/[pcre]/code/trunk/doc/pcrepattern.3
ViewVC logotype

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

Parent Directory Parent Directory | Revision Log Revision Log


Revision 175 - (hide annotations) (download)
Mon Jun 11 13:38:38 2007 UTC (7 years, 1 month ago) by ph10
File size: 79109 byte(s)
Added the Perl 5.10 (?| "branch reset" feature.

1 nigel 79 .TH PCREPATTERN 3
2 nigel 63 .SH NAME
3     PCRE - Perl-compatible regular expressions
4 nigel 75 .SH "PCRE REGULAR EXPRESSION DETAILS"
5 nigel 63 .rs
6     .sp
7     The syntax and semantics of the regular expressions supported by PCRE are
8     described below. Regular expressions are also described in the Perl
9 nigel 75 documentation and in a number of books, some of which have copious examples.
10     Jeffrey Friedl's "Mastering Regular Expressions", published by O'Reilly, covers
11     regular expressions in great detail. This description of PCRE's regular
12     expressions is intended as reference material.
13     .P
14     The original operation of PCRE was on strings of one-byte characters. However,
15     there is now also support for UTF-8 character strings. To use this, you must
16     build PCRE to include UTF-8 support, and then call \fBpcre_compile()\fP with
17     the PCRE_UTF8 option. How this affects pattern matching is mentioned in several
18     places below. There is also a summary of UTF-8 features in the
19 nigel 63 .\" HTML <a href="pcre.html#utf8support">
20     .\" </a>
21     section on UTF-8 support
22     .\"
23     in the main
24     .\" HREF
25 nigel 75 \fBpcre\fP
26 nigel 63 .\"
27     page.
28 nigel 75 .P
29 nigel 77 The remainder of this document discusses the patterns that are supported by
30     PCRE when its main matching function, \fBpcre_exec()\fP, is used.
31     From release 6.0, PCRE offers a second matching function,
32     \fBpcre_dfa_exec()\fP, which matches using a different algorithm that is not
33 ph10 172 Perl-compatible. Some of the features discussed below are not available when
34 ph10 168 \fBpcre_dfa_exec()\fP is used. The advantages and disadvantages of the
35     alternative function, and how it differs from the normal function, are
36     discussed in the
37 nigel 77 .\" HREF
38     \fBpcrematching\fP
39     .\"
40     page.
41 nigel 93 .
42     .
43     .SH "CHARACTERS AND METACHARACTERS"
44     .rs
45     .sp
46 nigel 63 A regular expression is a pattern that is matched against a subject string from
47     left to right. Most characters stand for themselves in a pattern, and match the
48     corresponding characters in the subject. As a trivial example, the pattern
49 nigel 75 .sp
50 nigel 63 The quick brown fox
51 nigel 75 .sp
52 nigel 77 matches a portion of a subject string that is identical to itself. When
53     caseless matching is specified (the PCRE_CASELESS option), letters are matched
54     independently of case. In UTF-8 mode, PCRE always understands the concept of
55     case for characters whose values are less than 128, so caseless matching is
56     always possible. For characters with higher values, the concept of case is
57     supported if PCRE is compiled with Unicode property support, but not otherwise.
58     If you want to use caseless matching for characters 128 and above, you must
59     ensure that PCRE is compiled with Unicode property support as well as with
60     UTF-8 support.
61     .P
62     The power of regular expressions comes from the ability to include alternatives
63     and repetitions in the pattern. These are encoded in the pattern by the use of
64 nigel 75 \fImetacharacters\fP, which do not stand for themselves but instead are
65 nigel 63 interpreted in some special way.
66 nigel 75 .P
67     There are two different sets of metacharacters: those that are recognized
68 nigel 63 anywhere in the pattern except within square brackets, and those that are
69 nigel 93 recognized within square brackets. Outside square brackets, the metacharacters
70     are as follows:
71 nigel 75 .sp
72     \e general escape character with several uses
73 nigel 63 ^ assert start of string (or line, in multiline mode)
74     $ assert end of string (or line, in multiline mode)
75     . match any character except newline (by default)
76     [ start character class definition
77     | start of alternative branch
78     ( start subpattern
79     ) end subpattern
80     ? extends the meaning of (
81     also 0 or 1 quantifier
82     also quantifier minimizer
83     * 0 or more quantifier
84     + 1 or more quantifier
85     also "possessive quantifier"
86     { start min/max quantifier
87 nigel 75 .sp
88 nigel 63 Part of a pattern that is in square brackets is called a "character class". In
89 nigel 75 a character class the only metacharacters are:
90     .sp
91     \e general escape character
92 nigel 63 ^ negate the class, but only if the first character
93     - indicates character range
94 nigel 75 .\" JOIN
95 nigel 63 [ POSIX character class (only if followed by POSIX
96     syntax)
97     ] terminates the character class
98 nigel 75 .sp
99     The following sections describe the use of each of the metacharacters.
100     .
101 nigel 93 .
102 nigel 63 .SH BACKSLASH
103     .rs
104     .sp
105     The backslash character has several uses. Firstly, if it is followed by a
106 nigel 91 non-alphanumeric character, it takes away any special meaning that character
107     may have. This use of backslash as an escape character applies both inside and
108 nigel 63 outside character classes.
109 nigel 75 .P
110     For example, if you want to match a * character, you write \e* in the pattern.
111 nigel 63 This escaping action applies whether or not the following character would
112 nigel 75 otherwise be interpreted as a metacharacter, so it is always safe to precede a
113     non-alphanumeric with backslash to specify that it stands for itself. In
114     particular, if you want to match a backslash, you write \e\e.
115     .P
116 nigel 63 If a pattern is compiled with the PCRE_EXTENDED option, whitespace in the
117     pattern (other than in a character class) and characters between a # outside
118 nigel 91 a character class and the next newline are ignored. An escaping backslash can
119     be used to include a whitespace or # character as part of the pattern.
120 nigel 75 .P
121 nigel 63 If you want to remove the special meaning from a sequence of characters, you
122 nigel 75 can do so by putting them between \eQ and \eE. This is different from Perl in
123     that $ and @ are handled as literals in \eQ...\eE sequences in PCRE, whereas in
124 nigel 63 Perl, $ and @ cause variable interpolation. Note the following examples:
125 nigel 75 .sp
126 nigel 63 Pattern PCRE matches Perl matches
127 nigel 75 .sp
128     .\" JOIN
129     \eQabc$xyz\eE abc$xyz abc followed by the
130 nigel 63 contents of $xyz
131 nigel 75 \eQabc\e$xyz\eE abc\e$xyz abc\e$xyz
132     \eQabc\eE\e$\eQxyz\eE abc$xyz abc$xyz
133     .sp
134     The \eQ...\eE sequence is recognized both inside and outside character classes.
135     .
136     .
137     .\" HTML <a name="digitsafterbackslash"></a>
138     .SS "Non-printing characters"
139     .rs
140     .sp
141 nigel 63 A second use of backslash provides a way of encoding non-printing characters
142     in patterns in a visible manner. There is no restriction on the appearance of
143     non-printing characters, apart from the binary zero that terminates a pattern,
144     but when a pattern is being prepared by text editing, it is usually easier to
145     use one of the following escape sequences than the binary character it
146     represents:
147 nigel 75 .sp
148     \ea alarm, that is, the BEL character (hex 07)
149     \ecx "control-x", where x is any character
150     \ee escape (hex 1B)
151     \ef formfeed (hex 0C)
152     \en newline (hex 0A)
153     \er carriage return (hex 0D)
154     \et tab (hex 09)
155     \eddd character with octal code ddd, or backreference
156     \exhh character with hex code hh
157 nigel 87 \ex{hhh..} character with hex code hhh..
158 nigel 75 .sp
159     The precise effect of \ecx is as follows: if x is a lower case letter, it
160 nigel 63 is converted to upper case. Then bit 6 of the character (hex 40) is inverted.
161 nigel 75 Thus \ecz becomes hex 1A, but \ec{ becomes hex 3B, while \ec; becomes hex
162 nigel 63 7B.
163 nigel 75 .P
164     After \ex, from zero to two hexadecimal digits are read (letters can be in
165 nigel 87 upper or lower case). Any number of hexadecimal digits may appear between \ex{
166     and }, but the value of the character code must be less than 256 in non-UTF-8
167     mode, and less than 2**31 in UTF-8 mode (that is, the maximum hexadecimal value
168     is 7FFFFFFF). If characters other than hexadecimal digits appear between \ex{
169     and }, or if there is no terminating }, this form of escape is not recognized.
170     Instead, the initial \ex will be interpreted as a basic hexadecimal escape,
171     with no following digits, giving a character whose value is zero.
172 nigel 75 .P
173 nigel 63 Characters whose value is less than 256 can be defined by either of the two
174 nigel 87 syntaxes for \ex. There is no difference in the way they are handled. For
175     example, \exdc is exactly the same as \ex{dc}.
176 nigel 75 .P
177 nigel 91 After \e0 up to two further octal digits are read. If there are fewer than two
178     digits, just those that are present are used. Thus the sequence \e0\ex\e07
179     specifies two binary zeros followed by a BEL character (code value 7). Make
180     sure you supply two digits after the initial zero if the pattern character that
181     follows is itself an octal digit.
182 nigel 75 .P
183 nigel 63 The handling of a backslash followed by a digit other than 0 is complicated.
184     Outside a character class, PCRE reads it and any following digits as a decimal
185     number. If the number is less than 10, or if there have been at least that many
186     previous capturing left parentheses in the expression, the entire sequence is
187 nigel 75 taken as a \fIback reference\fP. A description of how this works is given
188     .\" HTML <a href="#backreferences">
189     .\" </a>
190     later,
191     .\"
192     following the discussion of
193     .\" HTML <a href="#subpattern">
194     .\" </a>
195     parenthesized subpatterns.
196     .\"
197     .P
198 nigel 63 Inside a character class, or if the decimal number is greater than 9 and there
199     have not been that many capturing subpatterns, PCRE re-reads up to three octal
200 nigel 93 digits following the backslash, and uses them to generate a data character. Any
201 nigel 91 subsequent digits stand for themselves. In non-UTF-8 mode, the value of a
202     character specified in octal must be less than \e400. In UTF-8 mode, values up
203     to \e777 are permitted. For example:
204 nigel 75 .sp
205     \e040 is another way of writing a space
206     .\" JOIN
207     \e40 is the same, provided there are fewer than 40
208 nigel 63 previous capturing subpatterns
209 nigel 75 \e7 is always a back reference
210     .\" JOIN
211     \e11 might be a back reference, or another way of
212 nigel 63 writing a tab
213 nigel 75 \e011 is always a tab
214     \e0113 is a tab followed by the character "3"
215     .\" JOIN
216     \e113 might be a back reference, otherwise the
217 nigel 63 character with octal code 113
218 nigel 75 .\" JOIN
219     \e377 might be a back reference, otherwise
220 nigel 63 the byte consisting entirely of 1 bits
221 nigel 75 .\" JOIN
222     \e81 is either a back reference, or a binary zero
223 nigel 63 followed by the two characters "8" and "1"
224 nigel 75 .sp
225 nigel 63 Note that octal values of 100 or greater must not be introduced by a leading
226     zero, because no more than three octal digits are ever read.
227 nigel 75 .P
228 nigel 91 All the sequences that define a single character value can be used both inside
229     and outside character classes. In addition, inside a character class, the
230     sequence \eb is interpreted as the backspace character (hex 08), and the
231 nigel 93 sequences \eR and \eX are interpreted as the characters "R" and "X",
232     respectively. Outside a character class, these sequences have different
233     meanings
234 nigel 75 .\" HTML <a href="#uniextseq">
235     .\" </a>
236     (see below).
