ViewVC logotype

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

Parent Directory Parent Directory | Revision Log Revision Log

Revision 385 - (show annotations)
Sun Mar 8 16:56:58 2009 UTC (6 years, 6 months ago) by ph10
File size: 91571 byte(s)
Error occurred while calculating annotation data.
Document named subpattern behaviour in the presence of duplicate numbers.
3 PCRE - Perl-compatible regular expressions
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, you must
25 build PCRE to include UTF-8 support, and then call \fBpcre_compile()\fP with
26 the PCRE_UTF8 option. How this affects pattern matching is mentioned in several
27 places below. There is also a summary of UTF-8 features in the
28 .\" HTML <a href="pcre.html#utf8support">
29 .\" </a>
30 section on UTF-8 support
31 .\"
32 in the main
33 .\" HREF
34 \fBpcre\fP
35 .\"
36 page.
37 .P
38 The remainder of this document discusses the patterns that are supported by
39 PCRE when its main matching function, \fBpcre_exec()\fP, is used.
40 From release 6.0, PCRE offers a second matching function,
41 \fBpcre_dfa_exec()\fP, which matches using a different algorithm that is not
42 Perl-compatible. Some of the features discussed below are not available when
43 \fBpcre_dfa_exec()\fP is used. The advantages and disadvantages of the
44 alternative function, and how it differs from the normal function, are
45 discussed in the
46 .\" HREF
47 \fBpcrematching\fP
48 .\"
49 page.
50 .
51 .
53 .rs
54 .sp
55 PCRE supports five different conventions for indicating line breaks in
56 strings: a single CR (carriage return) character, a single LF (linefeed)
57 character, the two-character sequence CRLF, any of the three preceding, or any
58 Unicode newline sequence. The
59 .\" HREF
60 \fBpcreapi\fP
61 .\"
62 page has
63 .\" HTML <a href="pcreapi.html#newlines">
64 .\" </a>
65 further discussion
66 .\"
67 about newlines, and shows how to set the newline convention in the
68 \fIoptions\fP arguments for the compiling and matching functions.
69 .P
70 It is also possible to specify a newline convention by starting a pattern
71 string with one of the following five sequences:
72 .sp
73 (*CR) carriage return
74 (*LF) linefeed
75 (*CRLF) carriage return, followed by linefeed
76 (*ANYCRLF) any of the three above
77 (*ANY) all Unicode newline sequences
78 .sp
79 These override the default and the options given to \fBpcre_compile()\fP. For
80 example, on a Unix system where LF is the default newline sequence, the pattern
81 .sp
82 (*CR)a.b
83 .sp
84 changes the convention to CR. That pattern matches "a\enb" because LF is no
85 longer a newline. Note that these special settings, which are not
86 Perl-compatible, are recognized only at the very start of a pattern, and that
87 they must be in upper case. If more than one of them is present, the last one
88 is used.
89 .P
90 The newline convention does not affect what the \eR escape sequence matches. By
91 default, this is any Unicode newline sequence, for Perl compatibility. However,
92 this can be changed; see the description of \eR in the section entitled
93 .\" HTML <a href="#newlineseq">
94 .\" </a>
95 "Newline sequences"
96 .\"
97 below. A change of \eR setting can be combined with a change of newline
98 convention.
99 .
100 .
102 .rs
103 .sp
104 A regular expression is a pattern that is matched against a subject string from
105 left to right. Most characters stand for themselves in a pattern, and match the
106 corresponding characters in the subject. As a trivial example, the pattern
107 .sp
108 The quick brown fox
109 .sp
110 matches a portion of a subject string that is identical to itself. When
111 caseless matching is specified (the PCRE_CASELESS option), letters are matched
112 independently of case. In UTF-8 mode, PCRE always understands the concept of
113 case for characters whose values are less than 128, so caseless matching is
114 always possible. For characters with higher values, the concept of case is
115 supported if PCRE is compiled with Unicode property support, but not otherwise.
116 If you want to use caseless matching for characters 128 and above, you must
117 ensure that PCRE is compiled with Unicode property support as well as with
118 UTF-8 support.
119 .P
120 The power of regular expressions comes from the ability to include alternatives
121 and repetitions in the pattern. These are encoded in the pattern by the use of
122 \fImetacharacters\fP, which do not stand for themselves but instead are
123 interpreted in some special way.
124 .P
125 There are two different sets of metacharacters: those that are recognized
126 anywhere in the pattern except within square brackets, and those that are
127 recognized within square brackets. Outside square brackets, the metacharacters
128 are as follows:
129 .sp
130 \e general escape character with several uses
131 ^ assert start of string (or line, in multiline mode)
132 $ assert end of string (or line, in multiline mode)
133 . match any character except newline (by default)
134 [ start character class definition
135 | start of alternative branch
136 ( start subpattern
137 ) end subpattern
138 ? extends the meaning of (
139 also 0 or 1 quantifier
140 also quantifier minimizer
141 * 0 or more quantifier
142 + 1 or more quantifier
143 also "possessive quantifier"
144 { start min/max quantifier
145 .sp
146 Part of a pattern that is in square brackets is called a "character class". In
147 a character class the only metacharacters are:
148 .sp
149 \e general escape character
150 ^ negate the class, but only if the first character
151 - indicates character range
152 .\" JOIN
153 [ POSIX character class (only if followed by POSIX
154 syntax)
155 ] terminates the character class
156 .sp
157 The following sections describe the use of each of the metacharacters.
158 .
159 .
161 .rs
162 .sp
163 The backslash character has several uses. Firstly, if it is followed by a
164 non-alphanumeric character, it takes away any special meaning that character
165 may have. This use of backslash as an escape character applies both inside and
166 outside character classes.
167 .P
168 For example, if you want to match a * character, you write \e* in the pattern.
169 This escaping action applies whether or not the following character would
170 otherwise be interpreted as a metacharacter, so it is always safe to precede a
171 non-alphanumeric with backslash to specify that it stands for itself. In
172 particular, if you want to match a backslash, you write \e\e.
173 .P
174 If a pattern is compiled with the PCRE_EXTENDED option, whitespace in the
175 pattern (other than in a character class) and characters between a # outside
176 a character class and the next newline are ignored. An escaping backslash can
177 be used to include a whitespace or # character as part of the pattern.
178 .P
179 If you want to remove the special meaning from a sequence of characters, you
180 can do so by putting them between \eQ and \eE. This is different from Perl in
181 that $ and @ are handled as literals in \eQ...\eE sequences in PCRE, whereas in
182 Perl, $ and @ cause variable interpolation. Note the following examples:
183 .sp
184 Pattern PCRE matches Perl matches
185 .sp
186 .\" JOIN
187 \eQabc$xyz\eE abc$xyz abc followed by the
188 contents of $xyz
189 \eQabc\e$xyz\eE abc\e$xyz abc\e$xyz
190 \eQabc\eE\e$\eQxyz\eE abc$xyz abc$xyz
191 .sp
192 The \eQ...\eE sequence is recognized both inside and outside character classes.
193 .
194 .
195 .\" HTML <a name="digitsafterbackslash"></a>
196 .SS "Non-printing characters"
197 .rs
198 .sp
199 A second use of backslash provides a way of encoding non-printing characters
200 in patterns in a visible manner. There is no restriction on the appearance of
201 non-printing characters, apart from the binary zero that terminates a pattern,
202 but when a pattern is being prepared by text editing, it is usually easier to
203 use one of the following escape sequences than the binary character it
204 represents:
205 .sp
206 \ea alarm, that is, the BEL character (hex 07)
207 \ecx "control-x", where x is any character
208 \ee escape (hex 1B)
209 \ef formfeed (hex 0C)
210 \en linefeed (hex 0A)
211 \er carriage return (hex 0D)
212 \et tab (hex 09)
213 \eddd character with octal code ddd, or backreference
214 \exhh character with hex code hh
215 \ex{hhh..} character with hex code hhh..
216 .sp
217 The precise effect of \ecx is as follows: if x is a lower case letter, it
218 is converted to upper case. Then bit 6 of the character (hex 40) is inverted.
219 Thus \ecz becomes hex 1A, but \ec{ becomes hex 3B, while \ec; becomes hex
220 7B.
221 .P
222 After \ex, from zero to two hexadecimal digits are read (letters can be in
223 upper or lower case). Any number of hexadecimal digits may appear between \ex{
224 and }, but the value of the character code must be less than 256 in non-UTF-8
225 mode, and less than 2**31 in UTF-8 mode. That is, the maximum value in
226 hexadecimal is 7FFFFFFF. Note that this is bigger than the largest Unicode code
227 point, which is 10FFFF.
228 .P
229 If characters other than hexadecimal digits appear between \ex{ and }, or if
230 there is no terminating }, this form of escape is not recognized. Instead, the
231 initial \ex will be interpreted as a basic hexadecimal escape, with no
232 following digits, giving a character whose value is zero.
233 .P
234 Characters whose value is less than 256 can be defined by either of the two
235 syntaxes for \ex. There is no difference in the way they are handled. For
236 example, \exdc is exactly the same as \ex{dc}.
237 .P
238 After \e0 up to two further octal digits are read. If there are fewer than two
239 digits, just those that are present are used. Thus the sequence \e0\ex\e07
240 specifies two binary zeros followed by a BEL character (code value 7). Make
241 sure you supply two digits after the initial zero if the pattern character that
242 follows is itself an octal digit.
243 .P
244 The handling of a backslash followed by a digit other than 0 is complicated.
245 Outside a character class, PCRE reads it and any following digits as a decimal
246 number. If the number is less than 10, or if there have been at least that many
247 previous capturing left parentheses in the expression, the entire sequence is
248 taken as a \fIback reference\fP. A description of how this works is given
249 .\" HTML <a href="#backreferences">
250 .\" </a>
251 later,
252 .\"
253 following the discussion of
254 .\" HTML <a href="#subpattern">
255 .\" </a>
256 parenthesized subpatterns.
257 .\"
258 .P
259 Inside a character class, or if the decimal number is greater than 9 and there
260 have not been that many capturing subpatterns, PCRE re-reads up to three octal
261 digits following the backslash, and uses them to generate a data character. Any
262 subsequent digits stand for themselves. In non-UTF-8 mode, the value of a
263 character specified in octal must be less than \e400. In UTF-8 mode, values up
264 to \e777 are permitted. For example:
265 .sp
266 \e040 is another way of writing a space
267 .\" JOIN
268 \e40 is the same, provided there are fewer than 40
269 previous capturing subpatterns
270 \e7 is always a back reference
271 .\" JOIN
272 \e11 might be a back reference, or another way of
273 writing a tab
274 \e011 is always a tab
275 \e0113 is a tab followed by the character "3"
276 .\" JOIN
277 \e113 might be a back reference, otherwise the
278 character with octal code 113
279 .\" JOIN
280 \e377 might be a back reference, otherwise
281 the byte consisting entirely of 1 bits
282 .\" JOIN
283 \e81 is either a back reference, or a binary zero
284 followed by the two characters "8" and "1"
285 .sp
286 Note that octal values of 100 or greater must not be introduced by a leading
287 zero, because no more than three octal digits are ever read.
288 .P
289 All the sequences that define a single character value can be used both inside
290 and outside character classes. In addition, inside a character class, the
291 sequence \eb is interpreted as the backspace character (hex 08), and the
292 sequences \eR and \eX are interpreted as the characters "R" and "X",
293 respectively. Outside a character class, these sequences have different
294 meanings
295 .\" HTML <a href="#uniextseq">
296 .\" </a>
297 (see below).
298 .\"
299 .
300 .
301 .SS "Absolute and relative back references"
302 .rs
303 .sp
304 The sequence \eg followed by an unsigned or a negative number, optionally
305 enclosed in braces, is an absolute or relative back reference. A named back
306 reference can be coded as \eg{name}. Back references are discussed
307 .\" HTML <a href="#backreferences">
308 .\" </a>
309 later,
310 .\"
311 following the discussion of
312 .\" HTML <a href="#subpattern">
313 .\" </a>
314 parenthesized subpatterns.
