ViewVC logotype

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

Parent Directory Parent Directory | Revision Log Revision Log

Revision 1011 - (show annotations)
Sat Aug 25 11:36:15 2012 UTC (7 years, 7 months ago) by ph10
File size: 122918 byte(s)
Error occurred while calculating annotation data.
Upgrade \X to match an extended grapheme cluster
1 .TH PCREPATTERN 3 "04 May 2012" "PCRE 8.31"
3 PCRE - Perl-compatible regular expressions
5 .rs
6 .sp
7 The syntax and semantics of the regular expressions that are supported by PCRE
8 are described in detail below. There is a quick-reference syntax summary in the
9 .\" HREF
10 \fBpcresyntax\fP
11 .\"
12 page. PCRE tries to match Perl syntax and semantics as closely as it can. PCRE
13 also supports some alternative regular expression syntax (which does not
14 conflict with the Perl syntax) in order to provide some compatibility with
15 regular expressions in Python, .NET, and Oniguruma.
16 .P
17 Perl's regular expressions are described in its own documentation, and
18 regular expressions in general are covered in a number of books, some of which
19 have copious examples. Jeffrey Friedl's "Mastering Regular Expressions",
20 published by O'Reilly, covers regular expressions in great detail. This
21 description of PCRE's regular expressions is intended as reference material.
22 .P
23 The original operation of PCRE was on strings of one-byte characters. However,
24 there is now also support for UTF-8 strings in the original library, and a
25 second library that supports 16-bit and UTF-16 character strings. To use these
26 features, PCRE must be built to include appropriate support. When using UTF
27 strings you must either call the compiling function with the PCRE_UTF8 or
28 PCRE_UTF16 option, or the pattern must start with one of these special
29 sequences:
30 .sp
31 (*UTF8)
32 (*UTF16)
33 .sp
34 Starting a pattern with such a sequence is equivalent to setting the relevant
35 option. This feature is not Perl-compatible. How setting a UTF mode affects
36 pattern matching is mentioned in several places below. There is also a summary
37 of features in the
38 .\" HREF
39 \fBpcreunicode\fP
40 .\"
41 page.
42 .P
43 Another special sequence that may appear at the start of a pattern or in
44 combination with (*UTF8) or (*UTF16) is:
45 .sp
46 (*UCP)
47 .sp
48 This has the same effect as setting the PCRE_UCP option: it causes sequences
49 such as \ed and \ew to use Unicode properties to determine character types,
50 instead of recognizing only characters with codes less than 128 via a lookup
51 table.
52 .P
53 If a pattern starts with (*NO_START_OPT), it has the same effect as setting the
54 PCRE_NO_START_OPTIMIZE option either at compile or matching time. There are
55 also some more of these special sequences that are concerned with the handling
56 of newlines; they are described below.
57 .P
58 The remainder of this document discusses the patterns that are supported by
59 PCRE when one its main matching functions, \fBpcre_exec()\fP (8-bit) or
60 \fBpcre16_exec()\fP (16-bit), is used. PCRE also has alternative matching
61 functions, \fBpcre_dfa_exec()\fP and \fBpcre16_dfa_exec()\fP, which match using
62 a different algorithm that is not Perl-compatible. Some of the features
63 discussed below are not available when DFA matching is used. The advantages and
64 disadvantages of the alternative functions, and how they differ from the normal
65 functions, are discussed in the
66 .\" HREF
67 \fBpcrematching\fP
68 .\"
69 page.
70 .
71 .
72 .\" HTML <a name="newlines"></a>
74 .rs
75 .sp
76 PCRE supports five different conventions for indicating line breaks in
77 strings: a single CR (carriage return) character, a single LF (linefeed)
78 character, the two-character sequence CRLF, any of the three preceding, or any
79 Unicode newline sequence. The
80 .\" HREF
81 \fBpcreapi\fP
82 .\"
83 page has
84 .\" HTML <a href="pcreapi.html#newlines">
85 .\" </a>
86 further discussion
87 .\"
88 about newlines, and shows how to set the newline convention in the
89 \fIoptions\fP arguments for the compiling and matching functions.
90 .P
91 It is also possible to specify a newline convention by starting a pattern
92 string with one of the following five sequences:
93 .sp
94 (*CR) carriage return
95 (*LF) linefeed
96 (*CRLF) carriage return, followed by linefeed
97 (*ANYCRLF) any of the three above
98 (*ANY) all Unicode newline sequences
99 .sp
100 These override the default and the options given to the compiling function. For
101 example, on a Unix system where LF is the default newline sequence, the pattern
102 .sp
103 (*CR)a.b
104 .sp
105 changes the convention to CR. That pattern matches "a\enb" because LF is no
106 longer a newline. Note that these special settings, which are not
107 Perl-compatible, are recognized only at the very start of a pattern, and that
108 they must be in upper case. If more than one of them is present, the last one
109 is used.
110 .P
111 The newline convention affects the interpretation of the dot metacharacter when
112 PCRE_DOTALL is not set, and also the behaviour of \eN. However, it does not
113 affect what the \eR escape sequence matches. By default, this is any Unicode
114 newline sequence, for Perl compatibility. However, this can be changed; see the
115 description of \eR in the section entitled
116 .\" HTML <a href="#newlineseq">
117 .\" </a>
118 "Newline sequences"
119 .\"
120 below. A change of \eR setting can be combined with a change of newline
121 convention.
122 .
123 .
125 .rs
126 .sp
127 A regular expression is a pattern that is matched against a subject string from
128 left to right. Most characters stand for themselves in a pattern, and match the
129 corresponding characters in the subject. As a trivial example, the pattern
130 .sp
131 The quick brown fox
132 .sp
133 matches a portion of a subject string that is identical to itself. When
134 caseless matching is specified (the PCRE_CASELESS option), letters are matched
135 independently of case. In a UTF mode, PCRE always understands the concept of
136 case for characters whose values are less than 128, so caseless matching is
137 always possible. For characters with higher values, the concept of case is
138 supported if PCRE is compiled with Unicode property support, but not otherwise.
139 If you want to use caseless matching for characters 128 and above, you must
140 ensure that PCRE is compiled with Unicode property support as well as with
141 UTF support.
142 .P
143 The power of regular expressions comes from the ability to include alternatives
144 and repetitions in the pattern. These are encoded in the pattern by the use of
145 \fImetacharacters\fP, which do not stand for themselves but instead are
146 interpreted in some special way.
147 .P
148 There are two different sets of metacharacters: those that are recognized
149 anywhere in the pattern except within square brackets, and those that are
150 recognized within square brackets. Outside square brackets, the metacharacters
151 are as follows:
152 .sp
153 \e general escape character with several uses
154 ^ assert start of string (or line, in multiline mode)
155 $ assert end of string (or line, in multiline mode)
156 . match any character except newline (by default)
157 [ start character class definition
158 | start of alternative branch
159 ( start subpattern
160 ) end subpattern
161 ? extends the meaning of (
162 also 0 or 1 quantifier
163 also quantifier minimizer
164 * 0 or more quantifier
165 + 1 or more quantifier
166 also "possessive quantifier"
167 { start min/max quantifier
168 .sp
169 Part of a pattern that is in square brackets is called a "character class". In
170 a character class the only metacharacters are:
171 .sp
172 \e general escape character
173 ^ negate the class, but only if the first character
174 - indicates character range
175 .\" JOIN
176 [ POSIX character class (only if followed by POSIX
177 syntax)
178 ] terminates the character class
179 .sp
180 The following sections describe the use of each of the metacharacters.
181 .
182 .
184 .rs
185 .sp
186 The backslash character has several uses. Firstly, if it is followed by a
187 character that is not a number or a letter, it takes away any special meaning
188 that character may have. This use of backslash as an escape character applies
189 both inside and outside character classes.
190 .P
191 For example, if you want to match a * character, you write \e* in the pattern.
192 This escaping action applies whether or not the following character would
193 otherwise be interpreted as a metacharacter, so it is always safe to precede a
194 non-alphanumeric with backslash to specify that it stands for itself. In
195 particular, if you want to match a backslash, you write \e\e.
196 .P
197 In a UTF mode, only ASCII numbers and letters have any special meaning after a
198 backslash. All other characters (in particular, those whose codepoints are
199 greater than 127) are treated as literals.
200 .P
201 If a pattern is compiled with the PCRE_EXTENDED option, white space in the
202 pattern (other than in a character class) and characters between a # outside
203 a character class and the next newline are ignored. An escaping backslash can
204 be used to include a white space or # character as part of the pattern.
205 .P
206 If you want to remove the special meaning from a sequence of characters, you
207 can do so by putting them between \eQ and \eE. This is different from Perl in
208 that $ and @ are handled as literals in \eQ...\eE sequences in PCRE, whereas in
209 Perl, $ and @ cause variable interpolation. Note the following examples:
210 .sp
211 Pattern PCRE matches Perl matches
212 .sp
213 .\" JOIN
214 \eQabc$xyz\eE abc$xyz abc followed by the
215 contents of $xyz
216 \eQabc\e$xyz\eE abc\e$xyz abc\e$xyz
217 \eQabc\eE\e$\eQxyz\eE abc$xyz abc$xyz
218 .sp
219 The \eQ...\eE sequence is recognized both inside and outside character classes.
220 An isolated \eE that is not preceded by \eQ is ignored. If \eQ is not followed
221 by \eE later in the pattern, the literal interpretation continues to the end of
222 the pattern (that is, \eE is assumed at the end). If the isolated \eQ is inside
223 a character class, this causes an error, because the character class is not
224 terminated.
225 .
226 .
227 .\" HTML <a name="digitsafterbackslash"></a>
228 .SS "Non-printing characters"
229 .rs
230 .sp
231 A second use of backslash provides a way of encoding non-printing characters
232 in patterns in a visible manner. There is no restriction on the appearance of
233 non-printing characters, apart from the binary zero that terminates a pattern,
234 but when a pattern is being prepared by text editing, it is often easier to use
235 one of the following escape sequences than the binary character it represents:
236 .sp
237 \ea alarm, that is, the BEL character (hex 07)
238 \ecx "control-x", where x is any ASCII character
239 \ee escape (hex 1B)
240 \ef form feed (hex 0C)
241 \en linefeed (hex 0A)
242 \er carriage return (hex 0D)
243 \et tab (hex 09)
244 \eddd character with octal code ddd, or back reference
245 \exhh character with hex code hh
246 \ex{hhh..} character with hex code hhh.. (non-JavaScript mode)
247 \euhhhh character with hex code hhhh (JavaScript mode only)
248 .sp
249 The precise effect of \ecx on ASCII characters is as follows: if x is a lower
250 case letter, it is converted to upper case. Then bit 6 of the character (hex
251 40) is inverted. Thus \ecA to \ecZ become hex 01 to hex 1A (A is 41, Z is 5A),
252 but \ec{ becomes hex 3B ({ is 7B), and \ec; becomes hex 7B (; is 3B). If the
253 data item (byte or 16-bit value) following \ec has a value greater than 127, a
254 compile-time error occurs. This locks out non-ASCII characters in all modes.
255 .P
256 The \ec facility was designed for use with ASCII characters, but with the
257 extension to Unicode it is even less useful than it once was. It is, however,
258 recognized when PCRE is compiled in EBCDIC mode, where data items are always
259 bytes. In this mode, all values are valid after \ec. If the next character is a
260 lower case letter, it is converted to upper case. Then the 0xc0 bits of the
261 byte are inverted. Thus \ecA becomes hex 01, as in ASCII (A is C1), but because
262 the EBCDIC letters are disjoint, \ecZ becomes hex 29 (Z is E9), and other
263 characters also generate different values.
264 .P
265 By default, after \ex, from zero to two hexadecimal digits are read (letters
266 can be in upper or lower case). Any number of hexadecimal digits may appear
267 between \ex{ and }, but the character code is constrained as follows:
268 .sp
269 8-bit non-UTF mode less than 0x100
270 8-bit UTF-8 mode less than 0x10ffff and a valid codepoint
271 16-bit non-UTF mode less than 0x10000
272 16-bit UTF-16 mode less than 0x10ffff and a valid codepoint
273 .sp
274 Invalid Unicode codepoints are the range 0xd800 to 0xdfff (the so-called
275 "surrogate" codepoints).
276 .P
277 If characters other than hexadecimal digits appear between \ex{ and }, or if
278 there is no terminating }, this form of escape is not recognized. Instead, the
279 initial \ex will be interpreted as a basic hexadecimal escape, with no
280 following digits, giving a character whose value is zero.
281 .P
282 If the PCRE_JAVASCRIPT_COMPAT option is set, the interpretation of \ex is
283 as just described only when it is followed by two hexadecimal digits.
284 Otherwise, it matches a literal "x" character. In JavaScript mode, support for
285 code points greater than 256 is provided by \eu, which must be followed by
286 four hexadecimal digits; otherwise it matches a literal "u" character.
287 Character codes specified by \eu in JavaScript mode are constrained in the same
288 was as those specified by \ex in non-JavaScript mode.
289 .P
290 Characters whose value is less than 256 can be defined by either of the two
291 syntaxes for \ex (or by \eu in JavaScript mode). There is no difference in the
292 way they are handled. For example, \exdc is exactly the same as \ex{dc} (or
293 \eu00dc in JavaScript mode).
294 .P
295 After \e0 up to two further octal digits are read. If there are fewer than two
296 digits, just those that are present are used. Thus the sequence \e0\ex\e07
297 specifies two binary zeros followed by a BEL character (code value 7). Make
298 sure you supply two digits after the initial zero if the pattern character that
299 follows is itself an octal digit.
300 .P
301 The handling of a backslash followed by a digit other than 0 is complicated.
302 Outside a character class, PCRE reads it and any following digits as a decimal
303 number. If the number is less than 10, or if there have been at least that many
304 previous capturing left parentheses in the expression, the entire sequence is
305 taken as a \fIback reference\fP. A description of how this works is given
306 .\" HTML <a href="#backreferences">
307 .\" </a>
308 later,
309 .\"
310 following the discussion of
311 .\" HTML <a href="#subpattern">
312 .\" </a>
313 parenthesized subpatterns.
314 .\"
315 .P
316 Inside a character class, or if the decimal number is greater than 9 and there
317 have not been that many capturing subpatterns, PCRE re-reads up to three octal
318 digits following the backslash, and uses them to generate a data character. Any
319 subsequent digits stand for themselves. The value of the character is
320 constrained in the same way as characters specified in hexadecimal.
321 For example:
322 .sp
323 \e040 is another way of writing a space
324 .\" JOIN
325 \e40 is the same, provided there are fewer than 40
326 previous capturing subpatterns
327 \e7 is always a back reference
328 .\" JOIN
329 \e11 might be a back reference, or another way of
330 writing a tab
331 \e011 is always a tab
332 \e0113 is a tab followed by the character "3"
333 .\" JOIN
334 \e113 might be a back reference, otherwise the
335 character with octal code 113
336 .\" JOIN
337 \e377 might be a back reference, otherwise
338 the value 255 (decimal)
339 .\" JOIN
340 \e81 is either a back reference, or a binary zero
341 followed by the two characters "8" and "1"
342 .sp
343 Note that octal values of 100 or greater must not be introduced by a leading
344 zero, because no more than three octal digits are ever read.
345 .P
346 All the sequences that define a single character value can be used both inside
347 and outside character classes. In addition, inside a character class, \eb is
348 interpreted as the backspace character (hex 08).
349 .P
350 \eN is not allowed in a character class. \eB, \eR, and \eX are not special
351 inside a character class. Like other unrecognized escape sequences, they are
352 treated as the literal characters "B", "R", and "X" by default, but cause an
353 error if the PCRE_EXTRA option is set. Outside a character class, these
354 sequences have different meanings.
355 .
356 .
357 .SS "Unsupported escape sequences"
358 .rs
359 .sp
360 In Perl, the sequences \el, \eL, \eu, and \eU are recognized by its string
361 handler and used to modify the case of following characters. By default, PCRE
362 does not support these escape sequences. However, if the PCRE_JAVASCRIPT_COMPAT
363 option is set, \eU matches a "U" character, and \eu can be used to define a
364 character by code point, as described in the previous section.