237     .\"
238     .
239     .
240 nigel 93 .SS "Absolute and relative back references"
241     .rs
242     .sp
243     The sequence \eg followed by a positive or negative number, optionally enclosed
244 ph10 172 in braces, is an absolute or relative back reference. A named back reference
245 ph10 171 can be coded as \eg{name}. Back references are discussed
246 nigel 93 .\" HTML <a href="#backreferences">
247     .\" </a>
248     later,
249     .\"
250     following the discussion of
251     .\" HTML <a href="#subpattern">
252     .\" </a>
253     parenthesized subpatterns.
254     .\"
255     .
256     .
257 nigel 75 .SS "Generic character types"
258     .rs
259     .sp
260 nigel 93 Another use of backslash is for specifying generic character types. The
261 nigel 75 following are always recognized:
262     .sp
263     \ed any decimal digit
264     \eD any character that is not a decimal digit
265     \es any whitespace character
266     \eS any character that is not a whitespace character
267     \ew any "word" character
268     \eW any "non-word" character
269     .sp
270 nigel 63 Each pair of escape sequences partitions the complete set of characters into
271     two disjoint sets. Any given character matches one, and only one, of each pair.
272 nigel 75 .P
273     These character type sequences can appear both inside and outside character
274     classes. They each match one character of the appropriate type. If the current
275     matching point is at the end of the subject string, all of them fail, since
276     there is no character to match.
277     .P
278     For compatibility with Perl, \es does not match the VT character (code 11).
279     This makes it different from the the POSIX "space" class. The \es characters
280 nigel 91 are HT (9), LF (10), FF (12), CR (13), and space (32). (If "use locale;" is
281     included in a Perl script, \es may match the VT character. In PCRE, it never
282     does.)
283 nigel 75 .P
284     A "word" character is an underscore or any character less than 256 that is a
285     letter or digit. The definition of letters and digits is controlled by PCRE's
286     low-valued character tables, and may vary if locale-specific matching is taking
287     place (see
288 nigel 63 .\" HTML <a href="pcreapi.html#localesupport">
289     .\" </a>
290     "Locale support"
291     .\"
292     in the
293     .\" HREF
294 nigel 75 \fBpcreapi\fP
295 nigel 63 .\"
296 ph10 139 page). For example, in a French locale such as "fr_FR" in Unix-like systems,
297     or "french" in Windows, some character codes greater than 128 are used for
298     accented letters, and these are matched by \ew.
299 nigel 75 .P
300     In UTF-8 mode, characters with values greater than 128 never match \ed, \es, or
301     \ew, and always match \eD, \eS, and \eW. This is true even when Unicode
302 nigel 87 character property support is available. The use of locales with Unicode is
303     discouraged.
304 nigel 75 .
305     .
306 nigel 93 .SS "Newline sequences"
307     .rs
308     .sp
309     Outside a character class, the escape sequence \eR matches any Unicode newline
310     sequence. This is an extension to Perl. In non-UTF-8 mode \eR is equivalent to
311     the following:
312     .sp
313     (?>\er\en|\en|\ex0b|\ef|\er|\ex85)
314     .sp
315     This is an example of an "atomic group", details of which are given
316     .\" HTML <a href="#atomicgroup">
317     .\" </a>
318     below.
319     .\"
320     This particular group matches either the two-character sequence CR followed by
321     LF, or one of the single characters LF (linefeed, U+000A), VT (vertical tab,
322     U+000B), FF (formfeed, U+000C), CR (carriage return, U+000D), or NEL (next
323     line, U+0085). The two-character sequence is treated as a single unit that
324     cannot be split.
325     .P
326     In UTF-8 mode, two additional characters whose codepoints are greater than 255
327     are added: LS (line separator, U+2028) and PS (paragraph separator, U+2029).
328     Unicode character property support is not needed for these characters to be
329     recognized.
330     .P
331     Inside a character class, \eR matches the letter "R".
332     .
333     .
334 nigel 75 .\" HTML <a name="uniextseq"></a>
335     .SS Unicode character properties
336     .rs
337     .sp
338     When PCRE is built with Unicode character property support, three additional
339 nigel 87 escape sequences to match character properties are available when UTF-8 mode
340 nigel 75 is selected. They are:
341     .sp
342 nigel 87 \ep{\fIxx\fP} a character with the \fIxx\fP property
343     \eP{\fIxx\fP} a character without the \fIxx\fP property
344     \eX an extended Unicode sequence
345 nigel 75 .sp
346 nigel 87 The property names represented by \fIxx\fP above are limited to the Unicode
347     script names, the general category properties, and "Any", which matches any
348     character (including newline). Other properties such as "InMusicalSymbols" are
349     not currently supported by PCRE. Note that \eP{Any} does not match any
350     characters, so always causes a match failure.
351 nigel 75 .P
352 nigel 87 Sets of Unicode characters are defined as belonging to certain scripts. A
353     character from one of these sets can be matched using a script name. For
354     example:
355 nigel 75 .sp
356 nigel 87 \ep{Greek}
357     \eP{Han}
358     .sp
359     Those that are not part of an identified script are lumped together as
360     "Common". The current list of scripts is:
361     .P
362     Arabic,
363     Armenian,
364 nigel 93 Balinese,
365 nigel 87 Bengali,
366     Bopomofo,
367     Braille,
368     Buginese,
369     Buhid,
370     Canadian_Aboriginal,
371     Cherokee,
372     Common,
373     Coptic,
374 nigel 93 Cuneiform,
375 nigel 87 Cypriot,
376     Cyrillic,
377     Deseret,
378     Devanagari,
379     Ethiopic,
380     Georgian,
381     Glagolitic,
382     Gothic,
383     Greek,
384     Gujarati,
385     Gurmukhi,
386     Han,
387     Hangul,
388     Hanunoo,
389     Hebrew,
390     Hiragana,
391     Inherited,
392     Kannada,
393     Katakana,
394     Kharoshthi,
395     Khmer,
396     Lao,
397     Latin,
398     Limbu,
399     Linear_B,
400     Malayalam,
401     Mongolian,
402     Myanmar,
403     New_Tai_Lue,
404 nigel 93 Nko,
405 nigel 87 Ogham,
406     Old_Italic,
407     Old_Persian,
408     Oriya,
409     Osmanya,
410 nigel 93 Phags_Pa,
411     Phoenician,
412 nigel 87 Runic,
413     Shavian,
414     Sinhala,
415     Syloti_Nagri,
416     Syriac,
417     Tagalog,
418     Tagbanwa,
419     Tai_Le,
420     Tamil,
421     Telugu,
422     Thaana,
423     Thai,
424     Tibetan,
425     Tifinagh,
426     Ugaritic,
427     Yi.
428     .P
429     Each character has exactly one general category property, specified by a
430     two-letter abbreviation. For compatibility with Perl, negation can be specified
431     by including a circumflex between the opening brace and the property name. For
432     example, \ep{^Lu} is the same as \eP{Lu}.
433     .P
434     If only one letter is specified with \ep or \eP, it includes all the general
435     category properties that start with that letter. In this case, in the absence
436     of negation, the curly brackets in the escape sequence are optional; these two
437     examples have the same effect:
438     .sp
439 nigel 75 \ep{L}
440     \epL
441     .sp
442 nigel 87 The following general category property codes are supported:
443 nigel 75 .sp
444     C Other
445     Cc Control
446     Cf Format
447     Cn Unassigned
448     Co Private use
449     Cs Surrogate
450     .sp
451     L Letter
452     Ll Lower case letter
453     Lm Modifier letter
454     Lo Other letter
455     Lt Title case letter
456     Lu Upper case letter
457     .sp
458     M Mark
459     Mc Spacing mark
460     Me Enclosing mark
461     Mn Non-spacing mark
462     .sp
463     N Number
464     Nd Decimal number
465     Nl Letter number
466     No Other number
467     .sp
468     P Punctuation
469     Pc Connector punctuation
470     Pd Dash punctuation
471     Pe Close punctuation
472     Pf Final punctuation
473     Pi Initial punctuation
474     Po Other punctuation
475     Ps Open punctuation
476     .sp
477     S Symbol
478     Sc Currency symbol
479     Sk Modifier symbol
480     Sm Mathematical symbol
481     So Other symbol
482     .sp
483     Z Separator
484     Zl Line separator
485     Zp Paragraph separator
486     Zs Space separator
487     .sp
488 nigel 87 The special property L& is also supported: it matches a character that has
489     the Lu, Ll, or Lt property, in other words, a letter that is not classified as
490     a modifier or "other".
491 nigel 75 .P
492 nigel 87 The long synonyms for these properties that Perl supports (such as \ep{Letter})
493 nigel 91 are not supported by PCRE, nor is it permitted to prefix any of these
494 nigel 87 properties with "Is".
495     .P
496     No character that is in the Unicode table has the Cn (unassigned) property.
497     Instead, this property is assumed for any code point that is not in the
498     Unicode table.
499     .P
500 nigel 75 Specifying caseless matching does not affect these escape sequences. For
501     example, \ep{Lu} always matches only upper case letters.
502     .P
503     The \eX escape matches any number of Unicode characters that form an extended
504     Unicode sequence. \eX is equivalent to
505     .sp
506     (?>\ePM\epM*)
507     .sp
508     That is, it matches a character without the "mark" property, followed by zero
509     or more characters with the "mark" property, and treats the sequence as an
510     atomic group
511     .\" HTML <a href="#atomicgroup">
512     .\" </a>
513     (see below).
514     .\"
515     Characters with the "mark" property are typically accents that affect the
516     preceding character.
517     .P
518     Matching characters by Unicode property is not fast, because PCRE has to search
519     a structure that contains data for over fifteen thousand characters. That is
520     why the traditional escape sequences such as \ed and \ew do not use Unicode
521     properties in PCRE.
522     .
523     .
524 ph10 168 .\" HTML <a name="resetmatchstart"></a>
525     .SS "Resetting the match start"
526     .rs
527     .sp
528 ph10 172 The escape sequence \eK, which is a Perl 5.10 feature, causes any previously
529     matched characters not to be included in the final matched sequence. For
530 ph10 168 example, the pattern:
531     .sp
532     foo\eKbar
533     .sp
534 ph10 172 matches "foobar", but reports that it has matched "bar". This feature is
535 ph10 168 similar to a lookbehind assertion
536     .\" HTML <a href="#lookbehind">
537     .\" </a>
538     (described below).
539     .\"
540 ph10 172 However, in this case, the part of the subject before the real match does not
541     have to be of fixed length, as lookbehind assertions do. The use of \eK does
542 ph10 168 not interfere with the setting of
543     .\" HTML <a href="#subpattern">
544     .\" </a>
545     captured substrings.
546 ph10 172 .\"
547 ph10 168 For example, when the pattern
548     .sp
549     (foo)\eKbar
550     .sp
551 ph10 172 matches "foobar", the first substring is still set to "foo".
552 ph10 168 .
553     .
554 nigel 75 .\" HTML <a name="smallassertions"></a>
555     .SS "Simple assertions"
556     .rs
557     .sp
558 nigel 93 The final use of backslash is for certain simple assertions. An assertion
559 nigel 63 specifies a condition that has to be met at a particular point in a match,
560     without consuming any characters from the subject string. The use of
561 nigel 75 subpatterns for more complicated assertions is described
562     .\" HTML <a href="#bigassertions">
563     .\" </a>
564     below.
