (code
5043: 11). This makes it different from the the POSIX "space" class. The \s
5044: characters are HT (9), LF (10), FF (12), CR (13), and space (32). If
5045: "use locale;" is included in a Perl script, \s may match the VT charac-
5046: ter. In PCRE, it never does.
5047:
5048: A "word" character is an underscore or any character that is a letter
5049: or digit. By default, the definition of letters and digits is con-
5050: trolled by PCRE's low-valued character tables, and may vary if locale-
5051: specific matching is taking place (see "Locale support" in the pcreapi
5052: page). For example, in a French locale such as "fr_FR" in Unix-like
5053: systems, or "french" in Windows, some character codes greater than 128
5054: are used for accented letters, and these are then matched by \w. The
5055: use of locales with Unicode is discouraged.
5056:
1.1.1.2 misho 5057: By default, in a UTF mode, characters with values greater than 128
1.1 misho 5058: never match \d, \s, or \w, and always match \D, \S, and \W. These
1.1.1.2 misho 5059: sequences retain their original meanings from before UTF support was
1.1 misho 5060: available, mainly for efficiency reasons. However, if PCRE is compiled
5061: with Unicode property support, and the PCRE_UCP option is set, the be-
5062: haviour is changed so that Unicode properties are used to determine
5063: character types, as follows:
5064:
5065: \d any character that \p{Nd} matches (decimal digit)
5066: \s any character that \p{Z} matches, plus HT, LF, FF, CR
5067: \w any character that \p{L} or \p{N} matches, plus underscore
5068:
5069: The upper case escapes match the inverse sets of characters. Note that
5070: \d matches only decimal digits, whereas \w matches any Unicode digit,
5071: as well as any Unicode letter, and underscore. Note also that PCRE_UCP
5072: affects \b, and \B because they are defined in terms of \w and \W.
5073: Matching these sequences is noticeably slower when PCRE_UCP is set.
5074:
5075: The sequences \h, \H, \v, and \V are features that were added to Perl
5076: at release 5.10. In contrast to the other sequences, which match only
5077: ASCII characters by default, these always match certain high-valued
1.1.1.2 misho 5078: codepoints, whether or not PCRE_UCP is set. The horizontal space char-
5079: acters are:
1.1 misho 5080:
1.1.1.4 ! misho 5081: U+0009 Horizontal tab (HT)
1.1 misho 5082: U+0020 Space
5083: U+00A0 Non-break space
5084: U+1680 Ogham space mark
5085: U+180E Mongolian vowel separator
5086: U+2000 En quad
5087: U+2001 Em quad
5088: U+2002 En space
5089: U+2003 Em space
5090: U+2004 Three-per-em space
5091: U+2005 Four-per-em space
5092: U+2006 Six-per-em space
5093: U+2007 Figure space
5094: U+2008 Punctuation space
5095: U+2009 Thin space
5096: U+200A Hair space
5097: U+202F Narrow no-break space
5098: U+205F Medium mathematical space
5099: U+3000 Ideographic space
5100:
5101: The vertical space characters are:
5102:
1.1.1.4 ! misho 5103: U+000A Linefeed (LF)
! 5104: U+000B Vertical tab (VT)
! 5105: U+000C Form feed (FF)
! 5106: U+000D Carriage return (CR)
! 5107: U+0085 Next line (NEL)
1.1 misho 5108: U+2028 Line separator
5109: U+2029 Paragraph separator
5110:
1.1.1.2 misho 5111: In 8-bit, non-UTF-8 mode, only the characters with codepoints less than
5112: 256 are relevant.
5113:
1.1 misho 5114: Newline sequences
5115:
1.1.1.2 misho 5116: Outside a character class, by default, the escape sequence \R matches
5117: any Unicode newline sequence. In 8-bit non-UTF-8 mode \R is equivalent
5118: to the following:
1.1 misho 5119:
5120: (?>\r\n|\n|\x0b|\f|\r|\x85)
5121:
1.1.1.2 misho 5122: This is an example of an "atomic group", details of which are given
1.1 misho 5123: below. This particular group matches either the two-character sequence
1.1.1.2 misho 5124: CR followed by LF, or one of the single characters LF (linefeed,
1.1.1.3 misho 5125: U+000A), VT (vertical tab, U+000B), FF (form feed, U+000C), CR (car-
5126: riage return, U+000D), or NEL (next line, U+0085). The two-character
5127: sequence is treated as a single unit that cannot be split.
1.1 misho 5128:
1.1.1.2 misho 5129: In other modes, two additional characters whose codepoints are greater
1.1 misho 5130: than 255 are added: LS (line separator, U+2028) and PS (paragraph sepa-
1.1.1.2 misho 5131: rator, U+2029). Unicode character property support is not needed for
1.1 misho 5132: these characters to be recognized.
5133:
5134: It is possible to restrict \R to match only CR, LF, or CRLF (instead of
1.1.1.2 misho 5135: the complete set of Unicode line endings) by setting the option
1.1 misho 5136: PCRE_BSR_ANYCRLF either at compile time or when the pattern is matched.
5137: (BSR is an abbrevation for "backslash R".) This can be made the default
1.1.1.2 misho 5138: when PCRE is built; if this is the case, the other behaviour can be
5139: requested via the PCRE_BSR_UNICODE option. It is also possible to
5140: specify these settings by starting a pattern string with one of the
1.1 misho 5141: following sequences:
5142:
5143: (*BSR_ANYCRLF) CR, LF, or CRLF only
5144: (*BSR_UNICODE) any Unicode newline sequence
5145:
1.1.1.2 misho 5146: These override the default and the options given to the compiling func-
5147: tion, but they can themselves be overridden by options given to a
5148: matching function. Note that these special settings, which are not
5149: Perl-compatible, are recognized only at the very start of a pattern,
5150: and that they must be in upper case. If more than one of them is
5151: present, the last one is used. They can be combined with a change of
1.1 misho 5152: newline convention; for example, a pattern can start with:
5153:
5154: (*ANY)(*BSR_ANYCRLF)
5155:
1.1.1.4 ! misho 5156: They can also be combined with the (*UTF8), (*UTF16), (*UTF32), (*UTF)
! 5157: or (*UCP) special sequences. Inside a character class, \R is treated as
! 5158: an unrecognized escape sequence, and so matches the letter "R" by
! 5159: default, but causes an error if PCRE_EXTRA is set.
1.1 misho 5160:
5161: Unicode character properties
5162:
5163: When PCRE is built with Unicode character property support, three addi-
1.1.1.2 misho 5164: tional escape sequences that match characters with specific properties
5165: are available. When in 8-bit non-UTF-8 mode, these sequences are of
5166: course limited to testing characters whose codepoints are less than
5167: 256, but they do work in this mode. The extra escape sequences are:
1.1 misho 5168:
5169: \p{xx} a character with the xx property
5170: \P{xx} a character without the xx property
1.1.1.4 ! misho 5171: \X a Unicode extended grapheme cluster
1.1 misho 5172:
1.1.1.2 misho 5173: The property names represented by xx above are limited to the Unicode
1.1 misho 5174: script names, the general category properties, "Any", which matches any
1.1.1.2 misho 5175: character (including newline), and some special PCRE properties
5176: (described in the next section). Other Perl properties such as "InMu-
5177: sicalSymbols" are not currently supported by PCRE. Note that \P{Any}
1.1 misho 5178: does not match any characters, so always causes a match failure.
5179:
5180: Sets of Unicode characters are defined as belonging to certain scripts.
1.1.1.2 misho 5181: A character from one of these sets can be matched using a script name.
1.1 misho 5182: For example:
5183:
5184: \p{Greek}
5185: \P{Han}
5186:
1.1.1.2 misho 5187: Those that are not part of an identified script are lumped together as
1.1 misho 5188: "Common". The current list of scripts is:
5189:
1.1.1.3 misho 5190: Arabic, Armenian, Avestan, Balinese, Bamum, Batak, Bengali, Bopomofo,
5191: Brahmi, Braille, Buginese, Buhid, Canadian_Aboriginal, Carian, Chakma,
5192: Cham, Cherokee, Common, Coptic, Cuneiform, Cypriot, Cyrillic, Deseret,
5193: Devanagari, Egyptian_Hieroglyphs, Ethiopic, Georgian, Glagolitic,
5194: Gothic, Greek, Gujarati, Gurmukhi, Han, Hangul, Hanunoo, Hebrew, Hira-
5195: gana, Imperial_Aramaic, Inherited, Inscriptional_Pahlavi, Inscrip-
5196: tional_Parthian, Javanese, Kaithi, Kannada, Katakana, Kayah_Li,
5197: Kharoshthi, Khmer, Lao, Latin, Lepcha, Limbu, Linear_B, Lisu, Lycian,
5198: Lydian, Malayalam, Mandaic, Meetei_Mayek, Meroitic_Cursive,
5199: Meroitic_Hieroglyphs, Miao, Mongolian, Myanmar, New_Tai_Lue, Nko,
5200: Ogham, Old_Italic, Old_Persian, Old_South_Arabian, Old_Turkic,
5201: Ol_Chiki, Oriya, Osmanya, Phags_Pa, Phoenician, Rejang, Runic, Samari-
5202: tan, Saurashtra, Sharada, Shavian, Sinhala, Sora_Sompeng, Sundanese,
5203: Syloti_Nagri, Syriac, Tagalog, Tagbanwa, Tai_Le, Tai_Tham, Tai_Viet,
5204: Takri, Tamil, Telugu, Thaana, Thai, Tibetan, Tifinagh, Ugaritic, Vai,
5205: Yi.
1.1 misho 5206:
5207: Each character has exactly one Unicode general category property, spec-
1.1.1.2 misho 5208: ified by a two-letter abbreviation. For compatibility with Perl, nega-
5209: tion can be specified by including a circumflex between the opening
5210: brace and the property name. For example, \p{^Lu} is the same as
1.1 misho 5211: \P{Lu}.
5212:
5213: If only one letter is specified with \p or \P, it includes all the gen-
1.1.1.2 misho 5214: eral category properties that start with that letter. In this case, in
5215: the absence of negation, the curly brackets in the escape sequence are
1.1 misho 5216: optional; these two examples have the same effect:
5217:
5218: \p{L}
5219: \pL
5220:
5221: The following general category property codes are supported:
5222:
5223: C Other
5224: Cc Control
5225: Cf Format
5226: Cn Unassigned
5227: Co Private use
5228: Cs Surrogate
5229:
5230: L Letter
5231: Ll Lower case letter
5232: Lm Modifier letter
5233: Lo Other letter
5234: Lt Title case letter
5235: Lu Upper case letter
5236:
5237: M Mark
5238: Mc Spacing mark
5239: Me Enclosing mark
5240: Mn Non-spacing mark
5241:
5242: N Number
5243: Nd Decimal number
5244: Nl Letter number
5245: No Other number
5246:
5247: P Punctuation
5248: Pc Connector punctuation
5249: Pd Dash punctuation
5250: Pe Close punctuation
5251: Pf Final punctuation
5252: Pi Initial punctuation
5253: Po Other punctuation
5254: Ps Open punctuation
5255:
5256: S Symbol
5257: Sc Currency symbol
5258: Sk Modifier symbol
5259: Sm Mathematical symbol
5260: So Other symbol
5261:
5262: Z Separator
5263: Zl Line separator
5264: Zp Paragraph separator
5265: Zs Space separator
5266:
1.1.1.2 misho 5267: The special property L& is also supported: it matches a character that
5268: has the Lu, Ll, or Lt property, in other words, a letter that is not
1.1 misho 5269: classified as a modifier or "other".
5270:
1.1.1.2 misho 5271: The Cs (Surrogate) property applies only to characters in the range
5272: U+D800 to U+DFFF. Such characters are not valid in Unicode strings and
5273: so cannot be tested by PCRE, unless UTF validity checking has been
1.1.1.4 ! misho 5274: turned off (see the discussion of PCRE_NO_UTF8_CHECK,
! 5275: PCRE_NO_UTF16_CHECK and PCRE_NO_UTF32_CHECK in the pcreapi page). Perl
! 5276: does not support the Cs property.
1.1 misho 5277:
5278: The long synonyms for property names that Perl supports (such as
5279: \p{Letter}) are not supported by PCRE, nor is it permitted to prefix
5280: any of these properties with "Is".
5281:
5282: No character that is in the Unicode table has the Cn (unassigned) prop-
5283: erty. Instead, this property is assumed for any code point that is not
5284: in the Unicode table.
5285:
5286: Specifying caseless matching does not affect these escape sequences.
1.1.1.4 ! misho 5287: For example, \p{Lu} always matches only upper case letters. This is
! 5288: different from the behaviour of current versions of Perl.
! 5289:
! 5290: Matching characters by Unicode property is not fast, because PCRE has
! 5291: to do a multistage table lookup in order to find a character's prop-
! 5292: erty. That is why the traditional escape sequences such as \d and \w do
! 5293: not use Unicode properties in PCRE by default, though you can make them
! 5294: do so by setting the PCRE_UCP option or by starting the pattern with
! 5295: (*UCP).
! 5296:
! 5297: Extended grapheme clusters
1.1 misho 5298:
5299: The \X escape matches any number of Unicode characters that form an
1.1.1.4 ! misho 5300: "extended grapheme cluster", and treats the sequence as an atomic group
! 5301: (see below). Up to and including release 8.31, PCRE matched an ear-
! 5302: lier, simpler definition that was equivalent to
1.1 misho 5303:
5304: (?>\PM\pM*)
5305:
1.1.1.4 ! misho 5306: That is, it matched a character without the "mark" property, followed
! 5307: by zero or more characters with the "mark" property. Characters with
! 5308: the "mark" property are typically non-spacing accents that affect the
! 5309: preceding character.
! 5310:
! 5311: This simple definition was extended in Unicode to include more compli-
! 5312: cated kinds of composite character by giving each character a grapheme
! 5313: breaking property, and creating rules that use these properties to
! 5314: define the boundaries of extended grapheme clusters. In releases of
! 5315: PCRE later than 8.31, \X matches one of these clusters.
! 5316:
! 5317: \X always matches at least one character. Then it decides whether to
! 5318: add additional characters according to the following rules for ending a
! 5319: cluster:
! 5320:
! 5321: 1. End at the end of the subject string.
! 5322:
! 5323: 2. Do not end between CR and LF; otherwise end after any control char-
! 5324: acter.
! 5325:
! 5326: 3. Do not break Hangul (a Korean script) syllable sequences. Hangul
! 5327: characters are of five types: L, V, T, LV, and LVT. An L character may
! 5328: be followed by an L, V, LV, or LVT character; an LV or V character may
! 5329: be followed by a V or T character; an LVT or T character may be follwed
! 5330: only by a T character.
! 5331:
! 5332: 4. Do not end before extending characters or spacing marks. Characters
! 5333: with the "mark" property always have the "extend" grapheme breaking
! 5334: property.
! 5335:
! 5336: 5. Do not end after prepend characters.
! 5337:
! 5338: 6. Otherwise, end the cluster.
1.1 misho 5339:
5340: PCRE's additional properties
5341:
1.1.1.4 ! misho 5342: As well as the standard Unicode properties described above, PCRE sup-
! 5343: ports four more that make it possible to convert traditional escape
! 5344: sequences such as \w and \s and POSIX character classes to use Unicode
! 5345: properties. PCRE uses these non-standard, non-Perl properties inter-
! 5346: nally when PCRE_UCP is set. However, they may also be used explicitly.
! 5347: These properties are:
1.1 misho 5348:
5349: Xan Any alphanumeric character
5350: Xps Any POSIX space character
5351: Xsp Any Perl space character
5352: Xwd Any Perl "word" character
5353:
1.1.1.4 ! misho 5354: Xan matches characters that have either the L (letter) or the N (num-
! 5355: ber) property. Xps matches the characters tab, linefeed, vertical tab,
! 5356: form feed, or carriage return, and any other character that has the Z
1.1 misho 5357: (separator) property. Xsp is the same as Xps, except that vertical tab
5358: is excluded. Xwd matches the same characters as Xan, plus underscore.
5359:
1.1.1.4 ! misho 5360: There is another non-standard property, Xuc, which matches any charac-
! 5361: ter that can be represented by a Universal Character Name in C++ and
! 5362: other programming languages. These are the characters $, @, ` (grave
! 5363: accent), and all characters with Unicode code points greater than or
! 5364: equal to U+00A0, except for the surrogates U+D800 to U+DFFF. Note that
! 5365: most base (ASCII) characters are excluded. (Universal Character Names
! 5366: are of the form \uHHHH or \UHHHHHHHH where H is a hexadecimal digit.
! 5367: Note that the Xuc property does not match these sequences but the char-
! 5368: acters that they represent.)
! 5369:
1.1 misho 5370: Resetting the match start
5371:
1.1.1.4 ! misho 5372: The escape sequence \K causes any previously matched characters not to
1.1 misho 5373: be included in the final matched sequence. For example, the pattern:
5374:
5375: foo\Kbar
5376:
1.1.1.4 ! misho 5377: matches "foobar", but reports that it has matched "bar". This feature
! 5378: is similar to a lookbehind assertion (described below). However, in
! 5379: this case, the part of the subject before the real match does not have
! 5380: to be of fixed length, as lookbehind assertions do. The use of \K does
! 5381: not interfere with the setting of captured substrings. For example,
1.1 misho 5382: when the pattern
5383:
5384: (foo)\Kbar
5385:
5386: matches "foobar", the first substring is still set to "foo".
5387:
1.1.1.4 ! misho 5388: Perl documents that the use of \K within assertions is "not well
! 5389: defined". In PCRE, \K is acted upon when it occurs inside positive
1.1 misho 5390: assertions, but is ignored in negative assertions.
5391:
5392: Simple assertions
5393:
1.1.1.4 ! misho 5394: The final use of backslash is for certain simple assertions. An asser-
! 5395: tion specifies a condition that has to be met at a particular point in
! 5396: a match, without consuming any characters from the subject string. The
! 5397: use of subpatterns for more complicated assertions is described below.
1.1 misho 5398: The backslashed assertions are:
5399:
5400: \b matches at a word boundary
5401: \B matches when not at a word boundary
5402: \A matches at the start of the subject
5403: \Z matches at the end of the subject
5404: also matches before a newline at the end of the subject
5405: \z matches only at the end of the subject
5406: \G matches at the first matching position in the subject
5407:
1.1.1.4 ! misho 5408: Inside a character class, \b has a different meaning; it matches the
! 5409: backspace character. If any other of these assertions appears in a
! 5410: character class, by default it matches the corresponding literal char-
1.1 misho 5411: acter (for example, \B matches the letter B). However, if the
1.1.1.4 ! misho 5412: PCRE_EXTRA option is set, an "invalid escape sequence" error is gener-
1.1 misho 5413: ated instead.
5414:
1.1.1.4 ! misho 5415: A word boundary is a position in the subject string where the current
! 5416: character and the previous character do not both match \w or \W (i.e.
! 5417: one matches \w and the other matches \W), or the start or end of the
! 5418: string if the first or last character matches \w, respectively. In a
! 5419: UTF mode, the meanings of \w and \W can be changed by setting the
! 5420: PCRE_UCP option. When this is done, it also affects \b and \B. Neither
! 5421: PCRE nor Perl has a separate "start of word" or "end of word" metase-
! 5422: quence. However, whatever follows \b normally determines which it is.
1.1 misho 5423: For example, the fragment \ba matches "a" at the start of a word.
5424:
1.1.1.4 ! misho 5425: The \A, \Z, and \z assertions differ from the traditional circumflex
1.1 misho 5426: and dollar (described in the next section) in that they only ever match
1.1.1.4 ! misho 5427: at the very start and end of the subject string, whatever options are
! 5428: set. Thus, they are independent of multiline mode. These three asser-
1.1 misho 5429: tions are not affected by the PCRE_NOTBOL or PCRE_NOTEOL options, which
1.1.1.4 ! misho 5430: affect only the behaviour of the circumflex and dollar metacharacters.
! 5431: However, if the startoffset argument of pcre_exec() is non-zero, indi-
1.1 misho 5432: cating that matching is to start at a point other than the beginning of
1.1.1.4 ! misho 5433: the subject, \A can never match. The difference between \Z and \z is
1.1 misho 5434: that \Z matches before a newline at the end of the string as well as at
5435: the very end, whereas \z matches only at the end.
5436:
1.1.1.4 ! misho 5437: The \G assertion is true only when the current matching position is at
! 5438: the start point of the match, as specified by the startoffset argument
! 5439: of pcre_exec(). It differs from \A when the value of startoffset is
! 5440: non-zero. By calling pcre_exec() multiple times with appropriate argu-
1.1 misho 5441: ments, you can mimic Perl's /g option, and it is in this kind of imple-
5442: mentation where \G can be useful.
5443:
1.1.1.4 ! misho 5444: Note, however, that PCRE's interpretation of \G, as the start of the
1.1 misho 5445: current match, is subtly different from Perl's, which defines it as the
1.1.1.4 ! misho 5446: end of the previous match. In Perl, these can be different when the
! 5447: previously matched string was empty. Because PCRE does just one match
1.1 misho 5448: at a time, it cannot reproduce this behaviour.
5449:
1.1.1.4 ! misho 5450: If all the alternatives of a pattern begin with \G, the expression is
1.1 misho 5451: anchored to the starting match position, and the "anchored" flag is set
5452: in the compiled regular expression.
5453:
5454:
5455: CIRCUMFLEX AND DOLLAR
5456:
1.1.1.4 ! misho 5457: The circumflex and dollar metacharacters are zero-width assertions.
! 5458: That is, they test for a particular condition being true without con-
! 5459: suming any characters from the subject string.
! 5460:
1.1 misho 5461: Outside a character class, in the default matching mode, the circumflex
1.1.1.4 ! misho 5462: character is an assertion that is true only if the current matching
! 5463: point is at the start of the subject string. If the startoffset argu-
! 5464: ment of pcre_exec() is non-zero, circumflex can never match if the
! 5465: PCRE_MULTILINE option is unset. Inside a character class, circumflex
1.1 misho 5466: has an entirely different meaning (see below).
5467:
1.1.1.4 ! misho 5468: Circumflex need not be the first character of the pattern if a number
! 5469: of alternatives are involved, but it should be the first thing in each
! 5470: alternative in which it appears if the pattern is ever to match that
! 5471: branch. If all possible alternatives start with a circumflex, that is,
! 5472: if the pattern is constrained to match only at the start of the sub-
! 5473: ject, it is said to be an "anchored" pattern. (There are also other
1.1 misho 5474: constructs that can cause a pattern to be anchored.)
5475:
1.1.1.4 ! misho 5476: The dollar character is an assertion that is true only if the current
! 5477: matching point is at the end of the subject string, or immediately
! 5478: before a newline at the end of the string (by default). Note, however,
! 5479: that it does not actually match the newline. Dollar need not be the
! 5480: last character of the pattern if a number of alternatives are involved,
! 5481: but it should be the last item in any branch in which it appears. Dol-
! 5482: lar has no special meaning in a character class.
1.1 misho 5483:
5484: The meaning of dollar can be changed so that it matches only at the
5485: very end of the string, by setting the PCRE_DOLLAR_ENDONLY option at
5486: compile time. This does not affect the \Z assertion.
5487:
5488: The meanings of the circumflex and dollar characters are changed if the
5489: PCRE_MULTILINE option is set. When this is the case, a circumflex
5490: matches immediately after internal newlines as well as at the start of
5491: the subject string. It does not match after a newline that ends the
5492: string. A dollar matches before any newlines in the string, as well as
5493: at the very end, when PCRE_MULTILINE is set. When newline is specified
5494: as the two-character sequence CRLF, isolated CR and LF characters do
5495: not indicate newlines.
5496:
5497: For example, the pattern /^abc$/ matches the subject string "def\nabc"
5498: (where \n represents a newline) in multiline mode, but not otherwise.
5499: Consequently, patterns that are anchored in single line mode because
5500: all branches start with ^ are not anchored in multiline mode, and a
5501: match for circumflex is possible when the startoffset argument of
5502: pcre_exec() is non-zero. The PCRE_DOLLAR_ENDONLY option is ignored if
5503: PCRE_MULTILINE is set.
5504:
5505: Note that the sequences \A, \Z, and \z can be used to match the start
5506: and end of the subject in both modes, and if all branches of a pattern
5507: start with \A it is always anchored, whether or not PCRE_MULTILINE is
5508: set.
5509:
5510:
5511: FULL STOP (PERIOD, DOT) AND \N
5512:
5513: Outside a character class, a dot in the pattern matches any one charac-
5514: ter in the subject string except (by default) a character that signi-
1.1.1.2 misho 5515: fies the end of a line.
1.1 misho 5516:
1.1.1.2 misho 5517: When a line ending is defined as a single character, dot never matches
5518: that character; when the two-character sequence CRLF is used, dot does
5519: not match CR if it is immediately followed by LF, but otherwise it
5520: matches all characters (including isolated CRs and LFs). When any Uni-
5521: code line endings are being recognized, dot does not match CR or LF or
1.1 misho 5522: any of the other line ending characters.
5523:
1.1.1.2 misho 5524: The behaviour of dot with regard to newlines can be changed. If the
5525: PCRE_DOTALL option is set, a dot matches any one character, without
1.1 misho 5526: exception. If the two-character sequence CRLF is present in the subject
5527: string, it takes two dots to match it.
5528:
1.1.1.2 misho 5529: The handling of dot is entirely independent of the handling of circum-
5530: flex and dollar, the only relationship being that they both involve
1.1 misho 5531: newlines. Dot has no special meaning in a character class.
5532:
1.1.1.2 misho 5533: The escape sequence \N behaves like a dot, except that it is not
5534: affected by the PCRE_DOTALL option. In other words, it matches any
5535: character except one that signifies the end of a line. Perl also uses
1.1 misho 5536: \N to match characters by name; PCRE does not support this.
5537:
5538:
1.1.1.2 misho 5539: MATCHING A SINGLE DATA UNIT
1.1 misho 5540:
1.1.1.2 misho 5541: Outside a character class, the escape sequence \C matches any one data
5542: unit, whether or not a UTF mode is set. In the 8-bit library, one data
1.1.1.4 ! misho 5543: unit is one byte; in the 16-bit library it is a 16-bit unit; in the
! 5544: 32-bit library it is a 32-bit unit. Unlike a dot, \C always matches
! 5545: line-ending characters. The feature is provided in Perl in order to
! 5546: match individual bytes in UTF-8 mode, but it is unclear how it can use-
! 5547: fully be used. Because \C breaks up characters into individual data
! 5548: units, matching one unit with \C in a UTF mode means that the rest of
! 5549: the string may start with a malformed UTF character. This has undefined
! 5550: results, because PCRE assumes that it is dealing with valid UTF strings
! 5551: (and by default it checks this at the start of processing unless the
! 5552: PCRE_NO_UTF8_CHECK, PCRE_NO_UTF16_CHECK or PCRE_NO_UTF32_CHECK option
! 5553: is used).
1.1 misho 5554:
1.1.1.4 ! misho 5555: PCRE does not allow \C to appear in lookbehind assertions (described
! 5556: below) in a UTF mode, because this would make it impossible to calcu-
1.1 misho 5557: late the length of the lookbehind.
5558:
1.1.1.2 misho 5559: In general, the \C escape sequence is best avoided. However, one way of
1.1.1.4 ! misho 5560: using it that avoids the problem of malformed UTF characters is to use
! 5561: a lookahead to check the length of the next character, as in this pat-
! 5562: tern, which could be used with a UTF-8 string (ignore white space and
1.1.1.2 misho 5563: line breaks):
1.1 misho 5564:
5565: (?| (?=[\x00-\x7f])(\C) |
5566: (?=[\x80-\x{7ff}])(\C)(\C) |
5567: (?=[\x{800}-\x{ffff}])(\C)(\C)(\C) |
5568: (?=[\x{10000}-\x{1fffff}])(\C)(\C)(\C)(\C))
5569:
1.1.1.4 ! misho 5570: A group that starts with (?| resets the capturing parentheses numbers
! 5571: in each alternative (see "Duplicate Subpattern Numbers" below). The
! 5572: assertions at the start of each branch check the next UTF-8 character
! 5573: for values whose encoding uses 1, 2, 3, or 4 bytes, respectively. The
! 5574: character's individual bytes are then captured by the appropriate num-
1.1 misho 5575: ber of groups.
5576:
5577:
5578: SQUARE BRACKETS AND CHARACTER CLASSES
5579:
5580: An opening square bracket introduces a character class, terminated by a
5581: closing square bracket. A closing square bracket on its own is not spe-
5582: cial by default. However, if the PCRE_JAVASCRIPT_COMPAT option is set,
5583: a lone closing square bracket causes a compile-time error. If a closing
1.1.1.4 ! misho 5584: square bracket is required as a member of the class, it should be the
! 5585: first data character in the class (after an initial circumflex, if
1.1 misho 5586: present) or escaped with a backslash.
5587:
1.1.1.4 ! misho 5588: A character class matches a single character in the subject. In a UTF
! 5589: mode, the character may be more than one data unit long. A matched
1.1.1.2 misho 5590: character must be in the set of characters defined by the class, unless
1.1.1.4 ! misho 5591: the first character in the class definition is a circumflex, in which
1.1.1.2 misho 5592: case the subject character must not be in the set defined by the class.
1.1.1.4 ! misho 5593: If a circumflex is actually required as a member of the class, ensure
1.1.1.2 misho 5594: it is not the first character, or escape it with a backslash.
1.1 misho 5595:
1.1.1.4 ! misho 5596: For example, the character class [aeiou] matches any lower case vowel,
! 5597: while [^aeiou] matches any character that is not a lower case vowel.
1.1 misho 5598: Note that a circumflex is just a convenient notation for specifying the
1.1.1.4 ! misho 5599: characters that are in the class by enumerating those that are not. A
! 5600: class that starts with a circumflex is not an assertion; it still con-
! 5601: sumes a character from the subject string, and therefore it fails if
1.1 misho 5602: the current pointer is at the end of the string.
5603:
1.1.1.4 ! misho 5604: In UTF-8 (UTF-16, UTF-32) mode, characters with values greater than 255
! 5605: (0xffff) can be included in a class as a literal string of data units,
1.1.1.2 misho 5606: or by using the \x{ escaping mechanism.
5607:
1.1.1.4 ! misho 5608: When caseless matching is set, any letters in a class represent both
! 5609: their upper case and lower case versions, so for example, a caseless
! 5610: [aeiou] matches "A" as well as "a", and a caseless [^aeiou] does not
! 5611: match "A", whereas a caseful version would. In a UTF mode, PCRE always
! 5612: understands the concept of case for characters whose values are less
! 5613: than 128, so caseless matching is always possible. For characters with
! 5614: higher values, the concept of case is supported if PCRE is compiled
! 5615: with Unicode property support, but not otherwise. If you want to use
! 5616: caseless matching in a UTF mode for characters 128 and above, you must
! 5617: ensure that PCRE is compiled with Unicode property support as well as
1.1.1.2 misho 5618: with UTF support.
5619:
1.1.1.4 ! misho 5620: Characters that might indicate line breaks are never treated in any
! 5621: special way when matching character classes, whatever line-ending
! 5622: sequence is in use, and whatever setting of the PCRE_DOTALL and
1.1 misho 5623: PCRE_MULTILINE options is used. A class such as [^a] always matches one
5624: of these characters.
5625:
1.1.1.4 ! misho 5626: The minus (hyphen) character can be used to specify a range of charac-
! 5627: ters in a character class. For example, [d-m] matches any letter
! 5628: between d and m, inclusive. If a minus character is required in a
! 5629: class, it must be escaped with a backslash or appear in a position
! 5630: where it cannot be interpreted as indicating a range, typically as the
1.1 misho 5631: first or last character in the class.
5632:
5633: It is not possible to have the literal character "]" as the end charac-
1.1.1.4 ! misho 5634: ter of a range. A pattern such as [W-]46] is interpreted as a class of
! 5635: two characters ("W" and "-") followed by a literal string "46]", so it
! 5636: would match "W46]" or "-46]". However, if the "]" is escaped with a
! 5637: backslash it is interpreted as the end of range, so [W-\]46] is inter-
! 5638: preted as a class containing a range followed by two other characters.
! 5639: The octal or hexadecimal representation of "]" can also be used to end
1.1 misho 5640: a range.
5641:
1.1.1.4 ! misho 5642: Ranges operate in the collating sequence of character values. They can
! 5643: also be used for characters specified numerically, for example
! 5644: [\000-\037]. Ranges can include any characters that are valid for the
1.1.1.2 misho 5645: current mode.
1.1 misho 5646:
5647: If a range that includes letters is used when caseless matching is set,
5648: it matches the letters in either case. For example, [W-c] is equivalent
1.1.1.4 ! misho 5649: to [][\\^_`wxyzabc], matched caselessly, and in a non-UTF mode, if
! 5650: character tables for a French locale are in use, [\xc8-\xcb] matches
! 5651: accented E characters in both cases. In UTF modes, PCRE supports the
! 5652: concept of case for characters with values greater than 128 only when
1.1 misho 5653: it is compiled with Unicode property support.
