ViewVC logotype

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

Parent Directory Parent Directory | Revision Log Revision Log

Revision 254 - (show annotations) (download) (as text)
Mon Sep 17 11:05:37 2007 UTC (6 years, 7 months ago) by ph10
File MIME type: text/html
File size: 93197 byte(s)
Update files tidied by making a release.

1 <html>
2 <head>
3 <title>pcrepattern specification</title>
4 </head>
5 <body bgcolor="#FFFFFF" text="#00005A" link="#0066FF" alink="#3399FF" vlink="#2222BB">
6 <h1>pcrepattern man page</h1>
7 <p>
8 Return to the <a href="index.html">PCRE index page</a>.
9 </p>
10 <p>
11 This page is part of the PCRE HTML documentation. It was generated automatically
12 from the original man page. If there is any nonsense in it, please consult the
13 man page, in case the conversion went wrong.
14 <br>
15 <ul>
16 <li><a name="TOC1" href="#SEC1">PCRE REGULAR EXPRESSION DETAILS</a>
17 <li><a name="TOC2" href="#SEC2">NEWLINE CONVENTIONS</a>
18 <li><a name="TOC3" href="#SEC3">CHARACTERS AND METACHARACTERS</a>
19 <li><a name="TOC4" href="#SEC4">BACKSLASH</a>
20 <li><a name="TOC5" href="#SEC5">CIRCUMFLEX AND DOLLAR</a>
21 <li><a name="TOC6" href="#SEC6">FULL STOP (PERIOD, DOT)</a>
22 <li><a name="TOC7" href="#SEC7">MATCHING A SINGLE BYTE</a>
24 <li><a name="TOC9" href="#SEC9">POSIX CHARACTER CLASSES</a>
25 <li><a name="TOC10" href="#SEC10">VERTICAL BAR</a>
26 <li><a name="TOC11" href="#SEC11">INTERNAL OPTION SETTING</a>
27 <li><a name="TOC12" href="#SEC12">SUBPATTERNS</a>
28 <li><a name="TOC13" href="#SEC13">DUPLICATE SUBPATTERN NUMBERS</a>
29 <li><a name="TOC14" href="#SEC14">NAMED SUBPATTERNS</a>
30 <li><a name="TOC15" href="#SEC15">REPETITION</a>
32 <li><a name="TOC17" href="#SEC17">BACK REFERENCES</a>
33 <li><a name="TOC18" href="#SEC18">ASSERTIONS</a>
34 <li><a name="TOC19" href="#SEC19">CONDITIONAL SUBPATTERNS</a>
35 <li><a name="TOC20" href="#SEC20">COMMENTS</a>
36 <li><a name="TOC21" href="#SEC21">RECURSIVE PATTERNS</a>
37 <li><a name="TOC22" href="#SEC22">SUBPATTERNS AS SUBROUTINES</a>
38 <li><a name="TOC23" href="#SEC23">CALLOUTS</a>
39 <li><a name="TOC24" href="#SEC24">BACKTRACKING CONTROL</a>
40 <li><a name="TOC25" href="#SEC25">SEE ALSO</a>
41 <li><a name="TOC26" href="#SEC26">AUTHOR</a>
42 <li><a name="TOC27" href="#SEC27">REVISION</a>
43 </ul>
44 <br><a name="SEC1" href="#TOC1">PCRE REGULAR EXPRESSION DETAILS</a><br>
45 <P>
46 The syntax and semantics of the regular expressions that are supported by PCRE
47 are described in detail below. There is a quick-reference syntax summary in the
48 <a href="pcresyntax.html"><b>pcresyntax</b></a>
49 page. Perl's regular expressions are described in its own documentation, and
50 regular expressions in general are covered in a number of books, some of which
51 have copious examples. Jeffrey Friedl's "Mastering Regular Expressions",
52 published by O'Reilly, covers regular expressions in great detail. This
53 description of PCRE's regular expressions is intended as reference material.
54 </P>
55 <P>
56 The original operation of PCRE was on strings of one-byte characters. However,
57 there is now also support for UTF-8 character strings. To use this, you must
58 build PCRE to include UTF-8 support, and then call <b>pcre_compile()</b> with
59 the PCRE_UTF8 option. How this affects pattern matching is mentioned in several
60 places below. There is also a summary of UTF-8 features in the
61 <a href="pcre.html#utf8support">section on UTF-8 support</a>
62 in the main
63 <a href="pcre.html"><b>pcre</b></a>
64 page.
65 </P>
66 <P>
67 The remainder of this document discusses the patterns that are supported by
68 PCRE when its main matching function, <b>pcre_exec()</b>, is used.
69 From release 6.0, PCRE offers a second matching function,
70 <b>pcre_dfa_exec()</b>, which matches using a different algorithm that is not
71 Perl-compatible. Some of the features discussed below are not available when
72 <b>pcre_dfa_exec()</b> is used. The advantages and disadvantages of the
73 alternative function, and how it differs from the normal function, are
74 discussed in the
75 <a href="pcrematching.html"><b>pcrematching</b></a>
76 page.
77 </P>
78 <br><a name="SEC2" href="#TOC1">NEWLINE CONVENTIONS</a><br>
79 <P>
80 PCRE supports five different conventions for indicating line breaks in
81 strings: a single CR (carriage return) character, a single LF (linefeed)
82 character, the two-character sequence CRLF, any of the three preceding, or any
83 Unicode newline sequence. The
84 <a href="pcreapi.html"><b>pcreapi</b></a>
85 page has
86 <a href="pcreapi.html#newlines">further discussion</a>
87 about newlines, and shows how to set the newline convention in the
88 <i>options</i> arguments for the compiling and matching functions.
89 </P>
90 <P>
91 It is also possible to specify a newline convention by starting a pattern
92 string with one of the following five sequences:
93 <pre>
94 (*CR) carriage return
95 (*LF) linefeed
96 (*CRLF) carriage return, followed by linefeed
97 (*ANYCRLF) any of the three above
98 (*ANY) all Unicode newline sequences
99 </pre>
100 These override the default and the options given to <b>pcre_compile()</b>. For
101 example, on a Unix system where LF is the default newline sequence, the pattern
102 <pre>
103 (*CR)a.b
104 </pre>
105 changes the convention to CR. That pattern matches "a\nb" because LF is no
106 longer a newline. Note that these special settings, which are not
107 Perl-compatible, are recognized only at the very start of a pattern, and that
108 they must be in upper case. If more than one of them is present, the last one
109 is used.
110 </P>
111 <P>
112 The newline convention does not affect what the \R escape sequence matches. By
113 default, this is any Unicode newline sequence, for Perl compatibility. However,
114 this can be changed; see the description of \R in the section entitled
115 <a href="#newlineseq">"Newline sequences"</a>
116 below. A change of \R setting can be combined with a change of newline
117 convention.
118 </P>
119 <br><a name="SEC3" href="#TOC1">CHARACTERS AND METACHARACTERS</a><br>
120 <P>
121 A regular expression is a pattern that is matched against a subject string from
122 left to right. Most characters stand for themselves in a pattern, and match the
123 corresponding characters in the subject. As a trivial example, the pattern
124 <pre>
125 The quick brown fox
126 </pre>
127 matches a portion of a subject string that is identical to itself. When
128 caseless matching is specified (the PCRE_CASELESS option), letters are matched
129 independently of case. In UTF-8 mode, PCRE always understands the concept of
130 case for characters whose values are less than 128, so caseless matching is
131 always possible. For characters with higher values, the concept of case is
132 supported if PCRE is compiled with Unicode property support, but not otherwise.
133 If you want to use caseless matching for characters 128 and above, you must
134 ensure that PCRE is compiled with Unicode property support as well as with
135 UTF-8 support.
136 </P>
137 <P>
138 The power of regular expressions comes from the ability to include alternatives
139 and repetitions in the pattern. These are encoded in the pattern by the use of
140 <i>metacharacters</i>, which do not stand for themselves but instead are
141 interpreted in some special way.
142 </P>
143 <P>
144 There are two different sets of metacharacters: those that are recognized
145 anywhere in the pattern except within square brackets, and those that are
146 recognized within square brackets. Outside square brackets, the metacharacters
147 are as follows:
148 <pre>
149 \ general escape character with several uses
150 ^ assert start of string (or line, in multiline mode)
151 $ assert end of string (or line, in multiline mode)
152 . match any character except newline (by default)
153 [ start character class definition
154 | start of alternative branch
155 ( start subpattern
156 ) end subpattern
157 ? extends the meaning of (
158 also 0 or 1 quantifier
159 also quantifier minimizer
160 * 0 or more quantifier
161 + 1 or more quantifier
162 also "possessive quantifier"
163 { start min/max quantifier
164 </pre>
165 Part of a pattern that is in square brackets is called a "character class". In
166 a character class the only metacharacters are:
167 <pre>
168 \ general escape character
169 ^ negate the class, but only if the first character
170 - indicates character range
171 [ POSIX character class (only if followed by POSIX syntax)
172 ] terminates the character class
173 </pre>
174 The following sections describe the use of each of the metacharacters.
175 </P>
176 <br><a name="SEC4" href="#TOC1">BACKSLASH</a><br>
177 <P>
178 The backslash character has several uses. Firstly, if it is followed by a
179 non-alphanumeric character, it takes away any special meaning that character
180 may have. This use of backslash as an escape character applies both inside and
181 outside character classes.
182 </P>
183 <P>
184 For example, if you want to match a * character, you write \* in the pattern.
185 This escaping action applies whether or not the following character would
186 otherwise be interpreted as a metacharacter, so it is always safe to precede a
187 non-alphanumeric with backslash to specify that it stands for itself. In
188 particular, if you want to match a backslash, you write \\.
189 </P>
190 <P>
191 If a pattern is compiled with the PCRE_EXTENDED option, whitespace in the
192 pattern (other than in a character class) and characters between a # outside
193 a character class and the next newline are ignored. An escaping backslash can
194 be used to include a whitespace or # character as part of the pattern.
195 </P>
196 <P>
197 If you want to remove the special meaning from a sequence of characters, you
198 can do so by putting them between \Q and \E. This is different from Perl in
199 that $ and @ are handled as literals in \Q...\E sequences in PCRE, whereas in
200 Perl, $ and @ cause variable interpolation. Note the following examples:
201 <pre>
202 Pattern PCRE matches Perl matches
204 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
205 \Qabc\$xyz\E abc\$xyz abc\$xyz
206 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
207 </pre>
208 The \Q...\E sequence is recognized both inside and outside character classes.
209 <a name="digitsafterbackslash"></a></P>
210 <br><b>
211 Non-printing characters
212 </b><br>
213 <P>
214 A second use of backslash provides a way of encoding non-printing characters
215 in patterns in a visible manner. There is no restriction on the appearance of
216 non-printing characters, apart from the binary zero that terminates a pattern,
217 but when a pattern is being prepared by text editing, it is usually easier to
218 use one of the following escape sequences than the binary character it
219 represents:
220 <pre>
221 \a alarm, that is, the BEL character (hex 07)
222 \cx "control-x", where x is any character
223 \e escape (hex 1B)
224 \f formfeed (hex 0C)
225 \n linefeed (hex 0A)
226 \r carriage return (hex 0D)
227 \t tab (hex 09)
228 \ddd character with octal code ddd, or backreference
229 \xhh character with hex code hh
230 \x{hhh..} character with hex code hhh..
231 </pre>
232 The precise effect of \cx is as follows: if x is a lower case letter, it
233 is converted to upper case. Then bit 6 of the character (hex 40) is inverted.
234 Thus \cz becomes hex 1A, but \c{ becomes hex 3B, while \c; becomes hex
235 7B.
236 </P>
237 <P>
238 After \x, from zero to two hexadecimal digits are read (letters can be in
239 upper or lower case). Any number of hexadecimal digits may appear between \x{
240 and }, but the value of the character code must be less than 256 in non-UTF-8
241 mode, and less than 2**31 in UTF-8 mode. That is, the maximum value in
242 hexadecimal is 7FFFFFFF. Note that this is bigger than the largest Unicode code
243 point, which is 10FFFF.
244 </P>
245 <P>
246 If characters other than hexadecimal digits appear between \x{ and }, or if
247 there is no terminating }, this form of escape is not recognized. Instead, the
248 initial \x will be interpreted as a basic hexadecimal escape, with no
249 following digits, giving a character whose value is zero.
