/[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 - (show annotations) (download)
Mon Jun 11 13:38:38 2007 UTC (7 years, 4 months ago) by ph10
File size: 79109 byte(s)
Added the Perl 5.10 (?| "branch reset" feature.

1 .TH PCREPATTERN 3
2 .SH NAME
3 PCRE - Perl-compatible regular expressions
4 .SH "PCRE REGULAR EXPRESSION DETAILS"
5 .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 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 .\" HTML <a href="pcre.html#utf8support">
20 .\" </a>
21 section on UTF-8 support
22 .\"
23 in the main
24 .\" HREF
25 \fBpcre\fP
26 .\"
27 page.
28 .P
29 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 Perl-compatible. Some of the features discussed below are not available when
34 \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 .\" HREF
38 \fBpcrematching\fP
39 .\"
40 page.
41 .
42 .
43 .SH "CHARACTERS AND METACHARACTERS"
44 .rs
45 .sp
46 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 .sp
50 The quick brown fox
51 .sp
52 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 \fImetacharacters\fP, which do not stand for themselves but instead are
65 interpreted in some special way.
66 .P
67 There are two different sets of metacharacters: those that are recognized
68 anywhere in the pattern except within square brackets, and those that are
69 recognized within square brackets. Outside square brackets, the metacharacters
70 are as follows:
71 .sp
72 \e general escape character with several uses
73 ^ 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 .sp
88 Part of a pattern that is in square brackets is called a "character class". In
89 a character class the only metacharacters are:
90 .sp
91 \e general escape character
92 ^ negate the class, but only if the first character
93 - indicates character range
94 .\" JOIN
95 [ POSIX character class (only if followed by POSIX
96 syntax)
97 ] terminates the character class
98 .sp
99 The following sections describe the use of each of the metacharacters.
100 .
101 .
102 .SH BACKSLASH
103 .rs
104 .sp
105 The backslash character has several uses. Firstly, if it is followed by a
106 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 outside character classes.
109 .P
110 For example, if you want to match a * character, you write \e* in the pattern.
111 This escaping action applies whether or not the following character would
112 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 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 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 .P
121 If you want to remove the special meaning from a sequence of characters, you
122 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 Perl, $ and @ cause variable interpolation. Note the following examples:
125 .sp
126 Pattern PCRE matches Perl matches
127 .sp
128 .\" JOIN
129 \eQabc$xyz\eE abc$xyz abc followed by the
130 contents of $xyz
131 \eQabc\e$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 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 .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 \ex{hhh..} character with hex code hhh..
158 .sp
159 The precise effect of \ecx is as follows: if x is a lower case letter, it
160 is converted to upper case. Then bit 6 of the character (hex 40) is inverted.
161 Thus \ecz becomes hex 1A, but \ec{ becomes hex 3B, while \ec; becomes hex
162 7B.
163 .P
164 After \ex, from zero to two hexadecimal digits are read (letters can be in
165 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 .P
173 Characters whose value is less than 256 can be defined by either of the two
174 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 .P
177 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 .P
183 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 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 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 digits following the backslash, and uses them to generate a data character. Any
201 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 .sp
205 \e040 is another way of writing a space
206 .\" JOIN
207 \e40 is the same, provided there are fewer than 40
208 previous capturing subpatterns
209 \e7 is always a back reference
210 .\" JOIN
211 \e11 might be a back reference, or another way of
212 writing a tab
213 \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 character with octal code 113
218 .\" JOIN
219 \e377 might be a back reference, otherwise
220 the byte consisting entirely of 1 bits
221 .\" JOIN
222 \e81 is either a back reference, or a binary zero
223 followed by the two characters "8" and "1"
224 .sp
225 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 .P
228 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 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 .\" HTML <a href="#uniextseq">
235 .\" </a>
236 (see below).
237 .\"
238 .
239 .
