/[pcre]/code/trunk/doc/pcrepattern.3
ViewVC logotype

Contents of /code/trunk/doc/pcrepattern.3

Parent Directory Parent Directory | Revision Log Revision Log


Revision 512 - (show annotations) (download)
Tue Mar 30 11:11:52 2010 UTC (4 years, 3 months ago) by ph10
File size: 105334 byte(s)
Fix compile problems when heap is in use

1 .TH PCREPATTERN 3
2 .SH NAME
3 PCRE - Perl-compatible regular expressions
4 .SH "PCRE REGULAR EXPRESSION DETAILS"
5 .rs
6 .sp
7 The syntax and semantics of the regular expressions that are supported by PCRE
8 are described in detail below. There is a quick-reference syntax summary in the
9 .\" HREF
10 \fBpcresyntax\fP
11 .\"
12 page. PCRE tries to match Perl syntax and semantics as closely as it can. PCRE
13 also supports some alternative regular expression syntax (which does not
14 conflict with the Perl syntax) in order to provide some compatibility with
15 regular expressions in Python, .NET, and Oniguruma.
16 .P
17 Perl's regular expressions are described in its own documentation, and
18 regular expressions in general are covered in a number of books, some of which
19 have copious examples. Jeffrey Friedl's "Mastering Regular Expressions",
20 published by O'Reilly, covers regular expressions in great detail. This
21 description of PCRE's regular expressions is intended as reference material.
22 .P
23 The original operation of PCRE was on strings of one-byte characters. However,
24 there is now also support for UTF-8 character strings. To use this,
25 PCRE must be built to include UTF-8 support, and you must call
26 \fBpcre_compile()\fP or \fBpcre_compile2()\fP with the PCRE_UTF8 option. There
27 is also a special sequence that can be given at the start of a pattern:
28 .sp
29 (*UTF8)
30 .sp
31 Starting a pattern with this sequence is equivalent to setting the PCRE_UTF8
32 option. This feature is not Perl-compatible. How setting UTF-8 mode affects
33 pattern matching is mentioned in several places below. There is also a summary
34 of UTF-8 features in the
35 .\" HTML <a href="pcre.html#utf8support">
36 .\" </a>
37 section on UTF-8 support
38 .\"
39 in the main
40 .\" HREF
41 \fBpcre\fP
42 .\"
43 page.
44 .P
45 The remainder of this document discusses the patterns that are supported by
46 PCRE when its main matching function, \fBpcre_exec()\fP, is used.
47 From release 6.0, PCRE offers a second matching function,
48 \fBpcre_dfa_exec()\fP, which matches using a different algorithm that is not
49 Perl-compatible. Some of the features discussed below are not available when
50 \fBpcre_dfa_exec()\fP is used. The advantages and disadvantages of the
51 alternative function, and how it differs from the normal function, are
52 discussed in the
53 .\" HREF
54 \fBpcrematching\fP
55 .\"
56 page.
57 .
58 .
59 .SH "NEWLINE CONVENTIONS"
60 .rs
61 .sp
62 PCRE supports five different conventions for indicating line breaks in
63 strings: a single CR (carriage return) character, a single LF (linefeed)
64 character, the two-character sequence CRLF, any of the three preceding, or any
65 Unicode newline sequence. The
66 .\" HREF
67 \fBpcreapi\fP
68 .\"
69 page has
70 .\" HTML <a href="pcreapi.html#newlines">
71 .\" </a>
72 further discussion
73 .\"
74 about newlines, and shows how to set the newline convention in the
75 \fIoptions\fP arguments for the compiling and matching functions.
76 .P
77 It is also possible to specify a newline convention by starting a pattern
78 string with one of the following five sequences:
79 .sp
80 (*CR) carriage return
81 (*LF) linefeed
82 (*CRLF) carriage return, followed by linefeed
83 (*ANYCRLF) any of the three above
84 (*ANY) all Unicode newline sequences
85 .sp
86 These override the default and the options given to \fBpcre_compile()\fP or
87 \fBpcre_compile2()\fP. For example, on a Unix system where LF is the default
88 newline sequence, the pattern
89 .sp
90 (*CR)a.b
91 .sp
92 changes the convention to CR. That pattern matches "a\enb" because LF is no
93 longer a newline. Note that these special settings, which are not
94 Perl-compatible, are recognized only at the very start of a pattern, and that
95 they must be in upper case. If more than one of them is present, the last one
96 is used.
97 .P
98 The newline convention does not affect what the \eR escape sequence matches. By
99 default, this is any Unicode newline sequence, for Perl compatibility. However,
100 this can be changed; see the description of \eR in the section entitled
101 .\" HTML <a href="#newlineseq">
102 .\" </a>
103 "Newline sequences"
104 .\"
105 below. A change of \eR setting can be combined with a change of newline
106 convention.
107 .
108 .
109 .SH "CHARACTERS AND METACHARACTERS"
110 .rs
111 .sp
112 A regular expression is a pattern that is matched against a subject string from
113 left to right. Most characters stand for themselves in a pattern, and match the
114 corresponding characters in the subject. As a trivial example, the pattern
115 .sp
116 The quick brown fox
117 .sp
118 matches a portion of a subject string that is identical to itself. When
119 caseless matching is specified (the PCRE_CASELESS option), letters are matched
120 independently of case. In UTF-8 mode, PCRE always understands the concept of
121 case for characters whose values are less than 128, so caseless matching is
122 always possible. For characters with higher values, the concept of case is
123 supported if PCRE is compiled with Unicode property support, but not otherwise.
124 If you want to use caseless matching for characters 128 and above, you must
125 ensure that PCRE is compiled with Unicode property support as well as with
126 UTF-8 support.
127 .P
128 The power of regular expressions comes from the ability to include alternatives
129 and repetitions in the pattern. These are encoded in the pattern by the use of
130 \fImetacharacters\fP, which do not stand for themselves but instead are
131 interpreted in some special way.
132 .P
133 There are two different sets of metacharacters: those that are recognized
134 anywhere in the pattern except within square brackets, and those that are
135 recognized within square brackets. Outside square brackets, the metacharacters
136 are as follows:
137 .sp
138 \e general escape character with several uses
139 ^ assert start of string (or line, in multiline mode)
140 $ assert end of string (or line, in multiline mode)
141 . match any character except newline (by default)
142 [ start character class definition
143 | start of alternative branch
144 ( start subpattern
145 ) end subpattern
146 ? extends the meaning of (
147 also 0 or 1 quantifier
148 also quantifier minimizer
149 * 0 or more quantifier
150 + 1 or more quantifier
151 also "possessive quantifier"
152 { start min/max quantifier
153 .sp
154 Part of a pattern that is in square brackets is called a "character class". In
155 a character class the only metacharacters are:
156 .sp
157 \e general escape character
158 ^ negate the class, but only if the first character
159 - indicates character range
160 .\" JOIN
161 [ POSIX character class (only if followed by POSIX
162 syntax)
163 ] terminates the character class
164 .sp
165 The following sections describe the use of each of the metacharacters.
166 .
167 .
168 .SH BACKSLASH
169 .rs
170 .sp
171 The backslash character has several uses. Firstly, if it is followed by a
172 non-alphanumeric character, it takes away any special meaning that character
173 may have. This use of backslash as an escape character applies both inside and
174 outside character classes.
175 .P
176 For example, if you want to match a * character, you write \e* in the pattern.
177 This escaping action applies whether or not the following character would
178 otherwise be interpreted as a metacharacter, so it is always safe to precede a
179 non-alphanumeric with backslash to specify that it stands for itself. In
180 particular, if you want to match a backslash, you write \e\e.
181 .P
182 If a pattern is compiled with the PCRE_EXTENDED option, whitespace in the
183 pattern (other than in a character class) and characters between a # outside
184 a character class and the next newline are ignored. An escaping backslash can
185 be used to include a whitespace or # character as part of the pattern.
186 .P
187 If you want to remove the special meaning from a sequence of characters, you
188 can do so by putting them between \eQ and \eE. This is different from Perl in
189 that $ and @ are handled as literals in \eQ...\eE sequences in PCRE, whereas in
190 Perl, $ and @ cause variable interpolation. Note the following examples:
191 .sp
192 Pattern PCRE matches Perl matches
193 .sp
194 .\" JOIN
195 \eQabc$xyz\eE abc$xyz abc followed by the
196 contents of $xyz
197 \eQabc\e$xyz\eE abc\e$xyz abc\e$xyz
198 \eQabc\eE\e$\eQxyz\eE abc$xyz abc$xyz
199 .sp
200 The \eQ...\eE sequence is recognized both inside and outside character classes.
201 .
202 .
203 .\" HTML <a name="digitsafterbackslash"></a>
204 .SS "Non-printing characters"
205 .rs
206 .sp
207 A second use of backslash provides a way of encoding non-printing characters
208 in patterns in a visible manner. There is no restriction on the appearance of
209 non-printing characters, apart from the binary zero that terminates a pattern,
210 but when a pattern is being prepared by text editing, it is often easier to use
211 one of the following escape sequences than the binary character it represents:
212 .sp
213 \ea alarm, that is, the BEL character (hex 07)
214 \ecx "control-x", where x is any character
215 \ee escape (hex 1B)
216 \ef formfeed (hex 0C)
217 \en linefeed (hex 0A)
218 \er carriage return (hex 0D)
219 \et tab (hex 09)
220 \eddd character with octal code ddd, or back reference
221 \exhh character with hex code hh
222 \ex{hhh..} character with hex code hhh..
223 .sp
224 The precise effect of \ecx is as follows: if x is a lower case letter, it
225 is converted to upper case. Then bit 6 of the character (hex 40) is inverted.
226 Thus \ecz becomes hex 1A, but \ec{ becomes hex 3B, while \ec; becomes hex
227 7B.
228 .P
229 After \ex, from zero to two hexadecimal digits are read (letters can be in
230 upper or lower case). Any number of hexadecimal digits may appear between \ex{
231 and }, but the value of the character code must be less than 256 in non-UTF-8
232 mode, and less than 2**31 in UTF-8 mode. That is, the maximum value in
233 hexadecimal is 7FFFFFFF. Note that this is bigger than the largest Unicode code
234 point, which is 10FFFF.
235 .P
236 If characters other than hexadecimal digits appear between \ex{ and }, or if
237 there is no terminating }, this form of escape is not recognized. Instead, the
238 initial \ex will be interpreted as a basic hexadecimal escape, with no
239 following digits, giving a character whose value is zero.
240 .P
241 Characters whose value is less than 256 can be defined by either of the two
242 syntaxes for \ex. There is no difference in the way they are handled. For
243 example, \exdc is exactly the same as \ex{dc}.
244 .P
245 After \e0 up to two further octal digits are read. If there are fewer than two
246 digits, just those that are present are used. Thus the sequence \e0\ex\e07
247 specifies two binary zeros followed by a BEL character (code value 7). Make
248 sure you supply two digits after the initial zero if the pattern character that
249 follows is itself an octal digit.
250 .P
251 The handling of a backslash followed by a digit other than 0 is complicated.
252 Outside a character class, PCRE reads it and any following digits as a decimal
253 number. If the number is less than 10, or if there have been at least that many
254 previous capturing left parentheses in the expression, the entire sequence is
255 taken as a \fIback reference\fP. A description of how this works is given
256 .\" HTML <a href="#backreferences">
257 .\" </a>
258 later,
259 .\"
260 following the discussion of
261 .\" HTML <a href="#subpattern">
262 .\" </a>
263 parenthesized subpatterns.
264 .\"
265 .P
266 Inside a character class, or if the decimal number is greater than 9 and there
267 have not been that many capturing subpatterns, PCRE re-reads up to three octal
268 digits following the backslash, and uses them to generate a data character. Any
269 subsequent digits stand for themselves. In non-UTF-8 mode, the value of a
270 character specified in octal must be less than \e400. In UTF-8 mode, values up
271 to \e777 are permitted. For example:
272 .sp
273 \e040 is another way of writing a space
274 .\" JOIN
275 \e40 is the same, provided there are fewer than 40
276 previous capturing subpatterns
277 \e7 is always a back reference
278 .\" JOIN
279 \e11 might be a back reference, or another way of
280 writing a tab
281 \e011 is always a tab
282 \e0113 is a tab followed by the character "3"
283 .\" JOIN
284 \e113 might be a back reference, otherwise the
285 character with octal code 113
286 .\" JOIN
287 \e377 might be a back reference, otherwise
288 the byte consisting entirely of 1 bits
289 .\" JOIN
290 \e81 is either a back reference, or a binary zero
291 followed by the two characters "8" and "1"
292 .sp
293 Note that octal values of 100 or greater must not be introduced by a leading
294 zero, because no more than three octal digits are ever read.
295 .P
296 All the sequences that define a single character value can be used both inside
297 and outside character classes. In addition, inside a character class, the
298 sequence \eb is interpreted as the backspace character (hex 08), and the
299 sequences \eR and \eX are interpreted as the characters "R" and "X",
300 respectively. Outside a character class, these sequences have different
301 meanings
302 .\" HTML <a href="#uniextseq">
303 .\" </a>
304 (see below).
305 .\"
306 .
307 .
308 .SS "Absolute and relative back references"
309 .rs
310 .sp
311 The sequence \eg followed by an unsigned or a negative number, optionally
312 enclosed in braces, is an absolute or relative back reference. A named back
313 reference can be coded as \eg{name}. Back references are discussed
314 .\" HTML <a href="#backreferences">
315 .\" </a>
316 later,
317 .\"
318 following the discussion of
319 .\" HTML <a href="#subpattern">
320 .\" </a>
321 parenthesized subpatterns.
322 .\"
323 .
324 .
325 .SS "Absolute and relative subroutine calls"
326 .rs
327 .sp
328 For compatibility with Oniguruma, the non-Perl syntax \eg followed by a name or
329 a number enclosed either in angle brackets or single quotes, is an alternative
330 syntax for referencing a subpattern as a "subroutine". Details are discussed
331 .\" HTML <a href="#onigurumasubroutines">
332 .\" </a>
333 later.
