ViewVC logotype

Contents of /code/trunk/doc/pcrepattern.3

Parent Directory Parent Directory | Revision Log Revision Log

Revision 208 - (hide annotations) (download)
Mon Aug 6 15:23:29 2007 UTC (7 years, 8 months ago) by ph10
File size: 80926 byte(s)
Added a pcresyntax man page; tidied some others.

1 nigel 79 .TH PCREPATTERN 3
2 nigel 63 .SH NAME
3     PCRE - Perl-compatible regular expressions
5 nigel 63 .rs
6     .sp
7 ph10 208 The syntax and semantics of the regular expressions that are supported by PCRE
8     are described in detail below. There is a quick-reference syntax summary in the
9     .\" HREF
10     \fBpcresyntax\fP
11     .\"
12     page. Perl's regular expressions are described in its own documentation, and
13     regular expressions in general are covered in a number of books, some of which
14     have copious examples. Jeffrey Friedl's "Mastering Regular Expressions",
15     published by O'Reilly, covers regular expressions in great detail. This
16     description of PCRE's regular expressions is intended as reference material.
17 nigel 75 .P
18     The original operation of PCRE was on strings of one-byte characters. However,
19     there is now also support for UTF-8 character strings. To use this, you must
20     build PCRE to include UTF-8 support, and then call \fBpcre_compile()\fP with
21     the PCRE_UTF8 option. How this affects pattern matching is mentioned in several
22     places below. There is also a summary of UTF-8 features in the
23 nigel 63 .\" HTML <a href="pcre.html#utf8support">
24     .\" </a>
25     section on UTF-8 support
26     .\"
27     in the main
28     .\" HREF
29 nigel 75 \fBpcre\fP
30 nigel 63 .\"
31     page.
32 nigel 75 .P
33 nigel 77 The remainder of this document discusses the patterns that are supported by
34     PCRE when its main matching function, \fBpcre_exec()\fP, is used.
35     From release 6.0, PCRE offers a second matching function,
36     \fBpcre_dfa_exec()\fP, which matches using a different algorithm that is not
37 ph10 172 Perl-compatible. Some of the features discussed below are not available when
38 ph10 168 \fBpcre_dfa_exec()\fP is used. The advantages and disadvantages of the
39     alternative function, and how it differs from the normal function, are
40     discussed in the
41 nigel 77 .\" HREF
42     \fBpcrematching\fP
43     .\"
44     page.
45 nigel 93 .
46     .
48     .rs
49     .sp
50 nigel 63 A regular expression is a pattern that is matched against a subject string from
51     left to right. Most characters stand for themselves in a pattern, and match the
52     corresponding characters in the subject. As a trivial example, the pattern
53 nigel 75 .sp
54 nigel 63 The quick brown fox
55 nigel 75 .sp
56 nigel 77 matches a portion of a subject string that is identical to itself. When
57     caseless matching is specified (the PCRE_CASELESS option), letters are matched
58     independently of case. In UTF-8 mode, PCRE always understands the concept of
59     case for characters whose values are less than 128, so caseless matching is
60     always possible. For characters with higher values, the concept of case is
61     supported if PCRE is compiled with Unicode property support, but not otherwise.
62     If you want to use caseless matching for characters 128 and above, you must
63     ensure that PCRE is compiled with Unicode property support as well as with
64     UTF-8 support.
65     .P
66     The power of regular expressions comes from the ability to include alternatives
67     and repetitions in the pattern. These are encoded in the pattern by the use of
68 nigel 75 \fImetacharacters\fP, which do not stand for themselves but instead are
69 nigel 63 interpreted in some special way.
70 nigel 75 .P
71     There are two different sets of metacharacters: those that are recognized
72 nigel 63 anywhere in the pattern except within square brackets, and those that are
73 nigel 93 recognized within square brackets. Outside square brackets, the metacharacters
74     are as follows:
75 nigel 75 .sp
76     \e general escape character with several uses
77 nigel 63 ^ assert start of string (or line, in multiline mode)
78     $ assert end of string (or line, in multiline mode)
79     . match any character except newline (by default)
80     [ start character class definition
81     | start of alternative branch
82     ( start subpattern
83     ) end subpattern
84     ? extends the meaning of (
85     also 0 or 1 quantifier
86     also quantifier minimizer
87     * 0 or more quantifier
88     + 1 or more quantifier
89     also "possessive quantifier"
90     { start min/max quantifier
91 nigel 75 .sp
92 nigel 63 Part of a pattern that is in square brackets is called a "character class". In
93 nigel 75 a character class the only metacharacters are:
94     .sp
95     \e general escape character
96 nigel 63 ^ negate the class, but only if the first character
97     - indicates character range
98 nigel 75 .\" JOIN
99 nigel 63 [ POSIX character class (only if followed by POSIX
100     syntax)
101     ] terminates the character class
102 nigel 75 .sp
103     The following sections describe the use of each of the metacharacters.
104     .
105 nigel 93 .
106 nigel 63 .SH BACKSLASH
107     .rs
108     .sp
109     The backslash character has several uses. Firstly, if it is followed by a
110 nigel 91 non-alphanumeric character, it takes away any special meaning that character
111     may have. This use of backslash as an escape character applies both inside and
112 nigel 63 outside character classes.
113 nigel 75 .P
114     For example, if you want to match a * character, you write \e* in the pattern.
115 nigel 63 This escaping action applies whether or not the following character would
116 nigel 75 otherwise be interpreted as a metacharacter, so it is always safe to precede a
117     non-alphanumeric with backslash to specify that it stands for itself. In
118     particular, if you want to match a backslash, you write \e\e.
119     .P
120 nigel 63 If a pattern is compiled with the PCRE_EXTENDED option, whitespace in the
121     pattern (other than in a character class) and characters between a # outside
122 nigel 91 a character class and the next newline are ignored. An escaping backslash can
123     be used to include a whitespace or # character as part of the pattern.
124 nigel 75 .P
125 nigel 63 If you want to remove the special meaning from a sequence of characters, you
126 nigel 75 can do so by putting them between \eQ and \eE. This is different from Perl in
127     that $ and @ are handled as literals in \eQ...\eE sequences in PCRE, whereas in
128 nigel 63 Perl, $ and @ cause variable interpolation. Note the following examples:
129 nigel 75 .sp
130 nigel 63 Pattern PCRE matches Perl matches
131 nigel 75 .sp
132     .\" JOIN
133     \eQabc$xyz\eE abc$xyz abc followed by the
134 nigel 63 contents of $xyz
135 nigel 75 \eQabc\e$xyz\eE abc\e$xyz abc\e$xyz
136     \eQabc\eE\e$\eQxyz\eE abc$xyz abc$xyz
137     .sp
138     The \eQ...\eE sequence is recognized both inside and outside character classes.
139     .
140     .
141     .\" HTML <a name="digitsafterbackslash"></a>
142     .SS "Non-printing characters"
143     .rs
144     .sp
145 nigel 63 A second use of backslash provides a way of encoding non-printing characters
146     in patterns in a visible manner. There is no restriction on the appearance of
147     non-printing characters, apart from the binary zero that terminates a pattern,
148     but when a pattern is being prepared by text editing, it is usually easier to
149     use one of the following escape sequences than the binary character it
150     represents:
151 nigel 75 .sp
152     \ea alarm, that is, the BEL character (hex 07)
153     \ecx "control-x", where x is any character
154     \ee escape (hex 1B)
155     \ef formfeed (hex 0C)
156     \en newline (hex 0A)
157     \er carriage return (hex 0D)
158     \et tab (hex 09)
159     \eddd character with octal code ddd, or backreference
160     \exhh character with hex code hh
161 nigel 87 \ex{hhh..} character with hex code hhh..
162 nigel 75 .sp
163     The precise effect of \ecx is as follows: if x is a lower case letter, it
164 nigel 63 is converted to upper case. Then bit 6 of the character (hex 40) is inverted.
165 nigel 75 Thus \ecz becomes hex 1A, but \ec{ becomes hex 3B, while \ec; becomes hex
166 nigel 63 7B.
167 nigel 75 .P
168     After \ex, from zero to two hexadecimal digits are read (letters can be in
169 nigel 87 upper or lower case). Any number of hexadecimal digits may appear between \ex{
170     and }, but the value of the character code must be less than 256 in non-UTF-8
171     mode, and less than 2**31 in UTF-8 mode (that is, the maximum hexadecimal value
172     is 7FFFFFFF). If characters other than hexadecimal digits appear between \ex{
173     and }, or if there is no terminating }, this form of escape is not recognized.
174     Instead, the initial \ex will be interpreted as a basic hexadecimal escape,
175     with no following digits, giving a character whose value is zero.
176 nigel 75 .P
177 nigel 63 Characters whose value is less than 256 can be defined by either of the two
178 nigel 87 syntaxes for \ex. There is no difference in the way they are handled. For
179     example, \exdc is exactly the same as \ex{dc}.
180 nigel 75 .P
181 nigel 91 After \e0 up to two further octal digits are read. If there are fewer than two
182     digits, just those that are present are used. Thus the sequence \e0\ex\e07
183     specifies two binary zeros followed by a BEL character (code value 7). Make
184     sure you supply two digits after the initial zero if the pattern character that
185     follows is itself an octal digit.
186 nigel 75 .P
187 nigel 63 The handling of a backslash followed by a digit other than 0 is complicated.
188     Outside a character class, PCRE reads it and any following digits as a decimal
189     number. If the number is less than 10, or if there have been at least that many
190     previous capturing left parentheses in the expression, the entire sequence is
191 nigel 75 taken as a \fIback reference\fP. A description of how this works is given
192     .\" HTML <a href="#backreferences">
193     .\" </a>
194     later,
195     .\"
196     following the discussion of
197     .\" HTML <a href="#subpattern">
198     .\" </a>
199     parenthesized subpatterns.
200     .\"
201     .P
202 nigel 63 Inside a character class, or if the decimal number is greater than 9 and there
203     have not been that many capturing subpatterns, PCRE re-reads up to three octal
204 nigel 93 digits following the backslash, and uses them to generate a data character. Any
205 nigel 91 subsequent digits stand for themselves. In non-UTF-8 mode, the value of a
206     character specified in octal must be less than \e400. In UTF-8 mode, values up
207     to \e777 are permitted. For example:
208 nigel 75 .sp
209     \e040 is another way of writing a space
210     .\" JOIN
211     \e40 is the same, provided there are fewer than 40
212 nigel 63 previous capturing subpatterns
213 nigel 75 \e7 is always a back reference
214     .\" JOIN
215     \e11 might be a back reference, or another way of
216 nigel 63 writing a tab
217 nigel 75 \e011 is always a tab
218     \e0113 is a tab followed by the character "3"
219     .\" JOIN
220     \e113 might be a back reference, otherwise the
221 nigel 63 character with octal code 113
222 nigel 75 .\" JOIN
223     \e377 might be a back reference, otherwise
224 nigel 63 the byte consisting entirely of 1 bits
225 nigel 75 .\" JOIN
226     \e81 is either a back reference, or a binary zero
227 nigel 63 followed by the two characters "8" and "1"
228 nigel 75 .sp
229 nigel 63 Note that octal values of 100 or greater must not be introduced by a leading
230     zero, because no more than three octal digits are ever read.
231 nigel 75 .P
232 nigel 91 All the sequences that define a single character value can be used both inside
233     and outside character classes. In addition, inside a character class, the
234     sequence \eb is interpreted as the backspace character (hex 08), and the
235 nigel 93 sequences \eR and \eX are interpreted as the characters "R" and "X",
236     respectively. Outside a character class, these sequences have different
237     meanings
238 nigel 75 .\" HTML <a href="#uniextseq">
239     .\" </a>
240     (see below).
241     .\"
242     .
243     .
