ViewVC logotype

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

Parent Directory Parent Directory | Revision Log Revision Log

Revision 469 - (show annotations)
Mon Oct 19 14:38:48 2009 UTC (12 years ago) by ph10
File MIME type: text/html
File size: 102908 byte(s)
Error occurred while calculating annotation data.
Final doc and source tidies for 8.00
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">ONIGURUMA SUBROUTINE SYNTAX</a>
39 <li><a name="TOC24" href="#SEC24">CALLOUTS</a>
40 <li><a name="TOC25" href="#SEC25">BACKTRACKING CONTROL</a>
41 <li><a name="TOC26" href="#SEC26">SEE ALSO</a>
42 <li><a name="TOC27" href="#SEC27">AUTHOR</a>
43 <li><a name="TOC28" href="#SEC28">REVISION</a>
44 </ul>
45 <br><a name="SEC1" href="#TOC1">PCRE REGULAR EXPRESSION DETAILS</a><br>
46 <P>
47 The syntax and semantics of the regular expressions that are supported by PCRE
48 are described in detail below. There is a quick-reference syntax summary in the
49 <a href="pcresyntax.html"><b>pcresyntax</b></a>
50 page. PCRE tries to match Perl syntax and semantics as closely as it can. PCRE
51 also supports some alternative regular expression syntax (which does not
52 conflict with the Perl syntax) in order to provide some compatibility with
53 regular expressions in Python, .NET, and Oniguruma.
54 </P>
55 <P>
56 Perl's regular expressions are described in its own documentation, and
57 regular expressions in general are covered in a number of books, some of which
58 have copious examples. Jeffrey Friedl's "Mastering Regular Expressions",
59 published by O'Reilly, covers regular expressions in great detail. This
60 description of PCRE's regular expressions is intended as reference material.
61 </P>
62 <P>
63 The original operation of PCRE was on strings of one-byte characters. However,
64 there is now also support for UTF-8 character strings. To use this,
65 PCRE must be built to include UTF-8 support, and you must call
66 <b>pcre_compile()</b> or <b>pcre_compile2()</b> with the PCRE_UTF8 option. There
67 is also a special sequence that can be given at the start of a pattern:
68 <pre>
69 (*UTF8)
70 </pre>
71 Starting a pattern with this sequence is equivalent to setting the PCRE_UTF8
72 option. This feature is not Perl-compatible. How setting UTF-8 mode affects
73 pattern matching is mentioned in several places below. There is also a summary
74 of UTF-8 features in the
75 <a href="pcre.html#utf8support">section on UTF-8 support</a>
76 in the main
77 <a href="pcre.html"><b>pcre</b></a>
78 page.
79 </P>
80 <P>
81 The remainder of this document discusses the patterns that are supported by
82 PCRE when its main matching function, <b>pcre_exec()</b>, is used.
83 From release 6.0, PCRE offers a second matching function,
84 <b>pcre_dfa_exec()</b>, which matches using a different algorithm that is not
85 Perl-compatible. Some of the features discussed below are not available when
86 <b>pcre_dfa_exec()</b> is used. The advantages and disadvantages of the
87 alternative function, and how it differs from the normal function, are
88 discussed in the
89 <a href="pcrematching.html"><b>pcrematching</b></a>
90 page.
91 </P>
92 <br><a name="SEC2" href="#TOC1">NEWLINE CONVENTIONS</a><br>
93 <P>
94 PCRE supports five different conventions for indicating line breaks in
95 strings: a single CR (carriage return) character, a single LF (linefeed)
96 character, the two-character sequence CRLF, any of the three preceding, or any
97 Unicode newline sequence. The
98 <a href="pcreapi.html"><b>pcreapi</b></a>
99 page has
100 <a href="pcreapi.html#newlines">further discussion</a>
101 about newlines, and shows how to set the newline convention in the
102 <i>options</i> arguments for the compiling and matching functions.
103 </P>
104 <P>
105 It is also possible to specify a newline convention by starting a pattern
106 string with one of the following five sequences:
107 <pre>
108 (*CR) carriage return
109 (*LF) linefeed
110 (*CRLF) carriage return, followed by linefeed
111 (*ANYCRLF) any of the three above
112 (*ANY) all Unicode newline sequences
113 </pre>
114 These override the default and the options given to <b>pcre_compile()</b> or
115 <b>pcre_compile2()</b>. For example, on a Unix system where LF is the default
116 newline sequence, the pattern
117 <pre>
118 (*CR)a.b
119 </pre>
120 changes the convention to CR. That pattern matches "a\nb" because LF is no
121 longer a newline. Note that these special settings, which are not
122 Perl-compatible, are recognized only at the very start of a pattern, and that
123 they must be in upper case. If more than one of them is present, the last one
124 is used.
125 </P>
126 <P>
127 The newline convention does not affect what the \R escape sequence matches. By
128 default, this is any Unicode newline sequence, for Perl compatibility. However,
129 this can be changed; see the description of \R in the section entitled
130 <a href="#newlineseq">"Newline sequences"</a>
131 below. A change of \R setting can be combined with a change of newline
132 convention.
133 </P>
134 <br><a name="SEC3" href="#TOC1">CHARACTERS AND METACHARACTERS</a><br>
135 <P>
136 A regular expression is a pattern that is matched against a subject string from
137 left to right. Most characters stand for themselves in a pattern, and match the
138 corresponding characters in the subject. As a trivial example, the pattern
139 <pre>
140 The quick brown fox
141 </pre>
142 matches a portion of a subject string that is identical to itself. When
143 caseless matching is specified (the PCRE_CASELESS option), letters are matched
144 independently of case. In UTF-8 mode, PCRE always understands the concept of
145 case for characters whose values are less than 128, so caseless matching is
146 always possible. For characters with higher values, the concept of case is
147 supported if PCRE is compiled with Unicode property support, but not otherwise.
148 If you want to use caseless matching for characters 128 and above, you must
149 ensure that PCRE is compiled with Unicode property support as well as with
150 UTF-8 support.
151 </P>
152 <P>
153 The power of regular expressions comes from the ability to include alternatives
154 and repetitions in the pattern. These are encoded in the pattern by the use of
155 <i>metacharacters</i>, which do not stand for themselves but instead are
156 interpreted in some special way.
157 </P>
158 <P>
159 There are two different sets of metacharacters: those that are recognized
160 anywhere in the pattern except within square brackets, and those that are
161 recognized within square brackets. Outside square brackets, the metacharacters
162 are as follows:
163 <pre>
164 \ general escape character with several uses
165 ^ assert start of string (or line, in multiline mode)
166 $ assert end of string (or line, in multiline mode)
167 . match any character except newline (by default)
168 [ start character class definition
169 | start of alternative branch
170 ( start subpattern
171 ) end subpattern
172 ? extends the meaning of (
173 also 0 or 1 quantifier
174 also quantifier minimizer
175 * 0 or more quantifier
176 + 1 or more quantifier
177 also "possessive quantifier"
178 { start min/max quantifier
179 </pre>
180 Part of a pattern that is in square brackets is called a "character class". In
181 a character class the only metacharacters are:
182 <pre>
183 \ general escape character
184 ^ negate the class, but only if the first character
185 - indicates character range
186 [ POSIX character class (only if followed by POSIX syntax)
187 ] terminates the character class
188 </pre>
189 The following sections describe the use of each of the metacharacters.
190 </P>
191 <br><a name="SEC4" href="#TOC1">BACKSLASH</a><br>
192 <P>
193 The backslash character has several uses. Firstly, if it is followed by a
194 non-alphanumeric character, it takes away any special meaning that character
195 may have. This use of backslash as an escape character applies both inside and
196 outside character classes.
197 </P>
198 <P>
199 For example, if you want to match a * character, you write \* in the pattern.
200 This escaping action applies whether or not the following character would
201 otherwise be interpreted as a metacharacter, so it is always safe to precede a
202 non-alphanumeric with backslash to specify that it stands for itself. In
203 particular, if you want to match a backslash, you write \\.
204 </P>
205 <P>
206 If a pattern is compiled with the PCRE_EXTENDED option, whitespace in the
207 pattern (other than in a character class) and characters between a # outside
208 a character class and the next newline are ignored. An escaping backslash can
209 be used to include a whitespace or # character as part of the pattern.
210 </P>
211 <P>
212 If you want to remove the special meaning from a sequence of characters, you
213 can do so by putting them between \Q and \E. This is different from Perl in
214 that $ and @ are handled as literals in \Q...\E sequences in PCRE, whereas in
215 Perl, $ and @ cause variable interpolation. Note the following examples:
216 <pre>
217 Pattern PCRE matches Perl matches
219 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
220 \Qabc\$xyz\E abc\$xyz abc\$xyz
221 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
222 </pre>
223 The \Q...\E sequence is recognized both inside and outside character classes.
224 <a name="digitsafterbackslash"></a></P>
225 <br><b>
226 Non-printing characters
227 </b><br>
228 <P>
229 A second use of backslash provides a way of encoding non-printing characters
230 in patterns in a visible manner. There is no restriction on the appearance of
231 non-printing characters, apart from the binary zero that terminates a pattern,
232 but when a pattern is being prepared by text editing, it is often easier to use
233 one of the following escape sequences than the binary character it represents:
234 <pre>
235 \a alarm, that is, the BEL character (hex 07)
236 \cx "control-x", where x is any character
237 \e escape (hex 1B)
238 \f formfeed (hex 0C)
239 \n linefeed (hex 0A)
240 \r carriage return (hex 0D)
241 \t tab (hex 09)
242 \ddd character with octal code ddd, or backreference
243 \xhh character with hex code hh
244 \x{hhh..} character with hex code hhh..
245 </pre>
246 The precise effect of \cx is as follows: if x is a lower case letter, it
247 is converted to upper case. Then bit 6 of the character (hex 40) is inverted.
248 Thus \cz becomes hex 1A, but \c{ becomes hex 3B, while \c; becomes hex
249 7B.
250 </P>
251 <P>
252 After \x, from zero to two hexadecimal digits are read (letters can be in
253 upper or lower case). Any number of hexadecimal digits may appear between \x{
254 and }, but the value of the character code must be less than 256 in non-UTF-8
255 mode, and less than 2**31 in UTF-8 mode. That is, the maximum value in
256 hexadecimal is 7FFFFFFF. Note that this is bigger than the largest Unicode code
257 point, which is 10FFFF.
258 </P>
259 <P>
260 If characters other than hexadecimal digits appear between \x{ and }, or if
261 there is no terminating }, this form of escape is not recognized. Instead, the
262 initial \x will be interpreted as a basic hexadecimal escape, with no
263 following digits, giving a character whose value is zero.
264 </P>
265 <P>
266 Characters whose value is less than 256 can be defined by either of the two
267 syntaxes for \x. There is no difference in the way they are handled. For
268 example, \xdc is exactly the same as \x{dc}.
269 </P>
270 <P>
271 After \0 up to two further octal digits are read. If there are fewer than two
272 digits, just those that are present are used. Thus the sequence \0\x\07
273 specifies two binary zeros followed by a BEL character (code value 7). Make
274 sure you supply two digits after the initial zero if the pattern character that
275 follows is itself an octal digit.
276 </P>
277 <P>
278 The handling of a backslash followed by a digit other than 0 is complicated.
279 Outside a character class, PCRE reads it and any following digits as a decimal
280 number. If the number is less than 10, or if there have been at least that many
281 previous capturing left parentheses in the expression, the entire sequence is
282 taken as a <i>back reference</i>. A description of how this works is given
283 <a href="#backreferences">later,</a>
284 following the discussion of
285 <a href="#subpattern">parenthesized subpatterns.</a>
286 </P>
287 <P>
288 Inside a character class, or if the decimal number is greater than 9 and there
289 have not been that many capturing subpatterns, PCRE re-reads up to three octal
290 digits following the backslash, and uses them to generate a data character. Any
291 subsequent digits stand for themselves. In non-UTF-8 mode, the value of a
292 character specified in octal must be less than \400. In UTF-8 mode, values up
293 to \777 are permitted. For example:
294 <pre>
295 \040 is another way of writing a space
296 \40 is the same, provided there are fewer than 40 previous capturing subpatterns
297 \7 is always a back reference
298 \11 might be a back reference, or another way of writing a tab
299 \011 is always a tab
300 \0113 is a tab followed by the character "3"
301 \113 might be a back reference, otherwise the character with octal code 113
302 \377 might be a back reference, otherwise the byte consisting entirely of 1 bits
303 \81 is either a back reference, or a binary zero followed by the two characters "8" and "1"
304 </pre>
305 Note that octal values of 100 or greater must not be introduced by a leading
306 zero, because no more than three octal digits are ever read.