250 </P>
251 <P>
252 Characters whose value is less than 256 can be defined by either of the two
253 syntaxes for \x. There is no difference in the way they are handled. For
254 example, \xdc is exactly the same as \x{dc}.
255 </P>
256 <P>
257 After \0 up to two further octal digits are read. If there are fewer than two
258 digits, just those that are present are used. Thus the sequence \0\x\07
259 specifies two binary zeros followed by a BEL character (code value 7). Make
260 sure you supply two digits after the initial zero if the pattern character that
261 follows is itself an octal digit.
262 </P>
263 <P>
264 The handling of a backslash followed by a digit other than 0 is complicated.
265 Outside a character class, PCRE reads it and any following digits as a decimal
266 number. If the number is less than 10, or if there have been at least that many
267 previous capturing left parentheses in the expression, the entire sequence is
268 taken as a <i>back reference</i>. A description of how this works is given
269 <a href="#backreferences">later,</a>
270 following the discussion of
271 <a href="#subpattern">parenthesized subpatterns.</a>
272 </P>
273 <P>
274 Inside a character class, or if the decimal number is greater than 9 and there
275 have not been that many capturing subpatterns, PCRE re-reads up to three octal
276 digits following the backslash, and uses them to generate a data character. Any
277 subsequent digits stand for themselves. In non-UTF-8 mode, the value of a
278 character specified in octal must be less than \400. In UTF-8 mode, values up
279 to \777 are permitted. For example:
280 <pre>
281 \040 is another way of writing a space
282 \40 is the same, provided there are fewer than 40 previous capturing subpatterns
283 \7 is always a back reference
284 \11 might be a back reference, or another way of writing a tab
285 \011 is always a tab
286 \0113 is a tab followed by the character "3"
287 \113 might be a back reference, otherwise the character with octal code 113
288 \377 might be a back reference, otherwise the byte consisting entirely of 1 bits
289 \81 is either a back reference, or a binary zero followed by the two characters "8" and "1"
290 </pre>
291 Note that octal values of 100 or greater must not be introduced by a leading
292 zero, because no more than three octal digits are ever read.
293 </P>
294 <P>
295 All the sequences that define a single character value can be used both inside
296 and outside character classes. In addition, inside a character class, the
297 sequence \b is interpreted as the backspace character (hex 08), and the
298 sequences \R and \X are interpreted as the characters "R" and "X",
299 respectively. Outside a character class, these sequences have different
300 meanings
301 <a href="#uniextseq">(see below).</a>
302 </P>
303 <br><b>
304 Absolute and relative back references
305 </b><br>
306 <P>
307 The sequence \g followed by an unsigned or a negative number, optionally
308 enclosed in braces, is an absolute or relative back reference. A named back
309 reference can be coded as \g{name}. Back references are discussed
310 <a href="#backreferences">later,</a>
311 following the discussion of
312 <a href="#subpattern">parenthesized subpatterns.</a>
313 </P>
314 <br><b>
315 Generic character types
316 </b><br>
317 <P>
318 Another use of backslash is for specifying generic character types. The
319 following are always recognized:
320 <pre>
321 \d any decimal digit
322 \D any character that is not a decimal digit
323 \h any horizontal whitespace character
324 \H any character that is not a horizontal whitespace character
325 \s any whitespace character
326 \S any character that is not a whitespace character
327 \v any vertical whitespace character
328 \V any character that is not a vertical whitespace character
329 \w any "word" character
330 \W any "non-word" character
331 </pre>
332 Each pair of escape sequences partitions the complete set of characters into
333 two disjoint sets. Any given character matches one, and only one, of each pair.
334 </P>
335 <P>
336 These character type sequences can appear both inside and outside character
337 classes. They each match one character of the appropriate type. If the current
338 matching point is at the end of the subject string, all of them fail, since
339 there is no character to match.
340 </P>
341 <P>
342 For compatibility with Perl, \s does not match the VT character (code 11).
343 This makes it different from the the POSIX "space" class. The \s characters
344 are HT (9), LF (10), FF (12), CR (13), and space (32). If "use locale;" is
345 included in a Perl script, \s may match the VT character. In PCRE, it never
346 does.
347 </P>
348 <P>
349 In UTF-8 mode, characters with values greater than 128 never match \d, \s, or
350 \w, and always match \D, \S, and \W. This is true even when Unicode
351 character property support is available. These sequences retain their original
352 meanings from before UTF-8 support was available, mainly for efficiency
353 reasons.
354 </P>
355 <P>
356 The sequences \h, \H, \v, and \V are Perl 5.10 features. In contrast to the
357 other sequences, these do match certain high-valued codepoints in UTF-8 mode.
358 The horizontal space characters are:
359 <pre>
360 U+0009 Horizontal tab
361 U+0020 Space
362 U+00A0 Non-break space
363 U+1680 Ogham space mark
364 U+180E Mongolian vowel separator
365 U+2000 En quad
366 U+2001 Em quad
367 U+2002 En space
368 U+2003 Em space
369 U+2004 Three-per-em space
370 U+2005 Four-per-em space
371 U+2006 Six-per-em space
372 U+2007 Figure space
373 U+2008 Punctuation space
374 U+2009 Thin space
375 U+200A Hair space
376 U+202F Narrow no-break space
377 U+205F Medium mathematical space
378 U+3000 Ideographic space
379 </pre>
380 The vertical space characters are:
381 <pre>
382 U+000A Linefeed
383 U+000B Vertical tab
384 U+000C Formfeed
385 U+000D Carriage return
386 U+0085 Next line
387 U+2028 Line separator
388 U+2029 Paragraph separator
389 </PRE>
390 </P>
391 <P>
392 A "word" character is an underscore or any character less than 256 that is a
393 letter or digit. The definition of letters and digits is controlled by PCRE's
394 low-valued character tables, and may vary if locale-specific matching is taking
395 place (see
396 <a href="pcreapi.html#localesupport">"Locale support"</a>
397 in the
398 <a href="pcreapi.html"><b>pcreapi</b></a>
399 page). For example, in a French locale such as "fr_FR" in Unix-like systems,
400 or "french" in Windows, some character codes greater than 128 are used for
401 accented letters, and these are matched by \w. The use of locales with Unicode
402 is discouraged.
403 <a name="newlineseq"></a></P>
404 <br><b>
405 Newline sequences
406 </b><br>
407 <P>
408 Outside a character class, by default, the escape sequence \R matches any
409 Unicode newline sequence. This is a Perl 5.10 feature. In non-UTF-8 mode \R is
410 equivalent to the following:
411 <pre>
412 (?&#62;\r\n|\n|\x0b|\f|\r|\x85)
413 </pre>
414 This is an example of an "atomic group", details of which are given
415 <a href="#atomicgroup">below.</a>
416 This particular group matches either the two-character sequence CR followed by
417 LF, or one of the single characters LF (linefeed, U+000A), VT (vertical tab,
418 U+000B), FF (formfeed, U+000C), CR (carriage return, U+000D), or NEL (next
419 line, U+0085). The two-character sequence is treated as a single unit that
420 cannot be split.
421 </P>
422 <P>
423 In UTF-8 mode, two additional characters whose codepoints are greater than 255
424 are added: LS (line separator, U+2028) and PS (paragraph separator, U+2029).
425 Unicode character property support is not needed for these characters to be
426 recognized.
427 </P>
428 <P>
429 It is possible to restrict \R to match only CR, LF, or CRLF (instead of the
430 complete set of Unicode line endings) by setting the option PCRE_BSR_ANYCRLF
431 either at compile time or when the pattern is matched. (BSR is an abbrevation
432 for "backslash R".) This can be made the default when PCRE is built; if this is
433 the case, the other behaviour can be requested via the PCRE_BSR_UNICODE option.
434 It is also possible to specify these settings by starting a pattern string with
435 one of the following sequences:
436 <pre>
437 (*BSR_ANYCRLF) CR, LF, or CRLF only
438 (*BSR_UNICODE) any Unicode newline sequence
439 </pre>
440 These override the default and the options given to <b>pcre_compile()</b>, but
441 they can be overridden by options given to <b>pcre_exec()</b>. Note that these
442 special settings, which are not Perl-compatible, are recognized only at the
443 very start of a pattern, and that they must be in upper case. If more than one
444 of them is present, the last one is used. They can be combined with a change of
445 newline convention, for example, a pattern can start with:
446 <pre>
448 </pre>
449 Inside a character class, \R matches the letter "R".
450 <a name="uniextseq"></a></P>
451 <br><b>
452 Unicode character properties
453 </b><br>
454 <P>
455 When PCRE is built with Unicode character property support, three additional
456 escape sequences that match characters with specific properties are available.
457 When not in UTF-8 mode, these sequences are of course limited to testing
458 characters whose codepoints are less than 256, but they do work in this mode.
459 The extra escape sequences are:
460 <pre>
461 \p{<i>xx</i>} a character with the <i>xx</i> property
462 \P{<i>xx</i>} a character without the <i>xx</i> property
463 \X an extended Unicode sequence
464 </pre>
465 The property names represented by <i>xx</i> above are limited to the Unicode
466 script names, the general category properties, and "Any", which matches any
467 character (including newline). Other properties such as "InMusicalSymbols" are
468 not currently supported by PCRE. Note that \P{Any} does not match any
469 characters, so always causes a match failure.
470 </P>
471 <P>
472 Sets of Unicode characters are defined as belonging to certain scripts. A
473 character from one of these sets can be matched using a script name. For
474 example:
475 <pre>
476 \p{Greek}
477 \P{Han}
478 </pre>
479 Those that are not part of an identified script are lumped together as
480 "Common". The current list of scripts is:
481 </P>
482 <P>
483 Arabic,
484 Armenian,
485 Balinese,
486 Bengali,
487 Bopomofo,
488 Braille,
489 Buginese,
490 Buhid,
491 Canadian_Aboriginal,
492 Cherokee,
493 Common,
494 Coptic,
495 Cuneiform,
496 Cypriot,
497 Cyrillic,
498 Deseret,
499 Devanagari,
500 Ethiopic,
501 Georgian,
502 Glagolitic,
503 Gothic,
504 Greek,
505 Gujarati,
506 Gurmukhi,
507 Han,
508 Hangul,
509 Hanunoo,
510 Hebrew,
511 Hiragana,
512 Inherited,
513 Kannada,
514 Katakana,
515 Kharoshthi,
516 Khmer,
517 Lao,
518 Latin,
519 Limbu,
520 Linear_B,
521 Malayalam,
522 Mongolian,
523 Myanmar,
524 New_Tai_Lue,
525 Nko,
526 Ogham,
527 Old_Italic,
528 Old_Persian,
529 Oriya,
530 Osmanya,
531 Phags_Pa,
532 Phoenician,
533 Runic,
534 Shavian,
535 Sinhala,
536 Syloti_Nagri,
537 Syriac,
538 Tagalog,
539 Tagbanwa,
540 Tai_Le,
541 Tamil,
542 Telugu,
543 Thaana,
544 Thai,
545 Tibetan,
546 Tifinagh,
547 Ugaritic,
548 Yi.
549 </P>
550 <P>
551 Each character has exactly one general category property, specified by a
552 two-letter abbreviation. For compatibility with Perl, negation can be specified
553 by including a circumflex between the opening brace and the property name. For
554 example, \p{^Lu} is the same as \P{Lu}.
555 </P>
556 <P>
557 If only one letter is specified with \p or \P, it includes all the general
558 category properties that start with that letter. In this case, in the absence
559 of negation, the curly brackets in the escape sequence are optional; these two
560 examples have the same effect:
561 <pre>
562 \p{L}
563 \pL
564 </pre>
565 The following general category property codes are supported:
566 <pre>
567 C Other
568 Cc Control
569 Cf Format
570 Cn Unassigned
571 Co Private use
572 Cs Surrogate
574 L Letter
575 Ll Lower case letter
576 Lm Modifier letter
577 Lo Other letter
578 Lt Title case letter
579 Lu Upper case letter
581 M Mark
582 Mc Spacing mark
583 Me Enclosing mark
584 Mn Non-spacing mark
586 N Number
587 Nd Decimal number
588 Nl Letter number
589 No Other number
591 P Punctuation
592 Pc Connector punctuation
593 Pd Dash punctuation
594 Pe Close punctuation
595 Pf Final punctuation
596 Pi Initial punctuation
597 Po Other punctuation
598 Ps Open punctuation
600 S Symbol
601 Sc Currency symbol
602 Sk Modifier symbol
603 Sm Mathematical symbol
604 So Other symbol
606 Z Separator
607 Zl Line separator
608 Zp Paragraph separator
609 Zs Space separator
610 </pre>
611 The special property L& is also supported: it matches a character that has
612 the Lu, Ll, or Lt property, in other words, a letter that is not classified as
613 a modifier or "other".