334 .\"
335 Note that \eg{...} (Perl syntax) and \eg<...> (Oniguruma syntax) are \fInot\fP
336 synonymous. The former is a back reference; the latter is a
337 .\" HTML <a href="#subpatternsassubroutines">
338 .\" </a>
339 subroutine
340 .\"
341 call.
342 .
343 .
344 .SS "Generic character types"
345 .rs
346 .sp
347 Another use of backslash is for specifying generic character types. The
348 following are always recognized:
349 .sp
350 \ed any decimal digit
351 \eD any character that is not a decimal digit
352 \eh any horizontal whitespace character
353 \eH any character that is not a horizontal whitespace character
354 \es any whitespace character
355 \eS any character that is not a whitespace character
356 \ev any vertical whitespace character
357 \eV any character that is not a vertical whitespace character
358 \ew any "word" character
359 \eW any "non-word" character
360 .sp
361 Each pair of escape sequences partitions the complete set of characters into
362 two disjoint sets. Any given character matches one, and only one, of each pair.
363 .P
364 These character type sequences can appear both inside and outside character
365 classes. They each match one character of the appropriate type. If the current
366 matching point is at the end of the subject string, all of them fail, since
367 there is no character to match.
368 .P
369 For compatibility with Perl, \es does not match the VT character (code 11).
370 This makes it different from the the POSIX "space" class. The \es characters
371 are HT (9), LF (10), FF (12), CR (13), and space (32). If "use locale;" is
372 included in a Perl script, \es may match the VT character. In PCRE, it never
373 does.
374 .P
375 In UTF-8 mode, characters with values greater than 128 never match \ed, \es, or
376 \ew, and always match \eD, \eS, and \eW. This is true even when Unicode
377 character property support is available. These sequences retain their original
378 meanings from before UTF-8 support was available, mainly for efficiency
379 reasons. Note that this also affects \eb, because it is defined in terms of \ew
380 and \eW.
381 .P
382 The sequences \eh, \eH, \ev, and \eV are Perl 5.10 features. In contrast to the
383 other sequences, these do match certain high-valued codepoints in UTF-8 mode.
384 The horizontal space characters are:
385 .sp
386 U+0009 Horizontal tab
387 U+0020 Space
388 U+00A0 Non-break space
389 U+1680 Ogham space mark
390 U+180E Mongolian vowel separator
391 U+2000 En quad
392 U+2001 Em quad
393 U+2002 En space
394 U+2003 Em space
395 U+2004 Three-per-em space
396 U+2005 Four-per-em space
397 U+2006 Six-per-em space
398 U+2007 Figure space
399 U+2008 Punctuation space
400 U+2009 Thin space
401 U+200A Hair space
402 U+202F Narrow no-break space
403 U+205F Medium mathematical space
404 U+3000 Ideographic space
405 .sp
406 The vertical space characters are:
407 .sp
408 U+000A Linefeed
409 U+000B Vertical tab
410 U+000C Formfeed
411 U+000D Carriage return
412 U+0085 Next line
413 U+2028 Line separator
414 U+2029 Paragraph separator
415 .P
416 A "word" character is an underscore or any character less than 256 that is a
417 letter or digit. The definition of letters and digits is controlled by PCRE's
418 low-valued character tables, and may vary if locale-specific matching is taking
419 place (see
420 .\" HTML <a href="pcreapi.html#localesupport">
421 .\" </a>
422 "Locale support"
423 .\"
424 in the
425 .\" HREF
426 \fBpcreapi\fP
427 .\"
428 page). For example, in a French locale such as "fr_FR" in Unix-like systems,
429 or "french" in Windows, some character codes greater than 128 are used for
430 accented letters, and these are matched by \ew. The use of locales with Unicode
431 is discouraged.
432 .
433 .
434 .\" HTML <a name="newlineseq"></a>
435 .SS "Newline sequences"
436 .rs
437 .sp
438 Outside a character class, by default, the escape sequence \eR matches any
439 Unicode newline sequence. This is a Perl 5.10 feature. In non-UTF-8 mode \eR is
440 equivalent to the following:
441 .sp
442 (?>\er\en|\en|\ex0b|\ef|\er|\ex85)
443 .sp
444 This is an example of an "atomic group", details of which are given
445 .\" HTML <a href="#atomicgroup">
446 .\" </a>
447 below.
448 .\"
449 This particular group matches either the two-character sequence CR followed by
450 LF, or one of the single characters LF (linefeed, U+000A), VT (vertical tab,
451 U+000B), FF (formfeed, U+000C), CR (carriage return, U+000D), or NEL (next
452 line, U+0085). The two-character sequence is treated as a single unit that
453 cannot be split.
454 .P
455 In UTF-8 mode, two additional characters whose codepoints are greater than 255
456 are added: LS (line separator, U+2028) and PS (paragraph separator, U+2029).
457 Unicode character property support is not needed for these characters to be
458 recognized.
459 .P
460 It is possible to restrict \eR to match only CR, LF, or CRLF (instead of the
461 complete set of Unicode line endings) by setting the option PCRE_BSR_ANYCRLF
462 either at compile time or when the pattern is matched. (BSR is an abbrevation
463 for "backslash R".) This can be made the default when PCRE is built; if this is
464 the case, the other behaviour can be requested via the PCRE_BSR_UNICODE option.
465 It is also possible to specify these settings by starting a pattern string with
466 one of the following sequences:
467 .sp
468 (*BSR_ANYCRLF) CR, LF, or CRLF only
469 (*BSR_UNICODE) any Unicode newline sequence
470 .sp
471 These override the default and the options given to \fBpcre_compile()\fP or
472 \fBpcre_compile2()\fP, but they can be overridden by options given to
473 \fBpcre_exec()\fP or \fBpcre_dfa_exec()\fP. Note that these special settings,
474 which are not Perl-compatible, are recognized only at the very start of a
475 pattern, and that they must be in upper case. If more than one of them is
476 present, the last one is used. They can be combined with a change of newline
477 convention, for example, a pattern can start with:
478 .sp
479 (*ANY)(*BSR_ANYCRLF)
480 .sp
481 Inside a character class, \eR matches the letter "R".
482 .
483 .
484 .\" HTML <a name="uniextseq"></a>
485 .SS Unicode character properties
486 .rs
487 .sp
488 When PCRE is built with Unicode character property support, three additional
489 escape sequences that match characters with specific properties are available.
490 When not in UTF-8 mode, these sequences are of course limited to testing
491 characters whose codepoints are less than 256, but they do work in this mode.
492 The extra escape sequences are:
493 .sp
494 \ep{\fIxx\fP} a character with the \fIxx\fP property
495 \eP{\fIxx\fP} a character without the \fIxx\fP property
496 \eX an extended Unicode sequence
497 .sp
498 The property names represented by \fIxx\fP above are limited to the Unicode
499 script names, the general category properties, and "Any", which matches any
500 character (including newline). Other properties such as "InMusicalSymbols" are
501 not currently supported by PCRE. Note that \eP{Any} does not match any
502 characters, so always causes a match failure.
503 .P
504 Sets of Unicode characters are defined as belonging to certain scripts. A
505 character from one of these sets can be matched using a script name. For
506 example:
507 .sp
508 \ep{Greek}
509 \eP{Han}
510 .sp
511 Those that are not part of an identified script are lumped together as
512 "Common". The current list of scripts is:
513 .P
514 Arabic,
515 Armenian,
516 Avestan,
517 Balinese,
518 Bamum,
519 Bengali,
520 Bopomofo,
521 Braille,
522 Buginese,
523 Buhid,
524 Canadian_Aboriginal,
525 Carian,
526 Cham,
527 Cherokee,
528 Common,
529 Coptic,
530 Cuneiform,
531 Cypriot,
532 Cyrillic,
533 Deseret,
534 Devanagari,
535 Egyptian_Hieroglyphs,
536 Ethiopic,
537 Georgian,
538 Glagolitic,
539 Gothic,
540 Greek,
541 Gujarati,
542 Gurmukhi,
543 Han,
544 Hangul,
545 Hanunoo,
546 Hebrew,
547 Hiragana,
548 Imperial_Aramaic,
549 Inherited,
550 Inscriptional_Pahlavi,
551 Inscriptional_Parthian,
552 Javanese,
553 Kaithi,
554 Kannada,
555 Katakana,
556 Kayah_Li,
557 Kharoshthi,
558 Khmer,
559 Lao,
560 Latin,
561 Lepcha,
562 Limbu,
563 Linear_B,
564 Lisu,
565 Lycian,
566 Lydian,
567 Malayalam,
568 Meetei_Mayek,
569 Mongolian,
570 Myanmar,
571 New_Tai_Lue,
572 Nko,
573 Ogham,
574 Old_Italic,
575 Old_Persian,
576 Old_South_Arabian,
577 Old_Turkic,
578 Ol_Chiki,
579 Oriya,
580 Osmanya,
581 Phags_Pa,
582 Phoenician,
583 Rejang,
584 Runic,
585 Samaritan,
586 Saurashtra,
587 Shavian,
588 Sinhala,
589 Sundanese,
590 Syloti_Nagri,
591 Syriac,
592 Tagalog,
593 Tagbanwa,
594 Tai_Le,
595 Tai_Tham,
596 Tai_Viet,
597 Tamil,
598 Telugu,
599 Thaana,
600 Thai,
601 Tibetan,
602 Tifinagh,
603 Ugaritic,
604 Vai,
605 Yi.
606 .P
607 Each character has exactly one general category property, specified by a
608 two-letter abbreviation. For compatibility with Perl, negation can be specified
609 by including a circumflex between the opening brace and the property name. For
610 example, \ep{^Lu} is the same as \eP{Lu}.
611 .P
612 If only one letter is specified with \ep or \eP, it includes all the general
613 category properties that start with that letter. In this case, in the absence
614 of negation, the curly brackets in the escape sequence are optional; these two
615 examples have the same effect:
616 .sp
617 \ep{L}
618 \epL
619 .sp
620 The following general category property codes are supported:
621 .sp
622 C Other
623 Cc Control
624 Cf Format
625 Cn Unassigned
626 Co Private use
627 Cs Surrogate
628 .sp
629 L Letter
630 Ll Lower case letter
631 Lm Modifier letter
632 Lo Other letter
633 Lt Title case letter
634 Lu Upper case letter
635 .sp
636 M Mark
637 Mc Spacing mark
638 Me Enclosing mark
639 Mn Non-spacing mark
640 .sp
641 N Number
642 Nd Decimal number
643 Nl Letter number
644 No Other number
645 .sp
646 P Punctuation
647 Pc Connector punctuation
648 Pd Dash punctuation
649 Pe Close punctuation
650 Pf Final punctuation
651 Pi Initial punctuation
652 Po Other punctuation
653 Ps Open punctuation
654 .sp
655 S Symbol
656 Sc Currency symbol
657 Sk Modifier symbol
658 Sm Mathematical symbol
659 So Other symbol
660 .sp
661 Z Separator
662 Zl Line separator
663 Zp Paragraph separator
664 Zs Space separator
665 .sp
666 The special property L& is also supported: it matches a character that has
667 the Lu, Ll, or Lt property, in other words, a letter that is not classified as
668 a modifier or "other".
669 .P
670 The Cs (Surrogate) property applies only to characters in the range U+D800 to
671 U+DFFF. Such characters are not valid in UTF-8 strings (see RFC 3629) and so
672 cannot be tested by PCRE, unless UTF-8 validity checking has been turned off
673 (see the discussion of PCRE_NO_UTF8_CHECK in the
674 .\" HREF
675 \fBpcreapi\fP
676 .\"
677 page). Perl does not support the Cs property.
678 .P
679 The long synonyms for property names that Perl supports (such as \ep{Letter})
680 are not supported by PCRE, nor is it permitted to prefix any of these
681 properties with "Is".
682 .P
683 No character that is in the Unicode table has the Cn (unassigned) property.
684 Instead, this property is assumed for any code point that is not in the
685 Unicode table.
686 .P
687 Specifying caseless matching does not affect these escape sequences. For
688 example, \ep{Lu} always matches only upper case letters.
689 .P
690 The \eX escape matches any number of Unicode characters that form an extended
691 Unicode sequence. \eX is equivalent to
692 .sp
693 (?>\ePM\epM*)
694 .sp
695 That is, it matches a character without the "mark" property, followed by zero
696 or more characters with the "mark" property, and treats the sequence as an
697 atomic group
698 .\" HTML <a href="#atomicgroup">
699 .\" </a>
700 (see below).
701 .\"
702 Characters with the "mark" property are typically accents that affect the
703 preceding character. None of them have codepoints less than 256, so in
704 non-UTF-8 mode \eX matches any one character.
705 .P
706 Matching characters by Unicode property is not fast, because PCRE has to search
707 a structure that contains data for over fifteen thousand characters. That is
708 why the traditional escape sequences such as \ed and \ew do not use Unicode
709 properties in PCRE.
710 .
711 .
712 .\" HTML <a name="resetmatchstart"></a>
713 .SS "Resetting the match start"
714 .rs
715 .sp
716 The escape sequence \eK, which is a Perl 5.10 feature, causes any previously
717 matched characters not to be included in the final matched sequence. For
718 example, the pattern:
719 .sp
720 foo\eKbar
721 .sp
722 matches "foobar", but reports that it has matched "bar". This feature is
723 similar to a lookbehind assertion
724 .\" HTML <a href="#lookbehind">
725 .\" </a>
726 (described below).
727 .\"
728 However, in this case, the part of the subject before the real match does not
729 have to be of fixed length, as lookbehind assertions do. The use of \eK does
730 not interfere with the setting of
731 .\" HTML <a href="#subpattern">
732 .\" </a>
733 captured substrings.
734 .\"
735 For example, when the pattern
736 .sp
737 (foo)\eKbar
738 .sp
739 matches "foobar", the first substring is still set to "foo".
740 .P
741 Perl documents that the use of \eK within assertions is "not well defined". In
742 PCRE, \eK is acted upon when it occurs inside positive assertions, but is
743 ignored in negative assertions.
744 .
745 .
