ViewVC logotype

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

Parent Directory Parent Directory | Revision Log Revision Log

Revision 1364 - (show annotations)
Sat Oct 5 15:45:11 2013 UTC (7 years, 9 months ago) by ph10
File size: 131048 byte(s)
Add VT to the set of characters recognized as white space.
1 .TH PCREPATTERN 3 "05 October 2013" "PCRE 8.34"
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 This document discusses the patterns that are supported by PCRE when one its
24 main matching functions, \fBpcre_exec()\fP (8-bit) or \fBpcre[16|32]_exec()\fP
25 (16- or 32-bit), is used. PCRE also has alternative matching functions,
26 \fBpcre_dfa_exec()\fP and \fBpcre[16|32_dfa_exec()\fP, which match using a
27 different algorithm that is not Perl-compatible. Some of the features discussed
28 below are not available when DFA matching is used. The advantages and
29 disadvantages of the alternative functions, and how they differ from the normal
30 functions, are discussed in the
31 .\" HREF
32 \fBpcrematching\fP
33 .\"
34 page.
35 .
36 .
38 .rs
39 .sp
40 A number of options that can be passed to \fBpcre_compile()\fP can also be set
41 by special items at the start of a pattern. These are not Perl-compatible, but
42 are provided to make these options accessible to pattern writers who are not
43 able to change the program that processes the pattern. Any number of these
44 items may appear, but they must all be together right at the start of the
45 pattern string, and the letters must be in upper case.
46 .
47 .
48 .SS "UTF support"
49 .rs
50 .sp
51 The original operation of PCRE was on strings of one-byte characters. However,
52 there is now also support for UTF-8 strings in the original library, an
53 extra library that supports 16-bit and UTF-16 character strings, and a
54 third library that supports 32-bit and UTF-32 character strings. To use these
55 features, PCRE must be built to include appropriate support. When using UTF
56 strings you must either call the compiling function with the PCRE_UTF8,
57 PCRE_UTF16, or PCRE_UTF32 option, or the pattern must start with one of
58 these special sequences:
59 .sp
60 (*UTF8)
61 (*UTF16)
62 (*UTF32)
63 (*UTF)
64 .sp
65 (*UTF) is a generic sequence that can be used with any of the libraries.
66 Starting a pattern with such a sequence is equivalent to setting the relevant
67 option. How setting a UTF mode affects pattern matching is mentioned in several
68 places below. There is also a summary of features in the
69 .\" HREF
70 \fBpcreunicode\fP
71 .\"
72 page.
73 .P
74 Some applications that allow their users to supply patterns may wish to
75 restrict them to non-UTF data for security reasons. If the PCRE_NEVER_UTF
76 option is set at compile time, (*UTF) etc. are not allowed, and their
77 appearance causes an error.
78 .
79 .
80 .SS "Unicode property support"
81 .rs
82 .sp
83 Another special sequence that may appear at the start of a pattern is
84 .sp
85 (*UCP)
86 .sp
87 This has the same effect as setting the PCRE_UCP option: it causes sequences
88 such as \ed and \ew to use Unicode properties to determine character types,
89 instead of recognizing only characters with codes less than 128 via a lookup
90 table.
91 .
92 .
93 .SS "Disabling start-up optimizations"
94 .rs
95 .sp
96 If a pattern starts with (*NO_START_OPT), it has the same effect as setting the
97 PCRE_NO_START_OPTIMIZE option either at compile or matching time.
98 .
99 .
100 .\" HTML <a name="newlines"></a>
101 .SS "Newline conventions"
102 .rs
103 .sp
104 PCRE supports five different conventions for indicating line breaks in
105 strings: a single CR (carriage return) character, a single LF (linefeed)
106 character, the two-character sequence CRLF, any of the three preceding, or any
107 Unicode newline sequence. The
108 .\" HREF
109 \fBpcreapi\fP
110 .\"
111 page has
112 .\" HTML <a href="pcreapi.html#newlines">
113 .\" </a>
114 further discussion
115 .\"
116 about newlines, and shows how to set the newline convention in the
117 \fIoptions\fP arguments for the compiling and matching functions.
118 .P
119 It is also possible to specify a newline convention by starting a pattern
120 string with one of the following five sequences:
121 .sp
122 (*CR) carriage return
123 (*LF) linefeed
124 (*CRLF) carriage return, followed by linefeed
125 (*ANYCRLF) any of the three above
126 (*ANY) all Unicode newline sequences
127 .sp
128 These override the default and the options given to the compiling function. For
129 example, on a Unix system where LF is the default newline sequence, the pattern
130 .sp
131 (*CR)a.b
132 .sp
133 changes the convention to CR. That pattern matches "a\enb" because LF is no
134 longer a newline. If more than one of these settings is present, the last one
135 is used.
136 .P
137 The newline convention affects where the circumflex and dollar assertions are
138 true. It also affects the interpretation of the dot metacharacter when
139 PCRE_DOTALL is not set, and the behaviour of \eN. However, it does not affect
140 what the \eR escape sequence matches. By default, this is any Unicode newline
141 sequence, for Perl compatibility. However, this can be changed; see the
142 description of \eR in the section entitled
143 .\" HTML <a href="#newlineseq">
144 .\" </a>
145 "Newline sequences"
146 .\"
147 below. A change of \eR setting can be combined with a change of newline
148 convention.
149 .
150 .
151 .SS "Setting match and recursion limits"
152 .rs
153 .sp
154 The caller of \fBpcre_exec()\fP can set a limit on the number of times the
155 internal \fBmatch()\fP function is called and on the maximum depth of
156 recursive calls. These facilities are provided to catch runaway matches that
157 are provoked by patterns with huge matching trees (a typical example is a
158 pattern with nested unlimited repeats) and to avoid running out of system stack
159 by too much recursion. When one of these limits is reached, \fBpcre_exec()\fP
160 gives an error return. The limits can also be set by items at the start of the
161 pattern of the form
162 .sp
163 (*LIMIT_MATCH=d)
165 .sp
166 where d is any number of decimal digits. However, the value of the setting must
167 be less than the value set by the caller of \fBpcre_exec()\fP for it to have
168 any effect. In other words, the pattern writer can lower the limit set by the
169 programmer, but not raise it. If there is more than one setting of one of these
170 limits, the lower value is used.
171 .
172 .
174 .rs
175 .sp
176 PCRE can be compiled to run in an environment that uses EBCDIC as its character
177 code rather than ASCII or Unicode (typically a mainframe system). In the
178 sections below, character code values are ASCII or Unicode; in an EBCDIC
179 environment these characters may have different code values, and there are no
180 code points greater than 255.
181 .
182 .
184 .rs
185 .sp
186 A regular expression is a pattern that is matched against a subject string from
187 left to right. Most characters stand for themselves in a pattern, and match the
188 corresponding characters in the subject. As a trivial example, the pattern
189 .sp
190 The quick brown fox
191 .sp
192 matches a portion of a subject string that is identical to itself. When
193 caseless matching is specified (the PCRE_CASELESS option), letters are matched
194 independently of case. In a UTF mode, PCRE always understands the concept of
195 case for characters whose values are less than 128, so caseless matching is
196 always possible. For characters with higher values, the concept of case is
197 supported if PCRE is compiled with Unicode property support, but not otherwise.
198 If you want to use caseless matching for characters 128 and above, you must
199 ensure that PCRE is compiled with Unicode property support as well as with
200 UTF support.
201 .P
202 The power of regular expressions comes from the ability to include alternatives
203 and repetitions in the pattern. These are encoded in the pattern by the use of
204 \fImetacharacters\fP, which do not stand for themselves but instead are
205 interpreted in some special way.
206 .P
207 There are two different sets of metacharacters: those that are recognized
208 anywhere in the pattern except within square brackets, and those that are
209 recognized within square brackets. Outside square brackets, the metacharacters
210 are as follows:
211 .sp
212 \e general escape character with several uses
213 ^ assert start of string (or line, in multiline mode)
214 $ assert end of string (or line, in multiline mode)
215 . match any character except newline (by default)
216 [ start character class definition
217 | start of alternative branch
218 ( start subpattern
219 ) end subpattern
220 ? extends the meaning of (
221 also 0 or 1 quantifier
222 also quantifier minimizer
223 * 0 or more quantifier
224 + 1 or more quantifier
225 also "possessive quantifier"
226 { start min/max quantifier
227 .sp
228 Part of a pattern that is in square brackets is called a "character class". In
229 a character class the only metacharacters are:
230 .sp
231 \e general escape character
232 ^ negate the class, but only if the first character
233 - indicates character range
234 .\" JOIN
235 [ POSIX character class (only if followed by POSIX
236 syntax)
237 ] terminates the character class
238 .sp
239 The following sections describe the use of each of the metacharacters.
240 .
241 .
243 .rs
244 .sp
245 The backslash character has several uses. Firstly, if it is followed by a
246 character that is not a number or a letter, it takes away any special meaning
247 that character may have. This use of backslash as an escape character applies
248 both inside and outside character classes.
249 .P
250 For example, if you want to match a * character, you write \e* in the pattern.
251 This escaping action applies whether or not the following character would
252 otherwise be interpreted as a metacharacter, so it is always safe to precede a
253 non-alphanumeric with backslash to specify that it stands for itself. In
254 particular, if you want to match a backslash, you write \e\e.
255 .P
256 In a UTF mode, only ASCII numbers and letters have any special meaning after a
257 backslash. All other characters (in particular, those whose codepoints are
258 greater than 127) are treated as literals.
259 .P
260 If a pattern is compiled with the PCRE_EXTENDED option, white space in the
261 pattern (other than in a character class) and characters between a # outside
262 a character class and the next newline are ignored. An escaping backslash can
263 be used to include a white space or # character as part of the pattern.
264 .P
265 If you want to remove the special meaning from a sequence of characters, you
266 can do so by putting them between \eQ and \eE. This is different from Perl in
267 that $ and @ are handled as literals in \eQ...\eE sequences in PCRE, whereas in
268 Perl, $ and @ cause variable interpolation. Note the following examples:
269 .sp
270 Pattern PCRE matches Perl matches
271 .sp
272 .\" JOIN
273 \eQabc$xyz\eE abc$xyz abc followed by the
274 contents of $xyz
275 \eQabc\e$xyz\eE abc\e$xyz abc\e$xyz
276 \eQabc\eE\e$\eQxyz\eE abc$xyz abc$xyz
277 .sp
278 The \eQ...\eE sequence is recognized both inside and outside character classes.
279 An isolated \eE that is not preceded by \eQ is ignored. If \eQ is not followed
280 by \eE later in the pattern, the literal interpretation continues to the end of
281 the pattern (that is, \eE is assumed at the end). If the isolated \eQ is inside
282 a character class, this causes an error, because the character class is not
283 terminated.
284 .
285 .
286 .\" HTML <a name="digitsafterbackslash"></a>
287 .SS "Non-printing characters"
288 .rs
289 .sp
290 A second use of backslash provides a way of encoding non-printing characters
291 in patterns in a visible manner. There is no restriction on the appearance of
292 non-printing characters, apart from the binary zero that terminates a pattern,
293 but when a pattern is being prepared by text editing, it is often easier to use
294 one of the following escape sequences than the binary character it represents:
295 .sp
296 \ea alarm, that is, the BEL character (hex 07)
297 \ecx "control-x", where x is any ASCII character
298 \ee escape (hex 1B)
299 \ef form feed (hex 0C)
300 \en linefeed (hex 0A)
301 \er carriage return (hex 0D)
302 \et tab (hex 09)
303 \eddd character with octal code ddd, or back reference
304 \exhh character with hex code hh
305 \ex{hhh..} character with hex code hhh.. (non-JavaScript mode)
306 \euhhhh character with hex code hhhh (JavaScript mode only)
307 .sp
308 The precise effect of \ecx on ASCII characters is as follows: if x is a lower
309 case letter, it is converted to upper case. Then bit 6 of the character (hex
310 40) is inverted. Thus \ecA to \ecZ become hex 01 to hex 1A (A is 41, Z is 5A),
311 but \ec{ becomes hex 3B ({ is 7B), and \ec; becomes hex 7B (; is 3B). If the
312 data item (byte or 16-bit value) following \ec has a value greater than 127, a
313 compile-time error occurs. This locks out non-ASCII characters in all modes.
314 .P
315 The \ec facility was designed for use with ASCII characters, but with the
316 extension to Unicode it is even less useful than it once was. It is, however,
317 recognized when PCRE is compiled in EBCDIC mode, where data items are always
318 bytes. In this mode, all values are valid after \ec. If the next character is a
319 lower case letter, it is converted to upper case. Then the 0xc0 bits of the
320 byte are inverted. Thus \ecA becomes hex 01, as in ASCII (A is C1), but because
321 the EBCDIC letters are disjoint, \ecZ becomes hex 29 (Z is E9), and other
322 characters also generate different values.
323 .P
324 By default, after \ex, from zero to two hexadecimal digits are read (letters
325 can be in upper or lower case). Any number of hexadecimal digits may appear
326 between \ex{ and }, but the character code is constrained as follows:
327 .sp
328 8-bit non-UTF mode less than 0x100
329 8-bit UTF-8 mode less than 0x10ffff and a valid codepoint
330 16-bit non-UTF mode less than 0x10000
331 16-bit UTF-16 mode less than 0x10ffff and a valid codepoint
332 32-bit non-UTF mode less than 0x80000000
333 32-bit UTF-32 mode less than 0x10ffff and a valid codepoint
334 .sp
335 Invalid Unicode codepoints are the range 0xd800 to 0xdfff (the so-called
336 "surrogate" codepoints), and 0xffef.
337 .P
338 If characters other than hexadecimal digits appear between \ex{ and }, or if
339 there is no terminating }, this form of escape is not recognized. Instead, the
340 initial \ex will be interpreted as a basic hexadecimal escape, with no
341 following digits, giving a character whose value is zero.
342 .P
343 If the PCRE_JAVASCRIPT_COMPAT option is set, the interpretation of \ex is
344 as just described only when it is followed by two hexadecimal digits.
345 Otherwise, it matches a literal "x" character. In JavaScript mode, support for
346 code points greater than 256 is provided by \eu, which must be followed by
347 four hexadecimal digits; otherwise it matches a literal "u" character.
348 Character codes specified by \eu in JavaScript mode are constrained in the same
349 was as those specified by \ex in non-JavaScript mode.
350 .P
351 Characters whose value is less than 256 can be defined by either of the two
352 syntaxes for \ex (or by \eu in JavaScript mode). There is no difference in the
353 way they are handled. For example, \exdc is exactly the same as \ex{dc} (or
354 \eu00dc in JavaScript mode).
355 .P
356 After \e0 up to two further octal digits are read. If there are fewer than two
357 digits, just those that are present are used. Thus the sequence \e0\ex\e07
358 specifies two binary zeros followed by a BEL character (code value 7). Make
359 sure you supply two digits after the initial zero if the pattern character that
360 follows is itself an octal digit.
361 .P
362 The handling of a backslash followed by a digit other than 0 is complicated.
363 Outside a character class, PCRE reads it and any following digits as a decimal
364 number. If the number is less than 10, or if there have been at least that many
365 previous capturing left parentheses in the expression, the entire sequence is
366 taken as a \fIback reference\fP. A description of how this works is given
367 .\" HTML <a href="#backreferences">
368 .\" </a>
369 later,
370 .\"
371 following the discussion of
372 .\" HTML <a href="#subpattern">
373 .\" </a>
374 parenthesized subpatterns.
375 .\"
376 .P
377 Inside a character class, or if the decimal number is greater than 9 and there
378 have not been that many capturing subpatterns, PCRE re-reads up to three octal
379 digits following the backslash, and uses them to generate a data character. Any
380 subsequent digits stand for themselves. The value of the character is
381 constrained in the same way as characters specified in hexadecimal.
382 For example:
383 .sp
384 \e040 is another way of writing an ASCII space
385 .\" JOIN
386 \e40 is the same, provided there are fewer than 40
387 previous capturing subpatterns
388 \e7 is always a back reference
389 .\" JOIN
390 \e11 might be a back reference, or another way of
391 writing a tab
392 \e011 is always a tab
393 \e0113 is a tab followed by the character "3"
394 .\" JOIN
395 \e113 might be a back reference, otherwise the
396 character with octal code 113
397 .\" JOIN
398 \e377 might be a back reference, otherwise
399 the value 255 (decimal)
400 .\" JOIN
401 \e81 is either a back reference, or a binary zero
402 followed by the two characters "8" and "1"
403 .sp
404 Note that octal values of 100 or greater must not be introduced by a leading
405 zero, because no more than three octal digits are ever read.
406 .P
407 All the sequences that define a single character value can be used both inside
408 and outside character classes. In addition, inside a character class, \eb is
409 interpreted as the backspace character (hex 08).
410 .P
411 \eN is not allowed in a character class. \eB, \eR, and \eX are not special
412 inside a character class. Like other unrecognized escape sequences, they are
413 treated as the literal characters "B", "R", and "X" by default, but cause an
414 error if the PCRE_EXTRA option is set. Outside a character class, these
415 sequences have different meanings.
416 .
417 .