565     .\"
566 nigel 91 The backslashed assertions are:
567 nigel 75 .sp
568     \eb matches at a word boundary
569     \eB matches when not at a word boundary
570 nigel 93 \eA matches at the start of the subject
571     \eZ matches at the end of the subject
572     also matches before a newline at the end of the subject
573     \ez matches only at the end of the subject
574     \eG matches at the first matching position in the subject
575 nigel 75 .sp
576     These assertions may not appear in character classes (but note that \eb has a
577 nigel 63 different meaning, namely the backspace character, inside a character class).
578 nigel 75 .P
579 nigel 63 A word boundary is a position in the subject string where the current character
580 nigel 75 and the previous character do not both match \ew or \eW (i.e. one matches
581     \ew and the other matches \eW), or the start or end of the string if the
582     first or last character matches \ew, respectively.
583     .P
584     The \eA, \eZ, and \ez assertions differ from the traditional circumflex and
585     dollar (described in the next section) in that they only ever match at the very
586     start and end of the subject string, whatever options are set. Thus, they are
587     independent of multiline mode. These three assertions are not affected by the
588     PCRE_NOTBOL or PCRE_NOTEOL options, which affect only the behaviour of the
589     circumflex and dollar metacharacters. However, if the \fIstartoffset\fP
590     argument of \fBpcre_exec()\fP is non-zero, indicating that matching is to start
591     at a point other than the beginning of the subject, \eA can never match. The
592 nigel 91 difference between \eZ and \ez is that \eZ matches before a newline at the end
593     of the string as well as at the very end, whereas \ez matches only at the end.
594 nigel 75 .P
595     The \eG assertion is true only when the current matching position is at the
596     start point of the match, as specified by the \fIstartoffset\fP argument of
597     \fBpcre_exec()\fP. It differs from \eA when the value of \fIstartoffset\fP is
598     non-zero. By calling \fBpcre_exec()\fP multiple times with appropriate
599 nigel 63 arguments, you can mimic Perl's /g option, and it is in this kind of
600 nigel 75 implementation where \eG can be useful.
601     .P
602     Note, however, that PCRE's interpretation of \eG, as the start of the current
603 nigel 63 match, is subtly different from Perl's, which defines it as the end of the
604     previous match. In Perl, these can be different when the previously matched
605     string was empty. Because PCRE does just one match at a time, it cannot
606     reproduce this behaviour.
607 nigel 75 .P
608     If all the alternatives of a pattern begin with \eG, the expression is anchored
609 nigel 63 to the starting match position, and the "anchored" flag is set in the compiled
610     regular expression.
611 nigel 75 .
612     .
613     .SH "CIRCUMFLEX AND DOLLAR"
614 nigel 63 .rs
615     .sp
616     Outside a character class, in the default matching mode, the circumflex
617 nigel 75 character is an assertion that is true only if the current matching point is
618     at the start of the subject string. If the \fIstartoffset\fP argument of
619     \fBpcre_exec()\fP is non-zero, circumflex can never match if the PCRE_MULTILINE
620 nigel 63 option is unset. Inside a character class, circumflex has an entirely different
621 nigel 75 meaning
622     .\" HTML <a href="#characterclass">
623     .\" </a>
624     (see below).
625     .\"
626     .P
627 nigel 63 Circumflex need not be the first character of the pattern if a number of
628     alternatives are involved, but it should be the first thing in each alternative
629     in which it appears if the pattern is ever to match that branch. If all
630     possible alternatives start with a circumflex, that is, if the pattern is
631     constrained to match only at the start of the subject, it is said to be an
632     "anchored" pattern. (There are also other constructs that can cause a pattern
633     to be anchored.)
634 nigel 75 .P
635     A dollar character is an assertion that is true only if the current matching
636 nigel 63 point is at the end of the subject string, or immediately before a newline
637 nigel 91 at the end of the string (by default). Dollar need not be the last character of
638     the pattern if a number of alternatives are involved, but it should be the last
639     item in any branch in which it appears. Dollar has no special meaning in a
640     character class.
641 nigel 75 .P
642 nigel 63 The meaning of dollar can be changed so that it matches only at the very end of
643     the string, by setting the PCRE_DOLLAR_ENDONLY option at compile time. This
644 nigel 75 does not affect the \eZ assertion.
645     .P
646 nigel 63 The meanings of the circumflex and dollar characters are changed if the
647 nigel 91 PCRE_MULTILINE option is set. When this is the case, a circumflex matches
648     immediately after internal newlines as well as at the start of the subject
649     string. It does not match after a newline that ends the string. A dollar
650     matches before any newlines in the string, as well as at the very end, when
651     PCRE_MULTILINE is set. When newline is specified as the two-character
652     sequence CRLF, isolated CR and LF characters do not indicate newlines.
653 nigel 75 .P
654 nigel 91 For example, the pattern /^abc$/ matches the subject string "def\enabc" (where
655     \en represents a newline) in multiline mode, but not otherwise. Consequently,
656     patterns that are anchored in single line mode because all branches start with
657     ^ are not anchored in multiline mode, and a match for circumflex is possible
658     when the \fIstartoffset\fP argument of \fBpcre_exec()\fP is non-zero. The
659     PCRE_DOLLAR_ENDONLY option is ignored if PCRE_MULTILINE is set.
660     .P
661 nigel 75 Note that the sequences \eA, \eZ, and \ez can be used to match the start and
662 nigel 63 end of the subject in both modes, and if all branches of a pattern start with
663 nigel 91 \eA it is always anchored, whether or not PCRE_MULTILINE is set.
664 nigel 75 .
665     .
666     .SH "FULL STOP (PERIOD, DOT)"
667 nigel 63 .rs
668     .sp
669     Outside a character class, a dot in the pattern matches any one character in
670 nigel 91 the subject string except (by default) a character that signifies the end of a
671 nigel 93 line. In UTF-8 mode, the matched character may be more than one byte long.
672 nigel 91 .P
673 nigel 93 When a line ending is defined as a single character, dot never matches that
674     character; when the two-character sequence CRLF is used, dot does not match CR
675     if it is immediately followed by LF, but otherwise it matches all characters
676     (including isolated CRs and LFs). When any Unicode line endings are being
677     recognized, dot does not match CR or LF or any of the other line ending
678     characters.
679     .P
680 nigel 91 The behaviour of dot with regard to newlines can be changed. If the PCRE_DOTALL
681 nigel 93 option is set, a dot matches any one character, without exception. If the
682     two-character sequence CRLF is present in the subject string, it takes two dots
683     to match it.
684 nigel 91 .P
685     The handling of dot is entirely independent of the handling of circumflex and
686     dollar, the only relationship being that they both involve newlines. Dot has no
687     special meaning in a character class.
688 nigel 75 .
689     .
690     .SH "MATCHING A SINGLE BYTE"
691 nigel 63 .rs
692     .sp
693 nigel 75 Outside a character class, the escape sequence \eC matches any one byte, both
694 nigel 93 in and out of UTF-8 mode. Unlike a dot, it always matches any line-ending
695     characters. The feature is provided in Perl in order to match individual bytes
696     in UTF-8 mode. Because it breaks up UTF-8 characters into individual bytes,
697     what remains in the string may be a malformed UTF-8 string. For this reason,
698     the \eC escape sequence is best avoided.
699 nigel 75 .P
700     PCRE does not allow \eC to appear in lookbehind assertions
701     .\" HTML <a href="#lookbehind">
702     .\" </a>
703     (described below),
704     .\"
705     because in UTF-8 mode this would make it impossible to calculate the length of
706     the lookbehind.
707     .
708     .
709     .\" HTML <a name="characterclass"></a>
710     .SH "SQUARE BRACKETS AND CHARACTER CLASSES"
711 nigel 63 .rs
712     .sp
713     An opening square bracket introduces a character class, terminated by a closing
714     square bracket. A closing square bracket on its own is not special. If a
715     closing square bracket is required as a member of the class, it should be the
716     first data character in the class (after an initial circumflex, if present) or
717     escaped with a backslash.
718 nigel 75 .P
719 nigel 63 A character class matches a single character in the subject. In UTF-8 mode, the
720     character may occupy more than one byte. A matched character must be in the set
721     of characters defined by the class, unless the first character in the class
722     definition is a circumflex, in which case the subject character must not be in
723     the set defined by the class. If a circumflex is actually required as a member
724     of the class, ensure it is not the first character, or escape it with a
725     backslash.
726 nigel 75 .P
727 nigel 63 For example, the character class [aeiou] matches any lower case vowel, while
728     [^aeiou] matches any character that is not a lower case vowel. Note that a
729 nigel 75 circumflex is just a convenient notation for specifying the characters that
730     are in the class by enumerating those that are not. A class that starts with a
731     circumflex is not an assertion: it still consumes a character from the subject
732     string, and therefore it fails if the current pointer is at the end of the
733     string.
734     .P
735 nigel 63 In UTF-8 mode, characters with values greater than 255 can be included in a
736 nigel 75 class as a literal string of bytes, or by using the \ex{ escaping mechanism.
737     .P
738 nigel 63 When caseless matching is set, any letters in a class represent both their
739     upper case and lower case versions, so for example, a caseless [aeiou] matches
740     "A" as well as "a", and a caseless [^aeiou] does not match "A", whereas a
741 nigel 77 caseful version would. In UTF-8 mode, PCRE always understands the concept of
742     case for characters whose values are less than 128, so caseless matching is
743     always possible. For characters with higher values, the concept of case is
744     supported if PCRE is compiled with Unicode property support, but not otherwise.
745     If you want to use caseless matching for characters 128 and above, you must
746     ensure that PCRE is compiled with Unicode property support as well as with
747     UTF-8 support.
748 nigel 75 .P
749 nigel 93 Characters that might indicate line breaks are never treated in any special way
750     when matching character classes, whatever line-ending sequence is in use, and
751     whatever setting of the PCRE_DOTALL and PCRE_MULTILINE options is used. A class
752     such as [^a] always matches one of these characters.
753 nigel 75 .P
754 nigel 63 The minus (hyphen) character can be used to specify a range of characters in a
755     character class. For example, [d-m] matches any letter between d and m,
756     inclusive. If a minus character is required in a class, it must be escaped with
757     a backslash or appear in a position where it cannot be interpreted as
758     indicating a range, typically as the first or last character in the class.
759 nigel 75 .P
760 nigel 63 It is not possible to have the literal character "]" as the end character of a
761     range. A pattern such as [W-]46] is interpreted as a class of two characters
762     ("W" and "-") followed by a literal string "46]", so it would match "W46]" or
763     "-46]". However, if the "]" is escaped with a backslash it is interpreted as
764 nigel 75 the end of range, so [W-\e]46] is interpreted as a class containing a range
765     followed by two other characters. The octal or hexadecimal representation of
766     "]" can also be used to end a range.
767     .P
768 nigel 63 Ranges operate in the collating sequence of character values. They can also be
769 nigel 75 used for characters specified numerically, for example [\e000-\e037]. In UTF-8
770 nigel 63 mode, ranges can include characters whose values are greater than 255, for
771 nigel 75 example [\ex{100}-\ex{2ff}].
772     .P
773 nigel 63 If a range that includes letters is used when caseless matching is set, it
774     matches the letters in either case. For example, [W-c] is equivalent to
775 nigel 75 [][\e\e^_`wxyzabc], matched caselessly, and in non-UTF-8 mode, if character
776 ph10 139 tables for a French locale are in use, [\exc8-\excb] matches accented E
777 nigel 75 characters in both cases. In UTF-8 mode, PCRE supports the concept of case for
778     characters with values greater than 128 only when it is compiled with Unicode
779     property support.
