ViewVC logotype

Contents of /code/branches/pcre16/doc/html/pcrepattern.html

Parent Directory Parent Directory | Revision Log Revision Log

Revision 755 - (show annotations)
Mon Nov 21 10:41:54 2011 UTC (9 years, 5 months ago) by ph10
File MIME type: text/html
File size: 122927 byte(s)
Created a new branch for the development of 16-bit support.
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) AND \N</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="pcreunicode.html"><b>pcreunicode</b></a>
76 page.
77 </P>
78 <P>
79 Another special sequence that may appear at the start of a pattern or in
80 combination with (*UTF8) is:
81 <pre>
82 (*UCP)
83 </pre>
84 This has the same effect as setting the PCRE_UCP option: it causes sequences
85 such as \d and \w to use Unicode properties to determine character types,
86 instead of recognizing only characters with codes less than 128 via a lookup
87 table.
88 </P>
89 <P>
90 If a pattern starts with (*NO_START_OPT), it has the same effect as setting the
91 PCRE_NO_START_OPTIMIZE option either at compile or matching time. There are
92 also some more of these special sequences that are concerned with the handling
93 of newlines; they are described below.
94 </P>
95 <P>
96 The remainder of this document discusses the patterns that are supported by
97 PCRE when its main matching function, <b>pcre_exec()</b>, is used.
98 From release 6.0, PCRE offers a second matching function,
99 <b>pcre_dfa_exec()</b>, which matches using a different algorithm that is not
100 Perl-compatible. Some of the features discussed below are not available when
101 <b>pcre_dfa_exec()</b> is used. The advantages and disadvantages of the
102 alternative function, and how it differs from the normal function, are
103 discussed in the
104 <a href="pcrematching.html"><b>pcrematching</b></a>
105 page.
106 <a name="newlines"></a></P>
107 <br><a name="SEC2" href="#TOC1">NEWLINE CONVENTIONS</a><br>
108 <P>
109 PCRE supports five different conventions for indicating line breaks in
110 strings: a single CR (carriage return) character, a single LF (linefeed)
111 character, the two-character sequence CRLF, any of the three preceding, or any
112 Unicode newline sequence. The
113 <a href="pcreapi.html"><b>pcreapi</b></a>
114 page has
115 <a href="pcreapi.html#newlines">further discussion</a>
116 about newlines, and shows how to set the newline convention in the
117 <i>options</i> arguments for the compiling and matching functions.
118 </P>
119 <P>
120 It is also possible to specify a newline convention by starting a pattern
121 string with one of the following five sequences:
122 <pre>
123 (*CR) carriage return
124 (*LF) linefeed
125 (*CRLF) carriage return, followed by linefeed
126 (*ANYCRLF) any of the three above
127 (*ANY) all Unicode newline sequences
128 </pre>
129 These override the default and the options given to <b>pcre_compile()</b> or
130 <b>pcre_compile2()</b>. For example, on a Unix system where LF is the default
131 newline sequence, the pattern
132 <pre>
133 (*CR)a.b
134 </pre>
135 changes the convention to CR. That pattern matches "a\nb" because LF is no
136 longer a newline. Note that these special settings, which are not
137 Perl-compatible, are recognized only at the very start of a pattern, and that
138 they must be in upper case. If more than one of them is present, the last one
139 is used.
140 </P>
141 <P>
142 The newline convention affects the interpretation of the dot metacharacter when
143 PCRE_DOTALL is not set, and also the behaviour of \N. However, it does not
144 affect what the \R escape sequence matches. By default, this is any Unicode
145 newline sequence, for Perl compatibility. However, this can be changed; see the
146 description of \R in the section entitled
147 <a href="#newlineseq">"Newline sequences"</a>
148 below. A change of \R setting can be combined with a change of newline
149 convention.
150 </P>
151 <br><a name="SEC3" href="#TOC1">CHARACTERS AND METACHARACTERS</a><br>
152 <P>
153 A regular expression is a pattern that is matched against a subject string from
154 left to right. Most characters stand for themselves in a pattern, and match the
155 corresponding characters in the subject. As a trivial example, the pattern
156 <pre>
157 The quick brown fox
158 </pre>
159 matches a portion of a subject string that is identical to itself. When
160 caseless matching is specified (the PCRE_CASELESS option), letters are matched
161 independently of case. In UTF-8 mode, PCRE always understands the concept of
162 case for characters whose values are less than 128, so caseless matching is
163 always possible. For characters with higher values, the concept of case is
164 supported if PCRE is compiled with Unicode property support, but not otherwise.
165 If you want to use caseless matching for characters 128 and above, you must
166 ensure that PCRE is compiled with Unicode property support as well as with
167 UTF-8 support.
168 </P>
169 <P>
170 The power of regular expressions comes from the ability to include alternatives
171 and repetitions in the pattern. These are encoded in the pattern by the use of
172 <i>metacharacters</i>, which do not stand for themselves but instead are
173 interpreted in some special way.
174 </P>
175 <P>
176 There are two different sets of metacharacters: those that are recognized
177 anywhere in the pattern except within square brackets, and those that are
178 recognized within square brackets. Outside square brackets, the metacharacters
179 are as follows:
180 <pre>
181 \ general escape character with several uses
182 ^ assert start of string (or line, in multiline mode)
183 $ assert end of string (or line, in multiline mode)
184 . match any character except newline (by default)
185 [ start character class definition
186 | start of alternative branch
187 ( start subpattern
188 ) end subpattern
189 ? extends the meaning of (
190 also 0 or 1 quantifier
191 also quantifier minimizer
192 * 0 or more quantifier
193 + 1 or more quantifier
194 also "possessive quantifier"
195 { start min/max quantifier
196 </pre>
197 Part of a pattern that is in square brackets is called a "character class". In
198 a character class the only metacharacters are:
199 <pre>
200 \ general escape character
201 ^ negate the class, but only if the first character
202 - indicates character range
203 [ POSIX character class (only if followed by POSIX syntax)
204 ] terminates the character class
205 </pre>
206 The following sections describe the use of each of the metacharacters.
207 </P>
208 <br><a name="SEC4" href="#TOC1">BACKSLASH</a><br>
209 <P>
210 The backslash character has several uses. Firstly, if it is followed by a
211 character that is not a number or a letter, it takes away any special meaning
212 that character may have. This use of backslash as an escape character applies
213 both inside and outside character classes.
214 </P>
215 <P>
216 For example, if you want to match a * character, you write \* in the pattern.
217 This escaping action applies whether or not the following character would
218 otherwise be interpreted as a metacharacter, so it is always safe to precede a
219 non-alphanumeric with backslash to specify that it stands for itself. In
220 particular, if you want to match a backslash, you write \\.
221 </P>
222 <P>
223 In UTF-8 mode, only ASCII numbers and letters have any special meaning after a
224 backslash. All other characters (in particular, those whose codepoints are
225 greater than 127) are treated as literals.
226 </P>
227 <P>
228 If a pattern is compiled with the PCRE_EXTENDED option, whitespace in the
229 pattern (other than in a character class) and characters between a # outside
230 a character class and the next newline are ignored. An escaping backslash can
231 be used to include a whitespace or # character as part of the pattern.
232 </P>
233 <P>
234 If you want to remove the special meaning from a sequence of characters, you
235 can do so by putting them between \Q and \E. This is different from Perl in
236 that $ and @ are handled as literals in \Q...\E sequences in PCRE, whereas in
237 Perl, $ and @ cause variable interpolation. Note the following examples:
238 <pre>
239 Pattern PCRE matches Perl matches
241 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
242 \Qabc\$xyz\E abc\$xyz abc\$xyz
243 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
244 </pre>
245 The \Q...\E sequence is recognized both inside and outside character classes.
246 An isolated \E that is not preceded by \Q is ignored. If \Q is not followed
247 by \E later in the pattern, the literal interpretation continues to the end of
248 the pattern (that is, \E is assumed at the end). If the isolated \Q is inside
249 a character class, this causes an error, because the character class is not
250 terminated.
251 <a name="digitsafterbackslash"></a></P>
252 <br><b>
253 Non-printing characters
254 </b><br>
255 <P>
256 A second use of backslash provides a way of encoding non-printing characters
257 in patterns in a visible manner. There is no restriction on the appearance of
258 non-printing characters, apart from the binary zero that terminates a pattern,
259 but when a pattern is being prepared by text editing, it is often easier to use
260 one of the following escape sequences than the binary character it represents:
261 <pre>
262 \a alarm, that is, the BEL character (hex 07)
263 \cx "control-x", where x is any ASCII character
264 \e escape (hex 1B)
265 \f formfeed (hex 0C)
266 \n linefeed (hex 0A)
267 \r carriage return (hex 0D)
268 \t tab (hex 09)
269 \ddd character with octal code ddd, or back reference
270 \xhh character with hex code hh
271 \x{hhh..} character with hex code hhh..
272 </pre>
273 The precise effect of \cx is as follows: if x is a lower case letter, it
274 is converted to upper case. Then bit 6 of the character (hex 40) is inverted.
275 Thus \cz becomes hex 1A (z is 7A), but \c{ becomes hex 3B ({ is 7B), while
276 \c; becomes hex 7B (; is 3B). If the byte following \c has a value greater
277 than 127, a compile-time error occurs. This locks out non-ASCII characters in
278 both byte mode and UTF-8 mode. (When PCRE is compiled in EBCDIC mode, all byte
279 values are valid. A lower case letter is converted to upper case, and then the
280 0xc0 bits are flipped.)
281 </P>
282 <P>
283 After \x, from zero to two hexadecimal digits are read (letters can be in
284 upper or lower case). Any number of hexadecimal digits may appear between \x{
285 and }, but the value of the character code must be less than 256 in non-UTF-8
286 mode, and less than 2**31 in UTF-8 mode. That is, the maximum value in
287 hexadecimal is 7FFFFFFF. Note that this is bigger than the largest Unicode code
288 point, which is 10FFFF.
289 </P>
290 <P>
291 If characters other than hexadecimal digits appear between \x{ and }, or if
292 there is no terminating }, this form of escape is not recognized. Instead, the
293 initial \x will be interpreted as a basic hexadecimal escape, with no
294 following digits, giving a character whose value is zero.
295 </P>
296 <P>
297 Characters whose value is less than 256 can be defined by either of the two
298 syntaxes for \x. There is no difference in the way they are handled. For
299 example, \xdc is exactly the same as \x{dc}.
300 </P>
301 <P>
302 After \0 up to two further octal digits are read. If there are fewer than two
303 digits, just those that are present are used. Thus the sequence \0\x\07
304 specifies two binary zeros followed by a BEL character (code value 7). Make
305 sure you supply two digits after the initial zero if the pattern character that
306 follows is itself an octal digit.
307 </P>
308 <P>
309 The handling of a backslash followed by a digit other than 0 is complicated.
310 Outside a character class, PCRE reads it and any following digits as a decimal
311 number. If the number is less than 10, or if there have been at least that many
312 previous capturing left parentheses in the expression, the entire sequence is
313 taken as a <i>back reference</i>. A description of how this works is given
314 <a href="#backreferences">later,</a>
315 following the discussion of
316 <a href="#subpattern">parenthesized subpatterns.</a>
317 </P>
318 <P>
319 Inside a character class, or if the decimal number is greater than 9 and there
320 have not been that many capturing subpatterns, PCRE re-reads up to three octal
321 digits following the backslash, and uses them to generate a data character. Any
322 subsequent digits stand for themselves. In non-UTF-8 mode, the value of a
323 character specified in octal must be less than \400. In UTF-8 mode, values up
324 to \777 are permitted. For example:
325 <pre>
326 \040 is another way of writing a space
327 \40 is the same, provided there are fewer than 40 previous capturing subpatterns
328 \7 is always a back reference
329 \11 might be a back reference, or another way of writing a tab
330 \011 is always a tab
331 \0113 is a tab followed by the character "3"
332 \113 might be a back reference, otherwise the character with octal code 113
333 \377 might be a back reference, otherwise the byte consisting entirely of 1 bits
334 \81 is either a back reference, or a binary zero followed by the two characters "8" and "1"
335 </pre>
336 Note that octal values of 100 or greater must not be introduced by a leading
337 zero, because no more than three octal digits are ever read.
338 </P>
339 <P>
340 All the sequences that define a single character value can be used both inside
341 and outside character classes. In addition, inside a character class, the
342 sequence \b is interpreted as the backspace character (hex 08). The sequences
343 \B, \N, \R, and \X are not special inside a character class. Like any other
344 unrecognized escape sequences, they are treated as the literal characters "B",
345 "N", "R", and "X" by default, but cause an error if the PCRE_EXTRA option is
346 set. Outside a character class, these sequences have different meanings.
347 </P>
348 <br><b>
349 Absolute and relative back references
350 </b><br>
351 <P>
352 The sequence \g followed by an unsigned or a negative number, optionally
353 enclosed in braces, is an absolute or relative back reference. A named back
354 reference can be coded as \g{name}. Back references are discussed
355 <a href="#backreferences">later,</a>
356 following the discussion of
357 <a href="#subpattern">parenthesized subpatterns.</a>
358 </P>
359 <br><b>
360 Absolute and relative subroutine calls
361 </b><br>
362 <P>
363 For compatibility with Oniguruma, the non-Perl syntax \g followed by a name or
364 a number enclosed either in angle brackets or single quotes, is an alternative
365 syntax for referencing a subpattern as a "subroutine". Details are discussed
366 <a href="#onigurumasubroutines">later.</a>
367 Note that \g{...} (Perl syntax) and \g&#60;...&#62; (Oniguruma syntax) are <i>not</i>
368 synonymous. The former is a back reference; the latter is a
369 <a href="#subpatternsassubroutines">subroutine</a>
370 call.
371 <a name="genericchartypes"></a></P>
372 <br><b>
373 Generic character types
374 </b><br>
375 <P>
376 Another use of backslash is for specifying generic character types:
377 <pre>
378 \d any decimal digit
379 \D any character that is not a decimal digit
380 \h any horizontal whitespace character
381 \H any character that is not a horizontal whitespace character
382 \s any whitespace character
383 \S any character that is not a whitespace character
384 \v any vertical whitespace character
385 \V any character that is not a vertical whitespace character
386 \w any "word" character
387 \W any "non-word" character
388 </pre>
389 There is also the single sequence \N, which matches a non-newline character.
390 This is the same as
391 <a href="#fullstopdot">the "." metacharacter</a>
392 when PCRE_DOTALL is not set.
393 </P>
394 <P>
395 Each pair of lower and upper case escape sequences partitions the complete set
396 of characters into two disjoint sets. Any given character matches one, and only
397 one, of each pair. The sequences can appear both inside and outside character
398 classes. They each match one character of the appropriate type. If the current
399 matching point is at the end of the subject string, all of them fail, because
400 there is no character to match.
401 </P>
402 <P>
403 For compatibility with Perl, \s does not match the VT character (code 11).
404 This makes it different from the the POSIX "space" class. The \s characters
405 are HT (9), LF (10), FF (12), CR (13), and space (32). If "use locale;" is
406 included in a Perl script, \s may match the VT character. In PCRE, it never
407 does.