315 .\"
316 .
317 .
318 .SS "Absolute and relative subroutine calls"
319 .rs
320 .sp
321 For compatibility with Oniguruma, the non-Perl syntax \eg followed by a name or
322 a number enclosed either in angle brackets or single quotes, is an alternative
323 syntax for referencing a subpattern as a "subroutine". Details are discussed
324 .\" HTML <a href="#onigurumasubroutines">
325 .\" </a>
326 later.
327 .\"
328 Note that \eg{...} (Perl syntax) and \eg<...> (Oniguruma syntax) are \fInot\fP
329 synonymous. The former is a back reference; the latter is a subroutine call.
330 .
331 .
332 .SS "Generic character types"
333 .rs
334 .sp
335 Another use of backslash is for specifying generic character types. The
336 following are always recognized:
337 .sp
338 \ed any decimal digit
339 \eD any character that is not a decimal digit
340 \eh any horizontal whitespace character
341 \eH any character that is not a horizontal whitespace character
342 \es any whitespace character
343 \eS any character that is not a whitespace character
344 \ev any vertical whitespace character
345 \eV any character that is not a vertical whitespace character
346 \ew any "word" character
347 \eW any "non-word" character
348 .sp
349 Each pair of escape sequences partitions the complete set of characters into
350 two disjoint sets. Any given character matches one, and only one, of each pair.
351 .P
352 These character type sequences can appear both inside and outside character
353 classes. They each match one character of the appropriate type. If the current
354 matching point is at the end of the subject string, all of them fail, since
355 there is no character to match.
356 .P
357 For compatibility with Perl, \es does not match the VT character (code 11).
358 This makes it different from the the POSIX "space" class. The \es characters
359 are HT (9), LF (10), FF (12), CR (13), and space (32). If "use locale;" is
360 included in a Perl script, \es may match the VT character. In PCRE, it never
361 does.
362 .P
363 In UTF-8 mode, characters with values greater than 128 never match \ed, \es, or
364 \ew, and always match \eD, \eS, and \eW. This is true even when Unicode
365 character property support is available. These sequences retain their original
366 meanings from before UTF-8 support was available, mainly for efficiency
367 reasons.
368 .P
369 The sequences \eh, \eH, \ev, and \eV are Perl 5.10 features. In contrast to the
370 other sequences, these do match certain high-valued codepoints in UTF-8 mode.
371 The horizontal space characters are:
372 .sp
373 U+0009 Horizontal tab
374 U+0020 Space
375 U+00A0 Non-break space
376 U+1680 Ogham space mark
377 U+180E Mongolian vowel separator
378 U+2000 En quad
379 U+2001 Em quad
380 U+2002 En space
381 U+2003 Em space
382 U+2004 Three-per-em space
383 U+2005 Four-per-em space
384 U+2006 Six-per-em space
385 U+2007 Figure space
386 U+2008 Punctuation space
387 U+2009 Thin space
388 U+200A Hair space
389 U+202F Narrow no-break space
390 U+205F Medium mathematical space
391 U+3000 Ideographic space
392 .sp
393 The vertical space characters are:
394 .sp
395 U+000A Linefeed
396 U+000B Vertical tab
397 U+000C Formfeed
398 U+000D Carriage return
399 U+0085 Next line
400 U+2028 Line separator
401 U+2029 Paragraph separator
402 .P
403 A "word" character is an underscore or any character less than 256 that is a
404 letter or digit. The definition of letters and digits is controlled by PCRE's
405 low-valued character tables, and may vary if locale-specific matching is taking
406 place (see
407 .\" HTML <a href="pcreapi.html#localesupport">
408 .\" </a>
409 "Locale support"
410 .\"
411 in the
412 .\" HREF
413 \fBpcreapi\fP
414 .\"
415 page). For example, in a French locale such as "fr_FR" in Unix-like systems,
416 or "french" in Windows, some character codes greater than 128 are used for
417 accented letters, and these are matched by \ew. The use of locales with Unicode
418 is discouraged.
419 .
420 .
421 .\" HTML <a name="newlineseq"></a>
422 .SS "Newline sequences"
423 .rs
424 .sp
425 Outside a character class, by default, the escape sequence \eR matches any
426 Unicode newline sequence. This is a Perl 5.10 feature. In non-UTF-8 mode \eR is
427 equivalent to the following:
428 .sp
429 (?>\er\en|\en|\ex0b|\ef|\er|\ex85)
430 .sp
431 This is an example of an "atomic group", details of which are given
432 .\" HTML <a href="#atomicgroup">
433 .\" </a>
434 below.
435 .\"
436 This particular group matches either the two-character sequence CR followed by
437 LF, or one of the single characters LF (linefeed, U+000A), VT (vertical tab,
438 U+000B), FF (formfeed, U+000C), CR (carriage return, U+000D), or NEL (next
439 line, U+0085). The two-character sequence is treated as a single unit that
440 cannot be split.
441 .P
442 In UTF-8 mode, two additional characters whose codepoints are greater than 255
443 are added: LS (line separator, U+2028) and PS (paragraph separator, U+2029).
444 Unicode character property support is not needed for these characters to be
445 recognized.
446 .P
447 It is possible to restrict \eR to match only CR, LF, or CRLF (instead of the
448 complete set of Unicode line endings) by setting the option PCRE_BSR_ANYCRLF
449 either at compile time or when the pattern is matched. (BSR is an abbrevation
450 for "backslash R".) This can be made the default when PCRE is built; if this is
451 the case, the other behaviour can be requested via the PCRE_BSR_UNICODE option.
452 It is also possible to specify these settings by starting a pattern string with
453 one of the following sequences:
454 .sp
455 (*BSR_ANYCRLF) CR, LF, or CRLF only
456 (*BSR_UNICODE) any Unicode newline sequence
457 .sp
458 These override the default and the options given to \fBpcre_compile()\fP, but
459 they can be overridden by options given to \fBpcre_exec()\fP. Note that these
460 special settings, which are not Perl-compatible, are recognized only at the
461 very start of a pattern, and that they must be in upper case. If more than one
462 of them is present, the last one is used. They can be combined with a change of
463 newline convention, for example, a pattern can start with:
464 .sp
466 .sp
467 Inside a character class, \eR matches the letter "R".
468 .
469 .
470 .\" HTML <a name="uniextseq"></a>
471 .SS Unicode character properties
472 .rs
473 .sp
474 When PCRE is built with Unicode character property support, three additional
475 escape sequences that match characters with specific properties are available.
476 When not in UTF-8 mode, these sequences are of course limited to testing
477 characters whose codepoints are less than 256, but they do work in this mode.
478 The extra escape sequences are:
479 .sp
480 \ep{\fIxx\fP} a character with the \fIxx\fP property
481 \eP{\fIxx\fP} a character without the \fIxx\fP property
482 \eX an extended Unicode sequence
483 .sp
484 The property names represented by \fIxx\fP above are limited to the Unicode
485 script names, the general category properties, and "Any", which matches any
486 character (including newline). Other properties such as "InMusicalSymbols" are
487 not currently supported by PCRE. Note that \eP{Any} does not match any
488 characters, so always causes a match failure.
489 .P
490 Sets of Unicode characters are defined as belonging to certain scripts. A
491 character from one of these sets can be matched using a script name. For
492 example:
493 .sp
494 \ep{Greek}
495 \eP{Han}
496 .sp
497 Those that are not part of an identified script are lumped together as
498 "Common". The current list of scripts is:
499 .P
500 Arabic,
501 Armenian,
502 Balinese,
503 Bengali,
504 Bopomofo,
505 Braille,
506 Buginese,
507 Buhid,
508 Canadian_Aboriginal,
509 Cherokee,
510 Common,
511 Coptic,
512 Cuneiform,
513 Cypriot,
514 Cyrillic,
515 Deseret,
516 Devanagari,
517 Ethiopic,
518 Georgian,
519 Glagolitic,
520 Gothic,
521 Greek,
522 Gujarati,
523 Gurmukhi,
524 Han,
525 Hangul,
526 Hanunoo,
527 Hebrew,
528 Hiragana,
529 Inherited,
530 Kannada,
531 Katakana,
532 Kharoshthi,
533 Khmer,
534 Lao,
535 Latin,
536 Limbu,
537 Linear_B,
538 Malayalam,
539 Mongolian,
540 Myanmar,
541 New_Tai_Lue,
542 Nko,
543 Ogham,
544 Old_Italic,
545 Old_Persian,
546 Oriya,
547 Osmanya,
548 Phags_Pa,
549 Phoenician,
550 Runic,
551 Shavian,
552 Sinhala,
553 Syloti_Nagri,
554 Syriac,
555 Tagalog,
556 Tagbanwa,
557 Tai_Le,
558 Tamil,
559 Telugu,
560 Thaana,
561 Thai,
562 Tibetan,
563 Tifinagh,
564 Ugaritic,
565 Yi.
566 .P
567 Each character has exactly one general category property, specified by a
568 two-letter abbreviation. For compatibility with Perl, negation can be specified
569 by including a circumflex between the opening brace and the property name. For
570 example, \ep{^Lu} is the same as \eP{Lu}.
571 .P
572 If only one letter is specified with \ep or \eP, it includes all the general
573 category properties that start with that letter. In this case, in the absence
574 of negation, the curly brackets in the escape sequence are optional; these two
575 examples have the same effect:
576 .sp
577 \ep{L}
578 \epL
579 .sp
580 The following general category property codes are supported:
581 .sp
582 C Other
583 Cc Control
584 Cf Format
585 Cn Unassigned
586 Co Private use
587 Cs Surrogate
588 .sp
589 L Letter
590 Ll Lower case letter
591 Lm Modifier letter
592 Lo Other letter
593 Lt Title case letter
594 Lu Upper case letter
595 .sp
596 M Mark
597 Mc Spacing mark
598 Me Enclosing mark
599 Mn Non-spacing mark
600 .sp
601 N Number
602 Nd Decimal number
603 Nl Letter number
604 No Other number
605 .sp
606 P Punctuation
607 Pc Connector punctuation
608 Pd Dash punctuation
609 Pe Close punctuation
610 Pf Final punctuation
611 Pi Initial punctuation
612 Po Other punctuation
613 Ps Open punctuation
614 .sp
615 S Symbol
616 Sc Currency symbol
617 Sk Modifier symbol
618 Sm Mathematical symbol
619 So Other symbol
620 .sp
621 Z Separator
622 Zl Line separator
623 Zp Paragraph separator
624 Zs Space separator
625 .sp
626 The special property L& is also supported: it matches a character that has
627 the Lu, Ll, or Lt property, in other words, a letter that is not classified as
628 a modifier or "other".
629 .P
630 The Cs (Surrogate) property applies only to characters in the range U+D800 to
631 U+DFFF. Such characters are not valid in UTF-8 strings (see RFC 3629) and so
632 cannot be tested by PCRE, unless UTF-8 validity checking has been turned off
633 (see the discussion of PCRE_NO_UTF8_CHECK in the
634 .\" HREF
635 \fBpcreapi\fP
636 .\"
637 page).
638 .P
639 The long synonyms for these properties that Perl supports (such as \ep{Letter})
640 are not supported by PCRE, nor is it permitted to prefix any of these
641 properties with "Is".
642 .P
643 No character that is in the Unicode table has the Cn (unassigned) property.
644 Instead, this property is assumed for any code point that is not in the
645 Unicode table.
646 .P
647 Specifying caseless matching does not affect these escape sequences. For
648 example, \ep{Lu} always matches only upper case letters.