5654:
1.1.1.4 ! misho 5655: The character escape sequences \d, \D, \h, \H, \p, \P, \s, \S, \v, \V,
1.1 misho 5656: \w, and \W may appear in a character class, and add the characters that
1.1.1.4 ! misho 5657: they match to the class. For example, [\dABCDEF] matches any hexadeci-
! 5658: mal digit. In UTF modes, the PCRE_UCP option affects the meanings of
! 5659: \d, \s, \w and their upper case partners, just as it does when they
! 5660: appear outside a character class, as described in the section entitled
1.1 misho 5661: "Generic character types" above. The escape sequence \b has a different
1.1.1.4 ! misho 5662: meaning inside a character class; it matches the backspace character.
! 5663: The sequences \B, \N, \R, and \X are not special inside a character
! 5664: class. Like any other unrecognized escape sequences, they are treated
! 5665: as the literal characters "B", "N", "R", and "X" by default, but cause
1.1 misho 5666: an error if the PCRE_EXTRA option is set.
5667:
1.1.1.4 ! misho 5668: A circumflex can conveniently be used with the upper case character
! 5669: types to specify a more restricted set of characters than the matching
! 5670: lower case type. For example, the class [^\W_] matches any letter or
1.1 misho 5671: digit, but not underscore, whereas [\w] includes underscore. A positive
5672: character class should be read as "something OR something OR ..." and a
5673: negative class as "NOT something AND NOT something AND NOT ...".
5674:
1.1.1.4 ! misho 5675: The only metacharacters that are recognized in character classes are
! 5676: backslash, hyphen (only where it can be interpreted as specifying a
! 5677: range), circumflex (only at the start), opening square bracket (only
! 5678: when it can be interpreted as introducing a POSIX class name - see the
! 5679: next section), and the terminating closing square bracket. However,
1.1 misho 5680: escaping other non-alphanumeric characters does no harm.
5681:
5682:
5683: POSIX CHARACTER CLASSES
5684:
5685: Perl supports the POSIX notation for character classes. This uses names
1.1.1.4 ! misho 5686: enclosed by [: and :] within the enclosing square brackets. PCRE also
1.1 misho 5687: supports this notation. For example,
5688:
5689: [01[:alpha:]%]
5690:
5691: matches "0", "1", any alphabetic character, or "%". The supported class
5692: names are:
5693:
5694: alnum letters and digits
5695: alpha letters
5696: ascii character codes 0 - 127
5697: blank space or tab only
5698: cntrl control characters
5699: digit decimal digits (same as \d)
5700: graph printing characters, excluding space
5701: lower lower case letters
5702: print printing characters, including space
5703: punct printing characters, excluding letters and digits and space
5704: space white space (not quite the same as \s)
5705: upper upper case letters
5706: word "word" characters (same as \w)
5707: xdigit hexadecimal digits
5708:
1.1.1.4 ! misho 5709: The "space" characters are HT (9), LF (10), VT (11), FF (12), CR (13),
! 5710: and space (32). Notice that this list includes the VT character (code
1.1 misho 5711: 11). This makes "space" different to \s, which does not include VT (for
5712: Perl compatibility).
5713:
1.1.1.4 ! misho 5714: The name "word" is a Perl extension, and "blank" is a GNU extension
! 5715: from Perl 5.8. Another Perl extension is negation, which is indicated
1.1 misho 5716: by a ^ character after the colon. For example,
5717:
5718: [12[:^digit:]]
5719:
1.1.1.4 ! misho 5720: matches "1", "2", or any non-digit. PCRE (and Perl) also recognize the
1.1 misho 5721: POSIX syntax [.ch.] and [=ch=] where "ch" is a "collating element", but
5722: these are not supported, and an error is given if they are encountered.
5723:
1.1.1.4 ! misho 5724: By default, in UTF modes, characters with values greater than 128 do
! 5725: not match any of the POSIX character classes. However, if the PCRE_UCP
! 5726: option is passed to pcre_compile(), some of the classes are changed so
1.1 misho 5727: that Unicode character properties are used. This is achieved by replac-
5728: ing the POSIX classes by other sequences, as follows:
5729:
5730: [:alnum:] becomes \p{Xan}
5731: [:alpha:] becomes \p{L}
5732: [:blank:] becomes \h
5733: [:digit:] becomes \p{Nd}
5734: [:lower:] becomes \p{Ll}
5735: [:space:] becomes \p{Xps}
5736: [:upper:] becomes \p{Lu}
5737: [:word:] becomes \p{Xwd}
5738:
1.1.1.4 ! misho 5739: Negated versions, such as [:^alpha:] use \P instead of \p. The other
1.1 misho 5740: POSIX classes are unchanged, and match only characters with code points
5741: less than 128.
5742:
5743:
5744: VERTICAL BAR
5745:
1.1.1.4 ! misho 5746: Vertical bar characters are used to separate alternative patterns. For
1.1 misho 5747: example, the pattern
5748:
5749: gilbert|sullivan
5750:
1.1.1.4 ! misho 5751: matches either "gilbert" or "sullivan". Any number of alternatives may
! 5752: appear, and an empty alternative is permitted (matching the empty
1.1 misho 5753: string). The matching process tries each alternative in turn, from left
1.1.1.4 ! misho 5754: to right, and the first one that succeeds is used. If the alternatives
! 5755: are within a subpattern (defined below), "succeeds" means matching the
1.1 misho 5756: rest of the main pattern as well as the alternative in the subpattern.
5757:
5758:
5759: INTERNAL OPTION SETTING
5760:
1.1.1.4 ! misho 5761: The settings of the PCRE_CASELESS, PCRE_MULTILINE, PCRE_DOTALL, and
! 5762: PCRE_EXTENDED options (which are Perl-compatible) can be changed from
! 5763: within the pattern by a sequence of Perl option letters enclosed
1.1 misho 5764: between "(?" and ")". The option letters are
5765:
5766: i for PCRE_CASELESS
5767: m for PCRE_MULTILINE
5768: s for PCRE_DOTALL
5769: x for PCRE_EXTENDED
5770:
5771: For example, (?im) sets caseless, multiline matching. It is also possi-
5772: ble to unset these options by preceding the letter with a hyphen, and a
1.1.1.4 ! misho 5773: combined setting and unsetting such as (?im-sx), which sets PCRE_CASE-
! 5774: LESS and PCRE_MULTILINE while unsetting PCRE_DOTALL and PCRE_EXTENDED,
! 5775: is also permitted. If a letter appears both before and after the
1.1 misho 5776: hyphen, the option is unset.
5777:
1.1.1.4 ! misho 5778: The PCRE-specific options PCRE_DUPNAMES, PCRE_UNGREEDY, and PCRE_EXTRA
! 5779: can be changed in the same way as the Perl-compatible options by using
1.1 misho 5780: the characters J, U and X respectively.
5781:
1.1.1.4 ! misho 5782: When one of these option changes occurs at top level (that is, not
! 5783: inside subpattern parentheses), the change applies to the remainder of
1.1 misho 5784: the pattern that follows. If the change is placed right at the start of
5785: a pattern, PCRE extracts it into the global options (and it will there-
5786: fore show up in data extracted by the pcre_fullinfo() function).
5787:
1.1.1.4 ! misho 5788: An option change within a subpattern (see below for a description of
! 5789: subpatterns) affects only that part of the subpattern that follows it,
1.1 misho 5790: so
5791:
5792: (a(?i)b)c
5793:
5794: matches abc and aBc and no other strings (assuming PCRE_CASELESS is not
1.1.1.4 ! misho 5795: used). By this means, options can be made to have different settings
! 5796: in different parts of the pattern. Any changes made in one alternative
! 5797: do carry on into subsequent branches within the same subpattern. For
1.1 misho 5798: example,
5799:
5800: (a(?i)b|c)
5801:
1.1.1.4 ! misho 5802: matches "ab", "aB", "c", and "C", even though when matching "C" the
! 5803: first branch is abandoned before the option setting. This is because
! 5804: the effects of option settings happen at compile time. There would be
1.1 misho 5805: some very weird behaviour otherwise.
5806:
1.1.1.4 ! misho 5807: Note: There are other PCRE-specific options that can be set by the
! 5808: application when the compiling or matching functions are called. In
! 5809: some cases the pattern can contain special leading sequences such as
! 5810: (*CRLF) to override what the application has set or what has been
! 5811: defaulted. Details are given in the section entitled "Newline
! 5812: sequences" above. There are also the (*UTF8), (*UTF16),(*UTF32), and
! 5813: (*UCP) leading sequences that can be used to set UTF and Unicode prop-
! 5814: erty modes; they are equivalent to setting the PCRE_UTF8, PCRE_UTF16,
! 5815: PCRE_UTF32 and the PCRE_UCP options, respectively. The (*UTF) sequence
! 5816: is a generic version that can be used with any of the libraries. How-
! 5817: ever, the application can set the PCRE_NEVER_UTF option, which locks
! 5818: out the use of the (*UTF) sequences.
1.1 misho 5819:
5820:
5821: SUBPATTERNS
5822:
5823: Subpatterns are delimited by parentheses (round brackets), which can be
5824: nested. Turning part of a pattern into a subpattern does two things:
5825:
5826: 1. It localizes a set of alternatives. For example, the pattern
5827:
5828: cat(aract|erpillar|)
5829:
1.1.1.3 misho 5830: matches "cataract", "caterpillar", or "cat". Without the parentheses,
1.1 misho 5831: it would match "cataract", "erpillar" or an empty string.
5832:
1.1.1.3 misho 5833: 2. It sets up the subpattern as a capturing subpattern. This means
5834: that, when the whole pattern matches, that portion of the subject
1.1 misho 5835: string that matched the subpattern is passed back to the caller via the
1.1.1.3 misho 5836: ovector argument of the matching function. (This applies only to the
5837: traditional matching functions; the DFA matching functions do not sup-
1.1.1.2 misho 5838: port capturing.)
5839:
5840: Opening parentheses are counted from left to right (starting from 1) to
1.1.1.3 misho 5841: obtain numbers for the capturing subpatterns. For example, if the
1.1.1.2 misho 5842: string "the red king" is matched against the pattern
1.1 misho 5843:
5844: the ((red|white) (king|queen))
5845:
5846: the captured substrings are "red king", "red", and "king", and are num-
5847: bered 1, 2, and 3, respectively.
5848:
1.1.1.3 misho 5849: The fact that plain parentheses fulfil two functions is not always
5850: helpful. There are often times when a grouping subpattern is required
5851: without a capturing requirement. If an opening parenthesis is followed
5852: by a question mark and a colon, the subpattern does not do any captur-
5853: ing, and is not counted when computing the number of any subsequent
5854: capturing subpatterns. For example, if the string "the white queen" is
1.1 misho 5855: matched against the pattern
5856:
5857: the ((?:red|white) (king|queen))
5858:
5859: the captured substrings are "white queen" and "queen", and are numbered
5860: 1 and 2. The maximum number of capturing subpatterns is 65535.
5861:
1.1.1.3 misho 5862: As a convenient shorthand, if any option settings are required at the
5863: start of a non-capturing subpattern, the option letters may appear
1.1 misho 5864: between the "?" and the ":". Thus the two patterns
5865:
5866: (?i:saturday|sunday)
5867: (?:(?i)saturday|sunday)
5868:
5869: match exactly the same set of strings. Because alternative branches are
1.1.1.3 misho 5870: tried from left to right, and options are not reset until the end of
5871: the subpattern is reached, an option setting in one branch does affect
5872: subsequent branches, so the above patterns match "SUNDAY" as well as
1.1 misho 5873: "Saturday".
5874:
5875:
5876: DUPLICATE SUBPATTERN NUMBERS
5877:
5878: Perl 5.10 introduced a feature whereby each alternative in a subpattern
1.1.1.3 misho 5879: uses the same numbers for its capturing parentheses. Such a subpattern
5880: starts with (?| and is itself a non-capturing subpattern. For example,
1.1 misho 5881: consider this pattern:
5882:
5883: (?|(Sat)ur|(Sun))day
5884:
1.1.1.3 misho 5885: Because the two alternatives are inside a (?| group, both sets of cap-
5886: turing parentheses are numbered one. Thus, when the pattern matches,
5887: you can look at captured substring number one, whichever alternative
5888: matched. This construct is useful when you want to capture part, but
1.1 misho 5889: not all, of one of a number of alternatives. Inside a (?| group, paren-
1.1.1.3 misho 5890: theses are numbered as usual, but the number is reset at the start of
5891: each branch. The numbers of any capturing parentheses that follow the
5892: subpattern start after the highest number used in any branch. The fol-
1.1 misho 5893: lowing example is taken from the Perl documentation. The numbers under-
5894: neath show in which buffer the captured content will be stored.
5895:
5896: # before ---------------branch-reset----------- after
5897: / ( a ) (?| x ( y ) z | (p (q) r) | (t) u (v) ) ( z ) /x
5898: # 1 2 2 3 2 3 4
5899:
1.1.1.3 misho 5900: A back reference to a numbered subpattern uses the most recent value
5901: that is set for that number by any subpattern. The following pattern
1.1 misho 5902: matches "abcabc" or "defdef":
5903:
5904: /(?|(abc)|(def))\1/
5905:
1.1.1.3 misho 5906: In contrast, a subroutine call to a numbered subpattern always refers
5907: to the first one in the pattern with the given number. The following
1.1 misho 5908: pattern matches "abcabc" or "defabc":
5909:
5910: /(?|(abc)|(def))(?1)/
5911:
1.1.1.3 misho 5912: If a condition test for a subpattern's having matched refers to a non-
5913: unique number, the test is true if any of the subpatterns of that num-
1.1 misho 5914: ber have matched.
5915:
1.1.1.3 misho 5916: An alternative approach to using this "branch reset" feature is to use
1.1 misho 5917: duplicate named subpatterns, as described in the next section.
5918:
5919:
5920: NAMED SUBPATTERNS
5921:
1.1.1.3 misho 5922: Identifying capturing parentheses by number is simple, but it can be
5923: very hard to keep track of the numbers in complicated regular expres-
5924: sions. Furthermore, if an expression is modified, the numbers may
5925: change. To help with this difficulty, PCRE supports the naming of sub-
1.1 misho 5926: patterns. This feature was not added to Perl until release 5.10. Python
1.1.1.3 misho 5927: had the feature earlier, and PCRE introduced it at release 4.0, using
5928: the Python syntax. PCRE now supports both the Perl and the Python syn-
5929: tax. Perl allows identically numbered subpatterns to have different
1.1 misho 5930: names, but PCRE does not.
5931:
1.1.1.3 misho 5932: In PCRE, a subpattern can be named in one of three ways: (?<name>...)
5933: or (?'name'...) as in Perl, or (?P<name>...) as in Python. References
5934: to capturing parentheses from other parts of the pattern, such as back
5935: references, recursion, and conditions, can be made by name as well as
1.1 misho 5936: by number.
5937:
1.1.1.3 misho 5938: Names consist of up to 32 alphanumeric characters and underscores.
5939: Named capturing parentheses are still allocated numbers as well as
5940: names, exactly as if the names were not present. The PCRE API provides
1.1 misho 5941: function calls for extracting the name-to-number translation table from
5942: a compiled pattern. There is also a convenience function for extracting
5943: a captured substring by name.
5944:
1.1.1.3 misho 5945: By default, a name must be unique within a pattern, but it is possible
1.1 misho 5946: to relax this constraint by setting the PCRE_DUPNAMES option at compile
1.1.1.3 misho 5947: time. (Duplicate names are also always permitted for subpatterns with
5948: the same number, set up as described in the previous section.) Dupli-
5949: cate names can be useful for patterns where only one instance of the
5950: named parentheses can match. Suppose you want to match the name of a
5951: weekday, either as a 3-letter abbreviation or as the full name, and in
1.1 misho 5952: both cases you want to extract the abbreviation. This pattern (ignoring
5953: the line breaks) does the job:
5954:
5955: (?<DN>Mon|Fri|Sun)(?:day)?|
5956: (?<DN>Tue)(?:sday)?|
5957: (?<DN>Wed)(?:nesday)?|
5958: (?<DN>Thu)(?:rsday)?|
5959: (?<DN>Sat)(?:urday)?
5960:
1.1.1.3 misho 5961: There are five capturing substrings, but only one is ever set after a
1.1 misho 5962: match. (An alternative way of solving this problem is to use a "branch
5963: reset" subpattern, as described in the previous section.)
5964:
1.1.1.3 misho 5965: The convenience function for extracting the data by name returns the
5966: substring for the first (and in this example, the only) subpattern of
5967: that name that matched. This saves searching to find which numbered
1.1 misho 5968: subpattern it was.
5969:
1.1.1.3 misho 5970: If you make a back reference to a non-unique named subpattern from
5971: elsewhere in the pattern, the one that corresponds to the first occur-
1.1 misho 5972: rence of the name is used. In the absence of duplicate numbers (see the
1.1.1.3 misho 5973: previous section) this is the one with the lowest number. If you use a
5974: named reference in a condition test (see the section about conditions
5975: below), either to check whether a subpattern has matched, or to check
5976: for recursion, all subpatterns with the same name are tested. If the
5977: condition is true for any one of them, the overall condition is true.
1.1 misho 5978: This is the same behaviour as testing by number. For further details of
5979: the interfaces for handling named subpatterns, see the pcreapi documen-
5980: tation.
5981:
5982: Warning: You cannot use different names to distinguish between two sub-
1.1.1.3 misho 5983: patterns with the same number because PCRE uses only the numbers when
1.1 misho 5984: matching. For this reason, an error is given at compile time if differ-
1.1.1.3 misho 5985: ent names are given to subpatterns with the same number. However, you
5986: can give the same name to subpatterns with the same number, even when
1.1 misho 5987: PCRE_DUPNAMES is not set.
5988:
5989:
5990: REPETITION
5991:
1.1.1.3 misho 5992: Repetition is specified by quantifiers, which can follow any of the
1.1 misho 5993: following items:
5994:
5995: a literal data character
5996: the dot metacharacter
5997: the \C escape sequence
1.1.1.2 misho 5998: the \X escape sequence
1.1 misho 5999: the \R escape sequence
6000: an escape such as \d or \pL that matches a single character
6001: a character class
6002: a back reference (see next section)
6003: a parenthesized subpattern (including assertions)
6004: a subroutine call to a subpattern (recursive or otherwise)
6005:
1.1.1.3 misho 6006: The general repetition quantifier specifies a minimum and maximum num-
6007: ber of permitted matches, by giving the two numbers in curly brackets
6008: (braces), separated by a comma. The numbers must be less than 65536,
1.1 misho 6009: and the first must be less than or equal to the second. For example:
6010:
6011: z{2,4}
6012:
1.1.1.3 misho 6013: matches "zz", "zzz", or "zzzz". A closing brace on its own is not a
6014: special character. If the second number is omitted, but the comma is
6015: present, there is no upper limit; if the second number and the comma
6016: are both omitted, the quantifier specifies an exact number of required
1.1 misho 6017: matches. Thus
6018:
6019: [aeiou]{3,}
6020:
6021: matches at least 3 successive vowels, but may match many more, while
6022:
6023: \d{8}
6024:
1.1.1.3 misho 6025: matches exactly 8 digits. An opening curly bracket that appears in a
6026: position where a quantifier is not allowed, or one that does not match
6027: the syntax of a quantifier, is taken as a literal character. For exam-
1.1 misho 6028: ple, {,6} is not a quantifier, but a literal string of four characters.
6029:
1.1.1.2 misho 6030: In UTF modes, quantifiers apply to characters rather than to individual
1.1.1.3 misho 6031: data units. Thus, for example, \x{100}{2} matches two characters, each
1.1.1.2 misho 6032: of which is represented by a two-byte sequence in a UTF-8 string. Simi-
1.1.1.4 ! misho 6033: larly, \X{3} matches three Unicode extended grapheme clusters, each of
! 6034: which may be several data units long (and they may be of different
! 6035: lengths).
1.1 misho 6036:
6037: The quantifier {0} is permitted, causing the expression to behave as if
6038: the previous item and the quantifier were not present. This may be use-
1.1.1.4 ! misho 6039: ful for subpatterns that are referenced as subroutines from elsewhere
1.1 misho 6040: in the pattern (but see also the section entitled "Defining subpatterns
1.1.1.4 ! misho 6041: for use by reference only" below). Items other than subpatterns that
1.1 misho 6042: have a {0} quantifier are omitted from the compiled pattern.
6043:
1.1.1.4 ! misho 6044: For convenience, the three most common quantifiers have single-charac-
1.1 misho 6045: ter abbreviations:
6046:
6047: * is equivalent to {0,}
6048: + is equivalent to {1,}
6049: ? is equivalent to {0,1}
6050:
1.1.1.4 ! misho 6051: It is possible to construct infinite loops by following a subpattern
1.1 misho 6052: that can match no characters with a quantifier that has no upper limit,
6053: for example:
6054:
6055: (a?)*
6056:
6057: Earlier versions of Perl and PCRE used to give an error at compile time
1.1.1.4 ! misho 6058: for such patterns. However, because there are cases where this can be
! 6059: useful, such patterns are now accepted, but if any repetition of the
! 6060: subpattern does in fact match no characters, the loop is forcibly bro-
1.1 misho 6061: ken.
6062:
1.1.1.4 ! misho 6063: By default, the quantifiers are "greedy", that is, they match as much
! 6064: as possible (up to the maximum number of permitted times), without
! 6065: causing the rest of the pattern to fail. The classic example of where
1.1 misho 6066: this gives problems is in trying to match comments in C programs. These
1.1.1.4 ! misho 6067: appear between /* and */ and within the comment, individual * and /
! 6068: characters may appear. An attempt to match C comments by applying the
1.1 misho 6069: pattern
6070:
6071: /\*.*\*/
6072:
6073: to the string
6074:
6075: /* first comment */ not comment /* second comment */
6076:
1.1.1.4 ! misho 6077: fails, because it matches the entire string owing to the greediness of
1.1 misho 6078: the .* item.
6079:
1.1.1.4 ! misho 6080: However, if a quantifier is followed by a question mark, it ceases to
1.1 misho 6081: be greedy, and instead matches the minimum number of times possible, so
6082: the pattern
6083:
6084: /\*.*?\*/
6085:
1.1.1.4 ! misho 6086: does the right thing with the C comments. The meaning of the various
! 6087: quantifiers is not otherwise changed, just the preferred number of
! 6088: matches. Do not confuse this use of question mark with its use as a
! 6089: quantifier in its own right. Because it has two uses, it can sometimes
1.1 misho 6090: appear doubled, as in
6091:
6092: \d??\d
6093:
6094: which matches one digit by preference, but can match two if that is the
6095: only way the rest of the pattern matches.
6096:
1.1.1.4 ! misho 6097: If the PCRE_UNGREEDY option is set (an option that is not available in
! 6098: Perl), the quantifiers are not greedy by default, but individual ones
! 6099: can be made greedy by following them with a question mark. In other
1.1 misho 6100: words, it inverts the default behaviour.
6101:
1.1.1.4 ! misho 6102: When a parenthesized subpattern is quantified with a minimum repeat
! 6103: count that is greater than 1 or with a limited maximum, more memory is
! 6104: required for the compiled pattern, in proportion to the size of the
1.1 misho 6105: minimum or maximum.
6106:
6107: If a pattern starts with .* or .{0,} and the PCRE_DOTALL option (equiv-
1.1.1.4 ! misho 6108: alent to Perl's /s) is set, thus allowing the dot to match newlines,
! 6109: the pattern is implicitly anchored, because whatever follows will be
! 6110: tried against every character position in the subject string, so there
! 6111: is no point in retrying the overall match at any position after the
! 6112: first. PCRE normally treats such a pattern as though it were preceded
1.1 misho 6113: by \A.
6114:
1.1.1.4 ! misho 6115: In cases where it is known that the subject string contains no new-
! 6116: lines, it is worth setting PCRE_DOTALL in order to obtain this opti-
1.1 misho 6117: mization, or alternatively using ^ to indicate anchoring explicitly.
6118:
1.1.1.4 ! misho 6119: However, there are some cases where the optimization cannot be used.
1.1 misho 6120: When .* is inside capturing parentheses that are the subject of a back
6121: reference elsewhere in the pattern, a match at the start may fail where
6122: a later one succeeds. Consider, for example:
6123:
6124: (.*)abc\1
6125:
1.1.1.4 ! misho 6126: If the subject is "xyz123abc123" the match point is the fourth charac-
1.1 misho 6127: ter. For this reason, such a pattern is not implicitly anchored.
6128:
1.1.1.4 ! misho 6129: Another case where implicit anchoring is not applied is when the lead-
! 6130: ing .* is inside an atomic group. Once again, a match at the start may
! 6131: fail where a later one succeeds. Consider this pattern:
! 6132:
! 6133: (?>.*?a)b
! 6134:
! 6135: It matches "ab" in the subject "aab". The use of the backtracking con-
! 6136: trol verbs (*PRUNE) and (*SKIP) also disable this optimization.
! 6137:
1.1 misho 6138: When a capturing subpattern is repeated, the value captured is the sub-
6139: string that matched the final iteration. For example, after
6140:
6141: (tweedle[dume]{3}\s*)+
6142:
6143: has matched "tweedledum tweedledee" the value of the captured substring
1.1.1.3 misho 6144: is "tweedledee". However, if there are nested capturing subpatterns,
6145: the corresponding captured values may have been set in previous itera-
1.1 misho 6146: tions. For example, after
6147:
6148: /(a|(b))+/
6149:
6150: matches "aba" the value of the second captured substring is "b".
6151:
6152:
6153: ATOMIC GROUPING AND POSSESSIVE QUANTIFIERS
6154:
1.1.1.3 misho 6155: With both maximizing ("greedy") and minimizing ("ungreedy" or "lazy")
6156: repetition, failure of what follows normally causes the repeated item
6157: to be re-evaluated to see if a different number of repeats allows the
6158: rest of the pattern to match. Sometimes it is useful to prevent this,
6159: either to change the nature of the match, or to cause it fail earlier
6160: than it otherwise might, when the author of the pattern knows there is
1.1 misho 6161: no point in carrying on.
6162:
1.1.1.3 misho 6163: Consider, for example, the pattern \d+foo when applied to the subject
1.1 misho 6164: line
6165:
6166: 123456bar
6167:
6168: After matching all 6 digits and then failing to match "foo", the normal
1.1.1.3 misho 6169: action of the matcher is to try again with only 5 digits matching the
6170: \d+ item, and then with 4, and so on, before ultimately failing.
6171: "Atomic grouping" (a term taken from Jeffrey Friedl's book) provides
6172: the means for specifying that once a subpattern has matched, it is not
1.1 misho 6173: to be re-evaluated in this way.
6174:
1.1.1.3 misho 6175: If we use atomic grouping for the previous example, the matcher gives
6176: up immediately on failing to match "foo" the first time. The notation
1.1 misho 6177: is a kind of special parenthesis, starting with (?> as in this example:
6178:
6179: (?>\d+)foo
6180:
1.1.1.3 misho 6181: This kind of parenthesis "locks up" the part of the pattern it con-
6182: tains once it has matched, and a failure further into the pattern is
6183: prevented from backtracking into it. Backtracking past it to previous
1.1 misho 6184: items, however, works as normal.
6185:
1.1.1.3 misho 6186: An alternative description is that a subpattern of this type matches
6187: the string of characters that an identical standalone pattern would
1.1 misho 6188: match, if anchored at the current point in the subject string.
6189:
6190: Atomic grouping subpatterns are not capturing subpatterns. Simple cases
6191: such as the above example can be thought of as a maximizing repeat that
1.1.1.3 misho 6192: must swallow everything it can. So, while both \d+ and \d+? are pre-
6193: pared to adjust the number of digits they match in order to make the
1.1 misho 6194: rest of the pattern match, (?>\d+) can only match an entire sequence of
6195: digits.
6196:
1.1.1.3 misho 6197: Atomic groups in general can of course contain arbitrarily complicated
6198: subpatterns, and can be nested. However, when the subpattern for an
1.1 misho 6199: atomic group is just a single repeated item, as in the example above, a
1.1.1.3 misho 6200: simpler notation, called a "possessive quantifier" can be used. This
6201: consists of an additional + character following a quantifier. Using
1.1 misho 6202: this notation, the previous example can be rewritten as
6203:
6204: \d++foo
6205:
6206: Note that a possessive quantifier can be used with an entire group, for
6207: example:
6208:
6209: (abc|xyz){2,3}+
6210:
1.1.1.3 misho 6211: Possessive quantifiers are always greedy; the setting of the
1.1 misho 6212: PCRE_UNGREEDY option is ignored. They are a convenient notation for the
1.1.1.3 misho 6213: simpler forms of atomic group. However, there is no difference in the
6214: meaning of a possessive quantifier and the equivalent atomic group,
6215: though there may be a performance difference; possessive quantifiers
1.1 misho 6216: should be slightly faster.
6217:
1.1.1.3 misho 6218: The possessive quantifier syntax is an extension to the Perl 5.8 syn-
6219: tax. Jeffrey Friedl originated the idea (and the name) in the first
1.1 misho 6220: edition of his book. Mike McCloskey liked it, so implemented it when he
1.1.1.3 misho 6221: built Sun's Java package, and PCRE copied it from there. It ultimately
1.1 misho 6222: found its way into Perl at release 5.10.
6223:
6224: PCRE has an optimization that automatically "possessifies" certain sim-
1.1.1.3 misho 6225: ple pattern constructs. For example, the sequence A+B is treated as
6226: A++B because there is no point in backtracking into a sequence of A's
1.1 misho 6227: when B must follow.
6228:
1.1.1.3 misho 6229: When a pattern contains an unlimited repeat inside a subpattern that
6230: can itself be repeated an unlimited number of times, the use of an
6231: atomic group is the only way to avoid some failing matches taking a
1.1 misho 6232: very long time indeed. The pattern
6233:
6234: (\D+|<\d+>)*[!?]
6235:
1.1.1.3 misho 6236: matches an unlimited number of substrings that either consist of non-
6237: digits, or digits enclosed in <>, followed by either ! or ?. When it
1.1 misho 6238: matches, it runs quickly. However, if it is applied to
6239:
6240: aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
6241:
1.1.1.3 misho 6242: it takes a long time before reporting failure. This is because the
6243: string can be divided between the internal \D+ repeat and the external
6244: * repeat in a large number of ways, and all have to be tried. (The
6245: example uses [!?] rather than a single character at the end, because
6246: both PCRE and Perl have an optimization that allows for fast failure
6247: when a single character is used. They remember the last single charac-
6248: ter that is required for a match, and fail early if it is not present
6249: in the string.) If the pattern is changed so that it uses an atomic
1.1 misho 6250: group, like this:
6251:
6252: ((?>\D+)|<\d+>)*[!?]
6253:
6254: sequences of non-digits cannot be broken, and failure happens quickly.
6255:
6256:
6257: BACK REFERENCES
6258:
6259: Outside a character class, a backslash followed by a digit greater than
6260: 0 (and possibly further digits) is a back reference to a capturing sub-
1.1.1.3 misho 6261: pattern earlier (that is, to its left) in the pattern, provided there
1.1 misho 6262: have been that many previous capturing left parentheses.
6263:
6264: However, if the decimal number following the backslash is less than 10,
1.1.1.3 misho 6265: it is always taken as a back reference, and causes an error only if
6266: there are not that many capturing left parentheses in the entire pat-
6267: tern. In other words, the parentheses that are referenced need not be
6268: to the left of the reference for numbers less than 10. A "forward back
6269: reference" of this type can make sense when a repetition is involved
6270: and the subpattern to the right has participated in an earlier itera-
1.1 misho 6271: tion.
6272:
1.1.1.3 misho 6273: It is not possible to have a numerical "forward back reference" to a
6274: subpattern whose number is 10 or more using this syntax because a
6275: sequence such as \50 is interpreted as a character defined in octal.
1.1 misho 6276: See the subsection entitled "Non-printing characters" above for further
1.1.1.3 misho 6277: details of the handling of digits following a backslash. There is no
6278: such problem when named parentheses are used. A back reference to any
1.1 misho 6279: subpattern is possible using named parentheses (see below).
6280:
1.1.1.3 misho 6281: Another way of avoiding the ambiguity inherent in the use of digits
6282: following a backslash is to use the \g escape sequence. This escape
1.1 misho 6283: must be followed by an unsigned number or a negative number, optionally
6284: enclosed in braces. These examples are all identical:
6285:
6286: (ring), \1
6287: (ring), \g1
6288: (ring), \g{1}
6289:
1.1.1.3 misho 6290: An unsigned number specifies an absolute reference without the ambigu-
1.1 misho 6291: ity that is present in the older syntax. It is also useful when literal
6292: digits follow the reference. A negative number is a relative reference.
