ViewVC logotype

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

Parent Directory Parent Directory | Revision Log Revision Log

Revision 335 - (hide annotations) (download)
Sat Apr 12 14:36:14 2008 UTC (7 years ago) by ph10
File size: 91309 byte(s)
Do not discard subpatterns with {0} quantifiers, as they may be called as 

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 ph10 333 page. PCRE tries to match Perl syntax and semantics as closely as it can. PCRE
13     also supports some alternative regular expression syntax (which does not
14     conflict with the Perl syntax) in order to provide some compatibility with
15     regular expressions in Python, .NET, and Oniguruma.
16     .P
17     Perl's regular expressions are described in its own documentation, and
18 ph10 208 regular expressions in general are covered in a number of books, some of which
19     have copious examples. Jeffrey Friedl's "Mastering Regular Expressions",
20     published by O'Reilly, covers regular expressions in great detail. This
21     description of PCRE's regular expressions is intended as reference material.
22 nigel 75 .P
23     The original operation of PCRE was on strings of one-byte characters. However,
24     there is now also support for UTF-8 character strings. To use this, you must
25     build PCRE to include UTF-8 support, and then call \fBpcre_compile()\fP with
26     the PCRE_UTF8 option. How this affects pattern matching is mentioned in several
27     places below. There is also a summary of UTF-8 features in the
28 nigel 63 .\" HTML <a href="pcre.html#utf8support">
29     .\" </a>
30     section on UTF-8 support
31     .\"
32     in the main
33     .\" HREF
34 nigel 75 \fBpcre\fP
35 nigel 63 .\"
36     page.
37 nigel 75 .P
38 nigel 77 The remainder of this document discusses the patterns that are supported by
39     PCRE when its main matching function, \fBpcre_exec()\fP, is used.
40     From release 6.0, PCRE offers a second matching function,
41     \fBpcre_dfa_exec()\fP, which matches using a different algorithm that is not
42 ph10 172 Perl-compatible. Some of the features discussed below are not available when
43 ph10 168 \fBpcre_dfa_exec()\fP is used. The advantages and disadvantages of the
44     alternative function, and how it differs from the normal function, are
45     discussed in the
46 nigel 77 .\" HREF
47     \fBpcrematching\fP
48     .\"
49     page.
50 nigel 93 .
51     .
53     .rs
54     .sp
55     PCRE supports five different conventions for indicating line breaks in
56     strings: a single CR (carriage return) character, a single LF (linefeed)
57     character, the two-character sequence CRLF, any of the three preceding, or any
58     Unicode newline sequence. The
59     .\" HREF
60     \fBpcreapi\fP
61     .\"
62     page has
63     .\" HTML <a href="pcreapi.html#newlines">
64     .\" </a>
65     further discussion
66     .\"
67     about newlines, and shows how to set the newline convention in the
68     \fIoptions\fP arguments for the compiling and matching functions.
69     .P
70     It is also possible to specify a newline convention by starting a pattern
71     string with one of the following five sequences:
72     .sp
73     (*CR) carriage return
74     (*LF) linefeed
75     (*CRLF) carriage return, followed by linefeed
76     (*ANYCRLF) any of the three above
77     (*ANY) all Unicode newline sequences
78     .sp
79     These override the default and the options given to \fBpcre_compile()\fP. For
80     example, on a Unix system where LF is the default newline sequence, the pattern
81     .sp
82     (*CR)a.b
83     .sp
84     changes the convention to CR. That pattern matches "a\enb" because LF is no
85     longer a newline. Note that these special settings, which are not
86     Perl-compatible, are recognized only at the very start of a pattern, and that
87 ph10 231 they must be in upper case. If more than one of them is present, the last one
88     is used.
89     .P
90     The newline convention does not affect what the \eR escape sequence matches. By
91     default, this is any Unicode newline sequence, for Perl compatibility. However,
92     this can be changed; see the description of \eR in the section entitled
93     .\" HTML <a href="#newlineseq">
94     .\" </a>
95     "Newline sequences"
96     .\"
97 ph10 247 below. A change of \eR setting can be combined with a change of newline
98 ph10 246 convention.
99 ph10 227 .
100     .
102     .rs
103     .sp
104 nigel 63 A regular expression is a pattern that is matched against a subject string from
105     left to right. Most characters stand for themselves in a pattern, and match the
106     corresponding characters in the subject. As a trivial example, the pattern
107 nigel 75 .sp
108 nigel 63 The quick brown fox
109 nigel 75 .sp
110 nigel 77 matches a portion of a subject string that is identical to itself. When
111     caseless matching is specified (the PCRE_CASELESS option), letters are matched
112     independently of case. In UTF-8 mode, PCRE always understands the concept of
113     case for characters whose values are less than 128, so caseless matching is
114     always possible. For characters with higher values, the concept of case is
115     supported if PCRE is compiled with Unicode property support, but not otherwise.
116     If you want to use caseless matching for characters 128 and above, you must
117     ensure that PCRE is compiled with Unicode property support as well as with
118     UTF-8 support.
119     .P
120     The power of regular expressions comes from the ability to include alternatives
121     and repetitions in the pattern. These are encoded in the pattern by the use of
122 nigel 75 \fImetacharacters\fP, which do not stand for themselves but instead are
123 nigel 63 interpreted in some special way.
124 nigel 75 .P
125     There are two different sets of metacharacters: those that are recognized
126 nigel 63 anywhere in the pattern except within square brackets, and those that are
127 nigel 93 recognized within square brackets. Outside square brackets, the metacharacters
128     are as follows:
129 nigel 75 .sp
130     \e general escape character with several uses
131 nigel 63 ^ assert start of string (or line, in multiline mode)
132     $ assert end of string (or line, in multiline mode)
133     . match any character except newline (by default)
134     [ start character class definition
135     | start of alternative branch
136     ( start subpattern
137     ) end subpattern
138     ? extends the meaning of (
139     also 0 or 1 quantifier
140     also quantifier minimizer
141     * 0 or more quantifier
142     + 1 or more quantifier
143     also "possessive quantifier"
144     { start min/max quantifier
145 nigel 75 .sp
146 nigel 63 Part of a pattern that is in square brackets is called a "character class". In
147 nigel 75 a character class the only metacharacters are:
148     .sp
149     \e general escape character
150 nigel 63 ^ negate the class, but only if the first character
151     - indicates character range
152 nigel 75 .\" JOIN
153 nigel 63 [ POSIX character class (only if followed by POSIX
154     syntax)
155     ] terminates the character class
156 nigel 75 .sp
157     The following sections describe the use of each of the metacharacters.
158     .
159 nigel 93 .
160 nigel 63 .SH BACKSLASH
161     .rs
162     .sp
163     The backslash character has several uses. Firstly, if it is followed by a
164 nigel 91 non-alphanumeric character, it takes away any special meaning that character
165     may have. This use of backslash as an escape character applies both inside and
166 nigel 63 outside character classes.
167 nigel 75 .P
168     For example, if you want to match a * character, you write \e* in the pattern.
169 nigel 63 This escaping action applies whether or not the following character would
170 nigel 75 otherwise be interpreted as a metacharacter, so it is always safe to precede a
171     non-alphanumeric with backslash to specify that it stands for itself. In
172     particular, if you want to match a backslash, you write \e\e.
173     .P
174 nigel 63 If a pattern is compiled with the PCRE_EXTENDED option, whitespace in the
175     pattern (other than in a character class) and characters between a # outside
176 nigel 91 a character class and the next newline are ignored. An escaping backslash can
177     be used to include a whitespace or # character as part of the pattern.
178 nigel 75 .P
179 nigel 63 If you want to remove the special meaning from a sequence of characters, you
180 nigel 75 can do so by putting them between \eQ and \eE. This is different from Perl in
181     that $ and @ are handled as literals in \eQ...\eE sequences in PCRE, whereas in
182 nigel 63 Perl, $ and @ cause variable interpolation. Note the following examples:
183 nigel 75 .sp
184 nigel 63 Pattern PCRE matches Perl matches
185 nigel 75 .sp
186     .\" JOIN
187     \eQabc$xyz\eE abc$xyz abc followed by the
188 nigel 63 contents of $xyz
189 nigel 75 \eQabc\e$xyz\eE abc\e$xyz abc\e$xyz
190     \eQabc\eE\e$\eQxyz\eE abc$xyz abc$xyz
191     .sp
192     The \eQ...\eE sequence is recognized both inside and outside character classes.
193     .
194     .
195     .\" HTML <a name="digitsafterbackslash"></a>
196     .SS "Non-printing characters"
197     .rs
198     .sp
199 nigel 63 A second use of backslash provides a way of encoding non-printing characters
200     in patterns in a visible manner. There is no restriction on the appearance of
201     non-printing characters, apart from the binary zero that terminates a pattern,
202     but when a pattern is being prepared by text editing, it is usually easier to
203     use one of the following escape sequences than the binary character it
204     represents:
205 nigel 75 .sp
206     \ea alarm, that is, the BEL character (hex 07)
207     \ecx "control-x", where x is any character
208     \ee escape (hex 1B)
209     \ef formfeed (hex 0C)
210 ph10 227 \en linefeed (hex 0A)
211 nigel 75 \er carriage return (hex 0D)
212     \et tab (hex 09)
213     \eddd character with octal code ddd, or backreference
214     \exhh character with hex code hh
215 nigel 87 \ex{hhh..} character with hex code hhh..
216 nigel 75 .sp
217     The precise effect of \ecx is as follows: if x is a lower case letter, it
218 nigel 63 is converted to upper case. Then bit 6 of the character (hex 40) is inverted.
219 nigel 75 Thus \ecz becomes hex 1A, but \ec{ becomes hex 3B, while \ec; becomes hex
220 nigel 63 7B.
221 nigel 75 .P
222     After \ex, from zero to two hexadecimal digits are read (letters can be in
223 nigel 87 upper or lower case). Any number of hexadecimal digits may appear between \ex{
224     and }, but the value of the character code must be less than 256 in non-UTF-8
225 ph10 211 mode, and less than 2**31 in UTF-8 mode. That is, the maximum value in
226     hexadecimal is 7FFFFFFF. Note that this is bigger than the largest Unicode code
227     point, which is 10FFFF.
228 nigel 75 .P
229 ph10 211 If characters other than hexadecimal digits appear between \ex{ and }, or if
230     there is no terminating }, this form of escape is not recognized. Instead, the
231     initial \ex will be interpreted as a basic hexadecimal escape, with no
232     following digits, giving a character whose value is zero.
233     .P
234 nigel 63 Characters whose value is less than 256 can be defined by either of the two
235 nigel 87 syntaxes for \ex. There is no difference in the way they are handled. For
236     example, \exdc is exactly the same as \ex{dc}.
237 nigel 75 .P
238 nigel 91 After \e0 up to two further octal digits are read. If there are fewer than two
239     digits, just those that are present are used. Thus the sequence \e0\ex\e07
240     specifies two binary zeros followed by a BEL character (code value 7). Make
241     sure you supply two digits after the initial zero if the pattern character that
242     follows is itself an octal digit.
243 nigel 75 .P
244 nigel 63 The handling of a backslash followed by a digit other than 0 is complicated.
245     Outside a character class, PCRE reads it and any following digits as a decimal
246     number. If the number is less than 10, or if there have been at least that many
247     previous capturing left parentheses in the expression, the entire sequence is
248 nigel 75 taken as a \fIback reference\fP. A description of how this works is given
249     .\" HTML <a href="#backreferences">
250     .\" </a>
251     later,
252     .\"
253     following the discussion of
254     .\" HTML <a href="#subpattern">
255     .\" </a>
256     parenthesized subpatterns.
257     .\"
258     .P
259 nigel 63 Inside a character class, or if the decimal number is greater than 9 and there
260     have not been that many capturing subpatterns, PCRE re-reads up to three octal
261 nigel 93 digits following the backslash, and uses them to generate a data character. Any
262 nigel 91 subsequent digits stand for themselves. In non-UTF-8 mode, the value of a
263     character specified in octal must be less than \e400. In UTF-8 mode, values up
264     to \e777 are permitted. For example:
265 nigel 75 .sp
266     \e040 is another way of writing a space
267     .\" JOIN
268     \e40 is the same, provided there are fewer than 40
269 nigel 63 previous capturing subpatterns
270 nigel 75 \e7 is always a back reference
271     .\" JOIN
272     \e11 might be a back reference, or another way of
273 nigel 63 writing a tab
274 nigel 75 \e011 is always a tab
275     \e0113 is a tab followed by the character "3"
276     .\" JOIN
277     \e113 might be a back reference, otherwise the
278 nigel 63 character with octal code 113
279 nigel 75 .\" JOIN
280     \e377 might be a back reference, otherwise
281 nigel 63 the byte consisting entirely of 1 bits
282 nigel 75 .\" JOIN
283     \e81 is either a back reference, or a binary zero
284 nigel 63 followed by the two characters "8" and "1"
285 nigel 75 .sp
286 nigel 63 Note that octal values of 100 or greater must not be introduced by a leading
287     zero, because no more than three octal digits are ever read.
288 nigel 75 .P
289 nigel 91 All the sequences that define a single character value can be used both inside
290     and outside character classes. In addition, inside a character class, the
291     sequence \eb is interpreted as the backspace character (hex 08), and the
292 nigel 93 sequences \eR and \eX are interpreted as the characters "R" and "X",
293     respectively. Outside a character class, these sequences have different
294     meanings
295 nigel 75 .\" HTML <a href="#uniextseq">
296     .\" </a>
297     (see below).