408 </P>
409 <P>
410 A "word" character is an underscore or any character that is a letter or digit.
411 By default, the definition of letters and digits is controlled by PCRE's
412 low-valued character tables, and may vary if locale-specific matching is taking
413 place (see
414 <a href="pcreapi.html#localesupport">"Locale support"</a>
415 in the
416 <a href="pcreapi.html"><b>pcreapi</b></a>
417 page). For example, in a French locale such as "fr_FR" in Unix-like systems,
418 or "french" in Windows, some character codes greater than 128 are used for
419 accented letters, and these are then matched by \w. The use of locales with
420 Unicode is discouraged.
421 </P>
422 <P>
423 By default, in UTF-8 mode, characters with values greater than 128 never match
424 \d, \s, or \w, and always match \D, \S, and \W. These sequences retain
425 their original meanings from before UTF-8 support was available, mainly for
426 efficiency reasons. However, if PCRE is compiled with Unicode property support,
427 and the PCRE_UCP option is set, the behaviour is changed so that Unicode
428 properties are used to determine character types, as follows:
429 <pre>
430 \d any character that \p{Nd} matches (decimal digit)
431 \s any character that \p{Z} matches, plus HT, LF, FF, CR
432 \w any character that \p{L} or \p{N} matches, plus underscore
433 </pre>
434 The upper case escapes match the inverse sets of characters. Note that \d
435 matches only decimal digits, whereas \w matches any Unicode digit, as well as
436 any Unicode letter, and underscore. Note also that PCRE_UCP affects \b, and
437 \B because they are defined in terms of \w and \W. Matching these sequences
438 is noticeably slower when PCRE_UCP is set.
439 </P>
440 <P>
441 The sequences \h, \H, \v, and \V are features that were added to Perl at
442 release 5.10. In contrast to the other sequences, which match only ASCII
443 characters by default, these always match certain high-valued codepoints in
444 UTF-8 mode, whether or not PCRE_UCP is set. The horizontal space characters
445 are:
446 <pre>
447 U+0009 Horizontal tab
448 U+0020 Space
449 U+00A0 Non-break space
450 U+1680 Ogham space mark
451 U+180E Mongolian vowel separator
452 U+2000 En quad
453 U+2001 Em quad
454 U+2002 En space
455 U+2003 Em space
456 U+2004 Three-per-em space
457 U+2005 Four-per-em space
458 U+2006 Six-per-em space
459 U+2007 Figure space
460 U+2008 Punctuation space
461 U+2009 Thin space
462 U+200A Hair space
463 U+202F Narrow no-break space
464 U+205F Medium mathematical space
465 U+3000 Ideographic space
466 </pre>
467 The vertical space characters are:
468 <pre>
469 U+000A Linefeed
470 U+000B Vertical tab
471 U+000C Formfeed
472 U+000D Carriage return
473 U+0085 Next line
474 U+2028 Line separator
475 U+2029 Paragraph separator
476 <a name="newlineseq"></a></PRE>
477 </P>
478 <br><b>
479 Newline sequences
480 </b><br>
481 <P>
482 Outside a character class, by default, the escape sequence \R matches any
483 Unicode newline sequence. In non-UTF-8 mode \R is equivalent to the following:
484 <pre>
485 (?&#62;\r\n|\n|\x0b|\f|\r|\x85)
486 </pre>
487 This is an example of an "atomic group", details of which are given
488 <a href="#atomicgroup">below.</a>
489 This particular group matches either the two-character sequence CR followed by
490 LF, or one of the single characters LF (linefeed, U+000A), VT (vertical tab,
491 U+000B), FF (formfeed, U+000C), CR (carriage return, U+000D), or NEL (next
492 line, U+0085). The two-character sequence is treated as a single unit that
493 cannot be split.
494 </P>
495 <P>
496 In UTF-8 mode, two additional characters whose codepoints are greater than 255
497 are added: LS (line separator, U+2028) and PS (paragraph separator, U+2029).
498 Unicode character property support is not needed for these characters to be
499 recognized.
500 </P>
501 <P>
502 It is possible to restrict \R to match only CR, LF, or CRLF (instead of the
503 complete set of Unicode line endings) by setting the option PCRE_BSR_ANYCRLF
504 either at compile time or when the pattern is matched. (BSR is an abbrevation
505 for "backslash R".) This can be made the default when PCRE is built; if this is
506 the case, the other behaviour can be requested via the PCRE_BSR_UNICODE option.
507 It is also possible to specify these settings by starting a pattern string with
508 one of the following sequences:
509 <pre>
510 (*BSR_ANYCRLF) CR, LF, or CRLF only
511 (*BSR_UNICODE) any Unicode newline sequence
512 </pre>
513 These override the default and the options given to <b>pcre_compile()</b> or
514 <b>pcre_compile2()</b>, but they can be overridden by options given to
515 <b>pcre_exec()</b> or <b>pcre_dfa_exec()</b>. Note that these special settings,
516 which are not Perl-compatible, are recognized only at the very start of a
517 pattern, and that they must be in upper case. If more than one of them is
518 present, the last one is used. They can be combined with a change of newline
519 convention; for example, a pattern can start with:
520 <pre>
522 </pre>
523 They can also be combined with the (*UTF8) or (*UCP) special sequences. Inside
524 a character class, \R is treated as an unrecognized escape sequence, and so
525 matches the letter "R" by default, but causes an error if PCRE_EXTRA is set.
526 <a name="uniextseq"></a></P>
527 <br><b>
528 Unicode character properties
529 </b><br>
530 <P>
531 When PCRE is built with Unicode character property support, three additional
532 escape sequences that match characters with specific properties are available.
533 When not in UTF-8 mode, these sequences are of course limited to testing
534 characters whose codepoints are less than 256, but they do work in this mode.
535 The extra escape sequences are:
536 <pre>
537 \p{<i>xx</i>} a character with the <i>xx</i> property
538 \P{<i>xx</i>} a character without the <i>xx</i> property
539 \X an extended Unicode sequence
540 </pre>
541 The property names represented by <i>xx</i> above are limited to the Unicode
542 script names, the general category properties, "Any", which matches any
543 character (including newline), and some special PCRE properties (described
544 in the
545 <a href="#extraprops">next section).</a>
546 Other Perl properties such as "InMusicalSymbols" are not currently supported by
547 PCRE. Note that \P{Any} does not match any characters, so always causes a
548 match failure.
549 </P>
550 <P>
551 Sets of Unicode characters are defined as belonging to certain scripts. A
552 character from one of these sets can be matched using a script name. For
553 example:
554 <pre>
555 \p{Greek}
556 \P{Han}
557 </pre>
558 Those that are not part of an identified script are lumped together as
559 "Common". The current list of scripts is:
560 </P>
561 <P>
562 Arabic,
563 Armenian,
564 Avestan,
565 Balinese,
566 Bamum,
567 Bengali,
568 Bopomofo,
569 Braille,
570 Buginese,
571 Buhid,
572 Canadian_Aboriginal,
573 Carian,
574 Cham,
575 Cherokee,
576 Common,
577 Coptic,
578 Cuneiform,
579 Cypriot,
580 Cyrillic,
581 Deseret,
582 Devanagari,
583 Egyptian_Hieroglyphs,
584 Ethiopic,
585 Georgian,
586 Glagolitic,
587 Gothic,
588 Greek,
589 Gujarati,
590 Gurmukhi,
591 Han,
592 Hangul,
593 Hanunoo,
594 Hebrew,
595 Hiragana,
596 Imperial_Aramaic,
597 Inherited,
598 Inscriptional_Pahlavi,
599 Inscriptional_Parthian,
600 Javanese,
601 Kaithi,
602 Kannada,
603 Katakana,
604 Kayah_Li,
605 Kharoshthi,
606 Khmer,
607 Lao,
608 Latin,
609 Lepcha,
610 Limbu,
611 Linear_B,
612 Lisu,
613 Lycian,
614 Lydian,
615 Malayalam,
616 Meetei_Mayek,
617 Mongolian,
618 Myanmar,
619 New_Tai_Lue,
620 Nko,
621 Ogham,
622 Old_Italic,
623 Old_Persian,
624 Old_South_Arabian,
625 Old_Turkic,
626 Ol_Chiki,
627 Oriya,
628 Osmanya,
629 Phags_Pa,
630 Phoenician,
631 Rejang,
632 Runic,
633 Samaritan,
634 Saurashtra,
635 Shavian,
636 Sinhala,
637 Sundanese,
638 Syloti_Nagri,
639 Syriac,
640 Tagalog,
641 Tagbanwa,
642 Tai_Le,
643 Tai_Tham,
644 Tai_Viet,
645 Tamil,
646 Telugu,
647 Thaana,
648 Thai,
649 Tibetan,
650 Tifinagh,
651 Ugaritic,
652 Vai,
653 Yi.
654 </P>
655 <P>
656 Each character has exactly one Unicode general category property, specified by
657 a two-letter abbreviation. For compatibility with Perl, negation can be
658 specified by including a circumflex between the opening brace and the property
659 name. For example, \p{^Lu} is the same as \P{Lu}.
660 </P>
661 <P>
662 If only one letter is specified with \p or \P, it includes all the general
663 category properties that start with that letter. In this case, in the absence
664 of negation, the curly brackets in the escape sequence are optional; these two
665 examples have the same effect:
666 <pre>
667 \p{L}
668 \pL
669 </pre>
670 The following general category property codes are supported:
671 <pre>
672 C Other
673 Cc Control
674 Cf Format
675 Cn Unassigned
676 Co Private use
677 Cs Surrogate
679 L Letter
680 Ll Lower case letter
681 Lm Modifier letter
682 Lo Other letter
683 Lt Title case letter
684 Lu Upper case letter
686 M Mark
687 Mc Spacing mark
688 Me Enclosing mark
689 Mn Non-spacing mark
691 N Number
692 Nd Decimal number
693 Nl Letter number
694 No Other number
696 P Punctuation
697 Pc Connector punctuation
698 Pd Dash punctuation
699 Pe Close punctuation
700 Pf Final punctuation
701 Pi Initial punctuation
702 Po Other punctuation
703 Ps Open punctuation
705 S Symbol
706 Sc Currency symbol
707 Sk Modifier symbol
708 Sm Mathematical symbol
709 So Other symbol
711 Z Separator
712 Zl Line separator
713 Zp Paragraph separator
714 Zs Space separator
715 </pre>
716 The special property L& is also supported: it matches a character that has
717 the Lu, Ll, or Lt property, in other words, a letter that is not classified as
718 a modifier or "other".
719 </P>
720 <P>
721 The Cs (Surrogate) property applies only to characters in the range U+D800 to
722 U+DFFF. Such characters are not valid in UTF-8 strings (see RFC 3629) and so
723 cannot be tested by PCRE, unless UTF-8 validity checking has been turned off
724 (see the discussion of PCRE_NO_UTF8_CHECK in the
725 <a href="pcreapi.html"><b>pcreapi</b></a>
726 page). Perl does not support the Cs property.
727 </P>
728 <P>
729 The long synonyms for property names that Perl supports (such as \p{Letter})
730 are not supported by PCRE, nor is it permitted to prefix any of these
731 properties with "Is".
732 </P>
733 <P>
734 No character that is in the Unicode table has the Cn (unassigned) property.
735 Instead, this property is assumed for any code point that is not in the
736 Unicode table.
737 </P>
738 <P>
739 Specifying caseless matching does not affect these escape sequences. For
740 example, \p{Lu} always matches only upper case letters.
741 </P>
742 <P>
743 The \X escape matches any number of Unicode characters that form an extended
744 Unicode sequence. \X is equivalent to
745 <pre>
746 (?&#62;\PM\pM*)
747 </pre>
748 That is, it matches a character without the "mark" property, followed by zero
749 or more characters with the "mark" property, and treats the sequence as an
750 atomic group
751 <a href="#atomicgroup">(see below).</a>
752 Characters with the "mark" property are typically accents that affect the
753 preceding character. None of them have codepoints less than 256, so in
754 non-UTF-8 mode \X matches any one character.
755 </P>
756 <P>
757 Note that recent versions of Perl have changed \X to match what Unicode calls
758 an "extended grapheme cluster", which has a more complicated definition.
759 </P>
760 <P>
761 Matching characters by Unicode property is not fast, because PCRE has to search
762 a structure that contains data for over fifteen thousand characters. That is
763 why the traditional escape sequences such as \d and \w do not use Unicode
764 properties in PCRE by default, though you can make them do so by setting the
765 PCRE_UCP option for <b>pcre_compile()</b> or by starting the pattern with
766 (*UCP).
767 <a name="extraprops"></a></P>
768 <br><b>
769 PCRE's additional properties
770 </b><br>
771 <P>
772 As well as the standard Unicode properties described in the previous
773 section, PCRE supports four more that make it possible to convert traditional
774 escape sequences such as \w and \s and POSIX character classes to use Unicode
775 properties. PCRE uses these non-standard, non-Perl properties internally when
776 PCRE_UCP is set. They are:
777 <pre>
778 Xan Any alphanumeric character
779 Xps Any POSIX space character
780 Xsp Any Perl space character
781 Xwd Any Perl "word" character
782 </pre>
783 Xan matches characters that have either the L (letter) or the N (number)
784 property. Xps matches the characters tab, linefeed, vertical tab, formfeed, or
785 carriage return, and any other character that has the Z (separator) property.
786 Xsp is the same as Xps, except that vertical tab is excluded. Xwd matches the
787 same characters as Xan, plus underscore.
788 <a name="resetmatchstart"></a></P>
789 <br><b>
790 Resetting the match start
791 </b><br>
792 <P>
793 The escape sequence \K causes any previously matched characters not to be
794 included in the final matched sequence. For example, the pattern:
795 <pre>
796 foo\Kbar
797 </pre>
798 matches "foobar", but reports that it has matched "bar". This feature is
799 similar to a lookbehind assertion
800 <a href="#lookbehind">(described below).</a>
801 However, in this case, the part of the subject before the real match does not
802 have to be of fixed length, as lookbehind assertions do. The use of \K does
803 not interfere with the setting of
804 <a href="#subpattern">captured substrings.</a>
805 For example, when the pattern
806 <pre>
807 (foo)\Kbar
808 </pre>
809 matches "foobar", the first substring is still set to "foo".
810 </P>
811 <P>
812 Perl documents that the use of \K within assertions is "not well defined". In
813 PCRE, \K is acted upon when it occurs inside positive assertions, but is
814 ignored in negative assertions.
815 <a name="smallassertions"></a></P>
816 <br><b>
817 Simple assertions
818 </b><br>
819 <P>
820 The final use of backslash is for certain simple assertions. An assertion
821 specifies a condition that has to be met at a particular point in a match,
822 without consuming any characters from the subject string. The use of
823 subpatterns for more complicated assertions is described
824 <a href="#bigassertions">below.</a>
825 The backslashed assertions are:
826 <pre>
827 \b matches at a word boundary
828 \B matches when not at a word boundary
829 \A matches at the start of the subject
830 \Z matches at the end of the subject
831 also matches before a newline at the end of the subject
832 \z matches only at the end of the subject
833 \G matches at the first matching position in the subject
834 </pre>
835 Inside a character class, \b has a different meaning; it matches the backspace
836 character. If any other of these assertions appears in a character class, by
837 default it matches the corresponding literal character (for example, \B
838 matches the letter B). However, if the PCRE_EXTRA option is set, an "invalid
839 escape sequence" error is generated instead.
840 </P>
841 <P>
842 A word boundary is a position in the subject string where the current character
843 and the previous character do not both match \w or \W (i.e. one matches
844 \w and the other matches \W), or the start or end of the string if the
845 first or last character matches \w, respectively. In UTF-8 mode, the meanings
846 of \w and \W can be changed by setting the PCRE_UCP option. When this is
847 done, it also affects \b and \B. Neither PCRE nor Perl has a separate "start
848 of word" or "end of word" metasequence. However, whatever follows \b normally
849 determines which it is. For example, the fragment \ba matches "a" at the start
850 of a word.