746 .\" HTML <a name="smallassertions"></a>
747 .SS "Simple assertions"
748 .rs
749 .sp
750 The final use of backslash is for certain simple assertions. An assertion
751 specifies a condition that has to be met at a particular point in a match,
752 without consuming any characters from the subject string. The use of
753 subpatterns for more complicated assertions is described
754 .\" HTML <a href="#bigassertions">
755 .\" </a>
756 below.
757 .\"
758 The backslashed assertions are:
759 .sp
760 \eb matches at a word boundary
761 \eB matches when not at a word boundary
762 \eA matches at the start of the subject
763 \eZ matches at the end of the subject
764 also matches before a newline at the end of the subject
765 \ez matches only at the end of the subject
766 \eG matches at the first matching position in the subject
767 .sp
768 These assertions may not appear in character classes (but note that \eb has a
769 different meaning, namely the backspace character, inside a character class).
770 .P
771 A word boundary is a position in the subject string where the current character
772 and the previous character do not both match \ew or \eW (i.e. one matches
773 \ew and the other matches \eW), or the start or end of the string if the
774 first or last character matches \ew, respectively. Neither PCRE nor Perl has a
775 separte "start of word" or "end of word" metasequence. However, whatever
776 follows \eb normally determines which it is. For example, the fragment
777 \eba matches "a" at the start of a word.
778 .P
779 The \eA, \eZ, and \ez assertions differ from the traditional circumflex and
780 dollar (described in the next section) in that they only ever match at the very
781 start and end of the subject string, whatever options are set. Thus, they are
782 independent of multiline mode. These three assertions are not affected by the
783 PCRE_NOTBOL or PCRE_NOTEOL options, which affect only the behaviour of the
784 circumflex and dollar metacharacters. However, if the \fIstartoffset\fP
785 argument of \fBpcre_exec()\fP is non-zero, indicating that matching is to start
786 at a point other than the beginning of the subject, \eA can never match. The
787 difference between \eZ and \ez is that \eZ matches before a newline at the end
788 of the string as well as at the very end, whereas \ez matches only at the end.
789 .P
790 The \eG assertion is true only when the current matching position is at the
791 start point of the match, as specified by the \fIstartoffset\fP argument of
792 \fBpcre_exec()\fP. It differs from \eA when the value of \fIstartoffset\fP is
793 non-zero. By calling \fBpcre_exec()\fP multiple times with appropriate
794 arguments, you can mimic Perl's /g option, and it is in this kind of
795 implementation where \eG can be useful.
796 .P
797 Note, however, that PCRE's interpretation of \eG, as the start of the current
798 match, is subtly different from Perl's, which defines it as the end of the
799 previous match. In Perl, these can be different when the previously matched
800 string was empty. Because PCRE does just one match at a time, it cannot
801 reproduce this behaviour.
802 .P
803 If all the alternatives of a pattern begin with \eG, the expression is anchored
804 to the starting match position, and the "anchored" flag is set in the compiled
805 regular expression.
806 .
807 .
808 .SH "CIRCUMFLEX AND DOLLAR"
809 .rs
810 .sp
811 Outside a character class, in the default matching mode, the circumflex
812 character is an assertion that is true only if the current matching point is
813 at the start of the subject string. If the \fIstartoffset\fP argument of
814 \fBpcre_exec()\fP is non-zero, circumflex can never match if the PCRE_MULTILINE
815 option is unset. Inside a character class, circumflex has an entirely different
816 meaning
817 .\" HTML <a href="#characterclass">
818 .\" </a>
819 (see below).
820 .\"
821 .P
822 Circumflex need not be the first character of the pattern if a number of
823 alternatives are involved, but it should be the first thing in each alternative
824 in which it appears if the pattern is ever to match that branch. If all
825 possible alternatives start with a circumflex, that is, if the pattern is
826 constrained to match only at the start of the subject, it is said to be an
827 "anchored" pattern. (There are also other constructs that can cause a pattern
828 to be anchored.)
829 .P
830 A dollar character is an assertion that is true only if the current matching
831 point is at the end of the subject string, or immediately before a newline
832 at the end of the string (by default). Dollar need not be the last character of
833 the pattern if a number of alternatives are involved, but it should be the last
834 item in any branch in which it appears. Dollar has no special meaning in a
835 character class.
836 .P
837 The meaning of dollar can be changed so that it matches only at the very end of
838 the string, by setting the PCRE_DOLLAR_ENDONLY option at compile time. This
839 does not affect the \eZ assertion.
840 .P
841 The meanings of the circumflex and dollar characters are changed if the
842 PCRE_MULTILINE option is set. When this is the case, a circumflex matches
843 immediately after internal newlines as well as at the start of the subject
844 string. It does not match after a newline that ends the string. A dollar
845 matches before any newlines in the string, as well as at the very end, when
846 PCRE_MULTILINE is set. When newline is specified as the two-character
847 sequence CRLF, isolated CR and LF characters do not indicate newlines.
848 .P
849 For example, the pattern /^abc$/ matches the subject string "def\enabc" (where
850 \en represents a newline) in multiline mode, but not otherwise. Consequently,
851 patterns that are anchored in single line mode because all branches start with
852 ^ are not anchored in multiline mode, and a match for circumflex is possible
853 when the \fIstartoffset\fP argument of \fBpcre_exec()\fP is non-zero. The
854 PCRE_DOLLAR_ENDONLY option is ignored if PCRE_MULTILINE is set.
855 .P
856 Note that the sequences \eA, \eZ, and \ez can be used to match the start and
857 end of the subject in both modes, and if all branches of a pattern start with
858 \eA it is always anchored, whether or not PCRE_MULTILINE is set.
859 .
860 .
861 .SH "FULL STOP (PERIOD, DOT)"
862 .rs
863 .sp
864 Outside a character class, a dot in the pattern matches any one character in
865 the subject string except (by default) a character that signifies the end of a
866 line. In UTF-8 mode, the matched character may be more than one byte long.
867 .P
868 When a line ending is defined as a single character, dot never matches that
869 character; when the two-character sequence CRLF is used, dot does not match CR
870 if it is immediately followed by LF, but otherwise it matches all characters
871 (including isolated CRs and LFs). When any Unicode line endings are being
872 recognized, dot does not match CR or LF or any of the other line ending
873 characters.
874 .P
875 The behaviour of dot with regard to newlines can be changed. If the PCRE_DOTALL
876 option is set, a dot matches any one character, without exception. If the
877 two-character sequence CRLF is present in the subject string, it takes two dots
878 to match it.
879 .P
880 The handling of dot is entirely independent of the handling of circumflex and
881 dollar, the only relationship being that they both involve newlines. Dot has no
882 special meaning in a character class.
883 .
884 .
885 .SH "MATCHING A SINGLE BYTE"
886 .rs
887 .sp
888 Outside a character class, the escape sequence \eC matches any one byte, both
889 in and out of UTF-8 mode. Unlike a dot, it always matches any line-ending
890 characters. The feature is provided in Perl in order to match individual bytes
891 in UTF-8 mode. Because it breaks up UTF-8 characters into individual bytes,
892 what remains in the string may be a malformed UTF-8 string. For this reason,
893 the \eC escape sequence is best avoided.
894 .P
895 PCRE does not allow \eC to appear in lookbehind assertions
896 .\" HTML <a href="#lookbehind">
897 .\" </a>
898 (described below),
899 .\"
900 because in UTF-8 mode this would make it impossible to calculate the length of
901 the lookbehind.
902 .
903 .
904 .\" HTML <a name="characterclass"></a>
905 .SH "SQUARE BRACKETS AND CHARACTER CLASSES"
906 .rs
907 .sp
908 An opening square bracket introduces a character class, terminated by a closing
909 square bracket. A closing square bracket on its own is not special by default.
910 However, if the PCRE_JAVASCRIPT_COMPAT option is set, a lone closing square
911 bracket causes a compile-time error. If a closing square bracket is required as
912 a member of the class, it should be the first data character in the class
913 (after an initial circumflex, if present) or escaped with a backslash.
914 .P
915 A character class matches a single character in the subject. In UTF-8 mode, the
916 character may be more than one byte long. A matched character must be in the
917 set of characters defined by the class, unless the first character in the class
918 definition is a circumflex, in which case the subject character must not be in
919 the set defined by the class. If a circumflex is actually required as a member
920 of the class, ensure it is not the first character, or escape it with a
921 backslash.
922 .P
923 For example, the character class [aeiou] matches any lower case vowel, while
924 [^aeiou] matches any character that is not a lower case vowel. Note that a
925 circumflex is just a convenient notation for specifying the characters that
926 are in the class by enumerating those that are not. A class that starts with a
927 circumflex is not an assertion; it still consumes a character from the subject
928 string, and therefore it fails if the current pointer is at the end of the
929 string.
930 .P
931 In UTF-8 mode, characters with values greater than 255 can be included in a
932 class as a literal string of bytes, or by using the \ex{ escaping mechanism.
933 .P
934 When caseless matching is set, any letters in a class represent both their
935 upper case and lower case versions, so for example, a caseless [aeiou] matches
936 "A" as well as "a", and a caseless [^aeiou] does not match "A", whereas a
937 caseful version would. In UTF-8 mode, PCRE always understands the concept of
938 case for characters whose values are less than 128, so caseless matching is
939 always possible. For characters with higher values, the concept of case is
940 supported if PCRE is compiled with Unicode property support, but not otherwise.
941 If you want to use caseless matching in UTF8-mode for characters 128 and above,
942 you must ensure that PCRE is compiled with Unicode property support as well as
943 with UTF-8 support.
944 .P
945 Characters that might indicate line breaks are never treated in any special way
946 when matching character classes, whatever line-ending sequence is in use, and
947 whatever setting of the PCRE_DOTALL and PCRE_MULTILINE options is used. A class
948 such as [^a] always matches one of these characters.
949 .P
950 The minus (hyphen) character can be used to specify a range of characters in a
951 character class. For example, [d-m] matches any letter between d and m,
952 inclusive. If a minus character is required in a class, it must be escaped with
953 a backslash or appear in a position where it cannot be interpreted as
954 indicating a range, typically as the first or last character in the class.
955 .P
956 It is not possible to have the literal character "]" as the end character of a
957 range. A pattern such as [W-]46] is interpreted as a class of two characters
958 ("W" and "-") followed by a literal string "46]", so it would match "W46]" or
959 "-46]". However, if the "]" is escaped with a backslash it is interpreted as
960 the end of range, so [W-\e]46] is interpreted as a class containing a range
961 followed by two other characters. The octal or hexadecimal representation of
962 "]" can also be used to end a range.
963 .P
964 Ranges operate in the collating sequence of character values. They can also be
965 used for characters specified numerically, for example [\e000-\e037]. In UTF-8
966 mode, ranges can include characters whose values are greater than 255, for
967 example [\ex{100}-\ex{2ff}].
968 .P
969 If a range that includes letters is used when caseless matching is set, it
970 matches the letters in either case. For example, [W-c] is equivalent to
971 [][\e\e^_`wxyzabc], matched caselessly, and in non-UTF-8 mode, if character
972 tables for a French locale are in use, [\exc8-\excb] matches accented E
973 characters in both cases. In UTF-8 mode, PCRE supports the concept of case for
974 characters with values greater than 128 only when it is compiled with Unicode
975 property support.
976 .P
977 The character types \ed, \eD, \ep, \eP, \es, \eS, \ew, and \eW may also appear
978 in a character class, and add the characters that they match to the class. For
979 example, [\edABCDEF] matches any hexadecimal digit. A circumflex can
980 conveniently be used with the upper case character types to specify a more
981 restricted set of characters than the matching lower case type. For example,
982 the class [^\eW_] matches any letter or digit, but not underscore.
983 .P
984 The only metacharacters that are recognized in character classes are backslash,
985 hyphen (only where it can be interpreted as specifying a range), circumflex
986 (only at the start), opening square bracket (only when it can be interpreted as
987 introducing a POSIX class name - see the next section), and the terminating
988 closing square bracket. However, escaping other non-alphanumeric characters
989 does no harm.
990 .
991 .
992 .SH "POSIX CHARACTER CLASSES"
993 .rs
994 .sp
995 Perl supports the POSIX notation for character classes. This uses names
996 enclosed by [: and :] within the enclosing square brackets. PCRE also supports
997 this notation. For example,
998 .sp
999 [01[:alpha:]%]
1000 .sp
1001 matches "0", "1", any alphabetic character, or "%". The supported class names
1002 are
1003 .sp
1004 alnum letters and digits
1005 alpha letters
1006 ascii character codes 0 - 127
1007 blank space or tab only
1008 cntrl control characters
1009 digit decimal digits (same as \ed)
1010 graph printing characters, excluding space
1011 lower lower case letters
1012 print printing characters, including space
1013 punct printing characters, excluding letters and digits
1014 space white space (not quite the same as \es)
1015 upper upper case letters
1016 word "word" characters (same as \ew)
1017 xdigit hexadecimal digits
1018 .sp
1019 The "space" characters are HT (9), LF (10), VT (11), FF (12), CR (13), and
1020 space (32). Notice that this list includes the VT character (code 11). This
1021 makes "space" different to \es, which does not include VT (for Perl
1022 compatibility).
1023 .P
1024 The name "word" is a Perl extension, and "blank" is a GNU extension from Perl
1025 5.8. Another Perl extension is negation, which is indicated by a ^ character
1026 after the colon. For example,
1027 .sp
1028 [12[:^digit:]]
1029 .sp
1030 matches "1", "2", or any non-digit. PCRE (and Perl) also recognize the POSIX
1031 syntax [.ch.] and [=ch=] where "ch" is a "collating element", but these are not
1032 supported, and an error is given if they are encountered.
1033 .P
1034 In UTF-8 mode, characters with values greater than 128 do not match any of
1035 the POSIX character classes.
1036 .
1037 .
