ViewVC logotype

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

Parent Directory Parent Directory | Revision Log Revision Log

Revision 208 - (hide annotations) (download) (as text)
Mon Aug 6 15:23:29 2007 UTC (7 years, 8 months ago) by ph10
File MIME type: text/html
File size: 84388 byte(s)
Added a pcresyntax man page; tidied some others.

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 nigel 93 <li><a name="TOC2" href="#SEC2">CHARACTERS AND METACHARACTERS</a>
18     <li><a name="TOC3" href="#SEC3">BACKSLASH</a>
19     <li><a name="TOC4" href="#SEC4">CIRCUMFLEX AND DOLLAR</a>
20     <li><a name="TOC5" href="#SEC5">FULL STOP (PERIOD, DOT)</a>
21     <li><a name="TOC6" href="#SEC6">MATCHING A SINGLE BYTE</a>
22     <li><a name="TOC7" href="#SEC7">SQUARE BRACKETS AND CHARACTER CLASSES</a>
23     <li><a name="TOC8" href="#SEC8">POSIX CHARACTER CLASSES</a>
24     <li><a name="TOC9" href="#SEC9">VERTICAL BAR</a>
25     <li><a name="TOC10" href="#SEC10">INTERNAL OPTION SETTING</a>
26     <li><a name="TOC11" href="#SEC11">SUBPATTERNS</a>
27 ph10 182 <li><a name="TOC12" href="#SEC12">DUPLICATE SUBPATTERN NUMBERS</a>
28     <li><a name="TOC13" href="#SEC13">NAMED SUBPATTERNS</a>
29     <li><a name="TOC14" href="#SEC14">REPETITION</a>
31     <li><a name="TOC16" href="#SEC16">BACK REFERENCES</a>
32     <li><a name="TOC17" href="#SEC17">ASSERTIONS</a>
33     <li><a name="TOC18" href="#SEC18">CONDITIONAL SUBPATTERNS</a>
34     <li><a name="TOC19" href="#SEC19">COMMENTS</a>
35     <li><a name="TOC20" href="#SEC20">RECURSIVE PATTERNS</a>
36     <li><a name="TOC21" href="#SEC21">SUBPATTERNS AS SUBROUTINES</a>
37     <li><a name="TOC22" href="#SEC22">CALLOUTS</a>
38     <li><a name="TOC23" href="#SEC23">SEE ALSO</a>
39     <li><a name="TOC24" href="#SEC24">AUTHOR</a>
40     <li><a name="TOC25" href="#SEC25">REVISION</a>
41 nigel 63 </ul>
42     <br><a name="SEC1" href="#TOC1">PCRE REGULAR EXPRESSION DETAILS</a><br>
43     <P>
44 ph10 208 The syntax and semantics of the regular expressions that are supported by PCRE
45     are described in detail below. There is a quick-reference syntax summary in the
46     <a href="pcresyntax.html"><b>pcresyntax</b></a>
47     page. Perl's regular expressions are described in its own documentation, and
48     regular expressions in general are covered in a number of books, some of which
49     have copious examples. Jeffrey Friedl's "Mastering Regular Expressions",
50     published by O'Reilly, covers regular expressions in great detail. This
51     description of PCRE's regular expressions is intended as reference material.
52 nigel 63 </P>
53     <P>
54 nigel 75 The original operation of PCRE was on strings of one-byte characters. However,
55     there is now also support for UTF-8 character strings. To use this, you must
56     build PCRE to include UTF-8 support, and then call <b>pcre_compile()</b> with
57     the PCRE_UTF8 option. How this affects pattern matching is mentioned in several
58     places below. There is also a summary of UTF-8 features in the
59 nigel 63 <a href="pcre.html#utf8support">section on UTF-8 support</a>
60     in the main
61     <a href="pcre.html"><b>pcre</b></a>
62     page.
63     </P>
64     <P>
65 nigel 77 The remainder of this document discusses the patterns that are supported by
66     PCRE when its main matching function, <b>pcre_exec()</b>, is used.
67     From release 6.0, PCRE offers a second matching function,
68     <b>pcre_dfa_exec()</b>, which matches using a different algorithm that is not
69 ph10 172 Perl-compatible. Some of the features discussed below are not available when
70     <b>pcre_dfa_exec()</b> is used. The advantages and disadvantages of the
71     alternative function, and how it differs from the normal function, are
72     discussed in the
73 nigel 77 <a href="pcrematching.html"><b>pcrematching</b></a>
74     page.
75     </P>
76 nigel 93 <br><a name="SEC2" href="#TOC1">CHARACTERS AND METACHARACTERS</a><br>
77 nigel 77 <P>
78 nigel 63 A regular expression is a pattern that is matched against a subject string from
79     left to right. Most characters stand for themselves in a pattern, and match the
80     corresponding characters in the subject. As a trivial example, the pattern
81     <pre>
82     The quick brown fox
83 nigel 75 </pre>
84 nigel 77 matches a portion of a subject string that is identical to itself. When
85     caseless matching is specified (the PCRE_CASELESS option), letters are matched
86     independently of case. In UTF-8 mode, PCRE always understands the concept of
87     case for characters whose values are less than 128, so caseless matching is
88     always possible. For characters with higher values, the concept of case is
89     supported if PCRE is compiled with Unicode property support, but not otherwise.
90     If you want to use caseless matching for characters 128 and above, you must
91     ensure that PCRE is compiled with Unicode property support as well as with
92     UTF-8 support.
93     </P>
94     <P>
95     The power of regular expressions comes from the ability to include alternatives
96     and repetitions in the pattern. These are encoded in the pattern by the use of
97 nigel 75 <i>metacharacters</i>, which do not stand for themselves but instead are
98 nigel 63 interpreted in some special way.
99     </P>
100     <P>
101 nigel 75 There are two different sets of metacharacters: those that are recognized
102 nigel 63 anywhere in the pattern except within square brackets, and those that are
103 nigel 93 recognized within square brackets. Outside square brackets, the metacharacters
104     are as follows:
105 nigel 63 <pre>
106     \ general escape character with several uses
107     ^ assert start of string (or line, in multiline mode)
108     $ assert end of string (or line, in multiline mode)
109     . match any character except newline (by default)
110     [ start character class definition
111     | start of alternative branch
112     ( start subpattern
113     ) end subpattern
114     ? extends the meaning of (
115     also 0 or 1 quantifier
116     also quantifier minimizer
117     * 0 or more quantifier
118     + 1 or more quantifier
119     also "possessive quantifier"
120     { start min/max quantifier
121 nigel 75 </pre>
122 nigel 63 Part of a pattern that is in square brackets is called a "character class". In
123 nigel 75 a character class the only metacharacters are:
124 nigel 63 <pre>
125     \ general escape character
126     ^ negate the class, but only if the first character
127     - indicates character range
128 nigel 75 [ POSIX character class (only if followed by POSIX syntax)
129 nigel 63 ] terminates the character class
130 nigel 75 </pre>
131     The following sections describe the use of each of the metacharacters.
132 nigel 63 </P>
133 nigel 93 <br><a name="SEC3" href="#TOC1">BACKSLASH</a><br>
134 nigel 63 <P>
135     The backslash character has several uses. Firstly, if it is followed by a
136 nigel 91 non-alphanumeric character, it takes away any special meaning that character
137     may have. This use of backslash as an escape character applies both inside and
138 nigel 63 outside character classes.
139     </P>
140     <P>
141     For example, if you want to match a * character, you write \* in the pattern.
142     This escaping action applies whether or not the following character would
143 nigel 75 otherwise be interpreted as a metacharacter, so it is always safe to precede a
144     non-alphanumeric with backslash to specify that it stands for itself. In
145 nigel 63 particular, if you want to match a backslash, you write \\.
146     </P>
147     <P>
148     If a pattern is compiled with the PCRE_EXTENDED option, whitespace in the
149     pattern (other than in a character class) and characters between a # outside
150 nigel 91 a character class and the next newline are ignored. An escaping backslash can
151     be used to include a whitespace or # character as part of the pattern.
152 nigel 63 </P>
153     <P>
154     If you want to remove the special meaning from a sequence of characters, you
155     can do so by putting them between \Q and \E. This is different from Perl in
156     that $ and @ are handled as literals in \Q...\E sequences in PCRE, whereas in
157     Perl, $ and @ cause variable interpolation. Note the following examples:
158     <pre>
159     Pattern PCRE matches Perl matches
160 nigel 75
161     \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
162 nigel 63 \Qabc\$xyz\E abc\$xyz abc\$xyz
163     \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
164 nigel 75 </pre>
165 nigel 63 The \Q...\E sequence is recognized both inside and outside character classes.
166 nigel 75 <a name="digitsafterbackslash"></a></P>
167     <br><b>
168     Non-printing characters
169     </b><br>
170 nigel 63 <P>
171     A second use of backslash provides a way of encoding non-printing characters
172     in patterns in a visible manner. There is no restriction on the appearance of
173     non-printing characters, apart from the binary zero that terminates a pattern,
174     but when a pattern is being prepared by text editing, it is usually easier to
175     use one of the following escape sequences than the binary character it
176     represents:
177     <pre>
178     \a alarm, that is, the BEL character (hex 07)
179     \cx "control-x", where x is any character
180     \e escape (hex 1B)
181     \f formfeed (hex 0C)
182     \n newline (hex 0A)
183     \r carriage return (hex 0D)
184     \t tab (hex 09)
185     \ddd character with octal code ddd, or backreference
186     \xhh character with hex code hh
187 nigel 87 \x{hhh..} character with hex code hhh..
188 nigel 75 </pre>
189 nigel 63 The precise effect of \cx is as follows: if x is a lower case letter, it
190     is converted to upper case. Then bit 6 of the character (hex 40) is inverted.
191     Thus \cz becomes hex 1A, but \c{ becomes hex 3B, while \c; becomes hex
192     7B.
193     </P>
194     <P>
195     After \x, from zero to two hexadecimal digits are read (letters can be in
196 nigel 87 upper or lower case). Any number of hexadecimal digits may appear between \x{
197     and }, but the value of the character code must be less than 256 in non-UTF-8
198     mode, and less than 2**31 in UTF-8 mode (that is, the maximum hexadecimal value
199     is 7FFFFFFF). If characters other than hexadecimal digits appear between \x{
200     and }, or if there is no terminating }, this form of escape is not recognized.
201     Instead, the initial \x will be interpreted as a basic hexadecimal escape,
202     with no following digits, giving a character whose value is zero.
203 nigel 63 </P>
204     <P>
205     Characters whose value is less than 256 can be defined by either of the two
206 nigel 87 syntaxes for \x. There is no difference in the way they are handled. For
207     example, \xdc is exactly the same as \x{dc}.
208 nigel 63 </P>
209     <P>
210 nigel 91 After \0 up to two further octal digits are read. If there are fewer than two
211     digits, just those that are present are used. Thus the sequence \0\x\07
212     specifies two binary zeros followed by a BEL character (code value 7). Make
213     sure you supply two digits after the initial zero if the pattern character that
214     follows is itself an octal digit.
215 nigel 63 </P>
216     <P>
217     The handling of a backslash followed by a digit other than 0 is complicated.
218     Outside a character class, PCRE reads it and any following digits as a decimal
219     number. If the number is less than 10, or if there have been at least that many
220     previous capturing left parentheses in the expression, the entire sequence is
221     taken as a <i>back reference</i>. A description of how this works is given
222 nigel 75 <a href="#backreferences">later,</a>
223     following the discussion of
224     <a href="#subpattern">parenthesized subpatterns.</a>
225 nigel 63 </P>
226     <P>
227     Inside a character class, or if the decimal number is greater than 9 and there
228     have not been that many capturing subpatterns, PCRE re-reads up to three octal
229 nigel 93 digits following the backslash, and uses them to generate a data character. Any
230 nigel 91 subsequent digits stand for themselves. In non-UTF-8 mode, the value of a
231     character specified in octal must be less than \400. In UTF-8 mode, values up
232     to \777 are permitted. For example:
233 nigel 63 <pre>
234     \040 is another way of writing a space
235 nigel 75 \40 is the same, provided there are fewer than 40 previous capturing subpatterns
236 nigel 63 \7 is always a back reference
237 nigel 75 \11 might be a back reference, or another way of writing a tab
238 nigel 63 \011 is always a tab
239     \0113 is a tab followed by the character "3"
240 nigel 75 \113 might be a back reference, otherwise the character with octal code 113
241     \377 might be a back reference, otherwise the byte consisting entirely of 1 bits
242     \81 is either a back reference, or a binary zero followed by the two characters "8" and "1"
243     </pre>
244 nigel 63 Note that octal values of 100 or greater must not be introduced by a leading
245     zero, because no more than three octal digits are ever read.