365 .
366 .
367 .SS "Absolute and relative back references"
368 .rs
369 .sp
370 The sequence \eg followed by an unsigned or a negative number, optionally
371 enclosed in braces, is an absolute or relative back reference. A named back
372 reference can be coded as \eg{name}. Back references are discussed
373 .\" HTML <a href="#backreferences">
374 .\" </a>
375 later,
376 .\"
377 following the discussion of
378 .\" HTML <a href="#subpattern">
379 .\" </a>
380 parenthesized subpatterns.
381 .\"
382 .
383 .
384 .SS "Absolute and relative subroutine calls"
385 .rs
386 .sp
387 For compatibility with Oniguruma, the non-Perl syntax \eg followed by a name or
388 a number enclosed either in angle brackets or single quotes, is an alternative
389 syntax for referencing a subpattern as a "subroutine". Details are discussed
390 .\" HTML <a href="#onigurumasubroutines">
391 .\" </a>
392 later.
393 .\"
394 Note that \eg{...} (Perl syntax) and \eg<...> (Oniguruma syntax) are \fInot\fP
395 synonymous. The former is a back reference; the latter is a
396 .\" HTML <a href="#subpatternsassubroutines">
397 .\" </a>
398 subroutine
399 .\"
400 call.
401 .
402 .
403 .\" HTML <a name="genericchartypes"></a>
404 .SS "Generic character types"
405 .rs
406 .sp
407 Another use of backslash is for specifying generic character types:
408 .sp
409 \ed any decimal digit
410 \eD any character that is not a decimal digit
411 \eh any horizontal white space character
412 \eH any character that is not a horizontal white space character
413 \es any white space character
414 \eS any character that is not a white space character
415 \ev any vertical white space character
416 \eV any character that is not a vertical white space character
417 \ew any "word" character
418 \eW any "non-word" character
419 .sp
420 There is also the single sequence \eN, which matches a non-newline character.
421 This is the same as
422 .\" HTML <a href="#fullstopdot">
423 .\" </a>
424 the "." metacharacter
425 .\"
426 when PCRE_DOTALL is not set. Perl also uses \eN to match characters by name;
427 PCRE does not support this.
428 .P
429 Each pair of lower and upper case escape sequences partitions the complete set
430 of characters into two disjoint sets. Any given character matches one, and only
431 one, of each pair. The sequences can appear both inside and outside character
432 classes. They each match one character of the appropriate type. If the current
433 matching point is at the end of the subject string, all of them fail, because
434 there is no character to match.
435 .P
436 For compatibility with Perl, \es does not match the VT character (code 11).
437 This makes it different from the the POSIX "space" class. The \es characters
438 are HT (9), LF (10), FF (12), CR (13), and space (32). If "use locale;" is
439 included in a Perl script, \es may match the VT character. In PCRE, it never
440 does.
441 .P
442 A "word" character is an underscore or any character that is a letter or digit.
443 By default, the definition of letters and digits is controlled by PCRE's
444 low-valued character tables, and may vary if locale-specific matching is taking
445 place (see
446 .\" HTML <a href="pcreapi.html#localesupport">
447 .\" </a>
448 "Locale support"
449 .\"
450 in the
451 .\" HREF
452 \fBpcreapi\fP
453 .\"
454 page). For example, in a French locale such as "fr_FR" in Unix-like systems,
455 or "french" in Windows, some character codes greater than 128 are used for
456 accented letters, and these are then matched by \ew. The use of locales with
457 Unicode is discouraged.
458 .P
459 By default, in a UTF mode, characters with values greater than 128 never match
460 \ed, \es, or \ew, and always match \eD, \eS, and \eW. These sequences retain
461 their original meanings from before UTF support was available, mainly for
462 efficiency reasons. However, if PCRE is compiled with Unicode property support,
463 and the PCRE_UCP option is set, the behaviour is changed so that Unicode
464 properties are used to determine character types, as follows:
465 .sp
466 \ed any character that \ep{Nd} matches (decimal digit)
467 \es any character that \ep{Z} matches, plus HT, LF, FF, CR
468 \ew any character that \ep{L} or \ep{N} matches, plus underscore
469 .sp
470 The upper case escapes match the inverse sets of characters. Note that \ed
471 matches only decimal digits, whereas \ew matches any Unicode digit, as well as
472 any Unicode letter, and underscore. Note also that PCRE_UCP affects \eb, and
473 \eB because they are defined in terms of \ew and \eW. Matching these sequences
474 is noticeably slower when PCRE_UCP is set.
475 .P
476 The sequences \eh, \eH, \ev, and \eV are features that were added to Perl at
477 release 5.10. In contrast to the other sequences, which match only ASCII
478 characters by default, these always match certain high-valued codepoints,
479 whether or not PCRE_UCP is set. The horizontal space characters are:
480 .sp
481 U+0009 Horizontal tab
482 U+0020 Space
483 U+00A0 Non-break space
484 U+1680 Ogham space mark
485 U+180E Mongolian vowel separator
486 U+2000 En quad
487 U+2001 Em quad
488 U+2002 En space
489 U+2003 Em space
490 U+2004 Three-per-em space
491 U+2005 Four-per-em space
492 U+2006 Six-per-em space
493 U+2007 Figure space
494 U+2008 Punctuation space
495 U+2009 Thin space
496 U+200A Hair space
497 U+202F Narrow no-break space
498 U+205F Medium mathematical space
499 U+3000 Ideographic space
500 .sp
501 The vertical space characters are:
502 .sp
503 U+000A Linefeed
504 U+000B Vertical tab
505 U+000C Form feed
506 U+000D Carriage return
507 U+0085 Next line
508 U+2028 Line separator
509 U+2029 Paragraph separator
510 .sp
511 In 8-bit, non-UTF-8 mode, only the characters with codepoints less than 256 are
512 relevant.
513 .
514 .
515 .\" HTML <a name="newlineseq"></a>
516 .SS "Newline sequences"
517 .rs
518 .sp
519 Outside a character class, by default, the escape sequence \eR matches any
520 Unicode newline sequence. In 8-bit non-UTF-8 mode \eR is equivalent to the
521 following:
522 .sp
523 (?>\er\en|\en|\ex0b|\ef|\er|\ex85)
524 .sp
525 This is an example of an "atomic group", details of which are given
526 .\" HTML <a href="#atomicgroup">
527 .\" </a>
528 below.
529 .\"
530 This particular group matches either the two-character sequence CR followed by
531 LF, or one of the single characters LF (linefeed, U+000A), VT (vertical tab,
532 U+000B), FF (form feed, U+000C), CR (carriage return, U+000D), or NEL (next
533 line, U+0085). The two-character sequence is treated as a single unit that
534 cannot be split.
535 .P
536 In other modes, two additional characters whose codepoints are greater than 255
537 are added: LS (line separator, U+2028) and PS (paragraph separator, U+2029).
538 Unicode character property support is not needed for these characters to be
539 recognized.
540 .P
541 It is possible to restrict \eR to match only CR, LF, or CRLF (instead of the
542 complete set of Unicode line endings) by setting the option PCRE_BSR_ANYCRLF
543 either at compile time or when the pattern is matched. (BSR is an abbrevation
544 for "backslash R".) This can be made the default when PCRE is built; if this is
545 the case, the other behaviour can be requested via the PCRE_BSR_UNICODE option.
546 It is also possible to specify these settings by starting a pattern string with
547 one of the following sequences:
548 .sp
549 (*BSR_ANYCRLF) CR, LF, or CRLF only
550 (*BSR_UNICODE) any Unicode newline sequence
551 .sp
552 These override the default and the options given to the compiling function, but
553 they can themselves be overridden by options given to a matching function. Note
554 that these special settings, which are not Perl-compatible, are recognized only
555 at the very start of a pattern, and that they must be in upper case. If more
556 than one of them is present, the last one is used. They can be combined with a
557 change of newline convention; for example, a pattern can start with:
558 .sp
560 .sp
561 They can also be combined with the (*UTF8), (*UTF16), or (*UCP) special
562 sequences. Inside a character class, \eR is treated as an unrecognized escape
563 sequence, and so matches the letter "R" by default, but causes an error if
564 PCRE_EXTRA is set.
565 .
566 .
567 .\" HTML <a name="uniextseq"></a>
568 .SS Unicode character properties
569 .rs
570 .sp
571 When PCRE is built with Unicode character property support, three additional
572 escape sequences that match characters with specific properties are available.
573 When in 8-bit non-UTF-8 mode, these sequences are of course limited to testing
574 characters whose codepoints are less than 256, but they do work in this mode.
575 The extra escape sequences are:
576 .sp
577 \ep{\fIxx\fP} a character with the \fIxx\fP property
578 \eP{\fIxx\fP} a character without the \fIxx\fP property
579 \eX a Unicode extended grapheme cluster
580 .sp
581 The property names represented by \fIxx\fP above are limited to the Unicode
582 script names, the general category properties, "Any", which matches any
583 character (including newline), and some special PCRE properties (described
584 in the
585 .\" HTML <a href="#extraprops">
586 .\" </a>
587 next section).
588 .\"
589 Other Perl properties such as "InMusicalSymbols" are not currently supported by
590 PCRE. Note that \eP{Any} does not match any characters, so always causes a
591 match failure.
592 .P
593 Sets of Unicode characters are defined as belonging to certain scripts. A
594 character from one of these sets can be matched using a script name. For
595 example:
596 .sp
597 \ep{Greek}
598 \eP{Han}
599 .sp
600 Those that are not part of an identified script are lumped together as
601 "Common". The current list of scripts is:
602 .P
603 Arabic,
604 Armenian,
605 Avestan,
606 Balinese,
607 Bamum,
608 Batak,
609 Bengali,
610 Bopomofo,
611 Brahmi,
612 Braille,
613 Buginese,
614 Buhid,
615 Canadian_Aboriginal,
616 Carian,
617 Chakma,
618 Cham,
619 Cherokee,
620 Common,
621 Coptic,
622 Cuneiform,
623 Cypriot,
624 Cyrillic,
625 Deseret,
626 Devanagari,
627 Egyptian_Hieroglyphs,
628 Ethiopic,
629 Georgian,
630 Glagolitic,
631 Gothic,
632 Greek,
633 Gujarati,
634 Gurmukhi,
635 Han,
636 Hangul,
637 Hanunoo,
638 Hebrew,
639 Hiragana,
640 Imperial_Aramaic,
641 Inherited,
642 Inscriptional_Pahlavi,
643 Inscriptional_Parthian,
644 Javanese,
645 Kaithi,
646 Kannada,
647 Katakana,
648 Kayah_Li,
649 Kharoshthi,
650 Khmer,
651 Lao,
652 Latin,
653 Lepcha,
654 Limbu,
655 Linear_B,
656 Lisu,
657 Lycian,
658 Lydian,
659 Malayalam,
660 Mandaic,
661 Meetei_Mayek,
662 Meroitic_Cursive,
663 Meroitic_Hieroglyphs,
664 Miao,
665 Mongolian,
666 Myanmar,
667 New_Tai_Lue,
668 Nko,
669 Ogham,
670 Old_Italic,
671 Old_Persian,
672 Old_South_Arabian,
673 Old_Turkic,
674 Ol_Chiki,
675 Oriya,
676 Osmanya,
677 Phags_Pa,
678 Phoenician,
679 Rejang,
680 Runic,
681 Samaritan,
682 Saurashtra,
683 Sharada,
684 Shavian,
685 Sinhala,
686 Sora_Sompeng,
687 Sundanese,
688 Syloti_Nagri,
689 Syriac,
690 Tagalog,
691 Tagbanwa,
692 Tai_Le,
693 Tai_Tham,
694 Tai_Viet,
695 Takri,
696 Tamil,
697 Telugu,
698 Thaana,
699 Thai,
700 Tibetan,
701 Tifinagh,
702 Ugaritic,
703 Vai,
704 Yi.
705 .P
706 Each character has exactly one Unicode general category property, specified by
707 a two-letter abbreviation. For compatibility with Perl, negation can be
708 specified by including a circumflex between the opening brace and the property
709 name. For example, \ep{^Lu} is the same as \eP{Lu}.
710 .P
711 If only one letter is specified with \ep or \eP, it includes all the general
712 category properties that start with that letter. In this case, in the absence
713 of negation, the curly brackets in the escape sequence are optional; these two
714 examples have the same effect:
715 .sp
716 \ep{L}
717 \epL
718 .sp
719 The following general category property codes are supported:
720 .sp
721 C Other
722 Cc Control
723 Cf Format
724 Cn Unassigned
725 Co Private use
726 Cs Surrogate
727 .sp
728 L Letter
729 Ll Lower case letter
730 Lm Modifier letter
731 Lo Other letter
732 Lt Title case letter
733 Lu Upper case letter
734 .sp
735 M Mark
736 Mc Spacing mark
737 Me Enclosing mark
738 Mn Non-spacing mark
739 .sp
740 N Number
741 Nd Decimal number
742 Nl Letter number
743 No Other number
744 .sp
745 P Punctuation
746 Pc Connector punctuation
747 Pd Dash punctuation
748 Pe Close punctuation
749 Pf Final punctuation
750 Pi Initial punctuation
751 Po Other punctuation
752 Ps Open punctuation
753 .sp
754 S Symbol
755 Sc Currency symbol
756 Sk Modifier symbol
757 Sm Mathematical symbol
758 So Other symbol
759 .sp
760 Z Separator
761 Zl Line separator
762 Zp Paragraph separator
763 Zs Space separator
764 .sp
765 The special property L& is also supported: it matches a character that has
766 the Lu, Ll, or Lt property, in other words, a letter that is not classified as
767 a modifier or "other".
768 .P
769 The Cs (Surrogate) property applies only to characters in the range U+D800 to
770 U+DFFF. Such characters are not valid in Unicode strings and so
771 cannot be tested by PCRE, unless UTF validity checking has been turned off
772 (see the discussion of PCRE_NO_UTF8_CHECK and PCRE_NO_UTF16_CHECK in the
773 .\" HREF
774 \fBpcreapi\fP
775 .\"
776 page). Perl does not support the Cs property.
777 .P
778 The long synonyms for property names that Perl supports (such as \ep{Letter})
779 are not supported by PCRE, nor is it permitted to prefix any of these
780 properties with "Is".
781 .P
782 No character that is in the Unicode table has the Cn (unassigned) property.
783 Instead, this property is assumed for any code point that is not in the
784 Unicode table.
785 .P
786 Specifying caseless matching does not affect these escape sequences. For
787 example, \ep{Lu} always matches only upper case letters.
788 .P
789 Matching characters by Unicode property is not fast, because PCRE has to do a
790 multistage table lookup in order to find a character's property. That is why
791 the traditional escape sequences such as \ed and \ew do not use Unicode
792 properties in PCRE by default, though you can make them do so by setting the
793 PCRE_UCP option or by starting the pattern with (*UCP).
794 .
795 .
796 .SS Extended grapheme clusters
797 .rs
798 .sp
799 The \eX escape matches any number of Unicode characters that form an "extended
800 grapheme cluster", and treats the sequence as an atomic group
801 .\" HTML <a href="#atomicgroup">
802 .\" </a>
803 (see below).
804 .\"
805 Up to and including release 8.31, PCRE matched an earlier, simpler definition
806 that was equivalent to
807 .sp
808 (?>\ePM\epM*)
809 .sp
810 That is, it matched a character without the "mark" property, followed by zero
811 or more characters with the "mark" property. Characters with the "mark"
812 property are typically non-spacing accents that affect the preceding character.
813 .P
814 This simple definition was extended in Unicode to include more complicated
815 kinds of composite character by giving each character a grapheme breaking
816 property, and creating rules that use these properties to define the boundaries
817 of extended grapheme clusters. In releases of PCRE later than 8.31, \eX matches
818 one of these clusters.
819 .P
820 \eX always matches at least one character. Then it decides whether to add
821 additional characters according to the following rules for ending a cluster:
822 .P
823 1. End at the end of the subject string.
824 .P
825 2. Do not end between CR and LF; otherwise end after any control character.