1038 .SH "VERTICAL BAR"
1039 .rs
1040 .sp
1041 Vertical bar characters are used to separate alternative patterns. For example,
1042 the pattern
1043 .sp
1044 gilbert|sullivan
1045 .sp
1046 matches either "gilbert" or "sullivan". Any number of alternatives may appear,
1047 and an empty alternative is permitted (matching the empty string). The matching
1048 process tries each alternative in turn, from left to right, and the first one
1049 that succeeds is used. If the alternatives are within a subpattern
1050 .\" HTML <a href="#subpattern">
1051 .\" </a>
1052 (defined below),
1053 .\"
1054 "succeeds" means matching the rest of the main pattern as well as the
1055 alternative in the subpattern.
1056 .
1057 .
1058 .SH "INTERNAL OPTION SETTING"
1059 .rs
1060 .sp
1061 The settings of the PCRE_CASELESS, PCRE_MULTILINE, PCRE_DOTALL, and
1062 PCRE_EXTENDED options (which are Perl-compatible) can be changed from within
1063 the pattern by a sequence of Perl option letters enclosed between "(?" and ")".
1064 The option letters are
1065 .sp
1066 i for PCRE_CASELESS
1067 m for PCRE_MULTILINE
1068 s for PCRE_DOTALL
1069 x for PCRE_EXTENDED
1070 .sp
1071 For example, (?im) sets caseless, multiline matching. It is also possible to
1072 unset these options by preceding the letter with a hyphen, and a combined
1073 setting and unsetting such as (?im-sx), which sets PCRE_CASELESS and
1074 PCRE_MULTILINE while unsetting PCRE_DOTALL and PCRE_EXTENDED, is also
1075 permitted. If a letter appears both before and after the hyphen, the option is
1076 unset.
1077 .P
1078 The PCRE-specific options PCRE_DUPNAMES, PCRE_UNGREEDY, and PCRE_EXTRA can be
1079 changed in the same way as the Perl-compatible options by using the characters
1080 J, U and X respectively.
1081 .P
1082 When one of these option changes occurs at top level (that is, not inside
1083 subpattern parentheses), the change applies to the remainder of the pattern
1084 that follows. If the change is placed right at the start of a pattern, PCRE
1085 extracts it into the global options (and it will therefore show up in data
1086 extracted by the \fBpcre_fullinfo()\fP function).
1087 .P
1088 An option change within a subpattern (see below for a description of
1089 subpatterns) affects only that part of the current pattern that follows it, so
1090 .sp
1091 (a(?i)b)c
1092 .sp
1093 matches abc and aBc and no other strings (assuming PCRE_CASELESS is not used).
1094 By this means, options can be made to have different settings in different
1095 parts of the pattern. Any changes made in one alternative do carry on
1096 into subsequent branches within the same subpattern. For example,
1097 .sp
1098 (a(?i)b|c)
1099 .sp
1100 matches "ab", "aB", "c", and "C", even though when matching "C" the first
1101 branch is abandoned before the option setting. This is because the effects of
1102 option settings happen at compile time. There would be some very weird
1103 behaviour otherwise.
1104 .P
1105 \fBNote:\fP There are other PCRE-specific options that can be set by the
1106 application when the compile or match functions are called. In some cases the
1107 pattern can contain special leading sequences such as (*CRLF) to override what
1108 the application has set or what has been defaulted. Details are given in the
1109 section entitled
1110 .\" HTML <a href="#newlineseq">
1111 .\" </a>
1112 "Newline sequences"
1113 .\"
1114 above. There is also the (*UTF8) leading sequence that can be used to set UTF-8
1115 mode; this is equivalent to setting the PCRE_UTF8 option.
1116 .
1117 .
1118 .\" HTML <a name="subpattern"></a>
1119 .SH SUBPATTERNS
1120 .rs
1121 .sp
1122 Subpatterns are delimited by parentheses (round brackets), which can be nested.
1123 Turning part of a pattern into a subpattern does two things:
1124 .sp
1125 1. It localizes a set of alternatives. For example, the pattern
1126 .sp
1127 cat(aract|erpillar|)
1128 .sp
1129 matches one of the words "cat", "cataract", or "caterpillar". Without the
1130 parentheses, it would match "cataract", "erpillar" or an empty string.
1131 .sp
1132 2. It sets up the subpattern as a capturing subpattern. This means that, when
1133 the whole pattern matches, that portion of the subject string that matched the
1134 subpattern is passed back to the caller via the \fIovector\fP argument of
1135 \fBpcre_exec()\fP. Opening parentheses are counted from left to right (starting
1136 from 1) to obtain numbers for the capturing subpatterns.
1137 .P
1138 For example, if the string "the red king" is matched against the pattern
1139 .sp
1140 the ((red|white) (king|queen))
1141 .sp
1142 the captured substrings are "red king", "red", and "king", and are numbered 1,
1143 2, and 3, respectively.
1144 .P
1145 The fact that plain parentheses fulfil two functions is not always helpful.
1146 There are often times when a grouping subpattern is required without a
1147 capturing requirement. If an opening parenthesis is followed by a question mark
1148 and a colon, the subpattern does not do any capturing, and is not counted when
1149 computing the number of any subsequent capturing subpatterns. For example, if
1150 the string "the white queen" is matched against the pattern
1151 .sp
1152 the ((?:red|white) (king|queen))
1153 .sp
1154 the captured substrings are "white queen" and "queen", and are numbered 1 and
1155 2. The maximum number of capturing subpatterns is 65535.
1156 .P
1157 As a convenient shorthand, if any option settings are required at the start of
1158 a non-capturing subpattern, the option letters may appear between the "?" and
1159 the ":". Thus the two patterns
1160 .sp
1161 (?i:saturday|sunday)
1162 (?:(?i)saturday|sunday)
1163 .sp
1164 match exactly the same set of strings. Because alternative branches are tried
1165 from left to right, and options are not reset until the end of the subpattern
1166 is reached, an option setting in one branch does affect subsequent branches, so
1167 the above patterns match "SUNDAY" as well as "Saturday".
1168 .
1169 .
1170 .\" HTML <a name="dupsubpatternnumber"></a>
1171 .SH "DUPLICATE SUBPATTERN NUMBERS"
1172 .rs
1173 .sp
1174 Perl 5.10 introduced a feature whereby each alternative in a subpattern uses
1175 the same numbers for its capturing parentheses. Such a subpattern starts with
1176 (?| and is itself a non-capturing subpattern. For example, consider this
1177 pattern:
1178 .sp
1179 (?|(Sat)ur|(Sun))day
1180 .sp
1181 Because the two alternatives are inside a (?| group, both sets of capturing
1182 parentheses are numbered one. Thus, when the pattern matches, you can look
1183 at captured substring number one, whichever alternative matched. This construct
1184 is useful when you want to capture part, but not all, of one of a number of
1185 alternatives. Inside a (?| group, parentheses are numbered as usual, but the
1186 number is reset at the start of each branch. The numbers of any capturing
1187 buffers that follow the subpattern start after the highest number used in any
1188 branch. The following example is taken from the Perl documentation.
1189 The numbers underneath show in which buffer the captured content will be
1190 stored.
1191 .sp
1192 # before ---------------branch-reset----------- after
1193 / ( a ) (?| x ( y ) z | (p (q) r) | (t) u (v) ) ( z ) /x
1194 # 1 2 2 3 2 3 4
1195 .sp
1196 A back reference to a numbered subpattern uses the most recent value that is
1197 set for that number by any subpattern. The following pattern matches "abcabc"
1198 or "defdef":
1199 .sp
1200 /(?|(abc)|(def))\e1/
1201 .sp
1202 In contrast, a recursive or "subroutine" call to a numbered subpattern always
1203 refers to the first one in the pattern with the given number. The following
1204 pattern matches "abcabc" or "defabc":
1205 .sp
1206 /(?|(abc)|(def))(?1)/
1207 .sp
1208 If a
1209 .\" HTML <a href="#conditions">
1210 .\" </a>
1211 condition test
1212 .\"
1213 for a subpattern's having matched refers to a non-unique number, the test is
1214 true if any of the subpatterns of that number have matched.
1215 .P
1216 An alternative approach to using this "branch reset" feature is to use
1217 duplicate named subpatterns, as described in the next section.
1218 .
1219 .
1220 .SH "NAMED SUBPATTERNS"
1221 .rs
1222 .sp
1223 Identifying capturing parentheses by number is simple, but it can be very hard
1224 to keep track of the numbers in complicated regular expressions. Furthermore,
1225 if an expression is modified, the numbers may change. To help with this
1226 difficulty, PCRE supports the naming of subpatterns. This feature was not
1227 added to Perl until release 5.10. Python had the feature earlier, and PCRE
1228 introduced it at release 4.0, using the Python syntax. PCRE now supports both
1229 the Perl and the Python syntax. Perl allows identically numbered subpatterns to
1230 have different names, but PCRE does not.
1231 .P
1232 In PCRE, a subpattern can be named in one of three ways: (?<name>...) or
1233 (?'name'...) as in Perl, or (?P<name>...) as in Python. References to capturing
1234 parentheses from other parts of the pattern, such as
1235 .\" HTML <a href="#backreferences">
1236 .\" </a>
1237 back references,
1238 .\"
1239 .\" HTML <a href="#recursion">
1240 .\" </a>
1241 recursion,
1242 .\"
1243 and
1244 .\" HTML <a href="#conditions">
1245 .\" </a>
1246 conditions,
1247 .\"
1248 can be made by name as well as by number.
1249 .P
1250 Names consist of up to 32 alphanumeric characters and underscores. Named
1251 capturing parentheses are still allocated numbers as well as names, exactly as
1252 if the names were not present. The PCRE API provides function calls for
1253 extracting the name-to-number translation table from a compiled pattern. There
1254 is also a convenience function for extracting a captured substring by name.
1255 .P
1256 By default, a name must be unique within a pattern, but it is possible to relax
1257 this constraint by setting the PCRE_DUPNAMES option at compile time. (Duplicate
1258 names are also always permitted for subpatterns with the same number, set up as
1259 described in the previous section.) Duplicate names can be useful for patterns
1260 where only one instance of the named parentheses can match. Suppose you want to
1261 match the name of a weekday, either as a 3-letter abbreviation or as the full
1262 name, and in both cases you want to extract the abbreviation. This pattern
1263 (ignoring the line breaks) does the job:
1264 .sp
1265 (?<DN>Mon|Fri|Sun)(?:day)?|
1266 (?<DN>Tue)(?:sday)?|
1267 (?<DN>Wed)(?:nesday)?|
1268 (?<DN>Thu)(?:rsday)?|
1269 (?<DN>Sat)(?:urday)?
1270 .sp
1271 There are five capturing substrings, but only one is ever set after a match.
1272 (An alternative way of solving this problem is to use a "branch reset"
1273 subpattern, as described in the previous section.)
1274 .P
1275 The convenience function for extracting the data by name returns the substring
1276 for the first (and in this example, the only) subpattern of that name that
1277 matched. This saves searching to find which numbered subpattern it was.
1278 .P
1279 If you make a back reference to a non-unique named subpattern from elsewhere in
1280 the pattern, the one that corresponds to the first occurrence of the name is
1281 used. In the absence of duplicate numbers (see the previous section) this is
1282 the one with the lowest number. If you use a named reference in a condition
1283 test (see the
1284 .\"
1285 .\" HTML <a href="#conditions">
1286 .\" </a>
1287 section about conditions
1288 .\"
1289 below), either to check whether a subpattern has matched, or to check for
1290 recursion, all subpatterns with the same name are tested. If the condition is
1291 true for any one of them, the overall condition is true. This is the same
1292 behaviour as testing by number. For further details of the interfaces for
1293 handling named subpatterns, see the
1294 .\" HREF
1295 \fBpcreapi\fP
1296 .\"
1297 documentation.
1298 .P
1299 \fBWarning:\fP You cannot use different names to distinguish between two
1300 subpatterns with the same number because PCRE uses only the numbers when
1301 matching. For this reason, an error is given at compile time if different names
1302 are given to subpatterns with the same number. However, you can give the same
1303 name to subpatterns with the same number, even when PCRE_DUPNAMES is not set.
1304 .
1305 .
1306 .SH REPETITION
1307 .rs
1308 .sp
1309 Repetition is specified by quantifiers, which can follow any of the following
1310 items:
1311 .sp
1312 a literal data character
1313 the dot metacharacter
1314 the \eC escape sequence
1315 the \eX escape sequence (in UTF-8 mode with Unicode properties)
1316 the \eR escape sequence
1317 an escape such as \ed that matches a single character
1318 a character class
1319 a back reference (see next section)
1320 a parenthesized subpattern (unless it is an assertion)
1321 a recursive or "subroutine" call to a subpattern
1322 .sp
1323 The general repetition quantifier specifies a minimum and maximum number of
1324 permitted matches, by giving the two numbers in curly brackets (braces),
1325 separated by a comma. The numbers must be less than 65536, and the first must
1326 be less than or equal to the second. For example:
1327 .sp
1328 z{2,4}
1329 .sp
1330 matches "zz", "zzz", or "zzzz". A closing brace on its own is not a special
1331 character. If the second number is omitted, but the comma is present, there is
1332 no upper limit; if the second number and the comma are both omitted, the
1333 quantifier specifies an exact number of required matches. Thus
1334 .sp
1335 [aeiou]{3,}
1336 .sp
1337 matches at least 3 successive vowels, but may match many more, while
1338 .sp
1339 \ed{8}
1340 .sp
1341 matches exactly 8 digits. An opening curly bracket that appears in a position
1342 where a quantifier is not allowed, or one that does not match the syntax of a
1343 quantifier, is taken as a literal character. For example, {,6} is not a
1344 quantifier, but a literal string of four characters.
1345 .P
1346 In UTF-8 mode, quantifiers apply to UTF-8 characters rather than to individual
1347 bytes. Thus, for example, \ex{100}{2} matches two UTF-8 characters, each of
1348 which is represented by a two-byte sequence. Similarly, when Unicode property
1349 support is available, \eX{3} matches three Unicode extended sequences, each of
1350 which may be several bytes long (and they may be of different lengths).