298     .\"
299     .
300     .
301 nigel 93 .SS "Absolute and relative back references"
302     .rs
303     .sp
304 ph10 208 The sequence \eg followed by an unsigned or a negative number, optionally
305     enclosed in braces, is an absolute or relative back reference. A named back
306     reference can be coded as \eg{name}. Back references are discussed
307 nigel 93 .\" HTML <a href="#backreferences">
308     .\" </a>
309     later,
310     .\"
311     following the discussion of
312     .\" HTML <a href="#subpattern">
313     .\" </a>
314     parenthesized subpatterns.
315     .\"
316     .
317     .
318 ph10 333 .SS "Absolute and relative subroutine calls"
319     .rs
320     .sp
321     For compatibility with Oniguruma, the non-Perl syntax \eg followed by a name or
322     a number enclosed either in angle brackets or single quotes, is an alternative
323     syntax for referencing a subpattern as a "subroutine". Details are discussed
324     .\" HTML <a href="#onigurumasubroutines">
325     .\" </a>
326     later.
327     .\"
328     Note that \eg{...} (Perl syntax) and \eg<...> (Oniguruma syntax) are \fInot\fP
329     synonymous. The former is a back reference; the latter is a subroutine call.
330     .
331     .
332 nigel 75 .SS "Generic character types"
333     .rs
334     .sp
335 nigel 93 Another use of backslash is for specifying generic character types. The
336 nigel 75 following are always recognized:
337     .sp
338 ph10 182 \ed any decimal digit
339 nigel 75 \eD any character that is not a decimal digit
340 ph10 178 \eh any horizontal whitespace character
341 ph10 182 \eH any character that is not a horizontal whitespace character
342 nigel 75 \es any whitespace character
343     \eS any character that is not a whitespace character
344 ph10 178 \ev any vertical whitespace character
345 ph10 182 \eV any character that is not a vertical whitespace character
346 nigel 75 \ew any "word" character
347     \eW any "non-word" character
348     .sp
349 nigel 63 Each pair of escape sequences partitions the complete set of characters into
350     two disjoint sets. Any given character matches one, and only one, of each pair.
351 nigel 75 .P
352     These character type sequences can appear both inside and outside character
353     classes. They each match one character of the appropriate type. If the current
354     matching point is at the end of the subject string, all of them fail, since
355     there is no character to match.
356     .P
357     For compatibility with Perl, \es does not match the VT character (code 11).
358     This makes it different from the the POSIX "space" class. The \es characters
359 ph10 178 are HT (9), LF (10), FF (12), CR (13), and space (32). If "use locale;" is
360 nigel 91 included in a Perl script, \es may match the VT character. In PCRE, it never
361 ph10 178 does.
362 nigel 75 .P
363 ph10 178 In UTF-8 mode, characters with values greater than 128 never match \ed, \es, or
364     \ew, and always match \eD, \eS, and \eW. This is true even when Unicode
365 ph10 182 character property support is available. These sequences retain their original
366     meanings from before UTF-8 support was available, mainly for efficiency
367 ph10 178 reasons.
368     .P
369 ph10 182 The sequences \eh, \eH, \ev, and \eV are Perl 5.10 features. In contrast to the
370 ph10 178 other sequences, these do match certain high-valued codepoints in UTF-8 mode.
371     The horizontal space characters are:
372     .sp
373     U+0009 Horizontal tab
374     U+0020 Space
375     U+00A0 Non-break space
376     U+1680 Ogham space mark
377     U+180E Mongolian vowel separator
378     U+2000 En quad
379     U+2001 Em quad
380     U+2002 En space
381     U+2003 Em space
382     U+2004 Three-per-em space
383     U+2005 Four-per-em space
384     U+2006 Six-per-em space
385     U+2007 Figure space
386     U+2008 Punctuation space
387     U+2009 Thin space
388     U+200A Hair space
389     U+202F Narrow no-break space
390     U+205F Medium mathematical space
391     U+3000 Ideographic space
392     .sp
393     The vertical space characters are:
394     .sp
395     U+000A Linefeed
396     U+000B Vertical tab
397     U+000C Formfeed
398     U+000D Carriage return
399     U+0085 Next line
400     U+2028 Line separator
401     U+2029 Paragraph separator
402     .P
403 nigel 75 A "word" character is an underscore or any character less than 256 that is a
404     letter or digit. The definition of letters and digits is controlled by PCRE's
405     low-valued character tables, and may vary if locale-specific matching is taking
406     place (see
407 nigel 63 .\" HTML <a href="pcreapi.html#localesupport">
408     .\" </a>
409     "Locale support"
410     .\"
411     in the
412     .\" HREF
413 nigel 75 \fBpcreapi\fP
414 nigel 63 .\"
415 ph10 139 page). For example, in a French locale such as "fr_FR" in Unix-like systems,
416     or "french" in Windows, some character codes greater than 128 are used for
417 ph10 178 accented letters, and these are matched by \ew. The use of locales with Unicode
418     is discouraged.
419 nigel 75 .
420     .
421 ph10 231 .\" HTML <a name="newlineseq"></a>
422 nigel 93 .SS "Newline sequences"
423     .rs
424     .sp
425 ph10 231 Outside a character class, by default, the escape sequence \eR matches any
426     Unicode newline sequence. This is a Perl 5.10 feature. In non-UTF-8 mode \eR is
427     equivalent to the following:
428 nigel 93 .sp
429     (?>\er\en|\en|\ex0b|\ef|\er|\ex85)
430     .sp
431     This is an example of an "atomic group", details of which are given
432     .\" HTML <a href="#atomicgroup">
433     .\" </a>
434     below.
435     .\"
436     This particular group matches either the two-character sequence CR followed by
437     LF, or one of the single characters LF (linefeed, U+000A), VT (vertical tab,
438     U+000B), FF (formfeed, U+000C), CR (carriage return, U+000D), or NEL (next
439     line, U+0085). The two-character sequence is treated as a single unit that
440     cannot be split.
441     .P
442     In UTF-8 mode, two additional characters whose codepoints are greater than 255
443     are added: LS (line separator, U+2028) and PS (paragraph separator, U+2029).
444     Unicode character property support is not needed for these characters to be
445     recognized.
446     .P
447 ph10 231 It is possible to restrict \eR to match only CR, LF, or CRLF (instead of the
448     complete set of Unicode line endings) by setting the option PCRE_BSR_ANYCRLF
449 ph10 247 either at compile time or when the pattern is matched. (BSR is an abbrevation
450 ph10 246 for "backslash R".) This can be made the default when PCRE is built; if this is
451     the case, the other behaviour can be requested via the PCRE_BSR_UNICODE option.
452     It is also possible to specify these settings by starting a pattern string with
453     one of the following sequences:
454 ph10 231 .sp
455     (*BSR_ANYCRLF) CR, LF, or CRLF only
456     (*BSR_UNICODE) any Unicode newline sequence
457     .sp
458     These override the default and the options given to \fBpcre_compile()\fP, but
459     they can be overridden by options given to \fBpcre_exec()\fP. Note that these
460     special settings, which are not Perl-compatible, are recognized only at the
461     very start of a pattern, and that they must be in upper case. If more than one
462 ph10 247 of them is present, the last one is used. They can be combined with a change of
463 ph10 246 newline convention, for example, a pattern can start with:
464     .sp
465     (*ANY)(*BSR_ANYCRLF)
466     .sp
467 nigel 93 Inside a character class, \eR matches the letter "R".
468     .
469     .
470 nigel 75 .\" HTML <a name="uniextseq"></a>
471     .SS Unicode character properties
472     .rs
473     .sp
474     When PCRE is built with Unicode character property support, three additional
475 ph10 184 escape sequences that match characters with specific properties are available.
476     When not in UTF-8 mode, these sequences are of course limited to testing
477     characters whose codepoints are less than 256, but they do work in this mode.
478     The extra escape sequences are:
479 nigel 75 .sp
480 nigel 87 \ep{\fIxx\fP} a character with the \fIxx\fP property
481     \eP{\fIxx\fP} a character without the \fIxx\fP property
482     \eX an extended Unicode sequence
483 nigel 75 .sp
484 nigel 87 The property names represented by \fIxx\fP above are limited to the Unicode
485     script names, the general category properties, and "Any", which matches any
486     character (including newline). Other properties such as "InMusicalSymbols" are
487     not currently supported by PCRE. Note that \eP{Any} does not match any
488     characters, so always causes a match failure.
489 nigel 75 .P
490 nigel 87 Sets of Unicode characters are defined as belonging to certain scripts. A
491     character from one of these sets can be matched using a script name. For
492     example:
493 nigel 75 .sp
494 nigel 87 \ep{Greek}
495     \eP{Han}
496     .sp
497     Those that are not part of an identified script are lumped together as
498     "Common". The current list of scripts is:
499     .P
500     Arabic,
501     Armenian,
502 nigel 93 Balinese,
503 nigel 87 Bengali,
504     Bopomofo,
505     Braille,
506     Buginese,
507     Buhid,
508     Canadian_Aboriginal,
509     Cherokee,
510     Common,
511     Coptic,
512 nigel 93 Cuneiform,
513 nigel 87 Cypriot,
514     Cyrillic,
515     Deseret,
516     Devanagari,
517     Ethiopic,
518     Georgian,
519     Glagolitic,
520     Gothic,
521     Greek,
522     Gujarati,
523     Gurmukhi,
524     Han,
525     Hangul,
526     Hanunoo,
527     Hebrew,
528     Hiragana,
529     Inherited,
530     Kannada,
531     Katakana,
532     Kharoshthi,
533     Khmer,
534     Lao,
535     Latin,
536     Limbu,
537     Linear_B,
538     Malayalam,
539     Mongolian,
540     Myanmar,
541     New_Tai_Lue,
542 nigel 93 Nko,
543 nigel 87 Ogham,
544     Old_Italic,
545     Old_Persian,
546     Oriya,
547     Osmanya,
548 nigel 93 Phags_Pa,
549     Phoenician,
550 nigel 87 Runic,
551     Shavian,
552     Sinhala,
553     Syloti_Nagri,
554     Syriac,
555     Tagalog,
556     Tagbanwa,
557     Tai_Le,
558     Tamil,
559     Telugu,
560     Thaana,
561     Thai,
562     Tibetan,
563     Tifinagh,
564     Ugaritic,
565     Yi.
566     .P
567     Each character has exactly one general category property, specified by a
568     two-letter abbreviation. For compatibility with Perl, negation can be specified
569     by including a circumflex between the opening brace and the property name. For
570     example, \ep{^Lu} is the same as \eP{Lu}.
571     .P
572     If only one letter is specified with \ep or \eP, it includes all the general
573     category properties that start with that letter. In this case, in the absence
574     of negation, the curly brackets in the escape sequence are optional; these two
575     examples have the same effect:
576     .sp
577 nigel 75 \ep{L}
578     \epL
579     .sp
580 nigel 87 The following general category property codes are supported:
581 nigel 75 .sp
582     C Other
583     Cc Control
584     Cf Format
585     Cn Unassigned
586     Co Private use
587     Cs Surrogate
588     .sp
589     L Letter
590     Ll Lower case letter
591     Lm Modifier letter
592     Lo Other letter
593     Lt Title case letter
594     Lu Upper case letter
595     .sp
596     M Mark
597     Mc Spacing mark
598     Me Enclosing mark
599     Mn Non-spacing mark
600     .sp
601     N Number
602     Nd Decimal number
603     Nl Letter number
604     No Other number
605     .sp
606     P Punctuation
607     Pc Connector punctuation
608     Pd Dash punctuation
609     Pe Close punctuation
610     Pf Final punctuation
611     Pi Initial punctuation
612     Po Other punctuation
613     Ps Open punctuation
614     .sp
615     S Symbol
616     Sc Currency symbol
617     Sk Modifier symbol
618     Sm Mathematical symbol
619     So Other symbol
620     .sp
621     Z Separator
622     Zl Line separator
623     Zp Paragraph separator
624     Zs Space separator
625     .sp
626 nigel 87 The special property L& is also supported: it matches a character that has
627     the Lu, Ll, or Lt property, in other words, a letter that is not classified as
628     a modifier or "other".
629 nigel 75 .P
630 ph10 211 The Cs (Surrogate) property applies only to characters in the range U+D800 to
631     U+DFFF. Such characters are not valid in UTF-8 strings (see RFC 3629) and so
632     cannot be tested by PCRE, unless UTF-8 validity checking has been turned off
633     (see the discussion of PCRE_NO_UTF8_CHECK in the
634     .\" HREF
635     \fBpcreapi\fP
636     .\"
637     page).
638     .P
639 nigel 87 The long synonyms for these properties that Perl supports (such as \ep{Letter})
640 nigel 91 are not supported by PCRE, nor is it permitted to prefix any of these
641 nigel 87 properties with "Is".
642     .P
643     No character that is in the Unicode table has the Cn (unassigned) property.
644     Instead, this property is assumed for any code point that is not in the
645     Unicode table.
646     .P
647 nigel 75 Specifying caseless matching does not affect these escape sequences. For
648     example, \ep{Lu} always matches only upper case letters.