649 .P
650 The \eX escape matches any number of Unicode characters that form an extended
651 Unicode sequence. \eX is equivalent to
652 .sp
653 (?>\ePM\epM*)
654 .sp
655 That is, it matches a character without the "mark" property, followed by zero
656 or more characters with the "mark" property, and treats the sequence as an
657 atomic group
658 .\" HTML <a href="#atomicgroup">
659 .\" </a>
660 (see below).
661 .\"
662 Characters with the "mark" property are typically accents that affect the
663 preceding character. None of them have codepoints less than 256, so in
664 non-UTF-8 mode \eX matches any one character.
665 .P
666 Matching characters by Unicode property is not fast, because PCRE has to search
667 a structure that contains data for over fifteen thousand characters. That is
668 why the traditional escape sequences such as \ed and \ew do not use Unicode
669 properties in PCRE.
670 .
671 .
672 .\" HTML <a name="resetmatchstart"></a>
673 .SS "Resetting the match start"
674 .rs
675 .sp
676 The escape sequence \eK, which is a Perl 5.10 feature, causes any previously
677 matched characters not to be included in the final matched sequence. For
678 example, the pattern:
679 .sp
680 foo\eKbar
681 .sp
682 matches "foobar", but reports that it has matched "bar". This feature is
683 similar to a lookbehind assertion
684 .\" HTML <a href="#lookbehind">
685 .\" </a>
686 (described below).
687 .\"
688 However, in this case, the part of the subject before the real match does not
689 have to be of fixed length, as lookbehind assertions do. The use of \eK does
690 not interfere with the setting of
691 .\" HTML <a href="#subpattern">
692 .\" </a>
693 captured substrings.
694 .\"
695 For example, when the pattern
696 .sp
697 (foo)\eKbar
698 .sp
699 matches "foobar", the first substring is still set to "foo".
700 .
701 .
702 .\" HTML <a name="smallassertions"></a>
703 .SS "Simple assertions"
704 .rs
705 .sp
706 The final use of backslash is for certain simple assertions. An assertion
707 specifies a condition that has to be met at a particular point in a match,
708 without consuming any characters from the subject string. The use of
709 subpatterns for more complicated assertions is described
710 .\" HTML <a href="#bigassertions">
711 .\" </a>
712 below.
713 .\"
714 The backslashed assertions are:
715 .sp
716 \eb matches at a word boundary
717 \eB matches when not at a word boundary
718 \eA matches at the start of the subject
719 \eZ matches at the end of the subject
720 also matches before a newline at the end of the subject
721 \ez matches only at the end of the subject
722 \eG matches at the first matching position in the subject
723 .sp
724 These assertions may not appear in character classes (but note that \eb has a
725 different meaning, namely the backspace character, inside a character class).
726 .P
727 A word boundary is a position in the subject string where the current character
728 and the previous character do not both match \ew or \eW (i.e. one matches
729 \ew and the other matches \eW), or the start or end of the string if the
730 first or last character matches \ew, respectively.
731 .P
732 The \eA, \eZ, and \ez assertions differ from the traditional circumflex and
733 dollar (described in the next section) in that they only ever match at the very
734 start and end of the subject string, whatever options are set. Thus, they are
735 independent of multiline mode. These three assertions are not affected by the
736 PCRE_NOTBOL or PCRE_NOTEOL options, which affect only the behaviour of the
737 circumflex and dollar metacharacters. However, if the \fIstartoffset\fP
738 argument of \fBpcre_exec()\fP is non-zero, indicating that matching is to start
739 at a point other than the beginning of the subject, \eA can never match. The
740 difference between \eZ and \ez is that \eZ matches before a newline at the end
741 of the string as well as at the very end, whereas \ez matches only at the end.
742 .P
743 The \eG assertion is true only when the current matching position is at the
744 start point of the match, as specified by the \fIstartoffset\fP argument of
745 \fBpcre_exec()\fP. It differs from \eA when the value of \fIstartoffset\fP is
746 non-zero. By calling \fBpcre_exec()\fP multiple times with appropriate
747 arguments, you can mimic Perl's /g option, and it is in this kind of
748 implementation where \eG can be useful.
749 .P
750 Note, however, that PCRE's interpretation of \eG, as the start of the current
751 match, is subtly different from Perl's, which defines it as the end of the
752 previous match. In Perl, these can be different when the previously matched
753 string was empty. Because PCRE does just one match at a time, it cannot
754 reproduce this behaviour.
755 .P
756 If all the alternatives of a pattern begin with \eG, the expression is anchored
757 to the starting match position, and the "anchored" flag is set in the compiled
758 regular expression.
759 .
760 .
762 .rs
763 .sp
764 Outside a character class, in the default matching mode, the circumflex
765 character is an assertion that is true only if the current matching point is
766 at the start of the subject string. If the \fIstartoffset\fP argument of
767 \fBpcre_exec()\fP is non-zero, circumflex can never match if the PCRE_MULTILINE
768 option is unset. Inside a character class, circumflex has an entirely different
769 meaning
770 .\" HTML <a href="#characterclass">
771 .\" </a>
772 (see below).
773 .\"
774 .P
775 Circumflex need not be the first character of the pattern if a number of
776 alternatives are involved, but it should be the first thing in each alternative
777 in which it appears if the pattern is ever to match that branch. If all
778 possible alternatives start with a circumflex, that is, if the pattern is
779 constrained to match only at the start of the subject, it is said to be an
780 "anchored" pattern. (There are also other constructs that can cause a pattern
781 to be anchored.)
782 .P
783 A dollar character is an assertion that is true only if the current matching
784 point is at the end of the subject string, or immediately before a newline
785 at the end of the string (by default). Dollar need not be the last character of
786 the pattern if a number of alternatives are involved, but it should be the last
787 item in any branch in which it appears. Dollar has no special meaning in a
788 character class.
789 .P
790 The meaning of dollar can be changed so that it matches only at the very end of
791 the string, by setting the PCRE_DOLLAR_ENDONLY option at compile time. This
792 does not affect the \eZ assertion.
793 .P
794 The meanings of the circumflex and dollar characters are changed if the
795 PCRE_MULTILINE option is set. When this is the case, a circumflex matches
796 immediately after internal newlines as well as at the start of the subject
797 string. It does not match after a newline that ends the string. A dollar
798 matches before any newlines in the string, as well as at the very end, when
799 PCRE_MULTILINE is set. When newline is specified as the two-character
800 sequence CRLF, isolated CR and LF characters do not indicate newlines.
801 .P
802 For example, the pattern /^abc$/ matches the subject string "def\enabc" (where
803 \en represents a newline) in multiline mode, but not otherwise. Consequently,
804 patterns that are anchored in single line mode because all branches start with
805 ^ are not anchored in multiline mode, and a match for circumflex is possible
806 when the \fIstartoffset\fP argument of \fBpcre_exec()\fP is non-zero. The
807 PCRE_DOLLAR_ENDONLY option is ignored if PCRE_MULTILINE is set.
808 .P
809 Note that the sequences \eA, \eZ, and \ez can be used to match the start and
810 end of the subject in both modes, and if all branches of a pattern start with
811 \eA it is always anchored, whether or not PCRE_MULTILINE is set.
812 .
813 .
815 .rs
816 .sp
817 Outside a character class, a dot in the pattern matches any one character in
818 the subject string except (by default) a character that signifies the end of a
819 line. In UTF-8 mode, the matched character may be more than one byte long.
820 .P
821 When a line ending is defined as a single character, dot never matches that
822 character; when the two-character sequence CRLF is used, dot does not match CR
823 if it is immediately followed by LF, but otherwise it matches all characters
824 (including isolated CRs and LFs). When any Unicode line endings are being
825 recognized, dot does not match CR or LF or any of the other line ending
826 characters.
827 .P
828 The behaviour of dot with regard to newlines can be changed. If the PCRE_DOTALL
829 option is set, a dot matches any one character, without exception. If the
830 two-character sequence CRLF is present in the subject string, it takes two dots
831 to match it.
832 .P
833 The handling of dot is entirely independent of the handling of circumflex and
834 dollar, the only relationship being that they both involve newlines. Dot has no
835 special meaning in a character class.
836 .
837 .
839 .rs
840 .sp
841 Outside a character class, the escape sequence \eC matches any one byte, both
842 in and out of UTF-8 mode. Unlike a dot, it always matches any line-ending
843 characters. The feature is provided in Perl in order to match individual bytes
844 in UTF-8 mode. Because it breaks up UTF-8 characters into individual bytes,
845 what remains in the string may be a malformed UTF-8 string. For this reason,
846 the \eC escape sequence is best avoided.
847 .P
848 PCRE does not allow \eC to appear in lookbehind assertions
849 .\" HTML <a href="#lookbehind">
850 .\" </a>
851 (described below),
852 .\"
853 because in UTF-8 mode this would make it impossible to calculate the length of
854 the lookbehind.
855 .
856 .
857 .\" HTML <a name="characterclass"></a>
859 .rs
860 .sp
861 An opening square bracket introduces a character class, terminated by a closing
862 square bracket. A closing square bracket on its own is not special. If a
863 closing square bracket is required as a member of the class, it should be the
864 first data character in the class (after an initial circumflex, if present) or
865 escaped with a backslash.
866 .P
867 A character class matches a single character in the subject. In UTF-8 mode, the
868 character may occupy more than one byte. A matched character must be in the set
869 of characters defined by the class, unless the first character in the class
870 definition is a circumflex, in which case the subject character must not be in
871 the set defined by the class. If a circumflex is actually required as a member
872 of the class, ensure it is not the first character, or escape it with a
873 backslash.
874 .P
875 For example, the character class [aeiou] matches any lower case vowel, while
876 [^aeiou] matches any character that is not a lower case vowel. Note that a
877 circumflex is just a convenient notation for specifying the characters that
878 are in the class by enumerating those that are not. A class that starts with a
879 circumflex is not an assertion: it still consumes a character from the subject
880 string, and therefore it fails if the current pointer is at the end of the
881 string.
882 .P
883 In UTF-8 mode, characters with values greater than 255 can be included in a
884 class as a literal string of bytes, or by using the \ex{ escaping mechanism.
885 .P
886 When caseless matching is set, any letters in a class represent both their
887 upper case and lower case versions, so for example, a caseless [aeiou] matches
888 "A" as well as "a", and a caseless [^aeiou] does not match "A", whereas a
889 caseful version would. In UTF-8 mode, PCRE always understands the concept of
890 case for characters whose values are less than 128, so caseless matching is
891 always possible. For characters with higher values, the concept of case is
892 supported if PCRE is compiled with Unicode property support, but not otherwise.
893 If you want to use caseless matching for characters 128 and above, you must
894 ensure that PCRE is compiled with Unicode property support as well as with
895 UTF-8 support.
896 .P
897 Characters that might indicate line breaks are never treated in any special way
898 when matching character classes, whatever line-ending sequence is in use, and
899 whatever setting of the PCRE_DOTALL and PCRE_MULTILINE options is used. A class
900 such as [^a] always matches one of these characters.
901 .P
902 The minus (hyphen) character can be used to specify a range of characters in a
903 character class. For example, [d-m] matches any letter between d and m,
904 inclusive. If a minus character is required in a class, it must be escaped with
905 a backslash or appear in a position where it cannot be interpreted as
906 indicating a range, typically as the first or last character in the class.
907 .P
908 It is not possible to have the literal character "]" as the end character of a
909 range. A pattern such as [W-]46] is interpreted as a class of two characters
910 ("W" and "-") followed by a literal string "46]", so it would match "W46]" or
911 "-46]". However, if the "]" is escaped with a backslash it is interpreted as
912 the end of range, so [W-\e]46] is interpreted as a class containing a range
913 followed by two other characters. The octal or hexadecimal representation of
914 "]" can also be used to end a range.
915 .P
916 Ranges operate in the collating sequence of character values. They can also be
917 used for characters specified numerically, for example [\e000-\e037]. In UTF-8
918 mode, ranges can include characters whose values are greater than 255, for
919 example [\ex{100}-\ex{2ff}].