307 </P>
308 <P>
309 All the sequences that define a single character value can be used both inside
310 and outside character classes. In addition, inside a character class, the
311 sequence \b is interpreted as the backspace character (hex 08), and the
312 sequences \R and \X are interpreted as the characters "R" and "X",
313 respectively. Outside a character class, these sequences have different
314 meanings
315 <a href="#uniextseq">(see below).</a>
316 </P>
317 <br><b>
318 Absolute and relative back references
319 </b><br>
320 <P>
321 The sequence \g followed by an unsigned or a negative number, optionally
322 enclosed in braces, is an absolute or relative back reference. A named back
323 reference can be coded as \g{name}. Back references are discussed
324 <a href="#backreferences">later,</a>
325 following the discussion of
326 <a href="#subpattern">parenthesized subpatterns.</a>
327 </P>
328 <br><b>
329 Absolute and relative subroutine calls
330 </b><br>
331 <P>
332 For compatibility with Oniguruma, the non-Perl syntax \g followed by a name or
333 a number enclosed either in angle brackets or single quotes, is an alternative
334 syntax for referencing a subpattern as a "subroutine". Details are discussed
335 <a href="#onigurumasubroutines">later.</a>
336 Note that \g{...} (Perl syntax) and \g&#60;...&#62; (Oniguruma syntax) are <i>not</i>
337 synonymous. The former is a back reference; the latter is a
338 <a href="#subpatternsassubroutines">subroutine</a>
339 call.
340 </P>
341 <br><b>
342 Generic character types
343 </b><br>
344 <P>
345 Another use of backslash is for specifying generic character types. The
346 following are always recognized:
347 <pre>
348 \d any decimal digit
349 \D any character that is not a decimal digit
350 \h any horizontal whitespace character
351 \H any character that is not a horizontal whitespace character
352 \s any whitespace character
353 \S any character that is not a whitespace character
354 \v any vertical whitespace character
355 \V any character that is not a vertical whitespace character
356 \w any "word" character
357 \W any "non-word" character
358 </pre>
359 Each pair of escape sequences partitions the complete set of characters into
360 two disjoint sets. Any given character matches one, and only one, of each pair.
361 </P>
362 <P>
363 These character type sequences can appear both inside and outside character
364 classes. They each match one character of the appropriate type. If the current
365 matching point is at the end of the subject string, all of them fail, since
366 there is no character to match.
367 </P>
368 <P>
369 For compatibility with Perl, \s does not match the VT character (code 11).
370 This makes it different from the the POSIX "space" class. The \s characters
371 are HT (9), LF (10), FF (12), CR (13), and space (32). If "use locale;" is
372 included in a Perl script, \s may match the VT character. In PCRE, it never
373 does.
374 </P>
375 <P>
376 In UTF-8 mode, characters with values greater than 128 never match \d, \s, or
377 \w, and always match \D, \S, and \W. This is true even when Unicode
378 character property support is available. These sequences retain their original
379 meanings from before UTF-8 support was available, mainly for efficiency
380 reasons. Note that this also affects \b, because it is defined in terms of \w
381 and \W.
382 </P>
383 <P>
384 The sequences \h, \H, \v, and \V are Perl 5.10 features. In contrast to the
385 other sequences, these do match certain high-valued codepoints in UTF-8 mode.
386 The horizontal space characters are:
387 <pre>
388 U+0009 Horizontal tab
389 U+0020 Space
390 U+00A0 Non-break space
391 U+1680 Ogham space mark
392 U+180E Mongolian vowel separator
393 U+2000 En quad
394 U+2001 Em quad
395 U+2002 En space
396 U+2003 Em space
397 U+2004 Three-per-em space
398 U+2005 Four-per-em space
399 U+2006 Six-per-em space
400 U+2007 Figure space
401 U+2008 Punctuation space
402 U+2009 Thin space
403 U+200A Hair space
404 U+202F Narrow no-break space
405 U+205F Medium mathematical space
406 U+3000 Ideographic space
407 </pre>
408 The vertical space characters are:
409 <pre>
410 U+000A Linefeed
411 U+000B Vertical tab
412 U+000C Formfeed
413 U+000D Carriage return
414 U+0085 Next line
415 U+2028 Line separator
416 U+2029 Paragraph separator
417 </PRE>
418 </P>
419 <P>
420 A "word" character is an underscore or any character less than 256 that is a
421 letter or digit. The definition of letters and digits is controlled by PCRE's
422 low-valued character tables, and may vary if locale-specific matching is taking
423 place (see
424 <a href="pcreapi.html#localesupport">"Locale support"</a>
425 in the
426 <a href="pcreapi.html"><b>pcreapi</b></a>
427 page). For example, in a French locale such as "fr_FR" in Unix-like systems,
428 or "french" in Windows, some character codes greater than 128 are used for
429 accented letters, and these are matched by \w. The use of locales with Unicode
430 is discouraged.
431 <a name="newlineseq"></a></P>
432 <br><b>
433 Newline sequences
434 </b><br>
435 <P>
436 Outside a character class, by default, the escape sequence \R matches any
437 Unicode newline sequence. This is a Perl 5.10 feature. In non-UTF-8 mode \R is
438 equivalent to the following:
439 <pre>
440 (?&#62;\r\n|\n|\x0b|\f|\r|\x85)
441 </pre>
442 This is an example of an "atomic group", details of which are given
443 <a href="#atomicgroup">below.</a>
444 This particular group matches either the two-character sequence CR followed by
445 LF, or one of the single characters LF (linefeed, U+000A), VT (vertical tab,
446 U+000B), FF (formfeed, U+000C), CR (carriage return, U+000D), or NEL (next
447 line, U+0085). The two-character sequence is treated as a single unit that
448 cannot be split.
449 </P>
450 <P>
451 In UTF-8 mode, two additional characters whose codepoints are greater than 255
452 are added: LS (line separator, U+2028) and PS (paragraph separator, U+2029).
453 Unicode character property support is not needed for these characters to be
454 recognized.
455 </P>
456 <P>
457 It is possible to restrict \R to match only CR, LF, or CRLF (instead of the
458 complete set of Unicode line endings) by setting the option PCRE_BSR_ANYCRLF
459 either at compile time or when the pattern is matched. (BSR is an abbrevation
460 for "backslash R".) This can be made the default when PCRE is built; if this is
461 the case, the other behaviour can be requested via the PCRE_BSR_UNICODE option.
462 It is also possible to specify these settings by starting a pattern string with
463 one of the following sequences:
464 <pre>
465 (*BSR_ANYCRLF) CR, LF, or CRLF only
466 (*BSR_UNICODE) any Unicode newline sequence
467 </pre>
468 These override the default and the options given to <b>pcre_compile()</b> or
469 <b>pcre_compile2()</b>, but they can be overridden by options given to
470 <b>pcre_exec()</b> or <b>pcre_dfa_exec()</b>. Note that these special settings,
471 which are not Perl-compatible, are recognized only at the very start of a
472 pattern, and that they must be in upper case. If more than one of them is
473 present, the last one is used. They can be combined with a change of newline
474 convention, for example, a pattern can start with:
475 <pre>
477 </pre>
478 Inside a character class, \R matches the letter "R".
479 <a name="uniextseq"></a></P>
480 <br><b>
481 Unicode character properties
482 </b><br>
483 <P>
484 When PCRE is built with Unicode character property support, three additional
485 escape sequences that match characters with specific properties are available.
486 When not in UTF-8 mode, these sequences are of course limited to testing
487 characters whose codepoints are less than 256, but they do work in this mode.
488 The extra escape sequences are:
489 <pre>
490 \p{<i>xx</i>} a character with the <i>xx</i> property
491 \P{<i>xx</i>} a character without the <i>xx</i> property
492 \X an extended Unicode sequence
493 </pre>
494 The property names represented by <i>xx</i> above are limited to the Unicode
495 script names, the general category properties, and "Any", which matches any
496 character (including newline). Other properties such as "InMusicalSymbols" are
497 not currently supported by PCRE. Note that \P{Any} does not match any
498 characters, so always causes a match failure.
499 </P>
500 <P>
501 Sets of Unicode characters are defined as belonging to certain scripts. A
502 character from one of these sets can be matched using a script name. For
503 example:
504 <pre>
505 \p{Greek}
506 \P{Han}
507 </pre>
508 Those that are not part of an identified script are lumped together as
509 "Common". The current list of scripts is:
510 </P>
511 <P>
512 Arabic,
513 Armenian,
514 Balinese,
515 Bengali,
516 Bopomofo,
517 Braille,
518 Buginese,
519 Buhid,
520 Canadian_Aboriginal,
521 Cherokee,
522 Common,
523 Coptic,
524 Cuneiform,
525 Cypriot,
526 Cyrillic,
527 Deseret,
528 Devanagari,
529 Ethiopic,
530 Georgian,
531 Glagolitic,
532 Gothic,
533 Greek,
534 Gujarati,
535 Gurmukhi,
536 Han,
537 Hangul,
538 Hanunoo,
539 Hebrew,
540 Hiragana,
541 Inherited,
542 Kannada,
543 Katakana,
544 Kharoshthi,
545 Khmer,
546 Lao,
547 Latin,
548 Limbu,
549 Linear_B,
550 Malayalam,
551 Mongolian,
552 Myanmar,
553 New_Tai_Lue,
554 Nko,
555 Ogham,
556 Old_Italic,
557 Old_Persian,
558 Oriya,
559 Osmanya,
560 Phags_Pa,
561 Phoenician,
562 Runic,
563 Shavian,
564 Sinhala,
565 Syloti_Nagri,
566 Syriac,
567 Tagalog,
568 Tagbanwa,
569 Tai_Le,
570 Tamil,
571 Telugu,
572 Thaana,
573 Thai,
574 Tibetan,
575 Tifinagh,
576 Ugaritic,
577 Yi.
578 </P>
579 <P>
580 Each character has exactly one general category property, specified by a
581 two-letter abbreviation. For compatibility with Perl, negation can be specified
582 by including a circumflex between the opening brace and the property name. For
583 example, \p{^Lu} is the same as \P{Lu}.
584 </P>
585 <P>
586 If only one letter is specified with \p or \P, it includes all the general
587 category properties that start with that letter. In this case, in the absence
588 of negation, the curly brackets in the escape sequence are optional; these two
589 examples have the same effect:
590 <pre>
591 \p{L}
592 \pL
593 </pre>
594 The following general category property codes are supported:
595 <pre>
596 C Other
597 Cc Control
598 Cf Format
599 Cn Unassigned
600 Co Private use
601 Cs Surrogate
603 L Letter
604 Ll Lower case letter
605 Lm Modifier letter
606 Lo Other letter
607 Lt Title case letter
608 Lu Upper case letter
610 M Mark
611 Mc Spacing mark
612 Me Enclosing mark
613 Mn Non-spacing mark
615 N Number
616 Nd Decimal number
617 Nl Letter number
618 No Other number
620 P Punctuation
621 Pc Connector punctuation
622 Pd Dash punctuation
623 Pe Close punctuation
624 Pf Final punctuation
625 Pi Initial punctuation
626 Po Other punctuation
627 Ps Open punctuation
629 S Symbol
630 Sc Currency symbol
631 Sk Modifier symbol
632 Sm Mathematical symbol
633 So Other symbol
635 Z Separator
636 Zl Line separator
637 Zp Paragraph separator
638 Zs Space separator
639 </pre>
640 The special property L& is also supported: it matches a character that has
641 the Lu, Ll, or Lt property, in other words, a letter that is not classified as
642 a modifier or "other".
643 </P>
644 <P>
645 The Cs (Surrogate) property applies only to characters in the range U+D800 to
646 U+DFFF. Such characters are not valid in UTF-8 strings (see RFC 3629) and so
647 cannot be tested by PCRE, unless UTF-8 validity checking has been turned off
648 (see the discussion of PCRE_NO_UTF8_CHECK in the
649 <a href="pcreapi.html"><b>pcreapi</b></a>
650 page). Perl does not support the Cs property.
651 </P>
652 <P>
653 The long synonyms for property names that Perl supports (such as \p{Letter})
654 are not supported by PCRE, nor is it permitted to prefix any of these
655 properties with "Is".
656 </P>
657 <P>
658 No character that is in the Unicode table has the Cn (unassigned) property.
659 Instead, this property is assumed for any code point that is not in the
660 Unicode table.
661 </P>
662 <P>
663 Specifying caseless matching does not affect these escape sequences. For
664 example, \p{Lu} always matches only upper case letters.
665 </P>
666 <P>
667 The \X escape matches any number of Unicode characters that form an extended
668 Unicode sequence. \X is equivalent to
669 <pre>
670 (?&#62;\PM\pM*)
671 </pre>
672 That is, it matches a character without the "mark" property, followed by zero
673 or more characters with the "mark" property, and treats the sequence as an
674 atomic group
675 <a href="#atomicgroup">(see below).</a>
676 Characters with the "mark" property are typically accents that affect the
677 preceding character. None of them have codepoints less than 256, so in
678 non-UTF-8 mode \X matches any one character.
679 </P>
680 <P>
681 Matching characters by Unicode property is not fast, because PCRE has to search
682 a structure that contains data for over fifteen thousand characters. That is
683 why the traditional escape sequences such as \d and \w do not use Unicode
684 properties in PCRE.
685 <a name="resetmatchstart"></a></P>
686 <br><b>
687 Resetting the match start
688 </b><br>
689 <P>
690 The escape sequence \K, which is a Perl 5.10 feature, causes any previously
691 matched characters not to be included in the final matched sequence. For
692 example, the pattern:
693 <pre>
694 foo\Kbar
695 </pre>
696 matches "foobar", but reports that it has matched "bar". This feature is
697 similar to a lookbehind assertion
698 <a href="#lookbehind">(described below).</a>
699 However, in this case, the part of the subject before the real match does not
700 have to be of fixed length, as lookbehind assertions do. The use of \K does
701 not interfere with the setting of
702 <a href="#subpattern">captured substrings.</a>
703 For example, when the pattern
704 <pre>
705 (foo)\Kbar
706 </pre>
707 matches "foobar", the first substring is still set to "foo".
