/[pcre]/code/trunk/doc/html/pcrepattern.html
ViewVC logotype

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

Parent Directory Parent Directory | Revision Log Revision Log


Revision 243 - (hide annotations) (download) (as text)
Thu Sep 13 09:28:14 2007 UTC (7 years, 1 month ago) by ph10
File MIME type: text/html
File size: 92582 byte(s)
Detrailed files for 7.4-RC1 test release.

1 nigel 63 <html>
2     <head>
3     <title>pcrepattern specification</title>
4     </head>
5     <body bgcolor="#FFFFFF" text="#00005A" link="#0066FF" alink="#3399FF" vlink="#2222BB">
6 nigel 75 <h1>pcrepattern man page</h1>
7     <p>
8     Return to the <a href="index.html">PCRE index page</a>.
9     </p>
10 ph10 111 <p>
11 nigel 75 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 ph10 111 <br>
15 nigel 63 <ul>
16     <li><a name="TOC1" href="#SEC1">PCRE REGULAR EXPRESSION DETAILS</a>
17 ph10 227 <li><a name="TOC2" href="#SEC2">NEWLINE CONVENTIONS</a>
18     <li><a name="TOC3" href="#SEC3">CHARACTERS AND METACHARACTERS</a>
19     <li><a name="TOC4" href="#SEC4">BACKSLASH</a>
20     <li><a name="TOC5" href="#SEC5">CIRCUMFLEX AND DOLLAR</a>
21     <li><a name="TOC6" href="#SEC6">FULL STOP (PERIOD, DOT)</a>
22     <li><a name="TOC7" href="#SEC7">MATCHING A SINGLE BYTE</a>
23     <li><a name="TOC8" href="#SEC8">SQUARE BRACKETS AND CHARACTER CLASSES</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>
31     <li><a name="TOC16" href="#SEC16">ATOMIC GROUPING AND POSSESSIVE QUANTIFIERS</a>
32     <li><a name="TOC17" href="#SEC17">BACK REFERENCES</a>
33     <li><a name="TOC18" href="#SEC18">ASSERTIONS</a>
34     <li><a name="TOC19" href="#SEC19">CONDITIONAL SUBPATTERNS</a>
35     <li><a name="TOC20" href="#SEC20">COMMENTS</a>
36     <li><a name="TOC21" href="#SEC21">RECURSIVE PATTERNS</a>
37     <li><a name="TOC22" href="#SEC22">SUBPATTERNS AS SUBROUTINES</a>
38     <li><a name="TOC23" href="#SEC23">CALLOUTS</a>
39 ph10 243 <li><a name="TOC24" href="#SEC24">BACKTRACKING CONTROL</a>
40 ph10 227 <li><a name="TOC25" href="#SEC25">SEE ALSO</a>
41     <li><a name="TOC26" href="#SEC26">AUTHOR</a>
42     <li><a name="TOC27" href="#SEC27">REVISION</a>
43 nigel 63 </ul>
44     <br><a name="SEC1" href="#TOC1">PCRE REGULAR EXPRESSION DETAILS</a><br>
45     <P>
46 ph10 208 The syntax and semantics of the regular expressions that are supported by PCRE
47     are described in detail below. There is a quick-reference syntax summary in the
48     <a href="pcresyntax.html"><b>pcresyntax</b></a>
49     page. Perl's regular expressions are described in its own documentation, and
50     regular expressions in general are covered in a number of books, some of which
51     have copious examples. Jeffrey Friedl's "Mastering Regular Expressions",
52     published by O'Reilly, covers regular expressions in great detail. This
53     description of PCRE's regular expressions is intended as reference material.
54 nigel 63 </P>
55     <P>
56 nigel 75 The original operation of PCRE was on strings of one-byte characters. However,
57     there is now also support for UTF-8 character strings. To use this, you must
58     build PCRE to include UTF-8 support, and then call <b>pcre_compile()</b> with
59     the PCRE_UTF8 option. How this affects pattern matching is mentioned in several
60     places below. There is also a summary of UTF-8 features in the
61 nigel 63 <a href="pcre.html#utf8support">section on UTF-8 support</a>
62     in the main
63     <a href="pcre.html"><b>pcre</b></a>
64     page.
65     </P>
66     <P>
67 nigel 77 The remainder of this document discusses the patterns that are supported by
68     PCRE when its main matching function, <b>pcre_exec()</b>, is used.
69     From release 6.0, PCRE offers a second matching function,
70     <b>pcre_dfa_exec()</b>, which matches using a different algorithm that is not
71 ph10 172 Perl-compatible. Some of the features discussed below are not available when
72     <b>pcre_dfa_exec()</b> is used. The advantages and disadvantages of the
73     alternative function, and how it differs from the normal function, are
74     discussed in the
75 nigel 77 <a href="pcrematching.html"><b>pcrematching</b></a>
76     page.
77     </P>
78 ph10 227 <br><a name="SEC2" href="#TOC1">NEWLINE CONVENTIONS</a><br>
79 nigel 77 <P>
80 ph10 227 PCRE supports five different conventions for indicating line breaks in
81     strings: a single CR (carriage return) character, a single LF (linefeed)
82     character, the two-character sequence CRLF, any of the three preceding, or any
83     Unicode newline sequence. The
84     <a href="pcreapi.html"><b>pcreapi</b></a>
85     page has
86     <a href="pcreapi.html#newlines">further discussion</a>
87     about newlines, and shows how to set the newline convention in the
88     <i>options</i> arguments for the compiling and matching functions.
89     </P>
90     <P>
91     It is also possible to specify a newline convention by starting a pattern
92     string with one of the following five sequences:
93     <pre>
94     (*CR) carriage return
95     (*LF) linefeed
96     (*CRLF) carriage return, followed by linefeed
97     (*ANYCRLF) any of the three above
98     (*ANY) all Unicode newline sequences
99     </pre>
100     These override the default and the options given to <b>pcre_compile()</b>. For
101     example, on a Unix system where LF is the default newline sequence, the pattern
102     <pre>
103     (*CR)a.b
104     </pre>
105     changes the convention to CR. That pattern matches "a\nb" because LF is no
106     longer a newline. Note that these special settings, which are not
107     Perl-compatible, are recognized only at the very start of a pattern, and that
108 ph10 231 they must be in upper case. If more than one of them is present, the last one
109     is used.
110 ph10 227 </P>
111 ph10 231 <P>
112     The newline convention does not affect what the \R escape sequence matches. By
113     default, this is any Unicode newline sequence, for Perl compatibility. However,
114     this can be changed; see the description of \R in the section entitled
115     <a href="#newlineseq">"Newline sequences"</a>
116     below.
117     </P>
118 ph10 227 <br><a name="SEC3" href="#TOC1">CHARACTERS AND METACHARACTERS</a><br>
119     <P>
120 nigel 63 A regular expression is a pattern that is matched against a subject string from
121     left to right. Most characters stand for themselves in a pattern, and match the
122     corresponding characters in the subject. As a trivial example, the pattern
123     <pre>
124     The quick brown fox
125 nigel 75 </pre>
126 nigel 77 matches a portion of a subject string that is identical to itself. When
127     caseless matching is specified (the PCRE_CASELESS option), letters are matched
128     independently of case. In UTF-8 mode, PCRE always understands the concept of
129     case for characters whose values are less than 128, so caseless matching is
130     always possible. For characters with higher values, the concept of case is
131     supported if PCRE is compiled with Unicode property support, but not otherwise.
132     If you want to use caseless matching for characters 128 and above, you must
133     ensure that PCRE is compiled with Unicode property support as well as with
134     UTF-8 support.
135     </P>
136     <P>
137     The power of regular expressions comes from the ability to include alternatives
138     and repetitions in the pattern. These are encoded in the pattern by the use of
139 nigel 75 <i>metacharacters</i>, which do not stand for themselves but instead are
140 nigel 63 interpreted in some special way.
141     </P>
142     <P>
143 nigel 75 There are two different sets of metacharacters: those that are recognized
144 nigel 63 anywhere in the pattern except within square brackets, and those that are
145 nigel 93 recognized within square brackets. Outside square brackets, the metacharacters
146     are as follows:
147 nigel 63 <pre>
148     \ general escape character with several uses
149     ^ assert start of string (or line, in multiline mode)
150     $ assert end of string (or line, in multiline mode)
151     . match any character except newline (by default)
152     [ start character class definition
153     | start of alternative branch
154     ( start subpattern
155     ) end subpattern
156     ? extends the meaning of (
157     also 0 or 1 quantifier
158     also quantifier minimizer
159     * 0 or more quantifier
160     + 1 or more quantifier
161     also "possessive quantifier"
162     { start min/max quantifier
163 nigel 75 </pre>
164 nigel 63 Part of a pattern that is in square brackets is called a "character class". In
165 nigel 75 a character class the only metacharacters are:
166 nigel 63 <pre>
167     \ general escape character
168     ^ negate the class, but only if the first character
169     - indicates character range
170 nigel 75 [ POSIX character class (only if followed by POSIX syntax)
171 nigel 63 ] terminates the character class
172 nigel 75 </pre>
173     The following sections describe the use of each of the metacharacters.
174 nigel 63 </P>
175 ph10 227 <br><a name="SEC4" href="#TOC1">BACKSLASH</a><br>
176 nigel 63 <P>
177     The backslash character has several uses. Firstly, if it is followed by a
178 nigel 91 non-alphanumeric character, it takes away any special meaning that character
179     may have. This use of backslash as an escape character applies both inside and
180 nigel 63 outside character classes.
181     </P>
182     <P>
183     For example, if you want to match a * character, you write \* in the pattern.
184     This escaping action applies whether or not the following character would
185 nigel 75 otherwise be interpreted as a metacharacter, so it is always safe to precede a
186     non-alphanumeric with backslash to specify that it stands for itself. In
187 nigel 63 particular, if you want to match a backslash, you write \\.
188     </P>
189     <P>
190     If a pattern is compiled with the PCRE_EXTENDED option, whitespace in the
191     pattern (other than in a character class) and characters between a # outside
192 nigel 91 a character class and the next newline are ignored. An escaping backslash can
193     be used to include a whitespace or # character as part of the pattern.
194 nigel 63 </P>
195     <P>
196     If you want to remove the special meaning from a sequence of characters, you
197     can do so by putting them between \Q and \E. This is different from Perl in
198     that $ and @ are handled as literals in \Q...\E sequences in PCRE, whereas in
199     Perl, $ and @ cause variable interpolation. Note the following examples:
200     <pre>
201     Pattern PCRE matches Perl matches
202 nigel 75
203     \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
204 nigel 63 \Qabc\$xyz\E abc\$xyz abc\$xyz
205     \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
206 nigel 75 </pre>
207 nigel 63 The \Q...\E sequence is recognized both inside and outside character classes.
208 nigel 75 <a name="digitsafterbackslash"></a></P>
209     <br><b>
210     Non-printing characters
211     </b><br>
212 nigel 63 <P>
213     A second use of backslash provides a way of encoding non-printing characters
214     in patterns in a visible manner. There is no restriction on the appearance of
215     non-printing characters, apart from the binary zero that terminates a pattern,
216     but when a pattern is being prepared by text editing, it is usually easier to
217     use one of the following escape sequences than the binary character it
218     represents:
219     <pre>
220     \a alarm, that is, the BEL character (hex 07)
221     \cx "control-x", where x is any character
222     \e escape (hex 1B)
223     \f formfeed (hex 0C)
224 ph10 227 \n linefeed (hex 0A)
225 nigel 63 \r carriage return (hex 0D)
226     \t tab (hex 09)
227     \ddd character with octal code ddd, or backreference
228     \xhh character with hex code hh
229 nigel 87 \x{hhh..} character with hex code hhh..
230 nigel 75 </pre>
231 nigel 63 The precise effect of \cx is as follows: if x is a lower case letter, it
232     is converted to upper case. Then bit 6 of the character (hex 40) is inverted.
233     Thus \cz becomes hex 1A, but \c{ becomes hex 3B, while \c; becomes hex
234     7B.
235     </P>
236     <P>
237     After \x, from zero to two hexadecimal digits are read (letters can be in
238 nigel 87 upper or lower case). Any number of hexadecimal digits may appear between \x{
239     and }, but the value of the character code must be less than 256 in non-UTF-8
240 ph10 211 mode, and less than 2**31 in UTF-8 mode. That is, the maximum value in
241     hexadecimal is 7FFFFFFF. Note that this is bigger than the largest Unicode code
242     point, which is 10FFFF.
243 nigel 63 </P>
244     <P>
245 ph10 211 If characters other than hexadecimal digits appear between \x{ and }, or if
246     there is no terminating }, this form of escape is not recognized. Instead, the
247     initial \x will be interpreted as a basic hexadecimal escape, with no
248     following digits, giving a character whose value is zero.
249     </P>
250     <P>
251 nigel 63 Characters whose value is less than 256 can be defined by either of the two
252 nigel 87 syntaxes for \x. There is no difference in the way they are handled. For
253     example, \xdc is exactly the same as \x{dc}.
254 nigel 63 </P>
255     <P>
256 nigel 91 After \0 up to two further octal digits are read. If there are fewer than two
257     digits, just those that are present are used. Thus the sequence \0\x\07
258     specifies two binary zeros followed by a BEL character (code value 7). Make
259     sure you supply two digits after the initial zero if the pattern character that
260     follows is itself an octal digit.