240 .SS "Absolute and relative back references"
241 .rs
242 .sp
243 The sequence \eg followed by a positive or negative number, optionally enclosed
244 in braces, is an absolute or relative back reference. A named back reference
245 can be coded as \eg{name}. Back references are discussed
246 .\" 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 .SS "Generic character types"
258 .rs
259 .sp
260 Another use of backslash is for specifying generic character types. The
261 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 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 .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 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 .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 .\" HTML <a href="pcreapi.html#localesupport">
289 .\" </a>
290 "Locale support"
291 .\"
292 in the
293 .\" HREF
294 \fBpcreapi\fP
295 .\"
296 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 .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 character property support is available. The use of locales with Unicode is
303 discouraged.
304 .
305 .
306 .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 .\" 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 escape sequences to match character properties are available when UTF-8 mode
340 is selected. They are:
341 .sp
342 \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 .sp
346 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 .P
352 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 .sp
356 \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 Balinese,
365 Bengali,
366 Bopomofo,
367 Braille,
368 Buginese,
369 Buhid,
370 Canadian_Aboriginal,
371 Cherokee,
372 Common,
373 Coptic,
374 Cuneiform,
375 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 Nko,
405 Ogham,
406 Old_Italic,
407 Old_Persian,
408 Oriya,
409 Osmanya,
410 Phags_Pa,
411 Phoenician,
412 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 \ep{L}
440 \epL
441 .sp
442 The following general category property codes are supported:
443 .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 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 .P
492 The long synonyms for these properties that Perl supports (such as \ep{Letter})
493 are not supported by PCRE, nor is it permitted to prefix any of these
494 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 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 .\" HTML <a name="resetmatchstart"></a>
525 .SS "Resetting the match start"
526 .rs
527 .sp
528 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 example, the pattern:
531 .sp
532 foo\eKbar
533 .sp
534 matches "foobar", but reports that it has matched "bar". This feature is
535 similar to a lookbehind assertion
536 .\" HTML <a href="#lookbehind">
537 .\" </a>
538 (described below).
539 .\"
540 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 not interfere with the setting of
543 .\" HTML <a href="#subpattern">
544 .\" </a>
545 captured substrings.
546 .\"
547 For example, when the pattern
548 .sp
549 (foo)\eKbar
550 .sp
551 matches "foobar", the first substring is still set to "foo".
552 .
553 .
554 .\" HTML <a name="smallassertions"></a>
555 .SS "Simple assertions"
556 .rs
557 .sp
558 The final use of backslash is for certain simple assertions. An assertion
559 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 subpatterns for more complicated assertions is described
562 .\" HTML <a href="#bigassertions">
563 .\" </a>
564 below.
565 .\"
566 The backslashed assertions are:
567 .sp
568 \eb matches at a word boundary
569 \eB matches when not at a word boundary
570 \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 .sp
576 These assertions may not appear in character classes (but note that \eb has a
577 different meaning, namely the backspace character, inside a character class).
578 .P
579 A word boundary is a position in the subject string where the current character
580 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 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 .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 arguments, you can mimic Perl's /g option, and it is in this kind of
600 implementation where \eG can be useful.
601 .P
602 Note, however, that PCRE's interpretation of \eG, as the start of the current
603 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 .P
608 If all the alternatives of a pattern begin with \eG, the expression is anchored
609 to the starting match position, and the "anchored" flag is set in the compiled
610 regular expression.
611 .
612 .
613 .SH "CIRCUMFLEX AND DOLLAR"
614 .rs
615 .sp
616 Outside a character class, in the default matching mode, the circumflex
617 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 option is unset. Inside a character class, circumflex has an entirely different
621 meaning
622 .\" HTML <a href="#characterclass">
623 .\" </a>
624 (see below).
625 .\"
626 .P
627 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 .P
635 A dollar character is an assertion that is true only if the current matching
636 point is at the end of the subject string, or immediately before a newline
637 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 .P
642 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 does not affect the \eZ assertion.
645 .P
646 The meanings of the circumflex and dollar characters are changed if the
647 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 .P
654 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 Note that the sequences \eA, \eZ, and \ez can be used to match the start and
662 end of the subject in both modes, and if all branches of a pattern start with
663 \eA it is always anchored, whether or not PCRE_MULTILINE is set.
664 .