1351 .P
1352 The quantifier {0} is permitted, causing the expression to behave as if the
1353 previous item and the quantifier were not present. This may be useful for
1354 subpatterns that are referenced as
1355 .\" HTML <a href="#subpatternsassubroutines">
1356 .\" </a>
1357 subroutines
1358 .\"
1359 from elsewhere in the pattern. Items other than subpatterns that have a {0}
1360 quantifier are omitted from the compiled pattern.
1361 .P
1362 For convenience, the three most common quantifiers have single-character
1363 abbreviations:
1364 .sp
1365 * is equivalent to {0,}
1366 + is equivalent to {1,}
1367 ? is equivalent to {0,1}
1368 .sp
1369 It is possible to construct infinite loops by following a subpattern that can
1370 match no characters with a quantifier that has no upper limit, for example:
1371 .sp
1372 (a?)*
1373 .sp
1374 Earlier versions of Perl and PCRE used to give an error at compile time for
1375 such patterns. However, because there are cases where this can be useful, such
1376 patterns are now accepted, but if any repetition of the subpattern does in fact
1377 match no characters, the loop is forcibly broken.
1378 .P
1379 By default, the quantifiers are "greedy", that is, they match as much as
1380 possible (up to the maximum number of permitted times), without causing the
1381 rest of the pattern to fail. The classic example of where this gives problems
1382 is in trying to match comments in C programs. These appear between /* and */
1383 and within the comment, individual * and / characters may appear. An attempt to
1384 match C comments by applying the pattern
1385 .sp
1386 /\e*.*\e*/
1387 .sp
1388 to the string
1389 .sp
1390 /* first comment */ not comment /* second comment */
1391 .sp
1392 fails, because it matches the entire string owing to the greediness of the .*
1393 item.
1394 .P
1395 However, if a quantifier is followed by a question mark, it ceases to be
1396 greedy, and instead matches the minimum number of times possible, so the
1397 pattern
1398 .sp
1399 /\e*.*?\e*/
1400 .sp
1401 does the right thing with the C comments. The meaning of the various
1402 quantifiers is not otherwise changed, just the preferred number of matches.
1403 Do not confuse this use of question mark with its use as a quantifier in its
1404 own right. Because it has two uses, it can sometimes appear doubled, as in
1405 .sp
1406 \ed??\ed
1407 .sp
1408 which matches one digit by preference, but can match two if that is the only
1409 way the rest of the pattern matches.
1410 .P
1411 If the PCRE_UNGREEDY option is set (an option that is not available in Perl),
1412 the quantifiers are not greedy by default, but individual ones can be made
1413 greedy by following them with a question mark. In other words, it inverts the
1414 default behaviour.
1415 .P
1416 When a parenthesized subpattern is quantified with a minimum repeat count that
1417 is greater than 1 or with a limited maximum, more memory is required for the
1418 compiled pattern, in proportion to the size of the minimum or maximum.
1419 .P
1420 If a pattern starts with .* or .{0,} and the PCRE_DOTALL option (equivalent
1421 to Perl's /s) is set, thus allowing the dot to match newlines, the pattern is
1422 implicitly anchored, because whatever follows will be tried against every
1423 character position in the subject string, so there is no point in retrying the
1424 overall match at any position after the first. PCRE normally treats such a
1425 pattern as though it were preceded by \eA.
1426 .P
1427 In cases where it is known that the subject string contains no newlines, it is
1428 worth setting PCRE_DOTALL in order to obtain this optimization, or
1429 alternatively using ^ to indicate anchoring explicitly.
1430 .P
1431 However, there is one situation where the optimization cannot be used. When .*
1432 is inside capturing parentheses that are the subject of a back reference
1433 elsewhere in the pattern, a match at the start may fail where a later one
1434 succeeds. Consider, for example:
1435 .sp
1436 (.*)abc\e1
1437 .sp
1438 If the subject is "xyz123abc123" the match point is the fourth character. For
1439 this reason, such a pattern is not implicitly anchored.
1440 .P
1441 When a capturing subpattern is repeated, the value captured is the substring
1442 that matched the final iteration. For example, after
1443 .sp
1444 (tweedle[dume]{3}\es*)+
1445 .sp
1446 has matched "tweedledum tweedledee" the value of the captured substring is
1447 "tweedledee". However, if there are nested capturing subpatterns, the
1448 corresponding captured values may have been set in previous iterations. For
1449 example, after
1450 .sp
1451 /(a|(b))+/
1452 .sp
1453 matches "aba" the value of the second captured substring is "b".
1454 .
1455 .
1456 .\" HTML <a name="atomicgroup"></a>
1457 .SH "ATOMIC GROUPING AND POSSESSIVE QUANTIFIERS"
1458 .rs
1459 .sp
1460 With both maximizing ("greedy") and minimizing ("ungreedy" or "lazy")
1461 repetition, failure of what follows normally causes the repeated item to be
1462 re-evaluated to see if a different number of repeats allows the rest of the
1463 pattern to match. Sometimes it is useful to prevent this, either to change the
1464 nature of the match, or to cause it fail earlier than it otherwise might, when
1465 the author of the pattern knows there is no point in carrying on.
1466 .P
1467 Consider, for example, the pattern \ed+foo when applied to the subject line
1468 .sp
1469 123456bar
1470 .sp
1471 After matching all 6 digits and then failing to match "foo", the normal
1472 action of the matcher is to try again with only 5 digits matching the \ed+
1473 item, and then with 4, and so on, before ultimately failing. "Atomic grouping"
1474 (a term taken from Jeffrey Friedl's book) provides the means for specifying
1475 that once a subpattern has matched, it is not to be re-evaluated in this way.
1476 .P
1477 If we use atomic grouping for the previous example, the matcher gives up
1478 immediately on failing to match "foo" the first time. The notation is a kind of
1479 special parenthesis, starting with (?> as in this example:
1480 .sp
1481 (?>\ed+)foo
1482 .sp
1483 This kind of parenthesis "locks up" the part of the pattern it contains once
1484 it has matched, and a failure further into the pattern is prevented from
1485 backtracking into it. Backtracking past it to previous items, however, works as
1486 normal.
1487 .P
1488 An alternative description is that a subpattern of this type matches the string
1489 of characters that an identical standalone pattern would match, if anchored at
1490 the current point in the subject string.
1491 .P
1492 Atomic grouping subpatterns are not capturing subpatterns. Simple cases such as
1493 the above example can be thought of as a maximizing repeat that must swallow
1494 everything it can. So, while both \ed+ and \ed+? are prepared to adjust the
1495 number of digits they match in order to make the rest of the pattern match,
1496 (?>\ed+) can only match an entire sequence of digits.
1497 .P
1498 Atomic groups in general can of course contain arbitrarily complicated
1499 subpatterns, and can be nested. However, when the subpattern for an atomic
1500 group is just a single repeated item, as in the example above, a simpler
1501 notation, called a "possessive quantifier" can be used. This consists of an
1502 additional + character following a quantifier. Using this notation, the
1503 previous example can be rewritten as
1504 .sp
1505 \ed++foo
1506 .sp
1507 Note that a possessive quantifier can be used with an entire group, for
1508 example:
1509 .sp
1510 (abc|xyz){2,3}+
1511 .sp
1512 Possessive quantifiers are always greedy; the setting of the PCRE_UNGREEDY
1513 option is ignored. They are a convenient notation for the simpler forms of
1514 atomic group. However, there is no difference in the meaning of a possessive
1515 quantifier and the equivalent atomic group, though there may be a performance
1516 difference; possessive quantifiers should be slightly faster.
1517 .P
1518 The possessive quantifier syntax is an extension to the Perl 5.8 syntax.
1519 Jeffrey Friedl originated the idea (and the name) in the first edition of his
1520 book. Mike McCloskey liked it, so implemented it when he built Sun's Java
1521 package, and PCRE copied it from there. It ultimately found its way into Perl
1522 at release 5.10.
1523 .P
1524 PCRE has an optimization that automatically "possessifies" certain simple
1525 pattern constructs. For example, the sequence A+B is treated as A++B because
1526 there is no point in backtracking into a sequence of A's when B must follow.
1527 .P
1528 When a pattern contains an unlimited repeat inside a subpattern that can itself
1529 be repeated an unlimited number of times, the use of an atomic group is the
1530 only way to avoid some failing matches taking a very long time indeed. The
1531 pattern
1532 .sp
1533 (\eD+|<\ed+>)*[!?]
1534 .sp
1535 matches an unlimited number of substrings that either consist of non-digits, or
1536 digits enclosed in <>, followed by either ! or ?. When it matches, it runs
1537 quickly. However, if it is applied to
1538 .sp
1539 aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
1540 .sp
1541 it takes a long time before reporting failure. This is because the string can
1542 be divided between the internal \eD+ repeat and the external * repeat in a
1543 large number of ways, and all have to be tried. (The example uses [!?] rather
1544 than a single character at the end, because both PCRE and Perl have an
1545 optimization that allows for fast failure when a single character is used. They
1546 remember the last single character that is required for a match, and fail early
1547 if it is not present in the string.) If the pattern is changed so that it uses
1548 an atomic group, like this:
1549 .sp
1550 ((?>\eD+)|<\ed+>)*[!?]
1551 .sp
1552 sequences of non-digits cannot be broken, and failure happens quickly.
1553 .
1554 .
1555 .\" HTML <a name="backreferences"></a>
1556 .SH "BACK REFERENCES"
1557 .rs
1558 .sp
1559 Outside a character class, a backslash followed by a digit greater than 0 (and
1560 possibly further digits) is a back reference to a capturing subpattern earlier
1561 (that is, to its left) in the pattern, provided there have been that many
1562 previous capturing left parentheses.
1563 .P
1564 However, if the decimal number following the backslash is less than 10, it is
1565 always taken as a back reference, and causes an error only if there are not
1566 that many capturing left parentheses in the entire pattern. In other words, the
1567 parentheses that are referenced need not be to the left of the reference for
1568 numbers less than 10. A "forward back reference" of this type can make sense
1569 when a repetition is involved and the subpattern to the right has participated
1570 in an earlier iteration.
1571 .P
1572 It is not possible to have a numerical "forward back reference" to a subpattern
1573 whose number is 10 or more using this syntax because a sequence such as \e50 is
1574 interpreted as a character defined in octal. See the subsection entitled
1575 "Non-printing characters"
1576 .\" HTML <a href="#digitsafterbackslash">
1577 .\" </a>
1578 above
1579 .\"
1580 for further details of the handling of digits following a backslash. There is
1581 no such problem when named parentheses are used. A back reference to any
1582 subpattern is possible using named parentheses (see below).
1583 .P
1584 Another way of avoiding the ambiguity inherent in the use of digits following a
1585 backslash is to use the \eg escape sequence, which is a feature introduced in
1586 Perl 5.10. This escape must be followed by an unsigned number or a negative
1587 number, optionally enclosed in braces. These examples are all identical:
1588 .sp
1589 (ring), \e1
1590 (ring), \eg1
1591 (ring), \eg{1}
1592 .sp
1593 An unsigned number specifies an absolute reference without the ambiguity that
1594 is present in the older syntax. It is also useful when literal digits follow
1595 the reference. A negative number is a relative reference. Consider this
1596 example:
1597 .sp
1598 (abc(def)ghi)\eg{-1}
1599 .sp
1600 The sequence \eg{-1} is a reference to the most recently started capturing
1601 subpattern before \eg, that is, is it equivalent to \e2. Similarly, \eg{-2}
1602 would be equivalent to \e1. The use of relative references can be helpful in
1603 long patterns, and also in patterns that are created by joining together
1604 fragments that contain references within themselves.
1605 .P
1606 A back reference matches whatever actually matched the capturing subpattern in
1607 the current subject string, rather than anything matching the subpattern
1608 itself (see
1609 .\" HTML <a href="#subpatternsassubroutines">
1610 .\" </a>
1611 "Subpatterns as subroutines"
1612 .\"
1613 below for a way of doing that). So the pattern
1614 .sp
1615 (sens|respons)e and \e1ibility
1616 .sp
1617 matches "sense and sensibility" and "response and responsibility", but not
1618 "sense and responsibility". If caseful matching is in force at the time of the
1619 back reference, the case of letters is relevant. For example,
1620 .sp
1621 ((?i)rah)\es+\e1
1622 .sp
1623 matches "rah rah" and "RAH RAH", but not "RAH rah", even though the original
1624 capturing subpattern is matched caselessly.
1625 .P
1626 There are several different ways of writing back references to named
1627 subpatterns. The .NET syntax \ek{name} and the Perl syntax \ek<name> or
1628 \ek'name' are supported, as is the Python syntax (?P=name). Perl 5.10's unified
1629 back reference syntax, in which \eg can be used for both numeric and named
1630 references, is also supported. We could rewrite the above example in any of
1631 the following ways:
1632 .sp
1633 (?<p1>(?i)rah)\es+\ek<p1>
1634 (?'p1'(?i)rah)\es+\ek{p1}
1635 (?P<p1>(?i)rah)\es+(?P=p1)
1636 (?<p1>(?i)rah)\es+\eg{p1}
1637 .sp
1638 A subpattern that is referenced by name may appear in the pattern before or
1639 after the reference.
1640 .P
1641 There may be more than one back reference to the same subpattern. If a
1642 subpattern has not actually been used in a particular match, any back
1643 references to it always fail by default. For example, the pattern
1644 .sp
1645 (a|(bc))\e2
1646 .sp
1647 always fails if it starts to match "a" rather than "bc". However, if the
1648 PCRE_JAVASCRIPT_COMPAT option is set at compile time, a back reference to an
1649 unset value matches an empty string.
1650 .P
1651 Because there may be many capturing parentheses in a pattern, all digits
1652 following a backslash are taken as part of a potential back reference number.
1653 If the pattern continues with a digit character, some delimiter must be used to
1654 terminate the back reference. If the PCRE_EXTENDED option is set, this can be
1655 whitespace. Otherwise, the \eg{ syntax or an empty comment (see
1656 .\" HTML <a href="#comments">
1657 .\" </a>
1658 "Comments"
1659 .\"
1660 below) can be used.
1661 .