614 </P>
615 <P>
616 The Cs (Surrogate) property applies only to characters in the range U+D800 to
617 U+DFFF. Such characters are not valid in UTF-8 strings (see RFC 3629) and so
618 cannot be tested by PCRE, unless UTF-8 validity checking has been turned off
619 (see the discussion of PCRE_NO_UTF8_CHECK in the
620 <a href="pcreapi.html"><b>pcreapi</b></a>
621 page).
622 </P>
623 <P>
624 The long synonyms for these properties that Perl supports (such as \p{Letter})
625 are not supported by PCRE, nor is it permitted to prefix any of these
626 properties with "Is".
627 </P>
628 <P>
629 No character that is in the Unicode table has the Cn (unassigned) property.
630 Instead, this property is assumed for any code point that is not in the
631 Unicode table.
632 </P>
633 <P>
634 Specifying caseless matching does not affect these escape sequences. For
635 example, \p{Lu} always matches only upper case letters.
636 </P>
637 <P>
638 The \X escape matches any number of Unicode characters that form an extended
639 Unicode sequence. \X is equivalent to
640 <pre>
641 (?&#62;\PM\pM*)
642 </pre>
643 That is, it matches a character without the "mark" property, followed by zero
644 or more characters with the "mark" property, and treats the sequence as an
645 atomic group
646 <a href="#atomicgroup">(see below).</a>
647 Characters with the "mark" property are typically accents that affect the
648 preceding character. None of them have codepoints less than 256, so in
649 non-UTF-8 mode \X matches any one character.
650 </P>
651 <P>
652 Matching characters by Unicode property is not fast, because PCRE has to search
653 a structure that contains data for over fifteen thousand characters. That is
654 why the traditional escape sequences such as \d and \w do not use Unicode
655 properties in PCRE.
656 <a name="resetmatchstart"></a></P>
657 <br><b>
658 Resetting the match start
659 </b><br>
660 <P>
661 The escape sequence \K, which is a Perl 5.10 feature, causes any previously
662 matched characters not to be included in the final matched sequence. For
663 example, the pattern:
664 <pre>
665 foo\Kbar
666 </pre>
667 matches "foobar", but reports that it has matched "bar". This feature is
668 similar to a lookbehind assertion
669 <a href="#lookbehind">(described below).</a>
670 However, in this case, the part of the subject before the real match does not
671 have to be of fixed length, as lookbehind assertions do. The use of \K does
672 not interfere with the setting of
673 <a href="#subpattern">captured substrings.</a>
674 For example, when the pattern
675 <pre>
676 (foo)\Kbar
677 </pre>
678 matches "foobar", the first substring is still set to "foo".
679 <a name="smallassertions"></a></P>
680 <br><b>
681 Simple assertions
682 </b><br>
683 <P>
684 The final use of backslash is for certain simple assertions. An assertion
685 specifies a condition that has to be met at a particular point in a match,
686 without consuming any characters from the subject string. The use of
687 subpatterns for more complicated assertions is described
688 <a href="#bigassertions">below.</a>
689 The backslashed assertions are:
690 <pre>
691 \b matches at a word boundary
692 \B matches when not at a word boundary
693 \A matches at the start of the subject
694 \Z matches at the end of the subject
695 also matches before a newline at the end of the subject
696 \z matches only at the end of the subject
697 \G matches at the first matching position in the subject
698 </pre>
699 These assertions may not appear in character classes (but note that \b has a
700 different meaning, namely the backspace character, inside a character class).
701 </P>
702 <P>
703 A word boundary is a position in the subject string where the current character
704 and the previous character do not both match \w or \W (i.e. one matches
705 \w and the other matches \W), or the start or end of the string if the
706 first or last character matches \w, respectively.
707 </P>
708 <P>
709 The \A, \Z, and \z assertions differ from the traditional circumflex and
710 dollar (described in the next section) in that they only ever match at the very
711 start and end of the subject string, whatever options are set. Thus, they are
712 independent of multiline mode. These three assertions are not affected by the
713 PCRE_NOTBOL or PCRE_NOTEOL options, which affect only the behaviour of the
714 circumflex and dollar metacharacters. However, if the <i>startoffset</i>
715 argument of <b>pcre_exec()</b> is non-zero, indicating that matching is to start
716 at a point other than the beginning of the subject, \A can never match. The
717 difference between \Z and \z is that \Z matches before a newline at the end
718 of the string as well as at the very end, whereas \z matches only at the end.
719 </P>
720 <P>
721 The \G assertion is true only when the current matching position is at the
722 start point of the match, as specified by the <i>startoffset</i> argument of
723 <b>pcre_exec()</b>. It differs from \A when the value of <i>startoffset</i> is
724 non-zero. By calling <b>pcre_exec()</b> multiple times with appropriate
725 arguments, you can mimic Perl's /g option, and it is in this kind of
726 implementation where \G can be useful.
727 </P>
728 <P>
729 Note, however, that PCRE's interpretation of \G, as the start of the current
730 match, is subtly different from Perl's, which defines it as the end of the
731 previous match. In Perl, these can be different when the previously matched
732 string was empty. Because PCRE does just one match at a time, it cannot
733 reproduce this behaviour.
734 </P>
735 <P>
736 If all the alternatives of a pattern begin with \G, the expression is anchored
737 to the starting match position, and the "anchored" flag is set in the compiled
738 regular expression.
739 </P>
740 <br><a name="SEC5" href="#TOC1">CIRCUMFLEX AND DOLLAR</a><br>
741 <P>
742 Outside a character class, in the default matching mode, the circumflex
743 character is an assertion that is true only if the current matching point is
744 at the start of the subject string. If the <i>startoffset</i> argument of
745 <b>pcre_exec()</b> is non-zero, circumflex can never match if the PCRE_MULTILINE
746 option is unset. Inside a character class, circumflex has an entirely different
747 meaning
748 <a href="#characterclass">(see below).</a>
749 </P>
750 <P>
751 Circumflex need not be the first character of the pattern if a number of
752 alternatives are involved, but it should be the first thing in each alternative
753 in which it appears if the pattern is ever to match that branch. If all
754 possible alternatives start with a circumflex, that is, if the pattern is
755 constrained to match only at the start of the subject, it is said to be an
756 "anchored" pattern. (There are also other constructs that can cause a pattern
757 to be anchored.)
758 </P>
759 <P>
760 A dollar character is an assertion that is true only if the current matching
761 point is at the end of the subject string, or immediately before a newline
762 at the end of the string (by default). Dollar need not be the last character of
763 the pattern if a number of alternatives are involved, but it should be the last
764 item in any branch in which it appears. Dollar has no special meaning in a
765 character class.
766 </P>
767 <P>
768 The meaning of dollar can be changed so that it matches only at the very end of
769 the string, by setting the PCRE_DOLLAR_ENDONLY option at compile time. This
770 does not affect the \Z assertion.
771 </P>
772 <P>
773 The meanings of the circumflex and dollar characters are changed if the
774 PCRE_MULTILINE option is set. When this is the case, a circumflex matches
775 immediately after internal newlines as well as at the start of the subject
776 string. It does not match after a newline that ends the string. A dollar
777 matches before any newlines in the string, as well as at the very end, when
778 PCRE_MULTILINE is set. When newline is specified as the two-character
779 sequence CRLF, isolated CR and LF characters do not indicate newlines.
780 </P>
781 <P>
782 For example, the pattern /^abc$/ matches the subject string "def\nabc" (where
783 \n represents a newline) in multiline mode, but not otherwise. Consequently,
784 patterns that are anchored in single line mode because all branches start with
785 ^ are not anchored in multiline mode, and a match for circumflex is possible
786 when the <i>startoffset</i> argument of <b>pcre_exec()</b> is non-zero. The
787 PCRE_DOLLAR_ENDONLY option is ignored if PCRE_MULTILINE is set.
788 </P>
789 <P>
790 Note that the sequences \A, \Z, and \z can be used to match the start and
791 end of the subject in both modes, and if all branches of a pattern start with
792 \A it is always anchored, whether or not PCRE_MULTILINE is set.
793 </P>
794 <br><a name="SEC6" href="#TOC1">FULL STOP (PERIOD, DOT)</a><br>
795 <P>
796 Outside a character class, a dot in the pattern matches any one character in
797 the subject string except (by default) a character that signifies the end of a
798 line. In UTF-8 mode, the matched character may be more than one byte long.
799 </P>
800 <P>
801 When a line ending is defined as a single character, dot never matches that
802 character; when the two-character sequence CRLF is used, dot does not match CR
803 if it is immediately followed by LF, but otherwise it matches all characters
804 (including isolated CRs and LFs). When any Unicode line endings are being
805 recognized, dot does not match CR or LF or any of the other line ending
806 characters.
807 </P>
808 <P>
809 The behaviour of dot with regard to newlines can be changed. If the PCRE_DOTALL
810 option is set, a dot matches any one character, without exception. If the
811 two-character sequence CRLF is present in the subject string, it takes two dots
812 to match it.
813 </P>
814 <P>
815 The handling of dot is entirely independent of the handling of circumflex and
816 dollar, the only relationship being that they both involve newlines. Dot has no
817 special meaning in a character class.
818 </P>
819 <br><a name="SEC7" href="#TOC1">MATCHING A SINGLE BYTE</a><br>
820 <P>
821 Outside a character class, the escape sequence \C matches any one byte, both
822 in and out of UTF-8 mode. Unlike a dot, it always matches any line-ending
823 characters. The feature is provided in Perl in order to match individual bytes
824 in UTF-8 mode. Because it breaks up UTF-8 characters into individual bytes,
825 what remains in the string may be a malformed UTF-8 string. For this reason,
826 the \C escape sequence is best avoided.
827 </P>
828 <P>
829 PCRE does not allow \C to appear in lookbehind assertions
830 <a href="#lookbehind">(described below),</a>
831 because in UTF-8 mode this would make it impossible to calculate the length of
832 the lookbehind.
833 <a name="characterclass"></a></P>
834 <br><a name="SEC8" href="#TOC1">SQUARE BRACKETS AND CHARACTER CLASSES</a><br>
835 <P>
836 An opening square bracket introduces a character class, terminated by a closing
837 square bracket. A closing square bracket on its own is not special. If a
838 closing square bracket is required as a member of the class, it should be the
839 first data character in the class (after an initial circumflex, if present) or
840 escaped with a backslash.
841 </P>
842 <P>
843 A character class matches a single character in the subject. In UTF-8 mode, the
844 character may occupy more than one byte. A matched character must be in the set
845 of characters defined by the class, unless the first character in the class
846 definition is a circumflex, in which case the subject character must not be in
847 the set defined by the class. If a circumflex is actually required as a member
848 of the class, ensure it is not the first character, or escape it with a
849 backslash.
850 </P>
851 <P>
852 For example, the character class [aeiou] matches any lower case vowel, while
853 [^aeiou] matches any character that is not a lower case vowel. Note that a
854 circumflex is just a convenient notation for specifying the characters that
855 are in the class by enumerating those that are not. A class that starts with a
856 circumflex is not an assertion: it still consumes a character from the subject
857 string, and therefore it fails if the current pointer is at the end of the
858 string.
859 </P>
860 <P>
861 In UTF-8 mode, characters with values greater than 255 can be included in a
862 class as a literal string of bytes, or by using the \x{ escaping mechanism.
863 </P>
864 <P>
865 When caseless matching is set, any letters in a class represent both their
866 upper case and lower case versions, so for example, a caseless [aeiou] matches
867 "A" as well as "a", and a caseless [^aeiou] does not match "A", whereas a
868 caseful version would. In UTF-8 mode, PCRE always understands the concept of
869 case for characters whose values are less than 128, so caseless matching is
870 always possible. For characters with higher values, the concept of case is
871 supported if PCRE is compiled with Unicode property support, but not otherwise.