244 nigel 93 .SS "Absolute and relative back references"
245     .rs
246     .sp
247 ph10 208 The sequence \eg followed by an unsigned or a negative number, optionally
248     enclosed in braces, is an absolute or relative back reference. A named back
249     reference can be coded as \eg{name}. Back references are discussed
250 nigel 93 .\" HTML <a href="#backreferences">
251     .\" </a>
252     later,
253     .\"
254     following the discussion of
255     .\" HTML <a href="#subpattern">
256     .\" </a>
257     parenthesized subpatterns.
258     .\"
259     .
260     .
261 nigel 75 .SS "Generic character types"
262     .rs
263     .sp
264 nigel 93 Another use of backslash is for specifying generic character types. The
265 nigel 75 following are always recognized:
266     .sp
267 ph10 182 \ed any decimal digit
268 nigel 75 \eD any character that is not a decimal digit
269 ph10 178 \eh any horizontal whitespace character
270 ph10 182 \eH any character that is not a horizontal whitespace character
271 nigel 75 \es any whitespace character
272     \eS any character that is not a whitespace character
273 ph10 178 \ev any vertical whitespace character
274 ph10 182 \eV any character that is not a vertical whitespace character
275 nigel 75 \ew any "word" character
276     \eW any "non-word" character
277     .sp
278 nigel 63 Each pair of escape sequences partitions the complete set of characters into
279     two disjoint sets. Any given character matches one, and only one, of each pair.
280 nigel 75 .P
281     These character type sequences can appear both inside and outside character
282     classes. They each match one character of the appropriate type. If the current
283     matching point is at the end of the subject string, all of them fail, since
284     there is no character to match.
285     .P
286     For compatibility with Perl, \es does not match the VT character (code 11).
287     This makes it different from the the POSIX "space" class. The \es characters
288 ph10 178 are HT (9), LF (10), FF (12), CR (13), and space (32). If "use locale;" is
289 nigel 91 included in a Perl script, \es may match the VT character. In PCRE, it never
290 ph10 178 does.
291 nigel 75 .P
292 ph10 178 In UTF-8 mode, characters with values greater than 128 never match \ed, \es, or
293     \ew, and always match \eD, \eS, and \eW. This is true even when Unicode
294 ph10 182 character property support is available. These sequences retain their original
295     meanings from before UTF-8 support was available, mainly for efficiency
296 ph10 178 reasons.
297     .P
298 ph10 182 The sequences \eh, \eH, \ev, and \eV are Perl 5.10 features. In contrast to the
299 ph10 178 other sequences, these do match certain high-valued codepoints in UTF-8 mode.
300     The horizontal space characters are:
301     .sp
302     U+0009 Horizontal tab
303     U+0020 Space
304     U+00A0 Non-break space
305     U+1680 Ogham space mark
306     U+180E Mongolian vowel separator
307     U+2000 En quad
308     U+2001 Em quad
309     U+2002 En space
310     U+2003 Em space
311     U+2004 Three-per-em space
312     U+2005 Four-per-em space
313     U+2006 Six-per-em space
314     U+2007 Figure space
315     U+2008 Punctuation space
316     U+2009 Thin space
317     U+200A Hair space
318     U+202F Narrow no-break space
319     U+205F Medium mathematical space
320     U+3000 Ideographic space
321     .sp
322     The vertical space characters are:
323     .sp
324     U+000A Linefeed
325     U+000B Vertical tab
326     U+000C Formfeed
327     U+000D Carriage return
328     U+0085 Next line
329     U+2028 Line separator
330     U+2029 Paragraph separator
331     .P
332 nigel 75 A "word" character is an underscore or any character less than 256 that is a
333     letter or digit. The definition of letters and digits is controlled by PCRE's
334     low-valued character tables, and may vary if locale-specific matching is taking
335     place (see
336 nigel 63 .\" HTML <a href="pcreapi.html#localesupport">
337     .\" </a>
338     "Locale support"
339     .\"
340     in the
341     .\" HREF
342 nigel 75 \fBpcreapi\fP
343 nigel 63 .\"
344 ph10 139 page). For example, in a French locale such as "fr_FR" in Unix-like systems,
345     or "french" in Windows, some character codes greater than 128 are used for
346 ph10 178 accented letters, and these are matched by \ew. The use of locales with Unicode
347     is discouraged.
348 nigel 75 .
349     .
350 nigel 93 .SS "Newline sequences"
351     .rs
352     .sp
353     Outside a character class, the escape sequence \eR matches any Unicode newline
354 ph10 178 sequence. This is a Perl 5.10 feature. In non-UTF-8 mode \eR is equivalent to
355 nigel 93 the following:
356     .sp
357     (?>\er\en|\en|\ex0b|\ef|\er|\ex85)
358     .sp
359     This is an example of an "atomic group", details of which are given
360     .\" HTML <a href="#atomicgroup">
361     .\" </a>
362     below.
363     .\"
364     This particular group matches either the two-character sequence CR followed by
365     LF, or one of the single characters LF (linefeed, U+000A), VT (vertical tab,
366     U+000B), FF (formfeed, U+000C), CR (carriage return, U+000D), or NEL (next
367     line, U+0085). The two-character sequence is treated as a single unit that
368     cannot be split.
369     .P
370     In UTF-8 mode, two additional characters whose codepoints are greater than 255
371     are added: LS (line separator, U+2028) and PS (paragraph separator, U+2029).
372     Unicode character property support is not needed for these characters to be
373     recognized.
374     .P
375     Inside a character class, \eR matches the letter "R".
376     .
377     .
378 nigel 75 .\" HTML <a name="uniextseq"></a>
379     .SS Unicode character properties
380     .rs
381     .sp
382     When PCRE is built with Unicode character property support, three additional
383 ph10 184 escape sequences that match characters with specific properties are available.
384     When not in UTF-8 mode, these sequences are of course limited to testing
385     characters whose codepoints are less than 256, but they do work in this mode.
386     The extra escape sequences are:
387 nigel 75 .sp
388 nigel 87 \ep{\fIxx\fP} a character with the \fIxx\fP property
389     \eP{\fIxx\fP} a character without the \fIxx\fP property
390     \eX an extended Unicode sequence
391 nigel 75 .sp
392 nigel 87 The property names represented by \fIxx\fP above are limited to the Unicode
393     script names, the general category properties, and "Any", which matches any
394     character (including newline). Other properties such as "InMusicalSymbols" are
395     not currently supported by PCRE. Note that \eP{Any} does not match any
396     characters, so always causes a match failure.
397 nigel 75 .P
398 nigel 87 Sets of Unicode characters are defined as belonging to certain scripts. A
399     character from one of these sets can be matched using a script name. For
400     example:
401 nigel 75 .sp
402 nigel 87 \ep{Greek}
403     \eP{Han}
404     .sp
405     Those that are not part of an identified script are lumped together as
406     "Common". The current list of scripts is:
407     .P
408     Arabic,
409     Armenian,
410 nigel 93 Balinese,
411 nigel 87 Bengali,
412     Bopomofo,
413     Braille,
414     Buginese,
415     Buhid,
416     Canadian_Aboriginal,
417     Cherokee,
418     Common,
419     Coptic,
420 nigel 93 Cuneiform,
421 nigel 87 Cypriot,
422     Cyrillic,
423     Deseret,
424     Devanagari,
425     Ethiopic,
426     Georgian,
427     Glagolitic,
428     Gothic,
429     Greek,
430     Gujarati,
431     Gurmukhi,
432     Han,
433     Hangul,
434     Hanunoo,
435     Hebrew,
436     Hiragana,
437     Inherited,
438     Kannada,
439     Katakana,
440     Kharoshthi,
441     Khmer,
442     Lao,
443     Latin,
444     Limbu,
445     Linear_B,
446     Malayalam,
447     Mongolian,
448     Myanmar,
449     New_Tai_Lue,
450 nigel 93 Nko,
451 nigel 87 Ogham,
452     Old_Italic,
453     Old_Persian,
454     Oriya,
455     Osmanya,
456 nigel 93 Phags_Pa,
457     Phoenician,
458 nigel 87 Runic,
459     Shavian,
460     Sinhala,
461     Syloti_Nagri,
462     Syriac,
463     Tagalog,
464     Tagbanwa,
465     Tai_Le,
466     Tamil,
467     Telugu,
468     Thaana,
469     Thai,
470     Tibetan,
471     Tifinagh,
472     Ugaritic,
473     Yi.
474     .P
475     Each character has exactly one general category property, specified by a
476     two-letter abbreviation. For compatibility with Perl, negation can be specified
477     by including a circumflex between the opening brace and the property name. For
478     example, \ep{^Lu} is the same as \eP{Lu}.
479     .P
480     If only one letter is specified with \ep or \eP, it includes all the general
481     category properties that start with that letter. In this case, in the absence
482     of negation, the curly brackets in the escape sequence are optional; these two
483     examples have the same effect:
484     .sp
485 nigel 75 \ep{L}
486     \epL
487     .sp
488 nigel 87 The following general category property codes are supported:
489 nigel 75 .sp
490     C Other
491     Cc Control
492     Cf Format
493     Cn Unassigned
494     Co Private use
495     Cs Surrogate
496     .sp
497     L Letter
498     Ll Lower case letter
499     Lm Modifier letter
500     Lo Other letter
501     Lt Title case letter
502     Lu Upper case letter
503     .sp
504     M Mark
505     Mc Spacing mark
506     Me Enclosing mark
507     Mn Non-spacing mark
508     .sp
509     N Number
510     Nd Decimal number
511     Nl Letter number
512     No Other number
513     .sp
514     P Punctuation
515     Pc Connector punctuation
516     Pd Dash punctuation
517     Pe Close punctuation
518     Pf Final punctuation
519     Pi Initial punctuation
520     Po Other punctuation
521     Ps Open punctuation
522     .sp
523     S Symbol
524     Sc Currency symbol
525     Sk Modifier symbol
526     Sm Mathematical symbol
527     So Other symbol
528     .sp
529     Z Separator
530     Zl Line separator
531     Zp Paragraph separator
532     Zs Space separator
533     .sp
534 nigel 87 The special property L& is also supported: it matches a character that has
535     the Lu, Ll, or Lt property, in other words, a letter that is not classified as
536     a modifier or "other".
537 nigel 75 .P
538 nigel 87 The long synonyms for these properties that Perl supports (such as \ep{Letter})
539 nigel 91 are not supported by PCRE, nor is it permitted to prefix any of these
540 nigel 87 properties with "Is".
541     .P
542     No character that is in the Unicode table has the Cn (unassigned) property.
543     Instead, this property is assumed for any code point that is not in the
544     Unicode table.
545     .P
546 nigel 75 Specifying caseless matching does not affect these escape sequences. For
547     example, \ep{Lu} always matches only upper case letters.
548     .P
549     The \eX escape matches any number of Unicode characters that form an extended
550     Unicode sequence. \eX is equivalent to
551     .sp
552     (?>\ePM\epM*)
553     .sp
554     That is, it matches a character without the "mark" property, followed by zero
555     or more characters with the "mark" property, and treats the sequence as an
556     atomic group
557     .\" HTML <a href="#atomicgroup">
558     .\" </a>
559     (see below).
560     .\"
561     Characters with the "mark" property are typically accents that affect the
562 ph10 185 preceding character. None of them have codepoints less than 256, so in
563 ph10 184 non-UTF-8 mode \eX matches any one character.
564 nigel 75 .P
565     Matching characters by Unicode property is not fast, because PCRE has to search
566     a structure that contains data for over fifteen thousand characters. That is
567     why the traditional escape sequences such as \ed and \ew do not use Unicode
568     properties in PCRE.
569     .
570     .