851 </P>
852 <P>
853 The \A, \Z, and \z assertions differ from the traditional circumflex and
854 dollar (described in the next section) in that they only ever match at the very
855 start and end of the subject string, whatever options are set. Thus, they are
856 independent of multiline mode. These three assertions are not affected by the
857 PCRE_NOTBOL or PCRE_NOTEOL options, which affect only the behaviour of the
858 circumflex and dollar metacharacters. However, if the <i>startoffset</i>
859 argument of <b>pcre_exec()</b> is non-zero, indicating that matching is to start
860 at a point other than the beginning of the subject, \A can never match. The
861 difference between \Z and \z is that \Z matches before a newline at the end
862 of the string as well as at the very end, whereas \z matches only at the end.
863 </P>
864 <P>
865 The \G assertion is true only when the current matching position is at the
866 start point of the match, as specified by the <i>startoffset</i> argument of
867 <b>pcre_exec()</b>. It differs from \A when the value of <i>startoffset</i> is
868 non-zero. By calling <b>pcre_exec()</b> multiple times with appropriate
869 arguments, you can mimic Perl's /g option, and it is in this kind of
870 implementation where \G can be useful.
871 </P>
872 <P>
873 Note, however, that PCRE's interpretation of \G, as the start of the current
874 match, is subtly different from Perl's, which defines it as the end of the
875 previous match. In Perl, these can be different when the previously matched
876 string was empty. Because PCRE does just one match at a time, it cannot
877 reproduce this behaviour.
878 </P>
879 <P>
880 If all the alternatives of a pattern begin with \G, the expression is anchored
881 to the starting match position, and the "anchored" flag is set in the compiled
882 regular expression.
883 </P>
884 <br><a name="SEC5" href="#TOC1">CIRCUMFLEX AND DOLLAR</a><br>
885 <P>
886 Outside a character class, in the default matching mode, the circumflex
887 character is an assertion that is true only if the current matching point is
888 at the start of the subject string. If the <i>startoffset</i> argument of
889 <b>pcre_exec()</b> is non-zero, circumflex can never match if the PCRE_MULTILINE
890 option is unset. Inside a character class, circumflex has an entirely different
891 meaning
892 <a href="#characterclass">(see below).</a>
893 </P>
894 <P>
895 Circumflex need not be the first character of the pattern if a number of
896 alternatives are involved, but it should be the first thing in each alternative
897 in which it appears if the pattern is ever to match that branch. If all
898 possible alternatives start with a circumflex, that is, if the pattern is
899 constrained to match only at the start of the subject, it is said to be an
900 "anchored" pattern. (There are also other constructs that can cause a pattern
901 to be anchored.)
902 </P>
903 <P>
904 A dollar character is an assertion that is true only if the current matching
905 point is at the end of the subject string, or immediately before a newline
906 at the end of the string (by default). Dollar need not be the last character of
907 the pattern if a number of alternatives are involved, but it should be the last
908 item in any branch in which it appears. Dollar has no special meaning in a
909 character class.
910 </P>
911 <P>
912 The meaning of dollar can be changed so that it matches only at the very end of
913 the string, by setting the PCRE_DOLLAR_ENDONLY option at compile time. This
914 does not affect the \Z assertion.
915 </P>
916 <P>
917 The meanings of the circumflex and dollar characters are changed if the
918 PCRE_MULTILINE option is set. When this is the case, a circumflex matches
919 immediately after internal newlines as well as at the start of the subject
920 string. It does not match after a newline that ends the string. A dollar
921 matches before any newlines in the string, as well as at the very end, when
922 PCRE_MULTILINE is set. When newline is specified as the two-character
923 sequence CRLF, isolated CR and LF characters do not indicate newlines.
924 </P>
925 <P>
926 For example, the pattern /^abc$/ matches the subject string "def\nabc" (where
927 \n represents a newline) in multiline mode, but not otherwise. Consequently,
928 patterns that are anchored in single line mode because all branches start with
929 ^ are not anchored in multiline mode, and a match for circumflex is possible
930 when the <i>startoffset</i> argument of <b>pcre_exec()</b> is non-zero. The
931 PCRE_DOLLAR_ENDONLY option is ignored if PCRE_MULTILINE is set.
932 </P>
933 <P>
934 Note that the sequences \A, \Z, and \z can be used to match the start and
935 end of the subject in both modes, and if all branches of a pattern start with
936 \A it is always anchored, whether or not PCRE_MULTILINE is set.
937 <a name="fullstopdot"></a></P>
938 <br><a name="SEC6" href="#TOC1">FULL STOP (PERIOD, DOT) AND \N</a><br>
939 <P>
940 Outside a character class, a dot in the pattern matches any one character in
941 the subject string except (by default) a character that signifies the end of a
942 line. In UTF-8 mode, the matched character may be more than one byte long.
943 </P>
944 <P>
945 When a line ending is defined as a single character, dot never matches that
946 character; when the two-character sequence CRLF is used, dot does not match CR
947 if it is immediately followed by LF, but otherwise it matches all characters
948 (including isolated CRs and LFs). When any Unicode line endings are being
949 recognized, dot does not match CR or LF or any of the other line ending
950 characters.
951 </P>
952 <P>
953 The behaviour of dot with regard to newlines can be changed. If the PCRE_DOTALL
954 option is set, a dot matches any one character, without exception. If the
955 two-character sequence CRLF is present in the subject string, it takes two dots
956 to match it.
957 </P>
958 <P>
959 The handling of dot is entirely independent of the handling of circumflex and
960 dollar, the only relationship being that they both involve newlines. Dot has no
961 special meaning in a character class.
962 </P>
963 <P>
964 The escape sequence \N behaves like a dot, except that it is not affected by
965 the PCRE_DOTALL option. In other words, it matches any character except one
966 that signifies the end of a line.
967 </P>
968 <br><a name="SEC7" href="#TOC1">MATCHING A SINGLE BYTE</a><br>
969 <P>
970 Outside a character class, the escape sequence \C matches any one byte, both
971 in and out of UTF-8 mode. Unlike a dot, it always matches line-ending
972 characters. The feature is provided in Perl in order to match individual bytes
973 in UTF-8 mode, but it is unclear how it can usefully be used. Because \C
974 breaks up characters into individual bytes, matching one byte with \C in UTF-8
975 mode means that the rest of the string may start with a malformed UTF-8
976 character. This has undefined results, because PCRE assumes that it is dealing
977 with valid UTF-8 strings (and by default it checks this at the start of
978 processing unless the PCRE_NO_UTF8_CHECK option is used).
979 </P>
980 <P>
981 PCRE does not allow \C to appear in lookbehind assertions
982 <a href="#lookbehind">(described below),</a>
983 because in UTF-8 mode this would make it impossible to calculate the length of
984 the lookbehind.
985 </P>
986 <P>
987 In general, the \C escape sequence is best avoided in UTF-8 mode. However, one
988 way of using it that avoids the problem of malformed UTF-8 characters is to
989 use a lookahead to check the length of the next character, as in this pattern
990 (ignore white space and line breaks):
991 <pre>
992 (?| (?=[\x00-\x7f])(\C) |
993 (?=[\x80-\x{7ff}])(\C)(\C) |
994 (?=[\x{800}-\x{ffff}])(\C)(\C)(\C) |
995 (?=[\x{10000}-\x{1fffff}])(\C)(\C)(\C)(\C))
996 </pre>
997 A group that starts with (?| resets the capturing parentheses numbers in each
998 alternative (see
999 <a href="#dupsubpatternnumber">"Duplicate Subpattern Numbers"</a>
1000 below). The assertions at the start of each branch check the next UTF-8
1001 character for values whose encoding uses 1, 2, 3, or 4 bytes, respectively. The
1002 character's individual bytes are then captured by the appropriate number of
1003 groups.
1004 <a name="characterclass"></a></P>
1005 <br><a name="SEC8" href="#TOC1">SQUARE BRACKETS AND CHARACTER CLASSES</a><br>
1006 <P>
1007 An opening square bracket introduces a character class, terminated by a closing
1008 square bracket. A closing square bracket on its own is not special by default.
1009 However, if the PCRE_JAVASCRIPT_COMPAT option is set, a lone closing square
1010 bracket causes a compile-time error. If a closing square bracket is required as
1011 a member of the class, it should be the first data character in the class
1012 (after an initial circumflex, if present) or escaped with a backslash.
1013 </P>
1014 <P>
1015 A character class matches a single character in the subject. In UTF-8 mode, the
1016 character may be more than one byte long. A matched character must be in the
1017 set of characters defined by the class, unless the first character in the class
1018 definition is a circumflex, in which case the subject character must not be in
1019 the set defined by the class. If a circumflex is actually required as a member
1020 of the class, ensure it is not the first character, or escape it with a
1021 backslash.
1022 </P>
1023 <P>
1024 For example, the character class [aeiou] matches any lower case vowel, while
1025 [^aeiou] matches any character that is not a lower case vowel. Note that a
1026 circumflex is just a convenient notation for specifying the characters that
1027 are in the class by enumerating those that are not. A class that starts with a
1028 circumflex is not an assertion; it still consumes a character from the subject
1029 string, and therefore it fails if the current pointer is at the end of the
1030 string.
1031 </P>
1032 <P>
1033 In UTF-8 mode, characters with values greater than 255 can be included in a
1034 class as a literal string of bytes, or by using the \x{ escaping mechanism.
1035 </P>
1036 <P>
1037 When caseless matching is set, any letters in a class represent both their
1038 upper case and lower case versions, so for example, a caseless [aeiou] matches
1039 "A" as well as "a", and a caseless [^aeiou] does not match "A", whereas a
1040 caseful version would. In UTF-8 mode, PCRE always understands the concept of
1041 case for characters whose values are less than 128, so caseless matching is
1042 always possible. For characters with higher values, the concept of case is
1043 supported if PCRE is compiled with Unicode property support, but not otherwise.
1044 If you want to use caseless matching in UTF8-mode for characters 128 and above,
1045 you must ensure that PCRE is compiled with Unicode property support as well as
1046 with UTF-8 support.
1047 </P>
1048 <P>
1049 Characters that might indicate line breaks are never treated in any special way
1050 when matching character classes, whatever line-ending sequence is in use, and
1051 whatever setting of the PCRE_DOTALL and PCRE_MULTILINE options is used. A class
1052 such as [^a] always matches one of these characters.
1053 </P>
1054 <P>
1055 The minus (hyphen) character can be used to specify a range of characters in a
1056 character class. For example, [d-m] matches any letter between d and m,
1057 inclusive. If a minus character is required in a class, it must be escaped with
1058 a backslash or appear in a position where it cannot be interpreted as
1059 indicating a range, typically as the first or last character in the class.
1060 </P>
1061 <P>
1062 It is not possible to have the literal character "]" as the end character of a
1063 range. A pattern such as [W-]46] is interpreted as a class of two characters
1064 ("W" and "-") followed by a literal string "46]", so it would match "W46]" or
1065 "-46]". However, if the "]" is escaped with a backslash it is interpreted as
1066 the end of range, so [W-\]46] is interpreted as a class containing a range
1067 followed by two other characters. The octal or hexadecimal representation of
1068 "]" can also be used to end a range.
1069 </P>
1070 <P>
1071 Ranges operate in the collating sequence of character values. They can also be
1072 used for characters specified numerically, for example [\000-\037]. In UTF-8
1073 mode, ranges can include characters whose values are greater than 255, for
1074 example [\x{100}-\x{2ff}].
1075 </P>
1076 <P>
1077 If a range that includes letters is used when caseless matching is set, it
1078 matches the letters in either case. For example, [W-c] is equivalent to
1079 [][\\^_`wxyzabc], matched caselessly, and in non-UTF-8 mode, if character
1080 tables for a French locale are in use, [\xc8-\xcb] matches accented E
1081 characters in both cases. In UTF-8 mode, PCRE supports the concept of case for
1082 characters with values greater than 128 only when it is compiled with Unicode
1083 property support.
1084 </P>
1085 <P>
1086 The character escape sequences \d, \D, \h, \H, \p, \P, \s, \S, \v,
1087 \V, \w, and \W may appear in a character class, and add the characters that
1088 they match to the class. For example, [\dABCDEF] matches any hexadecimal
1089 digit. In UTF-8 mode, the PCRE_UCP option affects the meanings of \d, \s, \w
1090 and their upper case partners, just as it does when they appear outside a
1091 character class, as described in the section entitled
1092 <a href="#genericchartypes">"Generic character types"</a>
1093 above. The escape sequence \b has a different meaning inside a character
1094 class; it matches the backspace character. The sequences \B, \N, \R, and \X
1095 are not special inside a character class. Like any other unrecognized escape
1096 sequences, they are treated as the literal characters "B", "N", "R", and "X" by
1097 default, but cause an error if the PCRE_EXTRA option is set.
1098 </P>
1099 <P>
1100 A circumflex can conveniently be used with the upper case character types to
1101 specify a more restricted set of characters than the matching lower case type.
1102 For example, the class [^\W_] matches any letter or digit, but not underscore,
1103 whereas [\w] includes underscore. A positive character class should be read as
1104 "something OR something OR ..." and a negative class as "NOT something AND NOT
1105 something AND NOT ...".
1106 </P>
1107 <P>
1108 The only metacharacters that are recognized in character classes are backslash,
1109 hyphen (only where it can be interpreted as specifying a range), circumflex
1110 (only at the start), opening square bracket (only when it can be interpreted as
1111 introducing a POSIX class name - see the next section), and the terminating
1112 closing square bracket. However, escaping other non-alphanumeric characters
1113 does no harm.
1114 </P>
1115 <br><a name="SEC9" href="#TOC1">POSIX CHARACTER CLASSES</a><br>
1116 <P>
1117 Perl supports the POSIX notation for character classes. This uses names
1118 enclosed by [: and :] within the enclosing square brackets. PCRE also supports
1119 this notation. For example,
1120 <pre>
1121 [01[:alpha:]%]
1122 </pre>
1123 matches "0", "1", any alphabetic character, or "%". The supported class names
1124 are:
1125 <pre>
1126 alnum letters and digits
1127 alpha letters
1128 ascii character codes 0 - 127
1129 blank space or tab only
1130 cntrl control characters
1131 digit decimal digits (same as \d)
1132 graph printing characters, excluding space
1133 lower lower case letters
1134 print printing characters, including space
1135 punct printing characters, excluding letters and digits and space
1136 space white space (not quite the same as \s)
1137 upper upper case letters
1138 word "word" characters (same as \w)
1139 xdigit hexadecimal digits
1140 </pre>
1141 The "space" characters are HT (9), LF (10), VT (11), FF (12), CR (13), and
1142 space (32). Notice that this list includes the VT character (code 11). This
1143 makes "space" different to \s, which does not include VT (for Perl
1144 compatibility).
1145 </P>
1146 <P>
1147 The name "word" is a Perl extension, and "blank" is a GNU extension from Perl
1148 5.8. Another Perl extension is negation, which is indicated by a ^ character
1149 after the colon. For example,
1150 <pre>
1151 [12[:^digit:]]
1152 </pre>
1153 matches "1", "2", or any non-digit. PCRE (and Perl) also recognize the POSIX
1154 syntax [.ch.] and [=ch=] where "ch" is a "collating element", but these are not
1155 supported, and an error is given if they are encountered.
1156 </P>
1157 <P>
1158 By default, in UTF-8 mode, characters with values greater than 128 do not match
1159 any of the POSIX character classes. However, if the PCRE_UCP option is passed
1160 to <b>pcre_compile()</b>, some of the classes are changed so that Unicode
1161 character properties are used. This is achieved by replacing the POSIX classes
1162 by other sequences, as follows:
1163 <pre>
1164 [:alnum:] becomes \p{Xan}
1165 [:alpha:] becomes \p{L}
1166 [:blank:] becomes \h
1167 [:digit:] becomes \p{Nd}
1168 [:lower:] becomes \p{Ll}
1169 [:space:] becomes \p{Xps}
1170 [:upper:] becomes \p{Lu}
1171 [:word:] becomes \p{Xwd}
1172 </pre>
1173 Negated versions, such as [:^alpha:] use \P instead of \p. The other POSIX
1174 classes are unchanged, and match only characters with code points less than
1175 128.