6293: Consider this example:
6294:
6295: (abc(def)ghi)\g{-1}
6296:
6297: The sequence \g{-1} is a reference to the most recently started captur-
6298: ing subpattern before \g, that is, is it equivalent to \2 in this exam-
1.1.1.3 misho 6299: ple. Similarly, \g{-2} would be equivalent to \1. The use of relative
6300: references can be helpful in long patterns, and also in patterns that
6301: are created by joining together fragments that contain references
1.1 misho 6302: within themselves.
6303:
1.1.1.3 misho 6304: A back reference matches whatever actually matched the capturing sub-
6305: pattern in the current subject string, rather than anything matching
1.1 misho 6306: the subpattern itself (see "Subpatterns as subroutines" below for a way
6307: of doing that). So the pattern
6308:
6309: (sens|respons)e and \1ibility
6310:
1.1.1.3 misho 6311: matches "sense and sensibility" and "response and responsibility", but
6312: not "sense and responsibility". If caseful matching is in force at the
6313: time of the back reference, the case of letters is relevant. For exam-
1.1 misho 6314: ple,
6315:
6316: ((?i)rah)\s+\1
6317:
1.1.1.3 misho 6318: matches "rah rah" and "RAH RAH", but not "RAH rah", even though the
1.1 misho 6319: original capturing subpattern is matched caselessly.
6320:
1.1.1.3 misho 6321: There are several different ways of writing back references to named
6322: subpatterns. The .NET syntax \k{name} and the Perl syntax \k<name> or
6323: \k'name' are supported, as is the Python syntax (?P=name). Perl 5.10's
1.1 misho 6324: unified back reference syntax, in which \g can be used for both numeric
1.1.1.3 misho 6325: and named references, is also supported. We could rewrite the above
1.1 misho 6326: example in any of the following ways:
6327:
6328: (?<p1>(?i)rah)\s+\k<p1>
6329: (?'p1'(?i)rah)\s+\k{p1}
6330: (?P<p1>(?i)rah)\s+(?P=p1)
6331: (?<p1>(?i)rah)\s+\g{p1}
6332:
1.1.1.3 misho 6333: A subpattern that is referenced by name may appear in the pattern
1.1 misho 6334: before or after the reference.
6335:
1.1.1.3 misho 6336: There may be more than one back reference to the same subpattern. If a
6337: subpattern has not actually been used in a particular match, any back
1.1 misho 6338: references to it always fail by default. For example, the pattern
6339:
6340: (a|(bc))\2
6341:
1.1.1.3 misho 6342: always fails if it starts to match "a" rather than "bc". However, if
1.1 misho 6343: the PCRE_JAVASCRIPT_COMPAT option is set at compile time, a back refer-
6344: ence to an unset value matches an empty string.
6345:
1.1.1.3 misho 6346: Because there may be many capturing parentheses in a pattern, all dig-
6347: its following a backslash are taken as part of a potential back refer-
6348: ence number. If the pattern continues with a digit character, some
6349: delimiter must be used to terminate the back reference. If the
6350: PCRE_EXTENDED option is set, this can be white space. Otherwise, the
6351: \g{ syntax or an empty comment (see "Comments" below) can be used.
1.1 misho 6352:
6353: Recursive back references
6354:
1.1.1.3 misho 6355: A back reference that occurs inside the parentheses to which it refers
6356: fails when the subpattern is first used, so, for example, (a\1) never
6357: matches. However, such references can be useful inside repeated sub-
1.1 misho 6358: patterns. For example, the pattern
6359:
6360: (a|b\1)+
6361:
6362: matches any number of "a"s and also "aba", "ababbaa" etc. At each iter-
1.1.1.3 misho 6363: ation of the subpattern, the back reference matches the character
6364: string corresponding to the previous iteration. In order for this to
6365: work, the pattern must be such that the first iteration does not need
6366: to match the back reference. This can be done using alternation, as in
1.1 misho 6367: the example above, or by a quantifier with a minimum of zero.
6368:
1.1.1.3 misho 6369: Back references of this type cause the group that they reference to be
6370: treated as an atomic group. Once the whole group has been matched, a
6371: subsequent matching failure cannot cause backtracking into the middle
1.1 misho 6372: of the group.
6373:
6374:
6375: ASSERTIONS
6376:
1.1.1.3 misho 6377: An assertion is a test on the characters following or preceding the
6378: current matching point that does not actually consume any characters.
6379: The simple assertions coded as \b, \B, \A, \G, \Z, \z, ^ and $ are
1.1 misho 6380: described above.
6381:
1.1.1.3 misho 6382: More complicated assertions are coded as subpatterns. There are two
6383: kinds: those that look ahead of the current position in the subject
6384: string, and those that look behind it. An assertion subpattern is
6385: matched in the normal way, except that it does not cause the current
1.1 misho 6386: matching position to be changed.
6387:
1.1.1.3 misho 6388: Assertion subpatterns are not capturing subpatterns. If such an asser-
6389: tion contains capturing subpatterns within it, these are counted for
6390: the purposes of numbering the capturing subpatterns in the whole pat-
6391: tern. However, substring capturing is carried out only for positive
1.1.1.4 ! misho 6392: assertions. (Perl sometimes, but not always, does do capturing in nega-
! 6393: tive assertions.)
1.1 misho 6394:
1.1.1.4 ! misho 6395: For compatibility with Perl, assertion subpatterns may be repeated;
! 6396: though it makes no sense to assert the same thing several times, the
! 6397: side effect of capturing parentheses may occasionally be useful. In
1.1 misho 6398: practice, there only three cases:
6399:
1.1.1.4 ! misho 6400: (1) If the quantifier is {0}, the assertion is never obeyed during
! 6401: matching. However, it may contain internal capturing parenthesized
1.1 misho 6402: groups that are called from elsewhere via the subroutine mechanism.
6403:
1.1.1.4 ! misho 6404: (2) If quantifier is {0,n} where n is greater than zero, it is treated
! 6405: as if it were {0,1}. At run time, the rest of the pattern match is
1.1 misho 6406: tried with and without the assertion, the order depending on the greed-
6407: iness of the quantifier.
6408:
1.1.1.4 ! misho 6409: (3) If the minimum repetition is greater than zero, the quantifier is
! 6410: ignored. The assertion is obeyed just once when encountered during
1.1 misho 6411: matching.
6412:
6413: Lookahead assertions
6414:
6415: Lookahead assertions start with (?= for positive assertions and (?! for
6416: negative assertions. For example,
6417:
6418: \w+(?=;)
6419:
1.1.1.4 ! misho 6420: matches a word followed by a semicolon, but does not include the semi-
1.1 misho 6421: colon in the match, and
6422:
6423: foo(?!bar)
6424:
1.1.1.4 ! misho 6425: matches any occurrence of "foo" that is not followed by "bar". Note
1.1 misho 6426: that the apparently similar pattern
6427:
6428: (?!foo)bar
6429:
1.1.1.4 ! misho 6430: does not find an occurrence of "bar" that is preceded by something
! 6431: other than "foo"; it finds any occurrence of "bar" whatsoever, because
1.1 misho 6432: the assertion (?!foo) is always true when the next three characters are
6433: "bar". A lookbehind assertion is needed to achieve the other effect.
6434:
6435: If you want to force a matching failure at some point in a pattern, the
1.1.1.4 ! misho 6436: most convenient way to do it is with (?!) because an empty string
! 6437: always matches, so an assertion that requires there not to be an empty
1.1 misho 6438: string must always fail. The backtracking control verb (*FAIL) or (*F)
6439: is a synonym for (?!).
6440:
6441: Lookbehind assertions
6442:
1.1.1.4 ! misho 6443: Lookbehind assertions start with (?<= for positive assertions and (?<!
1.1 misho 6444: for negative assertions. For example,
6445:
6446: (?<!foo)bar
6447:
1.1.1.4 ! misho 6448: does find an occurrence of "bar" that is not preceded by "foo". The
! 6449: contents of a lookbehind assertion are restricted such that all the
1.1 misho 6450: strings it matches must have a fixed length. However, if there are sev-
1.1.1.4 ! misho 6451: eral top-level alternatives, they do not all have to have the same
1.1 misho 6452: fixed length. Thus
6453:
6454: (?<=bullock|donkey)
6455:
6456: is permitted, but
6457:
6458: (?<!dogs?|cats?)
6459:
1.1.1.4 ! misho 6460: causes an error at compile time. Branches that match different length
! 6461: strings are permitted only at the top level of a lookbehind assertion.
1.1 misho 6462: This is an extension compared with Perl, which requires all branches to
6463: match the same length of string. An assertion such as
6464:
6465: (?<=ab(c|de))
6466:
1.1.1.4 ! misho 6467: is not permitted, because its single top-level branch can match two
1.1 misho 6468: different lengths, but it is acceptable to PCRE if rewritten to use two
6469: top-level branches:
6470:
6471: (?<=abc|abde)
6472:
1.1.1.4 ! misho 6473: In some cases, the escape sequence \K (see above) can be used instead
1.1 misho 6474: of a lookbehind assertion to get round the fixed-length restriction.
6475:
1.1.1.4 ! misho 6476: The implementation of lookbehind assertions is, for each alternative,
! 6477: to temporarily move the current position back by the fixed length and
1.1 misho 6478: then try to match. If there are insufficient characters before the cur-
6479: rent position, the assertion fails.
6480:
1.1.1.4 ! misho 6481: In a UTF mode, PCRE does not allow the \C escape (which matches a sin-
! 6482: gle data unit even in a UTF mode) to appear in lookbehind assertions,
! 6483: because it makes it impossible to calculate the length of the lookbe-
! 6484: hind. The \X and \R escapes, which can match different numbers of data
1.1.1.2 misho 6485: units, are also not permitted.
1.1 misho 6486:
1.1.1.4 ! misho 6487: "Subroutine" calls (see below) such as (?2) or (?&X) are permitted in
! 6488: lookbehinds, as long as the subpattern matches a fixed-length string.
1.1 misho 6489: Recursion, however, is not supported.
6490:
1.1.1.4 ! misho 6491: Possessive quantifiers can be used in conjunction with lookbehind
1.1 misho 6492: assertions to specify efficient matching of fixed-length strings at the
6493: end of subject strings. Consider a simple pattern such as
6494:
6495: abcd$
6496:
1.1.1.4 ! misho 6497: when applied to a long string that does not match. Because matching
1.1 misho 6498: proceeds from left to right, PCRE will look for each "a" in the subject
1.1.1.4 ! misho 6499: and then see if what follows matches the rest of the pattern. If the
1.1 misho 6500: pattern is specified as
6501:
6502: ^.*abcd$
6503:
1.1.1.4 ! misho 6504: the initial .* matches the entire string at first, but when this fails
1.1 misho 6505: (because there is no following "a"), it backtracks to match all but the
1.1.1.4 ! misho 6506: last character, then all but the last two characters, and so on. Once
! 6507: again the search for "a" covers the entire string, from right to left,
1.1 misho 6508: so we are no better off. However, if the pattern is written as
6509:
6510: ^.*+(?<=abcd)
6511:
1.1.1.4 ! misho 6512: there can be no backtracking for the .*+ item; it can match only the
! 6513: entire string. The subsequent lookbehind assertion does a single test
! 6514: on the last four characters. If it fails, the match fails immediately.
! 6515: For long strings, this approach makes a significant difference to the
1.1 misho 6516: processing time.
6517:
6518: Using multiple assertions
6519:
6520: Several assertions (of any sort) may occur in succession. For example,
6521:
6522: (?<=\d{3})(?<!999)foo
6523:
1.1.1.4 ! misho 6524: matches "foo" preceded by three digits that are not "999". Notice that
! 6525: each of the assertions is applied independently at the same point in
! 6526: the subject string. First there is a check that the previous three
! 6527: characters are all digits, and then there is a check that the same
1.1 misho 6528: three characters are not "999". This pattern does not match "foo" pre-
1.1.1.4 ! misho 6529: ceded by six characters, the first of which are digits and the last
! 6530: three of which are not "999". For example, it doesn't match "123abc-
1.1 misho 6531: foo". A pattern to do that is
6532:
6533: (?<=\d{3}...)(?<!999)foo
6534:
1.1.1.4 ! misho 6535: This time the first assertion looks at the preceding six characters,
1.1 misho 6536: checking that the first three are digits, and then the second assertion
6537: checks that the preceding three characters are not "999".
6538:
6539: Assertions can be nested in any combination. For example,
6540:
6541: (?<=(?<!foo)bar)baz
6542:
1.1.1.4 ! misho 6543: matches an occurrence of "baz" that is preceded by "bar" which in turn
1.1 misho 6544: is not preceded by "foo", while
6545:
6546: (?<=\d{3}(?!999)...)foo
6547:
1.1.1.4 ! misho 6548: is another pattern that matches "foo" preceded by three digits and any
1.1 misho 6549: three characters that are not "999".
6550:
6551:
6552: CONDITIONAL SUBPATTERNS
6553:
1.1.1.4 ! misho 6554: It is possible to cause the matching process to obey a subpattern con-
! 6555: ditionally or to choose between two alternative subpatterns, depending
! 6556: on the result of an assertion, or whether a specific capturing subpat-
! 6557: tern has already been matched. The two possible forms of conditional
1.1 misho 6558: subpattern are:
6559:
6560: (?(condition)yes-pattern)
6561: (?(condition)yes-pattern|no-pattern)
6562:
1.1.1.4 ! misho 6563: If the condition is satisfied, the yes-pattern is used; otherwise the
! 6564: no-pattern (if present) is used. If there are more than two alterna-
! 6565: tives in the subpattern, a compile-time error occurs. Each of the two
1.1 misho 6566: alternatives may itself contain nested subpatterns of any form, includ-
6567: ing conditional subpatterns; the restriction to two alternatives
6568: applies only at the level of the condition. This pattern fragment is an
6569: example where the alternatives are complex:
6570:
6571: (?(1) (A|B|C) | (D | (?(2)E|F) | E) )
6572:
6573:
1.1.1.4 ! misho 6574: There are four kinds of condition: references to subpatterns, refer-
1.1 misho 6575: ences to recursion, a pseudo-condition called DEFINE, and assertions.
6576:
6577: Checking for a used subpattern by number
6578:
1.1.1.4 ! misho 6579: If the text between the parentheses consists of a sequence of digits,
1.1 misho 6580: the condition is true if a capturing subpattern of that number has pre-
1.1.1.4 ! misho 6581: viously matched. If there is more than one capturing subpattern with
! 6582: the same number (see the earlier section about duplicate subpattern
! 6583: numbers), the condition is true if any of them have matched. An alter-
! 6584: native notation is to precede the digits with a plus or minus sign. In
! 6585: this case, the subpattern number is relative rather than absolute. The
! 6586: most recently opened parentheses can be referenced by (?(-1), the next
! 6587: most recent by (?(-2), and so on. Inside loops it can also make sense
1.1 misho 6588: to refer to subsequent groups. The next parentheses to be opened can be
1.1.1.4 ! misho 6589: referenced as (?(+1), and so on. (The value zero in any of these forms
1.1 misho 6590: is not used; it provokes a compile-time error.)
6591:
1.1.1.4 ! misho 6592: Consider the following pattern, which contains non-significant white
1.1 misho 6593: space to make it more readable (assume the PCRE_EXTENDED option) and to
6594: divide it into three parts for ease of discussion:
6595:
6596: ( \( )? [^()]+ (?(1) \) )
6597:
1.1.1.4 ! misho 6598: The first part matches an optional opening parenthesis, and if that
1.1 misho 6599: character is present, sets it as the first captured substring. The sec-
1.1.1.4 ! misho 6600: ond part matches one or more characters that are not parentheses. The
! 6601: third part is a conditional subpattern that tests whether or not the
! 6602: first set of parentheses matched. If they did, that is, if subject
! 6603: started with an opening parenthesis, the condition is true, and so the
! 6604: yes-pattern is executed and a closing parenthesis is required. Other-
! 6605: wise, since no-pattern is not present, the subpattern matches nothing.
! 6606: In other words, this pattern matches a sequence of non-parentheses,
1.1 misho 6607: optionally enclosed in parentheses.
6608:
1.1.1.4 ! misho 6609: If you were embedding this pattern in a larger one, you could use a
1.1 misho 6610: relative reference:
6611:
6612: ...other stuff... ( \( )? [^()]+ (?(-1) \) ) ...
6613:
1.1.1.4 ! misho 6614: This makes the fragment independent of the parentheses in the larger
1.1 misho 6615: pattern.
6616:
6617: Checking for a used subpattern by name
6618:
1.1.1.4 ! misho 6619: Perl uses the syntax (?(<name>)...) or (?('name')...) to test for a
! 6620: used subpattern by name. For compatibility with earlier versions of
! 6621: PCRE, which had this facility before Perl, the syntax (?(name)...) is
! 6622: also recognized. However, there is a possible ambiguity with this syn-
! 6623: tax, because subpattern names may consist entirely of digits. PCRE
! 6624: looks first for a named subpattern; if it cannot find one and the name
! 6625: consists entirely of digits, PCRE looks for a subpattern of that num-
! 6626: ber, which must be greater than zero. Using subpattern names that con-
1.1 misho 6627: sist entirely of digits is not recommended.
6628:
6629: Rewriting the above example to use a named subpattern gives this:
6630:
6631: (?<OPEN> \( )? [^()]+ (?(<OPEN>) \) )
6632:
1.1.1.4 ! misho 6633: If the name used in a condition of this kind is a duplicate, the test
! 6634: is applied to all subpatterns of the same name, and is true if any one
1.1 misho 6635: of them has matched.
6636:
6637: Checking for pattern recursion
6638:
6639: If the condition is the string (R), and there is no subpattern with the
1.1.1.4 ! misho 6640: name R, the condition is true if a recursive call to the whole pattern
1.1 misho 6641: or any subpattern has been made. If digits or a name preceded by amper-
6642: sand follow the letter R, for example:
6643:
6644: (?(R3)...) or (?(R&name)...)
6645:
6646: the condition is true if the most recent recursion is into a subpattern
6647: whose number or name is given. This condition does not check the entire
1.1.1.4 ! misho 6648: recursion stack. If the name used in a condition of this kind is a
1.1 misho 6649: duplicate, the test is applied to all subpatterns of the same name, and
6650: is true if any one of them is the most recent recursion.
6651:
1.1.1.4 ! misho 6652: At "top level", all these recursion test conditions are false. The
1.1 misho 6653: syntax for recursive patterns is described below.
6654:
6655: Defining subpatterns for use by reference only
6656:
1.1.1.4 ! misho 6657: If the condition is the string (DEFINE), and there is no subpattern
! 6658: with the name DEFINE, the condition is always false. In this case,
! 6659: there may be only one alternative in the subpattern. It is always
! 6660: skipped if control reaches this point in the pattern; the idea of
! 6661: DEFINE is that it can be used to define subroutines that can be refer-
! 6662: enced from elsewhere. (The use of subroutines is described below.) For
! 6663: example, a pattern to match an IPv4 address such as "192.168.23.245"
1.1.1.3 misho 6664: could be written like this (ignore white space and line breaks):
1.1 misho 6665:
6666: (?(DEFINE) (?<byte> 2[0-4]\d | 25[0-5] | 1\d\d | [1-9]?\d) )
6667: \b (?&byte) (\.(?&byte)){3} \b
6668:
1.1.1.4 ! misho 6669: The first part of the pattern is a DEFINE group inside which a another
! 6670: group named "byte" is defined. This matches an individual component of
! 6671: an IPv4 address (a number less than 256). When matching takes place,
! 6672: this part of the pattern is skipped because DEFINE acts like a false
! 6673: condition. The rest of the pattern uses references to the named group
! 6674: to match the four dot-separated components of an IPv4 address, insist-
1.1 misho 6675: ing on a word boundary at each end.
6676:
6677: Assertion conditions
6678:
1.1.1.4 ! misho 6679: If the condition is not in any of the above formats, it must be an
! 6680: assertion. This may be a positive or negative lookahead or lookbehind
! 6681: assertion. Consider this pattern, again containing non-significant
1.1 misho 6682: white space, and with the two alternatives on the second line:
6683:
6684: (?(?=[^a-z]*[a-z])
6685: \d{2}-[a-z]{3}-\d{2} | \d{2}-\d{2}-\d{2} )
6686:
1.1.1.4 ! misho 6687: The condition is a positive lookahead assertion that matches an
! 6688: optional sequence of non-letters followed by a letter. In other words,
! 6689: it tests for the presence of at least one letter in the subject. If a
! 6690: letter is found, the subject is matched against the first alternative;
! 6691: otherwise it is matched against the second. This pattern matches
! 6692: strings in one of the two forms dd-aaa-dd or dd-dd-dd, where aaa are
1.1 misho 6693: letters and dd are digits.
6694:
6695:
6696: COMMENTS
6697:
6698: There are two ways of including comments in patterns that are processed
6699: by PCRE. In both cases, the start of the comment must not be in a char-
6700: acter class, nor in the middle of any other sequence of related charac-
1.1.1.4 ! misho 6701: ters such as (?: or a subpattern name or number. The characters that
1.1 misho 6702: make up a comment play no part in the pattern matching.
6703:
1.1.1.4 ! misho 6704: The sequence (?# marks the start of a comment that continues up to the
! 6705: next closing parenthesis. Nested parentheses are not permitted. If the
1.1 misho 6706: PCRE_EXTENDED option is set, an unescaped # character also introduces a
1.1.1.4 ! misho 6707: comment, which in this case continues to immediately after the next
! 6708: newline character or character sequence in the pattern. Which charac-
1.1 misho 6709: ters are interpreted as newlines is controlled by the options passed to
1.1.1.4 ! misho 6710: a compiling function or by a special sequence at the start of the pat-
1.1.1.2 misho 6711: tern, as described in the section entitled "Newline conventions" above.
6712: Note that the end of this type of comment is a literal newline sequence
1.1.1.4 ! misho 6713: in the pattern; escape sequences that happen to represent a newline do
! 6714: not count. For example, consider this pattern when PCRE_EXTENDED is
1.1.1.2 misho 6715: set, and the default newline convention is in force:
1.1 misho 6716:
6717: abc #comment \n still comment
6718:
1.1.1.4 ! misho 6719: On encountering the # character, pcre_compile() skips along, looking
! 6720: for a newline in the pattern. The sequence \n is still literal at this
! 6721: stage, so it does not terminate the comment. Only an actual character
1.1 misho 6722: with the code value 0x0a (the default newline) does so.
6723:
6724:
6725: RECURSIVE PATTERNS
6726:
1.1.1.4 ! misho 6727: Consider the problem of matching a string in parentheses, allowing for
! 6728: unlimited nested parentheses. Without the use of recursion, the best
! 6729: that can be done is to use a pattern that matches up to some fixed
! 6730: depth of nesting. It is not possible to handle an arbitrary nesting
1.1 misho 6731: depth.
6732:
6733: For some time, Perl has provided a facility that allows regular expres-
1.1.1.4 ! misho 6734: sions to recurse (amongst other things). It does this by interpolating
! 6735: Perl code in the expression at run time, and the code can refer to the
1.1 misho 6736: expression itself. A Perl pattern using code interpolation to solve the
6737: parentheses problem can be created like this:
6738:
6739: $re = qr{\( (?: (?>[^()]+) | (?p{$re}) )* \)}x;
6740:
6741: The (?p{...}) item interpolates Perl code at run time, and in this case
6742: refers recursively to the pattern in which it appears.
6743:
6744: Obviously, PCRE cannot support the interpolation of Perl code. Instead,
1.1.1.4 ! misho 6745: it supports special syntax for recursion of the entire pattern, and
! 6746: also for individual subpattern recursion. After its introduction in
! 6747: PCRE and Python, this kind of recursion was subsequently introduced
1.1 misho 6748: into Perl at release 5.10.
6749:
1.1.1.4 ! misho 6750: A special item that consists of (? followed by a number greater than
! 6751: zero and a closing parenthesis is a recursive subroutine call of the
! 6752: subpattern of the given number, provided that it occurs inside that
! 6753: subpattern. (If not, it is a non-recursive subroutine call, which is
! 6754: described in the next section.) The special item (?R) or (?0) is a
1.1 misho 6755: recursive call of the entire regular expression.
6756:
1.1.1.4 ! misho 6757: This PCRE pattern solves the nested parentheses problem (assume the
1.1 misho 6758: PCRE_EXTENDED option is set so that white space is ignored):
6759:
6760: \( ( [^()]++ | (?R) )* \)
6761:
1.1.1.4 ! misho 6762: First it matches an opening parenthesis. Then it matches any number of
! 6763: substrings which can either be a sequence of non-parentheses, or a
! 6764: recursive match of the pattern itself (that is, a correctly parenthe-
1.1 misho 6765: sized substring). Finally there is a closing parenthesis. Note the use
6766: of a possessive quantifier to avoid backtracking into sequences of non-
6767: parentheses.
6768:
1.1.1.4 ! misho 6769: If this were part of a larger pattern, you would not want to recurse
1.1 misho 6770: the entire pattern, so instead you could use this:
6771:
6772: ( \( ( [^()]++ | (?1) )* \) )
6773:
1.1.1.4 ! misho 6774: We have put the pattern into parentheses, and caused the recursion to
1.1 misho 6775: refer to them instead of the whole pattern.
6776:
1.1.1.4 ! misho 6777: In a larger pattern, keeping track of parenthesis numbers can be
! 6778: tricky. This is made easier by the use of relative references. Instead
1.1 misho 6779: of (?1) in the pattern above you can write (?-2) to refer to the second
1.1.1.4 ! misho 6780: most recently opened parentheses preceding the recursion. In other
! 6781: words, a negative number counts capturing parentheses leftwards from
1.1 misho 6782: the point at which it is encountered.
6783:
1.1.1.4 ! misho 6784: It is also possible to refer to subsequently opened parentheses, by
! 6785: writing references such as (?+2). However, these cannot be recursive
! 6786: because the reference is not inside the parentheses that are refer-
! 6787: enced. They are always non-recursive subroutine calls, as described in
1.1 misho 6788: the next section.
6789:
1.1.1.4 ! misho 6790: An alternative approach is to use named parentheses instead. The Perl
! 6791: syntax for this is (?&name); PCRE's earlier syntax (?P>name) is also
1.1 misho 6792: supported. We could rewrite the above example as follows:
6793:
6794: (?<pn> \( ( [^()]++ | (?&pn) )* \) )
6795:
1.1.1.4 ! misho 6796: If there is more than one subpattern with the same name, the earliest
1.1 misho 6797: one is used.
6798:
1.1.1.4 ! misho 6799: This particular example pattern that we have been looking at contains
1.1 misho 6800: nested unlimited repeats, and so the use of a possessive quantifier for
6801: matching strings of non-parentheses is important when applying the pat-
1.1.1.4 ! misho 6802: tern to strings that do not match. For example, when this pattern is
1.1 misho 6803: applied to
6804:
6805: (aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa()
6806:
1.1.1.4 ! misho 6807: it yields "no match" quickly. However, if a possessive quantifier is
! 6808: not used, the match runs for a very long time indeed because there are
! 6809: so many different ways the + and * repeats can carve up the subject,
1.1 misho 6810: and all have to be tested before failure can be reported.
6811:
1.1.1.4 ! misho 6812: At the end of a match, the values of capturing parentheses are those
! 6813: from the outermost level. If you want to obtain intermediate values, a
! 6814: callout function can be used (see below and the pcrecallout documenta-
1.1 misho 6815: tion). If the pattern above is matched against
6816:
6817: (ab(cd)ef)
6818:
1.1.1.4 ! misho 6819: the value for the inner capturing parentheses (numbered 2) is "ef",
! 6820: which is the last value taken on at the top level. If a capturing sub-
! 6821: pattern is not matched at the top level, its final captured value is
! 6822: unset, even if it was (temporarily) set at a deeper level during the
1.1 misho 6823: matching process.
6824:
1.1.1.4 ! misho 6825: If there are more than 15 capturing parentheses in a pattern, PCRE has
! 6826: to obtain extra memory to store data during a recursion, which it does
1.1 misho 6827: by using pcre_malloc, freeing it via pcre_free afterwards. If no memory
6828: can be obtained, the match fails with the PCRE_ERROR_NOMEMORY error.
6829:
1.1.1.4 ! misho 6830: Do not confuse the (?R) item with the condition (R), which tests for
! 6831: recursion. Consider this pattern, which matches text in angle brack-
! 6832: ets, allowing for arbitrary nesting. Only digits are allowed in nested
! 6833: brackets (that is, when recursing), whereas any characters are permit-
1.1 misho 6834: ted at the outer level.
6835:
6836: < (?: (?(R) \d++ | [^<>]*+) | (?R)) * >
6837:
1.1.1.4 ! misho 6838: In this pattern, (?(R) is the start of a conditional subpattern, with
! 6839: two different alternatives for the recursive and non-recursive cases.
1.1 misho 6840: The (?R) item is the actual recursive call.
6841:
6842: Differences in recursion processing between PCRE and Perl
6843:
1.1.1.4 ! misho 6844: Recursion processing in PCRE differs from Perl in two important ways.
! 6845: In PCRE (like Python, but unlike Perl), a recursive subpattern call is
1.1 misho 6846: always treated as an atomic group. That is, once it has matched some of
6847: the subject string, it is never re-entered, even if it contains untried
1.1.1.4 ! misho 6848: alternatives and there is a subsequent matching failure. This can be
! 6849: illustrated by the following pattern, which purports to match a palin-
! 6850: dromic string that contains an odd number of characters (for example,
1.1 misho 6851: "a", "aba", "abcba", "abcdcba"):
6852:
6853: ^(.|(.)(?1)\2)$
6854:
6855: The idea is that it either matches a single character, or two identical
1.1.1.4 ! misho 6856: characters surrounding a sub-palindrome. In Perl, this pattern works;
! 6857: in PCRE it does not if the pattern is longer than three characters.
1.1 misho 6858: Consider the subject string "abcba":
6859:
1.1.1.4 ! misho 6860: At the top level, the first character is matched, but as it is not at
1.1 misho 6861: the end of the string, the first alternative fails; the second alterna-
6862: tive is taken and the recursion kicks in. The recursive call to subpat-
1.1.1.4 ! misho 6863: tern 1 successfully matches the next character ("b"). (Note that the
1.1 misho 6864: beginning and end of line tests are not part of the recursion).
6865:
1.1.1.4 ! misho 6866: Back at the top level, the next character ("c") is compared with what
! 6867: subpattern 2 matched, which was "a". This fails. Because the recursion
! 6868: is treated as an atomic group, there are now no backtracking points,
! 6869: and so the entire match fails. (Perl is able, at this point, to re-
! 6870: enter the recursion and try the second alternative.) However, if the
1.1 misho 6871: pattern is written with the alternatives in the other order, things are
6872: different:
6873:
6874: ^((.)(?1)\2|.)$
6875:
1.1.1.4 ! misho 6876: This time, the recursing alternative is tried first, and continues to
! 6877: recurse until it runs out of characters, at which point the recursion
! 6878: fails. But this time we do have another alternative to try at the
! 6879: higher level. That is the big difference: in the previous case the
1.1 misho 6880: remaining alternative is at a deeper recursion level, which PCRE cannot
6881: use.
6882:
1.1.1.4 ! misho 6883: To change the pattern so that it matches all palindromic strings, not
! 6884: just those with an odd number of characters, it is tempting to change
1.1 misho 6885: the pattern to this:
6886:
6887: ^((.)(?1)\2|.?)$
6888:
1.1.1.4 ! misho 6889: Again, this works in Perl, but not in PCRE, and for the same reason.
! 6890: When a deeper recursion has matched a single character, it cannot be
! 6891: entered again in order to match an empty string. The solution is to
! 6892: separate the two cases, and write out the odd and even cases as alter-
1.1 misho 6893: natives at the higher level:
6894:
6895: ^(?:((.)(?1)\2|)|((.)(?3)\4|.))
6896:
1.1.1.4 ! misho 6897: If you want to match typical palindromic phrases, the pattern has to
1.1 misho 6898: ignore all non-word characters, which can be done like this:
6899:
6900: ^\W*+(?:((.)\W*+(?1)\W*+\2|)|((.)\W*+(?3)\W*+\4|\W*+.\W*+))\W*+$
6901:
6902: If run with the PCRE_CASELESS option, this pattern matches phrases such
6903: as "A man, a plan, a canal: Panama!" and it works well in both PCRE and
1.1.1.4 ! misho 6904: Perl. Note the use of the possessive quantifier *+ to avoid backtrack-
! 6905: ing into sequences of non-word characters. Without this, PCRE takes a
! 6906: great deal longer (ten times or more) to match typical phrases, and
1.1 misho 6907: Perl takes so long that you think it has gone into a loop.
6908:
1.1.1.4 ! misho 6909: WARNING: The palindrome-matching patterns above work only if the sub-
! 6910: ject string does not start with a palindrome that is shorter than the
! 6911: entire string. For example, although "abcba" is correctly matched, if
! 6912: the subject is "ababa", PCRE finds the palindrome "aba" at the start,
! 6913: then fails at top level because the end of the string does not follow.