1662 .SS "Recursive back references"
1663 .rs
1664 .sp
1665 A back reference that occurs inside the parentheses to which it refers fails
1666 when the subpattern is first used, so, for example, (a\e1) never matches.
1667 However, such references can be useful inside repeated subpatterns. For
1668 example, the pattern
1669 .sp
1670 (a|b\e1)+
1671 .sp
1672 matches any number of "a"s and also "aba", "ababbaa" etc. At each iteration of
1673 the subpattern, the back reference matches the character string corresponding
1674 to the previous iteration. In order for this to work, the pattern must be such
1675 that the first iteration does not need to match the back reference. This can be
1676 done using alternation, as in the example above, or by a quantifier with a
1677 minimum of zero.
1678 .P
1679 Back references of this type cause the group that they reference to be treated
1680 as an
1681 .\" HTML <a href="#atomicgroup">
1682 .\" </a>
1683 atomic group.
1684 .\"
1685 Once the whole group has been matched, a subsequent matching failure cannot
1686 cause backtracking into the middle of the group.
1687 .
1688 .
1689 .\" HTML <a name="bigassertions"></a>
1690 .SH ASSERTIONS
1691 .rs
1692 .sp
1693 An assertion is a test on the characters following or preceding the current
1694 matching point that does not actually consume any characters. The simple
1695 assertions coded as \eb, \eB, \eA, \eG, \eZ, \ez, ^ and $ are described
1696 .\" HTML <a href="#smallassertions">
1697 .\" </a>
1698 above.
1699 .\"
1700 .P
1701 More complicated assertions are coded as subpatterns. There are two kinds:
1702 those that look ahead of the current position in the subject string, and those
1703 that look behind it. An assertion subpattern is matched in the normal way,
1704 except that it does not cause the current matching position to be changed.
1705 .P
1706 Assertion subpatterns are not capturing subpatterns, and may not be repeated,
1707 because it makes no sense to assert the same thing several times. If any kind
1708 of assertion contains capturing subpatterns within it, these are counted for
1709 the purposes of numbering the capturing subpatterns in the whole pattern.
1710 However, substring capturing is carried out only for positive assertions,
1711 because it does not make sense for negative assertions.
1712 .
1713 .
1714 .SS "Lookahead assertions"
1715 .rs
1716 .sp
1717 Lookahead assertions start with (?= for positive assertions and (?! for
1718 negative assertions. For example,
1719 .sp
1720 \ew+(?=;)
1721 .sp
1722 matches a word followed by a semicolon, but does not include the semicolon in
1723 the match, and
1724 .sp
1725 foo(?!bar)
1726 .sp
1727 matches any occurrence of "foo" that is not followed by "bar". Note that the
1728 apparently similar pattern
1729 .sp
1730 (?!foo)bar
1731 .sp
1732 does not find an occurrence of "bar" that is preceded by something other than
1733 "foo"; it finds any occurrence of "bar" whatsoever, because the assertion
1734 (?!foo) is always true when the next three characters are "bar". A
1735 lookbehind assertion is needed to achieve the other effect.
1736 .P
1737 If you want to force a matching failure at some point in a pattern, the most
1738 convenient way to do it is with (?!) because an empty string always matches, so
1739 an assertion that requires there not to be an empty string must always fail.
1740 The Perl 5.10 backtracking control verb (*FAIL) or (*F) is essentially a
1741 synonym for (?!).
1742 .
1743 .
1744 .\" HTML <a name="lookbehind"></a>
1745 .SS "Lookbehind assertions"
1746 .rs
1747 .sp
1748 Lookbehind assertions start with (?<= for positive assertions and (?<! for
1749 negative assertions. For example,
1750 .sp
1751 (?<!foo)bar
1752 .sp
1753 does find an occurrence of "bar" that is not preceded by "foo". The contents of
1754 a lookbehind assertion are restricted such that all the strings it matches must
1755 have a fixed length. However, if there are several top-level alternatives, they
1756 do not all have to have the same fixed length. Thus
1757 .sp
1758 (?<=bullock|donkey)
1759 .sp
1760 is permitted, but
1761 .sp
1762 (?<!dogs?|cats?)
1763 .sp
1764 causes an error at compile time. Branches that match different length strings
1765 are permitted only at the top level of a lookbehind assertion. This is an
1766 extension compared with Perl (5.8 and 5.10), which requires all branches to
1767 match the same length of string. An assertion such as
1768 .sp
1769 (?<=ab(c|de))
1770 .sp
1771 is not permitted, because its single top-level branch can match two different
1772 lengths, but it is acceptable to PCRE if rewritten to use two top-level
1773 branches:
1774 .sp
1775 (?<=abc|abde)
1776 .sp
1777 In some cases, the Perl 5.10 escape sequence \eK
1778 .\" HTML <a href="#resetmatchstart">
1779 .\" </a>
1780 (see above)
1781 .\"
1782 can be used instead of a lookbehind assertion to get round the fixed-length
1783 restriction.
1784 .P
1785 The implementation of lookbehind assertions is, for each alternative, to
1786 temporarily move the current position back by the fixed length and then try to
1787 match. If there are insufficient characters before the current position, the
1788 assertion fails.
1789 .P
1790 PCRE does not allow the \eC escape (which matches a single byte in UTF-8 mode)
1791 to appear in lookbehind assertions, because it makes it impossible to calculate
1792 the length of the lookbehind. The \eX and \eR escapes, which can match
1793 different numbers of bytes, are also not permitted.
1794 .P
1795 .\" HTML <a href="#subpatternsassubroutines">
1796 .\" </a>
1797 "Subroutine"
1798 .\"
1799 calls (see below) such as (?2) or (?&X) are permitted in lookbehinds, as long
1800 as the subpattern matches a fixed-length string.
1801 .\" HTML <a href="#recursion">
1802 .\" </a>
1803 Recursion,
1804 .\"
1805 however, is not supported.
1806 .P
1807 Possessive quantifiers can be used in conjunction with lookbehind assertions to
1808 specify efficient matching of fixed-length strings at the end of subject
1809 strings. Consider a simple pattern such as
1810 .sp
1811 abcd$
1812 .sp
1813 when applied to a long string that does not match. Because matching proceeds
1814 from left to right, PCRE will look for each "a" in the subject and then see if
1815 what follows matches the rest of the pattern. If the pattern is specified as
1816 .sp
1817 ^.*abcd$
1818 .sp
1819 the initial .* matches the entire string at first, but when this fails (because
1820 there is no following "a"), it backtracks to match all but the last character,
1821 then all but the last two characters, and so on. Once again the search for "a"
1822 covers the entire string, from right to left, so we are no better off. However,
1823 if the pattern is written as
1824 .sp
1825 ^.*+(?<=abcd)
1826 .sp
1827 there can be no backtracking for the .*+ item; it can match only the entire
1828 string. The subsequent lookbehind assertion does a single test on the last four
1829 characters. If it fails, the match fails immediately. For long strings, this
1830 approach makes a significant difference to the processing time.
1831 .
1832 .
1833 .SS "Using multiple assertions"
1834 .rs
1835 .sp
1836 Several assertions (of any sort) may occur in succession. For example,
1837 .sp
1838 (?<=\ed{3})(?<!999)foo
1839 .sp
1840 matches "foo" preceded by three digits that are not "999". Notice that each of
1841 the assertions is applied independently at the same point in the subject
1842 string. First there is a check that the previous three characters are all
1843 digits, and then there is a check that the same three characters are not "999".
1844 This pattern does \fInot\fP match "foo" preceded by six characters, the first
1845 of which are digits and the last three of which are not "999". For example, it
1846 doesn't match "123abcfoo". A pattern to do that is
1847 .sp
1848 (?<=\ed{3}...)(?<!999)foo
1849 .sp
1850 This time the first assertion looks at the preceding six characters, checking
1851 that the first three are digits, and then the second assertion checks that the
1852 preceding three characters are not "999".
1853 .P
1854 Assertions can be nested in any combination. For example,
1855 .sp
1856 (?<=(?<!foo)bar)baz
1857 .sp
1858 matches an occurrence of "baz" that is preceded by "bar" which in turn is not
1859 preceded by "foo", while
1860 .sp
1861 (?<=\ed{3}(?!999)...)foo
1862 .sp
1863 is another pattern that matches "foo" preceded by three digits and any three
1864 characters that are not "999".
1865 .
1866 .
1867 .\" HTML <a name="conditions"></a>
1868 .SH "CONDITIONAL SUBPATTERNS"
1869 .rs
1870 .sp
1871 It is possible to cause the matching process to obey a subpattern
1872 conditionally or to choose between two alternative subpatterns, depending on
1873 the result of an assertion, or whether a specific capturing subpattern has
1874 already been matched. The two possible forms of conditional subpattern are:
1875 .sp
1876 (?(condition)yes-pattern)
1877 (?(condition)yes-pattern|no-pattern)
1878 .sp
1879 If the condition is satisfied, the yes-pattern is used; otherwise the
1880 no-pattern (if present) is used. If there are more than two alternatives in the
1881 subpattern, a compile-time error occurs.
1882 .P
1883 There are four kinds of condition: references to subpatterns, references to
1884 recursion, a pseudo-condition called DEFINE, and assertions.
1885 .
1886 .SS "Checking for a used subpattern by number"
1887 .rs
1888 .sp
1889 If the text between the parentheses consists of a sequence of digits, the
1890 condition is true if a capturing subpattern of that number has previously
1891 matched. If there is more than one capturing subpattern with the same number
1892 (see the earlier
1893 .\"
1894 .\" HTML <a href="#recursion">
1895 .\" </a>
1896 section about duplicate subpattern numbers),
1897 .\"
1898 the condition is true if any of them have been set. An alternative notation is
1899 to precede the digits with a plus or minus sign. In this case, the subpattern
1900 number is relative rather than absolute. The most recently opened parentheses
1901 can be referenced by (?(-1), the next most recent by (?(-2), and so on. In
1902 looping constructs it can also make sense to refer to subsequent groups with
1903 constructs such as (?(+2).
1904 .P
1905 Consider the following pattern, which contains non-significant white space to
1906 make it more readable (assume the PCRE_EXTENDED option) and to divide it into
1907 three parts for ease of discussion:
1908 .sp
1909 ( \e( )? [^()]+ (?(1) \e) )
1910 .sp
1911 The first part matches an optional opening parenthesis, and if that
1912 character is present, sets it as the first captured substring. The second part
1913 matches one or more characters that are not parentheses. The third part is a
1914 conditional subpattern that tests whether the first set of parentheses matched
1915 or not. If they did, that is, if subject started with an opening parenthesis,
1916 the condition is true, and so the yes-pattern is executed and a closing
1917 parenthesis is required. Otherwise, since no-pattern is not present, the
1918 subpattern matches nothing. In other words, this pattern matches a sequence of
1919 non-parentheses, optionally enclosed in parentheses.
1920 .P
1921 If you were embedding this pattern in a larger one, you could use a relative
1922 reference:
1923 .sp
1924 ...other stuff... ( \e( )? [^()]+ (?(-1) \e) ) ...
1925 .sp
1926 This makes the fragment independent of the parentheses in the larger pattern.
1927 .
1928 .SS "Checking for a used subpattern by name"
1929 .rs
1930 .sp
1931 Perl uses the syntax (?(<name>)...) or (?('name')...) to test for a used
1932 subpattern by name. For compatibility with earlier versions of PCRE, which had
1933 this facility before Perl, the syntax (?(name)...) is also recognized. However,
1934 there is a possible ambiguity with this syntax, because subpattern names may
1935 consist entirely of digits. PCRE looks first for a named subpattern; if it
1936 cannot find one and the name consists entirely of digits, PCRE looks for a
1937 subpattern of that number, which must be greater than zero. Using subpattern
1938 names that consist entirely of digits is not recommended.
1939 .P
1940 Rewriting the above example to use a named subpattern gives this:
1941 .sp
1942 (?<OPEN> \e( )? [^()]+ (?(<OPEN>) \e) )
1943 .sp
1944 If the name used in a condition of this kind is a duplicate, the test is
1945 applied to all subpatterns of the same name, and is true if any one of them has
1946 matched.
1947 .
1948 .SS "Checking for pattern recursion"
1949 .rs
1950 .sp
1951 If the condition is the string (R), and there is no subpattern with the name R,
1952 the condition is true if a recursive call to the whole pattern or any
1953 subpattern has been made. If digits or a name preceded by ampersand follow the
1954 letter R, for example:
1955 .sp
1956 (?(R3)...) or (?(R&name)...)
1957 .sp
1958 the condition is true if the most recent recursion is into a subpattern whose
1959 number or name is given. This condition does not check the entire recursion
1960 stack. If the name used in a condition of this kind is a duplicate, the test is
1961 applied to all subpatterns of the same name, and is true if any one of them is
1962 the most recent recursion.
1963 .P
1964 At "top level", all these recursion test conditions are false.
1965 .\" HTML <a href="#recursion">
1966 .\" </a>
1967 The syntax for recursive patterns
1968 .\"
1969 is described below.
1970 .
1971 .SS "Defining subpatterns for use by reference only"
1972 .rs
1973 .sp
1974 If the condition is the string (DEFINE), and there is no subpattern with the
1975 name DEFINE, the condition is always false. In this case, there may be only one
1976 alternative in the subpattern. It is always skipped if control reaches this
1977 point in the pattern; the idea of DEFINE is that it can be used to define
1978 "subroutines" that can be referenced from elsewhere. (The use of
1979 .\" HTML <a href="#subpatternsassubroutines">
1980 .\" </a>
1981 "subroutines"
1982 .\"
1983 is described below.) For example, a pattern to match an IPv4 address could be
1984 written like this (ignore whitespace and line breaks):
1985 .sp
1986 (?(DEFINE) (?<byte> 2[0-4]\ed | 25[0-5] | 1\ed\ed | [1-9]?\ed) )
1987 \eb (?&byte) (\e.(?&byte)){3} \eb
1988 .sp
1989 The first part of the pattern is a DEFINE group inside which a another group
1990 named "byte" is defined. This matches an individual component of an IPv4
1991 address (a number less than 256). When matching takes place, this part of the
1992 pattern is skipped because DEFINE acts like a false condition. The rest of the
1993 pattern uses references to the named group to match the four dot-separated
1994 components of an IPv4 address, insisting on a word boundary at each end.