1176 </P>
1177 <br><a name="SEC10" href="#TOC1">VERTICAL BAR</a><br>
1178 <P>
1179 Vertical bar characters are used to separate alternative patterns. For example,
1180 the pattern
1181 <pre>
1182 gilbert|sullivan
1183 </pre>
1184 matches either "gilbert" or "sullivan". Any number of alternatives may appear,
1185 and an empty alternative is permitted (matching the empty string). The matching
1186 process tries each alternative in turn, from left to right, and the first one
1187 that succeeds is used. If the alternatives are within a subpattern
1188 <a href="#subpattern">(defined below),</a>
1189 "succeeds" means matching the rest of the main pattern as well as the
1190 alternative in the subpattern.
1191 </P>
1192 <br><a name="SEC11" href="#TOC1">INTERNAL OPTION SETTING</a><br>
1193 <P>
1194 The settings of the PCRE_CASELESS, PCRE_MULTILINE, PCRE_DOTALL, and
1195 PCRE_EXTENDED options (which are Perl-compatible) can be changed from within
1196 the pattern by a sequence of Perl option letters enclosed between "(?" and ")".
1197 The option letters are
1198 <pre>
1199 i for PCRE_CASELESS
1201 s for PCRE_DOTALL
1202 x for PCRE_EXTENDED
1203 </pre>
1204 For example, (?im) sets caseless, multiline matching. It is also possible to
1205 unset these options by preceding the letter with a hyphen, and a combined
1206 setting and unsetting such as (?im-sx), which sets PCRE_CASELESS and
1207 PCRE_MULTILINE while unsetting PCRE_DOTALL and PCRE_EXTENDED, is also
1208 permitted. If a letter appears both before and after the hyphen, the option is
1209 unset.
1210 </P>
1211 <P>
1212 The PCRE-specific options PCRE_DUPNAMES, PCRE_UNGREEDY, and PCRE_EXTRA can be
1213 changed in the same way as the Perl-compatible options by using the characters
1214 J, U and X respectively.
1215 </P>
1216 <P>
1217 When one of these option changes occurs at top level (that is, not inside
1218 subpattern parentheses), the change applies to the remainder of the pattern
1219 that follows. If the change is placed right at the start of a pattern, PCRE
1220 extracts it into the global options (and it will therefore show up in data
1221 extracted by the <b>pcre_fullinfo()</b> function).
1222 </P>
1223 <P>
1224 An option change within a subpattern (see below for a description of
1225 subpatterns) affects only that part of the subpattern that follows it, so
1226 <pre>
1227 (a(?i)b)c
1228 </pre>
1229 matches abc and aBc and no other strings (assuming PCRE_CASELESS is not used).
1230 By this means, options can be made to have different settings in different
1231 parts of the pattern. Any changes made in one alternative do carry on
1232 into subsequent branches within the same subpattern. For example,
1233 <pre>
1234 (a(?i)b|c)
1235 </pre>
1236 matches "ab", "aB", "c", and "C", even though when matching "C" the first
1237 branch is abandoned before the option setting. This is because the effects of
1238 option settings happen at compile time. There would be some very weird
1239 behaviour otherwise.
1240 </P>
1241 <P>
1242 <b>Note:</b> There are other PCRE-specific options that can be set by the
1243 application when the compile or match functions are called. In some cases the
1244 pattern can contain special leading sequences such as (*CRLF) to override what
1245 the application has set or what has been defaulted. Details are given in the
1246 section entitled
1247 <a href="#newlineseq">"Newline sequences"</a>
1248 above. There are also the (*UTF8) and (*UCP) leading sequences that can be used
1249 to set UTF-8 and Unicode property modes; they are equivalent to setting the
1250 PCRE_UTF8 and the PCRE_UCP options, respectively.
1251 <a name="subpattern"></a></P>
1252 <br><a name="SEC12" href="#TOC1">SUBPATTERNS</a><br>
1253 <P>
1254 Subpatterns are delimited by parentheses (round brackets), which can be nested.
1255 Turning part of a pattern into a subpattern does two things:
1256 <br>
1257 <br>
1258 1. It localizes a set of alternatives. For example, the pattern
1259 <pre>
1260 cat(aract|erpillar|)
1261 </pre>
1262 matches "cataract", "caterpillar", or "cat". Without the parentheses, it would
1263 match "cataract", "erpillar" or an empty string.
1264 <br>
1265 <br>
1266 2. It sets up the subpattern as a capturing subpattern. This means that, when
1267 the whole pattern matches, that portion of the subject string that matched the
1268 subpattern is passed back to the caller via the <i>ovector</i> argument of
1269 <b>pcre_exec()</b>. Opening parentheses are counted from left to right (starting
1270 from 1) to obtain numbers for the capturing subpatterns. For example, if the
1271 string "the red king" is matched against the pattern
1272 <pre>
1273 the ((red|white) (king|queen))
1274 </pre>
1275 the captured substrings are "red king", "red", and "king", and are numbered 1,
1276 2, and 3, respectively.
1277 </P>
1278 <P>
1279 The fact that plain parentheses fulfil two functions is not always helpful.
1280 There are often times when a grouping subpattern is required without a
1281 capturing requirement. If an opening parenthesis is followed by a question mark
1282 and a colon, the subpattern does not do any capturing, and is not counted when
1283 computing the number of any subsequent capturing subpatterns. For example, if
1284 the string "the white queen" is matched against the pattern
1285 <pre>
1286 the ((?:red|white) (king|queen))
1287 </pre>
1288 the captured substrings are "white queen" and "queen", and are numbered 1 and
1289 2. The maximum number of capturing subpatterns is 65535.
1290 </P>
1291 <P>
1292 As a convenient shorthand, if any option settings are required at the start of
1293 a non-capturing subpattern, the option letters may appear between the "?" and
1294 the ":". Thus the two patterns
1295 <pre>
1296 (?i:saturday|sunday)
1297 (?:(?i)saturday|sunday)
1298 </pre>
1299 match exactly the same set of strings. Because alternative branches are tried
1300 from left to right, and options are not reset until the end of the subpattern
1301 is reached, an option setting in one branch does affect subsequent branches, so
1302 the above patterns match "SUNDAY" as well as "Saturday".
1303 <a name="dupsubpatternnumber"></a></P>
1304 <br><a name="SEC13" href="#TOC1">DUPLICATE SUBPATTERN NUMBERS</a><br>
1305 <P>
1306 Perl 5.10 introduced a feature whereby each alternative in a subpattern uses
1307 the same numbers for its capturing parentheses. Such a subpattern starts with
1308 (?| and is itself a non-capturing subpattern. For example, consider this
1309 pattern:
1310 <pre>
1311 (?|(Sat)ur|(Sun))day
1312 </pre>
1313 Because the two alternatives are inside a (?| group, both sets of capturing
1314 parentheses are numbered one. Thus, when the pattern matches, you can look
1315 at captured substring number one, whichever alternative matched. This construct
1316 is useful when you want to capture part, but not all, of one of a number of
1317 alternatives. Inside a (?| group, parentheses are numbered as usual, but the
1318 number is reset at the start of each branch. The numbers of any capturing
1319 parentheses that follow the subpattern start after the highest number used in
1320 any branch. The following example is taken from the Perl documentation. The
1321 numbers underneath show in which buffer the captured content will be stored.
1322 <pre>
1323 # before ---------------branch-reset----------- after
1324 / ( a ) (?| x ( y ) z | (p (q) r) | (t) u (v) ) ( z ) /x
1325 # 1 2 2 3 2 3 4
1326 </pre>
1327 A back reference to a numbered subpattern uses the most recent value that is
1328 set for that number by any subpattern. The following pattern matches "abcabc"
1329 or "defdef":
1330 <pre>
1331 /(?|(abc)|(def))\1/
1332 </pre>
1333 In contrast, a subroutine call to a numbered subpattern always refers to the
1334 first one in the pattern with the given number. The following pattern matches
1335 "abcabc" or "defabc":
1336 <pre>
1337 /(?|(abc)|(def))(?1)/
1338 </pre>
1339 If a
1340 <a href="#conditions">condition test</a>
1341 for a subpattern's having matched refers to a non-unique number, the test is
1342 true if any of the subpatterns of that number have matched.
1343 </P>
1344 <P>
1345 An alternative approach to using this "branch reset" feature is to use
1346 duplicate named subpatterns, as described in the next section.
1347 </P>
1348 <br><a name="SEC14" href="#TOC1">NAMED SUBPATTERNS</a><br>
1349 <P>
1350 Identifying capturing parentheses by number is simple, but it can be very hard
1351 to keep track of the numbers in complicated regular expressions. Furthermore,
1352 if an expression is modified, the numbers may change. To help with this
1353 difficulty, PCRE supports the naming of subpatterns. This feature was not
1354 added to Perl until release 5.10. Python had the feature earlier, and PCRE
1355 introduced it at release 4.0, using the Python syntax. PCRE now supports both
1356 the Perl and the Python syntax. Perl allows identically numbered subpatterns to
1357 have different names, but PCRE does not.
1358 </P>
1359 <P>
1360 In PCRE, a subpattern can be named in one of three ways: (?&#60;name&#62;...) or
1361 (?'name'...) as in Perl, or (?P&#60;name&#62;...) as in Python. References to capturing
1362 parentheses from other parts of the pattern, such as
1363 <a href="#backreferences">back references,</a>
1364 <a href="#recursion">recursion,</a>
1365 and
1366 <a href="#conditions">conditions,</a>
1367 can be made by name as well as by number.
1368 </P>
1369 <P>
1370 Names consist of up to 32 alphanumeric characters and underscores. Named
1371 capturing parentheses are still allocated numbers as well as names, exactly as
1372 if the names were not present. The PCRE API provides function calls for
1373 extracting the name-to-number translation table from a compiled pattern. There
1374 is also a convenience function for extracting a captured substring by name.
1375 </P>
1376 <P>
1377 By default, a name must be unique within a pattern, but it is possible to relax
1378 this constraint by setting the PCRE_DUPNAMES option at compile time. (Duplicate
1379 names are also always permitted for subpatterns with the same number, set up as
1380 described in the previous section.) Duplicate names can be useful for patterns
1381 where only one instance of the named parentheses can match. Suppose you want to
1382 match the name of a weekday, either as a 3-letter abbreviation or as the full
1383 name, and in both cases you want to extract the abbreviation. This pattern
1384 (ignoring the line breaks) does the job:
1385 <pre>
1386 (?&#60;DN&#62;Mon|Fri|Sun)(?:day)?|
1387 (?&#60;DN&#62;Tue)(?:sday)?|
1388 (?&#60;DN&#62;Wed)(?:nesday)?|
1389 (?&#60;DN&#62;Thu)(?:rsday)?|
1390 (?&#60;DN&#62;Sat)(?:urday)?
1391 </pre>
1392 There are five capturing substrings, but only one is ever set after a match.
1393 (An alternative way of solving this problem is to use a "branch reset"
1394 subpattern, as described in the previous section.)
1395 </P>
1396 <P>
1397 The convenience function for extracting the data by name returns the substring
1398 for the first (and in this example, the only) subpattern of that name that
1399 matched. This saves searching to find which numbered subpattern it was.
1400 </P>
1401 <P>
1402 If you make a back reference to a non-unique named subpattern from elsewhere in
1403 the pattern, the one that corresponds to the first occurrence of the name is
1404 used. In the absence of duplicate numbers (see the previous section) this is
1405 the one with the lowest number. If you use a named reference in a condition
1406 test (see the
1407 <a href="#conditions">section about conditions</a>
1408 below), either to check whether a subpattern has matched, or to check for
1409 recursion, all subpatterns with the same name are tested. If the condition is
1410 true for any one of them, the overall condition is true. This is the same
1411 behaviour as testing by number. For further details of the interfaces for
1412 handling named subpatterns, see the
1413 <a href="pcreapi.html"><b>pcreapi</b></a>
1414 documentation.
1415 </P>
1416 <P>
1417 <b>Warning:</b> You cannot use different names to distinguish between two
1418 subpatterns with the same number because PCRE uses only the numbers when
1419 matching. For this reason, an error is given at compile time if different names
1420 are given to subpatterns with the same number. However, you can give the same
1421 name to subpatterns with the same number, even when PCRE_DUPNAMES is not set.
1422 </P>
1423 <br><a name="SEC15" href="#TOC1">REPETITION</a><br>
1424 <P>
1425 Repetition is specified by quantifiers, which can follow any of the following
1426 items:
1427 <pre>
1428 a literal data character
1429 the dot metacharacter
1430 the \C escape sequence
1431 the \X escape sequence (in UTF-8 mode with Unicode properties)
1432 the \R escape sequence
1433 an escape such as \d or \pL that matches a single character
1434 a character class
1435 a back reference (see next section)
1436 a parenthesized subpattern (including assertions)
1437 a subroutine call to a subpattern (recursive or otherwise)
1438 </pre>
1439 The general repetition quantifier specifies a minimum and maximum number of
1440 permitted matches, by giving the two numbers in curly brackets (braces),
1441 separated by a comma. The numbers must be less than 65536, and the first must
1442 be less than or equal to the second. For example:
1443 <pre>
1444 z{2,4}
1445 </pre>
1446 matches "zz", "zzz", or "zzzz". A closing brace on its own is not a special
1447 character. If the second number is omitted, but the comma is present, there is
1448 no upper limit; if the second number and the comma are both omitted, the
1449 quantifier specifies an exact number of required matches. Thus
1450 <pre>
1451 [aeiou]{3,}
1452 </pre>
1453 matches at least 3 successive vowels, but may match many more, while
1454 <pre>
1455 \d{8}
1456 </pre>
1457 matches exactly 8 digits. An opening curly bracket that appears in a position
1458 where a quantifier is not allowed, or one that does not match the syntax of a
1459 quantifier, is taken as a literal character. For example, {,6} is not a
1460 quantifier, but a literal string of four characters.
1461 </P>
1462 <P>
1463 In UTF-8 mode, quantifiers apply to UTF-8 characters rather than to individual
1464 bytes. Thus, for example, \x{100}{2} matches two UTF-8 characters, each of
1465 which is represented by a two-byte sequence. Similarly, when Unicode property
1466 support is available, \X{3} matches three Unicode extended sequences, each of
1467 which may be several bytes long (and they may be of different lengths).
1468 </P>
1469 <P>
1470 The quantifier {0} is permitted, causing the expression to behave as if the
1471 previous item and the quantifier were not present. This may be useful for
1472 subpatterns that are referenced as
1473 <a href="#subpatternsassubroutines">subroutines</a>
1474 from elsewhere in the pattern (but see also the section entitled
1475 <a href="#subdefine">"Defining subpatterns for use by reference only"</a>
1476 below). Items other than subpatterns that have a {0} quantifier are omitted
1477 from the compiled pattern.
1478 </P>
1479 <P>
1480 For convenience, the three most common quantifiers have single-character
1481 abbreviations:
1482 <pre>
1483 * is equivalent to {0,}
1484 + is equivalent to {1,}
1485 ? is equivalent to {0,1}
1486 </pre>
1487 It is possible to construct infinite loops by following a subpattern that can
1488 match no characters with a quantifier that has no upper limit, for example:
1489 <pre>
1490 (a?)*
1491 </pre>
1492 Earlier versions of Perl and PCRE used to give an error at compile time for
1493 such patterns. However, because there are cases where this can be useful, such
1494 patterns are now accepted, but if any repetition of the subpattern does in fact
1495 match no characters, the loop is forcibly broken.