! 6914: Once again, it cannot jump back into the recursion to try other alter-
1.1 misho 6915: natives, so the entire match fails.
6916:
1.1.1.4 ! misho 6917: The second way in which PCRE and Perl differ in their recursion pro-
! 6918: cessing is in the handling of captured values. In Perl, when a subpat-
! 6919: tern is called recursively or as a subpattern (see the next section),
! 6920: it has no access to any values that were captured outside the recur-
! 6921: sion, whereas in PCRE these values can be referenced. Consider this
1.1 misho 6922: pattern:
6923:
6924: ^(.)(\1|a(?2))
6925:
1.1.1.4 ! misho 6926: In PCRE, this pattern matches "bab". The first capturing parentheses
! 6927: match "b", then in the second group, when the back reference \1 fails
! 6928: to match "b", the second alternative matches "a" and then recurses. In
! 6929: the recursion, \1 does now match "b" and so the whole match succeeds.
! 6930: In Perl, the pattern fails to match because inside the recursive call
1.1 misho 6931: \1 cannot access the externally set value.
6932:
6933:
6934: SUBPATTERNS AS SUBROUTINES
6935:
1.1.1.4 ! misho 6936: If the syntax for a recursive subpattern call (either by number or by
! 6937: name) is used outside the parentheses to which it refers, it operates
! 6938: like a subroutine in a programming language. The called subpattern may
! 6939: be defined before or after the reference. A numbered reference can be
1.1 misho 6940: absolute or relative, as in these examples:
6941:
6942: (...(absolute)...)...(?2)...
6943: (...(relative)...)...(?-1)...
6944: (...(?+1)...(relative)...
6945:
6946: An earlier example pointed out that the pattern
6947:
6948: (sens|respons)e and \1ibility
6949:
1.1.1.4 ! misho 6950: matches "sense and sensibility" and "response and responsibility", but
1.1 misho 6951: not "sense and responsibility". If instead the pattern
6952:
6953: (sens|respons)e and (?1)ibility
6954:
1.1.1.4 ! misho 6955: is used, it does match "sense and responsibility" as well as the other
! 6956: two strings. Another example is given in the discussion of DEFINE
1.1 misho 6957: above.
6958:
1.1.1.4 ! misho 6959: All subroutine calls, whether recursive or not, are always treated as
! 6960: atomic groups. That is, once a subroutine has matched some of the sub-
1.1 misho 6961: ject string, it is never re-entered, even if it contains untried alter-
1.1.1.4 ! misho 6962: natives and there is a subsequent matching failure. Any capturing
! 6963: parentheses that are set during the subroutine call revert to their
1.1 misho 6964: previous values afterwards.
6965:
1.1.1.4 ! misho 6966: Processing options such as case-independence are fixed when a subpat-
! 6967: tern is defined, so if it is used as a subroutine, such options cannot
1.1 misho 6968: be changed for different calls. For example, consider this pattern:
6969:
6970: (abc)(?i:(?-1))
6971:
1.1.1.4 ! misho 6972: It matches "abcabc". It does not match "abcABC" because the change of
1.1 misho 6973: processing option does not affect the called subpattern.
6974:
6975:
6976: ONIGURUMA SUBROUTINE SYNTAX
6977:
1.1.1.4 ! misho 6978: For compatibility with Oniguruma, the non-Perl syntax \g followed by a
1.1 misho 6979: name or a number enclosed either in angle brackets or single quotes, is
1.1.1.4 ! misho 6980: an alternative syntax for referencing a subpattern as a subroutine,
! 6981: possibly recursively. Here are two of the examples used above, rewrit-
1.1 misho 6982: ten using this syntax:
6983:
6984: (?<pn> \( ( (?>[^()]+) | \g<pn> )* \) )
6985: (sens|respons)e and \g'1'ibility
6986:
1.1.1.4 ! misho 6987: PCRE supports an extension to Oniguruma: if a number is preceded by a
1.1 misho 6988: plus or a minus sign it is taken as a relative reference. For example:
6989:
6990: (abc)(?i:\g<-1>)
6991:
1.1.1.4 ! misho 6992: Note that \g{...} (Perl syntax) and \g<...> (Oniguruma syntax) are not
! 6993: synonymous. The former is a back reference; the latter is a subroutine
1.1 misho 6994: call.
6995:
6996:
6997: CALLOUTS
6998:
6999: Perl has a feature whereby using the sequence (?{...}) causes arbitrary
1.1.1.4 ! misho 7000: Perl code to be obeyed in the middle of matching a regular expression.
1.1 misho 7001: This makes it possible, amongst other things, to extract different sub-
7002: strings that match the same pair of parentheses when there is a repeti-
7003: tion.
7004:
7005: PCRE provides a similar feature, but of course it cannot obey arbitrary
7006: Perl code. The feature is called "callout". The caller of PCRE provides
1.1.1.4 ! misho 7007: an external function by putting its entry point in the global variable
! 7008: pcre_callout (8-bit library) or pcre[16|32]_callout (16-bit or 32-bit
! 7009: library). By default, this variable contains NULL, which disables all
! 7010: calling out.
1.1 misho 7011:
1.1.1.3 misho 7012: Within a regular expression, (?C) indicates the points at which the
7013: external function is to be called. If you want to identify different
7014: callout points, you can put a number less than 256 after the letter C.
7015: The default value is zero. For example, this pattern has two callout
1.1 misho 7016: points:
7017:
7018: (?C1)abc(?C2)def
7019:
1.1.1.3 misho 7020: If the PCRE_AUTO_CALLOUT flag is passed to a compiling function, call-
7021: outs are automatically installed before each item in the pattern. They
1.1.1.4 ! misho 7022: are all numbered 255. If there is a conditional group in the pattern
! 7023: whose condition is an assertion, an additional callout is inserted just
! 7024: before the condition. An explicit callout may also be set at this posi-
! 7025: tion, as in this example:
! 7026:
! 7027: (?(?C9)(?=a)abc|def)
! 7028:
! 7029: Note that this applies only to assertion conditions, not to other types
! 7030: of condition.
1.1.1.2 misho 7031:
1.1.1.3 misho 7032: During matching, when PCRE reaches a callout point, the external func-
7033: tion is called. It is provided with the number of the callout, the
7034: position in the pattern, and, optionally, one item of data originally
7035: supplied by the caller of the matching function. The callout function
7036: may cause matching to proceed, to backtrack, or to fail altogether. A
7037: complete description of the interface to the callout function is given
1.1.1.2 misho 7038: in the pcrecallout documentation.
1.1 misho 7039:
7040:
7041: BACKTRACKING CONTROL
7042:
1.1.1.3 misho 7043: Perl 5.10 introduced a number of "Special Backtracking Control Verbs",
1.1.1.4 ! misho 7044: which are still described in the Perl documentation as "experimental
! 7045: and subject to change or removal in a future version of Perl". It goes
! 7046: on to say: "Their usage in production code should be noted to avoid
! 7047: problems during upgrades." The same remarks apply to the PCRE features
! 7048: described in this section.
1.1 misho 7049:
1.1.1.4 ! misho 7050: The new verbs make use of what was previously invalid syntax: an open-
1.1 misho 7051: ing parenthesis followed by an asterisk. They are generally of the form
1.1.1.4 ! misho 7052: (*VERB) or (*VERB:NAME). Some may take either form, possibly behaving
! 7053: differently depending on whether or not a name is present. A name is
1.1 misho 7054: any sequence of characters that does not include a closing parenthesis.
1.1.1.3 misho 7055: The maximum length of name is 255 in the 8-bit library and 65535 in the
1.1.1.4 ! misho 7056: 16-bit and 32-bit libraries. If the name is empty, that is, if the
! 7057: closing parenthesis immediately follows the colon, the effect is as if
! 7058: the colon were not there. Any number of these verbs may occur in a
! 7059: pattern.
! 7060:
! 7061: Since these verbs are specifically related to backtracking, most of
! 7062: them can be used only when the pattern is to be matched using one of
! 7063: the traditional matching functions, because these use a backtracking
! 7064: algorithm. With the exception of (*FAIL), which behaves like a failing
! 7065: negative assertion, the backtracking control verbs cause an error if
! 7066: encountered by a DFA matching function.
! 7067:
! 7068: The behaviour of these verbs in repeated groups, assertions, and in
! 7069: subpatterns called as subroutines (whether or not recursively) is docu-
! 7070: mented below.
1.1.1.3 misho 7071:
7072: Optimizations that affect backtracking verbs
1.1 misho 7073:
1.1.1.4 ! misho 7074: PCRE contains some optimizations that are used to speed up matching by
1.1 misho 7075: running some checks at the start of each match attempt. For example, it
1.1.1.4 ! misho 7076: may know the minimum length of matching subject, or that a particular
! 7077: character must be present. When one of these optimizations bypasses the
! 7078: running of a match, any included backtracking verbs will not, of
1.1 misho 7079: course, be processed. You can suppress the start-of-match optimizations
1.1.1.4 ! misho 7080: by setting the PCRE_NO_START_OPTIMIZE option when calling pcre_com-
1.1 misho 7081: pile() or pcre_exec(), or by starting the pattern with (*NO_START_OPT).
1.1.1.3 misho 7082: There is more discussion of this option in the section entitled "Option
7083: bits for pcre_exec()" in the pcreapi documentation.
1.1 misho 7084:
1.1.1.4 ! misho 7085: Experiments with Perl suggest that it too has similar optimizations,
1.1 misho 7086: sometimes leading to anomalous results.
7087:
7088: Verbs that act immediately
7089:
1.1.1.4 ! misho 7090: The following verbs act as soon as they are encountered. They may not
1.1 misho 7091: be followed by a name.
7092:
7093: (*ACCEPT)
7094:
1.1.1.4 ! misho 7095: This verb causes the match to end successfully, skipping the remainder
! 7096: of the pattern. However, when it is inside a subpattern that is called
! 7097: as a subroutine, only that subpattern is ended successfully. Matching
! 7098: then continues at the outer level. If (*ACCEPT) in triggered in a posi-
! 7099: tive assertion, the assertion succeeds; in a negative assertion, the
! 7100: assertion fails.
! 7101:
! 7102: If (*ACCEPT) is inside capturing parentheses, the data so far is cap-
! 7103: tured. For example:
1.1 misho 7104:
7105: A((?:A|B(*ACCEPT)|C)D)
7106:
1.1.1.4 ! misho 7107: This matches "AB", "AAD", or "ACD"; when it matches "AB", "B" is cap-
1.1 misho 7108: tured by the outer parentheses.
7109:
7110: (*FAIL) or (*F)
7111:
1.1.1.4 ! misho 7112: This verb causes a matching failure, forcing backtracking to occur. It
! 7113: is equivalent to (?!) but easier to read. The Perl documentation notes
! 7114: that it is probably useful only when combined with (?{}) or (??{}).
! 7115: Those are, of course, Perl features that are not present in PCRE. The
! 7116: nearest equivalent is the callout feature, as for example in this pat-
1.1 misho 7117: tern:
7118:
7119: a+(?C)(*FAIL)
7120:
1.1.1.4 ! misho 7121: A match with the string "aaaa" always fails, but the callout is taken
1.1 misho 7122: before each backtrack happens (in this example, 10 times).
7123:
7124: Recording which path was taken
7125:
1.1.1.4 ! misho 7126: There is one verb whose main purpose is to track how a match was
! 7127: arrived at, though it also has a secondary use in conjunction with
1.1 misho 7128: advancing the match starting point (see (*SKIP) below).
7129:
7130: (*MARK:NAME) or (*:NAME)
7131:
1.1.1.4 ! misho 7132: A name is always required with this verb. There may be as many
! 7133: instances of (*MARK) as you like in a pattern, and their names do not
1.1 misho 7134: have to be unique.
7135:
1.1.1.4 ! misho 7136: When a match succeeds, the name of the last-encountered (*MARK:NAME),
! 7137: (*PRUNE:NAME), or (*THEN:NAME) on the matching path is passed back to
! 7138: the caller as described in the section entitled "Extra data for
! 7139: pcre_exec()" in the pcreapi documentation. Here is an example of
! 7140: pcretest output, where the /K modifier requests the retrieval and out-
! 7141: putting of (*MARK) data:
1.1 misho 7142:
7143: re> /X(*MARK:A)Y|X(*MARK:B)Z/K
7144: data> XY
7145: 0: XY
7146: MK: A
7147: XZ
7148: 0: XZ
7149: MK: B
7150:
7151: The (*MARK) name is tagged with "MK:" in this output, and in this exam-
1.1.1.2 misho 7152: ple it indicates which of the two alternatives matched. This is a more
7153: efficient way of obtaining this information than putting each alterna-
1.1 misho 7154: tive in its own capturing parentheses.
7155:
1.1.1.4 ! misho 7156: If a verb with a name is encountered in a positive assertion that is
! 7157: true, the name is recorded and passed back if it is the last-encoun-
! 7158: tered. This does not happen for negative assertions or failing positive
! 7159: assertions.
1.1 misho 7160:
1.1.1.4 ! misho 7161: After a partial match or a failed match, the last encountered name in
! 7162: the entire match process is returned. For example:
1.1 misho 7163:
7164: re> /X(*MARK:A)Y|X(*MARK:B)Z/K
7165: data> XP
7166: No match, mark = B
7167:
1.1.1.4 ! misho 7168: Note that in this unanchored example the mark is retained from the
1.1.1.3 misho 7169: match attempt that started at the letter "X" in the subject. Subsequent
7170: match attempts starting at "P" and then with an empty string do not get
7171: as far as the (*MARK) item, but nevertheless do not reset it.
7172:
1.1.1.4 ! misho 7173: If you are interested in (*MARK) values after failed matches, you
! 7174: should probably set the PCRE_NO_START_OPTIMIZE option (see above) to
1.1.1.3 misho 7175: ensure that the match is always attempted.
1.1 misho 7176:
7177: Verbs that act after backtracking
7178:
7179: The following verbs do nothing when they are encountered. Matching con-
1.1.1.4 ! misho 7180: tinues with what follows, but if there is no subsequent match, causing
! 7181: a backtrack to the verb, a failure is forced. That is, backtracking
! 7182: cannot pass to the left of the verb. However, when one of these verbs
! 7183: appears inside an atomic group or an assertion that is true, its effect
! 7184: is confined to that group, because once the group has been matched,
! 7185: there is never any backtracking into it. In this situation, backtrack-
! 7186: ing can "jump back" to the left of the entire atomic group or asser-
! 7187: tion. (Remember also, as stated above, that this localization also
! 7188: applies in subroutine calls.)
1.1 misho 7189:
1.1.1.2 misho 7190: These verbs differ in exactly what kind of failure occurs when back-
1.1.1.4 ! misho 7191: tracking reaches them. The behaviour described below is what happens
! 7192: when the verb is not in a subroutine or an assertion. Subsequent sec-
! 7193: tions cover these special cases.
1.1 misho 7194:
7195: (*COMMIT)
7196:
1.1.1.2 misho 7197: This verb, which may not be followed by a name, causes the whole match
1.1.1.4 ! misho 7198: to fail outright if there is a later matching failure that causes back-
! 7199: tracking to reach it. Even if the pattern is unanchored, no further
! 7200: attempts to find a match by advancing the starting point take place. If
! 7201: (*COMMIT) is the only backtracking verb that is encountered, once it
! 7202: has been passed pcre_exec() is committed to finding a match at the cur-
! 7203: rent starting point, or not at all. For example:
1.1 misho 7204:
7205: a+(*COMMIT)b
7206:
1.1.1.4 ! misho 7207: This matches "xxaab" but not "aacaab". It can be thought of as a kind
1.1 misho 7208: of dynamic anchor, or "I've started, so I must finish." The name of the
1.1.1.4 ! misho 7209: most recently passed (*MARK) in the path is passed back when (*COMMIT)
1.1 misho 7210: forces a match failure.
7211:
1.1.1.4 ! misho 7212: If there is more than one backtracking verb in a pattern, a different
! 7213: one that follows (*COMMIT) may be triggered first, so merely passing
! 7214: (*COMMIT) during a match does not always guarantee that a match must be
! 7215: at this starting point.
! 7216:
1.1.1.2 misho 7217: Note that (*COMMIT) at the start of a pattern is not the same as an
7218: anchor, unless PCRE's start-of-match optimizations are turned off, as
1.1 misho 7219: shown in this pcretest example:
7220:
7221: re> /(*COMMIT)abc/
7222: data> xyzabc
7223: 0: abc
7224: xyzabc\Y
7225: No match
7226:
1.1.1.2 misho 7227: PCRE knows that any match must start with "a", so the optimization
7228: skips along the subject to "a" before running the first match attempt,
7229: which succeeds. When the optimization is disabled by the \Y escape in
1.1 misho 7230: the second subject, the match starts at "x" and so the (*COMMIT) causes
7231: it to fail without trying any other starting points.
7232:
7233: (*PRUNE) or (*PRUNE:NAME)
7234:
1.1.1.2 misho 7235: This verb causes the match to fail at the current starting position in
1.1.1.4 ! misho 7236: the subject if there is a later matching failure that causes backtrack-
! 7237: ing to reach it. If the pattern is unanchored, the normal "bumpalong"
! 7238: advance to the next starting character then happens. Backtracking can
! 7239: occur as usual to the left of (*PRUNE), before it is reached, or when
! 7240: matching to the right of (*PRUNE), but if there is no match to the
! 7241: right, backtracking cannot cross (*PRUNE). In simple cases, the use of
! 7242: (*PRUNE) is just an alternative to an atomic group or possessive quan-
! 7243: tifier, but there are some uses of (*PRUNE) that cannot be expressed in
! 7244: any other way. In an anchored pattern (*PRUNE) has the same effect as
! 7245: (*COMMIT).
! 7246:
! 7247: The behaviour of (*PRUNE:NAME) is the not the same as
! 7248: (*MARK:NAME)(*PRUNE). It is like (*MARK:NAME) in that the name is
! 7249: remembered for passing back to the caller. However, (*SKIP:NAME)
! 7250: searches only for names set with (*MARK).
1.1 misho 7251:
7252: (*SKIP)
7253:
1.1.1.4 ! misho 7254: This verb, when given without a name, is like (*PRUNE), except that if
! 7255: the pattern is unanchored, the "bumpalong" advance is not to the next
1.1 misho 7256: character, but to the position in the subject where (*SKIP) was encoun-
1.1.1.4 ! misho 7257: tered. (*SKIP) signifies that whatever text was matched leading up to
1.1 misho 7258: it cannot be part of a successful match. Consider:
7259:
7260: a+(*SKIP)b
7261:
1.1.1.4 ! misho 7262: If the subject is "aaaac...", after the first match attempt fails
! 7263: (starting at the first character in the string), the starting point
1.1 misho 7264: skips on to start the next attempt at "c". Note that a possessive quan-
1.1.1.4 ! misho 7265: tifer does not have the same effect as this example; although it would
! 7266: suppress backtracking during the first match attempt, the second
! 7267: attempt would start at the second character instead of skipping on to
1.1 misho 7268: "c".
7269:
7270: (*SKIP:NAME)
7271:
1.1.1.4 ! misho 7272: When (*SKIP) has an associated name, its behaviour is modified. When it
! 7273: is triggered, the previous path through the pattern is searched for the
! 7274: most recent (*MARK) that has the same name. If one is found, the
! 7275: "bumpalong" advance is to the subject position that corresponds to that
! 7276: (*MARK) instead of to where (*SKIP) was encountered. If no (*MARK) with
! 7277: a matching name is found, the (*SKIP) is ignored.
! 7278:
! 7279: Note that (*SKIP:NAME) searches only for names set by (*MARK:NAME). It
! 7280: ignores names that are set by (*PRUNE:NAME) or (*THEN:NAME).
1.1 misho 7281:
7282: (*THEN) or (*THEN:NAME)
7283:
1.1.1.4 ! misho 7284: This verb causes a skip to the next innermost alternative when back-
! 7285: tracking reaches it. That is, it cancels any further backtracking
! 7286: within the current alternative. Its name comes from the observation
! 7287: that it can be used for a pattern-based if-then-else block:
1.1 misho 7288:
7289: ( COND1 (*THEN) FOO | COND2 (*THEN) BAR | COND3 (*THEN) BAZ ) ...
7290:
1.1.1.2 misho 7291: If the COND1 pattern matches, FOO is tried (and possibly further items
7292: after the end of the group if FOO succeeds); on failure, the matcher
7293: skips to the second alternative and tries COND2, without backtracking
1.1.1.4 ! misho 7294: into COND1. If that succeeds and BAR fails, COND3 is tried. If subse-
! 7295: quently BAZ fails, there are no more alternatives, so there is a back-
! 7296: track to whatever came before the entire group. If (*THEN) is not
! 7297: inside an alternation, it acts like (*PRUNE).
! 7298:
! 7299: The behaviour of (*THEN:NAME) is the not the same as
! 7300: (*MARK:NAME)(*THEN). It is like (*MARK:NAME) in that the name is
! 7301: remembered for passing back to the caller. However, (*SKIP:NAME)
! 7302: searches only for names set with (*MARK).
! 7303:
! 7304: A subpattern that does not contain a | character is just a part of the
! 7305: enclosing alternative; it is not a nested alternation with only one
! 7306: alternative. The effect of (*THEN) extends beyond such a subpattern to
! 7307: the enclosing alternative. Consider this pattern, where A, B, etc. are
! 7308: complex pattern fragments that do not contain any | characters at this
! 7309: level:
1.1 misho 7310:
7311: A (B(*THEN)C) | D
7312:
1.1.1.2 misho 7313: If A and B are matched, but there is a failure in C, matching does not
1.1 misho 7314: backtrack into A; instead it moves to the next alternative, that is, D.
1.1.1.2 misho 7315: However, if the subpattern containing (*THEN) is given an alternative,
1.1 misho 7316: it behaves differently:
7317:
7318: A (B(*THEN)C | (*FAIL)) | D
7319:
1.1.1.2 misho 7320: The effect of (*THEN) is now confined to the inner subpattern. After a
1.1 misho 7321: failure in C, matching moves to (*FAIL), which causes the whole subpat-
1.1.1.2 misho 7322: tern to fail because there are no more alternatives to try. In this
1.1 misho 7323: case, matching does now backtrack into A.
7324:
1.1.1.4 ! misho 7325: Note that a conditional subpattern is not considered as having two
1.1.1.2 misho 7326: alternatives, because only one is ever used. In other words, the |
1.1 misho 7327: character in a conditional subpattern has a different meaning. Ignoring
7328: white space, consider:
7329:
7330: ^.*? (?(?=a) a | b(*THEN)c )
7331:
1.1.1.2 misho 7332: If the subject is "ba", this pattern does not match. Because .*? is
7333: ungreedy, it initially matches zero characters. The condition (?=a)
7334: then fails, the character "b" is matched, but "c" is not. At this
7335: point, matching does not backtrack to .*? as might perhaps be expected
7336: from the presence of the | character. The conditional subpattern is
1.1 misho 7337: part of the single alternative that comprises the whole pattern, and so
1.1.1.2 misho 7338: the match fails. (If there was a backtrack into .*?, allowing it to
1.1 misho 7339: match "b", the match would succeed.)
7340:
1.1.1.2 misho 7341: The verbs just described provide four different "strengths" of control
1.1 misho 7342: when subsequent matching fails. (*THEN) is the weakest, carrying on the
1.1.1.2 misho 7343: match at the next alternative. (*PRUNE) comes next, failing the match
7344: at the current starting position, but allowing an advance to the next
7345: character (for an unanchored pattern). (*SKIP) is similar, except that
1.1 misho 7346: the advance may be more than one character. (*COMMIT) is the strongest,
7347: causing the entire match to fail.
7348:
1.1.1.4 ! misho 7349: More than one backtracking verb
! 7350:
! 7351: If more than one backtracking verb is present in a pattern, the one
! 7352: that is backtracked onto first acts. For example, consider this pat-
! 7353: tern, where A, B, etc. are complex pattern fragments:
! 7354:
! 7355: (A(*COMMIT)B(*THEN)C|ABD)
! 7356:
! 7357: If A matches but B fails, the backtrack to (*COMMIT) causes the entire
! 7358: match to fail. However, if A and B match, but C fails, the backtrack to
! 7359: (*THEN) causes the next alternative (ABD) to be tried. This behaviour
! 7360: is consistent, but is not always the same as Perl's. It means that if
! 7361: two or more backtracking verbs appear in succession, all the the last
! 7362: of them has no effect. Consider this example:
! 7363:
! 7364: ...(*COMMIT)(*PRUNE)...
! 7365:
! 7366: If there is a matching failure to the right, backtracking onto (*PRUNE)
! 7367: cases it to be triggered, and its action is taken. There can never be a
! 7368: backtrack onto (*COMMIT).
! 7369:
! 7370: Backtracking verbs in repeated groups
! 7371:
! 7372: PCRE differs from Perl in its handling of backtracking verbs in
! 7373: repeated groups. For example, consider:
! 7374:
! 7375: /(a(*COMMIT)b)+ac/
! 7376:
! 7377: If the subject is "abac", Perl matches, but PCRE fails because the
! 7378: (*COMMIT) in the second repeat of the group acts.
! 7379:
! 7380: Backtracking verbs in assertions
! 7381:
! 7382: (*FAIL) in an assertion has its normal effect: it forces an immediate
! 7383: backtrack.
! 7384:
! 7385: (*ACCEPT) in a positive assertion causes the assertion to succeed with-
! 7386: out any further processing. In a negative assertion, (*ACCEPT) causes
! 7387: the assertion to fail without any further processing.
! 7388:
! 7389: The other backtracking verbs are not treated specially if they appear
! 7390: in a positive assertion. In particular, (*THEN) skips to the next
! 7391: alternative in the innermost enclosing group that has alternations,
! 7392: whether or not this is within the assertion.
! 7393:
! 7394: Negative assertions are, however, different, in order to ensure that
! 7395: changing a positive assertion into a negative assertion changes its
! 7396: result. Backtracking into (*COMMIT), (*SKIP), or (*PRUNE) causes a neg-
! 7397: ative assertion to be true, without considering any further alternative
! 7398: branches in the assertion. Backtracking into (*THEN) causes it to skip
! 7399: to the next enclosing alternative within the assertion (the normal be-
! 7400: haviour), but if the assertion does not have such an alternative,
! 7401: (*THEN) behaves like (*PRUNE).
! 7402:
! 7403: Backtracking verbs in subroutines
! 7404:
! 7405: These behaviours occur whether or not the subpattern is called recur-
! 7406: sively. Perl's treatment of subroutines is different in some cases.
! 7407:
! 7408: (*FAIL) in a subpattern called as a subroutine has its normal effect:
! 7409: it forces an immediate backtrack.
! 7410:
! 7411: (*ACCEPT) in a subpattern called as a subroutine causes the subroutine
! 7412: match to succeed without any further processing. Matching then contin-
! 7413: ues after the subroutine call.
! 7414:
! 7415: (*COMMIT), (*SKIP), and (*PRUNE) in a subpattern called as a subroutine
! 7416: cause the subroutine match to fail.
! 7417:
! 7418: (*THEN) skips to the next alternative in the innermost enclosing group
! 7419: within the subpattern that has alternatives. If there is no such group
! 7420: within the subpattern, (*THEN) causes the subroutine match to fail.
1.1 misho 7421:
7422:
7423: SEE ALSO
7424:
1.1.1.2 misho 7425: pcreapi(3), pcrecallout(3), pcrematching(3), pcresyntax(3), pcre(3),
1.1.1.4 ! misho 7426: pcre16(3), pcre32(3).
1.1 misho 7427:
7428:
7429: AUTHOR
7430:
7431: Philip Hazel
7432: University Computing Service
7433: Cambridge CB2 3QH, England.
7434:
7435:
7436: REVISION
7437:
1.1.1.4 ! misho 7438: Last updated: 26 April 2013
! 7439: Copyright (c) 1997-2013 University of Cambridge.
1.1 misho 7440: ------------------------------------------------------------------------------
7441:
7442:
1.1.1.4 ! misho 7443: PCRESYNTAX(3) Library Functions Manual PCRESYNTAX(3)
! 7444:
1.1 misho 7445:
7446:
7447: NAME
7448: PCRE - Perl-compatible regular expressions
7449:
7450: PCRE REGULAR EXPRESSION SYNTAX SUMMARY
7451:
7452: The full syntax and semantics of the regular expressions that are sup-
7453: ported by PCRE are described in the pcrepattern documentation. This
1.1.1.2 misho 7454: document contains a quick-reference summary of the syntax.
1.1 misho 7455:
7456:
7457: QUOTING
7458:
7459: \x where x is non-alphanumeric is a literal x
7460: \Q...\E treat enclosed characters as literal
7461:
7462:
7463: CHARACTERS
7464:
7465: \a alarm, that is, the BEL character (hex 07)
7466: \cx "control-x", where x is any ASCII character
7467: \e escape (hex 1B)
1.1.1.3 misho 7468: \f form feed (hex 0C)
1.1 misho 7469: \n newline (hex 0A)
7470: \r carriage return (hex 0D)
7471: \t tab (hex 09)
7472: \ddd character with octal code ddd, or backreference
7473: \xhh character with hex code hh
7474: \x{hhh..} character with hex code hhh..
7475:
7476:
7477: CHARACTER TYPES
7478:
7479: . any character except newline;
7480: in dotall mode, any character whatsoever
1.1.1.2 misho 7481: \C one data unit, even in UTF mode (best avoided)
1.1 misho 7482: \d a decimal digit
7483: \D a character that is not a decimal digit
1.1.1.3 misho 7484: \h a horizontal white space character
7485: \H a character that is not a horizontal white space character
1.1 misho 7486: \N a character that is not a newline
7487: \p{xx} a character with the xx property
7488: \P{xx} a character without the xx property
7489: \R a newline sequence
1.1.1.3 misho 7490: \s a white space character
7491: \S a character that is not a white space character
7492: \v a vertical white space character
7493: \V a character that is not a vertical white space character
1.1 misho 7494: \w a "word" character
7495: \W a "non-word" character
1.1.1.4 ! misho 7496: \X a Unicode extended grapheme cluster
1.1 misho 7497:
7498: In PCRE, by default, \d, \D, \s, \S, \w, and \W recognize only ASCII
1.1.1.2 misho 7499: characters, even in a UTF mode. However, this can be changed by setting
1.1 misho 7500: the PCRE_UCP option.
7501:
7502:
7503: GENERAL CATEGORY PROPERTIES FOR \p and \P
7504:
7505: C Other
7506: Cc Control
7507: Cf Format
7508: Cn Unassigned
7509: Co Private use
7510: Cs Surrogate
7511:
7512: L Letter
7513: Ll Lower case letter
7514: Lm Modifier letter
7515: Lo Other letter
7516: Lt Title case letter
7517: Lu Upper case letter
7518: L& Ll, Lu, or Lt
7519:
7520: M Mark
7521: Mc Spacing mark
7522: Me Enclosing mark
7523: Mn Non-spacing mark
7524:
7525: N Number
7526: Nd Decimal number
7527: Nl Letter number
7528: No Other number
7529:
7530: P Punctuation
7531: Pc Connector punctuation
7532: Pd Dash punctuation
7533: Pe Close punctuation
7534: Pf Final punctuation
7535: Pi Initial punctuation
7536: Po Other punctuation
7537: Ps Open punctuation
7538:
7539: S Symbol
7540: Sc Currency symbol
7541: Sk Modifier symbol
7542: Sm Mathematical symbol
7543: So Other symbol
7544:
7545: Z Separator
7546: Zl Line separator
7547: Zp Paragraph separator
7548: Zs Space separator
7549:
7550:
7551: PCRE SPECIAL CATEGORY PROPERTIES FOR \p and \P
7552:
7553: Xan Alphanumeric: union of properties L and N
7554: Xps POSIX space: property Z or tab, NL, VT, FF, CR
7555: Xsp Perl space: property Z or tab, NL, FF, CR
1.1.1.4 ! misho 7556: Xuc Univerally-named character: one that can be
! 7557: represented by a Universal Character Name
1.1 misho 7558: Xwd Perl word: property Xan or underscore
7559:
7560:
7561: SCRIPT NAMES FOR \p AND \P
7562:
1.1.1.3 misho 7563: Arabic, Armenian, Avestan, Balinese, Bamum, Batak, Bengali, Bopomofo,
7564: Brahmi, Braille, Buginese, Buhid, Canadian_Aboriginal, Carian, Chakma,
7565: Cham, Cherokee, Common, Coptic, Cuneiform, Cypriot, Cyrillic, Deseret,
7566: Devanagari, Egyptian_Hieroglyphs, Ethiopic, Georgian, Glagolitic,
7567: Gothic, Greek, Gujarati, Gurmukhi, Han, Hangul, Hanunoo, Hebrew, Hira-
7568: gana, Imperial_Aramaic, Inherited, Inscriptional_Pahlavi, Inscrip-
7569: tional_Parthian, Javanese, Kaithi, Kannada, Katakana, Kayah_Li,
7570: Kharoshthi, Khmer, Lao, Latin, Lepcha, Limbu, Linear_B, Lisu, Lycian,
7571: Lydian, Malayalam, Mandaic, Meetei_Mayek, Meroitic_Cursive,
7572: Meroitic_Hieroglyphs, Miao, Mongolian, Myanmar, New_Tai_Lue, Nko,
7573: Ogham, Old_Italic, Old_Persian, Old_South_Arabian, Old_Turkic,
7574: Ol_Chiki, Oriya, Osmanya, Phags_Pa, Phoenician, Rejang, Runic, Samari-
7575: tan, Saurashtra, Sharada, Shavian, Sinhala, Sora_Sompeng, Sundanese,
7576: Syloti_Nagri, Syriac, Tagalog, Tagbanwa, Tai_Le, Tai_Tham, Tai_Viet,
7577: Takri, Tamil, Telugu, Thaana, Thai, Tibetan, Tifinagh, Ugaritic, Vai,
7578: Yi.