920 .P
921 If a range that includes letters is used when caseless matching is set, it
922 matches the letters in either case. For example, [W-c] is equivalent to
923 [][\e\e^_`wxyzabc], matched caselessly, and in non-UTF-8 mode, if character
924 tables for a French locale are in use, [\exc8-\excb] matches accented E
925 characters in both cases. In UTF-8 mode, PCRE supports the concept of case for
926 characters with values greater than 128 only when it is compiled with Unicode
927 property support.
928 .P
929 The character types \ed, \eD, \ep, \eP, \es, \eS, \ew, and \eW may also appear
930 in a character class, and add the characters that they match to the class. For
931 example, [\edABCDEF] matches any hexadecimal digit. A circumflex can
932 conveniently be used with the upper case character types to specify a more
933 restricted set of characters than the matching lower case type. For example,
934 the class [^\eW_] matches any letter or digit, but not underscore.
935 .P
936 The only metacharacters that are recognized in character classes are backslash,
937 hyphen (only where it can be interpreted as specifying a range), circumflex
938 (only at the start), opening square bracket (only when it can be interpreted as
939 introducing a POSIX class name - see the next section), and the terminating
940 closing square bracket. However, escaping other non-alphanumeric characters
941 does no harm.
942 .
943 .
945 .rs
946 .sp
947 Perl supports the POSIX notation for character classes. This uses names
948 enclosed by [: and :] within the enclosing square brackets. PCRE also supports
949 this notation. For example,
950 .sp
951 [01[:alpha:]%]
952 .sp
953 matches "0", "1", any alphabetic character, or "%". The supported class names
954 are
955 .sp
956 alnum letters and digits
957 alpha letters
958 ascii character codes 0 - 127
959 blank space or tab only
960 cntrl control characters
961 digit decimal digits (same as \ed)
962 graph printing characters, excluding space
963 lower lower case letters
964 print printing characters, including space
965 punct printing characters, excluding letters and digits
966 space white space (not quite the same as \es)
967 upper upper case letters
968 word "word" characters (same as \ew)
969 xdigit hexadecimal digits
970 .sp
971 The "space" characters are HT (9), LF (10), VT (11), FF (12), CR (13), and
972 space (32). Notice that this list includes the VT character (code 11). This
973 makes "space" different to \es, which does not include VT (for Perl
974 compatibility).
975 .P
976 The name "word" is a Perl extension, and "blank" is a GNU extension from Perl
977 5.8. Another Perl extension is negation, which is indicated by a ^ character
978 after the colon. For example,
979 .sp
980 [12[:^digit:]]
981 .sp
982 matches "1", "2", or any non-digit. PCRE (and Perl) also recognize the POSIX
983 syntax [.ch.] and [=ch=] where "ch" is a "collating element", but these are not
984 supported, and an error is given if they are encountered.
985 .P
986 In UTF-8 mode, characters with values greater than 128 do not match any of
987 the POSIX character classes.
988 .
989 .
991 .rs
992 .sp
993 Vertical bar characters are used to separate alternative patterns. For example,
994 the pattern
995 .sp
996 gilbert|sullivan
997 .sp
998 matches either "gilbert" or "sullivan". Any number of alternatives may appear,
999 and an empty alternative is permitted (matching the empty string). The matching
1000 process tries each alternative in turn, from left to right, and the first one
1001 that succeeds is used. If the alternatives are within a subpattern
1002 .\" HTML <a href="#subpattern">
1003 .\" </a>
1004 (defined below),
1005 .\"
1006 "succeeds" means matching the rest of the main pattern as well as the
1007 alternative in the subpattern.
1008 .
1009 .
1011 .rs
1012 .sp
1013 The settings of the PCRE_CASELESS, PCRE_MULTILINE, PCRE_DOTALL, and
1014 PCRE_EXTENDED options (which are Perl-compatible) can be changed from within
1015 the pattern by a sequence of Perl option letters enclosed between "(?" and ")".
1016 The option letters are
1017 .sp
1018 i for PCRE_CASELESS
1020 s for PCRE_DOTALL
1021 x for PCRE_EXTENDED
1022 .sp
1023 For example, (?im) sets caseless, multiline matching. It is also possible to
1024 unset these options by preceding the letter with a hyphen, and a combined
1025 setting and unsetting such as (?im-sx), which sets PCRE_CASELESS and
1026 PCRE_MULTILINE while unsetting PCRE_DOTALL and PCRE_EXTENDED, is also
1027 permitted. If a letter appears both before and after the hyphen, the option is
1028 unset.
1029 .P
1030 The PCRE-specific options PCRE_DUPNAMES, PCRE_UNGREEDY, and PCRE_EXTRA can be
1031 changed in the same way as the Perl-compatible options by using the characters
1032 J, U and X respectively.
1033 .P
1034 When an option change occurs at top level (that is, not inside subpattern
1035 parentheses), the change applies to the remainder of the pattern that follows.
1036 If the change is placed right at the start of a pattern, PCRE extracts it into
1037 the global options (and it will therefore show up in data extracted by the
1038 \fBpcre_fullinfo()\fP function).
1039 .P
1040 An option change within a subpattern (see below for a description of
1041 subpatterns) affects only that part of the current pattern that follows it, so
1042 .sp
1043 (a(?i)b)c
1044 .sp
1045 matches abc and aBc and no other strings (assuming PCRE_CASELESS is not used).
1046 By this means, options can be made to have different settings in different
1047 parts of the pattern. Any changes made in one alternative do carry on
1048 into subsequent branches within the same subpattern. For example,
1049 .sp
1050 (a(?i)b|c)
1051 .sp
1052 matches "ab", "aB", "c", and "C", even though when matching "C" the first
1053 branch is abandoned before the option setting. This is because the effects of
1054 option settings happen at compile time. There would be some very weird
1055 behaviour otherwise.
1056 .P
1057 \fBNote:\fP There are other PCRE-specific options that can be set by the
1058 application when the compile or match functions are called. In some cases the
1059 pattern can contain special leading sequences to override what the application
1060 has set or what has been defaulted. Details are given in the section entitled
1061 .\" HTML <a href="#newlineseq">
1062 .\" </a>
1063 "Newline sequences"
1064 .\"
1065 above.
1066 .
1067 .
1068 .\" HTML <a name="subpattern"></a>
1070 .rs
1071 .sp
1072 Subpatterns are delimited by parentheses (round brackets), which can be nested.
1073 Turning part of a pattern into a subpattern does two things:
1074 .sp
1075 1. It localizes a set of alternatives. For example, the pattern
1076 .sp
1077 cat(aract|erpillar|)
1078 .sp
1079 matches one of the words "cat", "cataract", or "caterpillar". Without the
1080 parentheses, it would match "cataract", "erpillar" or an empty string.
1081 .sp
1082 2. It sets up the subpattern as a capturing subpattern. This means that, when
1083 the whole pattern matches, that portion of the subject string that matched the
1084 subpattern is passed back to the caller via the \fIovector\fP argument of
1085 \fBpcre_exec()\fP. Opening parentheses are counted from left to right (starting
1086 from 1) to obtain numbers for the capturing subpatterns.
1087 .P
1088 For example, if the string "the red king" is matched against the pattern
1089 .sp
1090 the ((red|white) (king|queen))
1091 .sp
1092 the captured substrings are "red king", "red", and "king", and are numbered 1,
1093 2, and 3, respectively.
1094 .P
1095 The fact that plain parentheses fulfil two functions is not always helpful.
1096 There are often times when a grouping subpattern is required without a
1097 capturing requirement. If an opening parenthesis is followed by a question mark
1098 and a colon, the subpattern does not do any capturing, and is not counted when
1099 computing the number of any subsequent capturing subpatterns. For example, if
1100 the string "the white queen" is matched against the pattern
1101 .sp
1102 the ((?:red|white) (king|queen))
1103 .sp
1104 the captured substrings are "white queen" and "queen", and are numbered 1 and
1105 2. The maximum number of capturing subpatterns is 65535.
1106 .P
1107 As a convenient shorthand, if any option settings are required at the start of
1108 a non-capturing subpattern, the option letters may appear between the "?" and
1109 the ":". Thus the two patterns
1110 .sp
1111 (?i:saturday|sunday)
1112 (?:(?i)saturday|sunday)
1113 .sp
1114 match exactly the same set of strings. Because alternative branches are tried
1115 from left to right, and options are not reset until the end of the subpattern
1116 is reached, an option setting in one branch does affect subsequent branches, so
1117 the above patterns match "SUNDAY" as well as "Saturday".
1118 .
1119 .
1121 .rs
1122 .sp
1123 Perl 5.10 introduced a feature whereby each alternative in a subpattern uses
1124 the same numbers for its capturing parentheses. Such a subpattern starts with
1125 (?| and is itself a non-capturing subpattern. For example, consider this
1126 pattern:
1127 .sp
1128 (?|(Sat)ur|(Sun))day
1129 .sp
1130 Because the two alternatives are inside a (?| group, both sets of capturing
1131 parentheses are numbered one. Thus, when the pattern matches, you can look
1132 at captured substring number one, whichever alternative matched. This construct
1133 is useful when you want to capture part, but not all, of one of a number of
1134 alternatives. Inside a (?| group, parentheses are numbered as usual, but the
1135 number is reset at the start of each branch. The numbers of any capturing
1136 buffers that follow the subpattern start after the highest number used in any
1137 branch. The following example is taken from the Perl documentation.
1138 The numbers underneath show in which buffer the captured content will be
1139 stored.
1140 .sp
1141 # before ---------------branch-reset----------- after
1142 / ( a ) (?| x ( y ) z | (p (q) r) | (t) u (v) ) ( z ) /x
1143 # 1 2 2 3 2 3 4
1144 .sp
1145 A backreference or a recursive call to a numbered subpattern always refers to
1146 the first one in the pattern with the given number.
1147 .P
1148 An alternative approach to using this "branch reset" feature is to use
1149 duplicate named subpatterns, as described in the next section.
1150 .
1151 .
1153 .rs
1154 .sp
1155 Identifying capturing parentheses by number is simple, but it can be very hard
1156 to keep track of the numbers in complicated regular expressions. Furthermore,
1157 if an expression is modified, the numbers may change. To help with this
1158 difficulty, PCRE supports the naming of subpatterns. This feature was not
1159 added to Perl until release 5.10. Python had the feature earlier, and PCRE
1160 introduced it at release 4.0, using the Python syntax. PCRE now supports both
1161 the Perl and the Python syntax.
1162 .P
1163 In PCRE, a subpattern can be named in one of three ways: (?<name>...) or
1164 (?'name'...) as in Perl, or (?P<name>...) as in Python. References to capturing
1165 parentheses from other parts of the pattern, such as
1166 .\" HTML <a href="#backreferences">
1167 .\" </a>
1168 backreferences,
1169 .\"
1170 .\" HTML <a href="#recursion">
1171 .\" </a>
1172 recursion,
1173 .\"
1174 and
1175 .\" HTML <a href="#conditions">
1176 .\" </a>
1177 conditions,
1178 .\"
1179 can be made by name as well as by number.
1180 .P
1181 Names consist of up to 32 alphanumeric characters and underscores. Named
1182 capturing parentheses are still allocated numbers as well as names, exactly as
1183 if the names were not present. The PCRE API provides function calls for
1184 extracting the name-to-number translation table from a compiled pattern. There
1185 is also a convenience function for extracting a captured substring by name.