418 .SS "Unsupported escape sequences"
419 .rs
420 .sp
421 In Perl, the sequences \el, \eL, \eu, and \eU are recognized by its string
422 handler and used to modify the case of following characters. By default, PCRE
423 does not support these escape sequences. However, if the PCRE_JAVASCRIPT_COMPAT
424 option is set, \eU matches a "U" character, and \eu can be used to define a
425 character by code point, as described in the previous section.
426 .
427 .
428 .SS "Absolute and relative back references"
429 .rs
430 .sp
431 The sequence \eg followed by an unsigned or a negative number, optionally
432 enclosed in braces, is an absolute or relative back reference. A named back
433 reference can be coded as \eg{name}. Back references are discussed
434 .\" HTML <a href="#backreferences">
435 .\" </a>
436 later,
437 .\"
438 following the discussion of
439 .\" HTML <a href="#subpattern">
440 .\" </a>
441 parenthesized subpatterns.
442 .\"
443 .
444 .
445 .SS "Absolute and relative subroutine calls"
446 .rs
447 .sp
448 For compatibility with Oniguruma, the non-Perl syntax \eg followed by a name or
449 a number enclosed either in angle brackets or single quotes, is an alternative
450 syntax for referencing a subpattern as a "subroutine". Details are discussed
451 .\" HTML <a href="#onigurumasubroutines">
452 .\" </a>
453 later.
454 .\"
455 Note that \eg{...} (Perl syntax) and \eg<...> (Oniguruma syntax) are \fInot\fP
456 synonymous. The former is a back reference; the latter is a
457 .\" HTML <a href="#subpatternsassubroutines">
458 .\" </a>
459 subroutine
460 .\"
461 call.
462 .
463 .
464 .\" HTML <a name="genericchartypes"></a>
465 .SS "Generic character types"
466 .rs
467 .sp
468 Another use of backslash is for specifying generic character types:
469 .sp
470 \ed any decimal digit
471 \eD any character that is not a decimal digit
472 \eh any horizontal white space character
473 \eH any character that is not a horizontal white space character
474 \es any white space character
475 \eS any character that is not a white space character
476 \ev any vertical white space character
477 \eV any character that is not a vertical white space character
478 \ew any "word" character
479 \eW any "non-word" character
480 .sp
481 There is also the single sequence \eN, which matches a non-newline character.
482 This is the same as
483 .\" HTML <a href="#fullstopdot">
484 .\" </a>
485 the "." metacharacter
486 .\"
487 when PCRE_DOTALL is not set. Perl also uses \eN to match characters by name;
488 PCRE does not support this.
489 .P
490 Each pair of lower and upper case escape sequences partitions the complete set
491 of characters into two disjoint sets. Any given character matches one, and only
492 one, of each pair. The sequences can appear both inside and outside character
493 classes. They each match one character of the appropriate type. If the current
494 matching point is at the end of the subject string, all of them fail, because
495 there is no character to match.
496 .P
497 For compatibility with Perl, \es did not used to match the VT character (code
498 11), which made it different from the the POSIX "space" class. However, Perl
499 added VT at release 5.18, and PCRE followed suit at release 8.34. The \es
500 characters are now HT (9), LF (10), VT (11), FF (12), CR (13), and space (32).
501 .P
502 A "word" character is an underscore or any character that is a letter or digit.
503 By default, the definition of letters and digits is controlled by PCRE's
504 low-valued character tables, and may vary if locale-specific matching is taking
505 place (see
506 .\" HTML <a href="pcreapi.html#localesupport">
507 .\" </a>
508 "Locale support"
509 .\"
510 in the
511 .\" HREF
512 \fBpcreapi\fP
513 .\"
514 page). For example, in a French locale such as "fr_FR" in Unix-like systems,
515 or "french" in Windows, some character codes greater than 128 are used for
516 accented letters, and these are then matched by \ew. The use of locales with
517 Unicode is discouraged.
518 .P
519 By default, in a UTF mode, characters with values greater than 128 never match
520 \ed, \es, or \ew, and always match \eD, \eS, and \eW. These sequences retain
521 their original meanings from before UTF support was available, mainly for
522 efficiency reasons. However, if PCRE is compiled with Unicode property support,
523 and the PCRE_UCP option is set, the behaviour is changed so that Unicode
524 properties are used to determine character types, as follows:
525 .sp
526 \ed any character that \ep{Nd} matches (decimal digit)
527 \es any character that \ep{Z} matches, plus HT, LF, FF, CR
528 \ew any character that \ep{L} or \ep{N} matches, plus underscore
529 .sp
530 The upper case escapes match the inverse sets of characters. Note that \ed
531 matches only decimal digits, whereas \ew matches any Unicode digit, as well as
532 any Unicode letter, and underscore. Note also that PCRE_UCP affects \eb, and
533 \eB because they are defined in terms of \ew and \eW. Matching these sequences
534 is noticeably slower when PCRE_UCP is set.
535 .P
536 The sequences \eh, \eH, \ev, and \eV are features that were added to Perl at
537 release 5.10. In contrast to the other sequences, which match only ASCII
538 characters by default, these always match certain high-valued codepoints,
539 whether or not PCRE_UCP is set. The horizontal space characters are:
540 .sp
541 U+0009 Horizontal tab (HT)
542 U+0020 Space
543 U+00A0 Non-break space
544 U+1680 Ogham space mark
545 U+180E Mongolian vowel separator
546 U+2000 En quad
547 U+2001 Em quad
548 U+2002 En space
549 U+2003 Em space
550 U+2004 Three-per-em space
551 U+2005 Four-per-em space
552 U+2006 Six-per-em space
553 U+2007 Figure space
554 U+2008 Punctuation space
555 U+2009 Thin space
556 U+200A Hair space
557 U+202F Narrow no-break space
558 U+205F Medium mathematical space
559 U+3000 Ideographic space
560 .sp
561 The vertical space characters are:
562 .sp
563 U+000A Linefeed (LF)
564 U+000B Vertical tab (VT)
565 U+000C Form feed (FF)
566 U+000D Carriage return (CR)
567 U+0085 Next line (NEL)
568 U+2028 Line separator
569 U+2029 Paragraph separator
570 .sp
571 In 8-bit, non-UTF-8 mode, only the characters with codepoints less than 256 are
572 relevant.
573 .
574 .
575 .\" HTML <a name="newlineseq"></a>
576 .SS "Newline sequences"
577 .rs
578 .sp
579 Outside a character class, by default, the escape sequence \eR matches any
580 Unicode newline sequence. In 8-bit non-UTF-8 mode \eR is equivalent to the
581 following:
582 .sp
583 (?>\er\en|\en|\ex0b|\ef|\er|\ex85)
584 .sp
585 This is an example of an "atomic group", details of which are given
586 .\" HTML <a href="#atomicgroup">
587 .\" </a>
588 below.
589 .\"
590 This particular group matches either the two-character sequence CR followed by
591 LF, or one of the single characters LF (linefeed, U+000A), VT (vertical tab,
592 U+000B), FF (form feed, U+000C), CR (carriage return, U+000D), or NEL (next
593 line, U+0085). The two-character sequence is treated as a single unit that
594 cannot be split.
595 .P
596 In other modes, two additional characters whose codepoints are greater than 255
597 are added: LS (line separator, U+2028) and PS (paragraph separator, U+2029).
598 Unicode character property support is not needed for these characters to be
599 recognized.
600 .P
601 It is possible to restrict \eR to match only CR, LF, or CRLF (instead of the
602 complete set of Unicode line endings) by setting the option PCRE_BSR_ANYCRLF
603 either at compile time or when the pattern is matched. (BSR is an abbrevation
604 for "backslash R".) This can be made the default when PCRE is built; if this is
605 the case, the other behaviour can be requested via the PCRE_BSR_UNICODE option.
606 It is also possible to specify these settings by starting a pattern string with
607 one of the following sequences:
608 .sp
609 (*BSR_ANYCRLF) CR, LF, or CRLF only
610 (*BSR_UNICODE) any Unicode newline sequence
611 .sp
612 These override the default and the options given to the compiling function, but
613 they can themselves be overridden by options given to a matching function. Note
614 that these special settings, which are not Perl-compatible, are recognized only
615 at the very start of a pattern, and that they must be in upper case. If more
616 than one of them is present, the last one is used. They can be combined with a
617 change of newline convention; for example, a pattern can start with:
618 .sp
620 .sp
621 They can also be combined with the (*UTF8), (*UTF16), (*UTF32), (*UTF) or
622 (*UCP) special sequences. Inside a character class, \eR is treated as an
623 unrecognized escape sequence, and so matches the letter "R" by default, but
624 causes an error if PCRE_EXTRA is set.
625 .
626 .
627 .\" HTML <a name="uniextseq"></a>
628 .SS Unicode character properties
629 .rs
630 .sp
631 When PCRE is built with Unicode character property support, three additional
632 escape sequences that match characters with specific properties are available.
633 When in 8-bit non-UTF-8 mode, these sequences are of course limited to testing
634 characters whose codepoints are less than 256, but they do work in this mode.
635 The extra escape sequences are:
636 .sp
637 \ep{\fIxx\fP} a character with the \fIxx\fP property
638 \eP{\fIxx\fP} a character without the \fIxx\fP property
639 \eX a Unicode extended grapheme cluster
640 .sp
641 The property names represented by \fIxx\fP above are limited to the Unicode
642 script names, the general category properties, "Any", which matches any
643 character (including newline), and some special PCRE properties (described
644 in the
645 .\" HTML <a href="#extraprops">
646 .\" </a>
647 next section).
648 .\"
649 Other Perl properties such as "InMusicalSymbols" are not currently supported by
650 PCRE. Note that \eP{Any} does not match any characters, so always causes a
651 match failure.
652 .P
653 Sets of Unicode characters are defined as belonging to certain scripts. A
654 character from one of these sets can be matched using a script name. For
655 example:
656 .sp
657 \ep{Greek}
658 \eP{Han}
659 .sp
660 Those that are not part of an identified script are lumped together as
661 "Common". The current list of scripts is:
662 .P
663 Arabic,
664 Armenian,
665 Avestan,
666 Balinese,
667 Bamum,
668 Batak,
669 Bengali,
670 Bopomofo,
671 Brahmi,
672 Braille,
673 Buginese,
674 Buhid,
675 Canadian_Aboriginal,
676 Carian,
677 Chakma,
678 Cham,
679 Cherokee,
680 Common,
681 Coptic,
682 Cuneiform,
683 Cypriot,
684 Cyrillic,
685 Deseret,
686 Devanagari,
687 Egyptian_Hieroglyphs,
688 Ethiopic,
689 Georgian,
690 Glagolitic,
691 Gothic,
692 Greek,
693 Gujarati,
694 Gurmukhi,
695 Han,
696 Hangul,
697 Hanunoo,
698 Hebrew,
699 Hiragana,
700 Imperial_Aramaic,
701 Inherited,
702 Inscriptional_Pahlavi,
703 Inscriptional_Parthian,
704 Javanese,
705 Kaithi,
706 Kannada,
707 Katakana,
708 Kayah_Li,
709 Kharoshthi,
710 Khmer,
711 Lao,
712 Latin,
713 Lepcha,
714 Limbu,
715 Linear_B,
716 Lisu,
717 Lycian,
718 Lydian,
719 Malayalam,
720 Mandaic,
721 Meetei_Mayek,
722 Meroitic_Cursive,
723 Meroitic_Hieroglyphs,
724 Miao,
725 Mongolian,
726 Myanmar,
727 New_Tai_Lue,
728 Nko,
729 Ogham,
730 Old_Italic,
731 Old_Persian,
732 Old_South_Arabian,
733 Old_Turkic,
734 Ol_Chiki,
735 Oriya,
736 Osmanya,
737 Phags_Pa,
738 Phoenician,
739 Rejang,
740 Runic,
741 Samaritan,
742 Saurashtra,
743 Sharada,
744 Shavian,
745 Sinhala,
746 Sora_Sompeng,
747 Sundanese,
748 Syloti_Nagri,
749 Syriac,
750 Tagalog,
751 Tagbanwa,
752 Tai_Le,
753 Tai_Tham,
754 Tai_Viet,
755 Takri,
756 Tamil,
757 Telugu,
758 Thaana,
759 Thai,
760 Tibetan,
761 Tifinagh,
762 Ugaritic,
763 Vai,
764 Yi.
765 .P
766 Each character has exactly one Unicode general category property, specified by
767 a two-letter abbreviation. For compatibility with Perl, negation can be
768 specified by including a circumflex between the opening brace and the property
769 name. For example, \ep{^Lu} is the same as \eP{Lu}.
770 .P
771 If only one letter is specified with \ep or \eP, it includes all the general
772 category properties that start with that letter. In this case, in the absence
773 of negation, the curly brackets in the escape sequence are optional; these two
774 examples have the same effect:
775 .sp
776 \ep{L}
777 \epL
778 .sp
779 The following general category property codes are supported:
780 .sp
781 C Other
782 Cc Control
783 Cf Format
784 Cn Unassigned
785 Co Private use
786 Cs Surrogate
787 .sp
788 L Letter
789 Ll Lower case letter
790 Lm Modifier letter
791 Lo Other letter
792 Lt Title case letter
793 Lu Upper case letter
794 .sp
795 M Mark
796 Mc Spacing mark
797 Me Enclosing mark
798 Mn Non-spacing mark
799 .sp
800 N Number
801 Nd Decimal number
802 Nl Letter number
803 No Other number
804 .sp
805 P Punctuation
806 Pc Connector punctuation
807 Pd Dash punctuation
808 Pe Close punctuation
809 Pf Final punctuation
810 Pi Initial punctuation
811 Po Other punctuation
812 Ps Open punctuation
813 .sp
814 S Symbol
815 Sc Currency symbol
816 Sk Modifier symbol
817 Sm Mathematical symbol
818 So Other symbol
819 .sp
820 Z Separator
821 Zl Line separator
822 Zp Paragraph separator
823 Zs Space separator
824 .sp
825 The special property L& is also supported: it matches a character that has
826 the Lu, Ll, or Lt property, in other words, a letter that is not classified as
827 a modifier or "other".
828 .P
829 The Cs (Surrogate) property applies only to characters in the range U+D800 to
830 U+DFFF. Such characters are not valid in Unicode strings and so
831 cannot be tested by PCRE, unless UTF validity checking has been turned off
832 (see the discussion of PCRE_NO_UTF8_CHECK, PCRE_NO_UTF16_CHECK and
833 PCRE_NO_UTF32_CHECK in the
834 .\" HREF
835 \fBpcreapi\fP
836 .\"
837 page). Perl does not support the Cs property.
838 .P
839 The long synonyms for property names that Perl supports (such as \ep{Letter})
840 are not supported by PCRE, nor is it permitted to prefix any of these
841 properties with "Is".
842 .P
843 No character that is in the Unicode table has the Cn (unassigned) property.
844 Instead, this property is assumed for any code point that is not in the
845 Unicode table.
846 .P
847 Specifying caseless matching does not affect these escape sequences. For
848 example, \ep{Lu} always matches only upper case letters. This is different from
849 the behaviour of current versions of Perl.
850 .P
851 Matching characters by Unicode property is not fast, because PCRE has to do a
852 multistage table lookup in order to find a character's property. That is why
853 the traditional escape sequences such as \ed and \ew do not use Unicode
854 properties in PCRE by default, though you can make them do so by setting the
855 PCRE_UCP option or by starting the pattern with (*UCP).
856 .
857 .
858 .SS Extended grapheme clusters
859 .rs
860 .sp
861 The \eX escape matches any number of Unicode characters that form an "extended
862 grapheme cluster", and treats the sequence as an atomic group
863 .\" HTML <a href="#atomicgroup">
864 .\" </a>
865 (see below).
866 .\"
867 Up to and including release 8.31, PCRE matched an earlier, simpler definition
868 that was equivalent to
869 .sp
870 (?>\ePM\epM*)
871 .sp
872 That is, it matched a character without the "mark" property, followed by zero
873 or more characters with the "mark" property. Characters with the "mark"
874 property are typically non-spacing accents that affect the preceding character.
875 .P
876 This simple definition was extended in Unicode to include more complicated
877 kinds of composite character by giving each character a grapheme breaking
878 property, and creating rules that use these properties to define the boundaries
879 of extended grapheme clusters. In releases of PCRE later than 8.31, \eX matches
880 one of these clusters.
881 .P
882 \eX always matches at least one character. Then it decides whether to add
883 additional characters according to the following rules for ending a cluster:
884 .P
885 1. End at the end of the subject string.
886 .P
887 2. Do not end between CR and LF; otherwise end after any control character.
888 .P
889 3. Do not break Hangul (a Korean script) syllable sequences. Hangul characters
890 are of five types: L, V, T, LV, and LVT. An L character may be followed by an
891 L, V, LV, or LVT character; an LV or V character may be followed by a V or T
892 character; an LVT or T character may be follwed only by a T character.
893 .P
894 4. Do not end before extending characters or spacing marks. Characters with
895 the "mark" property always have the "extend" grapheme breaking property.
896 .P
897 5. Do not end after prepend characters.
898 .P
899 6. Otherwise, end the cluster.
