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

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

Parent Directory Parent Directory | Revision Log Revision Log


Revision 167 - (show annotations) (download)
Wed May 9 15:53:54 2007 UTC (7 years, 7 months ago) by ph10
File size: 76101 byte(s)
Add (?(-n) and (?(+n) relative conditions.

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