665 .
666 .SH "FULL STOP (PERIOD, DOT)"
667 .rs
668 .sp
669 Outside a character class, a dot in the pattern matches any one character in
670 the subject string except (by default) a character that signifies the end of a
671 line. In UTF-8 mode, the matched character may be more than one byte long.
672 .P
673 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 The behaviour of dot with regard to newlines can be changed. If the PCRE_DOTALL
681 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 .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 .
689 .
690 .SH "MATCHING A SINGLE BYTE"
691 .rs
692 .sp
693 Outside a character class, the escape sequence \eC matches any one byte, both
694 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 .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 .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 .P
719 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 .P
727 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 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 In UTF-8 mode, characters with values greater than 255 can be included in a
736 class as a literal string of bytes, or by using the \ex{ escaping mechanism.
737 .P
738 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 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 .P
749 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 .P
754 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 .P
760 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 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 Ranges operate in the collating sequence of character values. They can also be
769 used for characters specified numerically, for example [\e000-\e037]. In UTF-8
770 mode, ranges can include characters whose values are greater than 255, for
771 example [\ex{100}-\ex{2ff}].
772 .P
773 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 [][\e\e^_`wxyzabc], matched caselessly, and in non-UTF-8 mode, if character
776 tables for a French locale are in use, [\exc8-\excb] matches accented E
777 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 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 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 .rs
798 .sp
799 Perl supports the POSIX notation for character classes. This uses names
800 enclosed by [: and :] within the enclosing square brackets. PCRE also supports
801 this notation. For example,
802 .sp
803 [01[:alpha:]%]
804 .sp
805 matches "0", "1", any alphabetic character, or "%". The supported class names
806 are
807 .sp
808 alnum letters and digits
809 alpha letters
810 ascii character codes 0 - 127
811 blank space or tab only
812 cntrl control characters
813 digit decimal digits (same as \ed)
814 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 space white space (not quite the same as \es)
819 upper upper case letters
820 word "word" characters (same as \ew)
821 xdigit hexadecimal digits
822 .sp
823 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 makes "space" different to \es, which does not include VT (for Perl
826 compatibility).
827 .P
828 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 .sp
832 [12[:^digit:]]
833 .sp
834 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 .P
838 In UTF-8 mode, characters with values greater than 128 do not match any of
839 the POSIX character classes.
840 .
841 .
842 .SH "VERTICAL BAR"
843 .rs
844 .sp
845 Vertical bar characters are used to separate alternative patterns. For example,
846 the pattern
847 .sp
848 gilbert|sullivan
849 .sp
850 matches either "gilbert" or "sullivan". Any number of alternatives may appear,
851 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 .\" 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 .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 .sp
869 i for PCRE_CASELESS
870 m for PCRE_MULTILINE
871 s for PCRE_DOTALL
872 x for PCRE_EXTENDED
873 .sp
874 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 .P
881 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 \fBpcre_fullinfo()\fP function).
886 .P
887 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 .sp
890 (a(?i)b)c
891 .sp
892 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 .sp
897 (a(?i)b|c)
898 .sp
899 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 .P
904 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 .
908 .
909 .\" HTML <a name="subpattern"></a>
910 .SH SUBPATTERNS
911 .rs
912 .sp
913 Subpatterns are delimited by parentheses (round brackets), which can be nested.
914 Turning part of a pattern into a subpattern does two things:
915 .sp
916 1. It localizes a set of alternatives. For example, the pattern
917 .sp
918 cat(aract|erpillar|)
919 .sp
920 matches one of the words "cat", "cataract", or "caterpillar". Without the
921 parentheses, it would match "cataract", "erpillar" or an empty string.
922 .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 For example, if the string "the red king" is matched against the pattern
930 .sp
931 the ((red|white) (king|queen))
932 .sp
933 the captured substrings are "red king", "red", and "king", and are numbered 1,
934 2, and 3, respectively.
935 .P
936 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 .sp
943 the ((?:red|white) (king|queen))
944 .sp
945 the captured substrings are "white queen" and "queen", and are numbered 1 and
946 2. The maximum number of capturing subpatterns is 65535.