900 .
901 .
902 .\" HTML <a name="extraprops"></a>
903 .SS PCRE's additional properties
904 .rs
905 .sp
906 As well as the standard Unicode properties described above, PCRE supports four
907 more that make it possible to convert traditional escape sequences such as \ew
908 and \es and POSIX character classes to use Unicode properties. PCRE uses these
909 non-standard, non-Perl properties internally when PCRE_UCP is set. However,
910 they may also be used explicitly. These properties are:
911 .sp
912 Xan Any alphanumeric character
913 Xps Any POSIX space character
914 Xsp Any Perl space character
915 Xwd Any Perl "word" character
916 .sp
917 Xan matches characters that have either the L (letter) or the N (number)
918 property. Xps matches the characters tab, linefeed, vertical tab, form feed, or
919 carriage return, and any other character that has the Z (separator) property.
920 Xsp is the same as Xps, except that vertical tab is excluded. Xwd matches the
921 same characters as Xan, plus underscore.
922 .P
923 There is another non-standard property, Xuc, which matches any character that
924 can be represented by a Universal Character Name in C++ and other programming
925 languages. These are the characters $, @, ` (grave accent), and all characters
926 with Unicode code points greater than or equal to U+00A0, except for the
927 surrogates U+D800 to U+DFFF. Note that most base (ASCII) characters are
928 excluded. (Universal Character Names are of the form \euHHHH or \eUHHHHHHHH
929 where H is a hexadecimal digit. Note that the Xuc property does not match these
930 sequences but the characters that they represent.)
931 .
932 .
933 .\" HTML <a name="resetmatchstart"></a>
934 .SS "Resetting the match start"
935 .rs
936 .sp
937 The escape sequence \eK causes any previously matched characters not to be
938 included in the final matched sequence. For example, the pattern:
939 .sp
940 foo\eKbar
941 .sp
942 matches "foobar", but reports that it has matched "bar". This feature is
943 similar to a lookbehind assertion
944 .\" HTML <a href="#lookbehind">
945 .\" </a>
946 (described below).
947 .\"
948 However, in this case, the part of the subject before the real match does not
949 have to be of fixed length, as lookbehind assertions do. The use of \eK does
950 not interfere with the setting of
951 .\" HTML <a href="#subpattern">
952 .\" </a>
953 captured substrings.
954 .\"
955 For example, when the pattern
956 .sp
957 (foo)\eKbar
958 .sp
959 matches "foobar", the first substring is still set to "foo".
960 .P
961 Perl documents that the use of \eK within assertions is "not well defined". In
962 PCRE, \eK is acted upon when it occurs inside positive assertions, but is
963 ignored in negative assertions.
964 .
965 .
966 .\" HTML <a name="smallassertions"></a>
967 .SS "Simple assertions"
968 .rs
969 .sp
970 The final use of backslash is for certain simple assertions. An assertion
971 specifies a condition that has to be met at a particular point in a match,
972 without consuming any characters from the subject string. The use of
973 subpatterns for more complicated assertions is described
974 .\" HTML <a href="#bigassertions">
975 .\" </a>
976 below.
977 .\"
978 The backslashed assertions are:
979 .sp
980 \eb matches at a word boundary
981 \eB matches when not at a word boundary
982 \eA matches at the start of the subject
983 \eZ matches at the end of the subject
984 also matches before a newline at the end of the subject
985 \ez matches only at the end of the subject
986 \eG matches at the first matching position in the subject
987 .sp
988 Inside a character class, \eb has a different meaning; it matches the backspace
989 character. If any other of these assertions appears in a character class, by
990 default it matches the corresponding literal character (for example, \eB
991 matches the letter B). However, if the PCRE_EXTRA option is set, an "invalid
992 escape sequence" error is generated instead.
993 .P
994 A word boundary is a position in the subject string where the current character
995 and the previous character do not both match \ew or \eW (i.e. one matches
996 \ew and the other matches \eW), or the start or end of the string if the
997 first or last character matches \ew, respectively. In a UTF mode, the meanings
998 of \ew and \eW can be changed by setting the PCRE_UCP option. When this is
999 done, it also affects \eb and \eB. Neither PCRE nor Perl has a separate "start
1000 of word" or "end of word" metasequence. However, whatever follows \eb normally
1001 determines which it is. For example, the fragment \eba matches "a" at the start
1002 of a word.
1003 .P
1004 The \eA, \eZ, and \ez assertions differ from the traditional circumflex and
1005 dollar (described in the next section) in that they only ever match at the very
1006 start and end of the subject string, whatever options are set. Thus, they are
1007 independent of multiline mode. These three assertions are not affected by the
1008 PCRE_NOTBOL or PCRE_NOTEOL options, which affect only the behaviour of the
1009 circumflex and dollar metacharacters. However, if the \fIstartoffset\fP
1010 argument of \fBpcre_exec()\fP is non-zero, indicating that matching is to start
1011 at a point other than the beginning of the subject, \eA can never match. The
1012 difference between \eZ and \ez is that \eZ matches before a newline at the end
1013 of the string as well as at the very end, whereas \ez matches only at the end.
1014 .P
1015 The \eG assertion is true only when the current matching position is at the
1016 start point of the match, as specified by the \fIstartoffset\fP argument of
1017 \fBpcre_exec()\fP. It differs from \eA when the value of \fIstartoffset\fP is
1018 non-zero. By calling \fBpcre_exec()\fP multiple times with appropriate
1019 arguments, you can mimic Perl's /g option, and it is in this kind of
1020 implementation where \eG can be useful.
1021 .P
1022 Note, however, that PCRE's interpretation of \eG, as the start of the current
1023 match, is subtly different from Perl's, which defines it as the end of the
1024 previous match. In Perl, these can be different when the previously matched
1025 string was empty. Because PCRE does just one match at a time, it cannot
1026 reproduce this behaviour.
1027 .P
1028 If all the alternatives of a pattern begin with \eG, the expression is anchored
1029 to the starting match position, and the "anchored" flag is set in the compiled
1030 regular expression.
1031 .
1032 .
1034 .rs
1035 .sp
1036 The circumflex and dollar metacharacters are zero-width assertions. That is,
1037 they test for a particular condition being true without consuming any
1038 characters from the subject string.
1039 .P
1040 Outside a character class, in the default matching mode, the circumflex
1041 character is an assertion that is true only if the current matching point is at
1042 the start of the subject string. If the \fIstartoffset\fP argument of
1043 \fBpcre_exec()\fP is non-zero, circumflex can never match if the PCRE_MULTILINE
1044 option is unset. Inside a character class, circumflex has an entirely different
1045 meaning
1046 .\" HTML <a href="#characterclass">
1047 .\" </a>
1048 (see below).
1049 .\"
1050 .P
1051 Circumflex need not be the first character of the pattern if a number of
1052 alternatives are involved, but it should be the first thing in each alternative
1053 in which it appears if the pattern is ever to match that branch. If all
1054 possible alternatives start with a circumflex, that is, if the pattern is
1055 constrained to match only at the start of the subject, it is said to be an
1056 "anchored" pattern. (There are also other constructs that can cause a pattern
1057 to be anchored.)
1058 .P
1059 The dollar character is an assertion that is true only if the current matching
1060 point is at the end of the subject string, or immediately before a newline at
1061 the end of the string (by default). Note, however, that it does not actually
1062 match the newline. Dollar need not be the last character of the pattern if a
1063 number of alternatives are involved, but it should be the last item in any
1064 branch in which it appears. Dollar has no special meaning in a character class.
1065 .P
1066 The meaning of dollar can be changed so that it matches only at the very end of
1067 the string, by setting the PCRE_DOLLAR_ENDONLY option at compile time. This
1068 does not affect the \eZ assertion.
1069 .P
1070 The meanings of the circumflex and dollar characters are changed if the
1071 PCRE_MULTILINE option is set. When this is the case, a circumflex matches
1072 immediately after internal newlines as well as at the start of the subject
1073 string. It does not match after a newline that ends the string. A dollar
1074 matches before any newlines in the string, as well as at the very end, when
1075 PCRE_MULTILINE is set. When newline is specified as the two-character
1076 sequence CRLF, isolated CR and LF characters do not indicate newlines.
1077 .P
1078 For example, the pattern /^abc$/ matches the subject string "def\enabc" (where
1079 \en represents a newline) in multiline mode, but not otherwise. Consequently,
1080 patterns that are anchored in single line mode because all branches start with
1081 ^ are not anchored in multiline mode, and a match for circumflex is possible
1082 when the \fIstartoffset\fP argument of \fBpcre_exec()\fP is non-zero. The
1083 PCRE_DOLLAR_ENDONLY option is ignored if PCRE_MULTILINE is set.
1084 .P
1085 Note that the sequences \eA, \eZ, and \ez can be used to match the start and
1086 end of the subject in both modes, and if all branches of a pattern start with
1087 \eA it is always anchored, whether or not PCRE_MULTILINE is set.
1088 .
1089 .
1090 .\" HTML <a name="fullstopdot"></a>
1092 .rs
1093 .sp
1094 Outside a character class, a dot in the pattern matches any one character in
1095 the subject string except (by default) a character that signifies the end of a
1096 line.
1097 .P
1098 When a line ending is defined as a single character, dot never matches that
1099 character; when the two-character sequence CRLF is used, dot does not match CR
1100 if it is immediately followed by LF, but otherwise it matches all characters
1101 (including isolated CRs and LFs). When any Unicode line endings are being
1102 recognized, dot does not match CR or LF or any of the other line ending
1103 characters.
1104 .P
1105 The behaviour of dot with regard to newlines can be changed. If the PCRE_DOTALL
1106 option is set, a dot matches any one character, without exception. If the
1107 two-character sequence CRLF is present in the subject string, it takes two dots
1108 to match it.
1109 .P
1110 The handling of dot is entirely independent of the handling of circumflex and
1111 dollar, the only relationship being that they both involve newlines. Dot has no
1112 special meaning in a character class.
1113 .P
1114 The escape sequence \eN behaves like a dot, except that it is not affected by
1115 the PCRE_DOTALL option. In other words, it matches any character except one
1116 that signifies the end of a line. Perl also uses \eN to match characters by
1117 name; PCRE does not support this.
1118 .
1119 .
1121 .rs
1122 .sp
1123 Outside a character class, the escape sequence \eC matches any one data unit,
1124 whether or not a UTF mode is set. In the 8-bit library, one data unit is one
1125 byte; in the 16-bit library it is a 16-bit unit; in the 32-bit library it is
1126 a 32-bit unit. Unlike a dot, \eC always
1127 matches line-ending characters. The feature is provided in Perl in order to
1128 match individual bytes in UTF-8 mode, but it is unclear how it can usefully be
1129 used. Because \eC breaks up characters into individual data units, matching one
1130 unit with \eC in a UTF mode means that the rest of the string may start with a
1131 malformed UTF character. This has undefined results, because PCRE assumes that
1132 it is dealing with valid UTF strings (and by default it checks this at the
1133 start of processing unless the PCRE_NO_UTF8_CHECK, PCRE_NO_UTF16_CHECK or
1134 PCRE_NO_UTF32_CHECK option is used).
1135 .P
1136 PCRE does not allow \eC to appear in lookbehind assertions
1137 .\" HTML <a href="#lookbehind">
1138 .\" </a>
1139 (described below)
1140 .\"
1141 in a UTF mode, because this would make it impossible to calculate the length of
1142 the lookbehind.
1143 .P
1144 In general, the \eC escape sequence is best avoided. However, one
1145 way of using it that avoids the problem of malformed UTF characters is to use a
1146 lookahead to check the length of the next character, as in this pattern, which
1147 could be used with a UTF-8 string (ignore white space and line breaks):
1148 .sp
1149 (?| (?=[\ex00-\ex7f])(\eC) |
1150 (?=[\ex80-\ex{7ff}])(\eC)(\eC) |
1151 (?=[\ex{800}-\ex{ffff}])(\eC)(\eC)(\eC) |
1152 (?=[\ex{10000}-\ex{1fffff}])(\eC)(\eC)(\eC)(\eC))
1153 .sp
1154 A group that starts with (?| resets the capturing parentheses numbers in each
1155 alternative (see
1156 .\" HTML <a href="#dupsubpatternnumber">
1157 .\" </a>
1158 "Duplicate Subpattern Numbers"
1159 .\"
1160 below). The assertions at the start of each branch check the next UTF-8
1161 character for values whose encoding uses 1, 2, 3, or 4 bytes, respectively. The
1162 character's individual bytes are then captured by the appropriate number of
1163 groups.
1164 .
1165 .
1166 .\" HTML <a name="characterclass"></a>
1168 .rs
1169 .sp
1170 An opening square bracket introduces a character class, terminated by a closing
1171 square bracket. A closing square bracket on its own is not special by default.
1172 However, if the PCRE_JAVASCRIPT_COMPAT option is set, a lone closing square
1173 bracket causes a compile-time error. If a closing square bracket is required as
1174 a member of the class, it should be the first data character in the class
1175 (after an initial circumflex, if present) or escaped with a backslash.
1176 .P
1177 A character class matches a single character in the subject. In a UTF mode, the
1178 character may be more than one data unit long. A matched character must be in
1179 the set of characters defined by the class, unless the first character in the
1180 class definition is a circumflex, in which case the subject character must not
1181 be in the set defined by the class. If a circumflex is actually required as a
1182 member of the class, ensure it is not the first character, or escape it with a
1183 backslash.
1184 .P
1185 For example, the character class [aeiou] matches any lower case vowel, while
1186 [^aeiou] matches any character that is not a lower case vowel. Note that a
1187 circumflex is just a convenient notation for specifying the characters that
1188 are in the class by enumerating those that are not. A class that starts with a
1189 circumflex is not an assertion; it still consumes a character from the subject
1190 string, and therefore it fails if the current pointer is at the end of the
1191 string.
1192 .P
1193 In UTF-8 (UTF-16, UTF-32) mode, characters with values greater than 255 (0xffff)
1194 can be included in a class as a literal string of data units, or by using the
1195 \ex{ escaping mechanism.
1196 .P
1197 When caseless matching is set, any letters in a class represent both their
1198 upper case and lower case versions, so for example, a caseless [aeiou] matches
1199 "A" as well as "a", and a caseless [^aeiou] does not match "A", whereas a
1200 caseful version would. In a UTF mode, PCRE always understands the concept of
1201 case for characters whose values are less than 128, so caseless matching is
1202 always possible. For characters with higher values, the concept of case is
1203 supported if PCRE is compiled with Unicode property support, but not otherwise.
1204 If you want to use caseless matching in a UTF mode for characters 128 and
1205 above, you must ensure that PCRE is compiled with Unicode property support as
1206 well as with UTF support.
1207 .P
1208 Characters that might indicate line breaks are never treated in any special way
1209 when matching character classes, whatever line-ending sequence is in use, and
1210 whatever setting of the PCRE_DOTALL and PCRE_MULTILINE options is used. A class
1211 such as [^a] always matches one of these characters.
1212 .P
1213 The minus (hyphen) character can be used to specify a range of characters in a
1214 character class. For example, [d-m] matches any letter between d and m,
1215 inclusive. If a minus character is required in a class, it must be escaped with
1216 a backslash or appear in a position where it cannot be interpreted as
1217 indicating a range, typically as the first or last character in the class.
1218 .P
1219 It is not possible to have the literal character "]" as the end character of a
1220 range. A pattern such as [W-]46] is interpreted as a class of two characters
1221 ("W" and "-") followed by a literal string "46]", so it would match "W46]" or
1222 "-46]". However, if the "]" is escaped with a backslash it is interpreted as
1223 the end of range, so [W-\e]46] is interpreted as a class containing a range
1224 followed by two other characters. The octal or hexadecimal representation of
1225 "]" can also be used to end a range.
1226 .P
1227 Ranges operate in the collating sequence of character values. They can also be
1228 used for characters specified numerically, for example [\e000-\e037]. Ranges
1229 can include any characters that are valid for the current mode.
1230 .P
1231 If a range that includes letters is used when caseless matching is set, it
1232 matches the letters in either case. For example, [W-c] is equivalent to
1233 [][\e\e^_`wxyzabc], matched caselessly, and in a non-UTF mode, if character
1234 tables for a French locale are in use, [\exc8-\excb] matches accented E
1235 characters in both cases. In UTF modes, PCRE supports the concept of case for
1236 characters with values greater than 128 only when it is compiled with Unicode
1237 property support.
1238 .P
1239 The character escape sequences \ed, \eD, \eh, \eH, \ep, \eP, \es, \eS, \ev,
1240 \eV, \ew, and \eW may appear in a character class, and add the characters that
1241 they match to the class. For example, [\edABCDEF] matches any hexadecimal
1242 digit. In UTF modes, the PCRE_UCP option affects the meanings of \ed, \es, \ew
1243 and their upper case partners, just as it does when they appear outside a
1244 character class, as described in the section entitled
1245 .\" HTML <a href="#genericchartypes">
1246 .\" </a>
1247 "Generic character types"
1248 .\"
1249 above. The escape sequence \eb has a different meaning inside a character
1250 class; it matches the backspace character. The sequences \eB, \eN, \eR, and \eX
1251 are not special inside a character class. Like any other unrecognized escape
1252 sequences, they are treated as the literal characters "B", "N", "R", and "X" by
1253 default, but cause an error if the PCRE_EXTRA option is set.
