ViewVC logotype

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

Parent Directory Parent Directory | Revision Log Revision Log

Revision 412 - (show annotations)
Sat Apr 11 10:34:37 2009 UTC (12 years, 6 months ago) by ph10
File size: 92046 byte(s)
Error occurred while calculating annotation data.
Add support for (*UTF8).
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. There is also a special sequence that can be given at the
27 start of a pattern:
28 .sp
29 (*UTF8)
30 .sp
31 Starting a pattern with this sequence is equivalent to setting the PCRE_UTF8
32 option. This feature is not Perl-compatible. How setting UTF-8 mode affects
33 pattern matching is mentioned in several places below. There is also a summary
34 of UTF-8 features in the
35 .\" HTML <a href="pcre.html#utf8support">
36 .\" </a>
37 section on UTF-8 support
38 .\"
39 in the main
40 .\" HREF
41 \fBpcre\fP
42 .\"
43 page.
44 .P
45 The remainder of this document discusses the patterns that are supported by
46 PCRE when its main matching function, \fBpcre_exec()\fP, is used.
47 From release 6.0, PCRE offers a second matching function,
48 \fBpcre_dfa_exec()\fP, which matches using a different algorithm that is not
49 Perl-compatible. Some of the features discussed below are not available when
50 \fBpcre_dfa_exec()\fP is used. The advantages and disadvantages of the
51 alternative function, and how it differs from the normal function, are
52 discussed in the
53 .\" HREF
54 \fBpcrematching\fP
55 .\"
56 page.
57 .
58 .
60 .rs
61 .sp
62 PCRE supports five different conventions for indicating line breaks in
63 strings: a single CR (carriage return) character, a single LF (linefeed)
64 character, the two-character sequence CRLF, any of the three preceding, or any
65 Unicode newline sequence. The
66 .\" HREF
67 \fBpcreapi\fP
68 .\"
69 page has
70 .\" HTML <a href="pcreapi.html#newlines">
71 .\" </a>
72 further discussion
73 .\"
74 about newlines, and shows how to set the newline convention in the
75 \fIoptions\fP arguments for the compiling and matching functions.
76 .P
77 It is also possible to specify a newline convention by starting a pattern
78 string with one of the following five sequences:
79 .sp
80 (*CR) carriage return
81 (*LF) linefeed
82 (*CRLF) carriage return, followed by linefeed
83 (*ANYCRLF) any of the three above
84 (*ANY) all Unicode newline sequences
85 .sp
86 These override the default and the options given to \fBpcre_compile()\fP. For
87 example, on a Unix system where LF is the default newline sequence, the pattern
88 .sp
89 (*CR)a.b
90 .sp
91 changes the convention to CR. That pattern matches "a\enb" because LF is no
92 longer a newline. Note that these special settings, which are not
93 Perl-compatible, are recognized only at the very start of a pattern, and that
94 they must be in upper case. If more than one of them is present, the last one
95 is used.
96 .P
97 The newline convention does not affect what the \eR escape sequence matches. By
98 default, this is any Unicode newline sequence, for Perl compatibility. However,
99 this can be changed; see the description of \eR in the section entitled
100 .\" HTML <a href="#newlineseq">
101 .\" </a>
102 "Newline sequences"
103 .\"
104 below. A change of \eR setting can be combined with a change of newline
105 convention.
106 .
107 .
109 .rs
110 .sp
111 A regular expression is a pattern that is matched against a subject string from
112 left to right. Most characters stand for themselves in a pattern, and match the
113 corresponding characters in the subject. As a trivial example, the pattern
114 .sp
115 The quick brown fox
116 .sp
117 matches a portion of a subject string that is identical to itself. When
118 caseless matching is specified (the PCRE_CASELESS option), letters are matched
119 independently of case. In UTF-8 mode, PCRE always understands the concept of
120 case for characters whose values are less than 128, so caseless matching is
121 always possible. For characters with higher values, the concept of case is
122 supported if PCRE is compiled with Unicode property support, but not otherwise.
123 If you want to use caseless matching for characters 128 and above, you must
124 ensure that PCRE is compiled with Unicode property support as well as with
125 UTF-8 support.
126 .P
127 The power of regular expressions comes from the ability to include alternatives
128 and repetitions in the pattern. These are encoded in the pattern by the use of
129 \fImetacharacters\fP, which do not stand for themselves but instead are
130 interpreted in some special way.
131 .P
132 There are two different sets of metacharacters: those that are recognized
133 anywhere in the pattern except within square brackets, and those that are
134 recognized within square brackets. Outside square brackets, the metacharacters
135 are as follows:
136 .sp
137 \e general escape character with several uses
138 ^ assert start of string (or line, in multiline mode)
139 $ assert end of string (or line, in multiline mode)
140 . match any character except newline (by default)
141 [ start character class definition
142 | start of alternative branch
143 ( start subpattern
144 ) end subpattern
145 ? extends the meaning of (
146 also 0 or 1 quantifier
147 also quantifier minimizer
148 * 0 or more quantifier
149 + 1 or more quantifier
150 also "possessive quantifier"
151 { start min/max quantifier
152 .sp
153 Part of a pattern that is in square brackets is called a "character class". In
154 a character class the only metacharacters are:
155 .sp
156 \e general escape character
157 ^ negate the class, but only if the first character
158 - indicates character range
159 .\" JOIN
160 [ POSIX character class (only if followed by POSIX
161 syntax)
162 ] terminates the character class
163 .sp
164 The following sections describe the use of each of the metacharacters.
165 .
166 .
168 .rs
169 .sp
170 The backslash character has several uses. Firstly, if it is followed by a
171 non-alphanumeric character, it takes away any special meaning that character
172 may have. This use of backslash as an escape character applies both inside and
173 outside character classes.
174 .P
175 For example, if you want to match a * character, you write \e* in the pattern.
176 This escaping action applies whether or not the following character would
177 otherwise be interpreted as a metacharacter, so it is always safe to precede a
178 non-alphanumeric with backslash to specify that it stands for itself. In
179 particular, if you want to match a backslash, you write \e\e.
180 .P
181 If a pattern is compiled with the PCRE_EXTENDED option, whitespace in the
182 pattern (other than in a character class) and characters between a # outside
183 a character class and the next newline are ignored. An escaping backslash can
184 be used to include a whitespace or # character as part of the pattern.
185 .P
186 If you want to remove the special meaning from a sequence of characters, you
187 can do so by putting them between \eQ and \eE. This is different from Perl in
188 that $ and @ are handled as literals in \eQ...\eE sequences in PCRE, whereas in
189 Perl, $ and @ cause variable interpolation. Note the following examples:
190 .sp
191 Pattern PCRE matches Perl matches
192 .sp
193 .\" JOIN
194 \eQabc$xyz\eE abc$xyz abc followed by the
195 contents of $xyz
196 \eQabc\e$xyz\eE abc\e$xyz abc\e$xyz
197 \eQabc\eE\e$\eQxyz\eE abc$xyz abc$xyz
198 .sp
199 The \eQ...\eE sequence is recognized both inside and outside character classes.
200 .
201 .
202 .\" HTML <a name="digitsafterbackslash"></a>
203 .SS "Non-printing characters"
204 .rs
205 .sp
206 A second use of backslash provides a way of encoding non-printing characters
207 in patterns in a visible manner. There is no restriction on the appearance of
208 non-printing characters, apart from the binary zero that terminates a pattern,
209 but when a pattern is being prepared by text editing, it is usually easier to
210 use one of the following escape sequences than the binary character it
211 represents:
212 .sp
213 \ea alarm, that is, the BEL character (hex 07)
214 \ecx "control-x", where x is any character
215 \ee escape (hex 1B)
216 \ef formfeed (hex 0C)
217 \en linefeed (hex 0A)
218 \er carriage return (hex 0D)
219 \et tab (hex 09)
220 \eddd character with octal code ddd, or backreference
221 \exhh character with hex code hh
222 \ex{hhh..} character with hex code hhh..
223 .sp
224 The precise effect of \ecx is as follows: if x is a lower case letter, it
225 is converted to upper case. Then bit 6 of the character (hex 40) is inverted.
226 Thus \ecz becomes hex 1A, but \ec{ becomes hex 3B, while \ec; becomes hex
227 7B.
228 .P
229 After \ex, from zero to two hexadecimal digits are read (letters can be in
230 upper or lower case). Any number of hexadecimal digits may appear between \ex{
231 and }, but the value of the character code must be less than 256 in non-UTF-8
232 mode, and less than 2**31 in UTF-8 mode. That is, the maximum value in
233 hexadecimal is 7FFFFFFF. Note that this is bigger than the largest Unicode code
234 point, which is 10FFFF.
235 .P
236 If characters other than hexadecimal digits appear between \ex{ and }, or if
237 there is no terminating }, this form of escape is not recognized. Instead, the
238 initial \ex will be interpreted as a basic hexadecimal escape, with no
239 following digits, giving a character whose value is zero.
240 .P
241 Characters whose value is less than 256 can be defined by either of the two
242 syntaxes for \ex. There is no difference in the way they are handled. For
243 example, \exdc is exactly the same as \ex{dc}.
244 .P
245 After \e0 up to two further octal digits are read. If there are fewer than two
246 digits, just those that are present are used. Thus the sequence \e0\ex\e07
247 specifies two binary zeros followed by a BEL character (code value 7). Make
248 sure you supply two digits after the initial zero if the pattern character that
249 follows is itself an octal digit.
250 .P
251 The handling of a backslash followed by a digit other than 0 is complicated.
252 Outside a character class, PCRE reads it and any following digits as a decimal
253 number. If the number is less than 10, or if there have been at least that many
254 previous capturing left parentheses in the expression, the entire sequence is
255 taken as a \fIback reference\fP. A description of how this works is given
256 .\" HTML <a href="#backreferences">
257 .\" </a>
258 later,
259 .\"
260 following the discussion of
261 .\" HTML <a href="#subpattern">
262 .\" </a>
263 parenthesized subpatterns.
264 .\"
265 .P
266 Inside a character class, or if the decimal number is greater than 9 and there
267 have not been that many capturing subpatterns, PCRE re-reads up to three octal
268 digits following the backslash, and uses them to generate a data character. Any
269 subsequent digits stand for themselves. In non-UTF-8 mode, the value of a
270 character specified in octal must be less than \e400. In UTF-8 mode, values up
271 to \e777 are permitted. For example:
272 .sp
273 \e040 is another way of writing a space
274 .\" JOIN
275 \e40 is the same, provided there are fewer than 40
276 previous capturing subpatterns
277 \e7 is always a back reference
278 .\" JOIN
279 \e11 might be a back reference, or another way of
280 writing a tab
281 \e011 is always a tab
282 \e0113 is a tab followed by the character "3"
283 .\" JOIN
284 \e113 might be a back reference, otherwise the
285 character with octal code 113
286 .\" JOIN
287 \e377 might be a back reference, otherwise
288 the byte consisting entirely of 1 bits
289 .\" JOIN
290 \e81 is either a back reference, or a binary zero
291 followed by the two characters "8" and "1"
292 .sp
293 Note that octal values of 100 or greater must not be introduced by a leading
294 zero, because no more than three octal digits are ever read.
295 .P
296 All the sequences that define a single character value can be used both inside
297 and outside character classes. In addition, inside a character class, the
298 sequence \eb is interpreted as the backspace character (hex 08), and the
299 sequences \eR and \eX are interpreted as the characters "R" and "X",
300 respectively. Outside a character class, these sequences have different
301 meanings
302 .\" HTML <a href="#uniextseq">
303 .\" </a>
304 (see below).
305 .\"
306 .
307 .