780     .P
781     The character types \ed, \eD, \ep, \eP, \es, \eS, \ew, and \eW may also appear
782     in a character class, and add the characters that they match to the class. For
783     example, [\edABCDEF] matches any hexadecimal digit. A circumflex can
784 nigel 63 conveniently be used with the upper case character types to specify a more
785     restricted set of characters than the matching lower case type. For example,
786 nigel 75 the class [^\eW_] matches any letter or digit, but not underscore.
787     .P
788     The only metacharacters that are recognized in character classes are backslash,
789     hyphen (only where it can be interpreted as specifying a range), circumflex
790     (only at the start), opening square bracket (only when it can be interpreted as
791     introducing a POSIX class name - see the next section), and the terminating
792     closing square bracket. However, escaping other non-alphanumeric characters
793     does no harm.
794     .
795     .
796     .SH "POSIX CHARACTER CLASSES"
797 nigel 63 .rs
798     .sp
799 nigel 75 Perl supports the POSIX notation for character classes. This uses names
800 nigel 63 enclosed by [: and :] within the enclosing square brackets. PCRE also supports
801     this notation. For example,
802 nigel 75 .sp
803 nigel 63 [01[:alpha:]%]
804 nigel 75 .sp
805 nigel 63 matches "0", "1", any alphabetic character, or "%". The supported class names
806     are
807 nigel 75 .sp
808 nigel 63 alnum letters and digits
809     alpha letters
810     ascii character codes 0 - 127
811     blank space or tab only
812     cntrl control characters
813 nigel 75 digit decimal digits (same as \ed)
814 nigel 63 graph printing characters, excluding space
815     lower lower case letters
816     print printing characters, including space
817     punct printing characters, excluding letters and digits
818 nigel 75 space white space (not quite the same as \es)
819 nigel 63 upper upper case letters
820 nigel 75 word "word" characters (same as \ew)
821 nigel 63 xdigit hexadecimal digits
822 nigel 75 .sp
823 nigel 63 The "space" characters are HT (9), LF (10), VT (11), FF (12), CR (13), and
824     space (32). Notice that this list includes the VT character (code 11). This
825 nigel 75 makes "space" different to \es, which does not include VT (for Perl
826 nigel 63 compatibility).
827 nigel 75 .P
828 nigel 63 The name "word" is a Perl extension, and "blank" is a GNU extension from Perl
829     5.8. Another Perl extension is negation, which is indicated by a ^ character
830     after the colon. For example,
831 nigel 75 .sp
832 nigel 63 [12[:^digit:]]
833 nigel 75 .sp
834 nigel 63 matches "1", "2", or any non-digit. PCRE (and Perl) also recognize the POSIX
835     syntax [.ch.] and [=ch=] where "ch" is a "collating element", but these are not
836     supported, and an error is given if they are encountered.
837 nigel 75 .P
838     In UTF-8 mode, characters with values greater than 128 do not match any of
839 nigel 63 the POSIX character classes.
840 nigel 75 .
841     .
842     .SH "VERTICAL BAR"
843 nigel 63 .rs
844     .sp
845     Vertical bar characters are used to separate alternative patterns. For example,
846     the pattern
847 nigel 75 .sp
848 nigel 63 gilbert|sullivan
849 nigel 75 .sp
850 nigel 63 matches either "gilbert" or "sullivan". Any number of alternatives may appear,
851 nigel 91 and an empty alternative is permitted (matching the empty string). The matching
852     process tries each alternative in turn, from left to right, and the first one
853     that succeeds is used. If the alternatives are within a subpattern
854 nigel 75 .\" HTML <a href="#subpattern">
855     .\" </a>
856     (defined below),
857     .\"
858     "succeeds" means matching the rest of the main pattern as well as the
859     alternative in the subpattern.
860     .
861     .
862     .SH "INTERNAL OPTION SETTING"
863 nigel 63 .rs
864     .sp
865     The settings of the PCRE_CASELESS, PCRE_MULTILINE, PCRE_DOTALL, and
866     PCRE_EXTENDED options can be changed from within the pattern by a sequence of
867     Perl option letters enclosed between "(?" and ")". The option letters are
868 nigel 75 .sp
869 nigel 63 i for PCRE_CASELESS
870     m for PCRE_MULTILINE
871     s for PCRE_DOTALL
872     x for PCRE_EXTENDED
873 nigel 75 .sp
874 nigel 63 For example, (?im) sets caseless, multiline matching. It is also possible to
875     unset these options by preceding the letter with a hyphen, and a combined
876     setting and unsetting such as (?im-sx), which sets PCRE_CASELESS and
877     PCRE_MULTILINE while unsetting PCRE_DOTALL and PCRE_EXTENDED, is also
878     permitted. If a letter appears both before and after the hyphen, the option is
879     unset.
880 nigel 75 .P
881 nigel 63 When an option change occurs at top level (that is, not inside subpattern
882     parentheses), the change applies to the remainder of the pattern that follows.
883     If the change is placed right at the start of a pattern, PCRE extracts it into
884     the global options (and it will therefore show up in data extracted by the
885 nigel 75 \fBpcre_fullinfo()\fP function).
886     .P
887 nigel 93 An option change within a subpattern (see below for a description of
888     subpatterns) affects only that part of the current pattern that follows it, so
889 nigel 75 .sp
890 nigel 63 (a(?i)b)c
891 nigel 75 .sp
892 nigel 63 matches abc and aBc and no other strings (assuming PCRE_CASELESS is not used).
893     By this means, options can be made to have different settings in different
894     parts of the pattern. Any changes made in one alternative do carry on
895     into subsequent branches within the same subpattern. For example,
896 nigel 75 .sp
897 nigel 63 (a(?i)b|c)
898 nigel 75 .sp
899 nigel 63 matches "ab", "aB", "c", and "C", even though when matching "C" the first
900     branch is abandoned before the option setting. This is because the effects of
901     option settings happen at compile time. There would be some very weird
902     behaviour otherwise.
903 nigel 75 .P
904 nigel 91 The PCRE-specific options PCRE_DUPNAMES, PCRE_UNGREEDY, and PCRE_EXTRA can be
905     changed in the same way as the Perl-compatible options by using the characters
906     J, U and X respectively.
907 nigel 75 .
908     .
909     .\" HTML <a name="subpattern"></a>
910 nigel 63 .SH SUBPATTERNS
911     .rs
912     .sp
913     Subpatterns are delimited by parentheses (round brackets), which can be nested.
914 nigel 75 Turning part of a pattern into a subpattern does two things:
915     .sp
916 nigel 63 1. It localizes a set of alternatives. For example, the pattern
917 nigel 75 .sp
918 nigel 63 cat(aract|erpillar|)
919 nigel 75 .sp
920 nigel 63 matches one of the words "cat", "cataract", or "caterpillar". Without the
921 nigel 93 parentheses, it would match "cataract", "erpillar" or an empty string.
922 nigel 75 .sp
923     2. It sets up the subpattern as a capturing subpattern. This means that, when
924     the whole pattern matches, that portion of the subject string that matched the
925     subpattern is passed back to the caller via the \fIovector\fP argument of
926     \fBpcre_exec()\fP. Opening parentheses are counted from left to right (starting
927     from 1) to obtain numbers for the capturing subpatterns.
928     .P
929 nigel 63 For example, if the string "the red king" is matched against the pattern
930 nigel 75 .sp
931 nigel 63 the ((red|white) (king|queen))
932 nigel 75 .sp
933 nigel 63 the captured substrings are "red king", "red", and "king", and are numbered 1,
934     2, and 3, respectively.
935 nigel 75 .P
936 nigel 63 The fact that plain parentheses fulfil two functions is not always helpful.
937     There are often times when a grouping subpattern is required without a
938     capturing requirement. If an opening parenthesis is followed by a question mark
939     and a colon, the subpattern does not do any capturing, and is not counted when
940     computing the number of any subsequent capturing subpatterns. For example, if
941     the string "the white queen" is matched against the pattern
942 nigel 75 .sp
943 nigel 63 the ((?:red|white) (king|queen))
944 nigel 75 .sp
945 nigel 63 the captured substrings are "white queen" and "queen", and are numbered 1 and
946 nigel 93 2. The maximum number of capturing subpatterns is 65535.
947 nigel 75 .P
948 nigel 63 As a convenient shorthand, if any option settings are required at the start of
949     a non-capturing subpattern, the option letters may appear between the "?" and
950     the ":". Thus the two patterns
951 nigel 75 .sp
952 nigel 63 (?i:saturday|sunday)
953     (?:(?i)saturday|sunday)
954 nigel 75 .sp
955 nigel 63 match exactly the same set of strings. Because alternative branches are tried
956     from left to right, and options are not reset until the end of the subpattern
957     is reached, an option setting in one branch does affect subsequent branches, so
958     the above patterns match "SUNDAY" as well as "Saturday".
959 nigel 75 .
960     .
961 ph10 175 .SH "DUPLICATE SUBPATTERN NUMBERS"
962     .rs
963     .sp
964     Perl 5.10 introduced a feature whereby each alternative in a subpattern uses
965     the same numbers for its capturing parentheses. Such a subpattern starts with
966     (?| and is itself a non-capturing subpattern. For example, consider this
967     pattern:
968     .sp
969     (?|(Sat)ur|(Sun))day
970     .sp
971     Because the two alternatives are inside a (?| group, both sets of capturing
972     parentheses are numbered one. Thus, when the pattern matches, you can look
973     at captured substring number one, whichever alternative matched. This construct
974     is useful when you want to capture part, but not all, of one of a number of
975     alternatives. Inside a (?| group, parentheses are numbered as usual, but the
976     number is reset at the start of each branch. The numbers of any capturing
977     buffers that follow the subpattern start after the highest number used in any
978     branch. The following example is taken from the Perl documentation.
979     The numbers underneath show in which buffer the captured content will be
980     stored.
981     .sp
982     # before ---------------branch-reset----------- after
983     / ( a ) (?| x ( y ) z | (p (q) r) | (t) u (v) ) ( z ) /x
984     # 1 2 2 3 2 3 4
985     .sp
986     A backreference or a recursive call to a numbered subpattern always refers to
987     the first one in the pattern with the given number.
988     .P
989     An alternative approach to using this "branch reset" feature is to use
990     duplicate named subpatterns, as described in the next section.
991     .
992     .
993 nigel 75 .SH "NAMED SUBPATTERNS"
994 nigel 63 .rs
995     .sp
996     Identifying capturing parentheses by number is simple, but it can be very hard
997     to keep track of the numbers in complicated regular expressions. Furthermore,
998 nigel 75 if an expression is modified, the numbers may change. To help with this
999 nigel 93 difficulty, PCRE supports the naming of subpatterns. This feature was not
1000     added to Perl until release 5.10. Python had the feature earlier, and PCRE
1001     introduced it at release 4.0, using the Python syntax. PCRE now supports both
1002     the Perl and the Python syntax.
1003     .P
1004     In PCRE, a subpattern can be named in one of three ways: (?<name>...) or
1005     (?'name'...) as in Perl, or (?P<name>...) as in Python. References to capturing
1006 nigel 91 parentheses from other parts of the pattern, such as
1007     .\" HTML <a href="#backreferences">
1008     .\" </a>
1009     backreferences,
1010     .\"
1011     .\" HTML <a href="#recursion">
1012     .\" </a>
1013     recursion,
1014     .\"
1015     and
1016     .\" HTML <a href="#conditions">
1017     .\" </a>
1018     conditions,
1019     .\"
1020     can be made by name as well as by number.
1021 nigel 75 .P
1022 nigel 91 Names consist of up to 32 alphanumeric characters and underscores. Named
1023 nigel 93 capturing parentheses are still allocated numbers as well as names, exactly as
1024     if the names were not present. The PCRE API provides function calls for
1025     extracting the name-to-number translation table from a compiled pattern. There
1026     is also a convenience function for extracting a captured substring by name.