571 ph10 168 .\" HTML <a name="resetmatchstart"></a>
572     .SS "Resetting the match start"
573     .rs
574     .sp
575 ph10 172 The escape sequence \eK, which is a Perl 5.10 feature, causes any previously
576     matched characters not to be included in the final matched sequence. For
577 ph10 168 example, the pattern:
578     .sp
579     foo\eKbar
580     .sp
581 ph10 172 matches "foobar", but reports that it has matched "bar". This feature is
582 ph10 168 similar to a lookbehind assertion
583     .\" HTML <a href="#lookbehind">
584     .\" </a>
585     (described below).
586     .\"
587 ph10 172 However, in this case, the part of the subject before the real match does not
588     have to be of fixed length, as lookbehind assertions do. The use of \eK does
589 ph10 168 not interfere with the setting of
590     .\" HTML <a href="#subpattern">
591     .\" </a>
592     captured substrings.
593 ph10 172 .\"
594 ph10 168 For example, when the pattern
595     .sp
596     (foo)\eKbar
597     .sp
598 ph10 172 matches "foobar", the first substring is still set to "foo".
599 ph10 168 .
600     .
601 nigel 75 .\" HTML <a name="smallassertions"></a>
602     .SS "Simple assertions"
603     .rs
604     .sp
605 nigel 93 The final use of backslash is for certain simple assertions. An assertion
606 nigel 63 specifies a condition that has to be met at a particular point in a match,
607     without consuming any characters from the subject string. The use of
608 nigel 75 subpatterns for more complicated assertions is described
609     .\" HTML <a href="#bigassertions">
610     .\" </a>
611     below.
612     .\"
613 nigel 91 The backslashed assertions are:
614 nigel 75 .sp
615     \eb matches at a word boundary
616     \eB matches when not at a word boundary
617 nigel 93 \eA matches at the start of the subject
618     \eZ matches at the end of the subject
619     also matches before a newline at the end of the subject
620     \ez matches only at the end of the subject
621     \eG matches at the first matching position in the subject
622 nigel 75 .sp
623     These assertions may not appear in character classes (but note that \eb has a
624 nigel 63 different meaning, namely the backspace character, inside a character class).
625 nigel 75 .P
626 nigel 63 A word boundary is a position in the subject string where the current character
627 nigel 75 and the previous character do not both match \ew or \eW (i.e. one matches
628     \ew and the other matches \eW), or the start or end of the string if the
629     first or last character matches \ew, respectively.
630     .P
631     The \eA, \eZ, and \ez assertions differ from the traditional circumflex and
632     dollar (described in the next section) in that they only ever match at the very
633     start and end of the subject string, whatever options are set. Thus, they are
634     independent of multiline mode. These three assertions are not affected by the
635     PCRE_NOTBOL or PCRE_NOTEOL options, which affect only the behaviour of the
636     circumflex and dollar metacharacters. However, if the \fIstartoffset\fP
637     argument of \fBpcre_exec()\fP is non-zero, indicating that matching is to start
638     at a point other than the beginning of the subject, \eA can never match. The
639 nigel 91 difference between \eZ and \ez is that \eZ matches before a newline at the end
640     of the string as well as at the very end, whereas \ez matches only at the end.
641 nigel 75 .P
642     The \eG assertion is true only when the current matching position is at the
643     start point of the match, as specified by the \fIstartoffset\fP argument of
644     \fBpcre_exec()\fP. It differs from \eA when the value of \fIstartoffset\fP is
645     non-zero. By calling \fBpcre_exec()\fP multiple times with appropriate
646 nigel 63 arguments, you can mimic Perl's /g option, and it is in this kind of
647 nigel 75 implementation where \eG can be useful.
648     .P
649     Note, however, that PCRE's interpretation of \eG, as the start of the current
650 nigel 63 match, is subtly different from Perl's, which defines it as the end of the
651     previous match. In Perl, these can be different when the previously matched
652     string was empty. Because PCRE does just one match at a time, it cannot
653     reproduce this behaviour.
654 nigel 75 .P
655     If all the alternatives of a pattern begin with \eG, the expression is anchored
656 nigel 63 to the starting match position, and the "anchored" flag is set in the compiled
657     regular expression.
658 nigel 75 .
659     .
661 nigel 63 .rs
662     .sp
663     Outside a character class, in the default matching mode, the circumflex
664 nigel 75 character is an assertion that is true only if the current matching point is
665     at the start of the subject string. If the \fIstartoffset\fP argument of
666     \fBpcre_exec()\fP is non-zero, circumflex can never match if the PCRE_MULTILINE
667 nigel 63 option is unset. Inside a character class, circumflex has an entirely different
668 nigel 75 meaning
669     .\" HTML <a href="#characterclass">
670     .\" </a>
671     (see below).
672     .\"
673     .P
674 nigel 63 Circumflex need not be the first character of the pattern if a number of
675     alternatives are involved, but it should be the first thing in each alternative
676     in which it appears if the pattern is ever to match that branch. If all
677     possible alternatives start with a circumflex, that is, if the pattern is
678     constrained to match only at the start of the subject, it is said to be an
679     "anchored" pattern. (There are also other constructs that can cause a pattern
680     to be anchored.)
681 nigel 75 .P
682     A dollar character is an assertion that is true only if the current matching
683 nigel 63 point is at the end of the subject string, or immediately before a newline
684 nigel 91 at the end of the string (by default). Dollar need not be the last character of
685     the pattern if a number of alternatives are involved, but it should be the last
686     item in any branch in which it appears. Dollar has no special meaning in a
687     character class.
688 nigel 75 .P
689 nigel 63 The meaning of dollar can be changed so that it matches only at the very end of
690     the string, by setting the PCRE_DOLLAR_ENDONLY option at compile time. This
691 nigel 75 does not affect the \eZ assertion.
692     .P
693 nigel 63 The meanings of the circumflex and dollar characters are changed if the
694 nigel 91 PCRE_MULTILINE option is set. When this is the case, a circumflex matches
695     immediately after internal newlines as well as at the start of the subject
696     string. It does not match after a newline that ends the string. A dollar
697     matches before any newlines in the string, as well as at the very end, when
698     PCRE_MULTILINE is set. When newline is specified as the two-character
699     sequence CRLF, isolated CR and LF characters do not indicate newlines.
700 nigel 75 .P
701 nigel 91 For example, the pattern /^abc$/ matches the subject string "def\enabc" (where
702     \en represents a newline) in multiline mode, but not otherwise. Consequently,
703     patterns that are anchored in single line mode because all branches start with
704     ^ are not anchored in multiline mode, and a match for circumflex is possible
705     when the \fIstartoffset\fP argument of \fBpcre_exec()\fP is non-zero. The
706     PCRE_DOLLAR_ENDONLY option is ignored if PCRE_MULTILINE is set.
707     .P
708 nigel 75 Note that the sequences \eA, \eZ, and \ez can be used to match the start and
709 nigel 63 end of the subject in both modes, and if all branches of a pattern start with
710 nigel 91 \eA it is always anchored, whether or not PCRE_MULTILINE is set.
711 nigel 75 .
712     .
714 nigel 63 .rs
715     .sp
716     Outside a character class, a dot in the pattern matches any one character in
717 nigel 91 the subject string except (by default) a character that signifies the end of a
718 nigel 93 line. In UTF-8 mode, the matched character may be more than one byte long.
719 nigel 91 .P
720 nigel 93 When a line ending is defined as a single character, dot never matches that
721     character; when the two-character sequence CRLF is used, dot does not match CR
722     if it is immediately followed by LF, but otherwise it matches all characters
723     (including isolated CRs and LFs). When any Unicode line endings are being
724     recognized, dot does not match CR or LF or any of the other line ending
725     characters.
726     .P
727 nigel 91 The behaviour of dot with regard to newlines can be changed. If the PCRE_DOTALL
728 nigel 93 option is set, a dot matches any one character, without exception. If the
729     two-character sequence CRLF is present in the subject string, it takes two dots
730     to match it.
731 nigel 91 .P
732     The handling of dot is entirely independent of the handling of circumflex and
733     dollar, the only relationship being that they both involve newlines. Dot has no
734     special meaning in a character class.
735 nigel 75 .
736     .
738 nigel 63 .rs
739     .sp
740 nigel 75 Outside a character class, the escape sequence \eC matches any one byte, both
741 nigel 93 in and out of UTF-8 mode. Unlike a dot, it always matches any line-ending
742     characters. The feature is provided in Perl in order to match individual bytes
743     in UTF-8 mode. Because it breaks up UTF-8 characters into individual bytes,
744     what remains in the string may be a malformed UTF-8 string. For this reason,
745     the \eC escape sequence is best avoided.
746 nigel 75 .P
747     PCRE does not allow \eC to appear in lookbehind assertions
748     .\" HTML <a href="#lookbehind">
749     .\" </a>
750     (described below),
751     .\"
752     because in UTF-8 mode this would make it impossible to calculate the length of
753     the lookbehind.
754     .
755     .
756     .\" HTML <a name="characterclass"></a>
758 nigel 63 .rs
759     .sp
760     An opening square bracket introduces a character class, terminated by a closing
761     square bracket. A closing square bracket on its own is not special. If a
762     closing square bracket is required as a member of the class, it should be the
763     first data character in the class (after an initial circumflex, if present) or
764     escaped with a backslash.
765 nigel 75 .P
766 nigel 63 A character class matches a single character in the subject. In UTF-8 mode, the
767     character may occupy more than one byte. A matched character must be in the set
768     of characters defined by the class, unless the first character in the class
769     definition is a circumflex, in which case the subject character must not be in
770     the set defined by the class. If a circumflex is actually required as a member
771     of the class, ensure it is not the first character, or escape it with a
772     backslash.
773 nigel 75 .P
774 nigel 63 For example, the character class [aeiou] matches any lower case vowel, while
775     [^aeiou] matches any character that is not a lower case vowel. Note that a
776 nigel 75 circumflex is just a convenient notation for specifying the characters that
777     are in the class by enumerating those that are not. A class that starts with a
778     circumflex is not an assertion: it still consumes a character from the subject
779     string, and therefore it fails if the current pointer is at the end of the
780     string.
781     .P
782 nigel 63 In UTF-8 mode, characters with values greater than 255 can be included in a
783 nigel 75 class as a literal string of bytes, or by using the \ex{ escaping mechanism.
784     .P
785 nigel 63 When caseless matching is set, any letters in a class represent both their
786     upper case and lower case versions, so for example, a caseless [aeiou] matches
787     "A" as well as "a", and a caseless [^aeiou] does not match "A", whereas a
788 nigel 77 caseful version would. In UTF-8 mode, PCRE always understands the concept of
789     case for characters whose values are less than 128, so caseless matching is
790     always possible. For characters with higher values, the concept of case is
791     supported if PCRE is compiled with Unicode property support, but not otherwise.
792     If you want to use caseless matching for characters 128 and above, you must
793     ensure that PCRE is compiled with Unicode property support as well as with
794     UTF-8 support.
795 nigel 75 .P
796 nigel 93 Characters that might indicate line breaks are never treated in any special way
797     when matching character classes, whatever line-ending sequence is in use, and
798     whatever setting of the PCRE_DOTALL and PCRE_MULTILINE options is used. A class
799     such as [^a] always matches one of these characters.
800 nigel 75 .P
801 nigel 63 The minus (hyphen) character can be used to specify a range of characters in a
802     character class. For example, [d-m] matches any letter between d and m,
803     inclusive. If a minus character is required in a class, it must be escaped with
804     a backslash or appear in a position where it cannot be interpreted as
805     indicating a range, typically as the first or last character in the class.
806 nigel 75 .P
807 nigel 63 It is not possible to have the literal character "]" as the end character of a
808     range. A pattern such as [W-]46] is interpreted as a class of two characters
809     ("W" and "-") followed by a literal string "46]", so it would match "W46]" or
810     "-46]". However, if the "]" is escaped with a backslash it is interpreted as
811 nigel 75 the end of range, so [W-\e]46] is interpreted as a class containing a range
812     followed by two other characters. The octal or hexadecimal representation of
813     "]" can also be used to end a range.