708 <a name="smallassertions"></a></P>
709 <br><b>
710 Simple assertions
711 </b><br>
712 <P>
713 The final use of backslash is for certain simple assertions. An assertion
714 specifies a condition that has to be met at a particular point in a match,
715 without consuming any characters from the subject string. The use of
716 subpatterns for more complicated assertions is described
717 <a href="#bigassertions">below.</a>
718 The backslashed assertions are:
719 <pre>
720 \b matches at a word boundary
721 \B matches when not at a word boundary
722 \A matches at the start of the subject
723 \Z matches at the end of the subject
724 also matches before a newline at the end of the subject
725 \z matches only at the end of the subject
726 \G matches at the first matching position in the subject
727 </pre>
728 These assertions may not appear in character classes (but note that \b has a
729 different meaning, namely the backspace character, inside a character class).
730 </P>
731 <P>
732 A word boundary is a position in the subject string where the current character
733 and the previous character do not both match \w or \W (i.e. one matches
734 \w and the other matches \W), or the start or end of the string if the
735 first or last character matches \w, respectively. Neither PCRE nor Perl has a
736 separte "start of word" or "end of word" metasequence. However, whatever
737 follows \b normally determines which it is. For example, the fragment
738 \ba matches "a" at the start of a word.
739 </P>
740 <P>
741 The \A, \Z, and \z assertions differ from the traditional circumflex and
742 dollar (described in the next section) in that they only ever match at the very
743 start and end of the subject string, whatever options are set. Thus, they are
744 independent of multiline mode. These three assertions are not affected by the
745 PCRE_NOTBOL or PCRE_NOTEOL options, which affect only the behaviour of the
746 circumflex and dollar metacharacters. However, if the <i>startoffset</i>
747 argument of <b>pcre_exec()</b> is non-zero, indicating that matching is to start
748 at a point other than the beginning of the subject, \A can never match. The
749 difference between \Z and \z is that \Z matches before a newline at the end
750 of the string as well as at the very end, whereas \z matches only at the end.
751 </P>
752 <P>
753 The \G assertion is true only when the current matching position is at the
754 start point of the match, as specified by the <i>startoffset</i> argument of
755 <b>pcre_exec()</b>. It differs from \A when the value of <i>startoffset</i> is
756 non-zero. By calling <b>pcre_exec()</b> multiple times with appropriate
757 arguments, you can mimic Perl's /g option, and it is in this kind of
758 implementation where \G can be useful.
759 </P>
760 <P>
761 Note, however, that PCRE's interpretation of \G, as the start of the current
762 match, is subtly different from Perl's, which defines it as the end of the
763 previous match. In Perl, these can be different when the previously matched
764 string was empty. Because PCRE does just one match at a time, it cannot
765 reproduce this behaviour.
766 </P>
767 <P>
768 If all the alternatives of a pattern begin with \G, the expression is anchored
769 to the starting match position, and the "anchored" flag is set in the compiled
770 regular expression.
771 </P>
772 <br><a name="SEC5" href="#TOC1">CIRCUMFLEX AND DOLLAR</a><br>
773 <P>
774 Outside a character class, in the default matching mode, the circumflex
775 character is an assertion that is true only if the current matching point is
776 at the start of the subject string. If the <i>startoffset</i> argument of
777 <b>pcre_exec()</b> is non-zero, circumflex can never match if the PCRE_MULTILINE
778 option is unset. Inside a character class, circumflex has an entirely different
779 meaning
780 <a href="#characterclass">(see below).</a>
781 </P>
782 <P>
783 Circumflex need not be the first character of the pattern if a number of
784 alternatives are involved, but it should be the first thing in each alternative
785 in which it appears if the pattern is ever to match that branch. If all
786 possible alternatives start with a circumflex, that is, if the pattern is
787 constrained to match only at the start of the subject, it is said to be an
788 "anchored" pattern. (There are also other constructs that can cause a pattern
789 to be anchored.)
790 </P>
791 <P>
792 A dollar character is an assertion that is true only if the current matching
793 point is at the end of the subject string, or immediately before a newline
794 at the end of the string (by default). Dollar need not be the last character of
795 the pattern if a number of alternatives are involved, but it should be the last
796 item in any branch in which it appears. Dollar has no special meaning in a
797 character class.
798 </P>
799 <P>
800 The meaning of dollar can be changed so that it matches only at the very end of
801 the string, by setting the PCRE_DOLLAR_ENDONLY option at compile time. This
802 does not affect the \Z assertion.
803 </P>
804 <P>
805 The meanings of the circumflex and dollar characters are changed if the
806 PCRE_MULTILINE option is set. When this is the case, a circumflex matches
807 immediately after internal newlines as well as at the start of the subject
808 string. It does not match after a newline that ends the string. A dollar
809 matches before any newlines in the string, as well as at the very end, when
810 PCRE_MULTILINE is set. When newline is specified as the two-character
811 sequence CRLF, isolated CR and LF characters do not indicate newlines.
812 </P>
813 <P>
814 For example, the pattern /^abc$/ matches the subject string "def\nabc" (where
815 \n represents a newline) in multiline mode, but not otherwise. Consequently,
816 patterns that are anchored in single line mode because all branches start with
817 ^ are not anchored in multiline mode, and a match for circumflex is possible
818 when the <i>startoffset</i> argument of <b>pcre_exec()</b> is non-zero. The
819 PCRE_DOLLAR_ENDONLY option is ignored if PCRE_MULTILINE is set.
820 </P>
821 <P>
822 Note that the sequences \A, \Z, and \z can be used to match the start and
823 end of the subject in both modes, and if all branches of a pattern start with
824 \A it is always anchored, whether or not PCRE_MULTILINE is set.
825 </P>
826 <br><a name="SEC6" href="#TOC1">FULL STOP (PERIOD, DOT)</a><br>
827 <P>
828 Outside a character class, a dot in the pattern matches any one character in
829 the subject string except (by default) a character that signifies the end of a
830 line. In UTF-8 mode, the matched character may be more than one byte long.
831 </P>
832 <P>
833 When a line ending is defined as a single character, dot never matches that
834 character; when the two-character sequence CRLF is used, dot does not match CR
835 if it is immediately followed by LF, but otherwise it matches all characters
836 (including isolated CRs and LFs). When any Unicode line endings are being
837 recognized, dot does not match CR or LF or any of the other line ending
838 characters.
839 </P>
840 <P>
841 The behaviour of dot with regard to newlines can be changed. If the PCRE_DOTALL
842 option is set, a dot matches any one character, without exception. If the
843 two-character sequence CRLF is present in the subject string, it takes two dots
844 to match it.
845 </P>
846 <P>
847 The handling of dot is entirely independent of the handling of circumflex and
848 dollar, the only relationship being that they both involve newlines. Dot has no
849 special meaning in a character class.
850 </P>
851 <br><a name="SEC7" href="#TOC1">MATCHING A SINGLE BYTE</a><br>
852 <P>
853 Outside a character class, the escape sequence \C matches any one byte, both
854 in and out of UTF-8 mode. Unlike a dot, it always matches any line-ending
855 characters. The feature is provided in Perl in order to match individual bytes
856 in UTF-8 mode. Because it breaks up UTF-8 characters into individual bytes,
857 what remains in the string may be a malformed UTF-8 string. For this reason,
858 the \C escape sequence is best avoided.
859 </P>
860 <P>
861 PCRE does not allow \C to appear in lookbehind assertions
862 <a href="#lookbehind">(described below),</a>
863 because in UTF-8 mode this would make it impossible to calculate the length of
864 the lookbehind.
865 <a name="characterclass"></a></P>
866 <br><a name="SEC8" href="#TOC1">SQUARE BRACKETS AND CHARACTER CLASSES</a><br>
867 <P>
868 An opening square bracket introduces a character class, terminated by a closing
869 square bracket. A closing square bracket on its own is not special by default.
870 However, if the PCRE_JAVASCRIPT_COMPAT option is set, a lone closing square
871 bracket causes a compile-time error. If a closing square bracket is required as
872 a member of the class, it should be the first data character in the class
873 (after an initial circumflex, if present) or escaped with a backslash.
874 </P>
875 <P>
876 A character class matches a single character in the subject. In UTF-8 mode, the
877 character may be more than one byte long. A matched character must be in the
878 set of characters defined by the class, unless the first character in the class
879 definition is a circumflex, in which case the subject character must not be in
880 the set defined by the class. If a circumflex is actually required as a member
881 of the class, ensure it is not the first character, or escape it with a
882 backslash.
883 </P>
884 <P>
885 For example, the character class [aeiou] matches any lower case vowel, while
886 [^aeiou] matches any character that is not a lower case vowel. Note that a
887 circumflex is just a convenient notation for specifying the characters that
888 are in the class by enumerating those that are not. A class that starts with a
889 circumflex is not an assertion; it still consumes a character from the subject
890 string, and therefore it fails if the current pointer is at the end of the
891 string.
892 </P>
893 <P>
894 In UTF-8 mode, characters with values greater than 255 can be included in a
895 class as a literal string of bytes, or by using the \x{ escaping mechanism.
896 </P>
897 <P>
898 When caseless matching is set, any letters in a class represent both their
899 upper case and lower case versions, so for example, a caseless [aeiou] matches
900 "A" as well as "a", and a caseless [^aeiou] does not match "A", whereas a
901 caseful version would. In UTF-8 mode, PCRE always understands the concept of
902 case for characters whose values are less than 128, so caseless matching is
903 always possible. For characters with higher values, the concept of case is
904 supported if PCRE is compiled with Unicode property support, but not otherwise.
905 If you want to use caseless matching in UTF8-mode for characters 128 and above,
906 you must ensure that PCRE is compiled with Unicode property support as well as
907 with UTF-8 support.
908 </P>
909 <P>
910 Characters that might indicate line breaks are never treated in any special way
911 when matching character classes, whatever line-ending sequence is in use, and
912 whatever setting of the PCRE_DOTALL and PCRE_MULTILINE options is used. A class
913 such as [^a] always matches one of these characters.
914 </P>
915 <P>
916 The minus (hyphen) character can be used to specify a range of characters in a
917 character class. For example, [d-m] matches any letter between d and m,
918 inclusive. If a minus character is required in a class, it must be escaped with
919 a backslash or appear in a position where it cannot be interpreted as
920 indicating a range, typically as the first or last character in the class.
921 </P>
922 <P>
923 It is not possible to have the literal character "]" as the end character of a
924 range. A pattern such as [W-]46] is interpreted as a class of two characters
925 ("W" and "-") followed by a literal string "46]", so it would match "W46]" or
926 "-46]". However, if the "]" is escaped with a backslash it is interpreted as
927 the end of range, so [W-\]46] is interpreted as a class containing a range
928 followed by two other characters. The octal or hexadecimal representation of
929 "]" can also be used to end a range.
930 </P>
931 <P>
932 Ranges operate in the collating sequence of character values. They can also be
933 used for characters specified numerically, for example [\000-\037]. In UTF-8
934 mode, ranges can include characters whose values are greater than 255, for
935 example [\x{100}-\x{2ff}].
936 </P>
937 <P>
938 If a range that includes letters is used when caseless matching is set, it
939 matches the letters in either case. For example, [W-c] is equivalent to
940 [][\\^_`wxyzabc], matched caselessly, and in non-UTF-8 mode, if character
941 tables for a French locale are in use, [\xc8-\xcb] matches accented E
942 characters in both cases. In UTF-8 mode, PCRE supports the concept of case for
943 characters with values greater than 128 only when it is compiled with Unicode
944 property support.
945 </P>
946 <P>
947 The character types \d, \D, \p, \P, \s, \S, \w, and \W may also appear
948 in a character class, and add the characters that they match to the class. For
949 example, [\dABCDEF] matches any hexadecimal digit. A circumflex can
950 conveniently be used with the upper case character types to specify a more
951 restricted set of characters than the matching lower case type. For example,
952 the class [^\W_] matches any letter or digit, but not underscore.
953 </P>
954 <P>
955 The only metacharacters that are recognized in character classes are backslash,
956 hyphen (only where it can be interpreted as specifying a range), circumflex
957 (only at the start), opening square bracket (only when it can be interpreted as
958 introducing a POSIX class name - see the next section), and the terminating
959 closing square bracket. However, escaping other non-alphanumeric characters
960 does no harm.
961 </P>
962 <br><a name="SEC9" href="#TOC1">POSIX CHARACTER CLASSES</a><br>
963 <P>
964 Perl supports the POSIX notation for character classes. This uses names
965 enclosed by [: and :] within the enclosing square brackets. PCRE also supports
966 this notation. For example,
967 <pre>
968 [01[:alpha:]%]
969 </pre>
970 matches "0", "1", any alphabetic character, or "%". The supported class names
971 are
972 <pre>
973 alnum letters and digits
974 alpha letters
975 ascii character codes 0 - 127
976 blank space or tab only
977 cntrl control characters
978 digit decimal digits (same as \d)
979 graph printing characters, excluding space
980 lower lower case letters
981 print printing characters, including space
982 punct printing characters, excluding letters and digits
983 space white space (not quite the same as \s)
984 upper upper case letters
985 word "word" characters (same as \w)
986 xdigit hexadecimal digits
987 </pre>
988 The "space" characters are HT (9), LF (10), VT (11), FF (12), CR (13), and
989 space (32). Notice that this list includes the VT character (code 11). This
990 makes "space" different to \s, which does not include VT (for Perl
991 compatibility).