1254 .P
1255 A circumflex can conveniently be used with the upper case character types to
1256 specify a more restricted set of characters than the matching lower case type.
1257 For example, the class [^\eW_] matches any letter or digit, but not underscore,
1258 whereas [\ew] includes underscore. A positive character class should be read as
1259 "something OR something OR ..." and a negative class as "NOT something AND NOT
1260 something AND NOT ...".
1261 .P
1262 The only metacharacters that are recognized in character classes are backslash,
1263 hyphen (only where it can be interpreted as specifying a range), circumflex
1264 (only at the start), opening square bracket (only when it can be interpreted as
1265 introducing a POSIX class name - see the next section), and the terminating
1266 closing square bracket. However, escaping other non-alphanumeric characters
1267 does no harm.
1268 .
1269 .
1271 .rs
1272 .sp
1273 Perl supports the POSIX notation for character classes. This uses names
1274 enclosed by [: and :] within the enclosing square brackets. PCRE also supports
1275 this notation. For example,
1276 .sp
1277 [01[:alpha:]%]
1278 .sp
1279 matches "0", "1", any alphabetic character, or "%". The supported class names
1280 are:
1281 .sp
1282 alnum letters and digits
1283 alpha letters
1284 ascii character codes 0 - 127
1285 blank space or tab only
1286 cntrl control characters
1287 digit decimal digits (same as \ed)
1288 graph printing characters, excluding space
1289 lower lower case letters
1290 print printing characters, including space
1291 punct printing characters, excluding letters and digits and space
1292 space white space (not quite the same as \es)
1293 upper upper case letters
1294 word "word" characters (same as \ew)
1295 xdigit hexadecimal digits
1296 .sp
1297 The "space" characters are HT (9), LF (10), VT (11), FF (12), CR (13), and
1298 space (32). "Space" used to be different to \es, which did not include VT, for
1299 Perl compatibility. However, Perl changed at release 5.18, and PCRE followed at
1300 release 8.34. "Space" and \es now match the same set of characters.
1301 .P
1302 The name "word" is a Perl extension, and "blank" is a GNU extension from Perl
1303 5.8. Another Perl extension is negation, which is indicated by a ^ character
1304 after the colon. For example,
1305 .sp
1306 [12[:^digit:]]
1307 .sp
1308 matches "1", "2", or any non-digit. PCRE (and Perl) also recognize the POSIX
1309 syntax [.ch.] and [=ch=] where "ch" is a "collating element", but these are not
1310 supported, and an error is given if they are encountered.
1311 .P
1312 By default, in UTF modes, characters with values greater than 128 do not match
1313 any of the POSIX character classes. However, if the PCRE_UCP option is passed
1314 to \fBpcre_compile()\fP, some of the classes are changed so that Unicode
1315 character properties are used. This is achieved by replacing the POSIX classes
1316 by other sequences, as follows:
1317 .sp
1318 [:alnum:] becomes \ep{Xan}
1319 [:alpha:] becomes \ep{L}
1320 [:blank:] becomes \eh
1321 [:digit:] becomes \ep{Nd}
1322 [:lower:] becomes \ep{Ll}
1323 [:space:] becomes \ep{Xps}
1324 [:upper:] becomes \ep{Lu}
1325 [:word:] becomes \ep{Xwd}
1326 .sp
1327 Negated versions, such as [:^alpha:] use \eP instead of \ep. The other POSIX
1328 classes are unchanged, and match only characters with code points less than
1329 128.
1330 .
1331 .
1333 .rs
1334 .sp
1335 Vertical bar characters are used to separate alternative patterns. For example,
1336 the pattern
1337 .sp
1338 gilbert|sullivan
1339 .sp
1340 matches either "gilbert" or "sullivan". Any number of alternatives may appear,
1341 and an empty alternative is permitted (matching the empty string). The matching
1342 process tries each alternative in turn, from left to right, and the first one
1343 that succeeds is used. If the alternatives are within a subpattern
1344 .\" HTML <a href="#subpattern">
1345 .\" </a>
1346 (defined below),
1347 .\"
1348 "succeeds" means matching the rest of the main pattern as well as the
1349 alternative in the subpattern.
1350 .
1351 .
1353 .rs
1354 .sp
1355 The settings of the PCRE_CASELESS, PCRE_MULTILINE, PCRE_DOTALL, and
1356 PCRE_EXTENDED options (which are Perl-compatible) can be changed from within
1357 the pattern by a sequence of Perl option letters enclosed between "(?" and ")".
1358 The option letters are
1359 .sp
1360 i for PCRE_CASELESS
1362 s for PCRE_DOTALL
1363 x for PCRE_EXTENDED
1364 .sp
1365 For example, (?im) sets caseless, multiline matching. It is also possible to
1366 unset these options by preceding the letter with a hyphen, and a combined
1367 setting and unsetting such as (?im-sx), which sets PCRE_CASELESS and
1368 PCRE_MULTILINE while unsetting PCRE_DOTALL and PCRE_EXTENDED, is also
1369 permitted. If a letter appears both before and after the hyphen, the option is
1370 unset.
1371 .P
1372 The PCRE-specific options PCRE_DUPNAMES, PCRE_UNGREEDY, and PCRE_EXTRA can be
1373 changed in the same way as the Perl-compatible options by using the characters
1374 J, U and X respectively.
1375 .P
1376 When one of these option changes occurs at top level (that is, not inside
1377 subpattern parentheses), the change applies to the remainder of the pattern
1378 that follows. If the change is placed right at the start of a pattern, PCRE
1379 extracts it into the global options (and it will therefore show up in data
1380 extracted by the \fBpcre_fullinfo()\fP function).
1381 .P
1382 An option change within a subpattern (see below for a description of
1383 subpatterns) affects only that part of the subpattern that follows it, so
1384 .sp
1385 (a(?i)b)c
1386 .sp
1387 matches abc and aBc and no other strings (assuming PCRE_CASELESS is not used).
1388 By this means, options can be made to have different settings in different
1389 parts of the pattern. Any changes made in one alternative do carry on
1390 into subsequent branches within the same subpattern. For example,
1391 .sp
1392 (a(?i)b|c)
1393 .sp
1394 matches "ab", "aB", "c", and "C", even though when matching "C" the first
1395 branch is abandoned before the option setting. This is because the effects of
1396 option settings happen at compile time. There would be some very weird
1397 behaviour otherwise.
1398 .P
1399 \fBNote:\fP There are other PCRE-specific options that can be set by the
1400 application when the compiling or matching functions are called. In some cases
1401 the pattern can contain special leading sequences such as (*CRLF) to override
1402 what the application has set or what has been defaulted. Details are given in
1403 the section entitled
1404 .\" HTML <a href="#newlineseq">
1405 .\" </a>
1406 "Newline sequences"
1407 .\"
1408 above. There are also the (*UTF8), (*UTF16),(*UTF32), and (*UCP) leading
1409 sequences that can be used to set UTF and Unicode property modes; they are
1410 equivalent to setting the PCRE_UTF8, PCRE_UTF16, PCRE_UTF32 and the PCRE_UCP
1411 options, respectively. The (*UTF) sequence is a generic version that can be
1412 used with any of the libraries. However, the application can set the
1413 PCRE_NEVER_UTF option, which locks out the use of the (*UTF) sequences.
1414 .
1415 .
1416 .\" HTML <a name="subpattern"></a>
1418 .rs
1419 .sp
1420 Subpatterns are delimited by parentheses (round brackets), which can be nested.
1421 Turning part of a pattern into a subpattern does two things:
1422 .sp
1423 1. It localizes a set of alternatives. For example, the pattern
1424 .sp
1425 cat(aract|erpillar|)
1426 .sp
1427 matches "cataract", "caterpillar", or "cat". Without the parentheses, it would
1428 match "cataract", "erpillar" or an empty string.
1429 .sp
1430 2. It sets up the subpattern as a capturing subpattern. This means that, when
1431 the whole pattern matches, that portion of the subject string that matched the
1432 subpattern is passed back to the caller via the \fIovector\fP argument of the
1433 matching function. (This applies only to the traditional matching functions;
1434 the DFA matching functions do not support capturing.)
1435 .P
1436 Opening parentheses are counted from left to right (starting from 1) to obtain
1437 numbers for the capturing subpatterns. For example, if the string "the red
1438 king" is matched against the pattern
1439 .sp
1440 the ((red|white) (king|queen))
1441 .sp
1442 the captured substrings are "red king", "red", and "king", and are numbered 1,
1443 2, and 3, respectively.
1444 .P
1445 The fact that plain parentheses fulfil two functions is not always helpful.
1446 There are often times when a grouping subpattern is required without a
1447 capturing requirement. If an opening parenthesis is followed by a question mark
1448 and a colon, the subpattern does not do any capturing, and is not counted when
1449 computing the number of any subsequent capturing subpatterns. For example, if
1450 the string "the white queen" is matched against the pattern
1451 .sp
1452 the ((?:red|white) (king|queen))
1453 .sp
1454 the captured substrings are "white queen" and "queen", and are numbered 1 and
1455 2. The maximum number of capturing subpatterns is 65535.
1456 .P
1457 As a convenient shorthand, if any option settings are required at the start of
1458 a non-capturing subpattern, the option letters may appear between the "?" and
1459 the ":". Thus the two patterns
1460 .sp
1461 (?i:saturday|sunday)
1462 (?:(?i)saturday|sunday)
1463 .sp
1464 match exactly the same set of strings. Because alternative branches are tried
1465 from left to right, and options are not reset until the end of the subpattern
1466 is reached, an option setting in one branch does affect subsequent branches, so
1467 the above patterns match "SUNDAY" as well as "Saturday".
1468 .
1469 .
1470 .\" HTML <a name="dupsubpatternnumber"></a>
1472 .rs
1473 .sp
1474 Perl 5.10 introduced a feature whereby each alternative in a subpattern uses
1475 the same numbers for its capturing parentheses. Such a subpattern starts with
1476 (?| and is itself a non-capturing subpattern. For example, consider this
1477 pattern:
1478 .sp
1479 (?|(Sat)ur|(Sun))day
1480 .sp
1481 Because the two alternatives are inside a (?| group, both sets of capturing
1482 parentheses are numbered one. Thus, when the pattern matches, you can look
1483 at captured substring number one, whichever alternative matched. This construct
1484 is useful when you want to capture part, but not all, of one of a number of
1485 alternatives. Inside a (?| group, parentheses are numbered as usual, but the
1486 number is reset at the start of each branch. The numbers of any capturing
1487 parentheses that follow the subpattern start after the highest number used in
1488 any branch. The following example is taken from the Perl documentation. The
1489 numbers underneath show in which buffer the captured content will be stored.
1490 .sp
1491 # before ---------------branch-reset----------- after
1492 / ( a ) (?| x ( y ) z | (p (q) r) | (t) u (v) ) ( z ) /x
1493 # 1 2 2 3 2 3 4
1494 .sp
1495 A back reference to a numbered subpattern uses the most recent value that is
1496 set for that number by any subpattern. The following pattern matches "abcabc"
1497 or "defdef":
1498 .sp
1499 /(?|(abc)|(def))\e1/
1500 .sp
1501 In contrast, a subroutine call to a numbered subpattern always refers to the
1502 first one in the pattern with the given number. The following pattern matches
1503 "abcabc" or "defabc":
1504 .sp
1505 /(?|(abc)|(def))(?1)/
1506 .sp
1507 If a
1508 .\" HTML <a href="#conditions">
1509 .\" </a>
1510 condition test
1511 .\"
1512 for a subpattern's having matched refers to a non-unique number, the test is
1513 true if any of the subpatterns of that number have matched.
1514 .P
1515 An alternative approach to using this "branch reset" feature is to use
1516 duplicate named subpatterns, as described in the next section.
1517 .
1518 .
1520 .rs
1521 .sp
1522 Identifying capturing parentheses by number is simple, but it can be very hard
1523 to keep track of the numbers in complicated regular expressions. Furthermore,
1524 if an expression is modified, the numbers may change. To help with this
1525 difficulty, PCRE supports the naming of subpatterns. This feature was not
1526 added to Perl until release 5.10. Python had the feature earlier, and PCRE
1527 introduced it at release 4.0, using the Python syntax. PCRE now supports both
1528 the Perl and the Python syntax. Perl allows identically numbered subpatterns to
1529 have different names, but PCRE does not.
1530 .P
1531 In PCRE, a subpattern can be named in one of three ways: (?<name>...) or
1532 (?'name'...) as in Perl, or (?P<name>...) as in Python. References to capturing
1533 parentheses from other parts of the pattern, such as
1534 .\" HTML <a href="#backreferences">
1535 .\" </a>
1536 back references,
1537 .\"
1538 .\" HTML <a href="#recursion">
1539 .\" </a>
1540 recursion,
1541 .\"
1542 and
1543 .\" HTML <a href="#conditions">
1544 .\" </a>
1545 conditions,
1546 .\"
1547 can be made by name as well as by number.
1548 .P
1549 Names consist of up to 32 alphanumeric characters and underscores. Named
1550 capturing parentheses are still allocated numbers as well as names, exactly as
1551 if the names were not present. The PCRE API provides function calls for
1552 extracting the name-to-number translation table from a compiled pattern. There
1553 is also a convenience function for extracting a captured substring by name.
1554 .P
1555 By default, a name must be unique within a pattern, but it is possible to relax
1556 this constraint by setting the PCRE_DUPNAMES option at compile time. (Duplicate
1557 names are also always permitted for subpatterns with the same number, set up as
1558 described in the previous section.) Duplicate names can be useful for patterns
1559 where only one instance of the named parentheses can match. Suppose you want to
1560 match the name of a weekday, either as a 3-letter abbreviation or as the full
1561 name, and in both cases you want to extract the abbreviation. This pattern
1562 (ignoring the line breaks) does the job:
1563 .sp
1564 (?<DN>Mon|Fri|Sun)(?:day)?|
1565 (?<DN>Tue)(?:sday)?|
1566 (?<DN>Wed)(?:nesday)?|
1567 (?<DN>Thu)(?:rsday)?|
1568 (?<DN>Sat)(?:urday)?
1569 .sp
1570 There are five capturing substrings, but only one is ever set after a match.
1571 (An alternative way of solving this problem is to use a "branch reset"
1572 subpattern, as described in the previous section.)
1573 .P
1574 The convenience function for extracting the data by name returns the substring
1575 for the first (and in this example, the only) subpattern of that name that
1576 matched. This saves searching to find which numbered subpattern it was.
1577 .P
1578 If you make a back reference to a non-unique named subpattern from elsewhere in
1579 the pattern, the subpatterns to which the name refers are checked in the order
1580 in which they appear in the overall pattern. The first one that is set is used
1581 for the reference. For example, this pattern matches both "foofoo" and
1582 "barbar" but not "foobar" or "barfoo":
1583 .sp
1584 (?:(?<n>foo)|(?<n>bar))\k<n>
1585 .sp
1586 .P
1587 If you make a subroutine call to a non-unique named subpattern, the one that
1588 corresponds to the first occurrence of the name is used. In the absence of
1589 duplicate numbers (see the previous section) this is the one with the lowest
1590 number.
1591 .P
1592 If you use a named reference in a condition
1593 test (see the
1594 .\"
1595 .\" HTML <a href="#conditions">
1596 .\" </a>
1597 section about conditions
1598 .\"
1599 below), either to check whether a subpattern has matched, or to check for
1600 recursion, all subpatterns with the same name are tested. If the condition is
1601 true for any one of them, the overall condition is true. This is the same
1602 behaviour as testing by number. For further details of the interfaces for
1603 handling named subpatterns, see the
1604 .\" HREF
1605 \fBpcreapi\fP
1606 .\"
1607 documentation.
1608 .P
1609 \fBWarning:\fP You cannot use different names to distinguish between two
1610 subpatterns with the same number because PCRE uses only the numbers when
1611 matching. For this reason, an error is given at compile time if different names
1612 are given to subpatterns with the same number. However, you can always give the
1613 same name to subpatterns with the same number, even when PCRE_DUPNAMES is not
1614 set.
1615 .
1616 .
1618 .rs
1619 .sp
1620 Repetition is specified by quantifiers, which can follow any of the following
1621 items:
1622 .sp
1623 a literal data character
1624 the dot metacharacter
1625 the \eC escape sequence
1626 the \eX escape sequence
1627 the \eR escape sequence
1628 an escape such as \ed or \epL that matches a single character
1629 a character class
1630 a back reference (see next section)
1631 a parenthesized subpattern (including assertions)
1632 a subroutine call to a subpattern (recursive or otherwise)
1633 .sp
1634 The general repetition quantifier specifies a minimum and maximum number of
1635 permitted matches, by giving the two numbers in curly brackets (braces),
1636 separated by a comma. The numbers must be less than 65536, and the first must
1637 be less than or equal to the second. For example:
1638 .sp
1639 z{2,4}
1640 .sp
1641 matches "zz", "zzz", or "zzzz". A closing brace on its own is not a special
1642 character. If the second number is omitted, but the comma is present, there is
1643 no upper limit; if the second number and the comma are both omitted, the
1644 quantifier specifies an exact number of required matches. Thus
1645 .sp
1646 [aeiou]{3,}
1647 .sp
1648 matches at least 3 successive vowels, but may match many more, while
1649 .sp
1650 \ed{8}
1651 .sp
1652 matches exactly 8 digits. An opening curly bracket that appears in a position
1653 where a quantifier is not allowed, or one that does not match the syntax of a
1654 quantifier, is taken as a literal character. For example, {,6} is not a
1655 quantifier, but a literal string of four characters.