1.1 misho 7579:
7580:
7581: CHARACTER CLASSES
7582:
7583: [...] positive character class
7584: [^...] negative character class
7585: [x-y] range (can be used for hex characters)
7586: [[:xxx:]] positive POSIX named set
7587: [[:^xxx:]] negative POSIX named set
7588:
7589: alnum alphanumeric
7590: alpha alphabetic
7591: ascii 0-127
7592: blank space or tab
7593: cntrl control character
7594: digit decimal digit
7595: graph printing, excluding space
7596: lower lower case letter
7597: print printing, including space
7598: punct printing, excluding alphanumeric
1.1.1.3 misho 7599: space white space
1.1 misho 7600: upper upper case letter
7601: word same as \w
7602: xdigit hexadecimal digit
7603:
7604: In PCRE, POSIX character set names recognize only ASCII characters by
7605: default, but some of them use Unicode properties if PCRE_UCP is set.
7606: You can use \Q...\E inside a character class.
7607:
7608:
7609: QUANTIFIERS
7610:
7611: ? 0 or 1, greedy
7612: ?+ 0 or 1, possessive
7613: ?? 0 or 1, lazy
7614: * 0 or more, greedy
7615: *+ 0 or more, possessive
7616: *? 0 or more, lazy
7617: + 1 or more, greedy
7618: ++ 1 or more, possessive
7619: +? 1 or more, lazy
7620: {n} exactly n
7621: {n,m} at least n, no more than m, greedy
7622: {n,m}+ at least n, no more than m, possessive
7623: {n,m}? at least n, no more than m, lazy
7624: {n,} n or more, greedy
7625: {n,}+ n or more, possessive
7626: {n,}? n or more, lazy
7627:
7628:
7629: ANCHORS AND SIMPLE ASSERTIONS
7630:
7631: \b word boundary
7632: \B not a word boundary
7633: ^ start of subject
7634: also after internal newline in multiline mode
7635: \A start of subject
7636: $ end of subject
7637: also before newline at end of subject
7638: also before internal newline in multiline mode
7639: \Z end of subject
7640: also before newline at end of subject
7641: \z end of subject
7642: \G first matching position in subject
7643:
7644:
7645: MATCH POINT RESET
7646:
7647: \K reset start of match
7648:
7649:
7650: ALTERNATION
7651:
7652: expr|expr|expr...
7653:
7654:
7655: CAPTURING
7656:
7657: (...) capturing group
7658: (?<name>...) named capturing group (Perl)
7659: (?'name'...) named capturing group (Perl)
7660: (?P<name>...) named capturing group (Python)
7661: (?:...) non-capturing group
7662: (?|...) non-capturing group; reset group numbers for
7663: capturing groups in each alternative
7664:
7665:
7666: ATOMIC GROUPS
7667:
7668: (?>...) atomic, non-capturing group
7669:
7670:
7671: COMMENT
7672:
7673: (?#....) comment (not nestable)
7674:
7675:
7676: OPTION SETTING
7677:
7678: (?i) caseless
7679: (?J) allow duplicate names
7680: (?m) multiline
7681: (?s) single line (dotall)
7682: (?U) default ungreedy (lazy)
7683: (?x) extended (ignore white space)
7684: (?-...) unset option(s)
7685:
7686: The following are recognized only at the start of a pattern or after
7687: one of the newline-setting options with similar syntax:
7688:
1.1.1.4 ! misho 7689: (*LIMIT_MATCH=d) set the match limit to d (decimal number)
! 7690: (*LIMIT_RECURSION=d) set the recursion limit to d (decimal number)
1.1 misho 7691: (*NO_START_OPT) no start-match optimization (PCRE_NO_START_OPTIMIZE)
1.1.1.2 misho 7692: (*UTF8) set UTF-8 mode: 8-bit library (PCRE_UTF8)
7693: (*UTF16) set UTF-16 mode: 16-bit library (PCRE_UTF16)
1.1.1.4 ! misho 7694: (*UTF32) set UTF-32 mode: 32-bit library (PCRE_UTF32)
! 7695: (*UTF) set appropriate UTF mode for the library in use
1.1 misho 7696: (*UCP) set PCRE_UCP (use Unicode properties for \d etc)
7697:
7698:
7699: LOOKAHEAD AND LOOKBEHIND ASSERTIONS
7700:
7701: (?=...) positive look ahead
7702: (?!...) negative look ahead
7703: (?<=...) positive look behind
7704: (?<!...) negative look behind
7705:
7706: Each top-level branch of a look behind must be of a fixed length.
7707:
7708:
7709: BACKREFERENCES
7710:
7711: \n reference by number (can be ambiguous)
7712: \gn reference by number
7713: \g{n} reference by number
7714: \g{-n} relative reference by number
7715: \k<name> reference by name (Perl)
7716: \k'name' reference by name (Perl)
7717: \g{name} reference by name (Perl)
7718: \k{name} reference by name (.NET)
7719: (?P=name) reference by name (Python)
7720:
7721:
7722: SUBROUTINE REFERENCES (POSSIBLY RECURSIVE)
7723:
7724: (?R) recurse whole pattern
7725: (?n) call subpattern by absolute number
7726: (?+n) call subpattern by relative number
7727: (?-n) call subpattern by relative number
7728: (?&name) call subpattern by name (Perl)
7729: (?P>name) call subpattern by name (Python)
7730: \g<name> call subpattern by name (Oniguruma)
7731: \g'name' call subpattern by name (Oniguruma)
7732: \g<n> call subpattern by absolute number (Oniguruma)
7733: \g'n' call subpattern by absolute number (Oniguruma)
7734: \g<+n> call subpattern by relative number (PCRE extension)
7735: \g'+n' call subpattern by relative number (PCRE extension)
7736: \g<-n> call subpattern by relative number (PCRE extension)
7737: \g'-n' call subpattern by relative number (PCRE extension)
7738:
7739:
7740: CONDITIONAL PATTERNS
7741:
7742: (?(condition)yes-pattern)
7743: (?(condition)yes-pattern|no-pattern)
7744:
7745: (?(n)... absolute reference condition
7746: (?(+n)... relative reference condition
7747: (?(-n)... relative reference condition
7748: (?(<name>)... named reference condition (Perl)
7749: (?('name')... named reference condition (Perl)
7750: (?(name)... named reference condition (PCRE)
7751: (?(R)... overall recursion condition
7752: (?(Rn)... specific group recursion condition
7753: (?(R&name)... specific recursion condition
7754: (?(DEFINE)... define subpattern for reference
7755: (?(assert)... assertion condition
7756:
7757:
7758: BACKTRACKING CONTROL
7759:
7760: The following act immediately they are reached:
7761:
7762: (*ACCEPT) force successful match
7763: (*FAIL) force backtrack; synonym (*F)
1.1.1.2 misho 7764: (*MARK:NAME) set name to be passed back; synonym (*:NAME)
1.1 misho 7765:
7766: The following act only when a subsequent match failure causes a back-
7767: track to reach them. They all force a match failure, but they differ in
7768: what happens afterwards. Those that advance the start-of-match point do
7769: so only if the pattern is not anchored.
7770:
7771: (*COMMIT) overall failure, no advance of starting point
7772: (*PRUNE) advance to next starting character
1.1.1.2 misho 7773: (*PRUNE:NAME) equivalent to (*MARK:NAME)(*PRUNE)
7774: (*SKIP) advance to current matching position
7775: (*SKIP:NAME) advance to position corresponding to an earlier
7776: (*MARK:NAME); if not found, the (*SKIP) is ignored
1.1 misho 7777: (*THEN) local failure, backtrack to next alternation
1.1.1.2 misho 7778: (*THEN:NAME) equivalent to (*MARK:NAME)(*THEN)
1.1 misho 7779:
7780:
7781: NEWLINE CONVENTIONS
7782:
7783: These are recognized only at the very start of the pattern or after a
1.1.1.4 ! misho 7784: (*BSR_...), (*UTF8), (*UTF16), (*UTF32) or (*UCP) option.
1.1 misho 7785:
7786: (*CR) carriage return only
7787: (*LF) linefeed only
7788: (*CRLF) carriage return followed by linefeed
7789: (*ANYCRLF) all three of the above
7790: (*ANY) any Unicode newline sequence
7791:
7792:
7793: WHAT \R MATCHES
7794:
7795: These are recognized only at the very start of the pattern or after a
1.1.1.2 misho 7796: (*...) option that sets the newline convention or a UTF or UCP mode.
1.1 misho 7797:
7798: (*BSR_ANYCRLF) CR, LF, or CRLF
7799: (*BSR_UNICODE) any Unicode newline sequence
7800:
7801:
7802: CALLOUTS
7803:
7804: (?C) callout
7805: (?Cn) callout with data n
7806:
7807:
7808: SEE ALSO
7809:
7810: pcrepattern(3), pcreapi(3), pcrecallout(3), pcrematching(3), pcre(3).
7811:
7812:
7813: AUTHOR
7814:
7815: Philip Hazel
7816: University Computing Service
7817: Cambridge CB2 3QH, England.
7818:
7819:
7820: REVISION
7821:
1.1.1.4 ! misho 7822: Last updated: 26 April 2013
! 7823: Copyright (c) 1997-2013 University of Cambridge.
1.1 misho 7824: ------------------------------------------------------------------------------
7825:
7826:
1.1.1.4 ! misho 7827: PCREUNICODE(3) Library Functions Manual PCREUNICODE(3)
! 7828:
1.1 misho 7829:
7830:
7831: NAME
7832: PCRE - Perl-compatible regular expressions
7833:
1.1.1.4 ! misho 7834: UTF-8, UTF-16, UTF-32, AND UNICODE PROPERTY SUPPORT
1.1 misho 7835:
1.1.1.4 ! misho 7836: As well as UTF-8 support, PCRE also supports UTF-16 (from release 8.30)
! 7837: and UTF-32 (from release 8.32), by means of two additional libraries.
! 7838: They can be built as well as, or instead of, the 8-bit library.
1.1.1.2 misho 7839:
7840:
7841: UTF-8 SUPPORT
1.1 misho 7842:
1.1.1.2 misho 7843: In order process UTF-8 strings, you must build PCRE's 8-bit library
7844: with UTF support, and, in addition, you must call pcre_compile() with
7845: the PCRE_UTF8 option flag, or the pattern must start with the sequence
1.1.1.4 ! misho 7846: (*UTF8) or (*UTF). When either of these is the case, both the pattern
! 7847: and any subject strings that are matched against it are treated as
! 7848: UTF-8 strings instead of strings of individual 1-byte characters.
! 7849:
! 7850:
! 7851: UTF-16 AND UTF-32 SUPPORT
! 7852:
! 7853: In order process UTF-16 or UTF-32 strings, you must build PCRE's 16-bit
! 7854: or 32-bit library with UTF support, and, in addition, you must call
! 7855: pcre16_compile() or pcre32_compile() with the PCRE_UTF16 or PCRE_UTF32
! 7856: option flag, as appropriate. Alternatively, the pattern must start with
! 7857: the sequence (*UTF16), (*UTF32), as appropriate, or (*UTF), which can
! 7858: be used with either library. When UTF mode is set, both the pattern and
! 7859: any subject strings that are matched against it are treated as UTF-16
! 7860: or UTF-32 strings instead of strings of individual 16-bit or 32-bit
! 7861: characters.
1.1.1.2 misho 7862:
7863:
7864: UTF SUPPORT OVERHEAD
7865:
1.1.1.4 ! misho 7866: If you compile PCRE with UTF support, but do not use it at run time,
! 7867: the library will be a bit bigger, but the additional run time overhead
! 7868: is limited to testing the PCRE_UTF[8|16|32] flag occasionally, so
! 7869: should not be very big.
1.1.1.2 misho 7870:
7871:
7872: UNICODE PROPERTY SUPPORT
1.1 misho 7873:
7874: If PCRE is built with Unicode character property support (which implies
1.1.1.4 ! misho 7875: UTF support), the escape sequences \p{..}, \P{..}, and \X can be used.
! 7876: The available properties that can be tested are limited to the general
! 7877: category properties such as Lu for an upper case letter or Nd for a
1.1.1.2 misho 7878: decimal number, the Unicode script names such as Arabic or Han, and the
1.1.1.4 ! misho 7879: derived properties Any and L&. Full lists is given in the pcrepattern
! 7880: and pcresyntax documentation. Only the short names for properties are
! 7881: supported. For example, \p{L} matches a letter. Its Perl synonym,
! 7882: \p{Letter}, is not supported. Furthermore, in Perl, many properties
! 7883: may optionally be prefixed by "Is", for compatibility with Perl 5.6.
! 7884: PCRE does not support this.
1.1 misho 7885:
7886: Validity of UTF-8 strings
7887:
1.1.1.4 ! misho 7888: When you set the PCRE_UTF8 flag, the byte strings passed as patterns
1.1.1.2 misho 7889: and subjects are (by default) checked for validity on entry to the rel-
1.1.1.3 misho 7890: evant functions. The entire string is checked before any other process-
1.1.1.4 ! misho 7891: ing takes place. From release 7.3 of PCRE, the check is according the
1.1.1.2 misho 7892: rules of RFC 3629, which are themselves derived from the Unicode speci-
1.1.1.4 ! misho 7893: fication. Earlier releases of PCRE followed the rules of RFC 2279,
! 7894: which allows the full range of 31-bit values (0 to 0x7FFFFFFF). The
! 7895: current check allows only values in the range U+0 to U+10FFFF, exclud-
! 7896: ing the surrogate area. (From release 8.33 the so-called "non-charac-
! 7897: ter" code points are no longer excluded because Unicode corrigendum #9
! 7898: makes it clear that they should not be.)
! 7899:
! 7900: Characters in the "Surrogate Area" of Unicode are reserved for use by
! 7901: UTF-16, where they are used in pairs to encode codepoints with values
! 7902: greater than 0xFFFF. The code points that are encoded by UTF-16 pairs
! 7903: are available independently in the UTF-8 and UTF-32 encodings. (In
! 7904: other words, the whole surrogate thing is a fudge for UTF-16 which
! 7905: unfortunately messes up UTF-8 and UTF-32.)
1.1 misho 7906:
7907: If an invalid UTF-8 string is passed to PCRE, an error return is given.
1.1.1.4 ! misho 7908: At compile time, the only additional information is the offset to the
1.1.1.3 misho 7909: first byte of the failing character. The run-time functions pcre_exec()
1.1.1.4 ! misho 7910: and pcre_dfa_exec() also pass back this information, as well as a more
! 7911: detailed reason code if the caller has provided memory in which to do
1.1 misho 7912: this.
7913:
1.1.1.4 ! misho 7914: In some situations, you may already know that your strings are valid,
! 7915: and therefore want to skip these checks in order to improve perfor-
! 7916: mance, for example in the case of a long subject string that is being
! 7917: scanned repeatedly. If you set the PCRE_NO_UTF8_CHECK flag at compile
! 7918: time or at run time, PCRE assumes that the pattern or subject it is
! 7919: given (respectively) contains only valid UTF-8 codes. In this case, it
! 7920: does not diagnose an invalid UTF-8 string.
! 7921:
! 7922: Note that passing PCRE_NO_UTF8_CHECK to pcre_compile() just disables
! 7923: the check for the pattern; it does not also apply to subject strings.
! 7924: If you want to disable the check for a subject string you must pass
! 7925: this option to pcre_exec() or pcre_dfa_exec().
! 7926:
! 7927: If you pass an invalid UTF-8 string when PCRE_NO_UTF8_CHECK is set, the
! 7928: result is undefined and your program may crash.
1.1 misho 7929:
1.1.1.2 misho 7930: Validity of UTF-16 strings
1.1 misho 7931:
1.1.1.2 misho 7932: When you set the PCRE_UTF16 flag, the strings of 16-bit data units that
7933: are passed as patterns and subjects are (by default) checked for valid-
1.1.1.4 ! misho 7934: ity on entry to the relevant functions. Values other than those in the
1.1.1.2 misho 7935: surrogate range U+D800 to U+DFFF are independent code points. Values in
7936: the surrogate range must be used in pairs in the correct manner.
7937:
1.1.1.4 ! misho 7938: If an invalid UTF-16 string is passed to PCRE, an error return is
! 7939: given. At compile time, the only additional information is the offset
1.1.1.3 misho 7940: to the first data unit of the failing character. The run-time functions
1.1.1.2 misho 7941: pcre16_exec() and pcre16_dfa_exec() also pass back this information, as
1.1.1.4 ! misho 7942: well as a more detailed reason code if the caller has provided memory
! 7943: in which to do this.
! 7944:
! 7945: In some situations, you may already know that your strings are valid,
! 7946: and therefore want to skip these checks in order to improve perfor-
! 7947: mance. If you set the PCRE_NO_UTF16_CHECK flag at compile time or at
! 7948: run time, PCRE assumes that the pattern or subject it is given (respec-
! 7949: tively) contains only valid UTF-16 sequences. In this case, it does not
! 7950: diagnose an invalid UTF-16 string. However, if an invalid string is
! 7951: passed, the result is undefined.
! 7952:
! 7953: Validity of UTF-32 strings
! 7954:
! 7955: When you set the PCRE_UTF32 flag, the strings of 32-bit data units that
! 7956: are passed as patterns and subjects are (by default) checked for valid-
! 7957: ity on entry to the relevant functions. This check allows only values
! 7958: in the range U+0 to U+10FFFF, excluding the surrogate area U+D800 to
! 7959: U+DFFF.
! 7960:
! 7961: If an invalid UTF-32 string is passed to PCRE, an error return is
! 7962: given. At compile time, the only additional information is the offset
! 7963: to the first data unit of the failing character. The run-time functions
! 7964: pcre32_exec() and pcre32_dfa_exec() also pass back this information, as
1.1.1.3 misho 7965: well as a more detailed reason code if the caller has provided memory
1.1.1.2 misho 7966: in which to do this.
7967:
1.1.1.3 misho 7968: In some situations, you may already know that your strings are valid,
7969: and therefore want to skip these checks in order to improve perfor-
1.1.1.4 ! misho 7970: mance. If you set the PCRE_NO_UTF32_CHECK flag at compile time or at
1.1.1.2 misho 7971: run time, PCRE assumes that the pattern or subject it is given (respec-
1.1.1.4 ! misho 7972: tively) contains only valid UTF-32 sequences. In this case, it does not
! 7973: diagnose an invalid UTF-32 string. However, if an invalid string is
! 7974: passed, the result is undefined.
1.1.1.2 misho 7975:
7976: General comments about UTF modes
7977:
1.1.1.4 ! misho 7978: 1. Codepoints less than 256 can be specified in patterns by either
! 7979: braced or unbraced hexadecimal escape sequences (for example, \x{b3} or
! 7980: \xb3). Larger values have to use braced sequences.
1.1.1.2 misho 7981:
1.1.1.4 ! misho 7982: 2. Octal numbers up to \777 are recognized, and in UTF-8 mode they
1.1.1.2 misho 7983: match two-byte characters for values greater than \177.
7984:
7985: 3. Repeat quantifiers apply to complete UTF characters, not to individ-
7986: ual data units, for example: \x{100}{3}.
7987:
1.1.1.4 ! misho 7988: 4. The dot metacharacter matches one UTF character instead of a single
1.1.1.2 misho 7989: data unit.
7990:
1.1.1.4 ! misho 7991: 5. The escape sequence \C can be used to match a single byte in UTF-8
! 7992: mode, or a single 16-bit data unit in UTF-16 mode, or a single 32-bit
! 7993: data unit in UTF-32 mode, but its use can lead to some strange effects
! 7994: because it breaks up multi-unit characters (see the description of \C
! 7995: in the pcrepattern documentation). The use of \C is not supported in
! 7996: the alternative matching function pcre[16|32]_dfa_exec(), nor is it
! 7997: supported in UTF mode by the JIT optimization of pcre[16|32]_exec(). If
! 7998: JIT optimization is requested for a UTF pattern that contains \C, it
! 7999: will not succeed, and so the matching will be carried out by the normal
! 8000: interpretive function.
1.1 misho 8001:
1.1.1.3 misho 8002: 6. The character escapes \b, \B, \d, \D, \s, \S, \w, and \W correctly
1.1 misho 8003: test characters of any code value, but, by default, the characters that
1.1.1.3 misho 8004: PCRE recognizes as digits, spaces, or word characters remain the same
8005: set as in non-UTF mode, all with values less than 256. This remains
8006: true even when PCRE is built to include Unicode property support,
1.1.1.2 misho 8007: because to do otherwise would slow down PCRE in many common cases. Note
1.1.1.3 misho 8008: in particular that this applies to \b and \B, because they are defined
1.1.1.2 misho 8009: in terms of \w and \W. If you really want to test for a wider sense of,
1.1.1.3 misho 8010: say, "digit", you can use explicit Unicode property tests such as
1.1.1.2 misho 8011: \p{Nd}. Alternatively, if you set the PCRE_UCP option, the way that the
1.1.1.3 misho 8012: character escapes work is changed so that Unicode properties are used
1.1.1.2 misho 8013: to determine which characters match. There are more details in the sec-
8014: tion on generic character types in the pcrepattern documentation.
1.1 misho 8015:
1.1.1.3 misho 8016: 7. Similarly, characters that match the POSIX named character classes
1.1 misho 8017: are all low-valued characters, unless the PCRE_UCP option is set.
8018:
1.1.1.3 misho 8019: 8. However, the horizontal and vertical white space matching escapes
8020: (\h, \H, \v, and \V) do match all the appropriate Unicode characters,
1.1 misho 8021: whether or not PCRE_UCP is set.
8022:
1.1.1.3 misho 8023: 9. Case-insensitive matching applies only to characters whose values
8024: are less than 128, unless PCRE is built with Unicode property support.
1.1.1.4 ! misho 8025: A few Unicode characters such as Greek sigma have more than two code-
! 8026: points that are case-equivalent. Up to and including PCRE release 8.31,
! 8027: only one-to-one case mappings were supported, but later releases (with
! 8028: Unicode property support) do treat as case-equivalent all versions of
! 8029: characters such as Greek sigma.
1.1 misho 8030:
8031:
8032: AUTHOR
8033:
8034: Philip Hazel
8035: University Computing Service
8036: Cambridge CB2 3QH, England.
8037:
8038:
8039: REVISION
8040:
1.1.1.4 ! misho 8041: Last updated: 27 February 2013
! 8042: Copyright (c) 1997-2013 University of Cambridge.
1.1 misho 8043: ------------------------------------------------------------------------------
8044:
8045:
1.1.1.4 ! misho 8046: PCREJIT(3) Library Functions Manual PCREJIT(3)
! 8047:
1.1 misho 8048:
8049:
8050: NAME
8051: PCRE - Perl-compatible regular expressions
8052:
8053: PCRE JUST-IN-TIME COMPILER SUPPORT
8054:
8055: Just-in-time compiling is a heavyweight optimization that can greatly
8056: speed up pattern matching. However, it comes at the cost of extra pro-
8057: cessing before the match is performed. Therefore, it is of most benefit
8058: when the same pattern is going to be matched many times. This does not
1.1.1.2 misho 8059: necessarily mean many calls of a matching function; if the pattern is
8060: not anchored, matching attempts may take place many times at various
8061: positions in the subject, even for a single call. Therefore, if the
1.1 misho 8062: subject string is very long, it may still pay to use JIT for one-off
8063: matches.
8064:
1.1.1.2 misho 8065: JIT support applies only to the traditional Perl-compatible matching
8066: function. It does not apply when the DFA matching function is being
8067: used. The code for this support was written by Zoltan Herczeg.
8068:
8069:
1.1.1.4 ! misho 8070: 8-BIT, 16-BIT AND 32-BIT SUPPORT
1.1.1.2 misho 8071:
1.1.1.4 ! misho 8072: JIT support is available for all of the 8-bit, 16-bit and 32-bit PCRE
! 8073: libraries. To keep this documentation simple, only the 8-bit interface
! 8074: is described in what follows. If you are using the 16-bit library, sub-
! 8075: stitute the 16-bit functions and 16-bit structures (for example,
! 8076: pcre16_jit_stack instead of pcre_jit_stack). If you are using the
! 8077: 32-bit library, substitute the 32-bit functions and 32-bit structures
! 8078: (for example, pcre32_jit_stack instead of pcre_jit_stack).
1.1 misho 8079:
8080:
8081: AVAILABILITY OF JIT SUPPORT
8082:
8083: JIT support is an optional feature of PCRE. The "configure" option
8084: --enable-jit (or equivalent CMake option) must be set when PCRE is
8085: built if you want to use JIT. The support is limited to the following
8086: hardware platforms:
8087:
8088: ARM v5, v7, and Thumb2
8089: Intel x86 32-bit and 64-bit
8090: MIPS 32-bit
1.1.1.2 misho 8091: Power PC 32-bit and 64-bit
1.1.1.4 ! misho 8092: SPARC 32-bit (experimental)
1.1 misho 8093:
1.1.1.3 misho 8094: If --enable-jit is set on an unsupported platform, compilation fails.
1.1 misho 8095:
8096: A program that is linked with PCRE 8.20 or later can tell if JIT sup-
8097: port is available by calling pcre_config() with the PCRE_CONFIG_JIT
8098: option. The result is 1 when JIT is available, and 0 otherwise. How-
8099: ever, a simple program does not need to check this in order to use JIT.
1.1.1.4 ! misho 8100: The normal API is implemented in a way that falls back to the interpre-
! 8101: tive code if JIT is not available. For programs that need the best pos-
! 8102: sible performance, there is also a "fast path" API that is JIT-spe-
! 8103: cific.
1.1 misho 8104:
8105: If your program may sometimes be linked with versions of PCRE that are
8106: older than 8.20, but you want to use JIT when it is available, you can
8107: test the values of PCRE_MAJOR and PCRE_MINOR, or the existence of a JIT
8108: macro such as PCRE_CONFIG_JIT, for compile-time control of your code.
8109:
8110:
8111: SIMPLE USE OF JIT
8112:
8113: You have to do two things to make use of the JIT support in the sim-
8114: plest way:
8115:
8116: (1) Call pcre_study() with the PCRE_STUDY_JIT_COMPILE option for
8117: each compiled pattern, and pass the resulting pcre_extra block to
8118: pcre_exec().
8119:
8120: (2) Use pcre_free_study() to free the pcre_extra block when it is
1.1.1.4 ! misho 8121: no longer needed, instead of just freeing it yourself. This
! 8122: ensures that
! 8123: any JIT data is also freed.
1.1 misho 8124:
1.1.1.4 ! misho 8125: For a program that may be linked with pre-8.20 versions of PCRE, you
1.1 misho 8126: can insert
8127:
8128: #ifndef PCRE_STUDY_JIT_COMPILE
8129: #define PCRE_STUDY_JIT_COMPILE 0
8130: #endif
8131:
1.1.1.4 ! misho 8132: so that no option is passed to pcre_study(), and then use something
1.1 misho 8133: like this to free the study data:
8134:
8135: #ifdef PCRE_CONFIG_JIT
8136: pcre_free_study(study_ptr);
8137: #else
8138: pcre_free(study_ptr);
8139: #endif
8140:
1.1.1.4 ! misho 8141: PCRE_STUDY_JIT_COMPILE requests the JIT compiler to generate code for
! 8142: complete matches. If you want to run partial matches using the
! 8143: PCRE_PARTIAL_HARD or PCRE_PARTIAL_SOFT options of pcre_exec(), you
! 8144: should set one or both of the following options in addition to, or
1.1.1.3 misho 8145: instead of, PCRE_STUDY_JIT_COMPILE when you call pcre_study():
8146:
8147: PCRE_STUDY_JIT_PARTIAL_HARD_COMPILE
8148: PCRE_STUDY_JIT_PARTIAL_SOFT_COMPILE
8149:
1.1.1.4 ! misho 8150: The JIT compiler generates different optimized code for each of the
! 8151: three modes (normal, soft partial, hard partial). When pcre_exec() is
! 8152: called, the appropriate code is run if it is available. Otherwise, the
1.1.1.3 misho 8153: pattern is matched using interpretive code.
8154:
1.1.1.4 ! misho 8155: In some circumstances you may need to call additional functions. These
! 8156: are described in the section entitled "Controlling the JIT stack"
1.1 misho 8157: below.
8158:
1.1.1.4 ! misho 8159: If JIT support is not available, PCRE_STUDY_JIT_COMPILE etc. are
1.1.1.3 misho 8160: ignored, and no JIT data is created. Otherwise, the compiled pattern is
1.1.1.4 ! misho 8161: passed to the JIT compiler, which turns it into machine code that exe-
! 8162: cutes much faster than the normal interpretive code. When pcre_exec()
! 8163: is passed a pcre_extra block containing a pointer to JIT code of the
! 8164: appropriate mode (normal or hard/soft partial), it obeys that code
! 8165: instead of running the interpreter. The result is identical, but the
1.1.1.3 misho 8166: compiled JIT code runs much faster.
1.1 misho 8167:
1.1.1.4 ! misho 8168: There are some pcre_exec() options that are not supported for JIT exe-
! 8169: cution. There are also some pattern items that JIT cannot handle.
! 8170: Details are given below. In both cases, execution automatically falls
! 8171: back to the interpretive code. If you want to know whether JIT was
! 8172: actually used for a particular match, you should arrange for a JIT
! 8173: callback function to be set up as described in the section entitled
! 8174: "Controlling the JIT stack" below, even if you do not need to supply a
! 8175: non-default JIT stack. Such a callback function is called whenever JIT
! 8176: code is about to be obeyed. If the execution options are not right for
1.1.1.3 misho 8177: JIT execution, the callback function is not obeyed.
1.1 misho 8178:
1.1.1.4 ! misho 8179: If the JIT compiler finds an unsupported item, no JIT data is gener-
! 8180: ated. You can find out if JIT execution is available after studying a
! 8181: pattern by calling pcre_fullinfo() with the PCRE_INFO_JIT option. A
! 8182: result of 1 means that JIT compilation was successful. A result of 0
1.1 misho 8183: means that JIT support is not available, or the pattern was not studied
1.1.1.4 ! misho 8184: with PCRE_STUDY_JIT_COMPILE etc., or the JIT compiler was not able to
1.1.1.3 misho 8185: handle the pattern.
1.1 misho 8186:
8187: Once a pattern has been studied, with or without JIT, it can be used as
8188: many times as you like for matching different subject strings.
8189:
8190:
8191: UNSUPPORTED OPTIONS AND PATTERN ITEMS
8192:
1.1.1.4 ! misho 8193: The only pcre_exec() options that are supported for JIT execution are
! 8194: PCRE_NO_UTF8_CHECK, PCRE_NO_UTF16_CHECK, PCRE_NO_UTF32_CHECK, PCRE_NOT-
! 8195: BOL, PCRE_NOTEOL, PCRE_NOTEMPTY, PCRE_NOTEMPTY_ATSTART, PCRE_PAR-
! 8196: TIAL_HARD, and PCRE_PARTIAL_SOFT.
! 8197:
! 8198: The only unsupported pattern items are \C (match a single data unit)
! 8199: when running in a UTF mode, and a callout immediately before an asser-
! 8200: tion condition in a conditional group.
1.1 misho 8201:
8202:
8203: RETURN VALUES FROM JIT EXECUTION
8204:
1.1.1.4 ! misho 8205: When a pattern is matched using JIT execution, the return values are
! 8206: the same as those given by the interpretive pcre_exec() code, with the
! 8207: addition of one new error code: PCRE_ERROR_JIT_STACKLIMIT. This means
! 8208: that the memory used for the JIT stack was insufficient. See "Control-
1.1 misho 8209: ling the JIT stack" below for a discussion of JIT stack usage. For com-
1.1.1.4 ! misho 8210: patibility with the interpretive pcre_exec() code, no more than two-
! 8211: thirds of the ovector argument is used for passing back captured sub-
1.1 misho 8212: strings.