992 </P>
993 <P>
994 The name "word" is a Perl extension, and "blank" is a GNU extension from Perl
995 5.8. Another Perl extension is negation, which is indicated by a ^ character
996 after the colon. For example,
997 <pre>
998 [12[:^digit:]]
999 </pre>
1000 matches "1", "2", or any non-digit. PCRE (and Perl) also recognize the POSIX
1001 syntax [.ch.] and [=ch=] where "ch" is a "collating element", but these are not
1002 supported, and an error is given if they are encountered.
1003 </P>
1004 <P>
1005 In UTF-8 mode, characters with values greater than 128 do not match any of
1006 the POSIX character classes.
1007 </P>
1008 <br><a name="SEC10" href="#TOC1">VERTICAL BAR</a><br>
1009 <P>
1010 Vertical bar characters are used to separate alternative patterns. For example,
1011 the pattern
1012 <pre>
1013 gilbert|sullivan
1014 </pre>
1015 matches either "gilbert" or "sullivan". Any number of alternatives may appear,
1016 and an empty alternative is permitted (matching the empty string). The matching
1017 process tries each alternative in turn, from left to right, and the first one
1018 that succeeds is used. If the alternatives are within a subpattern
1019 <a href="#subpattern">(defined below),</a>
1020 "succeeds" means matching the rest of the main pattern as well as the
1021 alternative in the subpattern.
1022 </P>
1023 <br><a name="SEC11" href="#TOC1">INTERNAL OPTION SETTING</a><br>
1024 <P>
1025 The settings of the PCRE_CASELESS, PCRE_MULTILINE, PCRE_DOTALL, and
1026 PCRE_EXTENDED options (which are Perl-compatible) can be changed from within
1027 the pattern by a sequence of Perl option letters enclosed between "(?" and ")".
1028 The option letters are
1029 <pre>
1030 i for PCRE_CASELESS
1032 s for PCRE_DOTALL
1033 x for PCRE_EXTENDED
1034 </pre>
1035 For example, (?im) sets caseless, multiline matching. It is also possible to
1036 unset these options by preceding the letter with a hyphen, and a combined
1037 setting and unsetting such as (?im-sx), which sets PCRE_CASELESS and
1038 PCRE_MULTILINE while unsetting PCRE_DOTALL and PCRE_EXTENDED, is also
1039 permitted. If a letter appears both before and after the hyphen, the option is
1040 unset.
1041 </P>
1042 <P>
1043 The PCRE-specific options PCRE_DUPNAMES, PCRE_UNGREEDY, and PCRE_EXTRA can be
1044 changed in the same way as the Perl-compatible options by using the characters
1045 J, U and X respectively.
1046 </P>
1047 <P>
1048 When one of these option changes occurs at top level (that is, not inside
1049 subpattern parentheses), the change applies to the remainder of the pattern
1050 that follows. If the change is placed right at the start of a pattern, PCRE
1051 extracts it into the global options (and it will therefore show up in data
1052 extracted by the <b>pcre_fullinfo()</b> function).
1053 </P>
1054 <P>
1055 An option change within a subpattern (see below for a description of
1056 subpatterns) affects only that part of the current pattern that follows it, so
1057 <pre>
1058 (a(?i)b)c
1059 </pre>
1060 matches abc and aBc and no other strings (assuming PCRE_CASELESS is not used).
1061 By this means, options can be made to have different settings in different
1062 parts of the pattern. Any changes made in one alternative do carry on
1063 into subsequent branches within the same subpattern. For example,
1064 <pre>
1065 (a(?i)b|c)
1066 </pre>
1067 matches "ab", "aB", "c", and "C", even though when matching "C" the first
1068 branch is abandoned before the option setting. This is because the effects of
1069 option settings happen at compile time. There would be some very weird
1070 behaviour otherwise.
1071 </P>
1072 <P>
1073 <b>Note:</b> There are other PCRE-specific options that can be set by the
1074 application when the compile or match functions are called. In some cases the
1075 pattern can contain special leading sequences such as (*CRLF) to override what
1076 the application has set or what has been defaulted. Details are given in the
1077 section entitled
1078 <a href="#newlineseq">"Newline sequences"</a>
1079 above. There is also the (*UTF8) leading sequence that can be used to set UTF-8
1080 mode; this is equivalent to setting the PCRE_UTF8 option.
1081 <a name="subpattern"></a></P>
1082 <br><a name="SEC12" href="#TOC1">SUBPATTERNS</a><br>
1083 <P>
1084 Subpatterns are delimited by parentheses (round brackets), which can be nested.
1085 Turning part of a pattern into a subpattern does two things:
1086 <br>
1087 <br>
1088 1. It localizes a set of alternatives. For example, the pattern
1089 <pre>
1090 cat(aract|erpillar|)
1091 </pre>
1092 matches one of the words "cat", "cataract", or "caterpillar". Without the
1093 parentheses, it would match "cataract", "erpillar" or an empty string.
1094 <br>
1095 <br>
1096 2. It sets up the subpattern as a capturing subpattern. This means that, when
1097 the whole pattern matches, that portion of the subject string that matched the
1098 subpattern is passed back to the caller via the <i>ovector</i> argument of
1099 <b>pcre_exec()</b>. Opening parentheses are counted from left to right (starting
1100 from 1) to obtain numbers for the capturing subpatterns.
1101 </P>
1102 <P>
1103 For example, if the string "the red king" is matched against the pattern
1104 <pre>
1105 the ((red|white) (king|queen))
1106 </pre>
1107 the captured substrings are "red king", "red", and "king", and are numbered 1,
1108 2, and 3, respectively.
1109 </P>
1110 <P>
1111 The fact that plain parentheses fulfil two functions is not always helpful.
1112 There are often times when a grouping subpattern is required without a
1113 capturing requirement. If an opening parenthesis is followed by a question mark
1114 and a colon, the subpattern does not do any capturing, and is not counted when
1115 computing the number of any subsequent capturing subpatterns. For example, if
1116 the string "the white queen" is matched against the pattern
1117 <pre>
1118 the ((?:red|white) (king|queen))
1119 </pre>
1120 the captured substrings are "white queen" and "queen", and are numbered 1 and
1121 2. The maximum number of capturing subpatterns is 65535.
1122 </P>
1123 <P>
1124 As a convenient shorthand, if any option settings are required at the start of
1125 a non-capturing subpattern, the option letters may appear between the "?" and
1126 the ":". Thus the two patterns
1127 <pre>
1128 (?i:saturday|sunday)
1129 (?:(?i)saturday|sunday)
1130 </pre>
1131 match exactly the same set of strings. Because alternative branches are tried
1132 from left to right, and options are not reset until the end of the subpattern
1133 is reached, an option setting in one branch does affect subsequent branches, so
1134 the above patterns match "SUNDAY" as well as "Saturday".
1135 <a name="dupsubpatternnumber"></a></P>
1136 <br><a name="SEC13" href="#TOC1">DUPLICATE SUBPATTERN NUMBERS</a><br>
1137 <P>
1138 Perl 5.10 introduced a feature whereby each alternative in a subpattern uses
1139 the same numbers for its capturing parentheses. Such a subpattern starts with
1140 (?| and is itself a non-capturing subpattern. For example, consider this
1141 pattern:
1142 <pre>
1143 (?|(Sat)ur|(Sun))day
1144 </pre>
1145 Because the two alternatives are inside a (?| group, both sets of capturing
1146 parentheses are numbered one. Thus, when the pattern matches, you can look
1147 at captured substring number one, whichever alternative matched. This construct
1148 is useful when you want to capture part, but not all, of one of a number of
1149 alternatives. Inside a (?| group, parentheses are numbered as usual, but the
1150 number is reset at the start of each branch. The numbers of any capturing
1151 buffers that follow the subpattern start after the highest number used in any
1152 branch. The following example is taken from the Perl documentation.
1153 The numbers underneath show in which buffer the captured content will be
1154 stored.
1155 <pre>
1156 # before ---------------branch-reset----------- after
1157 / ( a ) (?| x ( y ) z | (p (q) r) | (t) u (v) ) ( z ) /x
1158 # 1 2 2 3 2 3 4
1159 </pre>
1160 A backreference to a numbered subpattern uses the most recent value that is set
1161 for that number by any subpattern. The following pattern matches "abcabc" or
1162 "defdef":
1163 <pre>
1164 /(?|(abc)|(def))\1/
1165 </pre>
1166 In contrast, a recursive or "subroutine" call to a numbered subpattern always
1167 refers to the first one in the pattern with the given number. The following
1168 pattern matches "abcabc" or "defabc":
1169 <pre>
1170 /(?|(abc)|(def))(?1)/
1171 </pre>
1172 If a
1173 <a href="#conditions">condition test</a>
1174 for a subpattern's having matched refers to a non-unique number, the test is
1175 true if any of the subpatterns of that number have matched.
1176 </P>
1177 <P>
1178 An alternative approach to using this "branch reset" feature is to use
1179 duplicate named subpatterns, as described in the next section.
1180 </P>
1181 <br><a name="SEC14" href="#TOC1">NAMED SUBPATTERNS</a><br>
1182 <P>
1183 Identifying capturing parentheses by number is simple, but it can be very hard
1184 to keep track of the numbers in complicated regular expressions. Furthermore,
1185 if an expression is modified, the numbers may change. To help with this
1186 difficulty, PCRE supports the naming of subpatterns. This feature was not
1187 added to Perl until release 5.10. Python had the feature earlier, and PCRE
1188 introduced it at release 4.0, using the Python syntax. PCRE now supports both
1189 the Perl and the Python syntax. Perl allows identically numbered subpatterns to
1190 have different names, but PCRE does not.
1191 </P>
1192 <P>
1193 In PCRE, a subpattern can be named in one of three ways: (?&#60;name&#62;...) or
1194 (?'name'...) as in Perl, or (?P&#60;name&#62;...) as in Python. References to capturing
1195 parentheses from other parts of the pattern, such as
1196 <a href="#backreferences">backreferences,</a>
1197 <a href="#recursion">recursion,</a>
1198 and
1199 <a href="#conditions">conditions,</a>
1200 can be made by name as well as by number.
1201 </P>
1202 <P>
1203 Names consist of up to 32 alphanumeric characters and underscores. Named
1204 capturing parentheses are still allocated numbers as well as names, exactly as
1205 if the names were not present. The PCRE API provides function calls for
1206 extracting the name-to-number translation table from a compiled pattern. There
1207 is also a convenience function for extracting a captured substring by name.
1208 </P>
1209 <P>
1210 By default, a name must be unique within a pattern, but it is possible to relax
1211 this constraint by setting the PCRE_DUPNAMES option at compile time. (Duplicate
1212 names are also always permitted for subpatterns with the same number, set up as
1213 described in the previous section.) Duplicate names can be useful for patterns
1214 where only one instance of the named parentheses can match. Suppose you want to
1215 match the name of a weekday, either as a 3-letter abbreviation or as the full
1216 name, and in both cases you want to extract the abbreviation. This pattern
1217 (ignoring the line breaks) does the job:
1218 <pre>
1219 (?&#60;DN&#62;Mon|Fri|Sun)(?:day)?|
1220 (?&#60;DN&#62;Tue)(?:sday)?|
1221 (?&#60;DN&#62;Wed)(?:nesday)?|
1222 (?&#60;DN&#62;Thu)(?:rsday)?|
1223 (?&#60;DN&#62;Sat)(?:urday)?
1224 </pre>
1225 There are five capturing substrings, but only one is ever set after a match.
1226 (An alternative way of solving this problem is to use a "branch reset"
1227 subpattern, as described in the previous section.)
1228 </P>
1229 <P>
1230 The convenience function for extracting the data by name returns the substring
1231 for the first (and in this example, the only) subpattern of that name that
1232 matched. This saves searching to find which numbered subpattern it was.
1233 </P>
1234 <P>
1235 If you make a backreference to a non-unique named subpattern from elsewhere in
1236 the pattern, the one that corresponds to the first occurrence of the name is
1237 used. In the absence of duplicate numbers (see the previous section) this is
1238 the one with the lowest number. If you use a named reference in a condition
1239 test (see the
1240 <a href="#conditions">section about conditions</a>
1241 below), either to check whether a subpattern has matched, or to check for
1242 recursion, all subpatterns with the same name are tested. If the condition is
1243 true for any one of them, the overall condition is true. This is the same
1244 behaviour as testing by number. For further details of the interfaces for
1245 handling named subpatterns, see the
1246 <a href="pcreapi.html"><b>pcreapi</b></a>
1247 documentation.
1248 </P>
1249 <P>
1250 <b>Warning:</b> You cannot use different names to distinguish between two
1251 subpatterns with the same number because PCRE uses only the numbers when
1252 matching. For this reason, an error is given at compile time if different names
1253 are given to subpatterns with the same number. However, you can give the same
1254 name to subpatterns with the same number, even when PCRE_DUPNAMES is not set.