1995 .
1996 .SS "Assertion conditions"
1997 .rs
1998 .sp
1999 If the condition is not in any of the above formats, it must be an assertion.
2000 This may be a positive or negative lookahead or lookbehind assertion. Consider
2001 this pattern, again containing non-significant white space, and with the two
2002 alternatives on the second line:
2003 .sp
2004 (?(?=[^a-z]*[a-z])
2005 \ed{2}-[a-z]{3}-\ed{2} | \ed{2}-\ed{2}-\ed{2} )
2006 .sp
2007 The condition is a positive lookahead assertion that matches an optional
2008 sequence of non-letters followed by a letter. In other words, it tests for the
2009 presence of at least one letter in the subject. If a letter is found, the
2010 subject is matched against the first alternative; otherwise it is matched
2011 against the second. This pattern matches strings in one of the two forms
2012 dd-aaa-dd or dd-dd-dd, where aaa are letters and dd are digits.
2013 .
2014 .
2015 .\" HTML <a name="comments"></a>
2016 .SH COMMENTS
2017 .rs
2018 .sp
2019 The sequence (?# marks the start of a comment that continues up to the next
2020 closing parenthesis. Nested parentheses are not permitted. The characters
2021 that make up a comment play no part in the pattern matching at all.
2022 .P
2023 If the PCRE_EXTENDED option is set, an unescaped # character outside a
2024 character class introduces a comment that continues to immediately after the
2025 next newline in the pattern.
2026 .
2027 .
2028 .\" HTML <a name="recursion"></a>
2029 .SH "RECURSIVE PATTERNS"
2030 .rs
2031 .sp
2032 Consider the problem of matching a string in parentheses, allowing for
2033 unlimited nested parentheses. Without the use of recursion, the best that can
2034 be done is to use a pattern that matches up to some fixed depth of nesting. It
2035 is not possible to handle an arbitrary nesting depth.
2036 .P
2037 For some time, Perl has provided a facility that allows regular expressions to
2038 recurse (amongst other things). It does this by interpolating Perl code in the
2039 expression at run time, and the code can refer to the expression itself. A Perl
2040 pattern using code interpolation to solve the parentheses problem can be
2041 created like this:
2042 .sp
2043 $re = qr{\e( (?: (?>[^()]+) | (?p{$re}) )* \e)}x;
2044 .sp
2045 The (?p{...}) item interpolates Perl code at run time, and in this case refers
2046 recursively to the pattern in which it appears.
2047 .P
2048 Obviously, PCRE cannot support the interpolation of Perl code. Instead, it
2049 supports special syntax for recursion of the entire pattern, and also for
2050 individual subpattern recursion. After its introduction in PCRE and Python,
2051 this kind of recursion was subsequently introduced into Perl at release 5.10.
2052 .P
2053 A special item that consists of (? followed by a number greater than zero and a
2054 closing parenthesis is a recursive call of the subpattern of the given number,
2055 provided that it occurs inside that subpattern. (If not, it is a
2056 .\" HTML <a href="#subpatternsassubroutines">
2057 .\" </a>
2058 "subroutine"
2059 .\"
2060 call, which is described in the next section.) The special item (?R) or (?0) is
2061 a recursive call of the entire regular expression.
2062 .P
2063 This PCRE pattern solves the nested parentheses problem (assume the
2064 PCRE_EXTENDED option is set so that white space is ignored):
2065 .sp
2066 \e( ( [^()]++ | (?R) )* \e)
2067 .sp
2068 First it matches an opening parenthesis. Then it matches any number of
2069 substrings which can either be a sequence of non-parentheses, or a recursive
2070 match of the pattern itself (that is, a correctly parenthesized substring).
2071 Finally there is a closing parenthesis. Note the use of a possessive quantifier
2072 to avoid backtracking into sequences of non-parentheses.
2073 .P
2074 If this were part of a larger pattern, you would not want to recurse the entire
2075 pattern, so instead you could use this:
2076 .sp
2077 ( \e( ( [^()]++ | (?1) )* \e) )
2078 .sp
2079 We have put the pattern into parentheses, and caused the recursion to refer to
2080 them instead of the whole pattern.
2081 .P
2082 In a larger pattern, keeping track of parenthesis numbers can be tricky. This
2083 is made easier by the use of relative references (a Perl 5.10 feature).
2084 Instead of (?1) in the pattern above you can write (?-2) to refer to the second
2085 most recently opened parentheses preceding the recursion. In other words, a
2086 negative number counts capturing parentheses leftwards from the point at which
2087 it is encountered.
2088 .P
2089 It is also possible to refer to subsequently opened parentheses, by writing
2090 references such as (?+2). However, these cannot be recursive because the
2091 reference is not inside the parentheses that are referenced. They are always
2092 .\" HTML <a href="#subpatternsassubroutines">
2093 .\" </a>
2094 "subroutine"
2095 .\"
2096 calls, as described in the next section.
2097 .P
2098 An alternative approach is to use named parentheses instead. The Perl syntax
2099 for this is (?&name); PCRE's earlier syntax (?P>name) is also supported. We
2100 could rewrite the above example as follows:
2101 .sp
2102 (?<pn> \e( ( [^()]++ | (?&pn) )* \e) )
2103 .sp
2104 If there is more than one subpattern with the same name, the earliest one is
2105 used.
2106 .P
2107 This particular example pattern that we have been looking at contains nested
2108 unlimited repeats, and so the use of a possessive quantifier for matching
2109 strings of non-parentheses is important when applying the pattern to strings
2110 that do not match. For example, when this pattern is applied to
2111 .sp
2112 (aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa()
2113 .sp
2114 it yields "no match" quickly. However, if a possessive quantifier is not used,
2115 the match runs for a very long time indeed because there are so many different
2116 ways the + and * repeats can carve up the subject, and all have to be tested
2117 before failure can be reported.
2118 .P
2119 At the end of a match, the values of capturing parentheses are those from
2120 the outermost level. If you want to obtain intermediate values, a callout
2121 function can be used (see below and the
2122 .\" HREF
2123 \fBpcrecallout\fP
2124 .\"
2125 documentation). If the pattern above is matched against
2126 .sp
2127 (ab(cd)ef)
2128 .sp
2129 the value for the inner capturing parentheses (numbered 2) is "ef", which is
2130 the last value taken on at the top level. If a capturing subpattern is not
2131 matched at the top level, its final value is unset, even if it is (temporarily)
2132 set at a deeper level.
2133 .P
2134 If there are more than 15 capturing parentheses in a pattern, PCRE has to
2135 obtain extra memory to store data during a recursion, which it does by using
2136 \fBpcre_malloc\fP, freeing it via \fBpcre_free\fP afterwards. If no memory can
2137 be obtained, the match fails with the PCRE_ERROR_NOMEMORY error.
2138 .P
2139 Do not confuse the (?R) item with the condition (R), which tests for recursion.
2140 Consider this pattern, which matches text in angle brackets, allowing for
2141 arbitrary nesting. Only digits are allowed in nested brackets (that is, when
2142 recursing), whereas any characters are permitted at the outer level.
2143 .sp
2144 < (?: (?(R) \ed++ | [^<>]*+) | (?R)) * >
2145 .sp
2146 In this pattern, (?(R) is the start of a conditional subpattern, with two
2147 different alternatives for the recursive and non-recursive cases. The (?R) item
2148 is the actual recursive call.
2149 .
2150 .
2151 .\" HTML <a name="recursiondifference"></a>
2152 .SS "Recursion difference from Perl"
2153 .rs
2154 .sp
2155 In PCRE (like Python, but unlike Perl), a recursive subpattern call is always
2156 treated as an atomic group. That is, once it has matched some of the subject
2157 string, it is never re-entered, even if it contains untried alternatives and
2158 there is a subsequent matching failure. This can be illustrated by the
2159 following pattern, which purports to match a palindromic string that contains
2160 an odd number of characters (for example, "a", "aba", "abcba", "abcdcba"):
2161 .sp
2162 ^(.|(.)(?1)\e2)$
2163 .sp
2164 The idea is that it either matches a single character, or two identical
2165 characters surrounding a sub-palindrome. In Perl, this pattern works; in PCRE
2166 it does not if the pattern is longer than three characters. Consider the
2167 subject string "abcba":
2168 .P
2169 At the top level, the first character is matched, but as it is not at the end
2170 of the string, the first alternative fails; the second alternative is taken
2171 and the recursion kicks in. The recursive call to subpattern 1 successfully
2172 matches the next character ("b"). (Note that the beginning and end of line
2173 tests are not part of the recursion).
2174 .P
2175 Back at the top level, the next character ("c") is compared with what
2176 subpattern 2 matched, which was "a". This fails. Because the recursion is
2177 treated as an atomic group, there are now no backtracking points, and so the
2178 entire match fails. (Perl is able, at this point, to re-enter the recursion and
2179 try the second alternative.) However, if the pattern is written with the
2180 alternatives in the other order, things are different:
2181 .sp
2182 ^((.)(?1)\e2|.)$
2183 .sp
2184 This time, the recursing alternative is tried first, and continues to recurse
2185 until it runs out of characters, at which point the recursion fails. But this
2186 time we do have another alternative to try at the higher level. That is the big
2187 difference: in the previous case the remaining alternative is at a deeper
2188 recursion level, which PCRE cannot use.
2189 .P
2190 To change the pattern so that matches all palindromic strings, not just those
2191 with an odd number of characters, it is tempting to change the pattern to this:
2192 .sp
2193 ^((.)(?1)\e2|.?)$
2194 .sp
2195 Again, this works in Perl, but not in PCRE, and for the same reason. When a
2196 deeper recursion has matched a single character, it cannot be entered again in
2197 order to match an empty string. The solution is to separate the two cases, and
2198 write out the odd and even cases as alternatives at the higher level:
2199 .sp
2200 ^(?:((.)(?1)\e2|)|((.)(?3)\e4|.))
2201 .sp
2202 If you want to match typical palindromic phrases, the pattern has to ignore all
2203 non-word characters, which can be done like this:
2204 .sp
2205 ^\eW*+(?:((.)\eW*+(?1)\eW*+\e2|)|((.)\eW*+(?3)\eW*+\e4|\eW*+.\eW*+))\eW*+$
2206 .sp
2207 If run with the PCRE_CASELESS option, this pattern matches phrases such as "A
2208 man, a plan, a canal: Panama!" and it works well in both PCRE and Perl. Note
2209 the use of the possessive quantifier *+ to avoid backtracking into sequences of
2210 non-word characters. Without this, PCRE takes a great deal longer (ten times or
2211 more) to match typical phrases, and Perl takes so long that you think it has
2212 gone into a loop.
2213 .P
2214 \fBWARNING\fP: The palindrome-matching patterns above work only if the subject
2215 string does not start with a palindrome that is shorter than the entire string.
2216 For example, although "abcba" is correctly matched, if the subject is "ababa",
2217 PCRE finds the palindrome "aba" at the start, then fails at top level because
2218 the end of the string does not follow. Once again, it cannot jump back into the
2219 recursion to try other alternatives, so the entire match fails.
2220 .
2221 .
2222 .\" HTML <a name="subpatternsassubroutines"></a>
2223 .SH "SUBPATTERNS AS SUBROUTINES"
2224 .rs
2225 .sp
2226 If the syntax for a recursive subpattern reference (either by number or by
2227 name) is used outside the parentheses to which it refers, it operates like a
2228 subroutine in a programming language. The "called" subpattern may be defined
2229 before or after the reference. A numbered reference can be absolute or
2230 relative, as in these examples:
2231 .sp
2232 (...(absolute)...)...(?2)...
2233 (...(relative)...)...(?-1)...
2234 (...(?+1)...(relative)...
2235 .sp
2236 An earlier example pointed out that the pattern
2237 .sp
2238 (sens|respons)e and \e1ibility
2239 .sp
2240 matches "sense and sensibility" and "response and responsibility", but not
2241 "sense and responsibility". If instead the pattern
2242 .sp
2243 (sens|respons)e and (?1)ibility
2244 .sp
2245 is used, it does match "sense and responsibility" as well as the other two
2246 strings. Another example is given in the discussion of DEFINE above.
2247 .P
2248 Like recursive subpatterns, a subroutine call is always treated as an atomic
2249 group. That is, once it has matched some of the subject string, it is never
2250 re-entered, even if it contains untried alternatives and there is a subsequent
2251 matching failure. Any capturing parentheses that are set during the subroutine
2252 call revert to their previous values afterwards.
2253 .P
2254 When a subpattern is used as a subroutine, processing options such as
2255 case-independence are fixed when the subpattern is defined. They cannot be
2256 changed for different calls. For example, consider this pattern:
2257 .sp
2258 (abc)(?i:(?-1))
2259 .sp
2260 It matches "abcabc". It does not match "abcABC" because the change of
2261 processing option does not affect the called subpattern.
2262 .
2263 .
2264 .\" HTML <a name="onigurumasubroutines"></a>
2265 .SH "ONIGURUMA SUBROUTINE SYNTAX"
2266 .rs
2267 .sp
2268 For compatibility with Oniguruma, the non-Perl syntax \eg followed by a name or
2269 a number enclosed either in angle brackets or single quotes, is an alternative
2270 syntax for referencing a subpattern as a subroutine, possibly recursively. Here
2271 are two of the examples used above, rewritten using this syntax:
2272 .sp
2273 (?<pn> \e( ( (?>[^()]+) | \eg<pn> )* \e) )
2274 (sens|respons)e and \eg'1'ibility
2275 .sp
2276 PCRE supports an extension to Oniguruma: if a number is preceded by a
2277 plus or a minus sign it is taken as a relative reference. For example:
2278 .sp
2279 (abc)(?i:\eg<-1>)
2280 .sp
2281 Note that \eg{...} (Perl syntax) and \eg<...> (Oniguruma syntax) are \fInot\fP
2282 synonymous. The former is a back reference; the latter is a subroutine call.