246     </P>
247     <P>
248 nigel 91 All the sequences that define a single character value can be used both inside
249     and outside character classes. In addition, inside a character class, the
250     sequence \b is interpreted as the backspace character (hex 08), and the
251 nigel 93 sequences \R and \X are interpreted as the characters "R" and "X",
252     respectively. Outside a character class, these sequences have different
253     meanings
254 nigel 75 <a href="#uniextseq">(see below).</a>
255 nigel 63 </P>
256 nigel 75 <br><b>
257 nigel 93 Absolute and relative back references
258     </b><br>
259     <P>
260 ph10 208 The sequence \g followed by an unsigned or a negative number, optionally
261     enclosed in braces, is an absolute or relative back reference. A named back
262     reference can be coded as \g{name}. Back references are discussed
263 nigel 93 <a href="#backreferences">later,</a>
264     following the discussion of
265     <a href="#subpattern">parenthesized subpatterns.</a>
266     </P>
267     <br><b>
268 nigel 75 Generic character types
269     </b><br>
270 nigel 63 <P>
271 nigel 93 Another use of backslash is for specifying generic character types. The
272 nigel 75 following are always recognized:
273 nigel 63 <pre>
274     \d any decimal digit
275     \D any character that is not a decimal digit
276 ph10 182 \h any horizontal whitespace character
277     \H any character that is not a horizontal whitespace character
278 nigel 63 \s any whitespace character
279     \S any character that is not a whitespace character
280 ph10 182 \v any vertical whitespace character
281     \V any character that is not a vertical whitespace character
282 nigel 63 \w any "word" character
283     \W any "non-word" character
284 nigel 75 </pre>
285 nigel 63 Each pair of escape sequences partitions the complete set of characters into
286     two disjoint sets. Any given character matches one, and only one, of each pair.
287     </P>
288     <P>
289 nigel 75 These character type sequences can appear both inside and outside character
290     classes. They each match one character of the appropriate type. If the current
291     matching point is at the end of the subject string, all of them fail, since
292     there is no character to match.
293 nigel 63 </P>
294     <P>
295     For compatibility with Perl, \s does not match the VT character (code 11).
296     This makes it different from the the POSIX "space" class. The \s characters
297 ph10 182 are HT (9), LF (10), FF (12), CR (13), and space (32). If "use locale;" is
298 nigel 91 included in a Perl script, \s may match the VT character. In PCRE, it never
299 ph10 182 does.
300 nigel 63 </P>
301     <P>
302 ph10 182 In UTF-8 mode, characters with values greater than 128 never match \d, \s, or
303     \w, and always match \D, \S, and \W. This is true even when Unicode
304     character property support is available. These sequences retain their original
305     meanings from before UTF-8 support was available, mainly for efficiency
306     reasons.
307     </P>
308     <P>
309     The sequences \h, \H, \v, and \V are Perl 5.10 features. In contrast to the
310     other sequences, these do match certain high-valued codepoints in UTF-8 mode.
311     The horizontal space characters are:
312     <pre>
313     U+0009 Horizontal tab
314     U+0020 Space
315     U+00A0 Non-break space
316     U+1680 Ogham space mark
317     U+180E Mongolian vowel separator
318     U+2000 En quad
319     U+2001 Em quad
320     U+2002 En space
321     U+2003 Em space
322     U+2004 Three-per-em space
323     U+2005 Four-per-em space
324     U+2006 Six-per-em space
325     U+2007 Figure space
326     U+2008 Punctuation space
327     U+2009 Thin space
328     U+200A Hair space
329     U+202F Narrow no-break space
330     U+205F Medium mathematical space
331     U+3000 Ideographic space
332     </pre>
333     The vertical space characters are:
334     <pre>
335     U+000A Linefeed
336     U+000B Vertical tab
337     U+000C Formfeed
338     U+000D Carriage return
339     U+0085 Next line
340     U+2028 Line separator
341     U+2029 Paragraph separator
342     </PRE>
343     </P>
344     <P>
345 nigel 75 A "word" character is an underscore or any character less than 256 that is a
346     letter or digit. The definition of letters and digits is controlled by PCRE's
347     low-valued character tables, and may vary if locale-specific matching is taking
348     place (see
349 nigel 63 <a href="pcreapi.html#localesupport">"Locale support"</a>
350     in the
351     <a href="pcreapi.html"><b>pcreapi</b></a>
352 ph10 142 page). For example, in a French locale such as "fr_FR" in Unix-like systems,
353     or "french" in Windows, some character codes greater than 128 are used for
354 ph10 182 accented letters, and these are matched by \w. The use of locales with Unicode
355     is discouraged.
356 nigel 63 </P>
357 nigel 93 <br><b>
358     Newline sequences
359     </b><br>
360     <P>
361     Outside a character class, the escape sequence \R matches any Unicode newline
362 ph10 182 sequence. This is a Perl 5.10 feature. In non-UTF-8 mode \R is equivalent to
363 nigel 93 the following:
364     <pre>
365     (?&#62;\r\n|\n|\x0b|\f|\r|\x85)
366     </pre>
367     This is an example of an "atomic group", details of which are given
368     <a href="#atomicgroup">below.</a>
369     This particular group matches either the two-character sequence CR followed by
370     LF, or one of the single characters LF (linefeed, U+000A), VT (vertical tab,
371     U+000B), FF (formfeed, U+000C), CR (carriage return, U+000D), or NEL (next
372     line, U+0085). The two-character sequence is treated as a single unit that
373     cannot be split.
374     </P>
375     <P>
376     In UTF-8 mode, two additional characters whose codepoints are greater than 255
377     are added: LS (line separator, U+2028) and PS (paragraph separator, U+2029).
378     Unicode character property support is not needed for these characters to be
379     recognized.
380     </P>
381     <P>
382     Inside a character class, \R matches the letter "R".
383 nigel 75 <a name="uniextseq"></a></P>
384     <br><b>
385     Unicode character properties
386     </b><br>
387     <P>
388     When PCRE is built with Unicode character property support, three additional
389 ph10 185 escape sequences that match characters with specific properties are available.
390     When not in UTF-8 mode, these sequences are of course limited to testing
391     characters whose codepoints are less than 256, but they do work in this mode.
392     The extra escape sequences are:
393 nigel 75 <pre>
394 nigel 87 \p{<i>xx</i>} a character with the <i>xx</i> property
395     \P{<i>xx</i>} a character without the <i>xx</i> property
396     \X an extended Unicode sequence
397 nigel 75 </pre>
398 nigel 87 The property names represented by <i>xx</i> above are limited to the Unicode
399     script names, the general category properties, and "Any", which matches any
400     character (including newline). Other properties such as "InMusicalSymbols" are
401     not currently supported by PCRE. Note that \P{Any} does not match any
402     characters, so always causes a match failure.
403 nigel 63 </P>
404     <P>
405 nigel 87 Sets of Unicode characters are defined as belonging to certain scripts. A
406     character from one of these sets can be matched using a script name. For
407     example:
408 nigel 75 <pre>
409 nigel 87 \p{Greek}
410     \P{Han}
411     </pre>
412     Those that are not part of an identified script are lumped together as
413     "Common". The current list of scripts is:
414     </P>
415     <P>
416     Arabic,
417     Armenian,
418 nigel 93 Balinese,
419 nigel 87 Bengali,
420     Bopomofo,
421     Braille,
422     Buginese,
423     Buhid,
424     Canadian_Aboriginal,
425     Cherokee,
426     Common,
427     Coptic,
428 nigel 93 Cuneiform,
429 nigel 87 Cypriot,
430     Cyrillic,
431     Deseret,
432     Devanagari,
433     Ethiopic,
434     Georgian,
435     Glagolitic,
436     Gothic,
437     Greek,
438     Gujarati,
439     Gurmukhi,
440     Han,
441     Hangul,
442     Hanunoo,
443     Hebrew,
444     Hiragana,
445     Inherited,
446     Kannada,
447     Katakana,
448     Kharoshthi,
449     Khmer,
450     Lao,
451     Latin,
452     Limbu,
453     Linear_B,
454     Malayalam,
455     Mongolian,
456     Myanmar,
457     New_Tai_Lue,
458 nigel 93 Nko,
459 nigel 87 Ogham,
460     Old_Italic,
461     Old_Persian,
462     Oriya,
463     Osmanya,
464 nigel 93 Phags_Pa,
465     Phoenician,
466 nigel 87 Runic,
467     Shavian,
468     Sinhala,
469     Syloti_Nagri,
470     Syriac,
471     Tagalog,
472     Tagbanwa,
473     Tai_Le,
474     Tamil,
475     Telugu,
476     Thaana,
477     Thai,
478     Tibetan,
479     Tifinagh,
480     Ugaritic,
481     Yi.
482     </P>
483     <P>
484     Each character has exactly one general category property, specified by a
485     two-letter abbreviation. For compatibility with Perl, negation can be specified
486     by including a circumflex between the opening brace and the property name. For
487     example, \p{^Lu} is the same as \P{Lu}.
488     </P>
489     <P>
490     If only one letter is specified with \p or \P, it includes all the general
491     category properties that start with that letter. In this case, in the absence
492     of negation, the curly brackets in the escape sequence are optional; these two
493     examples have the same effect:
494     <pre>
495 nigel 75 \p{L}
496     \pL
497     </pre>
498 nigel 87 The following general category property codes are supported:
499 nigel 75 <pre>
500     C Other
501     Cc Control
502     Cf Format
503     Cn Unassigned
504     Co Private use
505     Cs Surrogate
507     L Letter
508     Ll Lower case letter
509     Lm Modifier letter
510     Lo Other letter
511     Lt Title case letter
512     Lu Upper case letter
514     M Mark
515     Mc Spacing mark
516     Me Enclosing mark
517     Mn Non-spacing mark
519     N Number
520     Nd Decimal number
521     Nl Letter number
522     No Other number
524     P Punctuation
525     Pc Connector punctuation
526     Pd Dash punctuation
527     Pe Close punctuation
528     Pf Final punctuation
529     Pi Initial punctuation
530     Po Other punctuation
531     Ps Open punctuation
533     S Symbol
534     Sc Currency symbol
535     Sk Modifier symbol
536     Sm Mathematical symbol
537     So Other symbol
539     Z Separator
540     Zl Line separator
541     Zp Paragraph separator
542     Zs Space separator
543     </pre>
544 nigel 87 The special property L& is also supported: it matches a character that has
545     the Lu, Ll, or Lt property, in other words, a letter that is not classified as
546     a modifier or "other".
547 nigel 75 </P>
548     <P>
549 nigel 87 The long synonyms for these properties that Perl supports (such as \p{Letter})
550 nigel 91 are not supported by PCRE, nor is it permitted to prefix any of these
551 nigel 87 properties with "Is".
552     </P>
553     <P>
554     No character that is in the Unicode table has the Cn (unassigned) property.
555     Instead, this property is assumed for any code point that is not in the
556     Unicode table.
557     </P>
558     <P>
559 nigel 75 Specifying caseless matching does not affect these escape sequences. For
560     example, \p{Lu} always matches only upper case letters.
561     </P>
562     <P>
563     The \X escape matches any number of Unicode characters that form an extended
564     Unicode sequence. \X is equivalent to
565     <pre>
566     (?&#62;\PM\pM*)
567     </pre>
568     That is, it matches a character without the "mark" property, followed by zero
569     or more characters with the "mark" property, and treats the sequence as an
570     atomic group
571     <a href="#atomicgroup">(see below).</a>
572     Characters with the "mark" property are typically accents that affect the
573 ph10 185 preceding character. None of them have codepoints less than 256, so in
574     non-UTF-8 mode \X matches any one character.