1496 </P>
1497 <P>
1498 By default, the quantifiers are "greedy", that is, they match as much as
1499 possible (up to the maximum number of permitted times), without causing the
1500 rest of the pattern to fail. The classic example of where this gives problems
1501 is in trying to match comments in C programs. These appear between /* and */
1502 and within the comment, individual * and / characters may appear. An attempt to
1503 match C comments by applying the pattern
1504 <pre>
1505 /\*.*\*/
1506 </pre>
1507 to the string
1508 <pre>
1509 /* first comment */ not comment /* second comment */
1510 </pre>
1511 fails, because it matches the entire string owing to the greediness of the .*
1512 item.
1513 </P>
1514 <P>
1515 However, if a quantifier is followed by a question mark, it ceases to be
1516 greedy, and instead matches the minimum number of times possible, so the
1517 pattern
1518 <pre>
1519 /\*.*?\*/
1520 </pre>
1521 does the right thing with the C comments. The meaning of the various
1522 quantifiers is not otherwise changed, just the preferred number of matches.
1523 Do not confuse this use of question mark with its use as a quantifier in its
1524 own right. Because it has two uses, it can sometimes appear doubled, as in
1525 <pre>
1526 \d??\d
1527 </pre>
1528 which matches one digit by preference, but can match two if that is the only
1529 way the rest of the pattern matches.
1530 </P>
1531 <P>
1532 If the PCRE_UNGREEDY option is set (an option that is not available in Perl),
1533 the quantifiers are not greedy by default, but individual ones can be made
1534 greedy by following them with a question mark. In other words, it inverts the
1535 default behaviour.
1536 </P>
1537 <P>
1538 When a parenthesized subpattern is quantified with a minimum repeat count that
1539 is greater than 1 or with a limited maximum, more memory is required for the
1540 compiled pattern, in proportion to the size of the minimum or maximum.
1541 </P>
1542 <P>
1543 If a pattern starts with .* or .{0,} and the PCRE_DOTALL option (equivalent
1544 to Perl's /s) is set, thus allowing the dot to match newlines, the pattern is
1545 implicitly anchored, because whatever follows will be tried against every
1546 character position in the subject string, so there is no point in retrying the
1547 overall match at any position after the first. PCRE normally treats such a
1548 pattern as though it were preceded by \A.
1549 </P>
1550 <P>
1551 In cases where it is known that the subject string contains no newlines, it is
1552 worth setting PCRE_DOTALL in order to obtain this optimization, or
1553 alternatively using ^ to indicate anchoring explicitly.
1554 </P>
1555 <P>
1556 However, there is one situation where the optimization cannot be used. When .*
1557 is inside capturing parentheses that are the subject of a back reference
1558 elsewhere in the pattern, a match at the start may fail where a later one
1559 succeeds. Consider, for example:
1560 <pre>
1561 (.*)abc\1
1562 </pre>
1563 If the subject is "xyz123abc123" the match point is the fourth character. For
1564 this reason, such a pattern is not implicitly anchored.
1565 </P>
1566 <P>
1567 When a capturing subpattern is repeated, the value captured is the substring
1568 that matched the final iteration. For example, after
1569 <pre>
1570 (tweedle[dume]{3}\s*)+
1571 </pre>
1572 has matched "tweedledum tweedledee" the value of the captured substring is
1573 "tweedledee". However, if there are nested capturing subpatterns, the
1574 corresponding captured values may have been set in previous iterations. For
1575 example, after
1576 <pre>
1577 /(a|(b))+/
1578 </pre>
1579 matches "aba" the value of the second captured substring is "b".
1580 <a name="atomicgroup"></a></P>
1581 <br><a name="SEC16" href="#TOC1">ATOMIC GROUPING AND POSSESSIVE QUANTIFIERS</a><br>
1582 <P>
1583 With both maximizing ("greedy") and minimizing ("ungreedy" or "lazy")
1584 repetition, failure of what follows normally causes the repeated item to be
1585 re-evaluated to see if a different number of repeats allows the rest of the
1586 pattern to match. Sometimes it is useful to prevent this, either to change the
1587 nature of the match, or to cause it fail earlier than it otherwise might, when
1588 the author of the pattern knows there is no point in carrying on.
1589 </P>
1590 <P>
1591 Consider, for example, the pattern \d+foo when applied to the subject line
1592 <pre>
1593 123456bar
1594 </pre>
1595 After matching all 6 digits and then failing to match "foo", the normal
1596 action of the matcher is to try again with only 5 digits matching the \d+
1597 item, and then with 4, and so on, before ultimately failing. "Atomic grouping"
1598 (a term taken from Jeffrey Friedl's book) provides the means for specifying
1599 that once a subpattern has matched, it is not to be re-evaluated in this way.
1600 </P>
1601 <P>
1602 If we use atomic grouping for the previous example, the matcher gives up
1603 immediately on failing to match "foo" the first time. The notation is a kind of
1604 special parenthesis, starting with (?&#62; as in this example:
1605 <pre>
1606 (?&#62;\d+)foo
1607 </pre>
1608 This kind of parenthesis "locks up" the part of the pattern it contains once
1609 it has matched, and a failure further into the pattern is prevented from
1610 backtracking into it. Backtracking past it to previous items, however, works as
1611 normal.
1612 </P>
1613 <P>
1614 An alternative description is that a subpattern of this type matches the string
1615 of characters that an identical standalone pattern would match, if anchored at
1616 the current point in the subject string.
1617 </P>
1618 <P>
1619 Atomic grouping subpatterns are not capturing subpatterns. Simple cases such as
1620 the above example can be thought of as a maximizing repeat that must swallow
1621 everything it can. So, while both \d+ and \d+? are prepared to adjust the
1622 number of digits they match in order to make the rest of the pattern match,
1623 (?&#62;\d+) can only match an entire sequence of digits.
1624 </P>
1625 <P>
1626 Atomic groups in general can of course contain arbitrarily complicated
1627 subpatterns, and can be nested. However, when the subpattern for an atomic
1628 group is just a single repeated item, as in the example above, a simpler
1629 notation, called a "possessive quantifier" can be used. This consists of an
1630 additional + character following a quantifier. Using this notation, the
1631 previous example can be rewritten as
1632 <pre>
1633 \d++foo
1634 </pre>
1635 Note that a possessive quantifier can be used with an entire group, for
1636 example:
1637 <pre>
1638 (abc|xyz){2,3}+
1639 </pre>
1640 Possessive quantifiers are always greedy; the setting of the PCRE_UNGREEDY
1641 option is ignored. They are a convenient notation for the simpler forms of
1642 atomic group. However, there is no difference in the meaning of a possessive
1643 quantifier and the equivalent atomic group, though there may be a performance
1644 difference; possessive quantifiers should be slightly faster.
1645 </P>
1646 <P>
1647 The possessive quantifier syntax is an extension to the Perl 5.8 syntax.
1648 Jeffrey Friedl originated the idea (and the name) in the first edition of his
1649 book. Mike McCloskey liked it, so implemented it when he built Sun's Java
1650 package, and PCRE copied it from there. It ultimately found its way into Perl
1651 at release 5.10.
1652 </P>
1653 <P>
1654 PCRE has an optimization that automatically "possessifies" certain simple
1655 pattern constructs. For example, the sequence A+B is treated as A++B because
1656 there is no point in backtracking into a sequence of A's when B must follow.
1657 </P>
1658 <P>
1659 When a pattern contains an unlimited repeat inside a subpattern that can itself
1660 be repeated an unlimited number of times, the use of an atomic group is the
1661 only way to avoid some failing matches taking a very long time indeed. The
1662 pattern
1663 <pre>
1664 (\D+|&#60;\d+&#62;)*[!?]
1665 </pre>
1666 matches an unlimited number of substrings that either consist of non-digits, or
1667 digits enclosed in &#60;&#62;, followed by either ! or ?. When it matches, it runs
1668 quickly. However, if it is applied to
1669 <pre>
1670 aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
1671 </pre>
1672 it takes a long time before reporting failure. This is because the string can
1673 be divided between the internal \D+ repeat and the external * repeat in a
1674 large number of ways, and all have to be tried. (The example uses [!?] rather
1675 than a single character at the end, because both PCRE and Perl have an
1676 optimization that allows for fast failure when a single character is used. They
1677 remember the last single character that is required for a match, and fail early
1678 if it is not present in the string.) If the pattern is changed so that it uses
1679 an atomic group, like this:
1680 <pre>
1681 ((?&#62;\D+)|&#60;\d+&#62;)*[!?]
1682 </pre>
1683 sequences of non-digits cannot be broken, and failure happens quickly.
1684 <a name="backreferences"></a></P>
1685 <br><a name="SEC17" href="#TOC1">BACK REFERENCES</a><br>
1686 <P>
1687 Outside a character class, a backslash followed by a digit greater than 0 (and
1688 possibly further digits) is a back reference to a capturing subpattern earlier
1689 (that is, to its left) in the pattern, provided there have been that many
1690 previous capturing left parentheses.
1691 </P>
1692 <P>
1693 However, if the decimal number following the backslash is less than 10, it is
1694 always taken as a back reference, and causes an error only if there are not
1695 that many capturing left parentheses in the entire pattern. In other words, the
1696 parentheses that are referenced need not be to the left of the reference for
1697 numbers less than 10. A "forward back reference" of this type can make sense
1698 when a repetition is involved and the subpattern to the right has participated
1699 in an earlier iteration.
1700 </P>
1701 <P>
1702 It is not possible to have a numerical "forward back reference" to a subpattern
1703 whose number is 10 or more using this syntax because a sequence such as \50 is
1704 interpreted as a character defined in octal. See the subsection entitled
1705 "Non-printing characters"
1706 <a href="#digitsafterbackslash">above</a>
1707 for further details of the handling of digits following a backslash. There is
1708 no such problem when named parentheses are used. A back reference to any
1709 subpattern is possible using named parentheses (see below).
1710 </P>
1711 <P>
1712 Another way of avoiding the ambiguity inherent in the use of digits following a
1713 backslash is to use the \g escape sequence. This escape must be followed by an
1714 unsigned number or a negative number, optionally enclosed in braces. These
1715 examples are all identical:
1716 <pre>
1717 (ring), \1
1718 (ring), \g1
1719 (ring), \g{1}
1720 </pre>
1721 An unsigned number specifies an absolute reference without the ambiguity that
1722 is present in the older syntax. It is also useful when literal digits follow
1723 the reference. A negative number is a relative reference. Consider this
1724 example:
1725 <pre>
1726 (abc(def)ghi)\g{-1}
1727 </pre>
1728 The sequence \g{-1} is a reference to the most recently started capturing
1729 subpattern before \g, that is, is it equivalent to \2 in this example.
1730 Similarly, \g{-2} would be equivalent to \1. The use of relative references
1731 can be helpful in long patterns, and also in patterns that are created by
1732 joining together fragments that contain references within themselves.
1733 </P>
1734 <P>
1735 A back reference matches whatever actually matched the capturing subpattern in
1736 the current subject string, rather than anything matching the subpattern
1737 itself (see
1738 <a href="#subpatternsassubroutines">"Subpatterns as subroutines"</a>
1739 below for a way of doing that). So the pattern
1740 <pre>
1741 (sens|respons)e and \1ibility
1742 </pre>
1743 matches "sense and sensibility" and "response and responsibility", but not
1744 "sense and responsibility". If caseful matching is in force at the time of the
1745 back reference, the case of letters is relevant. For example,
1746 <pre>
1747 ((?i)rah)\s+\1
1748 </pre>
1749 matches "rah rah" and "RAH RAH", but not "RAH rah", even though the original
1750 capturing subpattern is matched caselessly.
1751 </P>
1752 <P>
1753 There are several different ways of writing back references to named
1754 subpatterns. The .NET syntax \k{name} and the Perl syntax \k&#60;name&#62; or
1755 \k'name' are supported, as is the Python syntax (?P=name). Perl 5.10's unified
1756 back reference syntax, in which \g can be used for both numeric and named
1757 references, is also supported. We could rewrite the above example in any of
1758 the following ways:
1759 <pre>
1760 (?&#60;p1&#62;(?i)rah)\s+\k&#60;p1&#62;
1761 (?'p1'(?i)rah)\s+\k{p1}
1762 (?P&#60;p1&#62;(?i)rah)\s+(?P=p1)
1763 (?&#60;p1&#62;(?i)rah)\s+\g{p1}
1764 </pre>
1765 A subpattern that is referenced by name may appear in the pattern before or
1766 after the reference.
1767 </P>
1768 <P>
1769 There may be more than one back reference to the same subpattern. If a
1770 subpattern has not actually been used in a particular match, any back
1771 references to it always fail by default. For example, the pattern
1772 <pre>
1773 (a|(bc))\2
1774 </pre>
1775 always fails if it starts to match "a" rather than "bc". However, if the
1776 PCRE_JAVASCRIPT_COMPAT option is set at compile time, a back reference to an
1777 unset value matches an empty string.
1778 </P>
1779 <P>
1780 Because there may be many capturing parentheses in a pattern, all digits
1781 following a backslash are taken as part of a potential back reference number.
1782 If the pattern continues with a digit character, some delimiter must be used to
1783 terminate the back reference. If the PCRE_EXTENDED option is set, this can be
1784 whitespace. Otherwise, the \g{ syntax or an empty comment (see
1785 <a href="#comments">"Comments"</a>
1786 below) can be used.
1787 </P>
1788 <br><b>
1789 Recursive back references
1790 </b><br>
1791 <P>
1792 A back reference that occurs inside the parentheses to which it refers fails
1793 when the subpattern is first used, so, for example, (a\1) never matches.
1794 However, such references can be useful inside repeated subpatterns. For
1795 example, the pattern
1796 <pre>
1797 (a|b\1)+
1798 </pre>
1799 matches any number of "a"s and also "aba", "ababbaa" etc. At each iteration of
1800 the subpattern, the back reference matches the character string corresponding
1801 to the previous iteration. In order for this to work, the pattern must be such
1802 that the first iteration does not need to match the back reference. This can be
1803 done using alternation, as in the example above, or by a quantifier with a
1804 minimum of zero.
1805 </P>
1806 <P>
1807 Back references of this type cause the group that they reference to be treated
1808 as an
1809 <a href="#atomicgroup">atomic group.</a>
1810 Once the whole group has been matched, a subsequent matching failure cannot
1811 cause backtracking into the middle of the group.
1812 <a name="bigassertions"></a></P>
1813 <br><a name="SEC18" href="#TOC1">ASSERTIONS</a><br>
1814 <P>
1815 An assertion is a test on the characters following or preceding the current
1816 matching point that does not actually consume any characters. The simple
1817 assertions coded as \b, \B, \A, \G, \Z, \z, ^ and $ are described
1818 <a href="#smallassertions">above.</a>
1819 </P>
1820 <P>
1821 More complicated assertions are coded as subpatterns. There are two kinds:
1822 those that look ahead of the current position in the subject string, and those
1823 that look behind it. An assertion subpattern is matched in the normal way,
1824 except that it does not cause the current matching position to be changed.
1825 </P>
1826 <P>
1827 Assertion subpatterns are not capturing subpatterns. If such an assertion
1828 contains capturing subpatterns within it, these are counted for the purposes of
1829 numbering the capturing subpatterns in the whole pattern. However, substring
1830 capturing is carried out only for positive assertions, because it does not make
1831 sense for negative assertions.
1832 </P>
1833 <P>
1834 For compatibility with Perl, assertion subpatterns may be repeated; though
1835 it makes no sense to assert the same thing several times, the side effect of
1836 capturing parentheses may occasionally be useful. In practice, there only three
1837 cases:
1838 <br>
1839 <br>
1840 (1) If the quantifier is {0}, the assertion is never obeyed during matching.