814     .P
815 nigel 63 Ranges operate in the collating sequence of character values. They can also be
816 nigel 75 used for characters specified numerically, for example [\e000-\e037]. In UTF-8
817 nigel 63 mode, ranges can include characters whose values are greater than 255, for
818 nigel 75 example [\ex{100}-\ex{2ff}].
819     .P
820 nigel 63 If a range that includes letters is used when caseless matching is set, it
821     matches the letters in either case. For example, [W-c] is equivalent to
822 nigel 75 [][\e\e^_`wxyzabc], matched caselessly, and in non-UTF-8 mode, if character
823 ph10 139 tables for a French locale are in use, [\exc8-\excb] matches accented E
824 nigel 75 characters in both cases. In UTF-8 mode, PCRE supports the concept of case for
825     characters with values greater than 128 only when it is compiled with Unicode
826     property support.
827     .P
828     The character types \ed, \eD, \ep, \eP, \es, \eS, \ew, and \eW may also appear
829     in a character class, and add the characters that they match to the class. For
830     example, [\edABCDEF] matches any hexadecimal digit. A circumflex can
831 nigel 63 conveniently be used with the upper case character types to specify a more
832     restricted set of characters than the matching lower case type. For example,
833 nigel 75 the class [^\eW_] matches any letter or digit, but not underscore.
834     .P
835     The only metacharacters that are recognized in character classes are backslash,
836     hyphen (only where it can be interpreted as specifying a range), circumflex
837     (only at the start), opening square bracket (only when it can be interpreted as
838     introducing a POSIX class name - see the next section), and the terminating
839     closing square bracket. However, escaping other non-alphanumeric characters
840     does no harm.
841     .
842     .
844 nigel 63 .rs
845     .sp
846 nigel 75 Perl supports the POSIX notation for character classes. This uses names
847 nigel 63 enclosed by [: and :] within the enclosing square brackets. PCRE also supports
848     this notation. For example,
849 nigel 75 .sp
850 nigel 63 [01[:alpha:]%]
851 nigel 75 .sp
852 nigel 63 matches "0", "1", any alphabetic character, or "%". The supported class names
853     are
854 nigel 75 .sp
855 nigel 63 alnum letters and digits
856     alpha letters
857     ascii character codes 0 - 127
858     blank space or tab only
859     cntrl control characters
860 nigel 75 digit decimal digits (same as \ed)
861 nigel 63 graph printing characters, excluding space
862     lower lower case letters
863     print printing characters, including space
864     punct printing characters, excluding letters and digits
865 nigel 75 space white space (not quite the same as \es)
866 nigel 63 upper upper case letters
867 nigel 75 word "word" characters (same as \ew)
868 nigel 63 xdigit hexadecimal digits
869 nigel 75 .sp
870 nigel 63 The "space" characters are HT (9), LF (10), VT (11), FF (12), CR (13), and
871     space (32). Notice that this list includes the VT character (code 11). This
872 nigel 75 makes "space" different to \es, which does not include VT (for Perl
873 nigel 63 compatibility).
874 nigel 75 .P
875 nigel 63 The name "word" is a Perl extension, and "blank" is a GNU extension from Perl
876     5.8. Another Perl extension is negation, which is indicated by a ^ character
877     after the colon. For example,
878 nigel 75 .sp
879 nigel 63 [12[:^digit:]]
880 nigel 75 .sp
881 nigel 63 matches "1", "2", or any non-digit. PCRE (and Perl) also recognize the POSIX
882     syntax [.ch.] and [=ch=] where "ch" is a "collating element", but these are not
883     supported, and an error is given if they are encountered.
884 nigel 75 .P
885     In UTF-8 mode, characters with values greater than 128 do not match any of
886 nigel 63 the POSIX character classes.
887 nigel 75 .
888     .
889     .SH "VERTICAL BAR"
890 nigel 63 .rs
891     .sp
892     Vertical bar characters are used to separate alternative patterns. For example,
893     the pattern
894 nigel 75 .sp
895 nigel 63 gilbert|sullivan
896 nigel 75 .sp
897 nigel 63 matches either "gilbert" or "sullivan". Any number of alternatives may appear,
898 nigel 91 and an empty alternative is permitted (matching the empty string). The matching
899     process tries each alternative in turn, from left to right, and the first one
900     that succeeds is used. If the alternatives are within a subpattern
901 nigel 75 .\" HTML <a href="#subpattern">
902     .\" </a>
903     (defined below),
904     .\"
905     "succeeds" means matching the rest of the main pattern as well as the
906     alternative in the subpattern.
907     .
908     .
910 nigel 63 .rs
911     .sp
912     The settings of the PCRE_CASELESS, PCRE_MULTILINE, PCRE_DOTALL, and
913     PCRE_EXTENDED options can be changed from within the pattern by a sequence of
914     Perl option letters enclosed between "(?" and ")". The option letters are
915 nigel 75 .sp
916 nigel 63 i for PCRE_CASELESS
917     m for PCRE_MULTILINE
918     s for PCRE_DOTALL
919     x for PCRE_EXTENDED
920 nigel 75 .sp
921 nigel 63 For example, (?im) sets caseless, multiline matching. It is also possible to
922     unset these options by preceding the letter with a hyphen, and a combined
923     setting and unsetting such as (?im-sx), which sets PCRE_CASELESS and
924     PCRE_MULTILINE while unsetting PCRE_DOTALL and PCRE_EXTENDED, is also
925     permitted. If a letter appears both before and after the hyphen, the option is
926     unset.
927 nigel 75 .P
928 nigel 63 When an option change occurs at top level (that is, not inside subpattern
929     parentheses), the change applies to the remainder of the pattern that follows.
930     If the change is placed right at the start of a pattern, PCRE extracts it into
931     the global options (and it will therefore show up in data extracted by the
932 nigel 75 \fBpcre_fullinfo()\fP function).
933     .P
934 nigel 93 An option change within a subpattern (see below for a description of
935     subpatterns) affects only that part of the current pattern that follows it, so
936 nigel 75 .sp
937 nigel 63 (a(?i)b)c
938 nigel 75 .sp
939 nigel 63 matches abc and aBc and no other strings (assuming PCRE_CASELESS is not used).
940     By this means, options can be made to have different settings in different
941     parts of the pattern. Any changes made in one alternative do carry on
942     into subsequent branches within the same subpattern. For example,
943 nigel 75 .sp
944 nigel 63 (a(?i)b|c)
945 nigel 75 .sp
946 nigel 63 matches "ab", "aB", "c", and "C", even though when matching "C" the first
947     branch is abandoned before the option setting. This is because the effects of
948     option settings happen at compile time. There would be some very weird
949     behaviour otherwise.
950 nigel 75 .P
951 nigel 91 The PCRE-specific options PCRE_DUPNAMES, PCRE_UNGREEDY, and PCRE_EXTRA can be
952     changed in the same way as the Perl-compatible options by using the characters
953     J, U and X respectively.
954 nigel 75 .
955     .
956     .\" HTML <a name="subpattern"></a>
957 nigel 63 .SH SUBPATTERNS
958     .rs
959     .sp
960     Subpatterns are delimited by parentheses (round brackets), which can be nested.
961 nigel 75 Turning part of a pattern into a subpattern does two things:
962     .sp
963 nigel 63 1. It localizes a set of alternatives. For example, the pattern
964 nigel 75 .sp
965 nigel 63 cat(aract|erpillar|)
966 nigel 75 .sp
967 nigel 63 matches one of the words "cat", "cataract", or "caterpillar". Without the
968 nigel 93 parentheses, it would match "cataract", "erpillar" or an empty string.
969 nigel 75 .sp
970     2. It sets up the subpattern as a capturing subpattern. This means that, when
971     the whole pattern matches, that portion of the subject string that matched the
972     subpattern is passed back to the caller via the \fIovector\fP argument of
973     \fBpcre_exec()\fP. Opening parentheses are counted from left to right (starting
974     from 1) to obtain numbers for the capturing subpatterns.
975     .P
976 nigel 63 For example, if the string "the red king" is matched against the pattern
977 nigel 75 .sp
978 nigel 63 the ((red|white) (king|queen))
979 nigel 75 .sp
980 nigel 63 the captured substrings are "red king", "red", and "king", and are numbered 1,
981     2, and 3, respectively.
982 nigel 75 .P
983 nigel 63 The fact that plain parentheses fulfil two functions is not always helpful.
984     There are often times when a grouping subpattern is required without a
985     capturing requirement. If an opening parenthesis is followed by a question mark
986     and a colon, the subpattern does not do any capturing, and is not counted when
987     computing the number of any subsequent capturing subpatterns. For example, if
988     the string "the white queen" is matched against the pattern
989 nigel 75 .sp
990 nigel 63 the ((?:red|white) (king|queen))
991 nigel 75 .sp
992 nigel 63 the captured substrings are "white queen" and "queen", and are numbered 1 and
993 nigel 93 2. The maximum number of capturing subpatterns is 65535.
994 nigel 75 .P
995 nigel 63 As a convenient shorthand, if any option settings are required at the start of
996     a non-capturing subpattern, the option letters may appear between the "?" and
997     the ":". Thus the two patterns
998 nigel 75 .sp
999 nigel 63 (?i:saturday|sunday)
1000     (?:(?i)saturday|sunday)
1001 nigel 75 .sp
1002 nigel 63 match exactly the same set of strings. Because alternative branches are tried
1003     from left to right, and options are not reset until the end of the subpattern
1004     is reached, an option setting in one branch does affect subsequent branches, so
1005     the above patterns match "SUNDAY" as well as "Saturday".
1006 nigel 75 .
1007     .
1009     .rs
1010     .sp
1011 ph10 182 Perl 5.10 introduced a feature whereby each alternative in a subpattern uses
1012     the same numbers for its capturing parentheses. Such a subpattern starts with
1013     (?| and is itself a non-capturing subpattern. For example, consider this
1014 ph10 175 pattern:
1015     .sp
1016     (?|(Sat)ur|(Sun))day
1017 ph10 182 .sp
1018     Because the two alternatives are inside a (?| group, both sets of capturing
1019     parentheses are numbered one. Thus, when the pattern matches, you can look
1020     at captured substring number one, whichever alternative matched. This construct
1021     is useful when you want to capture part, but not all, of one of a number of
1022     alternatives. Inside a (?| group, parentheses are numbered as usual, but the
1023 ph10 175 number is reset at the start of each branch. The numbers of any capturing
1024 ph10 182 buffers that follow the subpattern start after the highest number used in any
1025     branch. The following example is taken from the Perl documentation.
1026 ph10 175 The numbers underneath show in which buffer the captured content will be
1027     stored.
1028     .sp
1029     # before ---------------branch-reset----------- after
1030     / ( a ) (?| x ( y ) z | (p (q) r) | (t) u (v) ) ( z ) /x
1031     # 1 2 2 3 2 3 4
1032 ph10 182 .sp
1033 ph10 175 A backreference or a recursive call to a numbered subpattern always refers to
1034     the first one in the pattern with the given number.
1035     .P
1036     An alternative approach to using this "branch reset" feature is to use
1037     duplicate named subpatterns, as described in the next section.
1038     .
1039     .
1040 nigel 75 .SH "NAMED SUBPATTERNS"
1041 nigel 63 .rs
1042     .sp
1043     Identifying capturing parentheses by number is simple, but it can be very hard
1044     to keep track of the numbers in complicated regular expressions. Furthermore,
1045 nigel 75 if an expression is modified, the numbers may change. To help with this
1046 nigel 93 difficulty, PCRE supports the naming of subpatterns. This feature was not
1047     added to Perl until release 5.10. Python had the feature earlier, and PCRE
1048     introduced it at release 4.0, using the Python syntax. PCRE now supports both
1049     the Perl and the Python syntax.