575 nigel 75 </P>
576     <P>
577     Matching characters by Unicode property is not fast, because PCRE has to search
578     a structure that contains data for over fifteen thousand characters. That is
579     why the traditional escape sequences such as \d and \w do not use Unicode
580     properties in PCRE.
581 ph10 172 <a name="resetmatchstart"></a></P>
582     <br><b>
583     Resetting the match start
584     </b><br>
585     <P>
586     The escape sequence \K, which is a Perl 5.10 feature, causes any previously
587     matched characters not to be included in the final matched sequence. For
588     example, the pattern:
589     <pre>
590     foo\Kbar
591     </pre>
592     matches "foobar", but reports that it has matched "bar". This feature is
593     similar to a lookbehind assertion
594     <a href="#lookbehind">(described below).</a>
595     However, in this case, the part of the subject before the real match does not
596     have to be of fixed length, as lookbehind assertions do. The use of \K does
597     not interfere with the setting of
598     <a href="#subpattern">captured substrings.</a>
599     For example, when the pattern
600     <pre>
601     (foo)\Kbar
602     </pre>
603     matches "foobar", the first substring is still set to "foo".
604 nigel 75 <a name="smallassertions"></a></P>
605     <br><b>
606     Simple assertions
607     </b><br>
608     <P>
609 nigel 93 The final use of backslash is for certain simple assertions. An assertion
610 nigel 63 specifies a condition that has to be met at a particular point in a match,
611     without consuming any characters from the subject string. The use of
612 nigel 75 subpatterns for more complicated assertions is described
613     <a href="#bigassertions">below.</a>
614 nigel 91 The backslashed assertions are:
615 nigel 63 <pre>
616     \b matches at a word boundary
617     \B matches when not at a word boundary
618 nigel 93 \A matches at the start of the subject
619     \Z matches at the end of the subject
620     also matches before a newline at the end of the subject
621     \z matches only at the end of the subject
622     \G matches at the first matching position in the subject
623 nigel 75 </pre>
624 nigel 63 These assertions may not appear in character classes (but note that \b has a
625     different meaning, namely the backspace character, inside a character class).
626     </P>
627     <P>
628     A word boundary is a position in the subject string where the current character
629     and the previous character do not both match \w or \W (i.e. one matches
630     \w and the other matches \W), or the start or end of the string if the
631     first or last character matches \w, respectively.
632     </P>
633     <P>
634     The \A, \Z, and \z assertions differ from the traditional circumflex and
635 nigel 75 dollar (described in the next section) in that they only ever match at the very
636     start and end of the subject string, whatever options are set. Thus, they are
637     independent of multiline mode. These three assertions are not affected by the
638     PCRE_NOTBOL or PCRE_NOTEOL options, which affect only the behaviour of the
639     circumflex and dollar metacharacters. However, if the <i>startoffset</i>
640     argument of <b>pcre_exec()</b> is non-zero, indicating that matching is to start
641     at a point other than the beginning of the subject, \A can never match. The
642 nigel 91 difference between \Z and \z is that \Z matches before a newline at the end
643     of the string as well as at the very end, whereas \z matches only at the end.
644 nigel 63 </P>
645     <P>
646     The \G assertion is true only when the current matching position is at the
647     start point of the match, as specified by the <i>startoffset</i> argument of
648     <b>pcre_exec()</b>. It differs from \A when the value of <i>startoffset</i> is
649     non-zero. By calling <b>pcre_exec()</b> multiple times with appropriate
650     arguments, you can mimic Perl's /g option, and it is in this kind of
651     implementation where \G can be useful.
652     </P>
653     <P>
654     Note, however, that PCRE's interpretation of \G, as the start of the current
655     match, is subtly different from Perl's, which defines it as the end of the
656     previous match. In Perl, these can be different when the previously matched
657     string was empty. Because PCRE does just one match at a time, it cannot
658     reproduce this behaviour.
659     </P>
660     <P>
661     If all the alternatives of a pattern begin with \G, the expression is anchored
662     to the starting match position, and the "anchored" flag is set in the compiled
663     regular expression.
664     </P>
665 nigel 93 <br><a name="SEC4" href="#TOC1">CIRCUMFLEX AND DOLLAR</a><br>
666 nigel 63 <P>
667     Outside a character class, in the default matching mode, the circumflex
668 nigel 75 character is an assertion that is true only if the current matching point is
669 nigel 63 at the start of the subject string. If the <i>startoffset</i> argument of
670     <b>pcre_exec()</b> is non-zero, circumflex can never match if the PCRE_MULTILINE
671     option is unset. Inside a character class, circumflex has an entirely different
672 nigel 75 meaning
673     <a href="#characterclass">(see below).</a>
674 nigel 63 </P>
675     <P>
676     Circumflex need not be the first character of the pattern if a number of
677     alternatives are involved, but it should be the first thing in each alternative
678     in which it appears if the pattern is ever to match that branch. If all
679     possible alternatives start with a circumflex, that is, if the pattern is
680     constrained to match only at the start of the subject, it is said to be an
681     "anchored" pattern. (There are also other constructs that can cause a pattern
682     to be anchored.)
683     </P>
684     <P>
685 nigel 75 A dollar character is an assertion that is true only if the current matching
686 nigel 63 point is at the end of the subject string, or immediately before a newline
687 nigel 91 at the end of the string (by default). Dollar need not be the last character of
688     the pattern if a number of alternatives are involved, but it should be the last
689     item in any branch in which it appears. Dollar has no special meaning in a
690     character class.
691 nigel 63 </P>
692     <P>
693     The meaning of dollar can be changed so that it matches only at the very end of
694     the string, by setting the PCRE_DOLLAR_ENDONLY option at compile time. This
695     does not affect the \Z assertion.
696     </P>
697     <P>
698     The meanings of the circumflex and dollar characters are changed if the
699 nigel 91 PCRE_MULTILINE option is set. When this is the case, a circumflex matches
700     immediately after internal newlines as well as at the start of the subject
701     string. It does not match after a newline that ends the string. A dollar
702     matches before any newlines in the string, as well as at the very end, when
703     PCRE_MULTILINE is set. When newline is specified as the two-character
704     sequence CRLF, isolated CR and LF characters do not indicate newlines.
705 nigel 63 </P>
706     <P>
707 nigel 91 For example, the pattern /^abc$/ matches the subject string "def\nabc" (where
708     \n represents a newline) in multiline mode, but not otherwise. Consequently,
709     patterns that are anchored in single line mode because all branches start with
710     ^ are not anchored in multiline mode, and a match for circumflex is possible
711     when the <i>startoffset</i> argument of <b>pcre_exec()</b> is non-zero. The
712     PCRE_DOLLAR_ENDONLY option is ignored if PCRE_MULTILINE is set.
713     </P>
714     <P>
715 nigel 63 Note that the sequences \A, \Z, and \z can be used to match the start and
716     end of the subject in both modes, and if all branches of a pattern start with
717 nigel 91 \A it is always anchored, whether or not PCRE_MULTILINE is set.
718 nigel 63 </P>
719 nigel 93 <br><a name="SEC5" href="#TOC1">FULL STOP (PERIOD, DOT)</a><br>
720 nigel 63 <P>
721     Outside a character class, a dot in the pattern matches any one character in
722 nigel 91 the subject string except (by default) a character that signifies the end of a
723 nigel 93 line. In UTF-8 mode, the matched character may be more than one byte long.
724 nigel 63 </P>
725 nigel 91 <P>
726 nigel 93 When a line ending is defined as a single character, dot never matches that
727     character; when the two-character sequence CRLF is used, dot does not match CR
728     if it is immediately followed by LF, but otherwise it matches all characters
729     (including isolated CRs and LFs). When any Unicode line endings are being
730     recognized, dot does not match CR or LF or any of the other line ending
731     characters.
732     </P>
733     <P>
734 nigel 91 The behaviour of dot with regard to newlines can be changed. If the PCRE_DOTALL
735 nigel 93 option is set, a dot matches any one character, without exception. If the
736     two-character sequence CRLF is present in the subject string, it takes two dots
737     to match it.
738 nigel 91 </P>
739     <P>
740     The handling of dot is entirely independent of the handling of circumflex and
741     dollar, the only relationship being that they both involve newlines. Dot has no
742     special meaning in a character class.
743     </P>
744 nigel 93 <br><a name="SEC6" href="#TOC1">MATCHING A SINGLE BYTE</a><br>
745 nigel 63 <P>
746     Outside a character class, the escape sequence \C matches any one byte, both
747 nigel 93 in and out of UTF-8 mode. Unlike a dot, it always matches any line-ending
748     characters. The feature is provided in Perl in order to match individual bytes
749     in UTF-8 mode. Because it breaks up UTF-8 characters into individual bytes,
750     what remains in the string may be a malformed UTF-8 string. For this reason,
751     the \C escape sequence is best avoided.
752 nigel 63 </P>
753     <P>
754 nigel 75 PCRE does not allow \C to appear in lookbehind assertions
755     <a href="#lookbehind">(described below),</a>
756     because in UTF-8 mode this would make it impossible to calculate the length of
757     the lookbehind.
758     <a name="characterclass"></a></P>
759 nigel 93 <br><a name="SEC7" href="#TOC1">SQUARE BRACKETS AND CHARACTER CLASSES</a><br>
760 nigel 63 <P>
761     An opening square bracket introduces a character class, terminated by a closing
762     square bracket. A closing square bracket on its own is not special. If a
763     closing square bracket is required as a member of the class, it should be the
764     first data character in the class (after an initial circumflex, if present) or
765     escaped with a backslash.
766     </P>
767     <P>
768     A character class matches a single character in the subject. In UTF-8 mode, the
769     character may occupy more than one byte. A matched character must be in the set
770     of characters defined by the class, unless the first character in the class
771     definition is a circumflex, in which case the subject character must not be in
772     the set defined by the class. If a circumflex is actually required as a member
773     of the class, ensure it is not the first character, or escape it with a
774     backslash.
775     </P>
776     <P>
777     For example, the character class [aeiou] matches any lower case vowel, while
778     [^aeiou] matches any character that is not a lower case vowel. Note that a
779 nigel 75 circumflex is just a convenient notation for specifying the characters that
780     are in the class by enumerating those that are not. A class that starts with a
781     circumflex is not an assertion: it still consumes a character from the subject
782     string, and therefore it fails if the current pointer is at the end of the
783     string.
784 nigel 63 </P>
785     <P>
786     In UTF-8 mode, characters with values greater than 255 can be included in a
787     class as a literal string of bytes, or by using the \x{ escaping mechanism.
788     </P>
789     <P>
790     When caseless matching is set, any letters in a class represent both their
791     upper case and lower case versions, so for example, a caseless [aeiou] matches
792     "A" as well as "a", and a caseless [^aeiou] does not match "A", whereas a
793 nigel 77 caseful version would. In UTF-8 mode, PCRE always understands the concept of
794     case for characters whose values are less than 128, so caseless matching is
795     always possible. For characters with higher values, the concept of case is
796     supported if PCRE is compiled with Unicode property support, but not otherwise.
797     If you want to use caseless matching for characters 128 and above, you must
798     ensure that PCRE is compiled with Unicode property support as well as with
799     UTF-8 support.
800 nigel 63 </P>
801     <P>
802 nigel 93 Characters that might indicate line breaks are never treated in any special way
803     when matching character classes, whatever line-ending sequence is in use, and
804     whatever setting of the PCRE_DOTALL and PCRE_MULTILINE options is used. A class
805     such as [^a] always matches one of these characters.
806 nigel 63 </P>
807     <P>
808     The minus (hyphen) character can be used to specify a range of characters in a
809     character class. For example, [d-m] matches any letter between d and m,
810     inclusive. If a minus character is required in a class, it must be escaped with
811     a backslash or appear in a position where it cannot be interpreted as
812     indicating a range, typically as the first or last character in the class.