826 .P
827 3. Do not break Hangul (a Korean script) syllable sequences. Hangul characters
828 are of five types: L, V, T, LV, and LVT. An L character may be followed by an
829 L, V, LV, or LVT character; an LV or V character may be followed by a V or T
830 character; an LVT or T character may be follwed only by a T character.
831 .P
832 4. Do not end before extending characters or spacing marks. Characters with
833 the "mark" property always have the "extend" grapheme breaking property.
834 .P
835 5. Do not end after prepend characters.
836 .P
837 6. Otherwise, end the cluster.
838 .
839 .
840 .\" HTML <a name="extraprops"></a>
841 .SS PCRE's additional properties
842 .rs
843 .sp
844 As well as the standard Unicode properties described above, PCRE supports four
845 more that make it possible to convert traditional escape sequences such as \ew
846 and \es and POSIX character classes to use Unicode properties. PCRE uses these
847 non-standard, non-Perl properties internally when PCRE_UCP is set. They are:
848 .sp
849 Xan Any alphanumeric character
850 Xps Any POSIX space character
851 Xsp Any Perl space character
852 Xwd Any Perl "word" character
853 .sp
854 Xan matches characters that have either the L (letter) or the N (number)
855 property. Xps matches the characters tab, linefeed, vertical tab, form feed, or
856 carriage return, and any other character that has the Z (separator) property.
857 Xsp is the same as Xps, except that vertical tab is excluded. Xwd matches the
858 same characters as Xan, plus underscore.
859 .
860 .
861 .\" HTML <a name="resetmatchstart"></a>
862 .SS "Resetting the match start"
863 .rs
864 .sp
865 The escape sequence \eK causes any previously matched characters not to be
866 included in the final matched sequence. For example, the pattern:
867 .sp
868 foo\eKbar
869 .sp
870 matches "foobar", but reports that it has matched "bar". This feature is
871 similar to a lookbehind assertion
872 .\" HTML <a href="#lookbehind">
873 .\" </a>
874 (described below).
875 .\"
876 However, in this case, the part of the subject before the real match does not
877 have to be of fixed length, as lookbehind assertions do. The use of \eK does
878 not interfere with the setting of
879 .\" HTML <a href="#subpattern">
880 .\" </a>
881 captured substrings.
882 .\"
883 For example, when the pattern
884 .sp
885 (foo)\eKbar
886 .sp
887 matches "foobar", the first substring is still set to "foo".
888 .P
889 Perl documents that the use of \eK within assertions is "not well defined". In
890 PCRE, \eK is acted upon when it occurs inside positive assertions, but is
891 ignored in negative assertions.
892 .
893 .
894 .\" HTML <a name="smallassertions"></a>
895 .SS "Simple assertions"
896 .rs
897 .sp
898 The final use of backslash is for certain simple assertions. An assertion
899 specifies a condition that has to be met at a particular point in a match,
900 without consuming any characters from the subject string. The use of
901 subpatterns for more complicated assertions is described
902 .\" HTML <a href="#bigassertions">
903 .\" </a>
904 below.
905 .\"
906 The backslashed assertions are:
907 .sp
908 \eb matches at a word boundary
909 \eB matches when not at a word boundary
910 \eA matches at the start of the subject
911 \eZ matches at the end of the subject
912 also matches before a newline at the end of the subject
913 \ez matches only at the end of the subject
914 \eG matches at the first matching position in the subject
915 .sp
916 Inside a character class, \eb has a different meaning; it matches the backspace
917 character. If any other of these assertions appears in a character class, by
918 default it matches the corresponding literal character (for example, \eB
919 matches the letter B). However, if the PCRE_EXTRA option is set, an "invalid
920 escape sequence" error is generated instead.
921 .P
922 A word boundary is a position in the subject string where the current character
923 and the previous character do not both match \ew or \eW (i.e. one matches
924 \ew and the other matches \eW), or the start or end of the string if the
925 first or last character matches \ew, respectively. In a UTF mode, the meanings
926 of \ew and \eW can be changed by setting the PCRE_UCP option. When this is
927 done, it also affects \eb and \eB. Neither PCRE nor Perl has a separate "start
928 of word" or "end of word" metasequence. However, whatever follows \eb normally
929 determines which it is. For example, the fragment \eba matches "a" at the start
930 of a word.
931 .P
932 The \eA, \eZ, and \ez assertions differ from the traditional circumflex and
933 dollar (described in the next section) in that they only ever match at the very
934 start and end of the subject string, whatever options are set. Thus, they are
935 independent of multiline mode. These three assertions are not affected by the
936 PCRE_NOTBOL or PCRE_NOTEOL options, which affect only the behaviour of the
937 circumflex and dollar metacharacters. However, if the \fIstartoffset\fP
938 argument of \fBpcre_exec()\fP is non-zero, indicating that matching is to start
939 at a point other than the beginning of the subject, \eA can never match. The
940 difference between \eZ and \ez is that \eZ matches before a newline at the end
941 of the string as well as at the very end, whereas \ez matches only at the end.
942 .P
943 The \eG assertion is true only when the current matching position is at the
944 start point of the match, as specified by the \fIstartoffset\fP argument of
945 \fBpcre_exec()\fP. It differs from \eA when the value of \fIstartoffset\fP is
946 non-zero. By calling \fBpcre_exec()\fP multiple times with appropriate
947 arguments, you can mimic Perl's /g option, and it is in this kind of
948 implementation where \eG can be useful.
949 .P
950 Note, however, that PCRE's interpretation of \eG, as the start of the current
951 match, is subtly different from Perl's, which defines it as the end of the
952 previous match. In Perl, these can be different when the previously matched
953 string was empty. Because PCRE does just one match at a time, it cannot
954 reproduce this behaviour.
955 .P
956 If all the alternatives of a pattern begin with \eG, the expression is anchored
957 to the starting match position, and the "anchored" flag is set in the compiled
958 regular expression.
959 .
960 .
962 .rs
963 .sp
964 Outside a character class, in the default matching mode, the circumflex
965 character is an assertion that is true only if the current matching point is
966 at the start of the subject string. If the \fIstartoffset\fP argument of
967 \fBpcre_exec()\fP is non-zero, circumflex can never match if the PCRE_MULTILINE
968 option is unset. Inside a character class, circumflex has an entirely different
969 meaning
970 .\" HTML <a href="#characterclass">
971 .\" </a>
972 (see below).
973 .\"
974 .P
975 Circumflex need not be the first character of the pattern if a number of
976 alternatives are involved, but it should be the first thing in each alternative
977 in which it appears if the pattern is ever to match that branch. If all
978 possible alternatives start with a circumflex, that is, if the pattern is
979 constrained to match only at the start of the subject, it is said to be an
980 "anchored" pattern. (There are also other constructs that can cause a pattern
981 to be anchored.)
982 .P
983 A dollar character is an assertion that is true only if the current matching
984 point is at the end of the subject string, or immediately before a newline
985 at the end of the string (by default). Dollar need not be the last character of
986 the pattern if a number of alternatives are involved, but it should be the last
987 item in any branch in which it appears. Dollar has no special meaning in a
988 character class.
989 .P
990 The meaning of dollar can be changed so that it matches only at the very end of
991 the string, by setting the PCRE_DOLLAR_ENDONLY option at compile time. This
992 does not affect the \eZ assertion.
993 .P
994 The meanings of the circumflex and dollar characters are changed if the
995 PCRE_MULTILINE option is set. When this is the case, a circumflex matches
996 immediately after internal newlines as well as at the start of the subject
997 string. It does not match after a newline that ends the string. A dollar
998 matches before any newlines in the string, as well as at the very end, when
999 PCRE_MULTILINE is set. When newline is specified as the two-character
1000 sequence CRLF, isolated CR and LF characters do not indicate newlines.
1001 .P
1002 For example, the pattern /^abc$/ matches the subject string "def\enabc" (where
1003 \en represents a newline) in multiline mode, but not otherwise. Consequently,
1004 patterns that are anchored in single line mode because all branches start with
1005 ^ are not anchored in multiline mode, and a match for circumflex is possible
1006 when the \fIstartoffset\fP argument of \fBpcre_exec()\fP is non-zero. The
1007 PCRE_DOLLAR_ENDONLY option is ignored if PCRE_MULTILINE is set.
1008 .P
1009 Note that the sequences \eA, \eZ, and \ez can be used to match the start and
1010 end of the subject in both modes, and if all branches of a pattern start with
1011 \eA it is always anchored, whether or not PCRE_MULTILINE is set.
1012 .
1013 .
1014 .\" HTML <a name="fullstopdot"></a>
1016 .rs
1017 .sp
1018 Outside a character class, a dot in the pattern matches any one character in
1019 the subject string except (by default) a character that signifies the end of a
1020 line.
1021 .P
1022 When a line ending is defined as a single character, dot never matches that
1023 character; when the two-character sequence CRLF is used, dot does not match CR
1024 if it is immediately followed by LF, but otherwise it matches all characters
1025 (including isolated CRs and LFs). When any Unicode line endings are being
1026 recognized, dot does not match CR or LF or any of the other line ending
1027 characters.
1028 .P
1029 The behaviour of dot with regard to newlines can be changed. If the PCRE_DOTALL
1030 option is set, a dot matches any one character, without exception. If the
1031 two-character sequence CRLF is present in the subject string, it takes two dots
1032 to match it.
1033 .P
1034 The handling of dot is entirely independent of the handling of circumflex and
1035 dollar, the only relationship being that they both involve newlines. Dot has no
1036 special meaning in a character class.
1037 .P
1038 The escape sequence \eN behaves like a dot, except that it is not affected by
1039 the PCRE_DOTALL option. In other words, it matches any character except one
1040 that signifies the end of a line. Perl also uses \eN to match characters by
1041 name; PCRE does not support this.
1042 .
1043 .
1045 .rs
1046 .sp
1047 Outside a character class, the escape sequence \eC matches any one data unit,
1048 whether or not a UTF mode is set. In the 8-bit library, one data unit is one
1049 byte; in the 16-bit library it is a 16-bit unit. Unlike a dot, \eC always
1050 matches line-ending characters. The feature is provided in Perl in order to
1051 match individual bytes in UTF-8 mode, but it is unclear how it can usefully be
1052 used. Because \eC breaks up characters into individual data units, matching one
1053 unit with \eC in a UTF mode means that the rest of the string may start with a
1054 malformed UTF character. This has undefined results, because PCRE assumes that
1055 it is dealing with valid UTF strings (and by default it checks this at the
1056 start of processing unless the PCRE_NO_UTF8_CHECK or PCRE_NO_UTF16_CHECK option
1057 is used).
1058 .P
1059 PCRE does not allow \eC to appear in lookbehind assertions
1060 .\" HTML <a href="#lookbehind">
1061 .\" </a>
1062 (described below)
1063 .\"
1064 in a UTF mode, because this would make it impossible to calculate the length of
1065 the lookbehind.
1066 .P
1067 In general, the \eC escape sequence is best avoided. However, one
1068 way of using it that avoids the problem of malformed UTF characters is to use a
1069 lookahead to check the length of the next character, as in this pattern, which
1070 could be used with a UTF-8 string (ignore white space and line breaks):
1071 .sp
1072 (?| (?=[\ex00-\ex7f])(\eC) |
1073 (?=[\ex80-\ex{7ff}])(\eC)(\eC) |
1074 (?=[\ex{800}-\ex{ffff}])(\eC)(\eC)(\eC) |
1075 (?=[\ex{10000}-\ex{1fffff}])(\eC)(\eC)(\eC)(\eC))
1076 .sp
1077 A group that starts with (?| resets the capturing parentheses numbers in each
1078 alternative (see
1079 .\" HTML <a href="#dupsubpatternnumber">
1080 .\" </a>
1081 "Duplicate Subpattern Numbers"
1082 .\"
1083 below). The assertions at the start of each branch check the next UTF-8
1084 character for values whose encoding uses 1, 2, 3, or 4 bytes, respectively. The
1085 character's individual bytes are then captured by the appropriate number of
1086 groups.
1087 .
1088 .
1089 .\" HTML <a name="characterclass"></a>
1091 .rs
1092 .sp
1093 An opening square bracket introduces a character class, terminated by a closing
1094 square bracket. A closing square bracket on its own is not special by default.
1095 However, if the PCRE_JAVASCRIPT_COMPAT option is set, a lone closing square
1096 bracket causes a compile-time error. If a closing square bracket is required as
1097 a member of the class, it should be the first data character in the class
1098 (after an initial circumflex, if present) or escaped with a backslash.
1099 .P
1100 A character class matches a single character in the subject. In a UTF mode, the
1101 character may be more than one data unit long. A matched character must be in
1102 the set of characters defined by the class, unless the first character in the
1103 class definition is a circumflex, in which case the subject character must not
1104 be in the set defined by the class. If a circumflex is actually required as a
1105 member of the class, ensure it is not the first character, or escape it with a
1106 backslash.
1107 .P
1108 For example, the character class [aeiou] matches any lower case vowel, while
1109 [^aeiou] matches any character that is not a lower case vowel. Note that a
1110 circumflex is just a convenient notation for specifying the characters that
1111 are in the class by enumerating those that are not. A class that starts with a
1112 circumflex is not an assertion; it still consumes a character from the subject
1113 string, and therefore it fails if the current pointer is at the end of the
1114 string.
1115 .P
1116 In UTF-8 (UTF-16) mode, characters with values greater than 255 (0xffff) can be
1117 included in a class as a literal string of data units, or by using the \ex{
1118 escaping mechanism.
1119 .P
1120 When caseless matching is set, any letters in a class represent both their
1121 upper case and lower case versions, so for example, a caseless [aeiou] matches
1122 "A" as well as "a", and a caseless [^aeiou] does not match "A", whereas a
1123 caseful version would. In a UTF mode, PCRE always understands the concept of
1124 case for characters whose values are less than 128, so caseless matching is
1125 always possible. For characters with higher values, the concept of case is
1126 supported if PCRE is compiled with Unicode property support, but not otherwise.
1127 If you want to use caseless matching in a UTF mode for characters 128 and
1128 above, you must ensure that PCRE is compiled with Unicode property support as
1129 well as with UTF support.
1130 .P
1131 Characters that might indicate line breaks are never treated in any special way
1132 when matching character classes, whatever line-ending sequence is in use, and
1133 whatever setting of the PCRE_DOTALL and PCRE_MULTILINE options is used. A class
1134 such as [^a] always matches one of these characters.
1135 .P
1136 The minus (hyphen) character can be used to specify a range of characters in a
1137 character class. For example, [d-m] matches any letter between d and m,
1138 inclusive. If a minus character is required in a class, it must be escaped with
1139 a backslash or appear in a position where it cannot be interpreted as
1140 indicating a range, typically as the first or last character in the class.
1141 .P
1142 It is not possible to have the literal character "]" as the end character of a
1143 range. A pattern such as [W-]46] is interpreted as a class of two characters
1144 ("W" and "-") followed by a literal string "46]", so it would match "W46]" or
1145 "-46]". However, if the "]" is escaped with a backslash it is interpreted as
1146 the end of range, so [W-\e]46] is interpreted as a class containing a range
1147 followed by two other characters. The octal or hexadecimal representation of
1148 "]" can also be used to end a range.
1149 .P
1150 Ranges operate in the collating sequence of character values. They can also be
1151 used for characters specified numerically, for example [\e000-\e037]. Ranges
1152 can include any characters that are valid for the current mode.
1153 .P
1154 If a range that includes letters is used when caseless matching is set, it
1155 matches the letters in either case. For example, [W-c] is equivalent to
1156 [][\e\e^_`wxyzabc], matched caselessly, and in a non-UTF mode, if character
1157 tables for a French locale are in use, [\exc8-\excb] matches accented E
1158 characters in both cases. In UTF modes, PCRE supports the concept of case for
1159 characters with values greater than 128 only when it is compiled with Unicode
1160 property support.