261 nigel 63 </P>
262     <P>
263     The handling of a backslash followed by a digit other than 0 is complicated.
264     Outside a character class, PCRE reads it and any following digits as a decimal
265     number. If the number is less than 10, or if there have been at least that many
266     previous capturing left parentheses in the expression, the entire sequence is
267     taken as a <i>back reference</i>. A description of how this works is given
268 nigel 75 <a href="#backreferences">later,</a>
269     following the discussion of
270     <a href="#subpattern">parenthesized subpatterns.</a>
271 nigel 63 </P>
272     <P>
273     Inside a character class, or if the decimal number is greater than 9 and there
274     have not been that many capturing subpatterns, PCRE re-reads up to three octal
275 nigel 93 digits following the backslash, and uses them to generate a data character. Any
276 nigel 91 subsequent digits stand for themselves. In non-UTF-8 mode, the value of a
277     character specified in octal must be less than \400. In UTF-8 mode, values up
278     to \777 are permitted. For example:
279 nigel 63 <pre>
280     \040 is another way of writing a space
281 nigel 75 \40 is the same, provided there are fewer than 40 previous capturing subpatterns
282 nigel 63 \7 is always a back reference
283 nigel 75 \11 might be a back reference, or another way of writing a tab
284 nigel 63 \011 is always a tab
285     \0113 is a tab followed by the character "3"
286 nigel 75 \113 might be a back reference, otherwise the character with octal code 113
287     \377 might be a back reference, otherwise the byte consisting entirely of 1 bits
288     \81 is either a back reference, or a binary zero followed by the two characters "8" and "1"
289     </pre>
290 nigel 63 Note that octal values of 100 or greater must not be introduced by a leading
291     zero, because no more than three octal digits are ever read.
292     </P>
293     <P>
294 nigel 91 All the sequences that define a single character value can be used both inside
295     and outside character classes. In addition, inside a character class, the
296     sequence \b is interpreted as the backspace character (hex 08), and the
297 nigel 93 sequences \R and \X are interpreted as the characters "R" and "X",
298     respectively. Outside a character class, these sequences have different
299     meanings
300 nigel 75 <a href="#uniextseq">(see below).</a>
301 nigel 63 </P>
302 nigel 75 <br><b>
303 nigel 93 Absolute and relative back references
304     </b><br>
305     <P>
306 ph10 208 The sequence \g followed by an unsigned or a negative number, optionally
307     enclosed in braces, is an absolute or relative back reference. A named back
308     reference can be coded as \g{name}. Back references are discussed
309 nigel 93 <a href="#backreferences">later,</a>
310     following the discussion of
311     <a href="#subpattern">parenthesized subpatterns.</a>
312     </P>
313     <br><b>
314 nigel 75 Generic character types
315     </b><br>
316 nigel 63 <P>
317 nigel 93 Another use of backslash is for specifying generic character types. The
318 nigel 75 following are always recognized:
319 nigel 63 <pre>
320     \d any decimal digit
321     \D any character that is not a decimal digit
322 ph10 182 \h any horizontal whitespace character
323     \H any character that is not a horizontal whitespace character
324 nigel 63 \s any whitespace character
325     \S any character that is not a whitespace character
326 ph10 182 \v any vertical whitespace character
327     \V any character that is not a vertical whitespace character
328 nigel 63 \w any "word" character
329     \W any "non-word" character
330 nigel 75 </pre>
331 nigel 63 Each pair of escape sequences partitions the complete set of characters into
332     two disjoint sets. Any given character matches one, and only one, of each pair.
333     </P>
334     <P>
335 nigel 75 These character type sequences can appear both inside and outside character
336     classes. They each match one character of the appropriate type. If the current
337     matching point is at the end of the subject string, all of them fail, since
338     there is no character to match.
339 nigel 63 </P>
340     <P>
341     For compatibility with Perl, \s does not match the VT character (code 11).
342     This makes it different from the the POSIX "space" class. The \s characters
343 ph10 182 are HT (9), LF (10), FF (12), CR (13), and space (32). If "use locale;" is
344 nigel 91 included in a Perl script, \s may match the VT character. In PCRE, it never
345 ph10 182 does.
346 nigel 63 </P>
347     <P>
348 ph10 182 In UTF-8 mode, characters with values greater than 128 never match \d, \s, or
349     \w, and always match \D, \S, and \W. This is true even when Unicode
350     character property support is available. These sequences retain their original
351     meanings from before UTF-8 support was available, mainly for efficiency
352     reasons.
353     </P>
354     <P>
355     The sequences \h, \H, \v, and \V are Perl 5.10 features. In contrast to the
356     other sequences, these do match certain high-valued codepoints in UTF-8 mode.
357     The horizontal space characters are:
358     <pre>
359     U+0009 Horizontal tab
360     U+0020 Space
361     U+00A0 Non-break space
362     U+1680 Ogham space mark
363     U+180E Mongolian vowel separator
364     U+2000 En quad
365     U+2001 Em quad
366     U+2002 En space
367     U+2003 Em space
368     U+2004 Three-per-em space
369     U+2005 Four-per-em space
370     U+2006 Six-per-em space
371     U+2007 Figure space
372     U+2008 Punctuation space
373     U+2009 Thin space
374     U+200A Hair space
375     U+202F Narrow no-break space
376     U+205F Medium mathematical space
377     U+3000 Ideographic space
378     </pre>
379     The vertical space characters are:
380     <pre>
381     U+000A Linefeed
382     U+000B Vertical tab
383     U+000C Formfeed
384     U+000D Carriage return
385     U+0085 Next line
386     U+2028 Line separator
387     U+2029 Paragraph separator
388     </PRE>
389     </P>
390     <P>
391 nigel 75 A "word" character is an underscore or any character less than 256 that is a
392     letter or digit. The definition of letters and digits is controlled by PCRE's
393     low-valued character tables, and may vary if locale-specific matching is taking
394     place (see
395 nigel 63 <a href="pcreapi.html#localesupport">"Locale support"</a>
396     in the
397     <a href="pcreapi.html"><b>pcreapi</b></a>
398 ph10 142 page). For example, in a French locale such as "fr_FR" in Unix-like systems,
399     or "french" in Windows, some character codes greater than 128 are used for
400 ph10 182 accented letters, and these are matched by \w. The use of locales with Unicode
401     is discouraged.
402 ph10 231 <a name="newlineseq"></a></P>
403 nigel 93 <br><b>
404     Newline sequences
405     </b><br>
406     <P>
407 ph10 231 Outside a character class, by default, the escape sequence \R matches any
408     Unicode newline sequence. This is a Perl 5.10 feature. In non-UTF-8 mode \R is
409     equivalent to the following:
410 nigel 93 <pre>
411     (?&#62;\r\n|\n|\x0b|\f|\r|\x85)
412     </pre>
413     This is an example of an "atomic group", details of which are given
414     <a href="#atomicgroup">below.</a>
415     This particular group matches either the two-character sequence CR followed by
416     LF, or one of the single characters LF (linefeed, U+000A), VT (vertical tab,
417     U+000B), FF (formfeed, U+000C), CR (carriage return, U+000D), or NEL (next
418     line, U+0085). The two-character sequence is treated as a single unit that
419     cannot be split.
420     </P>
421     <P>
422     In UTF-8 mode, two additional characters whose codepoints are greater than 255
423     are added: LS (line separator, U+2028) and PS (paragraph separator, U+2029).
424     Unicode character property support is not needed for these characters to be
425     recognized.
426     </P>
427     <P>
428 ph10 231 It is possible to restrict \R to match only CR, LF, or CRLF (instead of the
429     complete set of Unicode line endings) by setting the option PCRE_BSR_ANYCRLF
430     either at compile time or when the pattern is matched. This can be made the
431     default when PCRE is built; if this is the case, the other behaviour can be
432     requested via the PCRE_BSR_UNICODE option. It is also possible to specify these
433     settings by starting a pattern string with one of the following sequences:
434     <pre>
435     (*BSR_ANYCRLF) CR, LF, or CRLF only
436     (*BSR_UNICODE) any Unicode newline sequence
437     </pre>
438     These override the default and the options given to <b>pcre_compile()</b>, but
439     they can be overridden by options given to <b>pcre_exec()</b>. Note that these
440     special settings, which are not Perl-compatible, are recognized only at the
441     very start of a pattern, and that they must be in upper case. If more than one
442     of them is present, the last one is used.
443     </P>
444     <P>
445 nigel 93 Inside a character class, \R matches the letter "R".
446 nigel 75 <a name="uniextseq"></a></P>
447     <br><b>
448     Unicode character properties
449     </b><br>
450     <P>
451     When PCRE is built with Unicode character property support, three additional
452 ph10 185 escape sequences that match characters with specific properties are available.
453     When not in UTF-8 mode, these sequences are of course limited to testing
454     characters whose codepoints are less than 256, but they do work in this mode.
455     The extra escape sequences are:
456 nigel 75 <pre>
457 nigel 87 \p{<i>xx</i>} a character with the <i>xx</i> property
458     \P{<i>xx</i>} a character without the <i>xx</i> property
459     \X an extended Unicode sequence
460 nigel 75 </pre>
461 nigel 87 The property names represented by <i>xx</i> above are limited to the Unicode
462     script names, the general category properties, and "Any", which matches any
463     character (including newline). Other properties such as "InMusicalSymbols" are
464     not currently supported by PCRE. Note that \P{Any} does not match any
465     characters, so always causes a match failure.
466 nigel 63 </P>
467     <P>
468 nigel 87 Sets of Unicode characters are defined as belonging to certain scripts. A
469     character from one of these sets can be matched using a script name. For
470     example:
471 nigel 75 <pre>
472 nigel 87 \p{Greek}
473     \P{Han}
474     </pre>
475     Those that are not part of an identified script are lumped together as
476     "Common". The current list of scripts is:
477     </P>
478     <P>
479     Arabic,
480     Armenian,
481 nigel 93 Balinese,
482 nigel 87 Bengali,
483     Bopomofo,
484     Braille,
485     Buginese,
486     Buhid,
487     Canadian_Aboriginal,
488     Cherokee,
489     Common,
490     Coptic,
491 nigel 93 Cuneiform,
492 nigel 87 Cypriot,
493     Cyrillic,
494     Deseret,
495     Devanagari,
496     Ethiopic,
497     Georgian,
498     Glagolitic,
499     Gothic,
500     Greek,
501     Gujarati,
502     Gurmukhi,
503     Han,
504     Hangul,
505     Hanunoo,
506     Hebrew,
507     Hiragana,
508     Inherited,
509     Kannada,
510     Katakana,
511     Kharoshthi,
512     Khmer,
513     Lao,
514     Latin,
515     Limbu,
516     Linear_B,
517     Malayalam,
518     Mongolian,
519     Myanmar,
520     New_Tai_Lue,
521 nigel 93 Nko,
522 nigel 87 Ogham,
523     Old_Italic,
524     Old_Persian,
525     Oriya,
526     Osmanya,
527 nigel 93 Phags_Pa,
528     Phoenician,
529 nigel 87 Runic,
530     Shavian,
531     Sinhala,
532     Syloti_Nagri,
533     Syriac,
534     Tagalog,
535     Tagbanwa,
536     Tai_Le,
537     Tamil,
538     Telugu,
539     Thaana,
540     Thai,
541     Tibetan,
542     Tifinagh,
543     Ugaritic,
544     Yi.
545     </P>
546     <P>
547     Each character has exactly one general category property, specified by a
548     two-letter abbreviation. For compatibility with Perl, negation can be specified
549     by including a circumflex between the opening brace and the property name. For
550     example, \p{^Lu} is the same as \P{Lu}.
551     </P>
552     <P>
553     If only one letter is specified with \p or \P, it includes all the general
554     category properties that start with that letter. In this case, in the absence
555     of negation, the curly brackets in the escape sequence are optional; these two
556     examples have the same effect:
557     <pre>
558 nigel 75 \p{L}
559     \pL
560     </pre>
561 nigel 87 The following general category property codes are supported:
562 nigel 75 <pre>
563     C Other
564     Cc Control
565     Cf Format
566     Cn Unassigned
567     Co Private use
568     Cs Surrogate
569    
570     L Letter
571     Ll Lower case letter
572     Lm Modifier letter
573     Lo Other letter
574     Lt Title case letter
575     Lu Upper case letter
576    
577     M Mark
578     Mc Spacing mark
579     Me Enclosing mark
580     Mn Non-spacing mark
581    
582     N Number
583     Nd Decimal number
584     Nl Letter number
585     No Other number
586    
587     P Punctuation
588     Pc Connector punctuation
589     Pd Dash punctuation
590     Pe Close punctuation
591     Pf Final punctuation
592     Pi Initial punctuation
593     Po Other punctuation
594     Ps Open punctuation
595    
596     S Symbol
597     Sc Currency symbol
598     Sk Modifier symbol
599     Sm Mathematical symbol
600     So Other symbol
601    
602     Z Separator
603     Zl Line separator
604     Zp Paragraph separator
605     Zs Space separator
606     </pre>
607 nigel 87 The special property L& is also supported: it matches a character that has
608     the Lu, Ll, or Lt property, in other words, a letter that is not classified as
609     a modifier or "other".