1841 However, it may contain internal capturing parenthesized groups that are called
1842 from elsewhere via the
1843 <a href="#subpatternsassubroutines">subroutine mechanism.</a>
1844 <br>
1845 <br>
1846 (2) If quantifier is {0,n} where n is greater than zero, it is treated as if it
1847 were {0,1}. At run time, the rest of the pattern match is tried with and
1848 without the assertion, the order depending on the greediness of the quantifier.
1849 <br>
1850 <br>
1851 (3) If the minimum repetition is greater than zero, the quantifier is ignored.
1852 The assertion is obeyed just once when encountered during matching.
1853 </P>
1854 <br><b>
1855 Lookahead assertions
1856 </b><br>
1857 <P>
1858 Lookahead assertions start with (?= for positive assertions and (?! for
1859 negative assertions. For example,
1860 <pre>
1861 \w+(?=;)
1862 </pre>
1863 matches a word followed by a semicolon, but does not include the semicolon in
1864 the match, and
1865 <pre>
1866 foo(?!bar)
1867 </pre>
1868 matches any occurrence of "foo" that is not followed by "bar". Note that the
1869 apparently similar pattern
1870 <pre>
1871 (?!foo)bar
1872 </pre>
1873 does not find an occurrence of "bar" that is preceded by something other than
1874 "foo"; it finds any occurrence of "bar" whatsoever, because the assertion
1875 (?!foo) is always true when the next three characters are "bar". A
1876 lookbehind assertion is needed to achieve the other effect.
1877 </P>
1878 <P>
1879 If you want to force a matching failure at some point in a pattern, the most
1880 convenient way to do it is with (?!) because an empty string always matches, so
1881 an assertion that requires there not to be an empty string must always fail.
1882 The backtracking control verb (*FAIL) or (*F) is a synonym for (?!).
1883 <a name="lookbehind"></a></P>
1884 <br><b>
1885 Lookbehind assertions
1886 </b><br>
1887 <P>
1888 Lookbehind assertions start with (?&#60;= for positive assertions and (?&#60;! for
1889 negative assertions. For example,
1890 <pre>
1891 (?&#60;!foo)bar
1892 </pre>
1893 does find an occurrence of "bar" that is not preceded by "foo". The contents of
1894 a lookbehind assertion are restricted such that all the strings it matches must
1895 have a fixed length. However, if there are several top-level alternatives, they
1896 do not all have to have the same fixed length. Thus
1897 <pre>
1898 (?&#60;=bullock|donkey)
1899 </pre>
1900 is permitted, but
1901 <pre>
1902 (?&#60;!dogs?|cats?)
1903 </pre>
1904 causes an error at compile time. Branches that match different length strings
1905 are permitted only at the top level of a lookbehind assertion. This is an
1906 extension compared with Perl, which requires all branches to match the same
1907 length of string. An assertion such as
1908 <pre>
1909 (?&#60;=ab(c|de))
1910 </pre>
1911 is not permitted, because its single top-level branch can match two different
1912 lengths, but it is acceptable to PCRE if rewritten to use two top-level
1913 branches:
1914 <pre>
1915 (?&#60;=abc|abde)
1916 </pre>
1917 In some cases, the escape sequence \K
1918 <a href="#resetmatchstart">(see above)</a>
1919 can be used instead of a lookbehind assertion to get round the fixed-length
1920 restriction.
1921 </P>
1922 <P>
1923 The implementation of lookbehind assertions is, for each alternative, to
1924 temporarily move the current position back by the fixed length and then try to
1925 match. If there are insufficient characters before the current position, the
1926 assertion fails.
1927 </P>
1928 <P>
1929 PCRE does not allow the \C escape (which matches a single byte in UTF-8 mode)
1930 to appear in lookbehind assertions, because it makes it impossible to calculate
1931 the length of the lookbehind. The \X and \R escapes, which can match
1932 different numbers of bytes, are also not permitted.
1933 </P>
1934 <P>
1935 <a href="#subpatternsassubroutines">"Subroutine"</a>
1936 calls (see below) such as (?2) or (?&X) are permitted in lookbehinds, as long
1937 as the subpattern matches a fixed-length string.
1938 <a href="#recursion">Recursion,</a>
1939 however, is not supported.
1940 </P>
1941 <P>
1942 Possessive quantifiers can be used in conjunction with lookbehind assertions to
1943 specify efficient matching of fixed-length strings at the end of subject
1944 strings. Consider a simple pattern such as
1945 <pre>
1946 abcd$
1947 </pre>
1948 when applied to a long string that does not match. Because matching proceeds
1949 from left to right, PCRE will look for each "a" in the subject and then see if
1950 what follows matches the rest of the pattern. If the pattern is specified as
1951 <pre>
1952 ^.*abcd$
1953 </pre>
1954 the initial .* matches the entire string at first, but when this fails (because
1955 there is no following "a"), it backtracks to match all but the last character,
1956 then all but the last two characters, and so on. Once again the search for "a"
1957 covers the entire string, from right to left, so we are no better off. However,
1958 if the pattern is written as
1959 <pre>
1960 ^.*+(?&#60;=abcd)
1961 </pre>
1962 there can be no backtracking for the .*+ item; it can match only the entire
1963 string. The subsequent lookbehind assertion does a single test on the last four
1964 characters. If it fails, the match fails immediately. For long strings, this
1965 approach makes a significant difference to the processing time.
1966 </P>
1967 <br><b>
1968 Using multiple assertions
1969 </b><br>
1970 <P>
1971 Several assertions (of any sort) may occur in succession. For example,
1972 <pre>
1973 (?&#60;=\d{3})(?&#60;!999)foo
1974 </pre>
1975 matches "foo" preceded by three digits that are not "999". Notice that each of
1976 the assertions is applied independently at the same point in the subject
1977 string. First there is a check that the previous three characters are all
1978 digits, and then there is a check that the same three characters are not "999".
1979 This pattern does <i>not</i> match "foo" preceded by six characters, the first
1980 of which are digits and the last three of which are not "999". For example, it
1981 doesn't match "123abcfoo". A pattern to do that is
1982 <pre>
1983 (?&#60;=\d{3}...)(?&#60;!999)foo
1984 </pre>
1985 This time the first assertion looks at the preceding six characters, checking
1986 that the first three are digits, and then the second assertion checks that the
1987 preceding three characters are not "999".
1988 </P>
1989 <P>
1990 Assertions can be nested in any combination. For example,
1991 <pre>
1992 (?&#60;=(?&#60;!foo)bar)baz
1993 </pre>
1994 matches an occurrence of "baz" that is preceded by "bar" which in turn is not
1995 preceded by "foo", while
1996 <pre>
1997 (?&#60;=\d{3}(?!999)...)foo
1998 </pre>
1999 is another pattern that matches "foo" preceded by three digits and any three
2000 characters that are not "999".
2001 <a name="conditions"></a></P>
2002 <br><a name="SEC19" href="#TOC1">CONDITIONAL SUBPATTERNS</a><br>
2003 <P>
2004 It is possible to cause the matching process to obey a subpattern
2005 conditionally or to choose between two alternative subpatterns, depending on
2006 the result of an assertion, or whether a specific capturing subpattern has
2007 already been matched. The two possible forms of conditional subpattern are:
2008 <pre>
2009 (?(condition)yes-pattern)
2010 (?(condition)yes-pattern|no-pattern)
2011 </pre>
2012 If the condition is satisfied, the yes-pattern is used; otherwise the
2013 no-pattern (if present) is used. If there are more than two alternatives in the
2014 subpattern, a compile-time error occurs. Each of the two alternatives may
2015 itself contain nested subpatterns of any form, including conditional
2016 subpatterns; the restriction to two alternatives applies only at the level of
2017 the condition. This pattern fragment is an example where the alternatives are
2018 complex:
2019 <pre>
2020 (?(1) (A|B|C) | (D | (?(2)E|F) | E) )
2022 </PRE>
2023 </P>
2024 <P>
2025 There are four kinds of condition: references to subpatterns, references to
2026 recursion, a pseudo-condition called DEFINE, and assertions.
2027 </P>
2028 <br><b>
2029 Checking for a used subpattern by number
2030 </b><br>
2031 <P>
2032 If the text between the parentheses consists of a sequence of digits, the
2033 condition is true if a capturing subpattern of that number has previously
2034 matched. If there is more than one capturing subpattern with the same number
2035 (see the earlier
2036 <a href="#recursion">section about duplicate subpattern numbers),</a>
2037 the condition is true if any of them have matched. An alternative notation is
2038 to precede the digits with a plus or minus sign. In this case, the subpattern
2039 number is relative rather than absolute. The most recently opened parentheses
2040 can be referenced by (?(-1), the next most recent by (?(-2), and so on. Inside
2041 loops it can also make sense to refer to subsequent groups. The next
2042 parentheses to be opened can be referenced as (?(+1), and so on. (The value
2043 zero in any of these forms is not used; it provokes a compile-time error.)
2044 </P>
2045 <P>
2046 Consider the following pattern, which contains non-significant white space to
2047 make it more readable (assume the PCRE_EXTENDED option) and to divide it into
2048 three parts for ease of discussion:
2049 <pre>
2050 ( \( )? [^()]+ (?(1) \) )
2051 </pre>
2052 The first part matches an optional opening parenthesis, and if that
2053 character is present, sets it as the first captured substring. The second part
2054 matches one or more characters that are not parentheses. The third part is a
2055 conditional subpattern that tests whether or not the first set of parentheses
2056 matched. If they did, that is, if subject started with an opening parenthesis,
2057 the condition is true, and so the yes-pattern is executed and a closing
2058 parenthesis is required. Otherwise, since no-pattern is not present, the
2059 subpattern matches nothing. In other words, this pattern matches a sequence of
2060 non-parentheses, optionally enclosed in parentheses.
2061 </P>
2062 <P>
2063 If you were embedding this pattern in a larger one, you could use a relative
2064 reference:
2065 <pre>
2066 ...other stuff... ( \( )? [^()]+ (?(-1) \) ) ...
2067 </pre>
2068 This makes the fragment independent of the parentheses in the larger pattern.
2069 </P>
2070 <br><b>
2071 Checking for a used subpattern by name
2072 </b><br>
2073 <P>
2074 Perl uses the syntax (?(&#60;name&#62;)...) or (?('name')...) to test for a used
2075 subpattern by name. For compatibility with earlier versions of PCRE, which had
2076 this facility before Perl, the syntax (?(name)...) is also recognized. However,
2077 there is a possible ambiguity with this syntax, because subpattern names may
2078 consist entirely of digits. PCRE looks first for a named subpattern; if it
2079 cannot find one and the name consists entirely of digits, PCRE looks for a
2080 subpattern of that number, which must be greater than zero. Using subpattern
2081 names that consist entirely of digits is not recommended.
2082 </P>
2083 <P>
2084 Rewriting the above example to use a named subpattern gives this:
2085 <pre>
2086 (?&#60;OPEN&#62; \( )? [^()]+ (?(&#60;OPEN&#62;) \) )
2087 </pre>
2088 If the name used in a condition of this kind is a duplicate, the test is
2089 applied to all subpatterns of the same name, and is true if any one of them has
2090 matched.
2091 </P>
2092 <br><b>
2093 Checking for pattern recursion
2094 </b><br>
2095 <P>
2096 If the condition is the string (R), and there is no subpattern with the name R,
2097 the condition is true if a recursive call to the whole pattern or any
2098 subpattern has been made. If digits or a name preceded by ampersand follow the
2099 letter R, for example:
2100 <pre>
2101 (?(R3)...) or (?(R&name)...)
2102 </pre>
2103 the condition is true if the most recent recursion is into a subpattern whose
2104 number or name is given. This condition does not check the entire recursion
2105 stack. If the name used in a condition of this kind is a duplicate, the test is
2106 applied to all subpatterns of the same name, and is true if any one of them is
2107 the most recent recursion.
2108 </P>
2109 <P>
2110 At "top level", all these recursion test conditions are false.
2111 <a href="#recursion">The syntax for recursive patterns</a>
2112 is described below.
2113 <a name="subdefine"></a></P>
2114 <br><b>
2115 Defining subpatterns for use by reference only
2116 </b><br>
2117 <P>
2118 If the condition is the string (DEFINE), and there is no subpattern with the
2119 name DEFINE, the condition is always false. In this case, there may be only one
2120 alternative in the subpattern. It is always skipped if control reaches this
2121 point in the pattern; the idea of DEFINE is that it can be used to define
2122 subroutines that can be referenced from elsewhere. (The use of
2123 <a href="#subpatternsassubroutines">subroutines</a>
2124 is described below.) For example, a pattern to match an IPv4 address such as
2125 "" could be written like this (ignore whitespace and line
2126 breaks):
2127 <pre>
2128 (?(DEFINE) (?&#60;byte&#62; 2[0-4]\d | 25[0-5] | 1\d\d | [1-9]?\d) )
2129 \b (?&byte) (\.(?&byte)){3} \b
2130 </pre>
2131 The first part of the pattern is a DEFINE group inside which a another group
2132 named "byte" is defined. This matches an individual component of an IPv4
2133 address (a number less than 256). When matching takes place, this part of the
2134 pattern is skipped because DEFINE acts like a false condition. The rest of the
2135 pattern uses references to the named group to match the four dot-separated
2136 components of an IPv4 address, insisting on a word boundary at each end.
2137 </P>
2138 <br><b>
2139 Assertion conditions
2140 </b><br>
2141 <P>
2142 If the condition is not in any of the above formats, it must be an assertion.
2143 This may be a positive or negative lookahead or lookbehind assertion. Consider
2144 this pattern, again containing non-significant white space, and with the two
2145 alternatives on the second line:
2146 <pre>
2147 (?(?=[^a-z]*[a-z])
2148 \d{2}-[a-z]{3}-\d{2} | \d{2}-\d{2}-\d{2} )
2149 </pre>
2150 The condition is a positive lookahead assertion that matches an optional
2151 sequence of non-letters followed by a letter. In other words, it tests for the
2152 presence of at least one letter in the subject. If a letter is found, the
2153 subject is matched against the first alternative; otherwise it is matched
2154 against the second. This pattern matches strings in one of the two forms
2155 dd-aaa-dd or dd-dd-dd, where aaa are letters and dd are digits.
2156 <a name="comments"></a></P>
2157 <br><a name="SEC20" href="#TOC1">COMMENTS</a><br>
2158 <P>
2159 There are two ways of including comments in patterns that are processed by
2160 PCRE. In both cases, the start of the comment must not be in a character class,
2161 nor in the middle of any other sequence of related characters such as (?: or a
2162 subpattern name or number. The characters that make up a comment play no part
2163 in the pattern matching.
2164 </P>
2165 <P>
2166 The sequence (?# marks the start of a comment that continues up to the next
2167 closing parenthesis. Nested parentheses are not permitted. If the PCRE_EXTENDED
2168 option is set, an unescaped # character also introduces a comment, which in
2169 this case continues to immediately after the next newline character or
2170 character sequence in the pattern. Which characters are interpreted as newlines
2171 is controlled by the options passed to <b>pcre_compile()</b> or by a special
2172 sequence at the start of the pattern, as described in the section entitled
2173 <a href="#newlines">"Newline conventions"</a>
2174 above. Note that the end of this type of comment is a literal newline sequence
2175 in the pattern; escape sequences that happen to represent a newline do not
2176 count. For example, consider this pattern when PCRE_EXTENDED is set, and the
2177 default newline convention is in force:
2178 <pre>
2179 abc #comment \n still comment
2180 </pre>
2181 On encountering the # character, <b>pcre_compile()</b> skips along, looking for
2182 a newline in the pattern. The sequence \n is still literal at this stage, so
2183 it does not terminate the comment. Only an actual character with the code value
2184 0x0a (the default newline) does so.