308 .SS "Absolute and relative back references"
309 .rs
310 .sp
311 The sequence \eg followed by an unsigned or a negative number, optionally
312 enclosed in braces, is an absolute or relative back reference. A named back
313 reference can be coded as \eg{name}. Back references are discussed
314 .\" HTML <a href="#backreferences">
315 .\" </a>
316 later,
317 .\"
318 following the discussion of
319 .\" HTML <a href="#subpattern">
320 .\" </a>
321 parenthesized subpatterns.
322 .\"
323 .
324 .
325 .SS "Absolute and relative subroutine calls"
326 .rs
327 .sp
328 For compatibility with Oniguruma, the non-Perl syntax \eg followed by a name or
329 a number enclosed either in angle brackets or single quotes, is an alternative
330 syntax for referencing a subpattern as a "subroutine". Details are discussed
331 .\" HTML <a href="#onigurumasubroutines">
332 .\" </a>
333 later.
334 .\"
335 Note that \eg{...} (Perl syntax) and \eg<...> (Oniguruma syntax) are \fInot\fP
336 synonymous. The former is a back reference; the latter is a subroutine call.
337 .
338 .
339 .SS "Generic character types"
340 .rs
341 .sp
342 Another use of backslash is for specifying generic character types. The
343 following are always recognized:
344 .sp
345 \ed any decimal digit
346 \eD any character that is not a decimal digit
347 \eh any horizontal whitespace character
348 \eH any character that is not a horizontal whitespace character
349 \es any whitespace character
350 \eS any character that is not a whitespace character
351 \ev any vertical whitespace character
352 \eV any character that is not a vertical whitespace character
353 \ew any "word" character
354 \eW any "non-word" character
355 .sp
356 Each pair of escape sequences partitions the complete set of characters into
357 two disjoint sets. Any given character matches one, and only one, of each pair.
358 .P
359 These character type sequences can appear both inside and outside character
360 classes. They each match one character of the appropriate type. If the current
361 matching point is at the end of the subject string, all of them fail, since
362 there is no character to match.
363 .P
364 For compatibility with Perl, \es does not match the VT character (code 11).
365 This makes it different from the the POSIX "space" class. The \es characters
366 are HT (9), LF (10), FF (12), CR (13), and space (32). If "use locale;" is
367 included in a Perl script, \es may match the VT character. In PCRE, it never
368 does.
369 .P
370 In UTF-8 mode, characters with values greater than 128 never match \ed, \es, or
371 \ew, and always match \eD, \eS, and \eW. This is true even when Unicode
372 character property support is available. These sequences retain their original
373 meanings from before UTF-8 support was available, mainly for efficiency
374 reasons. Note that this also affects \eb, because it is defined in terms of \ew
375 and \eW.
376 .P
377 The sequences \eh, \eH, \ev, and \eV are Perl 5.10 features. In contrast to the
378 other sequences, these do match certain high-valued codepoints in UTF-8 mode.
379 The horizontal space characters are:
380 .sp
381 U+0009 Horizontal tab
382 U+0020 Space
383 U+00A0 Non-break space
384 U+1680 Ogham space mark
385 U+180E Mongolian vowel separator
386 U+2000 En quad
387 U+2001 Em quad
388 U+2002 En space
389 U+2003 Em space
390 U+2004 Three-per-em space
391 U+2005 Four-per-em space
392 U+2006 Six-per-em space
393 U+2007 Figure space
394 U+2008 Punctuation space
395 U+2009 Thin space
396 U+200A Hair space
397 U+202F Narrow no-break space
398 U+205F Medium mathematical space
399 U+3000 Ideographic space
400 .sp
401 The vertical space characters are:
402 .sp
403 U+000A Linefeed
404 U+000B Vertical tab
405 U+000C Formfeed
406 U+000D Carriage return
407 U+0085 Next line
408 U+2028 Line separator
409 U+2029 Paragraph separator
410 .P
411 A "word" character is an underscore or any character less than 256 that is a
412 letter or digit. The definition of letters and digits is controlled by PCRE's
413 low-valued character tables, and may vary if locale-specific matching is taking
414 place (see
415 .\" HTML <a href="pcreapi.html#localesupport">
416 .\" </a>
417 "Locale support"
418 .\"
419 in the
420 .\" HREF
421 \fBpcreapi\fP
422 .\"
423 page). For example, in a French locale such as "fr_FR" in Unix-like systems,
424 or "french" in Windows, some character codes greater than 128 are used for
425 accented letters, and these are matched by \ew. The use of locales with Unicode
426 is discouraged.
427 .
428 .
429 .\" HTML <a name="newlineseq"></a>
430 .SS "Newline sequences"
431 .rs
432 .sp
433 Outside a character class, by default, the escape sequence \eR matches any
434 Unicode newline sequence. This is a Perl 5.10 feature. In non-UTF-8 mode \eR is
435 equivalent to the following:
436 .sp
437 (?>\er\en|\en|\ex0b|\ef|\er|\ex85)
438 .sp
439 This is an example of an "atomic group", details of which are given
440 .\" HTML <a href="#atomicgroup">
441 .\" </a>
442 below.
443 .\"
444 This particular group matches either the two-character sequence CR followed by
445 LF, or one of the single characters LF (linefeed, U+000A), VT (vertical tab,
446 U+000B), FF (formfeed, U+000C), CR (carriage return, U+000D), or NEL (next
447 line, U+0085). The two-character sequence is treated as a single unit that
448 cannot be split.
449 .P
450 In UTF-8 mode, two additional characters whose codepoints are greater than 255
451 are added: LS (line separator, U+2028) and PS (paragraph separator, U+2029).
452 Unicode character property support is not needed for these characters to be
453 recognized.
454 .P
455 It is possible to restrict \eR to match only CR, LF, or CRLF (instead of the
456 complete set of Unicode line endings) by setting the option PCRE_BSR_ANYCRLF
457 either at compile time or when the pattern is matched. (BSR is an abbrevation
458 for "backslash R".) This can be made the default when PCRE is built; if this is
459 the case, the other behaviour can be requested via the PCRE_BSR_UNICODE option.
460 It is also possible to specify these settings by starting a pattern string with
461 one of the following sequences:
462 .sp
463 (*BSR_ANYCRLF) CR, LF, or CRLF only
464 (*BSR_UNICODE) any Unicode newline sequence
465 .sp
466 These override the default and the options given to \fBpcre_compile()\fP, but
467 they can be overridden by options given to \fBpcre_exec()\fP. Note that these
468 special settings, which are not Perl-compatible, are recognized only at the
469 very start of a pattern, and that they must be in upper case. If more than one
470 of them is present, the last one is used. They can be combined with a change of
471 newline convention, for example, a pattern can start with:
472 .sp
474 .sp
475 Inside a character class, \eR matches the letter "R".
476 .
477 .
478 .\" HTML <a name="uniextseq"></a>
479 .SS Unicode character properties
480 .rs
481 .sp
482 When PCRE is built with Unicode character property support, three additional
483 escape sequences that match characters with specific properties are available.
484 When not in UTF-8 mode, these sequences are of course limited to testing
485 characters whose codepoints are less than 256, but they do work in this mode.
486 The extra escape sequences are:
487 .sp
488 \ep{\fIxx\fP} a character with the \fIxx\fP property
489 \eP{\fIxx\fP} a character without the \fIxx\fP property
490 \eX an extended Unicode sequence
491 .sp
492 The property names represented by \fIxx\fP above are limited to the Unicode
493 script names, the general category properties, and "Any", which matches any
494 character (including newline). Other properties such as "InMusicalSymbols" are
495 not currently supported by PCRE. Note that \eP{Any} does not match any
496 characters, so always causes a match failure.
497 .P
498 Sets of Unicode characters are defined as belonging to certain scripts. A
499 character from one of these sets can be matched using a script name. For
500 example:
501 .sp
502 \ep{Greek}
503 \eP{Han}
504 .sp
505 Those that are not part of an identified script are lumped together as
506 "Common". The current list of scripts is:
507 .P
508 Arabic,
509 Armenian,
510 Balinese,
511 Bengali,
512 Bopomofo,
513 Braille,
514 Buginese,
515 Buhid,
516 Canadian_Aboriginal,
517 Cherokee,
518 Common,
519 Coptic,
520 Cuneiform,
521 Cypriot,
522 Cyrillic,
523 Deseret,
524 Devanagari,
525 Ethiopic,
526 Georgian,
527 Glagolitic,
528 Gothic,
529 Greek,
530 Gujarati,
531 Gurmukhi,
532 Han,
533 Hangul,
534 Hanunoo,
535 Hebrew,
536 Hiragana,
537 Inherited,
538 Kannada,
539 Katakana,
540 Kharoshthi,
541 Khmer,
542 Lao,
543 Latin,
544 Limbu,
545 Linear_B,
546 Malayalam,
547 Mongolian,
548 Myanmar,
549 New_Tai_Lue,
550 Nko,
551 Ogham,
552 Old_Italic,
553 Old_Persian,
554 Oriya,
555 Osmanya,
556 Phags_Pa,
557 Phoenician,
558 Runic,
559 Shavian,
560 Sinhala,
561 Syloti_Nagri,
562 Syriac,
563 Tagalog,
564 Tagbanwa,
565 Tai_Le,
566 Tamil,
567 Telugu,
568 Thaana,
569 Thai,
570 Tibetan,
571 Tifinagh,
572 Ugaritic,
573 Yi.
574 .P
575 Each character has exactly one general category property, specified by a
576 two-letter abbreviation. For compatibility with Perl, negation can be specified
577 by including a circumflex between the opening brace and the property name. For
578 example, \ep{^Lu} is the same as \eP{Lu}.
579 .P
580 If only one letter is specified with \ep or \eP, it includes all the general
581 category properties that start with that letter. In this case, in the absence
582 of negation, the curly brackets in the escape sequence are optional; these two
583 examples have the same effect:
584 .sp
585 \ep{L}
586 \epL
587 .sp
588 The following general category property codes are supported:
589 .sp
590 C Other
591 Cc Control
592 Cf Format
593 Cn Unassigned
594 Co Private use
595 Cs Surrogate
596 .sp
597 L Letter
598 Ll Lower case letter
599 Lm Modifier letter
600 Lo Other letter
601 Lt Title case letter
602 Lu Upper case letter
603 .sp
604 M Mark
605 Mc Spacing mark
606 Me Enclosing mark
607 Mn Non-spacing mark
608 .sp
609 N Number
610 Nd Decimal number
611 Nl Letter number
612 No Other number
613 .sp
614 P Punctuation
615 Pc Connector punctuation
616 Pd Dash punctuation
617 Pe Close punctuation
618 Pf Final punctuation
619 Pi Initial punctuation
620 Po Other punctuation
621 Ps Open punctuation
622 .sp
623 S Symbol
624 Sc Currency symbol
625 Sk Modifier symbol
626 Sm Mathematical symbol
627 So Other symbol
628 .sp
629 Z Separator
630 Zl Line separator
631 Zp Paragraph separator
632 Zs Space separator
633 .sp
634 The special property L& is also supported: it matches a character that has
635 the Lu, Ll, or Lt property, in other words, a letter that is not classified as
636 a modifier or "other".
637 .P
638 The Cs (Surrogate) property applies only to characters in the range U+D800 to
639 U+DFFF. Such characters are not valid in UTF-8 strings (see RFC 3629) and so
640 cannot be tested by PCRE, unless UTF-8 validity checking has been turned off
641 (see the discussion of PCRE_NO_UTF8_CHECK in the
642 .\" HREF
643 \fBpcreapi\fP
644 .\"
645 page).
646 .P
647 The long synonyms for these properties that Perl supports (such as \ep{Letter})
648 are not supported by PCRE, nor is it permitted to prefix any of these
649 properties with "Is".
650 .P
651 No character that is in the Unicode table has the Cn (unassigned) property.
652 Instead, this property is assumed for any code point that is not in the
653 Unicode table.
654 .P
655 Specifying caseless matching does not affect these escape sequences. For
656 example, \ep{Lu} always matches only upper case letters.