610 nigel 75 </P>
611     <P>
612 ph10 211 The Cs (Surrogate) property applies only to characters in the range U+D800 to
613     U+DFFF. Such characters are not valid in UTF-8 strings (see RFC 3629) and so
614     cannot be tested by PCRE, unless UTF-8 validity checking has been turned off
615     (see the discussion of PCRE_NO_UTF8_CHECK in the
616     <a href="pcreapi.html"><b>pcreapi</b></a>
617     page).
618     </P>
619     <P>
620 nigel 87 The long synonyms for these properties that Perl supports (such as \p{Letter})
621 nigel 91 are not supported by PCRE, nor is it permitted to prefix any of these
622 nigel 87 properties with "Is".
623     </P>
624     <P>
625     No character that is in the Unicode table has the Cn (unassigned) property.
626     Instead, this property is assumed for any code point that is not in the
627     Unicode table.
628     </P>
629     <P>
630 nigel 75 Specifying caseless matching does not affect these escape sequences. For
631     example, \p{Lu} always matches only upper case letters.
632     </P>
633     <P>
634     The \X escape matches any number of Unicode characters that form an extended
635     Unicode sequence. \X is equivalent to
636     <pre>
637     (?&#62;\PM\pM*)
638     </pre>
639     That is, it matches a character without the "mark" property, followed by zero
640     or more characters with the "mark" property, and treats the sequence as an
641     atomic group
642     <a href="#atomicgroup">(see below).</a>
643     Characters with the "mark" property are typically accents that affect the
644 ph10 185 preceding character. None of them have codepoints less than 256, so in
645     non-UTF-8 mode \X matches any one character.
646 nigel 75 </P>
647     <P>
648     Matching characters by Unicode property is not fast, because PCRE has to search
649     a structure that contains data for over fifteen thousand characters. That is
650     why the traditional escape sequences such as \d and \w do not use Unicode
651     properties in PCRE.
652 ph10 172 <a name="resetmatchstart"></a></P>
653     <br><b>
654     Resetting the match start
655     </b><br>
656     <P>
657     The escape sequence \K, which is a Perl 5.10 feature, causes any previously
658     matched characters not to be included in the final matched sequence. For
659     example, the pattern:
660     <pre>
661     foo\Kbar
662     </pre>
663     matches "foobar", but reports that it has matched "bar". This feature is
664     similar to a lookbehind assertion
665     <a href="#lookbehind">(described below).</a>
666     However, in this case, the part of the subject before the real match does not
667     have to be of fixed length, as lookbehind assertions do. The use of \K does
668     not interfere with the setting of
669     <a href="#subpattern">captured substrings.</a>
670     For example, when the pattern
671     <pre>
672     (foo)\Kbar
673     </pre>
674     matches "foobar", the first substring is still set to "foo".
675 nigel 75 <a name="smallassertions"></a></P>
676     <br><b>
677     Simple assertions
678     </b><br>
679     <P>
680 nigel 93 The final use of backslash is for certain simple assertions. An assertion
681 nigel 63 specifies a condition that has to be met at a particular point in a match,
682     without consuming any characters from the subject string. The use of
683 nigel 75 subpatterns for more complicated assertions is described
684     <a href="#bigassertions">below.</a>
685 nigel 91 The backslashed assertions are:
686 nigel 63 <pre>
687     \b matches at a word boundary
688     \B matches when not at a word boundary
689 nigel 93 \A matches at the start of the subject
690     \Z matches at the end of the subject
691     also matches before a newline at the end of the subject
692     \z matches only at the end of the subject
693     \G matches at the first matching position in the subject
694 nigel 75 </pre>
695 nigel 63 These assertions may not appear in character classes (but note that \b has a
696     different meaning, namely the backspace character, inside a character class).
697     </P>
698     <P>
699     A word boundary is a position in the subject string where the current character
700     and the previous character do not both match \w or \W (i.e. one matches
701     \w and the other matches \W), or the start or end of the string if the
702     first or last character matches \w, respectively.
703     </P>
704     <P>
705     The \A, \Z, and \z assertions differ from the traditional circumflex and
706 nigel 75 dollar (described in the next section) in that they only ever match at the very
707     start and end of the subject string, whatever options are set. Thus, they are
708     independent of multiline mode. These three assertions are not affected by the
709     PCRE_NOTBOL or PCRE_NOTEOL options, which affect only the behaviour of the
710     circumflex and dollar metacharacters. However, if the <i>startoffset</i>
711     argument of <b>pcre_exec()</b> is non-zero, indicating that matching is to start
712     at a point other than the beginning of the subject, \A can never match. The
713 nigel 91 difference between \Z and \z is that \Z matches before a newline at the end
714     of the string as well as at the very end, whereas \z matches only at the end.
715 nigel 63 </P>
716     <P>
717     The \G assertion is true only when the current matching position is at the
718     start point of the match, as specified by the <i>startoffset</i> argument of
719     <b>pcre_exec()</b>. It differs from \A when the value of <i>startoffset</i> is
720     non-zero. By calling <b>pcre_exec()</b> multiple times with appropriate
721     arguments, you can mimic Perl's /g option, and it is in this kind of
722     implementation where \G can be useful.
723     </P>
724     <P>
725     Note, however, that PCRE's interpretation of \G, as the start of the current
726     match, is subtly different from Perl's, which defines it as the end of the
727     previous match. In Perl, these can be different when the previously matched
728     string was empty. Because PCRE does just one match at a time, it cannot
729     reproduce this behaviour.
730     </P>
731     <P>
732     If all the alternatives of a pattern begin with \G, the expression is anchored
733     to the starting match position, and the "anchored" flag is set in the compiled
734     regular expression.
735     </P>
736 ph10 227 <br><a name="SEC5" href="#TOC1">CIRCUMFLEX AND DOLLAR</a><br>
737 nigel 63 <P>
738     Outside a character class, in the default matching mode, the circumflex
739 nigel 75 character is an assertion that is true only if the current matching point is
740 nigel 63 at the start of the subject string. If the <i>startoffset</i> argument of
741     <b>pcre_exec()</b> is non-zero, circumflex can never match if the PCRE_MULTILINE
742     option is unset. Inside a character class, circumflex has an entirely different
743 nigel 75 meaning
744     <a href="#characterclass">(see below).</a>
745 nigel 63 </P>
746     <P>
747     Circumflex need not be the first character of the pattern if a number of
748     alternatives are involved, but it should be the first thing in each alternative
749     in which it appears if the pattern is ever to match that branch. If all
750     possible alternatives start with a circumflex, that is, if the pattern is
751     constrained to match only at the start of the subject, it is said to be an
752     "anchored" pattern. (There are also other constructs that can cause a pattern
753     to be anchored.)
754     </P>
755     <P>
756 nigel 75 A dollar character is an assertion that is true only if the current matching
757 nigel 63 point is at the end of the subject string, or immediately before a newline
758 nigel 91 at the end of the string (by default). Dollar need not be the last character of
759     the pattern if a number of alternatives are involved, but it should be the last
760     item in any branch in which it appears. Dollar has no special meaning in a
761     character class.
762 nigel 63 </P>
763     <P>
764     The meaning of dollar can be changed so that it matches only at the very end of
765     the string, by setting the PCRE_DOLLAR_ENDONLY option at compile time. This
766     does not affect the \Z assertion.
767     </P>
768     <P>
769     The meanings of the circumflex and dollar characters are changed if the
770 nigel 91 PCRE_MULTILINE option is set. When this is the case, a circumflex matches
771     immediately after internal newlines as well as at the start of the subject
772     string. It does not match after a newline that ends the string. A dollar
773     matches before any newlines in the string, as well as at the very end, when
774     PCRE_MULTILINE is set. When newline is specified as the two-character
775     sequence CRLF, isolated CR and LF characters do not indicate newlines.
776 nigel 63 </P>
777     <P>
778 nigel 91 For example, the pattern /^abc$/ matches the subject string "def\nabc" (where
779     \n represents a newline) in multiline mode, but not otherwise. Consequently,
780     patterns that are anchored in single line mode because all branches start with
781     ^ are not anchored in multiline mode, and a match for circumflex is possible
782     when the <i>startoffset</i> argument of <b>pcre_exec()</b> is non-zero. The
783     PCRE_DOLLAR_ENDONLY option is ignored if PCRE_MULTILINE is set.
784     </P>
785     <P>
786 nigel 63 Note that the sequences \A, \Z, and \z can be used to match the start and
787     end of the subject in both modes, and if all branches of a pattern start with
788 nigel 91 \A it is always anchored, whether or not PCRE_MULTILINE is set.
789 nigel 63 </P>
790 ph10 227 <br><a name="SEC6" href="#TOC1">FULL STOP (PERIOD, DOT)</a><br>
791 nigel 63 <P>
792     Outside a character class, a dot in the pattern matches any one character in
793 nigel 91 the subject string except (by default) a character that signifies the end of a
794 nigel 93 line. In UTF-8 mode, the matched character may be more than one byte long.
795 nigel 63 </P>
796 nigel 91 <P>
797 nigel 93 When a line ending is defined as a single character, dot never matches that
798     character; when the two-character sequence CRLF is used, dot does not match CR
799     if it is immediately followed by LF, but otherwise it matches all characters
800     (including isolated CRs and LFs). When any Unicode line endings are being
801     recognized, dot does not match CR or LF or any of the other line ending
802     characters.
803     </P>
804     <P>
805 nigel 91 The behaviour of dot with regard to newlines can be changed. If the PCRE_DOTALL
806 nigel 93 option is set, a dot matches any one character, without exception. If the
807     two-character sequence CRLF is present in the subject string, it takes two dots
808     to match it.
809 nigel 91 </P>
810     <P>
811     The handling of dot is entirely independent of the handling of circumflex and
812     dollar, the only relationship being that they both involve newlines. Dot has no
813     special meaning in a character class.
814     </P>
815 ph10 227 <br><a name="SEC7" href="#TOC1">MATCHING A SINGLE BYTE</a><br>
816 nigel 63 <P>
817     Outside a character class, the escape sequence \C matches any one byte, both
818 nigel 93 in and out of UTF-8 mode. Unlike a dot, it always matches any line-ending
819     characters. The feature is provided in Perl in order to match individual bytes
820     in UTF-8 mode. Because it breaks up UTF-8 characters into individual bytes,
821     what remains in the string may be a malformed UTF-8 string. For this reason,
822     the \C escape sequence is best avoided.
823 nigel 63 </P>
824     <P>
825 nigel 75 PCRE does not allow \C to appear in lookbehind assertions
826     <a href="#lookbehind">(described below),</a>
827     because in UTF-8 mode this would make it impossible to calculate the length of
828     the lookbehind.
829     <a name="characterclass"></a></P>
830 ph10 227 <br><a name="SEC8" href="#TOC1">SQUARE BRACKETS AND CHARACTER CLASSES</a><br>
831 nigel 63 <P>
832     An opening square bracket introduces a character class, terminated by a closing
833     square bracket. A closing square bracket on its own is not special. If a
834     closing square bracket is required as a member of the class, it should be the
835     first data character in the class (after an initial circumflex, if present) or
836     escaped with a backslash.
837     </P>
838     <P>
839     A character class matches a single character in the subject. In UTF-8 mode, the
840     character may occupy more than one byte. A matched character must be in the set
841     of characters defined by the class, unless the first character in the class
842     definition is a circumflex, in which case the subject character must not be in
843     the set defined by the class. If a circumflex is actually required as a member
844     of the class, ensure it is not the first character, or escape it with a
845     backslash.
846     </P>
847     <P>
848     For example, the character class [aeiou] matches any lower case vowel, while
849     [^aeiou] matches any character that is not a lower case vowel. Note that a
850 nigel 75 circumflex is just a convenient notation for specifying the characters that
851     are in the class by enumerating those that are not. A class that starts with a
852     circumflex is not an assertion: it still consumes a character from the subject
853     string, and therefore it fails if the current pointer is at the end of the
854     string.
855 nigel 63 </P>
856     <P>
857     In UTF-8 mode, characters with values greater than 255 can be included in a
858     class as a literal string of bytes, or by using the \x{ escaping mechanism.
859     </P>
860     <P>
861     When caseless matching is set, any letters in a class represent both their
862     upper case and lower case versions, so for example, a caseless [aeiou] matches
863     "A" as well as "a", and a caseless [^aeiou] does not match "A", whereas a
864 nigel 77 caseful version would. In UTF-8 mode, PCRE always understands the concept of
865     case for characters whose values are less than 128, so caseless matching is
866     always possible. For characters with higher values, the concept of case is
867     supported if PCRE is compiled with Unicode property support, but not otherwise.
868     If you want to use caseless matching for characters 128 and above, you must
869     ensure that PCRE is compiled with Unicode property support as well as with
870     UTF-8 support.
871 nigel 63 </P>
872     <P>
873 nigel 93 Characters that might indicate line breaks are never treated in any special way
874     when matching character classes, whatever line-ending sequence is in use, and
875     whatever setting of the PCRE_DOTALL and PCRE_MULTILINE options is used. A class
876     such as [^a] always matches one of these characters.