1161 .P
1162 The character escape sequences \ed, \eD, \eh, \eH, \ep, \eP, \es, \eS, \ev,
1163 \eV, \ew, and \eW may appear in a character class, and add the characters that
1164 they match to the class. For example, [\edABCDEF] matches any hexadecimal
1165 digit. In UTF modes, the PCRE_UCP option affects the meanings of \ed, \es, \ew
1166 and their upper case partners, just as it does when they appear outside a
1167 character class, as described in the section entitled
1168 .\" HTML <a href="#genericchartypes">
1169 .\" </a>
1170 "Generic character types"
1171 .\"
1172 above. The escape sequence \eb has a different meaning inside a character
1173 class; it matches the backspace character. The sequences \eB, \eN, \eR, and \eX
1174 are not special inside a character class. Like any other unrecognized escape
1175 sequences, they are treated as the literal characters "B", "N", "R", and "X" by
1176 default, but cause an error if the PCRE_EXTRA option is set.
1177 .P
1178 A circumflex can conveniently be used with the upper case character types to
1179 specify a more restricted set of characters than the matching lower case type.
1180 For example, the class [^\eW_] matches any letter or digit, but not underscore,
1181 whereas [\ew] includes underscore. A positive character class should be read as
1182 "something OR something OR ..." and a negative class as "NOT something AND NOT
1183 something AND NOT ...".
1184 .P
1185 The only metacharacters that are recognized in character classes are backslash,
1186 hyphen (only where it can be interpreted as specifying a range), circumflex
1187 (only at the start), opening square bracket (only when it can be interpreted as
1188 introducing a POSIX class name - see the next section), and the terminating
1189 closing square bracket. However, escaping other non-alphanumeric characters
1190 does no harm.
1191 .
1192 .
1194 .rs
1195 .sp
1196 Perl supports the POSIX notation for character classes. This uses names
1197 enclosed by [: and :] within the enclosing square brackets. PCRE also supports
1198 this notation. For example,
1199 .sp
1200 [01[:alpha:]%]
1201 .sp
1202 matches "0", "1", any alphabetic character, or "%". The supported class names
1203 are:
1204 .sp
1205 alnum letters and digits
1206 alpha letters
1207 ascii character codes 0 - 127
1208 blank space or tab only
1209 cntrl control characters
1210 digit decimal digits (same as \ed)
1211 graph printing characters, excluding space
1212 lower lower case letters
1213 print printing characters, including space
1214 punct printing characters, excluding letters and digits and space
1215 space white space (not quite the same as \es)
1216 upper upper case letters
1217 word "word" characters (same as \ew)
1218 xdigit hexadecimal digits
1219 .sp
1220 The "space" characters are HT (9), LF (10), VT (11), FF (12), CR (13), and
1221 space (32). Notice that this list includes the VT character (code 11). This
1222 makes "space" different to \es, which does not include VT (for Perl
1223 compatibility).
1224 .P
1225 The name "word" is a Perl extension, and "blank" is a GNU extension from Perl
1226 5.8. Another Perl extension is negation, which is indicated by a ^ character
1227 after the colon. For example,
1228 .sp
1229 [12[:^digit:]]
1230 .sp
1231 matches "1", "2", or any non-digit. PCRE (and Perl) also recognize the POSIX
1232 syntax [.ch.] and [=ch=] where "ch" is a "collating element", but these are not
1233 supported, and an error is given if they are encountered.
1234 .P
1235 By default, in UTF modes, characters with values greater than 128 do not match
1236 any of the POSIX character classes. However, if the PCRE_UCP option is passed
1237 to \fBpcre_compile()\fP, some of the classes are changed so that Unicode
1238 character properties are used. This is achieved by replacing the POSIX classes
1239 by other sequences, as follows:
1240 .sp
1241 [:alnum:] becomes \ep{Xan}
1242 [:alpha:] becomes \ep{L}
1243 [:blank:] becomes \eh
1244 [:digit:] becomes \ep{Nd}
1245 [:lower:] becomes \ep{Ll}
1246 [:space:] becomes \ep{Xps}
1247 [:upper:] becomes \ep{Lu}
1248 [:word:] becomes \ep{Xwd}
1249 .sp
1250 Negated versions, such as [:^alpha:] use \eP instead of \ep. The other POSIX
1251 classes are unchanged, and match only characters with code points less than
1252 128.
1253 .
1254 .
1256 .rs
1257 .sp
1258 Vertical bar characters are used to separate alternative patterns. For example,
1259 the pattern
1260 .sp
1261 gilbert|sullivan
1262 .sp
1263 matches either "gilbert" or "sullivan". Any number of alternatives may appear,
1264 and an empty alternative is permitted (matching the empty string). The matching
1265 process tries each alternative in turn, from left to right, and the first one
1266 that succeeds is used. If the alternatives are within a subpattern
1267 .\" HTML <a href="#subpattern">
1268 .\" </a>
1269 (defined below),
1270 .\"
1271 "succeeds" means matching the rest of the main pattern as well as the
1272 alternative in the subpattern.
1273 .
1274 .
1276 .rs
1277 .sp
1278 The settings of the PCRE_CASELESS, PCRE_MULTILINE, PCRE_DOTALL, and
1279 PCRE_EXTENDED options (which are Perl-compatible) can be changed from within
1280 the pattern by a sequence of Perl option letters enclosed between "(?" and ")".
1281 The option letters are
1282 .sp
1283 i for PCRE_CASELESS
1285 s for PCRE_DOTALL
1286 x for PCRE_EXTENDED
1287 .sp
1288 For example, (?im) sets caseless, multiline matching. It is also possible to
1289 unset these options by preceding the letter with a hyphen, and a combined
1290 setting and unsetting such as (?im-sx), which sets PCRE_CASELESS and
1291 PCRE_MULTILINE while unsetting PCRE_DOTALL and PCRE_EXTENDED, is also
1292 permitted. If a letter appears both before and after the hyphen, the option is
1293 unset.
1294 .P
1295 The PCRE-specific options PCRE_DUPNAMES, PCRE_UNGREEDY, and PCRE_EXTRA can be
1296 changed in the same way as the Perl-compatible options by using the characters
1297 J, U and X respectively.
1298 .P
1299 When one of these option changes occurs at top level (that is, not inside
1300 subpattern parentheses), the change applies to the remainder of the pattern
1301 that follows. If the change is placed right at the start of a pattern, PCRE
1302 extracts it into the global options (and it will therefore show up in data
1303 extracted by the \fBpcre_fullinfo()\fP function).
1304 .P
1305 An option change within a subpattern (see below for a description of
1306 subpatterns) affects only that part of the subpattern that follows it, so
1307 .sp
1308 (a(?i)b)c
1309 .sp
1310 matches abc and aBc and no other strings (assuming PCRE_CASELESS is not used).
1311 By this means, options can be made to have different settings in different
1312 parts of the pattern. Any changes made in one alternative do carry on
1313 into subsequent branches within the same subpattern. For example,
1314 .sp
1315 (a(?i)b|c)
1316 .sp
1317 matches "ab", "aB", "c", and "C", even though when matching "C" the first
1318 branch is abandoned before the option setting. This is because the effects of
1319 option settings happen at compile time. There would be some very weird
1320 behaviour otherwise.
1321 .P
1322 \fBNote:\fP There are other PCRE-specific options that can be set by the
1323 application when the compiling or matching functions are called. In some cases
1324 the pattern can contain special leading sequences such as (*CRLF) to override
1325 what the application has set or what has been defaulted. Details are given in
1326 the section entitled
1327 .\" HTML <a href="#newlineseq">
1328 .\" </a>
1329 "Newline sequences"
1330 .\"
1331 above. There are also the (*UTF8), (*UTF16), and (*UCP) leading sequences that
1332 can be used to set UTF and Unicode property modes; they are equivalent to
1333 setting the PCRE_UTF8, PCRE_UTF16, and the PCRE_UCP options, respectively.
1334 .
1335 .
1336 .\" HTML <a name="subpattern"></a>
1338 .rs
1339 .sp
1340 Subpatterns are delimited by parentheses (round brackets), which can be nested.
1341 Turning part of a pattern into a subpattern does two things:
1342 .sp
1343 1. It localizes a set of alternatives. For example, the pattern
1344 .sp
1345 cat(aract|erpillar|)
1346 .sp
1347 matches "cataract", "caterpillar", or "cat". Without the parentheses, it would
1348 match "cataract", "erpillar" or an empty string.
1349 .sp
1350 2. It sets up the subpattern as a capturing subpattern. This means that, when
1351 the whole pattern matches, that portion of the subject string that matched the
1352 subpattern is passed back to the caller via the \fIovector\fP argument of the
1353 matching function. (This applies only to the traditional matching functions;
1354 the DFA matching functions do not support capturing.)
1355 .P
1356 Opening parentheses are counted from left to right (starting from 1) to obtain
1357 numbers for the capturing subpatterns. For example, if the string "the red
1358 king" is matched against the pattern
1359 .sp
1360 the ((red|white) (king|queen))
1361 .sp
1362 the captured substrings are "red king", "red", and "king", and are numbered 1,
1363 2, and 3, respectively.
1364 .P
1365 The fact that plain parentheses fulfil two functions is not always helpful.
1366 There are often times when a grouping subpattern is required without a
1367 capturing requirement. If an opening parenthesis is followed by a question mark
1368 and a colon, the subpattern does not do any capturing, and is not counted when
1369 computing the number of any subsequent capturing subpatterns. For example, if
1370 the string "the white queen" is matched against the pattern
1371 .sp
1372 the ((?:red|white) (king|queen))
1373 .sp
1374 the captured substrings are "white queen" and "queen", and are numbered 1 and
1375 2. The maximum number of capturing subpatterns is 65535.
1376 .P
1377 As a convenient shorthand, if any option settings are required at the start of
1378 a non-capturing subpattern, the option letters may appear between the "?" and
1379 the ":". Thus the two patterns
1380 .sp
1381 (?i:saturday|sunday)
1382 (?:(?i)saturday|sunday)
1383 .sp
1384 match exactly the same set of strings. Because alternative branches are tried
1385 from left to right, and options are not reset until the end of the subpattern
1386 is reached, an option setting in one branch does affect subsequent branches, so
1387 the above patterns match "SUNDAY" as well as "Saturday".
1388 .
1389 .
1390 .\" HTML <a name="dupsubpatternnumber"></a>
1392 .rs
1393 .sp
1394 Perl 5.10 introduced a feature whereby each alternative in a subpattern uses
1395 the same numbers for its capturing parentheses. Such a subpattern starts with
1396 (?| and is itself a non-capturing subpattern. For example, consider this
1397 pattern:
1398 .sp
1399 (?|(Sat)ur|(Sun))day
1400 .sp
1401 Because the two alternatives are inside a (?| group, both sets of capturing
1402 parentheses are numbered one. Thus, when the pattern matches, you can look
1403 at captured substring number one, whichever alternative matched. This construct
1404 is useful when you want to capture part, but not all, of one of a number of
1405 alternatives. Inside a (?| group, parentheses are numbered as usual, but the
1406 number is reset at the start of each branch. The numbers of any capturing
1407 parentheses that follow the subpattern start after the highest number used in
1408 any branch. The following example is taken from the Perl documentation. The
1409 numbers underneath show in which buffer the captured content will be stored.
1410 .sp
1411 # before ---------------branch-reset----------- after
1412 / ( a ) (?| x ( y ) z | (p (q) r) | (t) u (v) ) ( z ) /x
1413 # 1 2 2 3 2 3 4
1414 .sp
1415 A back reference to a numbered subpattern uses the most recent value that is
1416 set for that number by any subpattern. The following pattern matches "abcabc"
1417 or "defdef":
1418 .sp
1419 /(?|(abc)|(def))\e1/
1420 .sp
1421 In contrast, a subroutine call to a numbered subpattern always refers to the
1422 first one in the pattern with the given number. The following pattern matches
1423 "abcabc" or "defabc":
1424 .sp
1425 /(?|(abc)|(def))(?1)/
1426 .sp
1427 If a
1428 .\" HTML <a href="#conditions">
1429 .\" </a>
1430 condition test
1431 .\"
1432 for a subpattern's having matched refers to a non-unique number, the test is
1433 true if any of the subpatterns of that number have matched.
1434 .P
1435 An alternative approach to using this "branch reset" feature is to use
1436 duplicate named subpatterns, as described in the next section.
1437 .
1438 .
1440 .rs
1441 .sp
1442 Identifying capturing parentheses by number is simple, but it can be very hard
1443 to keep track of the numbers in complicated regular expressions. Furthermore,
1444 if an expression is modified, the numbers may change. To help with this
1445 difficulty, PCRE supports the naming of subpatterns. This feature was not
1446 added to Perl until release 5.10. Python had the feature earlier, and PCRE
1447 introduced it at release 4.0, using the Python syntax. PCRE now supports both
1448 the Perl and the Python syntax. Perl allows identically numbered subpatterns to
1449 have different names, but PCRE does not.
1450 .P
1451 In PCRE, a subpattern can be named in one of three ways: (?<name>...) or
1452 (?'name'...) as in Perl, or (?P<name>...) as in Python. References to capturing
1453 parentheses from other parts of the pattern, such as
1454 .\" HTML <a href="#backreferences">
1455 .\" </a>
1456 back references,
1457 .\"
1458 .\" HTML <a href="#recursion">
1459 .\" </a>
1460 recursion,
1461 .\"
1462 and
1463 .\" HTML <a href="#conditions">
1464 .\" </a>
1465 conditions,
1466 .\"
1467 can be made by name as well as by number.
1468 .P
1469 Names consist of up to 32 alphanumeric characters and underscores. Named
1470 capturing parentheses are still allocated numbers as well as names, exactly as
1471 if the names were not present. The PCRE API provides function calls for
1472 extracting the name-to-number translation table from a compiled pattern. There
1473 is also a convenience function for extracting a captured substring by name.
1474 .P
1475 By default, a name must be unique within a pattern, but it is possible to relax
1476 this constraint by setting the PCRE_DUPNAMES option at compile time. (Duplicate
1477 names are also always permitted for subpatterns with the same number, set up as
1478 described in the previous section.) Duplicate names can be useful for patterns
1479 where only one instance of the named parentheses can match. Suppose you want to
1480 match the name of a weekday, either as a 3-letter abbreviation or as the full
1481 name, and in both cases you want to extract the abbreviation. This pattern
1482 (ignoring the line breaks) does the job:
1483 .sp
1484 (?<DN>Mon|Fri|Sun)(?:day)?|
1485 (?<DN>Tue)(?:sday)?|
1486 (?<DN>Wed)(?:nesday)?|
1487 (?<DN>Thu)(?:rsday)?|
1488 (?<DN>Sat)(?:urday)?
1489 .sp
1490 There are five capturing substrings, but only one is ever set after a match.
1491 (An alternative way of solving this problem is to use a "branch reset"
1492 subpattern, as described in the previous section.)
1493 .P
1494 The convenience function for extracting the data by name returns the substring
1495 for the first (and in this example, the only) subpattern of that name that
1496 matched. This saves searching to find which numbered subpattern it was.
1497 .P
1498 If you make a back reference to a non-unique named subpattern from elsewhere in
1499 the pattern, the one that corresponds to the first occurrence of the name is
1500 used. In the absence of duplicate numbers (see the previous section) this is
1501 the one with the lowest number. If you use a named reference in a condition
1502 test (see the
1503 .\"
1504 .\" HTML <a href="#conditions">
1505 .\" </a>
1506 section about conditions
1507 .\"
1508 below), either to check whether a subpattern has matched, or to check for
1509 recursion, all subpatterns with the same name are tested. If the condition is
1510 true for any one of them, the overall condition is true. This is the same
1511 behaviour as testing by number. For further details of the interfaces for
1512 handling named subpatterns, see the
1513 .\" HREF
1514 \fBpcreapi\fP
1515 .\"
1516 documentation.
1517 .P
1518 \fBWarning:\fP You cannot use different names to distinguish between two
1519 subpatterns with the same number because PCRE uses only the numbers when
1520 matching. For this reason, an error is given at compile time if different names
1521 are given to subpatterns with the same number. However, you can give the same
1522 name to subpatterns with the same number, even when PCRE_DUPNAMES is not set.
1523 .
1524 .