1027 nigel 91 .P
1028     By default, a name must be unique within a pattern, but it is possible to relax
1029     this constraint by setting the PCRE_DUPNAMES option at compile time. This can
1030     be useful for patterns where only one instance of the named parentheses can
1031     match. Suppose you want to match the name of a weekday, either as a 3-letter
1032     abbreviation or as the full name, and in both cases you want to extract the
1033     abbreviation. This pattern (ignoring the line breaks) does the job:
1034     .sp
1035 nigel 93 (?<DN>Mon|Fri|Sun)(?:day)?|
1036     (?<DN>Tue)(?:sday)?|
1037     (?<DN>Wed)(?:nesday)?|
1038     (?<DN>Thu)(?:rsday)?|
1039     (?<DN>Sat)(?:urday)?
1040 nigel 91 .sp
1041     There are five capturing substrings, but only one is ever set after a match.
1042 ph10 175 (An alternative way of solving this problem is to use a "branch reset"
1043     subpattern, as described in the previous section.)
1044     .P
1045 nigel 91 The convenience function for extracting the data by name returns the substring
1046 nigel 93 for the first (and in this example, the only) subpattern of that name that
1047 nigel 91 matched. This saves searching to find which numbered subpattern it was. If you
1048     make a reference to a non-unique named subpattern from elsewhere in the
1049     pattern, the one that corresponds to the lowest number is used. For further
1050     details of the interfaces for handling named subpatterns, see the
1051 nigel 63 .\" HREF
1052 nigel 75 \fBpcreapi\fP
1053 nigel 63 .\"
1054     documentation.
1055 nigel 75 .
1056     .
1057 nigel 63 .SH REPETITION
1058     .rs
1059     .sp
1060     Repetition is specified by quantifiers, which can follow any of the following
1061     items:
1062 nigel 75 .sp
1063 nigel 63 a literal data character
1064 nigel 93 the dot metacharacter
1065 nigel 75 the \eC escape sequence
1066     the \eX escape sequence (in UTF-8 mode with Unicode properties)
1067 nigel 93 the \eR escape sequence
1068 nigel 75 an escape such as \ed that matches a single character
1069 nigel 63 a character class
1070     a back reference (see next section)
1071     a parenthesized subpattern (unless it is an assertion)
1072 nigel 75 .sp
1073 nigel 63 The general repetition quantifier specifies a minimum and maximum number of
1074     permitted matches, by giving the two numbers in curly brackets (braces),
1075     separated by a comma. The numbers must be less than 65536, and the first must
1076     be less than or equal to the second. For example:
1077 nigel 75 .sp
1078 nigel 63 z{2,4}
1079 nigel 75 .sp
1080 nigel 63 matches "zz", "zzz", or "zzzz". A closing brace on its own is not a special
1081     character. If the second number is omitted, but the comma is present, there is
1082     no upper limit; if the second number and the comma are both omitted, the
1083     quantifier specifies an exact number of required matches. Thus
1084 nigel 75 .sp
1085 nigel 63 [aeiou]{3,}
1086 nigel 75 .sp
1087 nigel 63 matches at least 3 successive vowels, but may match many more, while
1088 nigel 75 .sp
1089     \ed{8}
1090     .sp
1091 nigel 63 matches exactly 8 digits. An opening curly bracket that appears in a position
1092     where a quantifier is not allowed, or one that does not match the syntax of a
1093     quantifier, is taken as a literal character. For example, {,6} is not a
1094     quantifier, but a literal string of four characters.
1095 nigel 75 .P
1096 nigel 63 In UTF-8 mode, quantifiers apply to UTF-8 characters rather than to individual
1097 nigel 75 bytes. Thus, for example, \ex{100}{2} matches two UTF-8 characters, each of
1098     which is represented by a two-byte sequence. Similarly, when Unicode property
1099     support is available, \eX{3} matches three Unicode extended sequences, each of
1100     which may be several bytes long (and they may be of different lengths).
1101     .P
1102 nigel 63 The quantifier {0} is permitted, causing the expression to behave as if the
1103     previous item and the quantifier were not present.
1104 nigel 75 .P
1105 nigel 93 For convenience, the three most common quantifiers have single-character
1106     abbreviations:
1107 nigel 75 .sp
1108 nigel 63 * is equivalent to {0,}
1109     + is equivalent to {1,}
1110     ? is equivalent to {0,1}
1111 nigel 75 .sp
1112 nigel 63 It is possible to construct infinite loops by following a subpattern that can
1113     match no characters with a quantifier that has no upper limit, for example:
1114 nigel 75 .sp
1115 nigel 63 (a?)*
1116 nigel 75 .sp
1117 nigel 63 Earlier versions of Perl and PCRE used to give an error at compile time for
1118     such patterns. However, because there are cases where this can be useful, such
1119     patterns are now accepted, but if any repetition of the subpattern does in fact
1120     match no characters, the loop is forcibly broken.
1121 nigel 75 .P
1122 nigel 63 By default, the quantifiers are "greedy", that is, they match as much as
1123     possible (up to the maximum number of permitted times), without causing the
1124     rest of the pattern to fail. The classic example of where this gives problems
1125 nigel 75 is in trying to match comments in C programs. These appear between /* and */
1126     and within the comment, individual * and / characters may appear. An attempt to
1127     match C comments by applying the pattern
1128     .sp
1129     /\e*.*\e*/
1130     .sp
1131 nigel 63 to the string
1132 nigel 75 .sp
1133     /* first comment */ not comment /* second comment */
1134     .sp
1135 nigel 63 fails, because it matches the entire string owing to the greediness of the .*
1136     item.
1137 nigel 75 .P
1138 nigel 63 However, if a quantifier is followed by a question mark, it ceases to be
1139     greedy, and instead matches the minimum number of times possible, so the
1140     pattern
1141 nigel 75 .sp
1142     /\e*.*?\e*/
1143     .sp
1144 nigel 63 does the right thing with the C comments. The meaning of the various
1145     quantifiers is not otherwise changed, just the preferred number of matches.
1146     Do not confuse this use of question mark with its use as a quantifier in its
1147     own right. Because it has two uses, it can sometimes appear doubled, as in
1148 nigel 75 .sp
1149     \ed??\ed
1150     .sp
1151 nigel 63 which matches one digit by preference, but can match two if that is the only
1152     way the rest of the pattern matches.
1153 nigel 75 .P
1154 nigel 93 If the PCRE_UNGREEDY option is set (an option that is not available in Perl),
1155 nigel 63 the quantifiers are not greedy by default, but individual ones can be made
1156     greedy by following them with a question mark. In other words, it inverts the
1157     default behaviour.
1158 nigel 75 .P
1159 nigel 63 When a parenthesized subpattern is quantified with a minimum repeat count that
1160 nigel 75 is greater than 1 or with a limited maximum, more memory is required for the
1161 nigel 63 compiled pattern, in proportion to the size of the minimum or maximum.
1162 nigel 75 .P
1163 nigel 63 If a pattern starts with .* or .{0,} and the PCRE_DOTALL option (equivalent
1164 nigel 93 to Perl's /s) is set, thus allowing the dot to match newlines, the pattern is
1165 nigel 63 implicitly anchored, because whatever follows will be tried against every
1166     character position in the subject string, so there is no point in retrying the
1167     overall match at any position after the first. PCRE normally treats such a
1168 nigel 75 pattern as though it were preceded by \eA.
1169     .P
1170 nigel 63 In cases where it is known that the subject string contains no newlines, it is
1171     worth setting PCRE_DOTALL in order to obtain this optimization, or
1172     alternatively using ^ to indicate anchoring explicitly.
1173 nigel 75 .P
1174 nigel 63 However, there is one situation where the optimization cannot be used. When .*
1175     is inside capturing parentheses that are the subject of a backreference
1176 nigel 93 elsewhere in the pattern, a match at the start may fail where a later one
1177     succeeds. Consider, for example:
1178 nigel 75 .sp
1179     (.*)abc\e1
1180     .sp
1181 nigel 63 If the subject is "xyz123abc123" the match point is the fourth character. For
1182     this reason, such a pattern is not implicitly anchored.
1183 nigel 75 .P
1184 nigel 63 When a capturing subpattern is repeated, the value captured is the substring
1185     that matched the final iteration. For example, after
1186 nigel 75 .sp
1187     (tweedle[dume]{3}\es*)+
1188     .sp
1189 nigel 63 has matched "tweedledum tweedledee" the value of the captured substring is
1190     "tweedledee". However, if there are nested capturing subpatterns, the
1191     corresponding captured values may have been set in previous iterations. For
1192     example, after
1193 nigel 75 .sp
1194 nigel 63 /(a|(b))+/
1195 nigel 75 .sp
1196 nigel 63 matches "aba" the value of the second captured substring is "b".
1197 nigel 75 .
1198     .
1199     .\" HTML <a name="atomicgroup"></a>
1200     .SH "ATOMIC GROUPING AND POSSESSIVE QUANTIFIERS"
1201 nigel 63 .rs
1202     .sp
1203 nigel 93 With both maximizing ("greedy") and minimizing ("ungreedy" or "lazy")
1204     repetition, failure of what follows normally causes the repeated item to be
1205     re-evaluated to see if a different number of repeats allows the rest of the
1206     pattern to match. Sometimes it is useful to prevent this, either to change the
1207     nature of the match, or to cause it fail earlier than it otherwise might, when
1208     the author of the pattern knows there is no point in carrying on.
1209 nigel 75 .P
1210     Consider, for example, the pattern \ed+foo when applied to the subject line
1211     .sp
1212 nigel 63 123456bar
1213 nigel 75 .sp
1214 nigel 63 After matching all 6 digits and then failing to match "foo", the normal
1215 nigel 75 action of the matcher is to try again with only 5 digits matching the \ed+
1216 nigel 63 item, and then with 4, and so on, before ultimately failing. "Atomic grouping"
1217     (a term taken from Jeffrey Friedl's book) provides the means for specifying
1218     that once a subpattern has matched, it is not to be re-evaluated in this way.
1219 nigel 75 .P
1220 nigel 93 If we use atomic grouping for the previous example, the matcher gives up
1221 nigel 63 immediately on failing to match "foo" the first time. The notation is a kind of
1222     special parenthesis, starting with (?> as in this example:
1223 nigel 75 .sp
1224     (?>\ed+)foo
1225     .sp
1226 nigel 63 This kind of parenthesis "locks up" the part of the pattern it contains once
1227     it has matched, and a failure further into the pattern is prevented from
1228     backtracking into it. Backtracking past it to previous items, however, works as
1229     normal.
1230 nigel 75 .P
1231 nigel 63 An alternative description is that a subpattern of this type matches the string
1232     of characters that an identical standalone pattern would match, if anchored at
1233     the current point in the subject string.
1234 nigel 75 .P
1235 nigel 63 Atomic grouping subpatterns are not capturing subpatterns. Simple cases such as
1236     the above example can be thought of as a maximizing repeat that must swallow
1237 nigel 75 everything it can. So, while both \ed+ and \ed+? are prepared to adjust the
1238 nigel 63 number of digits they match in order to make the rest of the pattern match,
1239 nigel 75 (?>\ed+) can only match an entire sequence of digits.
1240     .P
1241 nigel 63 Atomic groups in general can of course contain arbitrarily complicated
1242     subpatterns, and can be nested. However, when the subpattern for an atomic
1243     group is just a single repeated item, as in the example above, a simpler
1244     notation, called a "possessive quantifier" can be used. This consists of an
1245     additional + character following a quantifier. Using this notation, the
1246     previous example can be rewritten as
1247 nigel 75 .sp
1248     \ed++foo
1249     .sp
1250 nigel 63 Possessive quantifiers are always greedy; the setting of the PCRE_UNGREEDY
1251     option is ignored. They are a convenient notation for the simpler forms of
1252 nigel 93 atomic group. However, there is no difference in the meaning of a possessive
1253     quantifier and the equivalent atomic group, though there may be a performance
1254     difference; possessive quantifiers should be slightly faster.