877 nigel 63 </P>
878     <P>
879     The minus (hyphen) character can be used to specify a range of characters in a
880     character class. For example, [d-m] matches any letter between d and m,
881     inclusive. If a minus character is required in a class, it must be escaped with
882     a backslash or appear in a position where it cannot be interpreted as
883     indicating a range, typically as the first or last character in the class.
884     </P>
885     <P>
886     It is not possible to have the literal character "]" as the end character of a
887     range. A pattern such as [W-]46] is interpreted as a class of two characters
888     ("W" and "-") followed by a literal string "46]", so it would match "W46]" or
889     "-46]". However, if the "]" is escaped with a backslash it is interpreted as
890 nigel 75 the end of range, so [W-\]46] is interpreted as a class containing a range
891     followed by two other characters. The octal or hexadecimal representation of
892     "]" can also be used to end a range.
893 nigel 63 </P>
894     <P>
895     Ranges operate in the collating sequence of character values. They can also be
896     used for characters specified numerically, for example [\000-\037]. In UTF-8
897     mode, ranges can include characters whose values are greater than 255, for
898     example [\x{100}-\x{2ff}].
899     </P>
900     <P>
901     If a range that includes letters is used when caseless matching is set, it
902     matches the letters in either case. For example, [W-c] is equivalent to
903 nigel 75 [][\\^_`wxyzabc], matched caselessly, and in non-UTF-8 mode, if character
904 ph10 142 tables for a French locale are in use, [\xc8-\xcb] matches accented E
905 nigel 75 characters in both cases. In UTF-8 mode, PCRE supports the concept of case for
906     characters with values greater than 128 only when it is compiled with Unicode
907     property support.
908 nigel 63 </P>
909     <P>
910 nigel 75 The character types \d, \D, \p, \P, \s, \S, \w, and \W may also appear
911     in a character class, and add the characters that they match to the class. For
912 nigel 63 example, [\dABCDEF] matches any hexadecimal digit. A circumflex can
913     conveniently be used with the upper case character types to specify a more
914     restricted set of characters than the matching lower case type. For example,
915     the class [^\W_] matches any letter or digit, but not underscore.
916     </P>
917     <P>
918 nigel 75 The only metacharacters that are recognized in character classes are backslash,
919     hyphen (only where it can be interpreted as specifying a range), circumflex
920     (only at the start), opening square bracket (only when it can be interpreted as
921     introducing a POSIX class name - see the next section), and the terminating
922     closing square bracket. However, escaping other non-alphanumeric characters
923     does no harm.
924 nigel 63 </P>
925 ph10 227 <br><a name="SEC9" href="#TOC1">POSIX CHARACTER CLASSES</a><br>
926 nigel 63 <P>
927 nigel 75 Perl supports the POSIX notation for character classes. This uses names
928 nigel 63 enclosed by [: and :] within the enclosing square brackets. PCRE also supports
929     this notation. For example,
930     <pre>
931     [01[:alpha:]%]
932 nigel 75 </pre>
933 nigel 63 matches "0", "1", any alphabetic character, or "%". The supported class names
934     are
935     <pre>
936     alnum letters and digits
937     alpha letters
938     ascii character codes 0 - 127
939     blank space or tab only
940     cntrl control characters
941     digit decimal digits (same as \d)
942     graph printing characters, excluding space
943     lower lower case letters
944     print printing characters, including space
945     punct printing characters, excluding letters and digits
946     space white space (not quite the same as \s)
947     upper upper case letters
948     word "word" characters (same as \w)
949     xdigit hexadecimal digits
950 nigel 75 </pre>
951 nigel 63 The "space" characters are HT (9), LF (10), VT (11), FF (12), CR (13), and
952     space (32). Notice that this list includes the VT character (code 11). This
953     makes "space" different to \s, which does not include VT (for Perl
954     compatibility).
955     </P>
956     <P>
957     The name "word" is a Perl extension, and "blank" is a GNU extension from Perl
958     5.8. Another Perl extension is negation, which is indicated by a ^ character
959     after the colon. For example,
960     <pre>
961     [12[:^digit:]]
962 nigel 75 </pre>
963 nigel 63 matches "1", "2", or any non-digit. PCRE (and Perl) also recognize the POSIX
964     syntax [.ch.] and [=ch=] where "ch" is a "collating element", but these are not
965     supported, and an error is given if they are encountered.
966     </P>
967     <P>
968 nigel 75 In UTF-8 mode, characters with values greater than 128 do not match any of
969 nigel 63 the POSIX character classes.
970     </P>
971 ph10 227 <br><a name="SEC10" href="#TOC1">VERTICAL BAR</a><br>
972 nigel 63 <P>
973     Vertical bar characters are used to separate alternative patterns. For example,
974     the pattern
975     <pre>
976     gilbert|sullivan
977 nigel 75 </pre>
978 nigel 63 matches either "gilbert" or "sullivan". Any number of alternatives may appear,
979 nigel 91 and an empty alternative is permitted (matching the empty string). The matching
980     process tries each alternative in turn, from left to right, and the first one
981     that succeeds is used. If the alternatives are within a subpattern
982 nigel 75 <a href="#subpattern">(defined below),</a>
983     "succeeds" means matching the rest of the main pattern as well as the
984     alternative in the subpattern.
985 nigel 63 </P>
986 ph10 227 <br><a name="SEC11" href="#TOC1">INTERNAL OPTION SETTING</a><br>
987 nigel 63 <P>
988     The settings of the PCRE_CASELESS, PCRE_MULTILINE, PCRE_DOTALL, and
989 ph10 243 PCRE_EXTENDED options (which are Perl-compatible) can be changed from within
990     the pattern by a sequence of Perl option letters enclosed between "(?" and ")".
991     The option letters are
992 nigel 63 <pre>
993     i for PCRE_CASELESS
994     m for PCRE_MULTILINE
995     s for PCRE_DOTALL
996     x for PCRE_EXTENDED
997 nigel 75 </pre>
998 nigel 63 For example, (?im) sets caseless, multiline matching. It is also possible to
999     unset these options by preceding the letter with a hyphen, and a combined
1000     setting and unsetting such as (?im-sx), which sets PCRE_CASELESS and
1001     PCRE_MULTILINE while unsetting PCRE_DOTALL and PCRE_EXTENDED, is also
1002     permitted. If a letter appears both before and after the hyphen, the option is
1003     unset.
1004     </P>
1005     <P>
1006 ph10 243 The PCRE-specific options PCRE_DUPNAMES, PCRE_UNGREEDY, and PCRE_EXTRA can be
1007     changed in the same way as the Perl-compatible options by using the characters
1008     J, U and X respectively.
1009     </P>
1010     <P>
1011 nigel 63 When an option change occurs at top level (that is, not inside subpattern
1012     parentheses), the change applies to the remainder of the pattern that follows.
1013     If the change is placed right at the start of a pattern, PCRE extracts it into
1014     the global options (and it will therefore show up in data extracted by the
1015     <b>pcre_fullinfo()</b> function).
1016     </P>
1017     <P>
1018 nigel 93 An option change within a subpattern (see below for a description of
1019     subpatterns) affects only that part of the current pattern that follows it, so
1020 nigel 63 <pre>
1021     (a(?i)b)c
1022 nigel 75 </pre>
1023 nigel 63 matches abc and aBc and no other strings (assuming PCRE_CASELESS is not used).
1024     By this means, options can be made to have different settings in different
1025     parts of the pattern. Any changes made in one alternative do carry on
1026     into subsequent branches within the same subpattern. For example,
1027     <pre>
1028     (a(?i)b|c)
1029 nigel 75 </pre>
1030 nigel 63 matches "ab", "aB", "c", and "C", even though when matching "C" the first
1031     branch is abandoned before the option setting. This is because the effects of
1032     option settings happen at compile time. There would be some very weird
1033     behaviour otherwise.
1034 nigel 75 <a name="subpattern"></a></P>
1035 ph10 227 <br><a name="SEC12" href="#TOC1">SUBPATTERNS</a><br>
1036 nigel 63 <P>
1037     Subpatterns are delimited by parentheses (round brackets), which can be nested.
1038 nigel 75 Turning part of a pattern into a subpattern does two things:
1039     <br>
1040     <br>
1041 nigel 63 1. It localizes a set of alternatives. For example, the pattern
1042     <pre>
1043     cat(aract|erpillar|)
1044 nigel 75 </pre>
1045 nigel 63 matches one of the words "cat", "cataract", or "caterpillar". Without the
1046 nigel 93 parentheses, it would match "cataract", "erpillar" or an empty string.
1047 nigel 75 <br>
1048     <br>
1049     2. It sets up the subpattern as a capturing subpattern. This means that, when
1050     the whole pattern matches, that portion of the subject string that matched the
1051     subpattern is passed back to the caller via the <i>ovector</i> argument of
1052 nigel 63 <b>pcre_exec()</b>. Opening parentheses are counted from left to right (starting
1053 nigel 75 from 1) to obtain numbers for the capturing subpatterns.
1054 nigel 63 </P>
1055     <P>
1056     For example, if the string "the red king" is matched against the pattern
1057     <pre>
1058     the ((red|white) (king|queen))
1059 nigel 75 </pre>
1060 nigel 63 the captured substrings are "red king", "red", and "king", and are numbered 1,
1061     2, and 3, respectively.
1062     </P>
1063     <P>
1064     The fact that plain parentheses fulfil two functions is not always helpful.
1065     There are often times when a grouping subpattern is required without a
1066     capturing requirement. If an opening parenthesis is followed by a question mark
1067     and a colon, the subpattern does not do any capturing, and is not counted when
1068     computing the number of any subsequent capturing subpatterns. For example, if
1069     the string "the white queen" is matched against the pattern
1070     <pre>
1071     the ((?:red|white) (king|queen))
1072 nigel 75 </pre>
1073 nigel 63 the captured substrings are "white queen" and "queen", and are numbered 1 and
1074 nigel 93 2. The maximum number of capturing subpatterns is 65535.
1075 nigel 63 </P>
1076     <P>
1077     As a convenient shorthand, if any option settings are required at the start of
1078     a non-capturing subpattern, the option letters may appear between the "?" and
1079     the ":". Thus the two patterns
1080     <pre>
1081     (?i:saturday|sunday)
1082     (?:(?i)saturday|sunday)
1083 nigel 75 </pre>
1084 nigel 63 match exactly the same set of strings. Because alternative branches are tried
1085     from left to right, and options are not reset until the end of the subpattern
1086     is reached, an option setting in one branch does affect subsequent branches, so
1087     the above patterns match "SUNDAY" as well as "Saturday".
1088     </P>
1089 ph10 227 <br><a name="SEC13" href="#TOC1">DUPLICATE SUBPATTERN NUMBERS</a><br>
1090 nigel 63 <P>
1091 ph10 182 Perl 5.10 introduced a feature whereby each alternative in a subpattern uses
1092     the same numbers for its capturing parentheses. Such a subpattern starts with
1093     (?| and is itself a non-capturing subpattern. For example, consider this
1094     pattern:
1095     <pre>
1096     (?|(Sat)ur|(Sun))day
1097     </pre>
1098     Because the two alternatives are inside a (?| group, both sets of capturing
1099     parentheses are numbered one. Thus, when the pattern matches, you can look
1100     at captured substring number one, whichever alternative matched. This construct
1101     is useful when you want to capture part, but not all, of one of a number of
1102     alternatives. Inside a (?| group, parentheses are numbered as usual, but the
1103     number is reset at the start of each branch. The numbers of any capturing
1104     buffers that follow the subpattern start after the highest number used in any
1105     branch. The following example is taken from the Perl documentation.
1106     The numbers underneath show in which buffer the captured content will be
1107     stored.
1108     <pre>
1109     # before ---------------branch-reset----------- after
1110     / ( a ) (?| x ( y ) z | (p (q) r) | (t) u (v) ) ( z ) /x
1111     # 1 2 2 3 2 3 4
1112     </pre>
1113     A backreference or a recursive call to a numbered subpattern always refers to
1114     the first one in the pattern with the given number.
1115     </P>
1116     <P>
1117     An alternative approach to using this "branch reset" feature is to use
1118     duplicate named subpatterns, as described in the next section.
1119     </P>
1120 ph10 227 <br><a name="SEC14" href="#TOC1">NAMED SUBPATTERNS</a><br>
1121 ph10 182 <P>
1122 nigel 63 Identifying capturing parentheses by number is simple, but it can be very hard
1123     to keep track of the numbers in complicated regular expressions. Furthermore,
1124 nigel 75 if an expression is modified, the numbers may change. To help with this
1125 nigel 93 difficulty, PCRE supports the naming of subpatterns. This feature was not
1126     added to Perl until release 5.10. Python had the feature earlier, and PCRE
1127     introduced it at release 4.0, using the Python syntax. PCRE now supports both
1128     the Perl and the Python syntax.
1129     </P>
1130     <P>
1131     In PCRE, a subpattern can be named in one of three ways: (?&#60;name&#62;...) or
1132     (?'name'...) as in Perl, or (?P&#60;name&#62;...) as in Python. References to capturing
1133 nigel 91 parentheses from other parts of the pattern, such as
1134     <a href="#backreferences">backreferences,</a>
1135     <a href="#recursion">recursion,</a>
1136     and
1137     <a href="#conditions">conditions,</a>
1138     can be made by name as well as by number.