657 .P
658 The \eX escape matches any number of Unicode characters that form an extended
659 Unicode sequence. \eX is equivalent to
660 .sp
661 (?>\ePM\epM*)
662 .sp
663 That is, it matches a character without the "mark" property, followed by zero
664 or more characters with the "mark" property, and treats the sequence as an
665 atomic group
666 .\" HTML <a href="#atomicgroup">
667 .\" </a>
668 (see below).
669 .\"
670 Characters with the "mark" property are typically accents that affect the
671 preceding character. None of them have codepoints less than 256, so in
672 non-UTF-8 mode \eX matches any one character.
673 .P
674 Matching characters by Unicode property is not fast, because PCRE has to search
675 a structure that contains data for over fifteen thousand characters. That is
676 why the traditional escape sequences such as \ed and \ew do not use Unicode
677 properties in PCRE.
678 .
679 .
680 .\" HTML <a name="resetmatchstart"></a>
681 .SS "Resetting the match start"
682 .rs
683 .sp
684 The escape sequence \eK, which is a Perl 5.10 feature, causes any previously
685 matched characters not to be included in the final matched sequence. For
686 example, the pattern:
687 .sp
688 foo\eKbar
689 .sp
690 matches "foobar", but reports that it has matched "bar". This feature is
691 similar to a lookbehind assertion
692 .\" HTML <a href="#lookbehind">
693 .\" </a>
694 (described below).
695 .\"
696 However, in this case, the part of the subject before the real match does not
697 have to be of fixed length, as lookbehind assertions do. The use of \eK does
698 not interfere with the setting of
699 .\" HTML <a href="#subpattern">
700 .\" </a>
701 captured substrings.
702 .\"
703 For example, when the pattern
704 .sp
705 (foo)\eKbar
706 .sp
707 matches "foobar", the first substring is still set to "foo".
708 .
709 .
710 .\" HTML <a name="smallassertions"></a>
711 .SS "Simple assertions"
712 .rs
713 .sp
714 The final use of backslash is for certain simple assertions. An assertion
715 specifies a condition that has to be met at a particular point in a match,
716 without consuming any characters from the subject string. The use of
717 subpatterns for more complicated assertions is described
718 .\" HTML <a href="#bigassertions">
719 .\" </a>
720 below.
721 .\"
722 The backslashed assertions are:
723 .sp
724 \eb matches at a word boundary
725 \eB matches when not at a word boundary
726 \eA matches at the start of the subject
727 \eZ matches at the end of the subject
728 also matches before a newline at the end of the subject
729 \ez matches only at the end of the subject
730 \eG matches at the first matching position in the subject
731 .sp
732 These assertions may not appear in character classes (but note that \eb has a
733 different meaning, namely the backspace character, inside a character class).
734 .P
735 A word boundary is a position in the subject string where the current character
736 and the previous character do not both match \ew or \eW (i.e. one matches
737 \ew and the other matches \eW), or the start or end of the string if the
738 first or last character matches \ew, respectively.
739 .P
740 The \eA, \eZ, and \ez assertions differ from the traditional circumflex and
741 dollar (described in the next section) in that they only ever match at the very
742 start and end of the subject string, whatever options are set. Thus, they are
743 independent of multiline mode. These three assertions are not affected by the
744 PCRE_NOTBOL or PCRE_NOTEOL options, which affect only the behaviour of the
745 circumflex and dollar metacharacters. However, if the \fIstartoffset\fP
746 argument of \fBpcre_exec()\fP is non-zero, indicating that matching is to start
747 at a point other than the beginning of the subject, \eA can never match. The
748 difference between \eZ and \ez is that \eZ matches before a newline at the end
749 of the string as well as at the very end, whereas \ez matches only at the end.
750 .P
751 The \eG assertion is true only when the current matching position is at the
752 start point of the match, as specified by the \fIstartoffset\fP argument of
753 \fBpcre_exec()\fP. It differs from \eA when the value of \fIstartoffset\fP is
754 non-zero. By calling \fBpcre_exec()\fP multiple times with appropriate
755 arguments, you can mimic Perl's /g option, and it is in this kind of
756 implementation where \eG can be useful.
757 .P
758 Note, however, that PCRE's interpretation of \eG, as the start of the current
759 match, is subtly different from Perl's, which defines it as the end of the
760 previous match. In Perl, these can be different when the previously matched
761 string was empty. Because PCRE does just one match at a time, it cannot
762 reproduce this behaviour.
763 .P
764 If all the alternatives of a pattern begin with \eG, the expression is anchored
765 to the starting match position, and the "anchored" flag is set in the compiled
766 regular expression.
767 .
768 .
770 .rs
771 .sp
772 Outside a character class, in the default matching mode, the circumflex
773 character is an assertion that is true only if the current matching point is
774 at the start of the subject string. If the \fIstartoffset\fP argument of
775 \fBpcre_exec()\fP is non-zero, circumflex can never match if the PCRE_MULTILINE
776 option is unset. Inside a character class, circumflex has an entirely different
777 meaning
778 .\" HTML <a href="#characterclass">
779 .\" </a>
780 (see below).
781 .\"
782 .P
783 Circumflex need not be the first character of the pattern if a number of
784 alternatives are involved, but it should be the first thing in each alternative
785 in which it appears if the pattern is ever to match that branch. If all
786 possible alternatives start with a circumflex, that is, if the pattern is
787 constrained to match only at the start of the subject, it is said to be an
788 "anchored" pattern. (There are also other constructs that can cause a pattern
789 to be anchored.)
790 .P
791 A dollar character is an assertion that is true only if the current matching
792 point is at the end of the subject string, or immediately before a newline
793 at the end of the string (by default). Dollar need not be the last character of
794 the pattern if a number of alternatives are involved, but it should be the last
795 item in any branch in which it appears. Dollar has no special meaning in a
796 character class.
797 .P
798 The meaning of dollar can be changed so that it matches only at the very end of
799 the string, by setting the PCRE_DOLLAR_ENDONLY option at compile time. This
800 does not affect the \eZ assertion.
801 .P
802 The meanings of the circumflex and dollar characters are changed if the
803 PCRE_MULTILINE option is set. When this is the case, a circumflex matches
804 immediately after internal newlines as well as at the start of the subject
805 string. It does not match after a newline that ends the string. A dollar
806 matches before any newlines in the string, as well as at the very end, when
807 PCRE_MULTILINE is set. When newline is specified as the two-character
808 sequence CRLF, isolated CR and LF characters do not indicate newlines.
809 .P
810 For example, the pattern /^abc$/ matches the subject string "def\enabc" (where
811 \en represents a newline) in multiline mode, but not otherwise. Consequently,
812 patterns that are anchored in single line mode because all branches start with
813 ^ are not anchored in multiline mode, and a match for circumflex is possible
814 when the \fIstartoffset\fP argument of \fBpcre_exec()\fP is non-zero. The
815 PCRE_DOLLAR_ENDONLY option is ignored if PCRE_MULTILINE is set.
816 .P
817 Note that the sequences \eA, \eZ, and \ez can be used to match the start and
818 end of the subject in both modes, and if all branches of a pattern start with
819 \eA it is always anchored, whether or not PCRE_MULTILINE is set.
820 .
821 .
823 .rs
824 .sp
825 Outside a character class, a dot in the pattern matches any one character in
826 the subject string except (by default) a character that signifies the end of a
827 line. In UTF-8 mode, the matched character may be more than one byte long.
828 .P
829 When a line ending is defined as a single character, dot never matches that
830 character; when the two-character sequence CRLF is used, dot does not match CR
831 if it is immediately followed by LF, but otherwise it matches all characters
832 (including isolated CRs and LFs). When any Unicode line endings are being
833 recognized, dot does not match CR or LF or any of the other line ending
834 characters.
835 .P
836 The behaviour of dot with regard to newlines can be changed. If the PCRE_DOTALL
837 option is set, a dot matches any one character, without exception. If the
838 two-character sequence CRLF is present in the subject string, it takes two dots
839 to match it.
840 .P
841 The handling of dot is entirely independent of the handling of circumflex and
842 dollar, the only relationship being that they both involve newlines. Dot has no
843 special meaning in a character class.
844 .
845 .
847 .rs
848 .sp
849 Outside a character class, the escape sequence \eC matches any one byte, both
850 in and out of UTF-8 mode. Unlike a dot, it always matches any line-ending
851 characters. The feature is provided in Perl in order to match individual bytes
852 in UTF-8 mode. Because it breaks up UTF-8 characters into individual bytes,
853 what remains in the string may be a malformed UTF-8 string. For this reason,
854 the \eC escape sequence is best avoided.
855 .P
856 PCRE does not allow \eC to appear in lookbehind assertions
857 .\" HTML <a href="#lookbehind">
858 .\" </a>
859 (described below),
860 .\"
861 because in UTF-8 mode this would make it impossible to calculate the length of
862 the lookbehind.
863 .
864 .
865 .\" HTML <a name="characterclass"></a>
867 .rs
868 .sp
869 An opening square bracket introduces a character class, terminated by a closing
870 square bracket. A closing square bracket on its own is not special. If a
871 closing square bracket is required as a member of the class, it should be the
872 first data character in the class (after an initial circumflex, if present) or
873 escaped with a backslash.
874 .P
875 A character class matches a single character in the subject. In UTF-8 mode, the
876 character may occupy more than one byte. A matched character must be in the set
877 of characters defined by the class, unless the first character in the class
878 definition is a circumflex, in which case the subject character must not be in
879 the set defined by the class. If a circumflex is actually required as a member
880 of the class, ensure it is not the first character, or escape it with a
881 backslash.
882 .P
883 For example, the character class [aeiou] matches any lower case vowel, while
884 [^aeiou] matches any character that is not a lower case vowel. Note that a
885 circumflex is just a convenient notation for specifying the characters that
886 are in the class by enumerating those that are not. A class that starts with a
887 circumflex is not an assertion: it still consumes a character from the subject
888 string, and therefore it fails if the current pointer is at the end of the
889 string.
890 .P
891 In UTF-8 mode, characters with values greater than 255 can be included in a
892 class as a literal string of bytes, or by using the \ex{ escaping mechanism.
893 .P
894 When caseless matching is set, any letters in a class represent both their
895 upper case and lower case versions, so for example, a caseless [aeiou] matches
896 "A" as well as "a", and a caseless [^aeiou] does not match "A", whereas a
897 caseful version would. In UTF-8 mode, PCRE always understands the concept of
898 case for characters whose values are less than 128, so caseless matching is
899 always possible. For characters with higher values, the concept of case is
900 supported if PCRE is compiled with Unicode property support, but not otherwise.
901 If you want to use caseless matching for characters 128 and above, you must
902 ensure that PCRE is compiled with Unicode property support as well as with
903 UTF-8 support.
904 .P
905 Characters that might indicate line breaks are never treated in any special way
906 when matching character classes, whatever line-ending sequence is in use, and
907 whatever setting of the PCRE_DOTALL and PCRE_MULTILINE options is used. A class
908 such as [^a] always matches one of these characters.
909 .P
910 The minus (hyphen) character can be used to specify a range of characters in a
911 character class. For example, [d-m] matches any letter between d and m,
912 inclusive. If a minus character is required in a class, it must be escaped with
913 a backslash or appear in a position where it cannot be interpreted as
914 indicating a range, typically as the first or last character in the class.
915 .P
916 It is not possible to have the literal character "]" as the end character of a
917 range. A pattern such as [W-]46] is interpreted as a class of two characters
918 ("W" and "-") followed by a literal string "46]", so it would match "W46]" or
919 "-46]". However, if the "]" is escaped with a backslash it is interpreted as
920 the end of range, so [W-\e]46] is interpreted as a class containing a range
921 followed by two other characters. The octal or hexadecimal representation of
922 "]" can also be used to end a range.