1656 .P
1657 In UTF modes, quantifiers apply to characters rather than to individual data
1658 units. Thus, for example, \ex{100}{2} matches two characters, each of
1659 which is represented by a two-byte sequence in a UTF-8 string. Similarly,
1660 \eX{3} matches three Unicode extended grapheme clusters, each of which may be
1661 several data units long (and they may be of different lengths).
1662 .P
1663 The quantifier {0} is permitted, causing the expression to behave as if the
1664 previous item and the quantifier were not present. This may be useful for
1665 subpatterns that are referenced as
1666 .\" HTML <a href="#subpatternsassubroutines">
1667 .\" </a>
1668 subroutines
1669 .\"
1670 from elsewhere in the pattern (but see also the section entitled
1671 .\" HTML <a href="#subdefine">
1672 .\" </a>
1673 "Defining subpatterns for use by reference only"
1674 .\"
1675 below). Items other than subpatterns that have a {0} quantifier are omitted
1676 from the compiled pattern.
1677 .P
1678 For convenience, the three most common quantifiers have single-character
1679 abbreviations:
1680 .sp
1681 * is equivalent to {0,}
1682 + is equivalent to {1,}
1683 ? is equivalent to {0,1}
1684 .sp
1685 It is possible to construct infinite loops by following a subpattern that can
1686 match no characters with a quantifier that has no upper limit, for example:
1687 .sp
1688 (a?)*
1689 .sp
1690 Earlier versions of Perl and PCRE used to give an error at compile time for
1691 such patterns. However, because there are cases where this can be useful, such
1692 patterns are now accepted, but if any repetition of the subpattern does in fact
1693 match no characters, the loop is forcibly broken.
1694 .P
1695 By default, the quantifiers are "greedy", that is, they match as much as
1696 possible (up to the maximum number of permitted times), without causing the
1697 rest of the pattern to fail. The classic example of where this gives problems
1698 is in trying to match comments in C programs. These appear between /* and */
1699 and within the comment, individual * and / characters may appear. An attempt to
1700 match C comments by applying the pattern
1701 .sp
1702 /\e*.*\e*/
1703 .sp
1704 to the string
1705 .sp
1706 /* first comment */ not comment /* second comment */
1707 .sp
1708 fails, because it matches the entire string owing to the greediness of the .*
1709 item.
1710 .P
1711 However, if a quantifier is followed by a question mark, it ceases to be
1712 greedy, and instead matches the minimum number of times possible, so the
1713 pattern
1714 .sp
1715 /\e*.*?\e*/
1716 .sp
1717 does the right thing with the C comments. The meaning of the various
1718 quantifiers is not otherwise changed, just the preferred number of matches.
1719 Do not confuse this use of question mark with its use as a quantifier in its
1720 own right. Because it has two uses, it can sometimes appear doubled, as in
1721 .sp
1722 \ed??\ed
1723 .sp
1724 which matches one digit by preference, but can match two if that is the only
1725 way the rest of the pattern matches.
1726 .P
1727 If the PCRE_UNGREEDY option is set (an option that is not available in Perl),
1728 the quantifiers are not greedy by default, but individual ones can be made
1729 greedy by following them with a question mark. In other words, it inverts the
1730 default behaviour.
1731 .P
1732 When a parenthesized subpattern is quantified with a minimum repeat count that
1733 is greater than 1 or with a limited maximum, more memory is required for the
1734 compiled pattern, in proportion to the size of the minimum or maximum.
1735 .P
1736 If a pattern starts with .* or .{0,} and the PCRE_DOTALL option (equivalent
1737 to Perl's /s) is set, thus allowing the dot to match newlines, the pattern is
1738 implicitly anchored, because whatever follows will be tried against every
1739 character position in the subject string, so there is no point in retrying the
1740 overall match at any position after the first. PCRE normally treats such a
1741 pattern as though it were preceded by \eA.
1742 .P
1743 In cases where it is known that the subject string contains no newlines, it is
1744 worth setting PCRE_DOTALL in order to obtain this optimization, or
1745 alternatively using ^ to indicate anchoring explicitly.
1746 .P
1747 However, there are some cases where the optimization cannot be used. When .*
1748 is inside capturing parentheses that are the subject of a back reference
1749 elsewhere in the pattern, a match at the start may fail where a later one
1750 succeeds. Consider, for example:
1751 .sp
1752 (.*)abc\e1
1753 .sp
1754 If the subject is "xyz123abc123" the match point is the fourth character. For
1755 this reason, such a pattern is not implicitly anchored.
1756 .P
1757 Another case where implicit anchoring is not applied is when the leading .* is
1758 inside an atomic group. Once again, a match at the start may fail where a later
1759 one succeeds. Consider this pattern:
1760 .sp
1761 (?>.*?a)b
1762 .sp
1763 It matches "ab" in the subject "aab". The use of the backtracking control verbs
1764 (*PRUNE) and (*SKIP) also disable this optimization.
1765 .P
1766 When a capturing subpattern is repeated, the value captured is the substring
1767 that matched the final iteration. For example, after
1768 .sp
1769 (tweedle[dume]{3}\es*)+
1770 .sp
1771 has matched "tweedledum tweedledee" the value of the captured substring is
1772 "tweedledee". However, if there are nested capturing subpatterns, the
1773 corresponding captured values may have been set in previous iterations. For
1774 example, after
1775 .sp
1776 /(a|(b))+/
1777 .sp
1778 matches "aba" the value of the second captured substring is "b".
1779 .
1780 .
1781 .\" HTML <a name="atomicgroup"></a>
1783 .rs
1784 .sp
1785 With both maximizing ("greedy") and minimizing ("ungreedy" or "lazy")
1786 repetition, failure of what follows normally causes the repeated item to be
1787 re-evaluated to see if a different number of repeats allows the rest of the
1788 pattern to match. Sometimes it is useful to prevent this, either to change the
1789 nature of the match, or to cause it fail earlier than it otherwise might, when
1790 the author of the pattern knows there is no point in carrying on.
1791 .P
1792 Consider, for example, the pattern \ed+foo when applied to the subject line
1793 .sp
1794 123456bar
1795 .sp
1796 After matching all 6 digits and then failing to match "foo", the normal
1797 action of the matcher is to try again with only 5 digits matching the \ed+
1798 item, and then with 4, and so on, before ultimately failing. "Atomic grouping"
1799 (a term taken from Jeffrey Friedl's book) provides the means for specifying
1800 that once a subpattern has matched, it is not to be re-evaluated in this way.
1801 .P
1802 If we use atomic grouping for the previous example, the matcher gives up
1803 immediately on failing to match "foo" the first time. The notation is a kind of
1804 special parenthesis, starting with (?> as in this example:
1805 .sp
1806 (?>\ed+)foo
1807 .sp
1808 This kind of parenthesis "locks up" the part of the pattern it contains once
1809 it has matched, and a failure further into the pattern is prevented from
1810 backtracking into it. Backtracking past it to previous items, however, works as
1811 normal.
1812 .P
1813 An alternative description is that a subpattern of this type matches the string
1814 of characters that an identical standalone pattern would match, if anchored at
1815 the current point in the subject string.
1816 .P
1817 Atomic grouping subpatterns are not capturing subpatterns. Simple cases such as
1818 the above example can be thought of as a maximizing repeat that must swallow
1819 everything it can. So, while both \ed+ and \ed+? are prepared to adjust the
1820 number of digits they match in order to make the rest of the pattern match,
1821 (?>\ed+) can only match an entire sequence of digits.
1822 .P
1823 Atomic groups in general can of course contain arbitrarily complicated
1824 subpatterns, and can be nested. However, when the subpattern for an atomic
1825 group is just a single repeated item, as in the example above, a simpler
1826 notation, called a "possessive quantifier" can be used. This consists of an
1827 additional + character following a quantifier. Using this notation, the
1828 previous example can be rewritten as
1829 .sp
1830 \ed++foo
1831 .sp
1832 Note that a possessive quantifier can be used with an entire group, for
1833 example:
1834 .sp
1835 (abc|xyz){2,3}+
1836 .sp
1837 Possessive quantifiers are always greedy; the setting of the PCRE_UNGREEDY
1838 option is ignored. They are a convenient notation for the simpler forms of
1839 atomic group. However, there is no difference in the meaning of a possessive
1840 quantifier and the equivalent atomic group, though there may be a performance
1841 difference; possessive quantifiers should be slightly faster.
1842 .P
1843 The possessive quantifier syntax is an extension to the Perl 5.8 syntax.
1844 Jeffrey Friedl originated the idea (and the name) in the first edition of his
1845 book. Mike McCloskey liked it, so implemented it when he built Sun's Java
1846 package, and PCRE copied it from there. It ultimately found its way into Perl
1847 at release 5.10.
1848 .P
1849 PCRE has an optimization that automatically "possessifies" certain simple
1850 pattern constructs. For example, the sequence A+B is treated as A++B because
1851 there is no point in backtracking into a sequence of A's when B must follow.
1852 .P
1853 When a pattern contains an unlimited repeat inside a subpattern that can itself
1854 be repeated an unlimited number of times, the use of an atomic group is the
1855 only way to avoid some failing matches taking a very long time indeed. The
1856 pattern
1857 .sp
1858 (\eD+|<\ed+>)*[!?]
1859 .sp
1860 matches an unlimited number of substrings that either consist of non-digits, or
1861 digits enclosed in <>, followed by either ! or ?. When it matches, it runs
1862 quickly. However, if it is applied to
1863 .sp
1864 aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
1865 .sp
1866 it takes a long time before reporting failure. This is because the string can
1867 be divided between the internal \eD+ repeat and the external * repeat in a
1868 large number of ways, and all have to be tried. (The example uses [!?] rather
1869 than a single character at the end, because both PCRE and Perl have an
1870 optimization that allows for fast failure when a single character is used. They
1871 remember the last single character that is required for a match, and fail early
1872 if it is not present in the string.) If the pattern is changed so that it uses
1873 an atomic group, like this:
1874 .sp
1875 ((?>\eD+)|<\ed+>)*[!?]
1876 .sp
1877 sequences of non-digits cannot be broken, and failure happens quickly.
1878 .
1879 .
1880 .\" HTML <a name="backreferences"></a>
1882 .rs
1883 .sp
1884 Outside a character class, a backslash followed by a digit greater than 0 (and
1885 possibly further digits) is a back reference to a capturing subpattern earlier
1886 (that is, to its left) in the pattern, provided there have been that many
1887 previous capturing left parentheses.
1888 .P
1889 However, if the decimal number following the backslash is less than 10, it is
1890 always taken as a back reference, and causes an error only if there are not
1891 that many capturing left parentheses in the entire pattern. In other words, the
1892 parentheses that are referenced need not be to the left of the reference for
1893 numbers less than 10. A "forward back reference" of this type can make sense
1894 when a repetition is involved and the subpattern to the right has participated
1895 in an earlier iteration.
1896 .P
1897 It is not possible to have a numerical "forward back reference" to a subpattern
1898 whose number is 10 or more using this syntax because a sequence such as \e50 is
1899 interpreted as a character defined in octal. See the subsection entitled
1900 "Non-printing characters"
1901 .\" HTML <a href="#digitsafterbackslash">
1902 .\" </a>
1903 above
1904 .\"
1905 for further details of the handling of digits following a backslash. There is
1906 no such problem when named parentheses are used. A back reference to any
1907 subpattern is possible using named parentheses (see below).
1908 .P
1909 Another way of avoiding the ambiguity inherent in the use of digits following a
1910 backslash is to use the \eg escape sequence. This escape must be followed by an
1911 unsigned number or a negative number, optionally enclosed in braces. These
1912 examples are all identical:
1913 .sp
1914 (ring), \e1
1915 (ring), \eg1
1916 (ring), \eg{1}
1917 .sp
1918 An unsigned number specifies an absolute reference without the ambiguity that
1919 is present in the older syntax. It is also useful when literal digits follow
1920 the reference. A negative number is a relative reference. Consider this
1921 example:
1922 .sp
1923 (abc(def)ghi)\eg{-1}
1924 .sp
1925 The sequence \eg{-1} is a reference to the most recently started capturing
1926 subpattern before \eg, that is, is it equivalent to \e2 in this example.
1927 Similarly, \eg{-2} would be equivalent to \e1. The use of relative references
1928 can be helpful in long patterns, and also in patterns that are created by
1929 joining together fragments that contain references within themselves.
1930 .P
1931 A back reference matches whatever actually matched the capturing subpattern in
1932 the current subject string, rather than anything matching the subpattern
1933 itself (see
1934 .\" HTML <a href="#subpatternsassubroutines">
1935 .\" </a>
1936 "Subpatterns as subroutines"
1937 .\"
1938 below for a way of doing that). So the pattern
1939 .sp
1940 (sens|respons)e and \e1ibility
1941 .sp
1942 matches "sense and sensibility" and "response and responsibility", but not
1943 "sense and responsibility". If caseful matching is in force at the time of the
1944 back reference, the case of letters is relevant. For example,
1945 .sp
1946 ((?i)rah)\es+\e1
1947 .sp
1948 matches "rah rah" and "RAH RAH", but not "RAH rah", even though the original
1949 capturing subpattern is matched caselessly.
1950 .P
1951 There are several different ways of writing back references to named
1952 subpatterns. The .NET syntax \ek{name} and the Perl syntax \ek<name> or
1953 \ek'name' are supported, as is the Python syntax (?P=name). Perl 5.10's unified
1954 back reference syntax, in which \eg can be used for both numeric and named
1955 references, is also supported. We could rewrite the above example in any of
1956 the following ways:
1957 .sp
1958 (?<p1>(?i)rah)\es+\ek<p1>
1959 (?'p1'(?i)rah)\es+\ek{p1}
1960 (?P<p1>(?i)rah)\es+(?P=p1)
1961 (?<p1>(?i)rah)\es+\eg{p1}
1962 .sp
1963 A subpattern that is referenced by name may appear in the pattern before or
1964 after the reference.
1965 .P
1966 There may be more than one back reference to the same subpattern. If a
1967 subpattern has not actually been used in a particular match, any back
1968 references to it always fail by default. For example, the pattern
1969 .sp
1970 (a|(bc))\e2
1971 .sp
1972 always fails if it starts to match "a" rather than "bc". However, if the
1973 PCRE_JAVASCRIPT_COMPAT option is set at compile time, a back reference to an
1974 unset value matches an empty string.
1975 .P
1976 Because there may be many capturing parentheses in a pattern, all digits
1977 following a backslash are taken as part of a potential back reference number.
1978 If the pattern continues with a digit character, some delimiter must be used to
1979 terminate the back reference. If the PCRE_EXTENDED option is set, this can be
1980 white space. Otherwise, the \eg{ syntax or an empty comment (see
1981 .\" HTML <a href="#comments">
1982 .\" </a>
1983 "Comments"
1984 .\"
1985 below) can be used.
1986 .
1987 .SS "Recursive back references"
1988 .rs
1989 .sp
1990 A back reference that occurs inside the parentheses to which it refers fails
1991 when the subpattern is first used, so, for example, (a\e1) never matches.
1992 However, such references can be useful inside repeated subpatterns. For
1993 example, the pattern
1994 .sp
1995 (a|b\e1)+
1996 .sp
1997 matches any number of "a"s and also "aba", "ababbaa" etc. At each iteration of
1998 the subpattern, the back reference matches the character string corresponding
1999 to the previous iteration. In order for this to work, the pattern must be such
2000 that the first iteration does not need to match the back reference. This can be
2001 done using alternation, as in the example above, or by a quantifier with a
2002 minimum of zero.
2003 .P
2004 Back references of this type cause the group that they reference to be treated
2005 as an
2006 .\" HTML <a href="#atomicgroup">
2007 .\" </a>
2008 atomic group.
2009 .\"
2010 Once the whole group has been matched, a subsequent matching failure cannot
2011 cause backtracking into the middle of the group.
2012 .
2013 .
2014 .\" HTML <a name="bigassertions"></a>
2016 .rs
2017 .sp
2018 An assertion is a test on the characters following or preceding the current
2019 matching point that does not actually consume any characters. The simple
2020 assertions coded as \eb, \eB, \eA, \eG, \eZ, \ez, ^ and $ are described
2021 .\" HTML <a href="#smallassertions">
2022 .\" </a>
2023 above.
2024 .\"
2025 .P
2026 More complicated assertions are coded as subpatterns. There are two kinds:
2027 those that look ahead of the current position in the subject string, and those
2028 that look behind it. An assertion subpattern is matched in the normal way,
2029 except that it does not cause the current matching position to be changed.