649     .P
650     The \eX escape matches any number of Unicode characters that form an extended
651     Unicode sequence. \eX is equivalent to
652     .sp
653     (?>\ePM\epM*)
654     .sp
655     That is, it matches a character without the "mark" property, followed by zero
656     or more characters with the "mark" property, and treats the sequence as an
657     atomic group
658     .\" HTML <a href="#atomicgroup">
659     .\" </a>
660     (see below).
661     .\"
662     Characters with the "mark" property are typically accents that affect the
663 ph10 185 preceding character. None of them have codepoints less than 256, so in
664 ph10 184 non-UTF-8 mode \eX matches any one character.
665 nigel 75 .P
666     Matching characters by Unicode property is not fast, because PCRE has to search
667     a structure that contains data for over fifteen thousand characters. That is
668     why the traditional escape sequences such as \ed and \ew do not use Unicode
669     properties in PCRE.
670     .
671     .
672 ph10 168 .\" HTML <a name="resetmatchstart"></a>
673     .SS "Resetting the match start"
674     .rs
675     .sp
676 ph10 172 The escape sequence \eK, which is a Perl 5.10 feature, causes any previously
677     matched characters not to be included in the final matched sequence. For
678 ph10 168 example, the pattern:
679     .sp
680     foo\eKbar
681     .sp
682 ph10 172 matches "foobar", but reports that it has matched "bar". This feature is
683 ph10 168 similar to a lookbehind assertion
684     .\" HTML <a href="#lookbehind">
685     .\" </a>
686     (described below).
687     .\"
688 ph10 172 However, in this case, the part of the subject before the real match does not
689     have to be of fixed length, as lookbehind assertions do. The use of \eK does
690 ph10 168 not interfere with the setting of
691     .\" HTML <a href="#subpattern">
692     .\" </a>
693     captured substrings.
694 ph10 172 .\"
695 ph10 168 For example, when the pattern
696     .sp
697     (foo)\eKbar
698     .sp
699 ph10 172 matches "foobar", the first substring is still set to "foo".
700 ph10 168 .
701     .
702 nigel 75 .\" HTML <a name="smallassertions"></a>
703     .SS "Simple assertions"
704     .rs
705     .sp
706 nigel 93 The final use of backslash is for certain simple assertions. An assertion
707 nigel 63 specifies a condition that has to be met at a particular point in a match,
708     without consuming any characters from the subject string. The use of
709 nigel 75 subpatterns for more complicated assertions is described
710     .\" HTML <a href="#bigassertions">
711     .\" </a>
712     below.
713     .\"
714 nigel 91 The backslashed assertions are:
715 nigel 75 .sp
716     \eb matches at a word boundary
717     \eB matches when not at a word boundary
718 nigel 93 \eA matches at the start of the subject
719     \eZ matches at the end of the subject
720     also matches before a newline at the end of the subject
721     \ez matches only at the end of the subject
722     \eG matches at the first matching position in the subject
723 nigel 75 .sp
724     These assertions may not appear in character classes (but note that \eb has a
725 nigel 63 different meaning, namely the backspace character, inside a character class).
726 nigel 75 .P
727 nigel 63 A word boundary is a position in the subject string where the current character
728 nigel 75 and the previous character do not both match \ew or \eW (i.e. one matches
729     \ew and the other matches \eW), or the start or end of the string if the
730     first or last character matches \ew, respectively.
731     .P
732     The \eA, \eZ, and \ez assertions differ from the traditional circumflex and
733     dollar (described in the next section) in that they only ever match at the very
734     start and end of the subject string, whatever options are set. Thus, they are
735     independent of multiline mode. These three assertions are not affected by the
736     PCRE_NOTBOL or PCRE_NOTEOL options, which affect only the behaviour of the
737     circumflex and dollar metacharacters. However, if the \fIstartoffset\fP
738     argument of \fBpcre_exec()\fP is non-zero, indicating that matching is to start
739     at a point other than the beginning of the subject, \eA can never match. The
740 nigel 91 difference between \eZ and \ez is that \eZ matches before a newline at the end
741     of the string as well as at the very end, whereas \ez matches only at the end.
742 nigel 75 .P
743     The \eG assertion is true only when the current matching position is at the
744     start point of the match, as specified by the \fIstartoffset\fP argument of
745     \fBpcre_exec()\fP. It differs from \eA when the value of \fIstartoffset\fP is
746     non-zero. By calling \fBpcre_exec()\fP multiple times with appropriate
747 nigel 63 arguments, you can mimic Perl's /g option, and it is in this kind of
748 nigel 75 implementation where \eG can be useful.
749     .P
750     Note, however, that PCRE's interpretation of \eG, as the start of the current
751 nigel 63 match, is subtly different from Perl's, which defines it as the end of the
752     previous match. In Perl, these can be different when the previously matched
753     string was empty. Because PCRE does just one match at a time, it cannot
754     reproduce this behaviour.
755 nigel 75 .P
756     If all the alternatives of a pattern begin with \eG, the expression is anchored
757 nigel 63 to the starting match position, and the "anchored" flag is set in the compiled
758     regular expression.
759 nigel 75 .
760     .
762 nigel 63 .rs
763     .sp
764     Outside a character class, in the default matching mode, the circumflex
765 nigel 75 character is an assertion that is true only if the current matching point is
766     at the start of the subject string. If the \fIstartoffset\fP argument of
767     \fBpcre_exec()\fP is non-zero, circumflex can never match if the PCRE_MULTILINE
768 nigel 63 option is unset. Inside a character class, circumflex has an entirely different
769 nigel 75 meaning
770     .\" HTML <a href="#characterclass">
771     .\" </a>
772     (see below).
773     .\"
774     .P
775 nigel 63 Circumflex need not be the first character of the pattern if a number of
776     alternatives are involved, but it should be the first thing in each alternative
777     in which it appears if the pattern is ever to match that branch. If all
778     possible alternatives start with a circumflex, that is, if the pattern is
779     constrained to match only at the start of the subject, it is said to be an
780     "anchored" pattern. (There are also other constructs that can cause a pattern
781     to be anchored.)
782 nigel 75 .P
783     A dollar character is an assertion that is true only if the current matching
784 nigel 63 point is at the end of the subject string, or immediately before a newline
785 nigel 91 at the end of the string (by default). Dollar need not be the last character of
786     the pattern if a number of alternatives are involved, but it should be the last
787     item in any branch in which it appears. Dollar has no special meaning in a
788     character class.
789 nigel 75 .P
790 nigel 63 The meaning of dollar can be changed so that it matches only at the very end of
791     the string, by setting the PCRE_DOLLAR_ENDONLY option at compile time. This
792 nigel 75 does not affect the \eZ assertion.
793     .P
794 nigel 63 The meanings of the circumflex and dollar characters are changed if the
795 nigel 91 PCRE_MULTILINE option is set. When this is the case, a circumflex matches
796     immediately after internal newlines as well as at the start of the subject
797     string. It does not match after a newline that ends the string. A dollar
798     matches before any newlines in the string, as well as at the very end, when
799     PCRE_MULTILINE is set. When newline is specified as the two-character
800     sequence CRLF, isolated CR and LF characters do not indicate newlines.
801 nigel 75 .P
802 nigel 91 For example, the pattern /^abc$/ matches the subject string "def\enabc" (where
803     \en represents a newline) in multiline mode, but not otherwise. Consequently,
804     patterns that are anchored in single line mode because all branches start with
805     ^ are not anchored in multiline mode, and a match for circumflex is possible
806     when the \fIstartoffset\fP argument of \fBpcre_exec()\fP is non-zero. The
807     PCRE_DOLLAR_ENDONLY option is ignored if PCRE_MULTILINE is set.
808     .P
809 nigel 75 Note that the sequences \eA, \eZ, and \ez can be used to match the start and
810 nigel 63 end of the subject in both modes, and if all branches of a pattern start with
811 nigel 91 \eA it is always anchored, whether or not PCRE_MULTILINE is set.
812 nigel 75 .
813     .
815 nigel 63 .rs
816     .sp
817     Outside a character class, a dot in the pattern matches any one character in
818 nigel 91 the subject string except (by default) a character that signifies the end of a
819 nigel 93 line. In UTF-8 mode, the matched character may be more than one byte long.
820 nigel 91 .P
821 nigel 93 When a line ending is defined as a single character, dot never matches that
822     character; when the two-character sequence CRLF is used, dot does not match CR
823     if it is immediately followed by LF, but otherwise it matches all characters
824     (including isolated CRs and LFs). When any Unicode line endings are being
825     recognized, dot does not match CR or LF or any of the other line ending
826     characters.
827     .P
828 nigel 91 The behaviour of dot with regard to newlines can be changed. If the PCRE_DOTALL
829 nigel 93 option is set, a dot matches any one character, without exception. If the
830     two-character sequence CRLF is present in the subject string, it takes two dots
831     to match it.
832 nigel 91 .P
833     The handling of dot is entirely independent of the handling of circumflex and
834     dollar, the only relationship being that they both involve newlines. Dot has no
835     special meaning in a character class.
836 nigel 75 .
837     .
839 nigel 63 .rs
840     .sp
841 nigel 75 Outside a character class, the escape sequence \eC matches any one byte, both
842 nigel 93 in and out of UTF-8 mode. Unlike a dot, it always matches any line-ending
843     characters. The feature is provided in Perl in order to match individual bytes
844     in UTF-8 mode. Because it breaks up UTF-8 characters into individual bytes,
845     what remains in the string may be a malformed UTF-8 string. For this reason,
846     the \eC escape sequence is best avoided.
847 nigel 75 .P
848     PCRE does not allow \eC to appear in lookbehind assertions
849     .\" HTML <a href="#lookbehind">
850     .\" </a>
851     (described below),
852     .\"
853     because in UTF-8 mode this would make it impossible to calculate the length of
854     the lookbehind.
855     .
856     .
857     .\" HTML <a name="characterclass"></a>
859 nigel 63 .rs
860     .sp
861     An opening square bracket introduces a character class, terminated by a closing
862     square bracket. A closing square bracket on its own is not special. If a
863     closing square bracket is required as a member of the class, it should be the
864     first data character in the class (after an initial circumflex, if present) or
865     escaped with a backslash.
866 nigel 75 .P
867 nigel 63 A character class matches a single character in the subject. In UTF-8 mode, the
868     character may occupy more than one byte. A matched character must be in the set
869     of characters defined by the class, unless the first character in the class
870     definition is a circumflex, in which case the subject character must not be in
871     the set defined by the class. If a circumflex is actually required as a member
872     of the class, ensure it is not the first character, or escape it with a
873     backslash.
874 nigel 75 .P
875 nigel 63 For example, the character class [aeiou] matches any lower case vowel, while
876     [^aeiou] matches any character that is not a lower case vowel. Note that a
877 nigel 75 circumflex is just a convenient notation for specifying the characters that
878     are in the class by enumerating those that are not. A class that starts with a
879     circumflex is not an assertion: it still consumes a character from the subject
880     string, and therefore it fails if the current pointer is at the end of the
881     string.
882     .P
883 nigel 63 In UTF-8 mode, characters with values greater than 255 can be included in a
884 nigel 75 class as a literal string of bytes, or by using the \ex{ escaping mechanism.
885     .P
886 nigel 63 When caseless matching is set, any letters in a class represent both their
887     upper case and lower case versions, so for example, a caseless [aeiou] matches
888     "A" as well as "a", and a caseless [^aeiou] does not match "A", whereas a
889 nigel 77 caseful version would. In UTF-8 mode, PCRE always understands the concept of
890     case for characters whose values are less than 128, so caseless matching is
891     always possible. For characters with higher values, the concept of case is
892     supported if PCRE is compiled with Unicode property support, but not otherwise.
893     If you want to use caseless matching for characters 128 and above, you must
894     ensure that PCRE is compiled with Unicode property support as well as with
895     UTF-8 support.
896 nigel 75 .P
897 nigel 93 Characters that might indicate line breaks are never treated in any special way
898     when matching character classes, whatever line-ending sequence is in use, and
899     whatever setting of the PCRE_DOTALL and PCRE_MULTILINE options is used. A class
900     such as [^a] always matches one of these characters.
901 nigel 75 .P
902 nigel 63 The minus (hyphen) character can be used to specify a range of characters in a
903     character class. For example, [d-m] matches any letter between d and m,
904     inclusive. If a minus character is required in a class, it must be escaped with
905     a backslash or appear in a position where it cannot be interpreted as
906     indicating a range, typically as the first or last character in the class.
907 nigel 75 .P
908 nigel 63 It is not possible to have the literal character "]" as the end character of a
909     range. A pattern such as [W-]46] is interpreted as a class of two characters
910     ("W" and "-") followed by a literal string "46]", so it would match "W46]" or
911     "-46]". However, if the "]" is escaped with a backslash it is interpreted as
912 nigel 75 the end of range, so [W-\e]46] is interpreted as a class containing a range
913     followed by two other characters. The octal or hexadecimal representation of
914     "]" can also be used to end a range.
915     .P
916 nigel 63 Ranges operate in the collating sequence of character values. They can also be
917 nigel 75 used for characters specified numerically, for example [\e000-\e037]. In UTF-8
918 nigel 63 mode, ranges can include characters whose values are greater than 255, for
919 nigel 75 example [\ex{100}-\ex{2ff}].