1255 nigel 75 .P
1256 nigel 93 The possessive quantifier syntax is an extension to the Perl 5.8 syntax.
1257     Jeffrey Friedl originated the idea (and the name) in the first edition of his
1258     book. Mike McCloskey liked it, so implemented it when he built Sun's Java
1259     package, and PCRE copied it from there. It ultimately found its way into Perl
1260     at release 5.10.
1261 nigel 75 .P
1262 nigel 93 PCRE has an optimization that automatically "possessifies" certain simple
1263     pattern constructs. For example, the sequence A+B is treated as A++B because
1264     there is no point in backtracking into a sequence of A's when B must follow.
1265     .P
1266 nigel 63 When a pattern contains an unlimited repeat inside a subpattern that can itself
1267     be repeated an unlimited number of times, the use of an atomic group is the
1268     only way to avoid some failing matches taking a very long time indeed. The
1269     pattern
1270 nigel 75 .sp
1271     (\eD+|<\ed+>)*[!?]
1272     .sp
1273 nigel 63 matches an unlimited number of substrings that either consist of non-digits, or
1274     digits enclosed in <>, followed by either ! or ?. When it matches, it runs
1275     quickly. However, if it is applied to
1276 nigel 75 .sp
1277 nigel 63 aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
1278 nigel 75 .sp
1279 nigel 63 it takes a long time before reporting failure. This is because the string can
1280 nigel 75 be divided between the internal \eD+ repeat and the external * repeat in a
1281     large number of ways, and all have to be tried. (The example uses [!?] rather
1282     than a single character at the end, because both PCRE and Perl have an
1283     optimization that allows for fast failure when a single character is used. They
1284     remember the last single character that is required for a match, and fail early
1285     if it is not present in the string.) If the pattern is changed so that it uses
1286     an atomic group, like this:
1287     .sp
1288     ((?>\eD+)|<\ed+>)*[!?]
1289     .sp
1290 nigel 63 sequences of non-digits cannot be broken, and failure happens quickly.
1291 nigel 75 .
1292     .
1293     .\" HTML <a name="backreferences"></a>
1294     .SH "BACK REFERENCES"
1295 nigel 63 .rs
1296     .sp
1297     Outside a character class, a backslash followed by a digit greater than 0 (and
1298     possibly further digits) is a back reference to a capturing subpattern earlier
1299     (that is, to its left) in the pattern, provided there have been that many
1300     previous capturing left parentheses.
1301 nigel 75 .P
1302 nigel 63 However, if the decimal number following the backslash is less than 10, it is
1303     always taken as a back reference, and causes an error only if there are not
1304     that many capturing left parentheses in the entire pattern. In other words, the
1305     parentheses that are referenced need not be to the left of the reference for
1306 nigel 91 numbers less than 10. A "forward back reference" of this type can make sense
1307     when a repetition is involved and the subpattern to the right has participated
1308     in an earlier iteration.
1309     .P
1310 nigel 93 It is not possible to have a numerical "forward back reference" to a subpattern
1311     whose number is 10 or more using this syntax because a sequence such as \e50 is
1312     interpreted as a character defined in octal. See the subsection entitled
1313 nigel 91 "Non-printing characters"
1314 nigel 75 .\" HTML <a href="#digitsafterbackslash">
1315     .\" </a>
1316     above
1317     .\"
1318 nigel 93 for further details of the handling of digits following a backslash. There is
1319     no such problem when named parentheses are used. A back reference to any
1320     subpattern is possible using named parentheses (see below).
1321 nigel 75 .P
1322 nigel 93 Another way of avoiding the ambiguity inherent in the use of digits following a
1323     backslash is to use the \eg escape sequence, which is a feature introduced in
1324     Perl 5.10. This escape must be followed by a positive or a negative number,
1325     optionally enclosed in braces. These examples are all identical:
1326     .sp
1327     (ring), \e1
1328     (ring), \eg1
1329     (ring), \eg{1}
1330     .sp
1331     A positive number specifies an absolute reference without the ambiguity that is
1332     present in the older syntax. It is also useful when literal digits follow the
1333     reference. A negative number is a relative reference. Consider this example:
1334     .sp
1335     (abc(def)ghi)\eg{-1}
1336     .sp
1337     The sequence \eg{-1} is a reference to the most recently started capturing
1338     subpattern before \eg, that is, is it equivalent to \e2. Similarly, \eg{-2}
1339     would be equivalent to \e1. The use of relative references can be helpful in
1340     long patterns, and also in patterns that are created by joining together
1341     fragments that contain references within themselves.
1342     .P
1343 nigel 63 A back reference matches whatever actually matched the capturing subpattern in
1344     the current subject string, rather than anything matching the subpattern
1345     itself (see
1346     .\" HTML <a href="#subpatternsassubroutines">
1347     .\" </a>
1348     "Subpatterns as subroutines"
1349     .\"
1350     below for a way of doing that). So the pattern
1351 nigel 75 .sp
1352     (sens|respons)e and \e1ibility
1353     .sp
1354 nigel 63 matches "sense and sensibility" and "response and responsibility", but not
1355     "sense and responsibility". If caseful matching is in force at the time of the
1356     back reference, the case of letters is relevant. For example,
1357 nigel 75 .sp
1358     ((?i)rah)\es+\e1
1359     .sp
1360 nigel 63 matches "rah rah" and "RAH RAH", but not "RAH rah", even though the original
1361     capturing subpattern is matched caselessly.
1362 nigel 75 .P
1363 ph10 171 There are several different ways of writing back references to named
1364     subpatterns. The .NET syntax \ek{name} and the Perl syntax \ek<name> or
1365     \ek'name' are supported, as is the Python syntax (?P=name). Perl 5.10's unified
1366     back reference syntax, in which \eg can be used for both numeric and named
1367     references, is also supported. We could rewrite the above example in any of
1368 nigel 93 the following ways:
1369 nigel 75 .sp
1370 nigel 93 (?<p1>(?i)rah)\es+\ek<p1>
1371 ph10 171 (?'p1'(?i)rah)\es+\ek{p1}
1372 nigel 91 (?P<p1>(?i)rah)\es+(?P=p1)
1373 ph10 171 (?<p1>(?i)rah)\es+\eg{p1}
1374 nigel 75 .sp
1375 nigel 91 A subpattern that is referenced by name may appear in the pattern before or
1376     after the reference.
1377     .P
1378 nigel 63 There may be more than one back reference to the same subpattern. If a
1379     subpattern has not actually been used in a particular match, any back
1380     references to it always fail. For example, the pattern
1381 nigel 75 .sp
1382     (a|(bc))\e2
1383     .sp
1384 nigel 63 always fails if it starts to match "a" rather than "bc". Because there may be
1385     many capturing parentheses in a pattern, all digits following the backslash are
1386     taken as part of a potential back reference number. If the pattern continues
1387     with a digit character, some delimiter must be used to terminate the back
1388     reference. If the PCRE_EXTENDED option is set, this can be whitespace.
1389 nigel 75 Otherwise an empty comment (see
1390     .\" HTML <a href="#comments">
1391     .\" </a>
1392     "Comments"
1393     .\"
1394     below) can be used.
1395     .P
1396 nigel 63 A back reference that occurs inside the parentheses to which it refers fails
1397 nigel 75 when the subpattern is first used, so, for example, (a\e1) never matches.
1398 nigel 63 However, such references can be useful inside repeated subpatterns. For
1399     example, the pattern
1400 nigel 75 .sp
1401     (a|b\e1)+
1402     .sp
1403 nigel 63 matches any number of "a"s and also "aba", "ababbaa" etc. At each iteration of
1404     the subpattern, the back reference matches the character string corresponding
1405     to the previous iteration. In order for this to work, the pattern must be such
1406     that the first iteration does not need to match the back reference. This can be
1407     done using alternation, as in the example above, or by a quantifier with a
1408     minimum of zero.
1409 nigel 75 .
1410     .
1411     .\" HTML <a name="bigassertions"></a>
1412 nigel 63 .SH ASSERTIONS
1413     .rs
1414     .sp
1415     An assertion is a test on the characters following or preceding the current
1416     matching point that does not actually consume any characters. The simple
1417 nigel 75 assertions coded as \eb, \eB, \eA, \eG, \eZ, \ez, ^ and $ are described
1418     .\" HTML <a href="#smallassertions">
1419     .\" </a>
1420     above.
1421     .\"
1422     .P
1423 nigel 63 More complicated assertions are coded as subpatterns. There are two kinds:
1424     those that look ahead of the current position in the subject string, and those
1425 nigel 75 that look behind it. An assertion subpattern is matched in the normal way,
1426     except that it does not cause the current matching position to be changed.
1427     .P
1428     Assertion subpatterns are not capturing subpatterns, and may not be repeated,
1429     because it makes no sense to assert the same thing several times. If any kind
1430     of assertion contains capturing subpatterns within it, these are counted for
1431     the purposes of numbering the capturing subpatterns in the whole pattern.
1432     However, substring capturing is carried out only for positive assertions,
1433     because it does not make sense for negative assertions.
1434     .
1435     .
1436     .SS "Lookahead assertions"
1437     .rs
1438     .sp
1439 nigel 91 Lookahead assertions start with (?= for positive assertions and (?! for
1440     negative assertions. For example,
1441 nigel 75 .sp
1442     \ew+(?=;)
1443     .sp
1444 nigel 63 matches a word followed by a semicolon, but does not include the semicolon in
1445     the match, and
1446 nigel 75 .sp
1447 nigel 63 foo(?!bar)
1448 nigel 75 .sp
1449 nigel 63 matches any occurrence of "foo" that is not followed by "bar". Note that the
1450     apparently similar pattern
1451 nigel 75 .sp
1452 nigel 63 (?!foo)bar
1453 nigel 75 .sp
1454 nigel 63 does not find an occurrence of "bar" that is preceded by something other than
1455     "foo"; it finds any occurrence of "bar" whatsoever, because the assertion
1456     (?!foo) is always true when the next three characters are "bar". A
1457 nigel 75 lookbehind assertion is needed to achieve the other effect.
1458     .P
1459 nigel 63 If you want to force a matching failure at some point in a pattern, the most
1460     convenient way to do it is with (?!) because an empty string always matches, so
1461     an assertion that requires there not to be an empty string must always fail.
1462 nigel 75 .
1463     .
1464     .\" HTML <a name="lookbehind"></a>
1465     .SS "Lookbehind assertions"
1466     .rs
1467     .sp
1468 nigel 63 Lookbehind assertions start with (?<= for positive assertions and (?<! for
1469     negative assertions. For example,
1470 nigel 75 .sp
1471 nigel 63 (?<!foo)bar
1472 nigel 75 .sp
1473 nigel 63 does find an occurrence of "bar" that is not preceded by "foo". The contents of
1474     a lookbehind assertion are restricted such that all the strings it matches must
1475 nigel 91 have a fixed length. However, if there are several top-level alternatives, they
1476     do not all have to have the same fixed length. Thus
1477 nigel 75 .sp
1478 nigel 63 (?<=bullock|donkey)
1479 nigel 75 .sp
1480 nigel 63 is permitted, but
1481 nigel 75 .sp
1482 nigel 63 (?<!dogs?|cats?)