1186 .P
1187 By default, a name must be unique within a pattern, but it is possible to relax
1188 this constraint by setting the PCRE_DUPNAMES option at compile time. This can
1189 be useful for patterns where only one instance of the named parentheses can
1190 match. Suppose you want to match the name of a weekday, either as a 3-letter
1191 abbreviation or as the full name, and in both cases you want to extract the
1192 abbreviation. This pattern (ignoring the line breaks) does the job:
1193 .sp
1194 (?<DN>Mon|Fri|Sun)(?:day)?|
1195 (?<DN>Tue)(?:sday)?|
1196 (?<DN>Wed)(?:nesday)?|
1197 (?<DN>Thu)(?:rsday)?|
1198 (?<DN>Sat)(?:urday)?
1199 .sp
1200 There are five capturing substrings, but only one is ever set after a match.
1201 (An alternative way of solving this problem is to use a "branch reset"
1202 subpattern, as described in the previous section.)
1203 .P
1204 The convenience function for extracting the data by name returns the substring
1205 for the first (and in this example, the only) subpattern of that name that
1206 matched. This saves searching to find which numbered subpattern it was. If you
1207 make a reference to a non-unique named subpattern from elsewhere in the
1208 pattern, the one that corresponds to the lowest number is used. For further
1209 details of the interfaces for handling named subpatterns, see the
1210 .\" HREF
1211 \fBpcreapi\fP
1212 .\"
1213 documentation.
1214 .P
1215 \fBWarning:\fP You cannot use different names to distinguish between two
1216 subpatterns with the same number (see the previous section) because PCRE uses
1217 only the numbers when matching.
1218 .
1219 .
1221 .rs
1222 .sp
1223 Repetition is specified by quantifiers, which can follow any of the following
1224 items:
1225 .sp
1226 a literal data character
1227 the dot metacharacter
1228 the \eC escape sequence
1229 the \eX escape sequence (in UTF-8 mode with Unicode properties)
1230 the \eR escape sequence
1231 an escape such as \ed that matches a single character
1232 a character class
1233 a back reference (see next section)
1234 a parenthesized subpattern (unless it is an assertion)
1235 .sp
1236 The general repetition quantifier specifies a minimum and maximum number of
1237 permitted matches, by giving the two numbers in curly brackets (braces),
1238 separated by a comma. The numbers must be less than 65536, and the first must
1239 be less than or equal to the second. For example:
1240 .sp
1241 z{2,4}
1242 .sp
1243 matches "zz", "zzz", or "zzzz". A closing brace on its own is not a special
1244 character. If the second number is omitted, but the comma is present, there is
1245 no upper limit; if the second number and the comma are both omitted, the
1246 quantifier specifies an exact number of required matches. Thus
1247 .sp
1248 [aeiou]{3,}
1249 .sp
1250 matches at least 3 successive vowels, but may match many more, while
1251 .sp
1252 \ed{8}
1253 .sp
1254 matches exactly 8 digits. An opening curly bracket that appears in a position
1255 where a quantifier is not allowed, or one that does not match the syntax of a
1256 quantifier, is taken as a literal character. For example, {,6} is not a
1257 quantifier, but a literal string of four characters.
1258 .P
1259 In UTF-8 mode, quantifiers apply to UTF-8 characters rather than to individual
1260 bytes. Thus, for example, \ex{100}{2} matches two UTF-8 characters, each of
1261 which is represented by a two-byte sequence. Similarly, when Unicode property
1262 support is available, \eX{3} matches three Unicode extended sequences, each of
1263 which may be several bytes long (and they may be of different lengths).
1264 .P
1265 The quantifier {0} is permitted, causing the expression to behave as if the
1266 previous item and the quantifier were not present. This may be useful for
1267 subpatterns that are referenced as
1268 .\" HTML <a href="#subpatternsassubroutines">
1269 .\" </a>
1270 subroutines
1271 .\"
1272 from elsewhere in the pattern. Items other than subpatterns that have a {0}
1273 quantifier are omitted from the compiled pattern.
1274 .P
1275 For convenience, the three most common quantifiers have single-character
1276 abbreviations:
1277 .sp
1278 * is equivalent to {0,}
1279 + is equivalent to {1,}
1280 ? is equivalent to {0,1}
1281 .sp
1282 It is possible to construct infinite loops by following a subpattern that can
1283 match no characters with a quantifier that has no upper limit, for example:
1284 .sp
1285 (a?)*
1286 .sp
1287 Earlier versions of Perl and PCRE used to give an error at compile time for
1288 such patterns. However, because there are cases where this can be useful, such
1289 patterns are now accepted, but if any repetition of the subpattern does in fact
1290 match no characters, the loop is forcibly broken.
1291 .P
1292 By default, the quantifiers are "greedy", that is, they match as much as
1293 possible (up to the maximum number of permitted times), without causing the
1294 rest of the pattern to fail. The classic example of where this gives problems
1295 is in trying to match comments in C programs. These appear between /* and */
1296 and within the comment, individual * and / characters may appear. An attempt to
1297 match C comments by applying the pattern
1298 .sp
1299 /\e*.*\e*/
1300 .sp
1301 to the string
1302 .sp
1303 /* first comment */ not comment /* second comment */
1304 .sp
1305 fails, because it matches the entire string owing to the greediness of the .*
1306 item.
1307 .P
1308 However, if a quantifier is followed by a question mark, it ceases to be
1309 greedy, and instead matches the minimum number of times possible, so the
1310 pattern
1311 .sp
1312 /\e*.*?\e*/
1313 .sp
1314 does the right thing with the C comments. The meaning of the various
1315 quantifiers is not otherwise changed, just the preferred number of matches.
1316 Do not confuse this use of question mark with its use as a quantifier in its
1317 own right. Because it has two uses, it can sometimes appear doubled, as in
1318 .sp
1319 \ed??\ed
1320 .sp
1321 which matches one digit by preference, but can match two if that is the only
1322 way the rest of the pattern matches.
1323 .P
1324 If the PCRE_UNGREEDY option is set (an option that is not available in Perl),
1325 the quantifiers are not greedy by default, but individual ones can be made
1326 greedy by following them with a question mark. In other words, it inverts the
1327 default behaviour.
1328 .P
1329 When a parenthesized subpattern is quantified with a minimum repeat count that
1330 is greater than 1 or with a limited maximum, more memory is required for the
1331 compiled pattern, in proportion to the size of the minimum or maximum.
1332 .P
1333 If a pattern starts with .* or .{0,} and the PCRE_DOTALL option (equivalent
1334 to Perl's /s) is set, thus allowing the dot to match newlines, the pattern is
1335 implicitly anchored, because whatever follows will be tried against every
1336 character position in the subject string, so there is no point in retrying the
1337 overall match at any position after the first. PCRE normally treats such a
1338 pattern as though it were preceded by \eA.
1339 .P
1340 In cases where it is known that the subject string contains no newlines, it is
1341 worth setting PCRE_DOTALL in order to obtain this optimization, or
1342 alternatively using ^ to indicate anchoring explicitly.
1343 .P
1344 However, there is one situation where the optimization cannot be used. When .*
1345 is inside capturing parentheses that are the subject of a backreference
1346 elsewhere in the pattern, a match at the start may fail where a later one
1347 succeeds. Consider, for example:
1348 .sp
1349 (.*)abc\e1
1350 .sp
1351 If the subject is "xyz123abc123" the match point is the fourth character. For
1352 this reason, such a pattern is not implicitly anchored.
1353 .P
1354 When a capturing subpattern is repeated, the value captured is the substring
1355 that matched the final iteration. For example, after
1356 .sp
1357 (tweedle[dume]{3}\es*)+
1358 .sp
1359 has matched "tweedledum tweedledee" the value of the captured substring is
1360 "tweedledee". However, if there are nested capturing subpatterns, the
1361 corresponding captured values may have been set in previous iterations. For
1362 example, after
1363 .sp
1364 /(a|(b))+/
1365 .sp
1366 matches "aba" the value of the second captured substring is "b".
1367 .
1368 .
1369 .\" HTML <a name="atomicgroup"></a>
1371 .rs
1372 .sp
1373 With both maximizing ("greedy") and minimizing ("ungreedy" or "lazy")
1374 repetition, failure of what follows normally causes the repeated item to be
1375 re-evaluated to see if a different number of repeats allows the rest of the
1376 pattern to match. Sometimes it is useful to prevent this, either to change the
1377 nature of the match, or to cause it fail earlier than it otherwise might, when
1378 the author of the pattern knows there is no point in carrying on.
1379 .P
1380 Consider, for example, the pattern \ed+foo when applied to the subject line
1381 .sp
1382 123456bar
1383 .sp
1384 After matching all 6 digits and then failing to match "foo", the normal
1385 action of the matcher is to try again with only 5 digits matching the \ed+
1386 item, and then with 4, and so on, before ultimately failing. "Atomic grouping"
1387 (a term taken from Jeffrey Friedl's book) provides the means for specifying
1388 that once a subpattern has matched, it is not to be re-evaluated in this way.
1389 .P
1390 If we use atomic grouping for the previous example, the matcher gives up
1391 immediately on failing to match "foo" the first time. The notation is a kind of
1392 special parenthesis, starting with (?> as in this example:
1393 .sp
1394 (?>\ed+)foo
1395 .sp
1396 This kind of parenthesis "locks up" the part of the pattern it contains once
1397 it has matched, and a failure further into the pattern is prevented from
1398 backtracking into it. Backtracking past it to previous items, however, works as
1399 normal.
1400 .P
1401 An alternative description is that a subpattern of this type matches the string
1402 of characters that an identical standalone pattern would match, if anchored at
1403 the current point in the subject string.
1404 .P
1405 Atomic grouping subpatterns are not capturing subpatterns. Simple cases such as
1406 the above example can be thought of as a maximizing repeat that must swallow
1407 everything it can. So, while both \ed+ and \ed+? are prepared to adjust the
1408 number of digits they match in order to make the rest of the pattern match,
1409 (?>\ed+) can only match an entire sequence of digits.
1410 .P
1411 Atomic groups in general can of course contain arbitrarily complicated
1412 subpatterns, and can be nested. However, when the subpattern for an atomic
1413 group is just a single repeated item, as in the example above, a simpler
1414 notation, called a "possessive quantifier" can be used. This consists of an
1415 additional + character following a quantifier. Using this notation, the
1416 previous example can be rewritten as
1417 .sp
1418 \ed++foo
1419 .sp
1420 Note that a possessive quantifier can be used with an entire group, for
1421 example:
1422 .sp
1423 (abc|xyz){2,3}+
1424 .sp
1425 Possessive quantifiers are always greedy; the setting of the PCRE_UNGREEDY
1426 option is ignored. They are a convenient notation for the simpler forms of
1427 atomic group. However, there is no difference in the meaning of a possessive
1428 quantifier and the equivalent atomic group, though there may be a performance
1429 difference; possessive quantifiers should be slightly faster.
1430 .P
1431 The possessive quantifier syntax is an extension to the Perl 5.8 syntax.
1432 Jeffrey Friedl originated the idea (and the name) in the first edition of his
1433 book. Mike McCloskey liked it, so implemented it when he built Sun's Java
1434 package, and PCRE copied it from there. It ultimately found its way into Perl
1435 at release 5.10.
1436 .P
1437 PCRE has an optimization that automatically "possessifies" certain simple
1438 pattern constructs. For example, the sequence A+B is treated as A++B because
1439 there is no point in backtracking into a sequence of A's when B must follow.
1440 .P
1441 When a pattern contains an unlimited repeat inside a subpattern that can itself
1442 be repeated an unlimited number of times, the use of an atomic group is the
1443 only way to avoid some failing matches taking a very long time indeed. The
1444 pattern
1445 .sp
1446 (\eD+|<\ed+>)*[!?]