813     </P>
814     <P>
815     It is not possible to have the literal character "]" as the end character of a
816     range. A pattern such as [W-]46] is interpreted as a class of two characters
817     ("W" and "-") followed by a literal string "46]", so it would match "W46]" or
818     "-46]". However, if the "]" is escaped with a backslash it is interpreted as
819 nigel 75 the end of range, so [W-\]46] is interpreted as a class containing a range
820     followed by two other characters. The octal or hexadecimal representation of
821     "]" can also be used to end a range.
822 nigel 63 </P>
823     <P>
824     Ranges operate in the collating sequence of character values. They can also be
825     used for characters specified numerically, for example [\000-\037]. In UTF-8
826     mode, ranges can include characters whose values are greater than 255, for
827     example [\x{100}-\x{2ff}].
828     </P>
829     <P>
830     If a range that includes letters is used when caseless matching is set, it
831     matches the letters in either case. For example, [W-c] is equivalent to
832 nigel 75 [][\\^_`wxyzabc], matched caselessly, and in non-UTF-8 mode, if character
833 ph10 142 tables for a French locale are in use, [\xc8-\xcb] matches accented E
834 nigel 75 characters in both cases. In UTF-8 mode, PCRE supports the concept of case for
835     characters with values greater than 128 only when it is compiled with Unicode
836     property support.
837 nigel 63 </P>
838     <P>
839 nigel 75 The character types \d, \D, \p, \P, \s, \S, \w, and \W may also appear
840     in a character class, and add the characters that they match to the class. For
841 nigel 63 example, [\dABCDEF] matches any hexadecimal digit. A circumflex can
842     conveniently be used with the upper case character types to specify a more
843     restricted set of characters than the matching lower case type. For example,
844     the class [^\W_] matches any letter or digit, but not underscore.
845     </P>
846     <P>
847 nigel 75 The only metacharacters that are recognized in character classes are backslash,
848     hyphen (only where it can be interpreted as specifying a range), circumflex
849     (only at the start), opening square bracket (only when it can be interpreted as
850     introducing a POSIX class name - see the next section), and the terminating
851     closing square bracket. However, escaping other non-alphanumeric characters
852     does no harm.
853 nigel 63 </P>
854 nigel 93 <br><a name="SEC8" href="#TOC1">POSIX CHARACTER CLASSES</a><br>
855 nigel 63 <P>
856 nigel 75 Perl supports the POSIX notation for character classes. This uses names
857 nigel 63 enclosed by [: and :] within the enclosing square brackets. PCRE also supports
858     this notation. For example,
859     <pre>
860     [01[:alpha:]%]
861 nigel 75 </pre>
862 nigel 63 matches "0", "1", any alphabetic character, or "%". The supported class names
863     are
864     <pre>
865     alnum letters and digits
866     alpha letters
867     ascii character codes 0 - 127
868     blank space or tab only
869     cntrl control characters
870     digit decimal digits (same as \d)
871     graph printing characters, excluding space
872     lower lower case letters
873     print printing characters, including space
874     punct printing characters, excluding letters and digits
875     space white space (not quite the same as \s)
876     upper upper case letters
877     word "word" characters (same as \w)
878     xdigit hexadecimal digits
879 nigel 75 </pre>
880 nigel 63 The "space" characters are HT (9), LF (10), VT (11), FF (12), CR (13), and
881     space (32). Notice that this list includes the VT character (code 11). This
882     makes "space" different to \s, which does not include VT (for Perl
883     compatibility).
884     </P>
885     <P>
886     The name "word" is a Perl extension, and "blank" is a GNU extension from Perl
887     5.8. Another Perl extension is negation, which is indicated by a ^ character
888     after the colon. For example,
889     <pre>
890     [12[:^digit:]]
891 nigel 75 </pre>
892 nigel 63 matches "1", "2", or any non-digit. PCRE (and Perl) also recognize the POSIX
893     syntax [.ch.] and [=ch=] where "ch" is a "collating element", but these are not
894     supported, and an error is given if they are encountered.
895     </P>
896     <P>
897 nigel 75 In UTF-8 mode, characters with values greater than 128 do not match any of
898 nigel 63 the POSIX character classes.
899     </P>
900 nigel 93 <br><a name="SEC9" href="#TOC1">VERTICAL BAR</a><br>
901 nigel 63 <P>
902     Vertical bar characters are used to separate alternative patterns. For example,
903     the pattern
904     <pre>
905     gilbert|sullivan
906 nigel 75 </pre>
907 nigel 63 matches either "gilbert" or "sullivan". Any number of alternatives may appear,
908 nigel 91 and an empty alternative is permitted (matching the empty string). The matching
909     process tries each alternative in turn, from left to right, and the first one
910     that succeeds is used. If the alternatives are within a subpattern
911 nigel 75 <a href="#subpattern">(defined below),</a>
912     "succeeds" means matching the rest of the main pattern as well as the
913     alternative in the subpattern.
914 nigel 63 </P>
915 nigel 93 <br><a name="SEC10" href="#TOC1">INTERNAL OPTION SETTING</a><br>
916 nigel 63 <P>
917     The settings of the PCRE_CASELESS, PCRE_MULTILINE, PCRE_DOTALL, and
918     PCRE_EXTENDED options can be changed from within the pattern by a sequence of
919     Perl option letters enclosed between "(?" and ")". The option letters are
920     <pre>
921     i for PCRE_CASELESS
922     m for PCRE_MULTILINE
923     s for PCRE_DOTALL
924     x for PCRE_EXTENDED
925 nigel 75 </pre>
926 nigel 63 For example, (?im) sets caseless, multiline matching. It is also possible to
927     unset these options by preceding the letter with a hyphen, and a combined
928     setting and unsetting such as (?im-sx), which sets PCRE_CASELESS and
929     PCRE_MULTILINE while unsetting PCRE_DOTALL and PCRE_EXTENDED, is also
930     permitted. If a letter appears both before and after the hyphen, the option is
931     unset.
932     </P>
933     <P>
934     When an option change occurs at top level (that is, not inside subpattern
935     parentheses), the change applies to the remainder of the pattern that follows.
936     If the change is placed right at the start of a pattern, PCRE extracts it into
937     the global options (and it will therefore show up in data extracted by the
938     <b>pcre_fullinfo()</b> function).
939     </P>
940     <P>
941 nigel 93 An option change within a subpattern (see below for a description of
942     subpatterns) affects only that part of the current pattern that follows it, so
943 nigel 63 <pre>
944     (a(?i)b)c
945 nigel 75 </pre>
946 nigel 63 matches abc and aBc and no other strings (assuming PCRE_CASELESS is not used).
947     By this means, options can be made to have different settings in different
948     parts of the pattern. Any changes made in one alternative do carry on
949     into subsequent branches within the same subpattern. For example,
950     <pre>
951     (a(?i)b|c)
952 nigel 75 </pre>
953 nigel 63 matches "ab", "aB", "c", and "C", even though when matching "C" the first
954     branch is abandoned before the option setting. This is because the effects of
955     option settings happen at compile time. There would be some very weird
956     behaviour otherwise.
957     </P>
958     <P>
959 nigel 91 The PCRE-specific options PCRE_DUPNAMES, PCRE_UNGREEDY, and PCRE_EXTRA can be
960     changed in the same way as the Perl-compatible options by using the characters
961     J, U and X respectively.
962 nigel 75 <a name="subpattern"></a></P>
963 nigel 93 <br><a name="SEC11" href="#TOC1">SUBPATTERNS</a><br>
964 nigel 63 <P>
965     Subpatterns are delimited by parentheses (round brackets), which can be nested.
966 nigel 75 Turning part of a pattern into a subpattern does two things:
967     <br>
968     <br>
969 nigel 63 1. It localizes a set of alternatives. For example, the pattern
970     <pre>
971     cat(aract|erpillar|)
972 nigel 75 </pre>
973 nigel 63 matches one of the words "cat", "cataract", or "caterpillar". Without the
974 nigel 93 parentheses, it would match "cataract", "erpillar" or an empty string.
975 nigel 75 <br>
976     <br>
977     2. It sets up the subpattern as a capturing subpattern. This means that, when
978     the whole pattern matches, that portion of the subject string that matched the
979     subpattern is passed back to the caller via the <i>ovector</i> argument of
980 nigel 63 <b>pcre_exec()</b>. Opening parentheses are counted from left to right (starting
981 nigel 75 from 1) to obtain numbers for the capturing subpatterns.
982 nigel 63 </P>
983     <P>
984     For example, if the string "the red king" is matched against the pattern
985     <pre>
986     the ((red|white) (king|queen))
987 nigel 75 </pre>
988 nigel 63 the captured substrings are "red king", "red", and "king", and are numbered 1,
989     2, and 3, respectively.
990     </P>
991     <P>
992     The fact that plain parentheses fulfil two functions is not always helpful.
993     There are often times when a grouping subpattern is required without a
994     capturing requirement. If an opening parenthesis is followed by a question mark
995     and a colon, the subpattern does not do any capturing, and is not counted when
996     computing the number of any subsequent capturing subpatterns. For example, if
997     the string "the white queen" is matched against the pattern
998     <pre>
999     the ((?:red|white) (king|queen))
1000 nigel 75 </pre>
1001 nigel 63 the captured substrings are "white queen" and "queen", and are numbered 1 and
1002 nigel 93 2. The maximum number of capturing subpatterns is 65535.
1003 nigel 63 </P>
1004     <P>
1005     As a convenient shorthand, if any option settings are required at the start of
1006     a non-capturing subpattern, the option letters may appear between the "?" and
1007     the ":". Thus the two patterns
1008     <pre>
1009     (?i:saturday|sunday)
1010     (?:(?i)saturday|sunday)
1011 nigel 75 </pre>
1012 nigel 63 match exactly the same set of strings. Because alternative branches are tried
1013     from left to right, and options are not reset until the end of the subpattern
1014     is reached, an option setting in one branch does affect subsequent branches, so
1015     the above patterns match "SUNDAY" as well as "Saturday".
1016     </P>
1017 ph10 182 <br><a name="SEC12" href="#TOC1">DUPLICATE SUBPATTERN NUMBERS</a><br>
1018 nigel 63 <P>
1019 ph10 182 Perl 5.10 introduced a feature whereby each alternative in a subpattern uses
1020     the same numbers for its capturing parentheses. Such a subpattern starts with
1021     (?| and is itself a non-capturing subpattern. For example, consider this
1022     pattern:
1023     <pre>
1024     (?|(Sat)ur|(Sun))day
1025     </pre>
1026     Because the two alternatives are inside a (?| group, both sets of capturing
1027     parentheses are numbered one. Thus, when the pattern matches, you can look
1028     at captured substring number one, whichever alternative matched. This construct
1029     is useful when you want to capture part, but not all, of one of a number of
1030     alternatives. Inside a (?| group, parentheses are numbered as usual, but the
1031     number is reset at the start of each branch. The numbers of any capturing
1032     buffers that follow the subpattern start after the highest number used in any
1033     branch. The following example is taken from the Perl documentation.
1034     The numbers underneath show in which buffer the captured content will be
1035     stored.
1036     <pre>
1037     # before ---------------branch-reset----------- after
1038     / ( a ) (?| x ( y ) z | (p (q) r) | (t) u (v) ) ( z ) /x
1039     # 1 2 2 3 2 3 4
1040     </pre>
1041     A backreference or a recursive call to a numbered subpattern always refers to
1042     the first one in the pattern with the given number.
1043     </P>
1044     <P>
1045     An alternative approach to using this "branch reset" feature is to use
1046     duplicate named subpatterns, as described in the next section.
1047     </P>
1048     <br><a name="SEC13" href="#TOC1">NAMED SUBPATTERNS</a><br>
1049     <P>
1050 nigel 63 Identifying capturing parentheses by number is simple, but it can be very hard
1051     to keep track of the numbers in complicated regular expressions. Furthermore,
1052 nigel 75 if an expression is modified, the numbers may change. To help with this
1053 nigel 93 difficulty, PCRE supports the naming of subpatterns. This feature was not
1054     added to Perl until release 5.10. Python had the feature earlier, and PCRE
1055     introduced it at release 4.0, using the Python syntax. PCRE now supports both
1056     the Perl and the Python syntax.