923 .P
924 Ranges operate in the collating sequence of character values. They can also be
925 used for characters specified numerically, for example [\e000-\e037]. In UTF-8
926 mode, ranges can include characters whose values are greater than 255, for
927 example [\ex{100}-\ex{2ff}].
928 .P
929 If a range that includes letters is used when caseless matching is set, it
930 matches the letters in either case. For example, [W-c] is equivalent to
931 [][\e\e^_`wxyzabc], matched caselessly, and in non-UTF-8 mode, if character
932 tables for a French locale are in use, [\exc8-\excb] matches accented E
933 characters in both cases. In UTF-8 mode, PCRE supports the concept of case for
934 characters with values greater than 128 only when it is compiled with Unicode
935 property support.
936 .P
937 The character types \ed, \eD, \ep, \eP, \es, \eS, \ew, and \eW may also appear
938 in a character class, and add the characters that they match to the class. For
939 example, [\edABCDEF] matches any hexadecimal digit. A circumflex can
940 conveniently be used with the upper case character types to specify a more
941 restricted set of characters than the matching lower case type. For example,
942 the class [^\eW_] matches any letter or digit, but not underscore.
943 .P
944 The only metacharacters that are recognized in character classes are backslash,
945 hyphen (only where it can be interpreted as specifying a range), circumflex
946 (only at the start), opening square bracket (only when it can be interpreted as
947 introducing a POSIX class name - see the next section), and the terminating
948 closing square bracket. However, escaping other non-alphanumeric characters
949 does no harm.
950 .
951 .
953 .rs
954 .sp
955 Perl supports the POSIX notation for character classes. This uses names
956 enclosed by [: and :] within the enclosing square brackets. PCRE also supports
957 this notation. For example,
958 .sp
959 [01[:alpha:]%]
960 .sp
961 matches "0", "1", any alphabetic character, or "%". The supported class names
962 are
963 .sp
964 alnum letters and digits
965 alpha letters
966 ascii character codes 0 - 127
967 blank space or tab only
968 cntrl control characters
969 digit decimal digits (same as \ed)
970 graph printing characters, excluding space
971 lower lower case letters
972 print printing characters, including space
973 punct printing characters, excluding letters and digits
974 space white space (not quite the same as \es)
975 upper upper case letters
976 word "word" characters (same as \ew)
977 xdigit hexadecimal digits
978 .sp
979 The "space" characters are HT (9), LF (10), VT (11), FF (12), CR (13), and
980 space (32). Notice that this list includes the VT character (code 11). This
981 makes "space" different to \es, which does not include VT (for Perl
982 compatibility).
983 .P
984 The name "word" is a Perl extension, and "blank" is a GNU extension from Perl
985 5.8. Another Perl extension is negation, which is indicated by a ^ character
986 after the colon. For example,
987 .sp
988 [12[:^digit:]]
989 .sp
990 matches "1", "2", or any non-digit. PCRE (and Perl) also recognize the POSIX
991 syntax [.ch.] and [=ch=] where "ch" is a "collating element", but these are not
992 supported, and an error is given if they are encountered.
993 .P
994 In UTF-8 mode, characters with values greater than 128 do not match any of
995 the POSIX character classes.
996 .
997 .
999 .rs
1000 .sp
1001 Vertical bar characters are used to separate alternative patterns. For example,
1002 the pattern
1003 .sp
1004 gilbert|sullivan
1005 .sp
1006 matches either "gilbert" or "sullivan". Any number of alternatives may appear,
1007 and an empty alternative is permitted (matching the empty string). The matching
1008 process tries each alternative in turn, from left to right, and the first one
1009 that succeeds is used. If the alternatives are within a subpattern
1010 .\" HTML <a href="#subpattern">
1011 .\" </a>
1012 (defined below),
1013 .\"
1014 "succeeds" means matching the rest of the main pattern as well as the
1015 alternative in the subpattern.
1016 .
1017 .
1019 .rs
1020 .sp
1021 The settings of the PCRE_CASELESS, PCRE_MULTILINE, PCRE_DOTALL, and
1022 PCRE_EXTENDED options (which are Perl-compatible) can be changed from within
1023 the pattern by a sequence of Perl option letters enclosed between "(?" and ")".
1024 The option letters are
1025 .sp
1026 i for PCRE_CASELESS
1028 s for PCRE_DOTALL
1029 x for PCRE_EXTENDED
1030 .sp
1031 For example, (?im) sets caseless, multiline matching. It is also possible to
1032 unset these options by preceding the letter with a hyphen, and a combined
1033 setting and unsetting such as (?im-sx), which sets PCRE_CASELESS and
1034 PCRE_MULTILINE while unsetting PCRE_DOTALL and PCRE_EXTENDED, is also
1035 permitted. If a letter appears both before and after the hyphen, the option is
1036 unset.
1037 .P
1038 The PCRE-specific options PCRE_DUPNAMES, PCRE_UNGREEDY, and PCRE_EXTRA can be
1039 changed in the same way as the Perl-compatible options by using the characters
1040 J, U and X respectively.
1041 .P
1042 When one of these option changes occurs at top level (that is, not inside
1043 subpattern parentheses), the change applies to the remainder of the pattern
1044 that follows. If the change is placed right at the start of a pattern, PCRE
1045 extracts it into the global options (and it will therefore show up in data
1046 extracted by the \fBpcre_fullinfo()\fP function).
1047 .P
1048 An option change within a subpattern (see below for a description of
1049 subpatterns) affects only that part of the current pattern that follows it, so
1050 .sp
1051 (a(?i)b)c
1052 .sp
1053 matches abc and aBc and no other strings (assuming PCRE_CASELESS is not used).
1054 By this means, options can be made to have different settings in different
1055 parts of the pattern. Any changes made in one alternative do carry on
1056 into subsequent branches within the same subpattern. For example,
1057 .sp
1058 (a(?i)b|c)
1059 .sp
1060 matches "ab", "aB", "c", and "C", even though when matching "C" the first
1061 branch is abandoned before the option setting. This is because the effects of
1062 option settings happen at compile time. There would be some very weird
1063 behaviour otherwise.
1064 .P
1065 \fBNote:\fP There are other PCRE-specific options that can be set by the
1066 application when the compile or match functions are called. In some cases the
1067 pattern can contain special leading sequences such as (*CRLF) to override what
1068 the application has set or what has been defaulted. Details are given in the
1069 section entitled
1070 .\" HTML <a href="#newlineseq">
1071 .\" </a>
1072 "Newline sequences"
1073 .\"
1074 above. There is also the (*UTF8) leading sequence that can be used to set UTF-8
1075 mode; this is equivalent to setting the PCRE_UTF8 option.
1076 .
1077 .
1078 .\" HTML <a name="subpattern"></a>
1080 .rs
1081 .sp
1082 Subpatterns are delimited by parentheses (round brackets), which can be nested.
1083 Turning part of a pattern into a subpattern does two things:
1084 .sp
1085 1. It localizes a set of alternatives. For example, the pattern
1086 .sp
1087 cat(aract|erpillar|)
1088 .sp
1089 matches one of the words "cat", "cataract", or "caterpillar". Without the
1090 parentheses, it would match "cataract", "erpillar" or an empty string.
1091 .sp
1092 2. It sets up the subpattern as a capturing subpattern. This means that, when
1093 the whole pattern matches, that portion of the subject string that matched the
1094 subpattern is passed back to the caller via the \fIovector\fP argument of
1095 \fBpcre_exec()\fP. Opening parentheses are counted from left to right (starting
1096 from 1) to obtain numbers for the capturing subpatterns.
1097 .P
1098 For example, if the string "the red king" is matched against the pattern
1099 .sp
1100 the ((red|white) (king|queen))
1101 .sp
1102 the captured substrings are "red king", "red", and "king", and are numbered 1,
1103 2, and 3, respectively.
1104 .P
1105 The fact that plain parentheses fulfil two functions is not always helpful.
1106 There are often times when a grouping subpattern is required without a
1107 capturing requirement. If an opening parenthesis is followed by a question mark
1108 and a colon, the subpattern does not do any capturing, and is not counted when
1109 computing the number of any subsequent capturing subpatterns. For example, if
1110 the string "the white queen" is matched against the pattern
1111 .sp
1112 the ((?:red|white) (king|queen))
1113 .sp
1114 the captured substrings are "white queen" and "queen", and are numbered 1 and
1115 2. The maximum number of capturing subpatterns is 65535.
1116 .P
1117 As a convenient shorthand, if any option settings are required at the start of
1118 a non-capturing subpattern, the option letters may appear between the "?" and
1119 the ":". Thus the two patterns
1120 .sp
1121 (?i:saturday|sunday)
1122 (?:(?i)saturday|sunday)
1123 .sp
1124 match exactly the same set of strings. Because alternative branches are tried
1125 from left to right, and options are not reset until the end of the subpattern
1126 is reached, an option setting in one branch does affect subsequent branches, so
1127 the above patterns match "SUNDAY" as well as "Saturday".
1128 .
1129 .
1131 .rs
1132 .sp
1133 Perl 5.10 introduced a feature whereby each alternative in a subpattern uses
1134 the same numbers for its capturing parentheses. Such a subpattern starts with
1135 (?| and is itself a non-capturing subpattern. For example, consider this
1136 pattern:
1137 .sp
1138 (?|(Sat)ur|(Sun))day
1139 .sp
1140 Because the two alternatives are inside a (?| group, both sets of capturing
1141 parentheses are numbered one. Thus, when the pattern matches, you can look
1142 at captured substring number one, whichever alternative matched. This construct
1143 is useful when you want to capture part, but not all, of one of a number of
1144 alternatives. Inside a (?| group, parentheses are numbered as usual, but the
1145 number is reset at the start of each branch. The numbers of any capturing
1146 buffers that follow the subpattern start after the highest number used in any
1147 branch. The following example is taken from the Perl documentation.
1148 The numbers underneath show in which buffer the captured content will be
1149 stored.
1150 .sp
1151 # before ---------------branch-reset----------- after
1152 / ( a ) (?| x ( y ) z | (p (q) r) | (t) u (v) ) ( z ) /x
1153 # 1 2 2 3 2 3 4
1154 .sp
1155 A backreference or a recursive call to a numbered subpattern always refers to
1156 the first one in the pattern with the given number.
1157 .P
1158 An alternative approach to using this "branch reset" feature is to use
1159 duplicate named subpatterns, as described in the next section.
1160 .
1161 .
1163 .rs
1164 .sp
1165 Identifying capturing parentheses by number is simple, but it can be very hard
1166 to keep track of the numbers in complicated regular expressions. Furthermore,
1167 if an expression is modified, the numbers may change. To help with this
1168 difficulty, PCRE supports the naming of subpatterns. This feature was not
1169 added to Perl until release 5.10. Python had the feature earlier, and PCRE
1170 introduced it at release 4.0, using the Python syntax. PCRE now supports both
1171 the Perl and the Python syntax.
1172 .P
1173 In PCRE, a subpattern can be named in one of three ways: (?<name>...) or
1174 (?'name'...) as in Perl, or (?P<name>...) as in Python. References to capturing
1175 parentheses from other parts of the pattern, such as
1176 .\" HTML <a href="#backreferences">
1177 .\" </a>
1178 backreferences,
1179 .\"
1180 .\" HTML <a href="#recursion">
1181 .\" </a>
1182 recursion,
1183 .\"
1184 and
1185 .\" HTML <a href="#conditions">
1186 .\" </a>
1187 conditions,
1188 .\"
1189 can be made by name as well as by number.
1190 .P
1191 Names consist of up to 32 alphanumeric characters and underscores. Named
1192 capturing parentheses are still allocated numbers as well as names, exactly as
1193 if the names were not present. The PCRE API provides function calls for
1194 extracting the name-to-number translation table from a compiled pattern. There
1195 is also a convenience function for extracting a captured substring by name.