872 If you want to use caseless matching for characters 128 and above, you must
873 ensure that PCRE is compiled with Unicode property support as well as with
874 UTF-8 support.
875 </P>
876 <P>
877 Characters that might indicate line breaks are never treated in any special way
878 when matching character classes, whatever line-ending sequence is in use, and
879 whatever setting of the PCRE_DOTALL and PCRE_MULTILINE options is used. A class
880 such as [^a] always matches one of these characters.
881 </P>
882 <P>
883 The minus (hyphen) character can be used to specify a range of characters in a
884 character class. For example, [d-m] matches any letter between d and m,
885 inclusive. If a minus character is required in a class, it must be escaped with
886 a backslash or appear in a position where it cannot be interpreted as
887 indicating a range, typically as the first or last character in the class.
888 </P>
889 <P>
890 It is not possible to have the literal character "]" as the end character of a
891 range. A pattern such as [W-]46] is interpreted as a class of two characters
892 ("W" and "-") followed by a literal string "46]", so it would match "W46]" or
893 "-46]". However, if the "]" is escaped with a backslash it is interpreted as
894 the end of range, so [W-\]46] is interpreted as a class containing a range
895 followed by two other characters. The octal or hexadecimal representation of
896 "]" can also be used to end a range.
897 </P>
898 <P>
899 Ranges operate in the collating sequence of character values. They can also be
900 used for characters specified numerically, for example [\000-\037]. In UTF-8
901 mode, ranges can include characters whose values are greater than 255, for
902 example [\x{100}-\x{2ff}].
903 </P>
904 <P>
905 If a range that includes letters is used when caseless matching is set, it
906 matches the letters in either case. For example, [W-c] is equivalent to
907 [][\\^_`wxyzabc], matched caselessly, and in non-UTF-8 mode, if character
908 tables for a French locale are in use, [\xc8-\xcb] matches accented E
909 characters in both cases. In UTF-8 mode, PCRE supports the concept of case for
910 characters with values greater than 128 only when it is compiled with Unicode
911 property support.
912 </P>
913 <P>
914 The character types \d, \D, \p, \P, \s, \S, \w, and \W may also appear
915 in a character class, and add the characters that they match to the class. For
916 example, [\dABCDEF] matches any hexadecimal digit. A circumflex can
917 conveniently be used with the upper case character types to specify a more
918 restricted set of characters than the matching lower case type. For example,
919 the class [^\W_] matches any letter or digit, but not underscore.
920 </P>
921 <P>
922 The only metacharacters that are recognized in character classes are backslash,
923 hyphen (only where it can be interpreted as specifying a range), circumflex
924 (only at the start), opening square bracket (only when it can be interpreted as
925 introducing a POSIX class name - see the next section), and the terminating
926 closing square bracket. However, escaping other non-alphanumeric characters
927 does no harm.
928 </P>
929 <br><a name="SEC9" href="#TOC1">POSIX CHARACTER CLASSES</a><br>
930 <P>
931 Perl supports the POSIX notation for character classes. This uses names
932 enclosed by [: and :] within the enclosing square brackets. PCRE also supports
933 this notation. For example,
934 <pre>
935 [01[:alpha:]%]
936 </pre>
937 matches "0", "1", any alphabetic character, or "%". The supported class names
938 are
939 <pre>
940 alnum letters and digits
941 alpha letters
942 ascii character codes 0 - 127
943 blank space or tab only
944 cntrl control characters
945 digit decimal digits (same as \d)
946 graph printing characters, excluding space
947 lower lower case letters
948 print printing characters, including space
949 punct printing characters, excluding letters and digits
950 space white space (not quite the same as \s)
951 upper upper case letters
952 word "word" characters (same as \w)
953 xdigit hexadecimal digits
954 </pre>
955 The "space" characters are HT (9), LF (10), VT (11), FF (12), CR (13), and
956 space (32). Notice that this list includes the VT character (code 11). This
957 makes "space" different to \s, which does not include VT (for Perl
958 compatibility).
959 </P>
960 <P>
961 The name "word" is a Perl extension, and "blank" is a GNU extension from Perl
962 5.8. Another Perl extension is negation, which is indicated by a ^ character
963 after the colon. For example,
964 <pre>
965 [12[:^digit:]]
966 </pre>
967 matches "1", "2", or any non-digit. PCRE (and Perl) also recognize the POSIX
968 syntax [.ch.] and [=ch=] where "ch" is a "collating element", but these are not
969 supported, and an error is given if they are encountered.
970 </P>
971 <P>
972 In UTF-8 mode, characters with values greater than 128 do not match any of
973 the POSIX character classes.
974 </P>
975 <br><a name="SEC10" href="#TOC1">VERTICAL BAR</a><br>
976 <P>
977 Vertical bar characters are used to separate alternative patterns. For example,
978 the pattern
979 <pre>
980 gilbert|sullivan
981 </pre>
982 matches either "gilbert" or "sullivan". Any number of alternatives may appear,
983 and an empty alternative is permitted (matching the empty string). The matching
984 process tries each alternative in turn, from left to right, and the first one
985 that succeeds is used. If the alternatives are within a subpattern
986 <a href="#subpattern">(defined below),</a>
987 "succeeds" means matching the rest of the main pattern as well as the
988 alternative in the subpattern.
989 </P>
990 <br><a name="SEC11" href="#TOC1">INTERNAL OPTION SETTING</a><br>
991 <P>
993 PCRE_EXTENDED options (which are Perl-compatible) can be changed from within
994 the pattern by a sequence of Perl option letters enclosed between "(?" and ")".
995 The option letters are
996 <pre>
999 s for PCRE_DOTALL
1000 x for PCRE_EXTENDED
1001 </pre>
1002 For example, (?im) sets caseless, multiline matching. It is also possible to
1003 unset these options by preceding the letter with a hyphen, and a combined
1004 setting and unsetting such as (?im-sx), which sets PCRE_CASELESS and
1005 PCRE_MULTILINE while unsetting PCRE_DOTALL and PCRE_EXTENDED, is also
1006 permitted. If a letter appears both before and after the hyphen, the option is
1007 unset.
1008 </P>
1009 <P>
1010 The PCRE-specific options PCRE_DUPNAMES, PCRE_UNGREEDY, and PCRE_EXTRA can be
1011 changed in the same way as the Perl-compatible options by using the characters
1012 J, U and X respectively.
1013 </P>
1014 <P>
1015 When an option change occurs at top level (that is, not inside subpattern
1016 parentheses), the change applies to the remainder of the pattern that follows.
1017 If the change is placed right at the start of a pattern, PCRE extracts it into
1018 the global options (and it will therefore show up in data extracted by the
1019 <b>pcre_fullinfo()</b> function).
1020 </P>
1021 <P>
1022 An option change within a subpattern (see below for a description of
1023 subpatterns) affects only that part of the current pattern that follows it, so
1024 <pre>
1025 (a(?i)b)c
1026 </pre>
1027 matches abc and aBc and no other strings (assuming PCRE_CASELESS is not used).
1028 By this means, options can be made to have different settings in different
1029 parts of the pattern. Any changes made in one alternative do carry on
1030 into subsequent branches within the same subpattern. For example,
1031 <pre>
1032 (a(?i)b|c)
1033 </pre>
1034 matches "ab", "aB", "c", and "C", even though when matching "C" the first
1035 branch is abandoned before the option setting. This is because the effects of
1036 option settings happen at compile time. There would be some very weird
1037 behaviour otherwise.
1038 </P>
1039 <P>
1040 <b>Note:</b> There are other PCRE-specific options that can be set by the
1041 application when the compile or match functions are called. In some cases the
1042 pattern can contain special leading sequences to override what the application
1043 has set or what has been defaulted. Details are given in the section entitled
1044 <a href="#newlineseq">"Newline sequences"</a>
1045 above.
1046 <a name="subpattern"></a></P>
1047 <br><a name="SEC12" href="#TOC1">SUBPATTERNS</a><br>
1048 <P>
1049 Subpatterns are delimited by parentheses (round brackets), which can be nested.
1050 Turning part of a pattern into a subpattern does two things:
1051 <br>
1052 <br>
1053 1. It localizes a set of alternatives. For example, the pattern
1054 <pre>
1055 cat(aract|erpillar|)
1056 </pre>
1057 matches one of the words "cat", "cataract", or "caterpillar". Without the
1058 parentheses, it would match "cataract", "erpillar" or an empty string.
1059 <br>
1060 <br>
1061 2. It sets up the subpattern as a capturing subpattern. This means that, when
1062 the whole pattern matches, that portion of the subject string that matched the
1063 subpattern is passed back to the caller via the <i>ovector</i> argument of
1064 <b>pcre_exec()</b>. Opening parentheses are counted from left to right (starting
1065 from 1) to obtain numbers for the capturing subpatterns.
1066 </P>
1067 <P>
1068 For example, if the string "the red king" is matched against the pattern
1069 <pre>
1070 the ((red|white) (king|queen))
1071 </pre>
1072 the captured substrings are "red king", "red", and "king", and are numbered 1,
1073 2, and 3, respectively.
1074 </P>
1075 <P>
1076 The fact that plain parentheses fulfil two functions is not always helpful.
1077 There are often times when a grouping subpattern is required without a
1078 capturing requirement. If an opening parenthesis is followed by a question mark
1079 and a colon, the subpattern does not do any capturing, and is not counted when
1080 computing the number of any subsequent capturing subpatterns. For example, if
1081 the string "the white queen" is matched against the pattern
1082 <pre>
1083 the ((?:red|white) (king|queen))
1084 </pre>
1085 the captured substrings are "white queen" and "queen", and are numbered 1 and
1086 2. The maximum number of capturing subpatterns is 65535.
1087 </P>
1088 <P>
1089 As a convenient shorthand, if any option settings are required at the start of
1090 a non-capturing subpattern, the option letters may appear between the "?" and
1091 the ":". Thus the two patterns
1092 <pre>
1093 (?i:saturday|sunday)
1094 (?:(?i)saturday|sunday)
1095 </pre>
1096 match exactly the same set of strings. Because alternative branches are tried
1097 from left to right, and options are not reset until the end of the subpattern
1098 is reached, an option setting in one branch does affect subsequent branches, so
1099 the above patterns match "SUNDAY" as well as "Saturday".
1100 </P>
1101 <br><a name="SEC13" href="#TOC1">DUPLICATE SUBPATTERN NUMBERS</a><br>
1102 <P>
1103 Perl 5.10 introduced a feature whereby each alternative in a subpattern uses
1104 the same numbers for its capturing parentheses. Such a subpattern starts with
1105 (?| and is itself a non-capturing subpattern. For example, consider this
1106 pattern:
1107 <pre>
1108 (?|(Sat)ur|(Sun))day
1109 </pre>
1110 Because the two alternatives are inside a (?| group, both sets of capturing
1111 parentheses are numbered one. Thus, when the pattern matches, you can look
1112 at captured substring number one, whichever alternative matched. This construct
1113 is useful when you want to capture part, but not all, of one of a number of
1114 alternatives. Inside a (?| group, parentheses are numbered as usual, but the
1115 number is reset at the start of each branch. The numbers of any capturing
1116 buffers that follow the subpattern start after the highest number used in any
1117 branch. The following example is taken from the Perl documentation.
1118 The numbers underneath show in which buffer the captured content will be
1119 stored.
1120 <pre>
1121 # before ---------------branch-reset----------- after
1122 / ( a ) (?| x ( y ) z | (p (q) r) | (t) u (v) ) ( z ) /x
1123 # 1 2 2 3 2 3 4
1124 </pre>
1125 A backreference or a recursive call to a numbered subpattern always refers to
1126 the first one in the pattern with the given number.
1127 </P>
1128 <P>
1129 An alternative approach to using this "branch reset" feature is to use
1130 duplicate named subpatterns, as described in the next section.
1131 </P>
1132 <br><a name="SEC14" href="#TOC1">NAMED SUBPATTERNS</a><br>
1133 <P>
1134 Identifying capturing parentheses by number is simple, but it can be very hard
1135 to keep track of the numbers in complicated regular expressions. Furthermore,
1136 if an expression is modified, the numbers may change. To help with this
1137 difficulty, PCRE supports the naming of subpatterns. This feature was not
1138 added to Perl until release 5.10. Python had the feature earlier, and PCRE
1139 introduced it at release 4.0, using the Python syntax. PCRE now supports both
1140 the Perl and the Python syntax.