1447 .sp
1448 matches an unlimited number of substrings that either consist of non-digits, or
1449 digits enclosed in <>, followed by either ! or ?. When it matches, it runs
1450 quickly. However, if it is applied to
1451 .sp
1452 aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
1453 .sp
1454 it takes a long time before reporting failure. This is because the string can
1455 be divided between the internal \eD+ repeat and the external * repeat in a
1456 large number of ways, and all have to be tried. (The example uses [!?] rather
1457 than a single character at the end, because both PCRE and Perl have an
1458 optimization that allows for fast failure when a single character is used. They
1459 remember the last single character that is required for a match, and fail early
1460 if it is not present in the string.) If the pattern is changed so that it uses
1461 an atomic group, like this:
1462 .sp
1463 ((?>\eD+)|<\ed+>)*[!?]
1464 .sp
1465 sequences of non-digits cannot be broken, and failure happens quickly.
1466 .
1467 .
1468 .\" HTML <a name="backreferences"></a>
1470 .rs
1471 .sp
1472 Outside a character class, a backslash followed by a digit greater than 0 (and
1473 possibly further digits) is a back reference to a capturing subpattern earlier
1474 (that is, to its left) in the pattern, provided there have been that many
1475 previous capturing left parentheses.
1476 .P
1477 However, if the decimal number following the backslash is less than 10, it is
1478 always taken as a back reference, and causes an error only if there are not
1479 that many capturing left parentheses in the entire pattern. In other words, the
1480 parentheses that are referenced need not be to the left of the reference for
1481 numbers less than 10. A "forward back reference" of this type can make sense
1482 when a repetition is involved and the subpattern to the right has participated
1483 in an earlier iteration.
1484 .P
1485 It is not possible to have a numerical "forward back reference" to a subpattern
1486 whose number is 10 or more using this syntax because a sequence such as \e50 is
1487 interpreted as a character defined in octal. See the subsection entitled
1488 "Non-printing characters"
1489 .\" HTML <a href="#digitsafterbackslash">
1490 .\" </a>
1491 above
1492 .\"
1493 for further details of the handling of digits following a backslash. There is
1494 no such problem when named parentheses are used. A back reference to any
1495 subpattern is possible using named parentheses (see below).
1496 .P
1497 Another way of avoiding the ambiguity inherent in the use of digits following a
1498 backslash is to use the \eg escape sequence, which is a feature introduced in
1499 Perl 5.10. This escape must be followed by an unsigned number or a negative
1500 number, optionally enclosed in braces. These examples are all identical:
1501 .sp
1502 (ring), \e1
1503 (ring), \eg1
1504 (ring), \eg{1}
1505 .sp
1506 An unsigned number specifies an absolute reference without the ambiguity that
1507 is present in the older syntax. It is also useful when literal digits follow
1508 the reference. A negative number is a relative reference. Consider this
1509 example:
1510 .sp
1511 (abc(def)ghi)\eg{-1}
1512 .sp
1513 The sequence \eg{-1} is a reference to the most recently started capturing
1514 subpattern before \eg, that is, is it equivalent to \e2. Similarly, \eg{-2}
1515 would be equivalent to \e1. The use of relative references can be helpful in
1516 long patterns, and also in patterns that are created by joining together
1517 fragments that contain references within themselves.
1518 .P
1519 A back reference matches whatever actually matched the capturing subpattern in
1520 the current subject string, rather than anything matching the subpattern
1521 itself (see
1522 .\" HTML <a href="#subpatternsassubroutines">
1523 .\" </a>
1524 "Subpatterns as subroutines"
1525 .\"
1526 below for a way of doing that). So the pattern
1527 .sp
1528 (sens|respons)e and \e1ibility
1529 .sp
1530 matches "sense and sensibility" and "response and responsibility", but not
1531 "sense and responsibility". If caseful matching is in force at the time of the
1532 back reference, the case of letters is relevant. For example,
1533 .sp
1534 ((?i)rah)\es+\e1
1535 .sp
1536 matches "rah rah" and "RAH RAH", but not "RAH rah", even though the original
1537 capturing subpattern is matched caselessly.
1538 .P
1539 There are several different ways of writing back references to named
1540 subpatterns. The .NET syntax \ek{name} and the Perl syntax \ek<name> or
1541 \ek'name' are supported, as is the Python syntax (?P=name). Perl 5.10's unified
1542 back reference syntax, in which \eg can be used for both numeric and named
1543 references, is also supported. We could rewrite the above example in any of
1544 the following ways:
1545 .sp
1546 (?<p1>(?i)rah)\es+\ek<p1>
1547 (?'p1'(?i)rah)\es+\ek{p1}
1548 (?P<p1>(?i)rah)\es+(?P=p1)
1549 (?<p1>(?i)rah)\es+\eg{p1}
1550 .sp
1551 A subpattern that is referenced by name may appear in the pattern before or
1552 after the reference.
1553 .P
1554 There may be more than one back reference to the same subpattern. If a
1555 subpattern has not actually been used in a particular match, any back
1556 references to it always fail. For example, the pattern
1557 .sp
1558 (a|(bc))\e2
1559 .sp
1560 always fails if it starts to match "a" rather than "bc". Because there may be
1561 many capturing parentheses in a pattern, all digits following the backslash are
1562 taken as part of a potential back reference number. If the pattern continues
1563 with a digit character, some delimiter must be used to terminate the back
1564 reference. If the PCRE_EXTENDED option is set, this can be whitespace.
1565 Otherwise an empty comment (see
1566 .\" HTML <a href="#comments">
1567 .\" </a>
1568 "Comments"
1569 .\"
1570 below) can be used.
1571 .P
1572 A back reference that occurs inside the parentheses to which it refers fails
1573 when the subpattern is first used, so, for example, (a\e1) never matches.
1574 However, such references can be useful inside repeated subpatterns. For
1575 example, the pattern
1576 .sp
1577 (a|b\e1)+
1578 .sp
1579 matches any number of "a"s and also "aba", "ababbaa" etc. At each iteration of
1580 the subpattern, the back reference matches the character string corresponding
1581 to the previous iteration. In order for this to work, the pattern must be such
1582 that the first iteration does not need to match the back reference. This can be
1583 done using alternation, as in the example above, or by a quantifier with a
1584 minimum of zero.
1585 .
1586 .
1587 .\" HTML <a name="bigassertions"></a>
1589 .rs
1590 .sp
1591 An assertion is a test on the characters following or preceding the current
1592 matching point that does not actually consume any characters. The simple
1593 assertions coded as \eb, \eB, \eA, \eG, \eZ, \ez, ^ and $ are described
1594 .\" HTML <a href="#smallassertions">
1595 .\" </a>
1596 above.
1597 .\"
1598 .P
1599 More complicated assertions are coded as subpatterns. There are two kinds:
1600 those that look ahead of the current position in the subject string, and those
1601 that look behind it. An assertion subpattern is matched in the normal way,
1602 except that it does not cause the current matching position to be changed.
1603 .P
1604 Assertion subpatterns are not capturing subpatterns, and may not be repeated,
1605 because it makes no sense to assert the same thing several times. If any kind
1606 of assertion contains capturing subpatterns within it, these are counted for
1607 the purposes of numbering the capturing subpatterns in the whole pattern.
1608 However, substring capturing is carried out only for positive assertions,
1609 because it does not make sense for negative assertions.
1610 .
1611 .
1612 .SS "Lookahead assertions"
1613 .rs
1614 .sp
1615 Lookahead assertions start with (?= for positive assertions and (?! for
1616 negative assertions. For example,
1617 .sp
1618 \ew+(?=;)
1619 .sp
1620 matches a word followed by a semicolon, but does not include the semicolon in
1621 the match, and
1622 .sp
1623 foo(?!bar)
1624 .sp
1625 matches any occurrence of "foo" that is not followed by "bar". Note that the
1626 apparently similar pattern
1627 .sp
1628 (?!foo)bar
1629 .sp
1630 does not find an occurrence of "bar" that is preceded by something other than
1631 "foo"; it finds any occurrence of "bar" whatsoever, because the assertion
1632 (?!foo) is always true when the next three characters are "bar". A
1633 lookbehind assertion is needed to achieve the other effect.
1634 .P
1635 If you want to force a matching failure at some point in a pattern, the most
1636 convenient way to do it is with (?!) because an empty string always matches, so
1637 an assertion that requires there not to be an empty string must always fail.
1638 .
1639 .
1640 .\" HTML <a name="lookbehind"></a>
1641 .SS "Lookbehind assertions"
1642 .rs
1643 .sp
1644 Lookbehind assertions start with (?<= for positive assertions and (?<! for
1645 negative assertions. For example,
1646 .sp
1647 (?<!foo)bar
1648 .sp
1649 does find an occurrence of "bar" that is not preceded by "foo". The contents of
1650 a lookbehind assertion are restricted such that all the strings it matches must
1651 have a fixed length. However, if there are several top-level alternatives, they
1652 do not all have to have the same fixed length. Thus
1653 .sp
1654 (?<=bullock|donkey)
1655 .sp
1656 is permitted, but
1657 .sp
1658 (?<!dogs?|cats?)
1659 .sp
1660 causes an error at compile time. Branches that match different length strings
1661 are permitted only at the top level of a lookbehind assertion. This is an
1662 extension compared with Perl (at least for 5.8), which requires all branches to
1663 match the same length of string. An assertion such as
1664 .sp
1665 (?<=ab(c|de))
1666 .sp
1667 is not permitted, because its single top-level branch can match two different
1668 lengths, but it is acceptable if rewritten to use two top-level branches:
1669 .sp
1670 (?<=abc|abde)
1671 .sp
1672 In some cases, the Perl 5.10 escape sequence \eK
1673 .\" HTML <a href="#resetmatchstart">
1674 .\" </a>
1675 (see above)
1676 .\"
1677 can be used instead of a lookbehind assertion; this is not restricted to a
1678 fixed-length.
1679 .P
1680 The implementation of lookbehind assertions is, for each alternative, to
1681 temporarily move the current position back by the fixed length and then try to
1682 match. If there are insufficient characters before the current position, the
1683 assertion fails.
1684 .P
1685 PCRE does not allow the \eC escape (which matches a single byte in UTF-8 mode)
1686 to appear in lookbehind assertions, because it makes it impossible to calculate
1687 the length of the lookbehind. The \eX and \eR escapes, which can match
1688 different numbers of bytes, are also not permitted.
1689 .P
1690 Possessive quantifiers can be used in conjunction with lookbehind assertions to
1691 specify efficient matching at the end of the subject string. Consider a simple
1692 pattern such as
1693 .sp
1694 abcd$
1695 .sp
1696 when applied to a long string that does not match. Because matching proceeds
1697 from left to right, PCRE will look for each "a" in the subject and then see if
1698 what follows matches the rest of the pattern. If the pattern is specified as
1699 .sp
1700 ^.*abcd$
1701 .sp
1702 the initial .* matches the entire string at first, but when this fails (because
1703 there is no following "a"), it backtracks to match all but the last character,
1704 then all but the last two characters, and so on. Once again the search for "a"
1705 covers the entire string, from right to left, so we are no better off. However,
1706 if the pattern is written as
1707 .sp
1708 ^.*+(?<=abcd)
1709 .sp
1710 there can be no backtracking for the .*+ item; it can match only the entire
1711 string. The subsequent lookbehind assertion does a single test on the last four
1712 characters. If it fails, the match fails immediately. For long strings, this
1713 approach makes a significant difference to the processing time.
1714 .
1715 .
1716 .SS "Using multiple assertions"
1717 .rs
1718 .sp
1719 Several assertions (of any sort) may occur in succession. For example,
1720 .sp
1721 (?<=\ed{3})(?<!999)foo
1722 .sp
1723 matches "foo" preceded by three digits that are not "999". Notice that each of
1724 the assertions is applied independently at the same point in the subject
1725 string. First there is a check that the previous three characters are all
1726 digits, and then there is a check that the same three characters are not "999".