2030 .P
2031 Assertion subpatterns are not capturing subpatterns. If such an assertion
2032 contains capturing subpatterns within it, these are counted for the purposes of
2033 numbering the capturing subpatterns in the whole pattern. However, substring
2034 capturing is carried out only for positive assertions. (Perl sometimes, but not
2035 always, does do capturing in negative assertions.)
2036 .P
2037 For compatibility with Perl, assertion subpatterns may be repeated; though
2038 it makes no sense to assert the same thing several times, the side effect of
2039 capturing parentheses may occasionally be useful. In practice, there only three
2040 cases:
2041 .sp
2042 (1) If the quantifier is {0}, the assertion is never obeyed during matching.
2043 However, it may contain internal capturing parenthesized groups that are called
2044 from elsewhere via the
2045 .\" HTML <a href="#subpatternsassubroutines">
2046 .\" </a>
2047 subroutine mechanism.
2048 .\"
2049 .sp
2050 (2) If quantifier is {0,n} where n is greater than zero, it is treated as if it
2051 were {0,1}. At run time, the rest of the pattern match is tried with and
2052 without the assertion, the order depending on the greediness of the quantifier.
2053 .sp
2054 (3) If the minimum repetition is greater than zero, the quantifier is ignored.
2055 The assertion is obeyed just once when encountered during matching.
2056 .
2057 .
2058 .SS "Lookahead assertions"
2059 .rs
2060 .sp
2061 Lookahead assertions start with (?= for positive assertions and (?! for
2062 negative assertions. For example,
2063 .sp
2064 \ew+(?=;)
2065 .sp
2066 matches a word followed by a semicolon, but does not include the semicolon in
2067 the match, and
2068 .sp
2069 foo(?!bar)
2070 .sp
2071 matches any occurrence of "foo" that is not followed by "bar". Note that the
2072 apparently similar pattern
2073 .sp
2074 (?!foo)bar
2075 .sp
2076 does not find an occurrence of "bar" that is preceded by something other than
2077 "foo"; it finds any occurrence of "bar" whatsoever, because the assertion
2078 (?!foo) is always true when the next three characters are "bar". A
2079 lookbehind assertion is needed to achieve the other effect.
2080 .P
2081 If you want to force a matching failure at some point in a pattern, the most
2082 convenient way to do it is with (?!) because an empty string always matches, so
2083 an assertion that requires there not to be an empty string must always fail.
2084 The backtracking control verb (*FAIL) or (*F) is a synonym for (?!).
2085 .
2086 .
2087 .\" HTML <a name="lookbehind"></a>
2088 .SS "Lookbehind assertions"
2089 .rs
2090 .sp
2091 Lookbehind assertions start with (?<= for positive assertions and (?<! for
2092 negative assertions. For example,
2093 .sp
2094 (?<!foo)bar
2095 .sp
2096 does find an occurrence of "bar" that is not preceded by "foo". The contents of
2097 a lookbehind assertion are restricted such that all the strings it matches must
2098 have a fixed length. However, if there are several top-level alternatives, they
2099 do not all have to have the same fixed length. Thus
2100 .sp
2101 (?<=bullock|donkey)
2102 .sp
2103 is permitted, but
2104 .sp
2105 (?<!dogs?|cats?)
2106 .sp
2107 causes an error at compile time. Branches that match different length strings
2108 are permitted only at the top level of a lookbehind assertion. This is an
2109 extension compared with Perl, which requires all branches to match the same
2110 length of string. An assertion such as
2111 .sp
2112 (?<=ab(c|de))
2113 .sp
2114 is not permitted, because its single top-level branch can match two different
2115 lengths, but it is acceptable to PCRE if rewritten to use two top-level
2116 branches:
2117 .sp
2118 (?<=abc|abde)
2119 .sp
2120 In some cases, the escape sequence \eK
2121 .\" HTML <a href="#resetmatchstart">
2122 .\" </a>
2123 (see above)
2124 .\"
2125 can be used instead of a lookbehind assertion to get round the fixed-length
2126 restriction.
2127 .P
2128 The implementation of lookbehind assertions is, for each alternative, to
2129 temporarily move the current position back by the fixed length and then try to
2130 match. If there are insufficient characters before the current position, the
2131 assertion fails.
2132 .P
2133 In a UTF mode, PCRE does not allow the \eC escape (which matches a single data
2134 unit even in a UTF mode) to appear in lookbehind assertions, because it makes
2135 it impossible to calculate the length of the lookbehind. The \eX and \eR
2136 escapes, which can match different numbers of data units, are also not
2137 permitted.
2138 .P
2139 .\" HTML <a href="#subpatternsassubroutines">
2140 .\" </a>
2141 "Subroutine"
2142 .\"
2143 calls (see below) such as (?2) or (?&X) are permitted in lookbehinds, as long
2144 as the subpattern matches a fixed-length string.
2145 .\" HTML <a href="#recursion">
2146 .\" </a>
2147 Recursion,
2148 .\"
2149 however, is not supported.
2150 .P
2151 Possessive quantifiers can be used in conjunction with lookbehind assertions to
2152 specify efficient matching of fixed-length strings at the end of subject
2153 strings. Consider a simple pattern such as
2154 .sp
2155 abcd$
2156 .sp
2157 when applied to a long string that does not match. Because matching proceeds
2158 from left to right, PCRE will look for each "a" in the subject and then see if
2159 what follows matches the rest of the pattern. If the pattern is specified as
2160 .sp
2161 ^.*abcd$
2162 .sp
2163 the initial .* matches the entire string at first, but when this fails (because
2164 there is no following "a"), it backtracks to match all but the last character,
2165 then all but the last two characters, and so on. Once again the search for "a"
2166 covers the entire string, from right to left, so we are no better off. However,
2167 if the pattern is written as
2168 .sp
2169 ^.*+(?<=abcd)
2170 .sp
2171 there can be no backtracking for the .*+ item; it can match only the entire
2172 string. The subsequent lookbehind assertion does a single test on the last four
2173 characters. If it fails, the match fails immediately. For long strings, this
2174 approach makes a significant difference to the processing time.
2175 .
2176 .
2177 .SS "Using multiple assertions"
2178 .rs
2179 .sp
2180 Several assertions (of any sort) may occur in succession. For example,
2181 .sp
2182 (?<=\ed{3})(?<!999)foo
2183 .sp
2184 matches "foo" preceded by three digits that are not "999". Notice that each of
2185 the assertions is applied independently at the same point in the subject
2186 string. First there is a check that the previous three characters are all
2187 digits, and then there is a check that the same three characters are not "999".
2188 This pattern does \fInot\fP match "foo" preceded by six characters, the first
2189 of which are digits and the last three of which are not "999". For example, it
2190 doesn't match "123abcfoo". A pattern to do that is
2191 .sp
2192 (?<=\ed{3}...)(?<!999)foo
2193 .sp
2194 This time the first assertion looks at the preceding six characters, checking
2195 that the first three are digits, and then the second assertion checks that the
2196 preceding three characters are not "999".
2197 .P
2198 Assertions can be nested in any combination. For example,
2199 .sp
2200 (?<=(?<!foo)bar)baz
2201 .sp
2202 matches an occurrence of "baz" that is preceded by "bar" which in turn is not
2203 preceded by "foo", while
2204 .sp
2205 (?<=\ed{3}(?!999)...)foo
2206 .sp
2207 is another pattern that matches "foo" preceded by three digits and any three
2208 characters that are not "999".
2209 .
2210 .
2211 .\" HTML <a name="conditions"></a>
2213 .rs
2214 .sp
2215 It is possible to cause the matching process to obey a subpattern
2216 conditionally or to choose between two alternative subpatterns, depending on
2217 the result of an assertion, or whether a specific capturing subpattern has
2218 already been matched. The two possible forms of conditional subpattern are:
2219 .sp
2220 (?(condition)yes-pattern)
2221 (?(condition)yes-pattern|no-pattern)
2222 .sp
2223 If the condition is satisfied, the yes-pattern is used; otherwise the
2224 no-pattern (if present) is used. If there are more than two alternatives in the
2225 subpattern, a compile-time error occurs. Each of the two alternatives may
2226 itself contain nested subpatterns of any form, including conditional
2227 subpatterns; the restriction to two alternatives applies only at the level of
2228 the condition. This pattern fragment is an example where the alternatives are
2229 complex:
2230 .sp
2231 (?(1) (A|B|C) | (D | (?(2)E|F) | E) )
2232 .sp
2233 .P
2234 There are four kinds of condition: references to subpatterns, references to
2235 recursion, a pseudo-condition called DEFINE, and assertions.
2236 .
2237 .SS "Checking for a used subpattern by number"
2238 .rs
2239 .sp
2240 If the text between the parentheses consists of a sequence of digits, the
2241 condition is true if a capturing subpattern of that number has previously
2242 matched. If there is more than one capturing subpattern with the same number
2243 (see the earlier
2244 .\"
2245 .\" HTML <a href="#recursion">
2246 .\" </a>
2247 section about duplicate subpattern numbers),
2248 .\"
2249 the condition is true if any of them have matched. An alternative notation is
2250 to precede the digits with a plus or minus sign. In this case, the subpattern
2251 number is relative rather than absolute. The most recently opened parentheses
2252 can be referenced by (?(-1), the next most recent by (?(-2), and so on. Inside
2253 loops it can also make sense to refer to subsequent groups. The next
2254 parentheses to be opened can be referenced as (?(+1), and so on. (The value
2255 zero in any of these forms is not used; it provokes a compile-time error.)
2256 .P
2257 Consider the following pattern, which contains non-significant white space to
2258 make it more readable (assume the PCRE_EXTENDED option) and to divide it into
2259 three parts for ease of discussion:
2260 .sp
2261 ( \e( )? [^()]+ (?(1) \e) )
2262 .sp
2263 The first part matches an optional opening parenthesis, and if that
2264 character is present, sets it as the first captured substring. The second part
2265 matches one or more characters that are not parentheses. The third part is a
2266 conditional subpattern that tests whether or not the first set of parentheses
2267 matched. If they did, that is, if subject started with an opening parenthesis,
2268 the condition is true, and so the yes-pattern is executed and a closing
2269 parenthesis is required. Otherwise, since no-pattern is not present, the
2270 subpattern matches nothing. In other words, this pattern matches a sequence of
2271 non-parentheses, optionally enclosed in parentheses.
2272 .P
2273 If you were embedding this pattern in a larger one, you could use a relative
2274 reference:
2275 .sp
2276 ...other stuff... ( \e( )? [^()]+ (?(-1) \e) ) ...
2277 .sp
2278 This makes the fragment independent of the parentheses in the larger pattern.
2279 .
2280 .SS "Checking for a used subpattern by name"
2281 .rs
2282 .sp
2283 Perl uses the syntax (?(<name>)...) or (?('name')...) to test for a used
2284 subpattern by name. For compatibility with earlier versions of PCRE, which had
2285 this facility before Perl, the syntax (?(name)...) is also recognized. However,
2286 there is a possible ambiguity with this syntax, because subpattern names may
2287 consist entirely of digits. PCRE looks first for a named subpattern; if it
2288 cannot find one and the name consists entirely of digits, PCRE looks for a
2289 subpattern of that number, which must be greater than zero. Using subpattern
2290 names that consist entirely of digits is not recommended.
2291 .P
2292 Rewriting the above example to use a named subpattern gives this:
2293 .sp
2294 (?<OPEN> \e( )? [^()]+ (?(<OPEN>) \e) )
2295 .sp
2296 If the name used in a condition of this kind is a duplicate, the test is
2297 applied to all subpatterns of the same name, and is true if any one of them has
2298 matched.
2299 .
2300 .SS "Checking for pattern recursion"
2301 .rs
2302 .sp
2303 If the condition is the string (R), and there is no subpattern with the name R,
2304 the condition is true if a recursive call to the whole pattern or any
2305 subpattern has been made. If digits or a name preceded by ampersand follow the
2306 letter R, for example:
2307 .sp
2308 (?(R3)...) or (?(R&name)...)
2309 .sp
2310 the condition is true if the most recent recursion is into a subpattern whose
2311 number or name is given. This condition does not check the entire recursion
2312 stack. If the name used in a condition of this kind is a duplicate, the test is
2313 applied to all subpatterns of the same name, and is true if any one of them is
2314 the most recent recursion.
2315 .P
2316 At "top level", all these recursion test conditions are false.
2317 .\" HTML <a href="#recursion">
2318 .\" </a>
2319 The syntax for recursive patterns
2320 .\"
2321 is described below.
2322 .
2323 .\" HTML <a name="subdefine"></a>
2324 .SS "Defining subpatterns for use by reference only"
2325 .rs
2326 .sp
2327 If the condition is the string (DEFINE), and there is no subpattern with the
2328 name DEFINE, the condition is always false. In this case, there may be only one
2329 alternative in the subpattern. It is always skipped if control reaches this
2330 point in the pattern; the idea of DEFINE is that it can be used to define
2331 subroutines that can be referenced from elsewhere. (The use of
2332 .\" HTML <a href="#subpatternsassubroutines">
2333 .\" </a>
2334 subroutines
2335 .\"
2336 is described below.) For example, a pattern to match an IPv4 address such as
2337 "" could be written like this (ignore white space and line
2338 breaks):
2339 .sp
2340 (?(DEFINE) (?<byte> 2[0-4]\ed | 25[0-5] | 1\ed\ed | [1-9]?\ed) )
2341 \eb (?&byte) (\e.(?&byte)){3} \eb
2342 .sp
2343 The first part of the pattern is a DEFINE group inside which a another group
2344 named "byte" is defined. This matches an individual component of an IPv4
2345 address (a number less than 256). When matching takes place, this part of the
2346 pattern is skipped because DEFINE acts like a false condition. The rest of the
2347 pattern uses references to the named group to match the four dot-separated
2348 components of an IPv4 address, insisting on a word boundary at each end.
2349 .
2350 .SS "Assertion conditions"
2351 .rs
2352 .sp
2353 If the condition is not in any of the above formats, it must be an assertion.
2354 This may be a positive or negative lookahead or lookbehind assertion. Consider
2355 this pattern, again containing non-significant white space, and with the two
2356 alternatives on the second line:
2357 .sp
2358 (?(?=[^a-z]*[a-z])
2359 \ed{2}-[a-z]{3}-\ed{2} | \ed{2}-\ed{2}-\ed{2} )
2360 .sp
2361 The condition is a positive lookahead assertion that matches an optional
2362 sequence of non-letters followed by a letter. In other words, it tests for the
2363 presence of at least one letter in the subject. If a letter is found, the
2364 subject is matched against the first alternative; otherwise it is matched
2365 against the second. This pattern matches strings in one of the two forms
2366 dd-aaa-dd or dd-dd-dd, where aaa are letters and dd are digits.
2367 .
2368 .
2369 .\" HTML <a name="comments"></a>
2371 .rs
2372 .sp
2373 There are two ways of including comments in patterns that are processed by
2374 PCRE. In both cases, the start of the comment must not be in a character class,
2375 nor in the middle of any other sequence of related characters such as (?: or a
2376 subpattern name or number. The characters that make up a comment play no part
2377 in the pattern matching.
2378 .P
2379 The sequence (?# marks the start of a comment that continues up to the next
2380 closing parenthesis. Nested parentheses are not permitted. If the PCRE_EXTENDED
2381 option is set, an unescaped # character also introduces a comment, which in
2382 this case continues to immediately after the next newline character or
2383 character sequence in the pattern. Which characters are interpreted as newlines
2384 is controlled by the options passed to a compiling function or by a special
2385 sequence at the start of the pattern, as described in the section entitled
2386 .\" HTML <a href="#newlines">
2387 .\" </a>
2388 "Newline conventions"
2389 .\"
2390 above. Note that the end of this type of comment is a literal newline sequence
2391 in the pattern; escape sequences that happen to represent a newline do not
2392 count. For example, consider this pattern when PCRE_EXTENDED is set, and the
2393 default newline convention is in force:
2394 .sp
2395 abc #comment \en still comment
2396 .sp
2397 On encountering the # character, \fBpcre_compile()\fP skips along, looking for
2398 a newline in the pattern. The sequence \en is still literal at this stage, so
2399 it does not terminate the comment. Only an actual character with the code value
2400 0x0a (the default newline) does so.
2401 .
2402 .
2403 .\" HTML <a name="recursion"></a>
2405 .rs
2406 .sp
2407 Consider the problem of matching a string in parentheses, allowing for
2408 unlimited nested parentheses. Without the use of recursion, the best that can
2409 be done is to use a pattern that matches up to some fixed depth of nesting. It
2410 is not possible to handle an arbitrary nesting depth.
2411 .P
2412 For some time, Perl has provided a facility that allows regular expressions to
2413 recurse (amongst other things). It does this by interpolating Perl code in the
2414 expression at run time, and the code can refer to the expression itself. A Perl
2415 pattern using code interpolation to solve the parentheses problem can be
2416 created like this:
2417 .sp
2418 $re = qr{\e( (?: (?>[^()]+) | (?p{$re}) )* \e)}x;
2419 .sp
2420 The (?p{...}) item interpolates Perl code at run time, and in this case refers
2421 recursively to the pattern in which it appears.