947 .P
948 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 .sp
952 (?i:saturday|sunday)
953 (?:(?i)saturday|sunday)
954 .sp
955 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 .
960 .
961 .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 .SH "NAMED SUBPATTERNS"
994 .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 if an expression is modified, the numbers may change. To help with this
999 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 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 .P
1022 Names consist of up to 32 alphanumeric characters and underscores. Named
1023 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 .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 (?<DN>Mon|Fri|Sun)(?:day)?|
1036 (?<DN>Tue)(?:sday)?|
1037 (?<DN>Wed)(?:nesday)?|
1038 (?<DN>Thu)(?:rsday)?|
1039 (?<DN>Sat)(?:urday)?
1040 .sp
1041 There are five capturing substrings, but only one is ever set after a match.
1042 (An alternative way of solving this problem is to use a "branch reset"
1043 subpattern, as described in the previous section.)
1044 .P
1045 The convenience function for extracting the data by name returns the substring
1046 for the first (and in this example, the only) subpattern of that name that
1047 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 .\" HREF
1052 \fBpcreapi\fP
1053 .\"
1054 documentation.
1055 .
1056 .
1057 .SH REPETITION
1058 .rs
1059 .sp
1060 Repetition is specified by quantifiers, which can follow any of the following
1061 items:
1062 .sp
1063 a literal data character
1064 the dot metacharacter
1065 the \eC escape sequence
1066 the \eX escape sequence (in UTF-8 mode with Unicode properties)
1067 the \eR escape sequence
1068 an escape such as \ed that matches a single character
1069 a character class
1070 a back reference (see next section)
1071 a parenthesized subpattern (unless it is an assertion)
1072 .sp
1073 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 .sp
1078 z{2,4}
1079 .sp
1080 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 .sp
1085 [aeiou]{3,}
1086 .sp
1087 matches at least 3 successive vowels, but may match many more, while
1088 .sp
1089 \ed{8}
1090 .sp
1091 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 .P
1096 In UTF-8 mode, quantifiers apply to UTF-8 characters rather than to individual
1097 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 The quantifier {0} is permitted, causing the expression to behave as if the
1103 previous item and the quantifier were not present.
1104 .P
1105 For convenience, the three most common quantifiers have single-character
1106 abbreviations:
1107 .sp
1108 * is equivalent to {0,}
1109 + is equivalent to {1,}
1110 ? is equivalent to {0,1}
1111 .sp
1112 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 .sp
1115 (a?)*
1116 .sp
1117 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 .P
1122 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 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 to the string
1132 .sp
1133 /* first comment */ not comment /* second comment */
1134 .sp
1135 fails, because it matches the entire string owing to the greediness of the .*
1136 item.
1137 .P
1138 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 .sp
1142 /\e*.*?\e*/
1143 .sp
1144 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 .sp
1149 \ed??\ed
1150 .sp
1151 which matches one digit by preference, but can match two if that is the only
1152 way the rest of the pattern matches.
1153 .P
1154 If the PCRE_UNGREEDY option is set (an option that is not available in Perl),
1155 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 .P
1159 When a parenthesized subpattern is quantified with a minimum repeat count that
1160 is greater than 1 or with a limited maximum, more memory is required for the
1161 compiled pattern, in proportion to the size of the minimum or maximum.
1162 .P
1163 If a pattern starts with .* or .{0,} and the PCRE_DOTALL option (equivalent
1164 to Perl's /s) is set, thus allowing the dot to match newlines, the pattern is
1165 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 pattern as though it were preceded by \eA.
1169 .P
1170 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 .P
1174 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 elsewhere in the pattern, a match at the start may fail where a later one
1177 succeeds. Consider, for example:
1178 .sp
1179 (.*)abc\e1
1180 .sp
1181 If the subject is "xyz123abc123" the match point is the fourth character. For
1182 this reason, such a pattern is not implicitly anchored.