1526 .rs
1527 .sp
1528 Repetition is specified by quantifiers, which can follow any of the following
1529 items:
1530 .sp
1531 a literal data character
1532 the dot metacharacter
1533 the \eC escape sequence
1534 the \eX escape sequence
1535 the \eR escape sequence
1536 an escape such as \ed or \epL that matches a single character
1537 a character class
1538 a back reference (see next section)
1539 a parenthesized subpattern (including assertions)
1540 a subroutine call to a subpattern (recursive or otherwise)
1541 .sp
1542 The general repetition quantifier specifies a minimum and maximum number of
1543 permitted matches, by giving the two numbers in curly brackets (braces),
1544 separated by a comma. The numbers must be less than 65536, and the first must
1545 be less than or equal to the second. For example:
1546 .sp
1547 z{2,4}
1548 .sp
1549 matches "zz", "zzz", or "zzzz". A closing brace on its own is not a special
1550 character. If the second number is omitted, but the comma is present, there is
1551 no upper limit; if the second number and the comma are both omitted, the
1552 quantifier specifies an exact number of required matches. Thus
1553 .sp
1554 [aeiou]{3,}
1555 .sp
1556 matches at least 3 successive vowels, but may match many more, while
1557 .sp
1558 \ed{8}
1559 .sp
1560 matches exactly 8 digits. An opening curly bracket that appears in a position
1561 where a quantifier is not allowed, or one that does not match the syntax of a
1562 quantifier, is taken as a literal character. For example, {,6} is not a
1563 quantifier, but a literal string of four characters.
1564 .P
1565 In UTF modes, quantifiers apply to characters rather than to individual data
1566 units. Thus, for example, \ex{100}{2} matches two characters, each of
1567 which is represented by a two-byte sequence in a UTF-8 string. Similarly,
1568 \eX{3} matches three Unicode extended grapheme clusters, each of which may be
1569 several data units long (and they may be of different lengths).
1570 .P
1571 The quantifier {0} is permitted, causing the expression to behave as if the
1572 previous item and the quantifier were not present. This may be useful for
1573 subpatterns that are referenced as
1574 .\" HTML <a href="#subpatternsassubroutines">
1575 .\" </a>
1576 subroutines
1577 .\"
1578 from elsewhere in the pattern (but see also the section entitled
1579 .\" HTML <a href="#subdefine">
1580 .\" </a>
1581 "Defining subpatterns for use by reference only"
1582 .\"
1583 below). Items other than subpatterns that have a {0} quantifier are omitted
1584 from the compiled pattern.
1585 .P
1586 For convenience, the three most common quantifiers have single-character
1587 abbreviations:
1588 .sp
1589 * is equivalent to {0,}
1590 + is equivalent to {1,}
1591 ? is equivalent to {0,1}
1592 .sp
1593 It is possible to construct infinite loops by following a subpattern that can
1594 match no characters with a quantifier that has no upper limit, for example:
1595 .sp
1596 (a?)*
1597 .sp
1598 Earlier versions of Perl and PCRE used to give an error at compile time for
1599 such patterns. However, because there are cases where this can be useful, such
1600 patterns are now accepted, but if any repetition of the subpattern does in fact
1601 match no characters, the loop is forcibly broken.
1602 .P
1603 By default, the quantifiers are "greedy", that is, they match as much as
1604 possible (up to the maximum number of permitted times), without causing the
1605 rest of the pattern to fail. The classic example of where this gives problems
1606 is in trying to match comments in C programs. These appear between /* and */
1607 and within the comment, individual * and / characters may appear. An attempt to
1608 match C comments by applying the pattern
1609 .sp
1610 /\e*.*\e*/
1611 .sp
1612 to the string
1613 .sp
1614 /* first comment */ not comment /* second comment */
1615 .sp
1616 fails, because it matches the entire string owing to the greediness of the .*
1617 item.
1618 .P
1619 However, if a quantifier is followed by a question mark, it ceases to be
1620 greedy, and instead matches the minimum number of times possible, so the
1621 pattern
1622 .sp
1623 /\e*.*?\e*/
1624 .sp
1625 does the right thing with the C comments. The meaning of the various
1626 quantifiers is not otherwise changed, just the preferred number of matches.
1627 Do not confuse this use of question mark with its use as a quantifier in its
1628 own right. Because it has two uses, it can sometimes appear doubled, as in
1629 .sp
1630 \ed??\ed
1631 .sp
1632 which matches one digit by preference, but can match two if that is the only
1633 way the rest of the pattern matches.
1634 .P
1635 If the PCRE_UNGREEDY option is set (an option that is not available in Perl),
1636 the quantifiers are not greedy by default, but individual ones can be made
1637 greedy by following them with a question mark. In other words, it inverts the
1638 default behaviour.
1639 .P
1640 When a parenthesized subpattern is quantified with a minimum repeat count that
1641 is greater than 1 or with a limited maximum, more memory is required for the
1642 compiled pattern, in proportion to the size of the minimum or maximum.
1643 .P
1644 If a pattern starts with .* or .{0,} and the PCRE_DOTALL option (equivalent
1645 to Perl's /s) is set, thus allowing the dot to match newlines, the pattern is
1646 implicitly anchored, because whatever follows will be tried against every
1647 character position in the subject string, so there is no point in retrying the
1648 overall match at any position after the first. PCRE normally treats such a
1649 pattern as though it were preceded by \eA.
1650 .P
1651 In cases where it is known that the subject string contains no newlines, it is
1652 worth setting PCRE_DOTALL in order to obtain this optimization, or
1653 alternatively using ^ to indicate anchoring explicitly.
1654 .P
1655 However, there are some cases where the optimization cannot be used. When .*
1656 is inside capturing parentheses that are the subject of a back reference
1657 elsewhere in the pattern, a match at the start may fail where a later one
1658 succeeds. Consider, for example:
1659 .sp
1660 (.*)abc\e1
1661 .sp
1662 If the subject is "xyz123abc123" the match point is the fourth character. For
1663 this reason, such a pattern is not implicitly anchored.
1664 .P
1665 Another case where implicit anchoring is not applied is when the leading .* is
1666 inside an atomic group. Once again, a match at the start may fail where a later
1667 one succeeds. Consider this pattern:
1668 .sp
1669 (?>.*?a)b
1670 .sp
1671 It matches "ab" in the subject "aab". The use of the backtracking control verbs
1672 (*PRUNE) and (*SKIP) also disable this optimization.
1673 .P
1674 When a capturing subpattern is repeated, the value captured is the substring
1675 that matched the final iteration. For example, after
1676 .sp
1677 (tweedle[dume]{3}\es*)+
1678 .sp
1679 has matched "tweedledum tweedledee" the value of the captured substring is
1680 "tweedledee". However, if there are nested capturing subpatterns, the
1681 corresponding captured values may have been set in previous iterations. For
1682 example, after
1683 .sp
1684 /(a|(b))+/
1685 .sp
1686 matches "aba" the value of the second captured substring is "b".
1687 .
1688 .
1689 .\" HTML <a name="atomicgroup"></a>
1691 .rs
1692 .sp
1693 With both maximizing ("greedy") and minimizing ("ungreedy" or "lazy")
1694 repetition, failure of what follows normally causes the repeated item to be
1695 re-evaluated to see if a different number of repeats allows the rest of the
1696 pattern to match. Sometimes it is useful to prevent this, either to change the
1697 nature of the match, or to cause it fail earlier than it otherwise might, when
1698 the author of the pattern knows there is no point in carrying on.
1699 .P
1700 Consider, for example, the pattern \ed+foo when applied to the subject line
1701 .sp
1702 123456bar
1703 .sp
1704 After matching all 6 digits and then failing to match "foo", the normal
1705 action of the matcher is to try again with only 5 digits matching the \ed+
1706 item, and then with 4, and so on, before ultimately failing. "Atomic grouping"
1707 (a term taken from Jeffrey Friedl's book) provides the means for specifying
1708 that once a subpattern has matched, it is not to be re-evaluated in this way.
1709 .P
1710 If we use atomic grouping for the previous example, the matcher gives up
1711 immediately on failing to match "foo" the first time. The notation is a kind of
1712 special parenthesis, starting with (?> as in this example:
1713 .sp
1714 (?>\ed+)foo
1715 .sp
1716 This kind of parenthesis "locks up" the part of the pattern it contains once
1717 it has matched, and a failure further into the pattern is prevented from
1718 backtracking into it. Backtracking past it to previous items, however, works as
1719 normal.
1720 .P
1721 An alternative description is that a subpattern of this type matches the string
1722 of characters that an identical standalone pattern would match, if anchored at
1723 the current point in the subject string.
1724 .P
1725 Atomic grouping subpatterns are not capturing subpatterns. Simple cases such as
1726 the above example can be thought of as a maximizing repeat that must swallow
1727 everything it can. So, while both \ed+ and \ed+? are prepared to adjust the
1728 number of digits they match in order to make the rest of the pattern match,
1729 (?>\ed+) can only match an entire sequence of digits.
1730 .P
1731 Atomic groups in general can of course contain arbitrarily complicated
1732 subpatterns, and can be nested. However, when the subpattern for an atomic
1733 group is just a single repeated item, as in the example above, a simpler
1734 notation, called a "possessive quantifier" can be used. This consists of an
1735 additional + character following a quantifier. Using this notation, the
1736 previous example can be rewritten as
1737 .sp
1738 \ed++foo
1739 .sp
1740 Note that a possessive quantifier can be used with an entire group, for
1741 example:
1742 .sp
1743 (abc|xyz){2,3}+
1744 .sp
1745 Possessive quantifiers are always greedy; the setting of the PCRE_UNGREEDY
1746 option is ignored. They are a convenient notation for the simpler forms of
1747 atomic group. However, there is no difference in the meaning of a possessive
1748 quantifier and the equivalent atomic group, though there may be a performance
1749 difference; possessive quantifiers should be slightly faster.
1750 .P
1751 The possessive quantifier syntax is an extension to the Perl 5.8 syntax.
1752 Jeffrey Friedl originated the idea (and the name) in the first edition of his
1753 book. Mike McCloskey liked it, so implemented it when he built Sun's Java
1754 package, and PCRE copied it from there. It ultimately found its way into Perl
1755 at release 5.10.
1756 .P
1757 PCRE has an optimization that automatically "possessifies" certain simple
1758 pattern constructs. For example, the sequence A+B is treated as A++B because
1759 there is no point in backtracking into a sequence of A's when B must follow.
1760 .P
1761 When a pattern contains an unlimited repeat inside a subpattern that can itself
1762 be repeated an unlimited number of times, the use of an atomic group is the
1763 only way to avoid some failing matches taking a very long time indeed. The
1764 pattern
1765 .sp
1766 (\eD+|<\ed+>)*[!?]
1767 .sp
1768 matches an unlimited number of substrings that either consist of non-digits, or
1769 digits enclosed in <>, followed by either ! or ?. When it matches, it runs
1770 quickly. However, if it is applied to
1771 .sp
1772 aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
1773 .sp
1774 it takes a long time before reporting failure. This is because the string can
1775 be divided between the internal \eD+ repeat and the external * repeat in a
1776 large number of ways, and all have to be tried. (The example uses [!?] rather
1777 than a single character at the end, because both PCRE and Perl have an
1778 optimization that allows for fast failure when a single character is used. They
1779 remember the last single character that is required for a match, and fail early
1780 if it is not present in the string.) If the pattern is changed so that it uses
1781 an atomic group, like this:
1782 .sp
1783 ((?>\eD+)|<\ed+>)*[!?]
1784 .sp
1785 sequences of non-digits cannot be broken, and failure happens quickly.
1786 .
1787 .
1788 .\" HTML <a name="backreferences"></a>
1790 .rs
1791 .sp
1792 Outside a character class, a backslash followed by a digit greater than 0 (and
1793 possibly further digits) is a back reference to a capturing subpattern earlier
1794 (that is, to its left) in the pattern, provided there have been that many
1795 previous capturing left parentheses.
1796 .P
1797 However, if the decimal number following the backslash is less than 10, it is
1798 always taken as a back reference, and causes an error only if there are not
1799 that many capturing left parentheses in the entire pattern. In other words, the
1800 parentheses that are referenced need not be to the left of the reference for
1801 numbers less than 10. A "forward back reference" of this type can make sense
1802 when a repetition is involved and the subpattern to the right has participated
1803 in an earlier iteration.
1804 .P
1805 It is not possible to have a numerical "forward back reference" to a subpattern
1806 whose number is 10 or more using this syntax because a sequence such as \e50 is
1807 interpreted as a character defined in octal. See the subsection entitled
1808 "Non-printing characters"
1809 .\" HTML <a href="#digitsafterbackslash">
1810 .\" </a>
1811 above
1812 .\"
1813 for further details of the handling of digits following a backslash. There is
1814 no such problem when named parentheses are used. A back reference to any
1815 subpattern is possible using named parentheses (see below).
1816 .P
1817 Another way of avoiding the ambiguity inherent in the use of digits following a
1818 backslash is to use the \eg escape sequence. This escape must be followed by an
1819 unsigned number or a negative number, optionally enclosed in braces. These
1820 examples are all identical:
1821 .sp
1822 (ring), \e1
1823 (ring), \eg1
1824 (ring), \eg{1}
1825 .sp
1826 An unsigned number specifies an absolute reference without the ambiguity that
1827 is present in the older syntax. It is also useful when literal digits follow
1828 the reference. A negative number is a relative reference. Consider this
1829 example:
1830 .sp
1831 (abc(def)ghi)\eg{-1}
1832 .sp
1833 The sequence \eg{-1} is a reference to the most recently started capturing
1834 subpattern before \eg, that is, is it equivalent to \e2 in this example.
1835 Similarly, \eg{-2} would be equivalent to \e1. The use of relative references
1836 can be helpful in long patterns, and also in patterns that are created by
1837 joining together fragments that contain references within themselves.
1838 .P
1839 A back reference matches whatever actually matched the capturing subpattern in
1840 the current subject string, rather than anything matching the subpattern
1841 itself (see
1842 .\" HTML <a href="#subpatternsassubroutines">
1843 .\" </a>
1844 "Subpatterns as subroutines"
1845 .\"
1846 below for a way of doing that). So the pattern
1847 .sp
1848 (sens|respons)e and \e1ibility
1849 .sp
1850 matches "sense and sensibility" and "response and responsibility", but not
1851 "sense and responsibility". If caseful matching is in force at the time of the
1852 back reference, the case of letters is relevant. For example,
1853 .sp
1854 ((?i)rah)\es+\e1
1855 .sp
1856 matches "rah rah" and "RAH RAH", but not "RAH rah", even though the original
1857 capturing subpattern is matched caselessly.
1858 .P
1859 There are several different ways of writing back references to named
1860 subpatterns. The .NET syntax \ek{name} and the Perl syntax \ek<name> or
1861 \ek'name' are supported, as is the Python syntax (?P=name). Perl 5.10's unified
1862 back reference syntax, in which \eg can be used for both numeric and named
1863 references, is also supported. We could rewrite the above example in any of
1864 the following ways:
1865 .sp
1866 (?<p1>(?i)rah)\es+\ek<p1>
1867 (?'p1'(?i)rah)\es+\ek{p1}
1868 (?P<p1>(?i)rah)\es+(?P=p1)
1869 (?<p1>(?i)rah)\es+\eg{p1}
1870 .sp
1871 A subpattern that is referenced by name may appear in the pattern before or
1872 after the reference.
1873 .P
1874 There may be more than one back reference to the same subpattern. If a
1875 subpattern has not actually been used in a particular match, any back
1876 references to it always fail by default. For example, the pattern
1877 .sp
1878 (a|(bc))\e2
1879 .sp
1880 always fails if it starts to match "a" rather than "bc". However, if the
1881 PCRE_JAVASCRIPT_COMPAT option is set at compile time, a back reference to an
1882 unset value matches an empty string.
1883 .P
1884 Because there may be many capturing parentheses in a pattern, all digits
1885 following a backslash are taken as part of a potential back reference number.
1886 If the pattern continues with a digit character, some delimiter must be used to
1887 terminate the back reference. If the PCRE_EXTENDED option is set, this can be
1888 white space. Otherwise, the \eg{ syntax or an empty comment (see
1889 .\" HTML <a href="#comments">
1890 .\" </a>
1891 "Comments"
1892 .\"
1893 below) can be used.