1483 nigel 75 .sp
1484 nigel 63 causes an error at compile time. Branches that match different length strings
1485     are permitted only at the top level of a lookbehind assertion. This is an
1486     extension compared with Perl (at least for 5.8), which requires all branches to
1487     match the same length of string. An assertion such as
1488 nigel 75 .sp
1489 nigel 63 (?<=ab(c|de))
1490 nigel 75 .sp
1491 nigel 63 is not permitted, because its single top-level branch can match two different
1492     lengths, but it is acceptable if rewritten to use two top-level branches:
1493 nigel 75 .sp
1494 nigel 63 (?<=abc|abde)
1495 nigel 75 .sp
1496 ph10 172 In some cases, the Perl 5.10 escape sequence \eK
1497 ph10 168 .\" HTML <a href="#resetmatchstart">
1498     .\" </a>
1499     (see above)
1500     .\"
1501     can be used instead of a lookbehind assertion; this is not restricted to a
1502     fixed-length.
1503     .P
1504 nigel 63 The implementation of lookbehind assertions is, for each alternative, to
1505 nigel 93 temporarily move the current position back by the fixed length and then try to
1506 nigel 63 match. If there are insufficient characters before the current position, the
1507 nigel 93 assertion fails.
1508 nigel 75 .P
1509     PCRE does not allow the \eC escape (which matches a single byte in UTF-8 mode)
1510 nigel 63 to appear in lookbehind assertions, because it makes it impossible to calculate
1511 nigel 93 the length of the lookbehind. The \eX and \eR escapes, which can match
1512     different numbers of bytes, are also not permitted.
1513 nigel 75 .P
1514 nigel 93 Possessive quantifiers can be used in conjunction with lookbehind assertions to
1515     specify efficient matching at the end of the subject string. Consider a simple
1516     pattern such as
1517 nigel 75 .sp
1518 nigel 63 abcd$
1519 nigel 75 .sp
1520 nigel 63 when applied to a long string that does not match. Because matching proceeds
1521     from left to right, PCRE will look for each "a" in the subject and then see if
1522     what follows matches the rest of the pattern. If the pattern is specified as
1523 nigel 75 .sp
1524 nigel 63 ^.*abcd$
1525 nigel 75 .sp
1526 nigel 63 the initial .* matches the entire string at first, but when this fails (because
1527     there is no following "a"), it backtracks to match all but the last character,
1528     then all but the last two characters, and so on. Once again the search for "a"
1529     covers the entire string, from right to left, so we are no better off. However,
1530     if the pattern is written as
1531 nigel 75 .sp
1532 nigel 63 ^.*+(?<=abcd)
1533 nigel 75 .sp
1534 nigel 93 there can be no backtracking for the .*+ item; it can match only the entire
1535 nigel 63 string. The subsequent lookbehind assertion does a single test on the last four
1536     characters. If it fails, the match fails immediately. For long strings, this
1537     approach makes a significant difference to the processing time.
1538 nigel 75 .
1539     .
1540     .SS "Using multiple assertions"
1541     .rs
1542     .sp
1543 nigel 63 Several assertions (of any sort) may occur in succession. For example,
1544 nigel 75 .sp
1545     (?<=\ed{3})(?<!999)foo
1546     .sp
1547 nigel 63 matches "foo" preceded by three digits that are not "999". Notice that each of
1548     the assertions is applied independently at the same point in the subject
1549     string. First there is a check that the previous three characters are all
1550     digits, and then there is a check that the same three characters are not "999".
1551 nigel 75 This pattern does \fInot\fP match "foo" preceded by six characters, the first
1552 nigel 63 of which are digits and the last three of which are not "999". For example, it
1553     doesn't match "123abcfoo". A pattern to do that is
1554 nigel 75 .sp
1555     (?<=\ed{3}...)(?<!999)foo
1556     .sp
1557 nigel 63 This time the first assertion looks at the preceding six characters, checking
1558     that the first three are digits, and then the second assertion checks that the
1559     preceding three characters are not "999".
1560 nigel 75 .P
1561 nigel 63 Assertions can be nested in any combination. For example,
1562 nigel 75 .sp
1563 nigel 63 (?<=(?<!foo)bar)baz
1564 nigel 75 .sp
1565 nigel 63 matches an occurrence of "baz" that is preceded by "bar" which in turn is not
1566     preceded by "foo", while
1567 nigel 75 .sp
1568     (?<=\ed{3}(?!999)...)foo
1569     .sp
1570     is another pattern that matches "foo" preceded by three digits and any three
1571 nigel 63 characters that are not "999".
1572 nigel 75 .
1573     .
1574 nigel 91 .\" HTML <a name="conditions"></a>
1575 nigel 75 .SH "CONDITIONAL SUBPATTERNS"
1576 nigel 63 .rs
1577     .sp
1578     It is possible to cause the matching process to obey a subpattern
1579     conditionally or to choose between two alternative subpatterns, depending on
1580     the result of an assertion, or whether a previous capturing subpattern matched
1581     or not. The two possible forms of conditional subpattern are
1582 nigel 75 .sp
1583 nigel 63 (?(condition)yes-pattern)
1584     (?(condition)yes-pattern|no-pattern)
1585 nigel 75 .sp
1586 nigel 63 If the condition is satisfied, the yes-pattern is used; otherwise the
1587     no-pattern (if present) is used. If there are more than two alternatives in the
1588     subpattern, a compile-time error occurs.
1589 nigel 75 .P
1590 nigel 93 There are four kinds of condition: references to subpatterns, references to
1591     recursion, a pseudo-condition called DEFINE, and assertions.
1592     .
1593     .SS "Checking for a used subpattern by number"
1594     .rs
1595     .sp
1596     If the text between the parentheses consists of a sequence of digits, the
1597     condition is true if the capturing subpattern of that number has previously
1598 ph10 172 matched. An alternative notation is to precede the digits with a plus or minus
1599 ph10 167 sign. In this case, the subpattern number is relative rather than absolute.
1600 ph10 172 The most recently opened parentheses can be referenced by (?(-1), the next most
1601 ph10 167 recent by (?(-2), and so on. In looping constructs it can also make sense to
1602     refer to subsequent groups with constructs such as (?(+2).
1603 nigel 91 .P
1604     Consider the following pattern, which contains non-significant white space to
1605     make it more readable (assume the PCRE_EXTENDED option) and to divide it into
1606     three parts for ease of discussion:
1607 nigel 75 .sp
1608     ( \e( )? [^()]+ (?(1) \e) )
1609     .sp
1610 nigel 63 The first part matches an optional opening parenthesis, and if that
1611     character is present, sets it as the first captured substring. The second part
1612     matches one or more characters that are not parentheses. The third part is a
1613     conditional subpattern that tests whether the first set of parentheses matched
1614     or not. If they did, that is, if subject started with an opening parenthesis,
1615     the condition is true, and so the yes-pattern is executed and a closing
1616     parenthesis is required. Otherwise, since no-pattern is not present, the
1617     subpattern matches nothing. In other words, this pattern matches a sequence of
1618 nigel 93 non-parentheses, optionally enclosed in parentheses.
1619 ph10 167 .P
1620 ph10 172 If you were embedding this pattern in a larger one, you could use a relative
1621 ph10 167 reference:
1622     .sp
1623     ...other stuff... ( \e( )? [^()]+ (?(-1) \e) ) ...
1624     .sp
1625     This makes the fragment independent of the parentheses in the larger pattern.
1626 nigel 93 .
1627     .SS "Checking for a used subpattern by name"
1628     .rs
1629 nigel 91 .sp
1630 nigel 93 Perl uses the syntax (?(<name>)...) or (?('name')...) to test for a used
1631     subpattern by name. For compatibility with earlier versions of PCRE, which had
1632     this facility before Perl, the syntax (?(name)...) is also recognized. However,
1633     there is a possible ambiguity with this syntax, because subpattern names may
1634     consist entirely of digits. PCRE looks first for a named subpattern; if it
1635     cannot find one and the name consists entirely of digits, PCRE looks for a
1636     subpattern of that number, which must be greater than zero. Using subpattern
1637     names that consist entirely of digits is not recommended.
1638     .P
1639     Rewriting the above example to use a named subpattern gives this:
1640 nigel 91 .sp
1641 nigel 93 (?<OPEN> \e( )? [^()]+ (?(<OPEN>) \e) )
1642     .sp
1643     .
1644     .SS "Checking for pattern recursion"
1645     .rs
1646     .sp
1647 nigel 91 If the condition is the string (R), and there is no subpattern with the name R,
1648 nigel 93 the condition is true if a recursive call to the whole pattern or any
1649     subpattern has been made. If digits or a name preceded by ampersand follow the
1650     letter R, for example:
1651     .sp
1652     (?(R3)...) or (?(R&name)...)
1653     .sp
1654     the condition is true if the most recent recursion is into the subpattern whose
1655     number or name is given. This condition does not check the entire recursion
1656     stack.
1657 nigel 75 .P
1658 nigel 93 At "top level", all these recursion test conditions are false. Recursive
1659     patterns are described below.
1660     .
1661     .SS "Defining subpatterns for use by reference only"
1662     .rs
1663     .sp
1664     If the condition is the string (DEFINE), and there is no subpattern with the
1665     name DEFINE, the condition is always false. In this case, there may be only one
1666     alternative in the subpattern. It is always skipped if control reaches this
1667     point in the pattern; the idea of DEFINE is that it can be used to define
1668     "subroutines" that can be referenced from elsewhere. (The use of "subroutines"
1669     is described below.) For example, a pattern to match an IPv4 address could be
1670     written like this (ignore whitespace and line breaks):
1671     .sp
1672     (?(DEFINE) (?<byte> 2[0-4]\ed | 25[0-5] | 1\ed\ed | [1-9]?\ed) )
1673     \eb (?&byte) (\e.(?&byte)){3} \eb
1674     .sp
1675     The first part of the pattern is a DEFINE group inside which a another group
1676     named "byte" is defined. This matches an individual component of an IPv4
1677     address (a number less than 256). When matching takes place, this part of the
1678     pattern is skipped because DEFINE acts like a false condition.
1679     .P
1680     The rest of the pattern uses references to the named group to match the four
1681     dot-separated components of an IPv4 address, insisting on a word boundary at
1682     each end.
1683     .
1684     .SS "Assertion conditions"
1685     .rs
1686     .sp
1687     If the condition is not in any of the above formats, it must be an assertion.
1688 nigel 63 This may be a positive or negative lookahead or lookbehind assertion. Consider
1689     this pattern, again containing non-significant white space, and with the two
1690     alternatives on the second line:
1691 nigel 75 .sp
1692 nigel 63 (?(?=[^a-z]*[a-z])
1693 nigel 75 \ed{2}-[a-z]{3}-\ed{2} | \ed{2}-\ed{2}-\ed{2} )
1694     .sp
1695 nigel 63 The condition is a positive lookahead assertion that matches an optional
1696     sequence of non-letters followed by a letter. In other words, it tests for the
1697     presence of at least one letter in the subject. If a letter is found, the
1698     subject is matched against the first alternative; otherwise it is matched
1699     against the second. This pattern matches strings in one of the two forms
1700     dd-aaa-dd or dd-dd-dd, where aaa are letters and dd are digits.
1701 nigel 75 .
1702     .
1703     .\" HTML <a name="comments"></a>
1704 nigel 63 .SH COMMENTS
1705     .rs
1706     .sp
1707 nigel 75 The sequence (?# marks the start of a comment that continues up to the next
1708 nigel 63 closing parenthesis. Nested parentheses are not permitted. The characters
1709     that make up a comment play no part in the pattern matching at all.