1183 .P
1184 When a capturing subpattern is repeated, the value captured is the substring
1185 that matched the final iteration. For example, after
1186 .sp
1187 (tweedle[dume]{3}\es*)+
1188 .sp
1189 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 .sp
1194 /(a|(b))+/
1195 .sp
1196 matches "aba" the value of the second captured substring is "b".
1197 .
1198 .
1199 .\" HTML <a name="atomicgroup"></a>
1200 .SH "ATOMIC GROUPING AND POSSESSIVE QUANTIFIERS"
1201 .rs
1202 .sp
1203 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 .P
1210 Consider, for example, the pattern \ed+foo when applied to the subject line
1211 .sp
1212 123456bar
1213 .sp
1214 After matching all 6 digits and then failing to match "foo", the normal
1215 action of the matcher is to try again with only 5 digits matching the \ed+
1216 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 .P
1220 If we use atomic grouping for the previous example, the matcher gives up
1221 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 .sp
1224 (?>\ed+)foo
1225 .sp
1226 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 .P
1231 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 .P
1235 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 everything it can. So, while both \ed+ and \ed+? are prepared to adjust the
1238 number of digits they match in order to make the rest of the pattern match,
1239 (?>\ed+) can only match an entire sequence of digits.
1240 .P
1241 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 .sp
1248 \ed++foo
1249 .sp
1250 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 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 .P
1256 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 .P
1262 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 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 .sp
1271 (\eD+|<\ed+>)*[!?]
1272 .sp
1273 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 .sp
1277 aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
1278 .sp
1279 it takes a long time before reporting failure. This is because the string can
1280 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 sequences of non-digits cannot be broken, and failure happens quickly.
1291 .
1292 .
1293 .\" HTML <a name="backreferences"></a>
1294 .SH "BACK REFERENCES"
1295 .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 .P
1302 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 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 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 "Non-printing characters"
1314 .\" HTML <a href="#digitsafterbackslash">
1315 .\" </a>
1316 above
1317 .\"
1318 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 .P
1322 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 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 .sp
1352 (sens|respons)e and \e1ibility
1353 .sp
1354 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 .sp
1358 ((?i)rah)\es+\e1
1359 .sp
1360 matches "rah rah" and "RAH RAH", but not "RAH rah", even though the original
1361 capturing subpattern is matched caselessly.
1362 .P
1363 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 the following ways:
1369 .sp
1370 (?<p1>(?i)rah)\es+\ek<p1>
1371 (?'p1'(?i)rah)\es+\ek{p1}
1372 (?P<p1>(?i)rah)\es+(?P=p1)
1373 (?<p1>(?i)rah)\es+\eg{p1}
1374 .sp
1375 A subpattern that is referenced by name may appear in the pattern before or
1376 after the reference.
1377 .P
1378 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 .sp
1382 (a|(bc))\e2
1383 .sp
1384 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 Otherwise an empty comment (see
1390 .\" HTML <a href="#comments">
1391 .\" </a>
1392 "Comments"
1393 .\"
1394 below) can be used.
1395 .P
1396 A back reference that occurs inside the parentheses to which it refers fails
1397 when the subpattern is first used, so, for example, (a\e1) never matches.
1398 However, such references can be useful inside repeated subpatterns. For
1399 example, the pattern
1400 .sp
1401 (a|b\e1)+
1402 .sp
1403 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 .
1410 .
1411 .\" HTML <a name="bigassertions"></a>
1412 .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 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 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 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 Lookahead assertions start with (?= for positive assertions and (?! for
1440 negative assertions. For example,
1441 .sp
1442 \ew+(?=;)
1443 .sp
1444 matches a word followed by a semicolon, but does not include the semicolon in
1445 the match, and
1446 .sp
1447 foo(?!bar)
1448 .sp
1449 matches any occurrence of "foo" that is not followed by "bar". Note that the
1450 apparently similar pattern
1451 .sp
1452 (?!foo)bar
1453 .sp
1454 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 lookbehind assertion is needed to achieve the other effect.
1458 .P
1459 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 .
1463 .