920     .P
921 nigel 63 If a range that includes letters is used when caseless matching is set, it
922     matches the letters in either case. For example, [W-c] is equivalent to
923 nigel 75 [][\e\e^_`wxyzabc], matched caselessly, and in non-UTF-8 mode, if character
924 ph10 139 tables for a French locale are in use, [\exc8-\excb] matches accented E
925 nigel 75 characters in both cases. In UTF-8 mode, PCRE supports the concept of case for
926     characters with values greater than 128 only when it is compiled with Unicode
927     property support.
928     .P
929     The character types \ed, \eD, \ep, \eP, \es, \eS, \ew, and \eW may also appear
930     in a character class, and add the characters that they match to the class. For
931     example, [\edABCDEF] matches any hexadecimal digit. A circumflex can
932 nigel 63 conveniently be used with the upper case character types to specify a more
933     restricted set of characters than the matching lower case type. For example,
934 nigel 75 the class [^\eW_] matches any letter or digit, but not underscore.
935     .P
936     The only metacharacters that are recognized in character classes are backslash,
937     hyphen (only where it can be interpreted as specifying a range), circumflex
938     (only at the start), opening square bracket (only when it can be interpreted as
939     introducing a POSIX class name - see the next section), and the terminating
940     closing square bracket. However, escaping other non-alphanumeric characters
941     does no harm.
942     .
943     .
945 nigel 63 .rs
946     .sp
947 nigel 75 Perl supports the POSIX notation for character classes. This uses names
948 nigel 63 enclosed by [: and :] within the enclosing square brackets. PCRE also supports
949     this notation. For example,
950 nigel 75 .sp
951 nigel 63 [01[:alpha:]%]
952 nigel 75 .sp
953 nigel 63 matches "0", "1", any alphabetic character, or "%". The supported class names
954     are
955 nigel 75 .sp
956 nigel 63 alnum letters and digits
957     alpha letters
958     ascii character codes 0 - 127
959     blank space or tab only
960     cntrl control characters
961 nigel 75 digit decimal digits (same as \ed)
962 nigel 63 graph printing characters, excluding space
963     lower lower case letters
964     print printing characters, including space
965     punct printing characters, excluding letters and digits
966 nigel 75 space white space (not quite the same as \es)
967 nigel 63 upper upper case letters
968 nigel 75 word "word" characters (same as \ew)
969 nigel 63 xdigit hexadecimal digits
970 nigel 75 .sp
971 nigel 63 The "space" characters are HT (9), LF (10), VT (11), FF (12), CR (13), and
972     space (32). Notice that this list includes the VT character (code 11). This
973 nigel 75 makes "space" different to \es, which does not include VT (for Perl
974 nigel 63 compatibility).
975 nigel 75 .P
976 nigel 63 The name "word" is a Perl extension, and "blank" is a GNU extension from Perl
977     5.8. Another Perl extension is negation, which is indicated by a ^ character
978     after the colon. For example,
979 nigel 75 .sp
980 nigel 63 [12[:^digit:]]
981 nigel 75 .sp
982 nigel 63 matches "1", "2", or any non-digit. PCRE (and Perl) also recognize the POSIX
983     syntax [.ch.] and [=ch=] where "ch" is a "collating element", but these are not
984     supported, and an error is given if they are encountered.
985 nigel 75 .P
986     In UTF-8 mode, characters with values greater than 128 do not match any of
987 nigel 63 the POSIX character classes.
988 nigel 75 .
989     .
990     .SH "VERTICAL BAR"
991 nigel 63 .rs
992     .sp
993     Vertical bar characters are used to separate alternative patterns. For example,
994     the pattern
995 nigel 75 .sp
996 nigel 63 gilbert|sullivan
997 nigel 75 .sp
998 nigel 63 matches either "gilbert" or "sullivan". Any number of alternatives may appear,
999 nigel 91 and an empty alternative is permitted (matching the empty string). The matching
1000     process tries each alternative in turn, from left to right, and the first one
1001     that succeeds is used. If the alternatives are within a subpattern
1002 nigel 75 .\" HTML <a href="#subpattern">
1003     .\" </a>
1004     (defined below),
1005     .\"
1006     "succeeds" means matching the rest of the main pattern as well as the
1007     alternative in the subpattern.
1008     .
1009     .
1011 nigel 63 .rs
1012     .sp
1013     The settings of the PCRE_CASELESS, PCRE_MULTILINE, PCRE_DOTALL, and
1014 ph10 231 PCRE_EXTENDED options (which are Perl-compatible) can be changed from within
1015     the pattern by a sequence of Perl option letters enclosed between "(?" and ")".
1016     The option letters are
1017 nigel 75 .sp
1018 nigel 63 i for PCRE_CASELESS
1019     m for PCRE_MULTILINE
1020     s for PCRE_DOTALL
1021     x for PCRE_EXTENDED
1022 nigel 75 .sp
1023 nigel 63 For example, (?im) sets caseless, multiline matching. It is also possible to
1024     unset these options by preceding the letter with a hyphen, and a combined
1025     setting and unsetting such as (?im-sx), which sets PCRE_CASELESS and
1026     PCRE_MULTILINE while unsetting PCRE_DOTALL and PCRE_EXTENDED, is also
1027     permitted. If a letter appears both before and after the hyphen, the option is
1028     unset.
1029 nigel 75 .P
1030 ph10 231 The PCRE-specific options PCRE_DUPNAMES, PCRE_UNGREEDY, and PCRE_EXTRA can be
1031     changed in the same way as the Perl-compatible options by using the characters
1032     J, U and X respectively.
1033     .P
1034 nigel 63 When an option change occurs at top level (that is, not inside subpattern
1035     parentheses), the change applies to the remainder of the pattern that follows.
1036     If the change is placed right at the start of a pattern, PCRE extracts it into
1037     the global options (and it will therefore show up in data extracted by the
1038 nigel 75 \fBpcre_fullinfo()\fP function).
1039     .P
1040 nigel 93 An option change within a subpattern (see below for a description of
1041     subpatterns) affects only that part of the current pattern that follows it, so
1042 nigel 75 .sp
1043 nigel 63 (a(?i)b)c
1044 nigel 75 .sp
1045 nigel 63 matches abc and aBc and no other strings (assuming PCRE_CASELESS is not used).
1046     By this means, options can be made to have different settings in different
1047     parts of the pattern. Any changes made in one alternative do carry on
1048     into subsequent branches within the same subpattern. For example,
1049 nigel 75 .sp
1050 nigel 63 (a(?i)b|c)
1051 nigel 75 .sp
1052 nigel 63 matches "ab", "aB", "c", and "C", even though when matching "C" the first
1053     branch is abandoned before the option setting. This is because the effects of
1054     option settings happen at compile time. There would be some very weird
1055     behaviour otherwise.
1056 ph10 251 .P
1057     \fBNote:\fP There are other PCRE-specific options that can be set by the
1058     application when the compile or match functions are called. In some cases the
1059     pattern can contain special leading sequences to override what the application
1060     has set or what has been defaulted. Details are given in the section entitled
1061     .\" HTML <a href="#newlineseq">
1062     .\" </a>
1063     "Newline sequences"
1064     .\"
1065 ph10 259 above.
1066 nigel 75 .
1067     .
1068     .\" HTML <a name="subpattern"></a>
1069 nigel 63 .SH SUBPATTERNS
1070     .rs
1071     .sp
1072     Subpatterns are delimited by parentheses (round brackets), which can be nested.
1073 nigel 75 Turning part of a pattern into a subpattern does two things:
1074     .sp
1075 nigel 63 1. It localizes a set of alternatives. For example, the pattern
1076 nigel 75 .sp
1077 nigel 63 cat(aract|erpillar|)
1078 nigel 75 .sp
1079 nigel 63 matches one of the words "cat", "cataract", or "caterpillar". Without the
1080 nigel 93 parentheses, it would match "cataract", "erpillar" or an empty string.
1081 nigel 75 .sp
1082     2. It sets up the subpattern as a capturing subpattern. This means that, when
1083     the whole pattern matches, that portion of the subject string that matched the
1084     subpattern is passed back to the caller via the \fIovector\fP argument of
1085     \fBpcre_exec()\fP. Opening parentheses are counted from left to right (starting
1086     from 1) to obtain numbers for the capturing subpatterns.
1087     .P
1088 nigel 63 For example, if the string "the red king" is matched against the pattern
1089 nigel 75 .sp
1090 nigel 63 the ((red|white) (king|queen))
1091 nigel 75 .sp
1092 nigel 63 the captured substrings are "red king", "red", and "king", and are numbered 1,
1093     2, and 3, respectively.
1094 nigel 75 .P
1095 nigel 63 The fact that plain parentheses fulfil two functions is not always helpful.
1096     There are often times when a grouping subpattern is required without a
1097     capturing requirement. If an opening parenthesis is followed by a question mark
1098     and a colon, the subpattern does not do any capturing, and is not counted when
1099     computing the number of any subsequent capturing subpatterns. For example, if
1100     the string "the white queen" is matched against the pattern
1101 nigel 75 .sp
1102 nigel 63 the ((?:red|white) (king|queen))
1103 nigel 75 .sp
1104 nigel 63 the captured substrings are "white queen" and "queen", and are numbered 1 and
1105 nigel 93 2. The maximum number of capturing subpatterns is 65535.
1106 nigel 75 .P
1107 nigel 63 As a convenient shorthand, if any option settings are required at the start of
1108     a non-capturing subpattern, the option letters may appear between the "?" and
1109     the ":". Thus the two patterns
1110 nigel 75 .sp
1111 nigel 63 (?i:saturday|sunday)
1112     (?:(?i)saturday|sunday)
1113 nigel 75 .sp
1114 nigel 63 match exactly the same set of strings. Because alternative branches are tried
1115     from left to right, and options are not reset until the end of the subpattern
1116     is reached, an option setting in one branch does affect subsequent branches, so
1117     the above patterns match "SUNDAY" as well as "Saturday".
1118 nigel 75 .
1119     .
1121     .rs
1122     .sp
1123 ph10 182 Perl 5.10 introduced a feature whereby each alternative in a subpattern uses
1124     the same numbers for its capturing parentheses. Such a subpattern starts with
1125     (?| and is itself a non-capturing subpattern. For example, consider this
1126 ph10 175 pattern:
1127     .sp
1128     (?|(Sat)ur|(Sun))day
1129 ph10 182 .sp
1130     Because the two alternatives are inside a (?| group, both sets of capturing
1131     parentheses are numbered one. Thus, when the pattern matches, you can look
1132     at captured substring number one, whichever alternative matched. This construct
1133     is useful when you want to capture part, but not all, of one of a number of
1134     alternatives. Inside a (?| group, parentheses are numbered as usual, but the
1135 ph10 175 number is reset at the start of each branch. The numbers of any capturing
1136 ph10 182 buffers that follow the subpattern start after the highest number used in any
1137     branch. The following example is taken from the Perl documentation.
1138 ph10 175 The numbers underneath show in which buffer the captured content will be
1139     stored.
1140     .sp
1141     # before ---------------branch-reset----------- after
1142     / ( a ) (?| x ( y ) z | (p (q) r) | (t) u (v) ) ( z ) /x
1143     # 1 2 2 3 2 3 4
1144 ph10 182 .sp
1145 ph10 175 A backreference or a recursive call to a numbered subpattern always refers to
1146     the first one in the pattern with the given number.
1147     .P
1148     An alternative approach to using this "branch reset" feature is to use
1149     duplicate named subpatterns, as described in the next section.
1150     .
1151     .
1152 nigel 75 .SH "NAMED SUBPATTERNS"
1153 nigel 63 .rs
1154     .sp
1155     Identifying capturing parentheses by number is simple, but it can be very hard
1156     to keep track of the numbers in complicated regular expressions. Furthermore,
1157 nigel 75 if an expression is modified, the numbers may change. To help with this
1158 nigel 93 difficulty, PCRE supports the naming of subpatterns. This feature was not
1159     added to Perl until release 5.10. Python had the feature earlier, and PCRE
1160     introduced it at release 4.0, using the Python syntax. PCRE now supports both
1161     the Perl and the Python syntax.
1162     .P
1163     In PCRE, a subpattern can be named in one of three ways: (?<name>...) or
1164     (?'name'...) as in Perl, or (?P<name>...) as in Python. References to capturing
1165 nigel 91 parentheses from other parts of the pattern, such as
1166     .\" HTML <a href="#backreferences">
1167     .\" </a>
1168     backreferences,
1169     .\"
1170     .\" HTML <a href="#recursion">
1171     .\" </a>
1172     recursion,
1173     .\"
1174     and
1175     .\" HTML <a href="#conditions">
1176     .\" </a>
1177     conditions,
1178     .\"
1179     can be made by name as well as by number.
1180 nigel 75 .P
1181 nigel 91 Names consist of up to 32 alphanumeric characters and underscores. Named
1182 nigel 93 capturing parentheses are still allocated numbers as well as names, exactly as
1183     if the names were not present. The PCRE API provides function calls for
1184     extracting the name-to-number translation table from a compiled pattern. There
1185     is also a convenience function for extracting a captured substring by name.