8213:
1.1.1.4 ! misho 8214: The error code PCRE_ERROR_MATCHLIMIT is returned by the JIT code if
! 8215: searching a very large pattern tree goes on for too long, as it is in
! 8216: the same circumstance when JIT is not used, but the details of exactly
! 8217: what is counted are not the same. The PCRE_ERROR_RECURSIONLIMIT error
1.1 misho 8218: code is never returned by JIT execution.
8219:
8220:
8221: SAVING AND RESTORING COMPILED PATTERNS
8222:
1.1.1.4 ! misho 8223: The code that is generated by the JIT compiler is architecture-spe-
! 8224: cific, and is also position dependent. For those reasons it cannot be
! 8225: saved (in a file or database) and restored later like the bytecode and
! 8226: other data of a compiled pattern. Saving and restoring compiled pat-
! 8227: terns is not something many people do. More detail about this facility
! 8228: is given in the pcreprecompile documentation. It should be possible to
! 8229: run pcre_study() on a saved and restored pattern, and thereby recreate
! 8230: the JIT data, but because JIT compilation uses significant resources,
! 8231: it is probably not worth doing this; you might as well recompile the
1.1 misho 8232: original pattern.
8233:
8234:
8235: CONTROLLING THE JIT STACK
8236:
8237: When the compiled JIT code runs, it needs a block of memory to use as a
1.1.1.4 ! misho 8238: stack. By default, it uses 32K on the machine stack. However, some
! 8239: large or complicated patterns need more than this. The error
! 8240: PCRE_ERROR_JIT_STACKLIMIT is given when there is not enough stack.
! 8241: Three functions are provided for managing blocks of memory for use as
! 8242: JIT stacks. There is further discussion about the use of JIT stacks in
1.1 misho 8243: the section entitled "JIT stack FAQ" below.
8244:
1.1.1.4 ! misho 8245: The pcre_jit_stack_alloc() function creates a JIT stack. Its arguments
! 8246: are a starting size and a maximum size, and it returns a pointer to an
! 8247: opaque structure of type pcre_jit_stack, or NULL if there is an error.
! 8248: The pcre_jit_stack_free() function can be used to free a stack that is
! 8249: no longer needed. (For the technically minded: the address space is
1.1 misho 8250: allocated by mmap or VirtualAlloc.)
8251:
1.1.1.4 ! misho 8252: JIT uses far less memory for recursion than the interpretive code, and
! 8253: a maximum stack size of 512K to 1M should be more than enough for any
1.1 misho 8254: pattern.
8255:
1.1.1.4 ! misho 8256: The pcre_assign_jit_stack() function specifies which stack JIT code
1.1 misho 8257: should use. Its arguments are as follows:
8258:
8259: pcre_extra *extra
8260: pcre_jit_callback callback
8261: void *data
8262:
1.1.1.4 ! misho 8263: The extra argument must be the result of studying a pattern with
1.1.1.3 misho 8264: PCRE_STUDY_JIT_COMPILE etc. There are three cases for the values of the
1.1 misho 8265: other two options:
8266:
8267: (1) If callback is NULL and data is NULL, an internal 32K block
8268: on the machine stack is used.
8269:
8270: (2) If callback is NULL and data is not NULL, data must be
8271: a valid JIT stack, the result of calling pcre_jit_stack_alloc().
8272:
1.1.1.3 misho 8273: (3) If callback is not NULL, it must point to a function that is
8274: called with data as an argument at the start of matching, in
8275: order to set up a JIT stack. If the return from the callback
8276: function is NULL, the internal 32K stack is used; otherwise the
8277: return value must be a valid JIT stack, the result of calling
8278: pcre_jit_stack_alloc().
8279:
1.1.1.4 ! misho 8280: A callback function is obeyed whenever JIT code is about to be run; it
! 8281: is not obeyed when pcre_exec() is called with options that are incom-
1.1.1.3 misho 8282: patible for JIT execution. A callback function can therefore be used to
1.1.1.4 ! misho 8283: determine whether a match operation was executed by JIT or by the
1.1.1.3 misho 8284: interpreter.
8285:
8286: You may safely use the same JIT stack for more than one pattern (either
1.1.1.4 ! misho 8287: by assigning directly or by callback), as long as the patterns are all
! 8288: matched sequentially in the same thread. In a multithread application,
! 8289: if you do not specify a JIT stack, or if you assign or pass back NULL
! 8290: from a callback, that is thread-safe, because each thread has its own
! 8291: machine stack. However, if you assign or pass back a non-NULL JIT
! 8292: stack, this must be a different stack for each thread so that the
1.1.1.3 misho 8293: application is thread-safe.
8294:
1.1.1.4 ! misho 8295: Strictly speaking, even more is allowed. You can assign the same non-
! 8296: NULL stack to any number of patterns as long as they are not used for
! 8297: matching by multiple threads at the same time. For example, you can
! 8298: assign the same stack to all compiled patterns, and use a global mutex
! 8299: in the callback to wait until the stack is available for use. However,
1.1.1.3 misho 8300: this is an inefficient solution, and not recommended.
1.1 misho 8301:
1.1.1.4 ! misho 8302: This is a suggestion for how a multithreaded program that needs to set
1.1.1.3 misho 8303: up non-default JIT stacks might operate:
1.1 misho 8304:
8305: During thread initalization
8306: thread_local_var = pcre_jit_stack_alloc(...)
8307:
8308: During thread exit
8309: pcre_jit_stack_free(thread_local_var)
8310:
8311: Use a one-line callback function
8312: return thread_local_var
8313:
1.1.1.4 ! misho 8314: All the functions described in this section do nothing if JIT is not
! 8315: available, and pcre_assign_jit_stack() does nothing unless the extra
! 8316: argument is non-NULL and points to a pcre_extra block that is the
1.1.1.3 misho 8317: result of a successful study with PCRE_STUDY_JIT_COMPILE etc.
1.1 misho 8318:
8319:
8320: JIT STACK FAQ
8321:
8322: (1) Why do we need JIT stacks?
8323:
1.1.1.4 ! misho 8324: PCRE (and JIT) is a recursive, depth-first engine, so it needs a stack
! 8325: where the local data of the current node is pushed before checking its
1.1 misho 8326: child nodes. Allocating real machine stack on some platforms is diffi-
8327: cult. For example, the stack chain needs to be updated every time if we
1.1.1.4 ! misho 8328: extend the stack on PowerPC. Although it is possible, its updating
1.1 misho 8329: time overhead decreases performance. So we do the recursion in memory.
8330:
8331: (2) Why don't we simply allocate blocks of memory with malloc()?
8332:
1.1.1.4 ! misho 8333: Modern operating systems have a nice feature: they can reserve an
1.1 misho 8334: address space instead of allocating memory. We can safely allocate mem-
1.1.1.4 ! misho 8335: ory pages inside this address space, so the stack could grow without
1.1 misho 8336: moving memory data (this is important because of pointers). Thus we can
1.1.1.4 ! misho 8337: allocate 1M address space, and use only a single memory page (usually
! 8338: 4K) if that is enough. However, we can still grow up to 1M anytime if
1.1 misho 8339: needed.
8340:
8341: (3) Who "owns" a JIT stack?
8342:
8343: The owner of the stack is the user program, not the JIT studied pattern
1.1.1.4 ! misho 8344: or anything else. The user program must ensure that if a stack is used
! 8345: by pcre_exec(), (that is, it is assigned to the pattern currently run-
1.1 misho 8346: ning), that stack must not be used by any other threads (to avoid over-
8347: writing the same memory area). The best practice for multithreaded pro-
1.1.1.4 ! misho 8348: grams is to allocate a stack for each thread, and return this stack
1.1 misho 8349: through the JIT callback function.
8350:
8351: (4) When should a JIT stack be freed?
8352:
8353: You can free a JIT stack at any time, as long as it will not be used by
1.1.1.4 ! misho 8354: pcre_exec() again. When you assign the stack to a pattern, only a
! 8355: pointer is set. There is no reference counting or any other magic. You
! 8356: can free the patterns and stacks in any order, anytime. Just do not
! 8357: call pcre_exec() with a pattern pointing to an already freed stack, as
! 8358: that will cause SEGFAULT. (Also, do not free a stack currently used by
! 8359: pcre_exec() in another thread). You can also replace the stack for a
! 8360: pattern at any time. You can even free the previous stack before
1.1 misho 8361: assigning a replacement.
8362:
1.1.1.4 ! misho 8363: (5) Should I allocate/free a stack every time before/after calling
1.1 misho 8364: pcre_exec()?
8365:
1.1.1.4 ! misho 8366: No, because this is too costly in terms of resources. However, you
! 8367: could implement some clever idea which release the stack if it is not
! 8368: used in let's say two minutes. The JIT callback can help to achieve
! 8369: this without keeping a list of the currently JIT studied patterns.
1.1 misho 8370:
1.1.1.4 ! misho 8371: (6) OK, the stack is for long term memory allocation. But what happens
! 8372: if a pattern causes stack overflow with a stack of 1M? Is that 1M kept
1.1 misho 8373: until the stack is freed?
8374:
1.1.1.4 ! misho 8375: Especially on embedded sytems, it might be a good idea to release mem-
! 8376: ory sometimes without freeing the stack. There is no API for this at
! 8377: the moment. Probably a function call which returns with the currently
! 8378: allocated memory for any stack and another which allows releasing mem-
1.1 misho 8379: ory (shrinking the stack) would be a good idea if someone needs this.
8380:
8381: (7) This is too much of a headache. Isn't there any better solution for
8382: JIT stack handling?
8383:
1.1.1.4 ! misho 8384: No, thanks to Windows. If POSIX threads were used everywhere, we could
1.1 misho 8385: throw out this complicated API.
8386:
8387:
8388: EXAMPLE CODE
8389:
1.1.1.4 ! misho 8390: This is a single-threaded example that specifies a JIT stack without
1.1 misho 8391: using a callback.
8392:
8393: int rc;
8394: int ovector[30];
8395: pcre *re;
8396: pcre_extra *extra;
8397: pcre_jit_stack *jit_stack;
8398:
8399: re = pcre_compile(pattern, 0, &error, &erroffset, NULL);
8400: /* Check for errors */
8401: extra = pcre_study(re, PCRE_STUDY_JIT_COMPILE, &error);
8402: jit_stack = pcre_jit_stack_alloc(32*1024, 512*1024);
8403: /* Check for error (NULL) */
8404: pcre_assign_jit_stack(extra, NULL, jit_stack);
8405: rc = pcre_exec(re, extra, subject, length, 0, 0, ovector, 30);
8406: /* Check results */
8407: pcre_free(re);
8408: pcre_free_study(extra);
8409: pcre_jit_stack_free(jit_stack);
8410:
8411:
1.1.1.4 ! misho 8412: JIT FAST PATH API
! 8413:
! 8414: Because the API described above falls back to interpreted execution
! 8415: when JIT is not available, it is convenient for programs that are writ-
! 8416: ten for general use in many environments. However, calling JIT via
! 8417: pcre_exec() does have a performance impact. Programs that are written
! 8418: for use where JIT is known to be available, and which need the best
! 8419: possible performance, can instead use a "fast path" API to call JIT
! 8420: execution directly instead of calling pcre_exec() (obviously only for
! 8421: patterns that have been successfully studied by JIT).
! 8422:
! 8423: The fast path function is called pcre_jit_exec(), and it takes exactly
! 8424: the same arguments as pcre_exec(), plus one additional argument that
! 8425: must point to a JIT stack. The JIT stack arrangements described above
! 8426: do not apply. The return values are the same as for pcre_exec().
! 8427:
! 8428: When you call pcre_exec(), as well as testing for invalid options, a
! 8429: number of other sanity checks are performed on the arguments. For exam-
! 8430: ple, if the subject pointer is NULL, or its length is negative, an
! 8431: immediate error is given. Also, unless PCRE_NO_UTF[8|16|32] is set, a
! 8432: UTF subject string is tested for validity. In the interests of speed,
! 8433: these checks do not happen on the JIT fast path, and if invalid data is
! 8434: passed, the result is undefined.
! 8435:
! 8436: Bypassing the sanity checks and the pcre_exec() wrapping can give
! 8437: speedups of more than 10%.
! 8438:
! 8439:
1.1 misho 8440: SEE ALSO
8441:
8442: pcreapi(3)
8443:
8444:
8445: AUTHOR
8446:
8447: Philip Hazel (FAQ by Zoltan Herczeg)
8448: University Computing Service
8449: Cambridge CB2 3QH, England.
8450:
8451:
8452: REVISION
8453:
1.1.1.4 ! misho 8454: Last updated: 17 March 2013
! 8455: Copyright (c) 1997-2013 University of Cambridge.
1.1 misho 8456: ------------------------------------------------------------------------------
8457:
8458:
1.1.1.4 ! misho 8459: PCREPARTIAL(3) Library Functions Manual PCREPARTIAL(3)
! 8460:
1.1 misho 8461:
8462:
8463: NAME
8464: PCRE - Perl-compatible regular expressions
8465:
8466: PARTIAL MATCHING IN PCRE
8467:
1.1.1.2 misho 8468: In normal use of PCRE, if the subject string that is passed to a match-
8469: ing function matches as far as it goes, but is too short to match the
8470: entire pattern, PCRE_ERROR_NOMATCH is returned. There are circumstances
8471: where it might be helpful to distinguish this case from other cases in
8472: which there is no match.
1.1 misho 8473:
8474: Consider, for example, an application where a human is required to type
8475: in data for a field with specific formatting requirements. An example
8476: might be a date in the form ddmmmyy, defined by this pattern:
8477:
8478: ^\d?\d(jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)\d\d$
8479:
8480: If the application sees the user's keystrokes one by one, and can check
8481: that what has been typed so far is potentially valid, it is able to
8482: raise an error as soon as a mistake is made, by beeping and not
8483: reflecting the character that has been typed, for example. This immedi-
8484: ate feedback is likely to be a better user interface than a check that
8485: is delayed until the entire string has been entered. Partial matching
8486: can also be useful when the subject string is very long and is not all
8487: available at once.
8488:
8489: PCRE supports partial matching by means of the PCRE_PARTIAL_SOFT and
1.1.1.2 misho 8490: PCRE_PARTIAL_HARD options, which can be set when calling any of the
8491: matching functions. For backwards compatibility, PCRE_PARTIAL is a syn-
8492: onym for PCRE_PARTIAL_SOFT. The essential difference between the two
8493: options is whether or not a partial match is preferred to an alterna-
8494: tive complete match, though the details differ between the two types of
8495: matching function. If both options are set, PCRE_PARTIAL_HARD takes
8496: precedence.
8497:
1.1.1.3 misho 8498: If you want to use partial matching with just-in-time optimized code,
1.1.1.4 ! misho 8499: you must call pcre_study(), pcre16_study() or pcre32_study() with one
! 8500: or both of these options:
1.1.1.3 misho 8501:
8502: PCRE_STUDY_JIT_PARTIAL_SOFT_COMPILE
8503: PCRE_STUDY_JIT_PARTIAL_HARD_COMPILE
8504:
8505: PCRE_STUDY_JIT_COMPILE should also be set if you are going to run non-
8506: partial matches on the same pattern. If the appropriate JIT study mode
8507: has not been set for a match, the interpretive matching code is used.
8508:
8509: Setting a partial matching option disables two of PCRE's standard opti-
8510: mizations. PCRE remembers the last literal data unit in a pattern, and
8511: abandons matching immediately if it is not present in the subject
1.1.1.2 misho 8512: string. This optimization cannot be used for a subject string that
8513: might match only partially. If the pattern was studied, PCRE knows the
8514: minimum length of a matching string, and does not bother to run the
8515: matching function on shorter strings. This optimization is also dis-
1.1 misho 8516: abled for partial matching.
8517:
8518:
1.1.1.4 ! misho 8519: PARTIAL MATCHING USING pcre_exec() OR pcre[16|32]_exec()
1.1 misho 8520:
1.1.1.4 ! misho 8521: A partial match occurs during a call to pcre_exec() or
! 8522: pcre[16|32]_exec() when the end of the subject string is reached suc-
! 8523: cessfully, but matching cannot continue because more characters are
! 8524: needed. However, at least one character in the subject must have been
! 8525: inspected. This character need not form part of the final matched
! 8526: string; lookbehind assertions and the \K escape sequence provide ways
! 8527: of inspecting characters before the start of a matched substring. The
! 8528: requirement for inspecting at least one character exists because an
! 8529: empty string can always be matched; without such a restriction there
! 8530: would always be a partial match of an empty string at the end of the
! 8531: subject.
1.1.1.2 misho 8532:
1.1.1.4 ! misho 8533: If there are at least two slots in the offsets vector when a partial
! 8534: match is returned, the first slot is set to the offset of the earliest
1.1.1.2 misho 8535: character that was inspected. For convenience, the second offset points
8536: to the end of the subject so that a substring can easily be identified.
1.1.1.4 ! misho 8537: If there are at least three slots in the offsets vector, the third slot
! 8538: is set to the offset of the character where matching started.
1.1 misho 8539:
1.1.1.4 ! misho 8540: For the majority of patterns, the contents of the first and third slots
! 8541: will be the same. However, for patterns that contain lookbehind asser-
! 8542: tions, or begin with \b or \B, characters before the one where matching
! 8543: started may have been inspected while carrying out the match. For exam-
! 8544: ple, consider this pattern:
1.1 misho 8545:
8546: /(?<=abc)123/
8547:
8548: This pattern matches "123", but only if it is preceded by "abc". If the
1.1.1.4 ! misho 8549: subject string is "xyzabc12", the first two offsets after a partial
! 8550: match are for the substring "abc12", because all these characters were
! 8551: inspected. However, the third offset is set to 6, because that is the
! 8552: offset where matching began.
1.1 misho 8553:
8554: What happens when a partial match is identified depends on which of the
8555: two partial matching options are set.
8556:
1.1.1.4 ! misho 8557: PCRE_PARTIAL_SOFT WITH pcre_exec() OR pcre[16|32]_exec()
1.1 misho 8558:
1.1.1.4 ! misho 8559: If PCRE_PARTIAL_SOFT is set when pcre_exec() or pcre[16|32]_exec()
! 8560: identifies a partial match, the partial match is remembered, but match-
! 8561: ing continues as normal, and other alternatives in the pattern are
! 8562: tried. If no complete match can be found, PCRE_ERROR_PARTIAL is
! 8563: returned instead of PCRE_ERROR_NOMATCH.
! 8564:
! 8565: This option is "soft" because it prefers a complete match over a par-
! 8566: tial match. All the various matching items in a pattern behave as if
! 8567: the subject string is potentially complete. For example, \z, \Z, and $
! 8568: match at the end of the subject, as normal, and for \b and \B the end
1.1 misho 8569: of the subject is treated as a non-alphanumeric.
8570:
1.1.1.4 ! misho 8571: If there is more than one partial match, the first one that was found
1.1 misho 8572: provides the data that is returned. Consider this pattern:
8573:
8574: /123\w+X|dogY/
8575:
1.1.1.4 ! misho 8576: If this is matched against the subject string "abc123dog", both alter-
! 8577: natives fail to match, but the end of the subject is reached during
! 8578: matching, so PCRE_ERROR_PARTIAL is returned. The offsets are set to 3
! 8579: and 9, identifying "123dog" as the first partial match that was found.
! 8580: (In this example, there are two partial matches, because "dog" on its
1.1 misho 8581: own partially matches the second alternative.)
8582:
1.1.1.4 ! misho 8583: PCRE_PARTIAL_HARD WITH pcre_exec() OR pcre[16|32]_exec()
1.1 misho 8584:
1.1.1.4 ! misho 8585: If PCRE_PARTIAL_HARD is set for pcre_exec() or pcre[16|32]_exec(),
! 8586: PCRE_ERROR_PARTIAL is returned as soon as a partial match is found,
1.1.1.2 misho 8587: without continuing to search for possible complete matches. This option
8588: is "hard" because it prefers an earlier partial match over a later com-
1.1.1.4 ! misho 8589: plete match. For this reason, the assumption is made that the end of
! 8590: the supplied subject string may not be the true end of the available
1.1.1.2 misho 8591: data, and so, if \z, \Z, \b, \B, or $ are encountered at the end of the
1.1.1.4 ! misho 8592: subject, the result is PCRE_ERROR_PARTIAL, provided that at least one
1.1.1.2 misho 8593: character in the subject has been inspected.
8594:
8595: Setting PCRE_PARTIAL_HARD also affects the way UTF-8 and UTF-16 subject
1.1.1.4 ! misho 8596: strings are checked for validity. Normally, an invalid sequence causes
! 8597: the error PCRE_ERROR_BADUTF8 or PCRE_ERROR_BADUTF16. However, in the
! 8598: special case of a truncated character at the end of the subject,
! 8599: PCRE_ERROR_SHORTUTF8 or PCRE_ERROR_SHORTUTF16 is returned when
1.1.1.2 misho 8600: PCRE_PARTIAL_HARD is set.
1.1 misho 8601:
8602: Comparing hard and soft partial matching
8603:
1.1.1.4 ! misho 8604: The difference between the two partial matching options can be illus-
1.1 misho 8605: trated by a pattern such as:
8606:
8607: /dog(sbody)?/
8608:
1.1.1.4 ! misho 8609: This matches either "dog" or "dogsbody", greedily (that is, it prefers
! 8610: the longer string if possible). If it is matched against the string
! 8611: "dog" with PCRE_PARTIAL_SOFT, it yields a complete match for "dog".
1.1 misho 8612: However, if PCRE_PARTIAL_HARD is set, the result is PCRE_ERROR_PARTIAL.
1.1.1.4 ! misho 8613: On the other hand, if the pattern is made ungreedy the result is dif-
1.1 misho 8614: ferent:
8615:
8616: /dog(sbody)??/
8617:
1.1.1.4 ! misho 8618: In this case the result is always a complete match because that is
! 8619: found first, and matching never continues after finding a complete
1.1.1.2 misho 8620: match. It might be easier to follow this explanation by thinking of the
8621: two patterns like this:
1.1 misho 8622:
8623: /dog(sbody)?/ is the same as /dogsbody|dog/
8624: /dog(sbody)??/ is the same as /dog|dogsbody/
8625:
1.1.1.4 ! misho 8626: The second pattern will never match "dogsbody", because it will always
1.1.1.2 misho 8627: find the shorter match first.
1.1 misho 8628:
8629:
1.1.1.4 ! misho 8630: PARTIAL MATCHING USING pcre_dfa_exec() OR pcre[16|32]_dfa_exec()
1.1 misho 8631:
1.1.1.2 misho 8632: The DFA functions move along the subject string character by character,
1.1.1.4 ! misho 8633: without backtracking, searching for all possible matches simultane-
! 8634: ously. If the end of the subject is reached before the end of the pat-
! 8635: tern, there is the possibility of a partial match, again provided that
1.1.1.2 misho 8636: at least one character has been inspected.
1.1 misho 8637:
1.1.1.4 ! misho 8638: When PCRE_PARTIAL_SOFT is set, PCRE_ERROR_PARTIAL is returned only if
! 8639: there have been no complete matches. Otherwise, the complete matches
! 8640: are returned. However, if PCRE_PARTIAL_HARD is set, a partial match
! 8641: takes precedence over any complete matches. The portion of the string
! 8642: that was inspected when the longest partial match was found is set as
1.1 misho 8643: the first matching string, provided there are at least two slots in the
8644: offsets vector.
8645:
1.1.1.4 ! misho 8646: Because the DFA functions always search for all possible matches, and
! 8647: there is no difference between greedy and ungreedy repetition, their
! 8648: behaviour is different from the standard functions when PCRE_PAR-
! 8649: TIAL_HARD is set. Consider the string "dog" matched against the
1.1.1.2 misho 8650: ungreedy pattern shown above:
1.1 misho 8651:
8652: /dog(sbody)??/
8653:
1.1.1.4 ! misho 8654: Whereas the standard functions stop as soon as they find the complete
! 8655: match for "dog", the DFA functions also find the partial match for
1.1.1.2 misho 8656: "dogsbody", and so return that when PCRE_PARTIAL_HARD is set.
1.1 misho 8657:
8658:
8659: PARTIAL MATCHING AND WORD BOUNDARIES
8660:
1.1.1.4 ! misho 8661: If a pattern ends with one of sequences \b or \B, which test for word
! 8662: boundaries, partial matching with PCRE_PARTIAL_SOFT can give counter-
1.1 misho 8663: intuitive results. Consider this pattern:
8664:
8665: /\bcat\b/
8666:
8667: This matches "cat", provided there is a word boundary at either end. If
8668: the subject string is "the cat", the comparison of the final "t" with a
1.1.1.4 ! misho 8669: following character cannot take place, so a partial match is found.
! 8670: However, normal matching carries on, and \b matches at the end of the
! 8671: subject when the last character is a letter, so a complete match is
! 8672: found. The result, therefore, is not PCRE_ERROR_PARTIAL. Using
! 8673: PCRE_PARTIAL_HARD in this case does yield PCRE_ERROR_PARTIAL, because
1.1.1.2 misho 8674: then the partial match takes precedence.
1.1 misho 8675:
8676:
8677: FORMERLY RESTRICTED PATTERNS
8678:
8679: For releases of PCRE prior to 8.00, because of the way certain internal
1.1.1.4 ! misho 8680: optimizations were implemented in the pcre_exec() function, the
! 8681: PCRE_PARTIAL option (predecessor of PCRE_PARTIAL_SOFT) could not be
! 8682: used with all patterns. From release 8.00 onwards, the restrictions no
! 8683: longer apply, and partial matching with can be requested for any pat-
1.1.1.2 misho 8684: tern.
1.1 misho 8685:
8686: Items that were formerly restricted were repeated single characters and
1.1.1.4 ! misho 8687: repeated metasequences. If PCRE_PARTIAL was set for a pattern that did
! 8688: not conform to the restrictions, pcre_exec() returned the error code
! 8689: PCRE_ERROR_BADPARTIAL (-13). This error code is no longer in use. The
! 8690: PCRE_INFO_OKPARTIAL call to pcre_fullinfo() to find out if a compiled
1.1 misho 8691: pattern can be used for partial matching now always returns 1.
8692:
8693:
8694: EXAMPLE OF PARTIAL MATCHING USING PCRETEST
8695:
1.1.1.4 ! misho 8696: If the escape sequence \P is present in a pcretest data line, the
! 8697: PCRE_PARTIAL_SOFT option is used for the match. Here is a run of
1.1 misho 8698: pcretest that uses the date example quoted above:
8699:
8700: re> /^\d?\d(jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)\d\d$/
8701: data> 25jun04\P
8702: 0: 25jun04
8703: 1: jun
8704: data> 25dec3\P
8705: Partial match: 23dec3
8706: data> 3ju\P
8707: Partial match: 3ju
8708: data> 3juj\P
8709: No match
8710: data> j\P
8711: No match
8712:
1.1.1.4 ! misho 8713: The first data string is matched completely, so pcretest shows the
! 8714: matched substrings. The remaining four strings do not match the com-
1.1 misho 8715: plete pattern, but the first two are partial matches. Similar output is
1.1.1.2 misho 8716: obtained if DFA matching is used.
1.1 misho 8717:
1.1.1.4 ! misho 8718: If the escape sequence \P is present more than once in a pcretest data
1.1 misho 8719: line, the PCRE_PARTIAL_HARD option is set for the match.
8720:
8721:
1.1.1.4 ! misho 8722: MULTI-SEGMENT MATCHING WITH pcre_dfa_exec() OR pcre[16|32]_dfa_exec()
1.1 misho 8723:
1.1.1.4 ! misho 8724: When a partial match has been found using a DFA matching function, it
! 8725: is possible to continue the match by providing additional subject data
! 8726: and calling the function again with the same compiled regular expres-
! 8727: sion, this time setting the PCRE_DFA_RESTART option. You must pass the
1.1 misho 8728: same working space as before, because this is where details of the pre-
1.1.1.4 ! misho 8729: vious partial match are stored. Here is an example using pcretest,
! 8730: using the \R escape sequence to set the PCRE_DFA_RESTART option (\D
1.1.1.2 misho 8731: specifies the use of the DFA matching function):
1.1 misho 8732:
8733: re> /^\d?\d(jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)\d\d$/
8734: data> 23ja\P\D
8735: Partial match: 23ja
8736: data> n05\R\D
8737: 0: n05
8738:
1.1.1.4 ! misho 8739: The first call has "23ja" as the subject, and requests partial match-
! 8740: ing; the second call has "n05" as the subject for the continued
! 8741: (restarted) match. Notice that when the match is complete, only the
! 8742: last part is shown; PCRE does not retain the previously partially-
! 8743: matched string. It is up to the calling program to do that if it needs
1.1 misho 8744: to.
8745:
1.1.1.4 ! misho 8746: You can set the PCRE_PARTIAL_SOFT or PCRE_PARTIAL_HARD options with
! 8747: PCRE_DFA_RESTART to continue partial matching over multiple segments.
! 8748: This facility can be used to pass very long subject strings to the DFA
1.1.1.2 misho 8749: matching functions.
8750:
8751:
1.1.1.4 ! misho 8752: MULTI-SEGMENT MATCHING WITH pcre_exec() OR pcre[16|32]_exec()
1.1.1.2 misho 8753:
1.1.1.4 ! misho 8754: From release 8.00, the standard matching functions can also be used to
1.1.1.2 misho 8755: do multi-segment matching. Unlike the DFA functions, it is not possible
1.1.1.4 ! misho 8756: to restart the previous match with a new segment of data. Instead, new
1.1.1.2 misho 8757: data must be added to the previous subject string, and the entire match
1.1.1.4 ! misho 8758: re-run, starting from the point where the partial match occurred. Ear-
1.1.1.2 misho 8759: lier data can be discarded.
8760:
1.1.1.4 ! misho 8761: It is best to use PCRE_PARTIAL_HARD in this situation, because it does
! 8762: not treat the end of a segment as the end of the subject when matching
! 8763: \z, \Z, \b, \B, and $. Consider an unanchored pattern that matches
1.1.1.2 misho 8764: dates:
1.1 misho 8765:
8766: re> /\d?\d(jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)\d\d/
8767: data> The date is 23ja\P\P
8768: Partial match: 23ja
8769:
1.1.1.4 ! misho 8770: At this stage, an application could discard the text preceding "23ja",
! 8771: add on text from the next segment, and call the matching function
! 8772: again. Unlike the DFA matching functions, the entire matching string
! 8773: must always be available, and the complete matching process occurs for
1.1.1.2 misho 8774: each call, so more memory and more processing time is needed.
8775:
1.1.1.4 ! misho 8776: Note: If the pattern contains lookbehind assertions, or \K, or starts
1.1.1.2 misho 8777: with \b or \B, the string that is returned for a partial match includes
1.1.1.4 ! misho 8778: characters that precede the start of what would be returned for a com-
! 8779: plete match, because it contains all the characters that were inspected
! 8780: during the partial match.
1.1 misho 8781:
8782:
8783: ISSUES WITH MULTI-SEGMENT MATCHING
8784:
8785: Certain types of pattern may give problems with multi-segment matching,
8786: whichever matching function is used.
8787:
8788: 1. If the pattern contains a test for the beginning of a line, you need
1.1.1.3 misho 8789: to pass the PCRE_NOTBOL option when the subject string for any call
8790: does start at the beginning of a line. There is also a PCRE_NOTEOL
1.1 misho 8791: option, but in practice when doing multi-segment matching you should be
8792: using PCRE_PARTIAL_HARD, which includes the effect of PCRE_NOTEOL.
8793:
1.1.1.3 misho 8794: 2. Lookbehind assertions that have already been obeyed are catered for
8795: in the offsets that are returned for a partial match. However a lookbe-
8796: hind assertion later in the pattern could require even earlier charac-
8797: ters to be inspected. You can handle this case by using the
8798: PCRE_INFO_MAXLOOKBEHIND option of the pcre_fullinfo() or
1.1.1.4 ! misho 8799: pcre[16|32]_fullinfo() functions to obtain the length of the longest
! 8800: lookbehind in the pattern. This length is given in characters, not
! 8801: bytes. If you always retain at least that many characters before the
! 8802: partially matched string, all should be well. (Of course, near the
! 8803: start of the subject, fewer characters may be present; in that case all
! 8804: characters should be retained.)
! 8805:
! 8806: From release 8.33, there is a more accurate way of deciding which char-
! 8807: acters to retain. Instead of subtracting the length of the longest
! 8808: lookbehind from the earliest inspected character (offsets[0]), the
! 8809: match start position (offsets[2]) should be used, and the next match
! 8810: attempt started at the offsets[2] character by setting the startoffset
! 8811: argument of pcre_exec() or pcre_dfa_exec().
! 8812:
! 8813: For example, if the pattern "(?<=123)abc" is partially matched against
! 8814: the string "xx123a", the three offset values returned are 2, 6, and 5.