1464 .\" HTML <a name="lookbehind"></a>
1465 .SS "Lookbehind assertions"
1466 .rs
1467 .sp
1468 Lookbehind assertions start with (?<= for positive assertions and (?<! for
1469 negative assertions. For example,
1470 .sp
1471 (?<!foo)bar
1472 .sp
1473 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 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 .sp
1478 (?<=bullock|donkey)
1479 .sp
1480 is permitted, but
1481 .sp
1482 (?<!dogs?|cats?)
1483 .sp
1484 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 .sp
1489 (?<=ab(c|de))
1490 .sp
1491 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 .sp
1494 (?<=abc|abde)
1495 .sp
1496 In some cases, the Perl 5.10 escape sequence \eK
1497 .\" 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 The implementation of lookbehind assertions is, for each alternative, to
1505 temporarily move the current position back by the fixed length and then try to
1506 match. If there are insufficient characters before the current position, the
1507 assertion fails.
1508 .P
1509 PCRE does not allow the \eC escape (which matches a single byte in UTF-8 mode)
1510 to appear in lookbehind assertions, because it makes it impossible to calculate
1511 the length of the lookbehind. The \eX and \eR escapes, which can match
1512 different numbers of bytes, are also not permitted.
1513 .P
1514 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 .sp
1518 abcd$
1519 .sp
1520 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 .sp
1524 ^.*abcd$
1525 .sp
1526 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 .sp
1532 ^.*+(?<=abcd)
1533 .sp
1534 there can be no backtracking for the .*+ item; it can match only the entire
1535 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 .
1539 .
1540 .SS "Using multiple assertions"
1541 .rs
1542 .sp
1543 Several assertions (of any sort) may occur in succession. For example,
1544 .sp
1545 (?<=\ed{3})(?<!999)foo
1546 .sp
1547 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 This pattern does \fInot\fP match "foo" preceded by six characters, the first
1552 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 .sp
1555 (?<=\ed{3}...)(?<!999)foo
1556 .sp
1557 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 .P
1561 Assertions can be nested in any combination. For example,
1562 .sp
1563 (?<=(?<!foo)bar)baz
1564 .sp
1565 matches an occurrence of "baz" that is preceded by "bar" which in turn is not
1566 preceded by "foo", while
1567 .sp
1568 (?<=\ed{3}(?!999)...)foo
1569 .sp
1570 is another pattern that matches "foo" preceded by three digits and any three
1571 characters that are not "999".
1572 .
1573 .
1574 .\" HTML <a name="conditions"></a>
1575 .SH "CONDITIONAL SUBPATTERNS"
1576 .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 .sp
1583 (?(condition)yes-pattern)
1584 (?(condition)yes-pattern|no-pattern)
1585 .sp
1586 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 .P
1590 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 matched. An alternative notation is to precede the digits with a plus or minus
1599 sign. In this case, the subpattern number is relative rather than absolute.
1600 The most recently opened parentheses can be referenced by (?(-1), the next most
1601 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 .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 .sp
1608 ( \e( )? [^()]+ (?(1) \e) )
1609 .sp
1610 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 non-parentheses, optionally enclosed in parentheses.
1619 .P
1620 If you were embedding this pattern in a larger one, you could use a relative
1621 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 .
1627 .SS "Checking for a used subpattern by name"
1628 .rs
1629 .sp
1630 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 .sp
1641 (?<OPEN> \e( )? [^()]+ (?(<OPEN>) \e) )
1642 .sp
1643 .
1644 .SS "Checking for pattern recursion"
1645 .rs
1646 .sp
1647 If the condition is the string (R), and there is no subpattern with the name R,
1648 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 .P
1658 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 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 .sp
1692 (?(?=[^a-z]*[a-z])
1693 \ed{2}-[a-z]{3}-\ed{2} | \ed{2}-\ed{2}-\ed{2} )
1694 .sp
1695 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 .
1702 .
1703 .\" HTML <a name="comments"></a>
1704 .SH COMMENTS
1705 .rs
1706 .sp
1707 The sequence (?# marks the start of a comment that continues up to the next
1708 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 .P
1711 If the PCRE_EXTENDED option is set, an unescaped # character outside a
1712 character class introduces a comment that continues to immediately after the
1713 next newline in the pattern.