1894 .
1895 .SS "Recursive back references"
1896 .rs
1897 .sp
1898 A back reference that occurs inside the parentheses to which it refers fails
1899 when the subpattern is first used, so, for example, (a\e1) never matches.
1900 However, such references can be useful inside repeated subpatterns. For
1901 example, the pattern
1902 .sp
1903 (a|b\e1)+
1904 .sp
1905 matches any number of "a"s and also "aba", "ababbaa" etc. At each iteration of
1906 the subpattern, the back reference matches the character string corresponding
1907 to the previous iteration. In order for this to work, the pattern must be such
1908 that the first iteration does not need to match the back reference. This can be
1909 done using alternation, as in the example above, or by a quantifier with a
1910 minimum of zero.
1911 .P
1912 Back references of this type cause the group that they reference to be treated
1913 as an
1914 .\" HTML <a href="#atomicgroup">
1915 .\" </a>
1916 atomic group.
1917 .\"
1918 Once the whole group has been matched, a subsequent matching failure cannot
1919 cause backtracking into the middle of the group.
1920 .
1921 .
1922 .\" HTML <a name="bigassertions"></a>
1924 .rs
1925 .sp
1926 An assertion is a test on the characters following or preceding the current
1927 matching point that does not actually consume any characters. The simple
1928 assertions coded as \eb, \eB, \eA, \eG, \eZ, \ez, ^ and $ are described
1929 .\" HTML <a href="#smallassertions">
1930 .\" </a>
1931 above.
1932 .\"
1933 .P
1934 More complicated assertions are coded as subpatterns. There are two kinds:
1935 those that look ahead of the current position in the subject string, and those
1936 that look behind it. An assertion subpattern is matched in the normal way,
1937 except that it does not cause the current matching position to be changed.
1938 .P
1939 Assertion subpatterns are not capturing subpatterns. If such an assertion
1940 contains capturing subpatterns within it, these are counted for the purposes of
1941 numbering the capturing subpatterns in the whole pattern. However, substring
1942 capturing is carried out only for positive assertions, because it does not make
1943 sense for negative assertions.
1944 .P
1945 For compatibility with Perl, assertion subpatterns may be repeated; though
1946 it makes no sense to assert the same thing several times, the side effect of
1947 capturing parentheses may occasionally be useful. In practice, there only three
1948 cases:
1949 .sp
1950 (1) If the quantifier is {0}, the assertion is never obeyed during matching.
1951 However, it may contain internal capturing parenthesized groups that are called
1952 from elsewhere via the
1953 .\" HTML <a href="#subpatternsassubroutines">
1954 .\" </a>
1955 subroutine mechanism.
1956 .\"
1957 .sp
1958 (2) If quantifier is {0,n} where n is greater than zero, it is treated as if it
1959 were {0,1}. At run time, the rest of the pattern match is tried with and
1960 without the assertion, the order depending on the greediness of the quantifier.
1961 .sp
1962 (3) If the minimum repetition is greater than zero, the quantifier is ignored.
1963 The assertion is obeyed just once when encountered during matching.
1964 .
1965 .
1966 .SS "Lookahead assertions"
1967 .rs
1968 .sp
1969 Lookahead assertions start with (?= for positive assertions and (?! for
1970 negative assertions. For example,
1971 .sp
1972 \ew+(?=;)
1973 .sp
1974 matches a word followed by a semicolon, but does not include the semicolon in
1975 the match, and
1976 .sp
1977 foo(?!bar)
1978 .sp
1979 matches any occurrence of "foo" that is not followed by "bar". Note that the
1980 apparently similar pattern
1981 .sp
1982 (?!foo)bar
1983 .sp
1984 does not find an occurrence of "bar" that is preceded by something other than
1985 "foo"; it finds any occurrence of "bar" whatsoever, because the assertion
1986 (?!foo) is always true when the next three characters are "bar". A
1987 lookbehind assertion is needed to achieve the other effect.
1988 .P
1989 If you want to force a matching failure at some point in a pattern, the most
1990 convenient way to do it is with (?!) because an empty string always matches, so
1991 an assertion that requires there not to be an empty string must always fail.
1992 The backtracking control verb (*FAIL) or (*F) is a synonym for (?!).
1993 .
1994 .
1995 .\" HTML <a name="lookbehind"></a>
1996 .SS "Lookbehind assertions"
1997 .rs
1998 .sp
1999 Lookbehind assertions start with (?<= for positive assertions and (?<! for
2000 negative assertions. For example,
2001 .sp
2002 (?<!foo)bar
2003 .sp
2004 does find an occurrence of "bar" that is not preceded by "foo". The contents of
2005 a lookbehind assertion are restricted such that all the strings it matches must
2006 have a fixed length. However, if there are several top-level alternatives, they
2007 do not all have to have the same fixed length. Thus
2008 .sp
2009 (?<=bullock|donkey)
2010 .sp
2011 is permitted, but
2012 .sp
2013 (?<!dogs?|cats?)
2014 .sp
2015 causes an error at compile time. Branches that match different length strings
2016 are permitted only at the top level of a lookbehind assertion. This is an
2017 extension compared with Perl, which requires all branches to match the same
2018 length of string. An assertion such as
2019 .sp
2020 (?<=ab(c|de))
2021 .sp
2022 is not permitted, because its single top-level branch can match two different
2023 lengths, but it is acceptable to PCRE if rewritten to use two top-level
2024 branches:
2025 .sp
2026 (?<=abc|abde)
2027 .sp
2028 In some cases, the escape sequence \eK
2029 .\" HTML <a href="#resetmatchstart">
2030 .\" </a>
2031 (see above)
2032 .\"
2033 can be used instead of a lookbehind assertion to get round the fixed-length
2034 restriction.
2035 .P
2036 The implementation of lookbehind assertions is, for each alternative, to
2037 temporarily move the current position back by the fixed length and then try to
2038 match. If there are insufficient characters before the current position, the
2039 assertion fails.
2040 .P
2041 In a UTF mode, PCRE does not allow the \eC escape (which matches a single data
2042 unit even in a UTF mode) to appear in lookbehind assertions, because it makes
2043 it impossible to calculate the length of the lookbehind. The \eX and \eR
2044 escapes, which can match different numbers of data units, are also not
2045 permitted.
2046 .P
2047 .\" HTML <a href="#subpatternsassubroutines">
2048 .\" </a>
2049 "Subroutine"
2050 .\"
2051 calls (see below) such as (?2) or (?&X) are permitted in lookbehinds, as long
2052 as the subpattern matches a fixed-length string.
2053 .\" HTML <a href="#recursion">
2054 .\" </a>
2055 Recursion,
2056 .\"
2057 however, is not supported.
2058 .P
2059 Possessive quantifiers can be used in conjunction with lookbehind assertions to
2060 specify efficient matching of fixed-length strings at the end of subject
2061 strings. Consider a simple pattern such as
2062 .sp
2063 abcd$
2064 .sp
2065 when applied to a long string that does not match. Because matching proceeds
2066 from left to right, PCRE will look for each "a" in the subject and then see if
2067 what follows matches the rest of the pattern. If the pattern is specified as
2068 .sp
2069 ^.*abcd$
2070 .sp
2071 the initial .* matches the entire string at first, but when this fails (because
2072 there is no following "a"), it backtracks to match all but the last character,
2073 then all but the last two characters, and so on. Once again the search for "a"
2074 covers the entire string, from right to left, so we are no better off. However,
2075 if the pattern is written as
2076 .sp
2077 ^.*+(?<=abcd)
2078 .sp
2079 there can be no backtracking for the .*+ item; it can match only the entire
2080 string. The subsequent lookbehind assertion does a single test on the last four
2081 characters. If it fails, the match fails immediately. For long strings, this
2082 approach makes a significant difference to the processing time.
2083 .
2084 .
2085 .SS "Using multiple assertions"
2086 .rs
2087 .sp
2088 Several assertions (of any sort) may occur in succession. For example,
2089 .sp
2090 (?<=\ed{3})(?<!999)foo
2091 .sp
2092 matches "foo" preceded by three digits that are not "999". Notice that each of
2093 the assertions is applied independently at the same point in the subject
2094 string. First there is a check that the previous three characters are all
2095 digits, and then there is a check that the same three characters are not "999".
2096 This pattern does \fInot\fP match "foo" preceded by six characters, the first
2097 of which are digits and the last three of which are not "999". For example, it
2098 doesn't match "123abcfoo". A pattern to do that is
2099 .sp
2100 (?<=\ed{3}...)(?<!999)foo
2101 .sp
2102 This time the first assertion looks at the preceding six characters, checking
2103 that the first three are digits, and then the second assertion checks that the
2104 preceding three characters are not "999".
2105 .P
2106 Assertions can be nested in any combination. For example,
2107 .sp
2108 (?<=(?<!foo)bar)baz
2109 .sp
2110 matches an occurrence of "baz" that is preceded by "bar" which in turn is not
2111 preceded by "foo", while
2112 .sp
2113 (?<=\ed{3}(?!999)...)foo
2114 .sp
2115 is another pattern that matches "foo" preceded by three digits and any three
2116 characters that are not "999".
2117 .
2118 .
2119 .\" HTML <a name="conditions"></a>
2121 .rs
2122 .sp
2123 It is possible to cause the matching process to obey a subpattern
2124 conditionally or to choose between two alternative subpatterns, depending on
2125 the result of an assertion, or whether a specific capturing subpattern has
2126 already been matched. The two possible forms of conditional subpattern are:
2127 .sp
2128 (?(condition)yes-pattern)
2129 (?(condition)yes-pattern|no-pattern)
2130 .sp
2131 If the condition is satisfied, the yes-pattern is used; otherwise the
2132 no-pattern (if present) is used. If there are more than two alternatives in the
2133 subpattern, a compile-time error occurs. Each of the two alternatives may
2134 itself contain nested subpatterns of any form, including conditional
2135 subpatterns; the restriction to two alternatives applies only at the level of
2136 the condition. This pattern fragment is an example where the alternatives are
2137 complex:
2138 .sp
2139 (?(1) (A|B|C) | (D | (?(2)E|F) | E) )
2140 .sp
2141 .P
2142 There are four kinds of condition: references to subpatterns, references to
2143 recursion, a pseudo-condition called DEFINE, and assertions.
2144 .
2145 .SS "Checking for a used subpattern by number"
2146 .rs
2147 .sp
2148 If the text between the parentheses consists of a sequence of digits, the
2149 condition is true if a capturing subpattern of that number has previously
2150 matched. If there is more than one capturing subpattern with the same number
2151 (see the earlier
2152 .\"
2153 .\" HTML <a href="#recursion">
2154 .\" </a>
2155 section about duplicate subpattern numbers),
2156 .\"
2157 the condition is true if any of them have matched. An alternative notation is
2158 to precede the digits with a plus or minus sign. In this case, the subpattern
2159 number is relative rather than absolute. The most recently opened parentheses
2160 can be referenced by (?(-1), the next most recent by (?(-2), and so on. Inside
2161 loops it can also make sense to refer to subsequent groups. The next
2162 parentheses to be opened can be referenced as (?(+1), and so on. (The value
2163 zero in any of these forms is not used; it provokes a compile-time error.)
2164 .P
2165 Consider the following pattern, which contains non-significant white space to
2166 make it more readable (assume the PCRE_EXTENDED option) and to divide it into
2167 three parts for ease of discussion:
2168 .sp
2169 ( \e( )? [^()]+ (?(1) \e) )
2170 .sp
2171 The first part matches an optional opening parenthesis, and if that
2172 character is present, sets it as the first captured substring. The second part
2173 matches one or more characters that are not parentheses. The third part is a
2174 conditional subpattern that tests whether or not the first set of parentheses
2175 matched. If they did, that is, if subject started with an opening parenthesis,
2176 the condition is true, and so the yes-pattern is executed and a closing
2177 parenthesis is required. Otherwise, since no-pattern is not present, the
2178 subpattern matches nothing. In other words, this pattern matches a sequence of
2179 non-parentheses, optionally enclosed in parentheses.
2180 .P
2181 If you were embedding this pattern in a larger one, you could use a relative
2182 reference:
2183 .sp
2184 ...other stuff... ( \e( )? [^()]+ (?(-1) \e) ) ...
2185 .sp
2186 This makes the fragment independent of the parentheses in the larger pattern.
2187 .
2188 .SS "Checking for a used subpattern by name"
2189 .rs
2190 .sp
2191 Perl uses the syntax (?(<name>)...) or (?('name')...) to test for a used
2192 subpattern by name. For compatibility with earlier versions of PCRE, which had
2193 this facility before Perl, the syntax (?(name)...) is also recognized. However,
2194 there is a possible ambiguity with this syntax, because subpattern names may
2195 consist entirely of digits. PCRE looks first for a named subpattern; if it
2196 cannot find one and the name consists entirely of digits, PCRE looks for a
2197 subpattern of that number, which must be greater than zero. Using subpattern
2198 names that consist entirely of digits is not recommended.
2199 .P
2200 Rewriting the above example to use a named subpattern gives this:
2201 .sp
2202 (?<OPEN> \e( )? [^()]+ (?(<OPEN>) \e) )
2203 .sp
2204 If the name used in a condition of this kind is a duplicate, the test is
2205 applied to all subpatterns of the same name, and is true if any one of them has
2206 matched.
2207 .
2208 .SS "Checking for pattern recursion"
2209 .rs
2210 .sp
2211 If the condition is the string (R), and there is no subpattern with the name R,
2212 the condition is true if a recursive call to the whole pattern or any
2213 subpattern has been made. If digits or a name preceded by ampersand follow the
2214 letter R, for example:
2215 .sp
2216 (?(R3)...) or (?(R&name)...)
2217 .sp
2218 the condition is true if the most recent recursion is into a subpattern whose
2219 number or name is given. This condition does not check the entire recursion
2220 stack. If the name used in a condition of this kind is a duplicate, the test is
2221 applied to all subpatterns of the same name, and is true if any one of them is
2222 the most recent recursion.
2223 .P
2224 At "top level", all these recursion test conditions are false.
2225 .\" HTML <a href="#recursion">
2226 .\" </a>
2227 The syntax for recursive patterns
2228 .\"
2229 is described below.
2230 .
2231 .\" HTML <a name="subdefine"></a>
2232 .SS "Defining subpatterns for use by reference only"
2233 .rs
2234 .sp
2235 If the condition is the string (DEFINE), and there is no subpattern with the
2236 name DEFINE, the condition is always false. In this case, there may be only one
2237 alternative in the subpattern. It is always skipped if control reaches this
2238 point in the pattern; the idea of DEFINE is that it can be used to define
2239 subroutines that can be referenced from elsewhere. (The use of
2240 .\" HTML <a href="#subpatternsassubroutines">
2241 .\" </a>
2242 subroutines
2243 .\"
2244 is described below.) For example, a pattern to match an IPv4 address such as
2245 "" could be written like this (ignore white space and line
2246 breaks):
2247 .sp
2248 (?(DEFINE) (?<byte> 2[0-4]\ed | 25[0-5] | 1\ed\ed | [1-9]?\ed) )
2249 \eb (?&byte) (\e.(?&byte)){3} \eb
2250 .sp
2251 The first part of the pattern is a DEFINE group inside which a another group
2252 named "byte" is defined. This matches an individual component of an IPv4
2253 address (a number less than 256). When matching takes place, this part of the
2254 pattern is skipped because DEFINE acts like a false condition. The rest of the
2255 pattern uses references to the named group to match the four dot-separated
2256 components of an IPv4 address, insisting on a word boundary at each end.
2257 .
2258 .SS "Assertion conditions"
2259 .rs
2260 .sp
2261 If the condition is not in any of the above formats, it must be an assertion.
2262 This may be a positive or negative lookahead or lookbehind assertion. Consider
2263 this pattern, again containing non-significant white space, and with the two
2264 alternatives on the second line:
2265 .sp
2266 (?(?=[^a-z]*[a-z])
2267 \ed{2}-[a-z]{3}-\ed{2} | \ed{2}-\ed{2}-\ed{2} )
2268 .sp
2269 The condition is a positive lookahead assertion that matches an optional
2270 sequence of non-letters followed by a letter. In other words, it tests for the
2271 presence of at least one letter in the subject. If a letter is found, the
2272 subject is matched against the first alternative; otherwise it is matched
2273 against the second. This pattern matches strings in one of the two forms
2274 dd-aaa-dd or dd-dd-dd, where aaa are letters and dd are digits.