1139 nigel 63 </P>
1140     <P>
1141 nigel 91 Names consist of up to 32 alphanumeric characters and underscores. Named
1142 nigel 93 capturing parentheses are still allocated numbers as well as names, exactly as
1143     if the names were not present. The PCRE API provides function calls for
1144     extracting the name-to-number translation table from a compiled pattern. There
1145     is also a convenience function for extracting a captured substring by name.
1146 nigel 91 </P>
1147     <P>
1148     By default, a name must be unique within a pattern, but it is possible to relax
1149     this constraint by setting the PCRE_DUPNAMES option at compile time. This can
1150     be useful for patterns where only one instance of the named parentheses can
1151     match. Suppose you want to match the name of a weekday, either as a 3-letter
1152     abbreviation or as the full name, and in both cases you want to extract the
1153     abbreviation. This pattern (ignoring the line breaks) does the job:
1154     <pre>
1155 nigel 93 (?&#60;DN&#62;Mon|Fri|Sun)(?:day)?|
1156     (?&#60;DN&#62;Tue)(?:sday)?|
1157     (?&#60;DN&#62;Wed)(?:nesday)?|
1158     (?&#60;DN&#62;Thu)(?:rsday)?|
1159     (?&#60;DN&#62;Sat)(?:urday)?
1160 nigel 91 </pre>
1161     There are five capturing substrings, but only one is ever set after a match.
1162 ph10 182 (An alternative way of solving this problem is to use a "branch reset"
1163     subpattern, as described in the previous section.)
1164     </P>
1165     <P>
1166 nigel 91 The convenience function for extracting the data by name returns the substring
1167 nigel 93 for the first (and in this example, the only) subpattern of that name that
1168 nigel 91 matched. This saves searching to find which numbered subpattern it was. If you
1169     make a reference to a non-unique named subpattern from elsewhere in the
1170     pattern, the one that corresponds to the lowest number is used. For further
1171     details of the interfaces for handling named subpatterns, see the
1172 nigel 63 <a href="pcreapi.html"><b>pcreapi</b></a>
1173     documentation.
1174     </P>
1175 ph10 227 <br><a name="SEC15" href="#TOC1">REPETITION</a><br>
1176 nigel 63 <P>
1177     Repetition is specified by quantifiers, which can follow any of the following
1178     items:
1179     <pre>
1180     a literal data character
1181 nigel 93 the dot metacharacter
1182 nigel 63 the \C escape sequence
1183 nigel 75 the \X escape sequence (in UTF-8 mode with Unicode properties)
1184 nigel 93 the \R escape sequence
1185 nigel 75 an escape such as \d that matches a single character
1186 nigel 63 a character class
1187     a back reference (see next section)
1188     a parenthesized subpattern (unless it is an assertion)
1189 nigel 75 </pre>
1190 nigel 63 The general repetition quantifier specifies a minimum and maximum number of
1191     permitted matches, by giving the two numbers in curly brackets (braces),
1192     separated by a comma. The numbers must be less than 65536, and the first must
1193     be less than or equal to the second. For example:
1194     <pre>
1195     z{2,4}
1196 nigel 75 </pre>
1197 nigel 63 matches "zz", "zzz", or "zzzz". A closing brace on its own is not a special
1198     character. If the second number is omitted, but the comma is present, there is
1199     no upper limit; if the second number and the comma are both omitted, the
1200     quantifier specifies an exact number of required matches. Thus
1201     <pre>
1202     [aeiou]{3,}
1203 nigel 75 </pre>
1204 nigel 63 matches at least 3 successive vowels, but may match many more, while
1205     <pre>
1206     \d{8}
1207 nigel 75 </pre>
1208 nigel 63 matches exactly 8 digits. An opening curly bracket that appears in a position
1209     where a quantifier is not allowed, or one that does not match the syntax of a
1210     quantifier, is taken as a literal character. For example, {,6} is not a
1211     quantifier, but a literal string of four characters.
1212     </P>
1213     <P>
1214     In UTF-8 mode, quantifiers apply to UTF-8 characters rather than to individual
1215     bytes. Thus, for example, \x{100}{2} matches two UTF-8 characters, each of
1216 nigel 75 which is represented by a two-byte sequence. Similarly, when Unicode property
1217     support is available, \X{3} matches three Unicode extended sequences, each of
1218     which may be several bytes long (and they may be of different lengths).
1219 nigel 63 </P>
1220     <P>
1221     The quantifier {0} is permitted, causing the expression to behave as if the
1222     previous item and the quantifier were not present.
1223     </P>
1224     <P>
1225 nigel 93 For convenience, the three most common quantifiers have single-character
1226     abbreviations:
1227 nigel 63 <pre>
1228     * is equivalent to {0,}
1229     + is equivalent to {1,}
1230     ? is equivalent to {0,1}
1231 nigel 75 </pre>
1232 nigel 63 It is possible to construct infinite loops by following a subpattern that can
1233     match no characters with a quantifier that has no upper limit, for example:
1234     <pre>
1235     (a?)*
1236 nigel 75 </pre>
1237 nigel 63 Earlier versions of Perl and PCRE used to give an error at compile time for
1238     such patterns. However, because there are cases where this can be useful, such
1239     patterns are now accepted, but if any repetition of the subpattern does in fact
1240     match no characters, the loop is forcibly broken.
1241     </P>
1242     <P>
1243     By default, the quantifiers are "greedy", that is, they match as much as
1244     possible (up to the maximum number of permitted times), without causing the
1245     rest of the pattern to fail. The classic example of where this gives problems
1246 nigel 75 is in trying to match comments in C programs. These appear between /* and */
1247     and within the comment, individual * and / characters may appear. An attempt to
1248     match C comments by applying the pattern
1249 nigel 63 <pre>
1250     /\*.*\*/
1251 nigel 75 </pre>
1252 nigel 63 to the string
1253     <pre>
1254 nigel 75 /* first comment */ not comment /* second comment */
1255     </pre>
1256 nigel 63 fails, because it matches the entire string owing to the greediness of the .*
1257     item.
1258     </P>
1259     <P>
1260     However, if a quantifier is followed by a question mark, it ceases to be
1261     greedy, and instead matches the minimum number of times possible, so the
1262     pattern
1263     <pre>
1264     /\*.*?\*/
1265 nigel 75 </pre>
1266 nigel 63 does the right thing with the C comments. The meaning of the various
1267     quantifiers is not otherwise changed, just the preferred number of matches.
1268     Do not confuse this use of question mark with its use as a quantifier in its
1269     own right. Because it has two uses, it can sometimes appear doubled, as in
1270     <pre>
1271     \d??\d
1272 nigel 75 </pre>
1273 nigel 63 which matches one digit by preference, but can match two if that is the only
1274     way the rest of the pattern matches.
1275     </P>
1276     <P>
1277 nigel 93 If the PCRE_UNGREEDY option is set (an option that is not available in Perl),
1278 nigel 63 the quantifiers are not greedy by default, but individual ones can be made
1279     greedy by following them with a question mark. In other words, it inverts the
1280     default behaviour.
1281     </P>
1282     <P>
1283     When a parenthesized subpattern is quantified with a minimum repeat count that
1284 nigel 75 is greater than 1 or with a limited maximum, more memory is required for the
1285 nigel 63 compiled pattern, in proportion to the size of the minimum or maximum.
1286     </P>
1287     <P>
1288     If a pattern starts with .* or .{0,} and the PCRE_DOTALL option (equivalent
1289 nigel 93 to Perl's /s) is set, thus allowing the dot to match newlines, the pattern is
1290 nigel 63 implicitly anchored, because whatever follows will be tried against every
1291     character position in the subject string, so there is no point in retrying the
1292     overall match at any position after the first. PCRE normally treats such a
1293     pattern as though it were preceded by \A.
1294     </P>
1295     <P>
1296     In cases where it is known that the subject string contains no newlines, it is
1297     worth setting PCRE_DOTALL in order to obtain this optimization, or
1298     alternatively using ^ to indicate anchoring explicitly.
1299     </P>
1300     <P>
1301     However, there is one situation where the optimization cannot be used. When .*
1302     is inside capturing parentheses that are the subject of a backreference
1303 nigel 93 elsewhere in the pattern, a match at the start may fail where a later one
1304     succeeds. Consider, for example:
1305 nigel 63 <pre>
1306     (.*)abc\1
1307 nigel 75 </pre>
1308 nigel 63 If the subject is "xyz123abc123" the match point is the fourth character. For
1309     this reason, such a pattern is not implicitly anchored.
1310     </P>
1311     <P>
1312     When a capturing subpattern is repeated, the value captured is the substring
1313     that matched the final iteration. For example, after
1314     <pre>
1315     (tweedle[dume]{3}\s*)+
1316 nigel 75 </pre>
1317 nigel 63 has matched "tweedledum tweedledee" the value of the captured substring is
1318     "tweedledee". However, if there are nested capturing subpatterns, the
1319     corresponding captured values may have been set in previous iterations. For
1320     example, after
1321     <pre>
1322     /(a|(b))+/
1323 nigel 75 </pre>
1324 nigel 63 matches "aba" the value of the second captured substring is "b".
1325 nigel 75 <a name="atomicgroup"></a></P>
1326 ph10 227 <br><a name="SEC16" href="#TOC1">ATOMIC GROUPING AND POSSESSIVE QUANTIFIERS</a><br>
1327 nigel 63 <P>
1328 nigel 93 With both maximizing ("greedy") and minimizing ("ungreedy" or "lazy")
1329     repetition, failure of what follows normally causes the repeated item to be
1330     re-evaluated to see if a different number of repeats allows the rest of the
1331     pattern to match. Sometimes it is useful to prevent this, either to change the
1332     nature of the match, or to cause it fail earlier than it otherwise might, when
1333     the author of the pattern knows there is no point in carrying on.
1334 nigel 63 </P>
1335     <P>
1336     Consider, for example, the pattern \d+foo when applied to the subject line
1337     <pre>
1338     123456bar
1339 nigel 75 </pre>
1340 nigel 63 After matching all 6 digits and then failing to match "foo", the normal
1341     action of the matcher is to try again with only 5 digits matching the \d+
1342     item, and then with 4, and so on, before ultimately failing. "Atomic grouping"
1343     (a term taken from Jeffrey Friedl's book) provides the means for specifying
1344     that once a subpattern has matched, it is not to be re-evaluated in this way.
1345     </P>
1346     <P>
1347 nigel 93 If we use atomic grouping for the previous example, the matcher gives up
1348 nigel 63 immediately on failing to match "foo" the first time. The notation is a kind of
1349     special parenthesis, starting with (?&#62; as in this example:
1350     <pre>
1351 nigel 73 (?&#62;\d+)foo
1352 nigel 75 </pre>
1353 nigel 63 This kind of parenthesis "locks up" the part of the pattern it contains once
1354     it has matched, and a failure further into the pattern is prevented from
1355     backtracking into it. Backtracking past it to previous items, however, works as
1356     normal.
1357     </P>
1358     <P>
1359     An alternative description is that a subpattern of this type matches the string
1360     of characters that an identical standalone pattern would match, if anchored at
1361     the current point in the subject string.
1362     </P>
1363     <P>
1364     Atomic grouping subpatterns are not capturing subpatterns. Simple cases such as
1365     the above example can be thought of as a maximizing repeat that must swallow
1366     everything it can. So, while both \d+ and \d+? are prepared to adjust the
1367     number of digits they match in order to make the rest of the pattern match,
1368     (?&#62;\d+) can only match an entire sequence of digits.
1369     </P>
1370     <P>
1371     Atomic groups in general can of course contain arbitrarily complicated
1372     subpatterns, and can be nested. However, when the subpattern for an atomic
1373     group is just a single repeated item, as in the example above, a simpler
1374     notation, called a "possessive quantifier" can be used. This consists of an
1375     additional + character following a quantifier. Using this notation, the
1376     previous example can be rewritten as
1377     <pre>
1378 nigel 75 \d++foo
1379     </pre>
1380 ph10 208 Note that a possessive quantifier can be used with an entire group, for
1381     example:
1382     <pre>
1383     (abc|xyz){2,3}+
1384     </pre>
1385 nigel 63 Possessive quantifiers are always greedy; the setting of the PCRE_UNGREEDY
1386     option is ignored. They are a convenient notation for the simpler forms of
1387 nigel 93 atomic group. However, there is no difference in the meaning of a possessive
1388     quantifier and the equivalent atomic group, though there may be a performance
1389     difference; possessive quantifiers should be slightly faster.
1390 nigel 63 </P>
1391     <P>
1392 nigel 93 The possessive quantifier syntax is an extension to the Perl 5.8 syntax.
1393     Jeffrey Friedl originated the idea (and the name) in the first edition of his
1394     book. Mike McCloskey liked it, so implemented it when he built Sun's Java
1395     package, and PCRE copied it from there. It ultimately found its way into Perl
1396     at release 5.10.
1397 nigel 63 </P>
1398     <P>
1399 nigel 93 PCRE has an optimization that automatically "possessifies" certain simple
1400     pattern constructs. For example, the sequence A+B is treated as A++B because
1401     there is no point in backtracking into a sequence of A's when B must follow.