1186 nigel 91 .P
1187     By default, a name must be unique within a pattern, but it is possible to relax
1188     this constraint by setting the PCRE_DUPNAMES option at compile time. This can
1189     be useful for patterns where only one instance of the named parentheses can
1190     match. Suppose you want to match the name of a weekday, either as a 3-letter
1191     abbreviation or as the full name, and in both cases you want to extract the
1192     abbreviation. This pattern (ignoring the line breaks) does the job:
1193     .sp
1194 nigel 93 (?<DN>Mon|Fri|Sun)(?:day)?|
1195     (?<DN>Tue)(?:sday)?|
1196     (?<DN>Wed)(?:nesday)?|
1197     (?<DN>Thu)(?:rsday)?|
1198     (?<DN>Sat)(?:urday)?
1199 nigel 91 .sp
1200     There are five capturing substrings, but only one is ever set after a match.
1201 ph10 182 (An alternative way of solving this problem is to use a "branch reset"
1202 ph10 175 subpattern, as described in the previous section.)
1203     .P
1204 nigel 91 The convenience function for extracting the data by name returns the substring
1205 nigel 93 for the first (and in this example, the only) subpattern of that name that
1206 nigel 91 matched. This saves searching to find which numbered subpattern it was. If you
1207     make a reference to a non-unique named subpattern from elsewhere in the
1208     pattern, the one that corresponds to the lowest number is used. For further
1209     details of the interfaces for handling named subpatterns, see the
1210 nigel 63 .\" HREF
1211 nigel 75 \fBpcreapi\fP
1212 nigel 63 .\"
1213     documentation.
1214 nigel 75 .
1215     .
1216 nigel 63 .SH REPETITION
1217     .rs
1218     .sp
1219     Repetition is specified by quantifiers, which can follow any of the following
1220     items:
1221 nigel 75 .sp
1222 nigel 63 a literal data character
1223 nigel 93 the dot metacharacter
1224 nigel 75 the \eC escape sequence
1225     the \eX escape sequence (in UTF-8 mode with Unicode properties)
1226 nigel 93 the \eR escape sequence
1227 nigel 75 an escape such as \ed that matches a single character
1228 nigel 63 a character class
1229     a back reference (see next section)
1230     a parenthesized subpattern (unless it is an assertion)
1231 nigel 75 .sp
1232 nigel 63 The general repetition quantifier specifies a minimum and maximum number of
1233     permitted matches, by giving the two numbers in curly brackets (braces),
1234     separated by a comma. The numbers must be less than 65536, and the first must
1235     be less than or equal to the second. For example:
1236 nigel 75 .sp
1237 nigel 63 z{2,4}
1238 nigel 75 .sp
1239 nigel 63 matches "zz", "zzz", or "zzzz". A closing brace on its own is not a special
1240     character. If the second number is omitted, but the comma is present, there is
1241     no upper limit; if the second number and the comma are both omitted, the
1242     quantifier specifies an exact number of required matches. Thus
1243 nigel 75 .sp
1244 nigel 63 [aeiou]{3,}
1245 nigel 75 .sp
1246 nigel 63 matches at least 3 successive vowels, but may match many more, while
1247 nigel 75 .sp
1248     \ed{8}
1249     .sp
1250 nigel 63 matches exactly 8 digits. An opening curly bracket that appears in a position
1251     where a quantifier is not allowed, or one that does not match the syntax of a
1252     quantifier, is taken as a literal character. For example, {,6} is not a
1253     quantifier, but a literal string of four characters.
1254 nigel 75 .P
1255 nigel 63 In UTF-8 mode, quantifiers apply to UTF-8 characters rather than to individual
1256 nigel 75 bytes. Thus, for example, \ex{100}{2} matches two UTF-8 characters, each of
1257     which is represented by a two-byte sequence. Similarly, when Unicode property
1258     support is available, \eX{3} matches three Unicode extended sequences, each of
1259     which may be several bytes long (and they may be of different lengths).
1260     .P
1261 nigel 63 The quantifier {0} is permitted, causing the expression to behave as if the
1262 ph10 335 previous item and the quantifier were not present. This may be useful for
1263     subpatterns that are referenced as
1264     .\" HTML <a href="#subpatternsassubroutines">
1265     .\" </a>
1266     subroutines
1267     .\"
1268     from elsewhere in the pattern. Items other than subpatterns that have a {0}
1269     quantifier are omitted from the compiled pattern.
1270 nigel 75 .P
1271 nigel 93 For convenience, the three most common quantifiers have single-character
1272     abbreviations:
1273 nigel 75 .sp
1274 nigel 63 * is equivalent to {0,}
1275     + is equivalent to {1,}
1276     ? is equivalent to {0,1}
1277 nigel 75 .sp
1278 nigel 63 It is possible to construct infinite loops by following a subpattern that can
1279     match no characters with a quantifier that has no upper limit, for example:
1280 nigel 75 .sp
1281 nigel 63 (a?)*
1282 nigel 75 .sp
1283 nigel 63 Earlier versions of Perl and PCRE used to give an error at compile time for
1284     such patterns. However, because there are cases where this can be useful, such
1285     patterns are now accepted, but if any repetition of the subpattern does in fact
1286     match no characters, the loop is forcibly broken.
1287 nigel 75 .P
1288 nigel 63 By default, the quantifiers are "greedy", that is, they match as much as
1289     possible (up to the maximum number of permitted times), without causing the
1290     rest of the pattern to fail. The classic example of where this gives problems
1291 nigel 75 is in trying to match comments in C programs. These appear between /* and */
1292     and within the comment, individual * and / characters may appear. An attempt to
1293     match C comments by applying the pattern
1294     .sp
1295     /\e*.*\e*/
1296     .sp
1297 nigel 63 to the string
1298 nigel 75 .sp
1299     /* first comment */ not comment /* second comment */
1300     .sp
1301 nigel 63 fails, because it matches the entire string owing to the greediness of the .*
1302     item.
1303 nigel 75 .P
1304 nigel 63 However, if a quantifier is followed by a question mark, it ceases to be
1305     greedy, and instead matches the minimum number of times possible, so the
1306     pattern
1307 nigel 75 .sp
1308     /\e*.*?\e*/
1309     .sp
1310 nigel 63 does the right thing with the C comments. The meaning of the various
1311     quantifiers is not otherwise changed, just the preferred number of matches.
1312     Do not confuse this use of question mark with its use as a quantifier in its
1313     own right. Because it has two uses, it can sometimes appear doubled, as in
1314 nigel 75 .sp
1315     \ed??\ed
1316     .sp
1317 nigel 63 which matches one digit by preference, but can match two if that is the only
1318     way the rest of the pattern matches.
1319 nigel 75 .P
1320 nigel 93 If the PCRE_UNGREEDY option is set (an option that is not available in Perl),
1321 nigel 63 the quantifiers are not greedy by default, but individual ones can be made
1322     greedy by following them with a question mark. In other words, it inverts the
1323     default behaviour.
1324 nigel 75 .P
1325 nigel 63 When a parenthesized subpattern is quantified with a minimum repeat count that
1326 nigel 75 is greater than 1 or with a limited maximum, more memory is required for the
1327 nigel 63 compiled pattern, in proportion to the size of the minimum or maximum.
1328 nigel 75 .P
1329 nigel 63 If a pattern starts with .* or .{0,} and the PCRE_DOTALL option (equivalent
1330 nigel 93 to Perl's /s) is set, thus allowing the dot to match newlines, the pattern is
1331 nigel 63 implicitly anchored, because whatever follows will be tried against every
1332     character position in the subject string, so there is no point in retrying the
1333     overall match at any position after the first. PCRE normally treats such a
1334 nigel 75 pattern as though it were preceded by \eA.
1335     .P
1336 nigel 63 In cases where it is known that the subject string contains no newlines, it is
1337     worth setting PCRE_DOTALL in order to obtain this optimization, or
1338     alternatively using ^ to indicate anchoring explicitly.
1339 nigel 75 .P
1340 nigel 63 However, there is one situation where the optimization cannot be used. When .*
1341     is inside capturing parentheses that are the subject of a backreference
1342 nigel 93 elsewhere in the pattern, a match at the start may fail where a later one
1343     succeeds. Consider, for example:
1344 nigel 75 .sp
1345     (.*)abc\e1
1346     .sp
1347 nigel 63 If the subject is "xyz123abc123" the match point is the fourth character. For
1348     this reason, such a pattern is not implicitly anchored.
1349 nigel 75 .P
1350 nigel 63 When a capturing subpattern is repeated, the value captured is the substring
1351     that matched the final iteration. For example, after
1352 nigel 75 .sp
1353     (tweedle[dume]{3}\es*)+
1354     .sp
1355 nigel 63 has matched "tweedledum tweedledee" the value of the captured substring is
1356     "tweedledee". However, if there are nested capturing subpatterns, the
1357     corresponding captured values may have been set in previous iterations. For
1358     example, after
1359 nigel 75 .sp
1360 nigel 63 /(a|(b))+/
1361 nigel 75 .sp
1362 nigel 63 matches "aba" the value of the second captured substring is "b".
1363 nigel 75 .
1364     .
1365     .\" HTML <a name="atomicgroup"></a>
1367 nigel 63 .rs
1368     .sp
1369 nigel 93 With both maximizing ("greedy") and minimizing ("ungreedy" or "lazy")
1370     repetition, failure of what follows normally causes the repeated item to be
1371     re-evaluated to see if a different number of repeats allows the rest of the
1372     pattern to match. Sometimes it is useful to prevent this, either to change the
1373     nature of the match, or to cause it fail earlier than it otherwise might, when
1374     the author of the pattern knows there is no point in carrying on.
1375 nigel 75 .P
1376     Consider, for example, the pattern \ed+foo when applied to the subject line
1377     .sp
1378 nigel 63 123456bar
1379 nigel 75 .sp
1380 nigel 63 After matching all 6 digits and then failing to match "foo", the normal
1381 nigel 75 action of the matcher is to try again with only 5 digits matching the \ed+
1382 nigel 63 item, and then with 4, and so on, before ultimately failing. "Atomic grouping"
1383     (a term taken from Jeffrey Friedl's book) provides the means for specifying
1384     that once a subpattern has matched, it is not to be re-evaluated in this way.
1385 nigel 75 .P
1386 nigel 93 If we use atomic grouping for the previous example, the matcher gives up
1387 nigel 63 immediately on failing to match "foo" the first time. The notation is a kind of
1388     special parenthesis, starting with (?> as in this example:
1389 nigel 75 .sp
1390     (?>\ed+)foo
1391     .sp
1392 nigel 63 This kind of parenthesis "locks up" the part of the pattern it contains once
1393     it has matched, and a failure further into the pattern is prevented from
1394     backtracking into it. Backtracking past it to previous items, however, works as
1395     normal.
1396 nigel 75 .P
1397 nigel 63 An alternative description is that a subpattern of this type matches the string
1398     of characters that an identical standalone pattern would match, if anchored at
1399     the current point in the subject string.
1400 nigel 75 .P
1401 nigel 63 Atomic grouping subpatterns are not capturing subpatterns. Simple cases such as
1402     the above example can be thought of as a maximizing repeat that must swallow
1403 nigel 75 everything it can. So, while both \ed+ and \ed+? are prepared to adjust the
1404 nigel 63 number of digits they match in order to make the rest of the pattern match,
1405 nigel 75 (?>\ed+) can only match an entire sequence of digits.
1406     .P
1407 nigel 63 Atomic groups in general can of course contain arbitrarily complicated
1408     subpatterns, and can be nested. However, when the subpattern for an atomic
1409     group is just a single repeated item, as in the example above, a simpler
1410     notation, called a "possessive quantifier" can be used. This consists of an
1411     additional + character following a quantifier. Using this notation, the
1412     previous example can be rewritten as
1413 nigel 75 .sp
1414     \ed++foo
1415     .sp
1416 ph10 208 Note that a possessive quantifier can be used with an entire group, for
1417     example:
1418     .sp
1419     (abc|xyz){2,3}+
1420     .sp
1421 nigel 63 Possessive quantifiers are always greedy; the setting of the PCRE_UNGREEDY
1422     option is ignored. They are a convenient notation for the simpler forms of
1423 nigel 93 atomic group. However, there is no difference in the meaning of a possessive
1424     quantifier and the equivalent atomic group, though there may be a performance
1425     difference; possessive quantifiers should be slightly faster.
1426 nigel 75 .P
1427 nigel 93 The possessive quantifier syntax is an extension to the Perl 5.8 syntax.
1428     Jeffrey Friedl originated the idea (and the name) in the first edition of his
1429     book. Mike McCloskey liked it, so implemented it when he built Sun's Java
1430     package, and PCRE copied it from there. It ultimately found its way into Perl
1431     at release 5.10.
1432 nigel 75 .P
1433 nigel 93 PCRE has an optimization that automatically "possessifies" certain simple
1434     pattern constructs. For example, the sequence A+B is treated as A++B because
1435     there is no point in backtracking into a sequence of A's when B must follow.
1436     .P
1437 nigel 63 When a pattern contains an unlimited repeat inside a subpattern that can itself
1438     be repeated an unlimited number of times, the use of an atomic group is the
1439     only way to avoid some failing matches taking a very long time indeed. The
1440     pattern
1441 nigel 75 .sp
1442     (\eD+|<\ed+>)*[!?]