2275 .
2276 .
2277 .\" HTML <a name="comments"></a>
2279 .rs
2280 .sp
2281 There are two ways of including comments in patterns that are processed by
2282 PCRE. In both cases, the start of the comment must not be in a character class,
2283 nor in the middle of any other sequence of related characters such as (?: or a
2284 subpattern name or number. The characters that make up a comment play no part
2285 in the pattern matching.
2286 .P
2287 The sequence (?# marks the start of a comment that continues up to the next
2288 closing parenthesis. Nested parentheses are not permitted. If the PCRE_EXTENDED
2289 option is set, an unescaped # character also introduces a comment, which in
2290 this case continues to immediately after the next newline character or
2291 character sequence in the pattern. Which characters are interpreted as newlines
2292 is controlled by the options passed to a compiling function or by a special
2293 sequence at the start of the pattern, as described in the section entitled
2294 .\" HTML <a href="#newlines">
2295 .\" </a>
2296 "Newline conventions"
2297 .\"
2298 above. Note that the end of this type of comment is a literal newline sequence
2299 in the pattern; escape sequences that happen to represent a newline do not
2300 count. For example, consider this pattern when PCRE_EXTENDED is set, and the
2301 default newline convention is in force:
2302 .sp
2303 abc #comment \en still comment
2304 .sp
2305 On encountering the # character, \fBpcre_compile()\fP skips along, looking for
2306 a newline in the pattern. The sequence \en is still literal at this stage, so
2307 it does not terminate the comment. Only an actual character with the code value
2308 0x0a (the default newline) does so.
2309 .
2310 .
2311 .\" HTML <a name="recursion"></a>
2313 .rs
2314 .sp
2315 Consider the problem of matching a string in parentheses, allowing for
2316 unlimited nested parentheses. Without the use of recursion, the best that can
2317 be done is to use a pattern that matches up to some fixed depth of nesting. It
2318 is not possible to handle an arbitrary nesting depth.
2319 .P
2320 For some time, Perl has provided a facility that allows regular expressions to
2321 recurse (amongst other things). It does this by interpolating Perl code in the
2322 expression at run time, and the code can refer to the expression itself. A Perl
2323 pattern using code interpolation to solve the parentheses problem can be
2324 created like this:
2325 .sp
2326 $re = qr{\e( (?: (?>[^()]+) | (?p{$re}) )* \e)}x;
2327 .sp
2328 The (?p{...}) item interpolates Perl code at run time, and in this case refers
2329 recursively to the pattern in which it appears.
2330 .P
2331 Obviously, PCRE cannot support the interpolation of Perl code. Instead, it
2332 supports special syntax for recursion of the entire pattern, and also for
2333 individual subpattern recursion. After its introduction in PCRE and Python,
2334 this kind of recursion was subsequently introduced into Perl at release 5.10.
2335 .P
2336 A special item that consists of (? followed by a number greater than zero and a
2337 closing parenthesis is a recursive subroutine call of the subpattern of the
2338 given number, provided that it occurs inside that subpattern. (If not, it is a
2339 .\" HTML <a href="#subpatternsassubroutines">
2340 .\" </a>
2341 non-recursive subroutine
2342 .\"
2343 call, which is described in the next section.) The special item (?R) or (?0) is
2344 a recursive call of the entire regular expression.
2345 .P
2346 This PCRE pattern solves the nested parentheses problem (assume the
2347 PCRE_EXTENDED option is set so that white space is ignored):
2348 .sp
2349 \e( ( [^()]++ | (?R) )* \e)
2350 .sp
2351 First it matches an opening parenthesis. Then it matches any number of
2352 substrings which can either be a sequence of non-parentheses, or a recursive
2353 match of the pattern itself (that is, a correctly parenthesized substring).
2354 Finally there is a closing parenthesis. Note the use of a possessive quantifier
2355 to avoid backtracking into sequences of non-parentheses.
2356 .P
2357 If this were part of a larger pattern, you would not want to recurse the entire
2358 pattern, so instead you could use this:
2359 .sp
2360 ( \e( ( [^()]++ | (?1) )* \e) )
2361 .sp
2362 We have put the pattern into parentheses, and caused the recursion to refer to
2363 them instead of the whole pattern.
2364 .P
2365 In a larger pattern, keeping track of parenthesis numbers can be tricky. This
2366 is made easier by the use of relative references. Instead of (?1) in the
2367 pattern above you can write (?-2) to refer to the second most recently opened
2368 parentheses preceding the recursion. In other words, a negative number counts
2369 capturing parentheses leftwards from the point at which it is encountered.
2370 .P
2371 It is also possible to refer to subsequently opened parentheses, by writing
2372 references such as (?+2). However, these cannot be recursive because the
2373 reference is not inside the parentheses that are referenced. They are always
2374 .\" HTML <a href="#subpatternsassubroutines">
2375 .\" </a>
2376 non-recursive subroutine
2377 .\"
2378 calls, as described in the next section.
2379 .P
2380 An alternative approach is to use named parentheses instead. The Perl syntax
2381 for this is (?&name); PCRE's earlier syntax (?P>name) is also supported. We
2382 could rewrite the above example as follows:
2383 .sp
2384 (?<pn> \e( ( [^()]++ | (?&pn) )* \e) )
2385 .sp
2386 If there is more than one subpattern with the same name, the earliest one is
2387 used.
2388 .P
2389 This particular example pattern that we have been looking at contains nested
2390 unlimited repeats, and so the use of a possessive quantifier for matching
2391 strings of non-parentheses is important when applying the pattern to strings
2392 that do not match. For example, when this pattern is applied to
2393 .sp
2394 (aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa()
2395 .sp
2396 it yields "no match" quickly. However, if a possessive quantifier is not used,
2397 the match runs for a very long time indeed because there are so many different
2398 ways the + and * repeats can carve up the subject, and all have to be tested
2399 before failure can be reported.
2400 .P
2401 At the end of a match, the values of capturing parentheses are those from
2402 the outermost level. If you want to obtain intermediate values, a callout
2403 function can be used (see below and the
2404 .\" HREF
2405 \fBpcrecallout\fP
2406 .\"
2407 documentation). If the pattern above is matched against
2408 .sp
2409 (ab(cd)ef)
2410 .sp
2411 the value for the inner capturing parentheses (numbered 2) is "ef", which is
2412 the last value taken on at the top level. If a capturing subpattern is not
2413 matched at the top level, its final captured value is unset, even if it was
2414 (temporarily) set at a deeper level during the matching process.
2415 .P
2416 If there are more than 15 capturing parentheses in a pattern, PCRE has to
2417 obtain extra memory to store data during a recursion, which it does by using
2418 \fBpcre_malloc\fP, freeing it via \fBpcre_free\fP afterwards. If no memory can
2419 be obtained, the match fails with the PCRE_ERROR_NOMEMORY error.
2420 .P
2421 Do not confuse the (?R) item with the condition (R), which tests for recursion.
2422 Consider this pattern, which matches text in angle brackets, allowing for
2423 arbitrary nesting. Only digits are allowed in nested brackets (that is, when
2424 recursing), whereas any characters are permitted at the outer level.
2425 .sp
2426 < (?: (?(R) \ed++ | [^<>]*+) | (?R)) * >
2427 .sp
2428 In this pattern, (?(R) is the start of a conditional subpattern, with two
2429 different alternatives for the recursive and non-recursive cases. The (?R) item
2430 is the actual recursive call.
2431 .
2432 .
2433 .\" HTML <a name="recursiondifference"></a>
2434 .SS "Differences in recursion processing between PCRE and Perl"
2435 .rs
2436 .sp
2437 Recursion processing in PCRE differs from Perl in two important ways. In PCRE
2438 (like Python, but unlike Perl), a recursive subpattern call is always treated
2439 as an atomic group. That is, once it has matched some of the subject string, it
2440 is never re-entered, even if it contains untried alternatives and there is a
2441 subsequent matching failure. This can be illustrated by the following pattern,
2442 which purports to match a palindromic string that contains an odd number of
2443 characters (for example, "a", "aba", "abcba", "abcdcba"):
2444 .sp
2445 ^(.|(.)(?1)\e2)$
2446 .sp
2447 The idea is that it either matches a single character, or two identical
2448 characters surrounding a sub-palindrome. In Perl, this pattern works; in PCRE
2449 it does not if the pattern is longer than three characters. Consider the
2450 subject string "abcba":
2451 .P
2452 At the top level, the first character is matched, but as it is not at the end
2453 of the string, the first alternative fails; the second alternative is taken
2454 and the recursion kicks in. The recursive call to subpattern 1 successfully
2455 matches the next character ("b"). (Note that the beginning and end of line
2456 tests are not part of the recursion).
2457 .P
2458 Back at the top level, the next character ("c") is compared with what
2459 subpattern 2 matched, which was "a". This fails. Because the recursion is
2460 treated as an atomic group, there are now no backtracking points, and so the
2461 entire match fails. (Perl is able, at this point, to re-enter the recursion and
2462 try the second alternative.) However, if the pattern is written with the
2463 alternatives in the other order, things are different:
2464 .sp
2465 ^((.)(?1)\e2|.)$
2466 .sp
2467 This time, the recursing alternative is tried first, and continues to recurse
2468 until it runs out of characters, at which point the recursion fails. But this
2469 time we do have another alternative to try at the higher level. That is the big
2470 difference: in the previous case the remaining alternative is at a deeper
2471 recursion level, which PCRE cannot use.
2472 .P
2473 To change the pattern so that it matches all palindromic strings, not just
2474 those with an odd number of characters, it is tempting to change the pattern to
2475 this:
2476 .sp
2477 ^((.)(?1)\e2|.?)$
2478 .sp
2479 Again, this works in Perl, but not in PCRE, and for the same reason. When a
2480 deeper recursion has matched a single character, it cannot be entered again in
2481 order to match an empty string. The solution is to separate the two cases, and
2482 write out the odd and even cases as alternatives at the higher level:
2483 .sp
2484 ^(?:((.)(?1)\e2|)|((.)(?3)\e4|.))
2485 .sp
2486 If you want to match typical palindromic phrases, the pattern has to ignore all
2487 non-word characters, which can be done like this:
2488 .sp
2489 ^\eW*+(?:((.)\eW*+(?1)\eW*+\e2|)|((.)\eW*+(?3)\eW*+\e4|\eW*+.\eW*+))\eW*+$
2490 .sp
2491 If run with the PCRE_CASELESS option, this pattern matches phrases such as "A
2492 man, a plan, a canal: Panama!" and it works well in both PCRE and Perl. Note
2493 the use of the possessive quantifier *+ to avoid backtracking into sequences of
2494 non-word characters. Without this, PCRE takes a great deal longer (ten times or
2495 more) to match typical phrases, and Perl takes so long that you think it has
2496 gone into a loop.
2497 .P
2498 \fBWARNING\fP: The palindrome-matching patterns above work only if the subject
2499 string does not start with a palindrome that is shorter than the entire string.
2500 For example, although "abcba" is correctly matched, if the subject is "ababa",
2501 PCRE finds the palindrome "aba" at the start, then fails at top level because
2502 the end of the string does not follow. Once again, it cannot jump back into the
2503 recursion to try other alternatives, so the entire match fails.
2504 .P
2505 The second way in which PCRE and Perl differ in their recursion processing is
2506 in the handling of captured values. In Perl, when a subpattern is called
2507 recursively or as a subpattern (see the next section), it has no access to any
2508 values that were captured outside the recursion, whereas in PCRE these values
2509 can be referenced. Consider this pattern:
2510 .sp
2511 ^(.)(\e1|a(?2))
2512 .sp
2513 In PCRE, this pattern matches "bab". The first capturing parentheses match "b",
2514 then in the second group, when the back reference \e1 fails to match "b", the
2515 second alternative matches "a" and then recurses. In the recursion, \e1 does
2516 now match "b" and so the whole match succeeds. In Perl, the pattern fails to
2517 match because inside the recursive call \e1 cannot access the externally set
2518 value.
2519 .
2520 .
2521 .\" HTML <a name="subpatternsassubroutines"></a>
2523 .rs
2524 .sp
2525 If the syntax for a recursive subpattern call (either by number or by
2526 name) is used outside the parentheses to which it refers, it operates like a
2527 subroutine in a programming language. The called subpattern may be defined
2528 before or after the reference. A numbered reference can be absolute or
2529 relative, as in these examples:
2530 .sp
2531 (...(absolute)...)...(?2)...
2532 (...(relative)...)...(?-1)...
2533 (...(?+1)...(relative)...
2534 .sp
2535 An earlier example pointed out that the pattern
2536 .sp
2537 (sens|respons)e and \e1ibility
2538 .sp
2539 matches "sense and sensibility" and "response and responsibility", but not
2540 "sense and responsibility". If instead the pattern
2541 .sp
2542 (sens|respons)e and (?1)ibility
2543 .sp
2544 is used, it does match "sense and responsibility" as well as the other two
2545 strings. Another example is given in the discussion of DEFINE above.
2546 .P
2547 All subroutine calls, whether recursive or not, are always treated as atomic
2548 groups. That is, once a subroutine has matched some of the subject string, it
2549 is never re-entered, even if it contains untried alternatives and there is a
2550 subsequent matching failure. Any capturing parentheses that are set during the
2551 subroutine call revert to their previous values afterwards.
2552 .P
2553 Processing options such as case-independence are fixed when a subpattern is
2554 defined, so if it is used as a subroutine, such options cannot be changed for
2555 different calls. For example, consider this pattern:
2556 .sp
2557 (abc)(?i:(?-1))
2558 .sp
2559 It matches "abcabc". It does not match "abcABC" because the change of
2560 processing option does not affect the called subpattern.
2561 .
2562 .
2563 .\" HTML <a name="onigurumasubroutines"></a>
2565 .rs
2566 .sp
2567 For compatibility with Oniguruma, the non-Perl syntax \eg followed by a name or
2568 a number enclosed either in angle brackets or single quotes, is an alternative
2569 syntax for referencing a subpattern as a subroutine, possibly recursively. Here
2570 are two of the examples used above, rewritten using this syntax:
2571 .sp
2572 (?<pn> \e( ( (?>[^()]+) | \eg<pn> )* \e) )
2573 (sens|respons)e and \eg'1'ibility
2574 .sp
2575 PCRE supports an extension to Oniguruma: if a number is preceded by a
2576 plus or a minus sign it is taken as a relative reference. For example:
2577 .sp
2578 (abc)(?i:\eg<-1>)
2579 .sp
2580 Note that \eg{...} (Perl syntax) and \eg<...> (Oniguruma syntax) are \fInot\fP
2581 synonymous. The former is a back reference; the latter is a subroutine call.
2582 .
2583 .
2585 .rs
2586 .sp
2587 Perl has a feature whereby using the sequence (?{...}) causes arbitrary Perl
2588 code to be obeyed in the middle of matching a regular expression. This makes it
2589 possible, amongst other things, to extract different substrings that match the
2590 same pair of parentheses when there is a repetition.
2591 .P
2592 PCRE provides a similar feature, but of course it cannot obey arbitrary Perl
2593 code. The feature is called "callout". The caller of PCRE provides an external
2594 function by putting its entry point in the global variable \fIpcre_callout\fP
2595 (8-bit library) or \fIpcre16_callout\fP (16-bit library). By default, this
2596 variable contains NULL, which disables all calling out.
2597 .P
2598 Within a regular expression, (?C) indicates the points at which the external
2599 function is to be called. If you want to identify different callout points, you
2600 can put a number less than 256 after the letter C. The default value is zero.
2601 For example, this pattern has two callout points:
2602 .sp
2603 (?C1)abc(?C2)def
2604 .sp
2605 If the PCRE_AUTO_CALLOUT flag is passed to a compiling function, callouts are
2606 automatically installed before each item in the pattern. They are all numbered
2607 255.