2422 .P
2423 Obviously, PCRE cannot support the interpolation of Perl code. Instead, it
2424 supports special syntax for recursion of the entire pattern, and also for
2425 individual subpattern recursion. After its introduction in PCRE and Python,
2426 this kind of recursion was subsequently introduced into Perl at release 5.10.
2427 .P
2428 A special item that consists of (? followed by a number greater than zero and a
2429 closing parenthesis is a recursive subroutine call of the subpattern of the
2430 given number, provided that it occurs inside that subpattern. (If not, it is a
2431 .\" HTML <a href="#subpatternsassubroutines">
2432 .\" </a>
2433 non-recursive subroutine
2434 .\"
2435 call, which is described in the next section.) The special item (?R) or (?0) is
2436 a recursive call of the entire regular expression.
2437 .P
2438 This PCRE pattern solves the nested parentheses problem (assume the
2439 PCRE_EXTENDED option is set so that white space is ignored):
2440 .sp
2441 \e( ( [^()]++ | (?R) )* \e)
2442 .sp
2443 First it matches an opening parenthesis. Then it matches any number of
2444 substrings which can either be a sequence of non-parentheses, or a recursive
2445 match of the pattern itself (that is, a correctly parenthesized substring).
2446 Finally there is a closing parenthesis. Note the use of a possessive quantifier
2447 to avoid backtracking into sequences of non-parentheses.
2448 .P
2449 If this were part of a larger pattern, you would not want to recurse the entire
2450 pattern, so instead you could use this:
2451 .sp
2452 ( \e( ( [^()]++ | (?1) )* \e) )
2453 .sp
2454 We have put the pattern into parentheses, and caused the recursion to refer to
2455 them instead of the whole pattern.
2456 .P
2457 In a larger pattern, keeping track of parenthesis numbers can be tricky. This
2458 is made easier by the use of relative references. Instead of (?1) in the
2459 pattern above you can write (?-2) to refer to the second most recently opened
2460 parentheses preceding the recursion. In other words, a negative number counts
2461 capturing parentheses leftwards from the point at which it is encountered.
2462 .P
2463 It is also possible to refer to subsequently opened parentheses, by writing
2464 references such as (?+2). However, these cannot be recursive because the
2465 reference is not inside the parentheses that are referenced. They are always
2466 .\" HTML <a href="#subpatternsassubroutines">
2467 .\" </a>
2468 non-recursive subroutine
2469 .\"
2470 calls, as described in the next section.
2471 .P
2472 An alternative approach is to use named parentheses instead. The Perl syntax
2473 for this is (?&name); PCRE's earlier syntax (?P>name) is also supported. We
2474 could rewrite the above example as follows:
2475 .sp
2476 (?<pn> \e( ( [^()]++ | (?&pn) )* \e) )
2477 .sp
2478 If there is more than one subpattern with the same name, the earliest one is
2479 used.
2480 .P
2481 This particular example pattern that we have been looking at contains nested
2482 unlimited repeats, and so the use of a possessive quantifier for matching
2483 strings of non-parentheses is important when applying the pattern to strings
2484 that do not match. For example, when this pattern is applied to
2485 .sp
2486 (aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa()
2487 .sp
2488 it yields "no match" quickly. However, if a possessive quantifier is not used,
2489 the match runs for a very long time indeed because there are so many different
2490 ways the + and * repeats can carve up the subject, and all have to be tested
2491 before failure can be reported.
2492 .P
2493 At the end of a match, the values of capturing parentheses are those from
2494 the outermost level. If you want to obtain intermediate values, a callout
2495 function can be used (see below and the
2496 .\" HREF
2497 \fBpcrecallout\fP
2498 .\"
2499 documentation). If the pattern above is matched against
2500 .sp
2501 (ab(cd)ef)
2502 .sp
2503 the value for the inner capturing parentheses (numbered 2) is "ef", which is
2504 the last value taken on at the top level. If a capturing subpattern is not
2505 matched at the top level, its final captured value is unset, even if it was
2506 (temporarily) set at a deeper level during the matching process.
2507 .P
2508 If there are more than 15 capturing parentheses in a pattern, PCRE has to
2509 obtain extra memory to store data during a recursion, which it does by using
2510 \fBpcre_malloc\fP, freeing it via \fBpcre_free\fP afterwards. If no memory can
2511 be obtained, the match fails with the PCRE_ERROR_NOMEMORY error.
2512 .P
2513 Do not confuse the (?R) item with the condition (R), which tests for recursion.
2514 Consider this pattern, which matches text in angle brackets, allowing for
2515 arbitrary nesting. Only digits are allowed in nested brackets (that is, when
2516 recursing), whereas any characters are permitted at the outer level.
2517 .sp
2518 < (?: (?(R) \ed++ | [^<>]*+) | (?R)) * >
2519 .sp
2520 In this pattern, (?(R) is the start of a conditional subpattern, with two
2521 different alternatives for the recursive and non-recursive cases. The (?R) item
2522 is the actual recursive call.
2523 .
2524 .
2525 .\" HTML <a name="recursiondifference"></a>
2526 .SS "Differences in recursion processing between PCRE and Perl"
2527 .rs
2528 .sp
2529 Recursion processing in PCRE differs from Perl in two important ways. In PCRE
2530 (like Python, but unlike Perl), a recursive subpattern call is always treated
2531 as an atomic group. That is, once it has matched some of the subject string, it
2532 is never re-entered, even if it contains untried alternatives and there is a
2533 subsequent matching failure. This can be illustrated by the following pattern,
2534 which purports to match a palindromic string that contains an odd number of
2535 characters (for example, "a", "aba", "abcba", "abcdcba"):
2536 .sp
2537 ^(.|(.)(?1)\e2)$
2538 .sp
2539 The idea is that it either matches a single character, or two identical
2540 characters surrounding a sub-palindrome. In Perl, this pattern works; in PCRE
2541 it does not if the pattern is longer than three characters. Consider the
2542 subject string "abcba":
2543 .P
2544 At the top level, the first character is matched, but as it is not at the end
2545 of the string, the first alternative fails; the second alternative is taken
2546 and the recursion kicks in. The recursive call to subpattern 1 successfully
2547 matches the next character ("b"). (Note that the beginning and end of line
2548 tests are not part of the recursion).
2549 .P
2550 Back at the top level, the next character ("c") is compared with what
2551 subpattern 2 matched, which was "a". This fails. Because the recursion is
2552 treated as an atomic group, there are now no backtracking points, and so the
2553 entire match fails. (Perl is able, at this point, to re-enter the recursion and
2554 try the second alternative.) However, if the pattern is written with the
2555 alternatives in the other order, things are different:
2556 .sp
2557 ^((.)(?1)\e2|.)$
2558 .sp
2559 This time, the recursing alternative is tried first, and continues to recurse
2560 until it runs out of characters, at which point the recursion fails. But this
2561 time we do have another alternative to try at the higher level. That is the big
2562 difference: in the previous case the remaining alternative is at a deeper
2563 recursion level, which PCRE cannot use.
2564 .P
2565 To change the pattern so that it matches all palindromic strings, not just
2566 those with an odd number of characters, it is tempting to change the pattern to
2567 this:
2568 .sp
2569 ^((.)(?1)\e2|.?)$
2570 .sp
2571 Again, this works in Perl, but not in PCRE, and for the same reason. When a
2572 deeper recursion has matched a single character, it cannot be entered again in
2573 order to match an empty string. The solution is to separate the two cases, and
2574 write out the odd and even cases as alternatives at the higher level:
2575 .sp
2576 ^(?:((.)(?1)\e2|)|((.)(?3)\e4|.))
2577 .sp
2578 If you want to match typical palindromic phrases, the pattern has to ignore all
2579 non-word characters, which can be done like this:
2580 .sp
2581 ^\eW*+(?:((.)\eW*+(?1)\eW*+\e2|)|((.)\eW*+(?3)\eW*+\e4|\eW*+.\eW*+))\eW*+$
2582 .sp
2583 If run with the PCRE_CASELESS option, this pattern matches phrases such as "A
2584 man, a plan, a canal: Panama!" and it works well in both PCRE and Perl. Note
2585 the use of the possessive quantifier *+ to avoid backtracking into sequences of
2586 non-word characters. Without this, PCRE takes a great deal longer (ten times or
2587 more) to match typical phrases, and Perl takes so long that you think it has
2588 gone into a loop.
2589 .P
2590 \fBWARNING\fP: The palindrome-matching patterns above work only if the subject
2591 string does not start with a palindrome that is shorter than the entire string.
2592 For example, although "abcba" is correctly matched, if the subject is "ababa",
2593 PCRE finds the palindrome "aba" at the start, then fails at top level because
2594 the end of the string does not follow. Once again, it cannot jump back into the
2595 recursion to try other alternatives, so the entire match fails.
2596 .P
2597 The second way in which PCRE and Perl differ in their recursion processing is
2598 in the handling of captured values. In Perl, when a subpattern is called
2599 recursively or as a subpattern (see the next section), it has no access to any
2600 values that were captured outside the recursion, whereas in PCRE these values
2601 can be referenced. Consider this pattern:
2602 .sp
2603 ^(.)(\e1|a(?2))
2604 .sp
2605 In PCRE, this pattern matches "bab". The first capturing parentheses match "b",
2606 then in the second group, when the back reference \e1 fails to match "b", the
2607 second alternative matches "a" and then recurses. In the recursion, \e1 does
2608 now match "b" and so the whole match succeeds. In Perl, the pattern fails to
2609 match because inside the recursive call \e1 cannot access the externally set
2610 value.
2611 .
2612 .
2613 .\" HTML <a name="subpatternsassubroutines"></a>
2615 .rs
2616 .sp
2617 If the syntax for a recursive subpattern call (either by number or by
2618 name) is used outside the parentheses to which it refers, it operates like a
2619 subroutine in a programming language. The called subpattern may be defined
2620 before or after the reference. A numbered reference can be absolute or
2621 relative, as in these examples:
2622 .sp
2623 (...(absolute)...)...(?2)...
2624 (...(relative)...)...(?-1)...
2625 (...(?+1)...(relative)...
2626 .sp
2627 An earlier example pointed out that the pattern
2628 .sp
2629 (sens|respons)e and \e1ibility
2630 .sp
2631 matches "sense and sensibility" and "response and responsibility", but not
2632 "sense and responsibility". If instead the pattern
2633 .sp
2634 (sens|respons)e and (?1)ibility
2635 .sp
2636 is used, it does match "sense and responsibility" as well as the other two
2637 strings. Another example is given in the discussion of DEFINE above.
2638 .P
2639 All subroutine calls, whether recursive or not, are always treated as atomic
2640 groups. That is, once a subroutine has matched some of the subject string, it
2641 is never re-entered, even if it contains untried alternatives and there is a
2642 subsequent matching failure. Any capturing parentheses that are set during the
2643 subroutine call revert to their previous values afterwards.
2644 .P
2645 Processing options such as case-independence are fixed when a subpattern is
2646 defined, so if it is used as a subroutine, such options cannot be changed for
2647 different calls. For example, consider this pattern:
2648 .sp
2649 (abc)(?i:(?-1))
2650 .sp
2651 It matches "abcabc". It does not match "abcABC" because the change of
2652 processing option does not affect the called subpattern.
2653 .
2654 .
2655 .\" HTML <a name="onigurumasubroutines"></a>
2657 .rs
2658 .sp
2659 For compatibility with Oniguruma, the non-Perl syntax \eg followed by a name or
2660 a number enclosed either in angle brackets or single quotes, is an alternative
2661 syntax for referencing a subpattern as a subroutine, possibly recursively. Here
2662 are two of the examples used above, rewritten using this syntax:
2663 .sp
2664 (?<pn> \e( ( (?>[^()]+) | \eg<pn> )* \e) )
2665 (sens|respons)e and \eg'1'ibility
2666 .sp
2667 PCRE supports an extension to Oniguruma: if a number is preceded by a
2668 plus or a minus sign it is taken as a relative reference. For example:
2669 .sp
2670 (abc)(?i:\eg<-1>)
2671 .sp
2672 Note that \eg{...} (Perl syntax) and \eg<...> (Oniguruma syntax) are \fInot\fP
2673 synonymous. The former is a back reference; the latter is a subroutine call.
2674 .
2675 .
2677 .rs
2678 .sp
2679 Perl has a feature whereby using the sequence (?{...}) causes arbitrary Perl
2680 code to be obeyed in the middle of matching a regular expression. This makes it
2681 possible, amongst other things, to extract different substrings that match the
2682 same pair of parentheses when there is a repetition.
2683 .P
2684 PCRE provides a similar feature, but of course it cannot obey arbitrary Perl
2685 code. The feature is called "callout". The caller of PCRE provides an external
2686 function by putting its entry point in the global variable \fIpcre_callout\fP
2687 (8-bit library) or \fIpcre[16|32]_callout\fP (16-bit or 32-bit library).
2688 By default, this variable contains NULL, which disables all calling out.
2689 .P
2690 Within a regular expression, (?C) indicates the points at which the external
2691 function is to be called. If you want to identify different callout points, you
2692 can put a number less than 256 after the letter C. The default value is zero.
2693 For example, this pattern has two callout points:
2694 .sp
2695 (?C1)abc(?C2)def
2696 .sp
2697 If the PCRE_AUTO_CALLOUT flag is passed to a compiling function, callouts are
2698 automatically installed before each item in the pattern. They are all numbered
2699 255. If there is a conditional group in the pattern whose condition is an
2700 assertion, an additional callout is inserted just before the condition. An
2701 explicit callout may also be set at this position, as in this example:
2702 .sp
2703 (?(?C9)(?=a)abc|def)
2704 .sp
2705 Note that this applies only to assertion conditions, not to other types of
2706 condition.
2707 .P
2708 During matching, when PCRE reaches a callout point, the external function is
2709 called. It is provided with the number of the callout, the position in the
2710 pattern, and, optionally, one item of data originally supplied by the caller of
2711 the matching function. The callout function may cause matching to proceed, to
2712 backtrack, or to fail altogether. A complete description of the interface to
2713 the callout function is given in the
2714 .\" HREF
2715 \fBpcrecallout\fP
2716 .\"
2717 documentation.
2718 .
2719 .
2720 .\" HTML <a name="backtrackcontrol"></a>
2722 .rs
2723 .sp
2724 Perl 5.10 introduced a number of "Special Backtracking Control Verbs", which
2725 are still described in the Perl documentation as "experimental and subject to
2726 change or removal in a future version of Perl". It goes on to say: "Their usage
2727 in production code should be noted to avoid problems during upgrades." The same
2728 remarks apply to the PCRE features described in this section.
2729 .P
2730 The new verbs make use of what was previously invalid syntax: an opening
2731 parenthesis followed by an asterisk. They are generally of the form
2732 (*VERB) or (*VERB:NAME). Some may take either form, possibly behaving
2733 differently depending on whether or not a name is present. A name is any
2734 sequence of characters that does not include a closing parenthesis. The maximum
2735 length of name is 255 in the 8-bit library and 65535 in the 16-bit and 32-bit
2736 libraries. If the name is empty, that is, if the closing parenthesis
2737 immediately follows the colon, the effect is as if the colon were not there.
2738 Any number of these verbs may occur in a pattern.
2739 .P
2740 Since these verbs are specifically related to backtracking, most of them can be
2741 used only when the pattern is to be matched using one of the traditional
2742 matching functions, because these use a backtracking algorithm. With the
2743 exception of (*FAIL), which behaves like a failing negative assertion, the
2744 backtracking control verbs cause an error if encountered by a DFA matching
2745 function.
2746 .P
2747 The behaviour of these verbs in
2748 .\" HTML <a href="#btrepeat">
2749 .\" </a>
2750 repeated groups,
2751 .\"
2752 .\" HTML <a href="#btassert">
2753 .\" </a>
2754 assertions,
2755 .\"
2756 and in
2757 .\" HTML <a href="#btsub">
2758 .\" </a>
2759 subpatterns called as subroutines
2760 .\"
2761 (whether or not recursively) is documented below.
2762 .
2763 .
2764 .\" HTML <a name="nooptimize"></a>
2765 .SS "Optimizations that affect backtracking verbs"
2766 .rs
2767 .sp
2768 PCRE contains some optimizations that are used to speed up matching by running
2769 some checks at the start of each match attempt. For example, it may know the
2770 minimum length of matching subject, or that a particular character must be
2771 present. When one of these optimizations bypasses the running of a match, any
2772 included backtracking verbs will not, of course, be processed. You can suppress
2773 the start-of-match optimizations by setting the PCRE_NO_START_OPTIMIZE option
2774 when calling \fBpcre_compile()\fP or \fBpcre_exec()\fP, or by starting the
2775 pattern with (*NO_START_OPT). There is more discussion of this option in the
2776 section entitled
2777 .\" HTML <a href="pcreapi.html#execoptions">
2778 .\" </a>
2779 "Option bits for \fBpcre_exec()\fP"
2780 .\"
2781 in the
2782 .\" HREF
2783 \fBpcreapi\fP
2784 .\"
2785 documentation.