1402     </P>
1403     <P>
1404 nigel 63 When a pattern contains an unlimited repeat inside a subpattern that can itself
1405     be repeated an unlimited number of times, the use of an atomic group is the
1406     only way to avoid some failing matches taking a very long time indeed. The
1407     pattern
1408     <pre>
1409     (\D+|&#60;\d+&#62;)*[!?]
1410 nigel 75 </pre>
1411 nigel 63 matches an unlimited number of substrings that either consist of non-digits, or
1412     digits enclosed in &#60;&#62;, followed by either ! or ?. When it matches, it runs
1413     quickly. However, if it is applied to
1414     <pre>
1415     aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
1416 nigel 75 </pre>
1417 nigel 63 it takes a long time before reporting failure. This is because the string can
1418 nigel 75 be divided between the internal \D+ repeat and the external * repeat in a
1419     large number of ways, and all have to be tried. (The example uses [!?] rather
1420     than a single character at the end, because both PCRE and Perl have an
1421     optimization that allows for fast failure when a single character is used. They
1422     remember the last single character that is required for a match, and fail early
1423     if it is not present in the string.) If the pattern is changed so that it uses
1424     an atomic group, like this:
1425 nigel 63 <pre>
1426     ((?&#62;\D+)|&#60;\d+&#62;)*[!?]
1427 nigel 75 </pre>
1428 nigel 63 sequences of non-digits cannot be broken, and failure happens quickly.
1429 nigel 75 <a name="backreferences"></a></P>
1430 ph10 227 <br><a name="SEC17" href="#TOC1">BACK REFERENCES</a><br>
1431 nigel 63 <P>
1432     Outside a character class, a backslash followed by a digit greater than 0 (and
1433     possibly further digits) is a back reference to a capturing subpattern earlier
1434     (that is, to its left) in the pattern, provided there have been that many
1435     previous capturing left parentheses.
1436     </P>
1437     <P>
1438     However, if the decimal number following the backslash is less than 10, it is
1439     always taken as a back reference, and causes an error only if there are not
1440     that many capturing left parentheses in the entire pattern. In other words, the
1441     parentheses that are referenced need not be to the left of the reference for
1442 nigel 91 numbers less than 10. A "forward back reference" of this type can make sense
1443     when a repetition is involved and the subpattern to the right has participated
1444     in an earlier iteration.
1445     </P>
1446     <P>
1447 nigel 93 It is not possible to have a numerical "forward back reference" to a subpattern
1448     whose number is 10 or more using this syntax because a sequence such as \50 is
1449     interpreted as a character defined in octal. See the subsection entitled
1450 nigel 91 "Non-printing characters"
1451 nigel 75 <a href="#digitsafterbackslash">above</a>
1452 nigel 93 for further details of the handling of digits following a backslash. There is
1453     no such problem when named parentheses are used. A back reference to any
1454     subpattern is possible using named parentheses (see below).
1455 nigel 63 </P>
1456     <P>
1457 nigel 93 Another way of avoiding the ambiguity inherent in the use of digits following a
1458     backslash is to use the \g escape sequence, which is a feature introduced in
1459 ph10 208 Perl 5.10. This escape must be followed by an unsigned number or a negative
1460     number, optionally enclosed in braces. These examples are all identical:
1461 nigel 93 <pre>
1462     (ring), \1
1463     (ring), \g1
1464     (ring), \g{1}
1465     </pre>
1466 ph10 208 An unsigned number specifies an absolute reference without the ambiguity that
1467     is present in the older syntax. It is also useful when literal digits follow
1468     the reference. A negative number is a relative reference. Consider this
1469     example:
1470 nigel 93 <pre>
1471     (abc(def)ghi)\g{-1}
1472     </pre>
1473     The sequence \g{-1} is a reference to the most recently started capturing
1474     subpattern before \g, that is, is it equivalent to \2. Similarly, \g{-2}
1475     would be equivalent to \1. The use of relative references can be helpful in
1476     long patterns, and also in patterns that are created by joining together
1477     fragments that contain references within themselves.
1478     </P>
1479     <P>
1480 nigel 63 A back reference matches whatever actually matched the capturing subpattern in
1481     the current subject string, rather than anything matching the subpattern
1482     itself (see
1483     <a href="#subpatternsassubroutines">"Subpatterns as subroutines"</a>
1484     below for a way of doing that). So the pattern
1485     <pre>
1486     (sens|respons)e and \1ibility
1487 nigel 75 </pre>
1488 nigel 63 matches "sense and sensibility" and "response and responsibility", but not
1489     "sense and responsibility". If caseful matching is in force at the time of the
1490     back reference, the case of letters is relevant. For example,
1491     <pre>
1492     ((?i)rah)\s+\1
1493 nigel 75 </pre>
1494 nigel 63 matches "rah rah" and "RAH RAH", but not "RAH rah", even though the original
1495     capturing subpattern is matched caselessly.
1496     </P>
1497     <P>
1498 ph10 172 There are several different ways of writing back references to named
1499     subpatterns. The .NET syntax \k{name} and the Perl syntax \k&#60;name&#62; or
1500     \k'name' are supported, as is the Python syntax (?P=name). Perl 5.10's unified
1501     back reference syntax, in which \g can be used for both numeric and named
1502     references, is also supported. We could rewrite the above example in any of
1503 nigel 93 the following ways:
1504 nigel 63 <pre>
1505 nigel 93 (?&#60;p1&#62;(?i)rah)\s+\k&#60;p1&#62;
1506 ph10 172 (?'p1'(?i)rah)\s+\k{p1}
1507 nigel 91 (?P&#60;p1&#62;(?i)rah)\s+(?P=p1)
1508 ph10 172 (?&#60;p1&#62;(?i)rah)\s+\g{p1}
1509 nigel 75 </pre>
1510 nigel 91 A subpattern that is referenced by name may appear in the pattern before or
1511     after the reference.
1512     </P>
1513     <P>
1514 nigel 63 There may be more than one back reference to the same subpattern. If a
1515     subpattern has not actually been used in a particular match, any back
1516     references to it always fail. For example, the pattern
1517     <pre>
1518     (a|(bc))\2
1519 nigel 75 </pre>
1520 nigel 63 always fails if it starts to match "a" rather than "bc". Because there may be
1521     many capturing parentheses in a pattern, all digits following the backslash are
1522     taken as part of a potential back reference number. If the pattern continues
1523     with a digit character, some delimiter must be used to terminate the back
1524     reference. If the PCRE_EXTENDED option is set, this can be whitespace.
1525 nigel 75 Otherwise an empty comment (see
1526     <a href="#comments">"Comments"</a>
1527     below) can be used.
1528 nigel 63 </P>
1529     <P>
1530     A back reference that occurs inside the parentheses to which it refers fails
1531     when the subpattern is first used, so, for example, (a\1) never matches.
1532     However, such references can be useful inside repeated subpatterns. For
1533     example, the pattern
1534     <pre>
1535     (a|b\1)+
1536 nigel 75 </pre>
1537 nigel 63 matches any number of "a"s and also "aba", "ababbaa" etc. At each iteration of
1538     the subpattern, the back reference matches the character string corresponding
1539     to the previous iteration. In order for this to work, the pattern must be such
1540     that the first iteration does not need to match the back reference. This can be
1541     done using alternation, as in the example above, or by a quantifier with a
1542     minimum of zero.
1543 nigel 75 <a name="bigassertions"></a></P>
1544 ph10 227 <br><a name="SEC18" href="#TOC1">ASSERTIONS</a><br>
1545 nigel 63 <P>
1546     An assertion is a test on the characters following or preceding the current
1547     matching point that does not actually consume any characters. The simple
1548 nigel 75 assertions coded as \b, \B, \A, \G, \Z, \z, ^ and $ are described
1549     <a href="#smallassertions">above.</a>
1550     </P>
1551     <P>
1552 nigel 63 More complicated assertions are coded as subpatterns. There are two kinds:
1553     those that look ahead of the current position in the subject string, and those
1554 nigel 75 that look behind it. An assertion subpattern is matched in the normal way,
1555     except that it does not cause the current matching position to be changed.
1556 nigel 63 </P>
1557     <P>
1558 nigel 75 Assertion subpatterns are not capturing subpatterns, and may not be repeated,
1559     because it makes no sense to assert the same thing several times. If any kind
1560     of assertion contains capturing subpatterns within it, these are counted for
1561     the purposes of numbering the capturing subpatterns in the whole pattern.
1562     However, substring capturing is carried out only for positive assertions,
1563     because it does not make sense for negative assertions.
1564 nigel 63 </P>
1565 nigel 75 <br><b>
1566     Lookahead assertions
1567     </b><br>
1568 nigel 63 <P>
1569 nigel 91 Lookahead assertions start with (?= for positive assertions and (?! for
1570     negative assertions. For example,
1571 nigel 63 <pre>
1572     \w+(?=;)
1573 nigel 75 </pre>
1574 nigel 63 matches a word followed by a semicolon, but does not include the semicolon in
1575     the match, and
1576     <pre>
1577     foo(?!bar)
1578 nigel 75 </pre>
1579 nigel 63 matches any occurrence of "foo" that is not followed by "bar". Note that the
1580     apparently similar pattern
1581     <pre>
1582     (?!foo)bar
1583 nigel 75 </pre>
1584 nigel 63 does not find an occurrence of "bar" that is preceded by something other than
1585     "foo"; it finds any occurrence of "bar" whatsoever, because the assertion
1586     (?!foo) is always true when the next three characters are "bar". A
1587 nigel 75 lookbehind assertion is needed to achieve the other effect.
1588 nigel 63 </P>
1589     <P>
1590     If you want to force a matching failure at some point in a pattern, the most
1591     convenient way to do it is with (?!) because an empty string always matches, so
1592     an assertion that requires there not to be an empty string must always fail.
1593 nigel 75 <a name="lookbehind"></a></P>
1594     <br><b>
1595     Lookbehind assertions
1596     </b><br>
1597 nigel 63 <P>
1598     Lookbehind assertions start with (?&#60;= for positive assertions and (?&#60;! for
1599     negative assertions. For example,
1600     <pre>
1601     (?&#60;!foo)bar
1602 nigel 75 </pre>
1603 nigel 63 does find an occurrence of "bar" that is not preceded by "foo". The contents of
1604     a lookbehind assertion are restricted such that all the strings it matches must
1605 nigel 91 have a fixed length. However, if there are several top-level alternatives, they
1606     do not all have to have the same fixed length. Thus
1607 nigel 63 <pre>
1608     (?&#60;=bullock|donkey)
1609 nigel 75 </pre>
1610 nigel 63 is permitted, but
1611     <pre>
1612     (?&#60;!dogs?|cats?)
1613 nigel 75 </pre>
1614 nigel 63 causes an error at compile time. Branches that match different length strings
1615     are permitted only at the top level of a lookbehind assertion. This is an
1616     extension compared with Perl (at least for 5.8), which requires all branches to
1617     match the same length of string. An assertion such as
1618     <pre>
1619     (?&#60;=ab(c|de))
1620 nigel 75 </pre>
1621 nigel 63 is not permitted, because its single top-level branch can match two different
1622     lengths, but it is acceptable if rewritten to use two top-level branches:
1623     <pre>
1624     (?&#60;=abc|abde)
1625 nigel 75 </pre>
1626 ph10 172 In some cases, the Perl 5.10 escape sequence \K
1627     <a href="#resetmatchstart">(see above)</a>
1628     can be used instead of a lookbehind assertion; this is not restricted to a
1629     fixed-length.
1630     </P>
1631     <P>
1632 nigel 63 The implementation of lookbehind assertions is, for each alternative, to
1633 nigel 93 temporarily move the current position back by the fixed length and then try to
1634 nigel 63 match. If there are insufficient characters before the current position, the
1635 nigel 93 assertion fails.
1636 nigel 63 </P>
1637     <P>
1638     PCRE does not allow the \C escape (which matches a single byte in UTF-8 mode)
1639     to appear in lookbehind assertions, because it makes it impossible to calculate
1640 nigel 93 the length of the lookbehind. The \X and \R escapes, which can match
1641     different numbers of bytes, are also not permitted.
1642 nigel 63 </P>
1643     <P>
1644 nigel 93 Possessive quantifiers can be used in conjunction with lookbehind assertions to
1645     specify efficient matching at the end of the subject string. Consider a simple
1646     pattern such as
1647 nigel 63 <pre>
1648     abcd$
1649 nigel 75 </pre>
1650 nigel 63 when applied to a long string that does not match. Because matching proceeds
1651     from left to right, PCRE will look for each "a" in the subject and then see if
1652     what follows matches the rest of the pattern. If the pattern is specified as
1653     <pre>
1654     ^.*abcd$
1655 nigel 75 </pre>
1656 nigel 63 the initial .* matches the entire string at first, but when this fails (because
1657     there is no following "a"), it backtracks to match all but the last character,
1658     then all but the last two characters, and so on. Once again the search for "a"
1659     covers the entire string, from right to left, so we are no better off. However,
1660     if the pattern is written as
1661     <pre>
1662     ^.*+(?&#60;=abcd)
1663 nigel 75 </pre>
1664 nigel 93 there can be no backtracking for the .*+ item; it can match only the entire
1665 nigel 63 string. The subsequent lookbehind assertion does a single test on the last four
1666     characters. If it fails, the match fails immediately. For long strings, this
1667     approach makes a significant difference to the processing time.