2608 .P
2609 During matching, when PCRE reaches a callout point, the external function is
2610 called. It is provided with the number of the callout, the position in the
2611 pattern, and, optionally, one item of data originally supplied by the caller of
2612 the matching function. The callout function may cause matching to proceed, to
2613 backtrack, or to fail altogether. A complete description of the interface to
2614 the callout function is given in the
2615 .\" HREF
2616 \fBpcrecallout\fP
2617 .\"
2618 documentation.
2619 .
2620 .
2621 .\" HTML <a name="backtrackcontrol"></a>
2623 .rs
2624 .sp
2625 Perl 5.10 introduced a number of "Special Backtracking Control Verbs", which
2626 are described in the Perl documentation as "experimental and subject to change
2627 or removal in a future version of Perl". It goes on to say: "Their usage in
2628 production code should be noted to avoid problems during upgrades." The same
2629 remarks apply to the PCRE features described in this section.
2630 .P
2631 Since these verbs are specifically related to backtracking, most of them can be
2632 used only when the pattern is to be matched using one of the traditional
2633 matching functions, which use a backtracking algorithm. With the exception of
2634 (*FAIL), which behaves like a failing negative assertion, they cause an error
2635 if encountered by a DFA matching function.
2636 .P
2637 If any of these verbs are used in an assertion or in a subpattern that is
2638 called as a subroutine (whether or not recursively), their effect is confined
2639 to that subpattern; it does not extend to the surrounding pattern, with one
2640 exception: the name from a *(MARK), (*PRUNE), or (*THEN) that is encountered in
2641 a successful positive assertion \fIis\fP passed back when a match succeeds
2642 (compare capturing parentheses in assertions). Note that such subpatterns are
2643 processed as anchored at the point where they are tested. Note also that Perl's
2644 treatment of subroutines and assertions is different in some cases.
2645 .P
2646 The new verbs make use of what was previously invalid syntax: an opening
2647 parenthesis followed by an asterisk. They are generally of the form
2648 (*VERB) or (*VERB:NAME). Some may take either form, with differing behaviour,
2649 depending on whether or not an argument is present. A name is any sequence of
2650 characters that does not include a closing parenthesis. The maximum length of
2651 name is 255 in the 8-bit library and 65535 in the 16-bit library. If the name
2652 is empty, that is, if the closing parenthesis immediately follows the colon,
2653 the effect is as if the colon were not there. Any number of these verbs may
2654 occur in a pattern.
2655 .
2656 .
2657 .\" HTML <a name="nooptimize"></a>
2658 .SS "Optimizations that affect backtracking verbs"
2659 .rs
2660 .sp
2661 PCRE contains some optimizations that are used to speed up matching by running
2662 some checks at the start of each match attempt. For example, it may know the
2663 minimum length of matching subject, or that a particular character must be
2664 present. When one of these optimizations suppresses the running of a match, any
2665 included backtracking verbs will not, of course, be processed. You can suppress
2666 the start-of-match optimizations by setting the PCRE_NO_START_OPTIMIZE option
2667 when calling \fBpcre_compile()\fP or \fBpcre_exec()\fP, or by starting the
2668 pattern with (*NO_START_OPT). There is more discussion of this option in the
2669 section entitled
2670 .\" HTML <a href="pcreapi.html#execoptions">
2671 .\" </a>
2672 "Option bits for \fBpcre_exec()\fP"
2673 .\"
2674 in the
2675 .\" HREF
2676 \fBpcreapi\fP
2677 .\"
2678 documentation.
2679 .P
2680 Experiments with Perl suggest that it too has similar optimizations, sometimes
2681 leading to anomalous results.
2682 .
2683 .
2684 .SS "Verbs that act immediately"
2685 .rs
2686 .sp
2687 The following verbs act as soon as they are encountered. They may not be
2688 followed by a name.
2689 .sp
2690 (*ACCEPT)
2691 .sp
2692 This verb causes the match to end successfully, skipping the remainder of the
2693 pattern. However, when it is inside a subpattern that is called as a
2694 subroutine, only that subpattern is ended successfully. Matching then continues
2695 at the outer level. If (*ACCEPT) is inside capturing parentheses, the data so
2696 far is captured. For example:
2697 .sp
2698 A((?:A|B(*ACCEPT)|C)D)
2699 .sp
2700 This matches "AB", "AAD", or "ACD"; when it matches "AB", "B" is captured by
2701 the outer parentheses.
2702 .sp
2703 (*FAIL) or (*F)
2704 .sp
2705 This verb causes a matching failure, forcing backtracking to occur. It is
2706 equivalent to (?!) but easier to read. The Perl documentation notes that it is
2707 probably useful only when combined with (?{}) or (??{}). Those are, of course,
2708 Perl features that are not present in PCRE. The nearest equivalent is the
2709 callout feature, as for example in this pattern:
2710 .sp
2711 a+(?C)(*FAIL)
2712 .sp
2713 A match with the string "aaaa" always fails, but the callout is taken before
2714 each backtrack happens (in this example, 10 times).
2715 .
2716 .
2717 .SS "Recording which path was taken"
2718 .rs
2719 .sp
2720 There is one verb whose main purpose is to track how a match was arrived at,
2721 though it also has a secondary use in conjunction with advancing the match
2722 starting point (see (*SKIP) below).
2723 .sp
2724 (*MARK:NAME) or (*:NAME)
2725 .sp
2726 A name is always required with this verb. There may be as many instances of
2727 (*MARK) as you like in a pattern, and their names do not have to be unique.
2728 .P
2729 When a match succeeds, the name of the last-encountered (*MARK) on the matching
2730 path is passed back to the caller as described in the section entitled
2731 .\" HTML <a href="pcreapi.html#extradata">
2732 .\" </a>
2733 "Extra data for \fBpcre_exec()\fP"
2734 .\"
2735 in the
2736 .\" HREF
2737 \fBpcreapi\fP
2738 .\"
2739 documentation. Here is an example of \fBpcretest\fP output, where the /K
2740 modifier requests the retrieval and outputting of (*MARK) data:
2741 .sp
2742 re> /X(*MARK:A)Y|X(*MARK:B)Z/K
2743 data> XY
2744 0: XY
2745 MK: A
2746 XZ
2747 0: XZ
2748 MK: B
2749 .sp
2750 The (*MARK) name is tagged with "MK:" in this output, and in this example it
2751 indicates which of the two alternatives matched. This is a more efficient way
2752 of obtaining this information than putting each alternative in its own
2753 capturing parentheses.
2754 .P
2755 If (*MARK) is encountered in a positive assertion, its name is recorded and
2756 passed back if it is the last-encountered. This does not happen for negative
2757 assertions.
2758 .P
2759 After a partial match or a failed match, the name of the last encountered
2760 (*MARK) in the entire match process is returned. For example:
2761 .sp
2762 re> /X(*MARK:A)Y|X(*MARK:B)Z/K
2763 data> XP
2764 No match, mark = B
2765 .sp
2766 Note that in this unanchored example the mark is retained from the match
2767 attempt that started at the letter "X" in the subject. Subsequent match
2768 attempts starting at "P" and then with an empty string do not get as far as the
2769 (*MARK) item, but nevertheless do not reset it.
2770 .P
2771 If you are interested in (*MARK) values after failed matches, you should
2772 probably set the PCRE_NO_START_OPTIMIZE option
2773 .\" HTML <a href="#nooptimize">
2774 .\" </a>
2775 (see above)
2776 .\"
2777 to ensure that the match is always attempted.
2778 .
2779 .
2780 .SS "Verbs that act after backtracking"
2781 .rs
2782 .sp
2783 The following verbs do nothing when they are encountered. Matching continues
2784 with what follows, but if there is no subsequent match, causing a backtrack to
2785 the verb, a failure is forced. That is, backtracking cannot pass to the left of
2786 the verb. However, when one of these verbs appears inside an atomic group, its
2787 effect is confined to that group, because once the group has been matched,
2788 there is never any backtracking into it. In this situation, backtracking can
2789 "jump back" to the left of the entire atomic group. (Remember also, as stated
2790 above, that this localization also applies in subroutine calls and assertions.)
2791 .P
2792 These verbs differ in exactly what kind of failure occurs when backtracking
2793 reaches them.
2794 .sp
2795 (*COMMIT)
2796 .sp
2797 This verb, which may not be followed by a name, causes the whole match to fail
2798 outright if the rest of the pattern does not match. Even if the pattern is
2799 unanchored, no further attempts to find a match by advancing the starting point
2800 take place. Once (*COMMIT) has been passed, \fBpcre_exec()\fP is committed to
2801 finding a match at the current starting point, or not at all. For example:
2802 .sp
2803 a+(*COMMIT)b
2804 .sp
2805 This matches "xxaab" but not "aacaab". It can be thought of as a kind of
2806 dynamic anchor, or "I've started, so I must finish." The name of the most
2807 recently passed (*MARK) in the path is passed back when (*COMMIT) forces a
2808 match failure.
2809 .P
2810 Note that (*COMMIT) at the start of a pattern is not the same as an anchor,
2811 unless PCRE's start-of-match optimizations are turned off, as shown in this
2812 \fBpcretest\fP example:
2813 .sp
2814 re> /(*COMMIT)abc/
2815 data> xyzabc
2816 0: abc
2817 xyzabc\eY
2818 No match
2819 .sp
2820 PCRE knows that any match must start with "a", so the optimization skips along
2821 the subject to "a" before running the first match attempt, which succeeds. When
2822 the optimization is disabled by the \eY escape in the second subject, the match
2823 starts at "x" and so the (*COMMIT) causes it to fail without trying any other
2824 starting points.
2825 .sp
2826 (*PRUNE) or (*PRUNE:NAME)
2827 .sp
2828 This verb causes the match to fail at the current starting position in the
2829 subject if the rest of the pattern does not match. If the pattern is
2830 unanchored, the normal "bumpalong" advance to the next starting character then
2831 happens. Backtracking can occur as usual to the left of (*PRUNE), before it is
2832 reached, or when matching to the right of (*PRUNE), but if there is no match to
2833 the right, backtracking cannot cross (*PRUNE). In simple cases, the use of
2834 (*PRUNE) is just an alternative to an atomic group or possessive quantifier,
2835 but there are some uses of (*PRUNE) that cannot be expressed in any other way.
2836 The behaviour of (*PRUNE:NAME) is the same as (*MARK:NAME)(*PRUNE). In an
2837 anchored pattern (*PRUNE) has the same effect as (*COMMIT).
2838 .sp
2839 (*SKIP)
2840 .sp
2841 This verb, when given without a name, is like (*PRUNE), except that if the
2842 pattern is unanchored, the "bumpalong" advance is not to the next character,
2843 but to the position in the subject where (*SKIP) was encountered. (*SKIP)
2844 signifies that whatever text was matched leading up to it cannot be part of a
2845 successful match. Consider:
2846 .sp
2847 a+(*SKIP)b
2848 .sp
2849 If the subject is "aaaac...", after the first match attempt fails (starting at
2850 the first character in the string), the starting point skips on to start the
2851 next attempt at "c". Note that a possessive quantifer does not have the same
2852 effect as this example; although it would suppress backtracking during the
2853 first match attempt, the second attempt would start at the second character
2854 instead of skipping on to "c".
2855 .sp
2856 (*SKIP:NAME)
2857 .sp
2858 When (*SKIP) has an associated name, its behaviour is modified. If the
2859 following pattern fails to match, the previous path through the pattern is
2860 searched for the most recent (*MARK) that has the same name. If one is found,
2861 the "bumpalong" advance is to the subject position that corresponds to that
2862 (*MARK) instead of to where (*SKIP) was encountered. If no (*MARK) with a
2863 matching name is found, the (*SKIP) is ignored.
2864 .sp
2865 (*THEN) or (*THEN:NAME)
2866 .sp
2867 This verb causes a skip to the next innermost alternative if the rest of the
2868 pattern does not match. That is, it cancels pending backtracking, but only
2869 within the current alternative. Its name comes from the observation that it can
2870 be used for a pattern-based if-then-else block:
2871 .sp
2872 ( COND1 (*THEN) FOO | COND2 (*THEN) BAR | COND3 (*THEN) BAZ ) ...
2873 .sp
2874 If the COND1 pattern matches, FOO is tried (and possibly further items after
2875 the end of the group if FOO succeeds); on failure, the matcher skips to the
2876 second alternative and tries COND2, without backtracking into COND1. The
2877 behaviour of (*THEN:NAME) is exactly the same as (*MARK:NAME)(*THEN).
2878 If (*THEN) is not inside an alternation, it acts like (*PRUNE).
2879 .P
2880 Note that a subpattern that does not contain a | character is just a part of
2881 the enclosing alternative; it is not a nested alternation with only one
2882 alternative. The effect of (*THEN) extends beyond such a subpattern to the
2883 enclosing alternative. Consider this pattern, where A, B, etc. are complex
2884 pattern fragments that do not contain any | characters at this level:
2885 .sp
2886 A (B(*THEN)C) | D
2887 .sp
2888 If A and B are matched, but there is a failure in C, matching does not
2889 backtrack into A; instead it moves to the next alternative, that is, D.
2890 However, if the subpattern containing (*THEN) is given an alternative, it
2891 behaves differently:
2892 .sp
2893 A (B(*THEN)C | (*FAIL)) | D
2894 .sp
2895 The effect of (*THEN) is now confined to the inner subpattern. After a failure
2896 in C, matching moves to (*FAIL), which causes the whole subpattern to fail
2897 because there are no more alternatives to try. In this case, matching does now
2898 backtrack into A.
2899 .P
2900 Note also that a conditional subpattern is not considered as having two
2901 alternatives, because only one is ever used. In other words, the | character in
2902 a conditional subpattern has a different meaning. Ignoring white space,
2903 consider:
2904 .sp
2905 ^.*? (?(?=a) a | b(*THEN)c )
2906 .sp
2907 If the subject is "ba", this pattern does not match. Because .*? is ungreedy,
2908 it initially matches zero characters. The condition (?=a) then fails, the
2909 character "b" is matched, but "c" is not. At this point, matching does not
2910 backtrack to .*? as might perhaps be expected from the presence of the |
2911 character. The conditional subpattern is part of the single alternative that
2912 comprises the whole pattern, and so the match fails. (If there was a backtrack
2913 into .*?, allowing it to match "b", the match would succeed.)
2914 .P
2915 The verbs just described provide four different "strengths" of control when
2916 subsequent matching fails. (*THEN) is the weakest, carrying on the match at the
2917 next alternative. (*PRUNE) comes next, failing the match at the current
2918 starting position, but allowing an advance to the next character (for an
2919 unanchored pattern). (*SKIP) is similar, except that the advance may be more
2920 than one character. (*COMMIT) is the strongest, causing the entire match to
2921 fail.
2922 .P
2923 If more than one such verb is present in a pattern, the "strongest" one wins.
2924 For example, consider this pattern, where A, B, etc. are complex pattern
2925 fragments:
2926 .sp
2927 (A(*COMMIT)B(*THEN)C|D)
2928 .sp
2929 Once A has matched, PCRE is committed to this match, at the current starting
2930 position. If subsequently B matches, but C does not, the normal (*THEN) action
2931 of trying the next alternative (that is, D) does not happen because (*COMMIT)
2932 overrides.
2933 .
2934 .
2935 .SH "SEE ALSO"
2936 .rs
2937 .sp
2938 \fBpcreapi\fP(3), \fBpcrecallout\fP(3), \fBpcrematching\fP(3),
2939 \fBpcresyntax\fP(3), \fBpcre\fP(3), \fBpcre16(3)\fP.
2940 .
2941 .
2943 .rs
2944 .sp
2945 .nf
2946 Philip Hazel
2947 University Computing Service
2948 Cambridge CB2 3QH, England.
2949 .fi
2950 .
2951 .
2953 .rs
2954 .sp
2955 .nf
2956 Last updated: 25 August 2012
2957 Copyright (c) 1997-2012 University of Cambridge.
2958 .fi


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

  ViewVC Help
Powered by ViewVC 1.1.5