! 8815: This indicates that the matching process that gave a partial match
! 8816: started at offset 5, but the characters "123a" were all inspected. The
! 8817: maximum lookbehind for that pattern is 3, so taking that away from 5
! 8818: shows that we need only keep "123a", and the next match attempt can be
! 8819: started at offset 3 (that is, at "a") when further characters have been
! 8820: added. When the match start is not the earliest inspected character,
! 8821: pcretest shows it explicitly:
! 8822:
! 8823: re> "(?<=123)abc"
! 8824: data> xx123a\P\P
! 8825: Partial match at offset 5: 123a
1.1.1.3 misho 8826:
1.1.1.4 ! misho 8827: 3. Because a partial match must always contain at least one character,
! 8828: what might be considered a partial match of an empty string actually
1.1.1.3 misho 8829: gives a "no match" result. For example:
8830:
8831: re> /c(?<=abc)x/
8832: data> ab\P
8833: No match
8834:
8835: If the next segment begins "cx", a match should be found, but this will
1.1.1.4 ! misho 8836: only happen if characters from the previous segment are retained. For
! 8837: this reason, a "no match" result should be interpreted as "partial
1.1.1.3 misho 8838: match of an empty string" when the pattern contains lookbehinds.
1.1 misho 8839:
1.1.1.4 ! misho 8840: 4. Matching a subject string that is split into multiple segments may
! 8841: not always produce exactly the same result as matching over one single
! 8842: long string, especially when PCRE_PARTIAL_SOFT is used. The section
! 8843: "Partial Matching and Word Boundaries" above describes an issue that
! 8844: arises if the pattern ends with \b or \B. Another kind of difference
! 8845: may occur when there are multiple matching possibilities, because (for
! 8846: PCRE_PARTIAL_SOFT) a partial match result is given only when there are
1.1 misho 8847: no completed matches. This means that as soon as the shortest match has
1.1.1.4 ! misho 8848: been found, continuation to a new subject segment is no longer possi-
1.1 misho 8849: ble. Consider again this pcretest example:
8850:
8851: re> /dog(sbody)?/
8852: data> dogsb\P
8853: 0: dog
8854: data> do\P\D
8855: Partial match: do
8856: data> gsb\R\P\D
8857: 0: g
8858: data> dogsbody\D
8859: 0: dogsbody
8860: 1: dog
8861:
1.1.1.4 ! misho 8862: The first data line passes the string "dogsb" to a standard matching
! 8863: function, setting the PCRE_PARTIAL_SOFT option. Although the string is
! 8864: a partial match for "dogsbody", the result is not PCRE_ERROR_PARTIAL,
! 8865: because the shorter string "dog" is a complete match. Similarly, when
! 8866: the subject is presented to a DFA matching function in several parts
! 8867: ("do" and "gsb" being the first two) the match stops when "dog" has
! 8868: been found, and it is not possible to continue. On the other hand, if
! 8869: "dogsbody" is presented as a single string, a DFA matching function
1.1.1.2 misho 8870: finds both matches.
1.1 misho 8871:
1.1.1.4 ! misho 8872: Because of these problems, it is best to use PCRE_PARTIAL_HARD when
! 8873: matching multi-segment data. The example above then behaves differ-
1.1 misho 8874: ently:
8875:
8876: re> /dog(sbody)?/
8877: data> dogsb\P\P
8878: Partial match: dogsb
8879: data> do\P\D
8880: Partial match: do
8881: data> gsb\R\P\P\D
8882: Partial match: gsb
8883:
1.1.1.3 misho 8884: 5. Patterns that contain alternatives at the top level which do not all
1.1.1.4 ! misho 8885: start with the same pattern item may not work as expected when
1.1.1.2 misho 8886: PCRE_DFA_RESTART is used. For example, consider this pattern:
1.1 misho 8887:
8888: 1234|3789
8889:
1.1.1.4 ! misho 8890: If the first part of the subject is "ABC123", a partial match of the
! 8891: first alternative is found at offset 3. There is no partial match for
1.1 misho 8892: the second alternative, because such a match does not start at the same
1.1.1.4 ! misho 8893: point in the subject string. Attempting to continue with the string
! 8894: "7890" does not yield a match because only those alternatives that
! 8895: match at one point in the subject are remembered. The problem arises
! 8896: because the start of the second alternative matches within the first
! 8897: alternative. There is no problem with anchored patterns or patterns
1.1 misho 8898: such as:
8899:
8900: 1234|ABCD
8901:
1.1.1.4 ! misho 8902: where no string can be a partial match for both alternatives. This is
! 8903: not a problem if a standard matching function is used, because the
1.1.1.2 misho 8904: entire match has to be rerun each time:
1.1 misho 8905:
8906: re> /1234|3789/
8907: data> ABC123\P\P
8908: Partial match: 123
8909: data> 1237890
8910: 0: 3789
8911:
8912: Of course, instead of using PCRE_DFA_RESTART, the same technique of re-
1.1.1.4 ! misho 8913: running the entire match can also be used with the DFA matching func-
! 8914: tions. Another possibility is to work with two buffers. If a partial
! 8915: match at offset n in the first buffer is followed by "no match" when
! 8916: PCRE_DFA_RESTART is used on the second buffer, you can then try a new
1.1.1.2 misho 8917: match starting at offset n+1 in the first buffer.
1.1 misho 8918:
8919:
8920: AUTHOR
8921:
8922: Philip Hazel
8923: University Computing Service
8924: Cambridge CB2 3QH, England.
8925:
8926:
8927: REVISION
8928:
1.1.1.4 ! misho 8929: Last updated: 20 February 2013
! 8930: Copyright (c) 1997-2013 University of Cambridge.
1.1 misho 8931: ------------------------------------------------------------------------------
8932:
8933:
1.1.1.4 ! misho 8934: PCREPRECOMPILE(3) Library Functions Manual PCREPRECOMPILE(3)
! 8935:
1.1 misho 8936:
8937:
8938: NAME
8939: PCRE - Perl-compatible regular expressions
8940:
8941: SAVING AND RE-USING PRECOMPILED PCRE PATTERNS
8942:
8943: If you are running an application that uses a large number of regular
8944: expression patterns, it may be useful to store them in a precompiled
8945: form instead of having to compile them every time the application is
8946: run. If you are not using any private character tables (see the
8947: pcre_maketables() documentation), this is relatively straightforward.
8948: If you are using private tables, it is a little bit more complicated.
1.1.1.2 misho 8949: However, if you are using the just-in-time optimization feature, it is
8950: not possible to save and reload the JIT data.
1.1 misho 8951:
8952: If you save compiled patterns to a file, you can copy them to a differ-
1.1.1.2 misho 8953: ent host and run them there. If the two hosts have different endianness
1.1.1.4 ! misho 8954: (byte order), you should run the pcre[16|32]_pat-
! 8955: tern_to_host_byte_order() function on the new host before trying to
! 8956: match the pattern. The matching functions return PCRE_ERROR_BADENDIAN-
! 8957: NESS if they detect a pattern with the wrong endianness.
1.1.1.2 misho 8958:
8959: Compiling regular expressions with one version of PCRE for use with a
8960: different version is not guaranteed to work and may cause crashes, and
8961: saving and restoring a compiled pattern loses any JIT optimization
8962: data.
1.1 misho 8963:
8964:
8965: SAVING A COMPILED PATTERN
8966:
1.1.1.4 ! misho 8967: The value returned by pcre[16|32]_compile() points to a single block of
1.1.1.2 misho 8968: memory that holds the compiled pattern and associated data. You can
1.1.1.4 ! misho 8969: find the length of this block in bytes by calling
! 8970: pcre[16|32]_fullinfo() with an argument of PCRE_INFO_SIZE. You can then
! 8971: save the data in any appropriate manner. Here is sample code for the
! 8972: 8-bit library that compiles a pattern and writes it to a file. It
! 8973: assumes that the variable fd refers to a file that is open for output:
1.1 misho 8974:
8975: int erroroffset, rc, size;
8976: char *error;
8977: pcre *re;
8978:
8979: re = pcre_compile("my pattern", 0, &error, &erroroffset, NULL);
8980: if (re == NULL) { ... handle errors ... }
8981: rc = pcre_fullinfo(re, NULL, PCRE_INFO_SIZE, &size);
8982: if (rc < 0) { ... handle errors ... }
8983: rc = fwrite(re, 1, size, fd);
8984: if (rc != size) { ... handle errors ... }
8985:
1.1.1.2 misho 8986: In this example, the bytes that comprise the compiled pattern are
8987: copied exactly. Note that this is binary data that may contain any of
8988: the 256 possible byte values. On systems that make a distinction
1.1 misho 8989: between binary and non-binary data, be sure that the file is opened for
8990: binary output.
8991:
1.1.1.2 misho 8992: If you want to write more than one pattern to a file, you will have to
8993: devise a way of separating them. For binary data, preceding each pat-
8994: tern with its length is probably the most straightforward approach.
8995: Another possibility is to write out the data in hexadecimal instead of
1.1 misho 8996: binary, one pattern to a line.
8997:
1.1.1.2 misho 8998: Saving compiled patterns in a file is only one possible way of storing
8999: them for later use. They could equally well be saved in a database, or
9000: in the memory of some daemon process that passes them via sockets to
1.1 misho 9001: the processes that want them.
9002:
9003: If the pattern has been studied, it is also possible to save the normal
9004: study data in a similar way to the compiled pattern itself. However, if
9005: the PCRE_STUDY_JIT_COMPILE was used, the just-in-time data that is cre-
1.1.1.2 misho 9006: ated cannot be saved because it is too dependent on the current envi-
9007: ronment. When studying generates additional information,
1.1.1.4 ! misho 9008: pcre[16|32]_study() returns a pointer to a pcre[16|32]_extra data
! 9009: block. Its format is defined in the section on matching a pattern in
! 9010: the pcreapi documentation. The study_data field points to the binary
! 9011: study data, and this is what you must save (not the pcre[16|32]_extra
! 9012: block itself). The length of the study data can be obtained by calling
! 9013: pcre[16|32]_fullinfo() with an argument of PCRE_INFO_STUDYSIZE. Remem-
! 9014: ber to check that pcre[16|32]_study() did return a non-NULL value
! 9015: before trying to save the study data.
1.1 misho 9016:
9017:
9018: RE-USING A PRECOMPILED PATTERN
9019:
9020: Re-using a precompiled pattern is straightforward. Having reloaded it
1.1.1.4 ! misho 9021: into main memory, called pcre[16|32]_pattern_to_host_byte_order() if
! 9022: necessary, you pass its pointer to pcre[16|32]_exec() or
! 9023: pcre[16|32]_dfa_exec() in the usual way.
1.1.1.2 misho 9024:
9025: However, if you passed a pointer to custom character tables when the
1.1.1.4 ! misho 9026: pattern was compiled (the tableptr argument of pcre[16|32]_compile()),
! 9027: you must now pass a similar pointer to pcre[16|32]_exec() or
! 9028: pcre[16|32]_dfa_exec(), because the value saved with the compiled pat-
! 9029: tern will obviously be nonsense. A field in a pcre[16|32]_extra() block
! 9030: is used to pass this data, as described in the section on matching a
! 9031: pattern in the pcreapi documentation.
1.1.1.2 misho 9032:
9033: If you did not provide custom character tables when the pattern was
9034: compiled, the pointer in the compiled pattern is NULL, which causes the
9035: matching functions to use PCRE's internal tables. Thus, you do not need
9036: to take any special action at run time in this case.
9037:
9038: If you saved study data with the compiled pattern, you need to create
1.1.1.4 ! misho 9039: your own pcre[16|32]_extra data block and set the study_data field to
1.1.1.2 misho 9040: point to the reloaded study data. You must also set the
9041: PCRE_EXTRA_STUDY_DATA bit in the flags field to indicate that study
1.1.1.4 ! misho 9042: data is present. Then pass the pcre[16|32]_extra block to the matching
1.1.1.2 misho 9043: function in the usual way. If the pattern was studied for just-in-time
9044: optimization, that data cannot be saved, and so is lost by a
9045: save/restore cycle.
1.1 misho 9046:
9047:
9048: COMPATIBILITY WITH DIFFERENT PCRE RELEASES
9049:
9050: In general, it is safest to recompile all saved patterns when you
9051: update to a new PCRE release, though not all updates actually require
9052: this.
9053:
9054:
9055: AUTHOR
9056:
9057: Philip Hazel
9058: University Computing Service
9059: Cambridge CB2 3QH, England.
9060:
9061:
9062: REVISION
9063:
1.1.1.4 ! misho 9064: Last updated: 24 June 2012
1.1.1.2 misho 9065: Copyright (c) 1997-2012 University of Cambridge.
1.1 misho 9066: ------------------------------------------------------------------------------
9067:
9068:
1.1.1.4 ! misho 9069: PCREPERFORM(3) Library Functions Manual PCREPERFORM(3)
! 9070:
1.1 misho 9071:
9072:
9073: NAME
9074: PCRE - Perl-compatible regular expressions
9075:
9076: PCRE PERFORMANCE
9077:
9078: Two aspects of performance are discussed below: memory usage and pro-
9079: cessing time. The way you express your pattern as a regular expression
9080: can affect both of them.
9081:
9082:
9083: COMPILED PATTERN MEMORY USAGE
9084:
1.1.1.2 misho 9085: Patterns are compiled by PCRE into a reasonably efficient interpretive
9086: code, so that most simple patterns do not use much memory. However,
9087: there is one case where the memory usage of a compiled pattern can be
9088: unexpectedly large. If a parenthesized subpattern has a quantifier with
9089: a minimum greater than 1 and/or a limited maximum, the whole subpattern
9090: is repeated in the compiled code. For example, the pattern
1.1 misho 9091:
9092: (abc|def){2,4}
9093:
9094: is compiled as if it were
9095:
9096: (abc|def)(abc|def)((abc|def)(abc|def)?)?
9097:
9098: (Technical aside: It is done this way so that backtrack points within
9099: each of the repetitions can be independently maintained.)
9100:
9101: For regular expressions whose quantifiers use only small numbers, this
9102: is not usually a problem. However, if the numbers are large, and par-
9103: ticularly if such repetitions are nested, the memory usage can become
9104: an embarrassment. For example, the very simple pattern
9105:
9106: ((ab){1,1000}c){1,3}
9107:
1.1.1.2 misho 9108: uses 51K bytes when compiled using the 8-bit library. When PCRE is com-
9109: piled with its default internal pointer size of two bytes, the size
9110: limit on a compiled pattern is 64K data units, and this is reached with
9111: the above pattern if the outer repetition is increased from 3 to 4.
9112: PCRE can be compiled to use larger internal pointers and thus handle
9113: larger compiled patterns, but it is better to try to rewrite your pat-
9114: tern to use less memory if you can.
1.1 misho 9115:
1.1.1.2 misho 9116: One way of reducing the memory usage for such patterns is to make use
1.1 misho 9117: of PCRE's "subroutine" facility. Re-writing the above pattern as
9118:
9119: ((ab)(?2){0,999}c)(?1){0,2}
9120:
9121: reduces the memory requirements to 18K, and indeed it remains under 20K
1.1.1.2 misho 9122: even with the outer repetition increased to 100. However, this pattern
9123: is not exactly equivalent, because the "subroutine" calls are treated
9124: as atomic groups into which there can be no backtracking if there is a
9125: subsequent matching failure. Therefore, PCRE cannot do this kind of
9126: rewriting automatically. Furthermore, there is a noticeable loss of
9127: speed when executing the modified pattern. Nevertheless, if the atomic
9128: grouping is not a problem and the loss of speed is acceptable, this
9129: kind of rewriting will allow you to process patterns that PCRE cannot
1.1 misho 9130: otherwise handle.
9131:
9132:
9133: STACK USAGE AT RUN TIME
9134:
1.1.1.4 ! misho 9135: When pcre_exec() or pcre[16|32]_exec() is used for matching, certain
! 9136: kinds of pattern can cause it to use large amounts of the process
! 9137: stack. In some environments the default process stack is quite small,
! 9138: and if it runs out the result is often SIGSEGV. This issue is probably
! 9139: the most frequently raised problem with PCRE. Rewriting your pattern
! 9140: can often help. The pcrestack documentation discusses this issue in
! 9141: detail.
1.1 misho 9142:
9143:
9144: PROCESSING TIME
9145:
1.1.1.4 ! misho 9146: Certain items in regular expression patterns are processed more effi-
1.1 misho 9147: ciently than others. It is more efficient to use a character class like
1.1.1.4 ! misho 9148: [aeiou] than a set of single-character alternatives such as
! 9149: (a|e|i|o|u). In general, the simplest construction that provides the
1.1 misho 9150: required behaviour is usually the most efficient. Jeffrey Friedl's book
1.1.1.4 ! misho 9151: contains a lot of useful general discussion about optimizing regular
! 9152: expressions for efficient performance. This document contains a few
1.1 misho 9153: observations about PCRE.
9154:
1.1.1.4 ! misho 9155: Using Unicode character properties (the \p, \P, and \X escapes) is
! 9156: slow, because PCRE has to use a multi-stage table lookup whenever it
! 9157: needs a character's property. If you can find an alternative pattern
! 9158: that does not use character properties, it will probably be faster.
1.1 misho 9159:
1.1.1.2 misho 9160: By default, the escape sequences \b, \d, \s, and \w, and the POSIX
9161: character classes such as [:alpha:] do not use Unicode properties,
1.1 misho 9162: partly for backwards compatibility, and partly for performance reasons.
1.1.1.2 misho 9163: However, you can set PCRE_UCP if you want Unicode character properties
9164: to be used. This can double the matching time for items such as \d,
9165: when matched with a traditional matching function; the performance loss
9166: is less with a DFA matching function, and in both cases there is not
9167: much difference for \b.
1.1 misho 9168:
9169: When a pattern begins with .* not in parentheses, or in parentheses
9170: that are not the subject of a backreference, and the PCRE_DOTALL option
9171: is set, the pattern is implicitly anchored by PCRE, since it can match
9172: only at the start of a subject string. However, if PCRE_DOTALL is not
9173: set, PCRE cannot make this optimization, because the . metacharacter
9174: does not then match a newline, and if the subject string contains new-
9175: lines, the pattern may match from the character immediately following
9176: one of them instead of from the very start. For example, the pattern
9177:
9178: .*second
9179:
9180: matches the subject "first\nand second" (where \n stands for a newline
9181: character), with the match starting at the seventh character. In order
9182: to do this, PCRE has to retry the match starting after every newline in
9183: the subject.
9184:
9185: If you are using such a pattern with subject strings that do not con-
9186: tain newlines, the best performance is obtained by setting PCRE_DOTALL,
9187: or starting the pattern with ^.* or ^.*? to indicate explicit anchor-
9188: ing. That saves PCRE from having to scan along the subject looking for
9189: a newline to restart at.
9190:
9191: Beware of patterns that contain nested indefinite repeats. These can
9192: take a long time to run when applied to a string that does not match.
9193: Consider the pattern fragment
9194:
9195: ^(a+)*
9196:
9197: This can match "aaaa" in 16 different ways, and this number increases
9198: very rapidly as the string gets longer. (The * repeat can match 0, 1,
9199: 2, 3, or 4 times, and for each of those cases other than 0 or 4, the +
9200: repeats can match different numbers of times.) When the remainder of
9201: the pattern is such that the entire match is going to fail, PCRE has in
9202: principle to try every possible variation, and this can take an
9203: extremely long time, even for relatively short strings.
9204:
9205: An optimization catches some of the more simple cases such as
9206:
9207: (a+)*b
9208:
9209: where a literal character follows. Before embarking on the standard
9210: matching procedure, PCRE checks that there is a "b" later in the sub-
9211: ject string, and if there is not, it fails the match immediately. How-
9212: ever, when there is no following literal this optimization cannot be
9213: used. You can see the difference by comparing the behaviour of
9214:
9215: (a+)*\d
9216:
9217: with the pattern above. The former gives a failure almost instantly
9218: when applied to a whole line of "a" characters, whereas the latter
9219: takes an appreciable time with strings longer than about 20 characters.
9220:
9221: In many cases, the solution to this kind of performance issue is to use
9222: an atomic group or a possessive quantifier.
9223:
9224:
9225: AUTHOR
9226:
9227: Philip Hazel
9228: University Computing Service
9229: Cambridge CB2 3QH, England.
9230:
9231:
9232: REVISION
9233:
1.1.1.4 ! misho 9234: Last updated: 25 August 2012
1.1.1.2 misho 9235: Copyright (c) 1997-2012 University of Cambridge.
1.1 misho 9236: ------------------------------------------------------------------------------
9237:
9238:
1.1.1.4 ! misho 9239: PCREPOSIX(3) Library Functions Manual PCREPOSIX(3)
! 9240:
1.1 misho 9241:
9242:
9243: NAME
9244: PCRE - Perl-compatible regular expressions.
9245:
9246: SYNOPSIS OF POSIX API
9247:
9248: #include <pcreposix.h>
9249:
9250: int regcomp(regex_t *preg, const char *pattern,
9251: int cflags);
9252:
9253: int regexec(regex_t *preg, const char *string,
9254: size_t nmatch, regmatch_t pmatch[], int eflags);
9255:
9256: size_t regerror(int errcode, const regex_t *preg,
9257: char *errbuf, size_t errbuf_size);
9258:
9259: void regfree(regex_t *preg);
9260:
9261:
9262: DESCRIPTION
9263:
1.1.1.2 misho 9264: This set of functions provides a POSIX-style API for the PCRE regular
9265: expression 8-bit library. See the pcreapi documentation for a descrip-
9266: tion of PCRE's native API, which contains much additional functional-
1.1.1.4 ! misho 9267: ity. There is no POSIX-style wrapper for PCRE's 16-bit and 32-bit
! 9268: library.
1.1 misho 9269:
9270: The functions described here are just wrapper functions that ultimately
9271: call the PCRE native API. Their prototypes are defined in the
1.1.1.4 ! misho 9272: pcreposix.h header file, and on Unix systems the library itself is
! 9273: called pcreposix.a, so can be accessed by adding -lpcreposix to the
! 9274: command for linking an application that uses them. Because the POSIX
1.1 misho 9275: functions call the native ones, it is also necessary to add -lpcre.
9276:
1.1.1.4 ! misho 9277: I have implemented only those POSIX option bits that can be reasonably
! 9278: mapped to PCRE native options. In addition, the option REG_EXTENDED is
! 9279: defined with the value zero. This has no effect, but since programs
! 9280: that are written to the POSIX interface often use it, this makes it
! 9281: easier to slot in PCRE as a replacement library. Other POSIX options
1.1 misho 9282: are not even defined.
9283:
1.1.1.4 ! misho 9284: There are also some other options that are not defined by POSIX. These
1.1 misho 9285: have been added at the request of users who want to make use of certain
9286: PCRE-specific features via the POSIX calling interface.
9287:
1.1.1.4 ! misho 9288: When PCRE is called via these functions, it is only the API that is
! 9289: POSIX-like in style. The syntax and semantics of the regular expres-
! 9290: sions themselves are still those of Perl, subject to the setting of
! 9291: various PCRE options, as described below. "POSIX-like in style" means
! 9292: that the API approximates to the POSIX definition; it is not fully
! 9293: POSIX-compatible, and in multi-byte encoding domains it is probably
1.1 misho 9294: even less compatible.
9295:
1.1.1.4 ! misho 9296: The header for these functions is supplied as pcreposix.h to avoid any
! 9297: potential clash with other POSIX libraries. It can, of course, be
1.1 misho 9298: renamed or aliased as regex.h, which is the "correct" name. It provides
1.1.1.4 ! misho 9299: two structure types, regex_t for compiled internal forms, and reg-
! 9300: match_t for returning captured substrings. It also defines some con-
! 9301: stants whose names start with "REG_"; these are used for setting
1.1 misho 9302: options and identifying error codes.
9303:
9304:
9305: COMPILING A PATTERN
9306:
1.1.1.4 ! misho 9307: The function regcomp() is called to compile a pattern into an internal
! 9308: form. The pattern is a C string terminated by a binary zero, and is
! 9309: passed in the argument pattern. The preg argument is a pointer to a
! 9310: regex_t structure that is used as a base for storing information about
1.1 misho 9311: the compiled regular expression.
9312:
9313: The argument cflags is either zero, or contains one or more of the bits
9314: defined by the following macros:
9315:
9316: REG_DOTALL
9317:
9318: The PCRE_DOTALL option is set when the regular expression is passed for
9319: compilation to the native function. Note that REG_DOTALL is not part of
9320: the POSIX standard.
9321:
9322: REG_ICASE
9323:
1.1.1.4 ! misho 9324: The PCRE_CASELESS option is set when the regular expression is passed
1.1 misho 9325: for compilation to the native function.
9326:
9327: REG_NEWLINE
9328:
1.1.1.4 ! misho 9329: The PCRE_MULTILINE option is set when the regular expression is passed
! 9330: for compilation to the native function. Note that this does not mimic
! 9331: the defined POSIX behaviour for REG_NEWLINE (see the following sec-
1.1 misho 9332: tion).
9333:
9334: REG_NOSUB
9335:
1.1.1.4 ! misho 9336: The PCRE_NO_AUTO_CAPTURE option is set when the regular expression is
1.1 misho 9337: passed for compilation to the native function. In addition, when a pat-
1.1.1.4 ! misho 9338: tern that is compiled with this flag is passed to regexec() for match-
! 9339: ing, the nmatch and pmatch arguments are ignored, and no captured
1.1 misho 9340: strings are returned.
9341:
9342: REG_UCP
9343:
1.1.1.4 ! misho 9344: The PCRE_UCP option is set when the regular expression is passed for
! 9345: compilation to the native function. This causes PCRE to use Unicode
! 9346: properties when matchine \d, \w, etc., instead of just recognizing
1.1 misho 9347: ASCII values. Note that REG_UTF8 is not part of the POSIX standard.
9348:
9349: REG_UNGREEDY
9350:
1.1.1.4 ! misho 9351: The PCRE_UNGREEDY option is set when the regular expression is passed
! 9352: for compilation to the native function. Note that REG_UNGREEDY is not
1.1 misho 9353: part of the POSIX standard.
9354:
9355: REG_UTF8
9356:
1.1.1.4 ! misho 9357: The PCRE_UTF8 option is set when the regular expression is passed for
! 9358: compilation to the native function. This causes the pattern itself and
! 9359: all data strings used for matching it to be treated as UTF-8 strings.
1.1 misho 9360: Note that REG_UTF8 is not part of the POSIX standard.
9361:
1.1.1.4 ! misho 9362: In the absence of these flags, no options are passed to the native
! 9363: function. This means the the regex is compiled with PCRE default
! 9364: semantics. In particular, the way it handles newline characters in the
! 9365: subject string is the Perl way, not the POSIX way. Note that setting
! 9366: PCRE_MULTILINE has only some of the effects specified for REG_NEWLINE.
! 9367: It does not affect the way newlines are matched by . (they are not) or
1.1 misho 9368: by a negative class such as [^a] (they are).
9369:
1.1.1.4 ! misho 9370: The yield of regcomp() is zero on success, and non-zero otherwise. The
1.1 misho 9371: preg structure is filled in on success, and one member of the structure
1.1.1.4 ! misho 9372: is public: re_nsub contains the number of capturing subpatterns in the
1.1 misho 9373: regular expression. Various error codes are defined in the header file.
9374:
1.1.1.4 ! misho 9375: NOTE: If the yield of regcomp() is non-zero, you must not attempt to
1.1 misho 9376: use the contents of the preg structure. If, for example, you pass it to
9377: regexec(), the result is undefined and your program is likely to crash.
9378:
9379:
9380: MATCHING NEWLINE CHARACTERS
9381:
9382: This area is not simple, because POSIX and Perl take different views of
1.1.1.4 ! misho 9383: things. It is not possible to get PCRE to obey POSIX semantics, but
! 9384: then PCRE was never intended to be a POSIX engine. The following table
! 9385: lists the different possibilities for matching newline characters in
1.1 misho 9386: PCRE:
9387:
9388: Default Change with
9389:
9390: . matches newline no PCRE_DOTALL
9391: newline matches [^a] yes not changeable
9392: $ matches \n at end yes PCRE_DOLLARENDONLY
9393: $ matches \n in middle no PCRE_MULTILINE
9394: ^ matches \n in middle no PCRE_MULTILINE
9395:
9396: This is the equivalent table for POSIX:
9397:
9398: Default Change with
9399:
9400: . matches newline yes REG_NEWLINE
9401: newline matches [^a] yes REG_NEWLINE
9402: $ matches \n at end no REG_NEWLINE
9403: $ matches \n in middle no REG_NEWLINE
9404: ^ matches \n in middle no REG_NEWLINE
9405:
9406: PCRE's behaviour is the same as Perl's, except that there is no equiva-
1.1.1.4 ! misho 9407: lent for PCRE_DOLLAR_ENDONLY in Perl. In both PCRE and Perl, there is
1.1 misho 9408: no way to stop newline from matching [^a].
9409:
1.1.1.4 ! misho 9410: The default POSIX newline handling can be obtained by setting
! 9411: PCRE_DOTALL and PCRE_DOLLAR_ENDONLY, but there is no way to make PCRE
1.1 misho 9412: behave exactly as for the REG_NEWLINE action.
9413:
9414:
9415: MATCHING A PATTERN
9416:
1.1.1.4 ! misho 9417: The function regexec() is called to match a compiled pattern preg
! 9418: against a given string, which is by default terminated by a zero byte
! 9419: (but see REG_STARTEND below), subject to the options in eflags. These
1.1 misho 9420: can be:
9421:
9422: REG_NOTBOL
9423:
9424: The PCRE_NOTBOL option is set when calling the underlying PCRE matching
9425: function.
9426:
9427: REG_NOTEMPTY
9428:
9429: The PCRE_NOTEMPTY option is set when calling the underlying PCRE match-
9430: ing function. Note that REG_NOTEMPTY is not part of the POSIX standard.
9431: However, setting this option can give more POSIX-like behaviour in some
9432: situations.
9433:
9434: REG_NOTEOL
9435:
9436: The PCRE_NOTEOL option is set when calling the underlying PCRE matching
9437: function.
9438:
9439: REG_STARTEND
9440:
1.1.1.4 ! misho 9441: The string is considered to start at string + pmatch[0].rm_so and to
! 9442: have a terminating NUL located at string + pmatch[0].rm_eo (there need
! 9443: not actually be a NUL at that location), regardless of the value of
! 9444: nmatch. This is a BSD extension, compatible with but not specified by
! 9445: IEEE Standard 1003.2 (POSIX.2), and should be used with caution in
1.1 misho 9446: software intended to be portable to other systems. Note that a non-zero
9447: rm_so does not imply REG_NOTBOL; REG_STARTEND affects only the location
9448: of the string, not how it is matched.
9449:
1.1.1.4 ! misho 9450: If the pattern was compiled with the REG_NOSUB flag, no data about any
! 9451: matched strings is returned. The nmatch and pmatch arguments of
1.1 misho 9452: regexec() are ignored.
9453:
9454: If the value of nmatch is zero, or if the value pmatch is NULL, no data
9455: about any matched strings is returned.
9456:
9457: Otherwise,the portion of the string that was matched, and also any cap-
9458: tured substrings, are returned via the pmatch argument, which points to
1.1.1.4 ! misho 9459: an array of nmatch structures of type regmatch_t, containing the mem-
! 9460: bers rm_so and rm_eo. These contain the offset to the first character
! 9461: of each substring and the offset to the first character after the end
! 9462: of each substring, respectively. The 0th element of the vector relates
! 9463: to the entire portion of string that was matched; subsequent elements
! 9464: relate to the capturing subpatterns of the regular expression. Unused
1.1 misho 9465: entries in the array have both structure members set to -1.
9466:
1.1.1.4 ! misho 9467: A successful match yields a zero return; various error codes are
! 9468: defined in the header file, of which REG_NOMATCH is the "expected"
1.1 misho 9469: failure code.
9470:
9471:
9472: ERROR MESSAGES
9473:
9474: The regerror() function maps a non-zero errorcode from either regcomp()
1.1.1.4 ! misho 9475: or regexec() to a printable message. If preg is not NULL, the error
1.1 misho 9476: should have arisen from the use of that structure. A message terminated
1.1.1.4 ! misho 9477: by a binary zero is placed in errbuf. The length of the message,
! 9478: including the zero, is limited to errbuf_size. The yield of the func-
1.1 misho 9479: tion is the size of buffer needed to hold the whole message.
9480:
9481:
9482: MEMORY USAGE
9483:
1.1.1.4 ! misho 9484: Compiling a regular expression causes memory to be allocated and asso-
! 9485: ciated with the preg structure. The function regfree() frees all such
! 9486: memory, after which preg may no longer be used as a compiled expres-
1.1 misho 9487: sion.
9488:
9489:
9490: AUTHOR
9491:
9492: Philip Hazel
9493: University Computing Service
9494: Cambridge CB2 3QH, England.
9495:
9496:
9497: REVISION
9498:
1.1.1.2 misho 9499: Last updated: 09 January 2012
9500: Copyright (c) 1997-2012 University of Cambridge.
1.1 misho 9501: ------------------------------------------------------------------------------
9502:
9503:
1.1.1.4 ! misho 9504: PCRECPP(3) Library Functions Manual PCRECPP(3)
! 9505:
1.1 misho 9506:
9507:
9508: NAME
9509: PCRE - Perl-compatible regular expressions.