1057     </P>
1058     <P>
1059     In PCRE, a subpattern can be named in one of three ways: (?&#60;name&#62;...) or
1060     (?'name'...) as in Perl, or (?P&#60;name&#62;...) as in Python. References to capturing
1061 nigel 91 parentheses from other parts of the pattern, such as
1062     <a href="#backreferences">backreferences,</a>
1063     <a href="#recursion">recursion,</a>
1064     and
1065     <a href="#conditions">conditions,</a>
1066     can be made by name as well as by number.
1067 nigel 63 </P>
1068     <P>
1069 nigel 91 Names consist of up to 32 alphanumeric characters and underscores. Named
1070 nigel 93 capturing parentheses are still allocated numbers as well as names, exactly as
1071     if the names were not present. The PCRE API provides function calls for
1072     extracting the name-to-number translation table from a compiled pattern. There
1073     is also a convenience function for extracting a captured substring by name.
1074 nigel 91 </P>
1075     <P>
1076     By default, a name must be unique within a pattern, but it is possible to relax
1077     this constraint by setting the PCRE_DUPNAMES option at compile time. This can
1078     be useful for patterns where only one instance of the named parentheses can
1079     match. Suppose you want to match the name of a weekday, either as a 3-letter
1080     abbreviation or as the full name, and in both cases you want to extract the
1081     abbreviation. This pattern (ignoring the line breaks) does the job:
1082     <pre>
1083 nigel 93 (?&#60;DN&#62;Mon|Fri|Sun)(?:day)?|
1084     (?&#60;DN&#62;Tue)(?:sday)?|
1085     (?&#60;DN&#62;Wed)(?:nesday)?|
1086     (?&#60;DN&#62;Thu)(?:rsday)?|
1087     (?&#60;DN&#62;Sat)(?:urday)?
1088 nigel 91 </pre>
1089     There are five capturing substrings, but only one is ever set after a match.
1090 ph10 182 (An alternative way of solving this problem is to use a "branch reset"
1091     subpattern, as described in the previous section.)
1092     </P>
1093     <P>
1094 nigel 91 The convenience function for extracting the data by name returns the substring
1095 nigel 93 for the first (and in this example, the only) subpattern of that name that
1096 nigel 91 matched. This saves searching to find which numbered subpattern it was. If you
1097     make a reference to a non-unique named subpattern from elsewhere in the
1098     pattern, the one that corresponds to the lowest number is used. For further
1099     details of the interfaces for handling named subpatterns, see the
1100 nigel 63 <a href="pcreapi.html"><b>pcreapi</b></a>
1101     documentation.
1102     </P>
1103 ph10 182 <br><a name="SEC14" href="#TOC1">REPETITION</a><br>
1104 nigel 63 <P>
1105     Repetition is specified by quantifiers, which can follow any of the following
1106     items:
1107     <pre>
1108     a literal data character
1109 nigel 93 the dot metacharacter
1110 nigel 63 the \C escape sequence
1111 nigel 75 the \X escape sequence (in UTF-8 mode with Unicode properties)
1112 nigel 93 the \R escape sequence
1113 nigel 75 an escape such as \d that matches a single character
1114 nigel 63 a character class
1115     a back reference (see next section)
1116     a parenthesized subpattern (unless it is an assertion)
1117 nigel 75 </pre>
1118 nigel 63 The general repetition quantifier specifies a minimum and maximum number of
1119     permitted matches, by giving the two numbers in curly brackets (braces),
1120     separated by a comma. The numbers must be less than 65536, and the first must
1121     be less than or equal to the second. For example:
1122     <pre>
1123     z{2,4}
1124 nigel 75 </pre>
1125 nigel 63 matches "zz", "zzz", or "zzzz". A closing brace on its own is not a special
1126     character. If the second number is omitted, but the comma is present, there is
1127     no upper limit; if the second number and the comma are both omitted, the
1128     quantifier specifies an exact number of required matches. Thus
1129     <pre>
1130     [aeiou]{3,}
1131 nigel 75 </pre>
1132 nigel 63 matches at least 3 successive vowels, but may match many more, while
1133     <pre>
1134     \d{8}
1135 nigel 75 </pre>
1136 nigel 63 matches exactly 8 digits. An opening curly bracket that appears in a position
1137     where a quantifier is not allowed, or one that does not match the syntax of a
1138     quantifier, is taken as a literal character. For example, {,6} is not a
1139     quantifier, but a literal string of four characters.
1140     </P>
1141     <P>
1142     In UTF-8 mode, quantifiers apply to UTF-8 characters rather than to individual
1143     bytes. Thus, for example, \x{100}{2} matches two UTF-8 characters, each of
1144 nigel 75 which is represented by a two-byte sequence. Similarly, when Unicode property
1145     support is available, \X{3} matches three Unicode extended sequences, each of
1146     which may be several bytes long (and they may be of different lengths).
1147 nigel 63 </P>
1148     <P>
1149     The quantifier {0} is permitted, causing the expression to behave as if the
1150     previous item and the quantifier were not present.
1151     </P>
1152     <P>
1153 nigel 93 For convenience, the three most common quantifiers have single-character
1154     abbreviations:
1155 nigel 63 <pre>
1156     * is equivalent to {0,}
1157     + is equivalent to {1,}
1158     ? is equivalent to {0,1}
1159 nigel 75 </pre>
1160 nigel 63 It is possible to construct infinite loops by following a subpattern that can
1161     match no characters with a quantifier that has no upper limit, for example:
1162     <pre>
1163     (a?)*
1164 nigel 75 </pre>
1165 nigel 63 Earlier versions of Perl and PCRE used to give an error at compile time for
1166     such patterns. However, because there are cases where this can be useful, such
1167     patterns are now accepted, but if any repetition of the subpattern does in fact
1168     match no characters, the loop is forcibly broken.
1169     </P>
1170     <P>
1171     By default, the quantifiers are "greedy", that is, they match as much as
1172     possible (up to the maximum number of permitted times), without causing the
1173     rest of the pattern to fail. The classic example of where this gives problems
1174 nigel 75 is in trying to match comments in C programs. These appear between /* and */
1175     and within the comment, individual * and / characters may appear. An attempt to
1176     match C comments by applying the pattern
1177 nigel 63 <pre>
1178     /\*.*\*/
1179 nigel 75 </pre>
1180 nigel 63 to the string
1181     <pre>
1182 nigel 75 /* first comment */ not comment /* second comment */
1183     </pre>
1184 nigel 63 fails, because it matches the entire string owing to the greediness of the .*
1185     item.
1186     </P>
1187     <P>
1188     However, if a quantifier is followed by a question mark, it ceases to be
1189     greedy, and instead matches the minimum number of times possible, so the
1190     pattern
1191     <pre>
1192     /\*.*?\*/
1193 nigel 75 </pre>
1194 nigel 63 does the right thing with the C comments. The meaning of the various
1195     quantifiers is not otherwise changed, just the preferred number of matches.
1196     Do not confuse this use of question mark with its use as a quantifier in its
1197     own right. Because it has two uses, it can sometimes appear doubled, as in
1198     <pre>
1199     \d??\d
1200 nigel 75 </pre>
1201 nigel 63 which matches one digit by preference, but can match two if that is the only
1202     way the rest of the pattern matches.
1203     </P>
1204     <P>
1205 nigel 93 If the PCRE_UNGREEDY option is set (an option that is not available in Perl),
1206 nigel 63 the quantifiers are not greedy by default, but individual ones can be made
1207     greedy by following them with a question mark. In other words, it inverts the
1208     default behaviour.
1209     </P>
1210     <P>
1211     When a parenthesized subpattern is quantified with a minimum repeat count that
1212 nigel 75 is greater than 1 or with a limited maximum, more memory is required for the
1213 nigel 63 compiled pattern, in proportion to the size of the minimum or maximum.
1214     </P>
1215     <P>
1216     If a pattern starts with .* or .{0,} and the PCRE_DOTALL option (equivalent
1217 nigel 93 to Perl's /s) is set, thus allowing the dot to match newlines, the pattern is
1218 nigel 63 implicitly anchored, because whatever follows will be tried against every
1219     character position in the subject string, so there is no point in retrying the
1220     overall match at any position after the first. PCRE normally treats such a
1221     pattern as though it were preceded by \A.
1222     </P>
1223     <P>
1224     In cases where it is known that the subject string contains no newlines, it is
1225     worth setting PCRE_DOTALL in order to obtain this optimization, or
1226     alternatively using ^ to indicate anchoring explicitly.
1227     </P>
1228     <P>
1229     However, there is one situation where the optimization cannot be used. When .*
1230     is inside capturing parentheses that are the subject of a backreference
1231 nigel 93 elsewhere in the pattern, a match at the start may fail where a later one
1232     succeeds. Consider, for example:
1233 nigel 63 <pre>
1234     (.*)abc\1
1235 nigel 75 </pre>
1236 nigel 63 If the subject is "xyz123abc123" the match point is the fourth character. For
1237     this reason, such a pattern is not implicitly anchored.
1238     </P>
1239     <P>
1240     When a capturing subpattern is repeated, the value captured is the substring
1241     that matched the final iteration. For example, after
1242     <pre>
1243     (tweedle[dume]{3}\s*)+
1244 nigel 75 </pre>
1245 nigel 63 has matched "tweedledum tweedledee" the value of the captured substring is
1246     "tweedledee". However, if there are nested capturing subpatterns, the
1247     corresponding captured values may have been set in previous iterations. For
1248     example, after
1249     <pre>
1250     /(a|(b))+/
1251 nigel 75 </pre>
1252 nigel 63 matches "aba" the value of the second captured substring is "b".
1253 nigel 75 <a name="atomicgroup"></a></P>
1254 ph10 182 <br><a name="SEC15" href="#TOC1">ATOMIC GROUPING AND POSSESSIVE QUANTIFIERS</a><br>
1255 nigel 63 <P>
1256 nigel 93 With both maximizing ("greedy") and minimizing ("ungreedy" or "lazy")
1257     repetition, failure of what follows normally causes the repeated item to be
1258     re-evaluated to see if a different number of repeats allows the rest of the
1259     pattern to match. Sometimes it is useful to prevent this, either to change the
1260     nature of the match, or to cause it fail earlier than it otherwise might, when
1261     the author of the pattern knows there is no point in carrying on.
1262 nigel 63 </P>
1263     <P>
1264     Consider, for example, the pattern \d+foo when applied to the subject line
1265     <pre>
1266     123456bar
1267 nigel 75 </pre>
1268 nigel 63 After matching all 6 digits and then failing to match "foo", the normal
1269     action of the matcher is to try again with only 5 digits matching the \d+
1270     item, and then with 4, and so on, before ultimately failing. "Atomic grouping"
1271     (a term taken from Jeffrey Friedl's book) provides the means for specifying
1272     that once a subpattern has matched, it is not to be re-evaluated in this way.
1273     </P>
1274     <P>
1275 nigel 93 If we use atomic grouping for the previous example, the matcher gives up
1276 nigel 63 immediately on failing to match "foo" the first time. The notation is a kind of
1277     special parenthesis, starting with (?&#62; as in this example:
1278     <pre>
1279 nigel 73 (?&#62;\d+)foo
1280 nigel 75 </pre>
1281 nigel 63 This kind of parenthesis "locks up" the part of the pattern it contains once
1282     it has matched, and a failure further into the pattern is prevented from
1283     backtracking into it. Backtracking past it to previous items, however, works as
1284     normal.
1285     </P>
1286     <P>
1287     An alternative description is that a subpattern of this type matches the string
1288     of characters that an identical standalone pattern would match, if anchored at
1289     the current point in the subject string.
1290     </P>
1291     <P>
1292     Atomic grouping subpatterns are not capturing subpatterns. Simple cases such as
1293     the above example can be thought of as a maximizing repeat that must swallow
1294     everything it can. So, while both \d+ and \d+? are prepared to adjust the
1295     number of digits they match in order to make the rest of the pattern match,
1296     (?&#62;\d+) can only match an entire sequence of digits.