1668     </P>
1669 nigel 75 <br><b>
1670     Using multiple assertions
1671     </b><br>
1672 nigel 63 <P>
1673     Several assertions (of any sort) may occur in succession. For example,
1674     <pre>
1675     (?&#60;=\d{3})(?&#60;!999)foo
1676 nigel 75 </pre>
1677 nigel 63 matches "foo" preceded by three digits that are not "999". Notice that each of
1678     the assertions is applied independently at the same point in the subject
1679     string. First there is a check that the previous three characters are all
1680     digits, and then there is a check that the same three characters are not "999".
1681     This pattern does <i>not</i> match "foo" preceded by six characters, the first
1682     of which are digits and the last three of which are not "999". For example, it
1683     doesn't match "123abcfoo". A pattern to do that is
1684     <pre>
1685     (?&#60;=\d{3}...)(?&#60;!999)foo
1686 nigel 75 </pre>
1687 nigel 63 This time the first assertion looks at the preceding six characters, checking
1688     that the first three are digits, and then the second assertion checks that the
1689     preceding three characters are not "999".
1690     </P>
1691     <P>
1692     Assertions can be nested in any combination. For example,
1693     <pre>
1694     (?&#60;=(?&#60;!foo)bar)baz
1695 nigel 75 </pre>
1696 nigel 63 matches an occurrence of "baz" that is preceded by "bar" which in turn is not
1697     preceded by "foo", while
1698     <pre>
1699     (?&#60;=\d{3}(?!999)...)foo
1700 nigel 75 </pre>
1701     is another pattern that matches "foo" preceded by three digits and any three
1702 nigel 63 characters that are not "999".
1703 nigel 91 <a name="conditions"></a></P>
1704 ph10 227 <br><a name="SEC19" href="#TOC1">CONDITIONAL SUBPATTERNS</a><br>
1705 nigel 63 <P>
1706     It is possible to cause the matching process to obey a subpattern
1707     conditionally or to choose between two alternative subpatterns, depending on
1708     the result of an assertion, or whether a previous capturing subpattern matched
1709     or not. The two possible forms of conditional subpattern are
1710     <pre>
1711     (?(condition)yes-pattern)
1712     (?(condition)yes-pattern|no-pattern)
1713 nigel 75 </pre>
1714 nigel 63 If the condition is satisfied, the yes-pattern is used; otherwise the
1715     no-pattern (if present) is used. If there are more than two alternatives in the
1716     subpattern, a compile-time error occurs.
1717     </P>
1718     <P>
1719 nigel 93 There are four kinds of condition: references to subpatterns, references to
1720     recursion, a pseudo-condition called DEFINE, and assertions.
1721 nigel 91 </P>
1722 nigel 93 <br><b>
1723     Checking for a used subpattern by number
1724     </b><br>
1725 nigel 91 <P>
1726 nigel 93 If the text between the parentheses consists of a sequence of digits, the
1727     condition is true if the capturing subpattern of that number has previously
1728 ph10 172 matched. An alternative notation is to precede the digits with a plus or minus
1729     sign. In this case, the subpattern number is relative rather than absolute.
1730     The most recently opened parentheses can be referenced by (?(-1), the next most
1731     recent by (?(-2), and so on. In looping constructs it can also make sense to
1732     refer to subsequent groups with constructs such as (?(+2).
1733 nigel 93 </P>
1734     <P>
1735 nigel 91 Consider the following pattern, which contains non-significant white space to
1736     make it more readable (assume the PCRE_EXTENDED option) and to divide it into
1737     three parts for ease of discussion:
1738 nigel 63 <pre>
1739     ( \( )? [^()]+ (?(1) \) )
1740 nigel 75 </pre>
1741 nigel 63 The first part matches an optional opening parenthesis, and if that
1742     character is present, sets it as the first captured substring. The second part
1743     matches one or more characters that are not parentheses. The third part is a
1744     conditional subpattern that tests whether the first set of parentheses matched
1745     or not. If they did, that is, if subject started with an opening parenthesis,
1746     the condition is true, and so the yes-pattern is executed and a closing
1747     parenthesis is required. Otherwise, since no-pattern is not present, the
1748     subpattern matches nothing. In other words, this pattern matches a sequence of
1749 nigel 93 non-parentheses, optionally enclosed in parentheses.
1750     </P>
1751 ph10 172 <P>
1752     If you were embedding this pattern in a larger one, you could use a relative
1753     reference:
1754     <pre>
1755     ...other stuff... ( \( )? [^()]+ (?(-1) \) ) ...
1756     </pre>
1757     This makes the fragment independent of the parentheses in the larger pattern.
1758     </P>
1759 nigel 93 <br><b>
1760     Checking for a used subpattern by name
1761     </b><br>
1762     <P>
1763     Perl uses the syntax (?(&#60;name&#62;)...) or (?('name')...) to test for a used
1764     subpattern by name. For compatibility with earlier versions of PCRE, which had
1765     this facility before Perl, the syntax (?(name)...) is also recognized. However,
1766     there is a possible ambiguity with this syntax, because subpattern names may
1767     consist entirely of digits. PCRE looks first for a named subpattern; if it
1768     cannot find one and the name consists entirely of digits, PCRE looks for a
1769     subpattern of that number, which must be greater than zero. Using subpattern
1770     names that consist entirely of digits is not recommended.
1771     </P>
1772     <P>
1773     Rewriting the above example to use a named subpattern gives this:
1774 nigel 91 <pre>
1775 nigel 93 (?&#60;OPEN&#62; \( )? [^()]+ (?(&#60;OPEN&#62;) \) )
1776    
1777     </PRE>
1778     </P>
1779     <br><b>
1780     Checking for pattern recursion
1781     </b><br>
1782     <P>
1783     If the condition is the string (R), and there is no subpattern with the name R,
1784     the condition is true if a recursive call to the whole pattern or any
1785     subpattern has been made. If digits or a name preceded by ampersand follow the
1786     letter R, for example:
1787     <pre>
1788     (?(R3)...) or (?(R&name)...)
1789 nigel 91 </pre>
1790 nigel 93 the condition is true if the most recent recursion is into the subpattern whose
1791     number or name is given. This condition does not check the entire recursion
1792     stack.
1793 nigel 63 </P>
1794     <P>
1795 nigel 93 At "top level", all these recursion test conditions are false. Recursive
1796     patterns are described below.
1797     </P>
1798     <br><b>
1799     Defining subpatterns for use by reference only
1800     </b><br>
1801     <P>
1802     If the condition is the string (DEFINE), and there is no subpattern with the
1803     name DEFINE, the condition is always false. In this case, there may be only one
1804     alternative in the subpattern. It is always skipped if control reaches this
1805     point in the pattern; the idea of DEFINE is that it can be used to define
1806     "subroutines" that can be referenced from elsewhere. (The use of "subroutines"
1807     is described below.) For example, a pattern to match an IPv4 address could be
1808     written like this (ignore whitespace and line breaks):
1809     <pre>
1810     (?(DEFINE) (?&#60;byte&#62; 2[0-4]\d | 25[0-5] | 1\d\d | [1-9]?\d) )
1811     \b (?&byte) (\.(?&byte)){3} \b
1812     </pre>
1813     The first part of the pattern is a DEFINE group inside which a another group
1814     named "byte" is defined. This matches an individual component of an IPv4
1815     address (a number less than 256). When matching takes place, this part of the
1816     pattern is skipped because DEFINE acts like a false condition.
1817     </P>
1818     <P>
1819     The rest of the pattern uses references to the named group to match the four
1820     dot-separated components of an IPv4 address, insisting on a word boundary at
1821     each end.
1822     </P>
1823     <br><b>
1824     Assertion conditions
1825     </b><br>
1826     <P>
1827     If the condition is not in any of the above formats, it must be an assertion.
1828 nigel 63 This may be a positive or negative lookahead or lookbehind assertion. Consider
1829     this pattern, again containing non-significant white space, and with the two
1830     alternatives on the second line:
1831     <pre>
1832     (?(?=[^a-z]*[a-z])
1833     \d{2}-[a-z]{3}-\d{2} | \d{2}-\d{2}-\d{2} )
1834 nigel 75 </pre>
1835 nigel 63 The condition is a positive lookahead assertion that matches an optional
1836     sequence of non-letters followed by a letter. In other words, it tests for the
1837     presence of at least one letter in the subject. If a letter is found, the
1838     subject is matched against the first alternative; otherwise it is matched
1839     against the second. This pattern matches strings in one of the two forms
1840     dd-aaa-dd or dd-dd-dd, where aaa are letters and dd are digits.
1841 nigel 75 <a name="comments"></a></P>
1842 ph10 227 <br><a name="SEC20" href="#TOC1">COMMENTS</a><br>
1843 nigel 63 <P>
1844 nigel 75 The sequence (?# marks the start of a comment that continues up to the next
1845 nigel 63 closing parenthesis. Nested parentheses are not permitted. The characters
1846     that make up a comment play no part in the pattern matching at all.
1847     </P>
1848     <P>
1849     If the PCRE_EXTENDED option is set, an unescaped # character outside a
1850 nigel 91 character class introduces a comment that continues to immediately after the
1851     next newline in the pattern.
1852     <a name="recursion"></a></P>
1853 ph10 227 <br><a name="SEC21" href="#TOC1">RECURSIVE PATTERNS</a><br>
1854 nigel 63 <P>
1855     Consider the problem of matching a string in parentheses, allowing for
1856     unlimited nested parentheses. Without the use of recursion, the best that can
1857     be done is to use a pattern that matches up to some fixed depth of nesting. It
1858 nigel 93 is not possible to handle an arbitrary nesting depth.
1859     </P>
1860     <P>
1861     For some time, Perl has provided a facility that allows regular expressions to
1862     recurse (amongst other things). It does this by interpolating Perl code in the
1863     expression at run time, and the code can refer to the expression itself. A Perl
1864     pattern using code interpolation to solve the parentheses problem can be
1865     created like this:
1866 nigel 63 <pre>
1867     $re = qr{\( (?: (?&#62;[^()]+) | (?p{$re}) )* \)}x;
1868 nigel 75 </pre>
1869 nigel 63 The (?p{...}) item interpolates Perl code at run time, and in this case refers
1870 nigel 93 recursively to the pattern in which it appears.
1871 nigel 63 </P>
1872     <P>
1873 nigel 93 Obviously, PCRE cannot support the interpolation of Perl code. Instead, it
1874     supports special syntax for recursion of the entire pattern, and also for
1875     individual subpattern recursion. After its introduction in PCRE and Python,
1876     this kind of recursion was introduced into Perl at release 5.10.
1877 nigel 63 </P>
1878     <P>
1879 nigel 93 A special item that consists of (? followed by a number greater than zero and a
1880     closing parenthesis is a recursive call of the subpattern of the given number,
1881     provided that it occurs inside that subpattern. (If not, it is a "subroutine"
1882     call, which is described in the next section.) The special item (?R) or (?0) is
1883     a recursive call of the entire regular expression.
1884 nigel 87 </P>
1885     <P>
1886 nigel 93 In PCRE (like Python, but unlike Perl), a recursive subpattern call is always
1887     treated as an atomic group. That is, once it has matched some of the subject
1888     string, it is never re-entered, even if it contains untried alternatives and
1889     there is a subsequent matching failure.
1890     </P>
1891     <P>
1892 nigel 87 This PCRE pattern solves the nested parentheses problem (assume the
1893     PCRE_EXTENDED option is set so that white space is ignored):
1894 nigel 63 <pre>
1895     \( ( (?&#62;[^()]+) | (?R) )* \)
1896 nigel 75 </pre>
1897 nigel 63 First it matches an opening parenthesis. Then it matches any number of
1898     substrings which can either be a sequence of non-parentheses, or a recursive
1899 nigel 87 match of the pattern itself (that is, a correctly parenthesized substring).
1900 nigel 63 Finally there is a closing parenthesis.
1901     </P>
1902     <P>
1903     If this were part of a larger pattern, you would not want to recurse the entire
1904     pattern, so instead you could use this:
1905     <pre>
1906     ( \( ( (?&#62;[^()]+) | (?1) )* \) )
1907 nigel 75 </pre>
1908 nigel 63 We have put the pattern into parentheses, and caused the recursion to refer to
1909 ph10 172 them instead of the whole pattern.
1910     </P>
1911     <P>
1912     In a larger pattern, keeping track of parenthesis numbers can be tricky. This
1913     is made easier by the use of relative references. (A Perl 5.10 feature.)
1914     Instead of (?1) in the pattern above you can write (?-2) to refer to the second
1915     most recently opened parentheses preceding the recursion. In other words, a
1916     negative number counts capturing parentheses leftwards from the point at which
1917     it is encountered.
1918     </P>
1919     <P>
1920     It is also possible to refer to subsequently opened parentheses, by writing
1921     references such as (?+2). However, these cannot be recursive because the
1922     reference is not inside the parentheses that are referenced. They are always
1923     "subroutine" calls, as described in the next section.
1924     </P>
1925     <P>
1926     An alternative approach is to use named parentheses instead. The Perl syntax
1927     for this is (?&name); PCRE's earlier syntax (?P&#62;name) is also supported. We
1928     could rewrite the above example as follows:
1929 nigel 63 <pre>
1930 nigel 93 (?&#60;pn&#62; \( ( (?&#62;[^()]+) | (?&pn) )* \) )
1931 nigel 75 </pre>
1932 nigel 93 If there is more than one subpattern with the same name, the earliest one is
1933 ph10 172 used.