1050     .P
1051     In PCRE, a subpattern can be named in one of three ways: (?<name>...) or
1052     (?'name'...) as in Perl, or (?P<name>...) as in Python. References to capturing
1053 nigel 91 parentheses from other parts of the pattern, such as
1054     .\" HTML <a href="#backreferences">
1055     .\" </a>
1056     backreferences,
1057     .\"
1058     .\" HTML <a href="#recursion">
1059     .\" </a>
1060     recursion,
1061     .\"
1062     and
1063     .\" HTML <a href="#conditions">
1064     .\" </a>
1065     conditions,
1066     .\"
1067     can be made by name as well as by number.
1068 nigel 75 .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     By default, a name must be unique within a pattern, but it is possible to relax
1076     this constraint by setting the PCRE_DUPNAMES option at compile time. This can
1077     be useful for patterns where only one instance of the named parentheses can
1078     match. Suppose you want to match the name of a weekday, either as a 3-letter
1079     abbreviation or as the full name, and in both cases you want to extract the
1080     abbreviation. This pattern (ignoring the line breaks) does the job:
1081     .sp
1082 nigel 93 (?<DN>Mon|Fri|Sun)(?:day)?|
1083     (?<DN>Tue)(?:sday)?|
1084     (?<DN>Wed)(?:nesday)?|
1085     (?<DN>Thu)(?:rsday)?|
1086     (?<DN>Sat)(?:urday)?
1087 nigel 91 .sp
1088     There are five capturing substrings, but only one is ever set after a match.
1089 ph10 182 (An alternative way of solving this problem is to use a "branch reset"
1090 ph10 175 subpattern, as described in the previous section.)
1091     .P
1092 nigel 91 The convenience function for extracting the data by name returns the substring
1093 nigel 93 for the first (and in this example, the only) subpattern of that name that
1094 nigel 91 matched. This saves searching to find which numbered subpattern it was. If you
1095     make a reference to a non-unique named subpattern from elsewhere in the
1096     pattern, the one that corresponds to the lowest number is used. For further
1097     details of the interfaces for handling named subpatterns, see the
1098 nigel 63 .\" HREF
1099 nigel 75 \fBpcreapi\fP
1100 nigel 63 .\"
1101     documentation.
1102 nigel 75 .
1103     .
1104 nigel 63 .SH REPETITION
1105     .rs
1106     .sp
1107     Repetition is specified by quantifiers, which can follow any of the following
1108     items:
1109 nigel 75 .sp
1110 nigel 63 a literal data character
1111 nigel 93 the dot metacharacter
1112 nigel 75 the \eC escape sequence
1113     the \eX escape sequence (in UTF-8 mode with Unicode properties)
1114 nigel 93 the \eR escape sequence
1115 nigel 75 an escape such as \ed that matches a single character
1116 nigel 63 a character class
1117     a back reference (see next section)
1118     a parenthesized subpattern (unless it is an assertion)
1119 nigel 75 .sp
1120 nigel 63 The general repetition quantifier specifies a minimum and maximum number of
1121     permitted matches, by giving the two numbers in curly brackets (braces),
1122     separated by a comma. The numbers must be less than 65536, and the first must
1123     be less than or equal to the second. For example:
1124 nigel 75 .sp
1125 nigel 63 z{2,4}
1126 nigel 75 .sp
1127 nigel 63 matches "zz", "zzz", or "zzzz". A closing brace on its own is not a special
1128     character. If the second number is omitted, but the comma is present, there is
1129     no upper limit; if the second number and the comma are both omitted, the
1130     quantifier specifies an exact number of required matches. Thus
1131 nigel 75 .sp
1132 nigel 63 [aeiou]{3,}
1133 nigel 75 .sp
1134 nigel 63 matches at least 3 successive vowels, but may match many more, while
1135 nigel 75 .sp
1136     \ed{8}
1137     .sp
1138 nigel 63 matches exactly 8 digits. An opening curly bracket that appears in a position
1139     where a quantifier is not allowed, or one that does not match the syntax of a
1140     quantifier, is taken as a literal character. For example, {,6} is not a
1141     quantifier, but a literal string of four characters.
1142 nigel 75 .P
1143 nigel 63 In UTF-8 mode, quantifiers apply to UTF-8 characters rather than to individual
1144 nigel 75 bytes. Thus, for example, \ex{100}{2} matches two UTF-8 characters, each of
1145     which is represented by a two-byte sequence. Similarly, when Unicode property
1146     support is available, \eX{3} matches three Unicode extended sequences, each of
1147     which may be several bytes long (and they may be of different lengths).
1148     .P
1149 nigel 63 The quantifier {0} is permitted, causing the expression to behave as if the
1150     previous item and the quantifier were not present.
1151 nigel 75 .P
1152 nigel 93 For convenience, the three most common quantifiers have single-character
1153     abbreviations:
1154 nigel 75 .sp
1155 nigel 63 * is equivalent to {0,}
1156     + is equivalent to {1,}
1157     ? is equivalent to {0,1}
1158 nigel 75 .sp
1159 nigel 63 It is possible to construct infinite loops by following a subpattern that can
1160     match no characters with a quantifier that has no upper limit, for example:
1161 nigel 75 .sp
1162 nigel 63 (a?)*
1163 nigel 75 .sp
1164 nigel 63 Earlier versions of Perl and PCRE used to give an error at compile time for
1165     such patterns. However, because there are cases where this can be useful, such
1166     patterns are now accepted, but if any repetition of the subpattern does in fact
1167     match no characters, the loop is forcibly broken.
1168 nigel 75 .P
1169 nigel 63 By default, the quantifiers are "greedy", that is, they match as much as
1170     possible (up to the maximum number of permitted times), without causing the
1171     rest of the pattern to fail. The classic example of where this gives problems
1172 nigel 75 is in trying to match comments in C programs. These appear between /* and */
1173     and within the comment, individual * and / characters may appear. An attempt to
1174     match C comments by applying the pattern
1175     .sp
1176     /\e*.*\e*/
1177     .sp
1178 nigel 63 to the string
1179 nigel 75 .sp
1180     /* first comment */ not comment /* second comment */
1181     .sp
1182 nigel 63 fails, because it matches the entire string owing to the greediness of the .*
1183     item.
1184 nigel 75 .P
1185 nigel 63 However, if a quantifier is followed by a question mark, it ceases to be
1186     greedy, and instead matches the minimum number of times possible, so the
1187     pattern
1188 nigel 75 .sp
1189     /\e*.*?\e*/
1190     .sp
1191 nigel 63 does the right thing with the C comments. The meaning of the various
1192     quantifiers is not otherwise changed, just the preferred number of matches.
1193     Do not confuse this use of question mark with its use as a quantifier in its
1194     own right. Because it has two uses, it can sometimes appear doubled, as in
1195 nigel 75 .sp
1196     \ed??\ed
1197     .sp
1198 nigel 63 which matches one digit by preference, but can match two if that is the only
1199     way the rest of the pattern matches.
1200 nigel 75 .P
1201 nigel 93 If the PCRE_UNGREEDY option is set (an option that is not available in Perl),
1202 nigel 63 the quantifiers are not greedy by default, but individual ones can be made
1203     greedy by following them with a question mark. In other words, it inverts the
1204     default behaviour.
1205 nigel 75 .P
1206 nigel 63 When a parenthesized subpattern is quantified with a minimum repeat count that
1207 nigel 75 is greater than 1 or with a limited maximum, more memory is required for the
1208 nigel 63 compiled pattern, in proportion to the size of the minimum or maximum.
1209 nigel 75 .P
1210 nigel 63 If a pattern starts with .* or .{0,} and the PCRE_DOTALL option (equivalent
1211 nigel 93 to Perl's /s) is set, thus allowing the dot to match newlines, the pattern is
1212 nigel 63 implicitly anchored, because whatever follows will be tried against every
1213     character position in the subject string, so there is no point in retrying the
1214     overall match at any position after the first. PCRE normally treats such a
1215 nigel 75 pattern as though it were preceded by \eA.
1216     .P
1217 nigel 63 In cases where it is known that the subject string contains no newlines, it is
1218     worth setting PCRE_DOTALL in order to obtain this optimization, or
1219     alternatively using ^ to indicate anchoring explicitly.
1220 nigel 75 .P
1221 nigel 63 However, there is one situation where the optimization cannot be used. When .*
1222     is inside capturing parentheses that are the subject of a backreference
1223 nigel 93 elsewhere in the pattern, a match at the start may fail where a later one
1224     succeeds. Consider, for example:
1225 nigel 75 .sp
1226     (.*)abc\e1
1227     .sp
1228 nigel 63 If the subject is "xyz123abc123" the match point is the fourth character. For
1229     this reason, such a pattern is not implicitly anchored.
1230 nigel 75 .P
1231 nigel 63 When a capturing subpattern is repeated, the value captured is the substring
1232     that matched the final iteration. For example, after
1233 nigel 75 .sp
1234     (tweedle[dume]{3}\es*)+
1235     .sp
1236 nigel 63 has matched "tweedledum tweedledee" the value of the captured substring is
1237     "tweedledee". However, if there are nested capturing subpatterns, the
1238     corresponding captured values may have been set in previous iterations. For
1239     example, after
1240 nigel 75 .sp
1241 nigel 63 /(a|(b))+/
1242 nigel 75 .sp
1243 nigel 63 matches "aba" the value of the second captured substring is "b".
1244 nigel 75 .
1245     .
1246     .\" HTML <a name="atomicgroup"></a>
1248 nigel 63 .rs
1249     .sp
1250 nigel 93 With both maximizing ("greedy") and minimizing ("ungreedy" or "lazy")
1251     repetition, failure of what follows normally causes the repeated item to be
1252     re-evaluated to see if a different number of repeats allows the rest of the
1253     pattern to match. Sometimes it is useful to prevent this, either to change the
1254     nature of the match, or to cause it fail earlier than it otherwise might, when
1255     the author of the pattern knows there is no point in carrying on.
1256 nigel 75 .P
1257     Consider, for example, the pattern \ed+foo when applied to the subject line
1258     .sp
1259 nigel 63 123456bar
1260 nigel 75 .sp
1261 nigel 63 After matching all 6 digits and then failing to match "foo", the normal
1262 nigel 75 action of the matcher is to try again with only 5 digits matching the \ed+
1263 nigel 63 item, and then with 4, and so on, before ultimately failing. "Atomic grouping"
1264     (a term taken from Jeffrey Friedl's book) provides the means for specifying
1265     that once a subpattern has matched, it is not to be re-evaluated in this way.
1266 nigel 75 .P
1267 nigel 93 If we use atomic grouping for the previous example, the matcher gives up
1268 nigel 63 immediately on failing to match "foo" the first time. The notation is a kind of
1269     special parenthesis, starting with (?> as in this example:
1270 nigel 75 .sp
1271     (?>\ed+)foo
1272     .sp
1273 nigel 63 This kind of parenthesis "locks up" the part of the pattern it contains once
1274     it has matched, and a failure further into the pattern is prevented from
1275     backtracking into it. Backtracking past it to previous items, however, works as
1276     normal.
1277 nigel 75 .P
1278 nigel 63 An alternative description is that a subpattern of this type matches the string
1279     of characters that an identical standalone pattern would match, if anchored at
1280     the current point in the subject string.
1281 nigel 75 .P
1282 nigel 63 Atomic grouping subpatterns are not capturing subpatterns. Simple cases such as
1283     the above example can be thought of as a maximizing repeat that must swallow
1284 nigel 75 everything it can. So, while both \ed+ and \ed+? are prepared to adjust the
1285 nigel 63 number of digits they match in order to make the rest of the pattern match,
1286 nigel 75 (?>\ed+) can only match an entire sequence of digits.