1196 .P
1197 By default, a name must be unique within a pattern, but it is possible to relax
1198 this constraint by setting the PCRE_DUPNAMES option at compile time. This can
1199 be useful for patterns where only one instance of the named parentheses can
1200 match. Suppose you want to match the name of a weekday, either as a 3-letter
1201 abbreviation or as the full name, and in both cases you want to extract the
1202 abbreviation. This pattern (ignoring the line breaks) does the job:
1203 .sp
1204 (?<DN>Mon|Fri|Sun)(?:day)?|
1205 (?<DN>Tue)(?:sday)?|
1206 (?<DN>Wed)(?:nesday)?|
1207 (?<DN>Thu)(?:rsday)?|
1208 (?<DN>Sat)(?:urday)?
1209 .sp
1210 There are five capturing substrings, but only one is ever set after a match.
1211 (An alternative way of solving this problem is to use a "branch reset"
1212 subpattern, as described in the previous section.)
1213 .P
1214 The convenience function for extracting the data by name returns the substring
1215 for the first (and in this example, the only) subpattern of that name that
1216 matched. This saves searching to find which numbered subpattern it was. If you
1217 make a reference to a non-unique named subpattern from elsewhere in the
1218 pattern, the one that corresponds to the lowest number is used. For further
1219 details of the interfaces for handling named subpatterns, see the
1220 .\" HREF
1221 \fBpcreapi\fP
1222 .\"
1223 documentation.
1224 .P
1225 \fBWarning:\fP You cannot use different names to distinguish between two
1226 subpatterns with the same number (see the previous section) because PCRE uses
1227 only the numbers when matching.
1228 .
1229 .
1231 .rs
1232 .sp
1233 Repetition is specified by quantifiers, which can follow any of the following
1234 items:
1235 .sp
1236 a literal data character
1237 the dot metacharacter
1238 the \eC escape sequence
1239 the \eX escape sequence (in UTF-8 mode with Unicode properties)
1240 the \eR escape sequence
1241 an escape such as \ed that matches a single character
1242 a character class
1243 a back reference (see next section)
1244 a parenthesized subpattern (unless it is an assertion)
1245 .sp
1246 The general repetition quantifier specifies a minimum and maximum number of
1247 permitted matches, by giving the two numbers in curly brackets (braces),
1248 separated by a comma. The numbers must be less than 65536, and the first must
1249 be less than or equal to the second. For example:
1250 .sp
1251 z{2,4}
1252 .sp
1253 matches "zz", "zzz", or "zzzz". A closing brace on its own is not a special
1254 character. If the second number is omitted, but the comma is present, there is
1255 no upper limit; if the second number and the comma are both omitted, the
1256 quantifier specifies an exact number of required matches. Thus
1257 .sp
1258 [aeiou]{3,}
1259 .sp
1260 matches at least 3 successive vowels, but may match many more, while
1261 .sp
1262 \ed{8}
1263 .sp
1264 matches exactly 8 digits. An opening curly bracket that appears in a position
1265 where a quantifier is not allowed, or one that does not match the syntax of a
1266 quantifier, is taken as a literal character. For example, {,6} is not a
1267 quantifier, but a literal string of four characters.
1268 .P
1269 In UTF-8 mode, quantifiers apply to UTF-8 characters rather than to individual
1270 bytes. Thus, for example, \ex{100}{2} matches two UTF-8 characters, each of
1271 which is represented by a two-byte sequence. Similarly, when Unicode property
1272 support is available, \eX{3} matches three Unicode extended sequences, each of
1273 which may be several bytes long (and they may be of different lengths).
1274 .P
1275 The quantifier {0} is permitted, causing the expression to behave as if the
1276 previous item and the quantifier were not present. This may be useful for
1277 subpatterns that are referenced as
1278 .\" HTML <a href="#subpatternsassubroutines">
1279 .\" </a>
1280 subroutines
1281 .\"
1282 from elsewhere in the pattern. Items other than subpatterns that have a {0}
1283 quantifier are omitted from the compiled pattern.
1284 .P
1285 For convenience, the three most common quantifiers have single-character
1286 abbreviations:
1287 .sp
1288 * is equivalent to {0,}
1289 + is equivalent to {1,}
1290 ? is equivalent to {0,1}
1291 .sp
1292 It is possible to construct infinite loops by following a subpattern that can
1293 match no characters with a quantifier that has no upper limit, for example:
1294 .sp
1295 (a?)*
1296 .sp
1297 Earlier versions of Perl and PCRE used to give an error at compile time for
1298 such patterns. However, because there are cases where this can be useful, such
1299 patterns are now accepted, but if any repetition of the subpattern does in fact
1300 match no characters, the loop is forcibly broken.
1301 .P
1302 By default, the quantifiers are "greedy", that is, they match as much as
1303 possible (up to the maximum number of permitted times), without causing the
1304 rest of the pattern to fail. The classic example of where this gives problems
1305 is in trying to match comments in C programs. These appear between /* and */
1306 and within the comment, individual * and / characters may appear. An attempt to
1307 match C comments by applying the pattern
1308 .sp
1309 /\e*.*\e*/
1310 .sp
1311 to the string
1312 .sp
1313 /* first comment */ not comment /* second comment */
1314 .sp
1315 fails, because it matches the entire string owing to the greediness of the .*
1316 item.
1317 .P
1318 However, if a quantifier is followed by a question mark, it ceases to be
1319 greedy, and instead matches the minimum number of times possible, so the
1320 pattern
1321 .sp
1322 /\e*.*?\e*/
1323 .sp
1324 does the right thing with the C comments. The meaning of the various
1325 quantifiers is not otherwise changed, just the preferred number of matches.
1326 Do not confuse this use of question mark with its use as a quantifier in its
1327 own right. Because it has two uses, it can sometimes appear doubled, as in
1328 .sp
1329 \ed??\ed
1330 .sp
1331 which matches one digit by preference, but can match two if that is the only
1332 way the rest of the pattern matches.
1333 .P
1334 If the PCRE_UNGREEDY option is set (an option that is not available in Perl),
1335 the quantifiers are not greedy by default, but individual ones can be made
1336 greedy by following them with a question mark. In other words, it inverts the
1337 default behaviour.
1338 .P
1339 When a parenthesized subpattern is quantified with a minimum repeat count that
1340 is greater than 1 or with a limited maximum, more memory is required for the
1341 compiled pattern, in proportion to the size of the minimum or maximum.
1342 .P
1343 If a pattern starts with .* or .{0,} and the PCRE_DOTALL option (equivalent
1344 to Perl's /s) is set, thus allowing the dot to match newlines, the pattern is
1345 implicitly anchored, because whatever follows will be tried against every
1346 character position in the subject string, so there is no point in retrying the
1347 overall match at any position after the first. PCRE normally treats such a
1348 pattern as though it were preceded by \eA.
1349 .P
1350 In cases where it is known that the subject string contains no newlines, it is
1351 worth setting PCRE_DOTALL in order to obtain this optimization, or
1352 alternatively using ^ to indicate anchoring explicitly.
1353 .P
1354 However, there is one situation where the optimization cannot be used. When .*
1355 is inside capturing parentheses that are the subject of a backreference
1356 elsewhere in the pattern, a match at the start may fail where a later one
1357 succeeds. Consider, for example:
1358 .sp
1359 (.*)abc\e1
1360 .sp
1361 If the subject is "xyz123abc123" the match point is the fourth character. For
1362 this reason, such a pattern is not implicitly anchored.
1363 .P
1364 When a capturing subpattern is repeated, the value captured is the substring
1365 that matched the final iteration. For example, after
1366 .sp
1367 (tweedle[dume]{3}\es*)+
1368 .sp
1369 has matched "tweedledum tweedledee" the value of the captured substring is
1370 "tweedledee". However, if there are nested capturing subpatterns, the
1371 corresponding captured values may have been set in previous iterations. For
1372 example, after
1373 .sp
1374 /(a|(b))+/
1375 .sp
1376 matches "aba" the value of the second captured substring is "b".
1377 .
1378 .
1379 .\" HTML <a name="atomicgroup"></a>
1381 .rs
1382 .sp
1383 With both maximizing ("greedy") and minimizing ("ungreedy" or "lazy")
1384 repetition, failure of what follows normally causes the repeated item to be
1385 re-evaluated to see if a different number of repeats allows the rest of the
1386 pattern to match. Sometimes it is useful to prevent this, either to change the
1387 nature of the match, or to cause it fail earlier than it otherwise might, when
1388 the author of the pattern knows there is no point in carrying on.
1389 .P
1390 Consider, for example, the pattern \ed+foo when applied to the subject line
1391 .sp
1392 123456bar
1393 .sp
1394 After matching all 6 digits and then failing to match "foo", the normal
1395 action of the matcher is to try again with only 5 digits matching the \ed+
1396 item, and then with 4, and so on, before ultimately failing. "Atomic grouping"
1397 (a term taken from Jeffrey Friedl's book) provides the means for specifying
1398 that once a subpattern has matched, it is not to be re-evaluated in this way.
1399 .P
1400 If we use atomic grouping for the previous example, the matcher gives up
1401 immediately on failing to match "foo" the first time. The notation is a kind of
1402 special parenthesis, starting with (?> as in this example:
1403 .sp
1404 (?>\ed+)foo
1405 .sp
1406 This kind of parenthesis "locks up" the part of the pattern it contains once
1407 it has matched, and a failure further into the pattern is prevented from
1408 backtracking into it. Backtracking past it to previous items, however, works as
1409 normal.
1410 .P
1411 An alternative description is that a subpattern of this type matches the string
1412 of characters that an identical standalone pattern would match, if anchored at
1413 the current point in the subject string.
1414 .P
1415 Atomic grouping subpatterns are not capturing subpatterns. Simple cases such as
1416 the above example can be thought of as a maximizing repeat that must swallow
1417 everything it can. So, while both \ed+ and \ed+? are prepared to adjust the
1418 number of digits they match in order to make the rest of the pattern match,
1419 (?>\ed+) can only match an entire sequence of digits.
1420 .P
1421 Atomic groups in general can of course contain arbitrarily complicated
1422 subpatterns, and can be nested. However, when the subpattern for an atomic
1423 group is just a single repeated item, as in the example above, a simpler
1424 notation, called a "possessive quantifier" can be used. This consists of an
1425 additional + character following a quantifier. Using this notation, the
1426 previous example can be rewritten as
1427 .sp
1428 \ed++foo
1429 .sp
1430 Note that a possessive quantifier can be used with an entire group, for
1431 example:
1432 .sp
1433 (abc|xyz){2,3}+
1434 .sp
1435 Possessive quantifiers are always greedy; the setting of the PCRE_UNGREEDY
1436 option is ignored. They are a convenient notation for the simpler forms of
1437 atomic group. However, there is no difference in the meaning of a possessive
1438 quantifier and the equivalent atomic group, though there may be a performance
1439 difference; possessive quantifiers should be slightly faster.
1440 .P
1441 The possessive quantifier syntax is an extension to the Perl 5.8 syntax.
1442 Jeffrey Friedl originated the idea (and the name) in the first edition of his
1443 book. Mike McCloskey liked it, so implemented it when he built Sun's Java
1444 package, and PCRE copied it from there. It ultimately found its way into Perl
1445 at release 5.10.
1446 .P
1447 PCRE has an optimization that automatically "possessifies" certain simple
1448 pattern constructs. For example, the sequence A+B is treated as A++B because
1449 there is no point in backtracking into a sequence of A's when B must follow.
1450 .P
1451 When a pattern contains an unlimited repeat inside a subpattern that can itself
1452 be repeated an unlimited number of times, the use of an atomic group is the
1453 only way to avoid some failing matches taking a very long time indeed. The
1454 pattern
1455 .sp
1456 (\eD+|<\ed+>)*[!?]