1255 </P>
1256 <br><a name="SEC15" href="#TOC1">REPETITION</a><br>
1257 <P>
1258 Repetition is specified by quantifiers, which can follow any of the following
1259 items:
1260 <pre>
1261 a literal data character
1262 the dot metacharacter
1263 the \C escape sequence
1264 the \X escape sequence (in UTF-8 mode with Unicode properties)
1265 the \R escape sequence
1266 an escape such as \d that matches a single character
1267 a character class
1268 a back reference (see next section)
1269 a parenthesized subpattern (unless it is an assertion)
1270 a recursive or "subroutine" call to a subpattern
1271 </pre>
1272 The general repetition quantifier specifies a minimum and maximum number of
1273 permitted matches, by giving the two numbers in curly brackets (braces),
1274 separated by a comma. The numbers must be less than 65536, and the first must
1275 be less than or equal to the second. For example:
1276 <pre>
1277 z{2,4}
1278 </pre>
1279 matches "zz", "zzz", or "zzzz". A closing brace on its own is not a special
1280 character. If the second number is omitted, but the comma is present, there is
1281 no upper limit; if the second number and the comma are both omitted, the
1282 quantifier specifies an exact number of required matches. Thus
1283 <pre>
1284 [aeiou]{3,}
1285 </pre>
1286 matches at least 3 successive vowels, but may match many more, while
1287 <pre>
1288 \d{8}
1289 </pre>
1290 matches exactly 8 digits. An opening curly bracket that appears in a position
1291 where a quantifier is not allowed, or one that does not match the syntax of a
1292 quantifier, is taken as a literal character. For example, {,6} is not a
1293 quantifier, but a literal string of four characters.
1294 </P>
1295 <P>
1296 In UTF-8 mode, quantifiers apply to UTF-8 characters rather than to individual
1297 bytes. Thus, for example, \x{100}{2} matches two UTF-8 characters, each of
1298 which is represented by a two-byte sequence. Similarly, when Unicode property
1299 support is available, \X{3} matches three Unicode extended sequences, each of
1300 which may be several bytes long (and they may be of different lengths).
1301 </P>
1302 <P>
1303 The quantifier {0} is permitted, causing the expression to behave as if the
1304 previous item and the quantifier were not present. This may be useful for
1305 subpatterns that are referenced as
1306 <a href="#subpatternsassubroutines">subroutines</a>
1307 from elsewhere in the pattern. Items other than subpatterns that have a {0}
1308 quantifier are omitted from the compiled pattern.
1309 </P>
1310 <P>
1311 For convenience, the three most common quantifiers have single-character
1312 abbreviations:
1313 <pre>
1314 * is equivalent to {0,}
1315 + is equivalent to {1,}
1316 ? is equivalent to {0,1}
1317 </pre>
1318 It is possible to construct infinite loops by following a subpattern that can
1319 match no characters with a quantifier that has no upper limit, for example:
1320 <pre>
1321 (a?)*
1322 </pre>
1323 Earlier versions of Perl and PCRE used to give an error at compile time for
1324 such patterns. However, because there are cases where this can be useful, such
1325 patterns are now accepted, but if any repetition of the subpattern does in fact
1326 match no characters, the loop is forcibly broken.
1327 </P>
1328 <P>
1329 By default, the quantifiers are "greedy", that is, they match as much as
1330 possible (up to the maximum number of permitted times), without causing the
1331 rest of the pattern to fail. The classic example of where this gives problems
1332 is in trying to match comments in C programs. These appear between /* and */
1333 and within the comment, individual * and / characters may appear. An attempt to
1334 match C comments by applying the pattern
1335 <pre>
1336 /\*.*\*/
1337 </pre>
1338 to the string
1339 <pre>
1340 /* first comment */ not comment /* second comment */
1341 </pre>
1342 fails, because it matches the entire string owing to the greediness of the .*
1343 item.
1344 </P>
1345 <P>
1346 However, if a quantifier is followed by a question mark, it ceases to be
1347 greedy, and instead matches the minimum number of times possible, so the
1348 pattern
1349 <pre>
1350 /\*.*?\*/
1351 </pre>
1352 does the right thing with the C comments. The meaning of the various
1353 quantifiers is not otherwise changed, just the preferred number of matches.
1354 Do not confuse this use of question mark with its use as a quantifier in its
1355 own right. Because it has two uses, it can sometimes appear doubled, as in
1356 <pre>
1357 \d??\d
1358 </pre>
1359 which matches one digit by preference, but can match two if that is the only
1360 way the rest of the pattern matches.
1361 </P>
1362 <P>
1363 If the PCRE_UNGREEDY option is set (an option that is not available in Perl),
1364 the quantifiers are not greedy by default, but individual ones can be made
1365 greedy by following them with a question mark. In other words, it inverts the
1366 default behaviour.
1367 </P>
1368 <P>
1369 When a parenthesized subpattern is quantified with a minimum repeat count that
1370 is greater than 1 or with a limited maximum, more memory is required for the
1371 compiled pattern, in proportion to the size of the minimum or maximum.
1372 </P>
1373 <P>
1374 If a pattern starts with .* or .{0,} and the PCRE_DOTALL option (equivalent
1375 to Perl's /s) is set, thus allowing the dot to match newlines, the pattern is
1376 implicitly anchored, because whatever follows will be tried against every
1377 character position in the subject string, so there is no point in retrying the
1378 overall match at any position after the first. PCRE normally treats such a
1379 pattern as though it were preceded by \A.
1380 </P>
1381 <P>
1382 In cases where it is known that the subject string contains no newlines, it is
1383 worth setting PCRE_DOTALL in order to obtain this optimization, or
1384 alternatively using ^ to indicate anchoring explicitly.
1385 </P>
1386 <P>
1387 However, there is one situation where the optimization cannot be used. When .*
1388 is inside capturing parentheses that are the subject of a backreference
1389 elsewhere in the pattern, a match at the start may fail where a later one
1390 succeeds. Consider, for example:
1391 <pre>
1392 (.*)abc\1
1393 </pre>
1394 If the subject is "xyz123abc123" the match point is the fourth character. For
1395 this reason, such a pattern is not implicitly anchored.
1396 </P>
1397 <P>
1398 When a capturing subpattern is repeated, the value captured is the substring
1399 that matched the final iteration. For example, after
1400 <pre>
1401 (tweedle[dume]{3}\s*)+
1402 </pre>
1403 has matched "tweedledum tweedledee" the value of the captured substring is
1404 "tweedledee". However, if there are nested capturing subpatterns, the
1405 corresponding captured values may have been set in previous iterations. For
1406 example, after
1407 <pre>
1408 /(a|(b))+/
1409 </pre>
1410 matches "aba" the value of the second captured substring is "b".
1411 <a name="atomicgroup"></a></P>
1412 <br><a name="SEC16" href="#TOC1">ATOMIC GROUPING AND POSSESSIVE QUANTIFIERS</a><br>
1413 <P>
1414 With both maximizing ("greedy") and minimizing ("ungreedy" or "lazy")
1415 repetition, failure of what follows normally causes the repeated item to be
1416 re-evaluated to see if a different number of repeats allows the rest of the
1417 pattern to match. Sometimes it is useful to prevent this, either to change the
1418 nature of the match, or to cause it fail earlier than it otherwise might, when
1419 the author of the pattern knows there is no point in carrying on.
1420 </P>
1421 <P>
1422 Consider, for example, the pattern \d+foo when applied to the subject line
1423 <pre>
1424 123456bar
1425 </pre>
1426 After matching all 6 digits and then failing to match "foo", the normal
1427 action of the matcher is to try again with only 5 digits matching the \d+
1428 item, and then with 4, and so on, before ultimately failing. "Atomic grouping"
1429 (a term taken from Jeffrey Friedl's book) provides the means for specifying
1430 that once a subpattern has matched, it is not to be re-evaluated in this way.
1431 </P>
1432 <P>
1433 If we use atomic grouping for the previous example, the matcher gives up
1434 immediately on failing to match "foo" the first time. The notation is a kind of
1435 special parenthesis, starting with (?&#62; as in this example:
1436 <pre>
1437 (?&#62;\d+)foo
1438 </pre>
1439 This kind of parenthesis "locks up" the part of the pattern it contains once
1440 it has matched, and a failure further into the pattern is prevented from
1441 backtracking into it. Backtracking past it to previous items, however, works as
1442 normal.
1443 </P>
1444 <P>
1445 An alternative description is that a subpattern of this type matches the string
1446 of characters that an identical standalone pattern would match, if anchored at
1447 the current point in the subject string.
1448 </P>
1449 <P>
1450 Atomic grouping subpatterns are not capturing subpatterns. Simple cases such as
1451 the above example can be thought of as a maximizing repeat that must swallow
1452 everything it can. So, while both \d+ and \d+? are prepared to adjust the
1453 number of digits they match in order to make the rest of the pattern match,
1454 (?&#62;\d+) can only match an entire sequence of digits.
1455 </P>
1456 <P>
1457 Atomic groups in general can of course contain arbitrarily complicated
1458 subpatterns, and can be nested. However, when the subpattern for an atomic
1459 group is just a single repeated item, as in the example above, a simpler
1460 notation, called a "possessive quantifier" can be used. This consists of an
1461 additional + character following a quantifier. Using this notation, the
1462 previous example can be rewritten as
1463 <pre>
1464 \d++foo
1465 </pre>
1466 Note that a possessive quantifier can be used with an entire group, for
1467 example:
1468 <pre>
1469 (abc|xyz){2,3}+
1470 </pre>
1471 Possessive quantifiers are always greedy; the setting of the PCRE_UNGREEDY
1472 option is ignored. They are a convenient notation for the simpler forms of
1473 atomic group. However, there is no difference in the meaning of a possessive
1474 quantifier and the equivalent atomic group, though there may be a performance
1475 difference; possessive quantifiers should be slightly faster.
1476 </P>
1477 <P>
1478 The possessive quantifier syntax is an extension to the Perl 5.8 syntax.
1479 Jeffrey Friedl originated the idea (and the name) in the first edition of his
1480 book. Mike McCloskey liked it, so implemented it when he built Sun's Java
1481 package, and PCRE copied it from there. It ultimately found its way into Perl
1482 at release 5.10.
1483 </P>
1484 <P>
1485 PCRE has an optimization that automatically "possessifies" certain simple
1486 pattern constructs. For example, the sequence A+B is treated as A++B because
1487 there is no point in backtracking into a sequence of A's when B must follow.
1488 </P>
1489 <P>
1490 When a pattern contains an unlimited repeat inside a subpattern that can itself
1491 be repeated an unlimited number of times, the use of an atomic group is the
1492 only way to avoid some failing matches taking a very long time indeed. The
1493 pattern
1494 <pre>
1495 (\D+|&#60;\d+&#62;)*[!?]
1496 </pre>
1497 matches an unlimited number of substrings that either consist of non-digits, or
1498 digits enclosed in &#60;&#62;, followed by either ! or ?. When it matches, it runs
1499 quickly. However, if it is applied to
1500 <pre>
1501 aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
1502 </pre>
1503 it takes a long time before reporting failure. This is because the string can
1504 be divided between the internal \D+ repeat and the external * repeat in a
1505 large number of ways, and all have to be tried. (The example uses [!?] rather
1506 than a single character at the end, because both PCRE and Perl have an
1507 optimization that allows for fast failure when a single character is used. They
1508 remember the last single character that is required for a match, and fail early
1509 if it is not present in the string.) If the pattern is changed so that it uses
1510 an atomic group, like this:
1511 <pre>
1512 ((?&#62;\D+)|&#60;\d+&#62;)*[!?]
1513 </pre>
1514 sequences of non-digits cannot be broken, and failure happens quickly.
1515 <a name="backreferences"></a></P>
1516 <br><a name="SEC17" href="#TOC1">BACK REFERENCES</a><br>
1517 <P>
1518 Outside a character class, a backslash followed by a digit greater than 0 (and
1519 possibly further digits) is a back reference to a capturing subpattern earlier
1520 (that is, to its left) in the pattern, provided there have been that many
1521 previous capturing left parentheses.
1522 </P>
1523 <P>
1524 However, if the decimal number following the backslash is less than 10, it is
1525 always taken as a back reference, and causes an error only if there are not
1526 that many capturing left parentheses in the entire pattern. In other words, the
1527 parentheses that are referenced need not be to the left of the reference for
1528 numbers less than 10. A "forward back reference" of this type can make sense
1529 when a repetition is involved and the subpattern to the right has participated
1530 in an earlier iteration.
1531 </P>
1532 <P>
1533 It is not possible to have a numerical "forward back reference" to a subpattern
1534 whose number is 10 or more using this syntax because a sequence such as \50 is
1535 interpreted as a character defined in octal. See the subsection entitled
1536 "Non-printing characters"
1537 <a href="#digitsafterbackslash">above</a>
1538 for further details of the handling of digits following a backslash. There is
1539 no such problem when named parentheses are used. A back reference to any
1540 subpattern is possible using named parentheses (see below).
1541 </P>
1542 <P>
1543 Another way of avoiding the ambiguity inherent in the use of digits following a
1544 backslash is to use the \g escape sequence, which is a feature introduced in
1545 Perl 5.10. This escape must be followed by an unsigned number or a negative
1546 number, optionally enclosed in braces. These examples are all identical:
1547 <pre>
1548 (ring), \1
1549 (ring), \g1
1550 (ring), \g{1}
1551 </pre>
1552 An unsigned number specifies an absolute reference without the ambiguity that
1553 is present in the older syntax. It is also useful when literal digits follow
1554 the reference. A negative number is a relative reference. Consider this
1555 example:
1556 <pre>
1557 (abc(def)ghi)\g{-1}
1558 </pre>
1559 The sequence \g{-1} is a reference to the most recently started capturing
1560 subpattern before \g, that is, is it equivalent to \2. Similarly, \g{-2}
1561 would be equivalent to \1. The use of relative references can be helpful in
1562 long patterns, and also in patterns that are created by joining together
1563 fragments that contain references within themselves.