1287     .P
1288 nigel 63 Atomic groups in general can of course contain arbitrarily complicated
1289     subpatterns, and can be nested. However, when the subpattern for an atomic
1290     group is just a single repeated item, as in the example above, a simpler
1291     notation, called a "possessive quantifier" can be used. This consists of an
1292     additional + character following a quantifier. Using this notation, the
1293     previous example can be rewritten as
1294 nigel 75 .sp
1295     \ed++foo
1296     .sp
1297 ph10 208 Note that a possessive quantifier can be used with an entire group, for
1298     example:
1299     .sp
1300     (abc|xyz){2,3}+
1301     .sp
1302 nigel 63 Possessive quantifiers are always greedy; the setting of the PCRE_UNGREEDY
1303     option is ignored. They are a convenient notation for the simpler forms of
1304 nigel 93 atomic group. However, there is no difference in the meaning of a possessive
1305     quantifier and the equivalent atomic group, though there may be a performance
1306     difference; possessive quantifiers should be slightly faster.
1307 nigel 75 .P
1308 nigel 93 The possessive quantifier syntax is an extension to the Perl 5.8 syntax.
1309     Jeffrey Friedl originated the idea (and the name) in the first edition of his
1310     book. Mike McCloskey liked it, so implemented it when he built Sun's Java
1311     package, and PCRE copied it from there. It ultimately found its way into Perl
1312     at release 5.10.
1313 nigel 75 .P
1314 nigel 93 PCRE has an optimization that automatically "possessifies" certain simple
1315     pattern constructs. For example, the sequence A+B is treated as A++B because
1316     there is no point in backtracking into a sequence of A's when B must follow.
1317     .P
1318 nigel 63 When a pattern contains an unlimited repeat inside a subpattern that can itself
1319     be repeated an unlimited number of times, the use of an atomic group is the
1320     only way to avoid some failing matches taking a very long time indeed. The
1321     pattern
1322 nigel 75 .sp
1323     (\eD+|<\ed+>)*[!?]
1324     .sp
1325 nigel 63 matches an unlimited number of substrings that either consist of non-digits, or
1326     digits enclosed in <>, followed by either ! or ?. When it matches, it runs
1327     quickly. However, if it is applied to
1328 nigel 75 .sp
1329 nigel 63 aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
1330 nigel 75 .sp
1331 nigel 63 it takes a long time before reporting failure. This is because the string can
1332 nigel 75 be divided between the internal \eD+ repeat and the external * repeat in a
1333     large number of ways, and all have to be tried. (The example uses [!?] rather
1334     than a single character at the end, because both PCRE and Perl have an
1335     optimization that allows for fast failure when a single character is used. They
1336     remember the last single character that is required for a match, and fail early
1337     if it is not present in the string.) If the pattern is changed so that it uses
1338     an atomic group, like this:
1339     .sp
1340     ((?>\eD+)|<\ed+>)*[!?]
1341     .sp
1342 nigel 63 sequences of non-digits cannot be broken, and failure happens quickly.
1343 nigel 75 .
1344     .
1345     .\" HTML <a name="backreferences"></a>
1347 nigel 63 .rs
1348     .sp
1349     Outside a character class, a backslash followed by a digit greater than 0 (and
1350     possibly further digits) is a back reference to a capturing subpattern earlier
1351     (that is, to its left) in the pattern, provided there have been that many
1352     previous capturing left parentheses.
1353 nigel 75 .P
1354 nigel 63 However, if the decimal number following the backslash is less than 10, it is
1355     always taken as a back reference, and causes an error only if there are not
1356     that many capturing left parentheses in the entire pattern. In other words, the
1357     parentheses that are referenced need not be to the left of the reference for
1358 nigel 91 numbers less than 10. A "forward back reference" of this type can make sense
1359     when a repetition is involved and the subpattern to the right has participated
1360     in an earlier iteration.
1361     .P
1362 nigel 93 It is not possible to have a numerical "forward back reference" to a subpattern
1363     whose number is 10 or more using this syntax because a sequence such as \e50 is
1364     interpreted as a character defined in octal. See the subsection entitled
1365 nigel 91 "Non-printing characters"
1366 nigel 75 .\" HTML <a href="#digitsafterbackslash">
1367     .\" </a>
1368     above
1369     .\"
1370 nigel 93 for further details of the handling of digits following a backslash. There is
1371     no such problem when named parentheses are used. A back reference to any
1372     subpattern is possible using named parentheses (see below).
1373 nigel 75 .P
1374 nigel 93 Another way of avoiding the ambiguity inherent in the use of digits following a
1375     backslash is to use the \eg escape sequence, which is a feature introduced in
1376 ph10 208 Perl 5.10. This escape must be followed by an unsigned number or a negative
1377     number, optionally enclosed in braces. These examples are all identical:
1378 nigel 93 .sp
1379     (ring), \e1
1380     (ring), \eg1
1381     (ring), \eg{1}
1382     .sp
1383 ph10 208 An unsigned number specifies an absolute reference without the ambiguity that
1384     is present in the older syntax. It is also useful when literal digits follow
1385     the reference. A negative number is a relative reference. Consider this
1386     example:
1387 nigel 93 .sp
1388     (abc(def)ghi)\eg{-1}
1389     .sp
1390     The sequence \eg{-1} is a reference to the most recently started capturing
1391     subpattern before \eg, that is, is it equivalent to \e2. Similarly, \eg{-2}
1392     would be equivalent to \e1. The use of relative references can be helpful in
1393     long patterns, and also in patterns that are created by joining together
1394     fragments that contain references within themselves.
1395     .P
1396 nigel 63 A back reference matches whatever actually matched the capturing subpattern in
1397     the current subject string, rather than anything matching the subpattern
1398     itself (see
1399     .\" HTML <a href="#subpatternsassubroutines">
1400     .\" </a>
1401     "Subpatterns as subroutines"
1402     .\"
1403     below for a way of doing that). So the pattern
1404 nigel 75 .sp
1405     (sens|respons)e and \e1ibility
1406     .sp
1407 nigel 63 matches "sense and sensibility" and "response and responsibility", but not
1408     "sense and responsibility". If caseful matching is in force at the time of the
1409     back reference, the case of letters is relevant. For example,
1410 nigel 75 .sp
1411     ((?i)rah)\es+\e1
1412     .sp
1413 nigel 63 matches "rah rah" and "RAH RAH", but not "RAH rah", even though the original
1414     capturing subpattern is matched caselessly.
1415 nigel 75 .P
1416 ph10 171 There are several different ways of writing back references to named
1417     subpatterns. The .NET syntax \ek{name} and the Perl syntax \ek<name> or
1418     \ek'name' are supported, as is the Python syntax (?P=name). Perl 5.10's unified
1419     back reference syntax, in which \eg can be used for both numeric and named
1420     references, is also supported. We could rewrite the above example in any of
1421 nigel 93 the following ways:
1422 nigel 75 .sp
1423 nigel 93 (?<p1>(?i)rah)\es+\ek<p1>
1424 ph10 171 (?'p1'(?i)rah)\es+\ek{p1}
1425 nigel 91 (?P<p1>(?i)rah)\es+(?P=p1)
1426 ph10 171 (?<p1>(?i)rah)\es+\eg{p1}
1427 nigel 75 .sp
1428 nigel 91 A subpattern that is referenced by name may appear in the pattern before or
1429     after the reference.
1430     .P
1431 nigel 63 There may be more than one back reference to the same subpattern. If a
1432     subpattern has not actually been used in a particular match, any back
1433     references to it always fail. For example, the pattern
1434 nigel 75 .sp
1435     (a|(bc))\e2
1436     .sp
1437 nigel 63 always fails if it starts to match "a" rather than "bc". Because there may be
1438     many capturing parentheses in a pattern, all digits following the backslash are
1439     taken as part of a potential back reference number. If the pattern continues
1440     with a digit character, some delimiter must be used to terminate the back
1441     reference. If the PCRE_EXTENDED option is set, this can be whitespace.
1442 nigel 75 Otherwise an empty comment (see
1443     .\" HTML <a href="#comments">
1444     .\" </a>
1445     "Comments"
1446     .\"
1447     below) can be used.
1448     .P
1449 nigel 63 A back reference that occurs inside the parentheses to which it refers fails
1450 nigel 75 when the subpattern is first used, so, for example, (a\e1) never matches.
1451 nigel 63 However, such references can be useful inside repeated subpatterns. For
1452     example, the pattern
1453 nigel 75 .sp
1454     (a|b\e1)+
1455     .sp
1456 nigel 63 matches any number of "a"s and also "aba", "ababbaa" etc. At each iteration of
1457     the subpattern, the back reference matches the character string corresponding
1458     to the previous iteration. In order for this to work, the pattern must be such
1459     that the first iteration does not need to match the back reference. This can be
1460     done using alternation, as in the example above, or by a quantifier with a
1461     minimum of zero.
1462 nigel 75 .
1463     .
1464     .\" HTML <a name="bigassertions"></a>
1465 nigel 63 .SH ASSERTIONS
1466     .rs
1467     .sp
1468     An assertion is a test on the characters following or preceding the current
1469     matching point that does not actually consume any characters. The simple
1470 nigel 75 assertions coded as \eb, \eB, \eA, \eG, \eZ, \ez, ^ and $ are described
1471     .\" HTML <a href="#smallassertions">
1472     .\" </a>
1473     above.
1474     .\"
1475     .P
1476 nigel 63 More complicated assertions are coded as subpatterns. There are two kinds:
1477     those that look ahead of the current position in the subject string, and those
1478 nigel 75 that look behind it. An assertion subpattern is matched in the normal way,
1479     except that it does not cause the current matching position to be changed.
1480     .P
1481     Assertion subpatterns are not capturing subpatterns, and may not be repeated,
1482     because it makes no sense to assert the same thing several times. If any kind
1483     of assertion contains capturing subpatterns within it, these are counted for
1484     the purposes of numbering the capturing subpatterns in the whole pattern.
1485     However, substring capturing is carried out only for positive assertions,
1486     because it does not make sense for negative assertions.
1487     .
1488     .
1489     .SS "Lookahead assertions"
1490     .rs
1491     .sp
1492 nigel 91 Lookahead assertions start with (?= for positive assertions and (?! for
1493     negative assertions. For example,
1494 nigel 75 .sp
1495     \ew+(?=;)
1496     .sp
1497 nigel 63 matches a word followed by a semicolon, but does not include the semicolon in
1498     the match, and
1499 nigel 75 .sp
1500 nigel 63 foo(?!bar)
1501 nigel 75 .sp
1502 nigel 63 matches any occurrence of "foo" that is not followed by "bar". Note that the
1503     apparently similar pattern
1504 nigel 75 .sp
1505 nigel 63 (?!foo)bar
1506 nigel 75 .sp
1507 nigel 63 does not find an occurrence of "bar" that is preceded by something other than
1508     "foo"; it finds any occurrence of "bar" whatsoever, because the assertion
1509     (?!foo) is always true when the next three characters are "bar". A
1510 nigel 75 lookbehind assertion is needed to achieve the other effect.
1511     .P
1512 nigel 63 If you want to force a matching failure at some point in a pattern, the most
1513     convenient way to do it is with (?!) because an empty string always matches, so
1514     an assertion that requires there not to be an empty string must always fail.
1515 nigel 75 .
1516     .
1517     .\" HTML <a name="lookbehind"></a>
1518     .SS "Lookbehind assertions"
1519     .rs
1520     .sp
1521 nigel 63 Lookbehind assertions start with (?<= for positive assertions and (?<! for
1522     negative assertions. For example,
1523 nigel 75 .sp
1524 nigel 63 (?<!foo)bar
1525 nigel 75 .sp
1526 nigel 63 does find an occurrence of "bar" that is not preceded by "foo". The contents of
1527     a lookbehind assertion are restricted such that all the strings it matches must
1528 nigel 91 have a fixed length. However, if there are several top-level alternatives, they
1529     do not all have to have the same fixed length. Thus
1530 nigel 75 .sp
1531 nigel 63 (?<=bullock|donkey)
1532 nigel 75 .sp
1533 nigel 63 is permitted, but
1534 nigel 75 .sp
1535 nigel 63 (?<!dogs?|cats?)