1714 .
1715 .
1716 .\" HTML <a name="recursion"></a>
1717 .SH "RECURSIVE PATTERNS"
1718 .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 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 .sp
1731 $re = qr{\e( (?: (?>[^()]+) | (?p{$re}) )* \e)}x;
1732 .sp
1733 The (?p{...}) item interpolates Perl code at run time, and in this case refers
1734 recursively to the pattern in which it appears.
1735 .P
1736 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 .P
1741 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 .P
1747 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 This PCRE pattern solves the nested parentheses problem (assume the
1753 PCRE_EXTENDED option is set so that white space is ignored):
1754 .sp
1755 \e( ( (?>[^()]+) | (?R) )* \e)
1756 .sp
1757 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 match of the pattern itself (that is, a correctly parenthesized substring).
1760 Finally there is a closing parenthesis.
1761 .P
1762 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 .sp
1765 ( \e( ( (?>[^()]+) | (?1) )* \e) )
1766 .sp
1767 We have put the pattern into parentheses, and caused the recursion to refer to
1768 them instead of the whole pattern.
1769 .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 .sp
1786 (?<pn> \e( ( (?>[^()]+) | (?&pn) )* \e) )
1787 .sp
1788 If there is more than one subpattern with the same name, the earliest one is
1789 used.
1790 .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 .sp
1796 (aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa()
1797 .sp
1798 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 .P
1803 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 below and the
1807 .\" HREF
1808 \fBpcrecallout\fP
1809 .\"
1810 documentation). If the pattern above is matched against
1811 .sp
1812 (ab(cd)ef)
1813 .sp
1814 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 .sp
1817 \e( ( ( (?>[^()]+) | (?R) )* ) \e)
1818 ^ ^
1819 ^ ^
1820 .sp
1821 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 using \fBpcre_malloc\fP, freeing it via \fBpcre_free\fP afterwards. If no
1825 memory can be obtained, the match fails with the PCRE_ERROR_NOMEMORY error.
1826 .P
1827 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 .sp
1832 < (?: (?(R) \ed++ | [^<>]*+) | (?R)) * >
1833 .sp
1834 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 .
1838 .
1839 .\" HTML <a name="subpatternsassubroutines"></a>
1840 .SH "SUBPATTERNS AS SUBROUTINES"
1841 .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 subroutine in a programming language. The "called" subpattern may be defined
1846 before or after the reference. A numbered reference can be absolute or
1847 relative, as in these examples:
1848 .sp
1849 (...(absolute)...)...(?2)...
1850 (...(relative)...)...(?-1)...
1851 (...(?+1)...(relative)...
1852 .sp
1853 An earlier example pointed out that the pattern
1854 .sp
1855 (sens|respons)e and \e1ibility
1856 .sp
1857 matches "sense and sensibility" and "response and responsibility", but not
1858 "sense and responsibility". If instead the pattern
1859 .sp
1860 (sens|respons)e and (?1)ibility
1861 .sp
1862 is used, it does match "sense and responsibility" as well as the other two
1863 strings. Another example is given in the discussion of DEFINE above.
1864 .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 .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 (abc)(?i:(?-1))
1875 .sp
1876 It matches "abcabc". It does not match "abcABC" because the change of
1877 processing option does not affect the called subpattern.
1878 .
1879 .
1880 .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 .P
1888 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 function by putting its entry point in the global variable \fIpcre_callout\fP.
1891 By default, this variable contains NULL, which disables all calling out.
1892 .P
1893 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 .sp
1898 (?C1)abc(?C2)def
1899 .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 set), the external function is called. It is provided with the number of the
1906 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 .\" HREF
1911 \fBpcrecallout\fP
1912 .\"
1913 documentation.
1914 .
1915 .
1916 .SH "SEE ALSO"
1917 .rs
1918 .sp
1919 \fBpcreapi\fP(3), \fBpcrecallout\fP(3), \fBpcrematching\fP(3), \fBpcre\fP(3).
1920 .
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 Last updated: 11 June 2007
1937 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