1457 .sp
1458 matches an unlimited number of substrings that either consist of non-digits, or
1459 digits enclosed in <>, followed by either ! or ?. When it matches, it runs
1460 quickly. However, if it is applied to
1461 .sp
1462 aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
1463 .sp
1464 it takes a long time before reporting failure. This is because the string can
1465 be divided between the internal \eD+ repeat and the external * repeat in a
1466 large number of ways, and all have to be tried. (The example uses [!?] rather
1467 than a single character at the end, because both PCRE and Perl have an
1468 optimization that allows for fast failure when a single character is used. They
1469 remember the last single character that is required for a match, and fail early
1470 if it is not present in the string.) If the pattern is changed so that it uses
1471 an atomic group, like this:
1472 .sp
1473 ((?>\eD+)|<\ed+>)*[!?]
1474 .sp
1475 sequences of non-digits cannot be broken, and failure happens quickly.
1476 .
1477 .
1478 .\" HTML <a name="backreferences"></a>
1480 .rs
1481 .sp
1482 Outside a character class, a backslash followed by a digit greater than 0 (and
1483 possibly further digits) is a back reference to a capturing subpattern earlier
1484 (that is, to its left) in the pattern, provided there have been that many
1485 previous capturing left parentheses.
1486 .P
1487 However, if the decimal number following the backslash is less than 10, it is
1488 always taken as a back reference, and causes an error only if there are not
1489 that many capturing left parentheses in the entire pattern. In other words, the
1490 parentheses that are referenced need not be to the left of the reference for
1491 numbers less than 10. A "forward back reference" of this type can make sense
1492 when a repetition is involved and the subpattern to the right has participated
1493 in an earlier iteration.
1494 .P
1495 It is not possible to have a numerical "forward back reference" to a subpattern
1496 whose number is 10 or more using this syntax because a sequence such as \e50 is
1497 interpreted as a character defined in octal. See the subsection entitled
1498 "Non-printing characters"
1499 .\" HTML <a href="#digitsafterbackslash">
1500 .\" </a>
1501 above
1502 .\"
1503 for further details of the handling of digits following a backslash. There is
1504 no such problem when named parentheses are used. A back reference to any
1505 subpattern is possible using named parentheses (see below).
1506 .P
1507 Another way of avoiding the ambiguity inherent in the use of digits following a
1508 backslash is to use the \eg escape sequence, which is a feature introduced in
1509 Perl 5.10. This escape must be followed by an unsigned number or a negative
1510 number, optionally enclosed in braces. These examples are all identical:
1511 .sp
1512 (ring), \e1
1513 (ring), \eg1
1514 (ring), \eg{1}
1515 .sp
1516 An unsigned number specifies an absolute reference without the ambiguity that
1517 is present in the older syntax. It is also useful when literal digits follow
1518 the reference. A negative number is a relative reference. Consider this
1519 example:
1520 .sp
1521 (abc(def)ghi)\eg{-1}
1522 .sp
1523 The sequence \eg{-1} is a reference to the most recently started capturing
1524 subpattern before \eg, that is, is it equivalent to \e2. Similarly, \eg{-2}
1525 would be equivalent to \e1. The use of relative references can be helpful in
1526 long patterns, and also in patterns that are created by joining together
1527 fragments that contain references within themselves.
1528 .P
1529 A back reference matches whatever actually matched the capturing subpattern in
1530 the current subject string, rather than anything matching the subpattern
1531 itself (see
1532 .\" HTML <a href="#subpatternsassubroutines">
1533 .\" </a>
1534 "Subpatterns as subroutines"
1535 .\"
1536 below for a way of doing that). So the pattern
1537 .sp
1538 (sens|respons)e and \e1ibility
1539 .sp
1540 matches "sense and sensibility" and "response and responsibility", but not
1541 "sense and responsibility". If caseful matching is in force at the time of the
1542 back reference, the case of letters is relevant. For example,
1543 .sp
1544 ((?i)rah)\es+\e1
1545 .sp
1546 matches "rah rah" and "RAH RAH", but not "RAH rah", even though the original
1547 capturing subpattern is matched caselessly.
1548 .P
1549 There are several different ways of writing back references to named
1550 subpatterns. The .NET syntax \ek{name} and the Perl syntax \ek<name> or
1551 \ek'name' are supported, as is the Python syntax (?P=name). Perl 5.10's unified
1552 back reference syntax, in which \eg can be used for both numeric and named
1553 references, is also supported. We could rewrite the above example in any of
1554 the following ways:
1555 .sp
1556 (?<p1>(?i)rah)\es+\ek<p1>
1557 (?'p1'(?i)rah)\es+\ek{p1}
1558 (?P<p1>(?i)rah)\es+(?P=p1)
1559 (?<p1>(?i)rah)\es+\eg{p1}
1560 .sp
1561 A subpattern that is referenced by name may appear in the pattern before or
1562 after the reference.
1563 .P
1564 There may be more than one back reference to the same subpattern. If a
1565 subpattern has not actually been used in a particular match, any back
1566 references to it always fail. For example, the pattern
1567 .sp
1568 (a|(bc))\e2
1569 .sp
1570 always fails if it starts to match "a" rather than "bc". Because there may be
1571 many capturing parentheses in a pattern, all digits following the backslash are
1572 taken as part of a potential back reference number. If the pattern continues
1573 with a digit character, some delimiter must be used to terminate the back
1574 reference. If the PCRE_EXTENDED option is set, this can be whitespace.
1575 Otherwise an empty comment (see
1576 .\" HTML <a href="#comments">
1577 .\" </a>
1578 "Comments"
1579 .\"
1580 below) can be used.
1581 .P
1582 A back reference that occurs inside the parentheses to which it refers fails
1583 when the subpattern is first used, so, for example, (a\e1) never matches.
1584 However, such references can be useful inside repeated subpatterns. For
1585 example, the pattern
1586 .sp
1587 (a|b\e1)+
1588 .sp
1589 matches any number of "a"s and also "aba", "ababbaa" etc. At each iteration of
1590 the subpattern, the back reference matches the character string corresponding
1591 to the previous iteration. In order for this to work, the pattern must be such
1592 that the first iteration does not need to match the back reference. This can be
1593 done using alternation, as in the example above, or by a quantifier with a
1594 minimum of zero.
1595 .
1596 .
1597 .\" HTML <a name="bigassertions"></a>
1599 .rs
1600 .sp
1601 An assertion is a test on the characters following or preceding the current
1602 matching point that does not actually consume any characters. The simple
1603 assertions coded as \eb, \eB, \eA, \eG, \eZ, \ez, ^ and $ are described
1604 .\" HTML <a href="#smallassertions">
1605 .\" </a>
1606 above.
1607 .\"
1608 .P
1609 More complicated assertions are coded as subpatterns. There are two kinds:
1610 those that look ahead of the current position in the subject string, and those
1611 that look behind it. An assertion subpattern is matched in the normal way,
1612 except that it does not cause the current matching position to be changed.
1613 .P
1614 Assertion subpatterns are not capturing subpatterns, and may not be repeated,
1615 because it makes no sense to assert the same thing several times. If any kind
1616 of assertion contains capturing subpatterns within it, these are counted for
1617 the purposes of numbering the capturing subpatterns in the whole pattern.
1618 However, substring capturing is carried out only for positive assertions,
1619 because it does not make sense for negative assertions.
1620 .
1621 .
1622 .SS "Lookahead assertions"
1623 .rs
1624 .sp
1625 Lookahead assertions start with (?= for positive assertions and (?! for
1626 negative assertions. For example,
1627 .sp
1628 \ew+(?=;)
1629 .sp
1630 matches a word followed by a semicolon, but does not include the semicolon in
1631 the match, and
1632 .sp
1633 foo(?!bar)
1634 .sp
1635 matches any occurrence of "foo" that is not followed by "bar". Note that the
1636 apparently similar pattern
1637 .sp
1638 (?!foo)bar
1639 .sp
1640 does not find an occurrence of "bar" that is preceded by something other than
1641 "foo"; it finds any occurrence of "bar" whatsoever, because the assertion
1642 (?!foo) is always true when the next three characters are "bar". A
1643 lookbehind assertion is needed to achieve the other effect.
1644 .P
1645 If you want to force a matching failure at some point in a pattern, the most
1646 convenient way to do it is with (?!) because an empty string always matches, so
1647 an assertion that requires there not to be an empty string must always fail.
1648 .
1649 .
1650 .\" HTML <a name="lookbehind"></a>
1651 .SS "Lookbehind assertions"
1652 .rs
1653 .sp
1654 Lookbehind assertions start with (?<= for positive assertions and (?<! for
1655 negative assertions. For example,
1656 .sp
1657 (?<!foo)bar
1658 .sp
1659 does find an occurrence of "bar" that is not preceded by "foo". The contents of
1660 a lookbehind assertion are restricted such that all the strings it matches must
1661 have a fixed length. However, if there are several top-level alternatives, they
1662 do not all have to have the same fixed length. Thus
1663 .sp
1664 (?<=bullock|donkey)
1665 .sp
1666 is permitted, but
1667 .sp
1668 (?<!dogs?|cats?)
1669 .sp
1670 causes an error at compile time. Branches that match different length strings
1671 are permitted only at the top level of a lookbehind assertion. This is an
1672 extension compared with Perl (at least for 5.8), which requires all branches to
1673 match the same length of string. An assertion such as
1674 .sp
1675 (?<=ab(c|de))
1676 .sp
1677 is not permitted, because its single top-level branch can match two different
1678 lengths, but it is acceptable if rewritten to use two top-level branches:
1679 .sp
1680 (?<=abc|abde)
1681 .sp
1682 In some cases, the Perl 5.10 escape sequence \eK
1683 .\" HTML <a href="#resetmatchstart">
1684 .\" </a>
1685 (see above)
1686 .\"
1687 can be used instead of a lookbehind assertion; this is not restricted to a
1688 fixed-length.
1689 .P
1690 The implementation of lookbehind assertions is, for each alternative, to
1691 temporarily move the current position back by the fixed length and then try to
1692 match. If there are insufficient characters before the current position, the
1693 assertion fails.
1694 .P
1695 PCRE does not allow the \eC escape (which matches a single byte in UTF-8 mode)
1696 to appear in lookbehind assertions, because it makes it impossible to calculate
1697 the length of the lookbehind. The \eX and \eR escapes, which can match
1698 different numbers of bytes, are also not permitted.
1699 .P
1700 Possessive quantifiers can be used in conjunction with lookbehind assertions to
1701 specify efficient matching at the end of the subject string. Consider a simple
1702 pattern such as
1703 .sp
1704 abcd$
1705 .sp
1706 when applied to a long string that does not match. Because matching proceeds
1707 from left to right, PCRE will look for each "a" in the subject and then see if
1708 what follows matches the rest of the pattern. If the pattern is specified as
1709 .sp
1710 ^.*abcd$
1711 .sp
1712 the initial .* matches the entire string at first, but when this fails (because
1713 there is no following "a"), it backtracks to match all but the last character,
1714 then all but the last two characters, and so on. Once again the search for "a"
1715 covers the entire string, from right to left, so we are no better off. However,
1716 if the pattern is written as
1717 .sp
1718 ^.*+(?<=abcd)
1719 .sp
1720 there can be no backtracking for the .*+ item; it can match only the entire
1721 string. The subsequent lookbehind assertion does a single test on the last four
1722 characters. If it fails, the match fails immediately. For long strings, this
1723 approach makes a significant difference to the processing time.
1724 .
1725 .
1726 .SS "Using multiple assertions"
1727 .rs
1728 .sp
1729 Several assertions (of any sort) may occur in succession. For example,
1730 .sp
1731 (?<=\ed{3})(?<!999)foo
1732 .sp
1733 matches "foo" preceded by three digits that are not "999". Notice that each of
1734 the assertions is applied independently at the same point in the subject
1735 string. First there is a check that the previous three characters are all
1736 digits, and then there is a check that the same three characters are not "999".