1564 </P>
1565 <P>
1566 A back reference matches whatever actually matched the capturing subpattern in
1567 the current subject string, rather than anything matching the subpattern
1568 itself (see
1569 <a href="#subpatternsassubroutines">"Subpatterns as subroutines"</a>
1570 below for a way of doing that). So the pattern
1571 <pre>
1572 (sens|respons)e and \1ibility
1573 </pre>
1574 matches "sense and sensibility" and "response and responsibility", but not
1575 "sense and responsibility". If caseful matching is in force at the time of the
1576 back reference, the case of letters is relevant. For example,
1577 <pre>
1578 ((?i)rah)\s+\1
1579 </pre>
1580 matches "rah rah" and "RAH RAH", but not "RAH rah", even though the original
1581 capturing subpattern is matched caselessly.
1582 </P>
1583 <P>
1584 There are several different ways of writing back references to named
1585 subpatterns. The .NET syntax \k{name} and the Perl syntax \k&#60;name&#62; or
1586 \k'name' are supported, as is the Python syntax (?P=name). Perl 5.10's unified
1587 back reference syntax, in which \g can be used for both numeric and named
1588 references, is also supported. We could rewrite the above example in any of
1589 the following ways:
1590 <pre>
1591 (?&#60;p1&#62;(?i)rah)\s+\k&#60;p1&#62;
1592 (?'p1'(?i)rah)\s+\k{p1}
1593 (?P&#60;p1&#62;(?i)rah)\s+(?P=p1)
1594 (?&#60;p1&#62;(?i)rah)\s+\g{p1}
1595 </pre>
1596 A subpattern that is referenced by name may appear in the pattern before or
1597 after the reference.
1598 </P>
1599 <P>
1600 There may be more than one back reference to the same subpattern. If a
1601 subpattern has not actually been used in a particular match, any back
1602 references to it always fail by default. For example, the pattern
1603 <pre>
1604 (a|(bc))\2
1605 </pre>
1606 always fails if it starts to match "a" rather than "bc". However, if the
1607 PCRE_JAVASCRIPT_COMPAT option is set at compile time, a back reference to an
1608 unset value matches an empty string.
1609 </P>
1610 <P>
1611 Because there may be many capturing parentheses in a pattern, all digits
1612 following a backslash are taken as part of a potential back reference number.
1613 If the pattern continues with a digit character, some delimiter must be used to
1614 terminate the back reference. If the PCRE_EXTENDED option is set, this can be
1615 whitespace. Otherwise, the \g{ syntax or an empty comment (see
1616 <a href="#comments">"Comments"</a>
1617 below) can be used.
1618 </P>
1619 <P>
1620 A back reference that occurs inside the parentheses to which it refers fails
1621 when the subpattern is first used, so, for example, (a\1) never matches.
1622 However, such references can be useful inside repeated subpatterns. For
1623 example, the pattern
1624 <pre>
1625 (a|b\1)+
1626 </pre>
1627 matches any number of "a"s and also "aba", "ababbaa" etc. At each iteration of
1628 the subpattern, the back reference matches the character string corresponding
1629 to the previous iteration. In order for this to work, the pattern must be such
1630 that the first iteration does not need to match the back reference. This can be
1631 done using alternation, as in the example above, or by a quantifier with a
1632 minimum of zero.
1633 <a name="bigassertions"></a></P>
1634 <br><a name="SEC18" href="#TOC1">ASSERTIONS</a><br>
1635 <P>
1636 An assertion is a test on the characters following or preceding the current
1637 matching point that does not actually consume any characters. The simple
1638 assertions coded as \b, \B, \A, \G, \Z, \z, ^ and $ are described
1639 <a href="#smallassertions">above.</a>
1640 </P>
1641 <P>
1642 More complicated assertions are coded as subpatterns. There are two kinds:
1643 those that look ahead of the current position in the subject string, and those
1644 that look behind it. An assertion subpattern is matched in the normal way,
1645 except that it does not cause the current matching position to be changed.
1646 </P>
1647 <P>
1648 Assertion subpatterns are not capturing subpatterns, and may not be repeated,
1649 because it makes no sense to assert the same thing several times. If any kind
1650 of assertion contains capturing subpatterns within it, these are counted for
1651 the purposes of numbering the capturing subpatterns in the whole pattern.
1652 However, substring capturing is carried out only for positive assertions,
1653 because it does not make sense for negative assertions.
1654 </P>
1655 <br><b>
1656 Lookahead assertions
1657 </b><br>
1658 <P>
1659 Lookahead assertions start with (?= for positive assertions and (?! for
1660 negative assertions. For example,
1661 <pre>
1662 \w+(?=;)
1663 </pre>
1664 matches a word followed by a semicolon, but does not include the semicolon in
1665 the match, and
1666 <pre>
1667 foo(?!bar)
1668 </pre>
1669 matches any occurrence of "foo" that is not followed by "bar". Note that the
1670 apparently similar pattern
1671 <pre>
1672 (?!foo)bar
1673 </pre>
1674 does not find an occurrence of "bar" that is preceded by something other than
1675 "foo"; it finds any occurrence of "bar" whatsoever, because the assertion
1676 (?!foo) is always true when the next three characters are "bar". A
1677 lookbehind assertion is needed to achieve the other effect.
1678 </P>
1679 <P>
1680 If you want to force a matching failure at some point in a pattern, the most
1681 convenient way to do it is with (?!) because an empty string always matches, so
1682 an assertion that requires there not to be an empty string must always fail.
1683 The Perl 5.10 backtracking control verb (*FAIL) or (*F) is essentially a
1684 synonym for (?!).
1685 <a name="lookbehind"></a></P>
1686 <br><b>
1687 Lookbehind assertions
1688 </b><br>
1689 <P>
1690 Lookbehind assertions start with (?&#60;= for positive assertions and (?&#60;! for
1691 negative assertions. For example,
1692 <pre>
1693 (?&#60;!foo)bar
1694 </pre>
1695 does find an occurrence of "bar" that is not preceded by "foo". The contents of
1696 a lookbehind assertion are restricted such that all the strings it matches must
1697 have a fixed length. However, if there are several top-level alternatives, they
1698 do not all have to have the same fixed length. Thus
1699 <pre>
1700 (?&#60;=bullock|donkey)
1701 </pre>
1702 is permitted, but
1703 <pre>
1704 (?&#60;!dogs?|cats?)
1705 </pre>
1706 causes an error at compile time. Branches that match different length strings
1707 are permitted only at the top level of a lookbehind assertion. This is an
1708 extension compared with Perl (5.8 and 5.10), which requires all branches to
1709 match the same length of string. An assertion such as
1710 <pre>
1711 (?&#60;=ab(c|de))
1712 </pre>
1713 is not permitted, because its single top-level branch can match two different
1714 lengths, but it is acceptable to PCRE if rewritten to use two top-level
1715 branches:
1716 <pre>
1717 (?&#60;=abc|abde)
1718 </pre>
1719 In some cases, the Perl 5.10 escape sequence \K
1720 <a href="#resetmatchstart">(see above)</a>
1721 can be used instead of a lookbehind assertion to get round the fixed-length
1722 restriction.
1723 </P>
1724 <P>
1725 The implementation of lookbehind assertions is, for each alternative, to
1726 temporarily move the current position back by the fixed length and then try to
1727 match. If there are insufficient characters before the current position, the
1728 assertion fails.
1729 </P>
1730 <P>
1731 PCRE does not allow the \C escape (which matches a single byte in UTF-8 mode)
1732 to appear in lookbehind assertions, because it makes it impossible to calculate
1733 the length of the lookbehind. The \X and \R escapes, which can match
1734 different numbers of bytes, are also not permitted.
1735 </P>
1736 <P>
1737 <a href="#subpatternsassubroutines">"Subroutine"</a>
1738 calls (see below) such as (?2) or (?&X) are permitted in lookbehinds, as long
1739 as the subpattern matches a fixed-length string.
1740 <a href="#recursion">Recursion,</a>
1741 however, is not supported.
1742 </P>
1743 <P>
1744 Possessive quantifiers can be used in conjunction with lookbehind assertions to
1745 specify efficient matching of fixed-length strings at the end of subject
1746 strings. Consider a simple pattern such as
1747 <pre>
1748 abcd$
1749 </pre>
1750 when applied to a long string that does not match. Because matching proceeds
1751 from left to right, PCRE will look for each "a" in the subject and then see if
1752 what follows matches the rest of the pattern. If the pattern is specified as
1753 <pre>
1754 ^.*abcd$
1755 </pre>
1756 the initial .* matches the entire string at first, but when this fails (because
1757 there is no following "a"), it backtracks to match all but the last character,
1758 then all but the last two characters, and so on. Once again the search for "a"
1759 covers the entire string, from right to left, so we are no better off. However,
1760 if the pattern is written as
1761 <pre>
1762 ^.*+(?&#60;=abcd)
1763 </pre>
1764 there can be no backtracking for the .*+ item; it can match only the entire
1765 string. The subsequent lookbehind assertion does a single test on the last four
1766 characters. If it fails, the match fails immediately. For long strings, this
1767 approach makes a significant difference to the processing time.
1768 </P>
1769 <br><b>
1770 Using multiple assertions
1771 </b><br>
1772 <P>
1773 Several assertions (of any sort) may occur in succession. For example,
1774 <pre>
1775 (?&#60;=\d{3})(?&#60;!999)foo
1776 </pre>
1777 matches "foo" preceded by three digits that are not "999". Notice that each of
1778 the assertions is applied independently at the same point in the subject
1779 string. First there is a check that the previous three characters are all
1780 digits, and then there is a check that the same three characters are not "999".
1781 This pattern does <i>not</i> match "foo" preceded by six characters, the first
1782 of which are digits and the last three of which are not "999". For example, it
1783 doesn't match "123abcfoo". A pattern to do that is
1784 <pre>
1785 (?&#60;=\d{3}...)(?&#60;!999)foo
1786 </pre>
1787 This time the first assertion looks at the preceding six characters, checking
1788 that the first three are digits, and then the second assertion checks that the
1789 preceding three characters are not "999".
1790 </P>
1791 <P>
1792 Assertions can be nested in any combination. For example,
1793 <pre>
1794 (?&#60;=(?&#60;!foo)bar)baz
1795 </pre>
1796 matches an occurrence of "baz" that is preceded by "bar" which in turn is not
1797 preceded by "foo", while
1798 <pre>
1799 (?&#60;=\d{3}(?!999)...)foo
1800 </pre>
1801 is another pattern that matches "foo" preceded by three digits and any three
1802 characters that are not "999".
1803 <a name="conditions"></a></P>
1804 <br><a name="SEC19" href="#TOC1">CONDITIONAL SUBPATTERNS</a><br>
1805 <P>
1806 It is possible to cause the matching process to obey a subpattern
1807 conditionally or to choose between two alternative subpatterns, depending on
1808 the result of an assertion, or whether a specific capturing subpattern has
1809 already been matched. The two possible forms of conditional subpattern are:
1810 <pre>
1811 (?(condition)yes-pattern)
1812 (?(condition)yes-pattern|no-pattern)
1813 </pre>
1814 If the condition is satisfied, the yes-pattern is used; otherwise the
1815 no-pattern (if present) is used. If there are more than two alternatives in the
1816 subpattern, a compile-time error occurs.
1817 </P>
1818 <P>
1819 There are four kinds of condition: references to subpatterns, references to
1820 recursion, a pseudo-condition called DEFINE, and assertions.
1821 </P>
1822 <br><b>
1823 Checking for a used subpattern by number
1824 </b><br>
1825 <P>
1826 If the text between the parentheses consists of a sequence of digits, the
1827 condition is true if a capturing subpattern of that number has previously
1828 matched. If there is more than one capturing subpattern with the same number
1829 (see the earlier
1830 <a href="#recursion">section about duplicate subpattern numbers),</a>
1831 the condition is true if any of them have been set. An alternative notation is
1832 to precede the digits with a plus or minus sign. In this case, the subpattern
1833 number is relative rather than absolute. The most recently opened parentheses
1834 can be referenced by (?(-1), the next most recent by (?(-2), and so on. In
1835 looping constructs it can also make sense to refer to subsequent groups with
1836 constructs such as (?(+2).
1837 </P>
1838 <P>
1839 Consider the following pattern, which contains non-significant white space to
1840 make it more readable (assume the PCRE_EXTENDED option) and to divide it into
1841 three parts for ease of discussion:
1842 <pre>
1843 ( \( )? [^()]+ (?(1) \) )
1844 </pre>
1845 The first part matches an optional opening parenthesis, and if that
1846 character is present, sets it as the first captured substring. The second part
1847 matches one or more characters that are not parentheses. The third part is a
1848 conditional subpattern that tests whether the first set of parentheses matched
1849 or not. If they did, that is, if subject started with an opening parenthesis,
1850 the condition is true, and so the yes-pattern is executed and a closing
1851 parenthesis is required. Otherwise, since no-pattern is not present, the
1852 subpattern matches nothing. In other words, this pattern matches a sequence of
1853 non-parentheses, optionally enclosed in parentheses.