1443     .sp
1444 nigel 63 matches an unlimited number of substrings that either consist of non-digits, or
1445     digits enclosed in <>, followed by either ! or ?. When it matches, it runs
1446     quickly. However, if it is applied to
1447 nigel 75 .sp
1448 nigel 63 aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
1449 nigel 75 .sp
1450 nigel 63 it takes a long time before reporting failure. This is because the string can
1451 nigel 75 be divided between the internal \eD+ repeat and the external * repeat in a
1452     large number of ways, and all have to be tried. (The example uses [!?] rather
1453     than a single character at the end, because both PCRE and Perl have an
1454     optimization that allows for fast failure when a single character is used. They
1455     remember the last single character that is required for a match, and fail early
1456     if it is not present in the string.) If the pattern is changed so that it uses
1457     an atomic group, like this:
1458     .sp
1459     ((?>\eD+)|<\ed+>)*[!?]
1460     .sp
1461 nigel 63 sequences of non-digits cannot be broken, and failure happens quickly.
1462 nigel 75 .
1463     .
1464     .\" HTML <a name="backreferences"></a>
1466 nigel 63 .rs
1467     .sp
1468     Outside a character class, a backslash followed by a digit greater than 0 (and
1469     possibly further digits) is a back reference to a capturing subpattern earlier
1470     (that is, to its left) in the pattern, provided there have been that many
1471     previous capturing left parentheses.
1472 nigel 75 .P
1473 nigel 63 However, if the decimal number following the backslash is less than 10, it is
1474     always taken as a back reference, and causes an error only if there are not
1475     that many capturing left parentheses in the entire pattern. In other words, the
1476     parentheses that are referenced need not be to the left of the reference for
1477 nigel 91 numbers less than 10. A "forward back reference" of this type can make sense
1478     when a repetition is involved and the subpattern to the right has participated
1479     in an earlier iteration.
1480     .P
1481 nigel 93 It is not possible to have a numerical "forward back reference" to a subpattern
1482     whose number is 10 or more using this syntax because a sequence such as \e50 is
1483     interpreted as a character defined in octal. See the subsection entitled
1484 nigel 91 "Non-printing characters"
1485 nigel 75 .\" HTML <a href="#digitsafterbackslash">
1486     .\" </a>
1487     above
1488     .\"
1489 nigel 93 for further details of the handling of digits following a backslash. There is
1490     no such problem when named parentheses are used. A back reference to any
1491     subpattern is possible using named parentheses (see below).
1492 nigel 75 .P
1493 nigel 93 Another way of avoiding the ambiguity inherent in the use of digits following a
1494     backslash is to use the \eg escape sequence, which is a feature introduced in
1495 ph10 208 Perl 5.10. This escape must be followed by an unsigned number or a negative
1496     number, optionally enclosed in braces. These examples are all identical:
1497 nigel 93 .sp
1498     (ring), \e1
1499     (ring), \eg1
1500     (ring), \eg{1}
1501     .sp
1502 ph10 208 An unsigned number specifies an absolute reference without the ambiguity that
1503     is present in the older syntax. It is also useful when literal digits follow
1504     the reference. A negative number is a relative reference. Consider this
1505     example:
1506 nigel 93 .sp
1507     (abc(def)ghi)\eg{-1}
1508     .sp
1509     The sequence \eg{-1} is a reference to the most recently started capturing
1510     subpattern before \eg, that is, is it equivalent to \e2. Similarly, \eg{-2}
1511     would be equivalent to \e1. The use of relative references can be helpful in
1512     long patterns, and also in patterns that are created by joining together
1513     fragments that contain references within themselves.
1514     .P
1515 nigel 63 A back reference matches whatever actually matched the capturing subpattern in
1516     the current subject string, rather than anything matching the subpattern
1517     itself (see
1518     .\" HTML <a href="#subpatternsassubroutines">
1519     .\" </a>
1520     "Subpatterns as subroutines"
1521     .\"
1522     below for a way of doing that). So the pattern
1523 nigel 75 .sp
1524     (sens|respons)e and \e1ibility
1525     .sp
1526 nigel 63 matches "sense and sensibility" and "response and responsibility", but not
1527     "sense and responsibility". If caseful matching is in force at the time of the
1528     back reference, the case of letters is relevant. For example,
1529 nigel 75 .sp
1530     ((?i)rah)\es+\e1
1531     .sp
1532 nigel 63 matches "rah rah" and "RAH RAH", but not "RAH rah", even though the original
1533     capturing subpattern is matched caselessly.
1534 nigel 75 .P
1535 ph10 171 There are several different ways of writing back references to named
1536     subpatterns. The .NET syntax \ek{name} and the Perl syntax \ek<name> or
1537     \ek'name' are supported, as is the Python syntax (?P=name). Perl 5.10's unified
1538     back reference syntax, in which \eg can be used for both numeric and named
1539     references, is also supported. We could rewrite the above example in any of
1540 nigel 93 the following ways:
1541 nigel 75 .sp
1542 nigel 93 (?<p1>(?i)rah)\es+\ek<p1>
1543 ph10 171 (?'p1'(?i)rah)\es+\ek{p1}
1544 nigel 91 (?P<p1>(?i)rah)\es+(?P=p1)
1545 ph10 171 (?<p1>(?i)rah)\es+\eg{p1}
1546 nigel 75 .sp
1547 nigel 91 A subpattern that is referenced by name may appear in the pattern before or
1548     after the reference.
1549     .P
1550 nigel 63 There may be more than one back reference to the same subpattern. If a
1551     subpattern has not actually been used in a particular match, any back
1552     references to it always fail. For example, the pattern
1553 nigel 75 .sp
1554     (a|(bc))\e2
1555     .sp
1556 nigel 63 always fails if it starts to match "a" rather than "bc". Because there may be
1557     many capturing parentheses in a pattern, all digits following the backslash are
1558     taken as part of a potential back reference number. If the pattern continues
1559     with a digit character, some delimiter must be used to terminate the back
1560     reference. If the PCRE_EXTENDED option is set, this can be whitespace.
1561 nigel 75 Otherwise an empty comment (see
1562     .\" HTML <a href="#comments">
1563     .\" </a>
1564     "Comments"
1565     .\"
1566     below) can be used.
1567     .P
1568 nigel 63 A back reference that occurs inside the parentheses to which it refers fails
1569 nigel 75 when the subpattern is first used, so, for example, (a\e1) never matches.
1570 nigel 63 However, such references can be useful inside repeated subpatterns. For
1571     example, the pattern
1572 nigel 75 .sp
1573     (a|b\e1)+
1574     .sp
1575 nigel 63 matches any number of "a"s and also "aba", "ababbaa" etc. At each iteration of
1576     the subpattern, the back reference matches the character string corresponding
1577     to the previous iteration. In order for this to work, the pattern must be such
1578     that the first iteration does not need to match the back reference. This can be
1579     done using alternation, as in the example above, or by a quantifier with a
1580     minimum of zero.
1581 nigel 75 .
1582     .
1583     .\" HTML <a name="bigassertions"></a>
1584 nigel 63 .SH ASSERTIONS
1585     .rs
1586     .sp
1587     An assertion is a test on the characters following or preceding the current
1588     matching point that does not actually consume any characters. The simple
1589 nigel 75 assertions coded as \eb, \eB, \eA, \eG, \eZ, \ez, ^ and $ are described
1590     .\" HTML <a href="#smallassertions">
1591     .\" </a>
1592     above.
1593     .\"
1594     .P
1595 nigel 63 More complicated assertions are coded as subpatterns. There are two kinds:
1596     those that look ahead of the current position in the subject string, and those
1597 nigel 75 that look behind it. An assertion subpattern is matched in the normal way,
1598     except that it does not cause the current matching position to be changed.
1599     .P
1600     Assertion subpatterns are not capturing subpatterns, and may not be repeated,
1601     because it makes no sense to assert the same thing several times. If any kind
1602     of assertion contains capturing subpatterns within it, these are counted for
1603     the purposes of numbering the capturing subpatterns in the whole pattern.
1604     However, substring capturing is carried out only for positive assertions,
1605     because it does not make sense for negative assertions.
1606     .
1607     .
1608     .SS "Lookahead assertions"
1609     .rs
1610     .sp
1611 nigel 91 Lookahead assertions start with (?= for positive assertions and (?! for
1612     negative assertions. For example,
1613 nigel 75 .sp
1614     \ew+(?=;)
1615     .sp
1616 nigel 63 matches a word followed by a semicolon, but does not include the semicolon in
1617     the match, and
1618 nigel 75 .sp
1619 nigel 63 foo(?!bar)
1620 nigel 75 .sp
1621 nigel 63 matches any occurrence of "foo" that is not followed by "bar". Note that the
1622     apparently similar pattern
1623 nigel 75 .sp
1624 nigel 63 (?!foo)bar
1625 nigel 75 .sp
1626 nigel 63 does not find an occurrence of "bar" that is preceded by something other than
1627     "foo"; it finds any occurrence of "bar" whatsoever, because the assertion
1628     (?!foo) is always true when the next three characters are "bar". A
1629 nigel 75 lookbehind assertion is needed to achieve the other effect.
1630     .P
1631 nigel 63 If you want to force a matching failure at some point in a pattern, the most
1632     convenient way to do it is with (?!) because an empty string always matches, so
1633     an assertion that requires there not to be an empty string must always fail.
1634 nigel 75 .
1635     .
1636     .\" HTML <a name="lookbehind"></a>
1637     .SS "Lookbehind assertions"
1638     .rs
1639     .sp
1640 nigel 63 Lookbehind assertions start with (?<= for positive assertions and (?<! for
1641     negative assertions. For example,
1642 nigel 75 .sp
1643 nigel 63 (?<!foo)bar
1644 nigel 75 .sp
1645 nigel 63 does find an occurrence of "bar" that is not preceded by "foo". The contents of
1646     a lookbehind assertion are restricted such that all the strings it matches must
1647 nigel 91 have a fixed length. However, if there are several top-level alternatives, they
1648     do not all have to have the same fixed length. Thus
1649 nigel 75 .sp
1650 nigel 63 (?<=bullock|donkey)
1651 nigel 75 .sp
1652 nigel 63 is permitted, but
1653 nigel 75 .sp
1654 nigel 63 (?<!dogs?|cats?)
1655 nigel 75 .sp
1656 nigel 63 causes an error at compile time. Branches that match different length strings
1657     are permitted only at the top level of a lookbehind assertion. This is an
1658     extension compared with Perl (at least for 5.8), which requires all branches to
1659     match the same length of string. An assertion such as
1660 nigel 75 .sp
1661 nigel 63 (?<=ab(c|de))
1662 nigel 75 .sp
1663 nigel 63 is not permitted, because its single top-level branch can match two different
1664     lengths, but it is acceptable if rewritten to use two top-level branches:
1665 nigel 75 .sp
1666 nigel 63 (?<=abc|abde)
1667 nigel 75 .sp
1668 ph10 172 In some cases, the Perl 5.10 escape sequence \eK
1669 ph10 168 .\" HTML <a href="#resetmatchstart">
1670     .\" </a>
1671     (see above)
1672     .\"
1673     can be used instead of a lookbehind assertion; this is not restricted to a
1674     fixed-length.
1675     .P
1676 nigel 63 The implementation of lookbehind assertions is, for each alternative, to
1677 nigel 93 temporarily move the current position back by the fixed length and then try to
1678 nigel 63 match. If there are insufficient characters before the current position, the
1679 nigel 93 assertion fails.
1680 nigel 75 .P
1681     PCRE does not allow the \eC escape (which matches a single byte in UTF-8 mode)
1682 nigel 63 to appear in lookbehind assertions, because it makes it impossible to calculate
1683 nigel 93 the length of the lookbehind. The \eX and \eR escapes, which can match
1684     different numbers of bytes, are also not permitted.
1685 nigel 75 .P
1686 nigel 93 Possessive quantifiers can be used in conjunction with lookbehind assertions to
1687     specify efficient matching at the end of the subject string. Consider a simple
1688     pattern such as
1689 nigel 75 .sp
1690 nigel 63 abcd$
1691 nigel 75 .sp
1692 nigel 63 when applied to a long string that does not match. Because matching proceeds
1693     from left to right, PCRE will look for each "a" in the subject and then see if
1694     what follows matches the rest of the pattern. If the pattern is specified as
1695 nigel 75 .sp
1696 nigel 63 ^.*abcd$
1697 nigel 75 .sp
1698 nigel 63 the initial .* matches the entire string at first, but when this fails (because
1699     there is no following "a"), it backtracks to match all but the last character,
1700     then all but the last two characters, and so on. Once again the search for "a"
1701     covers the entire string, from right to left, so we are no better off. However,
1702     if the pattern is written as
1703 nigel 75 .sp
1704 nigel 63 ^.*+(?<=abcd)
1705 nigel 75 .sp
1706 nigel 93 there can be no backtracking for the .*+ item; it can match only the entire
1707 nigel 63 string. The subsequent lookbehind assertion does a single test on the last four
1708     characters. If it fails, the match fails immediately. For long strings, this
1709     approach makes a significant difference to the processing time.
1710 nigel 75 .
1711     .
1712     .SS "Using multiple assertions"
1713     .rs
1714     .sp
1715 nigel 63 Several assertions (of any sort) may occur in succession. For example,
1716 nigel 75 .sp
1717     (?<=\ed{3})(?<!999)foo
1718     .sp
1719 nigel 63 matches "foo" preceded by three digits that are not "999". Notice that each of
1720     the assertions is applied independently at the same point in the subject
1721     string. First there is a check that the previous three characters are all
1722     digits, and then there is a check that the same three characters are not "999".