1297     </P>
1298     <P>
1299     Atomic groups in general can of course contain arbitrarily complicated
1300     subpatterns, and can be nested. However, when the subpattern for an atomic
1301     group is just a single repeated item, as in the example above, a simpler
1302     notation, called a "possessive quantifier" can be used. This consists of an
1303     additional + character following a quantifier. Using this notation, the
1304     previous example can be rewritten as
1305     <pre>
1306 nigel 75 \d++foo
1307     </pre>
1308 ph10 208 Note that a possessive quantifier can be used with an entire group, for
1309     example:
1310     <pre>
1311     (abc|xyz){2,3}+
1312     </pre>
1313 nigel 63 Possessive quantifiers are always greedy; the setting of the PCRE_UNGREEDY
1314     option is ignored. They are a convenient notation for the simpler forms of
1315 nigel 93 atomic group. However, there is no difference in the meaning of a possessive
1316     quantifier and the equivalent atomic group, though there may be a performance
1317     difference; possessive quantifiers should be slightly faster.
1318 nigel 63 </P>
1319     <P>
1320 nigel 93 The possessive quantifier syntax is an extension to the Perl 5.8 syntax.
1321     Jeffrey Friedl originated the idea (and the name) in the first edition of his
1322     book. Mike McCloskey liked it, so implemented it when he built Sun's Java
1323     package, and PCRE copied it from there. It ultimately found its way into Perl
1324     at release 5.10.
1325 nigel 63 </P>
1326     <P>
1327 nigel 93 PCRE has an optimization that automatically "possessifies" certain simple
1328     pattern constructs. For example, the sequence A+B is treated as A++B because
1329     there is no point in backtracking into a sequence of A's when B must follow.
1330     </P>
1331     <P>
1332 nigel 63 When a pattern contains an unlimited repeat inside a subpattern that can itself
1333     be repeated an unlimited number of times, the use of an atomic group is the
1334     only way to avoid some failing matches taking a very long time indeed. The
1335     pattern
1336     <pre>
1337     (\D+|&#60;\d+&#62;)*[!?]
1338 nigel 75 </pre>
1339 nigel 63 matches an unlimited number of substrings that either consist of non-digits, or
1340     digits enclosed in &#60;&#62;, followed by either ! or ?. When it matches, it runs
1341     quickly. However, if it is applied to
1342     <pre>
1343     aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
1344 nigel 75 </pre>
1345 nigel 63 it takes a long time before reporting failure. This is because the string can
1346 nigel 75 be divided between the internal \D+ repeat and the external * repeat in a
1347     large number of ways, and all have to be tried. (The example uses [!?] rather
1348     than a single character at the end, because both PCRE and Perl have an
1349     optimization that allows for fast failure when a single character is used. They
1350     remember the last single character that is required for a match, and fail early
1351     if it is not present in the string.) If the pattern is changed so that it uses
1352     an atomic group, like this:
1353 nigel 63 <pre>
1354     ((?&#62;\D+)|&#60;\d+&#62;)*[!?]
1355 nigel 75 </pre>
1356 nigel 63 sequences of non-digits cannot be broken, and failure happens quickly.
1357 nigel 75 <a name="backreferences"></a></P>
1358 ph10 182 <br><a name="SEC16" href="#TOC1">BACK REFERENCES</a><br>
1359 nigel 63 <P>
1360     Outside a character class, a backslash followed by a digit greater than 0 (and
1361     possibly further digits) is a back reference to a capturing subpattern earlier
1362     (that is, to its left) in the pattern, provided there have been that many
1363     previous capturing left parentheses.
1364     </P>
1365     <P>
1366     However, if the decimal number following the backslash is less than 10, it is
1367     always taken as a back reference, and causes an error only if there are not
1368     that many capturing left parentheses in the entire pattern. In other words, the
1369     parentheses that are referenced need not be to the left of the reference for
1370 nigel 91 numbers less than 10. A "forward back reference" of this type can make sense
1371     when a repetition is involved and the subpattern to the right has participated
1372     in an earlier iteration.
1373     </P>
1374     <P>
1375 nigel 93 It is not possible to have a numerical "forward back reference" to a subpattern
1376     whose number is 10 or more using this syntax because a sequence such as \50 is
1377     interpreted as a character defined in octal. See the subsection entitled
1378 nigel 91 "Non-printing characters"
1379 nigel 75 <a href="#digitsafterbackslash">above</a>
1380 nigel 93 for further details of the handling of digits following a backslash. There is
1381     no such problem when named parentheses are used. A back reference to any
1382     subpattern is possible using named parentheses (see below).
1383 nigel 63 </P>
1384     <P>
1385 nigel 93 Another way of avoiding the ambiguity inherent in the use of digits following a
1386     backslash is to use the \g escape sequence, which is a feature introduced in
1387 ph10 208 Perl 5.10. This escape must be followed by an unsigned number or a negative
1388     number, optionally enclosed in braces. These examples are all identical:
1389 nigel 93 <pre>
1390     (ring), \1
1391     (ring), \g1
1392     (ring), \g{1}
1393     </pre>
1394 ph10 208 An unsigned number specifies an absolute reference without the ambiguity that
1395     is present in the older syntax. It is also useful when literal digits follow
1396     the reference. A negative number is a relative reference. Consider this
1397     example:
1398 nigel 93 <pre>
1399     (abc(def)ghi)\g{-1}
1400     </pre>
1401     The sequence \g{-1} is a reference to the most recently started capturing
1402     subpattern before \g, that is, is it equivalent to \2. Similarly, \g{-2}
1403     would be equivalent to \1. The use of relative references can be helpful in
1404     long patterns, and also in patterns that are created by joining together
1405     fragments that contain references within themselves.
1406     </P>
1407     <P>
1408 nigel 63 A back reference matches whatever actually matched the capturing subpattern in
1409     the current subject string, rather than anything matching the subpattern
1410     itself (see
1411     <a href="#subpatternsassubroutines">"Subpatterns as subroutines"</a>
1412     below for a way of doing that). So the pattern
1413     <pre>
1414     (sens|respons)e and \1ibility
1415 nigel 75 </pre>
1416 nigel 63 matches "sense and sensibility" and "response and responsibility", but not
1417     "sense and responsibility". If caseful matching is in force at the time of the
1418     back reference, the case of letters is relevant. For example,
1419     <pre>
1420     ((?i)rah)\s+\1
1421 nigel 75 </pre>
1422 nigel 63 matches "rah rah" and "RAH RAH", but not "RAH rah", even though the original
1423     capturing subpattern is matched caselessly.
1424     </P>
1425     <P>
1426 ph10 172 There are several different ways of writing back references to named
1427     subpatterns. The .NET syntax \k{name} and the Perl syntax \k&#60;name&#62; or
1428     \k'name' are supported, as is the Python syntax (?P=name). Perl 5.10's unified
1429     back reference syntax, in which \g can be used for both numeric and named
1430     references, is also supported. We could rewrite the above example in any of
1431 nigel 93 the following ways:
1432 nigel 63 <pre>
1433 nigel 93 (?&#60;p1&#62;(?i)rah)\s+\k&#60;p1&#62;
1434 ph10 172 (?'p1'(?i)rah)\s+\k{p1}
1435 nigel 91 (?P&#60;p1&#62;(?i)rah)\s+(?P=p1)
1436 ph10 172 (?&#60;p1&#62;(?i)rah)\s+\g{p1}
1437 nigel 75 </pre>
1438 nigel 91 A subpattern that is referenced by name may appear in the pattern before or
1439     after the reference.
1440     </P>
1441     <P>
1442 nigel 63 There may be more than one back reference to the same subpattern. If a
1443     subpattern has not actually been used in a particular match, any back
1444     references to it always fail. For example, the pattern
1445     <pre>
1446     (a|(bc))\2
1447 nigel 75 </pre>
1448 nigel 63 always fails if it starts to match "a" rather than "bc". Because there may be
1449     many capturing parentheses in a pattern, all digits following the backslash are
1450     taken as part of a potential back reference number. If the pattern continues
1451     with a digit character, some delimiter must be used to terminate the back
1452     reference. If the PCRE_EXTENDED option is set, this can be whitespace.
1453 nigel 75 Otherwise an empty comment (see
1454     <a href="#comments">"Comments"</a>
1455     below) can be used.
1456 nigel 63 </P>
1457     <P>
1458     A back reference that occurs inside the parentheses to which it refers fails
1459     when the subpattern is first used, so, for example, (a\1) never matches.
1460     However, such references can be useful inside repeated subpatterns. For
1461     example, the pattern
1462     <pre>
1463     (a|b\1)+
1464 nigel 75 </pre>
1465 nigel 63 matches any number of "a"s and also "aba", "ababbaa" etc. At each iteration of
1466     the subpattern, the back reference matches the character string corresponding
1467     to the previous iteration. In order for this to work, the pattern must be such
1468     that the first iteration does not need to match the back reference. This can be
1469     done using alternation, as in the example above, or by a quantifier with a
1470     minimum of zero.
1471 nigel 75 <a name="bigassertions"></a></P>
1472 ph10 182 <br><a name="SEC17" href="#TOC1">ASSERTIONS</a><br>
1473 nigel 63 <P>
1474     An assertion is a test on the characters following or preceding the current
1475     matching point that does not actually consume any characters. The simple
1476 nigel 75 assertions coded as \b, \B, \A, \G, \Z, \z, ^ and $ are described
1477     <a href="#smallassertions">above.</a>
1478     </P>
1479     <P>
1480 nigel 63 More complicated assertions are coded as subpatterns. There are two kinds:
1481     those that look ahead of the current position in the subject string, and those
1482 nigel 75 that look behind it. An assertion subpattern is matched in the normal way,
1483     except that it does not cause the current matching position to be changed.
1484 nigel 63 </P>
1485     <P>
1486 nigel 75 Assertion subpatterns are not capturing subpatterns, and may not be repeated,
1487     because it makes no sense to assert the same thing several times. If any kind
1488     of assertion contains capturing subpatterns within it, these are counted for
1489     the purposes of numbering the capturing subpatterns in the whole pattern.
1490     However, substring capturing is carried out only for positive assertions,
1491     because it does not make sense for negative assertions.
1492 nigel 63 </P>
1493 nigel 75 <br><b>
1494     Lookahead assertions
1495     </b><br>
1496 nigel 63 <P>
1497 nigel 91 Lookahead assertions start with (?= for positive assertions and (?! for
1498     negative assertions. For example,
1499 nigel 63 <pre>
1500     \w+(?=;)
1501 nigel 75 </pre>
1502 nigel 63 matches a word followed by a semicolon, but does not include the semicolon in
1503     the match, and
1504     <pre>
1505     foo(?!bar)
1506 nigel 75 </pre>
1507 nigel 63 matches any occurrence of "foo" that is not followed by "bar". Note that the
1508     apparently similar pattern
1509     <pre>
1510     (?!foo)bar
1511 nigel 75 </pre>
1512 nigel 63 does not find an occurrence of "bar" that is preceded by something other than
1513     "foo"; it finds any occurrence of "bar" whatsoever, because the assertion
1514     (?!foo) is always true when the next three characters are "bar". A
1515 nigel 75 lookbehind assertion is needed to achieve the other effect.
1516 nigel 63 </P>
1517     <P>
1518     If you want to force a matching failure at some point in a pattern, the most
1519     convenient way to do it is with (?!) because an empty string always matches, so
1520     an assertion that requires there not to be an empty string must always fail.
1521 nigel 75 <a name="lookbehind"></a></P>
1522     <br><b>
1523     Lookbehind assertions
1524     </b><br>
1525 nigel 63 <P>
1526     Lookbehind assertions start with (?&#60;= for positive assertions and (?&#60;! for
1527     negative assertions. For example,
1528     <pre>
1529     (?&#60;!foo)bar
1530 nigel 75 </pre>
1531 nigel 63 does find an occurrence of "bar" that is not preceded by "foo". The contents of
1532     a lookbehind assertion are restricted such that all the strings it matches must
1533 nigel 91 have a fixed length. However, if there are several top-level alternatives, they
1534     do not all have to have the same fixed length. Thus
1535 nigel 63 <pre>
1536     (?&#60;=bullock|donkey)
1537 nigel 75 </pre>
1538 nigel 63 is permitted, but
1539     <pre>
1540     (?&#60;!dogs?|cats?)