1727 This pattern does \fInot\fP match "foo" preceded by six characters, the first
1728 of which are digits and the last three of which are not "999". For example, it
1729 doesn't match "123abcfoo". A pattern to do that is
1730 .sp
1731 (?<=\ed{3}...)(?<!999)foo
1732 .sp
1733 This time the first assertion looks at the preceding six characters, checking
1734 that the first three are digits, and then the second assertion checks that the
1735 preceding three characters are not "999".
1736 .P
1737 Assertions can be nested in any combination. For example,
1738 .sp
1739 (?<=(?<!foo)bar)baz
1740 .sp
1741 matches an occurrence of "baz" that is preceded by "bar" which in turn is not
1742 preceded by "foo", while
1743 .sp
1744 (?<=\ed{3}(?!999)...)foo
1745 .sp
1746 is another pattern that matches "foo" preceded by three digits and any three
1747 characters that are not "999".
1748 .
1749 .
1750 .\" HTML <a name="conditions"></a>
1752 .rs
1753 .sp
1754 It is possible to cause the matching process to obey a subpattern
1755 conditionally or to choose between two alternative subpatterns, depending on
1756 the result of an assertion, or whether a previous capturing subpattern matched
1757 or not. The two possible forms of conditional subpattern are
1758 .sp
1759 (?(condition)yes-pattern)
1760 (?(condition)yes-pattern|no-pattern)
1761 .sp
1762 If the condition is satisfied, the yes-pattern is used; otherwise the
1763 no-pattern (if present) is used. If there are more than two alternatives in the
1764 subpattern, a compile-time error occurs.
1765 .P
1766 There are four kinds of condition: references to subpatterns, references to
1767 recursion, a pseudo-condition called DEFINE, and assertions.
1768 .
1769 .SS "Checking for a used subpattern by number"
1770 .rs
1771 .sp
1772 If the text between the parentheses consists of a sequence of digits, the
1773 condition is true if the capturing subpattern of that number has previously
1774 matched. An alternative notation is to precede the digits with a plus or minus
1775 sign. In this case, the subpattern number is relative rather than absolute.
1776 The most recently opened parentheses can be referenced by (?(-1), the next most
1777 recent by (?(-2), and so on. In looping constructs it can also make sense to
1778 refer to subsequent groups with constructs such as (?(+2).
1779 .P
1780 Consider the following pattern, which contains non-significant white space to
1781 make it more readable (assume the PCRE_EXTENDED option) and to divide it into
1782 three parts for ease of discussion:
1783 .sp
1784 ( \e( )? [^()]+ (?(1) \e) )
1785 .sp
1786 The first part matches an optional opening parenthesis, and if that
1787 character is present, sets it as the first captured substring. The second part
1788 matches one or more characters that are not parentheses. The third part is a
1789 conditional subpattern that tests whether the first set of parentheses matched
1790 or not. If they did, that is, if subject started with an opening parenthesis,
1791 the condition is true, and so the yes-pattern is executed and a closing
1792 parenthesis is required. Otherwise, since no-pattern is not present, the
1793 subpattern matches nothing. In other words, this pattern matches a sequence of
1794 non-parentheses, optionally enclosed in parentheses.
1795 .P
1796 If you were embedding this pattern in a larger one, you could use a relative
1797 reference:
1798 .sp
1799 ...other stuff... ( \e( )? [^()]+ (?(-1) \e) ) ...
1800 .sp
1801 This makes the fragment independent of the parentheses in the larger pattern.
1802 .
1803 .SS "Checking for a used subpattern by name"
1804 .rs
1805 .sp
1806 Perl uses the syntax (?(<name>)...) or (?('name')...) to test for a used
1807 subpattern by name. For compatibility with earlier versions of PCRE, which had
1808 this facility before Perl, the syntax (?(name)...) is also recognized. However,
1809 there is a possible ambiguity with this syntax, because subpattern names may
1810 consist entirely of digits. PCRE looks first for a named subpattern; if it
1811 cannot find one and the name consists entirely of digits, PCRE looks for a
1812 subpattern of that number, which must be greater than zero. Using subpattern
1813 names that consist entirely of digits is not recommended.
1814 .P
1815 Rewriting the above example to use a named subpattern gives this:
1816 .sp
1817 (?<OPEN> \e( )? [^()]+ (?(<OPEN>) \e) )
1818 .sp
1819 .
1820 .SS "Checking for pattern recursion"
1821 .rs
1822 .sp
1823 If the condition is the string (R), and there is no subpattern with the name R,
1824 the condition is true if a recursive call to the whole pattern or any
1825 subpattern has been made. If digits or a name preceded by ampersand follow the
1826 letter R, for example:
1827 .sp
1828 (?(R3)...) or (?(R&name)...)
1829 .sp
1830 the condition is true if the most recent recursion is into the subpattern whose
1831 number or name is given. This condition does not check the entire recursion
1832 stack.
1833 .P
1834 At "top level", all these recursion test conditions are false. Recursive
1835 patterns are described below.
1836 .
1837 .SS "Defining subpatterns for use by reference only"
1838 .rs
1839 .sp
1840 If the condition is the string (DEFINE), and there is no subpattern with the
1841 name DEFINE, the condition is always false. In this case, there may be only one
1842 alternative in the subpattern. It is always skipped if control reaches this
1843 point in the pattern; the idea of DEFINE is that it can be used to define
1844 "subroutines" that can be referenced from elsewhere. (The use of "subroutines"
1845 is described below.) For example, a pattern to match an IPv4 address could be
1846 written like this (ignore whitespace and line breaks):
1847 .sp
1848 (?(DEFINE) (?<byte> 2[0-4]\ed | 25[0-5] | 1\ed\ed | [1-9]?\ed) )
1849 \eb (?&byte) (\e.(?&byte)){3} \eb
1850 .sp
1851 The first part of the pattern is a DEFINE group inside which a another group
1852 named "byte" is defined. This matches an individual component of an IPv4
1853 address (a number less than 256). When matching takes place, this part of the
1854 pattern is skipped because DEFINE acts like a false condition.
1855 .P
1856 The rest of the pattern uses references to the named group to match the four
1857 dot-separated components of an IPv4 address, insisting on a word boundary at
1858 each end.
1859 .
1860 .SS "Assertion conditions"
1861 .rs
1862 .sp
1863 If the condition is not in any of the above formats, it must be an assertion.
1864 This may be a positive or negative lookahead or lookbehind assertion. Consider
1865 this pattern, again containing non-significant white space, and with the two
1866 alternatives on the second line:
1867 .sp
1868 (?(?=[^a-z]*[a-z])
1869 \ed{2}-[a-z]{3}-\ed{2} | \ed{2}-\ed{2}-\ed{2} )
1870 .sp
1871 The condition is a positive lookahead assertion that matches an optional
1872 sequence of non-letters followed by a letter. In other words, it tests for the
1873 presence of at least one letter in the subject. If a letter is found, the
1874 subject is matched against the first alternative; otherwise it is matched
1875 against the second. This pattern matches strings in one of the two forms
1876 dd-aaa-dd or dd-dd-dd, where aaa are letters and dd are digits.
1877 .
1878 .
1879 .\" HTML <a name="comments"></a>
1881 .rs
1882 .sp
1883 The sequence (?# marks the start of a comment that continues up to the next
1884 closing parenthesis. Nested parentheses are not permitted. The characters
1885 that make up a comment play no part in the pattern matching at all.
1886 .P
1887 If the PCRE_EXTENDED option is set, an unescaped # character outside a
1888 character class introduces a comment that continues to immediately after the
1889 next newline in the pattern.
1890 .
1891 .
1892 .\" HTML <a name="recursion"></a>
1894 .rs
1895 .sp
1896 Consider the problem of matching a string in parentheses, allowing for
1897 unlimited nested parentheses. Without the use of recursion, the best that can
1898 be done is to use a pattern that matches up to some fixed depth of nesting. It
1899 is not possible to handle an arbitrary nesting depth.
1900 .P
1901 For some time, Perl has provided a facility that allows regular expressions to
1902 recurse (amongst other things). It does this by interpolating Perl code in the
1903 expression at run time, and the code can refer to the expression itself. A Perl
1904 pattern using code interpolation to solve the parentheses problem can be
1905 created like this:
1906 .sp
1907 $re = qr{\e( (?: (?>[^()]+) | (?p{$re}) )* \e)}x;
1908 .sp
1909 The (?p{...}) item interpolates Perl code at run time, and in this case refers
1910 recursively to the pattern in which it appears.
1911 .P
1912 Obviously, PCRE cannot support the interpolation of Perl code. Instead, it
1913 supports special syntax for recursion of the entire pattern, and also for
1914 individual subpattern recursion. After its introduction in PCRE and Python,
1915 this kind of recursion was introduced into Perl at release 5.10.
1916 .P
1917 A special item that consists of (? followed by a number greater than zero and a
1918 closing parenthesis is a recursive call of the subpattern of the given number,
1919 provided that it occurs inside that subpattern. (If not, it is a "subroutine"
1920 call, which is described in the next section.) The special item (?R) or (?0) is
1921 a recursive call of the entire regular expression.
1922 .P
1923 In PCRE (like Python, but unlike Perl), a recursive subpattern call is always
1924 treated as an atomic group. That is, once it has matched some of the subject
1925 string, it is never re-entered, even if it contains untried alternatives and
1926 there is a subsequent matching failure.
1927 .P
1928 This PCRE pattern solves the nested parentheses problem (assume the
1929 PCRE_EXTENDED option is set so that white space is ignored):
1930 .sp
1931 \e( ( (?>[^()]+) | (?R) )* \e)
1932 .sp
1933 First it matches an opening parenthesis. Then it matches any number of
1934 substrings which can either be a sequence of non-parentheses, or a recursive
1935 match of the pattern itself (that is, a correctly parenthesized substring).
1936 Finally there is a closing parenthesis.
1937 .P
1938 If this were part of a larger pattern, you would not want to recurse the entire
1939 pattern, so instead you could use this:
1940 .sp
1941 ( \e( ( (?>[^()]+) | (?1) )* \e) )
1942 .sp
1943 We have put the pattern into parentheses, and caused the recursion to refer to
1944 them instead of the whole pattern.
1945 .P
1946 In a larger pattern, keeping track of parenthesis numbers can be tricky. This
1947 is made easier by the use of relative references. (A Perl 5.10 feature.)
1948 Instead of (?1) in the pattern above you can write (?-2) to refer to the second
1949 most recently opened parentheses preceding the recursion. In other words, a
1950 negative number counts capturing parentheses leftwards from the point at which
1951 it is encountered.
1952 .P
1953 It is also possible to refer to subsequently opened parentheses, by writing
1954 references such as (?+2). However, these cannot be recursive because the
1955 reference is not inside the parentheses that are referenced. They are always
1956 "subroutine" calls, as described in the next section.
1957 .P
1958 An alternative approach is to use named parentheses instead. The Perl syntax
1959 for this is (?&name); PCRE's earlier syntax (?P>name) is also supported. We
1960 could rewrite the above example as follows:
1961 .sp
1962 (?<pn> \e( ( (?>[^()]+) | (?&pn) )* \e) )
1963 .sp
1964 If there is more than one subpattern with the same name, the earliest one is
1965 used.
1966 .P
1967 This particular example pattern that we have been looking at contains nested
1968 unlimited repeats, and so the use of atomic grouping for matching strings of
1969 non-parentheses is important when applying the pattern to strings that do not
1970 match. For example, when this pattern is applied to
1971 .sp
1972 (aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa()
1973 .sp
1974 it yields "no match" quickly. However, if atomic grouping is not used,
1975 the match runs for a very long time indeed because there are so many different
1976 ways the + and * repeats can carve up the subject, and all have to be tested
1977 before failure can be reported.
1978 .P
1979 At the end of a match, the values set for any capturing subpatterns are those
1980 from the outermost level of the recursion at which the subpattern value is set.