1737 This pattern does \fInot\fP match "foo" preceded by six characters, the first
1738 of which are digits and the last three of which are not "999". For example, it
1739 doesn't match "123abcfoo". A pattern to do that is
1740 .sp
1741 (?<=\ed{3}...)(?<!999)foo
1742 .sp
1743 This time the first assertion looks at the preceding six characters, checking
1744 that the first three are digits, and then the second assertion checks that the
1745 preceding three characters are not "999".
1746 .P
1747 Assertions can be nested in any combination. For example,
1748 .sp
1749 (?<=(?<!foo)bar)baz
1750 .sp
1751 matches an occurrence of "baz" that is preceded by "bar" which in turn is not
1752 preceded by "foo", while
1753 .sp
1754 (?<=\ed{3}(?!999)...)foo
1755 .sp
1756 is another pattern that matches "foo" preceded by three digits and any three
1757 characters that are not "999".
1758 .
1759 .
1760 .\" HTML <a name="conditions"></a>
1762 .rs
1763 .sp
1764 It is possible to cause the matching process to obey a subpattern
1765 conditionally or to choose between two alternative subpatterns, depending on
1766 the result of an assertion, or whether a previous capturing subpattern matched
1767 or not. The two possible forms of conditional subpattern are
1768 .sp
1769 (?(condition)yes-pattern)
1770 (?(condition)yes-pattern|no-pattern)
1771 .sp
1772 If the condition is satisfied, the yes-pattern is used; otherwise the
1773 no-pattern (if present) is used. If there are more than two alternatives in the
1774 subpattern, a compile-time error occurs.
1775 .P
1776 There are four kinds of condition: references to subpatterns, references to
1777 recursion, a pseudo-condition called DEFINE, and assertions.
1778 .
1779 .SS "Checking for a used subpattern by number"
1780 .rs
1781 .sp
1782 If the text between the parentheses consists of a sequence of digits, the
1783 condition is true if the capturing subpattern of that number has previously
1784 matched. An alternative notation is to precede the digits with a plus or minus
1785 sign. In this case, the subpattern number is relative rather than absolute.
1786 The most recently opened parentheses can be referenced by (?(-1), the next most
1787 recent by (?(-2), and so on. In looping constructs it can also make sense to
1788 refer to subsequent groups with constructs such as (?(+2).
1789 .P
1790 Consider the following pattern, which contains non-significant white space to
1791 make it more readable (assume the PCRE_EXTENDED option) and to divide it into
1792 three parts for ease of discussion:
1793 .sp
1794 ( \e( )? [^()]+ (?(1) \e) )
1795 .sp
1796 The first part matches an optional opening parenthesis, and if that
1797 character is present, sets it as the first captured substring. The second part
1798 matches one or more characters that are not parentheses. The third part is a
1799 conditional subpattern that tests whether the first set of parentheses matched
1800 or not. If they did, that is, if subject started with an opening parenthesis,
1801 the condition is true, and so the yes-pattern is executed and a closing
1802 parenthesis is required. Otherwise, since no-pattern is not present, the
1803 subpattern matches nothing. In other words, this pattern matches a sequence of
1804 non-parentheses, optionally enclosed in parentheses.
1805 .P
1806 If you were embedding this pattern in a larger one, you could use a relative
1807 reference:
1808 .sp
1809 ...other stuff... ( \e( )? [^()]+ (?(-1) \e) ) ...
1810 .sp
1811 This makes the fragment independent of the parentheses in the larger pattern.
1812 .
1813 .SS "Checking for a used subpattern by name"
1814 .rs
1815 .sp
1816 Perl uses the syntax (?(<name>)...) or (?('name')...) to test for a used
1817 subpattern by name. For compatibility with earlier versions of PCRE, which had
1818 this facility before Perl, the syntax (?(name)...) is also recognized. However,
1819 there is a possible ambiguity with this syntax, because subpattern names may
1820 consist entirely of digits. PCRE looks first for a named subpattern; if it
1821 cannot find one and the name consists entirely of digits, PCRE looks for a
1822 subpattern of that number, which must be greater than zero. Using subpattern
1823 names that consist entirely of digits is not recommended.
1824 .P
1825 Rewriting the above example to use a named subpattern gives this:
1826 .sp
1827 (?<OPEN> \e( )? [^()]+ (?(<OPEN>) \e) )
1828 .sp
1829 .
1830 .SS "Checking for pattern recursion"
1831 .rs
1832 .sp
1833 If the condition is the string (R), and there is no subpattern with the name R,
1834 the condition is true if a recursive call to the whole pattern or any
1835 subpattern has been made. If digits or a name preceded by ampersand follow the
1836 letter R, for example:
1837 .sp
1838 (?(R3)...) or (?(R&name)...)
1839 .sp
1840 the condition is true if the most recent recursion is into the subpattern whose
1841 number or name is given. This condition does not check the entire recursion
1842 stack.
1843 .P
1844 At "top level", all these recursion test conditions are false. Recursive
1845 patterns are described below.
1846 .
1847 .SS "Defining subpatterns for use by reference only"
1848 .rs
1849 .sp
1850 If the condition is the string (DEFINE), and there is no subpattern with the
1851 name DEFINE, the condition is always false. In this case, there may be only one
1852 alternative in the subpattern. It is always skipped if control reaches this
1853 point in the pattern; the idea of DEFINE is that it can be used to define
1854 "subroutines" that can be referenced from elsewhere. (The use of "subroutines"
1855 is described below.) For example, a pattern to match an IPv4 address could be
1856 written like this (ignore whitespace and line breaks):
1857 .sp
1858 (?(DEFINE) (?<byte> 2[0-4]\ed | 25[0-5] | 1\ed\ed | [1-9]?\ed) )
1859 \eb (?&byte) (\e.(?&byte)){3} \eb
1860 .sp
1861 The first part of the pattern is a DEFINE group inside which a another group
1862 named "byte" is defined. This matches an individual component of an IPv4
1863 address (a number less than 256). When matching takes place, this part of the
1864 pattern is skipped because DEFINE acts like a false condition.
1865 .P
1866 The rest of the pattern uses references to the named group to match the four
1867 dot-separated components of an IPv4 address, insisting on a word boundary at
1868 each end.
1869 .
1870 .SS "Assertion conditions"
1871 .rs
1872 .sp
1873 If the condition is not in any of the above formats, it must be an assertion.
1874 This may be a positive or negative lookahead or lookbehind assertion. Consider
1875 this pattern, again containing non-significant white space, and with the two
1876 alternatives on the second line:
1877 .sp
1878 (?(?=[^a-z]*[a-z])
1879 \ed{2}-[a-z]{3}-\ed{2} | \ed{2}-\ed{2}-\ed{2} )
1880 .sp
1881 The condition is a positive lookahead assertion that matches an optional
1882 sequence of non-letters followed by a letter. In other words, it tests for the
1883 presence of at least one letter in the subject. If a letter is found, the
1884 subject is matched against the first alternative; otherwise it is matched
1885 against the second. This pattern matches strings in one of the two forms
1886 dd-aaa-dd or dd-dd-dd, where aaa are letters and dd are digits.
1887 .
1888 .
1889 .\" HTML <a name="comments"></a>
1891 .rs
1892 .sp
1893 The sequence (?# marks the start of a comment that continues up to the next
1894 closing parenthesis. Nested parentheses are not permitted. The characters
1895 that make up a comment play no part in the pattern matching at all.
1896 .P
1897 If the PCRE_EXTENDED option is set, an unescaped # character outside a
1898 character class introduces a comment that continues to immediately after the
1899 next newline in the pattern.
1900 .
1901 .
1902 .\" HTML <a name="recursion"></a>
1904 .rs
1905 .sp
1906 Consider the problem of matching a string in parentheses, allowing for
1907 unlimited nested parentheses. Without the use of recursion, the best that can
1908 be done is to use a pattern that matches up to some fixed depth of nesting. It
1909 is not possible to handle an arbitrary nesting depth.
1910 .P
1911 For some time, Perl has provided a facility that allows regular expressions to
1912 recurse (amongst other things). It does this by interpolating Perl code in the
1913 expression at run time, and the code can refer to the expression itself. A Perl
1914 pattern using code interpolation to solve the parentheses problem can be
1915 created like this:
1916 .sp
1917 $re = qr{\e( (?: (?>[^()]+) | (?p{$re}) )* \e)}x;
1918 .sp
1919 The (?p{...}) item interpolates Perl code at run time, and in this case refers
1920 recursively to the pattern in which it appears.
1921 .P
1922 Obviously, PCRE cannot support the interpolation of Perl code. Instead, it
1923 supports special syntax for recursion of the entire pattern, and also for
1924 individual subpattern recursion. After its introduction in PCRE and Python,
1925 this kind of recursion was introduced into Perl at release 5.10.
1926 .P
1927 A special item that consists of (? followed by a number greater than zero and a
1928 closing parenthesis is a recursive call of the subpattern of the given number,
1929 provided that it occurs inside that subpattern. (If not, it is a "subroutine"
1930 call, which is described in the next section.) The special item (?R) or (?0) is
1931 a recursive call of the entire regular expression.
1932 .P
1933 In PCRE (like Python, but unlike Perl), a recursive subpattern call is always
1934 treated as an atomic group. That is, once it has matched some of the subject
1935 string, it is never re-entered, even if it contains untried alternatives and
1936 there is a subsequent matching failure.
1937 .P
1938 This PCRE pattern solves the nested parentheses problem (assume the
1939 PCRE_EXTENDED option is set so that white space is ignored):
1940 .sp
1941 \e( ( (?>[^()]+) | (?R) )* \e)
1942 .sp
1943 First it matches an opening parenthesis. Then it matches any number of
1944 substrings which can either be a sequence of non-parentheses, or a recursive
1945 match of the pattern itself (that is, a correctly parenthesized substring).
1946 Finally there is a closing parenthesis.
1947 .P
1948 If this were part of a larger pattern, you would not want to recurse the entire
1949 pattern, so instead you could use this:
1950 .sp
1951 ( \e( ( (?>[^()]+) | (?1) )* \e) )
1952 .sp
1953 We have put the pattern into parentheses, and caused the recursion to refer to
1954 them instead of the whole pattern.
1955 .P
1956 In a larger pattern, keeping track of parenthesis numbers can be tricky. This
1957 is made easier by the use of relative references. (A Perl 5.10 feature.)
1958 Instead of (?1) in the pattern above you can write (?-2) to refer to the second
1959 most recently opened parentheses preceding the recursion. In other words, a
1960 negative number counts capturing parentheses leftwards from the point at which
1961 it is encountered.
1962 .P
1963 It is also possible to refer to subsequently opened parentheses, by writing
1964 references such as (?+2). However, these cannot be recursive because the
1965 reference is not inside the parentheses that are referenced. They are always
1966 "subroutine" calls, as described in the next section.
1967 .P
1968 An alternative approach is to use named parentheses instead. The Perl syntax
1969 for this is (?&name); PCRE's earlier syntax (?P>name) is also supported. We
1970 could rewrite the above example as follows:
1971 .sp
1972 (?<pn> \e( ( (?>[^()]+) | (?&pn) )* \e) )
1973 .sp
1974 If there is more than one subpattern with the same name, the earliest one is
1975 used.
1976 .P
1977 This particular example pattern that we have been looking at contains nested
1978 unlimited repeats, and so the use of atomic grouping for matching strings of
1979 non-parentheses is important when applying the pattern to strings that do not
1980 match. For example, when this pattern is applied to
1981 .sp
1982 (aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa()
1983 .sp
1984 it yields "no match" quickly. However, if atomic grouping is not used,
1985 the match runs for a very long time indeed because there are so many different
1986 ways the + and * repeats can carve up the subject, and all have to be tested
1987 before failure can be reported.
1988 .P
1989 At the end of a match, the values set for any capturing subpatterns are those
1990 from the outermost level of the recursion at which the subpattern value is set.