2185 <a name="recursion"></a></P>
2186 <br><a name="SEC21" href="#TOC1">RECURSIVE PATTERNS</a><br>
2187 <P>
2188 Consider the problem of matching a string in parentheses, allowing for
2189 unlimited nested parentheses. Without the use of recursion, the best that can
2190 be done is to use a pattern that matches up to some fixed depth of nesting. It
2191 is not possible to handle an arbitrary nesting depth.
2192 </P>
2193 <P>
2194 For some time, Perl has provided a facility that allows regular expressions to
2195 recurse (amongst other things). It does this by interpolating Perl code in the
2196 expression at run time, and the code can refer to the expression itself. A Perl
2197 pattern using code interpolation to solve the parentheses problem can be
2198 created like this:
2199 <pre>
2200 $re = qr{\( (?: (?&#62;[^()]+) | (?p{$re}) )* \)}x;
2201 </pre>
2202 The (?p{...}) item interpolates Perl code at run time, and in this case refers
2203 recursively to the pattern in which it appears.
2204 </P>
2205 <P>
2206 Obviously, PCRE cannot support the interpolation of Perl code. Instead, it
2207 supports special syntax for recursion of the entire pattern, and also for
2208 individual subpattern recursion. After its introduction in PCRE and Python,
2209 this kind of recursion was subsequently introduced into Perl at release 5.10.
2210 </P>
2211 <P>
2212 A special item that consists of (? followed by a number greater than zero and a
2213 closing parenthesis is a recursive subroutine call of the subpattern of the
2214 given number, provided that it occurs inside that subpattern. (If not, it is a
2215 <a href="#subpatternsassubroutines">non-recursive subroutine</a>
2216 call, which is described in the next section.) The special item (?R) or (?0) is
2217 a recursive call of the entire regular expression.
2218 </P>
2219 <P>
2220 This PCRE pattern solves the nested parentheses problem (assume the
2221 PCRE_EXTENDED option is set so that white space is ignored):
2222 <pre>
2223 \( ( [^()]++ | (?R) )* \)
2224 </pre>
2225 First it matches an opening parenthesis. Then it matches any number of
2226 substrings which can either be a sequence of non-parentheses, or a recursive
2227 match of the pattern itself (that is, a correctly parenthesized substring).
2228 Finally there is a closing parenthesis. Note the use of a possessive quantifier
2229 to avoid backtracking into sequences of non-parentheses.
2230 </P>
2231 <P>
2232 If this were part of a larger pattern, you would not want to recurse the entire
2233 pattern, so instead you could use this:
2234 <pre>
2235 ( \( ( [^()]++ | (?1) )* \) )
2236 </pre>
2237 We have put the pattern into parentheses, and caused the recursion to refer to
2238 them instead of the whole pattern.
2239 </P>
2240 <P>
2241 In a larger pattern, keeping track of parenthesis numbers can be tricky. This
2242 is made easier by the use of relative references. Instead of (?1) in the
2243 pattern above you can write (?-2) to refer to the second most recently opened
2244 parentheses preceding the recursion. In other words, a negative number counts
2245 capturing parentheses leftwards from the point at which it is encountered.
2246 </P>
2247 <P>
2248 It is also possible to refer to subsequently opened parentheses, by writing
2249 references such as (?+2). However, these cannot be recursive because the
2250 reference is not inside the parentheses that are referenced. They are always
2251 <a href="#subpatternsassubroutines">non-recursive subroutine</a>
2252 calls, as described in the next section.
2253 </P>
2254 <P>
2255 An alternative approach is to use named parentheses instead. The Perl syntax
2256 for this is (?&name); PCRE's earlier syntax (?P&#62;name) is also supported. We
2257 could rewrite the above example as follows:
2258 <pre>
2259 (?&#60;pn&#62; \( ( [^()]++ | (?&pn) )* \) )
2260 </pre>
2261 If there is more than one subpattern with the same name, the earliest one is
2262 used.
2263 </P>
2264 <P>
2265 This particular example pattern that we have been looking at contains nested
2266 unlimited repeats, and so the use of a possessive quantifier for matching
2267 strings of non-parentheses is important when applying the pattern to strings
2268 that do not match. For example, when this pattern is applied to
2269 <pre>
2270 (aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa()
2271 </pre>
2272 it yields "no match" quickly. However, if a possessive quantifier is not used,
2273 the match runs for a very long time indeed because there are so many different
2274 ways the + and * repeats can carve up the subject, and all have to be tested
2275 before failure can be reported.
2276 </P>
2277 <P>
2278 At the end of a match, the values of capturing parentheses are those from
2279 the outermost level. If you want to obtain intermediate values, a callout
2280 function can be used (see below and the
2281 <a href="pcrecallout.html"><b>pcrecallout</b></a>
2282 documentation). If the pattern above is matched against
2283 <pre>
2284 (ab(cd)ef)
2285 </pre>
2286 the value for the inner capturing parentheses (numbered 2) is "ef", which is
2287 the last value taken on at the top level. If a capturing subpattern is not
2288 matched at the top level, its final captured value is unset, even if it was
2289 (temporarily) set at a deeper level during the matching process.
2290 </P>
2291 <P>
2292 If there are more than 15 capturing parentheses in a pattern, PCRE has to
2293 obtain extra memory to store data during a recursion, which it does by using
2294 <b>pcre_malloc</b>, freeing it via <b>pcre_free</b> afterwards. If no memory can
2295 be obtained, the match fails with the PCRE_ERROR_NOMEMORY error.
2296 </P>
2297 <P>
2298 Do not confuse the (?R) item with the condition (R), which tests for recursion.
2299 Consider this pattern, which matches text in angle brackets, allowing for
2300 arbitrary nesting. Only digits are allowed in nested brackets (that is, when
2301 recursing), whereas any characters are permitted at the outer level.
2302 <pre>
2303 &#60; (?: (?(R) \d++ | [^&#60;&#62;]*+) | (?R)) * &#62;
2304 </pre>
2305 In this pattern, (?(R) is the start of a conditional subpattern, with two
2306 different alternatives for the recursive and non-recursive cases. The (?R) item
2307 is the actual recursive call.
2308 <a name="recursiondifference"></a></P>
2309 <br><b>
2310 Differences in recursion processing between PCRE and Perl
2311 </b><br>
2312 <P>
2313 Recursion processing in PCRE differs from Perl in two important ways. In PCRE
2314 (like Python, but unlike Perl), a recursive subpattern call is always treated
2315 as an atomic group. That is, once it has matched some of the subject string, it
2316 is never re-entered, even if it contains untried alternatives and there is a
2317 subsequent matching failure. This can be illustrated by the following pattern,
2318 which purports to match a palindromic string that contains an odd number of
2319 characters (for example, "a", "aba", "abcba", "abcdcba"):
2320 <pre>
2321 ^(.|(.)(?1)\2)$
2322 </pre>
2323 The idea is that it either matches a single character, or two identical
2324 characters surrounding a sub-palindrome. In Perl, this pattern works; in PCRE
2325 it does not if the pattern is longer than three characters. Consider the
2326 subject string "abcba":
2327 </P>
2328 <P>
2329 At the top level, the first character is matched, but as it is not at the end
2330 of the string, the first alternative fails; the second alternative is taken
2331 and the recursion kicks in. The recursive call to subpattern 1 successfully
2332 matches the next character ("b"). (Note that the beginning and end of line
2333 tests are not part of the recursion).
2334 </P>
2335 <P>
2336 Back at the top level, the next character ("c") is compared with what
2337 subpattern 2 matched, which was "a". This fails. Because the recursion is
2338 treated as an atomic group, there are now no backtracking points, and so the
2339 entire match fails. (Perl is able, at this point, to re-enter the recursion and
2340 try the second alternative.) However, if the pattern is written with the
2341 alternatives in the other order, things are different:
2342 <pre>
2343 ^((.)(?1)\2|.)$
2344 </pre>
2345 This time, the recursing alternative is tried first, and continues to recurse
2346 until it runs out of characters, at which point the recursion fails. But this
2347 time we do have another alternative to try at the higher level. That is the big
2348 difference: in the previous case the remaining alternative is at a deeper
2349 recursion level, which PCRE cannot use.
2350 </P>
2351 <P>
2352 To change the pattern so that it matches all palindromic strings, not just
2353 those with an odd number of characters, it is tempting to change the pattern to
2354 this:
2355 <pre>
2356 ^((.)(?1)\2|.?)$
2357 </pre>
2358 Again, this works in Perl, but not in PCRE, and for the same reason. When a
2359 deeper recursion has matched a single character, it cannot be entered again in
2360 order to match an empty string. The solution is to separate the two cases, and
2361 write out the odd and even cases as alternatives at the higher level:
2362 <pre>
2363 ^(?:((.)(?1)\2|)|((.)(?3)\4|.))
2364 </pre>
2365 If you want to match typical palindromic phrases, the pattern has to ignore all
2366 non-word characters, which can be done like this:
2367 <pre>
2368 ^\W*+(?:((.)\W*+(?1)\W*+\2|)|((.)\W*+(?3)\W*+\4|\W*+.\W*+))\W*+$
2369 </pre>
2370 If run with the PCRE_CASELESS option, this pattern matches phrases such as "A
2371 man, a plan, a canal: Panama!" and it works well in both PCRE and Perl. Note
2372 the use of the possessive quantifier *+ to avoid backtracking into sequences of
2373 non-word characters. Without this, PCRE takes a great deal longer (ten times or
2374 more) to match typical phrases, and Perl takes so long that you think it has
2375 gone into a loop.
2376 </P>
2377 <P>
2378 <b>WARNING</b>: The palindrome-matching patterns above work only if the subject
2379 string does not start with a palindrome that is shorter than the entire string.
2380 For example, although "abcba" is correctly matched, if the subject is "ababa",
2381 PCRE finds the palindrome "aba" at the start, then fails at top level because
2382 the end of the string does not follow. Once again, it cannot jump back into the
2383 recursion to try other alternatives, so the entire match fails.
2384 </P>
2385 <P>
2386 The second way in which PCRE and Perl differ in their recursion processing is
2387 in the handling of captured values. In Perl, when a subpattern is called
2388 recursively or as a subpattern (see the next section), it has no access to any
2389 values that were captured outside the recursion, whereas in PCRE these values
2390 can be referenced. Consider this pattern:
2391 <pre>
2392 ^(.)(\1|a(?2))
2393 </pre>
2394 In PCRE, this pattern matches "bab". The first capturing parentheses match "b",
2395 then in the second group, when the back reference \1 fails to match "b", the
2396 second alternative matches "a" and then recurses. In the recursion, \1 does
2397 now match "b" and so the whole match succeeds. In Perl, the pattern fails to
2398 match because inside the recursive call \1 cannot access the externally set
2399 value.
2400 <a name="subpatternsassubroutines"></a></P>
2401 <br><a name="SEC22" href="#TOC1">SUBPATTERNS AS SUBROUTINES</a><br>
2402 <P>
2403 If the syntax for a recursive subpattern call (either by number or by
2404 name) is used outside the parentheses to which it refers, it operates like a
2405 subroutine in a programming language. The called subpattern may be defined
2406 before or after the reference. A numbered reference can be absolute or
2407 relative, as in these examples:
2408 <pre>
2409 (...(absolute)...)...(?2)...
2410 (...(relative)...)...(?-1)...
2411 (...(?+1)...(relative)...
2412 </pre>
2413 An earlier example pointed out that the pattern
2414 <pre>
2415 (sens|respons)e and \1ibility
2416 </pre>
2417 matches "sense and sensibility" and "response and responsibility", but not
2418 "sense and responsibility". If instead the pattern
2419 <pre>
2420 (sens|respons)e and (?1)ibility
2421 </pre>
2422 is used, it does match "sense and responsibility" as well as the other two
2423 strings. Another example is given in the discussion of DEFINE above.
2424 </P>
2425 <P>
2426 All subroutine calls, whether recursive or not, are always treated as atomic
2427 groups. That is, once a subroutine has matched some of the subject string, it
2428 is never re-entered, even if it contains untried alternatives and there is a
2429 subsequent matching failure. Any capturing parentheses that are set during the
2430 subroutine call revert to their previous values afterwards.
2431 </P>
2432 <P>
2433 Processing options such as case-independence are fixed when a subpattern is
2434 defined, so if it is used as a subroutine, such options cannot be changed for
2435 different calls. For example, consider this pattern:
2436 <pre>
2437 (abc)(?i:(?-1))
2438 </pre>
2439 It matches "abcabc". It does not match "abcABC" because the change of
2440 processing option does not affect the called subpattern.
2441 <a name="onigurumasubroutines"></a></P>
2442 <br><a name="SEC23" href="#TOC1">ONIGURUMA SUBROUTINE SYNTAX</a><br>
2443 <P>
2444 For compatibility with Oniguruma, the non-Perl syntax \g followed by a name or
2445 a number enclosed either in angle brackets or single quotes, is an alternative
2446 syntax for referencing a subpattern as a subroutine, possibly recursively. Here
2447 are two of the examples used above, rewritten using this syntax:
2448 <pre>
2449 (?&#60;pn&#62; \( ( (?&#62;[^()]+) | \g&#60;pn&#62; )* \) )
2450 (sens|respons)e and \g'1'ibility
2451 </pre>
2452 PCRE supports an extension to Oniguruma: if a number is preceded by a
2453 plus or a minus sign it is taken as a relative reference. For example:
2454 <pre>
2455 (abc)(?i:\g&#60;-1&#62;)
2456 </pre>
2457 Note that \g{...} (Perl syntax) and \g&#60;...&#62; (Oniguruma syntax) are <i>not</i>
2458 synonymous. The former is a back reference; the latter is a subroutine call.
2459 </P>
2460 <br><a name="SEC24" href="#TOC1">CALLOUTS</a><br>
2461 <P>
2462 Perl has a feature whereby using the sequence (?{...}) causes arbitrary Perl
2463 code to be obeyed in the middle of matching a regular expression. This makes it
2464 possible, amongst other things, to extract different substrings that match the
2465 same pair of parentheses when there is a repetition.
2466 </P>
2467 <P>
2468 PCRE provides a similar feature, but of course it cannot obey arbitrary Perl
2469 code. The feature is called "callout". The caller of PCRE provides an external
2470 function by putting its entry point in the global variable <i>pcre_callout</i>.
2471 By default, this variable contains NULL, which disables all calling out.
2472 </P>
2473 <P>
2474 Within a regular expression, (?C) indicates the points at which the external
2475 function is to be called. If you want to identify different callout points, you
2476 can put a number less than 256 after the letter C. The default value is zero.
2477 For example, this pattern has two callout points:
2478 <pre>
2479 (?C1)abc(?C2)def
2480 </pre>
2481 If the PCRE_AUTO_CALLOUT flag is passed to <b>pcre_compile()</b>, callouts are
2482 automatically installed before each item in the pattern. They are all numbered
2483 255.
2484 </P>
2485 <P>
2486 During matching, when PCRE reaches a callout point (and <i>pcre_callout</i> is
2487 set), the external function is called. It is provided with the number of the
2488 callout, the position in the pattern, and, optionally, one item of data
2489 originally supplied by the caller of <b>pcre_exec()</b>. The callout function
2490 may cause matching to proceed, to backtrack, or to fail altogether. A complete
2491 description of the interface to the callout function is given in the
2492 <a href="pcrecallout.html"><b>pcrecallout</b></a>
2493 documentation.
2494 <a name="backtrackcontrol"></a></P>
2495 <br><a name="SEC25" href="#TOC1">BACKTRACKING CONTROL</a><br>
2496 <P>
2497 Perl 5.10 introduced a number of "Special Backtracking Control Verbs", which
2498 are described in the Perl documentation as "experimental and subject to change
2499 or removal in a future version of Perl". It goes on to say: "Their usage in
2500 production code should be noted to avoid problems during upgrades." The same
2501 remarks apply to the PCRE features described in this section.