1723 nigel 75 This pattern does \fInot\fP match "foo" preceded by six characters, the first
1724 nigel 63 of which are digits and the last three of which are not "999". For example, it
1725     doesn't match "123abcfoo". A pattern to do that is
1726 nigel 75 .sp
1727     (?<=\ed{3}...)(?<!999)foo
1728     .sp
1729 nigel 63 This time the first assertion looks at the preceding six characters, checking
1730     that the first three are digits, and then the second assertion checks that the
1731     preceding three characters are not "999".
1732 nigel 75 .P
1733 nigel 63 Assertions can be nested in any combination. For example,
1734 nigel 75 .sp
1735 nigel 63 (?<=(?<!foo)bar)baz
1736 nigel 75 .sp
1737 nigel 63 matches an occurrence of "baz" that is preceded by "bar" which in turn is not
1738     preceded by "foo", while
1739 nigel 75 .sp
1740     (?<=\ed{3}(?!999)...)foo
1741     .sp
1742     is another pattern that matches "foo" preceded by three digits and any three
1743 nigel 63 characters that are not "999".
1744 nigel 75 .
1745     .
1746 nigel 91 .\" HTML <a name="conditions"></a>
1748 nigel 63 .rs
1749     .sp
1750     It is possible to cause the matching process to obey a subpattern
1751     conditionally or to choose between two alternative subpatterns, depending on
1752     the result of an assertion, or whether a previous capturing subpattern matched
1753     or not. The two possible forms of conditional subpattern are
1754 nigel 75 .sp
1755 nigel 63 (?(condition)yes-pattern)
1756     (?(condition)yes-pattern|no-pattern)
1757 nigel 75 .sp
1758 nigel 63 If the condition is satisfied, the yes-pattern is used; otherwise the
1759     no-pattern (if present) is used. If there are more than two alternatives in the
1760     subpattern, a compile-time error occurs.
1761 nigel 75 .P
1762 nigel 93 There are four kinds of condition: references to subpatterns, references to
1763     recursion, a pseudo-condition called DEFINE, and assertions.
1764     .
1765     .SS "Checking for a used subpattern by number"
1766     .rs
1767     .sp
1768     If the text between the parentheses consists of a sequence of digits, the
1769     condition is true if the capturing subpattern of that number has previously
1770 ph10 172 matched. An alternative notation is to precede the digits with a plus or minus
1771 ph10 167 sign. In this case, the subpattern number is relative rather than absolute.
1772 ph10 172 The most recently opened parentheses can be referenced by (?(-1), the next most
1773 ph10 167 recent by (?(-2), and so on. In looping constructs it can also make sense to
1774     refer to subsequent groups with constructs such as (?(+2).
1775 nigel 91 .P
1776     Consider the following pattern, which contains non-significant white space to
1777     make it more readable (assume the PCRE_EXTENDED option) and to divide it into
1778     three parts for ease of discussion:
1779 nigel 75 .sp
1780     ( \e( )? [^()]+ (?(1) \e) )
1781     .sp
1782 nigel 63 The first part matches an optional opening parenthesis, and if that
1783     character is present, sets it as the first captured substring. The second part
1784     matches one or more characters that are not parentheses. The third part is a
1785     conditional subpattern that tests whether the first set of parentheses matched
1786     or not. If they did, that is, if subject started with an opening parenthesis,
1787     the condition is true, and so the yes-pattern is executed and a closing
1788     parenthesis is required. Otherwise, since no-pattern is not present, the
1789     subpattern matches nothing. In other words, this pattern matches a sequence of
1790 nigel 93 non-parentheses, optionally enclosed in parentheses.
1791 ph10 167 .P
1792 ph10 172 If you were embedding this pattern in a larger one, you could use a relative
1793 ph10 167 reference:
1794     .sp
1795     ...other stuff... ( \e( )? [^()]+ (?(-1) \e) ) ...
1796     .sp
1797     This makes the fragment independent of the parentheses in the larger pattern.
1798 nigel 93 .
1799     .SS "Checking for a used subpattern by name"
1800     .rs
1801 nigel 91 .sp
1802 nigel 93 Perl uses the syntax (?(<name>)...) or (?('name')...) to test for a used
1803     subpattern by name. For compatibility with earlier versions of PCRE, which had
1804     this facility before Perl, the syntax (?(name)...) is also recognized. However,
1805     there is a possible ambiguity with this syntax, because subpattern names may
1806     consist entirely of digits. PCRE looks first for a named subpattern; if it
1807     cannot find one and the name consists entirely of digits, PCRE looks for a
1808     subpattern of that number, which must be greater than zero. Using subpattern
1809     names that consist entirely of digits is not recommended.
1810     .P
1811     Rewriting the above example to use a named subpattern gives this:
1812 nigel 91 .sp
1813 nigel 93 (?<OPEN> \e( )? [^()]+ (?(<OPEN>) \e) )
1814     .sp
1815     .
1816     .SS "Checking for pattern recursion"
1817     .rs
1818     .sp
1819 nigel 91 If the condition is the string (R), and there is no subpattern with the name R,
1820 nigel 93 the condition is true if a recursive call to the whole pattern or any
1821     subpattern has been made. If digits or a name preceded by ampersand follow the
1822     letter R, for example:
1823     .sp
1824     (?(R3)...) or (?(R&name)...)
1825     .sp
1826     the condition is true if the most recent recursion is into the subpattern whose
1827     number or name is given. This condition does not check the entire recursion
1828     stack.
1829 nigel 75 .P
1830 nigel 93 At "top level", all these recursion test conditions are false. Recursive
1831     patterns are described below.
1832     .
1833     .SS "Defining subpatterns for use by reference only"
1834     .rs
1835     .sp
1836     If the condition is the string (DEFINE), and there is no subpattern with the
1837     name DEFINE, the condition is always false. In this case, there may be only one
1838     alternative in the subpattern. It is always skipped if control reaches this
1839     point in the pattern; the idea of DEFINE is that it can be used to define
1840     "subroutines" that can be referenced from elsewhere. (The use of "subroutines"
1841     is described below.) For example, a pattern to match an IPv4 address could be
1842     written like this (ignore whitespace and line breaks):
1843     .sp
1844     (?(DEFINE) (?<byte> 2[0-4]\ed | 25[0-5] | 1\ed\ed | [1-9]?\ed) )
1845     \eb (?&byte) (\e.(?&byte)){3} \eb
1846     .sp
1847     The first part of the pattern is a DEFINE group inside which a another group
1848     named "byte" is defined. This matches an individual component of an IPv4
1849     address (a number less than 256). When matching takes place, this part of the
1850     pattern is skipped because DEFINE acts like a false condition.
1851     .P
1852     The rest of the pattern uses references to the named group to match the four
1853     dot-separated components of an IPv4 address, insisting on a word boundary at
1854     each end.
1855     .
1856     .SS "Assertion conditions"
1857     .rs
1858     .sp
1859     If the condition is not in any of the above formats, it must be an assertion.
1860 nigel 63 This may be a positive or negative lookahead or lookbehind assertion. Consider
1861     this pattern, again containing non-significant white space, and with the two
1862     alternatives on the second line:
1863 nigel 75 .sp
1864 nigel 63 (?(?=[^a-z]*[a-z])
1865 nigel 75 \ed{2}-[a-z]{3}-\ed{2} | \ed{2}-\ed{2}-\ed{2} )
1866     .sp
1867 nigel 63 The condition is a positive lookahead assertion that matches an optional
1868     sequence of non-letters followed by a letter. In other words, it tests for the
1869     presence of at least one letter in the subject. If a letter is found, the
1870     subject is matched against the first alternative; otherwise it is matched
1871     against the second. This pattern matches strings in one of the two forms
1872     dd-aaa-dd or dd-dd-dd, where aaa are letters and dd are digits.
1873 nigel 75 .
1874     .
1875     .\" HTML <a name="comments"></a>
1876 nigel 63 .SH COMMENTS
1877     .rs
1878     .sp
1879 nigel 75 The sequence (?# marks the start of a comment that continues up to the next
1880 nigel 63 closing parenthesis. Nested parentheses are not permitted. The characters
1881     that make up a comment play no part in the pattern matching at all.
1882 nigel 75 .P
1883 nigel 63 If the PCRE_EXTENDED option is set, an unescaped # character outside a
1884 nigel 91 character class introduces a comment that continues to immediately after the
1885     next newline in the pattern.
1886 nigel 75 .
1887     .
1888 nigel 91 .\" HTML <a name="recursion"></a>
1889 nigel 75 .SH "RECURSIVE PATTERNS"
1890 nigel 63 .rs
1891     .sp
1892     Consider the problem of matching a string in parentheses, allowing for
1893     unlimited nested parentheses. Without the use of recursion, the best that can
1894     be done is to use a pattern that matches up to some fixed depth of nesting. It
1895 nigel 93 is not possible to handle an arbitrary nesting depth.
1896     .P
1897     For some time, Perl has provided a facility that allows regular expressions to
1898     recurse (amongst other things). It does this by interpolating Perl code in the
1899     expression at run time, and the code can refer to the expression itself. A Perl
1900     pattern using code interpolation to solve the parentheses problem can be
1901     created like this:
1902 nigel 75 .sp
1903     $re = qr{\e( (?: (?>[^()]+) | (?p{$re}) )* \e)}x;
1904     .sp
1905 nigel 63 The (?p{...}) item interpolates Perl code at run time, and in this case refers
1906 nigel 93 recursively to the pattern in which it appears.
1907 nigel 75 .P
1908 nigel 93 Obviously, PCRE cannot support the interpolation of Perl code. Instead, it
1909     supports special syntax for recursion of the entire pattern, and also for
1910     individual subpattern recursion. After its introduction in PCRE and Python,
1911     this kind of recursion was introduced into Perl at release 5.10.
1912 nigel 75 .P
1913 nigel 93 A special item that consists of (? followed by a number greater than zero and a
1914     closing parenthesis is a recursive call of the subpattern of the given number,
1915     provided that it occurs inside that subpattern. (If not, it is a "subroutine"
1916     call, which is described in the next section.) The special item (?R) or (?0) is
1917     a recursive call of the entire regular expression.
1918 nigel 87 .P
1919 nigel 93 In PCRE (like Python, but unlike Perl), a recursive subpattern call is always
1920     treated as an atomic group. That is, once it has matched some of the subject
1921     string, it is never re-entered, even if it contains untried alternatives and
1922     there is a subsequent matching failure.
1923     .P
1924 nigel 87 This PCRE pattern solves the nested parentheses problem (assume the
1925     PCRE_EXTENDED option is set so that white space is ignored):
1926 nigel 75 .sp
1927     \e( ( (?>[^()]+) | (?R) )* \e)
1928     .sp
1929 nigel 63 First it matches an opening parenthesis. Then it matches any number of
1930     substrings which can either be a sequence of non-parentheses, or a recursive
1931 nigel 87 match of the pattern itself (that is, a correctly parenthesized substring).
1932 nigel 63 Finally there is a closing parenthesis.
1933 nigel 75 .P
1934 nigel 63 If this were part of a larger pattern, you would not want to recurse the entire
1935     pattern, so instead you could use this:
1936 nigel 75 .sp
1937     ( \e( ( (?>[^()]+) | (?1) )* \e) )
1938     .sp
1939 nigel 63 We have put the pattern into parentheses, and caused the recursion to refer to
1940 ph10 172 them instead of the whole pattern.
1941 ph10 166 .P
1942     In a larger pattern, keeping track of parenthesis numbers can be tricky. This
1943     is made easier by the use of relative references. (A Perl 5.10 feature.)
1944     Instead of (?1) in the pattern above you can write (?-2) to refer to the second
1945     most recently opened parentheses preceding the recursion. In other words, a
1946     negative number counts capturing parentheses leftwards from the point at which
1947     it is encountered.
1948     .P
1949     It is also possible to refer to subsequently opened parentheses, by writing
1950     references such as (?+2). However, these cannot be recursive because the
1951     reference is not inside the parentheses that are referenced. They are always
1952     "subroutine" calls, as described in the next section.
1953     .P
1954     An alternative approach is to use named parentheses instead. The Perl syntax
1955     for this is (?&name); PCRE's earlier syntax (?P>name) is also supported. We
1956     could rewrite the above example as follows:
1957 nigel 75 .sp
1958 nigel 93 (?<pn> \e( ( (?>[^()]+) | (?&pn) )* \e) )
1959 nigel 75 .sp
1960 nigel 93 If there is more than one subpattern with the same name, the earliest one is
1961 ph10 172 used.
1962 ph10 166 .P
1963     This particular example pattern that we have been looking at contains nested
1964     unlimited repeats, and so the use of atomic grouping for matching strings of
1965     non-parentheses is important when applying the pattern to strings that do not
1966     match. For example, when this pattern is applied to
1967 nigel 75 .sp
1968 nigel 63 (aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa()
1969 nigel 75 .sp
1970 nigel 63 it yields "no match" quickly. However, if atomic grouping is not used,
1971     the match runs for a very long time indeed because there are so many different
1972     ways the + and * repeats can carve up the subject, and all have to be tested
1973     before failure can be reported.
1974 nigel 75 .P
1975 nigel 63 At the end of a match, the values set for any capturing subpatterns are those
1976     from the outermost level of the recursion at which the subpattern value is set.