9510:
9511: SYNOPSIS OF C++ WRAPPER
9512:
9513: #include <pcrecpp.h>
9514:
9515:
9516: DESCRIPTION
9517:
9518: The C++ wrapper for PCRE was provided by Google Inc. Some additional
9519: functionality was added by Giuseppe Maxia. This brief man page was con-
9520: structed from the notes in the pcrecpp.h file, which should be con-
1.1.1.2 misho 9521: sulted for further details. Note that the C++ wrapper supports only the
1.1.1.4 ! misho 9522: original 8-bit PCRE library. There is no 16-bit or 32-bit support at
! 9523: present.
1.1 misho 9524:
9525:
9526: MATCHING INTERFACE
9527:
1.1.1.4 ! misho 9528: The "FullMatch" operation checks that supplied text matches a supplied
! 9529: pattern exactly. If pointer arguments are supplied, it copies matched
1.1 misho 9530: sub-strings that match sub-patterns into them.
9531:
9532: Example: successful match
9533: pcrecpp::RE re("h.*o");
9534: re.FullMatch("hello");
9535:
9536: Example: unsuccessful match (requires full match):
9537: pcrecpp::RE re("e");
9538: !re.FullMatch("hello");
9539:
9540: Example: creating a temporary RE object:
9541: pcrecpp::RE("h.*o").FullMatch("hello");
9542:
1.1.1.4 ! misho 9543: You can pass in a "const char*" or a "string" for "text". The examples
! 9544: below tend to use a const char*. You can, as in the different examples
! 9545: above, store the RE object explicitly in a variable or use a temporary
! 9546: RE object. The examples below use one mode or the other arbitrarily.
1.1 misho 9547: Either could correctly be used for any of these examples.
9548:
9549: You must supply extra pointer arguments to extract matched subpieces.
9550:
9551: Example: extracts "ruby" into "s" and 1234 into "i"
9552: int i;
9553: string s;
9554: pcrecpp::RE re("(\\w+):(\\d+)");
9555: re.FullMatch("ruby:1234", &s, &i);
9556:
9557: Example: does not try to extract any extra sub-patterns
9558: re.FullMatch("ruby:1234", &s);
9559:
9560: Example: does not try to extract into NULL
9561: re.FullMatch("ruby:1234", NULL, &i);
9562:
9563: Example: integer overflow causes failure
9564: !re.FullMatch("ruby:1234567891234", NULL, &i);
9565:
9566: Example: fails because there aren't enough sub-patterns:
9567: !pcrecpp::RE("\\w+:\\d+").FullMatch("ruby:1234", &s);
9568:
9569: Example: fails because string cannot be stored in integer
9570: !pcrecpp::RE("(.*)").FullMatch("ruby", &i);
9571:
1.1.1.4 ! misho 9572: The provided pointer arguments can be pointers to any scalar numeric
1.1 misho 9573: type, or one of:
9574:
9575: string (matched piece is copied to string)
9576: StringPiece (StringPiece is mutated to point to matched piece)
9577: T (where "bool T::ParseFrom(const char*, int)" exists)
9578: NULL (the corresponding matched sub-pattern is not copied)
9579:
1.1.1.4 ! misho 9580: The function returns true iff all of the following conditions are sat-
1.1 misho 9581: isfied:
9582:
9583: a. "text" matches "pattern" exactly;
9584:
9585: b. The number of matched sub-patterns is >= number of supplied
9586: pointers;
9587:
9588: c. The "i"th argument has a suitable type for holding the
9589: string captured as the "i"th sub-pattern. If you pass in
9590: void * NULL for the "i"th argument, or a non-void * NULL
9591: of the correct type, or pass fewer arguments than the
9592: number of sub-patterns, "i"th captured sub-pattern is
9593: ignored.
9594:
1.1.1.4 ! misho 9595: CAVEAT: An optional sub-pattern that does not exist in the matched
! 9596: string is assigned the empty string. Therefore, the following will
1.1 misho 9597: return false (because the empty string is not a valid number):
9598:
9599: int number;
9600: pcrecpp::RE::FullMatch("abc", "[a-z]+(\\d+)?", &number);
9601:
1.1.1.4 ! misho 9602: The matching interface supports at most 16 arguments per call. If you
! 9603: need more, consider using the more general interface
1.1 misho 9604: pcrecpp::RE::DoMatch. See pcrecpp.h for the signature for DoMatch.
9605:
1.1.1.4 ! misho 9606: NOTE: Do not use no_arg, which is used internally to mark the end of a
! 9607: list of optional arguments, as a placeholder for missing arguments, as
1.1 misho 9608: this can lead to segfaults.
9609:
9610:
9611: QUOTING METACHARACTERS
9612:
1.1.1.4 ! misho 9613: You can use the "QuoteMeta" operation to insert backslashes before all
! 9614: potentially meaningful characters in a string. The returned string,
1.1 misho 9615: used as a regular expression, will exactly match the original string.
9616:
9617: Example:
9618: string quoted = RE::QuoteMeta(unquoted);
9619:
1.1.1.4 ! misho 9620: Note that it's legal to escape a character even if it has no special
! 9621: meaning in a regular expression -- so this function does that. (This
! 9622: also makes it identical to the perl function of the same name; see
! 9623: "perldoc -f quotemeta".) For example, "1.5-2.0?" becomes
1.1 misho 9624: "1\.5\-2\.0\?".
9625:
9626:
9627: PARTIAL MATCHES
9628:
1.1.1.4 ! misho 9629: You can use the "PartialMatch" operation when you want the pattern to
1.1 misho 9630: match any substring of the text.
9631:
9632: Example: simple search for a string:
9633: pcrecpp::RE("ell").PartialMatch("hello");
9634:
9635: Example: find first number in a string:
9636: int number;
9637: pcrecpp::RE re("(\\d+)");
9638: re.PartialMatch("x*100 + 20", &number);
9639: assert(number == 100);
9640:
9641:
9642: UTF-8 AND THE MATCHING INTERFACE
9643:
1.1.1.4 ! misho 9644: By default, pattern and text are plain text, one byte per character.
! 9645: The UTF8 flag, passed to the constructor, causes both pattern and
1.1 misho 9646: string to be treated as UTF-8 text, still a byte stream but potentially
1.1.1.4 ! misho 9647: multiple bytes per character. In practice, the text is likelier to be
! 9648: UTF-8 than the pattern, but the match returned may depend on the UTF8
! 9649: flag, so always use it when matching UTF8 text. For example, "." will
! 9650: match one byte normally but with UTF8 set may match up to three bytes
1.1 misho 9651: of a multi-byte character.
9652:
9653: Example:
9654: pcrecpp::RE_Options options;
9655: options.set_utf8();
9656: pcrecpp::RE re(utf8_pattern, options);
9657: re.FullMatch(utf8_string);
9658:
9659: Example: using the convenience function UTF8():
9660: pcrecpp::RE re(utf8_pattern, pcrecpp::UTF8());
9661: re.FullMatch(utf8_string);
9662:
9663: NOTE: The UTF8 flag is ignored if pcre was not configured with the
9664: --enable-utf8 flag.
9665:
9666:
9667: PASSING MODIFIERS TO THE REGULAR EXPRESSION ENGINE
9668:
1.1.1.4 ! misho 9669: PCRE defines some modifiers to change the behavior of the regular
! 9670: expression engine. The C++ wrapper defines an auxiliary class,
! 9671: RE_Options, as a vehicle to pass such modifiers to a RE class. Cur-
1.1 misho 9672: rently, the following modifiers are supported:
9673:
9674: modifier description Perl corresponding
9675:
9676: PCRE_CASELESS case insensitive match /i
9677: PCRE_MULTILINE multiple lines match /m
9678: PCRE_DOTALL dot matches newlines /s
9679: PCRE_DOLLAR_ENDONLY $ matches only at end N/A
9680: PCRE_EXTRA strict escape parsing N/A
1.1.1.3 misho 9681: PCRE_EXTENDED ignore white spaces /x
1.1 misho 9682: PCRE_UTF8 handles UTF8 chars built-in
9683: PCRE_UNGREEDY reverses * and *? N/A
9684: PCRE_NO_AUTO_CAPTURE disables capturing parens N/A (*)
9685:
1.1.1.4 ! misho 9686: (*) Both Perl and PCRE allow non capturing parentheses by means of the
! 9687: "?:" modifier within the pattern itself. e.g. (?:ab|cd) does not cap-
1.1 misho 9688: ture, while (ab|cd) does.
9689:
1.1.1.4 ! misho 9690: For a full account on how each modifier works, please check the PCRE
1.1 misho 9691: API reference page.
9692:
1.1.1.4 ! misho 9693: For each modifier, there are two member functions whose name is made
! 9694: out of the modifier in lowercase, without the "PCRE_" prefix. For
1.1 misho 9695: instance, PCRE_CASELESS is handled by
9696:
9697: bool caseless()
9698:
9699: which returns true if the modifier is set, and
9700:
9701: RE_Options & set_caseless(bool)
9702:
9703: which sets or unsets the modifier. Moreover, PCRE_EXTRA_MATCH_LIMIT can
1.1.1.4 ! misho 9704: be accessed through the set_match_limit() and match_limit() member
! 9705: functions. Setting match_limit to a non-zero value will limit the exe-
! 9706: cution of pcre to keep it from doing bad things like blowing the stack
! 9707: or taking an eternity to return a result. A value of 5000 is good
! 9708: enough to stop stack blowup in a 2MB thread stack. Setting match_limit
! 9709: to zero disables match limiting. Alternatively, you can call
! 9710: match_limit_recursion() which uses PCRE_EXTRA_MATCH_LIMIT_RECURSION to
! 9711: limit how much PCRE recurses. match_limit() limits the number of
1.1 misho 9712: matches PCRE does; match_limit_recursion() limits the depth of internal
9713: recursion, and therefore the amount of stack that is used.
9714:
1.1.1.4 ! misho 9715: Normally, to pass one or more modifiers to a RE class, you declare a
1.1 misho 9716: RE_Options object, set the appropriate options, and pass this object to
9717: a RE constructor. Example:
9718:
9719: RE_Options opt;
9720: opt.set_caseless(true);
9721: if (RE("HELLO", opt).PartialMatch("hello world")) ...
9722:
9723: RE_options has two constructors. The default constructor takes no argu-
1.1.1.4 ! misho 9724: ments and creates a set of flags that are off by default. The optional
! 9725: parameter option_flags is to facilitate transfer of legacy code from C
1.1 misho 9726: programs. This lets you do
9727:
9728: RE(pattern,
9729: RE_Options(PCRE_CASELESS|PCRE_MULTILINE)).PartialMatch(str);
9730:
9731: However, new code is better off doing
9732:
9733: RE(pattern,
9734: RE_Options().set_caseless(true).set_multiline(true))
9735: .PartialMatch(str);
9736:
9737: If you are going to pass one of the most used modifiers, there are some
9738: convenience functions that return a RE_Options class with the appropri-
1.1.1.4 ! misho 9739: ate modifier already set: CASELESS(), UTF8(), MULTILINE(), DOTALL(),
1.1 misho 9740: and EXTENDED().
9741:
1.1.1.4 ! misho 9742: If you need to set several options at once, and you don't want to go
! 9743: through the pains of declaring a RE_Options object and setting several
! 9744: options, there is a parallel method that give you such ability on the
! 9745: fly. You can concatenate several set_xxxxx() member functions, since
! 9746: each of them returns a reference to its class object. For example, to
! 9747: pass PCRE_CASELESS, PCRE_EXTENDED, and PCRE_MULTILINE to a RE with one
1.1 misho 9748: statement, you may write:
9749:
9750: RE(" ^ xyz \\s+ .* blah$",
9751: RE_Options()
9752: .set_caseless(true)
9753: .set_extended(true)
9754: .set_multiline(true)).PartialMatch(sometext);
9755:
9756:
9757: SCANNING TEXT INCREMENTALLY
9758:
1.1.1.4 ! misho 9759: The "Consume" operation may be useful if you want to repeatedly match
1.1 misho 9760: regular expressions at the front of a string and skip over them as they
1.1.1.4 ! misho 9761: match. This requires use of the "StringPiece" type, which represents a
! 9762: sub-range of a real string. Like RE, StringPiece is defined in the
1.1 misho 9763: pcrecpp namespace.
9764:
9765: Example: read lines of the form "var = value" from a string.
9766: string contents = ...; // Fill string somehow
9767: pcrecpp::StringPiece input(contents); // Wrap in a StringPiece
9768:
9769: string var;
9770: int value;
9771: pcrecpp::RE re("(\\w+) = (\\d+)\n");
9772: while (re.Consume(&input, &var, &value)) {
9773: ...;
9774: }
9775:
1.1.1.4 ! misho 9776: Each successful call to "Consume" will set "var/value", and also
1.1 misho 9777: advance "input" so it points past the matched text.
9778:
1.1.1.4 ! misho 9779: The "FindAndConsume" operation is similar to "Consume" but does not
! 9780: anchor your match at the beginning of the string. For example, you
1.1 misho 9781: could extract all words from a string by repeatedly calling
9782:
9783: pcrecpp::RE("(\\w+)").FindAndConsume(&input, &word)
9784:
9785:
9786: PARSING HEX/OCTAL/C-RADIX NUMBERS
9787:
9788: By default, if you pass a pointer to a numeric value, the corresponding
1.1.1.4 ! misho 9789: text is interpreted as a base-10 number. You can instead wrap the
1.1 misho 9790: pointer with a call to one of the operators Hex(), Octal(), or CRadix()
1.1.1.4 ! misho 9791: to interpret the text in another base. The CRadix operator interprets
! 9792: C-style "0" (base-8) and "0x" (base-16) prefixes, but defaults to
1.1 misho 9793: base-10.
9794:
9795: Example:
9796: int a, b, c, d;
9797: pcrecpp::RE re("(.*) (.*) (.*) (.*)");
9798: re.FullMatch("100 40 0100 0x40",
9799: pcrecpp::Octal(&a), pcrecpp::Hex(&b),
9800: pcrecpp::CRadix(&c), pcrecpp::CRadix(&d));
9801:
9802: will leave 64 in a, b, c, and d.
9803:
9804:
9805: REPLACING PARTS OF STRINGS
9806:
1.1.1.4 ! misho 9807: You can replace the first match of "pattern" in "str" with "rewrite".
! 9808: Within "rewrite", backslash-escaped digits (\1 to \9) can be used to
! 9809: insert text matching corresponding parenthesized group from the pat-
1.1 misho 9810: tern. \0 in "rewrite" refers to the entire matching text. For example:
9811:
9812: string s = "yabba dabba doo";
9813: pcrecpp::RE("b+").Replace("d", &s);
9814:
1.1.1.4 ! misho 9815: will leave "s" containing "yada dabba doo". The result is true if the
1.1 misho 9816: pattern matches and a replacement occurs, false otherwise.
9817:
1.1.1.4 ! misho 9818: GlobalReplace is like Replace except that it replaces all occurrences
! 9819: of the pattern in the string with the rewrite. Replacements are not
1.1 misho 9820: subject to re-matching. For example:
9821:
9822: string s = "yabba dabba doo";
9823: pcrecpp::RE("b+").GlobalReplace("d", &s);
9824:
1.1.1.4 ! misho 9825: will leave "s" containing "yada dada doo". It returns the number of
1.1 misho 9826: replacements made.
9827:
1.1.1.4 ! misho 9828: Extract is like Replace, except that if the pattern matches, "rewrite"
! 9829: is copied into "out" (an additional argument) with substitutions. The
! 9830: non-matching portions of "text" are ignored. Returns true iff a match
1.1 misho 9831: occurred and the extraction happened successfully; if no match occurs,
9832: the string is left unaffected.
9833:
9834:
9835: AUTHOR
9836:
9837: The C++ wrapper was contributed by Google Inc.
9838: Copyright (c) 2007 Google Inc.
9839:
9840:
9841: REVISION
9842:
1.1.1.2 misho 9843: Last updated: 08 January 2012
1.1 misho 9844: ------------------------------------------------------------------------------
9845:
9846:
1.1.1.4 ! misho 9847: PCRESAMPLE(3) Library Functions Manual PCRESAMPLE(3)
! 9848:
1.1 misho 9849:
9850:
9851: NAME
9852: PCRE - Perl-compatible regular expressions
9853:
9854: PCRE SAMPLE PROGRAM
9855:
9856: A simple, complete demonstration program, to get you started with using
9857: PCRE, is supplied in the file pcredemo.c in the PCRE distribution. A
9858: listing of this program is given in the pcredemo documentation. If you
9859: do not have a copy of the PCRE distribution, you can save this listing
9860: to re-create pcredemo.c.
9861:
1.1.1.2 misho 9862: The demonstration program, which uses the original PCRE 8-bit library,
9863: compiles the regular expression that is its first argument, and matches
9864: it against the subject string in its second argument. No PCRE options
9865: are set, and default character tables are used. If matching succeeds,
9866: the program outputs the portion of the subject that matched, together
9867: with the contents of any captured substrings.
1.1 misho 9868:
9869: If the -g option is given on the command line, the program then goes on
9870: to check for further matches of the same regular expression in the same
1.1.1.2 misho 9871: subject string. The logic is a little bit tricky because of the possi-
9872: bility of matching an empty string. Comments in the code explain what
1.1 misho 9873: is going on.
9874:
1.1.1.2 misho 9875: If PCRE is installed in the standard include and library directories
1.1 misho 9876: for your operating system, you should be able to compile the demonstra-
9877: tion program using this command:
9878:
9879: gcc -o pcredemo pcredemo.c -lpcre
9880:
1.1.1.2 misho 9881: If PCRE is installed elsewhere, you may need to add additional options
9882: to the command line. For example, on a Unix-like system that has PCRE
9883: installed in /usr/local, you can compile the demonstration program
1.1 misho 9884: using a command like this:
9885:
9886: gcc -o pcredemo -I/usr/local/include pcredemo.c \
9887: -L/usr/local/lib -lpcre
9888:
1.1.1.2 misho 9889: In a Windows environment, if you want to statically link the program
1.1 misho 9890: against a non-dll pcre.a file, you must uncomment the line that defines
1.1.1.2 misho 9891: PCRE_STATIC before including pcre.h, because otherwise the pcre_mal-
1.1 misho 9892: loc() and pcre_free() exported functions will be declared
9893: __declspec(dllimport), with unwanted results.
9894:
1.1.1.2 misho 9895: Once you have compiled and linked the demonstration program, you can
1.1 misho 9896: run simple tests like this:
9897:
9898: ./pcredemo 'cat|dog' 'the cat sat on the mat'
9899: ./pcredemo -g 'cat|dog' 'the dog sat on the cat'
9900:
1.1.1.2 misho 9901: Note that there is a much more comprehensive test program, called
9902: pcretest, which supports many more facilities for testing regular
9903: expressions and both PCRE libraries. The pcredemo program is provided
9904: as a simple coding example.
1.1 misho 9905:
1.1.1.2 misho 9906: If you try to run pcredemo when PCRE is not installed in the standard
9907: library directory, you may get an error like this on some operating
1.1 misho 9908: systems (e.g. Solaris):
9909:
1.1.1.2 misho 9910: ld.so.1: a.out: fatal: libpcre.so.0: open failed: No such file or
1.1 misho 9911: directory
9912:
1.1.1.2 misho 9913: This is caused by the way shared library support works on those sys-
1.1 misho 9914: tems. You need to add
9915:
9916: -R/usr/local/lib
9917:
9918: (for example) to the compile command to get round this problem.
9919:
9920:
9921: AUTHOR
9922:
9923: Philip Hazel
9924: University Computing Service
9925: Cambridge CB2 3QH, England.
9926:
9927:
9928: REVISION
9929:
1.1.1.2 misho 9930: Last updated: 10 January 2012
9931: Copyright (c) 1997-2012 University of Cambridge.
1.1 misho 9932: ------------------------------------------------------------------------------
1.1.1.4 ! misho 9933: PCRELIMITS(3) Library Functions Manual PCRELIMITS(3)
! 9934:
1.1 misho 9935:
9936:
9937: NAME
9938: PCRE - Perl-compatible regular expressions
9939:
9940: SIZE AND OTHER LIMITATIONS
9941:
9942: There are some size limitations in PCRE but it is hoped that they will
9943: never in practice be relevant.
9944:
1.1.1.2 misho 9945: The maximum length of a compiled pattern is approximately 64K data
1.1.1.4 ! misho 9946: units (bytes for the 8-bit library, 32-bit units for the 32-bit
! 9947: library, and 32-bit units for the 32-bit library) if PCRE is compiled
! 9948: with the default internal linkage size of 2 bytes. If you want to
! 9949: process regular expressions that are truly enormous, you can compile
! 9950: PCRE with an internal linkage size of 3 or 4 (when building the 16-bit
! 9951: or 32-bit library, 3 is rounded up to 4). See the README file in the
! 9952: source distribution and the pcrebuild documentation for details. In
! 9953: these cases the limit is substantially larger. However, the speed of
! 9954: execution is slower.
1.1 misho 9955:
9956: All values in repeating quantifiers must be less than 65536.
9957:
9958: There is no limit to the number of parenthesized subpatterns, but there
9959: can be no more than 65535 capturing subpatterns.
9960:
9961: There is a limit to the number of forward references to subsequent sub-
1.1.1.4 ! misho 9962: patterns of around 200,000. Repeated forward references with fixed
! 9963: upper limits, for example, (?2){0,100} when subpattern number 2 is to
! 9964: the right, are included in the count. There is no limit to the number
1.1 misho 9965: of backward references.
9966:
9967: The maximum length of name for a named subpattern is 32 characters, and
9968: the maximum number of named subpatterns is 10000.
9969:
1.1.1.4 ! misho 9970: The maximum length of a name in a (*MARK), (*PRUNE), (*SKIP), or
! 9971: (*THEN) verb is 255 for the 8-bit library and 65535 for the 16-bit and
! 9972: 32-bit library.
1.1.1.3 misho 9973:
1.1.1.4 ! misho 9974: The maximum length of a subject string is the largest positive number
! 9975: that an integer variable can hold. However, when using the traditional
1.1 misho 9976: matching function, PCRE uses recursion to handle subpatterns and indef-
1.1.1.4 ! misho 9977: inite repetition. This means that the available stack space may limit
1.1 misho 9978: the size of a subject string that can be processed by certain patterns.
9979: For a discussion of stack issues, see the pcrestack documentation.
9980:
9981:
9982: AUTHOR
9983:
9984: Philip Hazel
9985: University Computing Service
9986: Cambridge CB2 3QH, England.
9987:
9988:
9989: REVISION
9990:
1.1.1.3 misho 9991: Last updated: 04 May 2012
1.1.1.2 misho 9992: Copyright (c) 1997-2012 University of Cambridge.
1.1 misho 9993: ------------------------------------------------------------------------------
9994:
9995:
1.1.1.4 ! misho 9996: PCRESTACK(3) Library Functions Manual PCRESTACK(3)
! 9997:
1.1 misho 9998:
9999:
10000: NAME
10001: PCRE - Perl-compatible regular expressions
10002:
10003: PCRE DISCUSSION OF STACK USAGE
10004:
1.1.1.4 ! misho 10005: When you call pcre[16|32]_exec(), it makes use of an internal function
1.1.1.2 misho 10006: called match(). This calls itself recursively at branch points in the
10007: pattern, in order to remember the state of the match so that it can
10008: back up and try a different alternative if the first one fails. As
10009: matching proceeds deeper and deeper into the tree of possibilities, the
10010: recursion depth increases. The match() function is also called in other
10011: circumstances, for example, whenever a parenthesized sub-pattern is
10012: entered, and in certain cases of repetition.
1.1 misho 10013:
10014: Not all calls of match() increase the recursion depth; for an item such
10015: as a* it may be called several times at the same level, after matching
10016: different numbers of a's. Furthermore, in a number of cases where the
10017: result of the recursive call would immediately be passed back as the
10018: result of the current call (a "tail recursion"), the function is just
10019: restarted instead.
10020:
1.1.1.4 ! misho 10021: The above comments apply when pcre[16|32]_exec() is run in its normal
1.1.1.2 misho 10022: interpretive manner. If the pattern was studied with the
10023: PCRE_STUDY_JIT_COMPILE option, and just-in-time compiling was success-
1.1.1.4 ! misho 10024: ful, and the options passed to pcre[16|32]_exec() were not incompati-
! 10025: ble, the matching process uses the JIT-compiled code instead of the
! 10026: match() function. In this case, the memory requirements are handled
! 10027: entirely differently. See the pcrejit documentation for details.
! 10028:
! 10029: The pcre[16|32]_dfa_exec() function operates in an entirely different
! 10030: way, and uses recursion only when there is a regular expression recur-
! 10031: sion or subroutine call in the pattern. This includes the processing of
! 10032: assertion and "once-only" subpatterns, which are handled like subrou-
! 10033: tine calls. Normally, these are never very deep, and the limit on the
! 10034: complexity of pcre[16|32]_dfa_exec() is controlled by the amount of
! 10035: workspace it is given. However, it is possible to write patterns with
! 10036: runaway infinite recursions; such patterns will cause
! 10037: pcre[16|32]_dfa_exec() to run out of stack. At present, there is no
! 10038: protection against this.
! 10039:
! 10040: The comments that follow do NOT apply to pcre[16|32]_dfa_exec(); they
! 10041: are relevant only for pcre[16|32]_exec() without the JIT optimization.
! 10042:
! 10043: Reducing pcre[16|32]_exec()'s stack usage
! 10044:
! 10045: Each time that match() is actually called recursively, it uses memory
! 10046: from the process stack. For certain kinds of pattern and data, very
! 10047: large amounts of stack may be needed, despite the recognition of "tail
! 10048: recursion". You can often reduce the amount of recursion, and there-
! 10049: fore the amount of stack used, by modifying the pattern that is being
1.1 misho 10050: matched. Consider, for example, this pattern:
10051:
10052: ([^<]|<(?!inet))+
10053:
1.1.1.4 ! misho 10054: It matches from wherever it starts until it encounters "<inet" or the
! 10055: end of the data, and is the kind of pattern that might be used when
1.1 misho 10056: processing an XML file. Each iteration of the outer parentheses matches
1.1.1.4 ! misho 10057: either one character that is not "<" or a "<" that is not followed by
! 10058: "inet". However, each time a parenthesis is processed, a recursion
1.1 misho 10059: occurs, so this formulation uses a stack frame for each matched charac-
1.1.1.4 ! misho 10060: ter. For a long string, a lot of stack is required. Consider now this
1.1 misho 10061: rewritten pattern, which matches exactly the same strings:
10062:
10063: ([^<]++|<(?!inet))+
10064:
1.1.1.4 ! misho 10065: This uses very much less stack, because runs of characters that do not
! 10066: contain "<" are "swallowed" in one item inside the parentheses. Recur-
! 10067: sion happens only when a "<" character that is not followed by "inet"
! 10068: is encountered (and we assume this is relatively rare). A possessive
! 10069: quantifier is used to stop any backtracking into the runs of non-"<"
1.1 misho 10070: characters, but that is not related to stack usage.
10071:
1.1.1.4 ! misho 10072: This example shows that one way of avoiding stack problems when match-
1.1 misho 10073: ing long subject strings is to write repeated parenthesized subpatterns
10074: to match more than one character whenever possible.
10075:
1.1.1.4 ! misho 10076: Compiling PCRE to use heap instead of stack for pcre[16|32]_exec()
1.1 misho 10077:
1.1.1.4 ! misho 10078: In environments where stack memory is constrained, you might want to
! 10079: compile PCRE to use heap memory instead of stack for remembering back-
! 10080: up points when pcre[16|32]_exec() is running. This makes it run a lot
! 10081: more slowly, however. Details of how to do this are given in the pcre-
! 10082: build documentation. When built in this way, instead of using the
! 10083: stack, PCRE obtains and frees memory by calling the functions that are
! 10084: pointed to by the pcre[16|32]_stack_malloc and pcre[16|32]_stack_free
! 10085: variables. By default, these point to malloc() and free(), but you can
! 10086: replace the pointers to cause PCRE to use your own functions. Since the
! 10087: block sizes are always the same, and are always freed in reverse order,
! 10088: it may be possible to implement customized memory handlers that are
! 10089: more efficient than the standard functions.
! 10090:
! 10091: Limiting pcre[16|32]_exec()'s stack usage
! 10092:
! 10093: You can set limits on the number of times that match() is called, both
! 10094: in total and recursively. If a limit is exceeded, pcre[16|32]_exec()
! 10095: returns an error code. Setting suitable limits should prevent it from
! 10096: running out of stack. The default values of the limits are very large,
! 10097: and unlikely ever to operate. They can be changed when PCRE is built,
! 10098: and they can also be set when pcre[16|32]_exec() is called. For details
! 10099: of these interfaces, see the pcrebuild documentation and the section on
! 10100: extra data for pcre[16|32]_exec() in the pcreapi documentation.
1.1 misho 10101:
10102: As a very rough rule of thumb, you should reckon on about 500 bytes per
1.1.1.4 ! misho 10103: recursion. Thus, if you want to limit your stack usage to 8Mb, you
! 10104: should set the limit at 16000 recursions. A 64Mb stack, on the other
1.1 misho 10105: hand, can support around 128000 recursions.
10106:
10107: In Unix-like environments, the pcretest test program has a command line
10108: option (-S) that can be used to increase the size of its stack. As long
1.1.1.4 ! misho 10109: as the stack is large enough, another option (-M) can be used to find
! 10110: the smallest limits that allow a particular pattern to match a given
! 10111: subject string. This is done by calling pcre[16|32]_exec() repeatedly
! 10112: with different limits.
1.1 misho 10113:
1.1.1.2 misho 10114: Obtaining an estimate of stack usage
10115:
1.1.1.4 ! misho 10116: The actual amount of stack used per recursion can vary quite a lot,
1.1.1.2 misho 10117: depending on the compiler that was used to build PCRE and the optimiza-
10118: tion or debugging options that were set for it. The rule of thumb value
1.1.1.4 ! misho 10119: of 500 bytes mentioned above may be larger or smaller than what is
1.1.1.2 misho 10120: actually needed. A better approximation can be obtained by running this
10121: command:
10122:
10123: pcretest -m -C
10124:
1.1.1.4 ! misho 10125: The -C option causes pcretest to output information about the options
1.1.1.2 misho 10126: with which PCRE was compiled. When -m is also given (before -C), infor-
10127: mation about stack use is given in a line like this:
10128:
10129: Match recursion uses stack: approximate frame size = 640 bytes
10130:
10131: The value is approximate because some recursions need a bit more (up to
10132: perhaps 16 more bytes).
10133:
1.1.1.4 ! misho 10134: If the above command is given when PCRE is compiled to use the heap
! 10135: instead of the stack for recursion, the value that is output is the
1.1.1.2 misho 10136: size of each block that is obtained from the heap.
10137:
1.1 misho 10138: Changing stack size in Unix-like systems
10139:
1.1.1.4 ! misho 10140: In Unix-like environments, there is not often a problem with the stack
! 10141: unless very long strings are involved, though the default limit on
! 10142: stack size varies from system to system. Values from 8Mb to 64Mb are
1.1 misho 10143: common. You can find your default limit by running the command:
10144:
10145: ulimit -s
10146:
1.1.1.4 ! misho 10147: Unfortunately, the effect of running out of stack is often SIGSEGV,
! 10148: though sometimes a more explicit error message is given. You can nor-
1.1 misho 10149: mally increase the limit on stack size by code such as this:
10150:
10151: struct rlimit rlim;
10152: getrlimit(RLIMIT_STACK, &rlim);
10153: rlim.rlim_cur = 100*1024*1024;
10154: setrlimit(RLIMIT_STACK, &rlim);
10155:
1.1.1.4 ! misho 10156: This reads the current limits (soft and hard) using getrlimit(), then
! 10157: attempts to increase the soft limit to 100Mb using setrlimit(). You
! 10158: must do this before calling pcre[16|32]_exec().
1.1 misho 10159:
10160: Changing stack size in Mac OS X
10161:
10162: Using setrlimit(), as described above, should also work on Mac OS X. It
10163: is also possible to set a stack size when linking a program. There is a
1.1.1.4 ! misho 10164: discussion about stack sizes in Mac OS X at this web site:
1.1 misho 10165: http://developer.apple.com/qa/qa2005/qa1419.html.
10166:
10167:
10168: AUTHOR
10169:
10170: Philip Hazel
10171: University Computing Service
10172: Cambridge CB2 3QH, England.
10173:
10174:
10175: REVISION
10176:
1.1.1.4 ! misho 10177: Last updated: 24 June 2012
1.1.1.2 misho 10178: Copyright (c) 1997-2012 University of Cambridge.
1.1 misho 10179: ------------------------------------------------------------------------------
10180:
10181:
FreeBSD-CVSweb <freebsd-cvsweb@FreeBSD.org>