1536 nigel 75 .sp
1537 nigel 63 causes an error at compile time. Branches that match different length strings
1538     are permitted only at the top level of a lookbehind assertion. This is an
1539     extension compared with Perl (at least for 5.8), which requires all branches to
1540     match the same length of string. An assertion such as
1541 nigel 75 .sp
1542 nigel 63 (?<=ab(c|de))
1543 nigel 75 .sp
1544 nigel 63 is not permitted, because its single top-level branch can match two different
1545     lengths, but it is acceptable if rewritten to use two top-level branches:
1546 nigel 75 .sp
1547 nigel 63 (?<=abc|abde)
1548 nigel 75 .sp
1549 ph10 172 In some cases, the Perl 5.10 escape sequence \eK
1550 ph10 168 .\" HTML <a href="#resetmatchstart">
1551     .\" </a>
1552     (see above)
1553     .\"
1554     can be used instead of a lookbehind assertion; this is not restricted to a
1555     fixed-length.
1556     .P
1557 nigel 63 The implementation of lookbehind assertions is, for each alternative, to
1558 nigel 93 temporarily move the current position back by the fixed length and then try to
1559 nigel 63 match. If there are insufficient characters before the current position, the
1560 nigel 93 assertion fails.
1561 nigel 75 .P
1562     PCRE does not allow the \eC escape (which matches a single byte in UTF-8 mode)
1563 nigel 63 to appear in lookbehind assertions, because it makes it impossible to calculate
1564 nigel 93 the length of the lookbehind. The \eX and \eR escapes, which can match
1565     different numbers of bytes, are also not permitted.
1566 nigel 75 .P
1567 nigel 93 Possessive quantifiers can be used in conjunction with lookbehind assertions to
1568     specify efficient matching at the end of the subject string. Consider a simple
1569     pattern such as
1570 nigel 75 .sp
1571 nigel 63 abcd$
1572 nigel 75 .sp
1573 nigel 63 when applied to a long string that does not match. Because matching proceeds
1574     from left to right, PCRE will look for each "a" in the subject and then see if
1575     what follows matches the rest of the pattern. If the pattern is specified as
1576 nigel 75 .sp
1577 nigel 63 ^.*abcd$
1578 nigel 75 .sp
1579 nigel 63 the initial .* matches the entire string at first, but when this fails (because
1580     there is no following "a"), it backtracks to match all but the last character,
1581     then all but the last two characters, and so on. Once again the search for "a"
1582     covers the entire string, from right to left, so we are no better off. However,
1583     if the pattern is written as
1584 nigel 75 .sp
1585 nigel 63 ^.*+(?<=abcd)
1586 nigel 75 .sp
1587 nigel 93 there can be no backtracking for the .*+ item; it can match only the entire
1588 nigel 63 string. The subsequent lookbehind assertion does a single test on the last four
1589     characters. If it fails, the match fails immediately. For long strings, this
1590     approach makes a significant difference to the processing time.
1591 nigel 75 .
1592     .
1593     .SS "Using multiple assertions"
1594     .rs
1595     .sp
1596 nigel 63 Several assertions (of any sort) may occur in succession. For example,
1597 nigel 75 .sp
1598     (?<=\ed{3})(?<!999)foo
1599     .sp
1600 nigel 63 matches "foo" preceded by three digits that are not "999". Notice that each of
1601     the assertions is applied independently at the same point in the subject
1602     string. First there is a check that the previous three characters are all
1603     digits, and then there is a check that the same three characters are not "999".
1604 nigel 75 This pattern does \fInot\fP match "foo" preceded by six characters, the first
1605 nigel 63 of which are digits and the last three of which are not "999". For example, it
1606     doesn't match "123abcfoo". A pattern to do that is
1607 nigel 75 .sp
1608     (?<=\ed{3}...)(?<!999)foo
1609     .sp
1610 nigel 63 This time the first assertion looks at the preceding six characters, checking
1611     that the first three are digits, and then the second assertion checks that the
1612     preceding three characters are not "999".
1613 nigel 75 .P
1614 nigel 63 Assertions can be nested in any combination. For example,
1615 nigel 75 .sp
1616 nigel 63 (?<=(?<!foo)bar)baz
1617 nigel 75 .sp
1618 nigel 63 matches an occurrence of "baz" that is preceded by "bar" which in turn is not
1619     preceded by "foo", while
1620 nigel 75 .sp
1621     (?<=\ed{3}(?!999)...)foo
1622     .sp
1623     is another pattern that matches "foo" preceded by three digits and any three
1624 nigel 63 characters that are not "999".
1625 nigel 75 .
1626     .
1627 nigel 91 .\" HTML <a name="conditions"></a>
1629 nigel 63 .rs
1630     .sp
1631     It is possible to cause the matching process to obey a subpattern
1632     conditionally or to choose between two alternative subpatterns, depending on
1633     the result of an assertion, or whether a previous capturing subpattern matched
1634     or not. The two possible forms of conditional subpattern are
1635 nigel 75 .sp
1636 nigel 63 (?(condition)yes-pattern)
1637     (?(condition)yes-pattern|no-pattern)
1638 nigel 75 .sp
1639 nigel 63 If the condition is satisfied, the yes-pattern is used; otherwise the
1640     no-pattern (if present) is used. If there are more than two alternatives in the
1641     subpattern, a compile-time error occurs.
1642 nigel 75 .P
1643 nigel 93 There are four kinds of condition: references to subpatterns, references to
1644     recursion, a pseudo-condition called DEFINE, and assertions.
1645     .
1646     .SS "Checking for a used subpattern by number"
1647     .rs
1648     .sp
1649     If the text between the parentheses consists of a sequence of digits, the
1650     condition is true if the capturing subpattern of that number has previously
1651 ph10 172 matched. An alternative notation is to precede the digits with a plus or minus
1652 ph10 167 sign. In this case, the subpattern number is relative rather than absolute.
1653 ph10 172 The most recently opened parentheses can be referenced by (?(-1), the next most
1654 ph10 167 recent by (?(-2), and so on. In looping constructs it can also make sense to
1655     refer to subsequent groups with constructs such as (?(+2).
1656 nigel 91 .P
1657     Consider the following pattern, which contains non-significant white space to
1658     make it more readable (assume the PCRE_EXTENDED option) and to divide it into
1659     three parts for ease of discussion:
1660 nigel 75 .sp
1661     ( \e( )? [^()]+ (?(1) \e) )
1662     .sp
1663 nigel 63 The first part matches an optional opening parenthesis, and if that
1664     character is present, sets it as the first captured substring. The second part
1665     matches one or more characters that are not parentheses. The third part is a
1666     conditional subpattern that tests whether the first set of parentheses matched
1667     or not. If they did, that is, if subject started with an opening parenthesis,
1668     the condition is true, and so the yes-pattern is executed and a closing
1669     parenthesis is required. Otherwise, since no-pattern is not present, the
1670     subpattern matches nothing. In other words, this pattern matches a sequence of
1671 nigel 93 non-parentheses, optionally enclosed in parentheses.
1672 ph10 167 .P
1673 ph10 172 If you were embedding this pattern in a larger one, you could use a relative
1674 ph10 167 reference:
1675     .sp
1676     ...other stuff... ( \e( )? [^()]+ (?(-1) \e) ) ...
1677     .sp
1678     This makes the fragment independent of the parentheses in the larger pattern.
1679 nigel 93 .
1680     .SS "Checking for a used subpattern by name"
1681     .rs
1682 nigel 91 .sp
1683 nigel 93 Perl uses the syntax (?(<name>)...) or (?('name')...) to test for a used
1684     subpattern by name. For compatibility with earlier versions of PCRE, which had
1685     this facility before Perl, the syntax (?(name)...) is also recognized. However,
1686     there is a possible ambiguity with this syntax, because subpattern names may
1687     consist entirely of digits. PCRE looks first for a named subpattern; if it
1688     cannot find one and the name consists entirely of digits, PCRE looks for a
1689     subpattern of that number, which must be greater than zero. Using subpattern
1690     names that consist entirely of digits is not recommended.
1691     .P
1692     Rewriting the above example to use a named subpattern gives this:
1693 nigel 91 .sp
1694 nigel 93 (?<OPEN> \e( )? [^()]+ (?(<OPEN>) \e) )
1695     .sp
1696     .
1697     .SS "Checking for pattern recursion"
1698     .rs
1699     .sp
1700 nigel 91 If the condition is the string (R), and there is no subpattern with the name R,
1701 nigel 93 the condition is true if a recursive call to the whole pattern or any
1702     subpattern has been made. If digits or a name preceded by ampersand follow the
1703     letter R, for example:
1704     .sp
1705     (?(R3)...) or (?(R&name)...)
1706     .sp
1707     the condition is true if the most recent recursion is into the subpattern whose
1708     number or name is given. This condition does not check the entire recursion
1709     stack.
1710 nigel 75 .P
1711 nigel 93 At "top level", all these recursion test conditions are false. Recursive
1712     patterns are described below.
1713     .
1714     .SS "Defining subpatterns for use by reference only"
1715     .rs
1716     .sp
1717     If the condition is the string (DEFINE), and there is no subpattern with the
1718     name DEFINE, the condition is always false. In this case, there may be only one
1719     alternative in the subpattern. It is always skipped if control reaches this
1720     point in the pattern; the idea of DEFINE is that it can be used to define
1721     "subroutines" that can be referenced from elsewhere. (The use of "subroutines"
1722     is described below.) For example, a pattern to match an IPv4 address could be
1723     written like this (ignore whitespace and line breaks):
1724     .sp
1725     (?(DEFINE) (?<byte> 2[0-4]\ed | 25[0-5] | 1\ed\ed | [1-9]?\ed) )
1726     \eb (?&byte) (\e.(?&byte)){3} \eb
1727     .sp
1728     The first part of the pattern is a DEFINE group inside which a another group
1729     named "byte" is defined. This matches an individual component of an IPv4
1730     address (a number less than 256). When matching takes place, this part of the
1731     pattern is skipped because DEFINE acts like a false condition.
1732     .P
1733     The rest of the pattern uses references to the named group to match the four
1734     dot-separated components of an IPv4 address, insisting on a word boundary at
1735     each end.
1736     .
1737     .SS "Assertion conditions"
1738     .rs
1739     .sp
1740     If the condition is not in any of the above formats, it must be an assertion.
1741 nigel 63 This may be a positive or negative lookahead or lookbehind assertion. Consider
1742     this pattern, again containing non-significant white space, and with the two
1743     alternatives on the second line:
1744 nigel 75 .sp
1745 nigel 63 (?(?=[^a-z]*[a-z])
1746 nigel 75 \ed{2}-[a-z]{3}-\ed{2} | \ed{2}-\ed{2}-\ed{2} )
1747     .sp
1748 nigel 63 The condition is a positive lookahead assertion that matches an optional
1749     sequence of non-letters followed by a letter. In other words, it tests for the
1750     presence of at least one letter in the subject. If a letter is found, the
1751     subject is matched against the first alternative; otherwise it is matched
1752     against the second. This pattern matches strings in one of the two forms
1753     dd-aaa-dd or dd-dd-dd, where aaa are letters and dd are digits.
1754 nigel 75 .
1755     .
1756     .\" HTML <a name="comments"></a>
1757 nigel 63 .SH COMMENTS
1758     .rs
1759     .sp
1760 nigel 75 The sequence (?# marks the start of a comment that continues up to the next
1761 nigel 63 closing parenthesis. Nested parentheses are not permitted. The characters
1762     that make up a comment play no part in the pattern matching at all.