1541 nigel 75 </pre>
1542 nigel 63 causes an error at compile time. Branches that match different length strings
1543     are permitted only at the top level of a lookbehind assertion. This is an
1544     extension compared with Perl (at least for 5.8), which requires all branches to
1545     match the same length of string. An assertion such as
1546     <pre>
1547     (?&#60;=ab(c|de))
1548 nigel 75 </pre>
1549 nigel 63 is not permitted, because its single top-level branch can match two different
1550     lengths, but it is acceptable if rewritten to use two top-level branches:
1551     <pre>
1552     (?&#60;=abc|abde)
1553 nigel 75 </pre>
1554 ph10 172 In some cases, the Perl 5.10 escape sequence \K
1555     <a href="#resetmatchstart">(see above)</a>
1556     can be used instead of a lookbehind assertion; this is not restricted to a
1557     fixed-length.
1558     </P>
1559     <P>
1560 nigel 63 The implementation of lookbehind assertions is, for each alternative, to
1561 nigel 93 temporarily move the current position back by the fixed length and then try to
1562 nigel 63 match. If there are insufficient characters before the current position, the
1563 nigel 93 assertion fails.
1564 nigel 63 </P>
1565     <P>
1566     PCRE does not allow the \C escape (which matches a single byte in UTF-8 mode)
1567     to appear in lookbehind assertions, because it makes it impossible to calculate
1568 nigel 93 the length of the lookbehind. The \X and \R escapes, which can match
1569     different numbers of bytes, are also not permitted.
1570 nigel 63 </P>
1571     <P>
1572 nigel 93 Possessive quantifiers can be used in conjunction with lookbehind assertions to
1573     specify efficient matching at the end of the subject string. Consider a simple
1574     pattern such as
1575 nigel 63 <pre>
1576     abcd$
1577 nigel 75 </pre>
1578 nigel 63 when applied to a long string that does not match. Because matching proceeds
1579     from left to right, PCRE will look for each "a" in the subject and then see if
1580     what follows matches the rest of the pattern. If the pattern is specified as
1581     <pre>
1582     ^.*abcd$
1583 nigel 75 </pre>
1584 nigel 63 the initial .* matches the entire string at first, but when this fails (because
1585     there is no following "a"), it backtracks to match all but the last character,
1586     then all but the last two characters, and so on. Once again the search for "a"
1587     covers the entire string, from right to left, so we are no better off. However,
1588     if the pattern is written as
1589     <pre>
1590     ^.*+(?&#60;=abcd)
1591 nigel 75 </pre>
1592 nigel 93 there can be no backtracking for the .*+ item; it can match only the entire
1593 nigel 63 string. The subsequent lookbehind assertion does a single test on the last four
1594     characters. If it fails, the match fails immediately. For long strings, this
1595     approach makes a significant difference to the processing time.
1596     </P>
1597 nigel 75 <br><b>
1598     Using multiple assertions
1599     </b><br>
1600 nigel 63 <P>
1601     Several assertions (of any sort) may occur in succession. For example,
1602     <pre>
1603     (?&#60;=\d{3})(?&#60;!999)foo
1604 nigel 75 </pre>
1605 nigel 63 matches "foo" preceded by three digits that are not "999". Notice that each of
1606     the assertions is applied independently at the same point in the subject
1607     string. First there is a check that the previous three characters are all
1608     digits, and then there is a check that the same three characters are not "999".
1609     This pattern does <i>not</i> match "foo" preceded by six characters, the first
1610     of which are digits and the last three of which are not "999". For example, it
1611     doesn't match "123abcfoo". A pattern to do that is
1612     <pre>
1613     (?&#60;=\d{3}...)(?&#60;!999)foo
1614 nigel 75 </pre>
1615 nigel 63 This time the first assertion looks at the preceding six characters, checking
1616     that the first three are digits, and then the second assertion checks that the
1617     preceding three characters are not "999".
1618     </P>
1619     <P>
1620     Assertions can be nested in any combination. For example,
1621     <pre>
1622     (?&#60;=(?&#60;!foo)bar)baz
1623 nigel 75 </pre>
1624 nigel 63 matches an occurrence of "baz" that is preceded by "bar" which in turn is not
1625     preceded by "foo", while
1626     <pre>
1627     (?&#60;=\d{3}(?!999)...)foo
1628 nigel 75 </pre>
1629     is another pattern that matches "foo" preceded by three digits and any three
1630 nigel 63 characters that are not "999".
1631 nigel 91 <a name="conditions"></a></P>
1632 ph10 182 <br><a name="SEC18" href="#TOC1">CONDITIONAL SUBPATTERNS</a><br>
1633 nigel 63 <P>
1634     It is possible to cause the matching process to obey a subpattern
1635     conditionally or to choose between two alternative subpatterns, depending on
1636     the result of an assertion, or whether a previous capturing subpattern matched
1637     or not. The two possible forms of conditional subpattern are
1638     <pre>
1639     (?(condition)yes-pattern)
1640     (?(condition)yes-pattern|no-pattern)
1641 nigel 75 </pre>
1642 nigel 63 If the condition is satisfied, the yes-pattern is used; otherwise the
1643     no-pattern (if present) is used. If there are more than two alternatives in the
1644     subpattern, a compile-time error occurs.
1645     </P>
1646     <P>
1647 nigel 93 There are four kinds of condition: references to subpatterns, references to
1648     recursion, a pseudo-condition called DEFINE, and assertions.
1649 nigel 91 </P>
1650 nigel 93 <br><b>
1651     Checking for a used subpattern by number
1652     </b><br>
1653 nigel 91 <P>
1654 nigel 93 If the text between the parentheses consists of a sequence of digits, the
1655     condition is true if the capturing subpattern of that number has previously
1656 ph10 172 matched. An alternative notation is to precede the digits with a plus or minus
1657     sign. In this case, the subpattern number is relative rather than absolute.
1658     The most recently opened parentheses can be referenced by (?(-1), the next most
1659     recent by (?(-2), and so on. In looping constructs it can also make sense to
1660     refer to subsequent groups with constructs such as (?(+2).
1661 nigel 93 </P>
1662     <P>
1663 nigel 91 Consider the following pattern, which contains non-significant white space to
1664     make it more readable (assume the PCRE_EXTENDED option) and to divide it into
1665     three parts for ease of discussion:
1666 nigel 63 <pre>
1667     ( \( )? [^()]+ (?(1) \) )
1668 nigel 75 </pre>
1669 nigel 63 The first part matches an optional opening parenthesis, and if that
1670     character is present, sets it as the first captured substring. The second part
1671     matches one or more characters that are not parentheses. The third part is a
1672     conditional subpattern that tests whether the first set of parentheses matched
1673     or not. If they did, that is, if subject started with an opening parenthesis,
1674     the condition is true, and so the yes-pattern is executed and a closing
1675     parenthesis is required. Otherwise, since no-pattern is not present, the
1676     subpattern matches nothing. In other words, this pattern matches a sequence of
1677 nigel 93 non-parentheses, optionally enclosed in parentheses.
1678     </P>
1679 ph10 172 <P>
1680     If you were embedding this pattern in a larger one, you could use a relative
1681     reference:
1682     <pre>
1683     ...other stuff... ( \( )? [^()]+ (?(-1) \) ) ...
1684     </pre>
1685     This makes the fragment independent of the parentheses in the larger pattern.
1686     </P>
1687 nigel 93 <br><b>
1688     Checking for a used subpattern by name
1689     </b><br>
1690     <P>
1691     Perl uses the syntax (?(&#60;name&#62;)...) or (?('name')...) to test for a used
1692     subpattern by name. For compatibility with earlier versions of PCRE, which had
1693     this facility before Perl, the syntax (?(name)...) is also recognized. However,
1694     there is a possible ambiguity with this syntax, because subpattern names may
1695     consist entirely of digits. PCRE looks first for a named subpattern; if it
1696     cannot find one and the name consists entirely of digits, PCRE looks for a
1697     subpattern of that number, which must be greater than zero. Using subpattern
1698     names that consist entirely of digits is not recommended.
1699     </P>
1700     <P>
1701     Rewriting the above example to use a named subpattern gives this:
1702 nigel 91 <pre>
1703 nigel 93 (?&#60;OPEN&#62; \( )? [^()]+ (?(&#60;OPEN&#62;) \) )
1705     </PRE>
1706     </P>
1707     <br><b>
1708     Checking for pattern recursion
1709     </b><br>
1710     <P>
1711     If the condition is the string (R), and there is no subpattern with the name R,
1712     the condition is true if a recursive call to the whole pattern or any
1713     subpattern has been made. If digits or a name preceded by ampersand follow the
1714     letter R, for example:
1715     <pre>
1716     (?(R3)...) or (?(R&name)...)
1717 nigel 91 </pre>
1718 nigel 93 the condition is true if the most recent recursion is into the subpattern whose
1719     number or name is given. This condition does not check the entire recursion
1720     stack.
1721 nigel 63 </P>
1722     <P>
1723 nigel 93 At "top level", all these recursion test conditions are false. Recursive
1724     patterns are described below.
1725     </P>
1726     <br><b>
1727     Defining subpatterns for use by reference only
1728     </b><br>
1729     <P>
1730     If the condition is the string (DEFINE), and there is no subpattern with the
1731     name DEFINE, the condition is always false. In this case, there may be only one
1732     alternative in the subpattern. It is always skipped if control reaches this
1733     point in the pattern; the idea of DEFINE is that it can be used to define
1734     "subroutines" that can be referenced from elsewhere. (The use of "subroutines"
1735     is described below.) For example, a pattern to match an IPv4 address could be
1736     written like this (ignore whitespace and line breaks):
1737     <pre>
1738     (?(DEFINE) (?&#60;byte&#62; 2[0-4]\d | 25[0-5] | 1\d\d | [1-9]?\d) )
1739     \b (?&byte) (\.(?&byte)){3} \b
1740     </pre>
1741     The first part of the pattern is a DEFINE group inside which a another group
1742     named "byte" is defined. This matches an individual component of an IPv4
1743     address (a number less than 256). When matching takes place, this part of the
1744     pattern is skipped because DEFINE acts like a false condition.
1745     </P>
1746     <P>
1747     The rest of the pattern uses references to the named group to match the four
1748     dot-separated components of an IPv4 address, insisting on a word boundary at
1749     each end.
1750     </P>
1751     <br><b>
1752     Assertion conditions
1753     </b><br>
1754     <P>
1755     If the condition is not in any of the above formats, it must be an assertion.
1756 nigel 63 This may be a positive or negative lookahead or lookbehind assertion. Consider
1757     this pattern, again containing non-significant white space, and with the two
1758     alternatives on the second line:
1759     <pre>
1760     (?(?=[^a-z]*[a-z])
1761     \d{2}-[a-z]{3}-\d{2} | \d{2}-\d{2}-\d{2} )
1762 nigel 75 </pre>
1763 nigel 63 The condition is a positive lookahead assertion that matches an optional
1764     sequence of non-letters followed by a letter. In other words, it tests for the
1765     presence of at least one letter in the subject. If a letter is found, the
1766     subject is matched against the first alternative; otherwise it is matched
1767     against the second. This pattern matches strings in one of the two forms
1768     dd-aaa-dd or dd-dd-dd, where aaa are letters and dd are digits.
1769 nigel 75 <a name="comments"></a></P>
1770 ph10 182 <br><a name="SEC19" href="#TOC1">COMMENTS</a><br>
1771 nigel 63 <P>
1772 nigel 75 The sequence (?# marks the start of a comment that continues up to the next
1773 nigel 63 closing parenthesis. Nested parentheses are not permitted. The characters
1774     that make up a comment play no part in the pattern matching at all.