2502 </P>
2503 <P>
2504 Since these verbs are specifically related to backtracking, most of them can be
2505 used only when the pattern is to be matched using <b>pcre_exec()</b>, which uses
2506 a backtracking algorithm. With the exception of (*FAIL), which behaves like a
2507 failing negative assertion, they cause an error if encountered by
2508 <b>pcre_dfa_exec()</b>.
2509 </P>
2510 <P>
2511 If any of these verbs are used in an assertion or in a subpattern that is
2512 called as a subroutine (whether or not recursively), their effect is confined
2513 to that subpattern; it does not extend to the surrounding pattern, with one
2514 exception: a *MARK that is encountered in a positive assertion <i>is</i> passed
2515 back (compare capturing parentheses in assertions). Note that such subpatterns
2516 are processed as anchored at the point where they are tested. Note also that
2517 Perl's treatment of subroutines is different in some cases.
2518 </P>
2519 <P>
2520 The new verbs make use of what was previously invalid syntax: an opening
2521 parenthesis followed by an asterisk. They are generally of the form
2522 (*VERB) or (*VERB:NAME). Some may take either form, with differing behaviour,
2523 depending on whether or not an argument is present. A name is any sequence of
2524 characters that does not include a closing parenthesis. If the name is empty,
2525 that is, if the closing parenthesis immediately follows the colon, the effect
2526 is as if the colon were not there. Any number of these verbs may occur in a
2527 pattern.
2528 </P>
2529 <P>
2530 PCRE contains some optimizations that are used to speed up matching by running
2531 some checks at the start of each match attempt. For example, it may know the
2532 minimum length of matching subject, or that a particular character must be
2533 present. When one of these optimizations suppresses the running of a match, any
2534 included backtracking verbs will not, of course, be processed. You can suppress
2535 the start-of-match optimizations by setting the PCRE_NO_START_OPTIMIZE option
2536 when calling <b>pcre_compile()</b> or <b>pcre_exec()</b>, or by starting the
2537 pattern with (*NO_START_OPT).
2538 </P>
2539 <br><b>
2540 Verbs that act immediately
2541 </b><br>
2542 <P>
2543 The following verbs act as soon as they are encountered. They may not be
2544 followed by a name.
2545 <pre>
2546 (*ACCEPT)
2547 </pre>
2548 This verb causes the match to end successfully, skipping the remainder of the
2549 pattern. However, when it is inside a subpattern that is called as a
2550 subroutine, only that subpattern is ended successfully. Matching then continues
2551 at the outer level. If (*ACCEPT) is inside capturing parentheses, the data so
2552 far is captured. For example:
2553 <pre>
2554 A((?:A|B(*ACCEPT)|C)D)
2555 </pre>
2556 This matches "AB", "AAD", or "ACD"; when it matches "AB", "B" is captured by
2557 the outer parentheses.
2558 <pre>
2559 (*FAIL) or (*F)
2560 </pre>
2561 This verb causes a matching failure, forcing backtracking to occur. It is
2562 equivalent to (?!) but easier to read. The Perl documentation notes that it is
2563 probably useful only when combined with (?{}) or (??{}). Those are, of course,
2564 Perl features that are not present in PCRE. The nearest equivalent is the
2565 callout feature, as for example in this pattern:
2566 <pre>
2567 a+(?C)(*FAIL)
2568 </pre>
2569 A match with the string "aaaa" always fails, but the callout is taken before
2570 each backtrack happens (in this example, 10 times).
2571 </P>
2572 <br><b>
2573 Recording which path was taken
2574 </b><br>
2575 <P>
2576 There is one verb whose main purpose is to track how a match was arrived at,
2577 though it also has a secondary use in conjunction with advancing the match
2578 starting point (see (*SKIP) below).
2579 <pre>
2580 (*MARK:NAME) or (*:NAME)
2581 </pre>
2582 A name is always required with this verb. There may be as many instances of
2583 (*MARK) as you like in a pattern, and their names do not have to be unique.
2584 </P>
2585 <P>
2586 When a match succeeds, the name of the last-encountered (*MARK) is passed back
2587 to the caller via the <i>pcre_extra</i> data structure, as described in the
2588 <a href="pcreapi.html#extradata">section on <i>pcre_extra</i></a>
2589 in the
2590 <a href="pcreapi.html"><b>pcreapi</b></a>
2591 documentation. No data is returned for a partial match. Here is an example of
2592 <b>pcretest</b> output, where the /K modifier requests the retrieval and
2593 outputting of (*MARK) data:
2594 <pre>
2595 /X(*MARK:A)Y|X(*MARK:B)Z/K
2596 XY
2597 0: XY
2598 MK: A
2599 XZ
2600 0: XZ
2601 MK: B
2602 </pre>
2603 The (*MARK) name is tagged with "MK:" in this output, and in this example it
2604 indicates which of the two alternatives matched. This is a more efficient way
2605 of obtaining this information than putting each alternative in its own
2606 capturing parentheses.
2607 </P>
2608 <P>
2609 If (*MARK) is encountered in a positive assertion, its name is recorded and
2610 passed back if it is the last-encountered. This does not happen for negative
2611 assertions.
2612 </P>
2613 <P>
2614 A name may also be returned after a failed match if the final path through the
2615 pattern involves (*MARK). However, unless (*MARK) used in conjunction with
2616 (*COMMIT), this is unlikely to happen for an unanchored pattern because, as the
2617 starting point for matching is advanced, the final check is often with an empty
2618 string, causing a failure before (*MARK) is reached. For example:
2619 <pre>
2620 /X(*MARK:A)Y|X(*MARK:B)Z/K
2621 XP
2622 No match
2623 </pre>
2624 There are three potential starting points for this match (starting with X,
2625 starting with P, and with an empty string). If the pattern is anchored, the
2626 result is different:
2627 <pre>
2628 /^X(*MARK:A)Y|^X(*MARK:B)Z/K
2629 XP
2630 No match, mark = B
2631 </pre>
2632 PCRE's start-of-match optimizations can also interfere with this. For example,
2633 if, as a result of a call to <b>pcre_study()</b>, it knows the minimum
2634 subject length for a match, a shorter subject will not be scanned at all.
2635 </P>
2636 <P>
2637 Note that similar anomalies (though different in detail) exist in Perl, no
2638 doubt for the same reasons. The use of (*MARK) data after a failed match of an
2639 unanchored pattern is not recommended, unless (*COMMIT) is involved.
2640 </P>
2641 <br><b>
2642 Verbs that act after backtracking
2643 </b><br>
2644 <P>
2645 The following verbs do nothing when they are encountered. Matching continues
2646 with what follows, but if there is no subsequent match, causing a backtrack to
2647 the verb, a failure is forced. That is, backtracking cannot pass to the left of
2648 the verb. However, when one of these verbs appears inside an atomic group, its
2649 effect is confined to that group, because once the group has been matched,
2650 there is never any backtracking into it. In this situation, backtracking can
2651 "jump back" to the left of the entire atomic group. (Remember also, as stated
2652 above, that this localization also applies in subroutine calls and assertions.)
2653 </P>
2654 <P>
2655 These verbs differ in exactly what kind of failure occurs when backtracking
2656 reaches them.
2657 <pre>
2658 (*COMMIT)
2659 </pre>
2660 This verb, which may not be followed by a name, causes the whole match to fail
2661 outright if the rest of the pattern does not match. Even if the pattern is
2662 unanchored, no further attempts to find a match by advancing the starting point
2663 take place. Once (*COMMIT) has been passed, <b>pcre_exec()</b> is committed to
2664 finding a match at the current starting point, or not at all. For example:
2665 <pre>
2666 a+(*COMMIT)b
2667 </pre>
2668 This matches "xxaab" but not "aacaab". It can be thought of as a kind of
2669 dynamic anchor, or "I've started, so I must finish." The name of the most
2670 recently passed (*MARK) in the path is passed back when (*COMMIT) forces a
2671 match failure.
2672 </P>
2673 <P>
2674 Note that (*COMMIT) at the start of a pattern is not the same as an anchor,
2675 unless PCRE's start-of-match optimizations are turned off, as shown in this
2676 <b>pcretest</b> example:
2677 <pre>
2678 /(*COMMIT)abc/
2679 xyzabc
2680 0: abc
2681 xyzabc\Y
2682 No match
2683 </pre>
2684 PCRE knows that any match must start with "a", so the optimization skips along
2685 the subject to "a" before running the first match attempt, which succeeds. When
2686 the optimization is disabled by the \Y escape in the second subject, the match
2687 starts at "x" and so the (*COMMIT) causes it to fail without trying any other
2688 starting points.
2689 <pre>
2690 (*PRUNE) or (*PRUNE:NAME)
2691 </pre>
2692 This verb causes the match to fail at the current starting position in the
2693 subject if the rest of the pattern does not match. If the pattern is
2694 unanchored, the normal "bumpalong" advance to the next starting character then
2695 happens. Backtracking can occur as usual to the left of (*PRUNE), before it is
2696 reached, or when matching to the right of (*PRUNE), but if there is no match to
2697 the right, backtracking cannot cross (*PRUNE). In simple cases, the use of
2698 (*PRUNE) is just an alternative to an atomic group or possessive quantifier,
2699 but there are some uses of (*PRUNE) that cannot be expressed in any other way.
2700 The behaviour of (*PRUNE:NAME) is the same as (*MARK:NAME)(*PRUNE) when the
2701 match fails completely; the name is passed back if this is the final attempt.
2702 (*PRUNE:NAME) does not pass back a name if the match succeeds. In an anchored
2703 pattern (*PRUNE) has the same effect as (*COMMIT).
2704 <pre>
2705 (*SKIP)
2706 </pre>
2707 This verb, when given without a name, is like (*PRUNE), except that if the
2708 pattern is unanchored, the "bumpalong" advance is not to the next character,
2709 but to the position in the subject where (*SKIP) was encountered. (*SKIP)
2710 signifies that whatever text was matched leading up to it cannot be part of a
2711 successful match. Consider:
2712 <pre>
2713 a+(*SKIP)b
2714 </pre>
2715 If the subject is "aaaac...", after the first match attempt fails (starting at
2716 the first character in the string), the starting point skips on to start the
2717 next attempt at "c". Note that a possessive quantifer does not have the same
2718 effect as this example; although it would suppress backtracking during the
2719 first match attempt, the second attempt would start at the second character
2720 instead of skipping on to "c".
2721 <pre>
2722 (*SKIP:NAME)
2723 </pre>
2724 When (*SKIP) has an associated name, its behaviour is modified. If the
2725 following pattern fails to match, the previous path through the pattern is
2726 searched for the most recent (*MARK) that has the same name. If one is found,
2727 the "bumpalong" advance is to the subject position that corresponds to that
2728 (*MARK) instead of to where (*SKIP) was encountered. If no (*MARK) with a
2729 matching name is found, normal "bumpalong" of one character happens (that is,
2730 the (*SKIP) is ignored).
2731 <pre>
2732 (*THEN) or (*THEN:NAME)
2733 </pre>
2734 This verb causes a skip to the next innermost alternative if the rest of the
2735 pattern does not match. That is, it cancels pending backtracking, but only
2736 within the current alternative. Its name comes from the observation that it can
2737 be used for a pattern-based if-then-else block:
2738 <pre>
2739 ( COND1 (*THEN) FOO | COND2 (*THEN) BAR | COND3 (*THEN) BAZ ) ...
2740 </pre>
2741 If the COND1 pattern matches, FOO is tried (and possibly further items after
2742 the end of the group if FOO succeeds); on failure, the matcher skips to the
2743 second alternative and tries COND2, without backtracking into COND1. The
2744 behaviour of (*THEN:NAME) is exactly the same as (*MARK:NAME)(*THEN) if the
2745 overall match fails. If (*THEN) is not inside an alternation, it acts like
2746 (*PRUNE).
2747 </P>
2748 <P>
2749 Note that a subpattern that does not contain a | character is just a part of
2750 the enclosing alternative; it is not a nested alternation with only one
2751 alternative. The effect of (*THEN) extends beyond such a subpattern to the
2752 enclosing alternative. Consider this pattern, where A, B, etc. are complex
2753 pattern fragments that do not contain any | characters at this level:
2754 <pre>
2755 A (B(*THEN)C) | D
2756 </pre>
2757 If A and B are matched, but there is a failure in C, matching does not
2758 backtrack into A; instead it moves to the next alternative, that is, D.
2759 However, if the subpattern containing (*THEN) is given an alternative, it
2760 behaves differently:
2761 <pre>
2762 A (B(*THEN)C | (*FAIL)) | D
2763 </pre>
2764 The effect of (*THEN) is now confined to the inner subpattern. After a failure
2765 in C, matching moves to (*FAIL), which causes the whole subpattern to fail
2766 because there are no more alternatives to try. In this case, matching does now
2767 backtrack into A.
2768 </P>
2769 <P>
2770 Note also that a conditional subpattern is not considered as having two
2771 alternatives, because only one is ever used. In other words, the | character in
2772 a conditional subpattern has a different meaning. Ignoring white space,
2773 consider:
2774 <pre>
2775 ^.*? (?(?=a) a | b(*THEN)c )
2776 </pre>
2777 If the subject is "ba", this pattern does not match. Because .*? is ungreedy,
2778 it initially matches zero characters. The condition (?=a) then fails, the
2779 character "b" is matched, but "c" is not. At this point, matching does not
2780 backtrack to .*? as might perhaps be expected from the presence of the |
2781 character. The conditional subpattern is part of the single alternative that
2782 comprises the whole pattern, and so the match fails. (If there was a backtrack
2783 into .*?, allowing it to match "b", the match would succeed.)
2784 </P>
2785 <P>
2786 The verbs just described provide four different "strengths" of control when
2787 subsequent matching fails. (*THEN) is the weakest, carrying on the match at the
2788 next alternative. (*PRUNE) comes next, failing the match at the current
2789 starting position, but allowing an advance to the next character (for an
2790 unanchored pattern). (*SKIP) is similar, except that the advance may be more
2791 than one character. (*COMMIT) is the strongest, causing the entire match to
2792 fail.
2793 </P>
2794 <P>
2795 If more than one such verb is present in a pattern, the "strongest" one wins.
2796 For example, consider this pattern, where A, B, etc. are complex pattern
2797 fragments:
2798 <pre>
2799 (A(*COMMIT)B(*THEN)C|D)
2800 </pre>
2801 Once A has matched, PCRE is committed to this match, at the current starting
2802 position. If subsequently B matches, but C does not, the normal (*THEN) action
2803 of trying the next alternative (that is, D) does not happen because (*COMMIT)
2804 overrides.
2805 </P>
2806 <br><a name="SEC26" href="#TOC1">SEE ALSO</a><br>
2807 <P>
2808 <b>pcreapi</b>(3), <b>pcrecallout</b>(3), <b>pcrematching</b>(3),
2809 <b>pcresyntax</b>(3), <b>pcre</b>(3).
2810 </P>
2811 <br><a name="SEC27" href="#TOC1">AUTHOR</a><br>
2812 <P>
2813 Philip Hazel
2814 <br>
2815 University Computing Service
2816 <br>
2817 Cambridge CB2 3QH, England.
2818 <br>
2819 </P>
2820 <br><a name="SEC28" href="#TOC1">REVISION</a><br>
2821 <P>
2822 Last updated: 19 October 2011
2823 <br>
2824 Copyright &copy; 1997-2011 University of Cambridge.
2825 <br>
2826 <p>
2827 Return to the <a href="index.html">PCRE index page</a>.
2828 </p>


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

  ViewVC Help
Powered by ViewVC 1.1.5