1854 </P>
1855 <P>
1856 If you were embedding this pattern in a larger one, you could use a relative
1857 reference:
1858 <pre>
1859 ...other stuff... ( \( )? [^()]+ (?(-1) \) ) ...
1860 </pre>
1861 This makes the fragment independent of the parentheses in the larger pattern.
1862 </P>
1863 <br><b>
1864 Checking for a used subpattern by name
1865 </b><br>
1866 <P>
1867 Perl uses the syntax (?(&#60;name&#62;)...) or (?('name')...) to test for a used
1868 subpattern by name. For compatibility with earlier versions of PCRE, which had
1869 this facility before Perl, the syntax (?(name)...) is also recognized. However,
1870 there is a possible ambiguity with this syntax, because subpattern names may
1871 consist entirely of digits. PCRE looks first for a named subpattern; if it
1872 cannot find one and the name consists entirely of digits, PCRE looks for a
1873 subpattern of that number, which must be greater than zero. Using subpattern
1874 names that consist entirely of digits is not recommended.
1875 </P>
1876 <P>
1877 Rewriting the above example to use a named subpattern gives this:
1878 <pre>
1879 (?&#60;OPEN&#62; \( )? [^()]+ (?(&#60;OPEN&#62;) \) )
1880 </pre>
1881 If the name used in a condition of this kind is a duplicate, the test is
1882 applied to all subpatterns of the same name, and is true if any one of them has
1883 matched.
1884 </P>
1885 <br><b>
1886 Checking for pattern recursion
1887 </b><br>
1888 <P>
1889 If the condition is the string (R), and there is no subpattern with the name R,
1890 the condition is true if a recursive call to the whole pattern or any
1891 subpattern has been made. If digits or a name preceded by ampersand follow the
1892 letter R, for example:
1893 <pre>
1894 (?(R3)...) or (?(R&name)...)
1895 </pre>
1896 the condition is true if the most recent recursion is into a subpattern whose
1897 number or name is given. This condition does not check the entire recursion
1898 stack. If the name used in a condition of this kind is a duplicate, the test is
1899 applied to all subpatterns of the same name, and is true if any one of them is
1900 the most recent recursion.
1901 </P>
1902 <P>
1903 At "top level", all these recursion test conditions are false.
1904 <a href="#recursion">The syntax for recursive patterns</a>
1905 is described below.
1906 </P>
1907 <br><b>
1908 Defining subpatterns for use by reference only
1909 </b><br>
1910 <P>
1911 If the condition is the string (DEFINE), and there is no subpattern with the
1912 name DEFINE, the condition is always false. In this case, there may be only one
1913 alternative in the subpattern. It is always skipped if control reaches this
1914 point in the pattern; the idea of DEFINE is that it can be used to define
1915 "subroutines" that can be referenced from elsewhere. (The use of
1916 <a href="#subpatternsassubroutines">"subroutines"</a>
1917 is described below.) For example, a pattern to match an IPv4 address could be
1918 written like this (ignore whitespace and line breaks):
1919 <pre>
1920 (?(DEFINE) (?&#60;byte&#62; 2[0-4]\d | 25[0-5] | 1\d\d | [1-9]?\d) )
1921 \b (?&byte) (\.(?&byte)){3} \b
1922 </pre>
1923 The first part of the pattern is a DEFINE group inside which a another group
1924 named "byte" is defined. This matches an individual component of an IPv4
1925 address (a number less than 256). When matching takes place, this part of the
1926 pattern is skipped because DEFINE acts like a false condition. The rest of the
1927 pattern uses references to the named group to match the four dot-separated
1928 components of an IPv4 address, insisting on a word boundary at each end.
1929 </P>
1930 <br><b>
1931 Assertion conditions
1932 </b><br>
1933 <P>
1934 If the condition is not in any of the above formats, it must be an assertion.
1935 This may be a positive or negative lookahead or lookbehind assertion. Consider
1936 this pattern, again containing non-significant white space, and with the two
1937 alternatives on the second line:
1938 <pre>
1939 (?(?=[^a-z]*[a-z])
1940 \d{2}-[a-z]{3}-\d{2} | \d{2}-\d{2}-\d{2} )
1941 </pre>
1942 The condition is a positive lookahead assertion that matches an optional
1943 sequence of non-letters followed by a letter. In other words, it tests for the
1944 presence of at least one letter in the subject. If a letter is found, the
1945 subject is matched against the first alternative; otherwise it is matched
1946 against the second. This pattern matches strings in one of the two forms
1947 dd-aaa-dd or dd-dd-dd, where aaa are letters and dd are digits.
1948 <a name="comments"></a></P>
1949 <br><a name="SEC20" href="#TOC1">COMMENTS</a><br>
1950 <P>
1951 The sequence (?# marks the start of a comment that continues up to the next
1952 closing parenthesis. Nested parentheses are not permitted. The characters
1953 that make up a comment play no part in the pattern matching at all.
1954 </P>
1955 <P>
1956 If the PCRE_EXTENDED option is set, an unescaped # character outside a
1957 character class introduces a comment that continues to immediately after the
1958 next newline in the pattern.
1959 <a name="recursion"></a></P>
1960 <br><a name="SEC21" href="#TOC1">RECURSIVE PATTERNS</a><br>
1961 <P>
1962 Consider the problem of matching a string in parentheses, allowing for
1963 unlimited nested parentheses. Without the use of recursion, the best that can
1964 be done is to use a pattern that matches up to some fixed depth of nesting. It
1965 is not possible to handle an arbitrary nesting depth.
1966 </P>
1967 <P>
1968 For some time, Perl has provided a facility that allows regular expressions to
1969 recurse (amongst other things). It does this by interpolating Perl code in the
1970 expression at run time, and the code can refer to the expression itself. A Perl
1971 pattern using code interpolation to solve the parentheses problem can be
1972 created like this:
1973 <pre>
1974 $re = qr{\( (?: (?&#62;[^()]+) | (?p{$re}) )* \)}x;
1975 </pre>
1976 The (?p{...}) item interpolates Perl code at run time, and in this case refers
1977 recursively to the pattern in which it appears.
1978 </P>
1979 <P>
1980 Obviously, PCRE cannot support the interpolation of Perl code. Instead, it
1981 supports special syntax for recursion of the entire pattern, and also for
1982 individual subpattern recursion. After its introduction in PCRE and Python,
1983 this kind of recursion was subsequently introduced into Perl at release 5.10.
1984 </P>
1985 <P>
1986 A special item that consists of (? followed by a number greater than zero and a
1987 closing parenthesis is a recursive call of the subpattern of the given number,
1988 provided that it occurs inside that subpattern. (If not, it is a
1989 <a href="#subpatternsassubroutines">"subroutine"</a>
1990 call, which is described in the next section.) The special item (?R) or (?0) is
1991 a recursive call of the entire regular expression.
1992 </P>
1993 <P>
1994 This PCRE pattern solves the nested parentheses problem (assume the
1995 PCRE_EXTENDED option is set so that white space is ignored):
1996 <pre>
1997 \( ( [^()]++ | (?R) )* \)
1998 </pre>
1999 First it matches an opening parenthesis. Then it matches any number of
2000 substrings which can either be a sequence of non-parentheses, or a recursive
2001 match of the pattern itself (that is, a correctly parenthesized substring).
2002 Finally there is a closing parenthesis. Note the use of a possessive quantifier
2003 to avoid backtracking into sequences of non-parentheses.
2004 </P>
2005 <P>
2006 If this were part of a larger pattern, you would not want to recurse the entire
2007 pattern, so instead you could use this:
2008 <pre>
2009 ( \( ( [^()]++ | (?1) )* \) )
2010 </pre>
2011 We have put the pattern into parentheses, and caused the recursion to refer to
2012 them instead of the whole pattern.
2013 </P>
2014 <P>
2015 In a larger pattern, keeping track of parenthesis numbers can be tricky. This
2016 is made easier by the use of relative references (a Perl 5.10 feature).
2017 Instead of (?1) in the pattern above you can write (?-2) to refer to the second
2018 most recently opened parentheses preceding the recursion. In other words, a
2019 negative number counts capturing parentheses leftwards from the point at which
2020 it is encountered.
2021 </P>
2022 <P>
2023 It is also possible to refer to subsequently opened parentheses, by writing
2024 references such as (?+2). However, these cannot be recursive because the
2025 reference is not inside the parentheses that are referenced. They are always
2026 <a href="#subpatternsassubroutines">"subroutine"</a>
2027 calls, as described in the next section.
2028 </P>
2029 <P>
2030 An alternative approach is to use named parentheses instead. The Perl syntax
2031 for this is (?&name); PCRE's earlier syntax (?P&#62;name) is also supported. We
2032 could rewrite the above example as follows:
2033 <pre>
2034 (?&#60;pn&#62; \( ( [^()]++ | (?&pn) )* \) )
2035 </pre>
2036 If there is more than one subpattern with the same name, the earliest one is
2037 used.
2038 </P>
2039 <P>
2040 This particular example pattern that we have been looking at contains nested
2041 unlimited repeats, and so the use of a possessive quantifier for matching
2042 strings of non-parentheses is important when applying the pattern to strings
2043 that do not match. For example, when this pattern is applied to
2044 <pre>
2045 (aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa()
2046 </pre>
2047 it yields "no match" quickly. However, if a possessive quantifier is not used,
2048 the match runs for a very long time indeed because there are so many different
2049 ways the + and * repeats can carve up the subject, and all have to be tested
2050 before failure can be reported.
2051 </P>
2052 <P>
2053 At the end of a match, the values of capturing parentheses are those from
2054 the outermost level. If you want to obtain intermediate values, a callout
2055 function can be used (see below and the
2056 <a href="pcrecallout.html"><b>pcrecallout</b></a>
2057 documentation). If the pattern above is matched against
2058 <pre>
2059 (ab(cd)ef)
2060 </pre>
2061 the value for the inner capturing parentheses (numbered 2) is "ef", which is
2062 the last value taken on at the top level. If a capturing subpattern is not
2063 matched at the top level, its final value is unset, even if it is (temporarily)
2064 set at a deeper level.
2065 </P>
2066 <P>
2067 If there are more than 15 capturing parentheses in a pattern, PCRE has to
2068 obtain extra memory to store data during a recursion, which it does by using
2069 <b>pcre_malloc</b>, freeing it via <b>pcre_free</b> afterwards. If no memory can
2070 be obtained, the match fails with the PCRE_ERROR_NOMEMORY error.
2071 </P>
2072 <P>
2073 Do not confuse the (?R) item with the condition (R), which tests for recursion.
2074 Consider this pattern, which matches text in angle brackets, allowing for
2075 arbitrary nesting. Only digits are allowed in nested brackets (that is, when
2076 recursing), whereas any characters are permitted at the outer level.
2077 <pre>
2078 &#60; (?: (?(R) \d++ | [^&#60;&#62;]*+) | (?R)) * &#62;
2079 </pre>
2080 In this pattern, (?(R) is the start of a conditional subpattern, with two
2081 different alternatives for the recursive and non-recursive cases. The (?R) item
2082 is the actual recursive call.
2083 <a name="recursiondifference"></a></P>
2084 <br><b>
2085 Recursion difference from Perl
2086 </b><br>
2087 <P>
2088 In PCRE (like Python, but unlike Perl), a recursive subpattern call is always
2089 treated as an atomic group. That is, once it has matched some of the subject
2090 string, it is never re-entered, even if it contains untried alternatives and
2091 there is a subsequent matching failure. This can be illustrated by the
2092 following pattern, which purports to match a palindromic string that contains
2093 an odd number of characters (for example, "a", "aba", "abcba", "abcdcba"):
2094 <pre>
2095 ^(.|(.)(?1)\2)$
2096 </pre>
2097 The idea is that it either matches a single character, or two identical
2098 characters surrounding a sub-palindrome. In Perl, this pattern works; in PCRE
2099 it does not if the pattern is longer than three characters. Consider the
2100 subject string "abcba":
2101 </P>
2102 <P>
2103 At the top level, the first character is matched, but as it is not at the end
2104 of the string, the first alternative fails; the second alternative is taken
2105 and the recursion kicks in. The recursive call to subpattern 1 successfully
2106 matches the next character ("b"). (Note that the beginning and end of line
2107 tests are not part of the recursion).
2108 </P>
2109 <P>
2110 Back at the top level, the next character ("c") is compared with what
2111 subpattern 2 matched, which was "a". This fails. Because the recursion is
2112 treated as an atomic group, there are now no backtracking points, and so the
2113 entire match fails. (Perl is able, at this point, to re-enter the recursion and
2114 try the second alternative.) However, if the pattern is written with the
2115 alternatives in the other order, things are different:
2116 <pre>
2117 ^((.)(?1)\2|.)$
2118 </pre>
2119 This time, the recursing alternative is tried first, and continues to recurse
2120 until it runs out of characters, at which point the recursion fails. But this
2121 time we do have another alternative to try at the higher level. That is the big
2122 difference: in the previous case the remaining alternative is at a deeper
2123 recursion level, which PCRE cannot use.