1934     </P>
1935     <P>
1936     This particular example pattern that we have been looking at contains nested
1937     unlimited repeats, and so the use of atomic grouping for matching strings of
1938     non-parentheses is important when applying the pattern to strings that do not
1939     match. For example, when this pattern is applied to
1940 nigel 63 <pre>
1941     (aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa()
1942 nigel 75 </pre>
1943 nigel 63 it yields "no match" quickly. However, if atomic grouping is not used,
1944     the match runs for a very long time indeed because there are so many different
1945     ways the + and * repeats can carve up the subject, and all have to be tested
1946     before failure can be reported.
1947     </P>
1948     <P>
1949     At the end of a match, the values set for any capturing subpatterns are those
1950     from the outermost level of the recursion at which the subpattern value is set.
1951     If you want to obtain intermediate values, a callout function can be used (see
1952 nigel 93 below and the
1953 nigel 63 <a href="pcrecallout.html"><b>pcrecallout</b></a>
1954     documentation). If the pattern above is matched against
1955     <pre>
1956     (ab(cd)ef)
1957 nigel 75 </pre>
1958 nigel 63 the value for the capturing parentheses is "ef", which is the last value taken
1959     on at the top level. If additional parentheses are added, giving
1960     <pre>
1961     \( ( ( (?&#62;[^()]+) | (?R) )* ) \)
1962     ^ ^
1963     ^ ^
1964 nigel 75 </pre>
1965 nigel 63 the string they capture is "ab(cd)ef", the contents of the top level
1966     parentheses. If there are more than 15 capturing parentheses in a pattern, PCRE
1967     has to obtain extra memory to store data during a recursion, which it does by
1968     using <b>pcre_malloc</b>, freeing it via <b>pcre_free</b> afterwards. If no
1969     memory can be obtained, the match fails with the PCRE_ERROR_NOMEMORY error.
1970     </P>
1971     <P>
1972     Do not confuse the (?R) item with the condition (R), which tests for recursion.
1973     Consider this pattern, which matches text in angle brackets, allowing for
1974     arbitrary nesting. Only digits are allowed in nested brackets (that is, when
1975     recursing), whereas any characters are permitted at the outer level.
1976     <pre>
1977     &#60; (?: (?(R) \d++ | [^&#60;&#62;]*+) | (?R)) * &#62;
1978 nigel 75 </pre>
1979 nigel 63 In this pattern, (?(R) is the start of a conditional subpattern, with two
1980     different alternatives for the recursive and non-recursive cases. The (?R) item
1981     is the actual recursive call.
1982 nigel 75 <a name="subpatternsassubroutines"></a></P>
1983 ph10 227 <br><a name="SEC22" href="#TOC1">SUBPATTERNS AS SUBROUTINES</a><br>
1984 nigel 63 <P>
1985     If the syntax for a recursive subpattern reference (either by number or by
1986     name) is used outside the parentheses to which it refers, it operates like a
1987 nigel 93 subroutine in a programming language. The "called" subpattern may be defined
1988 ph10 172 before or after the reference. A numbered reference can be absolute or
1989     relative, as in these examples:
1990 nigel 63 <pre>
1991 ph10 172 (...(absolute)...)...(?2)...
1992     (...(relative)...)...(?-1)...
1993     (...(?+1)...(relative)...
1994     </pre>
1995     An earlier example pointed out that the pattern
1996     <pre>
1997 nigel 63 (sens|respons)e and \1ibility
1998 nigel 75 </pre>
1999 nigel 63 matches "sense and sensibility" and "response and responsibility", but not
2000     "sense and responsibility". If instead the pattern
2001     <pre>
2002     (sens|respons)e and (?1)ibility
2003 nigel 75 </pre>
2004 nigel 63 is used, it does match "sense and responsibility" as well as the other two
2005 nigel 93 strings. Another example is given in the discussion of DEFINE above.
2006 nigel 63 </P>
2007 nigel 87 <P>
2008     Like recursive subpatterns, a "subroutine" call is always treated as an atomic
2009     group. That is, once it has matched some of the subject string, it is never
2010     re-entered, even if it contains untried alternatives and there is a subsequent
2011     matching failure.
2012     </P>
2013 nigel 63 <P>
2014 nigel 93 When a subpattern is used as a subroutine, processing options such as
2015     case-independence are fixed when the subpattern is defined. They cannot be
2016     changed for different calls. For example, consider this pattern:
2017     <pre>
2018 ph10 172 (abc)(?i:(?-1))
2019 nigel 93 </pre>
2020     It matches "abcabc". It does not match "abcABC" because the change of
2021     processing option does not affect the called subpattern.
2022     </P>
2023 ph10 227 <br><a name="SEC23" href="#TOC1">CALLOUTS</a><br>
2024 nigel 93 <P>
2025 nigel 63 Perl has a feature whereby using the sequence (?{...}) causes arbitrary Perl
2026     code to be obeyed in the middle of matching a regular expression. This makes it
2027     possible, amongst other things, to extract different substrings that match the
2028     same pair of parentheses when there is a repetition.
2029     </P>
2030     <P>
2031     PCRE provides a similar feature, but of course it cannot obey arbitrary Perl
2032     code. The feature is called "callout". The caller of PCRE provides an external
2033     function by putting its entry point in the global variable <i>pcre_callout</i>.
2034     By default, this variable contains NULL, which disables all calling out.
2035     </P>
2036     <P>
2037     Within a regular expression, (?C) indicates the points at which the external
2038     function is to be called. If you want to identify different callout points, you
2039     can put a number less than 256 after the letter C. The default value is zero.
2040     For example, this pattern has two callout points:
2041     <pre>
2042 ph10 155 (?C1)abc(?C2)def
2043 nigel 75 </pre>
2044     If the PCRE_AUTO_CALLOUT flag is passed to <b>pcre_compile()</b>, callouts are
2045     automatically installed before each item in the pattern. They are all numbered
2046     255.
2047 nigel 63 </P>
2048     <P>
2049     During matching, when PCRE reaches a callout point (and <i>pcre_callout</i> is
2050     set), the external function is called. It is provided with the number of the
2051 nigel 75 callout, the position in the pattern, and, optionally, one item of data
2052     originally supplied by the caller of <b>pcre_exec()</b>. The callout function
2053     may cause matching to proceed, to backtrack, or to fail altogether. A complete
2054     description of the interface to the callout function is given in the
2055 nigel 63 <a href="pcrecallout.html"><b>pcrecallout</b></a>
2056     documentation.
2057     </P>
2058 ph10 243 <br><a name="SEC24" href="#TOC1">BACKTRACKING CONTROL</a><br>
2059 nigel 63 <P>
2060 ph10 211 Perl 5.10 introduced a number of "Special Backtracking Control Verbs", which
2061     are described in the Perl documentation as "experimental and subject to change
2062     or removal in a future version of Perl". It goes on to say: "Their usage in
2063     production code should be noted to avoid problems during upgrades." The same
2064     remarks apply to the PCRE features described in this section.
2065     </P>
2066     <P>
2067     Since these verbs are specifically related to backtracking, they can be used
2068     only when the pattern is to be matched using <b>pcre_exec()</b>, which uses a
2069     backtracking algorithm. They cause an error if encountered by
2070     <b>pcre_dfa_exec()</b>.
2071     </P>
2072     <P>
2073     The new verbs make use of what was previously invalid syntax: an opening
2074     parenthesis followed by an asterisk. In Perl, they are generally of the form
2075     (*VERB:ARG) but PCRE does not support the use of arguments, so its general
2076     form is just (*VERB). Any number of these verbs may occur in a pattern. There
2077     are two kinds:
2078     </P>
2079     <br><b>
2080     Verbs that act immediately
2081     </b><br>
2082     <P>
2083     The following verbs act as soon as they are encountered:
2084     <pre>
2085     (*ACCEPT)
2086     </pre>
2087     This verb causes the match to end successfully, skipping the remainder of the
2088     pattern. When inside a recursion, only the innermost pattern is ended
2089     immediately. PCRE differs from Perl in what happens if the (*ACCEPT) is inside
2090     capturing parentheses. In Perl, the data so far is captured: in PCRE no data is
2091     captured. For example:
2092     <pre>
2093     A(A|B(*ACCEPT)|C)D
2094     </pre>
2095     This matches "AB", "AAD", or "ACD", but when it matches "AB", no data is
2096     captured.
2097     <pre>
2098     (*FAIL) or (*F)
2099     </pre>
2100     This verb causes the match to fail, forcing backtracking to occur. It is
2101     equivalent to (?!) but easier to read. The Perl documentation notes that it is
2102     probably useful only when combined with (?{}) or (??{}). Those are, of course,
2103     Perl features that are not present in PCRE. The nearest equivalent is the
2104     callout feature, as for example in this pattern:
2105     <pre>
2106     a+(?C)(*FAIL)
2107     </pre>
2108     A match with the string "aaaa" always fails, but the callout is taken before
2109     each backtrack happens (in this example, 10 times).
2110     </P>
2111     <br><b>
2112     Verbs that act after backtracking
2113     </b><br>
2114     <P>
2115     The following verbs do nothing when they are encountered. Matching continues
2116     with what follows, but if there is no subsequent match, a failure is forced.
2117     The verbs differ in exactly what kind of failure occurs.
2118     <pre>
2119     (*COMMIT)
2120     </pre>
2121     This verb causes the whole match to fail outright if the rest of the pattern
2122     does not match. Even if the pattern is unanchored, no further attempts to find
2123     a match by advancing the start point take place. Once (*COMMIT) has been
2124     passed, <b>pcre_exec()</b> is committed to finding a match at the current
2125     starting point, or not at all. For example:
2126     <pre>
2127     a+(*COMMIT)b
2128     </pre>
2129     This matches "xxaab" but not "aacaab". It can be thought of as a kind of
2130     dynamic anchor, or "I've started, so I must finish."
2131     <pre>
2132     (*PRUNE)
2133     </pre>
2134     This verb causes the match to fail at the current position if the rest of the
2135     pattern does not match. If the pattern is unanchored, the normal "bumpalong"
2136     advance to the next starting character then happens. Backtracking can occur as
2137     usual to the left of (*PRUNE), or when matching to the right of (*PRUNE), but
2138     if there is no match to the right, backtracking cannot cross (*PRUNE).
2139     In simple cases, the use of (*PRUNE) is just an alternative to an atomic
2140     group or possessive quantifier, but there are some uses of (*PRUNE) that cannot
2141     be expressed in any other way.
2142     <pre>
2143     (*SKIP)
2144     </pre>
2145     This verb is like (*PRUNE), except that if the pattern is unanchored, the
2146     "bumpalong" advance is not to the next character, but to the position in the
2147     subject where (*SKIP) was encountered. (*SKIP) signifies that whatever text
2148     was matched leading up to it cannot be part of a successful match. Consider:
2149     <pre>
2150     a+(*SKIP)b
2151     </pre>
2152     If the subject is "aaaac...", after the first match attempt fails (starting at
2153     the first character in the string), the starting point skips on to start the
2154     next attempt at "c". Note that a possessive quantifer does not have the same
2155     effect in this example; although it would suppress backtracking during the
2156     first match attempt, the second attempt would start at the second character
2157     instead of skipping on to "c".
2158     <pre>
2159     (*THEN)
2160     </pre>
2161     This verb causes a skip to the next alternation if the rest of the pattern does
2162     not match. That is, it cancels pending backtracking, but only within the
2163     current alternation. Its name comes from the observation that it can be used
2164     for a pattern-based if-then-else block:
2165     <pre>
2166     ( COND1 (*THEN) FOO | COND2 (*THEN) BAR | COND3 (*THEN) BAZ ) ...
2167     </pre>
2168     If the COND1 pattern matches, FOO is tried (and possibly further items after
2169     the end of the group if FOO succeeds); on failure the matcher skips to the
2170     second alternative and tries COND2, without backtracking into COND1. If (*THEN)
2171     is used outside of any alternation, it acts exactly like (*PRUNE).
2172     </P>
2173 ph10 227 <br><a name="SEC25" href="#TOC1">SEE ALSO</a><br>
2174 ph10 211 <P>
2175 nigel 93 <b>pcreapi</b>(3), <b>pcrecallout</b>(3), <b>pcrematching</b>(3), <b>pcre</b>(3).
2176     </P>
2177 ph10 227 <br><a name="SEC26" href="#TOC1">AUTHOR</a><br>
2178 nigel 93 <P>
2179 ph10 99 Philip Hazel
2180 nigel 63 <br>
2181 ph10 99 University Computing Service
2182     <br>
2183     Cambridge CB2 3QH, England.
2184     <br>
2185     </P>
2186 ph10 227 <br><a name="SEC27" href="#TOC1">REVISION</a><br>
2187 ph10 99 <P>
2188 ph10 231 Last updated: 11 September 2007
2189 ph10 99 <br>
2190     Copyright &copy; 1997-2007 University of Cambridge.
2191     <br>
2192 nigel 75 <p>
2193     Return to the <a href="index.html">PCRE index page</a>.
2194     </p>

Properties

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

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12