1977     If you want to obtain intermediate values, a callout function can be used (see
1978 nigel 93 below and the
1979 nigel 63 .\" HREF
1980 nigel 75 \fBpcrecallout\fP
1981 nigel 63 .\"
1982     documentation). If the pattern above is matched against
1983 nigel 75 .sp
1984 nigel 63 (ab(cd)ef)
1985 nigel 75 .sp
1986 nigel 63 the value for the capturing parentheses is "ef", which is the last value taken
1987     on at the top level. If additional parentheses are added, giving
1988 nigel 75 .sp
1989     \e( ( ( (?>[^()]+) | (?R) )* ) \e)
1990 nigel 63 ^ ^
1991     ^ ^
1992 nigel 75 .sp
1993 nigel 63 the string they capture is "ab(cd)ef", the contents of the top level
1994     parentheses. If there are more than 15 capturing parentheses in a pattern, PCRE
1995     has to obtain extra memory to store data during a recursion, which it does by
1996 nigel 75 using \fBpcre_malloc\fP, freeing it via \fBpcre_free\fP afterwards. If no
1997 nigel 63 memory can be obtained, the match fails with the PCRE_ERROR_NOMEMORY error.
1998 nigel 75 .P
1999 nigel 63 Do not confuse the (?R) item with the condition (R), which tests for recursion.
2000     Consider this pattern, which matches text in angle brackets, allowing for
2001     arbitrary nesting. Only digits are allowed in nested brackets (that is, when
2002     recursing), whereas any characters are permitted at the outer level.
2003 nigel 75 .sp
2004     < (?: (?(R) \ed++ | [^<>]*+) | (?R)) * >
2005     .sp
2006 nigel 63 In this pattern, (?(R) is the start of a conditional subpattern, with two
2007     different alternatives for the recursive and non-recursive cases. The (?R) item
2008     is the actual recursive call.
2009 nigel 75 .
2010     .
2011 nigel 63 .\" HTML <a name="subpatternsassubroutines"></a>
2013 nigel 63 .rs
2014     .sp
2015     If the syntax for a recursive subpattern reference (either by number or by
2016     name) is used outside the parentheses to which it refers, it operates like a
2017 nigel 93 subroutine in a programming language. The "called" subpattern may be defined
2018 ph10 166 before or after the reference. A numbered reference can be absolute or
2019     relative, as in these examples:
2020 nigel 75 .sp
2021 ph10 166 (...(absolute)...)...(?2)...
2022     (...(relative)...)...(?-1)...
2023 ph10 172 (...(?+1)...(relative)...
2024 ph10 166 .sp
2025     An earlier example pointed out that the pattern
2026     .sp
2027 nigel 75 (sens|respons)e and \e1ibility
2028     .sp
2029 nigel 63 matches "sense and sensibility" and "response and responsibility", but not
2030     "sense and responsibility". If instead the pattern
2031 nigel 75 .sp
2032 nigel 63 (sens|respons)e and (?1)ibility
2033 nigel 75 .sp
2034 nigel 63 is used, it does match "sense and responsibility" as well as the other two
2035 nigel 93 strings. Another example is given in the discussion of DEFINE above.
2036 nigel 87 .P
2037     Like recursive subpatterns, a "subroutine" call is always treated as an atomic
2038     group. That is, once it has matched some of the subject string, it is never
2039     re-entered, even if it contains untried alternatives and there is a subsequent
2040     matching failure.
2041 nigel 93 .P
2042     When a subpattern is used as a subroutine, processing options such as
2043     case-independence are fixed when the subpattern is defined. They cannot be
2044     changed for different calls. For example, consider this pattern:
2045     .sp
2046 ph10 166 (abc)(?i:(?-1))
2047 nigel 93 .sp
2048     It matches "abcabc". It does not match "abcABC" because the change of
2049     processing option does not affect the called subpattern.
2050 nigel 75 .
2051     .
2052 ph10 333 .\" HTML <a name="onigurumasubroutines"></a>
2054     .rs
2055     .sp
2056     For compatibility with Oniguruma, the non-Perl syntax \eg followed by a name or
2057     a number enclosed either in angle brackets or single quotes, is an alternative
2058     syntax for referencing a subpattern as a subroutine, possibly recursively. Here
2059     are two of the examples used above, rewritten using this syntax:
2060     .sp
2061     (?<pn> \e( ( (?>[^()]+) | \eg<pn> )* \e) )
2062     (sens|respons)e and \eg'1'ibility
2063     .sp
2064     PCRE supports an extension to Oniguruma: if a number is preceded by a
2065     plus or a minus sign it is taken as a relative reference. For example:
2066     .sp
2067     (abc)(?i:\eg<-1>)
2068     .sp
2069     Note that \eg{...} (Perl syntax) and \eg<...> (Oniguruma syntax) are \fInot\fP
2070     synonymous. The former is a back reference; the latter is a subroutine call.
2071     .
2072     .
2073 nigel 63 .SH CALLOUTS
2074     .rs
2075     .sp
2076     Perl has a feature whereby using the sequence (?{...}) causes arbitrary Perl
2077     code to be obeyed in the middle of matching a regular expression. This makes it
2078     possible, amongst other things, to extract different substrings that match the
2079     same pair of parentheses when there is a repetition.
2080 nigel 75 .P
2081 nigel 63 PCRE provides a similar feature, but of course it cannot obey arbitrary Perl
2082     code. The feature is called "callout". The caller of PCRE provides an external
2083 nigel 75 function by putting its entry point in the global variable \fIpcre_callout\fP.
2084 nigel 63 By default, this variable contains NULL, which disables all calling out.
2085 nigel 75 .P
2086 nigel 63 Within a regular expression, (?C) indicates the points at which the external
2087     function is to be called. If you want to identify different callout points, you
2088     can put a number less than 256 after the letter C. The default value is zero.
2089     For example, this pattern has two callout points:
2090 nigel 75 .sp
2091 ph10 155 (?C1)abc(?C2)def
2092 nigel 75 .sp
2093     If the PCRE_AUTO_CALLOUT flag is passed to \fBpcre_compile()\fP, callouts are
2094     automatically installed before each item in the pattern. They are all numbered
2095     255.
2096     .P
2097     During matching, when PCRE reaches a callout point (and \fIpcre_callout\fP is
2098 nigel 63 set), the external function is called. It is provided with the number of the
2099 nigel 75 callout, the position in the pattern, and, optionally, one item of data
2100     originally supplied by the caller of \fBpcre_exec()\fP. The callout function
2101     may cause matching to proceed, to backtrack, or to fail altogether. A complete
2102     description of the interface to the callout function is given in the
2103 nigel 63 .\" HREF
2104 nigel 75 \fBpcrecallout\fP
2105 nigel 63 .\"
2106     documentation.
2107 nigel 93 .
2108     .
2110 ph10 210 .rs
2111     .sp
2112 ph10 211 Perl 5.10 introduced a number of "Special Backtracking Control Verbs", which
2113 ph10 210 are described in the Perl documentation as "experimental and subject to change
2114 ph10 211 or removal in a future version of Perl". It goes on to say: "Their usage in
2115     production code should be noted to avoid problems during upgrades." The same
2116 ph10 210 remarks apply to the PCRE features described in this section.
2117     .P
2118 ph10 211 Since these verbs are specifically related to backtracking, they can be used
2119     only when the pattern is to be matched using \fBpcre_exec()\fP, which uses a
2120     backtracking algorithm. They cause an error if encountered by
2121 ph10 210 \fBpcre_dfa_exec()\fP.
2122     .P
2123 ph10 211 The new verbs make use of what was previously invalid syntax: an opening
2124 ph10 210 parenthesis followed by an asterisk. In Perl, they are generally of the form
2125     (*VERB:ARG) but PCRE does not support the use of arguments, so its general
2126     form is just (*VERB). Any number of these verbs may occur in a pattern. There
2127     are two kinds:
2128     .
2129     .SS "Verbs that act immediately"
2130     .rs
2131     .sp
2132     The following verbs act as soon as they are encountered:
2133     .sp
2134     (*ACCEPT)
2135     .sp
2136     This verb causes the match to end successfully, skipping the remainder of the
2137 ph10 211 pattern. When inside a recursion, only the innermost pattern is ended
2138     immediately. PCRE differs from Perl in what happens if the (*ACCEPT) is inside
2139 ph10 210 capturing parentheses. In Perl, the data so far is captured: in PCRE no data is
2140     captured. For example:
2141     .sp
2142     A(A|B(*ACCEPT)|C)D
2143     .sp
2144 ph10 211 This matches "AB", "AAD", or "ACD", but when it matches "AB", no data is
2145 ph10 210 captured.
2146     .sp
2147     (*FAIL) or (*F)
2148     .sp
2149 ph10 211 This verb causes the match to fail, forcing backtracking to occur. It is
2150 ph10 210 equivalent to (?!) but easier to read. The Perl documentation notes that it is
2151     probably useful only when combined with (?{}) or (??{}). Those are, of course,
2152     Perl features that are not present in PCRE. The nearest equivalent is the
2153     callout feature, as for example in this pattern:
2154     .sp
2155     a+(?C)(*FAIL)
2156     .sp
2157 ph10 211 A match with the string "aaaa" always fails, but the callout is taken before
2158     each backtrack happens (in this example, 10 times).
2159 ph10 210 .
2160     .SS "Verbs that act after backtracking"
2161     .rs
2162     .sp
2163 ph10 211 The following verbs do nothing when they are encountered. Matching continues
2164     with what follows, but if there is no subsequent match, a failure is forced.
2165 ph10 210 The verbs differ in exactly what kind of failure occurs.
2166     .sp
2167     (*COMMIT)
2168     .sp
2169 ph10 211 This verb causes the whole match to fail outright if the rest of the pattern
2170 ph10 210 does not match. Even if the pattern is unanchored, no further attempts to find
2171     a match by advancing the start point take place. Once (*COMMIT) has been
2172 ph10 211 passed, \fBpcre_exec()\fP is committed to finding a match at the current
2173 ph10 210 starting point, or not at all. For example:
2174     .sp
2175     a+(*COMMIT)b
2176     .sp
2177 ph10 211 This matches "xxaab" but not "aacaab". It can be thought of as a kind of
2178 ph10 210 dynamic anchor, or "I've started, so I must finish."
2179     .sp
2180     (*PRUNE)
2181     .sp
2182 ph10 211 This verb causes the match to fail at the current position if the rest of the
2183     pattern does not match. If the pattern is unanchored, the normal "bumpalong"
2184 ph10 210 advance to the next starting character then happens. Backtracking can occur as
2185     usual to the left of (*PRUNE), or when matching to the right of (*PRUNE), but
2186     if there is no match to the right, backtracking cannot cross (*PRUNE).
2187 ph10 211 In simple cases, the use of (*PRUNE) is just an alternative to an atomic
2188 ph10 210 group or possessive quantifier, but there are some uses of (*PRUNE) that cannot
2189     be expressed in any other way.
2190     .sp
2191     (*SKIP)
2192     .sp
2193 ph10 211 This verb is like (*PRUNE), except that if the pattern is unanchored, the
2194 ph10 210 "bumpalong" advance is not to the next character, but to the position in the
2195     subject where (*SKIP) was encountered. (*SKIP) signifies that whatever text
2196     was matched leading up to it cannot be part of a successful match. Consider:
2197     .sp
2198     a+(*SKIP)b
2199     .sp
2200 ph10 211 If the subject is "aaaac...", after the first match attempt fails (starting at
2201 ph10 210 the first character in the string), the starting point skips on to start the
2202 ph10 211 next attempt at "c". Note that a possessive quantifer does not have the same
2203 ph10 210 effect in this example; although it would suppress backtracking during the
2204     first match attempt, the second attempt would start at the second character
2205     instead of skipping on to "c".
2206     .sp
2207     (*THEN)
2208 ph10 211 .sp
2209 ph10 210 This verb causes a skip to the next alternation if the rest of the pattern does
2210     not match. That is, it cancels pending backtracking, but only within the
2211     current alternation. Its name comes from the observation that it can be used
2212     for a pattern-based if-then-else block:
2213     .sp
2214     ( COND1 (*THEN) FOO | COND2 (*THEN) BAR | COND3 (*THEN) BAZ ) ...
2215     .sp
2216 ph10 211 If the COND1 pattern matches, FOO is tried (and possibly further items after
2217 ph10 210 the end of the group if FOO succeeds); on failure the matcher skips to the
2218 ph10 211 second alternative and tries COND2, without backtracking into COND1. If (*THEN)
2219 ph10 210 is used outside of any alternation, it acts exactly like (*PRUNE).
2220     .
2221     .
2222 nigel 93 .SH "SEE ALSO"
2223     .rs
2224     .sp
2225     \fBpcreapi\fP(3), \fBpcrecallout\fP(3), \fBpcrematching\fP(3), \fBpcre\fP(3).
2226 ph10 99 .
2227     .
2228     .SH AUTHOR
2229     .rs
2230     .sp
2231     .nf
2232     Philip Hazel
2233     University Computing Service
2234     Cambridge CB2 3QH, England.
2235     .fi
2236     .
2237     .
2238     .SH REVISION
2239     .rs
2240     .sp
2241     .nf
2242 ph10 335 Last updated: 12 April 2008
2243 ph10 333 Copyright (c) 1997-2008 University of Cambridge.
2244 ph10 99 .fi


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

ViewVC Help
Powered by ViewVC 1.1.12