1991 If you want to obtain intermediate values, a callout function can be used (see
1992 below and the
1993 .\" HREF
1994 \fBpcrecallout\fP
1995 .\"
1996 documentation). If the pattern above is matched against
1997 .sp
1998 (ab(cd)ef)
1999 .sp
2000 the value for the capturing parentheses is "ef", which is the last value taken
2001 on at the top level. If additional parentheses are added, giving
2002 .sp
2003 \e( ( ( (?>[^()]+) | (?R) )* ) \e)
2004 ^ ^
2005 ^ ^
2006 .sp
2007 the string they capture is "ab(cd)ef", the contents of the top level
2008 parentheses. If there are more than 15 capturing parentheses in a pattern, PCRE
2009 has to obtain extra memory to store data during a recursion, which it does by
2010 using \fBpcre_malloc\fP, freeing it via \fBpcre_free\fP afterwards. If no
2011 memory can be obtained, the match fails with the PCRE_ERROR_NOMEMORY error.
2012 .P
2013 Do not confuse the (?R) item with the condition (R), which tests for recursion.
2014 Consider this pattern, which matches text in angle brackets, allowing for
2015 arbitrary nesting. Only digits are allowed in nested brackets (that is, when
2016 recursing), whereas any characters are permitted at the outer level.
2017 .sp
2018 < (?: (?(R) \ed++ | [^<>]*+) | (?R)) * >
2019 .sp
2020 In this pattern, (?(R) is the start of a conditional subpattern, with two
2021 different alternatives for the recursive and non-recursive cases. The (?R) item
2022 is the actual recursive call.
2023 .
2024 .
2025 .\" HTML <a name="subpatternsassubroutines"></a>
2027 .rs
2028 .sp
2029 If the syntax for a recursive subpattern reference (either by number or by
2030 name) is used outside the parentheses to which it refers, it operates like a
2031 subroutine in a programming language. The "called" subpattern may be defined
2032 before or after the reference. A numbered reference can be absolute or
2033 relative, as in these examples:
2034 .sp
2035 (...(absolute)...)...(?2)...
2036 (...(relative)...)...(?-1)...
2037 (...(?+1)...(relative)...
2038 .sp
2039 An earlier example pointed out that the pattern
2040 .sp
2041 (sens|respons)e and \e1ibility
2042 .sp
2043 matches "sense and sensibility" and "response and responsibility", but not
2044 "sense and responsibility". If instead the pattern
2045 .sp
2046 (sens|respons)e and (?1)ibility
2047 .sp
2048 is used, it does match "sense and responsibility" as well as the other two
2049 strings. Another example is given in the discussion of DEFINE above.
2050 .P
2051 Like recursive subpatterns, a "subroutine" call is always treated as an atomic
2052 group. That is, once it has matched some of the subject string, it is never
2053 re-entered, even if it contains untried alternatives and there is a subsequent
2054 matching failure.
2055 .P
2056 When a subpattern is used as a subroutine, processing options such as
2057 case-independence are fixed when the subpattern is defined. They cannot be
2058 changed for different calls. For example, consider this pattern:
2059 .sp
2060 (abc)(?i:(?-1))
2061 .sp
2062 It matches "abcabc". It does not match "abcABC" because the change of
2063 processing option does not affect the called subpattern.
2064 .
2065 .
2066 .\" HTML <a name="onigurumasubroutines"></a>
2068 .rs
2069 .sp
2070 For compatibility with Oniguruma, the non-Perl syntax \eg followed by a name or
2071 a number enclosed either in angle brackets or single quotes, is an alternative
2072 syntax for referencing a subpattern as a subroutine, possibly recursively. Here
2073 are two of the examples used above, rewritten using this syntax:
2074 .sp
2075 (?<pn> \e( ( (?>[^()]+) | \eg<pn> )* \e) )
2076 (sens|respons)e and \eg'1'ibility
2077 .sp
2078 PCRE supports an extension to Oniguruma: if a number is preceded by a
2079 plus or a minus sign it is taken as a relative reference. For example:
2080 .sp
2081 (abc)(?i:\eg<-1>)
2082 .sp
2083 Note that \eg{...} (Perl syntax) and \eg<...> (Oniguruma syntax) are \fInot\fP
2084 synonymous. The former is a back reference; the latter is a subroutine call.
2085 .
2086 .
2088 .rs
2089 .sp
2090 Perl has a feature whereby using the sequence (?{...}) causes arbitrary Perl
2091 code to be obeyed in the middle of matching a regular expression. This makes it
2092 possible, amongst other things, to extract different substrings that match the
2093 same pair of parentheses when there is a repetition.
2094 .P
2095 PCRE provides a similar feature, but of course it cannot obey arbitrary Perl
2096 code. The feature is called "callout". The caller of PCRE provides an external
2097 function by putting its entry point in the global variable \fIpcre_callout\fP.
2098 By default, this variable contains NULL, which disables all calling out.
2099 .P
2100 Within a regular expression, (?C) indicates the points at which the external
2101 function is to be called. If you want to identify different callout points, you
2102 can put a number less than 256 after the letter C. The default value is zero.
2103 For example, this pattern has two callout points:
2104 .sp
2105 (?C1)abc(?C2)def
2106 .sp
2107 If the PCRE_AUTO_CALLOUT flag is passed to \fBpcre_compile()\fP, callouts are
2108 automatically installed before each item in the pattern. They are all numbered
2109 255.
2110 .P
2111 During matching, when PCRE reaches a callout point (and \fIpcre_callout\fP is
2112 set), the external function is called. It is provided with the number of the
2113 callout, the position in the pattern, and, optionally, one item of data
2114 originally supplied by the caller of \fBpcre_exec()\fP. The callout function
2115 may cause matching to proceed, to backtrack, or to fail altogether. A complete
2116 description of the interface to the callout function is given in the
2117 .\" HREF
2118 \fBpcrecallout\fP
2119 .\"
2120 documentation.
2121 .
2122 .
2124 .rs
2125 .sp
2126 Perl 5.10 introduced a number of "Special Backtracking Control Verbs", which
2127 are described in the Perl documentation as "experimental and subject to change
2128 or removal in a future version of Perl". It goes on to say: "Their usage in
2129 production code should be noted to avoid problems during upgrades." The same
2130 remarks apply to the PCRE features described in this section.
2131 .P
2132 Since these verbs are specifically related to backtracking, most of them can be
2133 used only when the pattern is to be matched using \fBpcre_exec()\fP, which uses
2134 a backtracking algorithm. With the exception of (*FAIL), which behaves like a
2135 failing negative assertion, they cause an error if encountered by
2136 \fBpcre_dfa_exec()\fP.
2137 .P
2138 The new verbs make use of what was previously invalid syntax: an opening
2139 parenthesis followed by an asterisk. In Perl, they are generally of the form
2140 (*VERB:ARG) but PCRE does not support the use of arguments, so its general
2141 form is just (*VERB). Any number of these verbs may occur in a pattern. There
2142 are two kinds:
2143 .
2144 .SS "Verbs that act immediately"
2145 .rs
2146 .sp
2147 The following verbs act as soon as they are encountered:
2148 .sp
2149 (*ACCEPT)
2150 .sp
2151 This verb causes the match to end successfully, skipping the remainder of the
2152 pattern. When inside a recursion, only the innermost pattern is ended
2153 immediately. PCRE differs from Perl in what happens if the (*ACCEPT) is inside
2154 capturing parentheses. In Perl, the data so far is captured: in PCRE no data is
2155 captured. For example:
2156 .sp
2157 A(A|B(*ACCEPT)|C)D
2158 .sp
2159 This matches "AB", "AAD", or "ACD", but when it matches "AB", no data is
2160 captured.
2161 .sp
2162 (*FAIL) or (*F)
2163 .sp
2164 This verb causes the match to fail, forcing backtracking to occur. It is
2165 equivalent to (?!) but easier to read. The Perl documentation notes that it is
2166 probably useful only when combined with (?{}) or (??{}). Those are, of course,
2167 Perl features that are not present in PCRE. The nearest equivalent is the
2168 callout feature, as for example in this pattern:
2169 .sp
2170 a+(?C)(*FAIL)
2171 .sp
2172 A match with the string "aaaa" always fails, but the callout is taken before
2173 each backtrack happens (in this example, 10 times).
2174 .
2175 .SS "Verbs that act after backtracking"
2176 .rs
2177 .sp
2178 The following verbs do nothing when they are encountered. Matching continues
2179 with what follows, but if there is no subsequent match, a failure is forced.
2180 The verbs differ in exactly what kind of failure occurs.
2181 .sp
2182 (*COMMIT)
2183 .sp
2184 This verb causes the whole match to fail outright if the rest of the pattern
2185 does not match. Even if the pattern is unanchored, no further attempts to find
2186 a match by advancing the start point take place. Once (*COMMIT) has been
2187 passed, \fBpcre_exec()\fP is committed to finding a match at the current
2188 starting point, or not at all. For example:
2189 .sp
2190 a+(*COMMIT)b
2191 .sp
2192 This matches "xxaab" but not "aacaab". It can be thought of as a kind of
2193 dynamic anchor, or "I've started, so I must finish."
2194 .sp
2195 (*PRUNE)
2196 .sp
2197 This verb causes the match to fail at the current position if the rest of the
2198 pattern does not match. If the pattern is unanchored, the normal "bumpalong"
2199 advance to the next starting character then happens. Backtracking can occur as
2200 usual to the left of (*PRUNE), or when matching to the right of (*PRUNE), but
2201 if there is no match to the right, backtracking cannot cross (*PRUNE).
2202 In simple cases, the use of (*PRUNE) is just an alternative to an atomic
2203 group or possessive quantifier, but there are some uses of (*PRUNE) that cannot
2204 be expressed in any other way.
2205 .sp
2206 (*SKIP)
2207 .sp
2208 This verb is like (*PRUNE), except that if the pattern is unanchored, the
2209 "bumpalong" advance is not to the next character, but to the position in the
2210 subject where (*SKIP) was encountered. (*SKIP) signifies that whatever text
2211 was matched leading up to it cannot be part of a successful match. Consider:
2212 .sp
2213 a+(*SKIP)b
2214 .sp
2215 If the subject is "aaaac...", after the first match attempt fails (starting at
2216 the first character in the string), the starting point skips on to start the
2217 next attempt at "c". Note that a possessive quantifer does not have the same
2218 effect in this example; although it would suppress backtracking during the
2219 first match attempt, the second attempt would start at the second character
2220 instead of skipping on to "c".
2221 .sp
2222 (*THEN)
2223 .sp
2224 This verb causes a skip to the next alternation if the rest of the pattern does
2225 not match. That is, it cancels pending backtracking, but only within the
2226 current alternation. Its name comes from the observation that it can be used
2227 for a pattern-based if-then-else block:
2228 .sp
2229 ( COND1 (*THEN) FOO | COND2 (*THEN) BAR | COND3 (*THEN) BAZ ) ...
2230 .sp
2231 If the COND1 pattern matches, FOO is tried (and possibly further items after
2232 the end of the group if FOO succeeds); on failure the matcher skips to the
2233 second alternative and tries COND2, without backtracking into COND1. If (*THEN)
2234 is used outside of any alternation, it acts exactly like (*PRUNE).
2235 .
2236 .
2237 .SH "SEE ALSO"
2238 .rs
2239 .sp
2240 \fBpcreapi\fP(3), \fBpcrecallout\fP(3), \fBpcrematching\fP(3), \fBpcre\fP(3).
2241 .
2242 .
2244 .rs
2245 .sp
2246 .nf
2247 Philip Hazel
2248 University Computing Service
2249 Cambridge CB2 3QH, England.
2250 .fi
2251 .
2252 .
2254 .rs
2255 .sp
2256 .nf
2257 Last updated: 11 April 2009
2258 Copyright (c) 1997-2009 University of Cambridge.
2259 .fi


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

  ViewVC Help
Powered by ViewVC 1.1.5