1141 </P>
1142 <P>
1143 In PCRE, a subpattern can be named in one of three ways: (?&#60;name&#62;...) or
1144 (?'name'...) as in Perl, or (?P&#60;name&#62;...) as in Python. References to capturing
1145 parentheses from other parts of the pattern, such as
1146 <a href="#backreferences">backreferences,</a>
1147 <a href="#recursion">recursion,</a>
1148 and
1149 <a href="#conditions">conditions,</a>
1150 can be made by name as well as by number.
1151 </P>
1152 <P>
1153 Names consist of up to 32 alphanumeric characters and underscores. Named
1154 capturing parentheses are still allocated numbers as well as names, exactly as
1155 if the names were not present. The PCRE API provides function calls for
1156 extracting the name-to-number translation table from a compiled pattern. There
1157 is also a convenience function for extracting a captured substring by name.
1158 </P>
1159 <P>
1160 By default, a name must be unique within a pattern, but it is possible to relax
1161 this constraint by setting the PCRE_DUPNAMES option at compile time. This can
1162 be useful for patterns where only one instance of the named parentheses can
1163 match. Suppose you want to match the name of a weekday, either as a 3-letter
1164 abbreviation or as the full name, and in both cases you want to extract the
1165 abbreviation. This pattern (ignoring the line breaks) does the job:
1166 <pre>
1167 (?&#60;DN&#62;Mon|Fri|Sun)(?:day)?|
1168 (?&#60;DN&#62;Tue)(?:sday)?|
1169 (?&#60;DN&#62;Wed)(?:nesday)?|
1170 (?&#60;DN&#62;Thu)(?:rsday)?|
1171 (?&#60;DN&#62;Sat)(?:urday)?
1172 </pre>
1173 There are five capturing substrings, but only one is ever set after a match.
1174 (An alternative way of solving this problem is to use a "branch reset"
1175 subpattern, as described in the previous section.)
1176 </P>
1177 <P>
1178 The convenience function for extracting the data by name returns the substring
1179 for the first (and in this example, the only) subpattern of that name that
1180 matched. This saves searching to find which numbered subpattern it was. If you
1181 make a reference to a non-unique named subpattern from elsewhere in the
1182 pattern, the one that corresponds to the lowest number is used. For further
1183 details of the interfaces for handling named subpatterns, see the
1184 <a href="pcreapi.html"><b>pcreapi</b></a>
1185 documentation.
1186 </P>
1187 <br><a name="SEC15" href="#TOC1">REPETITION</a><br>
1188 <P>
1189 Repetition is specified by quantifiers, which can follow any of the following
1190 items:
1191 <pre>
1192 a literal data character
1193 the dot metacharacter
1194 the \C escape sequence
1195 the \X escape sequence (in UTF-8 mode with Unicode properties)
1196 the \R escape sequence
1197 an escape such as \d that matches a single character
1198 a character class
1199 a back reference (see next section)
1200 a parenthesized subpattern (unless it is an assertion)
1201 </pre>
1202 The general repetition quantifier specifies a minimum and maximum number of
1203 permitted matches, by giving the two numbers in curly brackets (braces),
1204 separated by a comma. The numbers must be less than 65536, and the first must
1205 be less than or equal to the second. For example:
1206 <pre>
1207 z{2,4}
1208 </pre>
1209 matches "zz", "zzz", or "zzzz". A closing brace on its own is not a special
1210 character. If the second number is omitted, but the comma is present, there is
1211 no upper limit; if the second number and the comma are both omitted, the
1212 quantifier specifies an exact number of required matches. Thus
1213 <pre>
1214 [aeiou]{3,}
1215 </pre>
1216 matches at least 3 successive vowels, but may match many more, while
1217 <pre>
1218 \d{8}
1219 </pre>
1220 matches exactly 8 digits. An opening curly bracket that appears in a position
1221 where a quantifier is not allowed, or one that does not match the syntax of a
1222 quantifier, is taken as a literal character. For example, {,6} is not a
1223 quantifier, but a literal string of four characters.
1224 </P>
1225 <P>
1226 In UTF-8 mode, quantifiers apply to UTF-8 characters rather than to individual
1227 bytes. Thus, for example, \x{100}{2} matches two UTF-8 characters, each of
1228 which is represented by a two-byte sequence. Similarly, when Unicode property
1229 support is available, \X{3} matches three Unicode extended sequences, each of
1230 which may be several bytes long (and they may be of different lengths).
1231 </P>
1232 <P>
1233 The quantifier {0} is permitted, causing the expression to behave as if the
1234 previous item and the quantifier were not present.
1235 </P>
1236 <P>
1237 For convenience, the three most common quantifiers have single-character
1238 abbreviations:
1239 <pre>
1240 * is equivalent to {0,}
1241 + is equivalent to {1,}
1242 ? is equivalent to {0,1}
1243 </pre>
1244 It is possible to construct infinite loops by following a subpattern that can
1245 match no characters with a quantifier that has no upper limit, for example:
1246 <pre>
1247 (a?)*
1248 </pre>
1249 Earlier versions of Perl and PCRE used to give an error at compile time for
1250 such patterns. However, because there are cases where this can be useful, such
1251 patterns are now accepted, but if any repetition of the subpattern does in fact
1252 match no characters, the loop is forcibly broken.
1253 </P>
1254 <P>
1255 By default, the quantifiers are "greedy", that is, they match as much as
1256 possible (up to the maximum number of permitted times), without causing the
1257 rest of the pattern to fail. The classic example of where this gives problems
1258 is in trying to match comments in C programs. These appear between /* and */
1259 and within the comment, individual * and / characters may appear. An attempt to
1260 match C comments by applying the pattern
1261 <pre>
1262 /\*.*\*/
1263 </pre>
1264 to the string
1265 <pre>
1266 /* first comment */ not comment /* second comment */
1267 </pre>
1268 fails, because it matches the entire string owing to the greediness of the .*
1269 item.
1270 </P>
1271 <P>
1272 However, if a quantifier is followed by a question mark, it ceases to be
1273 greedy, and instead matches the minimum number of times possible, so the
1274 pattern
1275 <pre>
1276 /\*.*?\*/
1277 </pre>
1278 does the right thing with the C comments. The meaning of the various
1279 quantifiers is not otherwise changed, just the preferred number of matches.
1280 Do not confuse this use of question mark with its use as a quantifier in its
1281 own right. Because it has two uses, it can sometimes appear doubled, as in
1282 <pre>
1283 \d??\d
1284 </pre>
1285 which matches one digit by preference, but can match two if that is the only
1286 way the rest of the pattern matches.
1287 </P>
1288 <P>
1289 If the PCRE_UNGREEDY option is set (an option that is not available in Perl),
1290 the quantifiers are not greedy by default, but individual ones can be made
1291 greedy by following them with a question mark. In other words, it inverts the
1292 default behaviour.
1293 </P>
1294 <P>
1295 When a parenthesized subpattern is quantified with a minimum repeat count that
1296 is greater than 1 or with a limited maximum, more memory is required for the
1297 compiled pattern, in proportion to the size of the minimum or maximum.
1298 </P>
1299 <P>
1300 If a pattern starts with .* or .{0,} and the PCRE_DOTALL option (equivalent
1301 to Perl's /s) is set, thus allowing the dot to match newlines, the pattern is
1302 implicitly anchored, because whatever follows will be tried against every
1303 character position in the subject string, so there is no point in retrying the
1304 overall match at any position after the first. PCRE normally treats such a
1305 pattern as though it were preceded by \A.
1306 </P>
1307 <P>
1308 In cases where it is known that the subject string contains no newlines, it is
1309 worth setting PCRE_DOTALL in order to obtain this optimization, or
1310 alternatively using ^ to indicate anchoring explicitly.
1311 </P>
1312 <P>
1313 However, there is one situation where the optimization cannot be used. When .*
1314 is inside capturing parentheses that are the subject of a backreference
1315 elsewhere in the pattern, a match at the start may fail where a later one
1316 succeeds. Consider, for example:
1317 <pre>
1318 (.*)abc\1
1319 </pre>
1320 If the subject is "xyz123abc123" the match point is the fourth character. For
1321 this reason, such a pattern is not implicitly anchored.
1322 </P>
1323 <P>
1324 When a capturing subpattern is repeated, the value captured is the substring
1325 that matched the final iteration. For example, after
1326 <pre>
1327 (tweedle[dume]{3}\s*)+
1328 </pre>
1329 has matched "tweedledum tweedledee" the value of the captured substring is
1330 "tweedledee". However, if there are nested capturing subpatterns, the
1331 corresponding captured values may have been set in previous iterations. For
1332 example, after
1333 <pre>
1334 /(a|(b))+/
1335 </pre>
1336 matches "aba" the value of the second captured substring is "b".
1337 <a name="atomicgroup"></a></P>
1338 <br><a name="SEC16" href="#TOC1">ATOMIC GROUPING AND POSSESSIVE QUANTIFIERS</a><br>
1339 <P>
1340 With both maximizing ("greedy") and minimizing ("ungreedy" or "lazy")
1341 repetition, failure of what follows normally causes the repeated item to be
1342 re-evaluated to see if a different number of repeats allows the rest of the
1343 pattern to match. Sometimes it is useful to prevent this, either to change the
1344 nature of the match, or to cause it fail earlier than it otherwise might, when
1345 the author of the pattern knows there is no point in carrying on.
1346 </P>
1347 <P>
1348 Consider, for example, the pattern \d+foo when applied to the subject line
1349 <pre>
1350 123456bar
1351 </pre>
1352 After matching all 6 digits and then failing to match "foo", the normal
1353 action of the matcher is to try again with only 5 digits matching the \d+
1354 item, and then with 4, and so on, before ultimately failing. "Atomic grouping"
1355 (a term taken from Jeffrey Friedl's book) provides the means for specifying
1356 that once a subpattern has matched, it is not to be re-evaluated in this way.
1357 </P>
1358 <P>
1359 If we use atomic grouping for the previous example, the matcher gives up
1360 immediately on failing to match "foo" the first time. The notation is a kind of
1361 special parenthesis, starting with (?&#62; as in this example:
1362 <pre>
1363 (?&#62;\d+)foo
1364 </pre>
1365 This kind of parenthesis "locks up" the part of the pattern it contains once
1366 it has matched, and a failure further into the pattern is prevented from
1367 backtracking into it. Backtracking past it to previous items, however, works as
1368 normal.
1369 </P>
1370 <P>
1371 An alternative description is that a subpattern of this type matches the string
1372 of characters that an identical standalone pattern would match, if anchored at
1373 the current point in the subject string.
1374 </P>
1375 <P>
1376 Atomic grouping subpatterns are not capturing subpatterns. Simple cases such as
1377 the above example can be thought of as a maximizing repeat that must swallow
1378 everything it can. So, while both \d+ and \d+? are prepared to adjust the
1379 number of digits they match in order to make the rest of the pattern match,
1380 (?&#62;\d+) can only match an entire sequence of digits.
1381 </P>
1382 <P>
1383 Atomic groups in general can of course contain arbitrarily complicated
1384 subpatterns, and can be nested. However, when the subpattern for an atomic
1385 group is just a single repeated item, as in the example above, a simpler
1386 notation, called a "possessive quantifier" can be used. This consists of an
1387 additional + character following a quantifier. Using this notation, the
1388 previous example can be rewritten as
1389 <pre>
1390 \d++foo
1391 </pre>
1392 Note that a possessive quantifier can be used with an entire group, for
1393 example:
1394 <pre>
1395 (abc|xyz){2,3}+
1396 </pre>
1397 Possessive quantifiers are always greedy; the setting of the PCRE_UNGREEDY
1398 option is ignored. They are a convenient notation for the simpler forms of
1399 atomic group. However, there is no difference in the meaning of a possessive
1400 quantifier and the equivalent atomic group, though there may be a performance
1401 difference; possessive quantifiers should be slightly faster.
1402 </P>
1403 <P>
1404 The possessive quantifier syntax is an extension to the Perl 5.8 syntax.
1405 Jeffrey Friedl originated the idea (and the name) in the first edition of his
1406 book. Mike McCloskey liked it, so implemented it when he built Sun's Java
1407 package, and PCRE copied it from there. It ultimately found its way into Perl
1408 at release 5.10.