1763 nigel 75 .P
1764 nigel 63 If the PCRE_EXTENDED option is set, an unescaped # character outside a
1765 nigel 91 character class introduces a comment that continues to immediately after the
1766     next newline in the pattern.
1767 nigel 75 .
1768     .
1769 nigel 91 .\" HTML <a name="recursion"></a>
1770 nigel 75 .SH "RECURSIVE PATTERNS"
1771 nigel 63 .rs
1772     .sp
1773     Consider the problem of matching a string in parentheses, allowing for
1774     unlimited nested parentheses. Without the use of recursion, the best that can
1775     be done is to use a pattern that matches up to some fixed depth of nesting. It
1776 nigel 93 is not possible to handle an arbitrary nesting depth.
1777     .P
1778     For some time, Perl has provided a facility that allows regular expressions to
1779     recurse (amongst other things). It does this by interpolating Perl code in the
1780     expression at run time, and the code can refer to the expression itself. A Perl
1781     pattern using code interpolation to solve the parentheses problem can be
1782     created like this:
1783 nigel 75 .sp
1784     $re = qr{\e( (?: (?>[^()]+) | (?p{$re}) )* \e)}x;
1785     .sp
1786 nigel 63 The (?p{...}) item interpolates Perl code at run time, and in this case refers
1787 nigel 93 recursively to the pattern in which it appears.
1788 nigel 75 .P
1789 nigel 93 Obviously, PCRE cannot support the interpolation of Perl code. Instead, it
1790     supports special syntax for recursion of the entire pattern, and also for
1791     individual subpattern recursion. After its introduction in PCRE and Python,
1792     this kind of recursion was introduced into Perl at release 5.10.
1793 nigel 75 .P
1794 nigel 93 A special item that consists of (? followed by a number greater than zero and a
1795     closing parenthesis is a recursive call of the subpattern of the given number,
1796     provided that it occurs inside that subpattern. (If not, it is a "subroutine"
1797     call, which is described in the next section.) The special item (?R) or (?0) is
1798     a recursive call of the entire regular expression.
1799 nigel 87 .P
1800 nigel 93 In PCRE (like Python, but unlike Perl), a recursive subpattern call is always
1801     treated as an atomic group. That is, once it has matched some of the subject
1802     string, it is never re-entered, even if it contains untried alternatives and
1803     there is a subsequent matching failure.
1804     .P
1805 nigel 87 This PCRE pattern solves the nested parentheses problem (assume the
1806     PCRE_EXTENDED option is set so that white space is ignored):
1807 nigel 75 .sp
1808     \e( ( (?>[^()]+) | (?R) )* \e)
1809     .sp
1810 nigel 63 First it matches an opening parenthesis. Then it matches any number of
1811     substrings which can either be a sequence of non-parentheses, or a recursive
1812 nigel 87 match of the pattern itself (that is, a correctly parenthesized substring).
1813 nigel 63 Finally there is a closing parenthesis.
1814 nigel 75 .P
1815 nigel 63 If this were part of a larger pattern, you would not want to recurse the entire
1816     pattern, so instead you could use this:
1817 nigel 75 .sp
1818     ( \e( ( (?>[^()]+) | (?1) )* \e) )
1819     .sp
1820 nigel 63 We have put the pattern into parentheses, and caused the recursion to refer to
1821 ph10 172 them instead of the whole pattern.
1822 ph10 166 .P
1823     In a larger pattern, keeping track of parenthesis numbers can be tricky. This
1824     is made easier by the use of relative references. (A Perl 5.10 feature.)
1825     Instead of (?1) in the pattern above you can write (?-2) to refer to the second
1826     most recently opened parentheses preceding the recursion. In other words, a
1827     negative number counts capturing parentheses leftwards from the point at which
1828     it is encountered.
1829     .P
1830     It is also possible to refer to subsequently opened parentheses, by writing
1831     references such as (?+2). However, these cannot be recursive because the
1832     reference is not inside the parentheses that are referenced. They are always
1833     "subroutine" calls, as described in the next section.
1834     .P
1835     An alternative approach is to use named parentheses instead. The Perl syntax
1836     for this is (?&name); PCRE's earlier syntax (?P>name) is also supported. We
1837     could rewrite the above example as follows:
1838 nigel 75 .sp
1839 nigel 93 (?<pn> \e( ( (?>[^()]+) | (?&pn) )* \e) )
1840 nigel 75 .sp
1841 nigel 93 If there is more than one subpattern with the same name, the earliest one is
1842 ph10 172 used.
1843 ph10 166 .P
1844     This particular example pattern that we have been looking at contains nested
1845     unlimited repeats, and so the use of atomic grouping for matching strings of
1846     non-parentheses is important when applying the pattern to strings that do not
1847     match. For example, when this pattern is applied to
1848 nigel 75 .sp
1849 nigel 63 (aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa()
1850 nigel 75 .sp
1851 nigel 63 it yields "no match" quickly. However, if atomic grouping is not used,
1852     the match runs for a very long time indeed because there are so many different
1853     ways the + and * repeats can carve up the subject, and all have to be tested
1854     before failure can be reported.
1855 nigel 75 .P
1856 nigel 63 At the end of a match, the values set for any capturing subpatterns are those
1857     from the outermost level of the recursion at which the subpattern value is set.
1858     If you want to obtain intermediate values, a callout function can be used (see
1859 nigel 93 below and the
1860 nigel 63 .\" HREF
1861 nigel 75 \fBpcrecallout\fP
1862 nigel 63 .\"
1863     documentation). If the pattern above is matched against
1864 nigel 75 .sp
1865 nigel 63 (ab(cd)ef)
1866 nigel 75 .sp
1867 nigel 63 the value for the capturing parentheses is "ef", which is the last value taken
1868     on at the top level. If additional parentheses are added, giving
1869 nigel 75 .sp
1870     \e( ( ( (?>[^()]+) | (?R) )* ) \e)
1871 nigel 63 ^ ^
1872     ^ ^
1873 nigel 75 .sp
1874 nigel 63 the string they capture is "ab(cd)ef", the contents of the top level
1875     parentheses. If there are more than 15 capturing parentheses in a pattern, PCRE
1876     has to obtain extra memory to store data during a recursion, which it does by
1877 nigel 75 using \fBpcre_malloc\fP, freeing it via \fBpcre_free\fP afterwards. If no
1878 nigel 63 memory can be obtained, the match fails with the PCRE_ERROR_NOMEMORY error.
1879 nigel 75 .P
1880 nigel 63 Do not confuse the (?R) item with the condition (R), which tests for recursion.
1881     Consider this pattern, which matches text in angle brackets, allowing for
1882     arbitrary nesting. Only digits are allowed in nested brackets (that is, when
1883     recursing), whereas any characters are permitted at the outer level.
1884 nigel 75 .sp
1885     < (?: (?(R) \ed++ | [^<>]*+) | (?R)) * >
1886     .sp
1887 nigel 63 In this pattern, (?(R) is the start of a conditional subpattern, with two
1888     different alternatives for the recursive and non-recursive cases. The (?R) item
1889     is the actual recursive call.
1890 nigel 75 .
1891     .
1892 nigel 63 .\" HTML <a name="subpatternsassubroutines"></a>
1894 nigel 63 .rs
1895     .sp
1896     If the syntax for a recursive subpattern reference (either by number or by
1897     name) is used outside the parentheses to which it refers, it operates like a
1898 nigel 93 subroutine in a programming language. The "called" subpattern may be defined
1899 ph10 166 before or after the reference. A numbered reference can be absolute or
1900     relative, as in these examples:
1901 nigel 75 .sp
1902 ph10 166 (...(absolute)...)...(?2)...
1903     (...(relative)...)...(?-1)...
1904 ph10 172 (...(?+1)...(relative)...
1905 ph10 166 .sp
1906     An earlier example pointed out that the pattern
1907     .sp
1908 nigel 75 (sens|respons)e and \e1ibility
1909     .sp
1910 nigel 63 matches "sense and sensibility" and "response and responsibility", but not
1911     "sense and responsibility". If instead the pattern
1912 nigel 75 .sp
1913 nigel 63 (sens|respons)e and (?1)ibility
1914 nigel 75 .sp
1915 nigel 63 is used, it does match "sense and responsibility" as well as the other two
1916 nigel 93 strings. Another example is given in the discussion of DEFINE above.
1917 nigel 87 .P
1918     Like recursive subpatterns, a "subroutine" call is always treated as an atomic
1919     group. That is, once it has matched some of the subject string, it is never
1920     re-entered, even if it contains untried alternatives and there is a subsequent
1921     matching failure.
1922 nigel 93 .P
1923     When a subpattern is used as a subroutine, processing options such as
1924     case-independence are fixed when the subpattern is defined. They cannot be
1925     changed for different calls. For example, consider this pattern:
1926     .sp
1927 ph10 166 (abc)(?i:(?-1))
1928 nigel 93 .sp
1929     It matches "abcabc". It does not match "abcABC" because the change of
1930     processing option does not affect the called subpattern.
1931 nigel 75 .
1932     .
1933 nigel 63 .SH CALLOUTS
1934     .rs
1935     .sp
1936     Perl has a feature whereby using the sequence (?{...}) causes arbitrary Perl
1937     code to be obeyed in the middle of matching a regular expression. This makes it
1938     possible, amongst other things, to extract different substrings that match the
1939     same pair of parentheses when there is a repetition.
1940 nigel 75 .P
1941 nigel 63 PCRE provides a similar feature, but of course it cannot obey arbitrary Perl
1942     code. The feature is called "callout". The caller of PCRE provides an external
1943 nigel 75 function by putting its entry point in the global variable \fIpcre_callout\fP.
1944 nigel 63 By default, this variable contains NULL, which disables all calling out.
1945 nigel 75 .P
1946 nigel 63 Within a regular expression, (?C) indicates the points at which the external
1947     function is to be called. If you want to identify different callout points, you
1948     can put a number less than 256 after the letter C. The default value is zero.
1949     For example, this pattern has two callout points:
1950 nigel 75 .sp
1951 ph10 155 (?C1)abc(?C2)def
1952 nigel 75 .sp
1953     If the PCRE_AUTO_CALLOUT flag is passed to \fBpcre_compile()\fP, callouts are
1954     automatically installed before each item in the pattern. They are all numbered
1955     255.
1956     .P
1957     During matching, when PCRE reaches a callout point (and \fIpcre_callout\fP is
1958 nigel 63 set), the external function is called. It is provided with the number of the
1959 nigel 75 callout, the position in the pattern, and, optionally, one item of data
1960     originally supplied by the caller of \fBpcre_exec()\fP. The callout function
1961     may cause matching to proceed, to backtrack, or to fail altogether. A complete
1962     description of the interface to the callout function is given in the
1963 nigel 63 .\" HREF
1964 nigel 75 \fBpcrecallout\fP
1965 nigel 63 .\"
1966     documentation.
1967 nigel 93 .
1968     .
1969     .SH "SEE ALSO"
1970     .rs
1971     .sp
1972     \fBpcreapi\fP(3), \fBpcrecallout\fP(3), \fBpcrematching\fP(3), \fBpcre\fP(3).
1973 ph10 99 .
1974     .
1975     .SH AUTHOR
1976     .rs
1977     .sp
1978     .nf
1979     Philip Hazel
1980     University Computing Service
1981     Cambridge CB2 3QH, England.
1982     .fi
1983     .
1984     .
1985     .SH REVISION
1986     .rs
1987     .sp
1988     .nf
1989 ph10 208 Last updated: 06 August 2007
1990 ph10 99 Copyright (c) 1997-2007 University of Cambridge.
1991     .fi


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

ViewVC Help
Powered by ViewVC 1.1.12