1981 If you want to obtain intermediate values, a callout function can be used (see
1982 below and the
1983 .\" HREF
1984 \fBpcrecallout\fP
1985 .\"
1986 documentation). If the pattern above is matched against
1987 .sp
1988 (ab(cd)ef)
1989 .sp
1990 the value for the capturing parentheses is "ef", which is the last value taken
1991 on at the top level. If additional parentheses are added, giving
1992 .sp
1993 \e( ( ( (?>[^()]+) | (?R) )* ) \e)
1994 ^ ^
1995 ^ ^
1996 .sp
1997 the string they capture is "ab(cd)ef", the contents of the top level
1998 parentheses. If there are more than 15 capturing parentheses in a pattern, PCRE
1999 has to obtain extra memory to store data during a recursion, which it does by
2000 using \fBpcre_malloc\fP, freeing it via \fBpcre_free\fP afterwards. If no
2001 memory can be obtained, the match fails with the PCRE_ERROR_NOMEMORY error.
2002 .P
2003 Do not confuse the (?R) item with the condition (R), which tests for recursion.
2004 Consider this pattern, which matches text in angle brackets, allowing for
2005 arbitrary nesting. Only digits are allowed in nested brackets (that is, when
2006 recursing), whereas any characters are permitted at the outer level.
2007 .sp
2008 < (?: (?(R) \ed++ | [^<>]*+) | (?R)) * >
2009 .sp
2010 In this pattern, (?(R) is the start of a conditional subpattern, with two
2011 different alternatives for the recursive and non-recursive cases. The (?R) item
2012 is the actual recursive call.
2013 .
2014 .
2015 .\" HTML <a name="subpatternsassubroutines"></a>
2017 .rs
2018 .sp
2019 If the syntax for a recursive subpattern reference (either by number or by
2020 name) is used outside the parentheses to which it refers, it operates like a
2021 subroutine in a programming language. The "called" subpattern may be defined
2022 before or after the reference. A numbered reference can be absolute or
2023 relative, as in these examples:
2024 .sp
2025 (...(absolute)...)...(?2)...
2026 (...(relative)...)...(?-1)...
2027 (...(?+1)...(relative)...
2028 .sp
2029 An earlier example pointed out that the pattern
2030 .sp
2031 (sens|respons)e and \e1ibility
2032 .sp
2033 matches "sense and sensibility" and "response and responsibility", but not
2034 "sense and responsibility". If instead the pattern
2035 .sp
2036 (sens|respons)e and (?1)ibility
2037 .sp
2038 is used, it does match "sense and responsibility" as well as the other two
2039 strings. Another example is given in the discussion of DEFINE above.
2040 .P
2041 Like recursive subpatterns, a "subroutine" call is always treated as an atomic
2042 group. That is, once it has matched some of the subject string, it is never
2043 re-entered, even if it contains untried alternatives and there is a subsequent
2044 matching failure.
2045 .P
2046 When a subpattern is used as a subroutine, processing options such as
2047 case-independence are fixed when the subpattern is defined. They cannot be
2048 changed for different calls. For example, consider this pattern:
2049 .sp
2050 (abc)(?i:(?-1))
2051 .sp
2052 It matches "abcabc". It does not match "abcABC" because the change of
2053 processing option does not affect the called subpattern.
2054 .
2055 .
2056 .\" HTML <a name="onigurumasubroutines"></a>
2058 .rs
2059 .sp
2060 For compatibility with Oniguruma, the non-Perl syntax \eg followed by a name or
2061 a number enclosed either in angle brackets or single quotes, is an alternative
2062 syntax for referencing a subpattern as a subroutine, possibly recursively. Here
2063 are two of the examples used above, rewritten using this syntax:
2064 .sp
2065 (?<pn> \e( ( (?>[^()]+) | \eg<pn> )* \e) )
2066 (sens|respons)e and \eg'1'ibility
2067 .sp
2068 PCRE supports an extension to Oniguruma: if a number is preceded by a
2069 plus or a minus sign it is taken as a relative reference. For example:
2070 .sp
2071 (abc)(?i:\eg<-1>)
2072 .sp
2073 Note that \eg{...} (Perl syntax) and \eg<...> (Oniguruma syntax) are \fInot\fP
2074 synonymous. The former is a back reference; the latter is a subroutine call.
2075 .
2076 .
2078 .rs
2079 .sp
2080 Perl has a feature whereby using the sequence (?{...}) causes arbitrary Perl
2081 code to be obeyed in the middle of matching a regular expression. This makes it
2082 possible, amongst other things, to extract different substrings that match the
2083 same pair of parentheses when there is a repetition.
2084 .P
2085 PCRE provides a similar feature, but of course it cannot obey arbitrary Perl
2086 code. The feature is called "callout". The caller of PCRE provides an external
2087 function by putting its entry point in the global variable \fIpcre_callout\fP.
2088 By default, this variable contains NULL, which disables all calling out.
2089 .P
2090 Within a regular expression, (?C) indicates the points at which the external
2091 function is to be called. If you want to identify different callout points, you
2092 can put a number less than 256 after the letter C. The default value is zero.
2093 For example, this pattern has two callout points:
2094 .sp
2095 (?C1)abc(?C2)def
2096 .sp
2097 If the PCRE_AUTO_CALLOUT flag is passed to \fBpcre_compile()\fP, callouts are
2098 automatically installed before each item in the pattern. They are all numbered
2099 255.
2100 .P
2101 During matching, when PCRE reaches a callout point (and \fIpcre_callout\fP is
2102 set), the external function is called. It is provided with the number of the
2103 callout, the position in the pattern, and, optionally, one item of data
2104 originally supplied by the caller of \fBpcre_exec()\fP. The callout function
2105 may cause matching to proceed, to backtrack, or to fail altogether. A complete
2106 description of the interface to the callout function is given in the
2107 .\" HREF
2108 \fBpcrecallout\fP
2109 .\"
2110 documentation.
2111 .
2112 .
2114 .rs
2115 .sp
2116 Perl 5.10 introduced a number of "Special Backtracking Control Verbs", which
2117 are described in the Perl documentation as "experimental and subject to change
2118 or removal in a future version of Perl". It goes on to say: "Their usage in
2119 production code should be noted to avoid problems during upgrades." The same
2120 remarks apply to the PCRE features described in this section.
2121 .P
2122 Since these verbs are specifically related to backtracking, most of them can be
2123 used only when the pattern is to be matched using \fBpcre_exec()\fP, which uses
2124 a backtracking algorithm. With the exception of (*FAIL), which behaves like a
2125 failing negative assertion, they cause an error if encountered by
2126 \fBpcre_dfa_exec()\fP.
2127 .P
2128 The new verbs make use of what was previously invalid syntax: an opening
2129 parenthesis followed by an asterisk. In Perl, they are generally of the form
2130 (*VERB:ARG) but PCRE does not support the use of arguments, so its general
2131 form is just (*VERB). Any number of these verbs may occur in a pattern. There
2132 are two kinds:
2133 .
2134 .SS "Verbs that act immediately"
2135 .rs
2136 .sp
2137 The following verbs act as soon as they are encountered:
2138 .sp
2139 (*ACCEPT)
2140 .sp
2141 This verb causes the match to end successfully, skipping the remainder of the
2142 pattern. When inside a recursion, only the innermost pattern is ended
2143 immediately. PCRE differs from Perl in what happens if the (*ACCEPT) is inside
2144 capturing parentheses. In Perl, the data so far is captured: in PCRE no data is
2145 captured. For example:
2146 .sp
2147 A(A|B(*ACCEPT)|C)D
2148 .sp
2149 This matches "AB", "AAD", or "ACD", but when it matches "AB", no data is
2150 captured.
2151 .sp
2152 (*FAIL) or (*F)
2153 .sp
2154 This verb causes the match to fail, forcing backtracking to occur. It is
2155 equivalent to (?!) but easier to read. The Perl documentation notes that it is
2156 probably useful only when combined with (?{}) or (??{}). Those are, of course,
2157 Perl features that are not present in PCRE. The nearest equivalent is the
2158 callout feature, as for example in this pattern:
2159 .sp
2160 a+(?C)(*FAIL)
2161 .sp
2162 A match with the string "aaaa" always fails, but the callout is taken before
2163 each backtrack happens (in this example, 10 times).
2164 .
2165 .SS "Verbs that act after backtracking"
2166 .rs
2167 .sp
2168 The following verbs do nothing when they are encountered. Matching continues
2169 with what follows, but if there is no subsequent match, a failure is forced.
2170 The verbs differ in exactly what kind of failure occurs.
2171 .sp
2172 (*COMMIT)
2173 .sp
2174 This verb causes the whole match to fail outright if the rest of the pattern
2175 does not match. Even if the pattern is unanchored, no further attempts to find
2176 a match by advancing the start point take place. Once (*COMMIT) has been
2177 passed, \fBpcre_exec()\fP is committed to finding a match at the current
2178 starting point, or not at all. For example:
2179 .sp
2180 a+(*COMMIT)b
2181 .sp
2182 This matches "xxaab" but not "aacaab". It can be thought of as a kind of
2183 dynamic anchor, or "I've started, so I must finish."
2184 .sp
2185 (*PRUNE)
2186 .sp
2187 This verb causes the match to fail at the current position if the rest of the
2188 pattern does not match. If the pattern is unanchored, the normal "bumpalong"
2189 advance to the next starting character then happens. Backtracking can occur as
2190 usual to the left of (*PRUNE), or when matching to the right of (*PRUNE), but
2191 if there is no match to the right, backtracking cannot cross (*PRUNE).
2192 In simple cases, the use of (*PRUNE) is just an alternative to an atomic
2193 group or possessive quantifier, but there are some uses of (*PRUNE) that cannot
2194 be expressed in any other way.
2195 .sp
2196 (*SKIP)
2197 .sp
2198 This verb is like (*PRUNE), except that if the pattern is unanchored, the
2199 "bumpalong" advance is not to the next character, but to the position in the
2200 subject where (*SKIP) was encountered. (*SKIP) signifies that whatever text
2201 was matched leading up to it cannot be part of a successful match. Consider:
2202 .sp
2203 a+(*SKIP)b
2204 .sp
2205 If the subject is "aaaac...", after the first match attempt fails (starting at
2206 the first character in the string), the starting point skips on to start the
2207 next attempt at "c". Note that a possessive quantifer does not have the same
2208 effect in this example; although it would suppress backtracking during the
2209 first match attempt, the second attempt would start at the second character
2210 instead of skipping on to "c".
2211 .sp
2212 (*THEN)
2213 .sp
2214 This verb causes a skip to the next alternation if the rest of the pattern does
2215 not match. That is, it cancels pending backtracking, but only within the
2216 current alternation. Its name comes from the observation that it can be used
2217 for a pattern-based if-then-else block:
2218 .sp
2219 ( COND1 (*THEN) FOO | COND2 (*THEN) BAR | COND3 (*THEN) BAZ ) ...
2220 .sp
2221 If the COND1 pattern matches, FOO is tried (and possibly further items after
2222 the end of the group if FOO succeeds); on failure the matcher skips to the
2223 second alternative and tries COND2, without backtracking into COND1. If (*THEN)
2224 is used outside of any alternation, it acts exactly like (*PRUNE).
2225 .
2226 .
2227 .SH "SEE ALSO"
2228 .rs
2229 .sp
2230 \fBpcreapi\fP(3), \fBpcrecallout\fP(3), \fBpcrematching\fP(3), \fBpcre\fP(3).
2231 .
2232 .
2234 .rs
2235 .sp
2236 .nf
2237 Philip Hazel
2238 University Computing Service
2239 Cambridge CB2 3QH, England.
2240 .fi
2241 .
2242 .
2244 .rs
2245 .sp
2246 .nf
2247 Last updated: 08 March 2009
2248 Copyright (c) 1997-2009 University of Cambridge.
2249 .fi


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

  ViewVC Help
Powered by ViewVC 1.1.5