1409 </P>
1410 <P>
1411 PCRE has an optimization that automatically "possessifies" certain simple
1412 pattern constructs. For example, the sequence A+B is treated as A++B because
1413 there is no point in backtracking into a sequence of A's when B must follow.
1414 </P>
1415 <P>
1416 When a pattern contains an unlimited repeat inside a subpattern that can itself
1417 be repeated an unlimited number of times, the use of an atomic group is the
1418 only way to avoid some failing matches taking a very long time indeed. The
1419 pattern
1420 <pre>
1421 (\D+|&#60;\d+&#62;)*[!?]
1422 </pre>
1423 matches an unlimited number of substrings that either consist of non-digits, or
1424 digits enclosed in &#60;&#62;, followed by either ! or ?. When it matches, it runs
1425 quickly. However, if it is applied to
1426 <pre>
1427 aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
1428 </pre>
1429 it takes a long time before reporting failure. This is because the string can
1430 be divided between the internal \D+ repeat and the external * repeat in a
1431 large number of ways, and all have to be tried. (The example uses [!?] rather
1432 than a single character at the end, because both PCRE and Perl have an
1433 optimization that allows for fast failure when a single character is used. They
1434 remember the last single character that is required for a match, and fail early
1435 if it is not present in the string.) If the pattern is changed so that it uses
1436 an atomic group, like this:
1437 <pre>
1438 ((?&#62;\D+)|&#60;\d+&#62;)*[!?]
1439 </pre>
1440 sequences of non-digits cannot be broken, and failure happens quickly.
1441 <a name="backreferences"></a></P>
1442 <br><a name="SEC17" href="#TOC1">BACK REFERENCES</a><br>
1443 <P>
1444 Outside a character class, a backslash followed by a digit greater than 0 (and
1445 possibly further digits) is a back reference to a capturing subpattern earlier
1446 (that is, to its left) in the pattern, provided there have been that many
1447 previous capturing left parentheses.
1448 </P>
1449 <P>
1450 However, if the decimal number following the backslash is less than 10, it is
1451 always taken as a back reference, and causes an error only if there are not
1452 that many capturing left parentheses in the entire pattern. In other words, the
1453 parentheses that are referenced need not be to the left of the reference for
1454 numbers less than 10. A "forward back reference" of this type can make sense
1455 when a repetition is involved and the subpattern to the right has participated
1456 in an earlier iteration.
1457 </P>
1458 <P>
1459 It is not possible to have a numerical "forward back reference" to a subpattern
1460 whose number is 10 or more using this syntax because a sequence such as \50 is
1461 interpreted as a character defined in octal. See the subsection entitled
1462 "Non-printing characters"
1463 <a href="#digitsafterbackslash">above</a>
1464 for further details of the handling of digits following a backslash. There is
1465 no such problem when named parentheses are used. A back reference to any
1466 subpattern is possible using named parentheses (see below).
1467 </P>
1468 <P>
1469 Another way of avoiding the ambiguity inherent in the use of digits following a
1470 backslash is to use the \g escape sequence, which is a feature introduced in
1471 Perl 5.10. This escape must be followed by an unsigned number or a negative
1472 number, optionally enclosed in braces. These examples are all identical:
1473 <pre>
1474 (ring), \1
1475 (ring), \g1
1476 (ring), \g{1}
1477 </pre>
1478 An unsigned number specifies an absolute reference without the ambiguity that
1479 is present in the older syntax. It is also useful when literal digits follow
1480 the reference. A negative number is a relative reference. Consider this
1481 example:
1482 <pre>
1483 (abc(def)ghi)\g{-1}
1484 </pre>
1485 The sequence \g{-1} is a reference to the most recently started capturing
1486 subpattern before \g, that is, is it equivalent to \2. Similarly, \g{-2}
1487 would be equivalent to \1. The use of relative references can be helpful in
1488 long patterns, and also in patterns that are created by joining together
1489 fragments that contain references within themselves.
1490 </P>
1491 <P>
1492 A back reference matches whatever actually matched the capturing subpattern in
1493 the current subject string, rather than anything matching the subpattern
1494 itself (see
1495 <a href="#subpatternsassubroutines">"Subpatterns as subroutines"</a>
1496 below for a way of doing that). So the pattern
1497 <pre>
1498 (sens|respons)e and \1ibility
1499 </pre>
1500 matches "sense and sensibility" and "response and responsibility", but not
1501 "sense and responsibility". If caseful matching is in force at the time of the
1502 back reference, the case of letters is relevant. For example,
1503 <pre>
1504 ((?i)rah)\s+\1
1505 </pre>
1506 matches "rah rah" and "RAH RAH", but not "RAH rah", even though the original
1507 capturing subpattern is matched caselessly.
1508 </P>
1509 <P>
1510 There are several different ways of writing back references to named
1511 subpatterns. The .NET syntax \k{name} and the Perl syntax \k&#60;name&#62; or
1512 \k'name' are supported, as is the Python syntax (?P=name). Perl 5.10's unified
1513 back reference syntax, in which \g can be used for both numeric and named
1514 references, is also supported. We could rewrite the above example in any of
1515 the following ways:
1516 <pre>
1517 (?&#60;p1&#62;(?i)rah)\s+\k&#60;p1&#62;
1518 (?'p1'(?i)rah)\s+\k{p1}
1519 (?P&#60;p1&#62;(?i)rah)\s+(?P=p1)
1520 (?&#60;p1&#62;(?i)rah)\s+\g{p1}
1521 </pre>
1522 A subpattern that is referenced by name may appear in the pattern before or
1523 after the reference.
1524 </P>
1525 <P>
1526 There may be more than one back reference to the same subpattern. If a
1527 subpattern has not actually been used in a particular match, any back
1528 references to it always fail. For example, the pattern
1529 <pre>
1530 (a|(bc))\2
1531 </pre>
1532 always fails if it starts to match "a" rather than "bc". Because there may be
1533 many capturing parentheses in a pattern, all digits following the backslash are
1534 taken as part of a potential back reference number. If the pattern continues
1535 with a digit character, some delimiter must be used to terminate the back
1536 reference. If the PCRE_EXTENDED option is set, this can be whitespace.
1537 Otherwise an empty comment (see
1538 <a href="#comments">"Comments"</a>
1539 below) can be used.
1540 </P>
1541 <P>
1542 A back reference that occurs inside the parentheses to which it refers fails
1543 when the subpattern is first used, so, for example, (a\1) never matches.
1544 However, such references can be useful inside repeated subpatterns. For
1545 example, the pattern
1546 <pre>
1547 (a|b\1)+
1548 </pre>
1549 matches any number of "a"s and also "aba", "ababbaa" etc. At each iteration of
1550 the subpattern, the back reference matches the character string corresponding
1551 to the previous iteration. In order for this to work, the pattern must be such
1552 that the first iteration does not need to match the back reference. This can be
1553 done using alternation, as in the example above, or by a quantifier with a
1554 minimum of zero.
1555 <a name="bigassertions"></a></P>
1556 <br><a name="SEC18" href="#TOC1">ASSERTIONS</a><br>
1557 <P>
1558 An assertion is a test on the characters following or preceding the current
1559 matching point that does not actually consume any characters. The simple
1560 assertions coded as \b, \B, \A, \G, \Z, \z, ^ and $ are described
1561 <a href="#smallassertions">above.</a>
1562 </P>
1563 <P>
1564 More complicated assertions are coded as subpatterns. There are two kinds:
1565 those that look ahead of the current position in the subject string, and those
1566 that look behind it. An assertion subpattern is matched in the normal way,
1567 except that it does not cause the current matching position to be changed.
1568 </P>
1569 <P>
1570 Assertion subpatterns are not capturing subpatterns, and may not be repeated,
1571 because it makes no sense to assert the same thing several times. If any kind
1572 of assertion contains capturing subpatterns within it, these are counted for
1573 the purposes of numbering the capturing subpatterns in the whole pattern.
1574 However, substring capturing is carried out only for positive assertions,
1575 because it does not make sense for negative assertions.
1576 </P>
1577 <br><b>
1578 Lookahead assertions
1579 </b><br>
1580 <P>
1581 Lookahead assertions start with (?= for positive assertions and (?! for
1582 negative assertions. For example,
1583 <pre>
1584 \w+(?=;)
1585 </pre>
1586 matches a word followed by a semicolon, but does not include the semicolon in
1587 the match, and
1588 <pre>
1589 foo(?!bar)
1590 </pre>
1591 matches any occurrence of "foo" that is not followed by "bar". Note that the
1592 apparently similar pattern
1593 <pre>
1594 (?!foo)bar
1595 </pre>
1596 does not find an occurrence of "bar" that is preceded by something other than
1597 "foo"; it finds any occurrence of "bar" whatsoever, because the assertion
1598 (?!foo) is always true when the next three characters are "bar". A
1599 lookbehind assertion is needed to achieve the other effect.
1600 </P>
1601 <P>
1602 If you want to force a matching failure at some point in a pattern, the most
1603 convenient way to do it is with (?!) because an empty string always matches, so
1604 an assertion that requires there not to be an empty string must always fail.
1605 <a name="lookbehind"></a></P>
1606 <br><b>
1607 Lookbehind assertions
1608 </b><br>
1609 <P>
1610 Lookbehind assertions start with (?&#60;= for positive assertions and (?&#60;! for
1611 negative assertions. For example,
1612 <pre>
1613 (?&#60;!foo)bar
1614 </pre>
1615 does find an occurrence of "bar" that is not preceded by "foo". The contents of
1616 a lookbehind assertion are restricted such that all the strings it matches must
1617 have a fixed length. However, if there are several top-level alternatives, they
1618 do not all have to have the same fixed length. Thus
1619 <pre>
1620 (?&#60;=bullock|donkey)
1621 </pre>
1622 is permitted, but
1623 <pre>
1624 (?&#60;!dogs?|cats?)
1625 </pre>
1626 causes an error at compile time. Branches that match different length strings
1627 are permitted only at the top level of a lookbehind assertion. This is an
1628 extension compared with Perl (at least for 5.8), which requires all branches to
1629 match the same length of string. An assertion such as
1630 <pre>
1631 (?&#60;=ab(c|de))
1632 </pre>
1633 is not permitted, because its single top-level branch can match two different
1634 lengths, but it is acceptable if rewritten to use two top-level branches:
1635 <pre>
1636 (?&#60;=abc|abde)
1637 </pre>
1638 In some cases, the Perl 5.10 escape sequence \K
1639 <a href="#resetmatchstart">(see above)</a>
1640 can be used instead of a lookbehind assertion; this is not restricted to a
1641 fixed-length.
1642 </P>
1643 <P>
1644 The implementation of lookbehind assertions is, for each alternative, to
1645 temporarily move the current position back by the fixed length and then try to
1646 match. If there are insufficient characters before the current position, the
1647 assertion fails.
1648 </P>
1649 <P>
1650 PCRE does not allow the \C escape (which matches a single byte in UTF-8 mode)
1651 to appear in lookbehind assertions, because it makes it impossible to calculate
1652 the length of the lookbehind. The \X and \R escapes, which can match
1653 different numbers of bytes, are also not permitted.
1654 </P>
1655 <P>
1656 Possessive quantifiers can be used in conjunction with lookbehind assertions to
1657 specify efficient matching at the end of the subject string. Consider a simple
1658 pattern such as
1659 <pre>
1660 abcd$
1661 </pre>
1662 when applied to a long string that does not match. Because matching proceeds
1663 from left to right, PCRE will look for each "a" in the subject and then see if
1664 what follows matches the rest of the pattern. If the pattern is specified as
1665 <pre>
1666 ^.*abcd$
1667 </pre>
1668 the initial .* matches the entire string at first, but when this fails (because
1669 there is no following "a"), it backtracks to match all but the last character,
1670 then all but the last two characters, and so on. Once again the search for "a"
1671 covers the entire string, from right to left, so we are no better off. However,
1672 if the pattern is written as
1673 <pre>
1674 ^.*+(?&#60;=abcd)
1675 </pre>
1676 there can be no backtracking for the .*+ item; it can match only the entire
1677 string. The subsequent lookbehind assertion does a single test on the last four
1678 characters. If it fails, the match fails immediately. For long strings, this
1679 approach makes a significant difference to the processing time.