1775     </P>
1776     <P>
1777     If the PCRE_EXTENDED option is set, an unescaped # character outside a
1778 nigel 91 character class introduces a comment that continues to immediately after the
1779     next newline in the pattern.
1780     <a name="recursion"></a></P>
1781 ph10 182 <br><a name="SEC20" href="#TOC1">RECURSIVE PATTERNS</a><br>
1782 nigel 63 <P>
1783     Consider the problem of matching a string in parentheses, allowing for
1784     unlimited nested parentheses. Without the use of recursion, the best that can
1785     be done is to use a pattern that matches up to some fixed depth of nesting. It
1786 nigel 93 is not possible to handle an arbitrary nesting depth.
1787     </P>
1788     <P>
1789     For some time, Perl has provided a facility that allows regular expressions to
1790     recurse (amongst other things). It does this by interpolating Perl code in the
1791     expression at run time, and the code can refer to the expression itself. A Perl
1792     pattern using code interpolation to solve the parentheses problem can be
1793     created like this:
1794 nigel 63 <pre>
1795     $re = qr{\( (?: (?&#62;[^()]+) | (?p{$re}) )* \)}x;
1796 nigel 75 </pre>
1797 nigel 63 The (?p{...}) item interpolates Perl code at run time, and in this case refers
1798 nigel 93 recursively to the pattern in which it appears.
1799 nigel 63 </P>
1800     <P>
1801 nigel 93 Obviously, PCRE cannot support the interpolation of Perl code. Instead, it
1802     supports special syntax for recursion of the entire pattern, and also for
1803     individual subpattern recursion. After its introduction in PCRE and Python,
1804     this kind of recursion was introduced into Perl at release 5.10.
1805 nigel 63 </P>
1806     <P>
1807 nigel 93 A special item that consists of (? followed by a number greater than zero and a
1808     closing parenthesis is a recursive call of the subpattern of the given number,
1809     provided that it occurs inside that subpattern. (If not, it is a "subroutine"
1810     call, which is described in the next section.) The special item (?R) or (?0) is
1811     a recursive call of the entire regular expression.
1812 nigel 87 </P>
1813     <P>
1814 nigel 93 In PCRE (like Python, but unlike Perl), a recursive subpattern call is always
1815     treated as an atomic group. That is, once it has matched some of the subject
1816     string, it is never re-entered, even if it contains untried alternatives and
1817     there is a subsequent matching failure.
1818     </P>
1819     <P>
1820 nigel 87 This PCRE pattern solves the nested parentheses problem (assume the
1821     PCRE_EXTENDED option is set so that white space is ignored):
1822 nigel 63 <pre>
1823     \( ( (?&#62;[^()]+) | (?R) )* \)
1824 nigel 75 </pre>
1825 nigel 63 First it matches an opening parenthesis. Then it matches any number of
1826     substrings which can either be a sequence of non-parentheses, or a recursive
1827 nigel 87 match of the pattern itself (that is, a correctly parenthesized substring).
1828 nigel 63 Finally there is a closing parenthesis.
1829     </P>
1830     <P>
1831     If this were part of a larger pattern, you would not want to recurse the entire
1832     pattern, so instead you could use this:
1833     <pre>
1834     ( \( ( (?&#62;[^()]+) | (?1) )* \) )
1835 nigel 75 </pre>
1836 nigel 63 We have put the pattern into parentheses, and caused the recursion to refer to
1837 ph10 172 them instead of the whole pattern.
1838     </P>
1839     <P>
1840     In a larger pattern, keeping track of parenthesis numbers can be tricky. This
1841     is made easier by the use of relative references. (A Perl 5.10 feature.)
1842     Instead of (?1) in the pattern above you can write (?-2) to refer to the second
1843     most recently opened parentheses preceding the recursion. In other words, a
1844     negative number counts capturing parentheses leftwards from the point at which
1845     it is encountered.
1846     </P>
1847     <P>
1848     It is also possible to refer to subsequently opened parentheses, by writing
1849     references such as (?+2). However, these cannot be recursive because the
1850     reference is not inside the parentheses that are referenced. They are always
1851     "subroutine" calls, as described in the next section.
1852     </P>
1853     <P>
1854     An alternative approach is to use named parentheses instead. The Perl syntax
1855     for this is (?&name); PCRE's earlier syntax (?P&#62;name) is also supported. We
1856     could rewrite the above example as follows:
1857 nigel 63 <pre>
1858 nigel 93 (?&#60;pn&#62; \( ( (?&#62;[^()]+) | (?&pn) )* \) )
1859 nigel 75 </pre>
1860 nigel 93 If there is more than one subpattern with the same name, the earliest one is
1861 ph10 172 used.
1862     </P>
1863     <P>
1864     This particular example pattern that we have been looking at contains nested
1865     unlimited repeats, and so the use of atomic grouping for matching strings of
1866     non-parentheses is important when applying the pattern to strings that do not
1867     match. For example, when this pattern is applied to
1868 nigel 63 <pre>
1869     (aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa()
1870 nigel 75 </pre>
1871 nigel 63 it yields "no match" quickly. However, if atomic grouping is not used,
1872     the match runs for a very long time indeed because there are so many different
1873     ways the + and * repeats can carve up the subject, and all have to be tested
1874     before failure can be reported.
1875     </P>
1876     <P>
1877     At the end of a match, the values set for any capturing subpatterns are those
1878     from the outermost level of the recursion at which the subpattern value is set.
1879     If you want to obtain intermediate values, a callout function can be used (see
1880 nigel 93 below and the
1881 nigel 63 <a href="pcrecallout.html"><b>pcrecallout</b></a>
1882     documentation). If the pattern above is matched against
1883     <pre>
1884     (ab(cd)ef)
1885 nigel 75 </pre>
1886 nigel 63 the value for the capturing parentheses is "ef", which is the last value taken
1887     on at the top level. If additional parentheses are added, giving
1888     <pre>
1889     \( ( ( (?&#62;[^()]+) | (?R) )* ) \)
1890     ^ ^
1891     ^ ^
1892 nigel 75 </pre>
1893 nigel 63 the string they capture is "ab(cd)ef", the contents of the top level
1894     parentheses. If there are more than 15 capturing parentheses in a pattern, PCRE
1895     has to obtain extra memory to store data during a recursion, which it does by
1896     using <b>pcre_malloc</b>, freeing it via <b>pcre_free</b> afterwards. If no
1897     memory can be obtained, the match fails with the PCRE_ERROR_NOMEMORY error.
1898     </P>
1899     <P>
1900     Do not confuse the (?R) item with the condition (R), which tests for recursion.
1901     Consider this pattern, which matches text in angle brackets, allowing for
1902     arbitrary nesting. Only digits are allowed in nested brackets (that is, when
1903     recursing), whereas any characters are permitted at the outer level.
1904     <pre>
1905     &#60; (?: (?(R) \d++ | [^&#60;&#62;]*+) | (?R)) * &#62;
1906 nigel 75 </pre>
1907 nigel 63 In this pattern, (?(R) is the start of a conditional subpattern, with two
1908     different alternatives for the recursive and non-recursive cases. The (?R) item
1909     is the actual recursive call.
1910 nigel 75 <a name="subpatternsassubroutines"></a></P>
1911 ph10 182 <br><a name="SEC21" href="#TOC1">SUBPATTERNS AS SUBROUTINES</a><br>
1912 nigel 63 <P>
1913     If the syntax for a recursive subpattern reference (either by number or by
1914     name) is used outside the parentheses to which it refers, it operates like a
1915 nigel 93 subroutine in a programming language. The "called" subpattern may be defined
1916 ph10 172 before or after the reference. A numbered reference can be absolute or
1917     relative, as in these examples:
1918 nigel 63 <pre>
1919 ph10 172 (...(absolute)...)...(?2)...
1920     (...(relative)...)...(?-1)...
1921     (...(?+1)...(relative)...
1922     </pre>
1923     An earlier example pointed out that the pattern
1924     <pre>
1925 nigel 63 (sens|respons)e and \1ibility
1926 nigel 75 </pre>
1927 nigel 63 matches "sense and sensibility" and "response and responsibility", but not
1928     "sense and responsibility". If instead the pattern
1929     <pre>
1930     (sens|respons)e and (?1)ibility
1931 nigel 75 </pre>
1932 nigel 63 is used, it does match "sense and responsibility" as well as the other two
1933 nigel 93 strings. Another example is given in the discussion of DEFINE above.
1934 nigel 63 </P>
1935 nigel 87 <P>
1936     Like recursive subpatterns, a "subroutine" call is always treated as an atomic
1937     group. That is, once it has matched some of the subject string, it is never
1938     re-entered, even if it contains untried alternatives and there is a subsequent
1939     matching failure.
1940     </P>
1941 nigel 63 <P>
1942 nigel 93 When a subpattern is used as a subroutine, processing options such as
1943     case-independence are fixed when the subpattern is defined. They cannot be
1944     changed for different calls. For example, consider this pattern:
1945     <pre>
1946 ph10 172 (abc)(?i:(?-1))
1947 nigel 93 </pre>
1948     It matches "abcabc". It does not match "abcABC" because the change of
1949     processing option does not affect the called subpattern.
1950     </P>
1951 ph10 182 <br><a name="SEC22" href="#TOC1">CALLOUTS</a><br>
1952 nigel 93 <P>
1953 nigel 63 Perl has a feature whereby using the sequence (?{...}) causes arbitrary Perl
1954     code to be obeyed in the middle of matching a regular expression. This makes it
1955     possible, amongst other things, to extract different substrings that match the
1956     same pair of parentheses when there is a repetition.
1957     </P>
1958     <P>
1959     PCRE provides a similar feature, but of course it cannot obey arbitrary Perl
1960     code. The feature is called "callout". The caller of PCRE provides an external
1961     function by putting its entry point in the global variable <i>pcre_callout</i>.
1962     By default, this variable contains NULL, which disables all calling out.
1963     </P>
1964     <P>
1965     Within a regular expression, (?C) indicates the points at which the external
1966     function is to be called. If you want to identify different callout points, you
1967     can put a number less than 256 after the letter C. The default value is zero.
1968     For example, this pattern has two callout points:
1969     <pre>
1970 ph10 155 (?C1)abc(?C2)def
1971 nigel 75 </pre>
1972     If the PCRE_AUTO_CALLOUT flag is passed to <b>pcre_compile()</b>, callouts are
1973     automatically installed before each item in the pattern. They are all numbered
1974     255.
1975 nigel 63 </P>
1976     <P>
1977     During matching, when PCRE reaches a callout point (and <i>pcre_callout</i> is
1978     set), the external function is called. It is provided with the number of the
1979 nigel 75 callout, the position in the pattern, and, optionally, one item of data
1980     originally supplied by the caller of <b>pcre_exec()</b>. The callout function
1981     may cause matching to proceed, to backtrack, or to fail altogether. A complete
1982     description of the interface to the callout function is given in the
1983 nigel 63 <a href="pcrecallout.html"><b>pcrecallout</b></a>
1984     documentation.
1985     </P>
1986 ph10 182 <br><a name="SEC23" href="#TOC1">SEE ALSO</a><br>
1987 nigel 63 <P>
1988 nigel 93 <b>pcreapi</b>(3), <b>pcrecallout</b>(3), <b>pcrematching</b>(3), <b>pcre</b>(3).
1989     </P>
1990 ph10 182 <br><a name="SEC24" href="#TOC1">AUTHOR</a><br>
1991 nigel 93 <P>
1992 ph10 99 Philip Hazel
1993 nigel 63 <br>
1994 ph10 99 University Computing Service
1995     <br>
1996     Cambridge CB2 3QH, England.
1997     <br>
1998     </P>
1999 ph10 182 <br><a name="SEC25" href="#TOC1">REVISION</a><br>
2000 ph10 99 <P>
2001 ph10 208 Last updated: 06 August 2007
2002 ph10 99 <br>
2003     Copyright &copy; 1997-2007 University of Cambridge.
2004     <br>
2005 nigel 75 <p>
2006     Return to the <a href="index.html">PCRE index page</a>.
2007     </p>


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

ViewVC Help
Powered by ViewVC 1.1.12