2124 </P>
2125 <P>
2126 To change the pattern so that matches all palindromic strings, not just those
2127 with an odd number of characters, it is tempting to change the pattern to this:
2128 <pre>
2129 ^((.)(?1)\2|.?)$
2130 </pre>
2131 Again, this works in Perl, but not in PCRE, and for the same reason. When a
2132 deeper recursion has matched a single character, it cannot be entered again in
2133 order to match an empty string. The solution is to separate the two cases, and
2134 write out the odd and even cases as alternatives at the higher level:
2135 <pre>
2136 ^(?:((.)(?1)\2|)|((.)(?3)\4|.))
2137 </pre>
2138 If you want to match typical palindromic phrases, the pattern has to ignore all
2139 non-word characters, which can be done like this:
2140 <pre>
2141 ^\W*+(?:((.)\W*+(?1)\W*+\2|)|((.)\W*+(?3)\W*+\4|\W*+.\W*+))\W*+$
2142 </pre>
2143 If run with the PCRE_CASELESS option, this pattern matches phrases such as "A
2144 man, a plan, a canal: Panama!" and it works well in both PCRE and Perl. Note
2145 the use of the possessive quantifier *+ to avoid backtracking into sequences of
2146 non-word characters. Without this, PCRE takes a great deal longer (ten times or
2147 more) to match typical phrases, and Perl takes so long that you think it has
2148 gone into a loop.
2149 </P>
2150 <P>
2151 <b>WARNING</b>: The palindrome-matching patterns above work only if the subject
2152 string does not start with a palindrome that is shorter than the entire string.
2153 For example, although "abcba" is correctly matched, if the subject is "ababa",
2154 PCRE finds the palindrome "aba" at the start, then fails at top level because
2155 the end of the string does not follow. Once again, it cannot jump back into the
2156 recursion to try other alternatives, so the entire match fails.
2157 <a name="subpatternsassubroutines"></a></P>
2158 <br><a name="SEC22" href="#TOC1">SUBPATTERNS AS SUBROUTINES</a><br>
2159 <P>
2160 If the syntax for a recursive subpattern reference (either by number or by
2161 name) is used outside the parentheses to which it refers, it operates like a
2162 subroutine in a programming language. The "called" subpattern may be defined
2163 before or after the reference. A numbered reference can be absolute or
2164 relative, as in these examples:
2165 <pre>
2166 (...(absolute)...)...(?2)...
2167 (...(relative)...)...(?-1)...
2168 (...(?+1)...(relative)...
2169 </pre>
2170 An earlier example pointed out that the pattern
2171 <pre>
2172 (sens|respons)e and \1ibility
2173 </pre>
2174 matches "sense and sensibility" and "response and responsibility", but not
2175 "sense and responsibility". If instead the pattern
2176 <pre>
2177 (sens|respons)e and (?1)ibility
2178 </pre>
2179 is used, it does match "sense and responsibility" as well as the other two
2180 strings. Another example is given in the discussion of DEFINE above.
2181 </P>
2182 <P>
2183 Like recursive subpatterns, a subroutine call is always treated as an atomic
2184 group. That is, once it has matched some of the subject string, it is never
2185 re-entered, even if it contains untried alternatives and there is a subsequent
2186 matching failure. Any capturing parentheses that are set during the subroutine
2187 call revert to their previous values afterwards.
2188 </P>
2189 <P>
2190 When a subpattern is used as a subroutine, processing options such as
2191 case-independence are fixed when the subpattern is defined. They cannot be
2192 changed for different calls. For example, consider this pattern:
2193 <pre>
2194 (abc)(?i:(?-1))
2195 </pre>
2196 It matches "abcabc". It does not match "abcABC" because the change of
2197 processing option does not affect the called subpattern.
2198 <a name="onigurumasubroutines"></a></P>
2199 <br><a name="SEC23" href="#TOC1">ONIGURUMA SUBROUTINE SYNTAX</a><br>
2200 <P>
2201 For compatibility with Oniguruma, the non-Perl syntax \g followed by a name or
2202 a number enclosed either in angle brackets or single quotes, is an alternative
2203 syntax for referencing a subpattern as a subroutine, possibly recursively. Here
2204 are two of the examples used above, rewritten using this syntax:
2205 <pre>
2206 (?&#60;pn&#62; \( ( (?&#62;[^()]+) | \g&#60;pn&#62; )* \) )
2207 (sens|respons)e and \g'1'ibility
2208 </pre>
2209 PCRE supports an extension to Oniguruma: if a number is preceded by a
2210 plus or a minus sign it is taken as a relative reference. For example:
2211 <pre>
2212 (abc)(?i:\g&#60;-1&#62;)
2213 </pre>
2214 Note that \g{...} (Perl syntax) and \g&#60;...&#62; (Oniguruma syntax) are <i>not</i>
2215 synonymous. The former is a back reference; the latter is a subroutine call.
2216 </P>
2217 <br><a name="SEC24" href="#TOC1">CALLOUTS</a><br>
2218 <P>
2219 Perl has a feature whereby using the sequence (?{...}) causes arbitrary Perl
2220 code to be obeyed in the middle of matching a regular expression. This makes it
2221 possible, amongst other things, to extract different substrings that match the
2222 same pair of parentheses when there is a repetition.
2223 </P>
2224 <P>
2225 PCRE provides a similar feature, but of course it cannot obey arbitrary Perl
2226 code. The feature is called "callout". The caller of PCRE provides an external
2227 function by putting its entry point in the global variable <i>pcre_callout</i>.
2228 By default, this variable contains NULL, which disables all calling out.
2229 </P>
2230 <P>
2231 Within a regular expression, (?C) indicates the points at which the external
2232 function is to be called. If you want to identify different callout points, you
2233 can put a number less than 256 after the letter C. The default value is zero.
2234 For example, this pattern has two callout points:
2235 <pre>
2236 (?C1)abc(?C2)def
2237 </pre>
2238 If the PCRE_AUTO_CALLOUT flag is passed to <b>pcre_compile()</b>, callouts are
2239 automatically installed before each item in the pattern. They are all numbered
2240 255.
2241 </P>
2242 <P>
2243 During matching, when PCRE reaches a callout point (and <i>pcre_callout</i> is
2244 set), the external function is called. It is provided with the number of the
2245 callout, the position in the pattern, and, optionally, one item of data
2246 originally supplied by the caller of <b>pcre_exec()</b>. The callout function
2247 may cause matching to proceed, to backtrack, or to fail altogether. A complete
2248 description of the interface to the callout function is given in the
2249 <a href="pcrecallout.html"><b>pcrecallout</b></a>
2250 documentation.
2251 </P>
2252 <br><a name="SEC25" href="#TOC1">BACKTRACKING CONTROL</a><br>
2253 <P>
2254 Perl 5.10 introduced a number of "Special Backtracking Control Verbs", which
2255 are described in the Perl documentation as "experimental and subject to change
2256 or removal in a future version of Perl". It goes on to say: "Their usage in
2257 production code should be noted to avoid problems during upgrades." The same
2258 remarks apply to the PCRE features described in this section.
2259 </P>
2260 <P>
2261 Since these verbs are specifically related to backtracking, most of them can be
2262 used only when the pattern is to be matched using <b>pcre_exec()</b>, which uses
2263 a backtracking algorithm. With the exception of (*FAIL), which behaves like a
2264 failing negative assertion, they cause an error if encountered by
2265 <b>pcre_dfa_exec()</b>.
2266 </P>
2267 <P>
2268 If any of these verbs are used in an assertion or subroutine subpattern
2269 (including recursive subpatterns), their effect is confined to that subpattern;
2270 it does not extend to the surrounding pattern. Note that such subpatterns are
2271 processed as anchored at the point where they are tested.
2272 </P>
2273 <P>
2274 The new verbs make use of what was previously invalid syntax: an opening
2275 parenthesis followed by an asterisk. In Perl, they are generally of the form
2276 (*VERB:ARG) but PCRE does not support the use of arguments, so its general
2277 form is just (*VERB). Any number of these verbs may occur in a pattern. There
2278 are two kinds:
2279 </P>
2280 <br><b>
2281 Verbs that act immediately
2282 </b><br>
2283 <P>
2284 The following verbs act as soon as they are encountered:
2285 <pre>
2286 (*ACCEPT)
2287 </pre>
2288 This verb causes the match to end successfully, skipping the remainder of the
2289 pattern. When inside a recursion, only the innermost pattern is ended
2290 immediately. If (*ACCEPT) is inside capturing parentheses, the data so far is
2291 captured. (This feature was added to PCRE at release 8.00.) For example:
2292 <pre>
2293 A((?:A|B(*ACCEPT)|C)D)
2294 </pre>
2295 This matches "AB", "AAD", or "ACD"; when it matches "AB", "B" is captured by
2296 the outer parentheses.
2297 <pre>
2298 (*FAIL) or (*F)
2299 </pre>
2300 This verb causes the match to fail, forcing backtracking to occur. It is
2301 equivalent to (?!) but easier to read. The Perl documentation notes that it is
2302 probably useful only when combined with (?{}) or (??{}). Those are, of course,
2303 Perl features that are not present in PCRE. The nearest equivalent is the
2304 callout feature, as for example in this pattern:
2305 <pre>
2306 a+(?C)(*FAIL)
2307 </pre>
2308 A match with the string "aaaa" always fails, but the callout is taken before
2309 each backtrack happens (in this example, 10 times).
2310 </P>
2311 <br><b>
2312 Verbs that act after backtracking
2313 </b><br>
2314 <P>
2315 The following verbs do nothing when they are encountered. Matching continues
2316 with what follows, but if there is no subsequent match, a failure is forced.
2317 The verbs differ in exactly what kind of failure occurs.
2318 <pre>
2319 (*COMMIT)
2320 </pre>
2321 This verb causes the whole match to fail outright if the rest of the pattern
2322 does not match. Even if the pattern is unanchored, no further attempts to find
2323 a match by advancing the starting point take place. Once (*COMMIT) has been
2324 passed, <b>pcre_exec()</b> is committed to finding a match at the current
2325 starting point, or not at all. For example:
2326 <pre>
2327 a+(*COMMIT)b
2328 </pre>
2329 This matches "xxaab" but not "aacaab". It can be thought of as a kind of
2330 dynamic anchor, or "I've started, so I must finish."
2331 <pre>
2332 (*PRUNE)
2333 </pre>
2334 This verb causes the match to fail at the current position if the rest of the
2335 pattern does not match. If the pattern is unanchored, the normal "bumpalong"
2336 advance to the next starting character then happens. Backtracking can occur as
2337 usual to the left of (*PRUNE), or when matching to the right of (*PRUNE), but
2338 if there is no match to the right, backtracking cannot cross (*PRUNE).
2339 In simple cases, the use of (*PRUNE) is just an alternative to an atomic
2340 group or possessive quantifier, but there are some uses of (*PRUNE) that cannot
2341 be expressed in any other way.
2342 <pre>
2343 (*SKIP)
2344 </pre>
2345 This verb is like (*PRUNE), except that if the pattern is unanchored, the
2346 "bumpalong" advance is not to the next character, but to the position in the
2347 subject where (*SKIP) was encountered. (*SKIP) signifies that whatever text
2348 was matched leading up to it cannot be part of a successful match. Consider:
2349 <pre>
2350 a+(*SKIP)b
2351 </pre>
2352 If the subject is "aaaac...", after the first match attempt fails (starting at
2353 the first character in the string), the starting point skips on to start the
2354 next attempt at "c". Note that a possessive quantifer does not have the same
2355 effect as this example; although it would suppress backtracking during the
2356 first match attempt, the second attempt would start at the second character
2357 instead of skipping on to "c".
2358 <pre>
2359 (*THEN)
2360 </pre>
2361 This verb causes a skip to the next alternation if the rest of the pattern does
2362 not match. That is, it cancels pending backtracking, but only within the
2363 current alternation. Its name comes from the observation that it can be used
2364 for a pattern-based if-then-else block:
2365 <pre>
2366 ( COND1 (*THEN) FOO | COND2 (*THEN) BAR | COND3 (*THEN) BAZ ) ...
2367 </pre>
2368 If the COND1 pattern matches, FOO is tried (and possibly further items after
2369 the end of the group if FOO succeeds); on failure the matcher skips to the
2370 second alternative and tries COND2, without backtracking into COND1. If (*THEN)
2371 is used outside of any alternation, it acts exactly like (*PRUNE).
2372 </P>
2373 <br><a name="SEC26" href="#TOC1">SEE ALSO</a><br>
2374 <P>
2375 <b>pcreapi</b>(3), <b>pcrecallout</b>(3), <b>pcrematching</b>(3),
2376 <b>pcresyntax</b>(3), <b>pcre</b>(3).
2377 </P>
2378 <br><a name="SEC27" href="#TOC1">AUTHOR</a><br>
2379 <P>
2380 Philip Hazel
2381 <br>
2382 University Computing Service
2383 <br>
2384 Cambridge CB2 3QH, England.
2385 <br>
2386 </P>
2387 <br><a name="SEC28" href="#TOC1">REVISION</a><br>
2388 <P>
2389 Last updated: 18 October 2009
2390 <br>
2391 Copyright &copy; 1997-2009 University of Cambridge.
2392 <br>
2393 <p>
2394 Return to the <a href="index.html">PCRE index page</a>.
2395 </p>


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

  ViewVC Help
Powered by ViewVC 1.1.5