2283 .
2284 .
2285 .SH CALLOUTS
2286 .rs
2287 .sp
2288 Perl has a feature whereby using the sequence (?{...}) causes arbitrary Perl
2289 code to be obeyed in the middle of matching a regular expression. This makes it
2290 possible, amongst other things, to extract different substrings that match the
2291 same pair of parentheses when there is a repetition.
2292 .P
2293 PCRE provides a similar feature, but of course it cannot obey arbitrary Perl
2294 code. The feature is called "callout". The caller of PCRE provides an external
2295 function by putting its entry point in the global variable \fIpcre_callout\fP.
2296 By default, this variable contains NULL, which disables all calling out.
2297 .P
2298 Within a regular expression, (?C) indicates the points at which the external
2299 function is to be called. If you want to identify different callout points, you
2300 can put a number less than 256 after the letter C. The default value is zero.
2301 For example, this pattern has two callout points:
2302 .sp
2303 (?C1)abc(?C2)def
2304 .sp
2305 If the PCRE_AUTO_CALLOUT flag is passed to \fBpcre_compile()\fP, callouts are
2306 automatically installed before each item in the pattern. They are all numbered
2307 255.
2308 .P
2309 During matching, when PCRE reaches a callout point (and \fIpcre_callout\fP is
2310 set), the external function is called. It is provided with the number of the
2311 callout, the position in the pattern, and, optionally, one item of data
2312 originally supplied by the caller of \fBpcre_exec()\fP. The callout function
2313 may cause matching to proceed, to backtrack, or to fail altogether. A complete
2314 description of the interface to the callout function is given in the
2315 .\" HREF
2316 \fBpcrecallout\fP
2317 .\"
2318 documentation.
2319 .
2320 .
2321 .\" HTML <a name="backtrackcontrol"></a>
2322 .SH "BACKTRACKING CONTROL"
2323 .rs
2324 .sp
2325 Perl 5.10 introduced a number of "Special Backtracking Control Verbs", which
2326 are described in the Perl documentation as "experimental and subject to change
2327 or removal in a future version of Perl". It goes on to say: "Their usage in
2328 production code should be noted to avoid problems during upgrades." The same
2329 remarks apply to the PCRE features described in this section.
2330 .P
2331 Since these verbs are specifically related to backtracking, most of them can be
2332 used only when the pattern is to be matched using \fBpcre_exec()\fP, which uses
2333 a backtracking algorithm. With the exception of (*FAIL), which behaves like a
2334 failing negative assertion, they cause an error if encountered by
2335 \fBpcre_dfa_exec()\fP.
2336 .P
2337 If any of these verbs are used in an assertion or subroutine subpattern
2338 (including recursive subpatterns), their effect is confined to that subpattern;
2339 it does not extend to the surrounding pattern. Note that such subpatterns are
2340 processed as anchored at the point where they are tested.
2341 .P
2342 The new verbs make use of what was previously invalid syntax: an opening
2343 parenthesis followed by an asterisk. They are generally of the form
2344 (*VERB) or (*VERB:NAME). Some may take either form, with differing behaviour,
2345 depending on whether or not an argument is present. An name is a sequence of
2346 letters, digits, and underscores. If the name is empty, that is, if the closing
2347 parenthesis immediately follows the colon, the effect is as if the colon were
2348 not there. Any number of these verbs may occur in a pattern.
2349 .P
2350 PCRE contains some optimizations that are used to speed up matching by running
2351 some checks at the start of each match attempt. For example, it may know the
2352 minimum length of matching subject, or that a particular character must be
2353 present. When one of these optimizations suppresses the running of a match, any
2354 included backtracking verbs will not, of course, be processed. You can suppress
2355 the start-of-match optimizations by setting the PCRE_NO_START_OPTIMIZE option
2356 when calling \fBpcre_exec()\fP.
2357 .
2358 .
2359 .SS "Verbs that act immediately"
2360 .rs
2361 .sp
2362 The following verbs act as soon as they are encountered. They may not be
2363 followed by a name.
2364 .sp
2365 (*ACCEPT)
2366 .sp
2367 This verb causes the match to end successfully, skipping the remainder of the
2368 pattern. When inside a recursion, only the innermost pattern is ended
2369 immediately. If (*ACCEPT) is inside capturing parentheses, the data so far is
2370 captured. (This feature was added to PCRE at release 8.00.) For example:
2371 .sp
2372 A((?:A|B(*ACCEPT)|C)D)
2373 .sp
2374 This matches "AB", "AAD", or "ACD"; when it matches "AB", "B" is captured by
2375 the outer parentheses.
2376 .sp
2377 (*FAIL) or (*F)
2378 .sp
2379 This verb causes the match to fail, forcing backtracking to occur. It is
2380 equivalent to (?!) but easier to read. The Perl documentation notes that it is
2381 probably useful only when combined with (?{}) or (??{}). Those are, of course,
2382 Perl features that are not present in PCRE. The nearest equivalent is the
2383 callout feature, as for example in this pattern:
2384 .sp
2385 a+(?C)(*FAIL)
2386 .sp
2387 A match with the string "aaaa" always fails, but the callout is taken before
2388 each backtrack happens (in this example, 10 times).
2389 .
2390 .
2391 .SS "Recording which path was taken"
2392 .rs
2393 .sp
2394 There is one verb whose main purpose is to track how a match was arrived at,
2395 though it also has a secondary use in conjunction with advancing the match
2396 starting point (see (*SKIP) below).
2397 .sp
2398 (*MARK:NAME) or (*:NAME)
2399 .sp
2400 A name is always required with this verb. There may be as many instances of
2401 (*MARK) as you like in a pattern, and their names do not have to be unique.
2402 .P
2403 When a match succeeds, the name of the last-encountered (*MARK) is passed back
2404 to the caller via the \fIpcre_extra\fP data structure, as described in the
2405 .\" HTML <a href="pcreapi.html#extradata">
2406 .\" </a>
2407 section on \fIpcre_extra\fP
2408 .\"
2409 in the
2410 .\" HREF
2411 \fBpcreapi\fP
2412 .\"
2413 documentation. No data is returned for a partial match. Here is an example of
2414 \fBpcretest\fP output, where the /K modifier requests the retrieval and
2415 outputting of (*MARK) data:
2416 .sp
2417 /X(*MARK:A)Y|X(*MARK:B)Z/K
2418 XY
2419 0: XY
2420 MK: A
2421 XZ
2422 0: XZ
2423 MK: B
2424 .sp
2425 The (*MARK) name is tagged with "MK:" in this output, and in this example it
2426 indicates which of the two alternatives matched. This is a more efficient way
2427 of obtaining this information than putting each alternative in its own
2428 capturing parentheses.
2429 .P
2430 A name may also be returned after a failed match if the final path through the
2431 pattern involves (*MARK). However, unless (*MARK) used in conjunction with
2432 (*COMMIT), this is unlikely to happen for an unanchored pattern because, as the
2433 starting point for matching is advanced, the final check is often with an empty
2434 string, causing a failure before (*MARK) is reached. For example:
2435 .sp
2436 /X(*MARK:A)Y|X(*MARK:B)Z/K
2437 XP
2438 No match
2439 .sp
2440 There are three potential starting points for this match (starting with X,
2441 starting with P, and with an empty string). If the pattern is anchored, the
2442 result is different:
2443 .sp
2444 /^X(*MARK:A)Y|^X(*MARK:B)Z/K
2445 XP
2446 No match, mark = B
2447 .sp
2448 PCRE's start-of-match optimizations can also interfere with this. For example,
2449 if, as a result of a call to \fBpcre_study()\fP, it knows the minimum
2450 subject length for a match, a shorter subject will not be scanned at all.
2451 .P
2452 Note that similar anomalies (though different in detail) exist in Perl, no
2453 doubt for the same reasons. The use of (*MARK) data after a failed match of an
2454 unanchored pattern is not recommended, unless (*COMMIT) is involved.
2455 .
2456 .
2457 .SS "Verbs that act after backtracking"
2458 .rs
2459 .sp
2460 The following verbs do nothing when they are encountered. Matching continues
2461 with what follows, but if there is no subsequent match, causing a backtrack to
2462 the verb, a failure is forced. That is, backtracking cannot pass to the left of
2463 the verb. However, when one of these verbs appears inside an atomic group, its
2464 effect is confined to that group, because once the group has been matched,
2465 there is never any backtracking into it. In this situation, backtracking can
2466 "jump back" to the left of the entire atomic group. (Remember also, as stated
2467 above, that this localization also applies in subroutine calls and assertions.)
2468 .P
2469 These verbs differ in exactly what kind of failure occurs when backtracking
2470 reaches them.
2471 .sp
2472 (*COMMIT)
2473 .sp
2474 This verb, which may not be followed by a name, causes the whole match to fail
2475 outright if the rest of the pattern does not match. Even if the pattern is
2476 unanchored, no further attempts to find a match by advancing the starting point
2477 take place. Once (*COMMIT) has been passed, \fBpcre_exec()\fP is committed to
2478 finding a match at the current starting point, or not at all. For example:
2479 .sp
2480 a+(*COMMIT)b
2481 .sp
2482 This matches "xxaab" but not "aacaab". It can be thought of as a kind of
2483 dynamic anchor, or "I've started, so I must finish." The name of the most
2484 recently passed (*MARK) in the path is passed back when (*COMMIT) forces a
2485 match failure.
2486 .P
2487 Note that (*COMMIT) at the start of a pattern is not the same as an anchor,
2488 unless PCRE's start-of-match optimizations are turned off, as shown in this
2489 \fBpcretest\fP example:
2490 .sp
2491 /(*COMMIT)abc/
2492 xyzabc
2493 0: abc
2494 xyzabc\eY
2495 No match
2496 .sp
2497 PCRE knows that any match must start with "a", so the optimization skips along
2498 the subject to "a" before running the first match attempt, which succeeds. When
2499 the optimization is disabled by the \eY escape in the second subject, the match
2500 starts at "x" and so the (*COMMIT) causes it to fail without trying any other
2501 starting points.
2502 .sp
2503 (*PRUNE) or (*PRUNE:NAME)
2504 .sp
2505 This verb causes the match to fail at the current starting position in the
2506 subject if the rest of the pattern does not match. If the pattern is
2507 unanchored, the normal "bumpalong" advance to the next starting character then
2508 happens. Backtracking can occur as usual to the left of (*PRUNE), before it is
2509 reached, or when matching to the right of (*PRUNE), but if there is no match to
2510 the right, backtracking cannot cross (*PRUNE). In simple cases, the use of
2511 (*PRUNE) is just an alternative to an atomic group or possessive quantifier,
2512 but there are some uses of (*PRUNE) that cannot be expressed in any other way.
2513 The behaviour of (*PRUNE:NAME) is the same as (*MARK:NAME)(*PRUNE) when the
2514 match fails completely; the name is passed back if this is the final attempt.
2515 (*PRUNE:NAME) does not pass back a name if the match succeeds. In an anchored
2516 pattern (*PRUNE) has the same effect as (*COMMIT).
2517 .sp
2518 (*SKIP)
2519 .sp
2520 This verb, when given without a name, is like (*PRUNE), except that if the
2521 pattern is unanchored, the "bumpalong" advance is not to the next character,
2522 but to the position in the subject where (*SKIP) was encountered. (*SKIP)
2523 signifies that whatever text was matched leading up to it cannot be part of a
2524 successful match. Consider:
2525 .sp
2526 a+(*SKIP)b
2527 .sp
2528 If the subject is "aaaac...", after the first match attempt fails (starting at
2529 the first character in the string), the starting point skips on to start the
2530 next attempt at "c". Note that a possessive quantifer does not have the same
2531 effect as this example; although it would suppress backtracking during the
2532 first match attempt, the second attempt would start at the second character
2533 instead of skipping on to "c".
2534 .sp
2535 (*SKIP:NAME)
2536 .sp
2537 When (*SKIP) has an associated name, its behaviour is modified. If the
2538 following pattern fails to match, the previous path through the pattern is
2539 searched for the most recent (*MARK) that has the same name. If one is found,
2540 the "bumpalong" advance is to the subject position that corresponds to that
2541 (*MARK) instead of to where (*SKIP) was encountered. If no (*MARK) with a
2542 matching name is found, normal "bumpalong" of one character happens (the
2543 (*SKIP) is ignored).
2544 .sp
2545 (*THEN) or (*THEN:NAME)
2546 .sp
2547 This verb causes a skip to the next alternation if the rest of the pattern does
2548 not match. That is, it cancels pending backtracking, but only within the
2549 current alternation. Its name comes from the observation that it can be used
2550 for a pattern-based if-then-else block:
2551 .sp
2552 ( COND1 (*THEN) FOO | COND2 (*THEN) BAR | COND3 (*THEN) BAZ ) ...
2553 .sp
2554 If the COND1 pattern matches, FOO is tried (and possibly further items after
2555 the end of the group if FOO succeeds); on failure the matcher skips to the
2556 second alternative and tries COND2, without backtracking into COND1. The
2557 behaviour of (*THEN:NAME) is exactly the same as (*MARK:NAME)(*THEN) if the
2558 overall match fails. If (*THEN) is not directly inside an alternation, it acts
2559 like (*PRUNE).
2560 .
2561 .
2562 .SH "SEE ALSO"
2563 .rs
2564 .sp
2565 \fBpcreapi\fP(3), \fBpcrecallout\fP(3), \fBpcrematching\fP(3),
2566 \fBpcresyntax\fP(3), \fBpcre\fP(3).
2567 .
2568 .
2569 .SH AUTHOR
2570 .rs
2571 .sp
2572 .nf
2573 Philip Hazel
2574 University Computing Service
2575 Cambridge CB2 3QH, England.
2576 .fi
2577 .
2578 .
2579 .SH REVISION
2580 .rs
2581 .sp
2582 .nf
2583 Last updated: 27 March 2010
2584 Copyright (c) 1997-2010 University of Cambridge.
2585 .fi

Properties

Name Value
svn:eol-style native
svn:keywords "Author Date Id Revision Url"

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12