1680 </P>
1681 <br><b>
1682 Using multiple assertions
1683 </b><br>
1684 <P>
1685 Several assertions (of any sort) may occur in succession. For example,
1686 <pre>
1687 (?&#60;=\d{3})(?&#60;!999)foo
1688 </pre>
1689 matches "foo" preceded by three digits that are not "999". Notice that each of
1690 the assertions is applied independently at the same point in the subject
1691 string. First there is a check that the previous three characters are all
1692 digits, and then there is a check that the same three characters are not "999".
1693 This pattern does <i>not</i> match "foo" preceded by six characters, the first
1694 of which are digits and the last three of which are not "999". For example, it
1695 doesn't match "123abcfoo". A pattern to do that is
1696 <pre>
1697 (?&#60;=\d{3}...)(?&#60;!999)foo
1698 </pre>
1699 This time the first assertion looks at the preceding six characters, checking
1700 that the first three are digits, and then the second assertion checks that the
1701 preceding three characters are not "999".
1702 </P>
1703 <P>
1704 Assertions can be nested in any combination. For example,
1705 <pre>
1706 (?&#60;=(?&#60;!foo)bar)baz
1707 </pre>
1708 matches an occurrence of "baz" that is preceded by "bar" which in turn is not
1709 preceded by "foo", while
1710 <pre>
1711 (?&#60;=\d{3}(?!999)...)foo
1712 </pre>
1713 is another pattern that matches "foo" preceded by three digits and any three
1714 characters that are not "999".
1715 <a name="conditions"></a></P>
1716 <br><a name="SEC19" href="#TOC1">CONDITIONAL SUBPATTERNS</a><br>
1717 <P>
1718 It is possible to cause the matching process to obey a subpattern
1719 conditionally or to choose between two alternative subpatterns, depending on
1720 the result of an assertion, or whether a previous capturing subpattern matched
1721 or not. The two possible forms of conditional subpattern are
1722 <pre>
1723 (?(condition)yes-pattern)
1724 (?(condition)yes-pattern|no-pattern)
1725 </pre>
1726 If the condition is satisfied, the yes-pattern is used; otherwise the
1727 no-pattern (if present) is used. If there are more than two alternatives in the
1728 subpattern, a compile-time error occurs.
1729 </P>
1730 <P>
1731 There are four kinds of condition: references to subpatterns, references to
1732 recursion, a pseudo-condition called DEFINE, and assertions.
1733 </P>
1734 <br><b>
1735 Checking for a used subpattern by number
1736 </b><br>
1737 <P>
1738 If the text between the parentheses consists of a sequence of digits, the
1739 condition is true if the capturing subpattern of that number has previously
1740 matched. An alternative notation is to precede the digits with a plus or minus
1741 sign. In this case, the subpattern number is relative rather than absolute.
1742 The most recently opened parentheses can be referenced by (?(-1), the next most
1743 recent by (?(-2), and so on. In looping constructs it can also make sense to
1744 refer to subsequent groups with constructs such as (?(+2).
1745 </P>
1746 <P>
1747 Consider the following pattern, which contains non-significant white space to
1748 make it more readable (assume the PCRE_EXTENDED option) and to divide it into
1749 three parts for ease of discussion:
1750 <pre>
1751 ( \( )? [^()]+ (?(1) \) )
1752 </pre>
1753 The first part matches an optional opening parenthesis, and if that
1754 character is present, sets it as the first captured substring. The second part
1755 matches one or more characters that are not parentheses. The third part is a
1756 conditional subpattern that tests whether the first set of parentheses matched
1757 or not. If they did, that is, if subject started with an opening parenthesis,
1758 the condition is true, and so the yes-pattern is executed and a closing
1759 parenthesis is required. Otherwise, since no-pattern is not present, the
1760 subpattern matches nothing. In other words, this pattern matches a sequence of
1761 non-parentheses, optionally enclosed in parentheses.
1762 </P>
1763 <P>
1764 If you were embedding this pattern in a larger one, you could use a relative
1765 reference:
1766 <pre>
1767 ...other stuff... ( \( )? [^()]+ (?(-1) \) ) ...
1768 </pre>
1769 This makes the fragment independent of the parentheses in the larger pattern.
1770 </P>
1771 <br><b>
1772 Checking for a used subpattern by name
1773 </b><br>
1774 <P>
1775 Perl uses the syntax (?(&#60;name&#62;)...) or (?('name')...) to test for a used
1776 subpattern by name. For compatibility with earlier versions of PCRE, which had
1777 this facility before Perl, the syntax (?(name)...) is also recognized. However,
1778 there is a possible ambiguity with this syntax, because subpattern names may
1779 consist entirely of digits. PCRE looks first for a named subpattern; if it
1780 cannot find one and the name consists entirely of digits, PCRE looks for a
1781 subpattern of that number, which must be greater than zero. Using subpattern
1782 names that consist entirely of digits is not recommended.
1783 </P>
1784 <P>
1785 Rewriting the above example to use a named subpattern gives this:
1786 <pre>
1787 (?&#60;OPEN&#62; \( )? [^()]+ (?(&#60;OPEN&#62;) \) )
1789 </PRE>
1790 </P>
1791 <br><b>
1792 Checking for pattern recursion
1793 </b><br>
1794 <P>
1795 If the condition is the string (R), and there is no subpattern with the name R,
1796 the condition is true if a recursive call to the whole pattern or any
1797 subpattern has been made. If digits or a name preceded by ampersand follow the
1798 letter R, for example:
1799 <pre>
1800 (?(R3)...) or (?(R&name)...)
1801 </pre>
1802 the condition is true if the most recent recursion is into the subpattern whose
1803 number or name is given. This condition does not check the entire recursion
1804 stack.
1805 </P>
1806 <P>
1807 At "top level", all these recursion test conditions are false. Recursive
1808 patterns are described below.
1809 </P>
1810 <br><b>
1811 Defining subpatterns for use by reference only
1812 </b><br>
1813 <P>
1814 If the condition is the string (DEFINE), and there is no subpattern with the
1815 name DEFINE, the condition is always false. In this case, there may be only one
1816 alternative in the subpattern. It is always skipped if control reaches this
1817 point in the pattern; the idea of DEFINE is that it can be used to define
1818 "subroutines" that can be referenced from elsewhere. (The use of "subroutines"
1819 is described below.) For example, a pattern to match an IPv4 address could be
1820 written like this (ignore whitespace and line breaks):
1821 <pre>
1822 (?(DEFINE) (?&#60;byte&#62; 2[0-4]\d | 25[0-5] | 1\d\d | [1-9]?\d) )
1823 \b (?&byte) (\.(?&byte)){3} \b
1824 </pre>
1825 The first part of the pattern is a DEFINE group inside which a another group
1826 named "byte" is defined. This matches an individual component of an IPv4
1827 address (a number less than 256). When matching takes place, this part of the
1828 pattern is skipped because DEFINE acts like a false condition.
1829 </P>
1830 <P>
1831 The rest of the pattern uses references to the named group to match the four
1832 dot-separated components of an IPv4 address, insisting on a word boundary at
1833 each end.
1834 </P>
1835 <br><b>
1836 Assertion conditions
1837 </b><br>
1838 <P>
1839 If the condition is not in any of the above formats, it must be an assertion.
1840 This may be a positive or negative lookahead or lookbehind assertion. Consider
1841 this pattern, again containing non-significant white space, and with the two
1842 alternatives on the second line:
1843 <pre>
1844 (?(?=[^a-z]*[a-z])
1845 \d{2}-[a-z]{3}-\d{2} | \d{2}-\d{2}-\d{2} )
1846 </pre>
1847 The condition is a positive lookahead assertion that matches an optional
1848 sequence of non-letters followed by a letter. In other words, it tests for the
1849 presence of at least one letter in the subject. If a letter is found, the
1850 subject is matched against the first alternative; otherwise it is matched
1851 against the second. This pattern matches strings in one of the two forms
1852 dd-aaa-dd or dd-dd-dd, where aaa are letters and dd are digits.
1853 <a name="comments"></a></P>
1854 <br><a name="SEC20" href="#TOC1">COMMENTS</a><br>
1855 <P>
1856 The sequence (?# marks the start of a comment that continues up to the next
1857 closing parenthesis. Nested parentheses are not permitted. The characters
1858 that make up a comment play no part in the pattern matching at all.
1859 </P>
1860 <P>
1861 If the PCRE_EXTENDED option is set, an unescaped # character outside a
1862 character class introduces a comment that continues to immediately after the
1863 next newline in the pattern.
1864 <a name="recursion"></a></P>
1865 <br><a name="SEC21" href="#TOC1">RECURSIVE PATTERNS</a><br>
1866 <P>
1867 Consider the problem of matching a string in parentheses, allowing for
1868 unlimited nested parentheses. Without the use of recursion, the best that can
1869 be done is to use a pattern that matches up to some fixed depth of nesting. It
1870 is not possible to handle an arbitrary nesting depth.
1871 </P>
1872 <P>
1873 For some time, Perl has provided a facility that allows regular expressions to
1874 recurse (amongst other things). It does this by interpolating Perl code in the
1875 expression at run time, and the code can refer to the expression itself. A Perl
1876 pattern using code interpolation to solve the parentheses problem can be
1877 created like this:
1878 <pre>
1879 $re = qr{\( (?: (?&#62;[^()]+) | (?p{$re}) )* \)}x;
1880 </pre>
1881 The (?p{...}) item interpolates Perl code at run time, and in this case refers
1882 recursively to the pattern in which it appears.
1883 </P>
1884 <P>
1885 Obviously, PCRE cannot support the interpolation of Perl code. Instead, it
1886 supports special syntax for recursion of the entire pattern, and also for
1887 individual subpattern recursion. After its introduction in PCRE and Python,
1888 this kind of recursion was introduced into Perl at release 5.10.
1889 </P>
1890 <P>
1891 A special item that consists of (? followed by a number greater than zero and a
1892 closing parenthesis is a recursive call of the subpattern of the given number,
1893 provided that it occurs inside that subpattern. (If not, it is a "subroutine"
1894 call, which is described in the next section.) The special item (?R) or (?0) is
1895 a recursive call of the entire regular expression.
1896 </P>
1897 <P>
1898 In PCRE (like Python, but unlike Perl), a recursive subpattern call is always
1899 treated as an atomic group. That is, once it has matched some of the subject
1900 string, it is never re-entered, even if it contains untried alternatives and
1901 there is a subsequent matching failure.
1902 </P>
1903 <P>
1904 This PCRE pattern solves the nested parentheses problem (assume the
1905 PCRE_EXTENDED option is set so that white space is ignored):
1906 <pre>
1907 \( ( (?&#62;[^()]+) | (?R) )* \)
1908 </pre>
1909 First it matches an opening parenthesis. Then it matches any number of
1910 substrings which can either be a sequence of non-parentheses, or a recursive
1911 match of the pattern itself (that is, a correctly parenthesized substring).
1912 Finally there is a closing parenthesis.
1913 </P>
1914 <P>
1915 If this were part of a larger pattern, you would not want to recurse the entire
1916 pattern, so instead you could use this:
1917 <pre>
1918 ( \( ( (?&#62;[^()]+) | (?1) )* \) )
1919 </pre>
1920 We have put the pattern into parentheses, and caused the recursion to refer to
1921 them instead of the whole pattern.
1922 </P>
1923 <P>
1924 In a larger pattern, keeping track of parenthesis numbers can be tricky. This
1925 is made easier by the use of relative references. (A Perl 5.10 feature.)
1926 Instead of (?1) in the pattern above you can write (?-2) to refer to the second
1927 most recently opened parentheses preceding the recursion. In other words, a
1928 negative number counts capturing parentheses leftwards from the point at which
1929 it is encountered.
1930 </P>
1931 <P>
1932 It is also possible to refer to subsequently opened parentheses, by writing
1933 references such as (?+2). However, these cannot be recursive because the
1934 reference is not inside the parentheses that are referenced. They are always
1935 "subroutine" calls, as described in the next section.
1936 </P>
1937 <P>
1938 An alternative approach is to use named parentheses instead. The Perl syntax
1939 for this is (?&name); PCRE's earlier syntax (?P&#62;name) is also supported. We
1940 could rewrite the above example as follows:
1941 <pre>
1942 (?&#60;pn&#62; \( ( (?&#62;[^()]+) | (?&pn) )* \) )
1943 </pre>
1944 If there is more than one subpattern with the same name, the earliest one is
1945 used.