2786 .P
2787 Experiments with Perl suggest that it too has similar optimizations, sometimes
2788 leading to anomalous results.
2789 .
2790 .
2791 .SS "Verbs that act immediately"
2792 .rs
2793 .sp
2794 The following verbs act as soon as they are encountered. They may not be
2795 followed by a name.
2796 .sp
2797 (*ACCEPT)
2798 .sp
2799 This verb causes the match to end successfully, skipping the remainder of the
2800 pattern. However, when it is inside a subpattern that is called as a
2801 subroutine, only that subpattern is ended successfully. Matching then continues
2802 at the outer level. If (*ACCEPT) in triggered in a positive assertion, the
2803 assertion succeeds; in a negative assertion, the assertion fails.
2804 .P
2805 If (*ACCEPT) is inside capturing parentheses, the data so far is captured. For
2806 example:
2807 .sp
2808 A((?:A|B(*ACCEPT)|C)D)
2809 .sp
2810 This matches "AB", "AAD", or "ACD"; when it matches "AB", "B" is captured by
2811 the outer parentheses.
2812 .sp
2813 (*FAIL) or (*F)
2814 .sp
2815 This verb causes a matching failure, forcing backtracking to occur. It is
2816 equivalent to (?!) but easier to read. The Perl documentation notes that it is
2817 probably useful only when combined with (?{}) or (??{}). Those are, of course,
2818 Perl features that are not present in PCRE. The nearest equivalent is the
2819 callout feature, as for example in this pattern:
2820 .sp
2821 a+(?C)(*FAIL)
2822 .sp
2823 A match with the string "aaaa" always fails, but the callout is taken before
2824 each backtrack happens (in this example, 10 times).
2825 .
2826 .
2827 .SS "Recording which path was taken"
2828 .rs
2829 .sp
2830 There is one verb whose main purpose is to track how a match was arrived at,
2831 though it also has a secondary use in conjunction with advancing the match
2832 starting point (see (*SKIP) below).
2833 .sp
2834 (*MARK:NAME) or (*:NAME)
2835 .sp
2836 A name is always required with this verb. There may be as many instances of
2837 (*MARK) as you like in a pattern, and their names do not have to be unique.
2838 .P
2839 When a match succeeds, the name of the last-encountered (*MARK:NAME),
2840 (*PRUNE:NAME), or (*THEN:NAME) on the matching path is passed back to the
2841 caller as described in the section entitled
2842 .\" HTML <a href="pcreapi.html#extradata">
2843 .\" </a>
2844 "Extra data for \fBpcre_exec()\fP"
2845 .\"
2846 in the
2847 .\" HREF
2848 \fBpcreapi\fP
2849 .\"
2850 documentation. Here is an example of \fBpcretest\fP output, where the /K
2851 modifier requests the retrieval and outputting of (*MARK) data:
2852 .sp
2853 re> /X(*MARK:A)Y|X(*MARK:B)Z/K
2854 data> XY
2855 0: XY
2856 MK: A
2857 XZ
2858 0: XZ
2859 MK: B
2860 .sp
2861 The (*MARK) name is tagged with "MK:" in this output, and in this example it
2862 indicates which of the two alternatives matched. This is a more efficient way
2863 of obtaining this information than putting each alternative in its own
2864 capturing parentheses.
2865 .P
2866 If a verb with a name is encountered in a positive assertion that is true, the
2867 name is recorded and passed back if it is the last-encountered. This does not
2868 happen for negative assertions or failing positive assertions.
2869 .P
2870 After a partial match or a failed match, the last encountered name in the
2871 entire match process is returned. For example:
2872 .sp
2873 re> /X(*MARK:A)Y|X(*MARK:B)Z/K
2874 data> XP
2875 No match, mark = B
2876 .sp
2877 Note that in this unanchored example the mark is retained from the match
2878 attempt that started at the letter "X" in the subject. Subsequent match
2879 attempts starting at "P" and then with an empty string do not get as far as the
2880 (*MARK) item, but nevertheless do not reset it.
2881 .P
2882 If you are interested in (*MARK) values after failed matches, you should
2883 probably set the PCRE_NO_START_OPTIMIZE option
2884 .\" HTML <a href="#nooptimize">
2885 .\" </a>
2886 (see above)
2887 .\"
2888 to ensure that the match is always attempted.
2889 .
2890 .
2891 .SS "Verbs that act after backtracking"
2892 .rs
2893 .sp
2894 The following verbs do nothing when they are encountered. Matching continues
2895 with what follows, but if there is no subsequent match, causing a backtrack to
2896 the verb, a failure is forced. That is, backtracking cannot pass to the left of
2897 the verb. However, when one of these verbs appears inside an atomic group or an
2898 assertion that is true, its effect is confined to that group, because once the
2899 group has been matched, there is never any backtracking into it. In this
2900 situation, backtracking can "jump back" to the left of the entire atomic group
2901 or assertion. (Remember also, as stated above, that this localization also
2902 applies in subroutine calls.)
2903 .P
2904 These verbs differ in exactly what kind of failure occurs when backtracking
2905 reaches them. The behaviour described below is what happens when the verb is
2906 not in a subroutine or an assertion. Subsequent sections cover these special
2907 cases.
2908 .sp
2909 (*COMMIT)
2910 .sp
2911 This verb, which may not be followed by a name, causes the whole match to fail
2912 outright if there is a later matching failure that causes backtracking to reach
2913 it. Even if the pattern is unanchored, no further attempts to find a match by
2914 advancing the starting point take place. If (*COMMIT) is the only backtracking
2915 verb that is encountered, once it has been passed \fBpcre_exec()\fP is
2916 committed to finding a match at the current starting point, or not at all. For
2917 example:
2918 .sp
2919 a+(*COMMIT)b
2920 .sp
2921 This matches "xxaab" but not "aacaab". It can be thought of as a kind of
2922 dynamic anchor, or "I've started, so I must finish." The name of the most
2923 recently passed (*MARK) in the path is passed back when (*COMMIT) forces a
2924 match failure.
2925 .P
2926 If there is more than one backtracking verb in a pattern, a different one that
2927 follows (*COMMIT) may be triggered first, so merely passing (*COMMIT) during a
2928 match does not always guarantee that a match must be at this starting point.
2929 .P
2930 Note that (*COMMIT) at the start of a pattern is not the same as an anchor,
2931 unless PCRE's start-of-match optimizations are turned off, as shown in this
2932 \fBpcretest\fP example:
2933 .sp
2934 re> /(*COMMIT)abc/
2935 data> xyzabc
2936 0: abc
2937 xyzabc\eY
2938 No match
2939 .sp
2940 PCRE knows that any match must start with "a", so the optimization skips along
2941 the subject to "a" before running the first match attempt, which succeeds. When
2942 the optimization is disabled by the \eY escape in the second subject, the match
2943 starts at "x" and so the (*COMMIT) causes it to fail without trying any other
2944 starting points.
2945 .sp
2946 (*PRUNE) or (*PRUNE:NAME)
2947 .sp
2948 This verb causes the match to fail at the current starting position in the
2949 subject if there is a later matching failure that causes backtracking to reach
2950 it. If the pattern is unanchored, the normal "bumpalong" advance to the next
2951 starting character then happens. Backtracking can occur as usual to the left of
2952 (*PRUNE), before it is reached, or when matching to the right of (*PRUNE), but
2953 if there is no match to the right, backtracking cannot cross (*PRUNE). In
2954 simple cases, the use of (*PRUNE) is just an alternative to an atomic group or
2955 possessive quantifier, but there are some uses of (*PRUNE) that cannot be
2956 expressed in any other way. In an anchored pattern (*PRUNE) has the same effect
2957 as (*COMMIT).
2958 .P
2959 The behaviour of (*PRUNE:NAME) is the not the same as (*MARK:NAME)(*PRUNE).
2960 It is like (*MARK:NAME) in that the name is remembered for passing back to the
2961 caller. However, (*SKIP:NAME) searches only for names set with (*MARK).
2962 .sp
2963 (*SKIP)
2964 .sp
2965 This verb, when given without a name, is like (*PRUNE), except that if the
2966 pattern is unanchored, the "bumpalong" advance is not to the next character,
2967 but to the position in the subject where (*SKIP) was encountered. (*SKIP)
2968 signifies that whatever text was matched leading up to it cannot be part of a
2969 successful match. Consider:
2970 .sp
2971 a+(*SKIP)b
2972 .sp
2973 If the subject is "aaaac...", after the first match attempt fails (starting at
2974 the first character in the string), the starting point skips on to start the
2975 next attempt at "c". Note that a possessive quantifer does not have the same
2976 effect as this example; although it would suppress backtracking during the
2977 first match attempt, the second attempt would start at the second character
2978 instead of skipping on to "c".
2979 .sp
2980 (*SKIP:NAME)
2981 .sp
2982 When (*SKIP) has an associated name, its behaviour is modified. When it is
2983 triggered, the previous path through the pattern is searched for the most
2984 recent (*MARK) that has the same name. If one is found, the "bumpalong" advance
2985 is to the subject position that corresponds to that (*MARK) instead of to where
2986 (*SKIP) was encountered. If no (*MARK) with a matching name is found, the
2987 (*SKIP) is ignored.
2988 .P
2989 Note that (*SKIP:NAME) searches only for names set by (*MARK:NAME). It ignores
2990 names that are set by (*PRUNE:NAME) or (*THEN:NAME).
2991 .sp
2992 (*THEN) or (*THEN:NAME)
2993 .sp
2994 This verb causes a skip to the next innermost alternative when backtracking
2995 reaches it. That is, it cancels any further backtracking within the current
2996 alternative. Its name comes from the observation that it can be used for a
2997 pattern-based if-then-else block:
2998 .sp
2999 ( COND1 (*THEN) FOO | COND2 (*THEN) BAR | COND3 (*THEN) BAZ ) ...
3000 .sp
3001 If the COND1 pattern matches, FOO is tried (and possibly further items after
3002 the end of the group if FOO succeeds); on failure, the matcher skips to the
3003 second alternative and tries COND2, without backtracking into COND1. If that
3004 succeeds and BAR fails, COND3 is tried. If subsequently BAZ fails, there are no
3005 more alternatives, so there is a backtrack to whatever came before the entire
3006 group. If (*THEN) is not inside an alternation, it acts like (*PRUNE).
3007 .P
3008 The behaviour of (*THEN:NAME) is the not the same as (*MARK:NAME)(*THEN).
3009 It is like (*MARK:NAME) in that the name is remembered for passing back to the
3010 caller. However, (*SKIP:NAME) searches only for names set with (*MARK).
3011 .P
3012 A subpattern that does not contain a | character is just a part of the
3013 enclosing alternative; it is not a nested alternation with only one
3014 alternative. The effect of (*THEN) extends beyond such a subpattern to the
3015 enclosing alternative. Consider this pattern, where A, B, etc. are complex
3016 pattern fragments that do not contain any | characters at this level:
3017 .sp
3018 A (B(*THEN)C) | D
3019 .sp
3020 If A and B are matched, but there is a failure in C, matching does not
3021 backtrack into A; instead it moves to the next alternative, that is, D.
3022 However, if the subpattern containing (*THEN) is given an alternative, it
3023 behaves differently:
3024 .sp
3025 A (B(*THEN)C | (*FAIL)) | D
3026 .sp
3027 The effect of (*THEN) is now confined to the inner subpattern. After a failure
3028 in C, matching moves to (*FAIL), which causes the whole subpattern to fail
3029 because there are no more alternatives to try. In this case, matching does now
3030 backtrack into A.
3031 .P
3032 Note that a conditional subpattern is not considered as having two
3033 alternatives, because only one is ever used. In other words, the | character in
3034 a conditional subpattern has a different meaning. Ignoring white space,
3035 consider:
3036 .sp
3037 ^.*? (?(?=a) a | b(*THEN)c )
3038 .sp
3039 If the subject is "ba", this pattern does not match. Because .*? is ungreedy,
3040 it initially matches zero characters. The condition (?=a) then fails, the
3041 character "b" is matched, but "c" is not. At this point, matching does not
3042 backtrack to .*? as might perhaps be expected from the presence of the |
3043 character. The conditional subpattern is part of the single alternative that
3044 comprises the whole pattern, and so the match fails. (If there was a backtrack
3045 into .*?, allowing it to match "b", the match would succeed.)
3046 .P
3047 The verbs just described provide four different "strengths" of control when
3048 subsequent matching fails. (*THEN) is the weakest, carrying on the match at the
3049 next alternative. (*PRUNE) comes next, failing the match at the current
3050 starting position, but allowing an advance to the next character (for an
3051 unanchored pattern). (*SKIP) is similar, except that the advance may be more
3052 than one character. (*COMMIT) is the strongest, causing the entire match to
3053 fail.
3054 .
3055 .
3056 .SS "More than one backtracking verb"
3057 .rs
3058 .sp
3059 If more than one backtracking verb is present in a pattern, the one that is
3060 backtracked onto first acts. For example, consider this pattern, where A, B,
3061 etc. are complex pattern fragments:
3062 .sp
3064 .sp
3065 If A matches but B fails, the backtrack to (*COMMIT) causes the entire match to
3066 fail. However, if A and B match, but C fails, the backtrack to (*THEN) causes
3067 the next alternative (ABD) to be tried. This behaviour is consistent, but is
3068 not always the same as Perl's. It means that if two or more backtracking verbs
3069 appear in succession, all the the last of them has no effect. Consider this
3070 example:
3071 .sp
3072 ...(*COMMIT)(*PRUNE)...
3073 .sp
3074 If there is a matching failure to the right, backtracking onto (*PRUNE) causes
3075 it to be triggered, and its action is taken. There can never be a backtrack
3076 onto (*COMMIT).
3077 .
3078 .
3079 .\" HTML <a name="btrepeat"></a>
3080 .SS "Backtracking verbs in repeated groups"
3081 .rs
3082 .sp
3083 PCRE differs from Perl in its handling of backtracking verbs in repeated
3084 groups. For example, consider:
3085 .sp
3086 /(a(*COMMIT)b)+ac/
3087 .sp
3088 If the subject is "abac", Perl matches, but PCRE fails because the (*COMMIT) in
3089 the second repeat of the group acts.
3090 .
3091 .
3092 .\" HTML <a name="btassert"></a>
3093 .SS "Backtracking verbs in assertions"
3094 .rs
3095 .sp
3096 (*FAIL) in an assertion has its normal effect: it forces an immediate backtrack.
3097 .P
3098 (*ACCEPT) in a positive assertion causes the assertion to succeed without any
3099 further processing. In a negative assertion, (*ACCEPT) causes the assertion to
3100 fail without any further processing.
3101 .P
3102 The other backtracking verbs are not treated specially if they appear in a
3103 positive assertion. In particular, (*THEN) skips to the next alternative in the
3104 innermost enclosing group that has alternations, whether or not this is within
3105 the assertion.
3106 .P
3107 Negative assertions are, however, different, in order to ensure that changing a
3108 positive assertion into a negative assertion changes its result. Backtracking
3109 into (*COMMIT), (*SKIP), or (*PRUNE) causes a negative assertion to be true,
3110 without considering any further alternative branches in the assertion.
3111 Backtracking into (*THEN) causes it to skip to the next enclosing alternative
3112 within the assertion (the normal behaviour), but if the assertion does not have
3113 such an alternative, (*THEN) behaves like (*PRUNE).
3114 .
3115 .
3116 .\" HTML <a name="btsub"></a>
3117 .SS "Backtracking verbs in subroutines"
3118 .rs
3119 .sp
3120 These behaviours occur whether or not the subpattern is called recursively.
3121 Perl's treatment of subroutines is different in some cases.
3122 .P
3123 (*FAIL) in a subpattern called as a subroutine has its normal effect: it forces
3124 an immediate backtrack.
3125 .P
3126 (*ACCEPT) in a subpattern called as a subroutine causes the subroutine match to
3127 succeed without any further processing. Matching then continues after the
3128 subroutine call.
3129 .P
3130 (*COMMIT), (*SKIP), and (*PRUNE) in a subpattern called as a subroutine cause
3131 the subroutine match to fail.
3132 .P
3133 (*THEN) skips to the next alternative in the innermost enclosing group within
3134 the subpattern that has alternatives. If there is no such group within the
3135 subpattern, (*THEN) causes the subroutine match to fail.
3136 .
3137 .
3138 .SH "SEE ALSO"
3139 .rs
3140 .sp
3141 \fBpcreapi\fP(3), \fBpcrecallout\fP(3), \fBpcrematching\fP(3),
3142 \fBpcresyntax\fP(3), \fBpcre\fP(3), \fBpcre16(3)\fP, \fBpcre32(3)\fP.
3143 .
3144 .
3146 .rs
3147 .sp
3148 .nf
3149 Philip Hazel
3150 University Computing Service
3151 Cambridge CB2 3QH, England.
3152 .fi
3153 .
3154 .
3156 .rs
3157 .sp
3158 .nf
3159 Last updated: 05 October 2013
3160 Copyright (c) 1997-2013 University of Cambridge.
3161 .fi


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

  ViewVC Help
Powered by ViewVC 1.1.5