1710 nigel 75 .P
1711 nigel 63 If the PCRE_EXTENDED option is set, an unescaped # character outside a
1712 nigel 91 character class introduces a comment that continues to immediately after the
1713     next newline in the pattern.
1714 nigel 75 .
1715     .
1716 nigel 91 .\" HTML <a name="recursion"></a>
1717 nigel 75 .SH "RECURSIVE PATTERNS"
1718 nigel 63 .rs
1719     .sp
1720     Consider the problem of matching a string in parentheses, allowing for
1721     unlimited nested parentheses. Without the use of recursion, the best that can
1722     be done is to use a pattern that matches up to some fixed depth of nesting. It
1723 nigel 93 is not possible to handle an arbitrary nesting depth.
1724     .P
1725     For some time, Perl has provided a facility that allows regular expressions to
1726     recurse (amongst other things). It does this by interpolating Perl code in the
1727     expression at run time, and the code can refer to the expression itself. A Perl
1728     pattern using code interpolation to solve the parentheses problem can be
1729     created like this:
1730 nigel 75 .sp
1731     $re = qr{\e( (?: (?>[^()]+) | (?p{$re}) )* \e)}x;
1732     .sp
1733 nigel 63 The (?p{...}) item interpolates Perl code at run time, and in this case refers
1734 nigel 93 recursively to the pattern in which it appears.
1735 nigel 75 .P
1736 nigel 93 Obviously, PCRE cannot support the interpolation of Perl code. Instead, it
1737     supports special syntax for recursion of the entire pattern, and also for
1738     individual subpattern recursion. After its introduction in PCRE and Python,
1739     this kind of recursion was introduced into Perl at release 5.10.
1740 nigel 75 .P
1741 nigel 93 A special item that consists of (? followed by a number greater than zero and a
1742     closing parenthesis is a recursive call of the subpattern of the given number,
1743     provided that it occurs inside that subpattern. (If not, it is a "subroutine"
1744     call, which is described in the next section.) The special item (?R) or (?0) is
1745     a recursive call of the entire regular expression.
1746 nigel 87 .P
1747 nigel 93 In PCRE (like Python, but unlike Perl), a recursive subpattern call is always
1748     treated as an atomic group. That is, once it has matched some of the subject
1749     string, it is never re-entered, even if it contains untried alternatives and
1750     there is a subsequent matching failure.
1751     .P
1752 nigel 87 This PCRE pattern solves the nested parentheses problem (assume the
1753     PCRE_EXTENDED option is set so that white space is ignored):
1754 nigel 75 .sp
1755     \e( ( (?>[^()]+) | (?R) )* \e)
1756     .sp
1757 nigel 63 First it matches an opening parenthesis. Then it matches any number of
1758     substrings which can either be a sequence of non-parentheses, or a recursive
1759 nigel 87 match of the pattern itself (that is, a correctly parenthesized substring).
1760 nigel 63 Finally there is a closing parenthesis.
1761 nigel 75 .P
1762 nigel 63 If this were part of a larger pattern, you would not want to recurse the entire
1763     pattern, so instead you could use this:
1764 nigel 75 .sp
1765     ( \e( ( (?>[^()]+) | (?1) )* \e) )
1766     .sp
1767 nigel 63 We have put the pattern into parentheses, and caused the recursion to refer to
1768 ph10 172 them instead of the whole pattern.
1769 ph10 166 .P
1770     In a larger pattern, keeping track of parenthesis numbers can be tricky. This
1771     is made easier by the use of relative references. (A Perl 5.10 feature.)
1772     Instead of (?1) in the pattern above you can write (?-2) to refer to the second
1773     most recently opened parentheses preceding the recursion. In other words, a
1774     negative number counts capturing parentheses leftwards from the point at which
1775     it is encountered.
1776     .P
1777     It is also possible to refer to subsequently opened parentheses, by writing
1778     references such as (?+2). However, these cannot be recursive because the
1779     reference is not inside the parentheses that are referenced. They are always
1780     "subroutine" calls, as described in the next section.
1781     .P
1782     An alternative approach is to use named parentheses instead. The Perl syntax
1783     for this is (?&name); PCRE's earlier syntax (?P>name) is also supported. We
1784     could rewrite the above example as follows:
1785 nigel 75 .sp
1786 nigel 93 (?<pn> \e( ( (?>[^()]+) | (?&pn) )* \e) )
1787 nigel 75 .sp
1788 nigel 93 If there is more than one subpattern with the same name, the earliest one is
1789 ph10 172 used.
1790 ph10 166 .P
1791     This particular example pattern that we have been looking at contains nested
1792     unlimited repeats, and so the use of atomic grouping for matching strings of
1793     non-parentheses is important when applying the pattern to strings that do not
1794     match. For example, when this pattern is applied to
1795 nigel 75 .sp
1796 nigel 63 (aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa()
1797 nigel 75 .sp
1798 nigel 63 it yields "no match" quickly. However, if atomic grouping is not used,
1799     the match runs for a very long time indeed because there are so many different
1800     ways the + and * repeats can carve up the subject, and all have to be tested
1801     before failure can be reported.
1802 nigel 75 .P
1803 nigel 63 At the end of a match, the values set for any capturing subpatterns are those
1804     from the outermost level of the recursion at which the subpattern value is set.
1805     If you want to obtain intermediate values, a callout function can be used (see
1806 nigel 93 below and the
1807 nigel 63 .\" HREF
1808 nigel 75 \fBpcrecallout\fP
1809 nigel 63 .\"
1810     documentation). If the pattern above is matched against
1811 nigel 75 .sp
1812 nigel 63 (ab(cd)ef)
1813 nigel 75 .sp
1814 nigel 63 the value for the capturing parentheses is "ef", which is the last value taken
1815     on at the top level. If additional parentheses are added, giving
1816 nigel 75 .sp
1817     \e( ( ( (?>[^()]+) | (?R) )* ) \e)
1818 nigel 63 ^ ^
1819     ^ ^
1820 nigel 75 .sp
1821 nigel 63 the string they capture is "ab(cd)ef", the contents of the top level
1822     parentheses. If there are more than 15 capturing parentheses in a pattern, PCRE
1823     has to obtain extra memory to store data during a recursion, which it does by
1824 nigel 75 using \fBpcre_malloc\fP, freeing it via \fBpcre_free\fP afterwards. If no
1825 nigel 63 memory can be obtained, the match fails with the PCRE_ERROR_NOMEMORY error.
1826 nigel 75 .P
1827 nigel 63 Do not confuse the (?R) item with the condition (R), which tests for recursion.
1828     Consider this pattern, which matches text in angle brackets, allowing for
1829     arbitrary nesting. Only digits are allowed in nested brackets (that is, when
1830     recursing), whereas any characters are permitted at the outer level.
1831 nigel 75 .sp
1832     < (?: (?(R) \ed++ | [^<>]*+) | (?R)) * >
1833     .sp
1834 nigel 63 In this pattern, (?(R) is the start of a conditional subpattern, with two
1835     different alternatives for the recursive and non-recursive cases. The (?R) item
1836     is the actual recursive call.
1837 nigel 75 .
1838     .
1839 nigel 63 .\" HTML <a name="subpatternsassubroutines"></a>
1840 nigel 75 .SH "SUBPATTERNS AS SUBROUTINES"
1841 nigel 63 .rs
1842     .sp
1843     If the syntax for a recursive subpattern reference (either by number or by
1844     name) is used outside the parentheses to which it refers, it operates like a
1845 nigel 93 subroutine in a programming language. The "called" subpattern may be defined
1846 ph10 166 before or after the reference. A numbered reference can be absolute or
1847     relative, as in these examples:
1848 nigel 75 .sp
1849 ph10 166 (...(absolute)...)...(?2)...
1850     (...(relative)...)...(?-1)...
1851 ph10 172 (...(?+1)...(relative)...
1852 ph10 166 .sp
1853     An earlier example pointed out that the pattern
1854     .sp
1855 nigel 75 (sens|respons)e and \e1ibility
1856     .sp
1857 nigel 63 matches "sense and sensibility" and "response and responsibility", but not
1858     "sense and responsibility". If instead the pattern
1859 nigel 75 .sp
1860 nigel 63 (sens|respons)e and (?1)ibility
1861 nigel 75 .sp
1862 nigel 63 is used, it does match "sense and responsibility" as well as the other two
1863 nigel 93 strings. Another example is given in the discussion of DEFINE above.
1864 nigel 87 .P
1865     Like recursive subpatterns, a "subroutine" call is always treated as an atomic
1866     group. That is, once it has matched some of the subject string, it is never
1867     re-entered, even if it contains untried alternatives and there is a subsequent
1868     matching failure.
1869 nigel 93 .P
1870     When a subpattern is used as a subroutine, processing options such as
1871     case-independence are fixed when the subpattern is defined. They cannot be
1872     changed for different calls. For example, consider this pattern:
1873     .sp
1874 ph10 166 (abc)(?i:(?-1))
1875 nigel 93 .sp
1876     It matches "abcabc". It does not match "abcABC" because the change of
1877     processing option does not affect the called subpattern.
1878 nigel 75 .
1879     .
1880 nigel 63 .SH CALLOUTS
1881     .rs
1882     .sp
1883     Perl has a feature whereby using the sequence (?{...}) causes arbitrary Perl
1884     code to be obeyed in the middle of matching a regular expression. This makes it
1885     possible, amongst other things, to extract different substrings that match the
1886     same pair of parentheses when there is a repetition.
1887 nigel 75 .P
1888 nigel 63 PCRE provides a similar feature, but of course it cannot obey arbitrary Perl
1889     code. The feature is called "callout". The caller of PCRE provides an external
1890 nigel 75 function by putting its entry point in the global variable \fIpcre_callout\fP.
1891 nigel 63 By default, this variable contains NULL, which disables all calling out.
1892 nigel 75 .P
1893 nigel 63 Within a regular expression, (?C) indicates the points at which the external
1894     function is to be called. If you want to identify different callout points, you
1895     can put a number less than 256 after the letter C. The default value is zero.
1896     For example, this pattern has two callout points:
1897 nigel 75 .sp
1898 ph10 155 (?C1)abc(?C2)def
1899 nigel 75 .sp
1900     If the PCRE_AUTO_CALLOUT flag is passed to \fBpcre_compile()\fP, callouts are
1901     automatically installed before each item in the pattern. They are all numbered
1902     255.
1903     .P
1904     During matching, when PCRE reaches a callout point (and \fIpcre_callout\fP is
1905 nigel 63 set), the external function is called. It is provided with the number of the
1906 nigel 75 callout, the position in the pattern, and, optionally, one item of data
1907     originally supplied by the caller of \fBpcre_exec()\fP. The callout function
1908     may cause matching to proceed, to backtrack, or to fail altogether. A complete
1909     description of the interface to the callout function is given in the
1910 nigel 63 .\" HREF
1911 nigel 75 \fBpcrecallout\fP
1912 nigel 63 .\"
1913     documentation.
1914 nigel 93 .
1915     .
1916     .SH "SEE ALSO"
1917     .rs
1918     .sp
1919     \fBpcreapi\fP(3), \fBpcrecallout\fP(3), \fBpcrematching\fP(3), \fBpcre\fP(3).
1920 ph10 99 .
1921     .
1922     .SH AUTHOR
1923     .rs
1924     .sp
1925     .nf
1926     Philip Hazel
1927     University Computing Service
1928     Cambridge CB2 3QH, England.
1929     .fi
1930     .
1931     .
1932     .SH REVISION
1933     .rs
1934     .sp
1935     .nf
1936 ph10 175 Last updated: 11 June 2007
1937 ph10 99 Copyright (c) 1997-2007 University of Cambridge.
1938     .fi

Properties

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

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12