1946 </P>
1947 <P>
1948 This particular example pattern that we have been looking at contains nested
1949 unlimited repeats, and so the use of atomic grouping for matching strings of
1950 non-parentheses is important when applying the pattern to strings that do not
1951 match. For example, when this pattern is applied to
1952 <pre>
1953 (aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa()
1954 </pre>
1955 it yields "no match" quickly. However, if atomic grouping is not used,
1956 the match runs for a very long time indeed because there are so many different
1957 ways the + and * repeats can carve up the subject, and all have to be tested
1958 before failure can be reported.
1959 </P>
1960 <P>
1961 At the end of a match, the values set for any capturing subpatterns are those
1962 from the outermost level of the recursion at which the subpattern value is set.
1963 If you want to obtain intermediate values, a callout function can be used (see
1964 below and the
1965 <a href="pcrecallout.html"><b>pcrecallout</b></a>
1966 documentation). If the pattern above is matched against
1967 <pre>
1968 (ab(cd)ef)
1969 </pre>
1970 the value for the capturing parentheses is "ef", which is the last value taken
1971 on at the top level. If additional parentheses are added, giving
1972 <pre>
1973 \( ( ( (?&#62;[^()]+) | (?R) )* ) \)
1974 ^ ^
1975 ^ ^
1976 </pre>
1977 the string they capture is "ab(cd)ef", the contents of the top level
1978 parentheses. If there are more than 15 capturing parentheses in a pattern, PCRE
1979 has to obtain extra memory to store data during a recursion, which it does by
1980 using <b>pcre_malloc</b>, freeing it via <b>pcre_free</b> afterwards. If no
1981 memory can be obtained, the match fails with the PCRE_ERROR_NOMEMORY error.
1982 </P>
1983 <P>
1984 Do not confuse the (?R) item with the condition (R), which tests for recursion.
1985 Consider this pattern, which matches text in angle brackets, allowing for
1986 arbitrary nesting. Only digits are allowed in nested brackets (that is, when
1987 recursing), whereas any characters are permitted at the outer level.
1988 <pre>
1989 &#60; (?: (?(R) \d++ | [^&#60;&#62;]*+) | (?R)) * &#62;
1990 </pre>
1991 In this pattern, (?(R) is the start of a conditional subpattern, with two
1992 different alternatives for the recursive and non-recursive cases. The (?R) item
1993 is the actual recursive call.
1994 <a name="subpatternsassubroutines"></a></P>
1995 <br><a name="SEC22" href="#TOC1">SUBPATTERNS AS SUBROUTINES</a><br>
1996 <P>
1997 If the syntax for a recursive subpattern reference (either by number or by
1998 name) is used outside the parentheses to which it refers, it operates like a
1999 subroutine in a programming language. The "called" subpattern may be defined
2000 before or after the reference. A numbered reference can be absolute or
2001 relative, as in these examples:
2002 <pre>
2003 (...(absolute)...)...(?2)...
2004 (...(relative)...)...(?-1)...
2005 (...(?+1)...(relative)...
2006 </pre>
2007 An earlier example pointed out that the pattern
2008 <pre>
2009 (sens|respons)e and \1ibility
2010 </pre>
2011 matches "sense and sensibility" and "response and responsibility", but not
2012 "sense and responsibility". If instead the pattern
2013 <pre>
2014 (sens|respons)e and (?1)ibility
2015 </pre>
2016 is used, it does match "sense and responsibility" as well as the other two
2017 strings. Another example is given in the discussion of DEFINE above.
2018 </P>
2019 <P>
2020 Like recursive subpatterns, a "subroutine" call is always treated as an atomic
2021 group. That is, once it has matched some of the subject string, it is never
2022 re-entered, even if it contains untried alternatives and there is a subsequent
2023 matching failure.
2024 </P>
2025 <P>
2026 When a subpattern is used as a subroutine, processing options such as
2027 case-independence are fixed when the subpattern is defined. They cannot be
2028 changed for different calls. For example, consider this pattern:
2029 <pre>
2030 (abc)(?i:(?-1))
2031 </pre>
2032 It matches "abcabc". It does not match "abcABC" because the change of
2033 processing option does not affect the called subpattern.
2034 </P>
2035 <br><a name="SEC23" href="#TOC1">CALLOUTS</a><br>
2036 <P>
2037 Perl has a feature whereby using the sequence (?{...}) causes arbitrary Perl
2038 code to be obeyed in the middle of matching a regular expression. This makes it
2039 possible, amongst other things, to extract different substrings that match the
2040 same pair of parentheses when there is a repetition.
2041 </P>
2042 <P>
2043 PCRE provides a similar feature, but of course it cannot obey arbitrary Perl
2044 code. The feature is called "callout". The caller of PCRE provides an external
2045 function by putting its entry point in the global variable <i>pcre_callout</i>.
2046 By default, this variable contains NULL, which disables all calling out.
2047 </P>
2048 <P>
2049 Within a regular expression, (?C) indicates the points at which the external
2050 function is to be called. If you want to identify different callout points, you
2051 can put a number less than 256 after the letter C. The default value is zero.
2052 For example, this pattern has two callout points:
2053 <pre>
2054 (?C1)abc(?C2)def
2055 </pre>
2056 If the PCRE_AUTO_CALLOUT flag is passed to <b>pcre_compile()</b>, callouts are
2057 automatically installed before each item in the pattern. They are all numbered
2058 255.
2059 </P>
2060 <P>
2061 During matching, when PCRE reaches a callout point (and <i>pcre_callout</i> is
2062 set), the external function is called. It is provided with the number of the
2063 callout, the position in the pattern, and, optionally, one item of data
2064 originally supplied by the caller of <b>pcre_exec()</b>. The callout function
2065 may cause matching to proceed, to backtrack, or to fail altogether. A complete
2066 description of the interface to the callout function is given in the
2067 <a href="pcrecallout.html"><b>pcrecallout</b></a>
2068 documentation.
2069 </P>
2070 <br><a name="SEC24" href="#TOC1">BACKTRACKING CONTROL</a><br>
2071 <P>
2072 Perl 5.10 introduced a number of "Special Backtracking Control Verbs", which
2073 are described in the Perl documentation as "experimental and subject to change
2074 or removal in a future version of Perl". It goes on to say: "Their usage in
2075 production code should be noted to avoid problems during upgrades." The same
2076 remarks apply to the PCRE features described in this section.
2077 </P>
2078 <P>
2079 Since these verbs are specifically related to backtracking, they can be used
2080 only when the pattern is to be matched using <b>pcre_exec()</b>, which uses a
2081 backtracking algorithm. They cause an error if encountered by
2082 <b>pcre_dfa_exec()</b>.
2083 </P>
2084 <P>
2085 The new verbs make use of what was previously invalid syntax: an opening
2086 parenthesis followed by an asterisk. In Perl, they are generally of the form
2087 (*VERB:ARG) but PCRE does not support the use of arguments, so its general
2088 form is just (*VERB). Any number of these verbs may occur in a pattern. There
2089 are two kinds:
2090 </P>
2091 <br><b>
2092 Verbs that act immediately
2093 </b><br>
2094 <P>
2095 The following verbs act as soon as they are encountered:
2096 <pre>
2097 (*ACCEPT)
2098 </pre>
2099 This verb causes the match to end successfully, skipping the remainder of the
2100 pattern. When inside a recursion, only the innermost pattern is ended
2101 immediately. PCRE differs from Perl in what happens if the (*ACCEPT) is inside
2102 capturing parentheses. In Perl, the data so far is captured: in PCRE no data is
2103 captured. For example:
2104 <pre>
2105 A(A|B(*ACCEPT)|C)D
2106 </pre>
2107 This matches "AB", "AAD", or "ACD", but when it matches "AB", no data is
2108 captured.
2109 <pre>
2110 (*FAIL) or (*F)
2111 </pre>
2112 This verb causes the match to fail, forcing backtracking to occur. It is
2113 equivalent to (?!) but easier to read. The Perl documentation notes that it is
2114 probably useful only when combined with (?{}) or (??{}). Those are, of course,
2115 Perl features that are not present in PCRE. The nearest equivalent is the
2116 callout feature, as for example in this pattern:
2117 <pre>
2118 a+(?C)(*FAIL)
2119 </pre>
2120 A match with the string "aaaa" always fails, but the callout is taken before
2121 each backtrack happens (in this example, 10 times).
2122 </P>
2123 <br><b>
2124 Verbs that act after backtracking
2125 </b><br>
2126 <P>
2127 The following verbs do nothing when they are encountered. Matching continues
2128 with what follows, but if there is no subsequent match, a failure is forced.
2129 The verbs differ in exactly what kind of failure occurs.
2130 <pre>
2131 (*COMMIT)
2132 </pre>
2133 This verb causes the whole match to fail outright if the rest of the pattern
2134 does not match. Even if the pattern is unanchored, no further attempts to find
2135 a match by advancing the start point take place. Once (*COMMIT) has been
2136 passed, <b>pcre_exec()</b> is committed to finding a match at the current
2137 starting point, or not at all. For example:
2138 <pre>
2139 a+(*COMMIT)b
2140 </pre>
2141 This matches "xxaab" but not "aacaab". It can be thought of as a kind of
2142 dynamic anchor, or "I've started, so I must finish."
2143 <pre>
2144 (*PRUNE)
2145 </pre>
2146 This verb causes the match to fail at the current position if the rest of the
2147 pattern does not match. If the pattern is unanchored, the normal "bumpalong"
2148 advance to the next starting character then happens. Backtracking can occur as
2149 usual to the left of (*PRUNE), or when matching to the right of (*PRUNE), but
2150 if there is no match to the right, backtracking cannot cross (*PRUNE).
2151 In simple cases, the use of (*PRUNE) is just an alternative to an atomic
2152 group or possessive quantifier, but there are some uses of (*PRUNE) that cannot
2153 be expressed in any other way.
2154 <pre>
2155 (*SKIP)
2156 </pre>
2157 This verb is like (*PRUNE), except that if the pattern is unanchored, the
2158 "bumpalong" advance is not to the next character, but to the position in the
2159 subject where (*SKIP) was encountered. (*SKIP) signifies that whatever text
2160 was matched leading up to it cannot be part of a successful match. Consider:
2161 <pre>
2162 a+(*SKIP)b
2163 </pre>
2164 If the subject is "aaaac...", after the first match attempt fails (starting at
2165 the first character in the string), the starting point skips on to start the
2166 next attempt at "c". Note that a possessive quantifer does not have the same
2167 effect in this example; although it would suppress backtracking during the
2168 first match attempt, the second attempt would start at the second character
2169 instead of skipping on to "c".
2170 <pre>
2171 (*THEN)
2172 </pre>
2173 This verb causes a skip to the next alternation if the rest of the pattern does
2174 not match. That is, it cancels pending backtracking, but only within the
2175 current alternation. Its name comes from the observation that it can be used
2176 for a pattern-based if-then-else block:
2177 <pre>
2178 ( COND1 (*THEN) FOO | COND2 (*THEN) BAR | COND3 (*THEN) BAZ ) ...
2179 </pre>
2180 If the COND1 pattern matches, FOO is tried (and possibly further items after
2181 the end of the group if FOO succeeds); on failure the matcher skips to the
2182 second alternative and tries COND2, without backtracking into COND1. If (*THEN)
2183 is used outside of any alternation, it acts exactly like (*PRUNE).
2184 </P>
2185 <br><a name="SEC25" href="#TOC1">SEE ALSO</a><br>
2186 <P>
2187 <b>pcreapi</b>(3), <b>pcrecallout</b>(3), <b>pcrematching</b>(3), <b>pcre</b>(3).
2188 </P>
2189 <br><a name="SEC26" href="#TOC1">AUTHOR</a><br>
2190 <P>
2191 Philip Hazel
2192 <br>
2193 University Computing Service
2194 <br>
2195 Cambridge CB2 3QH, England.
2196 <br>
2197 </P>
2198 <br><a name="SEC27" href="#TOC1">REVISION</a><br>
2199 <P>
2200 Last updated: 17 September 2007
2201 <br>
2202 Copyright &copy; 1997-2007 University of Cambridge.
2203 <br>
2204 <p>
2205 Return to the <a href="index.html">PCRE index page</a>.
2206 </p>


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

ViewVC Help
Powered by ViewVC 1.1.12